article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
recent advances in @xmath0-ray instrumentation have led to the discovery of gev and tev emission from a handful of high - mass x - ray binaries .
most of these binaries appear to be either wind accreting or colliding wind binaries in which a compact object ( a neutron star or a black hole ) is orbiting an early - b or late - o type massive companion .
however , for most of these high - mass @xmath0-ray binaries ( hmgbs ) the energy release mechanism and the type of compact object are unknown .
it is possible that systems with a young pulsar orbiting a massive star may dominate the population of hmgbs . in these hmgbs most of the high energy emission
is believed to be produced at the shock located between the pulsar wind and the massive - star wind .
although the spatially - integrated multiwavelength spectra and variability have been the subject of multiple studies , they only provide indirect diagnostics of the complex wind interaction .
theoretical models suggest that the shocked flow can be accelerated to relativistic velocities , forming a spiral which might then be disrupted by instabilities .
these predictions are yet to be verified observationally because it is challenging to achieve the high angular resolution required to directly resolve the structure of the interacting winds in these systems .
some recent intriguing evidence of extended emission comes from vlbi observations which indicate variable , subarcsecond - scale structures in a few known hmgbs @xcite .
the only hmgb in which the compact source has been detected as a radio pulsar @xcite is b125963/ls2883 ( hereafter b1259 ) .
the radio pulsations are seen during a large fraction of the 3.4 yr orbital period , but they disappear for about four weeks @xcite when the pulsar is moving through the dense wind of its massive companion around periastron .
it is possible that other hmgbs , whose orbits are more compact , also host young pulsars which can not be seen in the radio because the radio emission is blocked by the dense stellar wind . studying the high - energy emission from b1259 allows one to probe the properties of both the massive - star and pulsar winds and study the mechanisms of their interaction . in particular
, the pulsar wind properties can be studied in a regime inaccessible in observations of isolated pulsars . in this paper
we report the discovery of variable extended x - ray emission around b1259 and discuss the implications for colliding wind models of such binaries for the theory of pulsar winds and for the nature of high - energy emission in hmgbs .
the radio pulsar psr b125963 has the following properties : period @xmath3ms , characteristic age @xmath4kyr , spin - down power @xmath5ergs@xmath6 , magnetic field @xmath7 g , and distance @xmath8kpc @xcite .
the orbit of the b1259 binary is highly eccentric ( @xmath9 ) , with semimajor axis @xmath10au , period @xmath11d , and inclination @xmath12 @xcite .
the massive companion , a fast - rotating late o - type star with a luminosity @xmath13 @xcite , has an equatorial excretion disk and is a source of a strong stellar wind .
the disk is thought to be inclined to the orbital plane at an angle of @xmath14 @xcite .
x - ray and radio data suggest that the pulsar passes through the disk during time intervals of @xmath15 days before periastron and @xmath16 days after periastron @xcite .
it has been known from multiple observations that the x - ray flux and spectrum of b1259 vary with the orbital phase @xcite , but no x - ray pulsations with the pulsar s period have been detected @xcite .
this and the relatively high x - ray luminosity of b1259 , compared to isolated pulsars with similar parameters , suggest that the observed x - ray emission is mainly caused by the collision of the pulsar wind with the massive companion wind , which produces a shock whose properties are modulated with the orbital period in this highly eccentric binary .
although most of the x - ray emission is likely generated within the binary and in its immediate vicinity , the shocked matter blown out from the binary can be visible at larger distances in high - resolution observations .
extended radio emission of 1050 mas ( @xmath17120 au ) size has been detected near periastron of b1259 and interpreted as synchrotron radiation of shocked pulsar wind ejected from the binary @xcite .
such a size is too small to be resolved in x - ray observations , which can , however , probe the blown - out matter at larger distances .
the only previous high - resolution x - ray image of b125963 was obtained in a 28 ks _
chandra _ x - ray observatory ( _ cxo _ ) observation carried out on 2009 may 14 , when the pulsar was very close to apastron .
very faint , amorphous , asymmetric extended emission ( up to @xmath1 south - southwest from the pulsar position ) was detected at a @xmath18 confidence level @xcite .
this emission was tentatively interpreted as a pulsar wind nebula ( pwn ) confined by a shock between the two winds and ultimately blown out from the binary by the wind of the massive companion . here
we report the results of two deeper observations .
we observed b1259 with the advanced ccd imaging spectrometer ( acis ) on board on 2009 may 14 , 2011 december 17 , and 2013 may 19 ( see table 1 for obsids and mjd dates ) .
as the first observation was already analyzed and the results were published @xcite , we only refer to it as needed to supplement our analysis and discussion of the results of the last two observations .
details of the binary phases and exposure times are listed in table 1 .
the target was imaged on the front - illuminated acis - i3 chip , in timed exposure mode .
we used ` very faint ' ( vfaint ) telemetry format to reduce the detector background and optimize the sensitivity for the faint extended emission .
a 1/8 subarray was used to reduce the frame time to 0.4s and mitigate the effect of pile - up .
the highest count rate ( see table 1 ) corresponds to the pile - up fraction of @xmath19% and hence can be neglected . for our analysis
we downloaded the pipeline - produced level 2 event files , which have been cleaned of various artifacts .
no episodes of anomalously high background rates occurred in the observations .
no randomization was applied to the event positions in the level 2 event files we used . to further improve the accuracy of the event positioning , the energy - dependent subpixel event repositioning ( edser ) procedure was applied during the pipeline processing .
below we consider only events with energies 0.58kev in order to minimize the background contribution .
the detector responses for spectral analysis were produced with ciao tools , following the standard procedure and using the calibration database caldb 4.4.1 .
spectral fitting was done for 0.58kev with ciao s sherpa modeling and fitting package .
lcccccccccccc obsid & mjd & @xmath20 & @xmath21 & exp . &
cts & @xmath22 & @xmath23 & @xmath24 & @xmath25 & @xmath26 & @xmath27 & @xmath28dof + & & deg & days & ks & & @xmath29 cgs & @xmath29 cgs & @xmath30 @xmath31 & & @xmath32 & arcsec@xmath33 & + 10089 & 54965 & 182 & 667 & 25.6 & 1825 & 139(5 ) & 158(6 ) & 1.5(7 ) & 1.51(10 ) & 2.3(3 ) & 2.5 & 20.9/30 + & & & & & 61 & 2.8(13 ) & 3.1(13 ) & 1.5@xmath34 & 1.3(5 ) & 0.039(16 ) & 22.1 & 3.1/9 + 14205 & 55912 & 169 & 370 & 56.3 & 6551 & 249(4 ) & 296(5 ) & 2.9(3 ) & 1.39(5 ) & 3.8(2 ) & 2.5 & 76.6/87 + & & & & & 343 & 7.5(8 ) & 9.1(9 ) & 2.9@xmath34 & 1.49(17 ) & 0.13(2 ) & 22.1 & 21.2/16 + 14206 & 56431 & 192 & 886 & 56.3 & 4162 & 137(5 ) & 176(7 ) & 3.1(3 ) & 1.68(6 ) & 3.07(2 ) & 2.5 & 146/169 + & & & & & 144 & 3.3(5 ) & 3.9(6 ) & 3.1@xmath34 & 1.4(3 ) & 0.051(14 ) & 12.8 & 7.1/13 +
the exquisite angular resolution of reveals the diffuse x - ray emission which is clearly seen to up to @xmath35 from the binary ( figure [ images ] ) . moreover , the appearance of the extended emission changed dramatically between the two observations which occurred 248 days before and 268 days after the apastron passage ( see figure [ orb ] ) . while the diffuse emission looks like a southward extension of the point - like source in the first observation , it is clearly detached from the compact source in the second observation .
this change appears as an outward motion of an arc - like structure which is particularly well seen in the difference image shown in figure [ difference ] .
the @xmath36 shift in the position of the arc - like structure corresponds to the proper motion @xmath37 arcsec yr@xmath38 mas day@xmath6 and the projected velocity @xmath39 at the distance of 2.3 kpc .
if the extended emission detected in the two observations were emitted by the same object moving with a constant velocity , this object left the binary @xmath40 days before the second observation , i.e. , somewhere around the apastron passage in the previous binary cycle . ] .
each pixel is 0@xmath41 , no smoothing or randomization is applied .
bottom panel shows the radial profile extracted from the difference image within the angular sector shown in the top panel .
[ difference ] ] the regions used for spectral extraction in obsids 14205 and 14206 are shown in figure [ specs ] .
for completeness we also re - extracted the obsid 10089 spectra using exactly the same extraction regions as for obsids 14205 .
we then binned the spectra and fit each of them to an absorbed power - law ( pl ) model with the interstellar absorption cross sections from @xcite . for each observation
the extended emission spectra were fit with the extinction column fixed at the value found for the the core component ( within @xmath42 ; includes the pulsar and unresolved pwn ) in the same observation .
an example of the extended emission spectrum ( from obsid 14205 ) is shown in figure [ 2011spec ] .
the fitting parameters the two observations are presented in table 1 and figure [ cont ] .
the best - fit photon indices , @xmath431.5 , of the extended emission are similar to those obtained for pwne around isolated pulsars @xcite .
the quality of the fit to the spectrum of the extended emission from the 2011 observation ( see fig . [ 2011spec ] ) is not perfect ( @xmath44 for 16 degrees of freedom ) , but this can be explained by the noise fluctuations due to small number of counts .
the 0.58 kev luminosities of the extended structures , on the order of @xmath45 erg s@xmath6 , are small fractions ( 2%3% ) of the compact source luminosity .
the brightness of the core and extended emission components , and the spectral slope for the core emission show appreciable variations .
the absorbed flux of the extended emission was brighter by a factor of @xmath46 in the 2011 observation ( obsid 14205 ) compared to the 2009 and 2013 observations while the pl slope of the core component became steeper by @xmath47 in 2013 compared to 2011 .
at the same time no statistically significant spectral changes are measured for the extended emission in all three observations .
the limited number of counts also does not allow us to say whether the spectrum of the extended emission differs from that of the core component .
the changes in the core flux agree with those reported from multi - year observations at low angular resolution @xcite .
the spectra of the extended emission also fit the thermal plasma model ( vmekal ) , with high plasma temperatures ( best - fit @xmath48 kev , the upper boundary is unconstrained , the @xmath49 lower limit is @xmath50 kev ) , for which the plasma radiation is dominated by thermal bremsstrahlung .
the vmekal normalization of @xmath51 @xmath52 corresponds to the number density of @xmath53 @xmath54 , assuming solar abundances and the emitting volume @xmath55 @xmath56 , where @xmath57 .
the quality of these fits is similar to that of the pl fits . ]
@xmath58 planes for the acis spectra of the unresolved core ( _ top _ ) and extended nebula ( _ bottom _ ; see figure [ specs ] for region definitions ) computed for the absorbed pl model .
the pl normalization @xmath58 is in units of @xmath59 photons @xmath31 s@xmath6 kev@xmath6 at 1 kev .
the dashed lines are lines of constant unabsorbed flux in the 0.58 kev band ( labeled in erg@xmath31s@xmath6 ) .
[ fonts][cont],title="fig : " ] @xmath58 planes for the acis spectra of the unresolved core ( _ top _ ) and extended nebula ( _ bottom _ ; see figure [ specs ] for region definitions ) computed for the absorbed pl model .
the pl normalization @xmath58 is in units of @xmath59 photons @xmath31 s@xmath6 kev@xmath6 at 1 kev .
the dashed lines are lines of constant unabsorbed flux in the 0.58 kev band ( labeled in erg@xmath31s@xmath6 ) .
[ fonts][cont],title="fig : " ]
the observed extended x - ray emission is clearly asymmetric with respect to the binary while the resolved size and separation from the binary greatly exceed the size of the binary orbit .
it does not resemble the flow structures expected in such systems based on the current models . the location and the shape of the intrabinary shock depend on the momentum fluxes of the winds and their angular distributions .
as we know from observations of pwne of isolated pulsars , the pulsar wind is predominantly concentrated around the pulsar s equatorial plane .
if this plane coincides with the orbital plane , and the high - mass companion wind is nearly isotropic , the location of the contact discontinuity , separating the shocked winds on the line connecting the two stars ( the stagnation point ) is at the distance @xmath60 from the pulsar , where @xmath61 is the separation between the binary companions , @xmath62 is the ratio of the momentum fluxes of the pulsar wind and stellar wind at the half - distance between the pulsar and the star , @xmath63 is the mass - loss rate of the high - mass companion , @xmath64 is terminal wind velocity , and @xmath65 is the half - opening angle around the equatorial plane into which the pulsar s wind is blowing ( @xmath66 for an isotropic wind ) . at @xmath67
the stellar wind wraps around the pulsar and collimates the pulsar wind into a diverging tail directed away from the high mass companion , creating a spiral pattern due to the pulsar s orbital motion , which mixes with the denser stellar wind at the contact discontinuity via kelvin - helmholtz instabilities . at @xmath68
the pulsar wind takes over and collimates the stellar wind within a shock of a complex shape . according to and ,
late - type o stars with @xmath69 produce `` weak '' winds , with @xmath70 yr@xmath6 and @xmath71@xmath72 km s@xmath6 .
therefore , for the nearly isotropic polar component of the ls 2883 wind , the parameter @xmath73 is on the order of unity or even larger .
the pl spectrum of the resolved extended emission could be interpreted as synchrotron emission of the shocked pulsar wind , while the thermal plasma emission could be produced by the massive companion wind heated to x - ray temperatures by the interaction with the energetic pulsar wind ( e.g. , in oblique shocks and instabilities ; ) and ejected from the system .
the latter possibility , however , requires a very large mass of the ejected plasma cloud , @xmath74 g , where @xmath75 @xmath54 is the proton number density estimated from the spectral fit ( see section 3.2 ) .
this mass is much larger than @xmath76 g lost by the high - mass companion wind during the whole orbital period .
therefore , we conclude that thermal emission from the heated o - star ejecta alone can not explain the observed extended x - ray emission
. the other possible interpretation could be synchrotron emission from pulsar wind particles .
using standard formulae for synchtrotron radiation ( e.g. , ) , we can estimate the lorentz factors of the emitting electrons as @xmath77 .
one can also express the magnetic field in a synchrotron - emitting region in terms of observable parameters @xcite : @xmath78
\frac{\mathcal{b}_{-7}}{\bar{s}_{16}}\right\}^{2/7}\,\,\mu{\rm g}\,.\ ] ] here @xmath79 is the ratio of the magnetic field energy density to the energy density of the relativistic particles , @xmath80 photons ( scm@xmath33kevarcsec@xmath33)@xmath6 is the average spectral surface brightness at @xmath82 kev ( see table 1 ) , @xmath26 is the normalization of photon spectral flux measured from the area @xmath27 ( in arcsec@xmath33 ) , @xmath83 cm is the average length of the radiating region along the line of sight , @xmath84 , @xmath85 , @xmath86 and @xmath87 are the lower and upper energies of the photon pl spectrum ( in kev ) , and @xmath88 , @xmath89 and @xmath90 are the numerical coefficients whose values depend on the slope @xmath91 of the electron pl spectral energy distribution ( ) . taking @xmath25 , @xmath26 , and @xmath27 values from table 1 , and assuming @xmath92 ( equivalent to @xmath93 at @xmath94 kpc ) , we obtain @xmath95 @xmath96 g for the extended emission seen in the 2011 and 2013 observations ( for @xmath97 kev , @xmath98 kev ) . note that for the measured @xmath99@xmath100 the magnetic field dependence on the ( unknown ) values of @xmath86 and @xmath87 is extremely weak ( e.g. , expanding the energy range by five orders of magnitude up and down increases the magnetic field value by a factor of 1.6 ) .
the lower limit on magnetic field can be obtained by requiring the electron larmor radius to be smaller than the transverse size ( @xmath101 ) of the arc - like extended emission seen in the 2013 image , @xmath102 @xmath96 g ( which implies @xmath103 ) .
the corresponding magnetic and electron energies are @xmath104 erg and @xmath105 erg , respectively , in the volume @xmath106 @xmath56 .
the total energy , @xmath107 , in the emitting volume is much smaller than the energy lost by the pulsar spin - down during one orbital period , @xmath108 , in a broad range of magnetization parameter , @xmath109 , which means that the cloud energy can be supplied by the pulsar in a fraction of the binary period . in the following subsections we describe two scenarios that could explain the origin of the extended synchrotron - emitting structure .
the first scenario assumes that we observed the same cloud in december 2011 and may 2013 , which had been moving with the average velocity @xmath110 since the launch .
however , if such a cloud is created and ejected from the binary in the course of interaction of the pulsar wind with the dense excretion disk close to the nearest preceding periastron ( i.e. , about one year before the first observation ) , it should decelerate very rapidly .
the drag force exerted on the light , fast - moving cloud by the ambient medium is very large : @xmath111 , where @xmath112 cm@xmath33 is the frontal area of the cloud , and @xmath113 is the ambient density .
the deceleration caused by this force can be estimated as @xmath114 cm s@xmath115 , where @xmath116 erg is the total ( particle and magnetic field ) energy in the emitting volume .
the characteristic deceleration time is then @xmath117 s. this means that if the circumbinary medium is filled with the stellar wind of the high - mass companion , the cloud of relativistic electrons would be decelerated and/or destroyed in a time much shorter than the interval of 519 days between our two observations .
the deceleration time could be larger if the cloud is loaded with a substantial amount of the o - star wind .
for instance , it will exceed the apparent travel time from the binary , @xmath118 d , if @xmath119 g , which still requires a rather low ambient density , @xmath120 @xmath54 , to be consistent with the mass and energy budget .
therefore , this scenario seems rather unlikely
. the drag force could be much smaller if the matter ahead of the cloud is mostly comprised of the pulsar wind particles and is moving away from the binary with a high velocity .
the presence of such a lepton - dominated medium would be more plausible on the apastron side of the orbit where the pulsar spends most of the time .
overall , the moving cloud scenario seems to require somewhat artificial assumptions , but it can not be firmly ruled out without further investigation ( e.g. , another _ cxo _ observation could confirm or rule out the translational motion of the cloud ) . in the second scenario , which we consider more plausible , the extended structures are formed at the _ extrabinary termination shock _ in the pulsar wind at large distances from the binary , comparable to those observed in pwne around isolated pulsars @xcite , so that the different locations of the structures in december 2011 and may 2013 are due to different conditions at the epochs of their formation , rather than to the motion of a cloud of relativistic electrons from the binary .
formation of such a distant termination shock can be understood as follows . in a high mass binary
the pulsar s wind collides with the companion s wind and forms a double - shock structure within the binary and in its vicinity @xcite .
if the pulsar wind is dynamically dominant , it escapes freely in the directions away from the high - mass companion and , similar to an isolated pulsar wind , shocks in the ambient medium at a distance @xmath121 , where the wind s ram pressure becomes comparable to the ambient pressure , @xmath122 .
synchrotron radiation from the shocked pulsar wind can be observed as an extrabinary pwn . as the pulsar orbits the heavy companion , this extrabinary shock rotates around the binary , seeding the medium with x - ray emitting relativistic electrons ( positrons ) .
if their cooling time is larger than the orbital period , one can expect a uniformly bright extended structure to form around a circular binary at distances @xmath123 ; such a structure would be brighter on the apastron side for an eccentric binary because the pulsar spends more time around apastron than around periastron .
the location of the structure can vary due to variations in the ambient pressure or because of instabilities that can destroy the contact discontinuity , separating the two winds , and mix the stellar wind matter into the pulsar wind .
the extended x - ray structures detected near the b1259 binary could be associated with such an extrabinary termination shock formed at a distance @xmath124 cm , where @xmath125 dyn @xmath31 .
the corresponding angular separation from the binary , @xmath126 ( where @xmath127 is the angle between the line of sight and the line connecting the binary and the emitting region ) is close to the observed separations of the detected extended structures at @xmath128 .
such high pressures are possible in wind - blown bubbles around massive stars @xcite .
the magnetic field immediately beyond the shock can be estimated as @xmath129 @xmath96 g for @xmath130 , where @xmath131 is the pre - shock magnetization parameter @xcite .
this field is consistent with @xmath132 @xmath96 g , inferred assuming a synchrotron origin of the detected emission .
the synchrotron cooling time in such a field , @xmath133 yr , exceeds the orbital period , which means that the pwn structure created in a given orbital cycle could be seen in a few following cycles , and its azimuthal extent around the binary can be rather large .
the latter conclusion does not contradict to the small azimuthal extent ( @xmath134 ) of the observed structures because the pwn brightness is proportional to the amount of pulsar wind particles supplied in a given direction .
this amount strongly varies throughout the eccentric orbit , with a maximum around apastron .
it should be noted that a major restructuring of both the extrabinary and intrabinary pwn components is expected when the pulsar moves through the dense excretion disk of the o star around periastron .
however , since the pulsar spends only a small fraction ( @xmath135 ) of the orbital period in this stage but sweeps a large range ( @xmath136 ) of true anomalies ( see figure [ orb ] ) , we do not expect a significant contribution of this stage to the extrabianry pwn .
the observed shift of the extended structure during 1.5 years can be reconciled with the large @xmath137 if the apparent motion is interpreted as a shift of the termination shock caused by a 30% change of the ratio of the pulsar - wind pressure to the ambient pressure .
the extended emission was the faintest during the 2009 observation , which occurred near apastron at the true anomaly intermediate between those of the 2011 and 2013 observations .
therefore , the extended emission flux is not a periodic function of true anomaly , even for adjacent orbital cycles , which could be explained either by variations in the density / velocity of the o - star wind or by instabilities developing at the interface between the winds .
these arguments may also explain a factor of three different x - ray fluxes from the entire binary ( at nearly the same true anomalies ) seen in the earlier observations with other satellites ( see figure 2 from @xcite ) . in this interpretation we neglected the role of possible mixing of the stellar wind into the pulsar wind , which could lead to mass loading and deceleration of the pulsar wind at large distances from the binary
the unexpected morphology and transient behavior of the resolved extended emission have not been captured in the existing models of such systems .
most likely , the observed structures can be interpreted as the synchrotron emission from the pulsar wind shocked by the interaction with the circumbinary medium or with the heavy stellar wind entrained into the pulsar wind outflow .
the appearance of the pwn and the parameter estimates suggest that in either case most of the observed extended emission comes from the pulsar wind ejected from the binary near apastron .
if our interpretation of the discovered x - ray emission is correct , then b1259 can be used to probe the properties of stellar and pulsar winds , to study baryonic mass loading of pulsar winds , and to shed light on the complex relativistic shock dynamics .
systematic monitoring of b1259 , using high - resolution x - ray , uv , optical , ir , and mm - band observations , is needed to fully utilize this unique laboratory offered by nature .
* acknowledgments : * this work was supported by national aeronautics space administration grants nnx09ac81 g and nnx09ac84 g , and through _ cxo _ award go2 - 13085 issued by the _ cxo _
x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of the national aeronautics space administration under contract nas8 - 03060 . , o. , & pavlov , g. g. 2008 , in american institute of physics conference series , vol .
983 , 40 years of pulsars : millisecond pulsars , magnetars and more , ed . c. bassa , z. wang , a. cumming , & v. m. kaspi , 171 , j. , rib , m. , & paredes , j. m. 2012 , in american institute of physics conference series , vol . 1505 , american institute of physics conference series , ed . f. a. aharonian , w. hofmann , & f. m. rieger , 386 , a. j. , langer , n. , & garca - segura , g. 2004 , in revista mexicana de astronomia y astrofisica conference series , vol .
22 , gravitational collapse : from massive stars to planets , ed .
g. garcia - segura , g. tenorio - tagle , j. franco , & h. w. yorke , 136 | we present observations of the eccentric @xmath0-ray binary b125963/ls2883 with the _ chandra _ x - ray observatory .
the images reveal a variable , extended ( about @xmath1 , or @xmath2 times the binary orbit size ) structure , which appears to be moving away from the binary with the velocity of 0.05 of the speed of light .
the observed emission is interpreted as synchrotron radiation from relativistic particles supplied by the pulsar .
however , the fast motion through the circumbinary medium would require the emitting cloud to be loaded with a large amount of baryonic matter .
alternatively , the extended emission can be interpreted as a variable extrabinary shock in the pulsar wind outflow launched near binary apastron .
the resolved variable x - ray nebula provides an opportunity to probe pulsar winds and their interaction with stellar winds in a previously inaccessible way . |
the ising model @xcite remains probably the most cited model in statistical physics .
today , the _ isi web of knowledge _ abstracting and indexing service returns over eleven thousands records for the query on `` ising '' for time span from 1996 to 2010 . for _ inspec
_ database ( for years 1969 - 2010 ) this number is almost doubled and reaches 19 thousands for _ scopus _ database ( for data range 1960 - 2010 ) .
the latter means that during the last half of century @xmath18 papers refer to the ising model every year .
the _ google _
search engine indicates over 279 thousands web pages which contain `` ising model '' phrase .
the beauty and the popularity of this model lies in both its simplicity and possible applications from pure and applied physics , via life sciences to social sciences . in the way similar to the percolation phenomenon ,
the ising model is one of the most convenient way of numerical investigations of second order phase transitions . in the simplest case
, the ising model may be used to simulate the system of interacting spins which are placed at the nodes of graphs or regular lattices . in its basic version only two values of the spin variable
are available , i.e. @xmath19 and @xmath20 .
this is the classical ising @xmath21 model . for a square lattice
this model defines the universality class of phase transitions with analytically known critical exponents which describe the system behaviour near the critical point .
the critical point separates two ordered and disordered phases .
one of possible generalisation of the ising model is to enlarge the set of possible spin values ( like in the potts model @xcite ) .
the ising @xmath1 model corresponds to three possible spin values , i.e. @xmath22 , ising @xmath23 allows for four spin variables @xmath24 , _ etc_. the ising @xmath25 model on various networks and lattices may form universality classes other than the classical square lattice ising model .
the spin models for @xmath1 were extensively studied by several approximate techniques in two and three dimensions and their phase diagrams are well known @xcite .
the case @xmath26 has also been investigated according to several procedures @xcite .
the ising model @xmath1 on directed barabsi
albert network was studied by lima in 2006 @xcite .
it was shown , that the system exhibits first - order phase transition .
the result is qualitatively different from the results for this model on a square lattice , where a second - order phase transition is observed . in this paper
we study the ising @xmath1 model on two archimedean lattices ( al ) , namely on @xmath2 and @xmath3 .
the topologies of @xmath2 and @xmath3 al are presented in fig .
[ fig - al ] .
critical properties of these lattices were investigated in terms of site percolation in ref .
topologies of all eleven existing al are given there as well .
also the critical temperatures for ising @xmath21 model @xcite and voter model @xcite on those al were estimated numerically . here , with extensive monte carlo simulations
we show that the ising @xmath1 model on @xmath2 and @xmath3 al exhibits a second - order phase transition with critical exponents that _ do not _ fall into universality class of the square lattice ising @xmath21 model .
we consider the two - dimensional ising @xmath1 model on @xmath2 and @xmath3 al lattices .
the hamiltonian of the system can be written as @xmath27 where spin variable @xmath28 takes values @xmath29 ,
@xmath30 , @xmath31 and decorates every @xmath32 vertex of the al . in eq .
@xmath33 is the magnetic exchange coupling parameter .
the simulations have been performed for different lattice sizes @xmath34 , 16 , 32 , 64 and 128 . for each system with @xmath32 spins and given temperature @xmath35 we performed monte carlo simulation in order to evaluate the system magnetisation @xmath36 .
the simulations start with a uniform configuration of spins ( @xmath37 , but the results are independent on the initial configuration ) .
it takes @xmath38 monte carlo steps ( mcs ) per spin for reaching the steady state , and then the time average over the next @xmath38 mcs are estimated .
one mcs is accomplished when all @xmath39 spins are investigated whether they should flip or not .
we carried out @xmath40 to @xmath41 independent simulations for each lattice and for given set of parameters @xmath42 .
we have employed the heat bath algorithm for the spin dynamic .
we evaluate the average magnetisation @xmath43 , the susceptibility @xmath44 , and the magnetic 4-th order cumulant @xmath45 : [ eq - mchiu ]
@xmath46 @xmath47 @xmath48 where @xmath49 and @xmath50 is the boltzmann constant . in the above equations @xmath51 stands for thermodynamic average
. in the infinite - volume limit these quantities exhibit singularities at the transition point @xmath52 . in finite systems the singularities are smeared out and scale in the critical region according to [ ffs ] @xmath53 @xmath54 where @xmath6 , @xmath55 and @xmath56 are the usual critical exponents , and @xmath57 are finite size scaling ( fss ) functions with @xmath58 being the scaling variable . therefore , from the size dependence of @xmath43 and @xmath44
one can obtain the exponents @xmath5 and @xmath4 , respectively .
the maximum value of susceptibility also scales as @xmath59 .
moreover , the value of temperature @xmath60 for which @xmath44 has a maximum , is expected to scale with the system size as @xmath61 where the constant @xmath62 is close to unity @xcite .
therefore , the eq . @xmath63 may be used to determine the exponent @xmath9 .
we have checked also if the calculated exponents satisfy the hyper - scaling hypothesis @xmath64 in order to get the effective dimensionality , @xmath65 , for both investigated al lattices .
the dependence of the magnetisation @xmath43 on the temperature @xmath35 , obtained from simulations on @xmath2 and @xmath3 al with @xmath32 ranging from @xmath66 to @xmath67 sites is presented in fig .
[ fig - m ] .
the shape of magnetisation curve versus temperature , for a given value of @xmath39 , suggests the presents of the second - order transition phase in the system .
the phase transition occurs at the critical value @xmath52 of temperature . in order to estimate the critical temperature @xmath52
we calculate the fourth - order binder cumulants given by eq . .
it is well known that these quantities are independent of the system size at @xmath52 and should intercept there @xcite . in fig .
4 the fourth - order binder cumulant is shown as a function of the temperature for several values of @xmath68 . taking two largest lattices ( for @xmath69 and @xmath70 ) we have @xmath71 and @xmath72 for @xmath2 and @xmath3 al , respectively . in order to go
further in our analysis we also computed the modulus of the magnetisation at the inflection @xmath73 .
the estimated exponents @xmath5 values are @xmath74 and @xmath75 for @xmath2 and @xmath3 al , respectively . to obtain the critical exponent @xmath9
, we used the scaling relation .
the calculated values of the exponents @xmath9 are @xmath79 for @xmath2 and @xmath80 for @xmath3 .
yields effective dimensionality of the systems @xmath16 for @xmath2 and @xmath17 for @xmath3 .
the above results , indicate that the ising @xmath1 model on @xmath2 and @xmath3 al _ does not fall _ in the same universality class as the square lattice ising model , for which the critical exponents are known analytically i.e. @xmath81 , @xmath82 and @xmath83 .
we have checked numerically , that ising @xmath21 model reproduces these critical exponents with reasonable accuracy for both studied lattices @xcite .
we improved the value of the critical temperature @xmath52 for these two lattices and @xmath21 as well , with respect to ref .
@xcite .
fwsl acknowledges the brazilian agency fapepi ( teresina , piau , brazil ) for its financial support .
this work also was supported ( in part ) from the agh - ust project 11.11.220.01 .
part of the calculations was carried out on _ sgi altix 1350 _ system at the computational park cenapad.unicamp-usp , so paulo , brazil and on _ sgi altix 3700 _ system at academic computer centre cyfronet - agh in cracow , poland .
m. blume , phys .
rev . * 141 * , 517 ( 1966 ) .
h. w. capel , physica ( amsterdam ) * 32 * , 966 ( 1966 ) .
d. m. saul , m. wortis , d. staufer , phys .
b * 9 * , 4964 ( 1974 ) . a. k. jain , d. p. landau , phys . rev .
b * 22 * , 445 ( 1980 ) . o. f. de alcantara bonfim , physica a * 130 * , 367 ( 1985 ) . a. n. berker , m. wortis , phys . rev .
b * 14 * , 4946 ( 1976 ) .
s. moss de oliveira , p. m. c. de oliveira , f. c. s barreto , j. stat .
phys . * 78 * , 1619 ( 1995 ) .
j. a. plascak , j. g. moreira , f. c. s barreto , phys .
a * 173 * , 360 ( 1993 ) .
m. n. tamashiro , s. r. salinas , physica a * 211 * , 124 ( 1994 ) .
j. c. xavier , f. c. alcaraz , d. pea lara , j. a. plascak , phys .
b * 57 * , 11575 ( 1998 )
. d. pea lara , j. a. plascak , int . j. mod .
b * 12 * , 2045 ( 1998 ) .
f. c. s barreto , o. f. alcantara bonfim , physica a * 172 * , 378 ( 1991 ) .
a. bakchinch , a. bassir , a. benyoussef , physica a * 195 * , 188 ( 1993 ) . | we investigate the critical properties of the ising @xmath0 and @xmath1 model on @xmath2 and @xmath3 archimedean lattices .
the system is studied through the extensive monte carlo simulations .
we calculate the critical temperature as well as the critical point exponents @xmath4 , @xmath5 and @xmath6 basing on finite size scaling analysis .
the calculated values of the critical temperature for @xmath1 are @xmath7 and @xmath8 for @xmath2 and @xmath3 archimedean lattices , respectively .
the critical exponents @xmath5 , @xmath4 and @xmath9 for @xmath1 are @xmath10 , @xmath11 and @xmath12 for @xmath2 and @xmath13 , @xmath14 and @xmath15 for @xmath3 archimedean lattices . obtained results differ from the ising @xmath0 model on @xmath2 , @xmath3 and square lattice .
the evaluated effective dimensionality of the system for @xmath1 are @xmath16 for @xmath2 and @xmath17 for @xmath3 . |
there is growing evidence that soft gamma repeaters ( sgrs ) could be magnetars experiencing starquakes that are connected ( through the intense magnetic field ) to gamma ray flare activity .
magnetars are thought to be neutron stars with very strong magnetic fields , greatly exceeding @xmath2 g @xcite .
the flare activity consist of an initial short peak in the hard part of the spectrum , followed by a decaying softer part ( tail ) which last for hundreds of seconds . recent analysis of detailed observations from several sgrs @xcite revealed that the decaying part of the spectrum exhibits a number of oscillations with frequencies in the range of a few tenths of hz to a few hundred hz .
there are three events detected up to now which can be associated with crust oscillations of magnetars .
the first event was detected already in 1979 from the source sgr 0526 - 66 @xcite , the second in 1998 from sgr 1900 + 14 @xcite , while the third and most energetic one was observed in december 2004 from the source sgr 1802 - 20 @xcite .
analysis of the tail oscillations of sgr 1806 - 20 revealed the presence of oscillations at approximately 18 , 26 , 29 , 92.5 , 150 , 626.5 , and 1837 hz @xcite , for sgr 1900 + 14 the detected frequencies were 28 , 54 , 84 , and 155 hz @xcite , while for the case of sgr 0526 - 66 there was only one frequency identified at 43.5 hz @xcite .
magnetar starquakes may be driven by the evolving intense magnetic field , which accumulates stress and eventually leads to crust fracturing .
the excited seismic oscillations are known to be of two types : _ shear modes _ , which are polar - type oscillations and _ torsional modes _ , which are axial - type oscillations .
the latter are thought to be more easily excited during a fracturing of the crust , since they only involve oscillations of the velocity , with the shear modulus of the crust acting as the necessary restoring force .
slowly rotating stars still maintain a nearly spherical shape and the velocity field of torsional oscillations is then divergence - free , with no radial component .
torsional modes ( _ t_-modes ) are labeled as @xmath3 , where @xmath0 is the angular index , while the index @xmath4 corresponds to the number of radial nodes in the eigenfunctions of the overtones for a specific @xmath0 .
shear and torsional modes have been studied mainly in newtonian theory , see e.g. @xcite , while there are only a few studies of torsional modes in general relativity @xcite . in the newtonian
, non - magnetized limit @xcite found that the period @xmath5 of the fundamental torsional modes depends mostly on the radius of the star , @xmath6 , the speed of shear waves , @xmath7 , and the angular index @xmath0 , via @xmath8^{-1/2 } r / u_s , \ ] ] where @xmath9 , with @xmath10 and @xmath11 being the shear modulus and the density , respectively .
furthermore , it was found that the period of overtones of index @xmath4 is essentially independent of the angular index @xmath12 ( provided @xmath12 does not greatly exceed @xmath4 ) and is basically determined by the crust thickness @xmath13 : @xmath14 from the above two relations it follows that the relative crust thickness , @xmath15 , is given by @xmath16 which is independent of the details of the equation of state in the crust and basically only depends on the ratio of the frequency of the overtones to the fundamental frequency , see also @xcite .
it thus becomes obvious that the successful identification of both the fundamental torsional frequencies and their overtones could allow for the determination of the crust thickness , which in turn could lead to information on the high - density part of the equation of state .
in such a determination , of course , general relativistic and magnetic field effects on the frequencies of torsional modes should properly be taken into account , which is one of the aims of the present paper .
relativistic effects have been found to significantly increase the fundamental @xmath17 torsional mode period by roughly 30% for a typical @xmath18 , @xmath19 model .
@xcite derived the following empirical formula for the period of the fundamental @xmath17 torsional mode of a relativistic star @xmath20 where @xmath21 and @xmath22 . in this relation ,
two effects have been taken into account , the ( relativistic ) gravitational redshift and the ( newtonian ) acceleration of gravity @xmath23 at the surface .
more details can be found in @xcite , see also @xcite .
in addition , the formula ( [ eq : duncan ] ) agrees well with the fully relativistic empirical relation found by @xcite @xmath24 the latter is an extension of the newtonian hansen & cioffi formula , based on a particular @xmath25 model . on the other hand ,
the effect of a strong magnetic field , @xmath26 , on torsional modes was considered by @xcite and @xcite . assuming that the shear modulus @xmath10 is augmented by the magnetic field tension , @xmath27
, one can derive the following estimate for the period of a torsional mode @xmath28^{-1/2 } , \label{pb}\ ] ] where @xmath29 is the period in the limit of vanishing magnetic field and @xmath30 .
however , @xcite found that if the structure of the magnetic field is close to monopolar in the crust , the period of the fundamental @xmath17 torsional mode does not follow the above relation .
this underlines the importance of properly taking into account the structure of the magnetic field , so that equation ( [ pb ] ) must be generalized to particular magnetic field multipoles .
an essential ingredient in computing torsional mode frequencies is the assumption one makes about the shear modulus @xmath10 .
here , we adopt the zero - temperature limit of the approximate formula derived by @xcite @xmath31 where @xmath32 is the ion number density , @xmath33 is the average ion spacing , and @xmath34 is the ion charge .
notice that this approximate formula for the shear modulus is derived on the assumption of a bcc crystalline state , which is averaged over all directions . for the detailed composition of the crust
, we consider a modern eos by @xcite ( hereafter dh ) and , for comparison , an older eos by @xcite ( hereafter nv ) .
the two eoss for the crust differ both in composition as well as in the value of the density at the base of the crust . as the crust thickness depends on the compactness of the star , we explore the allowed parameter space for the unknown high - density eos by selecting several tabulated high - density eoss that yield neutron star models with significantly different macroscopic properties . in the present work ,
we derive the perturbation equations for torsional oscillations of magnetized stars in general relativity following the approach by @xcite and neglecting spacetime perturbations ( cowling approximation ) .
we specialize the magnetic field to a dipole structure , which allows us to reduce the system of equations to only one dimension . a large number of torsional modes ( fundamental and overtones )
is obtained for selected magnetar models differing in the high - density eos , the crust eos and mass .
our study is valid for moderate magnetic field strengths , up to roughly @xmath35 g , assuming that the magnetic field is confined to the crust and neglecting the possible presence of a thin fluid ocean .
the reason for this is twofold : on one hand , for significantly stronger magnetic fields the distortion of the equilibrium shape of the star as well the magnetic coupling to higher harmonics should also be taken into account ; on the other hand , the alfvn velocity @xmath36 @xmath37 cm / s @xmath38 becomes comparable to the speed of shear waves in the crust when the magnetic field reaches values of 10@xmath39 - 10@xmath40 g , which implies that global magnetosonic waves will be strongly coupled to shear waves and the torsional oscillations may no longer be confined to the crust .
the effect of including global magnetosonic waves even at smaller magnetic field strengths should also be studied in more detail , see @xcite .
the article is structured as follows : in the next two sections we describe the general - relativistic ideal mhd equations , the background stellar configuration and the way that we have derived the dipole geometry of the magnetic field . in the fourth section
we derive the perturbation equations for torsional oscillations of magnetized stars in the cowling approximation , while in the fifth section we present the obtained frequencies for selected magnetar models along with fitting formulae for the influence of the magnetic field in the oscillation spectrum .
the article closes with a summary and discussion .
unless otherwise noted , we adopt units of @xmath41 , where @xmath42 and @xmath43 denote the speed of light and the gravitational constant , respectively , while the metric signature is @xmath44 .
the energy momentum tensor @xmath45 for a magnetized relativistic star in equilibrium , is the sum of the stress - energy tensors of a perfect fluid @xmath46 and the magnetic field @xmath47 @xmath48 where @xmath49 above , @xmath50 , @xmath51 , and @xmath52 are the energy density , the pressure , and a normalized magnetic field , which we define by absorbing a factor of @xmath53 , i.e. , @xmath54 , and @xmath55 . here and throughout the paper we will assume the _ ideal mhd _ approximation .
in addition , we assume the shear modulus to be isotropic , so that there is no contribution of the shear stress in equilibrium . using the stress - energy tensor ( [ t_munu ] ) , one obtains the equations of motion of the fluid by projecting the conservation of the energy - momentum tensor on to the hypersurface normal to @xmath56 , i.e. , @xmath57 where @xmath58 is the projection tensor .
the equations of motion are @xmath59 due to the ideal mhd approximation the electric field 4-vector vanishes , @xmath60 so that the electric field is zero for a comoving observer . in this approximation
, the maxwell s equations @xmath61}=0 $ ] , where @xmath62 is the faraday tensor , can be written in the simple form : @xmath63 in deriving the above equation the following definitions are used @xmath64 from equation ( [ maxwell0 - 1 ] ) one derives the magnetic induction equation @xmath65 where @xmath66 is the rate of shear tensor , @xmath67 is the twist tensor and @xmath68 is the expansion , defined as @xmath69 in a following section , the above equations will be linearized , in order to obtain the perturbed equations of motion describing small - amplitude oscillations . but
first , we describe the construction of the background equilibrium models .
a strongly magnetized relativistic star has a non - spherical shape , due to the tension of the magnetic field , which is expressed via the off - diagonal components in the stress - energy tensor @xmath45 .
however , the deformations from spherical symmetry induced by the magnetic field are small for neutron stars with magnetic field usually assumed for magnetars , because the energy of the magnetic field is considerably smaller then the gravitational energy , i.e. , @xmath70}\right)^2.\ ] ] for this reason , we neglect the deformations due to magnetic fields in the construction of equilibrium models . since magnetars are also extremely slowly rotating ( the intense magnetic field spins down the star on a short timescale ) we can also neglect any rotational deformations , so that the equilibrium models can be considered as spherically symmetric solutions of the well - known tov equations described by a metric of the form @xmath71 where @xmath72 and @xmath73 are the function of the schwarzschild radial coordinate @xmath74 .
the four - velocity of the equilibrium model is thus @xmath75 we supplement the equilibrium model by a dipole magnetic field ( but see @xcite for other assumptions ) , following the approach of @xcite i.e. we consider an axisymmetric , poloidal magnetic field , which is created by a 4-current @xmath76 . in ideal mhd
the electromagnetic 4-potential @xmath77 is very simple and has only one component in spherical polar coordinates @xmath78 with @xmath79 @xcite . for the metric chosen above , maxwell s equations @xmath80 lead to the following elliptic equation for @xmath81 @xmath82 the 4-vectors @xmath77 and
@xmath83 can be expanded in vector spherical harmonics , so that @xmath84 where @xmath85 and @xmath86 are functions of the radial coordinate only .
this leads to a one - dimensional form of equation ( [ maxwell - j ] ) @xmath87 for a dipole magnetic field , i.e. , @xmath88 , an analytic exterior solution for @xmath89 exists , setting @xmath90 and @xmath91 @xcite @xmath92,\ ] ] where @xmath93 is the magnetic dipole moment for an observer at infinity . in the interior
the potential @xmath94 is found by solving numerically equation ( [ maxwell - j1 ] ) .
the current distribution @xmath95 , which is the source term of equation ( [ maxwell - j1 ] ) must satisfy the integrability condition that the solution is stationary @xcite .
here we adopt the same , simplified form for the current distribution as the one used in @xcite , i.e. , @xmath96 where @xmath97 is an arbitrary constant . moreover , regularity at the center of the star implies that @xmath94 has the form @xmath98 where @xmath99 is a constant , which is determined by the junction condition with @xmath100 at stellar surface .
finally , the components of the vector @xmath101 describing the normalized magnetic field are given by the following set of equations @xmath102 figure [ fig : dipole_field ] shows the magnetic field lines of the complete solution , for a typical equilibrimum configuration .
we emphasize that , even though the magnetic field configuration has been computed throughout the star , the boundary conditions we will assume for the torsional oscillations will neglect the presence of the magnetic field outside the crust region , so that , effectively , that magnetic field will be assumed to be confined to the crust only .
the perturbation equations are derived by linearizing the equations of motion ( [ conservation ] ) and the magnetic induction equation ( [ maxwell1 ] ) .
since we only consider torsional oscillations , which are of axial type and do not induce density variations in spherical stars , there is no significant variation in the radiative part of the metric describing the gravitational field . in addition , since torsional modes are essentially material velocity oscillations , the imaginary part of their frequency due to the emission of current - multipole gravitational radiation is much smaller than the real part . for these reasons
the frequency of torsional oscillations is determined with satisfactory accuracy even when neglecting entirely the metric perturbations by setting @xmath103 ( i.e. adopting the relativistic cowling approximation ) , see @xcite and we adopt this approximation here .
the linearized form of the equation of motion ( [ conservation ] ) is @xmath104_{;\nu } \nonumber \\ & + & h^{\mu}_{\ \alpha}\left[h^{\alpha } \delta h^{\nu } + \delta h^{\alpha}h^{\nu } - g^{\alpha\nu}\left(\delta p + h_{\beta}\delta h^{\beta}\right)\right]_{;\nu } - h^{\mu}_{\ \alpha}\delta t^{\alpha\nu(s)}_{\ \ ; \nu}. \label{perturbation1}\end{aligned}\ ] ] where @xmath105 is the linearized shear stress tensor .
the latter is assumed to be related to the linearized shear tensor @xmath106 , through @xmath107 see @xcite .
in addition , linearization of the magnetic induction equation ( [ maxwell1 ] ) yields @xmath108 where we used @xmath109 .
notice that the above relation differs in some terms from equation ( 12 ) in @xcite , due to a sign error in the latter reference ) . ] . in appendix
[ sec : iv ] we also give the corrected equations and numerical results for the special magnetic field configuration considered in @xcite . the usual method for studying matter perturbations of spherically symmetric backgrounds is to decompose the perturbed quantities into scalar and vector spherical harmonics of definite indices @xmath0 and @xmath110 . in this way ,
the different oscillation modes characterized by @xmath0 and @xmath110 decouple and one can study them independently . here
we restrict attention to axial - type perturbations , for which the perturbed matter quantities can be written as @xmath111 where @xmath112 where @xmath113 denotes the partial derivative with respect to @xmath114 . above , @xmath115 describes the radial dependence of the angular displacement of the stellar material , @xmath116 is the legendre polynomial of order @xmath0 and we have set @xmath117 , due to the degeneracy in @xmath110 for spherically symmetric backgrounds .
the above assumptions lead to the following form of the @xmath118-component of the linearized equation of motion ( [ perturbation1 ] ) @xmath119 + \delta h^{\theta } \left[h^{\phi}_{\ , \theta } + 3\cot\theta h^{\phi } - \frac{h^{\theta}_{\ , \phi}}{\sin^2\theta}\right ] \nonumber \\ & + & \delta h^{\phi } \left[2\cot\theta h^{\theta } + \left(\phi_{,r } + \frac{2}{r}\right)h^r\right ] + h^{\phi}\left(\delta h^t_{\ , t } + \delta h^r_{\ , r } + \delta h^{\theta}_{\ , \theta } + \delta h^{\phi}_{\ , \phi}\right ) + h^r \delta h^{\phi}_{\ , r } + h^{\theta}\delta h^{\phi}_{\ , \theta } \nonumber \\ & - & \frac{h^r \delta h_{r,\phi}}{r^2\sin^2\theta } - \frac{h^{\theta}\delta h_{\theta,\phi}}{r^2\sin^2\theta } - \delta t^{r\phi(s)}_{\ \ , r } - \delta t^{\theta\phi(s)}_{\ \ ,
\theta } - \left(\frac{4}{r } + \phi_{,r } + \lambda_{,r}\right)\delta t^{r\phi(s ) } - 3\cot\theta \delta t^{\theta\phi(s ) } , \label{perturbation - equation1}\end{aligned}\ ] ] where we used the relation @xmath120 , which is derived from the maxwell s equations . for the derivation of the linearized shear tensor @xmath106
, we adopt the relationship @xmath121 @xcite , which becomes @xmath122 in the cowling approximation the only nonvanishing components of @xmath123 are written as @xmath124 it follows that the only non - vanishing components of the linearized shear stress tensor are @xmath125 the linearized induction equation ( [ perturbation2 ] ) yields that following relations for the first time - derivative of the components of @xmath126 @xmath127.\end{aligned}\ ] ] substituting the previous relations into the linearized equations of motion ( [ perturbation - equation1 ] ) and assuming that all perturbed variables have a harmonic time dependence , such that @xmath128 , one obtains @xmath129 \omega^2 e^{-2\phi } { \cal y } & = & e^{-2\lambda}\mu { \cal y } '' + \left[\left(\frac{4}{r } + \phi ' - \lambda ' \right)\mu + \mu'\right]e^{-2\lambda}{\cal y } ' - \frac{(\ell+2)(\ell-1)}{r^2}\mu { \cal y } \nonumber \\ & - & \bigg\{h^r_{\ , \phi } h^{\phi}_{\ , r } + h^r_{\ , \phi } h^{\phi } \left(\phi ' + \lambda ' + \frac{4}{r}\right ) + h^{\theta}_{\ , \phi } h^{\phi}_{\ , \theta } + 3\cot\theta h^{\theta}_{\ , \phi } h^{\phi } \nonumber \\ & - & \frac{1}{r^2\sin^2 \theta}e^{2\lambda}\left[(h^r_{\ , \phi})^2 + h^rh^r_{\ , \phi\phi}\right ] - \frac{1}{\sin^2\theta}\left[(h^{\theta}_{\ , \phi})^2 + h^{\theta}h^{\theta}_{\ , \phi\phi}\right ] \nonumber \\ & + & h^{\phi}_{\ , \phi } \left[2\cot\theta h^{\theta } + \left(\phi ' + \frac{2}{r}\right)h^r\right ] + h^{\phi}\left(h^r_{\ , r\phi } + h^{\theta}_{\ , \theta\phi } + h^{\phi}_{\ , \phi\phi}\right ) + h^rh^{\phi}_{\ , r\phi } \nonumber \\ & + & h^{\theta}h^{\phi}_{\ , \theta\phi}\bigg\ } { \cal y } + \left[2\cot\theta h^rh^{\theta } + \left(\phi ' + \frac{2}{r}\right)(h^r)^2 + h^r \left(h^r_{\ , r } - h^{\phi}_{\ , \phi}\right ) + h^{\theta}h^r_{\ , \theta}\right ] { \cal y } ' \nonumber \\ & + & ( h^r)^2 { \cal y } '' + \bigg\{\left[\left(\phi ' + \frac{2}{r}\right)h^rh^{\theta } + h^rh^{\theta}_{\ , r } + 2\cot\theta ( h^{\theta})^2
+ h^{\theta } \left(h^{\theta}_{\ , \theta } - h^{\phi}_{\ , \phi}\right)\right]{\cal y } \nonumber \\ & + & 2h^rh^{\theta}{\cal y}'\bigg\}\frac{b_{,\theta}}{b } + \left(h^{\theta}\right)^2 { \cal y } \frac{b_{,\theta\theta}}{b } , \label{eigen}\end{aligned}\ ] ] where a prime @xmath130 denotes the derivative with respect to @xmath74 .
the above eigenvalue equation for the mode - frequency @xmath131 is still written for a general equilibrium magnetic field @xmath132 .
restricting attention to an axisymmetric poloidal magnetic field , the eigenvalue equation ( [ eigen ] ) becomes @xmath133\omega^2 e^{-2\phi } { \cal y } & = & \ e^{-2\lambda}\mu { \cal y } '' + \left[\left(\frac{4}{r } + \phi ' - \lambda ' \right)\mu + \mu'\right]e^{-2\lambda}{\cal y } ' - ( \ell+2)(\ell-1 ) \left[\frac{\mu}{r^2 } + ( h^{\theta})^2\right ] { \cal y } \nonumber \\ & + & \left[2\cot\theta h^rh^{\theta } + \left(\phi ' + \frac{2}{r}\right)(h^r)^2 + h^rh^r_{\ , r } + h^{\theta}h^r_{\ , \theta}\right ] { \cal y } ' + ( h^r)^2 { \cal y } '' \nonumber \\ & + & \left\{\left[\left(\phi ' + \frac{2}{r}\right)h^rh^{\theta } + h^rh^{\theta}_{\ , r } - \cot\theta ( h^{\theta})^2 + h^{\theta}h^{\theta}_{\ , \theta}\right]{\cal y } + 2h^rh^{\theta}{\cal y}'\right\}\frac{b_{,\theta}}{b } , \label{master - equation}\end{aligned}\ ] ] where we used the following equation satisfied by the angular function @xmath134 @xmath135 the above eigenvalue equation can be simplified further by assuming a dipole magnetic field of the form ( [ eq : h_r ] ) , ( [ eq : h_theta ] ) .
we achieve this by first writing equation ( [ master - equation ] ) in the form @xmath136 \partial_{\theta}y_{\ell 0 } + [ { \cal c}_\ell(r ) - { \cal b}_\ell(r ) ] \sin^2 \theta \partial_{\theta}y_{\ell 0 } + { \cal d}_\ell(r ) \sin
\theta \cos \theta y_{\ell 0 } = 0 , \label{master - equation1}\end{aligned}\ ] ] where @xmath137 is the spherical harmonic function with @xmath117 , which is related to @xmath138 via the normalization @xmath139 .
the four radial functions @xmath140 , @xmath141 , @xmath142 , and @xmath143 are @xmath144 { \cal y } ' + \left[\left(\epsilon + p \right ) \omega^2 e^{-2\phi } - \frac{(\ell+2)(\ell-1)}{r^2}\mu\right]e^{2\lambda } { \cal y } , \\
{ \cal b}_\ell(r ) & = & \frac{a_1}{\pi r^4 } \left\{a_1 { \cal y } '' + \left[\left(\phi ' - \lambda ' \right)a_1 + 2{a_1 } ' \right ] { \cal y } ' + \left[\omega^2 e^{-2\phi + 2\lambda } a_1 + \left(\phi ' - \lambda ' \right){a_1 } ' + { a_1 } '' \right ] { \cal y}\right\ } , \\ { \cal c}_\ell(r ) & = & \frac{{a_1}'}{4\pi r^4}\left\{2a_1 { \cal y } ' + \left[\omega^2 e^{-2\phi}r^2 - ( \ell+2)(\ell-1)\right ] { a_1 } ' { \cal y } \right\ } , \\ { \cal d}_\ell(r ) & = & \frac{\ell(\ell+1)}{2\pi r^4}{a_1}\left\{2{a_1 } ' { \cal y } ' + \left[\left(\phi ' - \lambda ' \right){a_1 } ' + { a_1}''\right ] { \cal y } \right\},\end{aligned}\ ] ] in which the magnetic field now appears through the radial function @xmath89 of equations ( [ eq : h_r ] ) and ( [ eq : h_theta ] ) . multiplying equation ( [ master - equation1 ] ) with @xmath145 and integrating over the sphere , the eigenvalue equations takes the 1-dimensional form @xmath146 in deriving this equation we used the following properties of spherical harmonic functions : @xmath147 { \cal a}_{\ell } - ( \ell-2)(\ell+1)q_\ell q_{\ell-1}{\cal a}_{\ell-2 } , \\ { \cal l}^{\pm 2}_2 { \cal a}_\ell & : = & \sum_{\ell'}{\cal a}_{\ell ' } \int ( \partial_{\theta } y^*_{\ell 0})\sin\theta\cos\theta y_{\ell ' 0 } d\omega \nonumber \label{l2}\\ & = & \ell q_{\ell+1}q_{\ell+2}{\cal a}_{\ell+2 } + \left[\ell q^2_{\ell+1 } - ( \ell + 1)q^2_\ell \right ] { \cal a}_{\ell } - ( \ell+1)q_\ell q_{\ell-1}{\cal a}_{\ell-2 } , \\ q_\ell & : = & \sqrt{\frac{\ell^2}{(2\ell-1)(2\ell+1)}}.\end{aligned}\ ] ] from equations ( [ system ] ) , ( [ l1 ] ) and ( [ l2 ] ) it is evident that the presence of the dipole magnetic field causes the eigenfunction of a specific torsional @xmath148 mode to be no longer described by an angular part that corresponds to a single value of @xmath0 ( as in the nonmagnetized case ) . instead , the eigenfunction of a specific mode also acquires pieces with @xmath149 , @xmath150 and higher order dependence ( through successive couplings ) . in order to solve the full problem to a desired accuracy
, one has to solve a linear system of equations of the type ( [ system ] ) for required values of @xmath0 , up to some finite @xmath151 . since the coupling is purely a magnetic field effect , in the absence of a magnetic field the radial functions @xmath152 , @xmath153 , and @xmath154 vanish and the different @xmath0 dependencies decouple .
as mentioned in the introduction , we limit our current study to moderate magnetic field strengths of a few times @xmath1 g , since for larger values one would also have to take into account global magnetosonic oscillations .
we further simplify the numerical problem , by neglecting the couplings to @xmath155 terms in equation ( [ system ] ) .
this is motivated by the expectation that higher - order contributions will be less significant than the lowest - order piece of an eigenfunction , since the higher - order pieces are sampling the background magnetic field at smaller scales .
this expectation may be valid for the study of low - order modes .
however , it is difficult to predict the effect of @xmath155 couplings on high - order modes without solving the full two - dimensional problem .
our numerical results will thus be valid only to the extend that magnetic - field - induced couplings to @xmath155 terms can be neglected .
neglecting @xmath155 couplings leads to the eigenvalue equation @xmath156{\cal y } '' & + & \left\{\left(\frac{4}{r } + \phi ' - \lambda'\right)\mu + \mu ' + ( 1 + 2\lambda_1)\frac{a_1}{\pi r^4}\left[\left(\phi ' - \lambda'\right)a_1 + 2{a_1}'\right]\right\}{\cal y}'\nonumber \\ & + & \bigg\{\left[\left(\epsilon + p + ( 1 + 2\lambda_1)\frac{{a_1}^2}{\pi r^4 } \right)e^{2\lambda } - \frac{\lambda_1 { { a_1}'}^2}{2\pi r^2}\right]\omega^2 e^{-2\phi } \nonumber \\ & & \hspace{1 cm } - ( \lambda-2)\left(\frac { \mu e^{2\lambda}}{r^2 } - \frac{\lambda_1{{a_1}'}^2}{2\pi r^4}\right ) + ( 2 + 5\lambda_1)\frac{a_1}{2\pi r^4}\left\{\left(\phi ' - \lambda'\right){a_1 } ' + { a_1}''\right\ } \bigg\}{\cal y } = 0 , \label{system1}\end{aligned}\ ] ] where @xmath157 in order to solve a system of first - order odes , we define new variables @xmath158 and @xmath159 , through @xmath160e^{\phi-\lambda}{{\cal y}}'r^{2-\ell}. \label{y2}\end{aligned}\ ] ] in defining the new variables , we took into account the form of the eigenfunction @xmath161 near the center , where @xmath162 .
the final first - order system of equations to be solved numerically for the real eigenvalues @xmath131 is then @xmath163e^{\phi - \lambda}{\cal y}_1 - \frac{\ell+2}{r}{\cal y}_2 , \label{eq : master3b}\end{aligned}\ ] ] where equation ( [ maxwell - j1 ] ) was used in order to eliminate the term of @xmath164 .
if one would solve the above system of equations from the center of the star to the surface , then regularity at the center implies the boundary condition @xmath165e^{\phi}{{\cal y}_1}\ , .\end{aligned}\ ] ] if the equations ( [ eq : master3a])-([eq : master3b ] ) are used for the study of global oscillations , i.e. magnetosonic waves in the core and the crust and shear waves in the crust , an appropriate boundary condition at the base of the fluid / crust interface will be @xmath166 which leads to the condition @xmath167{\cal y}^{'(+ ) } .\end{aligned}\ ] ] this is analogous to the one used by @xcite in their study of a corresponding toy problem .
however , here , we only consider torsional modes confined to the crust and impose a zero traction condition at the base of the crust , which implies that @xmath168 there . at the stellar surface , the zero - torque - at - surface " condition
is imposed , i.e. @xmath169 @xcite , which again implies @xmath168 with our choice of variables .
we neglect the possible presence of a thin fluid ocean .
we have studied torsional modes for a variety of neutron star models using four different equations of state for the core , ranging from a very soft eos ( eos a ) @xcite to a very stiff ( eos l ) @xcite , with two intermediate ones , eos wff3 @xcite and apr @xcite . for each eos
we constructed a number of models , starting from a gravitational mass of @xmath25 and reaching close to the maximum mass limit in increments of @xmath170 . in order to separately investigate the effect of the composition of the crust
, we matched the various high - density eos to two different proposed equations of state for the crust , one recent derived by @xcite ( dh ) and , for reference , and older eos by @xcite ( nv ) .
figure [ fig : eos ] displays the mass - radius relationship of nonrotating equilibrium models constructed with the various high - density eos , in combination with our two choices for the crust eos . the individual models for which we compute torsional modes are shown with symbols , while their detailed properties are listed in table [ tab : eos_models ] . from figure
[ fig : eos ] it is evident that for the soft equations of state ( a and wff3 ) the choice of the crust eos does not affect significantly the bulk properties of the star , while for the stiffest eos l models with the same mass have considerably different radii .
the reason for this is that the models constructed with the stiff eos have central densities that are much closer to the density of the base of the crust than for soft eos .
the two crust eos differ significantly both in the detailed composition , as well as in the density at the base of the crust , which is at @xmath171gr/@xmath172 for @xcite and at @xmath173gr/@xmath172 for @xcite . for the stiff eos these different properties of the crust
have a considerable effect on the bulk properties .
for the two crust eos the shear modulus @xmath10 can be computed using equation ( [ eq : mu ] ) . for the nv
eos a simple fitting formula was already derived by @xcite @xmath174 which we also adopt here . for the dh eos
, we derive the following fitting ( see also @xcite ) @xmath175 where @xmath176 is the fraction of nucleons in the neutron gas outside nuclei . [ cols="<,>,>,>,^,<,>,>,>",options="header " , ] and the first overtone for @xmath177 and 4 torsional modes as functions of the normalized magnetic field ( @xmath178 ) .
the neutron star mass is 1.4@xmath179 and we show results for eos a+dh . , width=377 ] and @xmath17 torsional mode as functions of the normalized magnetic field ( @xmath178 ) .
the neutron star masses are 1.4@xmath179 and we show only results for four eos i.e. a+dh , a+nv , l+dh and l+nv .
the lines correspond to fits according to the empirical formula , eq .
( [ eq : fit_l2 ] ) , with coefficient values from table [ tab : fit_factors ] . as seen here , our empirical formula , eq .
( 79 ) , agrees very well with the numerical results .
we stress that these results were obtained in the approximation of neglecting magnetic - field - induced deformations of the background star and couplings to @xmath155 terms.,width=377 ]
we have derived the formalism for computing torsional oscillations of relativistic stars endowed with a strong dipole magnetic field , confined to the crust .
our equations are valid in the cowling approximation ( no spacetime perturbations ) and we have neglected the effect of the magnetic field on the equilibrium configuration and on the coupling of torsional modes to @xmath155 terms and to global magnetosonic modes . under these approximations , our formalism allows us to obtain an estimate of the magnetic field effects on torsional modes , up to moderate values of the magnetic field strength ( up to a few times @xmath180 g ) .
these moderate values of the magnetic field strength are appropriate for models of magnetars , for which there is strong evidence that they are at the heart of the sgr phenomenon .
we have done a systematic search of parameter space by computing torsional mode frequencies for various values of the harmonic index @xmath0 and for various overtones , using an extended sample of models of compact stars .
these models vary in mass , high - density equation of state and crust model , uniformly covering the allowed mass vs. radius parameter space .
our numerical results have shown that torsional mode frequencies are sensitive to the crust model if the high - density equation of state is very stiff ( such as eos l ) .
in addition , torsional mode frequencies are drastically affected by a dipole magnetic field , if the latter has a strength exceeding roughly @xmath1 g .
the effect of the magnetic field is surprisingly sensitive to the adopted crust model .
using our extended numerical results we have derived empirical relations for the effect of the magnetic field on torsional modes as well as for the crust thickness .
we compare our numerical results to observed frequencies in sgrs and find that certain high - density eos and mass values are favored over others in the non - magnetized limit . on the other hand ,
if the magnetic field is strong , then its effect has to be taken into account in attempts to formulate a theory of asteroseismology for magnetars .
this topic , as well as the inclusion of global magnetosonic modes will be discussed in a separate publication @xcite .
we are grateful to nils andersson , demetrios papadopoulos , adamantios stavridis and miltos vavoulidis for useful discussions .
this work was supported by the marie - curie grant mif1-ct-2005 - 021979 and the pythagoras ii program of gsrt . 99 akmal a. , pandharipande v.r . , ravenhall d.g . , 1998 ,
c 58 , 1804 barat et al . , 1983 , a&a , 126 , 400 bocquet m. , bonazzola s. , gourgoulhon e. , novak j. , 1995 , a&a , 301 , 757 bonazzola s. , gourgoulhon e. , salgado m. , marck j.a . , 1993 , a&a , 278 , 421 braithwaite j. , spruit h.c . , a&a , 450 , 1097 carroll b.w . , zweibel e.g. , hansen c .. , mcdermot p.n .
, savedoff m.p .
, thomas j.h . , van horn h.m . , 1986 ,
apj , 305 , 767 carter b. , quintana h. , 1973 , proc .
r. soc . lond . , a331 , 57 douchin f. , haensel p. , 2001
, a & a , 380 , 151 duncan r.c.,1998 , apj , 1998 , 498 , l45 duncan r.c . , thompson c. , 1992 , apj , 392 , l9 glampedakis k. , samuelson l. , andersson n. , 2006 , astro - ph/0605461 hansen c. , cioffi d.f , 1980 , apj , 238 , 740 hurley k. et al .
, 1999 , nature , 397 , 41 israel g. et al .
, 2005 , apj , 628 , l53 konno k. , obata t. , kojima y. , 1999 , a&a , 352 , 211 lee , u. , 2006 , astro - ph/0610182 leins m. , 1994 , phd thesis , university of tbingen mazets e.p .
et al . , 1979 ,
nature 282 , 587 mcdermott p.n .
, van horn h.m .
, hansen c.j . , 1988 ,
apj , 325 , 725 messios n. , papadopoulos d.b . , stergioulas n. , 2001 , mnras , 328 , 1161 .
negele j.w .
, vautherin d. , 1973 , nucl .
phys . , a207 , 298 palmer d.m .
et al . , 2005 ,
nature , 434 , 1107 pandharipande v.r .
, 1971 , nucl .
a , 178 , 123 pandharipande v.r . , smith r.a . , 1975 ,
59b , 15 piro a.l . , 2005 , apj , 634 , l153 samuelsson l. , andersson n. , astro - ph/0609265 schumaker b.l .
, thorne k.s .
, 1983 , mnras , 203 , 457 strohmayer t.e . ,
1991 , apj , 372 , 591 sotani h. , kokkotas , k.d . , stergioulas , n. , vavoulidis , m. , preprint ( astro - ph/0611666 ) strohmayer t.e . , etal . , 1991 , apj , 375 , 679 strohmayer t.e . ,
watts a.l . , 2005 , apj , 632 , l111 strohmayer t.e . ,
watts a.l . , astro - ph/0608463 ( accepted in apj ) terasawa t. , et al . , 2005 ,
nature 434 , 1110 unno w. , osaki y. , ando h. , saio h. , shibahashi h. , 1989 , nonradial oscillations of stars , university of tokyo press wasserman i. , shapiro s.l .
, 1983 , apj , 265 , 1036 watts a.l . , strohmayer t.e . , 2006 , apj , 637 , l117 wiringa r.b . , fiks v. , farbrocini a. , 1988 , phys . rev
. c 38 , 1010
if the magnetic field is assumed to be axisymmetric , maxwell s equations ( [ maxwell0 - 1 ] ) become @xmath181 as a toy model , we consider a monopolar magnetic field in the crust @xmath182 for this case , the solution of maxwell s equation ( [ maxwell - b ] ) we is @xcite @xmath183 where @xmath184 is some constant . for this case the perturbation equation ( [ master - equation ] ) is rewritten as @xmath185{\cal y } ' - \frac{(\ell+2)(\ell-1)}{r^2}\mu e^{2\lambda } { \cal y } \label{perturbation - equation - s}\end{aligned}\ ] ] if we introduce the new variables , such as @xmath186 then equations ( [ perturbation - equation - s ] ) reduce @xmath187e^{-\phi+\lambda } { \cal y}_1 - \frac{4}{r } { \cal y}_2\end{aligned}\ ] ] and @xmath188 torsional modes as functions of the normalized magnetic field ( @xmath178 ) for eos a+dh and a+nv are shown .
the magnetic field has monopole geometry while the neutron stars have masses 1.4@xmath179 .
, width=377 ] in figure [ fig : radial ] we show the fundamental and first overtone frequencies for neutron star models with mass 1.4@xmath179 .
it is apparent that in the case of a monopole magnetic field the frequencies ( especially of the fundamental mode ) will be marginally affected by the magnetic field and only for stengths on the order of 10@xmath189 g its influence becomes significant .
this is in qualitative agreement with the conclusions in @xcite .
in addition , we confirm the finding in @xcite that the frequency of the fundamental mode decreases for a monopole magnetic field , as is apparent in figure [ fig : radial ] . | we present the formalism and numerical results for torsional oscillations of relativistic stars endowed with a strong dipole magnetic field , assumed to be confined to the crust . in our approach ,
we focus on axisymmetric modes and neglect higher - order couplings induced by the magnetic field .
we do a systematic search of parameter space by computing torsional mode frequencies for various values of the harmonic index @xmath0 and for various overtones , using an extended sample of models of compact stars , varying in mass , high - density equation of state and crust model .
we show that torsional mode frequencies are sensitive to the crust model if the high - density equation of state is very stiff .
in addition , torsional mode frequencies are drastically affected by a dipole magnetic field , if the latter has a strength exceeding roughly @xmath1 g and we find that the magnetic field effects are sensitive to the adopted crust model . using our extended numerical results we derive empirical relations for the effect of the magnetic field on torsional modes as well as for the crust thickness .
we compare our numerical results to observed frequencies in sgrs and find that certain high - density eos and mass values are favored over others in the non - magnetized limit . on the other hand ,
if the magnetic field is strong , then its effect has to be taken into account in attempts to formulate a theory of asteroseismology for magnetars .
[ firstpage ] relativity mhd stars : neutron stars : oscillations stars : magnetic fields gamma rays : theory |
the microscopic cluster model is one of the successful models to study the structure and reactions of light nuclei @xcite . in the conventional cluster model
, one assumes that the nucleus is composed of several simple clusters with @xmath11 which are described by ( 0@xmath5 ) harmonic - oscillator shell model functions , and use an effective @xmath12-@xmath12 interaction which is appropriate for such a model space .
however , it is well known that the ground states of the typical clusters @xmath2 , @xmath3 , @xmath8 and @xmath10he have non - negligible admixtures of @xmath13-wave component due to the tensor interaction . since the conventional cluster model does not directly treat the @xmath13-wave component , the strong attraction of the nucleon - nucleon interaction due to the tensor term
is assumed to be renormalized into the central term of the effective interaction .
recently , @xmath0-@xmath1 structure calculations @xcite have been successfully developed : stochastic variational method ( svm ) @xcite , global vector representation method ( gvr ) @xcite , green s function monte carlo method @xcite , no core shell model @xcite , correlated hyperspherical harmonics method @xcite , unitary correlation operator method @xcite , and so on .
although the application of @xmath0-@xmath1 reaction calculations with a realistic interaction are restricted so much in comparison with structure calculations , it has been intensively applied to the four - nucleon systems @xmath3+@xmath9 and @xmath8+@xmath4 @xcite .
especially @xmath2+@xmath2 scattering states , which couple to @xmath3+@xmath4 and @xmath8+@xmath9 channels , have attracted much attention , because the @xmath2+@xmath2 radiative capture is one of the mechanisms making @xmath10he through electro - magnetic transitions @xcite and also have posed intriguing puzzles for analyzing powers @xcite , which are motivated by the famous @xmath14 problem in the three - nucleon system .
furthermore , the @xmath2+@xmath2 elastic - scattering phase shifts are interesting because the astrophysical s - factor of the @xmath2(@xmath2,@xmath15)@xmath10he reaction is not explained by any calculation using an effective interaction that contains no tensor term , and is expected to be contributed by the @xmath13-wave components of the clusters through @xmath16 transitions @xcite .
also , thanks to recent developments of the microscopic cluster model , the simple model using the ( 0@xmath5 ) harmonic - oscillator wave function with an effective interaction is not mandatory any more , at least , in light nuclei .
we can use a kind of @xmath0-@xmath1 cluster model which employs more realistic cluster wave functions with realistic interactions .
therefore , it is interesting to see the difference between the @xmath0-@xmath1 reaction calculations with a tensor term and the conventional cluster model calculations without a tensor term in few - body systems .
the microscopic @xmath17-matrix method ( mrm ) with a cluster model ( gcm or rgm ) has been applied to studies of many nuclei @xcite .
it is now used in @xmath0 @xmath1 descriptions of collisions @xcite .
we have also applied the mrm to the @xmath8+@xmath4 scattering problem with more realistic cluster wave functions by using a realistic interaction @xcite .
the gaussian basis functions for the expansion of the cluster wave functions are chosen by a technique of the svm @xcite . in the mrm , as will be shown later , the relative wave function between clusters ( @xmath18 and @xmath19 ) is connected to the boundary condition at a channel radius .
the problem is how to calculate the matrix elements . in this paper , we develop a method called the triple global vector representation method ( tgvr ) , by which we calculate the matrix elements in a unified way .
although we restricted ourselves to four nucleon systems in the present paper , the formulation of the tgvr itself can be applied to more than four - body systems as in the previous studies of the global vector representation methods ( gvr ) @xcite . furthermore , for scattering problems , the tgvr can deal with more complicated systems than the double ( or single ) global vector which was given in the previous papers @xcite , because we need three representative orbital angular momenta , the total internal orbital momenta of both clusters and the orbital momentum of their relative motion , in order to reasonably describe the scattering states . in other words
, the first global vector represents the angular momentum of cluster @xmath18 , the second global vector represents the angular momentum of cluster @xmath19 , and the third global vector represents the relative angular momentum between the clusters . in this paper , we will investigate the effect of the distortion of clusters on the @xmath2+@xmath2 elastic - scattering by comparing the phase shifts calculated with a realistic and an effective interaction . in section 2
, we explain the mrm in brief . in section 3 , the correlated gaussian ( cg )
method with the tgvr , which has newly been developed for the present analysis , will be presented . in section 4
, we will explain how to calculate the matrix elements with tgvr basis functions .
the typical matrix elements are also given in the appendix . in section 5 , we will present and discuss the calculated scattering phase shifts in detail .
finally , summary and conclusions are given in section 6 .
in the present study we calculate @xmath2+@xmath2 and @xmath3+@xmath4 ( and @xmath8+@xmath9 ) elastic scattering phase shifts with the microscopic @xmath17-matrix method . though the method is well documented in e.g. refs .
@xcite , we briefly explain it below in order to present definitions and equations needed in the subsequent sections . since our interest is on low - energy scattering , we consider only two - body channels .
a channel @xmath20 is specified by the two nuclei ( clusters ) @xmath21 , their angular momenta , @xmath22 , the channel spin @xmath23 that is a resultant of the coupling of @xmath24 and @xmath25 , and the orbital angular momentum @xmath26 for the relative motion of @xmath18 and @xmath19 .
the wave function of channel @xmath20 with the total angular momentum @xmath27 , its projection @xmath28 , and the parity @xmath29 takes the form @xmath30_i
\chi_{\alpha } ( \mbox{\boldmath$\rho$}_{\alpha})\right]_{jm } , \label{wf - ch1}\end{aligned}\ ] ] where @xmath31 and @xmath32 are respectively antisymmetrized intrinsic wave functions of @xmath18 and @xmath19 , and @xmath33 is an operator that antisymmetrizes between the clusters .
the square bracket @xmath34_i$ ] denotes the angular momentum coupling .
the coordinate @xmath35 in the relative motion function @xmath36 is the relative distance vector of the clusters .
the channel spin @xmath23 and the relative angular momentum @xmath26 in @xmath20 are coupled to give the total angular momentum @xmath27 .
the relative - motion functions @xmath37 also depend on @xmath27 and @xmath29 . for simplicity
, this dependence is not displayed explicitly in the notation for @xmath37 as well as for some other quantities below .
the configuration space is divided into two regions , internal and external , by the channel radius @xmath18 . in the internal region ( @xmath38 )
, the total wave function may be expressed in terms of a combination of various @xmath39s @xmath40 with @xmath41_i y_{\ell}(\widehat{{\mbox{\boldmath $ \rho$}}}_\alpha)\right]_{jm } \right .
\nonumber \\ \left .
+ ( -1)^{a_a+i_a+i_b - i+\ell } \left[\left[\phi^{b}_{i_b}\phi^{a}_{i_a}\right]_i y_{\ell}(\widehat{{\mbox{\boldmath $ \rho$}}}_\alpha)\right]_{jm } \delta_{ab } \right\ } , \label{wf.channel}\end{aligned}\ ] ] where @xmath42 is the number of nucleons in cluster @xmath18 , @xmath43 is unity if @xmath18 and @xmath19 are identical clusters and zero otherwise , and @xmath44 is unity if the clusters are in identical states and zero otherwise . in the second line of eq .
( [ wf.int ] ) , the relative motion functions of eq .
( [ wf - ch1 ] ) are expanded in terms of some basis functions as @xmath45 in what follows we take @xmath46 with a suitable set of @xmath47s . in the external region ( @xmath48 ) , the total wave function takes the form @xmath49 note that the antisymmetrization between the clusters is dropped in the external region under the condition that the channel radius @xmath18 is large enough . the function @xmath50 of eq .
( [ wf.ext ] ) is a solution of the equation @xmath51 g_{\alpha}(\rho_\alpha ) = e_{\alpha } g_{\alpha}(\rho_\alpha),\end{aligned}\ ] ] where @xmath52 is the reduced mass for the relative motion in channel @xmath20 , @xmath53 and @xmath54 are the charges of @xmath18 and @xmath19 , and @xmath55 is the energy for the relative motion , where @xmath56 is the total energy , and @xmath57 and @xmath58 are the internal energies for the clusters @xmath18 and @xmath19 , respectively . for the scattering initiated through the channel @xmath59 , the asymptotic form of @xmath60 for the open channel
@xmath20 @xmath61 is @xmath62,\end{aligned}\ ] ] where @xmath63 , @xmath64 and @xmath65 is an element of the @xmath66-matrix ( or collision matrix ) to be determined . here
@xmath67 and @xmath68 are the incoming and outgoing waves defined by @xmath69 with the regular and irregular coulomb functions @xmath70 and @xmath71 .
the sommerfeld parameter @xmath72 is @xmath73 . for a closed channel @xmath20 @xmath74 , the asymptotic form of @xmath75
is given by the whittaker function @xmath76 the matrix elements @xmath77 are determined by solving a schrdinger equation with a microscopic hamiltonian @xmath78 involving the @xmath79 nucleons , @xmath80 with the bloch operator @xmath81 @xmath82 where the channel radius @xmath18 is chosen to be the same for all channels , and the @xmath83 are arbitrary constants .
here , we choose @xmath84 for the open channels and @xmath85 for the closed channels .
the results do not depend on the choices for @xmath83 but these values simplify the calculations .
notice that the projector on @xmath86 in eq .
( [ bloch ] ) is not essential in a microscopic calculation and can be dropped since the various channels are orthogonal at the channel radius .
the bloch operator ensures that the logarithmic derivative of the wave function is continuous at the channel radius .
in addition , @xmath87 must be equal to @xmath88 at @xmath89 .
projecting the schrdinger equation on a basis state , one obtains @xmath90 with @xmath91 and @xmath92 here @xmath93 indicates that the integration with respect to @xmath94 is to be carried out in the internal region . actually @xmath93 is obtained by calculating the matrix element @xmath95 in the entire space and subtracting the corresponding external matrix element @xmath96 that is easily obtained because no intercluster antisymmetrization is needed .
the @xmath17-matrix and @xmath97-matrix are defined by @xmath98 the @xmath66-matrix is finally obtained as @xmath99 in this paper we focus on the elastic phase shifts @xmath100 that are defined by the diagonal elements of the @xmath66-matrix , @xmath101 we study four - nucleon scattering involving the @xmath2+@xmath2 , @xmath3+@xmath4 and @xmath8+@xmath9 channels in the energy region around and below the @xmath2+@xmath2 threshold . in table
[ chan0 ] we list all possible labels @xmath102 of physical channels for @xmath103 , @xmath104 , and @xmath105 , assuming @xmath106 . here
`` physical '' means that the channels involve the cluster bound states that appear in the external region as well .
non - physical channels involving excited pseudo states will also be included in most calculations .
note that the @xmath2+@xmath2 channel must satisfy the condition of @xmath107 even ( see eq .
( [ wf.channel ] ) ) .
the channel spin @xmath108 or 2 can couple with only even @xmath26 , but @xmath109 with only odd @xmath26 .
it is noted that the relative motion for the @xmath2+@xmath2 scattering can have @xmath110 only when @xmath111 is equal to @xmath112 and @xmath113 .
.channel spins ( @xmath102 ) of physical @xmath2+@xmath2 , @xmath3+@xmath4 , and @xmath8+@xmath9 channels for @xmath114 and @xmath115 . [ cols="^,^,^,^,^,^,^",options="header " , ] [ chan1 ] though it is expected that the pseudo channels with low threshold energies contribute more strongly to the scattering phase shift , we take into account all of these @xmath116+@xmath116 channels that include a vanishing total isospin as given in fig .
[ fig : threshold ] .
the total isospin of the @xmath117+@xmath12 channel is mixed in the present calculation . because the @xmath118=1 component of the scattering wave function only weakly couples to the @xmath119+@xmath119 elastic - channel , the channel @xmath119+@xmath120 is not employed in the calculation .
we also include the excited deuteron channels that comprise the @xmath121 and @xmath122 clusters .
the energies of these lowest thresholds are above 10 mev .
these channels are therefore expected not to be very important , but that is not always the case as will be discussed in the case of the @xmath123s@xmath124 @xmath2+@xmath2 phase shift .
table [ chan1 ] summarizes all the channels that are used in our calculation .
the 2@xmath12 + 2@xmath12 channels are distinguished by roman numerals , while the 3@xmath12+@xmath12 channels are labeled by arabic numerals . in the following ,
we use an abbreviation `` 2@xmath12 + 2@xmath12 '' or `` 3@xmath12+@xmath12 '' to indicate calculations including all 2@xmath12 + 2@xmath12 channels i - v or all 3@xmath12+@xmath12 channels ( 1 - 2 in table [ chan1 ] ) , respectively . here
@xmath125 and @xmath126 are excited 3@xmath12 continuum states .
a `` full '' calculation indicates that all the channels in the table are included to set up the @xmath66-matrix . in the case of the mn potential channels
iii and iv are not included because this potential contains no tensor force .
the relative wave functions @xmath127 are expanded with 15 basis functions .
we checked the stability of the @xmath66-matrix against the choice of the channel radius .
the channel radius employed in this calculation is about 15 fm .
[ fig : d+d - g3 ] displays the @xmath7 @xmath2+@xmath2 elastic - scattering phase shift obtained with the av8@xmath6 potential .
the dash - dotted line is the phase shift calculated with channel i ( @xmath128 ) , and the dash - dot - dotted line is the phase shift with channels i and ii ( @xmath129 ) .
the phase shifts calculated by including further excited deuterons are also plotted by the dashed and dotted lines that correspond to the channels i - iii ( @xmath130 ) and i - iv ( @xmath131 ) , respectively . a naive expectation that the @xmath7 @xmath2+@xmath2 elastic - scattering phase shift might be well described in channel i ( @xmath132 , @xmath133 and @xmath134 ) alone completely breaks down in the case of the av8@xmath6 potential .
@xmath2+@xmath2 elastic - scattering phase shift calculated with the av8@xmath6 potential .
the phase shifts are all obtained within the @xmath2+@xmath2 channels .
the set of included channels is successively increased from i to iv .
see table [ chan1 ] for the deuteron states included in each channel .
, width=7,height=6 ] because the deuteron has a virtual state @xmath135 with @xmath112 at low excitation energy , it is reasonable that the inclusion of channel ii gives rise to a considerable attractive effect of several tens of degrees on the phase shift , as shown by the dash - dot - dotted line of fig .
[ fig : d+d - g3 ] .
however , the phase shift exhibits no converging behavior even when the higher spin states such as @xmath136(2@xmath137 ) and @xmath136(3@xmath137 ) are taken into account in the calculation .
the additional attractions by these channels are of the same order as that of channel ii .
one may conclude that the deuteron is strongly distorted even in the low energy @xmath7 @xmath2+@xmath2 elastic scattering but more physically we have to realize that there exist two observed @xmath112 states below the @xmath2+@xmath2 threshold .
obviously the @xmath2+@xmath2 scattering wave function is subject to the structure of those states in the internal region .
the second 0@xmath137 state of @xmath10he lying about 4 mev below the @xmath2+@xmath2 threshold is known to have a @xmath117+@xmath12 cluster structure @xcite .
thus this state together with the ground state of @xmath10he can not be described well in the @xmath116+@xmath116 model space alone .
as seen in table [ chan0 ] , the @xmath117+@xmath12 channel contains a @xmath7 component , which is the dominant component of the @xmath138 state .
since the realistic force strongly couples the @xmath116+@xmath116 channel to the @xmath117+@xmath12 channel and the @xmath2+@xmath2 scattering wave function has to be orthogonal to the main component of the underlying @xmath112 states , we expect that the deuteron in the incoming @xmath2+@xmath2 channel never remains in its ground state but has to be distorted largely due to the @xmath117+@xmath12 channel .
the phase shift for the channel i - iv ( dotted line ) shows a resonant pattern .
this resonant state is expected to be the second 0@xmath137 state because of the restricted model space within the @xmath2+@xmath2 channel .
state energies between calculations with the av8@xmath6 ( left ) and mn ( middle ) potentials and experiment ( right ) .
the model space for av8@xmath6 is i - iv , @xmath117+@xmath12 and full and the model space for mn is i - ii , @xmath117+@xmath12 and full . ,
width=10,height=8 ] fig .
[ fig:2nd0 + ] displays the calculated ground state energy and the second 0@xmath137 energy for the av8@xmath6 ( left ) and mn ( middle ) potentials .
the model spaces of the calculations are i - iv , @xmath117+@xmath12 and full for av8@xmath6 and i - ii , @xmath117+@xmath12 and full for mn .
we also plot experimental energies ( right ) @xcite .
for the av8@xmath6 potential , the energies of the two lowest @xmath112 states do not change very much between the full and @xmath117+@xmath12 models .
but the second 0@xmath137 state with the @xmath2+@xmath2 model ( channels i - iv ) is not bound with respect to the @xmath2+@xmath2 threshold as expected before . on the contrary , for the mn potential , the second 0@xmath137 state with the @xmath2+@xmath2 model ( channels i - ii ) is bound with respect to the @xmath2+@xmath2 threshold .
we consider that this difference makes the drastic change of the @xmath2+@xmath2 phase shifts , between the av8@xmath6 and mn potentials .
it is also interesting to see that the energies of the two lowest @xmath112 states for the mn potential are almost the same between the full and @xmath117+@xmath12 models .
plotted in fig .
[ fig : d+d-3n+n ] are the @xmath7 @xmath2+@xmath2 elastic - scattering phase shifts obtained with the av8@xmath6 potential ( left ) and the mn potential ( right ) .
the full calculation ( solid line ) couples all 2@xmath12 + 2@xmath12 and 3@xmath12+@xmath12 channels that are listed in table [ chan1 ] .
the @xmath17-matrix analysis ( crosses ) @xcite is reproduced well by both the av8@xmath6 and mn potential with the full calculation .
compared to the uncoupled phase shift ( dotted line ) , one clearly sees that the 3@xmath12+@xmath12 channel produces a very large effect on the @xmath2+@xmath2 elastic phase shift , especially in the case of the av8@xmath6 potential .
we also verified that a calculation excluding the channels iii , iv or v from the full channel calculation gives only negligible change in the phase shift . the slow convergence seen in fig .
[ fig : d+d - g3 ] is thus attributed to the neglect of the @xmath117+@xmath12 channel , indicating that a proper account of the @xmath7 @xmath2+@xmath2 elastic phase shift at low energy can be possible only when the coupled channels \{@xmath2(1@xmath137)+@xmath2(1@xmath137 ) } + \{@xmath2(0@xmath137)+@xmath2(0@xmath137)}+ \{@xmath3(1/2@xmath137)+@xmath139 } + \{@xmath8(1/2@xmath137)+@xmath140 } are considered . thus , the slow convergence in fig .
[ fig : d+d - g3 ] suggests that the @xmath116+@xmath116 partition is not an economical way to include the effects of the @xmath117+@xmath12 channel . in the case of the mn potential ( right panel in fig .
[ fig : d+d-3n+n ] ) , the situation is very different from the av8@xmath6 case .
the channel coupling effect is rather modest , and the size of the @xmath7 @xmath2+@xmath2 elastic phase shift is already accounted for mostly in the @xmath2+@xmath2 channel calculation .
all these results are very consistent with the @xmath112 spectrum in fig .
[ fig:2nd0 + ] .
the large distortion effect of the deuteron clusters on the @xmath7 @xmath2+@xmath2 scattering phase shift is expected to appear in the 3@xmath12+@xmath12 phase shift as well because of the coupling between the 3@xmath12+@xmath12 and 2@xmath12 + 2@xmath12 channels .
we display in fig .
[ fig:3n+n - g3mn ] the @xmath7 @xmath3+@xmath4 elastic - scattering phase shift at energies below the @xmath2+@xmath2 threshold .
the 0@xmath141 state of @xmath10he is observed as a sharp resonance with a proton decay width of 0.5 mev at about 0.4 mev above the @xmath3+@xmath4 threshold .
the present energies ( @xmath142 mev for av8@xmath6 , @xmath143 mev for mn ) calculated with a bound state approximation are slightly smaller than the experimental value , but they are consistent with a calculation ( @xmath144 mev and @xmath145 mev for av18+uix , @xmath146 mev and @xmath147 mev for av18+uix+v@xmath148 ) with another realistic interaction ( av18 ) with three nucleon forces by hofmann and hale @xcite .
the calculated phase shifts appear slightly larger than that in the @xmath17-matrix analysis ( crosses in fig .
[ fig:3n+n - g3mn ] ) @xcite .
it is noted that the phase shift changes so much even for a small change of the 0@xmath141 resonant pole position ( @xmath1490.1 mev ) because it is very near to the threshold .
the phase shifts in the full calculation , for both av8@xmath6 and mn cases , show a resonance pattern in a small energy interval and the overall energy dependencies of the phase shifts are similar to each other
. however , the phase shifts obtained only in the @xmath117+@xmath12 channel are quite different as indicated by the dotted lines in fig .
[ fig:3n+n - g3mn ] . in the case of the mn potential
( right ) the phase shift is already close to the full phase shift , while in the case of the av8@xmath6 potential ( left ) the phase shift is much smaller ( by almost 90 degrees ) and moreover shows no resonance pattern . by looking into the wave functions in more detail ,
we argue that the large distortion effect in the @xmath7 @xmath2+@xmath2 and @xmath117+@xmath12 coupled channels is really brought about by the tensor force .
as shown in table [ sub1 ] , the av8@xmath6 potential with tnf gives 5.8% and 8.4% ( 8.3% ) @xmath13-state probability for @xmath2 and @xmath3 @xmath150 , respectively .
thus the @xmath2+@xmath2 state in the @xmath7 state contains @xmath151 components ( 89% ) as well as @xmath152 components ( 11% ) , where @xmath153 and @xmath66 are the total orbital and spin angular momenta of the four - nucleon system .
similarly the @xmath117+@xmath12 state in the @xmath7 state contains an @xmath151 component ( 92% ) and an @xmath152 component ( 8% ) .
thus the tensor force couples both states with @xmath154 and @xmath155 couplings , which are in fact very large compared to the central matrix element ( @xmath156 , @xmath157 ) .
an analysis of this type was performed for some levels of @xmath10he in refs .
the mn potential contains no tensor force , so that the @xmath2+@xmath2 and @xmath117+@xmath12 channel coupling is modest . as listed in table
[ chan0 ] , there are four channels , @xmath158 , @xmath159 , @xmath160 and @xmath161 , for @xmath162 at energies around the @xmath2+@xmath2 threshold . among these states , we expect that the effect of the coupling between the 3@xmath12+@xmath12 and 2@xmath12 + 2@xmath12 channels occurs most strongly in @xmath159 as it appears in all physical channels
. however , no sharp resonance is observed in @xmath10he up to 28mev of excitation energy , so that the coupling effect , if any , might be weaker than that observed in the @xmath7 case . fig .
[ fig:2 + ] displays the @xmath159 elastic - scattering phase shifts obtained in three types of calculations , @xmath117+@xmath12 ( dashed line ) , @xmath116+@xmath116 ( dotted line ) , and full ( solid line ) . the @xmath3+@xmath4 and @xmath2+@xmath2 phase shifts start from the @xmath3+@xmath4 ( @xmath163 ) and @xmath2+@xmath2 thresholds , respectively .
the phase shifts of the @xmath117+@xmath12 and @xmath116+@xmath116 calculations are both slightly positive , indicating a weak attraction in the @xmath3+@xmath4 and @xmath2+@xmath2 interactions . in the full calculation ,
the @xmath3+@xmath4 phase shift becomes more attractive and the @xmath2+@xmath2 phase shift turns to be negative ( repulsive ) .
the present full calculation reproduces the calculation of ref . @xcite as expected .
though the effect of the coupling is slightly larger in the av8@xmath6 potential than in the mn potential , it is much less compared to the case of the @xmath7 phase shift .
this is understood as follows . in the @xmath159 state ,
the main component of the wave function is given by the @xmath164 , @xmath165 state : its probability is the same as that of @xmath7 , that is , 92% in @xmath3+@xmath4 and 89% in @xmath2+@xmath2 .
however , the probability of finding the state with @xmath166 , @xmath167 , which causes a strong tensor coupling , is more than one order of magnitude smaller than in the case of @xmath7 , namely 0.23% in @xmath3+@xmath4 and 0.44% in @xmath2+@xmath2 , respectively .
the reason for this small percentage is that , to obtain @xmath166 , the incoming @xmath13-wave in the @xmath159 channel must couple with the @xmath13-components in the clusters , but this coupling leads to several fragmented components with different @xmath153 values .
this relatively weaker coupling of the tensor force explains the phase shift behavior in fig .
[ fig:2 + ] . in fig .
[ fig:2+full ] we plot the @xmath3+@xmath4 and @xmath2+@xmath2 elastic - scattering phase shifts for other channels , @xmath158 ( solid line ) , @xmath160 ( dashed line ) , and @xmath161 ( dotted line ) .
we show only the full result , because the phase shifts with the truncated basis do not change visibly at the scale of the figure .
the obtained phase shifts are not that different between the av8@xmath6 and mn potentials , and also consistent with the previous calculation @xcite .
thus , the effect of the distortion of the clusters is very small for 2@xmath137 except for @xmath159 .
we have three channels for @xmath168 , @xmath169 , @xmath170 and @xmath171 .
no sharp @xmath172 resonance of @xmath10he is observed experimentally up to 28 mev of excitation energy .
another theoretical calculation neither predicts it @xcite , so that the coupling between the 2@xmath12 + 2@xmath12 and 3@xmath12+@xmath12 channels is expected to be weak .
[ fig:1 + ] exhibits the @xmath3+@xmath4 and @xmath2+@xmath2 elastic - scattering phase shifts in the full calculation : @xmath169 @xmath2+@xmath2 ( solid line ) , @xmath170 @xmath3+@xmath4 ( dashed line ) , and @xmath171 @xmath3+@xmath4 ( dotted line ) .
only the full result is displayed because the phase shift change in other calculations is small .
both av8@xmath6 and mn potentials produce phase shifts quite similar to each other .
as seen from table [ chan0 ] , the main components of these negative parity states are considered to be @xmath173 .
we compare in fig .
[ fig:0- ] the @xmath174 elastic - scattering phase shifts calculated with the av8@xmath6 ( left ) and mn ( right ) potentials . the truncated @xmath117+@xmath12 ( dashed line ) and @xmath116+@xmath116 ( dotted line ) calculations are shown together with the full result ( solid line ) .
the @xmath3+@xmath4 phase shift of the @xmath117+@xmath12 calculation is similar with both av8@xmath6 and mn potentials , while the @xmath2+@xmath2 phase shift of the @xmath116+@xmath116 calculation behaves quite differently between the two potentials : the @xmath2+@xmath2 phase shift is weakly attractive with av8@xmath6 but is very strongly attractive with mn .
no typical resonance behavior shows up below the @xmath2+@xmath2 threshold , which is in contradiction to experiment . in the full model that combines both 3@xmath12+@xmath12 and 2@xmath12 + 2@xmath12 configurations ,
however , the two potentials predict quite different phase shifts especially in the @xmath3+@xmath4 channel .
the @xmath3+@xmath4 phase shift with av8@xmath6 becomes so attractive that it crosses @xmath175 , indicating a resonance at about 1 mev above the @xmath3+@xmath4 threshold .
the @xmath2+@xmath2 phase shift changes sign from attractive to repulsive .
the result based on the av8@xmath6 potential is thus consistent with experiment .
furthermore , we reproduce the flat structure of the @xmath174 phase shift around several mev above the @xmath3+@xmath4 threshold which was discussed as the coupling to the @xmath8+@xmath9 channel @xcite . on the other hand , the mn potential changes the @xmath3+@xmath4 phase shift only mildly and produces no sharp resonance behavior .
the @xmath2+@xmath2 phase shift changes drastically to the repulsive side .
as seen in the above figure , the sharp 0@xmath176 state appears provided a full model space with a realistic potential is employed .
the mechanism to produce this resonance is unambiguously attributed to the tensor force as discussed in ref .
@xcite for the realistic interaction g3rs @xcite .
according to it , the @xmath177 state consists of only two components , @xmath178 ( 95.5% ) and @xmath152 ( 4.5% ) , ignoring a tiny component with @xmath151 .
the @xmath152 component arises from the coupling of the incoming @xmath179-wave with the @xmath13-states contained in the @xmath117 and @xmath2 clusters .
all the pieces of the hamiltonian but the tensor force have no coupling matrix element between the two components .
the uncoupled hamiltonian thus gives a too high energy to accommodate a resonance .
the tensor force , however , couples the two components very strongly , bringing down its energy to a right position . the second lowest negative parity state has spin - parity 2@xmath176 .
the physical channel for this state is only @xmath180 as seen in table [ chan0 ] .
[ fig:2- ] compares the @xmath180 elastic - scattering phase shifts in a manner similar to fig .
[ fig:0- ] .
the phase shift obtained with the mn potential is almost the same as the @xmath174 phase shift , which is consistent with the previous result @xcite that the energies of the negative parity states calculated with the mn potential are found to be degenerate . in the case of the av8@xmath6 potential ,
the @xmath180 phase shifts grows significantly in the full calculation , indicating a resonant behavior .
the coupling effect between the @xmath117+@xmath12 and @xmath116+@xmath116 channels is however much less compared to the @xmath177 state .
this is because the incoming @xmath179-wave coupled to the @xmath13-states in the clusters gives rise to several @xmath153 values to produce the @xmath181 state and therefore the tensor coupling does not concentrate sufficiently to produce a sharp resonance .
[ fig:1- ] displays the @xmath182 and @xmath183 elastic - scattering phase shifts calculated with the av8@xmath6 ( left ) and mn ( right ) potentials .
note that no physical @xmath2+@xmath2 channel exists in the case of the @xmath183 state .
because both full and @xmath117+@xmath12 calculations give almost the same phase shifts , only the full result is shown in the figure .
the @xmath182 phase shift calculated with the mn potential is again almost the same as those of the @xmath174 and @xmath180 cases , supporting that the three negative parity states become almost degenerate .
the @xmath182 elastic - scattering phase shift calculated with the av8@xmath6 potential is qualitatively similar to that of @xmath180 .
the attractive nature of the @xmath3+@xmath4 phase shift becomes further weaker , and to identify a resonance appears to be very hard . even though it is possible in some way
, its width would be a few mev , which is not in contradiction to experiment .
the @xmath183 phase shifts are very small in both av8@xmath6 and mn cases . for the negative parity states , the full model with the av8@xmath6 potential gives results that are consistent with both experiment and the theoretical calculation of ref .
we have pointed out that the phase shift behavior reveals the importance of the tensor force particularly in the case of @xmath177 .
its effect is often masked however by the coupling between the @xmath13 states in the clusters and the incoming partial wave . in this subsection
, we investigate the phase shifts of the negative parity states which have dominant @xmath184 components . in fig .
[ fig.13 ] , we represent three experimental negative parity @xmath184 energies ( left ) .
the states are observed at @xmath185 ( @xmath177 ) , @xmath186 ( @xmath181 ) and @xmath187 ( @xmath188)mev below the four - nucleon threshold @xcite .
the former two are located below the @xmath2+@xmath2 threshold and their widths are 0.84 and 2.01mev , respectively , whereas the last one is above the @xmath2+@xmath2 threshold and its width is fairly broad ( 6.1 mev ) . here , we calculate these energies as @xmath189 ( @xmath177 ) , @xmath190 ( @xmath181 ) and @xmath191 ( @xmath188)mev , which are approximated by the half - value position from the maximum phase shifts .
the present calculation is not projected out to @xmath184 , but the dominant configurations of @xmath3+@xmath4 , @xmath8+@xmath9 and @xmath2+@xmath2 elastic scattering are @xmath184 .
our calculated energies with av8@xmath6 reproduce the ordering of @xmath177 , @xmath181 and @xmath188 ( middle in fig .
[ fig.13 ] ) .
the splitting between the two lower states @xmath177 and @xmath192 is reproduced , but the experimental @xmath188 energy is higher than the calculation .
however , the determination of the energy for such a high energy state with large decay width ( 6.1 mev ) is very difficult from both experimental and theoretical side , and it usually has a large ambiguity .
this type of analysis was done by horiuchi and suzuki , who applied the correlated gaussian basis with two global vectors to study the energy spectrum of @xmath10he @xcite . because the results of ref .
@xcite are based on approximate solutions that impose no proper resonance boundary condition , it is interesting to see how the tensor force changes the phase shifts in the negative parity states as shown in this subsection .
these authors also found that the negative parity states with @xmath184 turn out to be almost degenerate when the mn potential that contains no tensor force is employed . in the present calculation , three states ( @xmath177 , @xmath188 , @xmath181 ) completely degenerate at the same energy , @xmath193 mev ( right ) , and the same phase shift pattern ( solid lines in figs .
[ fig:0- ] , [ fig:2- ] , [ fig:1- ] ) .
thus we can expect to see a clear evidence for the tensor force in the scattering involving the negative parity states .
we have investigated the distortion of clusters appearing in the low - energy @xmath2+@xmath2 and @xmath3+@xmath4 elastic scattering using a microscopic cluster model with the triple global vector method .
we showed that the tensor interaction changes the phase shifts very much by comparing a realistic interaction and an effective interaction . in the present @xmath0-@xmath1 type cluster model
, the description of the cluster wave functions is extended from a simple ( 0@xmath5 ) harmonic - oscillator shell model to a few - body model . to compare distortion effects of the clusters with realistic and effective interactions , we employed the av8@xmath6 potential as a realistic interaction and the mn potential as an effective interaction .
for the realistic interaction , the calculated @xmath7 phase shift shows that the @xmath3+@xmath4 and @xmath8+@xmath9 channels strongly couple with the @xmath2+@xmath2 channel .
these channels are coupled because of the tensor interaction . on the contrary ,
the coupling of these @xmath117+@xmath12 channels plays a relatively minor role for the case of the effective interaction because of the absence of tensor term .
in other words , the @xmath117+@xmath12 channels strongly affect the @xmath2+@xmath2 elastic phase shift with the realistic interaction , but not with the effective interaction . for the 2@xmath137 phase shifts ,
there is a @xmath159 component in all physical channels ( @xmath2+@xmath2 , @xmath3+@xmath4 and @xmath8+@xmath9 ) .
the coupling of the @xmath116+@xmath116 and @xmath117+@xmath12 channels in @xmath161 is weaker than in @xmath194 because of a weaker tensor coupling as discussed in section [ results ] , and the calculated phase shifts are very similar for the realistic and effective potentials . for other positive parity cases ,
the phase shift behavior of the realistic and effective potentials are very similar , and the coupling between the @xmath116+@xmath116 and @xmath117+@xmath12 channels can be neglected or is very small . furthermore , the tensor interaction makes the energy splitting of the @xmath177 , @xmath181 and @xmath188 negative parity states of @xmath10he consistent with experiment .
no such splitting is however reproduced with the effective interaction .
we believe that the physical picture obtained in the large model space with the realistic interaction should be close to the real physical situation .
it is needless to say that @xmath0-@xmath1 reaction calculations are very important to understand the underlying reaction dynamics involving continuum states .
simpler calculations using effective interactions in the same framework , as carried out in the present paper , are also meaningful because we can understand more clearly the effect of the tensor force by comparing both calculations .
the reaction calculations with the microscopic cluster model , whose model space and interactions are restricted , have been successfully applied to many heavier nuclei .
therefore , it is instructive to see the difference from the realistic interaction by employing a simple conventional effective interaction as mn in the few - nucleon systems
. it will be quite interesting to see the importance of the tensor force in reaction observables of four nucleons . as a direct application of the present study the radiative capture reaction @xmath195he at energies of astrophysical interest is of prime importance .
it is expected to take place predominantly via @xmath16 transitions @xcite . as is seen from table [ chan0 ]
, the two deuterons can approach each other in the @xmath66-wave only when @xmath111 is either 0@xmath137 ( @xmath7 ) or 2@xmath137 ( @xmath158 ) .
the former case is excluded because a radiative capture reaction of @xmath196 is forbidden in the lowest - order electromagnetic interaction , and hence the @xmath16 transition should be predominant .
if there were no tensor force present , the radiative capture would be suppressed near @xmath197 because neither @xmath2 nor @xmath10he would have a @xmath13-wave component in contradiction with the flat behavior of the astrophysical @xmath66-factor @xcite .
the tensor force strongly changes this story because it can couple @xmath66- and @xmath13-waves , bringing a significant amount of @xmath13-state probability in both @xmath10he and @xmath2 .
details of this analysis will be reported elsewhere .
+ + * acknowledgment * + we thank dr .
r. kamouni for helpful discussions based on his phd thesis ( in french ) .
this work presents research results of bilateral joint research projects of the jsps ( japan ) and the fnrs ( belgium ) .
y. s. is supported by a grant - in - aid for scientific research ( no .
21540261 ) .
this text presents research results of the belgian program p6/23 on interuniversity attraction poles initiated by the belgian - state federal services for scientific , technical and cultural affairs ( fstc ) .
d. b. and p. d. also acknowledge travel support of the fonds de la recherche scientifique collective ( frsc ) .
the part of computational calculations were carried out in t2k - tsukuba .
we define an auxiliary coefficient @xmath97 that appears in the coupling @xmath198_0]_a \to z(ab)[y_{a+b}({{\mbox{\boldmath $ e$}}}_1)y_b({{\mbox{\boldmath $ e$}}}_2)]_a.\ ] ] by introducing a coefficient @xmath199 for the coupling @xmath200_{c}=c(ab , c)y_{c}({{\mbox{\boldmath $ e$}}}_1)$ ] , we can express @xmath97 as @xmath201 note that @xmath202 vanishes unless @xmath203 is even .
the coefficients that appear in sect .
[ sect.4 ] are given as follows : @xmath204 \left[\begin{array}{ccc } a & b & q\\ a ' & b ' & q'\\ a\!+\!a'&b\!+\!b ' & \ell'\\ \end{array}\right]c(aa ' , a\!+\!a')c(bb ' , b\!+\!b')c(cc ' , c\!+\!c').\end{aligned}\ ] ]
the purpose of this appendix is to collect formulas for various matrix elements .
the main procedure to derive the formulas is sketched in sect .
[ sect.4 ] .
more details for the case of two global vectors are given in ref .
@xcite .
[ app.b ] let @xmath205 denote the momentum operator conjugate to @xmath206 , @xmath207 .
the total kinetic energy operator for the @xmath12-nucleon system with its center of mass kinetic energy being subtracted takes the form @xmath208 where @xmath209 is the total momentum , @xmath210 , and @xmath211 is an @xmath212 symmetric mass matrix . defining @xmath213-dimensional column vectors @xmath214 as @xmath215 and an @xmath212 matrix @xmath216 @xmath217
we can calculate the matrix element for the kinetic energy through the overlap matrix element @xmath218 where @xmath219 the @xmath220 values are defined in eq .
( [ def.rho ] ) .
a two - body interaction @xmath222 can be expressed as @xmath223 once the matrix element of @xmath224 is obtained , the matrix element of the interaction is calculated by integrating over @xmath225 the @xmath221-function matrix element weighted with the form factor @xmath226 .
similarly , for a one - body operator @xmath227 its matrix element can be obtained from that of the @xmath221-function . because both @xmath228 and @xmath229 can be expressed in terms of a linear combination of the relative coordinate @xmath230 , it is enough to calculate the matrix element of @xmath231 , where @xmath232 is a combination constant to express @xmath228 or @xmath229 .
the matrix element of the @xmath221-function is given by @xmath233 with @xmath234 the summation over non - negative integers @xmath235 and @xmath236 is restricted by the following equations @xmath237 here @xmath238 is defined as a coefficient that appears in the coupling of a product of six terms @xmath239_{00 } = \sum_{\ell_{12}\ell_{45}\ell \ell ' \kappa}\overline{o}(p_i ; \ell_{12 } \ell_{45 } \ell \ell ' \kappa ) \nonumber \\ & \times & [ [ [ [ y_{p_1}({{\mbox{\boldmath $ e$}}}_1)y_{p_2}({{\mbox{\boldmath $ e$}}}_2)]_{\ell_{12}}y_{p_3}({{\mbox{\boldmath
$ e$}}}_3)]_{\ell}\ [ [ y_{p_4}({{\mbox{\boldmath $ e$}}}_4)y_{p_5}({{\mbox{\boldmath $ e$}}}_5)]_{\ell_{45}}y_{p_6}({{\mbox{\boldmath $ e$}}}_6)]_{\ell'}]_{\kappa}\ y_{\kappa}(\widehat{{\mbox{\boldmath $ r$}}})]_{00},\end{aligned}\ ] ] and it is given by @xmath240 the @xmath225-dependence of the matrix element ( [ me.del ] ) is @xmath241 for a central interaction , @xmath226 is a scalar function , and the sum over @xmath242 in eq .
( [ me.del ] ) is limited to 0 .
for a tensor interaction , the angular dependence of @xmath226 is proportional to @xmath243 , and @xmath242 is limited to 2 .
the electric multipole operator is a special case of one - body operator , so that one can make use of the formula ( [ me.del ] ) to calculate its matrix element .
more explicitly , we give the matrix element of @xmath244 that includes all the cases mentioned above : @xmath245 with the integral of the potential form factor @xmath246 in case @xmath247 takes the form of @xmath248 @xmath249 , the integral @xmath250 can be obtained analytically , giving a closed form for the matrix element .
it should be noted that the matrix element for a special class of a three - body force can be evaluated with ease .
for example , if the radial part of the three - body force has a form @xmath251 the exponent can be rewritten as @xmath252 with an @xmath212 symmetric matrix @xmath253 , where @xmath254 , @xmath255 and @xmath256 are defined by @xmath257 , @xmath258 and @xmath259 .
thus the matrix element reduces to that of the overlap with @xmath260 being replaced with @xmath261 @xmath262 the spatial form of a spin - orbit interaction reads @xmath263 where @xmath264 @xmath265 stands for the @xmath266th component of a vector product of @xmath267 and @xmath268 . as in the @xmath221-function matrix element ,
the spin - orbit potential is written as @xmath269 where @xmath270 is expressed in terms of the momentum operators @xmath271 , @xmath272 .
the matrix element of the spin - orbit potential is given by @xmath273 where @xmath274 @xmath275 is @xmath276 and where the non - negative integers @xmath235 and @xmath236 are constrained to satisfy the equations @xmath277 the symbol @xmath278 stands for the factor @xmath279 where @xmath280 is the @xmath281th element of the column vector @xmath282 defined in eq .
( [ def.gamma ] ) .
the coefficient @xmath283 appears in the coupling @xmath284_{\ell_{12}}y_{p_3}({{\mbox{\boldmath $ e$}}}_3)]_{\ell}\ [ [ y_{p_4}({{\mbox{\boldmath $ e$}}}_4)y_{p_5}({{\mbox{\boldmath $ e$}}}_5)]_{\ell_{45}}y_{p_6}({{\mbox{\boldmath $ e$}}}_6)]_{\ell'}]_{1}]_{1\mu } \nonumber \\ & & \to \sum_{\overline{\ell}_{12 } \overline{\ell}_{45 } \overline{\ell}\ \overline{\ell}'}t_k(p_i,\ell_{12}\ell_{45}\ell \ell ' , \overline{\ell}_{12 } \overline{\ell}_{45 } \overline{\ell}\ \overline{\ell } ' ) \nonumber \\ & & \qquad \times [ [ [ y_{{p}_1^k}({{\mbox{\boldmath $ e$}}}_1)y_{{p}_2^k}({{\mbox{\boldmath $ e$}}}_2)]_{\overline{\ell}_{12}}y_{{p}_3^k}({{\mbox{\boldmath
$ e$}}}_3)]_{\overline{\ell}}\ [ [ y_{{p}_4^k}({{\mbox{\boldmath $ e$}}}_4)y_{{p}_5^k}({{\mbox{\boldmath $ e$}}}_5)]_{\overline{\ell}_{45}}y_{{p}_6^k}({{\mbox{\boldmath $ e$}}}_6)]_{\overline{\ell}'}]_{1\mu}.\end{aligned}\ ] ] the coefficients @xmath285 are given below : @xmath286
the fourier transform of the correlated gaussian function ( [ cgtgv ] ) defines the corresponding basis function in momentum space .
the momentum space function is useful to evaluate those matrix elements which depend on the momentum operator @xcite .
suppose that we want to evaluate the matrix element of a two - body operator @xmath287 or a one - body operator @xmath288 . obviously evaluating the matrix element can be done more easily in momentum space . for this purpose
we need to obtain the fourier transform of the coordinate space function .
a great advantage in the correlated gaussian function @xmath289 is that its fourier transform is a linear combination of the correlated gaussian functions in the momentum space .
thus by expressing @xmath290 or @xmath291 as @xmath292 , we can calculate the matrix element of the momentum - dependent operators in exactly the same way as in the coordinate space . as in the case with two global vectors @xcite
, the transformation from the coordinate to momentum space is achieved by a function @xmath293 where @xmath294 is an @xmath295-dimensional column vector whose @xmath281th element is @xmath296 . with a straightforward integration together with the recoupling of angular momenta , we can show that @xmath297 where the coefficient @xmath298 is given by @xmath299 \nonumber \\ & & \quad \times ( \widetilde{u_1}a^{-1}u_2)^{\ell_1}\ ( \widetilde{u_1}a^{-1}u_3)^{\ell_2}\ ( \widetilde{u_2}a^{-1}u_3)^{\ell_3},\end{aligned}\ ] ] where @xmath97 and @xmath300 are defined in appendix a. non - negative integers @xmath301 run over all possible values that satisfy @xmath302 .
the value of @xmath303 is restricted by the triangular relations among ( @xmath304 ) and ( @xmath305 ) .
99 wildermuth k , tang y c ( 1977 ) a unified theory of the nucleus ( vieweg , braunschweig ) .
kamada h , nogga a , glckle w , hiyama e , kamimura m _
( 2001 ) phys rev 64:044001 varga k , suzuki y , r. g. lovas ( 1994 ) nucl phys a 571:447 varga k , ohbayasi k , suzuki y ( 1997 ) phys lett b 396:1 ; varga k , usukura j , suzuki y ( 1998 ) phys rev lett 80:1876 ; usukura j , varga k , suzuki y ( 1998 ) phys rev a58:1918 suzuki y , varga k ( 1998 ) stochastic variational approach to quantum - mechanical few - body problems ( lecture notes in physics , vol .
springer , berlin heidelberg new york varga k , suzuki y ( 1995 ) phys rev c 52:2885 suzuki y , horiuchi w , orabi m , arai k ( 2008 ) few - body syst 42:33 varga k , suzuki y , usukura j ( 1998 ) few - body syst 24:81 carlson j , schiavilla r ( 2008 ) rev mod phys 70:743 ; pudliner b.s , pandharipande v.r , carlson j , pieper s.c , wiringa r.b ( 1997 ) phys rev c 56:1720 navratil p , kamuntavicius g.p , barrett b.r ( 2000 ) phys rev c 61:044001 viviani m ( 1998 ) few - body syst 25:197 feldmeier h , neff t , roth r , schnack j ( 1998 ) nucl phys a632:61 ; neff t , feldmeier h ( 2003 ) nucl phys a 713:311 arai k , aoyama s , suzuki y ( 2010 ) phys rev c 81:037301 phitzinger b , hofmann m , hale g.m ( 2001 ) phys rev c 64:044003 deltuva a , fonseca a.c ( 2007 ) phys rev c 75:014005 ; deltuva a , fonseca a.c ( 2007 ) phys rev lett 98:162502 quaglioni s , navratil p ( 2009 ) phys rev c 79:044606 ; quaglioni s , navratil p ( 2008 ) phys rev lett 08:092501 viviani m , rosati s , kievsky a ( 1998 ) phys rev lett 81:1580 ; viviani m , kievsky a , rosati s , george e.a , knulson l.d ( 2001 ) phys rev lett 86:3739;viviani m , kievsky a , girlanda l , marcucci l.e , rosati s ( 2009 ) few - body syst 45:119 lazauskas r , carbonell j , fonseca a.c , viviani m , kievsky a , rosati s ( 2005 ) phys rev c 71:034004 fisher b.m @xmath306 @xmath307 ( 2006 ) phys rev c 74:034001 arriaga a , pandharipande v.r , schiavilla ( 1991 ) phys rev c 43:983 sabourov k @xmath306 @xmath307 ( 2004 ) phys rev c 70:064601 hofmann h.m , hale g.m ( 2008 ) phys rev c 77:044002 hofmann h.m , hale g.m ( 1997 ) nucl phys a 613:69 ; hofmann h.m , hale g.m ( 2003 ) phys rev c 68:021002 deltuva a , fonseca a.c ( 2007 ) phys rev c 76:021001 ; deltuva a , fonseca a.c , sauer p.u ( 2008 ) phys lett b 660:471 lazauskas r , carbonell j ( 2004 ) few - body syst 34:105 ciesielski f , carbonell j , gignoux c ( 1999 ) phys lett b 447:199 assenbaum h.j , langanke k ( 1987 ) phys rev c 36:17 fowler w.a , caughlan , zimmenrman ( 1967 ) annu rev astron astrophys 5:525 baye d ,
heenen p h , libert - heinemann m ( 1977 ) nucl phys a 291:230 kanada h , kaneko t , saito s , tang y c ( 1985 ) nucl phys a 444:209 arai k , descouvemont p , baye d , ( 2001 ) phys rev c 63:044611 descouvemont p , baye d ( 2010 ) rep prog phys 73:036301 pudliner b s , pandharipande v r , carlson j , pieper s c , wiringa r b ( 1997 ) : phys rev c 56 , 1720 hiyama e , gibson b f , kamimura m ( 2003 ) phys rev c 70:031001 thompson d r , lemere m , tang y c ( 1977 ) nucl phys a 286:53 boys s f ( 1960 ) proc r soc london ser a 258:402 ; singer k ( 1960 ) _ ibid . _
258:412 suzuki y , usukura j ( 2000 ) nucl inst method b 171:67 suzuki y , usukura j , varga k ( 1998 ) j phys b 31:31 horiuchi w , suzuki y ( 2008 ) phys rev c 78:034305 tilley d r , weller h r , hale g m ( 1992 ) , nucl phys a 541:1 santos f d , arriaga a , eir a m , tostevin j a ( 1985 ) phys rev c 31:707 wachter b , mertelmeier t , hofmann h m ( 1988 ) phys lett b 200:246 angulo c , arnould m , rayet m , descouvemont p , baye d _ et al . _ ( 1999 ) nucl phys a 656:3 | elastic - scattering phase shifts for four - nucleon systems are studied in an @xmath0-@xmath1 type cluster model in order to clarify the role of the tensor force and to investigate cluster distortions in low energy @xmath2+@xmath2 and @xmath3+@xmath4 scattering . in the present method ,
the description of the cluster wave function is extended from a simple ( 0@xmath5 ) harmonic - oscillator shell model to a few - body model with a realistic interaction , in which the wave function of the subsystems are determined with the stochastic variational method . in order to calculate the matrix elements of the four - body system ,
we have developed a triple global vector representation method for the correlated gaussian basis functions . to compare effects of the cluster distortion with realistic and effective interactions ,
we employ the av8@xmath6 potential as a realistic interaction and the minnesota potential as an effective interaction . especially for @xmath7 ,
the calculated phase shifts show that the @xmath3+@xmath4 and @xmath8+@xmath9 channels are strongly coupled to the @xmath2+@xmath2 channel for the case of the realistic interaction . on the contrary , the coupling of these channels plays a relatively minor role for the case of the effective interaction .
this difference between both potentials originates from the tensor term in the realistic interaction .
furthermore , the tensor interaction makes the energy splitting of the negative parity states of @xmath10he consistent with experiments .
no such splitting is however reproduced with the effective interaction . |
the detection of x - ray emission from wolf - rayet ( wr ) stars was one of the many notable discoveries of the _ einstein observatory _
( seward et al . 1979 ; pollock 1987 ) .
subsequent studies have broadened our knowledge of the x - ray properties of wr stars but the underlying physical mechanisms ultimately responsible for their x - ray emission are still not well understood .
most current theoretical models attribute the x - ray emission of wr stars ( and o - type stars ) to thermal plasma produced in shocks that form in their powerful supersonic winds .
wr stars have typical terminal wind speeds v@xmath1 @xmath6 1000 - 5000 km s@xmath7 and mass - loss rates @xmath4 @xmath6 10@xmath8 m@xmath9 yr@xmath7 .
it has also been suggested that wr stars might produce nonthermal x - rays via such processes as inverse - compton scattering , synchrotron emission , or nonthermal bremsstrahlung ( pollock 1987 ; chen & white 1991 ) .
however , most wr stars detected so far show line - dominated x - ray spectra that can be acceptably fitted by thermal models .
but , nonthermal emission has not yet been ruled out in a few cases ( e.g. wr 142 ; sokal et al .
2010 ) .
several different shock - related processes could in principle lead to thermal x - ray emission in wr stars .
first , soft x - rays ( kt @xmath10 1 kev ) may be produced in shocks embedded in the outflowing wind that form as a result of the line - driven instability ( lucy & white 1980 ; lucy 1982 ; feldmeier et al .
1997 ; gayley & owocki 1995 ) .
this mechanism may be responsible for the soft x - ray emission detected in some o - type stars ( cassinelli et al .
2001 ; kahn et al .
but , all wr stars detected so far show a harder component ( kt @xmath11 2 kev ) in addition to a soft component that is usually present .
the hard component is not expected if the emission arises solely from line - driven instability shocks .
second , a sufficiently strong magnetic field could confine the wind and channel it into two opposing streams which collide near the magnetic equator , forming a magnetically - confined wind shock ( mcws ) .
this mechanism is capable of producing hard x - rays ( kt @xmath11 2 kev ) and may be relevant to x - ray production in young magnetic ob stars such as @xmath12 ori c ( babel & montmerle 1997b ; gagn et al .
2005 ) and @xmath13 ori e ( skinner et al . 2008 ) .
however , very strong surface magnetic fields would be needed to confine the more powerful winds of wr stars ( ud - doula & owocki 2002 ; skinner et al .
2010 , hereafter s10 ) , and observational evidence for strong magnetic fields on wr stars is currently lacking .
thus , the question of whether the mcws mechanism operates in wr stars remains open .
finally , in the case of wr @xmath14 ob binary systems , hard x - rays ( kt @xmath11 2 kev ) can originate in the region between the two stars where the interaction of their winds results in a colliding wind ( cw ) shock ( prilutskii & usov 1976 ; luo et al .
1990 ; stevens et al . 1992 ; usov 1992 ) .
hot x - ray plasma that is thought to arise at least partially in a cw shock has now been detected in grating spectra of several x - ray bright wr @xmath14 ob binaries ( @xmath15 vel : skinner et al .
2001 ; schild et al .
2004 ; wr 140 : pollock et al . 2005 ; wr 147 : zhekov & park 2010 ) .
x - ray spectra of wr stars are needed to distinguish between the above possibilities . although good - quality x - ray grating spectra of a few x - ray bright wr@xmath14ob binaries have been obtained ( as noted above ) , the study of putatively single wr stars as a class has been hampered by a general lack of x - ray spectral data .
the observational situation has improved considerably during the past few years as a result of the completion of an exploratory x - ray survey of single wr stars with _ xmm - newton _ and _ chandra _
moderate - resolution ccd x - ray spectra are now available for several wn stars but attempts to detect carbon - rich wc stars have so far only yielded upper limits on their x - ray luminosities with typical constraints of log l@xmath0 @xmath10 31.0 ergs s@xmath7 ( skinner et al .
2006 ; see also fig . 9 of s10 ) .
a surprising result of the initial survey was that all detected wn stars showed evidence of hot plasma ( kt@xmath16 @xmath11 2 kev ) in their x - ray spectra .
this result is at odds with x - ray production via the line - driven instability mechanism , for which only soft emission is expected .
it is thus apparent that other mechanisms are needed to fully explain the origin of x - ray emission in wn stars and , within the current theoretical framework , either the mcws or cw shock processes could be responsible .
however , there is no compelling justification to invoke the cw interpretation because the wn stars in the initial survey sample were selected on the basis of lack of evidence for binarity .
the presumption that these stars are indeed single is of course subject to the usual caveat that close companions can escape detection . in the initial survey , _ xmm - newton _ and _
chandra _ data for the following detected wn stars were analyzed : wr 2 ( wn2 ) , wr 18 ( wn4 ) , wr 20b ( wn6h ) , wr 24 ( wn6 ha ) , wr 134 ( wn6 ) , wr 136 ( wn6h ) , and wr 79a ( wn9
in addition , archival _ rosat _ data were analyzed for wr 16 ( wn8h ) and wr 78 ( wn7h ) .
neither of these two wnl stars showed significant x - ray emission ( see also sanders et al .
1985 ) , although a marginal ( 2@xmath13 ) excess was noted in hard - band _ rosat _ pspc0.1 - 2.35 kev energy range but lacked sensitivity at higher energies e @xmath11 2.35 kev .
pspc hard - band images cover the range e @xmath17 0.5 - 2 kev .
in contrast , _ xmm - newton _ provides coverage to higher energies and we use the energy range e = 2 - 8 kev to refer to the _ xmm - newton _ hard - band . ]
images of wr 78 .
we present here more sensitive _ xmm - newton _ observations of wr 16 and wr 78 that were obtained after the results of our initial wn survey were published ( s10 ) .
these observations are important because they provide additional information on the x - ray properties of wnl subtypes , which comprise @xmath1724% of all wr stars listed in the _ viith catalog of wolf - rayet stars _
( van der hucht 2001 ) .
wnl stars are believed to be the least chemically - evolved wr stars and some o - type stars probably enter the wr phase as wnl stars ( crowther & bohannan 1997 ) .
wnl stars typically have higher mass - loss rates but lower terminal wind speeds than wne stars ( table 2 of crowther 2007 ) .
we report here a faint x - ray detection of wr 16 ( wn8h ) and bright emission including a hot plasma component from wr 78 ( wn7h ) .
the _ xmm - newton _ observations are listed in table 1 .
the general properties of the wnl stars discussed here are summarized in table 2 and their x - ray properties are given in table 3 .
the targets were selected from the _ viith wr catalog _ ( van der hucht 2001 ) based on proximity , low visual extinction , and lack of evidence for binarity . in addition to the new detections of wr 16 and wr 78 , we also include data from our previous _
xmm - newton _ observation ( s10 ) of the wnl star wr 79a for comparison .
data were acquired with the european photon imaging camera ( epic ) in full - window mode .
epic provides charge - coupled device ( ccd ) imaging spectroscopy from the pn camera ( strder et al .
2001 ) and two nearly identical mos cameras ( mos1 and mos2 ; turner et al .
the epic cameras provide energy coverage in the range e @xmath17 0.2 - 15 kev with energy resolution e/@xmath18e @xmath17 20 - 50 .
the mos cameras provide the best on - axis angular resolution with fwhm @xmath17 4.3@xmath19 at 1.5 kev .
data were reduced using the _ xmm - newton _ science analysis system ( sas vers .
11.0 ) using standard procedures including the filtering of raw event data to select good event patterns and removal of data within time intervals of high background radiation .
spectra and light curves were extracted from a circular region of radius r = 15@xmath19 ( @xmath1768% encircled energy ) centered on the source , thus capturing most of the source counts while reducing background counts .
background analysis was conducted on circular source - free regions near the source .
the sas tasks _ rmfgen _ and _ arfgen _ were used to generate source - specific rmfs and arfs for spectral analysis .
the data were analyzed using the heasoft _ xanadu _ software package .
this wn8h star has the lowest terminal wind speed ( v@xmath1 @xmath17 630 - 740 km s@xmath7 ) of any wn star in our exploratory x - ray survey .
the star is surrounded by a cocoon of molecular gas which was probably ejected during a previous evolutionary phase ( marston et al .
optical variability is present but its cause is not fully understood ( antokhin et al .
an early study attributed the variability to an unseen low - mass companion ( moffat & niemela 1982 ) but this interpretation was later called into question ( manfroid et al .
the stellar distance is somewhat uncertain with published values ranging from 2.37 - 4.36 kpc ( crowther et al .
1995 ; nugis & lamers 2000 ; van der hucht 2001 , hamann et al .
2006 ) and we adopt here an intermediate value d @xmath17 3.6 kpc
. visual inspection of energy - filtered epic images showed no obvious signs of an x - ray source at the stellar optical position .
however , more in - depth image analysis revealed a weak emission excess above the local background level in both the pn and mos images .
we extracted counts ( 0.3 - 8 kev ) within a circular region of radius 15@xmath19 ( 68% encircled energy ) centered on the star in each epic detector , and background counts in different source - free regions near the star .
after background subtraction we obtained 31@xmath206 net pn counts in 30.0 ks of usable exposure and 24@xmath205 net mos1@xmath142 ( sum of mos1 and mos2 ) in 66.0 ks of usable exposure ( 33.0 ks per mos ) .
the faint emission is clearly visible in the contoured mos1@xmath142 image in figure 1 .
although the statistical significance of the excess is low ( @xmath172@xmath13 ) , its presence in all three epic detectors at an offset of @xmath101@xmath19 from the _ hst _ gsc optical position ( table 3 ) strengthens the conclusion that weak x - ray emission from wr 16 is detected .
the emission is too faint for spectral and timing analysis .
however , we used the pimms simulator to estimate the x - ray luminosity of wr 16 based on the count rates .
the pn rate of 1.03 c ks@xmath7 gives an unabsorbed luminosity log l@xmath21(0.3 - 8 kev ) = 31.14 [ 30.77 - 31.30 ] ergs s@xmath7 for an assumed distance d = 3.6 kpc , where the range in square brackets reflects the range of published distances d = 2.37 - 4.36 kpc .
a pimms estimate based on the mos count rate gives a nearly identical unabsorbed value log l@xmath21(0.3 - 8 kev ) = 31.17 [ 30.80 - 31.33 ] ergs s@xmath7 .
for the underlying spectrum , we assumed an absorbed two - temperature optically thin thermal plasma ( @xmath22 ) with kt@xmath23 = 0.6 kev and kt@xmath24 = 3.5 kev , typical of detected wn stars ( s10 ) .
we adopted an absorption column density n@xmath25 = 3.7 @xmath26 10@xmath27 @xmath28 based on a@xmath29 = 1.67 ( table 1 ) and the gorenstein ( 1975 ) conversion from a@xmath29 to n@xmath25 . the slightly different conversion of vuong et al .
( 2003 ) decreases the upper limit on l@xmath21 by 0.05 dex .
the above l@xmath21 is the lowest of any wn star detected so far , but the derived value is dependent on the assumed spectral model and absorption .
most wn stars show x - ray absorption in excess of that expected from a@xmath29 ( s10 ; see also the results for wr 78 in sec .
3.2 below ) .
if that is the case for wr 16 , then its unabsorbed l@xmath21 could be larger than the value inferred above . in this regard , we note that the pn counts for wr 16 are distributed roughly equally between the soft ( 0.3 - 2 kev ) and hard ( 2 - 8 kev ) bands .
but , xspec simulations show that if the absorption were as low as expected based on a@xmath29 , only about @xmath1715% - 20% of the counts would emerge in the hard band .
this suggests that excess absorption above that expected from a@xmath29 is present in wr 16 , but a higher signal - to - noise spectrum would be needed to confirm this . for those wn stars where sufficient counts are available for spectral analysis , the best - fit n@xmath25 values are typically @xmath172.0 - 2.5 times larger than expected from estimates based on a@xmath29 ( table 4 of s10 ) .
if we assume that is the case for wr 16 then its unabsorbed l@xmath21 determined from pimms is @xmath170.10 - 0.15 dex larger than the values quoted above .
this wn7h star has been included in numerous optical studies and there is fairly good agreement in published stellar parameters when different studies are compared ( table 2 ) . optical variability has been reported ( vreux et al .
1987 ; bratschi & blecha 1996 ) but no periodicity that would hint of a companion has been found .
_ xmm - newton _ provides the first unambiguous detection of wr 78 ( fig .
the source is quite hard as is apparent from its median photon energy e@xmath30 = 2.47 kev ( table 3 ) .
clear evidence for high - temperature plasma is provided by the detection of the fe k@xmath31 emission line complex at 6.67 kev ( fig .
this line forms at a characteristic temperature t @xmath6 40 mk . using pn events in the 0.3 - 8 kev range
, the kolmogorov - smirnov ( ks ) test gives a probability of constant count rate p@xmath32 = 0.12 and a @xmath33 test applied to the background - subtracted pn light curve binned at 1000 s intervals gives p@xmath32 = 0.97 .
the ks test operates on all detected events within the source extraction region , and thus includes both background and source events .
the lower value of p@xmath32 based on the ks test may be due to background effects .
but , neither the ks test nor @xmath33 analysis suggests statistically significant variability for wr 78 during the 26.26 ks ( 7.3 hours ) of usable exposure .
likewise , the other wn stars in the survey have failed to show any clear x - ray variability in observations spanning @xmath173 - 10 hours ( s10 ) .
acceptable fits of the ccd x - ray spectra of wn stars detected in the initial survey were obtained using a two - temperature ( 2 t ) optically thin plasma model consisting of cool ( kt@xmath23 ) and hot ( kt@xmath24 ) plasma components plus a single absorption component ( n@xmath25 ) .
this model is expressed symbolically as n@xmath25@xmath34(kt@xmath23 @xmath14 kt@xmath24 ) , where parentheses denote that both plasma components are viewed through the same absorption column .
interestingly , this model gives an unacceptable fit of the wr 78 x - ray spectrum as does a simpler 1 t model of the form n@xmath25@xmath34kt@xmath23 .
these single - absorption models fit the spectrum above 2 kev ( including the fe k@xmath31 line ) reasonably well , but generally underestimate the observed flux at energies below 1 kev .
a similarly poor fit below 1 kev occurs for other single - absorption models such as a plane - parallel shock model ( @xmath35 in xspec ) . in order to obtain an acceptable fit for wr 78 , different absorption components were required for the cool and hot components , i.e. a model of the form n@xmath36@xmath34kt@xmath23 @xmath14 n@xmath37@xmath34kt@xmath24 . the acceptable two - absorber model is summarized in table 4 .
the inferred plasma temperatures kt@xmath23 = 0.64 [ 0.48 - 0.88 ; 90% conf . ] kev and kt@xmath24 = 2.27 [ 1.63 - 2.90 ] kev are not substantially different from the wn9h star wr 79a ( sec .
3.3 ) when 90% confidence ranges are taken into account , but the value of kt@xmath24 is at the low end of the range for all wn stars observed so far ( s10 ) .
most of the x - ray flux and emission measure ( as gauged by the value of @xmath38 in table 4 ) of wr 78 is associated with the hotter plasma component , which is evidently viewed through much higher absorption than the cool component .
the separate contributions of the cool and hot components are shown in the bottom panel of figure 3 .
the best - fit absorption values of n@xmath36 = 1.38 [ 1.20 - 1.60 ] @xmath26 10@xmath39 @xmath28 and n@xmath37 = 6.55 [ 4.10 - 10.7 ] @xmath26 10@xmath39 @xmath28 are substantially larger than expected on the basis of visual extinction esimates , since a@xmath29 = 1.55 mag ( table 2 ) corresponds to n@xmath25 = 3.4 @xmath26 10@xmath27 @xmath28 using the conversion of gorenstein ( 1975 ) . the two - absorber model for wr 78
gives an unabsorbed x - ray luminosity log l@xmath0(0.3 - 8 kev ) = 32.84 ergs s@xmath7 ( d = 1.99 kpc ; table 4 ) , which is at the high end of the range for detected wn stars ( fig .
4 ; s10 ) . however , this value is quite uncertain because of the large dereddening correction .
all spectral models that we considered , including simplistic 1 t models , give values log l@xmath0(0.3 - 8 kev ) @xmath40 32.1 ergs s@xmath7 which we consider to be a reliable lower limit .
the most likely cause of the anomalous x - ray absorption is the wr wind . in that case
, the higher absorption associated with the hot plasma component suggests that it originates deeper in the wind ( closer to the star ) where the wind density is higher .
a similar conclusion was reached for the wn6h star wr 136 ( s10 ) and for o - type stars studied by waldron & cassinelli ( 2007 ) and naz ( 2009 ) . in the ideal case of a smooth unclumped wind , the radius of optical depth unity ( as measured from the star ) decreases toward higher x - ray energies so harder photons are able to escape from smaller radii near the star .
softer x - ray photons produced far out in the wind suffer less wind absorption and can escape whereas any soft photons produced in deeper denser layers of the wind are absorbed and remain undetected . in the more complex case of an inhomogenous ( clumped ) wind ,
the above picture is not altered substantially if the clumps are optically thin to x - rays but could well be different if dense optically thick clumps are present .
this wn9 ha star is the latest wn subtype detected in x - rays so far .
it is discussed in more detail in s10 but for comparison we summarize its x - ray properties in table 3 and give its best - fit spectral parameters in figure 3 .
the epic pn spectrum of wr 79a is softer than that of wr 78 as evidenced by its lower median photon energy ( table 3 ) and it is apparent from figure 3 that the x - ray absorption of wr 79a is less than that of wr 78 .
the pn spectra of both stars show the high - temperature s xv ( 2.46 kev ) emission line but the very high temperature fe k@xmath31 complex ( 6.67 kev ) detected in wr 78 is not seen in wr 79a .
the fundamental stellar parameters that govern intrinsic x - ray luminosity levels in wr stars have not yet been identified .
good quality x - ray data or useful upper limits exist for only about a dozen putatively single wn stars , so robust statistics and stringent tests for correlations of l@xmath0 against stellar parameters are still limited by small sample size .
this is in contrast to massive o - type stars , from which wr stars evolve .
x - ray data now exist for hundreds of o - type stars and the empirical relation l@xmath0
@xmath6 10@xmath41l@xmath42 was already becoming apparent from a large sample of o and early b stars detected in the _ rosat _ all - sky survey ( berghfer et al . 1997 ) .
based on existing _ xmm - newton _ and _ chandra _ x - ray data , there is no convincing evidence for a similar l@xmath0 @xmath43 l@xmath42 relation in putatively single wn stars .
likewise , no evidence for such a relation was found using _ rosat
_ data ( wessolowski 1996 ) .
as we have already shown ( s10 ) , wn stars with similar values of l@xmath42 can have l@xmath0 values differing by an order of magnitude or more .
most detected wn stars , including wr 78 ( table 4 ) , have l@xmath0 @xmath11 10@xmath41l@xmath42 with typical values being l@xmath0 @xmath6 10@xmath44l@xmath42 , as shown in figure 4 .
however there are a few wnl stars whose l@xmath0/l@xmath42 ratios are at least an order of magnitude lower .
most notable among these is wr 16 ( wn8h ) , for which we obtain log l@xmath0/l@xmath42 = @xmath458.2 @xmath20 0.2 when taking the range of published distances and stellar luminosities into account ( table 2 ) .
a similarly low l@xmath0/l@xmath42 ratio is also assured for wr 40 ( wn8h ) by virtue of its _ xmm _ non - detection with an x - ray upper limit log l@xmath0(0.5 - 10 kev ) @xmath46 31.6 ergs s@xmath7 at an assumed distance d = 3 kpc ( gosset et al .
2005 ) .
wind shock models of x - ray emission from massive stars predict that l@xmath0 will depend on mass - loss parameters . in the line - driven instability shock model , l@xmath0 @xmath43 @xmath47 where @xmath31 = 1 for the radiative shock case , @xmath31 = 2 for the adiabatic case , and @xmath31 = 0.6 if the observed l@xmath0 @xmath43 l@xmath42 scaling relation in o - type stars is to be reproduced ( owocki et al .
2011 ) . from an observational standpoint
, there is no clear dependence of l@xmath0 on @xmath4 for wn stars .
in particular wr 2 ( wn2 ) has one of the lowest mass - loss rates tabulated for wn stars ( log @xmath4 = @xmath455.4 - @xmath455.3 m@xmath9 yr@xmath7 ; nugis & lamers 2000 ; hamann et al .
2006 ) but its l@xmath0 is comparable to that of other wn stars having @xmath4 values an order of magnitude larger . in addition
, the two wn8 stars wr 16 and wr 40 are underluminous in x - rays by about an order of magnitude compared to other wn stars with similar mass - loss rates .
thus , mass - loss rate alone is not a reliable predictor of l@xmath0 in wn stars .
there is one clue in the existing data that suggests l@xmath0 in wn stars may depend on terminal wind speed v@xmath1 or some function thereof .
specifically , we note that the two observed wn8 stars wr 16 ( v@xmath1 @xmath17 630 - 740 km s@xmath7 ) and wr 40 ( v@xmath1 @xmath17 650 - 910 km s@xmath7 ) have the lowest terminal wind speeds of all wn stars ( hamann et al . 2006
; see also gosset al .
2005 for wr 40 ) .
these two stars also have the lowest l@xmath0 values , even though only an upper limit is currently available for wr 40 .
the wn8h star wr 156 also has a similar low wind speed ( v@xmath1 @xmath17 660 km s@xmath7 ; hamann et al . 2006 ) but it has not yet been observed in x - rays . in the absence of a binary companion we expect that any x - ray detection of wr 156 would be quite faint .
the apparent falloff in l@xmath0 with decreasing v@xmath1 suggests that l@xmath0 may be related to wind momentum (= @xmath4v@xmath1 ) or wind luminosity l@xmath2
(= @xmath3@xmath4v@xmath5 ) . using data in the initial wn x - ray survey ( s10 )
, we already noted a trend of increasing l@xmath0 with l@xmath2 .
the new detections of wr 16 and wr 78 presented here conform to and strengthen this trend ( fig .
this trend was not seen in _ rosat _
wn - star data analyzed by wesselowski ( 1996 ) , perhaps because _ rosat _ had little sensitivity above @xmath62 kev where a significant fraction of the x - ray flux in wn stars emerges . using the data in figure 5 , the generalized kendall s tau test gives a correlation probability p@xmath48 = 0.93 .
a linear regression fit gives log l@xmath0 = 0.89(@xmath200.47)log l@xmath2 @xmath14 26.41(@xmath200.55 ) where l@xmath0 is in units of ergs s@xmath7 and l@xmath2 in units of 10@xmath8m@xmath9 yr@xmath7 @xmath26 ( km s@xmath7)@xmath49 .
thus , the regression fit gives a nearly linear dependence l@xmath0
@xmath43 l@xmath50 but the slope is quite uncertain due to scatter in the data .
factors that could contribute to this scatter include uncertainties in distances , x - ray luminosity dereddening corrections , wr star mass - loss rates ( which are sensitive to clumping effects ) , and terminal wind speeds .
a notable outlier is the wn6h star wr 136 , which appears to be underluminous in x - rays compared to other wn stars with comparable l@xmath2 values ( fig .
5 ) and to other wn stars with comparable stellar luminosities l@xmath51 ( fig .
its l@xmath51 value is also low compared to other wnh stars ( fig .
6 of hamann et al .
this raises a question as to whether current distance estimates may be too low .
the estimates range from d = 1.26 kpc ( hamann et al .
2006 ) to d = 1.82 kpc ( nugis & lamers 2000 ) , whereas our derived x - ray luminosity log l@xmath0 = 31.51 ergs s@xmath7 is based on the _ hipparcos _ distance d = 1.64 kpc ( s10 ) .
we note that there is also a large uncertainty in the l@xmath0 value of wr 136 because of poorly - constrained x - ray absorption ( n@xmath25 ) in spectral fits .
some statistically - acceptable spectral models require higher absorption and yield dereddened values as high as log l@xmath0 @xmath17 32.2 ergs s@xmath7 ( d = 1.64 kpc )
. such higher values , if correct , would shift the position of wr 136 upward in figures 4 and 5 ( as shown by the l@xmath0 error bar ) , in better agreement with other wn stars .
if the above l@xmath0 @xmath43 l@xmath2 relation reflects a true physical dependence , then the obvious conclusion would be that a relatively constant fraction of the kinetic energy in wn star winds is being converted into x - ray energy .
interestingly , the mcws model of babel & montmerle ( 1997a ) predicts such a dependence .
specifically , they obtain l@xmath0 @xmath17 @xmath3@xmath4v@xmath52 , and they infer v@xmath53 @xmath6 v@xmath1 in order to reproduce the x - ray luminosity of the a0p star iq aur from their model .
thus , there is some theoretical support for a l@xmath0 @xmath43 l@xmath2 relation in early - type stars with magnetic fields .
even so , some observational scatter in this relation would be expected due to differences in b - field strengths between stars ( eq . [ 10 ] of babel & montmerle 1997a ) and bearing in mind that collision of the two hemispherical wind components could in some cases occur at subterminal speeds .
although the existence of strong surface magnetic fields on single wr stars is not yet established , some wr @xmath14 ob binaries do exhibit nonthermal radio emission which may be associated with weak magnetic fields above the surface ( abbott et al .
1986 ; skinner et al . 1999
; dougherty & williams 2000 ) .
since the mcws mechanism operates in single stars there is no need to postulate unseen companions in this interpretation .
if close unseen companions are present in our sample of putatively single wn stars then hot plasma could be created via the colliding wind mechanism .
if the wr wind momentum dominates that of the companion , then the cw shock would form at or near the companion surface .
this could also give rise to a l@xmath0 @xmath43 l@xmath2 dependence , but considerable scatter would be expected since in this case the predicted l@xmath0 also depends on binary separation and the radius of the companion star ( usov 1992 ; luo et al .
1990 ) . since we have no specific information on the existence of such companions , their stellar or wind properties , or their orbital parameters , a meaningful comparison of the observed l@xmath0 values with cw model predictions can not yet be made .
all wn stars for which good - quality ccd x - ray spectra exist show a high - temperature plasma component ( kt@xmath16 @xmath11 2 kev ) and almost always a cooler component ( kt@xmath54 @xmath10 1 kev )
. the cool component could be due to radiative instability wind shocks , but the origin of the hotter component is not yet known .
analysis of _ xmm _ epic and _ chandra _ acis spectra of wn stars reveals a broad range of hot - component temperatures from kt@xmath16 @xmath17 2 kev ( wr 1 , wr 78 ) up to kt@xmath16 @xmath17 4 - 8 kev ( wr 20b , wr 110 , wr 134 ; s10 ) .
such high - temperature plasma could be produced by supersonic wr winds shocking onto denser downstream material , or the wind from the opposite hemisphere ( as in the mcws model ) , or the wind or surface of as yet undetected companions ( cw shocks ) .
these models all predict that the _ maximum _ shock temperature should scale proportionally with the pre - shock wind - speed as kt@xmath55 @xmath17 1.96@xmath56v@xmath57 kev , where @xmath56 is the mean atomic weight per particle and v@xmath58 is the wind speed in units of 1000 km s@xmath7 ( babel & montmerle 1997a ; luo et al . 1990 ; , stevens et al .
1992 ) . for solar abundances @xmath56 @xmath17 0.6 but for metal - enriched wn star winds @xmath56 @xmath17 1.06 ( van der hucht et al . 1986 ) .
for the wn star case of interest here one thus obtains kt@xmath55 @xmath17 2v@xmath59@xmath49 kev , assuming that the wind has reached terminal speed at the shock interface .
terminal wind speeds of x - ray detected wn stars lie in the range @xmath60 @xmath17 700 - 2500 km s@xmath7 ( table 2 ; s10 ) .
thus , the above shock models predict maximum shock temperatures kt@xmath55 @xmath17 1 - 12 kev .
the values of kt@xmath16 determined from x - ray spectra lie comfortably within this range .
however , there is no clear trend for a proportionality relation kt@xmath16 @xmath43 v@xmath5 , as one might naively expect from shock - model predictions .
the wn star with the lowest v@xmath1 for which a good - quality ccd spectrum is available is the wn9 ha star wr 79a ( v@xmath1 @xmath17 955 km s@xmath7 , table 2 ) and its best - fit temperature is kt@xmath16 = 2.66 [ 1.66 - 5.60 ; 90% conf . ]
kev ( s10 ) . by comparison ,
the wn4 star wr 1 has a higher terminal speed ( v@xmath1 @xmath17 1600 - 2135 km s@xmath7 ; nugis & lamers 2000 ; ignace et al .
2001 ; hamann et al .
2006 ) but a lower best - fit temperature kt@xmath16 = 1.72 [ 1.36 - 2.21 ; 90% conf . ] kev based on our analysis of _ xmm _ archive data ( obsid 0552220101 ; see also ignace et al .
2003a for analysis of an earlier _ xmm _ observation with a shorter exposure ) . given the above results , it is not yet clear that terminal wind speed is the primary parameter governing hot - plasma component temperatures in wn stars .
other factors such as magnetic fields ( if present ) could play a role in plasma heating . even if the hot - component plasma temperature is governed by v@xmath1 ,
establishing such a dependence from the existing data will not be straightforward .
as noted above , shock models predict that the maximum shock temperature kt@xmath55 @xmath43 v@xmath5 , but shocks will in general produce plasma over a range of temperatures .
x - ray spectral fits of ccd spectra using simplistic 2 t models only provide an estimate of the _ average _ hot - component temperature kt@xmath16 , which will in general be less than kt@xmath61 . furthermore , the inferred values of kt@xmath16 from fits of moderate resolution ccd spectra can be uncertain by a factor of @xmath62 ( s10 ) . by comparison ,
values of v@xmath1 are typically uncertain by @xmath1710% - 20% .
more accurate temperature discrimination is needed to search for any correlation of maximum plasma temperature with wind speed .
high - resolution x - ray grating spectra can provide more detailed information on the distribution of x - ray plasma emission measure as a function of temperature .
but improvements in x - ray telescope sensitivity will be needed to acquire grating spectra for a sample of wn stars that is sufficiently large to undertake correlation studies .
the main results of this study are the following : 1 .
we report new x - ray detections of the wnl stars wr 16 ( wn8h ) and wr 78 ( wn7h ) .
these new detections , when combined with previous detections show that x - ray emission is present in all wn subtypes from wn2 - wn9 .
this contrasts sharply with single carbon - rich wc stars which so far remain undetected in x - rays .
the _ xmm _ detection of wr 16 is of low significance and the emission is too faint for spectral analysis .
however , the x - ray spectrum of wr 78 reveals a high - temperature plasma component ( kt@xmath16 @xmath17 2 kev ) , similar to that observed in other wn stars . such hotter plasma is not predicted for x - ray emission arising solely in line - driven instability shocks .
the origin of the hot plasma in wn stars is not yet known .
shock models predict that the maximum plasma temperature will scale with terminal wind speed according to kt@xmath55 @xmath43 v@xmath5 .
such a dependence is not yet seen in the existing data but could be masked because temperatures determined from ccd x - ray spectra are average ( not maximum ) values and are in many cases uncertain by a factor of @xmath62 .
the x - ray luminosities of the wn8 stars wr 16 ( a faint detection ) and wr 40 ( a non - detection ) are an order of magnitude less than typical values observed for other detected wn stars .
these wn8 stars also have the lowest terminal wind speeds of any wn stars observed so far , suggesting a possible link between terminal wind speed ( or a parameter that depends on wind speed ) and x - ray luminosity in single wn stars .
the two new x - ray detections of wr 16 and wr 78 conform to and strengthen a previously noted trend of increasing l@xmath62 with wind luminosity l@xmath2 = @xmath3@xmath4v@xmath5 in wn stars .
a regression fit based on existing data for 11 wn stars gives l@xmath0 @xmath43 l@xmath50 and a correlation test yields a correlation probability p@xmath48 = 0.93 .
this trend needs to be confirmed in a larger sample of wn stars but tentatively suggests that wind kinetic energy may be an important factor in governing the x - ray luminosity levels of wn stars .
this work was supported by nasa / gsfc award nnx09ar25 g .
this work was based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and the usa ( nasa ) .
sz acknowledges financial support from bulgarian national science fund grant do-02 - 85 .
lll date & 2009 dec .
28 - 29 & 2010 mar . 2 - 3 obsid & 0602020301 & 0602020201 start time ( utc ) & 19:57:31 & 14:08:02 stop time ( utc ) & 06:35:01 & 00:45:08 optical filter & medium & thick pn exposure ( ks ) & 30.0 & 26.3 mos exposure ( ks ) & 33.0 & 26.5 llccccc wr 16 & wn8h & 3.60 ( 2.37 - 4.36 ) & 1.67 & 670 ( 630 - 740 ) & @xmath454.35 ( @xmath454.55 - @xmath454.20 ) & 5.8 ( 5.50 - 6.15 ) wr 78 & wn7h & 2.00 ( 1.58 - 2.00 ) & 1.55 & 1375 ( 1365 - 1385 ) & @xmath454.25 ( @xmath454.42 - @xmath454.10 ) & 5.9 ( 5.63 - 6.20 ) wr 79a & wn9 ha & 1.99 & 1.28 & 955 & @xmath454.60 ( @xmath454.57 - @xmath454.62 ) & 5.78 wr 16 & 09 54 52.94 & @xmath4557 43 39.2 & 31@xmath206 & [email protected] & ... & ... & 31.14 & gsc [email protected] ( 0.96 ) + wr 78 & 16 52 19.18 & @xmath4541 51 15.8 & 800@xmath2028 & [email protected] & 2.47 & 0.12 ( 0.97 ) & 32.84 & gsc [email protected] ( 0.93 ) + wr 79a & 16 54 58.52 & @xmath4541 09 02.9 & 805@xmath2029 & [email protected] & 1.35 & 0.63 ( 0.33 ) & 32.14 & gsc [email protected] ( 0.30 ) + ll model & n@xmath36*kt@xmath23 @xmath14 n@xmath37*kt@xmath24 abundances & wn n@xmath36 ( 10@xmath39 @xmath28 ) & 1.38 [ 1.20 - 1.60 ] kt@xmath23 ( kev ) & 0.64 [ 0.48 - 0.88 ] norm@xmath23 ( 10@xmath63 ) & 3.59 [ 2.40 - 6.30 ] n@xmath37 ( 10@xmath39 @xmath28 ) & 6.55 [ 4.10 - 10.7 ] kt@xmath24 ( kev ) & 2.27 [ 1.63 - 2.90 ] norm@xmath24 ( 10@xmath63 ) & 14.5 [ 8.30 - 34.4 ] @xmath33/dof & 49.2/51 @xmath64 & 0.96 f@xmath21 ( 10@xmath65 ergs @xmath28 s@xmath7 ) & 2.15 ( 14.5 ) f@xmath66 ( 10@xmath65 ergs @xmath28 s@xmath7 ) & 0.33 ( 4.27 ) log l@xmath21 ( ergs s@xmath7 ) & 32.84 log [ l@xmath21/l@xmath51 ] & @xmath456.64 log [ l@xmath21/l@xmath2 ] & @xmath454.68 abbott , d.c . ,
bieging , j.h . ,
churchwell , e. , & torres , a.v .
1986 , , 303 , 239 anders , e. , & ebihara , m. 1982 , , 46 , 2363 anders , e. , & grevesse , n. 1989 , , 53 , 197 antokhin , i. , bertrand , j .- f . ,
lamontagne , r. , & moffat , a.f.j .
1995 , , 109 , 817 babel , j. & montmerle , t. , 1997a , , 323 , 121 babel , j. & montmerle , t. , 1997b , , 485 , l29 berghfer , t.w . ,
schmitt , j.h.m.m .
, danner , r. , & cassinelli , j.p . , 1997 , , 322 , 167 bratschi , p. & blecha , a. 1996 , , 313 , 537 cassinelli , j.p . , miller , n.a . ,
waldron , w.a . ,
macfarlane , j.j .
, & cohen , d.h . , 2001 , , 554 , l55 chen , w. & white , r.l . 1991 , , 366 , 512 crowther , p.a . 2007 , , 45 , 177 crowther , p.a .
& bohannan , b. 1997 , , 317 , 532 crowther , p.a . , smith , l.j . , hillier , d.j .
, & schmutz , w. 1995 , , 293 , 427 dougherty , s.m . & williams , p.m. 2000 , , 319 , 1005 feldmeier , a. , kudritzki .
p . , palsa , r. pauldrach , a.w.a . , & puls , j. , 1997 , , 320 , 899 gagn , m. et al .
2005 , , 628 , 986 gayley , k.g . &
owocki , s.p . , 1995 , , 446 , 801 gorenstein , p. 1975
, , 198 , 95 gosset , e. , naz , y. , claeskens , j .- f . ,
rauw , g. , vreux , j .-
, & sana , h. 2005 , , 429 , 685 hamann , w .- r . ,
grfener , g. & liermann , a. 2006 , , 457 , 1015 howarth , i.d . & schmutz , w. 1992 , , 261 , 503 ignace , r. , cassinelli , j.p .
, quigley , m. , & babler , b. 2001 , , 558 , 771 ignace , r. , oskinova , l.m . , & brown , j.c .
2003a , , 408 , 353 ignace , r. , quigley , m.f . ,
& cassinelli , j.p .
2003b , , 596 , 538 kahn , s.m .
et al . , 2001 , , 365 , l312 lucy , l.b . , 1982 , , 255 , 286 lucy , l.b
& white , r.l . , 1980 , , 241 , 300 luo , d. , mccray , r. , & maclow , m - m . , 1990 , , 362 , 267 manfroid , j. , gosset , e. , & vreux , j.m . 1987 , , 185 , l7 marston , a.p . , welzmiller , j. , bransford , m.a . , black ,
bergman , p. 1999 , , 518 , 769 moffat , a.f.j .
& niemela , v.s . 1982 , , 108 , 326 morrison , r. & mccammon , d. 1983 , , 270 , 119 naz , y. 2009 , , 506 , 1055 nugis , t. & lamers , h.j.g.l.m .
2000 , , 360 , 227 owocki , s.p . , sundqvist , j. , cohen , d. , & gayley , k. 2011 , in four decades of research on massive stars , eds .
l. drissen , c. robert , & n. st .- louis ( san francisco : asp ) , in press pollock , a.m.t .
1987 , , 320 , 283 pollock , a.m.t . ,
corcoran , m.f . , stevens , i.r . , & williams , p.m. 2005 , , 629 , 482 prilutskii , o.f . &
usov , v.v .
1976 , , 20 , 2 sanders , w.t . ,
cassinelli , j.p . , myers , r.v . ,
& van der hucht , k.a .
1985 , , 288 , 756 schild , h. et al .
2004 , , 422 , 177 seward , f.d . , forman , w.r . ,
giacconi , r. , griffiths , r.e .
, harnden , f.r . ,
jones , c. , & pye , j.p .
1979 , , 234 , 55 skinner , s.l . ,
gdel , m. , schmutz , w. , & stevens , i.r .
, 2001 , , 558 , l113 skinner , s.l .
, gdel , m. , schmutz , w. , & zhekov , s. 2006 , astrophys .
space sci .
, 304 , 97 skinner , s.l . , itoh , m. , nagase , f. , & zhekov , s.a .
1999 , , 524 , 394 skinner , s.l . ,
sokal , k.r . ,
cohen , d.h . ,
gagn , m. , owocki , s.p . , & townsend , r.h .
2008 , , 683 , 796 skinner , s.l . ,
zhekov , s. , gdel , m. , & schmutz , w. 2002 , , 572 , 477 skinner , s.l . ,
zhekov , s. , gdel , m. , & schmutz , w. , & sokal , k. 2010 , , 139 , 825 ( s10 ) sokal , k.r . , skinner , s.l . ,
zhekov , s.a . ,
gdel , m. , & schmutz , w. 2010 , , 715 , 1327 stevens , i.r . ,
blondin , j.m . , &
pollock , a.m.t . , 1992 , , 386 , 265 strder , l. et al . , 2001 , , 365 , l18 turner , m.j.l . et al .
, 2001 , , 365 , l27 ud - doula , a. & owocki , s.p . , 2002 , , 576 , 413 usov , v.v . 1992 , , 389 , 635 van der hucht , k.a . 2001 , new ast . rev .
, 45 , 135 van der hucht , k.a . ,
cassinelli , j.p . , & williams , p.m. , 1986 , , 168 , 111 ( vcw86 ) vreux , j.m . ,
magain , p. , manfroid , j. , & scuflaire , r. 1987 , , 180 , l17 vuong , m.h . ,
montmerle , t. , grosso , n. , feigelson , e.d . , verstraete , l. , & ozawa , h. 2003 , , 408 , 581 waldron , w.l .
& cassinelli , j.p .
2007 , , 668 , 456 wessolowski , u. 1996 , in rntgenstrahlung from the universe ( mpe report 263 ) , eds . h.u .
zimmerman , j.e .
trmper , & h. york ( garching : max - planck - inst .
fr extraterrestrische physik ) , 75 zhekov , s.a .
& park , s. 2010 , , 721 , 518 | previous studies have demonstrated that putatively single nitrogen - type wolf - rayet stars ( wn stars ) without known companions are x - ray sources .
however , almost all wn star x - ray detections so far have been of earlier wn2 - wn6 spectral subtypes . later wn7 - wn9 subtypes ( also known as wnl stars ) have proved more difficult to detect , an important exception being wr 79a ( wn9 ha ) .
we present here new x - ray detections of the wnl stars wr 16 ( wn8h ) and wr 78 ( wn7h ) .
these new results , when combined with previous detections , demonstrate that x - ray emission is present in wn stars across the full range of spectral types , including later wnl stars
. the two wn8 stars observed to date ( wr 16 and wr 40 ) show unusually low x - ray luminosities ( l@xmath0 ) compared to other wn stars , and it is noteworthy that they also have the lowest terminal wind speeds ( v@xmath1 ) .
existing x - ray detections of about a dozen wn stars reveal a trend of increasing l@xmath0 with wind luminosity l@xmath2 = @xmath3@xmath4v@xmath5 , suggesting that wind kinetic energy may play a key role in establishing x - ray luminosity levels in wn stars . |
quantum teleportation @xcite is the most fundamental protocol in quantum information science , and indeed has always played a crucial role in the progress of quantum information theory and technology ( for a good review , see @xcite ) .
the standard teleportation scheme ( sts ) @xcite transfers an unknown quantum state from alice to bob as follows : alice performs a joint measurement on the state to be teleported and half of the previously shared entangled state , tells the outcome of the measurement to bob , and bob applies a unitary transformation , depending on the outcome , to the remaining half of the entangled state . to ensure no - signaling ( no faster than light communication ) ,
every teleportation scheme must accompany some sort of communication and bob s operation depending on the communication content . in so - called port - based teleportation ( pbt )
@xcite , bob s operation is quite simple : he regards his half of ( large ) entangled state as a collection of @xmath0 output ports , and he only picks up an output port ( and discards all the other ports ) .
the output port contains the teleported state as it is , without any correcting operation on the output port .
the absence of the correcting operation leads to an application of pbt as a universal programmable quantum processor @xcite , which is a device to play back the record of the ( past ) experiences of a quantum object .
the universality of the device is so powerful that arbitrary experiences can be recoded and played back ( just like a machine in science fiction to relive your childhood ) , including not only those described as unitary evolution but also measurements ( working as a quantum multimeter @xcite in this case ) .
the drawback is that huge amount of entanglement is necessary to increase the fidelity or success probability .
it has been shown , however , that the most of the huge amount of entanglement can be recycled for subsequent pbt @xcite .
moreover , it has been shown in @xcite that a combined protocol of pbt and sts works as if it could break the barrier of spacetime as follows : suppose that alice and bob , separated in spacetime , each has a quantum system @xmath1 and @xmath2 , respectively .
alice can then consider , _ without waiting for communication _ , that a port of her half of the shared entangled state contains the state of @xmath3 , i.e. she already has the non - local state of @xmath3 somewhere in her hand ( though she can know which port contains the state only after the communication from bob ) @xcite .
this technique is used for attacking position - based cryptography and for instantaneous non - local quantum computation @xcite .
moreover , pbt has been used as a tool to investigate the relation of quantum communication complexity and the bell non - locality @xcite .
however , the properties of pbt have not been completely clarified yet , in particular , for teleporting a high dimensional quantum state .
this is because , in contrast to sts , the simple multiple use of pbt for a qubit ( quantum bit ) does not result in pbt for higher dimension . for teleporting a state of a qudit ( @xmath4-dimensional system ) , only a lower bound of the teleportation fidelity of deterministic pbt @xcite and an upper bound of the success probability of probabilistic pbt @xcite
have been obtained so far .
more studies will be necessary to clarify the properties of pbt . in this paper , we make some remarks on pbt . after recalling the formulation of pbt in sec .
[ sec : formulation of pbt ] , in sec .
[ sec : remarks for @xmath5 , we pay attention to the fact that , in most cases of @xmath6 , the optimal measurements of alice agree with each other . in sec . [
sec : recoverable pbt ] , we propose a hybrid protocol between pbt and sts ( say recoverable pbt ) , where bob has another choice ( in addition to adopt usual pbt ) to adopt a faithful teleportation by utilizing all the @xmath0 output ports . in sec .
[ sec : rederivationo of probability bound ] , we consider the setting of the port - based superdense coding , a dual protocol to pbt , and rederive the upper bound of success probability of probabilistic pbt
. this bound is tight even for @xmath7 and @xmath8 . in sec .
[ sec : fidelity bound due to monogamy ] , we obtain an upper bound of the teleportation fidelity by using the entanglement monogamy relation in asymmetric @xmath9 universal cloning . in sec .
[ sec : port - based superdense coding ] , we finally remak that the superdense coding capacity can be asymptotically achieved in a limit different from the fidelity , and hence port - based superdense coding is possible .
a summary is given in sec .
[ sec : summary ] .
to begin with , let us recall the formulation of ( deterministic ) pbt , where bob has @xmath0 output ports and a teleported state appears in one of the @xmath0 ports without any correcting operation on each port . as a preparation of pbt , bob has @xmath0 qudits : @xmath10 , @xmath11 , @xmath12 , @xmath13 , where each corresponds to the output port of pbt . in this paper , @xmath10 , @xmath12 , @xmath13
are denoted by @xmath14 as a whole .
alice also has @xmath0 qudits : @xmath15 , @xmath16 , @xmath12 , @xmath17 , which are denoted by @xmath18 as a whole .
let us then describe an entangled state between @xmath18 and @xmath14 used for pbt as @xmath19-spins ( @xmath20 ) , and @xmath4 and @xmath21 will be used interchangeably .
the spin basis is denoted by @xmath22 ( @xmath23 , @xmath12 , @xmath24 ) .
then , @xmath25 is a state of spin 0 in two @xmath24-spins , which is maximally entangled between the two .
the operator @xmath26 specifies the actual form of @xmath27 , and @xmath28 so that @xmath27 is normalized .
note that , in pbt , the teleportation fidelity is maximized when @xmath29 in general , i.e. when @xmath27 is not maximally entangled . to teleport the state of the @xmath30 qudit , alice performs a joint measurement with @xmath0 possible outcomes ( @xmath31 , @xmath32 , @xmath12 , @xmath0 ) on the @xmath18 and @xmath30 qudits .
let us denote the positive operator valued measure ( povm ) of her measurement by @xmath33 and hence @xmath34 . when alice obtains the outcome @xmath35 , the state of @xmath36 qudit is close to the state of the @xmath30 qudit as it is .
it is then found that the entanglement fidelity of pbt is given by @xmath37,\ ] ] where @xmath38 @xmath39 , and @xmath40 denotes the @xmath18 qudits except for @xmath41 ( i.e. @xmath42 ) . in analyzing the properties of pbt , the following operator : @xmath43 frequently plays a crucial role . indeed , alice s optimal measurement in the case of @xmath6 and @xmath44 is the square - root measurement ( srm ) @xcite [ also known as a pretty good measurement ( pgm ) or least - squares measurement ( lsm ) @xcite ] for distinguishing the quantum signals @xmath45 , and the corresponding entanglement fidelity is given by @xmath46 to investigate the general properties of @xmath47 , let us decompose @xmath48 into the spin components : @xmath49 where @xmath50 is an identity on the subspace where the total spin angular momentum of @xmath51 is @xmath52 ( @xmath53 is the possible minimum value ) .
then , @xmath47 is also decomposed into @xmath54 since the addition of spin 0 and spin @xmath52 results in spin @xmath52 only , each term @xmath55 is an operator on the subspace of total spin @xmath52 ( though the spin function constructed by the addition is different , depending on @xmath35 ) , and hence @xmath56 is also an operator on the subspace of total spin @xmath52 .
therefore , @xmath57 for @xmath58 , i.e. @xmath47 is block diagonal with respect to the total spin angular momentum ( and clearly its @xmath59-component also ) of @xmath60 spins ( @xmath61 ) , but the maximum momentum is limited to @xmath62 . as far as we know ,
the eigenvalues and eigenstates of @xmath47 for @xmath63 have not been obtained yet , which leads to the difficulty in analyzing pbt in higher dimension .
the only exception is the maximum eigenvalue , which is proved in appendix [ sec : maximum eigenvalue of rho ] to be @xmath64 this leads to a known monogamy relation of singlet fraction @xcite for a multipartite state @xmath65 such that @xmath66 which explains the fidelity limit of symmetric @xmath67 universal cloning @xcite in a simpler way .
fortunately , all the eigenvalues and eigenstates of @xmath47 for @xmath6 can be analytically obtained as we showed in @xcite , but the results are not simple . therefore , it may be worth to summarize it again in a tractable way .
a special property held for @xmath6 is that the spin angular mometum of the @xmath18 spins ( denoted by @xmath68 hereafter ) is a good quantum number , i.e. @xmath47 is block diagonal with respect to @xmath68 also .
indeed , @xmath56 for @xmath6 is written as @xmath69 where @xmath70 is an identity on the subspace where total spin angular momentum of @xmath61 is @xmath52 and @xmath71 .
note that , since the total spin of @xmath52 is the result of the addition of @xmath68 and @xmath72-spin of @xmath30 , @xmath73 holds .
when @xmath27 used for pbt is fixed to a maximally entangled state ( @xmath44 ) , the optimal measurement of alice to provide the maximum entanglement fidelity is srm , i.e. the povm elements are given by @xmath74 here , it is implicitly assumed that the excess term @xmath75 is added to every @xmath76 so that @xmath77 . since @xmath52 in @xmath56 takes a value only for @xmath78 as mentioned before , we have @xmath79 where @xmath80 is an identity on the support of @xmath47 . the corresponding entanglement fidelity can be more increased by optimizing @xmath26 .
the optimization result is @xmath81}(j)(n+2)}}\sin\frac{\pi(2j+1)}{n+2}\openone(j)_a , \label{eq : deterministic optimal o}\ ] ] where @xmath82}(j)$ ] is the number of states with total spin @xmath52 in @xmath0 @xmath72-spins and hence @xmath82}(j)=(2j+1)^2\binom{n}{n/2+j}$ ] .
we showed in @xcite the corresponding optimal measurement of alice in the form of @xmath83 .
however , when the actual povm elements @xmath76 are derived from @xmath84 , it will be found that those agree with eq .
( [ eq : srm ] ) . note that , since @xmath47 is block diagonal with respect to @xmath68 and the optimal @xmath26 is an identity on each subspace
, we have @xmath85=0 $ ] and as a result the measurement can also be considered as srm for distinguishing @xmath86 . in the probabilistic version of pbt , the optimal @xmath26 which provides the maximum success probability of faithful teleportation @xmath87
is give by @xmath88}(j)}}\openone(j)_a .
\label{eq : probabilistic optimal o}\ ] ] we showed in @xcite the corresponding optimal measurement in the form of @xmath89 , but it will be found that the actual povm elements again agree with eq .
( [ eq : srm ] ) without the implicit excess term of @xmath75 . in this probabilistic case , @xmath90 of eq .
( [ eq : srm0 ] ) by itself constitutes a povm element , such that @xmath90 indicates the failure of faithful teleportation . in this way
, the optimal measurement of alice for @xmath6 is given by eq .
( [ eq : srm ] ) in many cases : both @xmath44 and optimal @xmath26 in the deterministic version , and optimal @xmath26 in the probabilistic version .
the only exception is the case of @xmath44 in the probabilistic version .
this seems to rely on the property that @xmath47 is block diagonal with respect to @xmath68 .
unfortunately , this property does not hold for general @xmath4 as shown in appendix [ sec : s=1 ] , where the result of @xmath7 and @xmath8 is explicitly shown .
it is quit interesting that , even in this case , the optimal measurement of probabilistic pbt again agrees with srm for distinguishing @xmath45 .
according to the no - go theorem for the faithful and deterministic universal programmable processor @xcite , a deterministic pbt protocol is inevitably forced to be an approximate one for finite @xmath0 .
therefore , it may be convenient if , for the same measurement of alice , bob can lately choose between two choices : ( 1 ) usual pbt ( with non - unit fidelity ) by selecting one of the @xmath0 output ports or ( 2 ) faithful teleportation ( with unit fidelity ) by utilizing all the @xmath0 output ports .
this protocol , say recoverable pbt , is indeed possible as shown below . to this end , let us consider the optimal probabilistic pbt for @xmath6 to teleport the @xmath30 qubit of @xmath91 .
when alice obtains the outcome @xmath76 with @xmath92 in her measurement , the state of the @xmath30 qubit is faithfully teleported to the @xmath36 qubit , and hence the resulting state is @xmath93 . when alice obtains @xmath90 that indicates the failure of faithful teleportation , the state of @xmath94 is give by @xmath95 where we used eq .
( [ eq : srm0 ] ) and ( [ eq : probabilistic optimal o ] ) , and @xmath96 now , suppose that , after alice obtains @xmath90 , she further measures the @xmath59-component of the total spin of @xmath61 to determine @xmath97 in eq .
( [ eq : state of bd ] ) .
when she obtains @xmath97 , the state of @xmath94 becomes proportional to @xmath98 , which is not maximally entangled between @xmath99 and @xmath14 unless @xmath100 from eq .
( [ eq : old basis ] ) . in this case , the initial entanglement of @xmath30 is not transfered to @xmath14 and bob can not recover the lost entanglement anymore . instead of this measurement , therefore , suppose that alice performs the measurement in the basis of @xmath101 it is not difficult to see that , when @xmath0 is odd and thus @xmath97 is an integer , all the above states are maximally entangled between @xmath99 and @xmath14 , and hence the entanglement of @xmath30 is completely transfered to @xmath14 in this case .
this implies that , if bob knows the outcome of the measurement [ denoted by @xmath102 and he applies an appropriate unitary transformation on @xmath14 according to the outcome , he can completely recover the state of @xmath30 in his hand .
note that , this does not work well for even @xmath0 , because @xmath103 is not maximally entangled for @xmath104 .
to summarize , the explicit protocol of recoverable pbt is as follows : alice performs a measurement @xmath105 on @xmath61 as in the case of optimal probabilistic pbt and obtains the outcome @xmath35 . when she obtains @xmath106 , she further performs the measurement in the basis of eq .
( [ eq : new basis ] ) and obtains @xmath107 .
she then send the outcome @xmath35 and @xmath107 to bob . for @xmath92 , the state of the @xmath30 qubit
is faithfully teleported to the @xmath36 qubit . for @xmath108 , bob has two choices .
if he ignores @xmath107 and randomly picks up one of the @xmath14 qubits as an output port , the protocol works as deterministic pbt .
the entanglement fidelity is equal to the probability of obtaining @xmath92 , because @xmath109 , and hence @xmath110 .
if bob utilizes the information of @xmath107 to apply an appropriate unitary transformation to the whole of the @xmath14 qubits , he can obtain the state of the @xmath30 qubit faithfully .
the recoverable pbt is considered to be a hybrid of pbt and sts .
indeed , the protocol completely agrees with sts for @xmath111 , where @xmath112 and @xmath113 in the standard notation of the bell basis .
[ 0.4 ] qudit ( the case of @xmath114 is shown in this figure ) of previously shared @xmath27 to alice , and alice performs an joint measurement on @xmath115 to know the value of @xmath116 .
, title="fig : " ]
it has been shown that the success probability of probabilistic pbt for any @xmath4 is upper bounded by @xcite @xmath117 it seems very plausible that this bound is indeed reachable , because the bound agrees with the optimal probability for @xmath6 with any @xmath0 , and even for the case of @xmath7 with @xmath8 , where @xmath47 is _ not _ block diagonal with respect to @xmath68 , as shown in appendix [ sec : s=1 ] . here
, we rederive the bound in a way different from @xcite , which is convenient for the later discussions . to this end , let us consider the setting of port - based superdense coding as shown in fig .
[ fig : pbsdc ] , where alice and bob previously share @xmath27 , bob sends the @xmath118 qudit to alice , and alice performs a measurement on @xmath115 to know the actual value of @xmath116 .
note that the roles of alice and bob are opposite to the usual setting of superdense coding , and note that @xmath27 is not necessarily a maximally entangled state .
suppose that , to know @xmath116 , alice performs the same measurement as probabilistic pbt , whose povm elements are @xmath119 , and let @xmath120 be the probability that alice obtains outcome @xmath35 ( @xmath121 ) when bob sent @xmath118 to alice . since the state that alice measures is obtained by projecting @xmath122 to @xmath123 , we have @xmath124p^-_{db_k}.\end{aligned}\ ] ] here , @xmath125 is nothing but the post - measurement state in probabilistic pbt to teleport the half of @xmath91 , and hence equal to @xmath126 , where @xmath127 is a normalized state and @xmath128 is the probability that alice obtains the outcome @xmath35 in pbt .
we then have @xmath129 the success probability of pbt is given by @xmath130 . since @xmath131 and @xmath132 we
obtain the bound of eq .
( [ eq : probability bound ] ) .
note that , in this derivation , we only used the fact that the state of @xmath30 is faithfully teleported to @xmath36 in probabilistic pbt . note further that @xmath133 must hold so that the bound of eq .
( [ eq : probability bound ] ) is tight .
in the same setting as fig . [ fig : pbsdc ] , let us now suppose that alice performs the same measurement as deterministic pbt .
the post - measurement state , denoted by @xmath134 hereafter , is close to but not equal to @xmath93 .
then , we have @xmath135 where @xmath136 is the entanglement fidelity with respect to @xmath122 when the state of @xmath30 is teleported to @xmath36 . for
the sake of simplicity , let us consider the symmetric case such that @xmath137 , @xmath138 ( irrespective of @xmath116 ) , and @xmath139 for @xmath140 , as this permutation symmetry generally holds in pbt .
namely , @xmath141 stands for the ( usual ) entanglement fidelity of the correct output port , and @xmath142 stands for the fidelity of the other output port .
then , from the condition @xmath143 in this deterministic case , we have @xmath144 this equality already implies that faithful and deterministic pbt is impossible for finite @xmath0 . indeed , when @xmath145 for the output port @xmath36 , the reduced post - measurement state for the other port @xmath118 must have the form of @xmath146 , and hence @xmath147 , but those @xmath141 and @xmath142 can not satisfy eq .
( [ eq : singlet equality ] ) for finite @xmath0 . in this way ,
( [ eq : singlet equality ] ) is a constraint on entanglement monogamy in pbt .
note that , for @xmath148 with small @xmath149 , @xmath150 and hence @xmath142 approaches to @xmath151 from below for @xmath152 .
let us then derive the upper bound of @xmath141 from the monogamy relation . to this end
, we regard pbt as a kind of asymmetric @xmath9 universal cloning , and consider the monogamy relation derived in @xcite : @xmath153 where @xmath154 is the fully entangled fraction of the @xmath35-th cloner . the fully entangled fraction is obtained by maximizing singlet fraction among local unitary transformations such as @xmath155 . when the following twirling operation is applied to the post - measurement state @xmath134 : @xmath156 the resulting reduced states @xmath157 are all isotropic states , whose fully entangled fraction has been obtained in @xcite .
we then have @xmath158 for the output port @xmath36 and @xmath159+{\cal o}(\epsilon / n)$ ] for the other output port @xmath118 because @xmath160 @xcite .
namely , the fully entangled fraction of @xmath161 can take this value , at least .
putting @xmath162 and @xmath163 into eq .
( [ eq : singlet monogamy ] ) , we obtain @xmath164 in this way , the monogamy relation in asymmetric universal cloning bounds the entanglement fidelity of pbt from above by @xmath165 .
note that this bound is tight ( leaving for the coefficient ) for @xmath6 , where @xmath166 @xcite .
superdense coding is a protocol dual to quantum teleportation , where the classical information capacity of @xmath167 bits is achieved per qudit sent from bob to alice . in this section
, we remark that the capacity @xmath167 bits can be asymptotically achieved , i.e. port - based superdense coding is possible in the setting of fig .
[ fig : pbsdc ] . when bob sends @xmath118 qudit to alice , the probability that alice can obtain the outcome @xmath35 by the same measurement as deterministic pbt is given by eq .
( [ eq : probablity from fidelity ] ) .
the entanglement fidelity employing srm and maximally entangled @xmath27 is lower bounded by @xmath168 @xcite , but this bound has been slighly improved in @xcite as @xmath169 the derivation of this bound using a convenient property of @xmath47 , instead of using @xmath170 , is given in appendix [ sec : derivation of fidelity lower bound ] .
we then have @xmath171 using this no - error probability , the mutual information between bob and alice , which takes maximum for bob s equal prior probability @xmath172 , is @xmath173 at first glance , port - based superdense coding seems impossible because @xmath174 in the limit of @xmath175 , in quite contrast to @xmath176 in the same limit .
however , @xmath177 takes the maximum at @xmath178 $ ] , and therefore with keeping @xmath179 in the limit of @xmath180 , we have @xmath181 in this way , the mutual information asymptotically approaches to the superdense coding capacity , in the limit different from the fidelity of pbt .
although the application of port - based superdense coding is unknown , this may provide an intriguing example to investigate the duality @xcite between teleportation and superdense coding .
in this paper , we first recalled the optimal protocols of pbt for @xmath6 and paid attention to the fact that , in most cases of @xmath6 , the optimal measurements of alice agree with srm for distinguishing @xmath45 .
we showed that , even in the higher dimension of @xmath7 , the optimal measurement of probabilistic pbt for @xmath8 is srm .
it might be conjectured that this holds for any @xmath4 and @xmath0 .
next , we proposed a hybrid protocol between pbt and sts . in this protocol of recoverable pbt , bob has two choices , to adopt pbt with an approximate fidelity by selecting one of @xmath0 output ports , or to adopt faithful teleportation by applying a unitary transformation to all the @xmath0 output ports as sts .
we showed that recoverable pbt is possible at least for @xmath6 and odd @xmath0 .
moreover , we considered the setting of the port - based superdense coding as shown in fig .
[ fig : pbsdc ] , and rederived the upper bound of success probability of probabilistic pbt @xcite .
in the same setting , we obtained a constraint between the entanglement fidelities of the output ports in pbt .
we then regarded pbt as asymmetric @xmath9 universal cloning , and derived the upper bound of the fidelity expected from the entanglement monogamy relation in the asymmetric cloning .
the obtained bound can explain why the entanglement fidelity of pbt is limited to @xmath182 . finally , we remaked that port - based superdense coding is possible .
indeed , the capacity of @xmath167 bits per qudit sent is asymptotically achieved in the limit of @xmath183 with keeping @xmath179 , while @xmath176 in the limit of @xmath184 in pbt .
namely , in spite that port - based superdense coding and pbt are dual to each other , the perfect transmission of classical and quantum information , respectively , is achieved in the different limiting conditions .
this will be a good example to deepen our understanding about the duality between superdense coding and teleportation .
this work was supported by jsps kakenhi grants no .
23246071 and no.24540405 .
let @xmath185 be a spin state of @xmath51 where @xmath186 distinguishes the permutation degeneracy , and @xmath187 by this , the block submatrix of @xmath47 with total spin angular momentum @xmath52 and its @xmath59-component @xmath97 is written as @xmath188 using @xmath189 , i.e. the matrix elements of @xmath190 are given by @xmath191 when the spin state @xmath185 of @xmath51 for @xmath192 is defined such that @xmath193_{a_i\rightarrow a_1},\ ] ] and hence @xmath194 with @xmath195 begin a swap operator , it is found that the matrix elements of @xmath190 are given by @xmath196 then , @xmath190 is a real symmetric matrix because the clebsch - gordan ( cg ) coefficients are all real .
when @xmath197 is an eigenvalue of @xmath190 and the corresponding normalized eigenvector is @xmath198 , it is not difficult to see that @xmath199 is an eigenstate of @xmath200 and the eigenvalue is @xmath201 .
moreover , @xmath202 is not normalized and @xmath203 .
now , let us rewrite @xmath202 as @xmath204 where @xmath205 , and hence @xmath206 and every @xmath207 is normalized .
then , @xmath208_{a_i\rightarrow a_k } \cr & \le & \sum_i d_{i}^2 + \frac{1}{2s+1}\sum_{i\ne k}|d_i d_k| \cr & = & \frac{2s}{2s+1}\sum_i d_{i}^2 + \frac{1}{2s+1}(\sum_{i}|d_i|)^2 \cr & \le & \frac{2s}{2s+1 } + \frac{n}{2s+1 } = \frac{n+d-1}{d},\end{aligned}\ ] ] where the cauchy - schwarz inequality was used in the second inequality . as a result
, it is found that the maximum eigenvalue of @xmath47 is upper bounded by @xmath209 , which is indeed achieved when @xmath210 and every @xmath207 is the same symmetric function , e.g. when @xmath52 takes the maximum spin angular momentum @xmath211 .
in this case , @xmath47 has only one spin component @xmath212 .
let us denote the spin state on @xmath213 by @xmath214 , which is constructed by the addition of @xmath215 ( @xmath216 ) and @xmath31-spin of @xmath30 . using the standard relation of the cg coefficients @xcite , we have @xmath217 and therefore the matrix elements of @xmath218 in the basis of @xmath214 is @xmath219 where the rows and columns are indexed by @xmath220 .
now , by choosing @xmath26 as @xmath221 the matrix elements of @xmath222 in the basis of @xmath214 are @xmath223 because @xmath224
. then , the povm elements of @xmath225 satisfy the constraints of probabilistic pbt @xcite cvas @xmath226 and @xmath227 .
the corresponding success probability is @xmath228 , which agrees with the upper bound of eq .
( [ eq : probability bound ] ) .
it is intriguing that , even in this case , we have from eq .
( [ eq : p- s=1 ] ) , ( [ eq : rho s=1 ] ) and ( [ eq : x s=1 ] ) , @xmath229 and hence the optimal measurement is srm .
note that @xmath85=0 $ ] also holds , and @xmath230 where @xmath231 is the eigenstate with a zero eigenvalue of @xmath218
for the operator @xmath47 , the following convenient relations hold : @xmath232 by using @xmath233 for @xmath234 , and by using @xmath235 defined in appendix [ sec : maximum eigenvalue of rho ] , we have @xmath236|\xi^{(1)}(j , m,\alpha)\rangle \right|^2 \cr & = \frac{na}{d^{2n } } \sum_{jm\alpha}\left| \langle j , m,\alpha|[\frac{3}{2}\openone-\frac{a(n+d^2 - 1)}{2d^{n+1}}\openone]|j , m,\alpha)\rangle \right|^2 \cr & = \frac{na}{d^{n+1}}\left(\frac{3}{2}-\frac{a(n+d^2 - 1)}{2d^{n+1}}\right)^2,\end{aligned}\ ] ] where see eq .
( [ eq : gram matrix ] ) for the second equality .
this lower bound is maximized when @xmath237 , and hence we obtain eq .
( [ eq : fidelity bound ] ) . | port - based teleportation ( pbt ) is a teleportation scheme such that the teleported state appears in one of receiver s multiple output ports without any correcting operation on the output port . in this paper
, we make some remarks on pbt .
those include the possibility of recoverable pbt ( a hybrid protocol between pbt and the standard teleportation scheme ) , the possibility of port - based superdense coding ( a dual protocol to pbt ) , and the fidelily upper bound expected from the entanglement monogamy relation in asymmetric universal cloning . |
the surprising discovery of x - ray emission from ob stars by the satellite in the late 1970s ( harnden 1979 ; and anticipated by uv superionization observations [ cassinelli & olson 1979 ] ) is now a 25-year - old mystery . because magnetic dynamos and the associated coronal x - rays were not believed to exist in hot stars , alternative models have been sought since the discovery of x - rays from hot stars .
most of these center on some type of wind - shock mechanism that taps the copious energy in the massive radiation - driven winds of hot stars ( lucy & solomon 1970 ; castor , abbott , & klein 1975 , hereafter cak ) .
wind - shock models , especially the line - force instability shock model ( owocki , castor , & rybicki 1988 ; feldmeier , puls , & pauldrach 1997 ; feldmeier 1997 ) , have become the favored explanations for the observed ob star x - ray properties . however , various pieces of evidence have recently emerged indicating that magnetic fields exist on some hot stars ( henrichs 2000 ; donati 2001 , 2002 ) . there is also mounting evidence that the standard wind - shock scenario is inadequate for explaining the observed levels of x - ray production in many b stars ( berghfer & schmitt 1994 ; cohen , cassinelli , & macfarlane 1997a ; cohen , cassinelli , & waldron 1997b ) , and perhaps in some o stars as well ( schulz 2001 ; waldron & cassinelli 2001 ; miller 2002 ) , leading to a revival in the idea that some type of dynamo mechanism might be operating on hot stars .
theoretical investigations of magnetized ob stars have followed ( charbonneau & macgregor 2001 ; macgregor & cassinelli 2002 ) , including studies of hybrid wind - magnetic models ( babel & montmerle 1997 ; ud - doula & owocki 2002 ; cassinelli 2002 ) .
discriminating between the two broad pictures of x - ray emission in hot stars ( magnetic coronal vs. wind shock ) has been very difficult during the era of low - resolution x - ray spectroscopy .
but with the launch of in 1999 , a new , high - resolution window has opened , and we can now apply quantitative diagnostics in our attempts to understand the high - energy mechanisms operating in the most massive and luminous stars in our galaxy .
coronal heating mechanisms are complex and , even in our own sun , not fully understood .
their possible application to hot stars is certainly not clear , even in those hot stars for which magnetic fields have been detected .
however , we can assume that any type of coronal x - ray - production mechanism involves the confinement of relatively stationary hot plasma near the photosphere .
this is in contrast to wind shocks , which , whatever their nature ( and many different specific models have been proposed ) , involve low - density hot plasma farther from the photosphere and moving at an appreciable velocity .
this broad dichotomy naturally leads to several possible observational discriminants that might be implemented using high - resolution x - ray spectroscopy .
these include the analysis of line widths ( or even profiles ) , which could reflect doppler broadening from the bulk outflow of winds ( which in normal ob stars have terminal velocities of up to 3000 km s@xmath1 ) and emission - line ratios that are sensitive to density , the local radiation field , or ( unfortunately ) in some cases , both .
there are numerous other spectral diagnostics available , including temperature - sensitive line ratios , spectral signatures of wind absorption , time variability of individual lines , and differential emission measure ( dem ) fitting ( multiline plus continuum global spectral modeling of x - ray emission from plasma with a continuous distribution of temperatures ) .
all of these can be useful for determining the physical properties of the hot plasma on ob stars , and , ultimately for constraining models , but none can provide the same level of direct discrimination between the wind - shock and coronal scenarios as line profiles and ratios .
in this paper we report on the very high signal - to - noise , high - resolution spectrum of the b0.2 v star .
we chose to observe this particular hot star because it is one of the most unusual single hot stars , with indications already from various low - resolution x - ray observations that its x - ray properties are extreme and present a severe test to the standard wind - shock scenario .
as such , it makes an interesting contrast to the five single o stars that have been observed at high resolution with and up to this point .
in fact , some of the unusual properties of ( discussed below in 3 ) have already inspired the development of alternate models of x - ray production , notably including the `` clump infall '' model of howk ( 2000 ) . in this scenario
overdense regions condense out of the wind , and the reduced line force due to the increased optical depth in the clumps results in these clumps falling back toward the star , generating strong shocks via interactions with the outflowing mean wind .
the data presented in this paper were taken between 2000 september 17 and 18 , over two different orbits , with an effective total exposure time of 72 ks .
we used the acis - s / hetgs configuration and report on the medium - energy grating ( meg ) and high - energy grating ( heg ) spectra , although we concentrate primarily on the meg 1st order spectrum , which contains a total of 19348 counts ( while the heg 1st order spectrum has 5230 total counts ) .
we extracted and analyzed the spectra with ciao tools ( version 2.1.2 ) , including the sherpa model - fitting engine .
we used the caldb libraries , version 2.6 , and the osip files from 2000 august to build auxiliary response files and grating response matrices .
we fit line - profile models using sherpa , but astrophysical modeling of the spectral data took place outside of the ciao environment with a custom non - lte spectral synthesis code , described in section 5.2 .
a member of the upper sco - cen association and a morgan and keenan system standard , the b0.2 v star has been the subject of much study over the past several decades . at a distance of @xmath2 pc ( perryman 1997 ) , this @xmath3 star is closer to the earth than any hotter star .
it has an effective temperature of @xmath4 k and @xmath5 ( kilian 1992 ) .
uv observations show evidence for a wind with a terminal velocity of roughly 2000 ( lamers & rogerson 1978 ) .
the mass - loss rate is difficult to determine because of the uncertain ionization correction , but it seems consistent with the theoretically expected value of @xmath6 ( calculations based on abbott 1982 )
. published values based on uv absorption lines ( e.g. , lamers & rogerson 1978 ) and a recently determined upper limit based on ir emission lines ( zaal 1999 ) are roughly half this value .
these wind properties contrast with those of the prototypical early o supergiant , which has a mass - loss rate at least 100 times larger ( because of its significantly larger luminosity ) .
observations of were recently reported by cassinelli ( 2001 ) , and clear signatures of the stellar wind ( broadening and absorption effects ) are seen in those data .
there are several unusual properties of that might be relevant for our attempts to understand its x - ray properties : ( 1 ) it has a quite highly ionized wind , since the star is the coolest to show in its _ copernicus _ spectrum ( lamers & rogerson 1978 ) . ( 2 ) the wind is also stronger than normal for the star s spectral type ( walborn , parker , & nichols 1995 ) .
( 3 ) it has a very low @xmath7 of @xmath8 ( wolff , edwards , & preston 1982 ) .
( 4 ) it shows unusual redshifted absorption of about 200 in some of the higher ionization uv lines ( which also show much more extended blue absorption , as one would expect ; lamers & rogerson 1978 ) .
( 5 ) it is extremely young ; perhaps being only 1 million years on the main sequence ( kilian 1994 ) .
( 6 ) it shows unusually strong photospheric turbulence , as seen in optical and uv absorption lines ( smith & karp 1979 ) and in the ir ( zaal 1999 ) .
in addition to these properties , it is already known from , , and measurements that has an unusually hard x - ray spectrum and a high x - ray luminosity ( cohen 1997a , 1997b ) .
it is this fact , along with the redshifted uv absorption ( item 4 above ) , that prompted howk ( 2000 ) to develop their model of infalling clumps for .
many of these properties have been cited as evidence that may not fit into the usual wind - shock picture of hot star x - ray production , or at least that its wind emission mechanism may be modified by the presence of a magnetic field or perhaps involve some other , less standard , wind - shock model .
we address these issues in detail in the next sections .
the overall quality and richness of the data can be seen in the meg first - order spectrum shown in figure [ fig : spec_main ] .
there are numerous lines present in this high signal - to - noise spectrum , and they are much better separated than in any of the other hot stars thus far observed ( because of the small wind broadening seen in compared to these stars ) . in figure
[ fig : spec_comp ] we show a portion of the meg spectrum , centered on the neon ly@xmath9 line , compared with data from and capella , a typical o supergiant wind source and a late - type coronal source , respectively .
it is immediately obvious that is superficially more similar to capella than to . to quantify the properties of the spectrum , we fitted every strong line with a simple model ( gaussian plus a polynomial for the nearby continuum ) , providing values of the line flux , characteristic width , and central wavelength .
we employed the appropriate spectral response matrices in these fits , so the derived line widths are intrinsic and do _ not _ include instrumental broadening .
when there was sufficient signal , we fit the meg and heg data simultaneously .
these results are summarized in table [ tab : lines ] , with the several strongly blended spectral features listed in table [ tab : blends ] .
the astrophysical analysis of these data is presented in 5 .
time variability properties of are also of interest in discriminating between the two primary x - ray - production mechanisms . in figure
[ fig : lightcurve_main ] we show a lightcurve formed of all the photons in the -1,-2,-3,+3,+2,+1 orders of both the meg and the heg .
the bins are 1000 seconds in length .
the data are not consistent with a constant source ( at @xmath10 % ) , but no large flare - type events , with a sudden increase in flux followed by a gradual decline back to the basal level , are seen .
the standard deviation of the data is slightly less than 10 % of the mean .
only one individual line feature , the oxygen ly@xmath9 line near 19 , shows evidence of significant variability .
its lightcurve is shown in figure [ fig : lightcurve_oxygen ] .
we tested other strong lines , and only the line at 15.01 shows even marginal variability .
the emission lines of are obviously narrow compared to the same lines observed in hot - star wind x - ray sources like , but how intrinsically narrow are they ? might they be no broader than thermal broadening would dictate ? to answer these questions ,
we fitted gaussian line models , convolved with the instrumental response function , so the width of the model gaussian reflects the actual width of the emitted line .
we do not know what actual line profile shape is warranted , but because the gaussian model is simple and relatively general and the observed line broadening is not dramatic , we did not attempt to fit other profile shapes .
the line widths half - width at half maximum ( hwhm)listed in table [ tab : lines ] generally exceed by a factor of about 3 those expected from thermal broadening ( assuming a temperature equal to that at which the emissivity for a given line peaks ) .
in addition , fits we performed to capella and ab dor ( both late - type coronal sources ) using the same procedures found narrower lines ( consistent with thermal broadening ) than we found for . in figure [ fig : nex_width ] we show the line of in the meg , along with an intrinsically narrow line ( gaussian model with @xmath11 ; i.e. a delta function ) convolved with the instrumental response .
the data clearly are broader than the model .
we also note that some of the features in the spectrum are doublets .
we generally fitted these with single - gaussian models . in each case
in which we fitted a doublet with a single - line model , the separation of the two components of the doublet was far less than the typical ( and derived ) line width . thus the derived line widths using this procedure are only marginally elevated because of the blended doublet .
the velocities of most of the lines in the spectrum of are a few hundred ( hwhm ) , which , while greater than the thermal velocity , is significantly less than the wind terminal velocity .
the x - ray line velocities seen in are thus larger than those in coronal sources and smaller than those in wind sources , a trend that is shown in figure [ fig : line_width_comparison ] , and importantly , the velocities derived for the wind of are significantly less than what would be seen if the x - ray - emitting gas had the same velocity distribution as the uv - absorbing gas .
the mean hwhms of the strongest , unblended lines are roughly 300 km s@xmath1 , or 15% of the terminal velocity of the wind of ( see fig .
[ fig : hwhms ] ) .
there is a slight trend of decreasing line width with wavelength , which is the opposite of what is seen in the o stars ( kahn 2001 ; waldron & cassinelli 2001 ; cassinelli 2001 ; miller 2002 ) . in the o stars , the trend of increasing line width with wavelength
can be understood in terms either of temperature stratification ( lower temperature plasma farther out in the wind , traveling at higher velocities ) or an absorption effect ( low energies are preferentially attenuated , so these lines tend to be broader because only photons from the outer wind , where velocities tend to be high , can escape ) .
the lack of such a trend in might be evidence of the lack of any wind continuum absorption effects on the observed spectral lines .
the wind is expected to be optically thin down to @xmath12 near 0.5 kev and effectively completely optically thin at photon energies of 1 kev and above .
the he - like forbidden - to - intercombination line strength ratio is known to be sensitive to the local mean intensity of the uv radiation field ( gabriel & jordan 1969 ; blumenthal , drake , & tucker 1972 ; waldron & cassinelli 2001 ; kahn 2001 ) .
if the photospheric flux is strong enough , then electrons can be radiatively excited out of the metastable upper level of the forbidden line ( @xmath13 ) to the upper level of the intercombination line ( @xmath14 ) , weakening the forbidden line and strengthening the intercombination line .
thus a measurement of the ratio [ using the more standard spectroscopic notation , @xmath15 , potentially yields the radius of the x - ray emitting plasma ( via sensitivity of the mean intensity to the dilution factor ) .
this assumes that we know or can ignore limb darkening and other geometrical effects and that we know the uv luminosity of the star at the relevant wavelengths .
more traditionally , this line ratio has been used as a density diagnostic in coronal sources ( where , because of the low temperature of the stars , very little uv radiation is present ) .
of course , in hot stars , this line ratio is still sensitive to density ( via collisional excitation between the same two levels described above ) .
however , the densities that would be required to explain the small observed values are very large , while the radiation field requirements are much more reasonable . in order to extract information from these line ratios , we model the oxygen , neon , magnesium , and silicon excitation / ionization using nltert , a non - lte radiation transport and statistical equilibrium code ( described in macfarlane 1993 ; macfarlane , cohen , and wang 1994 ) .
we use a kurucz model atmosphere , constrained by observational data in the spacecraft uv ( relevant for o and ne ) , but unconstrained by direct observation in the far - uv ( mg ) and extreme - uv ( euv ) ( si and s ) .
we note that the model atmosphere fluxes are sensitive to the adopted temperature only in the euv , in which a change of 1000 k can lead to about a factor of 2 change in emergent flux .
the he@xmath9 line complexes are shown in figure [ fig : healpha ] .
when the metastable level is neither collisionally nor radiatively destroyed ( i.e. , the `` low - density limit '' ) , the forbidden lines tend to be about a factor of 3 stronger than the intercombination lines .
we can see in the figure that the forbidden line is heavily affected in oxygen and neon , less so in magnesium , and less still in silicon .
the data are marginal , but are shown for the sake of completeness . in figure
[ fig : fir_model_mg ] we show the results of our model calculation for he - like magnesium .
note that we model the effects of both radiation and collisions ( i.e. , showing both the radial sensitivity and the density sensitivity ) in these calculations .
similar calculations were carried out for the he - like features for the other four species .
the results of these calculations and the constraints they imply for the location of the x - ray emitting plasma are summarized in table [ tab : fir ] .
the quoted errors on the determinations of the radii of line formation include statistical errors only .
the uncertainties in the atmosphere model and the associated uv luminosity will certainly affect the results for ( and ) more than for the other three line ratios .
for example , if the flux were actually twice as high as in the model atmosphere , then the lower limit to the radius of line formation would move out from @xmath16 to almost 1.5 .
we also point out that , where other lines were present ( primarily iron lines near the neon complex ) , we fitted them at the same time we fitted the he - like features , and upper limits to the nondetected forbidden lines were determined by fitting a gaussian model with its centroid position fixed at the laboratory wavelength and its width fixed at the value derived for the corresponding he - like resonance line . in general , it can be seen that most of these line ratio diagnostics are consistent with the x - ray emitting plasma being within several stellar radii of the photosphere .
the oxygen and sulfur complexes do not really provide interesting constraints , as they are consistent with the x - ray emitting plasma being either right above the photosphere or far out in the wind flow .
the neon ratio implies that the plasma giving rise to this complex is within a stellar radius of the photosphere , but confusion and blending with nearby iron lines may make this result somewhat less firm than the formal errors imply .
the strongest constraints are provided by the and features .
the ratio requires the x - ray emitting plasma be at least 1.6 above the photosphere ( but no more than 1.9 ) .
the ratio indicates that the hotter plasma giving rise to this feature is closer to the star , a trend that is also seen in the o stars observed with ( waldron & cassinelli 2001 ; cassinelli 2001 ; miller 2002 ) .
finally , we note that the intercombination line fluxes exceed the resonance line fluxes in the case of oxygen and neon . while the intercombination line strength is generally expected to be less than that of the resonance line in collisional plasmas , the observed g ratios ( @xmath17 ) for these complexes are not inconsistent with a collisional plasma , since the enhanced intercombination line flux comes at the expense of the weakened forbidden line ( pradhan & shull 1981 ) .
while the helium - like forbidden - to - intercombination ratios are sensitive to both density ( via the collisional destruction of the @xmath18 level ) and the local radiation field ( through photoexcitation from the same level ) , we have argued in 5.2 that in hot stars like , the ratio is dominated by radiative effects .
this , of course , makes them relatively useless as density diagnostics , although , for he - like ions for which the forbidden line is strong , an upper limit can be placed on the density of the x - ray - emitting plasma ( as well as on the local value of the mean intensity ) .
the he - like complexes we observe imply upper limits of @xmath19 - @xmath20 @xmath21 .
it is possible to determine an upper limit for the density via the 17.10-to-17.05 line ratio .
the upper level of the 17.10 line is metastable , and so this ratio is density sensitive with a critical density of about @xmath22 @xmath21 ( mauche , liedahl , & fournier 2001 ) , but the energy spacing of this deexciting transition is larger than that for the he - like diagnostics , so that photoexcitation requires a photon with a wavelength below 400 .
thus , for all practical purposes , in a b star this diagnostic is _ not _ sensitive to photospheric uv photoexcitation . in the spectrum of
, we see the 17.10-to-17.05 ratio in the low - density limit .
thus there is no evidence for very dense ( @xmath23 @xmath21 ) x - ray emitting plasma on . in fitting the lines in the grating spectrum , we allowed the line centroid wavelength to be a free parameter .
these results , along with line strengths and widths , are listed in table [ tab : lines ] .
although each individual line centroid is roughly consistent with the laboratory value of the wavelength , when taken as a whole , there is the potential to see an aggregate wavelength shift in the ensemble of lines .
we calculated the weighted mean centroid velocity for the roughly two dozen strongest unblended lines and found a value of @xmath24 km s@xmath1 ( see fig . [
fig : centroid ] ) .
the heliocentric correction actually shifts this number to the blue by about @xmath25 km s@xmath1 , making the heliocentric weighted - mean line centroid velocity @xmath26 km s@xmath1 .
thus , there is no direct evidence for redshifted emission that might be expected from the shocking of infalling clumps , as in the howk ( 2000 ) model , which could be associated with the wind component that gives rise to the observed redshifted uv absorption .
we can not , however , rule out redshifted x - ray emission with net velocities of the order of 100 km s@xmath1 .
we are deferring detailed dem modeling of the global hetg spectrum to a future paper .
the three - temperature mekal ( mewe , kaastra , & liedahl 1995 ) collisional equilibrium model that fitted the spectrum of this star ( cohen 1997b ) , does an adequate job of matching the gross properties of the spectrum . of course , with the vast increase in spectral resolution with , this model does not fit in detail .
however , it is clear that a detailed , global fit to the spectrum will not drastically alter the results from the fit to the spectrum .
one question left open by the analysis of the medium - resolution data is the extent to which the high - temperature component of the spectrum is present , in terms of both the overall emission measure of the hot component and the actual maximum plasma temperature . we find no obvious evidence for plasma with temperatures above the 27 mk lower limit to the hottest temperature component that was found in the spectrum , and in fact , our detection of the feature , which is relatively prominent in the spectrum , is quite marginal in the spectrum .
this is perhaps not surprising , however , given the much lower detector effective area available with + acis - s compared to solid - state imaging spectrometer .
we have made a rough assessment of the temperature distribution implied by the h - like to he - like line ratios from o , ne , mg , and si .
we modeled these ratios using the same nltert non - lte ionization / excitation code ( macfarlane 1993 ) that we used to model the he - like _ f / i _ ratios .
the best - fit temperature for each of these ions can be determined from the observed line ratios under the assumption of coronal equilibrium .
we found temperatures ranging from less than 3 mk for oxygen to slightly more than 10 mk for the silicon .
this is in agreement with the expectations from the off - the - shelf collisional equilibrium codes , such as apec ( smith 2001 ) , and it represents a wider temperature range than is seen in the o stars and based on this same diagnostic .
of course , the presence of a wide range of plasma temperatures is not surprising , and by itself is not a strong discriminant among the various physical models of x - ray production in hot stars .
finally , the wavelength region between 10 and 12 contains several lines from high ionization stages of iron , including at 11.02 and 11.74 and and at 10.62 .
the equilibrium temperatures of peak emissivity of these lines are roughly 20 mk ( aped ; smith 2001 ) .
these lines are quite strong in the spectrum of . of the other hot stars
observed thus far with , only shows these features as strongly as does .
thus , there is evidence in the spectrum of for a not - inconsequential amount of very hot plasma , qualitatively confirming the earlier results , and demonstrating a similarity between and that sets both of these young hot stars apart from the more standard o supergiants .
active coronal sources , including pre - main - sequence stars , show frequent and strong flaring activity , often in conjunction with very high plasma temperatures ( e.g. , ab dor : linsky & gagn 2001 ; pre - main - sequence stars : montmerle 2002 ) .
wind - shock sources show very little variability , and whatever variability is seen has been interpreted as being stochastic and reflecting a poisson statistical description of the number of individual sites of x - ray emission within the wind ( oskinova 2001 ) .
variability levels of 10% , as are seen in the spectra of , could then be interpreted as an indication that there are roughly 100 individual sites of x - ray emission at any given time on this star .
of course , arguments about characteristic timescales ( of , e.g. , cooling ) , as well as the evolution of the physical properties of individual emission sites , have a bearing on the details of this interpretation . our detection of variability in the oxygen ly@xmath9 line at 18.97 ,
but in no other line , could indicate that the relatively cooler gas that gives rise to this feature has a more variable emission measure than do other lines , and thus might be present in fewer discrete sites within the x - ray emitting region , whether it is wind shocks or magnetically confined coronal plasma .
it could , on the other hand , be an indication that the level of absorption by the overlying cool wind is variable .
the long - wavelength lines like are the most likely lines to be subject to wind absorption . however , the expectation is that there is very little continuum absorption of the x - rays in a hot star like , which has a relatively low - density wind .
we note that there is enough signal in several other lines ( , , several lines , for example ) that variability at the level seen in the line would be detected if it were present .
we also note that the time dependence of the variability in this line tracks the overall x - ray variability in the spectrum as a whole and also that the even longer wavelength lines of and do not have enough signal for any variability to be detected .
the qualitative impression one gets from comparing the spectrum of to that of other stars is that this early b star has more in common with late - type coronal x - ray sources than with early - type wind sources .
this impression is borne out primarily in the analysis of the line widths .
the emission lines are indeed narrow , although they are not as narrow as the same lines seen in capella and ab dor .
the bulk flow velocities derived from these lines , while supersonic , are only 10% - 20% of the wind terminal velocity .
if embedded in a standard cak wind , the x - ray - emitting regions would have to be within just a few tenths of a stellar radius from the photosphere based on the observed line widths .
however , the he - like @xmath27 ratios put constraints on the proximity of the x - ray emitting plasma to the photosphere .
these results are consistent with a height of 1 - 2 ( @xmath28 ) , and for the plasma must be at least 1.6 above the photosphere .
there is thus a contradiction , that would seem to imply the presence of hot plasma out to a few stellar radii above the surface of , but moving with very small ( but nonzero ) line - of - sight velocity .
this picture is inconsistent with standard wind - shock scenarios ( which seem to apply to o supergiants like ) in which hot plasma is embedded within the massive line - driven wind .
our data show lines far too narrow for this to be the case .
a similar situation is seen in the o9.5 star @xmath29 ori , though the emission lines in that star are significantly broader than what is seen in the spectrum of ( miller 2002 ) .
it is also difficult , however , to reconcile these data with a generic , solar - type picture of magnetic x - ray activity , in which small magnetic loops confine plasma heated by magnetic reconnection and the deposition of mechanical energy . in this picture , the plasma would be closer to the uv - bright photosphere of than can be accounted for by the observed @xmath27 value for magnesium .
furthermore , we see no evidence for coronal x - ray flaring in our data . it is conceivable that relatively stable , large scale magnetic loops could exist on the surface of , with hot plasma confined at the tops of the loops , several stellar radii from the surface , but the standard picture of energetic , relatively small spatial scale magnetic activity , as seen in the sun , is not consistent with these observations of .
one scenario that could explain the two primary results from the analysis of the spectrum of is the `` clump infall '' model ( howk 2000 ) , which was put forward to explain the hard spectrum and the redshifted uv absorption features seen in this star . in this model
it is _ assumed _ that density enhancements ( `` clumps '' ) form in the wind of a hot star .
the authors explored the dynamics of such clumps under the combined influence of radiation driving , gravity , and wind drag .
they found that for the parameters appropriate to , such density enhancements stall at heights of several stellar radii , and fall back toward the star .
a bow shock would develop around each clump , leading to strong shock heating and presumably hard x - ray emission .
the infalling clumps then also explain the observed redshifted uv absorption features .
since the outflowing wind , when it passes through the shock front associated with one of these density enhancements , would be slowed by a factor of 4 , this scenario provides a means for explaining the lack of strong line - broadening seen in our spectrum , and the heights at which these clumps stall are consistent with the he - like ratios .
it should be stressed , however , that howk ( 2000 ) present no specific mechanism for the condensation of these density enhancements out of the wind .
the relative youth of ( kilian 1994 ) has suggested that a fossil magnetic field may be present in this star .
the very low projected rotational velocity is also consistent with magnetic spin - down ( or , of course , with a pole - on viewing angle ) .
recent modeling of radiation - driven winds in the presence of a dipole magnetic field ( babel & montmerle 1997 ; donati 2002 ; ud - doula & owocki 2002 ) has suggested another scenario that may also be relevant for and could explain the formation of density enhancements at several stellar radii in this star s wind .
the magnetically confined wind shock ( mcws ) model predicts the confinement of the stellar wind by a strong enough dipole field and the subsequent shock heating of wind material trapped in the field and forced to collide at the magnetic equator . in the case of ,
a field of less than 100 g is required to provide the necessary confinement ( ud - doula & owocki 2002 ) .
this model has been successfully applied to ( donati 2002 ) , which is a very young o star with x - ray properties similar to those of .
dynamic simulations have shown that plasma in the post - shock region in the mcws model can eventually fall back toward the star when it can no longer be supported by radiation pressure and magnetic tension ( ud - doula & owocki 2002 ) . by coupling this mechanism to the howk ( 2000 ) scenario ,
we may have a self - consistent , physical picture of the strong shock heating of wind plasma , in a nearly stationary equatorial zone , at a height of several stellar radii , with periodic infalling of material .
we await dynamical simulations of this particular star in order to make quantitative comparisons with our observational data .
in addition , synthesis of x - ray line profiles from the dynamical simulations would be very important , since it is not clear if the specific observed small line widths can be produced within this context .
the other properties we derived from the spectrum can perhaps also be explained in this scenario of magnetic wind shock heating and confinement combined with the infall of density enhancements in the wind .
the high plasma temperatures are naturally explained by the large shock velocities inherent in the collision between the fast stellar wind and the equatorial shocked disk ( as well as with the infalling clumps ) .
the lack of very dense plasma argues against very dense coronal - type magnetic structures near the photosphere , although some level and type of coronal heating mechanism is certainly not ruled out .
in fact , recent work ( charbonneau & macgregor 2001 ; macgregor & cassinelli 2002 ) suggests that a dynamo mechanism could be operating in the interiors of normal ob stars and that if a means of bringing the dynamo - generated magnetic fields to the surface could be identified , then solar - type coronal activity on hot stars may be quite plausible .
the modest time variability seen in the spectrum is more significant than that seen in o stars that are supposed to be wind x - ray sources .
this variability level ( rms variations at the 10% level ) may simply be indicative of a relatively small number of sites of x - ray production on as compared with o stars showing a more standard mode of wind - shock x - ray production ( oskinova 2001 ) , and no large flare events were seen in this observation of ( or in any of the earlier x - ray observations of this star ) .
therefore , any coronal model applied to would need to explain the high temperatures ( presumably with a strong dynamo mechanism of some sort ) but also the absence of frequent strong flaring .
in addition , such a coronal mechanism would need to explain the presence of bulk plasma velocities of several hundred km s@xmath1 , along with at least some x - ray - emitting plasma a significant height above the surface of the star . in summary ,
the unusual x - ray properties of , elucidated in great detail for the first time by the superior spectral resolution of the , now require detailed dynamical modeling in order to test the viability of the mcws model , deduce the role of clump infall , and test the alternate scenario of dynamo activity . however
, we can say with certainty that the standard picture of numerous shocks embedded in a fast stellar wind can not explain the narrow x - ray lines and the formation heights of several stellar radii of the very hot x - ray emitting plasma on this young hot star .
we wish to thank eric levy , carolin cardamone , james ciccarelli , and prudence schran for help with the data analysis and marc gagn for fruitful science discussions .
this research was supported in part by nasa grant go0 - 1089a to swarthmore college , prism computational sciences , and the university of wisconsin .
work was performed under the auspices of the us department of energy by the university of california lawrence livermore national laboratory under contract no .
w-7405-eng-48 . &
@xmath30 & & @xmath31 & & & 7.1 + & @xmath32 & & @xmath33 & & & 7.1 + & @xmath34 & & @xmath35 & & & 7.1 + & @xmath36 & @xmath37 & @xmath38 & @xmath39 & @xmath40 & 7.2 + & @xmath41 & @xmath42 & @xmath43 & @xmath44 & @xmath45 & 7.0 + & @xmath46 & @xmath47 & @xmath48 & @xmath49 & @xmath50 & 7.0 + & @xmath51 & @xmath52 & @xmath53 & @xmath54 & @xmath55 & 7.0 + & @xmath56 & @xmath57 & @xmath58 & @xmath59 & @xmath60 & 7.2 + & @xmath61 & @xmath62 & @xmath63 & @xmath64 & @xmath65 & 7.0 + & @xmath66 & @xmath67 & @xmath68 & @xmath69 & @xmath70 & 6.8 + & @xmath71 & @xmath72 & @xmath73 & @xmath74 & @xmath75 & 6.9 + & @xmath76 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & 7.0 + & @xmath81 & @xmath82 & @xmath83 & @xmath84 & @xmath85 & 6.8 + & @xmath86 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & 6.8 + & @xmath91 & @xmath92 & @xmath93 & @xmath94 & @xmath95 & 6.8 + & @xmath96 & @xmath97 & @xmath98 & @xmath99 & @xmath100 & 7.0 + & @xmath101 & @xmath102 & @xmath103 & @xmath104 & @xmath105 & 6.8 + & @xmath106 & @xmath107 & @xmath108 & @xmath109 & @xmath110 & 6.9 + & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & 6.9 + & @xmath116 & @xmath117 & @xmath118 & @xmath119 & @xmath120 & 7.3 + & @xmath121 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & 6.8 + & @xmath126 & @xmath127 & @xmath128 & @xmath129 & @xmath130 & 6.8 + & @xmath131 & @xmath132 & @xmath133 & @xmath134 & @xmath135 & 6.9 + & @xmath136 & @xmath137 & @xmath138 & @xmath139 & @xmath140 & 6.6 + & @xmath141 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & 7.2 + & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & 7.1 + & @xmath151 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & 6.8 + & @xmath156 & @xmath157 & @xmath158 & @xmath159 & @xmath160 & 7.0 + & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & 7.0 + & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 & 7.0 + & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & 6.6 + & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 & 6.9 + & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & 6.6 + & @xmath186 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & 6.9 + & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 & 6.8 + & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & 6.9 + & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & 6.9 + & @xmath206 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & 6.9 + & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & 6.9 + & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & 6.7 + & @xmath221 & @xmath222 & @xmath223 & @xmath224 & @xmath225 & 6.7 + & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & 6.7 + & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & 6.5 + & @xmath236 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & 6.8 + & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & 6.7 + & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & 6.7 + & @xmath251 & @xmath252 & @xmath253 & @xmath254 & @xmath250 & 6.7 + & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & 6.5 + & @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & 6.3 + & @xmath265 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & 6.3 + & @xmath270 & @xmath271 & @xmath272 & @xmath273 & @xmath274 & 6.3 + | long known to be an unusual early - type star by virtue of its hard and strong x - ray emission , @xmath0 scorpii poses a severe challenge to the standard picture of o - star wind - shock x - ray emission .
the spectrum now provides significant direct evidence that this b0.2 star does not fit this standard wind - shock framework .
the many emission lines detected with the gratings are significantly narrower than what would be expected from a star with the known wind properties of , although they are broader than the corresponding lines seen in late - type coronal sources . while line ratios are consistent with the hot plasma on this star being within a few stellar radii of the photosphere , from at least one he - like complex there is evidence that the x - ray emitting plasma is located more than a stellar radius above the photosphere . the spectrum of is harder and more variable than those of other hot stars , with the exception of the young magnetized o star .
we discuss these new results in the context of wind , coronal , and hybrid wind - magnetic models of hot - star x - ray emission . |
at frankfurt university a storage ring for low energy high current density ion beams is proposed @xcite
. a longitudinal magnetic field component will provide a quite homogenous transverse beam focusing conditions along the whole structure .
the main advantage of a stellarator type ring against the conventional one is the higher transverse momentum acceptance which opens a possibility to accumulate high beam currents .
the planned ring comprises of curved sectors with longitudinal magnetic field to form the figure-8 geometry .
the continual and coupled longitudinal magnetic field should provide focusing and guiding forces . in curved magnetic fields ,
charged particles tend to drift leaving the desired trajectories . to compensate the drift arising from curved magnetic fields
a twisted geometry is preferred which looks like a classical stellarator with a figure-8 shape .
high magnetic field of about @xmath2 produced by superconducting coils is desired to minimize drift forces and to achieve high current densities .
an example of ring formed with toroidal segments depicting single magnetic surface .
magnetic field strength ( normalized ) is colour coded .
room temperature prototype segment is shown on the right .
, width=302 ] this ring can be built using multiple toroidal segments . due to the 3-dimensional geometry the magnetic field lines do not close in a single turn around , but tend to form a surface , known as a magnetic flux surface .
the torsion of the field lines , called rotational transform , is assumed to increase the stability for confined charged particles .
one example of such a magnetic surface is shown in fig . [ fig:01 ] .
the room temperature experiments scaled down to @xmath3 were planned to investigate the drift dynamics , space charge effects , beam diagnostics and multiturn beam injection system @xcite .
the properties of prototype magnets are stated in table [ tab : table1 ] .
the measurements from these experiments are compared with the numerical model which will be used to design the injection system .
the dynamics of charged particle beams in magnetic fields is characterized by gyro motion .
the beam size undergoes through periodic maxima and minima in longitudinal magnetic field when injected from field free region . in curved magnetic fields
the ion beam is guided on a circular path , additionally dominated by the drift motion , namely @xmath4 drift due to curved magnetic field lines , @xmath5 drift due to the inhomogeneous field and @xmath6 drift due to the space charge @xcite .
the curvature drift is given by , @xmath7 this equation implies that a particle when injected into the magnetic field will experience a drift force along the normal formed by the radius vector @xmath8 and magnetic field @xmath9 . for a proton beam with an energy of @xmath10 , injected in the magnetic field of @xmath11 and @xmath12 ,
it is @xmath13 .
along with the time of flight one can calculate this drift to be @xmath14 along the beam path over an arc distance of @xmath15 .
the @xmath16 drift arises due to the difference of coil density on the inner side and outer side of the toroid .
the drift velocity due to this inhomogeneous field is given by , @xmath17 where @xmath18 is larmor radius .
in the case of used toroidal segments the magnetic field gradient is @xmath19 . for a proton beam with a transversal velocity spread up to @xmath20 ( corresponding to @xmath21 measured from experiments )
, the drift velocity can be maximum @xmath22 .
an additional drift arises due to the self electric field of the beam .
the crossed terms in the lorentz force equation implies the particles will experience additional drift velocity given by , @xmath23 the homogeneous ion beam has a maximum coulomb force on the boundary .
thus a cross product of radial component of electric field ( @xmath24 ) and longitudinal magnetic field ( @xmath25 ) gives rise to rotation of the beam around its axis . for a proton beam ( @xmath26 ) with current @xmath27 at @xmath10 energy
this drift velocity is @xmath28 .
thus we conclude in our case the curvature drift is the most dominating over the others .
a computer code ( tbt ) was written to simulate the dynamics of particles in given fields .
tbt can simulate a realistic external fields e.g. magnetic field from coil or electric field from electrodes .
the biot - savart law is used to calculate the magnetic field at toroidal grid points from current carrying pancake coils as a source .
the standard toroidal coordinate system ( @xmath29 ) was used @xcite .
a space - charge subroutine uses a particle in cell ( pic ) method in 3-dimensional toroidal geometry @xcite .
a first order charge distribution scheme uses relative toroidal volume as a weighing factor . the potential at grid points
is solved using the poisson equation .
an iterative method bi - conjugate gradient stabilized ( bicgs ) is used to solve resulting two dimensional matrix with @xmath30 linear equations , where @xmath31 s represent the number of grid points in each direction @xcite .
flow chart of the tbt algorithm.,width=302 ] a middle step finite - difference time - domain scheme calculates position advance in the time domain .
the flowchart in fig .
[ fig:02 ] shows the complete algorithm .
for example , consider a parallel beam of proton at an energy of @xmath10 and a nominal beam current of @xmath32 injected into the single toroidal segments .
[ fig:03 ] shows the beam envelope is projected on the @xmath33 and the @xmath34 planes . for a theoretical analysis a homogeneous input distribution of particle with a beam radius of @xmath35 and angle spread of @xmath36
was defined .
the graph on the left shows beam curved with @xmath37 arc degree angle .
and the graph on the right shows beam drifted vertically in the negative x - direction from central axis about @xmath38 .
multiple beam waist can be located along the beam path . along with the time of flight
the number of beam waist during the transport can be calculated .
projections of proton beam envelope with energy @xmath10 transported into the magnetic field of @xmath11 .
a ) projection on the @xmath33 plane showing curved beam ,
b ) projection on the @xmath34 plane showing the beam drift.,width=302 ]
the experimental setup was designed by taking into account three main aspects , ion beam production , matching into toroidal field and diagnostics .
a volume type ion source was constructed to produce light ion species .
the high voltage terminal supports up to @xmath39 .
up to @xmath40 beams can be extracted using triode extractor .
pure @xmath41 beam can be extracted from @xmath42plasma , whereas @xmath43 , @xmath44 and @xmath45 - beams can be formed by a proper choice of @xmath46plasma parameters @xcite .
the ion source was mounted on the vacuum chamber with a faraday cup .
the solenoid was installed downstream for beam matching .
it provides a maximum field of @xmath47 on the axis .
.[tab : table1]experimental parameters [ cols="<,^",options="header " , ] an optical assembly used for the beam diagnostics , consists of a phosphor screen and a digital camera .
it was installed in the flange downstream of toroidal segment with a ring electrode in front of it .
the potential up to @xmath48 can be put to this electrode to attract or repel any secondary electrons .
phosphor screen was composed of @xmath49 , known as @xmath50 , emits a light within a range of @xmath51 and with a peak emission at @xmath52 , a yellow - green colour .
the screen had a diameter of about @xmath53 .
a digital camera produces @xmath54 image in the @xmath55 format .
a separate image manipulation routine was written to analyse the image produced . in the first stage
the dynamics in single toroidal segment was studied .
the optical assembly was fixed downstream of the magnet with a proper magnetic shielding for the camera . a proton beam at the energy of @xmath56 detected downstream of the toroid for three different values of magnetic field .
the measured vertical drifted position as a function of b - field is plotted and compared with the simulation.,width=302 ] the beam drifts were measured with this method as a function of the magnetic field . in the case of proton beam , the beam is composed of three different ion species ( see table [ tab : table1 ] ) , the drifted position of the beam do not follow simple @xmath57 dependence .
[ fig:04 ] shows that results are in the good agreement with simulations . in this case ,
a proton beam at the energy of @xmath56 was injected in the toroidal magnetic field .
the drifted position was calculated with respect to geometrical axis calibrated before measurements .
the electrons are produced along the beam line due to beam loss ( secondary electrons ) or ion collision with rest gas atoms .
it is assumed that some of the electrons are longitudinally confined in the transport channel .
along the beam line , perturbed negative potential from screening electrode of the ion source and the potential due to the repeller electrode when biased negative form a longitudinal trap for electrons .
these electrons can be observed to produce a small spot at the center of the screen ( see fig .
[ fig:05 ] )
. perturbed negative potential from screening electrode of the ion source and the potential due to the repeller electrode biased negative form a longitudinal trap for electrons .
secondary electrons produced at the center of the phosphor screen giving a magnetic center . on the right central electrons can be seen along with a drifted ion beam.,width=302 ] these electrons may have energy in the range of @xmath58 . since electrons have @xmath59 times the mass of proton , experience very low curvature drift less than @xmath60 .
this is practically negligible , giving a magnetic center of the system .
the experimental setup showing ion source , solenoid , toroidal sector magnet and the movable probe.,width=302 ] an innovative `` pigging technique '' was developed for the beam diagnostics .
the optical assembly was upgraded to manoeuvre the optical probe along the beam path .
the phosphor screen and the camera were installed on a movable cylinder ( see fig .
[ fig:06 ] ) .
this assembly is able to sustain a magnetic field up to @xmath11 and high vacuum conditions .
a repeller ring ( @xmath61 ) in front of the screen was fixed .
this detection probe gives an opportunity to detect a transversal beam profile along the longitudinal axis .
transversal beam profile of the ion beam ( @xmath41-beam at @xmath62 ) along the beam path , simulations ( left ) and measurements ( right ) .
the locations of the beam waist are indicated with arrows.,width=302 ] fig .
[ fig:07 ] compares the simulation results with a transversal profile measurement along the longitudinal axis .
@xmath41-beam with energy of @xmath63 was injected into the toroid with a magnetic field of @xmath11 . as the beam
is transported into the segment it shows periodic beam waists .
the number of beam waists along the path can be calculated from beam energy and field strengths .
simulations show a good agreement with measurements with respect to drifted position and number of beam waists .
the errors are caused mostly due to inaccuracy and low resolution in the probe positioning along the longitudinal axis .
when the beam is diverging the spot size is larger than the probe diameter .
the secondary electron production is also high at these positions .
this leads to band like structure in the measurements .
the deviation in the location of the beam waist from the simulation results can also be noticed .
this is presumed due to electron entrapment between solenoid and toroid .
early experiments using langmuir probe have shown the presence of electron cloud with a density of about @xmath64 in this region . the longitudinal focus shift due to a plasma column
is known @xcite .
two segments were coupled in the next stage ( see fig . [ fig:08 ] ) .
the diagnostic probe can be moved along the beam path over the total distance of @xmath65 from position-1 to position-6 as marked in the photograph .
this kind of setup provides a possibility for various experiments not only in terms of parameters but also with respect to the geometrical arrangement of magnet components . in the first step
a `` straight section '' , an intermediate separation with a length of @xmath66 was chosen .
the whole arrangement forms a @xmath67 part of the ring . along the longitudinal axis the magnetic field forms a ripple like structure .
magnetic field along the longitudinal axis.,width=302,height=151 ] the magnetic field drops down to @xmath68 constituting about @xmath69 from maximum field strength of @xmath11 ( see fig . [ fig:09 ] ) .
there have been indications of electron cloud entrapment between positions 3 - 4 due to mirror configuration of the magnetic field , according to the observations from the beam transport experiments .
measurement of proton beam with an energy of @xmath70 at different positions along the beam path when transported through two coupled segments .
, width=302 ] fig .
[ fig:10 ] shows a low energy proton beam ( @xmath70 , @xmath71 ) detected at various positions along the beam path .
position-1 shows clearly defined circular beam spot at the entrance of the first toroid .
it can be seen as quite homogenous beam with a well defined boundary .
position 2 shows a beam that is drifted vertically downward due to the curvature drift . from position 3 to position 4 the magnetic field drops down drastically and the field lines expand in the space .
the ions following the field lines are forced toward the vessel wall .
this leads to beam losses in the straight section .
thus the beam at these positions is not well defined rather influenced by the secondary electrons . at position 5
the beam is recaptured into the acceptance of the second toroid .
while the position 6 shows , the beam detected just at the exit of the second toroid .
overall , vertically drifted position along the beam path ( 1 - 6 ) can be seen due to the curvature drift .
the space charge effects due to electrons on the ion beam will be investigated in detail especially in the ripple region .
a ) transmission along the longitudinal axis simulated and measured ; b ) vertical position of beam plotted along longitudinal path.,width=302,height=302 ] fig .
[ fig:11 ] a ) shows the transmission ( detected by optical screen intensity ) as a function of longitudinal position of the probe .
maximum losses are seen in the straight section due to weak coupling .
[ fig:11 ] b ) shows the vertically drifted position of an ion beam .
the measurement is in reasonable agreement with the analytical value resulting in @xmath72 .
the simulation shows discrepancy near the exit of first toroid .
but the presence of high peak in the straight section and large oscillations in the second toroid can be readily spotted .
the differences between the simulation and measurement are blamed on rotation of the probe , lack of correct positioning system and camera sensitivity . in further experiments
the effect of separation distance between two segments will be investigated .
theoretically zero distance is the best case for beam transport . but some space is required for injection experiment to install second beam line . the best suitable case of @xmath73 as predicted by simulation will be chosen for the final setup . the beam loss is expected to reduce by about factor @xmath74 comparative to the earlier case .
this improvement is sufficient to show the effects of separation .
the drift compensation experiments are planned by installation of second segment in the opposite direction forming an s - shape .
the challenges in injecting charged particles from field free region into the confinement region has been addressed and investigated especially using electron beams @xcite .
the experimental setup has been designed to study injection experiments for such a type of accumulator ring .
two segments will be arranged on circular arc to form a @xmath75 part of storage ring ( see fig .
[ fig:12 ] ) .
the injection line will be installed between two of these sector magnets .
identical ion source and solenoid will be installed to upgrade the existing beam line .
setup for the beam injection experiments.,width=302 ] the beam optics for injection is calculated with the simulation tool tbt .
a special magnetic field combination is designed to inflect the proton beam into the acceptance of toroid-2 .
the 3d - phase space portrait is used to describe beam quality . to extract the information of guidance of beam along magnetic field lines a parameter _ velocity ratio _ ( @xmath76 ) was defined as , @xmath77 where @xmath78 is defined as a velocity component parallel to the magnetic field at a particular position and magnetic field strength distribution .
this ratio is also referred as _ pitch _ in some of the publications .
three dimensional map then can be produced by plotting @xmath76 as a function of relevant space coordinate ( e.g. @xmath34 at the input of the injection system and @xmath34 at the output , the exit of toroid-2 ) .
a kicker system which make use of @xmath6 drift force would move the injected beam onto the main ( ring ) field lines .
the whole geometry is optimized for a proton beam with the energy of @xmath10 . left : @xmath34 positions of the injected beam at the input colour coded with @xmath79 at the output plane representing acceptance of the system .
right : output @xmath80 distribution showing two beams.,width=302 ] fig .
[ fig:13 ] on the left shows the @xmath76 mapping at the input plane @xmath34 i.e at the entrance of the injection coil .
it can be seen that @xmath81 beam can be injected into the second toroid within acceptable pitch limit @xmath82 .
[ fig:13 ] on the right shows the mapping at the exit of toroid-2 .
two beams can be seen .
these two beams were simulated to match conditions at the output .
in this paper we have described the experimental activities under the project of figure-8 storage ring ( f8sr ) " for accumulation of intense low energy ion beams .
the experimental setup is very versatile for different beam transport experiments .
the ion beam transport in toroidal magnetic fields has been successfully demonstrated with _ in situ _ detection method .
further investigations would target to gain more knowledge about beam transport and to overcome the challenges poses by extreme conditions on beam diagnostics .
the main point will be experimental components operating in high magnetic field and high vacuum conditions .
the high magnetic field imposes limitations on electronics used for diagnostics e.g. digital camera , signal cables , inductors .
moreover a magnetic shielding must be designed for any component that operates with rotating metal e.g. rotary pumps .
the future experiment would stress on the optimal beam matching conditions .
the beam losses produce secondary electrons these can be trapped in the magnetic field .
the secondary electrons are observed to produce unwanted background structures influencing the beam signal . in further steps
, the magnetic field geometry will be tested by forming s - shape .
the drifts are expected to compensate due to the reversed direction of radius vector @xmath8 .
the experiments will be concluded by installation of the injection beam line and demonstration of ion beam injection in magnetic fields . | the stellarator - type storage ring for accumulation of multi- ampere proton and ion beams with energies in the range of @xmath0 to @xmath1 is designed at frankfurt university . the main idea for beam confinement with high transversal momentum acceptance
was presented in epac2006 .
this ring is typically suited for experiments in plasma physics and nuclear astrophysics .
the accumulator ring with a closed longitudinal magnetic field is foreseen with a strength up to @xmath2 .
the experiments with two room temperature 30 degree toroids are needed .
the beam transport experiments in toroidal magnetic fields were first described in epac2008 within the framework of a proposed low energy ion storage ring .
the test setup aims on developing a ring injection system with two beam lines representing the main beam line and the injection line .
the primary beam line for the experiments was installed and successfully commissioned in 2009 . a special diagnostics probe for _ `` in situ '' _ ion beam detection was installed .
this modular technique allows online diagnostics of the ion beam along the beam path . in this paper , we present new results on beam transport experiments and discuss transport and transverse beam injection properties of that system . |
the ` breeding method ' is a well - established and computationally inexpensive procedure for generating perturbations for ensemble integrations @xcite .
bred vectors ( bvs ) are finite perturbations periodically rescaled to a certain magnitude that have been prominently used in probabilistic weather forecasting with ensembles @xcite . the breeding method and variants of it
are applied in operative ensemble forecast systems , such as national centers for environmental predictions ( ncep , usa ) , see e.g. @xcite .
moreover , breeding continues to be a popular tool to study the predictability of a variety of systems such as the baroclinic rotating annulus @xcite or the atmosphere of mars @xcite .
different initial bv perturbations all generally tend to become aligned with the fastest growing modes .
if different bvs were globally quasi - orthogonal to each other @xcite , one might expect they would automatically provide a good sample of the different dominant growing error directions , without the need for additional computation .
a closer inspection reveals that the bv perturbations are often locally rather similar in shape , differing only in sign and amplitude @xcite .
in fact , a major modification of the bv implementation at ncep has recently been implemented by replacing the bvs given by the ensemble forecast with some ` ensemble transform ' that orthogonalizes the ensemble with respect to the metric defined by the inverse covariance matrix @xcite .
other metrics can be used and lead to different ensembles of bvs @xcite .
orthogonalization with respect to a given metric generally enhances the statistical diversity of the ensemble by making the bv perturbations globally more dissimilar @xcite . in this paper
we show how the ensemble diversity can be enhanced by using the geometric norm with no further transforms or orthogonalizations needed .
we first show that the bvs dynamics and the statistical properties of the ensemble strongly depend on the norm definition used to construct them .
so far euclidean - type norms are widely used in applications . however , our results demonstrate that , among a spectrum of studied norms , the geometric norm is the most convenient because it provides a greater statistical diversity of the ensemble , while it enhances the projection of the ensemble as a whole on the most unstable direction . with other norm choices , like the standard euclidean one , a good projection on the leading lyapunov vector ( lv )
is always associated with the collapse of all the bvs , i.e. the complete loss of the ensemble diversity .
we illustrate our study with numerical integrations of the well - known lorenz-96 model @xcite that has been used by various authors as a low order testbed for atmospheric prediction and assimilation studies @xcite .
this model is defined by the set of variables @xmath1 and evolves according to @xmath2 \nonumber\\ & - & u(x , t)+
f,\quad \mbox{with}\quad x=1, ... ,l . \label{lorenz}\end{aligned}\ ] ] with periodic boundary conditions in the discrete spatial variable @xmath3 . hereafter we adopt a system size of @xmath4 and a forcing constant @xmath5 .
for these values the system exhibits well developed chaos @xcite .
a good description of the chaotic dynamics can be achieved by understanding the behavior of initial infinitesimal perturbations , which are governed by the ` tangent linear model ' : @xmath6 \nonumber\\ - u(x-1,t)\left[\delta u(x-2,t)- \delta u(x+1,t)\right ] -\delta u(x , t ) .
\label{tangent}\end{aligned}\ ] ] after some transient any infinitesimal perturbation @xmath7 becomes permanently aligned along the most unstable direction
. this direction defines , disregarding an arbitrary nonzero constant factor , the leading lv , and hereafter denoted @xmath8 .
obtaining the tangent linear ( and adjoint ) models can be however extremely difficult in operative weather models and one has to resort to analyzing finite perturbations , which are evolved with the full nonlinear model .
this is for instance the situation at ncep , where ensembles of bvs are used .
bvs are finite perturbations obtained after periodic rescaling , say at times @xmath9 ( @xmath10 ) . a control trajectory @xmath11 and a perturbed one @xmath12 , are simultaneously integrated [ via eq . ( [ lorenz ] ) ] and at the scheduled times the difference between them
is calculated @xmath13 and rescaled to a given amplitude @xmath14 , obtaining the bv @xmath15 this bv is then used to redefine the perturbed system : @xmath16 with @xmath17 .
the perturbed @xmath12 and control @xmath11 states are then evolved in time according to the model equations , eq .
( [ lorenz ] ) , until the next scheduled rescaling . at the next scheduled time
@xmath18 the breeding cycle , eqs .
( [ dif])-([pert ] ) , is repeated .
after several breeding cycles , the perturbations generated by this procedure acquire a large growth rate , which makes them suitable for ensemble forecasting .
usually a set of bvs is evolved from different initial random perturbations and this constitutes the ensemble .
ideally a good ensemble of bvs should span the most unstable directions in phase space well enough to capture the main instabilities .
there are three basic ingredients in the definition of the bv : ( i ) the rescaling interval @xmath19 , ( ii ) the perturbation amplitude @xmath14 , and ( iii ) the choice of the norm @xmath20 used in eq .
( [ scaling ] ) .
the rescaling interval @xmath19 has a negligible influence in the results as long as it remains small say , smaller than the doubling time , which is on the order of @xmath21 time units ( t.u . ) for the lorenz-96 model .
we have used @xmath22 t.u . , which corresponds to @xmath23 day in the time scale assumed by @xcite .
the perturbation amplitude @xmath14 controls the `` finiteness '' of the perturbations ; a sufficiently small @xmath14 makes the perturbation quasi - infinitesimal , and in the limit @xmath24 the bv perfectly aligns with the leading lv of the system .
however , very little is known about the effect of the norm choice on the properties of the resulting ensemble and we discuss this issue in detail in the incoming sections .
the choice of the norm is probably the more obscure element determining the bvs nature .
bvs have often been claimed to be insensitive to the choice of norm @xcite .
however , this belief is not actually based on any rigorous argument . here
we show that the effect of changing the norm type has a dramatic impact on bvs .
we will show that different norms lead to different ensemble properties and it is not a mere change of the ` ruler ' or metrics .
there are intrinsic and genuine effects on the statistics of the bvs for each particular norm type .
intuitively , for a homogeneous system like the lorenz-96 model , any definition for the norm one wants to use should be homogeneous in the sense that it weights equally all sites . to see why this constraint is relevant
let us consider a particularly illustrative example .
think of a norm arising from some scalar product @xmath25 with a very `` unbalanced '' metric matrix @xmath26 , e.g. @xmath27 .
this choice would result in very dissimilar bvs depending if the site @xmath28 is more or less unstable at a given time . for a given @xmath14
, at some times the vector dynamics could be infinitesimal - like while at other moments it would be clearly finite . for spatially homogeneous systems ,
it is reasonable to restrict ourselves to homogeneous norms that produce a bv that is statistically equivalent up to a high degree at different times and we do so in our study . in this work
we compare the performance of @xmath0-norms , which are defined as @xmath29^{1/q}\ ] ] note that for @xmath30 the norm is an energy - like norm , analogous to those used in atmospheric models . in the limit
@xmath31 the @xmath0-norm becomes the supremum norm : @xmath32 moreover , the geometric mean is obtained in the limit= \lim_{q\to 0 } \exp[q^{-1 } \ln(l^{-1 } \sum_{x_=1}^l e^{q\ln|\delta u(x)|})]$ ] = eq .
( [ 0norm ] ) . ]
@xmath33 : @xmath34 the use of the geometric norm yields the so - called _ logarithmic _ bvs @xcite . for clarity of presentation
we will add the subscript @xmath0 to the notation ( for the bv , @xmath35 , and the amplitude , @xmath36 ) to emphasize which @xmath0-norm is being used . for all @xmath0-norms
the bvs look like very similar to the naked eye is strongly localized in space , and are strongly localized in space for small @xmath36 .
( see for instance @xcite and @xcite for typical snapshots of bvs with @xmath37 and @xmath30 ) .
along this paper we shall be considering an ensemble of @xmath38 bv members , @xmath39 .
all members of the ensemble are simultaneously rescaled , and they are initiated with independent random initial conditions , which is expected to result in some degree of diversity in the ensemble of bvs @xcite .
we define ensemble diversity as the degree of linear independence or transversality among the ensemble members .
diversity can be quantified by calculating the ` ensemble dimension ' , which measures the effective dimension of the sub - space spanned by the ensemble .
the higher the ensemble dimension the greater the statistical diversity .
higher ensemble dimension would imply larger dissimilarities among the ensemble members .
since bv perturbations tend to align with the fastest growing modes , a greater diversity indicates that the ensemble is able to actually sample not only the main lv but also other , less unstable , directions .
our goal here is to show , by means of several numerical calculations with a simple model , that the 0-norm is more convenient than other norms for constructing ensembles of bvs as far as ensemble diversity enhancement is concerned .
we arrive to this conclusion measuring the ensemble dimension and its temporal fluctuations , the average growth rate , and the alignment of the ensemble members with the main lv . in this subsection
we will analyze the statistical diversity in an ensemble of @xmath40 bvs .
clearly , for all @xmath0 values , in the limit @xmath41 all bvs become aligned with the leading lv and there is no diversity in the ensemble ( other than a global sign for the orientation of the vectors ) .
if @xmath36 is finite some degree of transversality between ensemble members can be expected , and to measure this diversity of the ensemble we resort to the so - called ensemble dimension @xcite . for ensembles of @xmath38 bvs , with different @xmath0-norms.,title="fig :
" ] + the ensemble dimension @xcite or bv dimension @xcite was proposed as a way to account for the number of effective degrees of freedom that explains most of the total ensemble variance ( in the spirit of principal component analysis ) , see e.g. @xcite and references therein . to compute the ensemble dimension at a given time one computes the @xmath42 covariance matrix @xmath43 with elements : @xmath44 where the standard scalar product is used in the numerator @xmath45 .
if we denote by @xmath46 the set of eigenvalues of @xmath43 , the ensemble dimension is : @xmath47 where the denominator ( @xmath48 ) equals @xmath40 due to the normalizing terms in the denominator of ( [ c ] ) .
the statistic ( [ den ] ) typically returns a real number between two limit values : @xmath49 ( if all vectors are orthogonal ) and @xmath50 ( if all vectors are aligned ) .
therefore @xmath51 measures the instantaneous degree of `` transversality '' of the ensemble .
figure [ epsilon ] depicts the results of the time - average ensemble dimension @xmath52 for different @xmath0-norms . depending on the value of @xmath0 the amplitude @xmath36
is varied in a different range .
the largest value of @xmath36 in each data set corresponds ( approximately ) to the value of the average distance ( for the corresponding @xmath0-norm ) between independent realizations of the model ( i.e. random climatological values ) . in applications
@xmath36 is much smaller than this value ( typically of the order of the analysis error ) . in the small @xmath36 region of the plots
, @xmath52 becomes equal to 1 below a particular value of @xmath36 , though for the 0-norm the convergence to @xmath53 appears to be much less abrupt .
figure [ epsilon ] shows that all @xmath0-norms allow to obtain ensembles with a certain @xmath52 after tuning @xmath36 to a particular value .
however the ensemble dimension is a time - fluctuating quantity and one should wish to minimize its fluctuations .
of course some degree of fluctuations is unavoidable due to ( i ) finiteness of the ensemble and ( ii ) intrinsic fluctuations in the state of the system ( which progressively average out for large enough systems ) .
we characterize the fluctuations of @xmath54 by means of the standard deviation @xmath55 ^ 2 \right>^{1/2}$ ] where the brackets denote a temporal average .
the results are depicted in fig .
[ sigma ] , where we plot the relative fluctuations of the ensemble dimension versus @xmath52 to better compare different @xmath0-norms .
one can readily see that the 0-norm produces the ensemble with the smallest fluctuations for most @xmath52 values .
+ ideally ( i.e. disregarding limitations by numerical accuracy ) the bvs become perfectly aligned with the main lv as @xmath56 . to determine quantitatively the degree of alignment with the lv , @xmath57 , we have measured the instantaneous angle between each bv of the ensemble , @xmath58 , and @xmath57 at breeding times @xmath59 as customary in a @xmath60-dimensional euclidean space in the range @xmath61 $ ] .
] : @xmath62 the ensemble and time average angle @xmath63 is shown in fig .
[ phiq ] , and demonstrates that the logarithmic bvs ( @xmath37 ) are able to achieve a considerable degree of alignment with the lv on average , while retaining some degree of diversity .
one clearly sees that bvs constructed with norms @xmath64 become strongly aligned among themselves while still keep a high degree of transversality with the main lv , as reflected by the high average angle of the ensembles ( @xmath65 ) in fig .
[ phiq ] for @xmath66 .
in contrast , the ` logarithmic ensemble ' ( @xmath37 ) exhibits a lower angle with the main lv , even if the statistical diversity is high .
we claim that the higher diversity and lower @xmath63 exhibited by the ensemble of logarithmic bvs ( @xmath37 ) , as compared with the ensembles with @xmath67 , indicates that this ensemble is spanning a sub - space formed by a narrow hyper - cone around the main lv , while ensembles with @xmath67 tend to lie in a lower dimension subspace that is more transverse to the lv .
+ also the growth rate of the ensemble members can be used to compare with that of the main lv , reflecting again the different behavior for different norm choices .
the average exponential growth rate of the bred vectors is @xmath68 notice that , for the sake of clarity we are using the same norm type ( @xmath30 ) to measure the exponential growth rate in all cases ( nevertheless due to the long averaging the norm type is irrelevant ) .
figure [ leq ] shows the dependence of @xmath69 on the ensemble dimension .
one can see that the logarithmic bvs ( @xmath37 ) exhibit the largest amplification rate for a given ensemble dimension , which is in agreement with the results discussed in the preceding subsections showing that the logarithmic ensemble ( @xmath37 ) , among all ensemble choices , exhibited the greatest projection on the lv .
conversely , given an exponential growth rate @xmath69 , using the @xmath70-norm will result in the most diverse ensemble . as a function of the average ensemble dimension .
the dotted line indicates the value of the lyapunov exponent .
( b ) zoom of panel ( a).,title="fig : " ] +
we have studied the effect of different norms on the construction of ensembles of bvs .
the geometric ( @xmath37 ) norm outperforms other norms ( like the euclidean one , @xmath30 ) for constructing ensembles of bvs in spatially extended systems .
the enhancement of performance ( in terms of root - mean square error , ensemble spread , and calibration time ) of ensembles of logarithmic ( @xmath37 ) bred vectors with respect to standard `` euclidean '' ( @xmath30 ) bred vectors was already uncovered by @xcite . in the present work we give a rationale behind those results .
we show that an ensemble of logarithmic bvs ( obtained with the 0-norm ) exhibits greater diversity larger ensemble dimension while its members are more strongly projected on the leading lv and have growth rates that rapidly approach the leading lyapunov exponent . in comparison ,
ensembles based on bvs with @xmath64 perform rather poorly .
they tend to collapse in one single direction ( i.e. , @xmath71 ) very abruptly as the bv amplitude is diminished and , even when all the statistical diversity is lost , they remain rather transverse to the leading lv as demonstrated by the angle with the main lv shown in fig .
moreover , the geometric norm also leads to the least fluctuating ensemble dimension among all the possible @xmath0-norms . in the view of these results
two prominent questions remain open . on the one hand
, it would be very interesting to evaluate the performance of @xmath70-norm bvs in real applications .
the study by @xcite of 0-norm bvs already showed promissing , albeit preliminary , results .
clearly , more research is needed in this direction . on the other hand ,
there is the problem of analyzing the potential advantages of ensemble kalman filters based on @xmath70-norm bvs .
our results show that logarithmic bvs have very nice properties regarding statistical diversity , growth rates , and projection onto the main lv .
therefore , a natural question that arises is : to what extent can these features translate into a better performance of ensemble kalman filtering methods ? .
we believe our results may serve as a basis for future research along these lines .
+ d.p . acknowledges support by csic under the junta de ampliacin de estudios programme ( jae - doc ) .
financial support from the ministerio de ciencia e innovacin ( spain ) under projects no .
fis2009 - 12964-c05 - 05 and no .
cgl2010 - 21869/cli is acknowledged .
wei , m. , z. toth , r. wobus , y. zhu , c. bishop , and x. wang , 2006 : ensemble transform kalman filter - based ensemble perturbations in an operational global prediction system at ncep .
_ tellus _ , * 58a * , 128 . | we show that the choice of the norm has a great impact on the construction of ensembles of bred vectors .
the geometric norm maximizes ( in comparison with other norms like the euclidean one ) the statistical diversity of the ensemble while , at the same time , enhances the growth rate of the bred vector and its projection on the linearly most unstable direction , _
i.e. _ the lyapunov vector .
the geometric norm is also optimal in providing the least fluctuating ensemble dimension among all the spectrum of @xmath0-norms studied .
we exemplify our results with numerical integrations of a toy model of the atmosphere ( the lorenz-96 model ) , but our findings are expected to be generic for spatially extended chaotic systems . |
neutrinos interact very feebly , only through weak interactions . despite this fact , they were in thermal equilibrium in the very early universe because the density of particles was much higher at that time , and so the mean free path ( or the average reaction time ) was small . at that stage ,
their number density differed from that of the photons only because one is a fermion and the other is a boson , and the relation was @xmath0 this age of thermal equilibrium ended once the density of the particles became small enough in an expanding universe , and it happened when the temperature was about @xmath1mev .
after that , the neutrinos followed only the overall expansion .
the photon number density also fell by the cube of the scale factor , so the ratio remained the same .
the only exception to this statement occurred when the temperature dropped somewhat below the electron mass , and @xmath2 annihilations produced new photons .
this event increased the photon number density by a factor of @xmath3 , so that in the present universe @xmath4 if the neutrinos are massive , the energy density is obtained by multiplying this by the mass .
the total energy density of the universe is parametrized by @xmath5 , where @xmath6 is the hubble parameter in units of @xmath7 .
thus , the fraction of energy density contributed by light neutrinos would be @xmath8 atmospheric neutrino data indicate neutrino masses of order @xmath9ev .
solar neutrino data indicate even smaller masses .
although these do not apply for all neutrino species , it surely is suggestive of the fact that @xmath10 .
= .3 before discarding light neutrinos as dark matter candidates on this ground , i would like to emphasize that there is a very strong assumption implicit in this entire argument , viz .
that the neutrinos have no chemical potential so that their number is the same as that of the antineutrinos .
if chemical potential is nonzero , the number density of neutrinos depend on it as well as the temperature , and one can not even start with the simple relation of eq .
( [ 3/4 ] ) .
if the neutrinos have chemical potential , they can contribute a larger fraction to the energy density of the universe .
this issue has been analyzed recently @xcite .
we show the results in fig .
[ f : kp ] .
the horizontal axis corresponds to neutrino mass .
the vertical axis is @xmath11 in units of the cmbr @xmath12 in the present universe .
the lines in this plot correspond to the values @xmath13 and @xmath14 respectively , starting from the inner line .
so far we talked about light neutrinos . they are known to exist .
the other candidates to be discussed have not yet been detected in any experiment .
thus , i need some motivation for talking about them . for the case at hand ,
the motivation comes from strong interactions .
these interactions are believed to be described by the gauge theory called qcd . in a gauge theory ,
one decides on some symmetry of the lagrangian and writes down all terms that are cosistent with this symmetry . for strong interactions ,
this symmetry was found in the 1970s .
if one cares to do only perturbative calculations with the lagrangian obtained from this symmetry , there is no problem .
if , however , one includes non - perturbative effects , one encounters one term which violates parity and time - reversal symmetries : @xmath15 where @xmath16 is the generalization of the electromagnetic field - strength tensor for the relevant symmetry . here
@xmath17 is the gauge coupling constant , and @xmath18 is an arbitrary parameter .
the term involving this parameter is the generalization of a term @xmath19 , which violates parity and time - reversal symmetries . among other things ,
this term will give rise to an electric dipole moment for the neutron .
the experimental bounds are very strong on this , which restricts this parameter severely : @xmath20 .
very small parameters always present certain conceptual difficulties in quantum field theories .
if they are not zero , one has to justify them by some symmetry .
peccei and quinn discovered such a symmetry by introducing a higgs field @xmath21 which has an interaction @xmath22 where @xmath23 has the dimensions of mass .
the field , of course , has other interactions , including a potential of its own , which is minimum at @xmath24 .
in general , now , we can write @xmath25 .
this field @xmath26 vanishes at the minimum and is qualified to be a quantum field .
it is called the _
axion field_. notice how it has solved our original problem .
if we add the terms shown in eqs .
( [ thgg ] ) and ( [ agg ] ) and use the definition of the quantum field , we see that the parameter @xmath18 cancels out . said another way , @xmath18 is forced to equal to zero by this procedure .
we have an interaction of the axion obtained by replacing @xmath27 by @xmath28 in eq .
( [ agg ] ) , but that does not violate parity and time reversal provided the axion is intrinsically negative under these symmetries .
we now examine how the axion can be important as the dark matter .
the equation of motion of the axion field , neglecting its interactions , is @xmath29 assuming the field to be spatially homogeneous .
this is similar to the klein - gordon equation , but there are two differences .
one of them is the second term on the left side , which shows the effects of the expansion of the universe .
the other is hidden in the notation , and is the fact that the ` mass ' @xmath30 appearing in the equation is not a constant .
it arises due to co - operative effects , and depends on the temperature . at early times ,
when the temperature was very high , @xmath30 was zero .
this is because all non - perturbative effects involve a factor @xmath31 , which is negligible since at high energies @xmath17 is very small for qcd . in this stage ,
the solution of eq .
( [ aeq ] ) is given by @xmath32 . as the universe cools down , @xmath17 increases , and becomes large enough to make the non - perturbative effects important when the temperature drops down sufficiently , say below a value @xmath33 .
to find the solution after this time , let us first forget about @xmath34 and also the time - dependence of @xmath30 . in that case , the solution would be @xmath35 where @xmath36 is a constant .
if @xmath37 and also the time variation of @xmath30 is small , we can still try a solution like this , where now both @xmath36 and @xmath30 should be regarded as functions of time .
substituting the solution in eq .
( [ aeq ] ) , we now obtain @xmath38 or @xmath39 where @xmath40 is the scale factor of the universe , i.e. , @xmath41 .
the energy density in these oscillations , at the present time , would be given by @xmath42 the axion mass in the present universe is determined by @xmath43 .
it depends mildly on the axion models , and is given by @xmath44 where the quantities on the right sides relate to properties of the pions , and are very well - known
. however , we do nt know the magnitude of the oscillations in the present universe . therefore , we use eq.([constant ] ) to write @xmath45 where @xmath40 is the value of the scale factor at any arbitrary temperature @xmath46 .
the oscillations started , as we said , around @xmath47mev . at that time , @xmath48 , and @xmath49 . using these estimates and utilizing the fact that @xmath50 , we obtain @xmath51 which gives @xmath52 thus
, axions can be an important source of dark matter if @xmath23 is close to @xmath53gev .
let us now see what is experimentally known about these axions . in fig .
[ f : axsearch ] , we show the present status @xcite .
the horizontal axis is the axion mass .
the vertical axis denotes its coupling with two photons , which allows the decay mode @xmath54 . as i said earlier , there is slight model dependence in these quantities .
the predictions of two compelling models are shown in the graph by the lines marked dfsz and ksvz .
if the axion mass is in the ev range , the decay photons could be detected through telescopes , and the failure to do so has been indicated in the figure . however , if @xmath23 indeed equals @xmath53gev and we take eq .
( [ curralg ] ) as a strict equality , the axion mass comes out to be @xmath55ev .
so , if the axions has to contribute significantly to the energy density of the universe , we should find them around that value of mass . in this case , the two photon decay mode is so slow that the axion can be considered stable on cosmological time scales .
however , because of the coupling , the axions may be resonantly converted into a microwave signal in a cavity where a strong magnetic field is present .
these experiments are very promising .
as we see , they are already probing very close to the interesting region .
we now come to the wimps , which is an acronym for `` weakly interaction massive particles . ''
there are a variety of ways in which such particles may be motivated .
the most recent and most compelling of them is supersymmetry .
supersymmetric models have new particles bosonic partners of known fermions and fermionic partners of known bosons .
there are phenomenolgically strong reasons to suspect that these new particles , or `` superpartners '' , can be produced or annihilated only in pairs .
if that is true , the lightest of these superpartners will be a stable particle .
it will not decay spontaneously
. these can be examples of wimps .
there are other models also which predict wimps .
the number density of these wimps in the early universe is governed by the boltzmann equation . in the homogeneous and isotropic universe , this takes the form @xmath56 where @xmath57 is the equilibrium number density for the temperature at any given time .
the decrease of number density due to the expansion of the universe is represented in the first term in this equation , and the same due to pair annihilations goes in the second term .
it is convenient to switch to dimensionless variables defined by @xmath58 where @xmath59 is the mass of the wimp and @xmath60 is the entropy of the background photons at a temperature @xmath46 .
the evolution equation now becomes @xmath61 \ , , \label{yeqn}\end{aligned}\ ] ] where @xmath62 , putting in the equilibrium number density for a non - relativistic particle .
the important thing to notice here is that @xmath63 hardly changes if @xmath64 at any given era characterized by a temperature determined by @xmath65 . using @xmath66
, we find @xmath67 as @xmath68 , i.e. , for large time , this ratio goes to zero .
it means that after a while , the pair annihilation rate will be negligible , because the universe has meanwhile become large enough so that the particles do not find each other .
the remnant density of the wimps can be obtained by numerical integration of eq .
( [ yeqn ] ) , but a rough estimate can be obtained by the density at the era when @xmath69 .
the solution of @xmath65 eq .
( [ g / h ] ) is now determined once the mass and the annihilation cross section are known .
after this era , the number density just scales inversely as the volume of the universe .
let me now turn to the experimental searches @xcite .
these are summarized it in fig .
[ f : klapdor ] . as i just said ,
the parameters which determine the density of wimps in the present universe are the mass and the annihilation cross section .
the two axes in this plot are precisely these two parameters . for the cross section ,
only the cross section with the nucleons is probed in the experiments .
the solid lines are exclusions plots from already published data , and the dotted lines are projections for future experiments .
the scatter plots are expected values of the parameters for various choices of the basic parameters in the supersymmetric model .
it does seem that the experiments in the near future will be able to decide whether wimps really dominate the energy density of the universe . | i discuss some compelling suggestions about particles which could be the dark matter in the universe , with special attention to experimental searches for them .
= -0.5 in = 16.5 cm = 2ex = 5ex = -5 mm = -5 mm = 16.5 cm plus.3em minus.5em plus.3em minus.5em plus.2em minus.4em plus.2em minus.4em = 10000
@=11 addtoresetequationsection
@=12 this is the write - up of a talk given at a conference where i was probably the only particle physicist .
my task was to give an overview of one area of cosmology which has a symbiotic relation with particle physics viz . , the candidates for dark matter .
i could proceed in two ways .
one is to give very brief outlines of all different suggestions ( or all i know about ) for dark matter candidate particles .
the second is to focus on a few compelling ones , and give some details of them .
after some thought , i chose the second path .
i will divide this talk into three sections , by three different kinds of particles which are potential candidates for dark matter . |
since predicted by the standard model ( sm)@xcite , top quark has become an important role in high energy physics due to its large mass , which is close to the electroweak symmetry breaking scale @xcite .
a great deal of researches focusing on top quark physics have been performed after its discovery in 1995 in the fermilab @xcite . on the experiment aspect , with the running of the tevatron and forthcoming lhc , the lack of adequate events will not be an obstacle for the top quark physics study . according to ref .
@xcite , at the lhc @xmath2 @xmath3 pairs can be obtained per year , so this enables people to measure various top quark decay channels .
meanwhile , the copious production of the top quarks supplies also a great number of bottom quark mesons since the dominant top quark decay channel is @xmath4 .
therefore , the bottom quark meson production in top quark decays may stand as an important and independent means for the study of heavy meson physics and the test of perturbative qcd ( pqcd ) .
as the known heaviest mesons , bottomonia and @xmath5 ( @xmath0 or @xmath6 ) possess particular meaning in the study of heavy flavor physics .
the lhcb as a detector specifically for the heavy flavor study at the lhc will supply copious @xmath5 and @xmath7 data for this aim .
theoretically , the direct hadroproduction of @xmath5 and @xmath7 was studied in the literature @xcite .
in addition to the `` direct '' production , `` indirect '' process as in top quark decays may stand as an independent and important source for @xmath5 and @xmath7 production .
since the top quark s lifetime is too short to form a bound state @xcite , the @xmath5 and @xmath7 production involved scheme in top quark decays is less affected by the non - pertubative effects than in other processes . in ref .
@xcite , the top quark decays into @xmath7 and @xmath8 at the born level was evaluated .
recently , the s- and p - wave @xmath5 meson productions in top quark decays were fully evaluated , including the color - octet contributions , at the leading order accuracy of qcd by chang @xmath9 @xmath10 @xcite . considering the importance of investigating @xmath5 and @xmath7 in the study of perturbative quantum chromodynamics ( pqcd ) and potential model ,
it is reasonable and interesting to evaluate the production rates of these mesons in top quark decays at the next - to - leading order ( nlo ) accuracy of pqcd . at the bottom quark and charm
quark mass scales the strong coupling is not very small , therefore the higher order corrections are usually large . on the other hand , in the processes of top quark
decays into @xmath11 , the @xmath12 , there exist large scale uncertainties in the tree level calculation @xcite .
the nlo corrections should in principle minimize it and give a more precise prediction . to calculate the @xmath13 and @xmath7 production rates in top quark decays at the nlo accuracy are the aims of this work . in our calculation ,
both of the s - wave spin - singlet and -triplet states are taken into account , i.e. , @xmath14 , @xmath15 , @xmath7 and @xmath16 . to deal with
the non - perturbative effects , the non - relativistic qcd ( nrqcd ) @xcite effective theory is employed .
the calculation will be performed at the nlo in pqcd expansion , but at leading order in relativistic expansion , that is in the expansion of @xmath17 , the relativistic velocity of heavy quarks inside bound states .
the paper is organized as follows : after the introduction , in section ii we explain the calculation of leading order decay width . in section iii ,
virtual and real qcd corrections to born level result are evaluated . in section iv
, the numerical calculation for concerned processes at nlo accuracy of pqcd is performed , and the scale dependence of the results is shown .
the last section is remained for a brief summary and conclusions .
at the leading order in @xmath18 , there are two feynman diagrams for each meson production , which are shown in figure 1 . for the convenience of analytical calculation ,
taking @xmath13 as an example , the momentum of each particle is assigned as : @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath25 .
for bottomonium , the only difference is that @xmath26 and @xmath27 represent the momenta of anti - bottom quark and bottom quark which are produced in gluon splitting . of the @xmath13 and @xmath28 production in top
quark decays , i.e. @xmath29 we employ the following commonly used projection operators for quarks hadronization : @xmath30 and @xmath31 here , @xmath32 is the polarization vector of @xmath8 with @xmath33 , @xmath34 stands for the unit color matrix , and @xmath35 for qcd . the nonperturbative parameters @xmath36 and @xmath37 are the schrdinger wave functions at the origin of @xmath0 bound states , and in the non - relativistic limit @xmath38 . in our calculation , the non - relativistic relation @xmath39 is also adopted . and @xmath40 production in top quark decays.,width=566,height=226 ] the lo amplitudes for @xmath13 production can then be readily obtained with above preparations .
they are : @xmath41 and @xmath42 here , @xmath43 , @xmath44 are color indices , @xmath45 belongs to the @xmath46 color structure . for @xmath47 production ,
the amplitudes can be obtained by simply substituting @xmath48 with @xmath49 in above expressions .
the born amplitude of the processes shown in fig.1 is then @xmath50 , and subsequently , the decay width at leading order reads : @xmath51 here , @xmath52 represents the sum over polarizations and colors of the initial and final particles , @xmath53 and @xmath54 come from spin and color average of initial t quark , @xmath55 stands for the integrants of three - body phase space , whose concrete form is @xmath56 where @xmath57 and @xmath58 are mandelstam variables .
the upper and lower bounds of the above integration are @xmath59\cdot f[s_2,m_c^2,m^2_{w}]}+[m_t^2-s_2-(m_b+m_c)^2](s_2+m_c^2-m^2_{w } ) } { 2s_2}\nonumber\\ & & + m_{\bar{b}_c}^2+m_c^2\ ; , \label{eq:9}\end{aligned}\ ] ] @xmath60\cdot f[s_2,m_c^2,m^2_{w}]}-[m_t^2-s_2-(m_b+m_c)^2](s_2+m_c^2-m^2_{w } ) } { 2s_2}\nonumber\\ & & + m_{\bar{b}_c}^2+m_c^2 \label{eq:10}\end{aligned}\ ] ] and @xmath61^{2}\ ; , \ ; s_{2}^{min } = ( m_{c}+m_{w})^{2}\end{aligned}\ ] ] with @xmath62&=&(x - y - z)^2 - 4yz\ ; .\end{aligned}\ ] ]
at the next - to - leading order , the top quark decays to @xmath13 and @xmath7 include the virtual and real qcd corrections to the leading order process , as shown in figs.[graph2]-[graph5 ] . with virtual corrections ,
the decay widths at the nlo can be formulated as @xmath63 the ultraviolet(uv ) and infrared(ir ) divergences usually exist in virtual corrections .
we use the dimensional regularization scheme to regularize the uv and ir divergences , similar as performed in ref.@xcite , and the coulomb divergence is regularized by the relative velocity @xmath17 . in dimensional regularization , @xmath64 is difficult to deal with . in this calculation , we adopt the naive scheme , that is , @xmath64 anticommutates with each @xmath65 matrix in d - dimension space - time , @xmath66 .
the uv divergences exist merely in self - energy and triangle diagrams , which can be renormalized by counter terms .
the renormalization constants include @xmath67 , @xmath68 , @xmath69 , and @xmath70 , corresponding to quark field , gluon field , quark mass , and strong coupling constant @xmath18 , respectively . here , in our calculation
the @xmath70 is defined in the modified - minimal - subtraction ( @xmath71 ) scheme , while for the other three the on - shell ( @xmath72 ) scheme is adopted , which tells @xmath73\ ; , \nonumber \\ & & \hspace{-0.3cm}\delta z_2^{os}=-c_f \frac{\alpha_s}{4\pi}\left[\frac{1 } { \epsilon_{uv}}+\frac{2}{\epsilon_{ir}}-3\gamma_{e}+3\ln \frac{4\pi\mu^{2}}{m^{2}}+4+{\mathcal{o}}(\epsilon)\right]\ ; , \nonumber\\ & & \hspace{-0.3cm}\delta z_3^{os}= \frac{\alpha_s}{4\pi } \left[(\beta_0 - 2c_a)(\frac{1}{\epsilon_{uv}}- \frac{1}{\epsilon_{ir}})+{\mathcal{o}}(\epsilon)\right]\ ; , \nonumber\\ & & \hspace{-0.3cm}\delta z_g^{\overline{ms}}=-\frac{\beta_0}{2 } \frac{\alpha_s}{4\pi}\left[\frac{1}{\epsilon_{uv}}-\gamma_{e}+ \ln4\pi+{\mathcal{o}}(\epsilon)\right]\ ; .\label{eq:13}\end{aligned}\ ] ] here , @xmath74 is the one - loop coefficient of the qcd beta function ; @xmath75 is the number of active quarks in our calculation ; @xmath76 and @xmath77 attribute to the su(3 ) group ; @xmath78 is the renormalization scale . in virtual corrections ,
ir divergences remain in the triangle and box diagrams .
of all the triangle diagrams , only two have ir divergences , which are denoted as @xmath79 and @xmath80 in fig.[graph3 ] . of the diagrams in fig.[graph4 ]
, @xmath81 has no ir singularity , while @xmath82 and @xmath83 have coulomb singularities and @xmath83 possesses ordinary ir singularity as well .
the remaining diagrams all have ir singularities , while the combinations @xmath84 , @xmath85 , @xmath86 are ir finite .
the coulomb singularities belonging to @xmath82 and @xmath83 can be regularized by the relative velocity @xmath17 .
after regularization procedure , the @xmath87 term will be canceled out by the counter terms of external quarks which form the @xmath13 or @xmath7 , while the @xmath88 term will be mapped onto the wave functions of the concerned heavy mesons .
the remaining ir singularities in @xmath89 and @xmath90 are canceled by the corresponding parts in real corrections . in the end ,
the ir and coulomb divergences in virtual corrections can be expressed as @xmath91\ ; , \label{eq:14}\end{aligned}\ ] ] with @xmath92 , @xmath93 and @xmath94 . here ,
in this work @xmath87 in fact represents @xmath95 . or @xmath40.,width=566,height=340 ] of our concerned processes
, there are @xmath96 different diagrams in real correction , as shown in fig.[graph5 ] . among them , @xmath97 , @xmath98 , @xmath99 , and @xmath100 are ir - finite , meanwhile the combinations of @xmath101 and @xmath102 exibit no ir singularities as well , due to the reasons of gluon connecting to the @xmath103 or @xmath104 quark of final @xmath13 or @xmath7 . the remaining diagrams , @xmath105 , @xmath106 , @xmath107 , and @xmath108 are not ir singularity free .
to regularize the ir divergence , we enforce a cut on the gluon momentum , the @xmath109 .
the gluon with energy @xmath110 is considered to be soft , while @xmath111 is thought to be hard .
the @xmath112 is a small quantity with energy - momentum unit . in this case , the ir term of the decay width can then be written as : @xmath113 where @xmath114 is the four - body phase space integrants for real correction . under the condition of @xmath115 ,
in the eikonal approximation we obtain @xmath116 in the small @xmath112 limit , the ir divergent terms in real correction can therefore be expressed as @xmath117 + \rm{finite\ ; terms}\right\}\ ; . \label{eq:17}\end{aligned}\ ] ] here , the @xmath118 involved terms will be canceled out by the @xmath112-dependent terms in the hard sector of real corrections . referring to the eq.([eq:14 ] ) , it is obvious that the ir divergent terms in real and virtual corrections cancel with each other . in case of hard gluons in real correction ,
the decay width reads @xmath119 in this case the phase space @xmath120 can be written as @xmath121 with @xmath122\ ; , \\ & & y_{+}=\frac{1}{s}[(\sqrt{s}-{p_{0}}^{0}-{p_{7}}^{0})^{2}-| \overrightarrow{{p_{0}}}|^{2}-({p_{7}}^{0})^{2 } + 2|\overrightarrow{p_{0}}|{p_{7}}^{0}]\ ; , \end{aligned}\ ] ] where @xmath123 is a dimensionless parameter defined as @xmath124 with @xmath125 , and @xmath126 the sum of the soft and hard sectors gives the total contribution of real corrections , i.e. , @xmath127 . with the real and virtual corrections , we then obtain the total decay width of @xmath128 quark to @xmath13 and @xmath7 at the nlo accuracy of qcd @xmath129 in above expression , the decay width is uv and ir finite . in our calculation
the feynarts @xcite was used to generate the feynman diagrams , the amplitudes were generated by the feyncalc @xcite , and the looptools @xcite was employed to calculate the passarino - veltman integrations .
the numerical integrations of the phase space were performed by the mathematica .
to complete the numerical calculation , the following ordinarily accepted input parameters are taken into account : @xmath130 @xmath131 @xmath132 here , @xmath133 is the cabibbo - kobayashi - maskawa(ckm ) matrix element and @xmath134 is weak interaction fermi constant . in above numerical calculation inputs ,
the radial wave function at the origin for s - wave @xmath135 system is estimated by potential model @xcite , while the corresponding @xmath136 nonperturbative parameter is determined from its electronic decay rate @xcite .
one loop result of strong coupling constant is taken into account , i.e. @xmath137 with the above preparation , one can readily obtain the decay widths of top quark to @xmath0 and @xmath1 mesons , as listed in table [ ratio2 ] . to see the scale dependence of the lo and nlo results , the ratios @xmath138 for @xmath0 system and @xmath139 for @xmath1 system
are showed in figures [ lpty1 ] and [ lpty2 ] , respectively .
calculation tells that after including the nlo corrections , the energy scale dependence of the results is reduced , as expected .
.the decay widths of the processes @xmath140 , @xmath141 , @xmath142 and @xmath143 at the tree level and with the nlo qcd corrections are presented in two renormalization scale @xmath78 limits , those are @xmath144 and @xmath145 for the first two processes and @xmath146 and @xmath145 for the other two . [ cols="^,^,^,^,^,^,^,^,^ " , ] [ ratio2 ] versus renormalization scale @xmath78 in @xmath128 quark decays . the left diagram for the @xmath0 spin - singlet state @xmath13 and the right diagram for the spin - triplet state @xmath28.,title="fig:",scaledwidth=48.0% ] versus renormalization scale @xmath78 in @xmath128 quark decays .
the left diagram for the @xmath0 spin - singlet state @xmath13 and the right diagram for the spin - triplet state @xmath28.,title="fig:",scaledwidth=48.0% ] versus renormalization scale @xmath78 in @xmath128 quark decays .
the left diagram for the @xmath1 spin - singlet state @xmath16 and the right diagram for the spin - triplet state @xmath7.,title="fig:",scaledwidth=48.0% ] versus renormalization scale @xmath78 in @xmath128 quark decays . the left diagram for the @xmath1 spin - singlet state @xmath16 and the right diagram for the spin - triplet state @xmath7.,title="fig:",scaledwidth=48.0% ]
in this work we have calculated the decay widths of top quark to s - wave @xmath0 and @xmath1 bound states at the nlo accuracy of perturbative qcd .
considering that there will be copious @xmath3 data in the near future at the lhc , our results are helpful to the study of the indirect production of these states
. they may be also useful to the future study of nlo heavy quark to @xmath0 and @xmath1 bound states fragmentation functions .
numerical results indicate that the nlo corrections greatly enhance the lo results for @xmath1 system , while slightly decrease the @xmath0 states production widths .
the main reason for this difference is that the nlo wave function for bottomonium is much larger than that of lo one , while for the calculation of @xmath13 meson , the same wave function given by potential model is used .
although from table i , superficially the number of indirectly produced @xmath13 overshoots that of @xmath7 , experimentally to detect the latter is much easier than the former . since top
quark dominantly decays into @xmath103 and @xmath147 final state with a width of 1.5 gev or so , numerical results remind us that the @xmath7 indirect production from top quark decay is detectable , while it is hard to pin down the @xmath13 states by this way .
the numerical calculation also shows that the next - to - leading order qcd corrections to processes @xmath148 decrease the energy scale dependence of the decay widths as expected , and hence the uncertainties in theoretical estimation .
future precise experiment on the concerned processes may provide a test on the theoretical framework for heavy quarkonium production and the reliability of perturbative calculation for them .
this work was supported in part by the national natural science foundation of china(nsfc ) under the grants 10935012 , 10928510 , 10821063 and 10775179 , by the cas key projects kjcx2-yw - n29 and h92a0200s2 .
99 w. hollik , in proceedings of the xvi international symposium on lepton - photon interactions , connell university , ithaca , n.y . , aug .
10 - 15 1993 ; m. swartz , in proceedings of the xvi international symposium on lepton - photon interactions , connell university , ithaca , n.y .
10 - 15 1993 .
e. braaten and t.c .
yuan , phys .
71 , 1673 ( 1993 ) ; ibid , phys . rev .
d50 , 3176 ( 1994 ) ; c .- h . chang and y .- q .
chen , phys .
b284 , 127 ( 1992 ) ; ibid , phys .
d46 , 3845 ( 1992 ) ; y .- q .
chen , phys .
d48 , 5181 ( 1993 ) ; t.c .
yuan , phys .
d50 , 5664 ( 1994 ) .
k. kolodzidj , a. leike and r. ruckl , phys .
b355 , 337 ( 1995 ) ; c .- h .
chang , y .- q .
chen , phys .
b364 , 78 ( 1995 ) ; c .- h .
chang , y .- q .
chen , phys .
d48 , 4086 ( 1993 ) ; c .- h .
chang , int .
a21 , 777 ( 2006 ) ; c .- h .
chang , j .- x .
wang and x .- g .
wu , phys .
d70 , 114019 ( 2004 ) ; c .- h .
chang , c .- f .
qiao , j. x. wang and x .- g .
wu , phys .
d71 , 074012 ( 2005 ) ; c .- h .
chang , x .-
wu , eur.phys.j.c38 , 267(2004 ) ; c .- h . chang , c. driouichi , p. eerola and x .- g .
wu , comput .
159 , 192(2004 ) ; c .- h .
chang , j .- x .
wang and x .- g .
wu , comput .
174 , 241(2006 ) . | the decay widths of top quark to s - wave @xmath0 and @xmath1 bound states are evaluated at the next - to - leading(nlo ) accuracy in strong interaction .
numerical calculation shows that the nlo corrections to these processes are remarkable
. the quantum chromodynamics(qcd ) renormalization scale dependence of the results is obviously depressed , and hence the uncertainties lying in the leading order calculation are reduced .
* pacs number(s ) : * 13.85.ni , 14.40.lb , 12.39.hg , 12.38.bx . |
the formation and the properties of the first stars in the universe are probably among the most fascinating problems in present - day cosmology . numerical simulations and analytical studies in a cosmological context ( tegmark 1997 , abel 1998 , bromm , coppi & larson 2000 , machacek , bryan & abel 2000 )
have reached a consensus on the fact that the fragments potentially leading to star formation were already available inside very small dark matter halos at redshift @xmath10 in most cosmological models .
after that point the requirements concerning the dynamic range and number of physical processes to be treated become so demanding that one has to turn to simplified and/or lower dimensional studies .
although considerably refined , these studies ( omukai & nishi 1999 , susa & umemura 2000 , nakamura & umemura 2000 , ripamonti 2000 ) have not yet reached an agreement on the mass of the stellar seed ( core ) and on the accretion rate by which the former grows into a star . determining if and when the accretion is quenched seems to hold the key for the understanding of the properties of the first stars .
however , the large variety of physical effects to be considered ( for a review see ferrara 2000 ) again complicate the problem tremendously . on general grounds
some authors have speculated that the imf of the first stars can be skewed towards high stellar masses ( larson 1998 ) or even be bimodal ( nakamura & umemura 2000 ) .
one of the consequences of this assumption is an increase of the relative number of massive , short - lived stars .
these stars will soon end their life as supernovae , leaving either a compact remnant ( a neutron stars or a black hole ) or nothing at all ( for masses above @xmath11 , see umeda 2000 and fryer , woosley & heger 2000 ) , and possibly releasing detectable bursts of gravitational waves ( schneider 2000 ) . in turn , this implies a strong decrease of the amount of zero- or very low - metallicity stars presumably included in the present - day milky way ( mw ) , according to the hierarchical paradigm .
given the above discussion it appears that number counts of very low - metallicity mw stars can not only provide evidences for living fossils " , stars formed immediately after the end of dark ages , but also probe the primordial imf and hierarchical models to cosmic epochs ( @xmath12 ) currently unreachable to essentially all experiments .
we propose here that the first stars are born inside the so - called popiii objects ( haiman & loeb 1996 , gnedin & ostriker 1997 , ciardi 2000 ) systems with total mass @xmath13 , which strongly rely on molecular hydrogen cooling in order to collapse , given their low virial temperatures ; in short , we define as popiii stars those formed inside such mini galaxies . as ciardi
have pointed out , these objects are extremely fragile to the energy release by their own supernovae ; as a result , their gas is quickly evacuated and they can only witness a single burst of star formation .
the remnant is a `` naked stellar object '' ( ciardi 2000 ) , a tiny agglomerate of long lived low - mass stars , essentially retaining the metallicity of the gas from which they formed .
these are the stars that we are considering as candidates for the very low - metallicity stars observed in the mw halo .
our paper is structured as follows : the next section presents the press - schechter formalism used to calculate the mass functions of mw sub - galactic clumps at different redshifts . in section 3
we describe the treatment of popiii halos as mini galaxies to estimate the number of stars each contributes to the final mw .
section 4 gives a comparison with recent observational results of the numbers and metallicity distribution of extremely low metallicity halo stars .
finally , a brief summary is given in section 5 .
we use the extended press - schechter formalism to calculate the conditional probability that a virialized halo of mass @xmath14 at @xmath15 will become part of a more massive halo of mass @xmath16 at a later time , @xmath17 .
this allows us to calculate the mass functions of the fragments which through mergers will end up as part of a mw sized system today . following nusser & sheth ( 1999 ) , we identify the mass contained in progenitors of mass between @xmath14 and @xmath18 with : m = m_0 _ c(z_1-z_0 ) exp| ds_1 | dm_1 in the above expression @xmath19 is the rms density fluctuation in a top hat window function of radius @xmath20 , and @xmath21 the critical overdensity for collapse , with @xmath22 being the present mean mass density of the universe .
both of the above quantities were calculated using a @xmath23 power spectrum normalized to @xmath24 , in a @xmath25 flat universe ( @xmath26 ) , with @xmath27 km s@xmath28mpc@xmath28 , which we used throughout , percival & miller(1999 ) .
a baryonic fraction of 0.05 of the total halo mass was assumed for all halos .
current direct estimates of the mass of the mw rely on distant galactic satellites used as probes of the mass internal to their current positions , using their line of sight velocities and reasonable assumptions on their orbits .
these methods give values for the total mass of the mw of around @xmath29 , kulessa & lynden - bell ( 1992 ) , binney & merrifield ( 1998 ) .
however , those estimates yield only lower limits , as the total extent of the galactic dark halo could be greater than the current positions of suitably observable satellites .
consistently with the press - schechter approach we are using , we estimate the total mass of the mw as the total mass contained within a sphere of radius @xmath30 , the radius inside which the average density is 200 times the present background density , @xmath31 .
this in fact is the only meaningful definition of `` galaxy '' within the press - schechter approach .
the circular velocity at this radius would be around @xmath32 , where @xmath33 km s@xmath28 and the factor of @xmath34 between @xmath35 and @xmath36 is deduced from cdm simulations avila - reese ( 1999 ) , steinmetz & navarro ( 1999 ) .
this gives : m_mw= ( 3 4 ) ^1/2 v_200 ^ 3 which for the adopted cosmological model gives @xmath37 , somewhat above the observational lower limits . in
what follows we shall use @xmath38 , within the uncertainties associated with going from @xmath35 to @xmath36 , and in determining @xmath35 .
1 shows the mass of sub - halos which at @xmath39 find themselves included in mw sized halos , as a function of sub - halo mass , at different redshifts ; the dotted lines show the limits within which halos at the redshifts shown will form popiii halos .
we associate a maximum metallicity to the stars to be formed and contained within these halos through the mean mass weighted metallicity of the universe at each redshift , as calculated by gnedin & ostriker ( 1997 ) using fully hydrodynamic simulations . as a result of the merger process , these halos will probably contain a few stars with yet lower metallicities formed in halos of the preceding hierarchies ; in this sense we take the metallicities of gnedin & ostriker ( 1997 ) as an upper limit for the stars of these halos .
the lower mass limit , @xmath40 , is taken from tegmark ( 1997 ) ( see however fuller & couchman 2000 ) . systems having virial temperatures below this curve can not cool in a hubble time .
the upper limit , @xmath41 , corresponds to the upper mass at which the stellar feedback due to supernova can expel the baryonic component entirely , thus quenching further star formation .
following ciardi ( 2000 ) , we define this as a `` blowaway '' , and we adopt their determination for the evolution of @xmath41 .
thus , larger systems where self enrichment can continue and stars of metallicities higher than the gnedin & ostriker ( 1997 ) limit will be formed are excluded .
however , given the sharply decreasing mass functions , our results are rather insensitive to this upper limit .
these two limits change slightly with redshift , the ones shown are typical for the redshifts displayed , although the more detailed redshift dependent estimates discussed above were used . fig .
2 shows the integral of the popiii halos mass functions between the selected limits , @xmath42 , as a function of redshift .
the upper axis shows the mean mass weighted metallicity corresponding to each redshift , discussed above .
2 also gives the typical mass of a single popiii halo , @xmath43 , showing that within the redshift interval displayed , each present day mw sized system will have had between @xmath44 and @xmath45 popiii halos as progenitors , contributing stars with metallicities in @xmath46 < -2.4 $ ] .
this means that the mass functions of eq . 1 will be sampled densely , and the spread between different present day systems will be minimal i.e. , all mw sized systems should contain the same numbers and distribution of stars having metallicities in the above range .
the two curves of fig . 2 intersect at a redshift of @xmath47 , corresponding to a metallicity of @xmath1=-4.3 $ ] in the gnedin & ostriker picture .
at this limit many present day mw sized systems will have had only one popiii halo coming from beyond this redshift , and many will have had none at all . in this way
we can expect that the numbers of stars in the milky way with metallicities below @xmath1=-4.3 $ ] should be minimal , indeed , no such stars have been found to date . for even lower metallicities , one goes to larger limiting redshifts to reach @xmath48 , @xmath49 and @xmath50 , the objects responsible for the appearance of the very first stars and of starting the reionization of the universe .
however , as commented above , the chance of finding any such object as part of the mw is slim , they will be found as part of larger systems , perhaps as intracluster stars . for simplicity
though , we will use the term first stars to identify the earliest stars of our galaxy .
we model each of the popiii halos discussed above as a small galactic system .
systems with masses larger than @xmath40 cool and the gas forms a disk in centrifugal equilibrium ( often seen in the numerical simulations quoted above ) , with an exponential surface density profile fixed by the total baryonic mass and angular momentum .
this last is given by the @xmath51 parameter , which is chosen at random from the distribution : @xmath52 { { d\lambda } \over { \lambda}},\ ] ] with @xmath53 and @xmath54 , dalcanton ( 1997 ) and references therein .
the initial dark halo profile is chosen such to have a central constant density core , as seen in present day dark halos from dwarf to cluster scales , firmani ( 2000 ) .
the formation of the disk modifies the gravitational potential in the central regions , we include the reaction on the dark halo to this process through an adiabatic invariance hypothesis ( flores 1993 ; hernandez & gilmore 1998 ) .
3 shows a typical popiii proto - galaxy , having a total mass of @xmath55 , with a close - to - average angular momentum , @xmath56 .
the final rotation curve is given by the solid curve , with the dotted line giving the surface density profile of the disk .
[ fig3 ] [ fig4 ] the final temperature of the gas in the disk is taken to be @xmath57 , as appropriate for gas where the main cooling mechanism is h@xmath58 ( ferrara 1998 , susa & umemura 2000 , machacek 2000 ) .
this temperature determines @xmath59 , the sound speed within the disk , which together with @xmath60 , the epicycle frequency obtained from the detailed final rotation curve , yields the toomre s stability parameter for the disk : @xmath61 in the regions where @xmath62 the total tidal shears , together with the velocity dispersion , are sufficient to stabilize the disk locally against its self gravity .
these regions are therefore not subject to star formation . the dashed curve in fig .
3 gives the profile of @xmath63 in the disk , showing the regions interior to the intersection with the @xmath64 line to be unstable towards star formation , in this case @xmath65 of the disk mass .
4 is analogous to fig . 3 , but for a system having a larger angular momentum , @xmath66 .
the changes in disk surface density and rotation curve imply than now only a smaller fraction ( @xmath67 ) of the disk will form stars . a value for the gas temperature lower than the @xmath57 we are assuming , due to the small presence of heavy elements in these objects , would only increase slightly the value of the unstable fraction .
given that the unstable fractions for systems with typical values of @xmath51 are already close to unity , this detail does not affect our conclusions .
finally , a gas to star conversion efficiency @xmath68 is used to turn the total mass within the instability region into a final mass of stars , for each object .
although we have calibrated this parameter from local observations ( ciardi et al . 2000 ) , since there is no firm theoretical prediction for any evolution with redshift of it , we have not explored such possibility .
although the above instability criterion might be satisfied , it does not necessarily follow that the considered popiii objects will be able to form stars .
in fact , it is possible that cloud - cloud collisions might disrupt the gas clouds formed by the gravitational instability before they have time to collapse ( a point originally made by kashlinsky & rees 1983 ) . to explore this possibility
we calculate the time for collisions between clouds , @xmath69 and for cloud collapse , @xmath70 .
this last will be given by : t_ff ( 1 g ) ^1/2 , where @xmath71 is the density of the gas clouds .
the collisional time scale is t_coll= ( _ j^2 c_s)^-1 , with @xmath72 the number density of clouds , and @xmath73 the jeans length in the disk ( in general a function of radius ) which we take as a representative size for the clouds . taking @xmath74 , where @xmath75 is the volume fraction of the disk filled with clouds , a filling factor , and the jeans mass m_j= ( 6 ) ( c_s^2 )
^3/2 , since the jeans length is : _
j= ( c_c^2 ) ^1/2 , we can calculate the ratio of the collision time to the collapse time by substituting the values above into eqs . 5 and 6 to yield , = 2.9 f_c .
9 shows that even if the filling factor for the fragmenting disk were unity , which is unlikely , the collapse time of the clouds is always shorter than the collision time , at all radii , and independent of the value of @xmath59 .
this is a result of having taken a constant value of @xmath59 throughout the disk , which is a consequence of h@xmath76 being the dominant cooling mechanism operating in the primordial metallicity gas . as eq .
9 shows , for all the scales of our problem , gravitational cloud collapse can proceed unimpeded by cloud - cloud collisions , which we shall not consider further . once the total gas mass turned into stars as a function of the metallicity has been determined
, we use the imf of larson ( 1998 ) to turn this into a total number of stars .
this function offers a convenient parameterization of the imf : @xmath77 in the above equation @xmath78 is a characteristic mass scale , of order @xmath79 for a present day solar neighbourhood imf .
this scale mass can be increased to explore the consequences of a top heavy imf applying to stars of metallicities much lower than solar , as the ones we are treating .
we combine this sampling of the imf using @xmath80 with the most recent padova stellar models of primordial metallicities of girardi ( 2000 ) to exclude those stars which due to their age have by now evolved off the tip of the rgb , and are hence no longer visible . in this way
we obtain a sample of present day stars .
sampling the imf from a fixed lower luminosity up to the tip of the rgb of the theoretical isochrones allows us to obtain the number of stars visible today upwards of a certain limiting luminosity , as a function of metallicity .
our final samplings include only stars between @xmath81 and the tip of the rgb of the padova isochrones ( of around @xmath82 ) , which is the range over which stellar models were calculated .
as the total mass in popiii objects is high , both the popiii halos mass functions and the @xmath51 distributions are sampled densely , yielding a low spread for the predicted final number of stars . to summarize , each of the popiii object selected from the mass function of eq .
1 is carefully modeled as a mini galaxy , with a value of @xmath51 selected at random from eq .
the resulting system is then treated as a mini galaxy where the cooling of the baryons leads to disk formation ; subsequently a disk instability criterion is used to obtain the total mass in stars contributed by each object .
finally , through the use of an imf and new low metallicity stellar models , this total mass is turned into a number of stars visible today .
this is repeated until the total popiii object mass equals that within the bounded region described in section 2 , for a given limiting redshift , which corresponds to a fixed maximum metallicity .
this process is repeated for different limiting redshifts , to construct a total mass distribution function for the low metallicity stars for the present day mw .
the curve in fig .
5 shows the number of mw stars with mass in the above range and having metallicities lower than a given value .
to compare our results with observations we take the sample of ryan & norris ( 1991 ) of low metallicity halo stars in the local disk .
they include a magnitude limited sample , ( their nltt sample ) which is complete down to @xmath81 and @xmath83 , the same limit we use to sample our imf . from the theoretical isochrones , this corresponds to a distance of 290 pc , which we take as the radius of the sample . in that sample
5 stars were found with @xmath1<-3.0 $ ] . to obtain the total number of stars in the galactic halo at this metallicity , we must introduce an assumption regarding the distribution of these stars .
[ fig5 ] if the popiii stellar systems are disrupted while they are incorporated into the growing mw , their stars will today be found almost exclusively in the galactic halo , while if the initial stellar systems form tight clusters , they might survive disruption and through dynamical friction end up in the galactic bulge . which of these two processes dominates will depend on the mass ratio of the mergers through which the popiii systems reach the mw .
if these mass ratios are close to one , disruption will be important , while if they are accreted late into a fully formed mw popiii objects might remain bound . from n - body simulations white & springel ( 2000 ) have argued that most of the lowest metallicity stars in a present day galaxy will be found in the stellar halo , and not in the bulge .
we therefore make the assumption that the lowest metallicity stars share the isothermal distribution of the dark halo , out to a distance of @xmath45 kpc , representative of the total extent of the halo .
the actual density profile of these objects might more closely resemble that of the old globular cluster population , having an @xmath84 profile .
fortunately , fixing the normalization point at the solar radius ( 8.5 kpc ) yields total numbers which are largely independent of this change in the assumed profile , e.g. for core radii in the @xmath84 profile going from 0.2 to 2 kpc ( a core radius must be introduced in the @xmath84 profile to avoid the central divergence ) , the total numbers of stars change by a factor 1.24 to 0.34 , respectively , small compared to our error bars in figures 5 and 6 .
it is worth pointing out that the assumption of a particular density profile for the current low metallicity halo stars is not intrinsic to our results , but becomes necessary only for data comparison purposes .
[ fig6 ] taking the ryan & norris ( 1991 ) data plus the hypothesis of an isothermal distribution discussed above , we derive the total number of stars with @xmath1<-3.0 $ ] in the galaxy as @xmath85 . to obtain numbers at other metallicities we use the relative metallicity distribution function of the more recent beers ( 1998 ) work , which combines several new samples .
these observations are shown in fig . 5 by the points with error bars , which reflect the uncertainties in the normalization .
the stellar halo might not extend as far out as @xmath45 kpc , but some popiii stars might be found in the bulge , so the interpretation of data used in fig .
5 is probably a lower limit . on the other hand ,
given the uncertainties in the primordial imf , our prediction is a strict upper limit .
it is clear that our results fall somewhat above the observational measurements , implying that our assumption of a constant @xmath80 , essentially a present day solar neighbourhood imf , is not valid .
reconciling our estimates with the observations requires not only the assumption of a higher @xmath78 in the past , but throughout the redshift range studied , an increasing trend for this value .
note that our derived numbers of popiii stars scale linearly with @xmath86 [ fig7 ] several authors , larson ( 1998 ) have pointed out that the increase of the cmb temperature with redshift alone , implies an increase in the jeans mass for the star forming clouds , once one is beyond @xmath87 , where the cmb temperature equals the @xmath88 k of cold star forming clouds .
it has therefore been suggested that at high redshift the imf should be increasingly weighted towards larger stellar masses . the labeled arrows in fig 5 .
show the values of @xmath78 which we require at @xmath89 and @xmath90 , to match the observed points . in fig .
6 we show the values of @xmath78 which we infer for the imf of popiii objects , by imposing that model and observations agree . the solid curve indicates the redshift evolution of the jeans mass in star forming clouds , which we identify with @xmath78 . although for @xmath91 our results are consistent with the increase in @xmath78 coming solely from that in the jeans mass due to the cmb , the trend for @xmath92 is clearly for a rise in @xmath78 beyond what this effect can account for . in fact , as larson ( 1998 ) points out , this is only the _ minimum _ increase to be expected , as other heating sources , massive stars themselves or a uv background , were likely to be present .
furthermore , a reduced opacity consistent with these low metal abundances would have the effect of skewing the imf towards larger masses , thus operating in the same direction . for @xmath93 , the values of the characteristic mass appear to stabilize around the value @xmath94 , indicative of a very top heavy imf at these early epochs , although the increase of our error bars makes determining trends in this redshift range harder .
finally , fig .
7 presents an hr diagram containing over 2000 stars , produced from the low metallicity stellar models used , convolved with errors appropriate to hst observations of halo stars ( gallart 1999 ; hernandez 2000 ) , for @xmath95 .
the inclusion of detailed atmosphere calculations and spatial information to yield detailed observational predictions will form part of a forthcoming paper .
from our study we conclude that : * if the standard hierarchical merger scenario is correct , in the mw one should not expect to find stars having metallicities lower than @xmath1=-4.3 $ ] , consistent with current observational limits . *
we infer the imf of popiii stars to have been strongly weighted towards high masses , and increasingly more so in the range @xmath92 , this levelling off at @xmath96 at a characteristic mass scale @xmath97 .
* observational limits on the numbers and metallicity distribution of very low metallicity stars in the mw can place sharp restrictions on the imf of the first stars .
this can be achieved through detailed modelling of the full range of phenomena giving rise to such stars , of the type attempted here , thus necessarily establishing a connection with the underlying structure formation scenario .
the authors wish to thank the referee , volker bromm , for his careful reading of the manuscript and helpful comments which improved the presentation of this paper , as well as richard larson for a detailed look at a first draft and leo girardi for making his new isochrones available electronically .
this work was completed as one of us ( af ) was a visiting professor at the center for computational physics , tsukuba university , whose support is gratefully acknowledged .
abel , t. , anninos , p. , norman , m. l. & zhang , y. 1998 , apj , 508 , 518 avila - reese , v. , firmani , c. , klypin , a. & kravtsov , a. v. , 1999 , mnras , 310 , 527 beers , t. c. , , 1998 , ssrv , 84 , 139 binney , j. , & merrifield , m. 1998 , galactic astronomy , ( princeton : unipress ) bromm , v. , coppi , p. s. & larson , r. b. 1999 , apj , 527 , l5 ciardi , b. , ferrara , a. , governato , f. , & jenkins , a. , 2000 , mnras 314 , 611 dalcanton , j. , spergel , d. n. , & summers , f.j . , 1997 , apj , 482 , 659 ferrara , a. 1998 , apjl , 499 , 17 ferrara , a. 2000 , preprint ( astro - ph/0007179 ) firmani , c. , donghia , e. , avila - reese , v. , chincarini , g. , hernandez , x. , 2000 , mnras , 315 , 29 flores r. , primack j. r. , blumenthal g. r. , faber s. m. , 1993 , apj , 412 , 443 fryer , c. l. , woosley , s. e. & heger , a. , 2000 , preprint ( astro - ph/0007176 ) fuller , t. m. & couchman , h. m. p. 2000 , preprint ( astro - ph/0003079 ) gallart c. , freedman w. , aparicio a. , bertelli g. , chiosi c. , 1999 , aj , 118 , 2245 girardi , l. , bressan , a. , bertelli , g. , & chiosi , c. , 2000 , a&as 141 , 371 gnedin , n. y. , & ostriker , j. p. , 1997 ,
apj , 486 , 581 haiman , z. & loeb , a. 1997 , apj , 483 , 21 hernandez , x. , gilmore , g. , 1998 , mnras , 294 , 595 hernandez , x. , gilmore , g. , valls - gabaud , d. , 2000 , mnras , in press kashlinsky , a. & rees , m. j. 1983 , mnras , 205 , 955 kulessa , a. s. & lynden - bell , d. , 1992 , mnras , 255 , 105 lacey , c. , & cole , s. , 1993 , mnras , 262 , 627 larson , r. b. , 1998 , mnras , 301 , 569 machacek , m. e. , bryan , g. l. & abel , t. 2000 , preprint ( astro - ph/0007198 ) nakamura , t. & umemura , m. 2000 , preprint nusser , a. , & sheth , r. k. , 1999 , mnras , 303 , 179 omukai , k. & nishi , r. , 1999 , apj , 518 , 64 percival , w. & miller , l. , 1999 , mnras , 309 , 823 ripamonti , e. , haardt , f. , colpi , m. & ferrara , a. 2000 , in preparation ryan , s. g. , & norris , j. e. , 1991 , aj , 101 , 1865 schneider , r. , ferrara , a. , ciardi , b. , ferrari , v. & matarrese , s. 2000 , mnras , in press ( astro - ph/9909419 ) steinmetz , m. & navarro , j. 1999 , apj , 513 , 555 susa , h. & umemura , m. 2000 , apj , 537 , 578 tegmark , m. 1997 , apj , 474 , 1 umeda , h. , nomoto , k , & nakamura , t. 2000 , in the first stars , ed .
weiss ( berlin : springer ) , 126 white , s. d. m. , & springel , v. 2000 , preprint ( astro - ph/9911378 ) | we explore the predictions of the standard hierarchical clustering scenario of galaxy formation , regarding the numbers and metallicities of popiii stars likely to be found within our galaxy today . by popiii
we shall be referring to stars formed at large redshift ( @xmath0 ) , with low metallicities ( @xmath1<-2.5 $ ] ) and in small systems ( total mass @xmath2 @xmath3 ) that are extremely sensitive to stellar feedback , and which through a prescribed merging history ( lacey & cole 1993 ) end up becoming part of the milky way today .
an analytic , extended press - schechter formalism is used to get the mass functions of halos which will host popiii stars at a given redshift , and which will end up in milky way sized systems today .
each of these is modeled as a mini galaxy , with a detailed treatment of the dark halo structure , angular momentum distribution , final gas temperature and disk instabilities , all of which determine the fraction of the baryons which are subject to star formation .
use of new primordial metallicity stellar evolutionary models allows us to trace the history of the stars formed , give accurate estimates of their expected numbers today , and their location in @xmath4 vs. @xmath5 hr diagrams .
a first comparison with observational data suggests that the imf of the first stars was increasingly high mass weighted towards high redshifts , levelling off at @xmath6 at a characteristic stellar mass scale @xmath7 .
# 1@xmath8 # 1@xmath9 # 1equation ( [ # 1 ] ) # 1equation ( [ # 1 ] ) # 1 ( [ # 1 ] ) # 110^#1 galaxies : formation galaxies : evolution galaxy : formation galaxy : abundances stars : luminosity function , mass function |
most of the difficulty of quantum many - body physics stems from the complexity of its fundamental mathematical objects : many - body wavefunctions and density matrices . in the simplest case , where we have @xmath0 qubits , a wavefunction ( pure state )
can be considered as a function mapping @xmath1 .
therefore , it is characterized by @xmath2 complex parameters .
density matrices ( mixed states ) have even greater mathematical complexity , mapping @xmath3 , i.e. @xmath4 complex parameters .
the aim of this work is to describe a pictorial representation of quantum many - body wavefunctions , in which a wavefunction characterizing a chain of @xmath0 qubits maps into an image with @xmath5 pixels .
thus , an increase in the number of qubits reflects itself in an increase in the resolution of the image .
these images are typically fractal , and sometimes self - similar .
extension to higher spin _ qudits _
is straightforward , and is also explored . some physical properties of the wavefunction become visually apprehensible : magnetization ( ferro or antiferromagnetic character ) , criticality , entanglement , translation invariance , permutation invariance , etc .
visualization of complex data is a common problem in many branches of science and technology .
let us review here some of the relevant hallmarks that preceded our work .
historically , it can be argued that the single most relevant advance in calculus was the discovery of the relation between algebraic functions and curves in the plane in the xvii century .
function visualization provided an insight which guided most of the subsequent development of calculus , not only by helping solve established problems , but also by suggesting new ones . with the advent of the new information technologies
, complex data visualization has developed into a full - fledged field of research .
the reader is directed to @xcite for a recent review of state - of - the - art techniques , and @xcite for a historical perspective . as a relevant example , the problem of visualization of dna and protein sequences was addressed in 1990 by jeffrey making use of the so - called _ chaos game representation _ ( cgr ) @xcite .
dna sequences are long , highly correlated strings of four symbols , @xmath6 .
let us label the four corners of a square with them .
now , select the central point of the square and proceed as follows .
pick the next symbol from the string .
find the point midway between the selected point and the corner which corresponds to the symbol .
mark that point , and make it your new selected point .
if the sequence is genuinely random , the points will cover the square uniformly .
otherwise , patterns will emerge , very often with fractal structure .
the original purpose of the technique was mere visualization , but it evolved @xcite to provide quantitative measurements , such as shannon entropies , which help researchers to characterize dna and protein sequences @xcite . in 2000 ,
hao and coworkers @xcite developed a different representation technique for long dna sequences that also had fractal properties . given a certain value of @xmath0 , they computed the frequency of every subsequence of length @xmath0 within the global sequence , thus obtaining a mathematical object which is similar to a many - body wavefunction , only mapping from @xmath7 .
the number of different subsequences of length @xmath0 is @xmath8 .
hao and coworkers represented the subsequence probability distribution by dividing a unit square in a recursive way , into @xmath8 small squares , and attaching a color to each of them .
the resulting images have fractal appearance , as remarked by the authors , but their quantification is not pursued .
their purpose is to identify which types of subsequences are under - represented , and to this end they analyse the corresponding patterns of low frequency . in 2005
latorre @xcite , unaware of the work of hao et al . , developed independently a mapping between bitmap images and many - body wavefunctions which has a similar philosophy , and applied quantum information techniques in order to develop an image compression algorithm .
although the compression rate was not competitive with standard _
, the insight provided by the mapping was of high value @xcite . a crucial insight for the present work was the idea that latorre s mapping might be inverted , in order to obtain bitmap images out of many - body wavefunctions .
focusing on quantum mechanics , the simplest visualization technique is provided by the representation of a qubit as a point on a bloch sphere .
early work of ettore majorana @xcite proved that a permutation - invariant system of @xmath0 spins-1/2 can be represented as a set of @xmath9 points on the bloch sphere .
this majorana representation has proved very useful in characterizations of entanglement @xcite . a different approach that can provide visualization schemes of quantum many - body systems was introduced by wootters and coworkers in 2004 @xcite .
the idea is to set a bidimensional array of operators which fulfill certain properties , and measure their expectation values in the given state .
those values , displayed in a 2d lattice , generate a discrete analogue of a wigner function . in this work ,
we describe a set of techniques which provide graphical representations of many - body wavefunctions , which share many features with the schemes of latorre and hao and coworkers .
the main insight is that the increase in complexity as we add more qubits is mapped into an increase in the resolution of the corresponding image .
thus , the thermodynamic limit , when the number of qubits tends to infinity , corresponds to the continuum limit for the images .
the scheme is recursive in scales , and this makes the images look fractal in a natural way .
in fact , as we will discuss , exact self - similarity of the image implies that the wavefunction is factorizable . in section [ 2dplots ]
we describe the basic wavefunction plotting scheme , while section [ examples ] is devoted to providing several examples ( heisenberg , itf , dicke states , product states , etc . ) emphasizing how physical features map into plot features .
the procedure is generalized in section [ otherplots ] , and some alternative plotting schemes are described , which allow us to try states of spin-1 systems , such as the aklt state .
section [ selfsim ] , on the other hand , deals with the fractal properties of the plots and extracts useful information from them . in section [ entanglement ] discusses how to recognize entangled states in a wavefunction plot , along with a simple technique to estimate entanglement by inspection .
a different plotting scheme , based upon the frame representation and related to the wootters group approach is succintly described in section [ pauli ] , and a few pictures are provided for the sake of comparison .
the article finishes with conclusions and a description of future work .
let us consider a couple of qubits .
the tensor basis is composed of four states : @xmath10 , @xmath11 , @xmath12 and @xmath13 .
consider also a unit square , @xmath14\times[0,1]$ ] , and divide it into four `` level-1 '' squares .
we can associate each of the basis states to one of the squares , as shown in figure [ fig.2dplot ] ( top ) .
( 45,0)(-10,-20)level 1 ( -10,-25)2 qubits ( 0,0)(40,-40)(0,-20)(40,-20)(20,0)(20,-40)(10,-10)00 ( 10,-30)10 ( 30,-10)01 ( 30,-30)11 ( 45,-42.5)(-10,-20)level 2 ( -10,-25)4 qubits ( 0,0)(40,-40 ) ( 0,-20)(40,-20)(20,0)(20,-40)(10,0)(10,-40)(30,0)(30,-40)(0,-10)(40,-10)(0,-30)(40,-30)(5,-5)(0,0)0000 ( 10,0)0001 ( 20,0)0100 ( 30,0)0101 ( 5,-15)(0,0)0010 ( 10,0)0011 ( 20,0)0110 ( 30,0)0111 ( 5,-25)(0,0)1000 ( 10,0)1001 ( 20,0)1100 ( 30,0)1101 ( 5,-35)(0,0)1010 ( 10,0)1011 ( 20,0)1110 ( 30,0)1111 the basic mapping is , therefore : @xmath15 the splitting of squares can be iterated , obtaining level-2 squares , etc .
, as it is shown in figure [ fig.2dplot ] ( bottom ) . for a wavefunction with @xmath0 qubits , we will have to descend down to level-@xmath16 squares .
each of them will correspond to one of the tensor basis states .
if the number of qubits @xmath0 is odd , the same scheme can be applied with a rectangular plot .
the last step is straightforward : attach a color , or a gray level , to each of the level-@xmath16 squares , depending on the value of the wavefunction . obviously , using only levels of gray ( or color intensity )
, only real values can be attached easily to each tensor basis state . in order to show phases
, we recourse to a color - cycle scheme .
figure [ fig.2dplot.features ] shows some features of this mapping .
the ferromagnetic ( fm ) states , @xmath17 and @xmath18 correspond , respectively , to the upper - left ( nw ) and lower - right ( se ) corners of the image , while the nel antiferromagnetic ( af ) states correspond to the other two corners : @xmath19 is the upper - right ( ne ) corner , and @xmath20 is the lower - left one ( sw ) .
it is straightforward to realize that the @xmath21 symmetry operation @xmath22 corresponds to a rotation of @xmath23 around the center of the plot .
( 40,-3)(0,0)(40,-40)(0,-20)(40,-20)(20,0)(20,-40)(10,0)(10,-40)(30,0)(30,-40)(0,-10)(40,-10)(0,-30)(40,-30)(0,0)(5,-5 ) ( 40,-40)(35,-35 ) ( 2,-2)(-5,-2)(-5,-2)fm 0000 ... ( 38,-38)(45,-38)(45,-38)fm 1111 ...
( 35,0)(40,-5 ) ( 0,-35)(5,-40 ) ( 2,-38)(-5,-38)(-5,-38)af 1010 ... ( 38,-2)(45,-2)(45,-2)af 0101 ... let us consider any state @xmath24 and denote its bits by @xmath25 , with @xmath26 .
in order to find the point in the unit square where this state will be mapped , build the following numbers : @xmath27 those are the coordinates of the upper - left corner of the corresponding level-@xmath9 square , if @xmath28 is the upper - left corner of the square , and the @xmath29-coordinate grows downwards . in our plots ,
unless otherwise stated , each cell is filled with a color corresponding with its wave - function amplitude according to the following scheme : color intensity corresponds to the modulus ( white means zero ) , and hue is used as a phase indicator .
concretely , red is used for positive values and green for negative ones , with a smooth interpolation scheme .
figure [ fig.n4 ] provides some simple examples of states with @xmath30 qubits .
panel ( a ) is the qubistic plot for the factorizable state @xmath31 , in which only the upper - left corner cell of the plot is marked . in ( b ) the greenberger
zeilinger ( ghz ) state is shown , @xmath32 . in this case , two opposite corner cells are marked , with the same color since their relative phase is positive .
the third panel , fig .
[ fig.n4 ] ( c ) corresponds to the so - called w state for @xmath30 qubits , i.e. : @xmath33 .
the plot consists of @xmath0 marked cells distributed along the upper and leftmost rows of the plot .
for larger values of @xmath0 the spacing among these marked cells becomes exponential .
the fourth plot in figure [ fig.n4 ] is the dicke state at half - filling , i.e. : the linear combination , with equal weights , of all basis states with half the qubits @xmath34 . [
cols="^,^,^,^,^ " , ] on the other hand , taking the half - filling dicke states of figure [ fig.dicke ] , it is evident that , at every magnification level , the number of different subimages increases by two .
thus , @xmath35 in terms of levels , or @xmath36 for qubits , if @xmath37 .
this bound is found to be fulfilled by the numerical calculations shown in figure [ fig.entropy.dicke ] .
the reason for the difference between the estimate and the actual values of entanglement in fig .
[ fig.entropy.dicke ] is twofold .
first , the number of different level-@xmath38 quadrants is , in general terms , a very poor way to estimate the dimension of the subspace spanned by them .
second , the value estimated this way is just the schmidt rank , whose logarithm is just an upper bound to the actual entanglement entropy .
both problems can be handled within the more comprehensive framework , described in the next section . given a wavefunction plot and a level @xmath39 ,
let us divide the full region into a grid of @xmath40 sub - plot .
moreover , let @xmath41 be an index running through all such sub - plots and @xmath42 be the actual image displayed in it , as in equation [ final.image ] .
now we define a _ cross - correlation matrix _ for the plot image , @xmath43 , as @xmath44 this cross - correlation matrix bears full information about entanglement of the first @xmath45 qubits within the wavefunction , as we proceed now to show . according to equation [ final.image ] , the image on quadrant @xmath41
is given by a linear combination of the right - states .
using the orthogonality property assumed for them we obtain @xmath46 thus , we recognize that @xmath43 is just the _ density matrix for the left part_. in other terms : @xmath47 therefore , _ the cross - correlation matrix of the wavefunction plot holds full information related to entanglement_. for example , for a product state , all subimages are equivalent modulo normalization .
thus , we can assume that @xmath48 with @xmath49 is the norm for each subimage .
obviously , @xmath50 and , thus , the matrix @xmath51 is just a projector on a line .
its spectrum is , in decreasing order , @xmath52 .
therefore , its entanglement entropy is zero .
in this last section we describe a rather different approach to the problem of providing a graphical representation of a quantum many - body system , but still self - similar by design . instead of plotting wavefunction amplitudes , or probabilities , we can plot the expectation values of a bidimensional array of operators , chosen in such a way that the full information contained in the wavefunction is preserved .
this is called a _ frame representation _ of the quantum state @xcite . according to wootters and coworkers @xcite
, the final representation may correspond to a discrete analogue of a _ wigner function _
@xcite , with very interesting properties in order to characterize _ non - classicality _ , such as its _ negativity _ @xcite .
let us consider a system of @xmath9 qubits , described by a certain density matrix @xmath53 .
now , let us consider the unit square @xmath14 \times [ 0,1]$ ] and any two numbers , @xmath41 and @xmath29 , characterized by their binary expansion : @xmath54 , @xmath55 .
the value attached to the point @xmath56 in the plot will be given by the expectation value in @xmath53 of the operator @xmath57 : @xmath58 \label{tracea}\ ] ] where @xmath57 is given by : @xmath59 in other words , we plot the expected value of every combination of tensor products of @xmath60 . in particular , on the @xmath61 line we get uniquely correlations in @xmath62 ; on @xmath63 , those in @xmath64 , and on @xmath65 those corresponding to @xmath66 .
such representation is unique for every density matrix , and can be reverted as follows : @xmath67 in order to attain some intuition about the representation , figure [ fig.frame ] illustrates it for one and two qubits . at each cell , we depict the expected value of a `` string '' operator , as shown .
( 30,0)(-7,-20 ) ( 47,-20 ) ( 0,0)(40,-40)(0,-20)(40,-20)(20,0)(20,-40)(10,-10 ) ( 10,-30 ) ( 30,-10 ) ( 30,-30 ) ( 30,-42.5)(-7,-20 ) ( 47,-20 ) ( 0,0)(40,-40 ) ( 0,-20)(40,-20)(20,0)(20,-40)(10,0)(10,-40)(30,0)(30,-40)(0,-10)(40,-10)(0,-30)(40,-30)(5,-5)(0,0)@xmath68 ( 10,0)@xmath69 ( 20,0)@xmath70 ( 30,0)@xmath71 ( 5,-15)(0,0)@xmath72 ( 10,0)@xmath73 ( 20,0)@xmath74 ( 30,0)@xmath75 ( 5,-25)(0,0)@xmath76 ( 10,0)@xmath77 ( 20,0)@xmath78 ( 30,0)@xmath79 ( 5,-35)(0,0)@xmath80 ( 10,0)@xmath81 ( 20,0)@xmath82 ( 30,0)@xmath83 figure [ fig.pauliprod ] shows our first example : the frame representation of a product state given by @xmath84 i.e. : a spin pointing half - way between the @xmath85 and @xmath86 axes . the plot shows a striking sierpiski - like structure , which can be fully understood by noticing that , in this state , @xmath87 and @xmath88 are nonzero , while @xmath89 . if , in figure [ fig.frame ] ( bottom ) we cross out all elements with a @xmath64 , the sierpiski - like structure will appear .
self - similarity , therefore , is rooted in the plotting scheme , as in the previous case . as an example , we provide in figure [ fig.pauliitf ] images illustrating the itf quantum phase transition : above , @xmath90 is small and only correlations in the @xmath86-axis are relevant .
below , @xmath90 is large and correlations appear only in the @xmath91-axis .
the middle panel shows the critical case .
in this work we have described a family of schemes which allow visualization of the information contained in quantum many - body wavefunctions , focusing on systems of many qubits .
the schemes are self - similar by design : addition of new qubits results in a higher resolution of the plots .
the thermodynamic limit , therefore , corresponds to the continuum limit .
the philosophy behind the schemes is to start out with a region @xmath92 and divide it into several congruent subdomains , all of them similar to @xmath92 .
this subdivision procedure can be iterated as many times as needed , producing an exponentially large amount of subdomains , each of them characterized by a geometrical index .
this index can be now associated to an element of the tensor - basis of the hilbert space , and its corresponding wavefunction amplitude goes , through a certain color code , into that subdomain .
the most simple example is with @xmath92 a square which splits into four equal quadrants , but we can also start with a right triangle , or even with a line segment . physical features of the wavefunctions translate naturally into visual features of the plot . for example , within the scheme in section [ 2dplots ] , the spin - liquid character of the ground state of the heisenberg model shows itself in a characteristic pattern of diagonal lines .
this pattern is able to distinguish between open and periodic boundary conditions .
other features which show up in the plots is magnetization , criticality , invariance under translations or permutation of the qubits , and marshall s sign rule .
we have analysed the characteristic features of product states , the ground states of the ising model in a transverse field , the majumdar - ghosh hamiltonian or dicke states .
we have also studied spin-1 systems , such as the aklt state .
a very relevant physical feature which becomes apparent in the plots is entanglement .
factorizability is straightforward to spot : a wavefunction is factorizable if all sub - images at a certain division level are equal , modulo normalization .
the schmidt rank of a given left - right partition of the system is related to the dimension of the subspace spanned by all sub - images within the corresponding subdivision of the plot and , so , a crude method to obtain an upper bound is to count the number of different sub - images .
the full information about entanglement is contained in the matrix that we have termed as cross - correlation , which contains the overlap between all subimages at a certain division level . in a very different spirit
, we have illustrated the frame representations of quantum states of many qubits .
this approach is related to wooters group ideas . in it ,
the expectation values of a selected set of operators are shown in a 2d array , which is again displayed in a self - similar manner . in this work
we have taken the first steps in the exploration of an alternative strategy in the study of quantum many - body sytems , which can provide support to the corpus of methods in the field .
regarding further work , we would like to stress the further exploration of interesting quantum many - body states which we have not done here , for example the ground states of fermionic hamiltonians , the hubbard model , the mott transition or the bec - bcs crossover .
understanding the plotting structure of matrix product states of low dimension might also result profitable .
moreover , the mathematical properties of the mapping itself are worth studying by themselves . as a final remark
, we would like to announce that source code and further images can be found at http://qubism.wikidot.com , a webpage dedicated to qubism - related resources .
this work has been supported by the spanish government grants fis2009 - 11654 , fis2009 - 12964-c05 - 01 , fis2008 - 00784 ( toqata ) and quitemad , and by erc grant quagatua .
m.l . acknowledges the alexander von humboldt foundation and the hamburg theory award .
l . acknowledges d. peralta and s.n .
santalla for very useful discussions . | a visualization scheme for quantum many - body wavefunctions is described , which we have termed _
its main property is its _ recursivity _ : increasing the number of qubits reflects in an increase in the image resolution .
thus , the plots are typically fractal . as examples ,
we provide images for the ground states of commonly used hamiltonians in condensed matter and cold atom physics , such as heisenberg or itf .
many features of the wavefunction , such as magnetization , correlations and criticality , can be visualized as properties of the images .
in particular , factorizability can be easily spotted , and a way to estimate the entanglement entropy from the image is provided . |
inverse compton ( ic ) scattering is one of the most common radiative processes in astrophysics .
the papers of jones ( 1968 ) and blumenthal & gould ( 1970 ) provide the basic theory , while both exact expressions and useful approximations for interaction rates , mean scattered energies , and dispersion about the mean energies are given by coppi & blandford ( 1990 ) .
the computations reported in these papers have been performed by supposing an isotropic distribution of the incident photon momenta .
however , there are astrophysical situations in which the photon momenta are anisotropically distributed so that the standard equations are not sufficient . from a theoretical standpoint , baylis et al.(1967 ) showed that the compton spectrum ( emitted over the total solid angle ) from a beam of ultra
relativistic electron population ( @xmath1 ) in the thomson approximation is independent on the incoming photon distribution being for all considered cases @xmath2 .
the anisotropic inverse compton ( aic ) equations describing the scattering between a photon beam and isotropic relativistic electrons in the thomson approximation were obtained by bonometto et al.(1970 ) for a power law energy distribution of ultra relativistic electrons .
aharonian & atoyan ( 1981 ) obtained the general _ redistribution function _ describing the scattering probability into a given direction of a monochromatic incoming photon beam scattered by isotropic monochromatic relativistic electrons .
furthermore , by integrating the _ redistribution function _ over a population of ultra
relativistic electrons with a power law energy distribution ( @xmath1 ) they obtained the analytical expression of the spectrum ( in terms of hypergeometric series ) and showed that it presents the same shape of that derived from an isotropic illumination ( i.e. @xmath2 in the thomson approximation and @xmath3 in the klein
nishina regime ; @xmath4 being the energy of the seed photons ) .
more recently , nagirner & poutanen ( 1993 ) have obtained the _ redistribution matrix _ for compton scattering of polarized incoming photon beams and , by integrating the _ redistribution function _ over a power law energy spectrum of ultra
relativistic electrons in the thomson approximation , they showed the equivalence between their results and those of bonometto et al.(1970 ) .
all the formulae of astrophysical interest reported in these papers are obtained by making use of the ultra relativistic limit . as a consequence , they are not useful to the detailed study of astrophysical situations in which also trans
relativistic and mildly
relativistic electrons are involved .
more recently brunetti et al.(1997 ) have treated the aic thomson scattering of a photon beam by an isotropic electron distribution not restricted to the ultra relativistic limit .
the aic has been considered in several specific astrophysical situations .
the compton losses suffered by electrons moving in a bath of anisotropic seed photons was studied both in the thomson approximation and in the klein
nishina case and was used to produce optical and gamma ray emission from the pulsars ( see morini 1981 , 83 and reference therein ) . with reference to the hard x ray properties of compact sources ( agns and galactic black hole candidates ) , ghisellini et al .
( 1991 ) have derived ( thomson approximation ) the angular distribution of the power and the spectrum of the hard ic radiation emitted by a relativistic electron located at the center of an hemispherical bowl , or cap , uniformly radiating soft x
rays ( the emission region was assumed to be optically thin and the aic scattering is taken to be isotropic in the electron rest frame ) .
dermer & schlickeiser ( 1993 ) and bttcher et al . (
1997 ) have derived approximate solutions describing the aic scattering of incoming photons from an accretion disc with the ultra
relativistic electrons in the relativistic jets of agns .
recently , moskalenko & strong ( 1999 ) have discussed the effect due to the anisotropic distribution of the incoming photons on the gamma
ray emission from interstellar photons ic scattered by cosmic
ray electrons in the galaxy .
this has been achieved by numerically integrating a _ redistribution function _
, in the klein
nishina regime , that was obtained by considering the scattered photons as unidirectionally emitted in the direction of the electrons .
however , this approximation is strictly valid only in the ultra relativistic case , which is the one of interest in the mentioned paper .
the inverse compton scattering of collimated streaming particles with a surrounding bath of soft photons emitted by the disc torus would produce high energy photons in agns and/or galactic x ray binaries ( bednarek & kirk 1995 , protheroe et al .
furthermore the same mechanism has been used to model the gamma ray emission in the mev blazar objects ( bednarek 1998 ) . in order to better describe the emission pattern of all these astrophysical situations
general exact analytical expressions would be useful .
the aim of the present paper is to investigate the aic single scattering in the general case : from the trans
relativistic to the ultra
relativistic limit , both in the klein
nishina case and in the thomson approximation .
general formulae of astrophysical interest are derived .
specifically , we will consider the scattering of a photon beam with an electron beam ( sect.3 ) and with an isotropic distribution of electron momenta ( sect.4 ) .
in sect.2 we define the basic geometry of the scattering and obtain general relationships from the scattering between a monochromatic photon beam and an electron . in sect.4 we also compare our results with those of bonometto et al.(1970 ) and aharonian & atonyan ( 1981 ) for the ultra relativistic aic and with the isotropic ultra relativistic inverse compton case ( blumenthal & gould 1970 ) . as an application , in sect.5 we use our general aic equations of sect.4 to study the x to gamma ray spectrum of the powerful radio galaxies deriving from the aic scattering of the nuclear ir to uv photons from a hidden quasar by the relativistic electrons in the radio lobes .
in particular we argue that the detection of such emission with the new generation x ray telescopes ( chandra , xmm ) can provide important information about the magnetic field strength , the energy density , distribution and aging of the relativistic particles in the radio galaxies .
we assume a scattering geometry such that the momenta of the incoming and scattered photon and of the electron are described by the four - vectors ( fig.1 ) : @xmath5 @xmath6 @xmath7 @xmath8 and @xmath9 being the incoming and scattered photon energies , @xmath10 ( @xmath11 ) the unit vector components in spherical coordinates , @xmath12 the electron mass and @xmath13 the electron velocity . with these assumptions
the incoming photon density is given by : @xmath14 @xmath4 being the assumed incoming photon energy , @xmath15 the _ dirac function_. it is well known that in the compton scattering the tetra vector components of the incident and scattered photons in the lab frame and in the electron frame ( primed ) are related by ( rybicki & lightman 1979 , pozdnyakov et al .
1983 ) : @xmath16 @xmath17 @xmath18 where @xmath0 is the lorentz factor of the electron . in general the inverse compton emissivity in the lab frame may be obtained by the manifold integral ( blumenthal & gould 1970 ) : @xmath19 where @xmath20 , the photon number density in the electron frame , can be easily derived from the relativistic invariant @xmath21 ( blumenthal & gould 1970 ) . for a monochromatic ( @xmath4 ) photon beam moving along the z axis ( eq.4 ) it is : @xmath22 @xmath23 is the differential compton cross section , in our case the unpolarized klein
nishina cross section ( berestetskii et al . 1982 ) :
@xmath24 @xmath25 is the electron number per energy and solid angle . in this case
it is formally written as : @xmath26 and @xmath27 , the scattering angle in the electron frame , is derived by a lorentz transformation of eqs.(12 ) along the electron direction : @xmath28 the velocity of the electron that scatters a photon from energy @xmath8 to @xmath9 is derived by combining eqs.(57 ) and making use of the relativistic transformation of the angles ( @xmath29 ) : @xmath30 that in the thomson approximation ( i.e. @xmath31 ) becomes @xmath32 the compton emissivity is readily obtained by eqs.(812 ) and ( 57 ) and well known relativistic transformations : @xmath33 ^ 2 \nonumber\\ + { { \epsilon_1}\over{mc^2 } } { { k_3 - 1}\over{\gamma_0 l } } \}\end{aligned}\ ] ] where @xmath34 , @xmath35 , @xmath36 is given by eq.(13 ) with @xmath37 and @xmath38 replaced with the assumed electron coordinates ( @xmath39 and @xmath40 respectively ) and @xmath41 in the thomson approximation eq.(15 ) becomes the much more simple : @xmath42 with @xmath36 given by eq.(14 ) and @xmath43
as previously assumed the photons propagate along the z axis . in this section
the direction of the electron beam is defined by the coordinates @xmath44 and @xmath45 . the electron number density is formally given by : @xmath46 where @xmath47 and @xmath48 are the number density and energy distribution of the electrons . in this case
given @xmath49 , the electron direction @xmath37 and the scattered photon direction @xmath50 , only one electron energy ( @xmath51 ) satisfies the conservation laws . as a consequence the @xmath15function appearing in eq.(15 ) can be written as : @xmath52 where @xmath53 is given by eq.(13 ) with @xmath4 instead of @xmath8 , while the jacobian of @xmath54 with respect to @xmath8 is : @xmath55 + { { 1-k_3}\over { 2mc^2 } } ( { { \epsilon d\phi / d\epsilon}\over { \phi^{1\over 2 } } } + 2 \phi^{1\over 2 } ) -{{e_jk_j+e_3 - 2{{\epsilon e_3}\over{\epsilon_1 } } } \over { \epsilon_1 } } \}\end{aligned}\ ] ] with @xmath56 . the compton emissivity is obtained by integrating eq.(15 ) over the electron distribution ( eq.19 ) . by making use of eqs.(2021 ) and well known relativistic transformations , the integration yields : @xmath57 where @xmath58 is given by eq.(13 ) with the assumed electron coordinates @xmath59 and the incoming photon energy @xmath4 replacing the general @xmath37 and @xmath8 respectively .
@xmath60 is given by eq.(21 ) with @xmath61 , @xmath40 , and @xmath58 replacing @xmath62 , @xmath38 and @xmath54 , respectively , and where @xmath63 eq.(22 ) can be used to calculate the emission due to aic scattering between a streaming population of relativistic electrons and seed photons emitted by a small region ( e.g. , a star close to the jet or a hot spot in the accretion disk ) in agns or galactic x ray binaries . to describe with high accuracy much more common astrophysical situations the solution
can also be easily obtained by numerically integrating over the distributions of the photon and electron momenta . in order to describe the aic emitted power and spectrum and the changes introduced by the non ultra
relativistic scattering kinematics , we have to assume an energy distribution of the electron population .
let us assume a simple power law energy distribution , @xmath64 , so that the results of this section can be compared with those in which the electron momenta are isotropically distributed ( next section ) .
it should be born in mind , however , that there are no known simple mechanisms providing a power law energy distribution in accelerated electron beams ( fermi mechanisms in general work with isotropic electrons momenta ) . the aic emissivity is simply obtained by replacing @xmath65 with @xmath66 in eq.(22 ) . it can be noticed that the emitted power vanishes when @xmath67 goes to infinity ( @xmath68 ) .
electrons of increasing energy are required to scatter the photons to a given energy @xmath9 at larger angles from the electron beam , up to a maximum angular distance for which @xmath69 .
therefore photons of energy @xmath9 are scattered in the line of sight only if the angular distance of the electron beam is sufficiently small . since the scattering problem is symmetric with respect to the z axis , it is simpler to study the spatial distribution of the scattered radiation by assuming that the electron beam lies on the ( y z ) plane . from eq.(13 ) , with @xmath70 , and @xmath71 , one finds that the limiting condition @xmath72 is obtained at two extremes in @xmath73 ( fig.2 ) : @xmath74 \pm \ { ( 1-(e_3 ^ 0)^2 ) \nonumber\\ \cdot ( 1-(1-{{\epsilon_0}\over{\epsilon_1}}(1-e_3 ^ 0))^2 ) \}^{1/2}\end{aligned}\ ] ] similarly , by releasing the condition @xmath75 , one obtains the extremes in @xmath76 by imposing the condition @xmath77 .
if @xmath78 is large enough , the emission cone may comprise the z axis .
the angle of the electron velocity at which this happens is derived from eq.(24 ) by setting @xmath79 : @xmath80 for @xmath81 the scattered directions on the ( y z ) plane are comprised in the interval @xmath82 and @xmath83 , while for @xmath84 the scattered directions are bounded by @xmath85 and @xmath86 and @xmath87 and @xmath88 ( fig.2 ) . in the thomson limit eq.(22 ) becomes : @xmath89 where @xmath67 and @xmath90 are given by eq.(14 ) with @xmath37 replaced by the assumed electron beam coordinates @xmath39 and with @xmath8 replaced by @xmath4 .
furthermore it can be readily shown that in the thomson approximation the emitted pattern has the same boundaries of the klein
nishina case .
an example of the emitted power as a function of the scattering angle , both in the thomson approximation and in the klein
nishina case , is represented in fig.3 for two different values of the electron spectral index .
relatively more power is channeled at larger scattering angles with decreasing @xmath15 .
the broadening in the emitted power distribution in the case of low values of @xmath15 is due to the contribution of the high energy electrons that emit along relatively large values of @xmath91 . due to the compton
recoil the energies of the electrons that are required to scatter incoming photons ( @xmath4 ) up to an energy @xmath9 in the klein
nishina regime are larger than those in the thomson approximation . as a consequence the emitted power in the klein
nishina regime is more depressed in the case of a steeper energy distribution of the electrons ( as it is @xmath92 in fig.3 ) .
the emitted power per unit frequency and solid angle as a function of the emitted energy is represented in fig.4 for different scattering angles both in the thomson approximation and in the klein
nishina case .
the very sharp cut - off in the emitted spectrum is simply due to the kinematics of the scattering and only weakly depends on the electron differential spectrum .
such feature represents the portion of the emitted spectrum contributed by the electrons with @xmath54 approaching to unity and with the emitted photons moving along the boundaries of the scattered cones .
in fig.4 it is seen that the difference between the thomson approximation and the klein
nishina scattering becomes larger at small scattering angles ( with respect to the direction of the electron beam ) and increases with the energy of the scattered photons .
actually , for each scattering configuration this difference follows a quite complicated trend being smaller both at low and at high emitted energies and larger at intermediate energies .
the trend is due to the difference between the thomson and the kn cross sections but , because of the compton recoil , it is also sensitive to the electron energy distribution ( see brunetti 1998 for further details ) . a relevant example is the thomson spectrum in the direction of the electron beam ( @xmath93 and @xmath94 ) . in this case from eq.(26 ) the compton emissivity is : @xmath95 in the ultra
relativistic limit eq.(27 ) can be expanded in series for @xmath96 to get the much more simple : @xmath97 now , as one can notice from eqs.(57 ) if @xmath98 then @xmath99 , however , the ultra
relativistic condition @xmath100 is not satisfied . as a consequence
although in the case @xmath101 is @xmath102 , the spectrum differs from the so called ultra relativistic case as it can be noticed from eqs.(2728 ) . the spectrum per unit solid angle in the mildly
relativistic case ( eq.27 ) is rather complicated . in fig.5
its slope is shown as a function of the electron beam direction and of @xmath103 ; it is also compared with the ultra relativistic value ( eq.28 ) .
the power emitted by the electron beam over @xmath104 can be compared with the formulae giving the emitted power in the case of isotropic inverse compton scattering given in the literature ( jones 1968 , blumenthal & gould 1970 , rybicki & lightman 1979 ) . in the ultra
relativistic case the integration of eq.(26 ) over @xmath105 yields : @xmath106 the spectrum is a power law as in baylis et al.(1967 ) . in the mildly
relativistic case the spectrum is harder for large angles between the incoming photon and the electron beam , while it is softer at small angles ( fig.6 ) .
in the case of isotropic distribution the electron number density per energy and solid angle is : @xmath107 the emissivity is given by integrating eq.(15 ) over the angular and energy distribution of the scattering electrons .
the integral over @xmath108 can be performed by following the method given in aharonian & atoyan ( 1981 ) ; one finds : @xmath109^{3/2 } } } \nonumber\\ + { { ( \gamma + \epsilon_0/mc^2 ) \epsilon_0/mc^2 + \gamma \epsilon_1/mc^2 - \epsilon_1 \epsilon_0 k_3/(mc^2)^2 } \over { [ ( \gamma+\epsilon_0/mc^2)^2 ( 1-k_3)^2 + 1-k_3 ^ 2 ] ^{3/2 } } } ) \nonumber\\ \cdot { { 1-k_3}\over{\epsilon_0\epsilon_1 ^ 2 } } ( mc^2)^3 \ } n d\gamma\end{aligned}\ ] ] where the function @xmath110^{-{{1}\over2}}\nonumber\\ - [ ( \gamma+\epsilon_0/mc^2)^2 ( 1-k_3)^2 + 1-k_3 ^ 2 ] ^{-{{1}\over 2}}\end{aligned}\ ] ] the lower limit of the integral eq.(31 ) is given by the minimum energy of the electrons that can scatter an incident photon ( @xmath4 ) to an energy @xmath9 in the direction @xmath105 .
the minimum velocity of the electrons is obtained by minimizing eq.(13 ) with respect to the electron directions ( @xmath37 , i.e. @xmath111 ) .
given that the distribution of the scattering electrons is isotropic , it is equivalent to describe the emitted spectrum with reference to any plane containing the incoming photon momenta ( the z axis ) . for simplicity of notation
we study the scattering properties on the ( y z ) plane , so that @xmath112 , @xmath113 , and the electrons having the lowest energy to make the scattering are those moving on the ( y z ) plane i.e. with @xmath114 ( i.e. @xmath115 ) .
we find that eq.(13 ) is minimized by electrons having directions : @xmath116 with @xmath117 for @xmath118 and @xmath119 for @xmath120 . by combining eqs.(13 ) and
( 33 ) one finds @xmath121 and @xmath122 .
it should be noticed that eq.(33 ) holds whatever the electron and the photon energy may be , i.e. both in the thomson approximation and in the klein nishina case , from the ultra
relativistic limit down to the trans relativistic case . in the thomson approximation , from eqs.(14 ) and ( 33 ) , we find : @xmath123 that becomes the simplest @xmath124 in the ultra
relativistic case ( bonometto et al . 1970 ) , and the well known @xmath125 in the ultra relativistic isotropic thomson scattering ( taking @xmath126 ) . in the literature
it is usually assumed a momentum power law energy distribution of the relativistic electrons @xmath127 since it is that expected from fermi acceleration mechanisms .
it deviates from a simple energy power law in the non ultra
relativistic case . in this section
our aim is to derive aic equations from the trans
relativistic to the ultra
relativistic regime and to compare our findings with the ultra
relativistic equations in the literature . as a consequence , in order to evidence the changes in the aic emitted spectrum and pattern introduced only by the non ultra
relativistic scattering kinematics , in this section , we adopt a simple energy power law distribution @xmath128 down to trans relativistic energies . the case of a general @xmath48 is discussed in appendix a where a simple semi analytical solution is given .
the case of the distribution @xmath127 is discussed in appendix b and is compared with the results of this section . from eq.(31 ) the integration in the thomson case yields : @xmath129 { \cal i}_{5/2 } \nonumber\\ + { 5\over 2}(1-k_3)^{-2 } [ 3(1+{{\epsilon_0}\over{\epsilon_1}})+({{\epsilon_1}\over { \epsilon_0}}+ ( { { \epsilon_0}\over{\epsilon_1}})^2 ) ] { \cal i}_{7/2 } \}\end{aligned}\ ] ] where @xmath130 @xmath131 @xmath132 @xmath133 where @xmath134 , @xmath135 , and @xmath136 is given by eq.(34).+1/2)/ \gamma(i[m]+1)$ ] in eqs.(3739 ) are also equivalent to @xmath137 } b(i[m]+1/2 , i[m]+1/2 ) / \pi^{1/2}$ ] ; @xmath138 being the _ beta function_. ] a solution in a simple closed form for eqs.(3539 ) that holds down to the mildly
relativistic case ( for @xmath139 with @xmath140 approximation ) is : @xmath141 where @xmath142 and @xmath143 this very simple formula could be very useful in astrophysical problems where non ultra
relativistic electrons are involved in the scattering process . by considering eqs.(36 - 39 ) in the ultra relativistic limit @xmath96 , after simple algebraic manipulation from eq.(35 ) ( or simply from eqs.(4042 ) ) we find the emissivity in the ultra
relativistic case : @xmath144 that is equivalent to the result given in bonometto et al.(1970 ) .
furthermore by integrating eq.(43 ) over @xmath145 ( it should be noticed that due to the isotropic distribution of the electron momenta it is equivalent to integrate over @xmath105 ) one finds the well known solution in the isotropic ultra relativistic thomson scattering case ( blumenthal & gould 1970 )
. the emitted power is obtained by eq.(35 ) in the general case or by the much more simple eq.(43 ) in the ultra
relativistic case ; in the ultra relativistic case , it is shown in fig.7 . as expected
the scattered power has a maximum at @xmath146 and goes to zero for small scattering angles .
the electrons that are mainly responsible for the scattered power at a given @xmath9 are those with energies close to the minimum energy that increases with the scattering angle .
it follows that by increasing @xmath15 there are relatively fewer electrons at higher energies and , consequently , the emitted power goes to zero more rapidly toward smaller scattering angles . in order to compare our general results with the aic formulae in the literature we give in fig.(8 ) the ratios between the aic emission ( eq.35 ) and that of the ultra
relativistic case ( eq.43 ) as a function of the scattering angle for different electron energy distributions and energies ; the differences are @xmath147 also in the mildly relativistic case ( i.e. @xmath148 ) . the main effect due to the non relativistic kinematics
is that the percentage of radiation emitted at large scattering angles decreases with decreasing @xmath149 .
the emitted spectrum for different scattering angles and @xmath15=2.5 is shown in fig.9 . as already noticed ( eqs.35 and 42 ) , at variance with the ultra
relativistic standard equations in the literature , the aic spectral index calculated with our general equations is a function of the scattering angle and of the energy .
with decreasing @xmath149 the spectrum flattens at large @xmath150 and steepens at small @xmath150 ; this causes the progressive decrease of the percentage of radiation emitted at large @xmath150 ( fig.8 ) .
the difference from the standard ultra
relativistic spectrum ( that is @xmath151 ) , vanishes at higher energies due to the relativistic aberration .
it is well known that the emitted ic spectrum , in the case of an isotropic distribution of the seed photons momenta , typically steepens at an energy ( @xmath152 ) such that @xmath153 . in the case of the aic scattering @xmath152
is expected to depend on the scattered directions .
eq.(31 ) can be used to investigate the klein
nishina steepening at different scattering angles both in the ultra
relativistic and in the general case .
this point is illustrated in fig.10 where we have plotted the aic spectra for different values of @xmath150 by assuming a power law energy distribution of the scattering electrons ( @xmath154 , @xmath15=2.5 and @xmath4=1 kev ) .
the behaviour in fig.10 is readily understood by keeping in mind the kinematics of the compton scattering ( eqs .
( 5 ) and ( 7 ) ) .
one has that @xmath155 and that @xmath0 , the typical ( or minimum ) energy of the electron required to scatter a photon ( @xmath4 ) to @xmath9 increases for large @xmath150 . as a consequence the kn
phase starts at smaller emitted energies with increasing @xmath150 . on the other hand
, we have shown that in the thomson approximation the emitted power is greatly enhanced at large @xmath150 ( fig.7 ) .
these two effects can largely compensate each other .
this is clearly illustrated in fig.11 where we have plotted the emitted power as a function of the scattering angle ( with @xmath15=2.5 and @xmath4=1 kev ) for several values of the scattered photon energy @xmath9 .
when the kn regime is well established ( larger @xmath9 ) the scattered power is much more isotropically distributed ( a sort of kn isotropization ) . as shown by aharonian & atoyan ( 1981 ) , the shape of the aic emitted spectrum in the ultra
relativistic case is identical to the isotropic case .
aharonian & atoyan ( 1981 ) equation can be obtained by integrating eq.(31 ) over the electron spectrum in the ultra relativistic limit . however , since the kn regime may become important at relatively low energies of the scattering electrons , both the non ultra
relativistic effects previously discussed ( fig.9 ) and the classical kn steepening may show up in the emitted spectrum .
this happens to be in the energy range that simultaneously satisfies @xmath156 ( for which the spectrum becomes sensitive to the kn effects ) and @xmath157 ( i.e. non ultra
relativistic ) .
it is illustrated in fig.12 where we have plotted the spectral index as a function of the scattered photon energy for two representative values of @xmath150 ( with @xmath15=2.5 and @xmath4=1 kev ) .
our calculations show that at low energies the spectral index is much steep in the case of the smaller @xmath150 and starts to converge toward the classical ultra
relativistic limit ( @xmath158=0.75 in this example ) but the progressive kn steepening of the spectra takes over before the ultra
relativistic regime is established .
when the kn regime is well established the spectral index is much steep in the case of the large @xmath150 .
by adopting the unification scheme linking radio loud quasars and fr ii radio galaxies ( barthel 1989 ) , it has been shown that the aic scattering of the nuclear photons by the relativistic electrons in the lobes may produce detectable x
ray fluxes well in excess of those calculated by the ic of the cmb photons ( brunetti et al.1997 ) . roughly speaking , the inverse compton emissivity is @xmath159 , @xmath160 and @xmath161 being the distance from the hidden quasar and the far ir
uv quasar luminosity respectively . by constraining @xmath161 ,
a comparison of the x ray properties with the radio synchrotron emission can provide important information on the magnetic field strength and relativistic particle distribution in the lobes . in a recent rosat hri observation of the fr ii radio galaxy 3c 219 brunetti et al.(1999 )
have found extended non thermal emission from the lobes most likely due to aic of nuclear photons by relativistic electrons .
it was found that the average magnetic field strength of 3c 219 is a factor @xmath1623 lower than that computed under equipartition assumption . in order to illustrate the possibility of obtaining detailed physical information from x ray observations , which will be possible with the future chandra and xmm missions , in this section we apply the general aic equations of sect.4 to a specific model .
we assume that the far ir optical spectrum of a typical quasar hidden in a powerful radio galaxy is represented by four power laws ( @xmath163 ) with @xmath158 = 0.2 , 0.9 , 1.7 , and 0.6 respectively in the intervals 10050 , 506 , 60.65 , and 0.650.35 ( as in brunetti et al.1997 ) .
since the uv photons may be important in the ic production of the hard x and gamma
rays , in addition we assume a uv spectrum simply modelled by a single power law with @xmath158 = 0.0 in the 0.350.03@xmath164 m band , consistently with walter & fink ( 1994 ) findings .
we further assume that the relativistic electrons are injected in the radio volume , approximated with a prolate ellipsoid of @xmath165 kpc semiaxis , with an energy differential spectrum @xmath166 ( the calculated aic fluxes are typically contributed by electrons by @xmath167 ) and that the magnetic fields , of average constant intensity @xmath138 , and particles momenta are randomly distributed on a sufficiently small scale
. the time evolution of the particle spectrum , injected at the hot spots and no reacceleration , due to synchrotron and compton losses may be described by ( jaffe & perola 1973 ) : @xmath168 being @xmath169 for @xmath170 and @xmath171 the break energy ( kardashev 1962 ) .
this is crucial in order to compare the predictions with detailed spatially resolved x ray spectroscopy .
deep radio observations of the hot spot regions of the radio galaxy cyg a have suggested the possibility of a low energy cut off , or a turn over in the particle distribution ( carilli et al .
we take into account this possibility by applying an exponential cut off to the power law spectrum below @xmath172 . in general
, @xmath171 is expected to be a function of the position @xmath160 in the radio galaxy . in a very simple model ,
in which the radio hot spots separate with constant velocity ( @xmath173 ) and the particles remain approximatively in the same place in which they are injected , the age of the relativistic particles is @xmath174 , d being the distance of the hot spot ( for which @xmath175=0 ) from the nucleus . at variance with the previous literature , in addition to the standard synchrotron and compton ( with the cmb photons ) losses , we consider also the radiative losses due to the aic scattering of the nuclear photons . with these assumptions the break energy of the time
/ spatially evolved electron population in the radio galaxy is : @xmath176 where @xmath177 is the equivalent magnetic field associated to the cmb and @xmath178 is the luminosity of the hidden quasar in units of @xmath179erg s@xmath180 .
the aic scattering with the nuclear photons dominates the radiative losses at relative small distance from the nucleus .
this distance depends on the quasar luminosity and on the magnetic field strength in the lobes ( fig.13 ) . due to the decrease of the break energy with decreasing distance from the nucleus ,
the innermost regions of the radio volume contribute less to the radiation spectrum of the radio galaxies at high frequencies .
spatially resolved spectroscopy in the x ray domain can provide important information about @xmath181 , especially in the innermost regions of the radio galaxies .
the break energy can also be constrained by deep radio spectroscopy .
however , it should be noticed that in the innermost parts , where @xmath182 , the expected repentine decrease of the radio brightness makes these investigations very difficult .
if the source is well resolved both in the radio and in the x ray bands , then @xmath161 , @xmath138 and @xmath183 can in principle be estimated by comparing the radio and the x ray spectral indices with the model predictions . in fig.14 , for a given radio galaxy model
, we have reported the spectral index from the aic scattering of the nuclear photons at different distances from the nucleus .
14 is obtained by integrating eq.(31 ) over the electron distribution ( eq.44 ) , with @xmath171 given by eq.(45 ) , and over the assumed spectrum of the nuclear photons . for simplicity , the axis of the quasar radiation cone , of half opening angle 45@xmath184 ,
is assumed to be coincident with that of the radio source and to lie on the plane of the sky . in order to give theoretical predictions to be compared with the future
x ray observations , a typical radio galaxy of total size 200 kpc is placed at a distance ( z=0.15 ) such that its angular size is much larger than the chandra psf . since the 0.210 kev luminosity of a strong fr ii radio galaxy
is expected to be close to @xmath185erg s@xmath180 , the estimated chandra count rate from the two regions indicated in fig.14 is @xmath186 counts s@xmath180 ( @xmath187=75 km s@xmath180 mpc@xmath180 , @xmath188=0.0 ) .
the complex spectral index behaviour is due to the interplay of the ir and uv bumps with the time / spatially evolved particle energy distribution .
the x ray spectral index of the innermost regions of the radio galaxies is significantly steeper than those pertaining to the extended regions and to the radio synchrotron spectral index of the high brightness lobes .
we have also computed the x ray spectral index and emitted power by imposing a low energy cut off in the particle energy distribution ( dashed lines ) . in this case the soft x
ray spectral index ( 0.1 1 kev ) is expected to be considerably harder than the synchrotron spectral index ( = 0.75 ) . as a consequence
detailed spatially resolved spectroscopy provided by chandra may yield unprecedented information about the energy distribution of the relativistic particles in the low energy portion of the spectrum . if the radio galaxy is inclined with respect to the plane of the sky
the nuclear photons in the far lobe are on average scattered toward the observer at angles larger than those in the near lobe .
since in the case of aic scattering the typical @xmath0 of the scattering electrons depends on the scattering angle ( eq.34 ) , one finds that in the far lobe the nuclear photons are scattered at a given @xmath9 by less energetic electrons than in the near lobe . as a consequence ,
depending on the inclination of the radio galaxy on the sky plane , all the spectral features in fig.14 would appear shifted at lower energies in the case of the near lobe and at larger energies in the case of the far lobe .
typically this results in a spectrum of the far lobe significantly harder than that of the near lobe . in the case of distant ( or faint ) radio galaxies
spatially resolved spectroscopy will not be possible .
however , due to their large effective area and spectral resolution , chandra and xmm could be used to study the integrated spectrum .
the total emitted power is obtained by integrating the emissivity over the emitting volume . the contribution from the ic scattering of the cmb photons should be taken into account in the case of high redshift and/or very extended radio galaxies .
the radio volume can be modelled in detail when one deals with a specific radio galaxy . in figs.14 and 15 the calculated aic spectral index and emitted power ( in arbitrary units ) from powerful radio galaxies
are shown from the soft x
rays to the gamma
rays for a number of parameters .
the spectral index from unresolved radio galaxy is close to the radio spectral index in the soft x
ray band , being steeper at higher energies . in the case of a powerful radio galaxy at a redshift @xmath189=0.5 with radio spectral index
@xmath158= 0.75 the expected spectral index in the xmm band would be @xmath190= 0.750.90 depending on the parameters ( i.e. quasar luminosity , age , magnetic field strength ) .
if the radio galaxy is inclined with respect to plane of the sky the spectrum from the far lobe would be harder than that of the near lobe . since the far lobe is also expected to be the most luminous ( brunetti et al.1997 ) , then the integrated spectrum from an unresolved radio galaxy would be harder with increasing inclination . from fig.16
we also notice that a large fraction of the energy emitted by the aic scattering of the far ir uv nuclear photons is channeled in the gamma ray band .
this will provide a new class of extragalactic gamma
ray sources to be revealed by future high energy experiments such as glast .
the anisotropic inverse compton ( aic ) scattering problem has been solved from the trans
relativistic to the ultra
relativistic regime , without introducing any approximation .
the equations are integrated over the energies of the scattering electrons to give the spectrum from the aic scattering of an unpolarized monochromatic photon beam by a beam of unpolarized relativistic electrons . in the case of a power law energy distribution of the electrons , we find that the spectrum is inverted peaking toward high energies before being sharply cut off at the emitted energies at which the highest energetic electrons ( @xmath191 ) are involved in the scattering process . the emitted power and spectrum in the thomson approximation is compared with those in the klein nishina regime . in the ultra
relativistic thomson approximation the radiation scattered in the direction of the electron beam has a power law spectrum with slope @xmath192 .
we find that in the mildly
relativistic case it is much harder at large scattering angles and much softer at small scattering angles .
the aic scattering between unpolarized isotropic electrons and a monochromatic photon beam is discussed from the trans
relativistic to the ultra relativistic case . in agreement with the previous studies
, we find that , toward the ultra
relativistic limit , the distribution of the scattered power is highly anisotropic for the astrophysically common case ( where the differential electron spectral index @xmath193 2 ) and that the spectrum has the same shape of the isotropic ultra relativistic case .
in addition to previous studies we find that in the mildly relativistic case the spectrum emitted at large scattering angles is considerably harder than the classical ultra relativistic one being softer at small scattering angles .
furthermore , we show that the emitted power in the klein nishina phase is distributed much more isotropically with increasing energy of the scattered photons .
we have also derived simple formulae that give the emissivity from the ultra
relativistic case down to @xmath194 .
in sect.5 we have applied the general aic equations derived in sect.4 to the calculation of the emitted power and spectrum from the scattering of the far ir
uv nuclear photons from a hidden quasar by the relativistic electrons in the lobes of powerful fr ii radio galaxies . in our simplified model
we have taken into account the effect of the radiation losses on the energy distribution of the relativistic electrons . due to the losses the calculated spectrum steepens with decreasing distance from the nucleus where the aic brightness increases .
this may provide an important test of the model to be performed by future x ray satellites such as chandra and , conversely , it may provide extremely important information on the magnetic field strengths and on the time evolution of the relativistic particles in the fr ii radio galaxies . in the case of resolved radio galaxies inclined with respect to the plane of the sky , due to the anisotropic illumination from the nuclear photons
, we predict that the x ray spectrum of the far lobe is on the average harder than that of the near one ; since the far lobe is expected to be the more luminous , also the spectrum of unresolved sources will be harder with increasing inclination .
i am indebted to prof .
g.setti for many discussions and for suggestions on the manuscript .
i would like to thank prof .
a.treves for providing useful references and for useful discussions .
i also thanks the anonymous referee for helpful comments that have improved the presentation of the paper .
this work was partly supported by the italian ministry for university and research ( murst ) under grant cofin98 - 02 - 32 .
in section 4.1 we have derived the thomson aic emissivity by an isotropic electron population with energy distribution @xmath195 ( eqs.3539 ) . in several astrophysical situations
@xmath48 may be more complicated than a simple infinite power law ( e.g. sect.5 ) . in these cases
the thomson aic emissivity is given by : @xmath196 { \cal i}_{5/2 } \nonumber\\ + { 5\over 2}(1-k_3)^{5 } [ 3(1+{{\epsilon_0}\over{\epsilon_1}})+({{\epsilon_1}\over { \epsilon_0}}+ ( { { \epsilon_0}\over{\epsilon_1}})^2 ) ] { \cal i}_{7/2 } \}\end{aligned}\ ] ] with @xmath197 @xmath198 @xmath199 @xmath200 which , in general , can be readily numerically calculated with @xmath136 given by eq.(34 ) .
it is well known that fermi acceleration mechanisms lead to an energy distribution of the relativistic particles that is a power law in momentum and not in @xmath0 . as a consequence , in some astrophysical situations in which trans and mildly relativistic aic scattering produces radiation at relatively large scattering angles ( typically @xmath202 ) one should use @xmath203 . | the problem of the anisotropic inverse compton scattering between a monochromatic photon beam and relativistic electrons is revisited and formally solved without approximations .
solutions are given for the single scattering with an electron beam and with a population of electrons isotropically distributed , under the assumption that the energy distribution of the relativistic particles follows a simple power law as it is the case in many astrophysical applications .
both the thomson approximation and the klein - nishina regime are considered for the scattering of an unpolarized photon beam .
the equations are obtained without the ultra
relativistic approximation and are compared with the ultra
relativistic solutions given in the literature .
the main characteristics of the power distribution and spectra of the scattered radiation are discussed for relevant examples . in the thomson case for an isotropic electron population a simple formula holding down to the mildly
relativistic energies is given .
as an application the formulae of the anisotropic inverse compton scattering are used to predict the properties of the x and @xmath0ray spectra from the radio lobes of strong fr ii radio galaxies due to the interaction of the relativistic electrons with the incoming photons from the nucleus .
the dependence of the emitted power on the relativistic electron energy distribution and on its evolution with time is discussed .
_ pacs _ : 95.30.cq ; 95.30.gv ; 98.38.am ; 98.54.gr ; 98.70.qy 27a 40l elementary particle processes radiation mechanisms physical properties radio galaxies x - ray sources |
in a recent paper , werner @xcite has shown there exists a measure on simple loops on any riemann surface which has the property of conformal restriction .
this is to say that if @xmath4 are any two subdomains of the manifold , the measures on loops in @xmath5 obtained by ( a ) restriction of the measure on loops in @xmath6 to those in @xmath5 , and ( b ) conformally mapping @xmath7 , are the same .
moreover this measure is unique up to multiplication by a constant .
we refer to these loops throughout the paper as self - avoiding loops and to the mass of any subset under werner s measure as the @xmath8-mass .
one may also consider the set of self - avoiding polygons on some regular lattice embedded in the manifold .
the total number of such polygons of length @xmath9 is known to grow as @xmath10 where @xmath8 is lattice dependent .
the measure which weights each polygon with a factor @xmath11 ( where @xmath12 ) has the restriction property , and is commonly conjectured also to be conformally invariant in the limit of vanishing lattice spacing , which means that it should give a particular example of werner s measure . in this paper
we assume this to be true , and hence conjecture values of the @xmath8-mass of certain loop subsets , using ( non - rigorous ) coulomb gas and cft methods applied to a generalised model , the o@xmath13 model .
the o(@xmath0 ) model encompasses a large group of physically interesting models , including the ising spin model , the percolation problem and self - avoiding loops .
the theory may be written as a spin model with nearest neighbour interactions , or alternatively in the loop gas picture , in which the states of the model are non - intersecting loops constructed on the edges of the spin model .
each loop is weighted with a factor @xmath14 , where @xmath15 is related to the reduced coupling in the model and @xmath9 is the length of the loop . in the ising model , these loops are the cluster boundaries of a spin model defined on the dual lattice . at the critical point of the model , @xmath16 ,
it is conjectured to flow under the renormalization group to a gaussian free field @xcite , which is a well known example of a conformal field theory ( cft ) .
one of us @xcite showed the existence of operators in this theory whose two point correlation functions count loops around one of two points with a weight @xmath17 rather than @xmath0 , hence giving information about the distribution of loops .
the choice @xmath18 is of particular interest ; these operators are then called twist operators and are at @xmath19 in the kac classification @xcite .
hence , they have null states at level two .
the two point correlation function is then equivalent to the expectation value of @xmath20 in the loop ensemble , where @xmath21 is the number of intersections of loops with a defect line , a continuous curve connecting the two points .
the twist operators may therefore be thought of as a source and a sink for this defect line .
self - avoiding loops are described by the loop gas picture of the o(@xmath0 ) model with @xmath22 .
the partition function and correlation functions are determined ( to order @xmath23 ) by graphs with just a single loop .
the two point correlation function of twist operators therefore leads to an analytic expression for the number of single loops which separate the locations of the operators , weighted by @xmath11 as in figure [ figintrotwopoints ] .
this number is logarithmically divergent in the continuum limit , however , due to the contribution from vanishingly small loops around each of the points . to the power of their length .
notice that the sum contains loops with vanishingly small perimeter in the continuum limit.,scaledwidth=60.0% ] in this paper , we consider the four point correlation function of these twist operators on the riemann sphere . in the loop gas picture
, we may think of each operator as again being a source ( or sink ) of a defect line .
hence there is a defect line running between each pair of points .
we argue that the choice of paths for the defect lines is unimportant and that they may run between any two distinct pairs of points . the four point correlation function
can then be shown to be the expectation value of @xmath20 where @xmath21 is now the total number of crossings of loops with defect lines .
particular semi - connected parts of this four point function yield the expected numbers of weighted loops ( the @xmath8-mass of loops ) which wind around two of the four locations of the operators , as in figure [ figintrofourpoints ] .
cft may be used to derive the form of the four point function , since the null states of the operators imply that the correlation functions obey a set of belavin , polyakov and zamalodchikov ( bpz ) type partial differential equations @xcite .
we solve these equations and hence find analytic expressions for the @xmath8-mass of loops which separate one pair of points from the other . for loops which separate @xmath24 from @xmath25 , for example
, we obtain the expression @xmath26 is the cross - ratio of the points .
these expected numbers of weighted loops are finite in the continuum limit because there is no contribution from vanishingly small loops , and are invariant under conformal transformations .
we show that the non - leading behaviour of these expressions as @xmath27 reveals the derivative of the central charge with respect to @xmath0 at @xmath28 .
this is consistent with the interpretation @xcite of the stress tensor as the spin - two component of the relative probability that a loop passes between two points in the limit @xmath29 .
we use similar arguments for the o(@xmath0 ) model in a simply connected domain .
such a domain may always be mapped via a conformal transformation to the upper half plane with the real axis as the boundary . the two point function in this domain satisfies the same partial differential equations as the four point function in the bulk , with the locations of the two operators and the reflections of these points in the real axis being the locations of the four operators in the bulk theory .
we show that the connected two point function of twist operators in the @xmath22 limit counts the mass of loops around the locations of both operators , and find the explicit expression @xmath30 where @xmath31 is now the cross ratio of the two points together with their conformal images in the boundary . to the power of their length .
note that , in contrast to the two point correlation function above , there are no vanishingly small loops contributing to the sum .
, scaledwidth=60.0% ] the layout of this paper is as follows . in the next section ,
we recall the arguments leading to the coulomb gas picture of the o(@xmath0 ) model . in section [ sectwistoperators ]
we discuss twist operators in the o(@xmath0 ) model and derive the form of the two point correlation function of these operators .
the small @xmath0 limit of the theory is explored and the correlation function in this limit is shown to depend ( to order @xmath23 ) only on the configurations of a single loop . in section [ sec4ptfncft ]
we discuss the four point correlation function of twist operators . because the twist operators have null states at level two , this correlation function satisfies a set of partial differential equations , which we solve analytically .
the interpretation of the four point function in terms of a pair of defect lines is explained , followed by the example of the four point function of twist operators as the four point function of spin operators in the ising model
. section [ sec4ptinterpretation ] is devoted to the small @xmath0 limit of the four point function and contains a derivation of the mass of loops winding around two of the four points .
we then explain how these numbers , as functions of the positions of the four points , hold a key to measuring the effective central charge of the o(@xmath0 ) model as @xmath22 . in section [ sec2ptbcft ]
we apply the theory of twist operators to the o(@xmath0 ) model in a simply connected domain and show that the mass of loops around two points in such a domain is invariant under conformal transformations and finite in the limit of vanishing lattice spacing .
finally , in section [ secstochasticprocess ] we show that the mass of loops around two of four points on the riemann sphere can be thought of in terms of an sle@xmath32 process .
let us start with the following partition function for the o(@xmath0 ) model @xcite : @xmath33 the @xmath34 are @xmath0-component spins on a lattice @xmath35 and the product is over pairs of nearest neighbours .
the product in the partition function can be expanded into a sum of @xmath36 terms , where @xmath21 is the number of nearest neighbours .
each term can be represented by a graph of open and closed loops in the following way : the neighbouring sites , @xmath37 and @xmath38 , are joined together by a line if the @xmath39 term was chosen in the expansion of equation ( [ eqzon ] ) , or left disconnected if the number @xmath40 was chosen instead .
the trace over spins may then be done for each graph individually . since @xmath41 because of the symmetry under the transformation @xmath42 , all graphs containing an odd number of @xmath34 make no contribution to the partition function .
these are the graphs with open loops , hence the partition function will only contain contributions from graphs with closed loops .
let us consider the honeycomb lattice for simplicity , then there may be 0,1,2 or 3 powers of a given @xmath34 .
the trace over spins of @xmath43 will contribute a factor of 1 , whereas , from @xmath44 , instances of @xmath45 lead to a factor of @xmath0 for each closed loop in the graph .
the partition function is therefore equivalent to @xmath46 where @xmath9 is the total length of all the loops in a given graph , @xmath47 is the number of closed loops and the trace is now over all graphs of closed , non - intersecting loops .
this form of the partition function may be used for general values of @xmath0 , whereas equation ( [ eqzon ] ) is applicable only to positive , integer @xmath0 .
there exists a critical point in the theory , at @xmath48 , where the mean loop length diverges and the model is supposed to become conformally invariant . in order to formulate a field theory , the non - local factors of @xmath0 must be made local .
this can be done by assigning orientations to the loops ( either clockwise or anticlockwise ) and inserting local factors of @xmath49 at each vertex where the curve turns to the right and @xmath50 where the curve turns to the left .
summing over the two possible orientations for each loop leads to a contribution of @xmath51 from each closed loop on the honeycomb lattice , where closed loops have a difference in the number of left and right turns of six with the sign depending on the orientation of the loop . in order to obtain the desired factor of @xmath0 for each closed loop , @xmath52
must therefore be chosen such that @xmath53 this may now be transformed into a height model .
the height variables are taken to be integer multiples of @xmath54 and are assigned to sites on the dual lattice , such that the loops are contours of the landscape . crossing a loop running from left to right leads to a decrease in the height of @xmath54 , whilst crossing a loop running from right to left leads to an increase in the height of @xmath54 .
a given configuration of the heights corresponds to a unique graph of orientated loops .
the assumption of the coulomb gas is that this height model flows under the renormalisation group ( rg ) into a free field theory with action @xmath55=\frac{g}{4\pi}\int{(\partial h(\textbf{r}))^2 d^{2}\textbf{r}}\,,\ ] ] for some @xmath56 .
there is an additional subtlety regarding topology , which may be seen by considering the model defined on the cylinder . on a cylinder of circumference
@xmath57 , the loops wrapping around the circumference are counted incorrectly ( for @xmath58 ) .
loops wrapping around have the same number of left turns as right turns .
these loops are therefore not counted correctly . the situation is remedied by placing a vertex operator ( also known as an electric charge ) @xmath59 at one end of the cylinder and an operator @xmath60 at the other end .
let these points be @xmath61 .
if there is a single loop wrapping around the cylinder , then there will be a height difference of @xmath54 between the ends , the sign of which depends on the orientation of the loop .
summing over the two possibilities for the orientation then leads to the required contribution of @xmath62 .
these charges at the ends of the cylinder lead to a modification of the partition function to @xmath63 where the abbreviation ffc signifies that the above expectation values and partition functions are for the free field theory on the cylinder and where we have used @xmath64 .
the free energy per unit length on the cylinder for this coulomb gas ( cg ) partition function may be calculated as @xmath65 this is the form of the free energy on the cylinder obtained via the schwartzian derivative from a theory on the plane with central charge @xmath66 @xcite given by @xmath67 where the first term comes from the known behaviour of @xmath68 @xcite .
all that remains is to fix the value of @xmath69 , which may be obtained from the following argument : adding a term @xmath70 to the action should not affect the critical behaviour , since the height variables were defined to be integer multiples of @xmath54 and @xmath71 for all integer @xmath72 .
hence , this term should be marginal under renormalisation group flow and therefore must have scaling dimension @xmath73 .
this determines @xcite @xmath74
in the loop ( coulomb ) gas picture of the @xmath75 model , loops wrapping around the cylinder may be counted with a weight @xmath17 different from @xmath0 by placing additional charges @xmath76 at the ends of the cylinder , with @xmath77 chosen according to equation ( [ eqchin ] ) .
the scaling dimension of these operators may be obtained from their two point correlation function in the loop gas ensemble . placing the charges at @xmath78 , @xmath79 these free field expectation values
may be calculated explicitly since they are gaussian moments .
then , in the limit @xmath80 , the two point function is seen to be @xmath81 this expression is of the general form of a two point function of operators on the cylinder with scaling dimension @xcite @xmath82 a particular choice of interest is @xmath18 , for which the scaling dimension is @xmath83 ; this is the scaling dimension of @xmath84 operators with level two null states in the kac classification : @xmath85 these correspond , in string theory language , to operators which insert orbifold points corresponding to the global symmetry @xmath86 of the hamiltonian and are known as twist operators
. they will be the focus of this paper .
their correlation functions may be considered in geometries other than the cylinder .
for example , on the riemann sphere , the two point function of such operators is fixed by scale invariance to be @xmath87 where we have inserted explicit factors of the lattice spacing , @xmath88 , so as to make @xmath89 dimensionless . the two point correlation function in the loop gas
may also be calculated on the sphere using the ensemble of equation ( [ eqlooppartition ] ) .
each loop separating the points @xmath90 and @xmath91 will be counted with weight @xmath92 rather than @xmath0 , hence graphs in @xmath93 will be weighted with an additional factor of @xmath94 to the power of the number of loops separating the two points .
graphs in @xmath93 with an odd number of loops separating point @xmath90 from point @xmath91 will be weighted by an additional factor of @xmath95 , whilst those with an even number of loops will be unaffected , since @xmath96 .
these two different types of graphs can be separated in the sum over graphs as follows @xmath97\big / z_{\textrm{loop gas}}\,,\ ] ] where @xmath98 is the set of graphs of closed , non - intersecting loops with an odd number of loops separating point @xmath90 and @xmath91 and the @xmath99 is the set with an even number of loops separating the two points .
this two point correlation function has a natural interpretation in terms of a defect line joining points @xmath90 and @xmath91 .
if @xmath100 is the number of times loops cross the defect line in a given graph , then equation ( [ eq2ptloops ] ) may be rewritten as @xmath101 the defect line can take any path between the two points because , for a given loop configuration , @xmath102 will be the same for all paths .
this can be best understood with reference to figure [ figdefectline ] . with one loop ( solid line ) and
two choices of path for the defect line ( dashed line ) . in ( a ) @xmath103
so that @xmath104 . in ( b ) @xmath105 so that @xmath104 also . any choice of path for the defect line will lead to @xmath104 since @xmath100 is always odd .
, scaledwidth=70.0% ] combining equations ( [ eq2ptcft ] ) and ( [ eqloopsdefect ] ) , we find @xmath106 self - avoiding loops correspond to @xmath107 , which is the dilute phase of the @xmath22 limit of the @xmath75 model .
the results of section [ sectwistoperators ] are summarized by equation ( [ eq2ptinterpretation ] ) , which for @xmath108 may be expanded in powers of @xmath0 .
the order @xmath109 term in the loop gas expansion ( the left hand side of the equation ) comes from graphs with a single loop . by equating this with the order @xmath109 term of the right hand side
, we shall demonstrate a property of the sum over graphs with a single loop .
the left hand side of equation ( [ eq2ptinterpretation ] ) for @xmath108 becomes @xmath110 where @xmath111 is the set of all graphs with no loops and @xmath112 is the set of all graphs with one loop , hence their respective powers of @xmath113 and @xmath109 .
@xmath111 contains a single graph with no edges , hence the first term in both brackets is equal to @xmath40 .
the set @xmath112 contains graphs with only a single loop .
this set is composed of two subsets with no overlap , the set @xmath114 with a single loop separating the point @xmath90 from @xmath91 and the set @xmath115 which contains no loops separating the points .
it should be noted that a single loop surrounding the point @xmath90 is classified in the same group as a loop around @xmath91 since on the riemann sphere one can be continuously deformed into the other .
both have an odd number of intersections with a defect line between the two points , hence @xmath104 .
for all graphs belonging to @xmath115 , it may be seen that @xmath116 .
this is shown in figure [ fig2pt1loop ] . and one belonging to @xmath114 for a choice of @xmath90 and @xmath91 on the riemann sphere .
notice that a loop in @xmath114 can be thought of as being either around @xmath90 or around @xmath91 .
similarly , a loop in @xmath115 can be thought of as being around both points or neither .
, scaledwidth=70.0% ] equation ( [ eq2ptsmallnlhs ] ) can therefore be rewritten as @xmath117 equations ( [ eqxfnofchi ] ) , ( [ eqchin ] ) and ( [ eqgchi ] ) may be combined and then expanded in powers of @xmath0 to find the following form for the scaling dimension of the twist operators in the small @xmath0 limit : @xmath118 therefore , the expansion of the right hand side of equation ( [ eq2ptinterpretation ] ) is @xmath119 hence , from comparing the coefficients of @xmath109 in equations ( [ eq2ptsmallnloopexpansion ] ) and ( [ eq2ptcftsmalln ] ) , we may conclude that @xmath120 in words , for a given pair of points @xmath90 and @xmath91 , the number of weighted loops which separate the two points , weighted by @xmath11 , diverges as @xmath121 in the limit of the lattice spacing tending to zero .
the cause of this is the diverging contribution from vanishingly small loops in the continuum limit .
the factor @xmath11 may be thought of as the measure , hence this weighted number of loops is equivalent to the expected number of loops around one point .
it may be noted that the coefficient @xmath122 differs from that seen on the annulus with shrinking internal radius . in the case of the annulus with vanishingly small modulus @xcite a coefficient of @xmath123
is seen , because there are only diverging contributions from small loops around one point , as opposed to the plane where small loops around both points contribute .
twist operators were introduced in section [ sectwistoperators ] as the operators responsible for counting loops with weight @xmath92 rather than @xmath0 in the loop gas picture of the @xmath75 model .
their scaling dimension was calculated and it was seen that it corresponds to the scaling dimension of operators with null states at level two .
conformal field theory may be used to derive a set of partial differential equations satisfied by the correlation functions of such null state operators . in the case of the four point function
, these differential equations may be solved analytically , as will be seen in this section . to make contact with the language of schramm ( stochastic)-loewner evolution ( sle ) ,
@xmath69 is dropped in favour of the parameter @xmath124 used in sle @xcite .
the two are related by the formula @xmath125 .
it was shown in section [ sectwistoperators ] that the twist operators have a null state at level two @xcite .
this null state is itself a highest weight state ; it is annihilated by all raising operators , @xmath126 with @xmath127 .
this is not the same as saying that the null state decouples from all other states in the theory , as it does in unitary theories .
that said , it can be shown by a modular invariance argument that physical results are seen only if the null state decouples on the torus .
this is because , barring unforeseen cancellations , the nondecoupling of the null states would lead to a density of states at high energies corresponding to @xmath128 rather than @xmath129 @xcite .
we shall therefore set the null state to zero in this theory and use the resulting partial differential equations for the correlation functions of the twist operators . in the complex plane ,
the four point function satisfies the following set of partial differential equations ( for @xmath130 ) @xcite : @xmath131{\langle}\phi({z_{1}},{\overline{z_{1}}})\phi({z_{2}},{\overline{z_{2 } } } ) \phi({z_{3}},{\overline{z_{3}}})\phi({z_{4}},{\overline{z_{4}}}){\rangle}=0\label{eq4ptbpzholomorphic}\\ & \big[{\partial_{\overline{{z_{j}}}}}^2-\frac{\kappa}{4}\sum_{i\neq j}{\frac{\overline{h_{2}}}{({\overline{z_{i}}}-{\overline{z_{j}}})^2}-\frac{{\partial_{\overline{{z_{i}}}}}}{{\overline{z_{i}}}-{\overline{z_{j}}}}}\big]{\langle}\phi({z_{1}},{\overline{z_{1}}})\phi({z_{2}},{\overline{z_{2 } } } ) \phi({z_{3}},{\overline{z_{3}}})\phi({z_{4}},{\overline{z_{4}}}){\rangle}=0.\label{eq4ptbpzantiholomorphic}\end{aligned}\ ] ] in terms of @xmath132 , the cross ratios are defined as @xmath133 these cross ratios are invariant under global conformal transformations .
the partial differential equations above have the solutions @xmath134 where @xmath135 is @xmath136 @xmath137 and @xmath138 is the value of the scaling dimension of an operator with a null state at level three .
the form of @xmath139 can be derived by requiring consistency under the permutation of the labels @xmath140 and is found to be @xcite @xmath141 { \gamma}(\frac{3\kappa}{4}-1)^{2}}{{\gamma}(\frac{\kappa}{2})^2{\gamma}(\frac{\kappa}{2}-1)^2{\gamma}(1-\frac{\kappa}{4})^2}\,.\end{aligned}\ ] ] the form of @xmath142 is dependent on the normalisation of the two point function since , as @xmath143 , the four point function becomes @xmath144 recall from section [ sectwistoperators ] that the correlation function of two @xmath89 operators has an interpretation in terms of a defect line in the loop gas ensemble .
the operators are the source and sink of the defect line , which can take any path between the two points .
each graph in the sum is weighted by an additional factor @xmath102 where @xmath100 is the number of intersections of loops in that graph with the defect line .
there is a similar interpretation of the four point function in the loop gas ensemble , but now there are two defect lines .
the four points are split into two pairs and defect lines run between the points in each pair . for the choice of pairing ( @xmath145 ) with ( @xmath146 ) and ( @xmath147 ) with ( @xmath148 )
, the four point function can be seen to correspond to the following expectation value in the loop gas picture @xmath149 where @xmath100 is the number of crossings of loops across a defect line from @xmath150 to @xmath151 and @xmath152 is similarly defined .
there are three ways in which the pairs may be chosen , but each choice leads to the same set of weights for the graphs in the partition function .
the reason for this may be seen in figure [ figdeformdefects ] .
thus , the four point function may also be considered to be the expectation value in the loop gas ensemble of @xmath153 or @xmath154 . in the loop gas partition function
, a pair of defect lines leads to a factor @xmath155 and hence makes no contribution to the weight of any graph .
hence , the defect lines shown in ( b ) are equivalent to the choice in ( c ) .
a similar construction may be used to show the equivalence to a pair of defect lines drawn from @xmath90 to @xmath91 and from @xmath156 to @xmath157 .
note that this feature is a result of the choice @xmath18 in section [ sectwistoperators ] .
, scaledwidth=100.0% ] comparing equations ( [ eq4ptcftresult ] ) and ( [ eq4ptloopresult ] ) leads to the main result of this section @xmath158 where @xmath159 is defined below equation ( [ eq4ptcftresult ] ) in section [ sec4ptfncft ] .
boundaries between ising spin clusters correspond to the loops of the @xmath75 model at @xmath160 , or @xmath161 .
the twist operators in this case are equivalent to magnetisation operators because the parity of the number of loops separating two spins depends on whether they are parallel or anti - parallel .
that is to say that @xmath162 similarly , the four point correlation function of the twist operators is the four point function of the magnetisation operators .
the following function is the result of substituting @xmath161 into the results of section [ sec4ptfncft ] @xmath163 equation ( [ eqising4ptresult ] ) is the well known four point correlation of spin operators in the ising model at criticality @xcite .
in section [ sectwistoperators ] , we examined the small @xmath0 limit of the two point function of twist operators , equation ( [ eq2ptinterpretation ] ) .
we found an exact expression for the @xmath8-mass of loops surrounding one of the two points .
this result is summarised by equation ( [ eq2ptweightresult ] ) . a similar expansion to order @xmath109 of the four point correlation function ( equation [ eq4ptinterpretation ] ) yields information about the configurations of a single self - avoiding loop around four points . from the equations in section [ sec4ptfncft ] ,
the following small @xmath0 expansions may be deduced @xmath164 the expansion of @xmath165 around @xmath28 is @xmath166 where @xmath167 is a constant independent of @xmath0 , since the correlation function should tend to @xmath40 as @xmath22 .
the taylor series expansion of the right hand side of equation ( [ eq4ptinterpretation ] ) is then @xmath168 ) may also be expanded in powers of @xmath0 .
the graphs up to order @xmath109 in @xmath93 may be split into the graph with no loops @xmath169 and eight distinct sets of graphs with a single loop , as will be shown in the next section . in this section
, we expand the left hand side of equation ( [ eq4ptinterpretation ] ) for @xmath108 to order @xmath109 . the coefficient of @xmath109
may then be equated with the result for expansion of the right hand side , described in the previous section and leading to equation ( [ eq4ptsmallncft ] ) .
recall that the partition function for the loop gas expansion is equation ( [ eqlooppartition ] ) .
the sum may be decomposed into a term of order @xmath170 coming from the graph with no loops ( hence the zeroth power of @xmath0 ) and graphs with a single loop weighted by @xmath109 .
other graphs will contribute terms of order @xmath171 or smaller .
the graphs with a single loop on the riemann sphere may be further split up into eight distinct subsets , which are shown in figure [ fig4pts1loop ] .
each configuration has a unique set of @xmath172 where @xmath173 and @xmath174 .
let @xmath57 be defined as @xmath175 the numerator of @xmath57 can be calculated as @xmath176+o(n^2)\,.\end{aligned}\ ] ] the denominator takes the same form as the expression above , but without the factors of @xmath177 and @xmath94 .
for ease of notation , define @xmath178 and similar expressions for the other graphs in figure [ fig4pts1loop ] .
the elements of @xmath179 are sums over all loops of a single configuration type @xmath37 , weighted by @xmath180 to the power of their length .
they therefore represent the @xmath8-masses of loops of configuration @xmath37 .
it may then be seen that @xmath181+o(n^2)\label{eq4ptcorrfnwexpansion}\end{aligned}\ ] ] the two point functions @xmath182 and products of pairs of two point functions can be expanded in terms of the @xmath179 also : @xmath183 thus , it can be seen from equations ( [ eq4ptcorrfnwexpansion]),([eqc12c34wexpansion]),([eqc13c24wexpansion ] ) and ( [ eqc14c23wexpansion ] ) that a subset @xmath184 of the @xmath179 can be expressed in terms of partly - connected four point functions : @xmath185 these @xmath184 correspond to graphs with a loop winding around two of the four points and hence do not include contributions from vanishingly small loops in the continuum limit . by comparison with the small @xmath0 expansions of @xmath57 ( section [ sec4ptsmalln ] ) and @xmath186 ( section [ sec2ptsmalln ] with the same choice for the normalisation of the twist operator as in the four point function ) from conformal field theory , these weights are the following functions of @xmath187 @xmath188 where @xmath189 above are finite in the continuum limit of vanishing lattice spacing .
in fact , the expressions in equations ( [ eqw14cft]),([eqw24cft ] ) and ( [ eqw34cft ] ) are independent of @xmath88 .
they are finite and non - zero in the limit @xmath22 and are invariant under conformal transformations , being functions only of the cross ratios .
these expressions for the @xmath8-masses @xmath184 constitute one of the main results of this paper .
it is a standard result of conformal field theory that , given the explicit form of any four - point function , the central charge @xmath66 may be determined .
this is because the operator product expansion ( ope ) of any scalar primary operator @xmath89 with itself takes the form @xmath190 where @xmath191 is the holomorphic component of the stress tensor .
the coefficient follows from consideration of the limit @xmath29 in the three point function @xmath192 and the two - point function @xmath193 .
if this result is applied to the explicit expression ( [ eq4ptcftresult ] ) , we find that the coefficient of @xmath194 is @xmath195 and so obtain the known expression for the central charge in terms of @xmath124 @xmath196 this result is of interest in the limit @xmath197 for the light it sheds on the interpretation of the stress tensor @xmath191 in terms of an observable of a random curve given by doyon , riva and cardy @xcite . in that paper it was shown that , for any measure on simple random curves which satisfies conformal restriction , one may identify @xmath198 as being proportional to the spin-2 angular fourier component of the probability that the curve intersects a small line segment of length @xmath199 centred at @xmath200 .
while in that paper the focus was on curves which connect two points on the boundary of a simply connected domain , described in the case when conformal restriction holds by sle@xmath201 , the theorem should equally well apply to werner s measure on self - avoiding loops , if suitably re - interpreted in terms of @xmath8-masses rather than probabilities .
however our results in the present paper do not directly apply to the intersection with a line segment , but rather to the event that the loop passes ( or does not pass ) between pairs of points ( which , however , may be taken to mark the ends of the line segment ) .
nevertheless , one might expect these to differ only by a constant of proportionality , and indeed our results in this paper support this , and suggest that the two - point function of the object @xmath191 introduced in ref .
@xcite is indeed given by the central charge for small @xmath0 as expected .
indeed , let us consider the quantity @xmath202 as given in equation [ eqw14cft ] .
this is the @xmath8-mass of loops which separate @xmath203 from @xmath204 . in the limit @xmath29 ,
writing @xmath205 , we can define @xmath206 where the numerical prefactor is @xmath207 .
our result equation [ eq4ptcftresult ] then implies that @xmath208 has the same dependence on its arguments as does the o@xmath13 term in the cft correlation function @xmath209 this result generalises to other correlation functions , and implies that we may interpret the spin-2 component of the @xmath8-mass of loops which pass between two nearby points as being , in some sense , the derivative @xmath210 of the cft stress tensor with respect to @xmath0 at @xmath28 . with this definition
we then find the result @xmath211 as expected , where the left hand side is @xmath212 it would , of course , be important to establish this interpretation directly from the restriction property .
the model may also be considered in a simply connected domain .
all such domains may be conformally transformed to the upper half plane , with the real axis being the boundary .
we may therefore derive the results for the upper half plane .
the two point function of twist operators in the presence of this boundary satisfies the same set of partial differential equations as the four point function in the bulk , with @xmath156,@xmath157 assigned as the complex conjugates of the positions of the operators .
this is a well known result in the theory of boundary conformal field theory @xcite ( bcft ) and follows from the condition of @xmath213 on the boundary .
the solution for the two point function is @xmath214\label{eqbcftresult}\,,\end{aligned}\ ] ] where now there is only one cross - ratio @xmath215 @xmath139 can be determined by looking at the boundary conditions as the two points approach the boundary , which is @xmath216 .
equation ( [ eqbcftresult ] ) may be analytically continued to large @xmath31 : @xmath217\label{eqbcftlargeeta}\,.\end{aligned}\ ] ] the two natural choices of @xmath139 are those which pick out one or the other conformal block as the operators approach the boundary . in the limit @xmath22 , there are no loops in the loop gas picture so the correlation function should tend to unity for all @xmath31 .
this is only possible if the first term in the square brackets of equation ( [ eqbcftlargeeta ] ) is not present .
a ( non - unique ) choice of @xmath139 which satisfies this requirement is that which picks out the second conformal block only : @xmath218 it is important to note also that this boundary condition is not compatible with the vanishing of the four point function in the limit @xmath216 .
( see text).,scaledwidth=110.0% ] the form of @xmath142 is dependent on the choice of normalisation of the twist operators .
this can be seen from the @xmath143 limit of equation ( [ eqbcftresult ] ) : @xmath219 @xmath142 must be of the form @xmath220 due to the constraint that the four point function tends to unity as @xmath22 . for small @xmath0 ,
the correlation function may be expanded as @xmath221+o(n^2)\,,\end{aligned}\ ] ] where @xmath222 in analogy with the interpretation of the four point function in the bulk in the loop gas picture described in section [ sec4ptinterpretation ] , the following is the interpretation of the two point function in the presence of a boundary @xmath223 the left hand side is an expectation value in the ensemble of lattice loops , as before .
@xmath224 is the number of times loops cross a defect line from @xmath225 and @xmath226 .
the set of all graphs may be decomposed into the set with no loops and the set with one loop ; the other graphs are of order @xmath171 .
the possible configurations of a single loop with a boundary are shown in figure [ fig2ptboundary1loop ] .
the two and one point functions are then found from the loop gas to be @xmath227 just as for the case of the loop gas in the bulk , the @xmath8-masses @xmath179 are the number of weighted loops belonging to the configuration @xmath37 ( see figure [ fig2ptboundary1loop ] ) .
they are defined as @xmath228 of the four possible configurations , the only one finite in the limit @xmath229 is @xmath230 : @xmath231 in terms of @xmath31 , @xmath232 note , as for the bulk case , that the term involving @xmath233 in @xmath234 is cancelled by the subtraction of the product of one point functions .
we have seen in section [ sec4ptinterpretation ] that the @xmath8-mass of loops which wind around two of four points is invariant under conformal transformations . in this section ,
we show that the differential equations they satisfy can be interpreted in terms of an sle process .
recall that these functions were identified as semi - connected four point functions , for example equation ( [ eqw14fnlcs ] ) @xmath235\,,\end{aligned}\ ] ] where @xmath186 are the correlation functions of a pair of twist operators at the points @xmath236 and @xmath237 .
the two point function is fixed by scale invariance and the four point function was determined as a solution to the partial differential equations ( [ eq4ptbpzholomorphic ] ) and ( [ eq4ptbpzantiholomorphic ] ) . using these
, we find that @xmath238 satisfies the pde @xmath239w_{{z_{1}}{z_{4}}|{z_{2}}{z_{3}}}=\frac{1}{24\pi}\big [ \frac{1}{({z_{4}}-{z_{1}})^2}+\frac{1}{{z_{3}}-{z_{1}}}\big(\frac{1}{{z_{3}}-{z_{4}}}+\frac{1}{{z_{2}}-{z_{3}}}\big)\nonumber\\ & + \frac{1}{{z_{2}}-{z_{1}}}\big(\frac{1}{{z_{2}}-{z_{4}}}+\frac{1}{{z_{3}}-{z_{2}}}\big ) + \frac{1}{{z_{4}}-{z_{1}}}\big(\frac{1}{{z_{4}}-{z_{3}}}+\frac{1}{{z_{4}}-{z_{2}}}\big ) \big]+o(n)\label{eqnwholo}\,,\end{aligned}\ ] ] together with a similar equation in which all the @xmath240s are replaced by @xmath241 .
note that in the bpz equations @xmath200 and @xmath242 can be taken as independent complex numbers , and it is only in applying these equations to physical quantities that one needs to impose reality conditions .
these equations have almost the second order linear form which would result from applying the it formula to a martingale of an sle process started at @xmath243 .
one difference is that they are complex .
we can obtain a real equation by , for example , taking the real part of the sum of the holomorphic and antiholomorphic equations ( the more general case will be discussed below ) , but note that when we do this we have the sum @xmath244 where @xmath245 and @xmath246 is the laplacian operator @xmath247 .
we also note that we can remove the inhomogenous part in ( [ eqnwholo ] ) by defining @xmath248\ ] ] then we can rewrite the real part of ( [ eqnwholo ] ) as @xmath249\widetilde w=\frac32\delta_{z_1 } \widetilde w\,.\label{eqwpde}\end{aligned}\ ] ] consider now a sequence of conformal maps @xmath250 which satisfy the stochastic chordal loewner equation @xmath251 where @xmath252 . then if @xmath253 were a martingale , its expectation value would satisfy ( [ eqwpde ] ) with the right - hand side set equal to zero .
the coefficient of @xmath254 would in general be @xmath255 , so we should take @xmath3 . the right - hand side of ( [ eqwpde ] ) , together with the subtractions in ( [ eqnsubtract ] ) , therefore express the degree to which the weighted number of loops separating @xmath203 from @xmath204 fails to be a martingale . growing in the half - plane @xmath256 with its reflection in @xmath257,scaledwidth=80.0% ] eq .
[ csle ] describes a chordal sle growing into the half - plane @xmath256 , together with its exact reflection in the line @xmath257 as in figure [ figsle ] @xcite . for @xmath258 the hull , or outer perimeter , of this curve
encloses a growing region of the plane .
the entire interior of this hull at time @xmath259 is mapped into the point @xmath260 .
because of this , we would not expect the weighted number of loops to be a martingale .
given some initial population of loops , each time that part of a loop is swallowed by the growing hull , it disappears from the population .
the right - hand side of ( [ eqwpde ] ) must therefore express the rate at which this happens .
it is reasonable that this should be proportional to @xmath261 : if @xmath262 were constant in some region , this would mean there were no loops passing through it .
the leading non - zero rotationally invariant contribution should therefore be proportional to the laplacian . apart from this , the left - hand side expresses the fact that the @xmath8-mass of loops which separate @xmath263 and @xmath264 and which have not yet been ( partly or wholly ) swallowed by the hull is the same as that in the conformally equivalent case of loops which separate @xmath265 from @xmath266 .
this breaks down when the hull swallows any of the other three points .
we conjecture that the subtractions in ( [ eqnsubtract ] ) take this into account . of course
, this is only suggestive and a number of important issues would have to be resolved before one could actually derive our results from sle . in particular
, one should explain why it is necessary to consider sle@xmath267 rather than some other value of @xmath124 .
this is presumably related to the fact that the hull of sle@xmath267 corresponds locally to sle@xmath201 , and that both correspond to cfts with central charge @xmath268 .
the choice of a chordal sle reflected in @xmath257 is clearly arbitrary ; other choices correspond to taking different linear combinations of the holomorphic and anti - holomorphic equations . perhaps using radial or whole - plane sle would make the formula look more symmetrical .
in this paper we have studied scaling properties of loops in the loop gas picture of the o(@xmath0 ) model . in the picture ,
the partition function is a sum over all graphs of non - intersecting closed loops , weighted by @xmath269 where @xmath15 is a function of the reduced coupling and @xmath9 is the length of the loop .
we introduced twist operators , whose correlation functions count the loops separating the locations of the operators with weight @xmath17 different to the usual weight @xmath0 , or equivalently count the minimum number of crossings of defect lines running between these locations . for the particular choice @xmath18 , the twist operators have level two null states and their correlation functions satisfy bpz type partial differential equations on the riemann sphere .
thus , conformal field theory may be used to determine the analytic form of the two and four point functions . in the loop gas picture
, the choice @xmath18 means that loops are counted are weighted by an additional factor of @xmath1 to the power of the number of defect lines crossed , and the choice of path for the defect lines between the locations of the operators is unimportant .
the limit @xmath22 describes the theory of self - avoiding loops . in this limit ,
the partition function and correlation functions are dependent to first order in @xmath0 only on the configurations of a single loop .
hence , by equating particular semi - connected parts of the four point function calculated using conformal field theory to the result from the coulomb gas picture , we have deduced the expected number of weighted loops winding around two of the four points and shown that this number is invariant under conformal transformations .
this is presumably the mass of such a subset under the measure on loops introduced by werner .
other configurations of loops receive contributions from vanishingly small loops around a single point and are not finite in the limit of vanishing lattice spacing .
the central charge of the o(@xmath0 ) model for small @xmath0 may be found from the dependence of the @xmath8-mass of loops around two of the four points as a function of the cross ratio of the positions of the four operators .
the result @xmath270 from the analytic results agrees with other methods of calculating the central charge of the o(@xmath0 ) model , and lends support to the interpretation of the stress tensor for curves satisfying conformal restriction given in ref.@xcite .
a similar calculation was also carried out for the model defined in a simply connected domain .
a general simply connected domain may be mapped via a conformal transformation to the upper half plane with the real axis as the boundary .
the two point function of operators in the upper half plane satisfies the same partial differential equations as the four point function on the riemann sphere , with two additional operators positioned at the complex conjugates of the original operators . by again equating the results from conformal field theory with those from the coulomb gas picture , we have deduced the @xmath8-mass of loops around the two points in the upper half plane .
finally , we have shown that the differential equations satisfied by the above quantities should have a stochastic interpretation in terms of a chordal sle@xmath32 process starting from one of the points @xmath225 , along with its reflection in a fixed line .
20 werner w 2005 the conformally invariant measure on self - avoiding loops _ preprint
_ math.pr/0511605 nienhuis b 1987 _ phase transitions and critical phenomena _ vol 11 domb c and lebowitz j l eds .
( academic ) dotsenko vl s and fateev v a 1984 _ nucl .
b _ * 240 * 312 cardy j 2000 _ phys . rev . lett . _
* 84 * issue 16 3507 ( _ preprint _ cond - mat/9911457 ) di francesco p , mathieu p and senechal d 1997 _ conformal field theory _ ( springer ) belavin a a , polyakov a m and zamolodchikov a b 1984 _ j. stat .
* 34 * 763 doyon b , cardy j and riva v 2005 identification of the stress - energy tensor through conformal restriction in sle and related processes _ preprint _ math - ph/0511054 kondev j 1997 _ phys . rev
_ * 78 * 4320 ( _ preprint _ cond - mat/9703113 ) cardy j 1987 _ phase transitions and critical phenomena _ vol 11 domb c and lebowitz j l eds .
( academic ) cardy j 2006 the o(n ) model on the annulus _ preprint _ math - ph/0604043 cardy j 2005 _ annals of physics _ * 318 * 81 ( _ preprint _ cond - mat/0503313 ) cardy j 1986 _ nucl . phys . _
* b270 * 186 gradsteyn i s and rhyzhik i m 2000 _ table of integrals , series and products _
( academic press ) abramowitz m and stegun i a 1970 _ handbook of mathematical functions _ ( dover ) duplantier b and saleur h 1988 _ phys .
_ * 60 * 23 | the o(@xmath0 ) spin model in two dimensions may equivalently be formulated as a loop model , and then mapped to a height model which is conjectured to flow under the renormalization group to a conformal field theory ( cft ) . at the critical point , the order @xmath0 terms in the partition function and correlation functions describe single self - avoiding loops .
we investigate the ensemble of these self - avoiding loops using twist operators , which count loops which wind non - trivially around them with a factor @xmath1 .
these turn out to have level two null states and hence their correlators satisfy a set of partial differential equations . we show that partly - connected parts of the four point function count the expected number of loops which separate one pair of points from the other pair , and find an explicit expression for this .
we argue that the differential equation satisfied by these expectation values should have an interpretation in terms of a stochastic(schramm)-loewner evolution ( sle@xmath2 ) process with @xmath3 .
the two point function in a simply connected domain satisfies a closely related set of equations .
we solve these and hence calculate the expected number of single loops which separate both points from the boundary . |
recurrent neural networks are a prominent model for information processing and memory in the brain . @xcite .
traditionally , these models assume synapses that may change on the time scale of learning , but that can be assumed constant during memory retrieval .
however , synapses are reported to exhibit rapid time variations , and it is likely that this finding has important implications for our understanding of the way information is processed in the brain @xcite .
for instance , hopfield like networks in which synapses undergo rather generic fluctuations have been shown to significantly improve the associative process , e.g. , @xcite .
in addition , motivated by specific neurobiological observations and their theoretical interpretation @xcite , activity dependent synaptic changes which induce _ depression _ of the response have been considered @xcite .
it was shown that synaptic depression induces , in addition to memories as stable attractors , special sensitivity of the network to changing stimuli as well as rapid switching of the activity among the stored patterns @xcite .
this behaviour has been observed experimentally to occur during the processing of sensory information @xcite . in this paper , we present and study networks that are inspired in the observation of certain , more complex synaptic changes .
that is , we assume that repeated presynaptic activation induces at short times not only depression but also _ facilitation _ of the postsynaptic potential @xcite .
the question , which has not been quite addressed yet , is how a competition between depression and facilitation will affect the network performance .
we here conclude that , as for the case of only depression @xcite , the system may exhibit up to three different _ phases _ or regimes , namely , one with standard associative memory , a disordered phase in which the network lacks this property , and an oscillatory phase in which activity switches between different memories . depending on the balance between facilitation and depression , novel intriguing behavior results in the oscillatory regime .
in particular , as the degree of facilitation increases , both the sensitivity to external stimuli is enhanced and the frequency of the oscillations increases .
it then follows that facilitation allows for recovering of information with less error , at least during a short interval of time and can therefore play an important role in short term memory processes .
we are concerned in this paper with a network of binary neurons .
previous studies have shown that the behaviour of such a simple network dynamics agree qualitatively with the behaviour that is observed in more realistic networks , such as integrate and fire neuron models of pyramidal cells @xcite .
let us consider @xmath0 binary neurons , @xmath1 @xmath2 endowed of a probabilistic dynamics , namely,@xmath3 \right\ } , \label{s}\]]which is controlled by a _ temperature _
parameter , @xmath4 see , for instance , @xcite for details .
the function @xmath5 denotes a time
dependent _ local field _ ,
i.e. , the total presynaptic current arriving to the postsynaptic neuron @xmath6 this will be determined in the model following the phenomenological description of nonlinear synapses reported in @xcite , which was shown to capture well the experimentally observed properties of neocortical connections .
accordingly , we assume that @xmath7where @xmath8 is a constant threshold associated to the firing of neuron @xmath9 and @xmath10 and @xmath11 are functions to be determined which describe the effect on the neuron activity of short
term synaptic depression and facilitation , respectively .
we further assume that the weight @xmath12 of the connection between the ( presynaptic ) neuron @xmath13 and the ( postsynaptic ) neuron @xmath14 are static and _ store _ a set of patterns of the network activity , namely , the familiar _ covariance rule _ : @xmath15here , @xmath16 with @xmath17 are different binary patterns of average activity @xmath18 .
the standard hopfield model is recovered for @xmath19 , @xmath20 we next implement a dynamics for @xmath21 and @xmath22 after the prescription in @xcite . a description of varying synapses requires ,
at least , three local variables , say @xmath23 , @xmath24 and @xmath25 to be associated to the fractions of neurotransmitters in recovered , active , and inactive states , respectively .
a simpler picture consists in dealing with only the @xmath26 variable . this simplification , which seems to describe accurately both interpyramidal and pyramidal interneuron synapses , corresponds to the fact that the time in which the postsynaptic current decays is much shorter than the recovery time for synaptic depression , say @xmath27 @xcite ( time intervals are in milliseconds hereafter ) . within this approach
, one may write that @xmath28where@xmath29and@xmath30the interpretation of this ansatz is as follows .
concerning any presynaptic neuron @xmath31 the product @xmath32 stands for the total fraction of neurotransmitters in the recovered state which are activated either by incoming spikes , @xmath33 or by facilitation mechanisms , @xmath34 for simplicity , we are assuming that @xmath35 $ ] @xmath36 the additional variable @xmath37 is assumed to satisfy , as in the quantal model of transmitter release in @xcite ,
that@xmath38 s_{j}\left ( t\right ) , \label{u}\]]which describes an increase with each presynaptic spike and a decay to the resting value with relaxation time @xmath39 ( that is given in milliseconds ) consequently , facilitation washes out ( @xmath40 ) as @xmath41 and each presynaptic spike uses a fraction @xmath42 of the available resources @xmath43 the effect of facilitation increases with decreasing @xmath44 because this will leave more neurotransmitters available to be activated by facilitation .
therefore , facilitation is not controlled only by @xmath39 but also by @xmath45 the hopfield case with static synapses is recovered after using @xmath46 in eq.([d ] ) and @xmath47 in eq.([f ] ) or , equivalently , @xmath48 in eqs .
( [ x ] ) and ( [ u ] ) .
in fact , the latter imply fields @xmath49 so that one may simply rescale both @xmath50 and the threshold . the above interesting phenomenological description of dynamic changes has already been implemented in attractor neural networks @xcite for pure depressing synapses between excitatory pyramidal neurons @xcite .
we are here interested in the consequences of a competition between depression and facilitation .
therefore , we shall use @xmath51 and @xmath39 in the following as relevant control parameters .
let us consider the mean activities associated , respectively , with active and inert neurons in a given pattern @xmath52 namely,@xmath53it follows for the overlap of the network activity with pattern @xmath54 that@xmath55@xmath56 one may also define the averages of @xmath57 and @xmath58 over the sites that are active and inert , respectively , in a given pattern @xmath52 namely,@xmath59@xmath60 which describe depression ( the @xmath61s ) and facilitation ( the @xmath62s ) , each concerning a subset of neurons , e.g. , @xmath63 neurons for @xmath64 the local fields then ensue as @xmath65@xmath66 \,x_{+}^{\nu } ( t)\,m_{+}^{\nu } ( t)-\left [ u-\left ( 1-u\right ) \,u_{-}^{\nu } ( t)\right ] \,x_{-}^{\nu } ( t)\,m_{-}^{\nu } ( t).$ ] one may solve the model ( [ s])([u ] ) in the thermodynamic limit @xmath67 under the standard mean - field assumption that @xmath68 within this approximation , we may also substitute @xmath57 ( @xmath58 ) by the mean field values @xmath69
( @xmath70 ) .
( notice that one expects , and it will be confirmed below by comparisons with direct simulation results , that the mean field approximation is accurate away from any possible critical point . ) assuming further that patterns are random with mean activity @xmath71 one obtains the set of dynamic equations:@xmath72 \,x_{\pm } ^{\nu } ( t)\,m_{\pm } ^{\nu } ( t),\]]@xmath73\,m_{\pm } ^{\nu } ( t),\qquad \qquad\]]@xmath74 \right ) \right\ } , \]]@xmath75 , \qquad \qquad \qquad \label{dyn}\]]where @xmath76 this is a @xmath77dimensional coupled map whose analytical treatment is difficult for large @xmath78 but it may be integrated numerically , at least for not too large @xmath79 one may also find the fixed
point equations for the coupled dynamics of neurons and synapses ; these are@xmath80 \,\tau _ { \mathrm{rec}}\,m_{\pm } ^{\nu } \right\ } ^{-1 } , \notag \\
u_{\pm } ^{\nu } & = & u\,\,\tau _
{ \mathrm{fac}}\,\,m_{\pm } ^{\nu } \left ( 1\,\,+\,\,u\,\,\tau _ { \mathrm{fac}}\,\,m_{\pm } ^{\nu } \right ) ^{-1 } , \notag \\
2m_{\pm } ^{\nu } & = & 1\pm \frac{2}{n}{\sum_{i}}\tanh
\left [ \beta \left ( m^{\nu } \pm \sum_{\mu \neq \nu } \epsilon _ { i}^{\mu } m^{\mu } \right ) \right ] , \notag \\ m^{\nu } & = & \frac{1}{n}\sum_{i}\epsilon _ { i}^{\nu }
\tanh \left ( \beta \sum_{\mu } \epsilon _ { i}^{\mu } m^{\mu } \right ) .
\label{steady}\end{aligned}\]]the numerical solution of these transcendental equations describes the resulting order as a function of the relevant parameters . determining the stability of these solutions for @xmath81 is a more difficult task ,
because it requires to linearize ( [ dyn ] ) and the dimensionality diverges in the thermodynamical limit ( see however @xcite ) . in the next section we therefore deal with the case @xmath82
consider a finite number of stored patterns @xmath83 i.e. , @xmath84 in the thermodynamic limit . in practice , it is sufficient to deal with @xmath85 to illustrate the main results ( therefore , we shall suppress the index @xmath54 hereafter ) .
let us define the vectors of order parameters @xmath86 , its stationary value @xmath87 that is given by the solution of eq .
[ steady ] , and @xmath88 whose components are the functions on the right hand side of ( [ dyn ] ) .
the stability of ( [ dyn ] ) around the steady state ( [ steady ] ) follows from the first derivative matrix @xmath89 this is @xmath90where @xmath91 @xmath92 \,x_{\pm } , $ ] @xmath93 \,m_{\pm } , $ ] @xmath94 @xmath95 @xmath96 and @xmath97 after noticing that @xmath98 one may numerically diagonalize @xmath99 and obtain the eigenvalues @xmath100 for a given set of control parameters @xmath101 for @xmath102 @xmath103 the system is stable ( unstable ) close to the fixed point @xmath104 .
the maximum of @xmath105 determines the local stability : for @xmath106 the system ( [ dyn ] ) is locally stable , while for @xmath107 there is at least one direction of instability , and the system consequently becomes locally unstable .
therefore , varying the control parameters one crosses the line @xmath108 that signals the bifurcation points . the resulting situation is summarized in figure [ fig1 ] for specific values of @xmath44 @xmath109 and @xmath110 .
( [ steady ] ) have three solutions , two of which are memory states corresponding to the pattern and anti - pattern and the other a so - called paramagnetic state that has no overlap with the memory pattern .
the stability of the two solutions depends on @xmath111 .
the region @xmath112 corresponds to the non - retrieval phase , where the paramagnetic solution is stable and the memory solutions are unstable . in this phase
, the average network behaviour has no significant overlap with the stored memory pattern .
the region @xmath113 corresponds to the memory phase , where the paramagnetic solution is unstable and the memory solutions are stable .
the network retrieves one of the stored memory patterns . for @xmath114 ( denoted `` o '' in the figure )
none of the solutions is stable .
the activity of the network in this regime keeps moving from one to the other fixed points neighborhood ( the pattern and anti - pattern in this simple example ) .
this rapid switching behaviour is typical for dynamical synapses and does not occur for static synapses .
a similar oscillatory behavior was reported in @xcite for the case of only synaptic depression .
a main novelty is that the inclusion of facilitation importantly modifies the _ phase diagram _ , as discussed below ( figure [ fig2 ] ) . on the other hand ,
the phases for @xmath115 ( f ) and @xmath116 ( p ) correspond , respectively , to a locally stable regime with associative memory ( @xmath117 ) and to a disordered regime without memory ( i.e. , @xmath118 ) .
the values @xmath119 and @xmath120 which , as a function of @xmath121 @xmath42 and @xmath122 determine the limits of the oscillatory phase correspond to the onset of condition @xmath123 this condition defines lines in the parameter space ( @xmath124 ) that are illustrated in figure [ fig2 ] .
this reveals that @xmath125 ( separation between the f and o regions ) in general decreases with increasing facilitation , which implies a larger oscillatory region and consequently a reduction of the memory phase . on the other hand , @xmath126 ( separation between o and p regions ) in general increases with facilitation , thus broadening further the width of the oscillatory phase @xmath127 the behavior of this quantity under different conditions is illustrated in the insets of figure [ fig2 ] .
another interesting consequence of facilitation are the changes in the phase diagram as one varies the facilitation parameter @xmath42 which measures the fraction of neurotransmitter that are not activated by the facilitating mechanism**. * * in order to discuss this , we define the ratio between the time scales , @xmath128 and monitor the phase diagram ( @xmath129 ) for varying @xmath130 the result is also in figure [ fig2 ] see the bottom graphs for @xmath131 ( left ) and @xmath132 ( right ) which correspond , respectively , to a situation in which depression and facilitation occur in the same time scale and to a situation in which facilitation is four times faster .
the two cases exhibit a similar behavior for large @xmath44 but they are qualitatively different for small @xmath45 in the case of faster facilitation , there is a range of @xmath42 values for which @xmath133 increases , in such a way that one passes from the oscillatory to the memory phase by slightly increasing @xmath45 this means that facilitation tries to drive the network activity to one of the attractors ( @xmath134 ) and , for weak depression ( @xmath42 small ) , the activity will remain there .
decreasing @xmath42 further has then the effect of increasing effectively the system temperature , which destabilizes the attractor .
this only requires small @xmath42 because the dynamics ( [ u ] ) rapidly decreases the second term in @xmath21 to zero .
figure [ fig3 ] shows the variation with both @xmath27 and @xmath135 of the stationary locally
stable solution with associative memory , @xmath117 , computed this time both in the mean field approximation and using monte carlo simulation .
this monte carlo simulation consists of iterating eqs .
( [ s ] ) , ( [ x ] ) and ( [ u ] ) using parallel dynamics .
this shows a perfect agreement between our mean
field approach above and monte carlo simulations as long as one is far from the transition , a fact which is confirmed below ( in figure [ figure5 ] ) .
this is because , near @xmath136 the simulations describe hops between positive and negative @xmath137 which do not compare well with the mean field absolute value @xmath138 the most interesting behavior is perhaps the one revealed by the phase diagram @xmath139 in figure [ figure4 ] .
here we depict a case with @xmath140 in order to clearly visualize the effect of facilitation facilitation has practically no effect for any @xmath141 as shown above and @xmath142ms in order to compare with the situation of only depression in @xcite .
a main result here is that , for appropriate values of the working temperature @xmath109 , one may force the system to undergo different types of transitions by simply varying @xmath143 first note , that the line @xmath144 corresponds roughly to the case of static synapses , since @xmath111 is very small . in this limit
the transition between retrieval ( f ) and non - retrieval ( p ) phases is at @xmath145 at low enough @xmath122 there is transition between the non retrieval ( p ) and retrieval phases ( f ) as facilitation is increased .
this reveals a positive effect of facilitation on memory at low temperature , and suggests improvement of the network storage capacity which is usually measured at @xmath146 a prediction that we have confirmed in preliminary simulations . at intermediate temperatures , e.g. ,
@xmath147 for @xmath140 the systems shows no memory in the absence of facilitation , but increasing @xmath148 one may describe consecutive transitions to a retrieval phase ( f ) , to a disordered phase ( p ) , and then to an oscillatory phase ( o ) .
the latter is associated to a new instability induced by a strong depression effect due to the further increase of facilitation . at higher @xmath122 facilitation
may drive the system directly from complete disorder to an oscillatory regime .
in addition to its influence on the onset and width of the oscillatory region , @xmath149 determines the frequency of the oscillations of @xmath150 in order to study this effect , we computed the average time between consecutive minimum and maximum of these oscillations , i.e. , a half period .
the result is illustrated in the left graph of figure [ figure5 ] .
this shows that the frequency of the oscillations increases with the facilitation time .
this means that the access of the network activity to the attractors is faster with increasing facilitation , though
the system then remains a shorter time near each attractor due to an stronger depression .
on the other hand , we also computed the maximum of @xmath137 during oscillations , namely , @xmath151 this , which is depicted in the right graph of figure [ figure5 ] , also increases with @xmath152 the overall conclusion is that not only the access to the stored information is faster under facilitation but that increasing facilitation will also help to retrieve information with less error . in order to deepen further on some aspects of the system behavior
, we present in figures [ figure6 ] and [ figure7 ] a detailed study of specific time series .
the middle graph in figure [ figure6 ] corresponds to a simulation of the system evolution for increasing values of @xmath153 as one describes the horizontal line for @xmath154 in figure figure4 .
the system thus visits consecutively the different regions ( separated by vertical lines ) as time goes on .
that is , the simulation starts with the system in the stable _ paramagnetic _ phase , denoted p1 in the figure , and then successively moves by varying @xmath149 into the stable _ ferromagnetic _ phase f , into another paramagnetic phase , p2 , and , finally , into the oscillatory phase o. we interpret that the observed behavior in p2 is due to competition between the facilitation mechanism , which tries to bring the system to the fixed point attractors , and the depression mechanism , which tends to desestabilize the attractors .
the result is a sort of intermittent behavior in which oscillations and convergence to a fixed point alternates , in a way which resembles ( but is not ) chaos .
the top graph in figure [ figure6 ] , which corresponds to an average over independent runs , illustrates the typical behaviour of the system in these simulations ; the middle run depicts an individual run .
further interesting behavior is shown in the bottom graph of figure figure6 .
this corresponds to an individual run in the presence of a very small and irregular external stimulus which is represented by the ( green ) line around @xmath155 this consist of an irregular series of positive and negative pulses of intensity @xmath156 and duration of 20 ms .
in addition to a great sensibility to weak inputs from the environment , this reveals that increasing facilitation tends to significantly enhance the system response .
figure [ figure7 ] shows the power spectra of typical time series such as the ones in figure [ figure6 ] , namely , describing the horizontal line for @xmath154 in figure [ figure4 ] to visit the different regimes .
we plot here time series @xmath157 obtained , respectively , for @xmath158 20 , 50 and 100 and , on top of each of them , the corresponding spectra .
this reveals a flat , white noise spectra for the p1 phase and also for the stable fixed
point solution in the f regime .
however , the case for the intermittent p2 phase depicts a small peak around 65 hz .
the peak is much sharper and it occurs at 70 hz in the oscillatory case .
we have shown that the dynamical properties of synapses have profound consequences on the behaviour , and the possible functional role , of recurrent neural networks . depending on the relative strength of the depression , the facilitation and the noise in the network , one observes attractor dynamics to one of the stored patterns , non - retrieval where the neurons fire largely at random in a fashion that is uncorrelated to the stored memory patterns , or switching where none of the stored patterns is stable and the network switches rapidly between ( the neighborhoods of ) all of them .
these three behaviours were also observed in our previous work where we studied the role of depression .
the particular role of facilitation is the following .
the transitions between these possible phases are controlled by two facilitation parameters , namely , @xmath159 and @xmath45 analysis of the oscillatory phase reveals that the frequency of the oscillations , as well as the maximum retrieval during oscillations increase when the degree of facilitation increases .
that is , facilitation favours in the model a faster access to the stored information with a noticeably smaller error .
this suggests that synaptic facilitation might have an important role in short
term memory processes
. there is increasing evidence in the literature that similar jumping processes could be at the origin of the animals ability to adapt and rapidly response to the continuously changing stimuli in their environment .
we therefore believe that the network behaviour that is the consequence of dynamic synapses as presented in this paper may have important functional implications .
this work was supported by the _ meyc
feder _ project fis2005 - 00791 , the _ junta de andaluca _ project fqm 165 and the epsrc - funded colamn project ref .
ep / co 10841/1 .
we thank useful discussion with jorge f. mejas .
g. laurent , m. stopfer , r. w. friedrich , m. i. rabinovich , a. volkovskii , and h. d. i. abarbanel .
odor encoding as an active , dynamical process : experiments , computation and theory .
, 24:0 263297 , 2001 . | we study the effect of competition between short - term synaptic depression and facilitation on the dynamical properties of attractor neural networks , using monte carlo simulation and a mean field analysis . depending on the balance between depression , facilitation and the noise ,
the network displays different behaviours , including associative _ memory _ and switching of the activity between different attractors .
we conclude that synaptic facilitation enhances the attractor instability in a way that ( _ i _ ) intensifies the system adaptability to external stimuli , which is in agreement with experiments , and ( _ ii _ ) favours the retrieval of information with less error during short time intervals . |
hunting for the primordial non - gaussianity in cosmological observations has been a focus of attention as a big impact on the cosmology . in march 2013
, planck collaboration has reported tight constraints on the non - linearity parameters which characterize the amplitude of the deviation from pure gaussian statistics @xcite .
this result apparently implies that the cosmic microwave background ( cmb ) anisotropies measured by the planck mission are very close to gaussian , and may support the standard scenario of structure formation that the large - scale structure ( lss ) of the universe has emerged from tiny gaussian fluctuations .
nevertheless , the constraints derived from lss observations is still limited ( e.g. , @xcite ) , and at least as a cross check of the cmb results , it is worthwhile to further investigate the validity of this hypothesis precisely and independently from the lss observations .
it is recently known that large - scale halo / galaxy distributions that trace the lss provide a distinct information on the primordial non - gaussianity . in particular ,
the scale dependence of halo / galaxy bias has been found to be a very powerful probe to search for a primordial non - gaussianity ( e.g. , @xcite ) .
the most striking feature of the scale - dependent bias is that the effect appears even in the halo / galaxy power spectrum , and drastically change its shape and amplitude on large scales .
this is in marked contrast to the case of cmb observations , where the fluctuations are still in the linear regime , and hence the bispectrum and other higher - order statistics of the cmb anisotropies are the direct probe of primordial non - gaussianity . in this paper , we are particularly interested in the halo / galaxy bispectrum .
notice that the influence of scale - dependent bias also appears in the halo / galaxy bispectrum and other polyspectra .
although the late - time gravitational evolution is known to induce the non - gaussianity which inevitably dominates the primordial non - gaussianity on small scales , a characteristic feature of the gravity - induced bispectrum basically differs from the one originating from the primordial non - gaussianity .
further , due to the scale - dependent bias , the amplitude of halo / galaxy bispectrum may be enhanced , leading to a detectable signal on large scales ( see , e.g. , ref .
@xcite ) . in ref .
@xcite , using the peak formalism and the local bias picture , the authors have derived the analytic expression for halo / galaxy bispectrum . on the other hand , the authors of ref .
@xcite make use of the peak - background split picture , and specifically studied the halo bispectrum in the presence of primordial local - type non - gaussianity , characterized by the two constant parameters @xmath0 and @xmath1 .
numerical study on the halo bispectrum has been also made with cosmological @xmath2-body simulation ( see , e.g. , ref .
still , however , most of works has restricted their attention to the local - type non - gaussianity . with advent of high - precision lss observations such as des @xcite , bigboss @xcite , lsst @xcite , euclid @xcite , and hsc / pfs ( sumire ) @xcite , it is important to give a systematic study on the prediction of bispectrum for various types of primordial non - gaussianity . in this paper
, we systematically study the effect of the primordial non - gaussianity on the bispectrum of the halos / galaxies , on the basis of the integrated perturbation theory ( ipt ) @xcite , which helps us to systematically derive a precise formula to connect the halo / galaxy clustering with the initial matter density field @xcite . with the ipt formalism
, the non - local biasing effect can be incorporated into the statistical calculation of galaxies and halos in a straightforward manner .
further , in deriviing the effect of the primordial non - gaussianity on the clustering of the halos , we do not need to introduce the peak - background split picture . in ref .
@xcite , the authors have compared an analytic formula for the correlation function of the halos based on the ipt formalism with numerical @xmath2-body simulation and found the agreement of those results with a high precision .
based on this formalism , we compute the halo bispectrum in the presence of local- , equilateral- , and orthogonal - type non - gaussianities .
further , in case of the local - type non - gaussianity , we include the effect of the primordial trispectrum characterized by two non - linearity parameters , @xmath1 and @xmath3 .
while we mainly present the results of ipt calculation at one - loop order ( i.e. , next - to - leading order ) , the two - loop order ( i.e. , next - to - next - leading order ) contributions turns out to be important in several case .
we will investigate in detail the impact of such higher - order contributions on the expected bispectrum signal on large scales .
this paper is organized as follows . in sec .
[ sec : generalform ] , we begin by presenting a general formula for the bispectrum of the biased objects , which includes the effect of the primordial bispectrum and trispectrum up to the one - loop order in terms of ipt .
then , in sec .
[ sec : result ] , we study in detail the formula for the halo bispectrum and separately consider the local- , equilateral and orthogonal - type non - gaussianities . in section [ sec : higherorder ] , we investigate the contributions of the higher order loops in terms of ipt .
we discuss the comparison with the previous works where the bispectrum is obtained by other approaches and stress the utility of our systematic approach based on the ipt diagrammatic picture in section [ sec : discuss ] .
we also investigate the dependences of the redshift and the mass of halos .
we devote the final section to summary .
we plot the figures of this paper with adopting the best fit cosmological parameters taken from wmap 9-year data @xcite .
in this section , we give the formula for the bispectrum of galaxies and halos , based on the integrated perturbation theory ( ipt ) . as we mentioned in the introduction , advantage of the integrated perturbation theory ( ipt )
is that the effect of non - local biasing can be incorporated into the statistical calculation of galaxies and halos in a straightforward manner . ref .
@xcite has shown that an analytic formula for the correlation function of the halos based on the ipt agrees with @xmath2-body simulations in a good accuracy .
further , we do not need to introduce the peak - background split picture or peak formalism to derive the effect of the primordial non - gaussianity in the clustering of the biased objects .
@xcite have shown that the ipt formalism reproduces the previously known results for the scale - dependent bias as limiting cases of a general formula , and thus ipt gives a more general description for the scale - dependent bias . in sec .
[ subsec : bispec_1loop ] , we first present the general expressions for the bispectrum at the one - loop order .
we then derive the expressions of multi - point propagators in the large - scale limit in sec .
[ subsec : gamma_x ] , which will be the important building blocks to study the scale - dependent behavior of the bispectrum .
we begin by defining the bispectrum of biased objects , @xmath4 : @xmath5 the quantity @xmath6 is a fourier transform of the number density field of the biased objects . in the ipt formalism
, the multi - point propagators constitute the building blocks , and the perturbative expansion of the statistical quantities such as power spectrum and bispectrum are made with these propagators and the linear polyspectra . denoting the @xmath7-point propagator of the biased objects by @xmath8
, we define @xcite @xmath9 where @xmath10 represents the ( initial ) linear density field .
the multi - point propagator represents the influence on @xmath6 due to the infinitesimal variation for the initial density fields @xmath11 through the non - linear mode coupling .
it corresponds to the summation of all the loop contributions which are attached to each external vertex . in order to discuss the effect of the primordial non - gaussianity on the bispectrum of the biased objects , we here consider the perturbative expansion up to the one - loop order in ipt , which includes the contributions from the primordial trispectrum . following the notation in the previous papers @xcite ,
the bispectrum of the biased objects is expanded as @xmath12 with @xmath13 , \cr\cr b_{\rm bis}^{\rm tree } & = & \gamma_x^{(1 ) } ( { { \bf k}}_1 ) \gamma_x^{(1)}({{\bf k}}_2 ) \gamma_x^{(1)}({{\bf k}}_3 ) b_{{\rm l}}({{\bf k}}_1 , { { \bf k}}_2,{{\bf k}}_3 ) , \cr\cr b_{\rm tris } & = & { 1 \over 2 } \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(1)}({{\bf k}}_2 ) \int { d^3 p \over ( 2\pi)^3 } \gamma_x^{(2 ) } ( { { \bf p } } , { { \bf k}}_3 - { { \bf p } } ) t_{{\rm l}}({{\bf k}}_1,{{\bf k}}_2,{{\bf p}},{{\bf k}}_3-{{\bf p } } ) + 2~{\rm perms . } , \cr\cr b_{\rm grav}^{{\rm loop},1 } & = & \int { d^3 p \over ( 2 \pi)^3 } \gamma_x^{(2)}({{\bf p } } , { { \bf k}}_1 - { { \bf p } } ) \gamma_x^{(2)}(-{{\bf p } } , { { \bf k}}_2 + { { \bf p } } ) \gamma_x^{(2)}(-{{\bf k}}_1+{{\bf p } } , -{{\bf k}}_2 - { { \bf p } } ) p_{{\rm l}}(p)p_{{\rm l}}(|{{\bf k}}_1 - { { \bf p}}| ) p_{{\rm l}}(|{{\bf k}}_2 + { { \bf p}}| ) , \cr\cr b_{\rm grav}^{{\rm loop},2 } & = & { 1 \over 2 } \gamma_x^{(1)}({{\bf k}}_1 ) p_{{\rm l}}(k_1)\int { d^3 p\over ( 2\pi)^3 } \gamma_x^{(2)}({{\bf p } } , { { \bf k}}_2 - { { \bf p } } ) \gamma_x^{(3)}(-{{\bf k}}_1 , -{{\bf p } } , - { { \bf k}}_2 + { { \bf p}})p_{{\rm l}}(p)p_{{\rm l}}(| { { \bf k}}_2-{{\bf p}}| ) + 5~{\rm perms.},\cr\cr b_{\rm bis}^{{\rm loop},1 } & = & { 1 \over 2 } \gamma_x^{(1)}({{\bf k}}_1 ) \gamma_x^{(1)}({{\bf k}}_2 ) \int { d^3 p\over ( 2\pi)^3 } \gamma_x^{(3)}(-{{\bf k}}_1,{{\bf p}},-{{\bf k}}_2-{{\bf p}})p_{{\rm l}}(k_1)b_{{\rm l}}({{\bf k}}_2,{{\bf p } } , - { { \bf k}}_2-{{\bf p } } ) + 5~{\rm perms . } , \cr\cr b_{\rm bis}^{{\rm loop},2 } & = & { 1 \over 2 } \gamma_x^{(1)}({{\bf k}}_1 ) \gamma_x^{(2)}(-{{\bf k}}_1,-{{\bf k}}_2 ) \int { d^3 p\over ( 2\pi)^3 } \gamma_x^{(2)}({{\bf p } } , { { \bf k}}_2-{{\bf p}})p_{{\rm l}}(k_1)b_{{\rm l}}(- { { \bf k}}_2 , { { \bf p } } , { { \bf k}}_2-{{\bf p } } ) + 5~{\rm perms . } , \cr\cr b_{\rm bis}^{{\rm loop},3 } & = & \gamma_x^{(1 ) } ( { { \bf k}}_1 ) \int { d^3 p \over ( 2 \pi)^3 } \gamma_x^{(2 ) } ( { { \bf p } } , { { \bf k}}_2 - { { \bf p } } ) \gamma_x^{(2 ) } ( -{{\bf p } } , { { \bf k}}_3 + { { \bf p } } ) p_{{\rm l}}(p ) b_{{\rm l}}({{\bf k}}_1 , { { \bf k}}_2+{{\bf p } } , { { \bf k}}_3 - { { \bf p } } ) + 2~{\rm perms . } .
\label{eq : ourform}\end{aligned}\ ] ] the functions @xmath14 , @xmath15 and @xmath16 respectively denote the power- , bi- and tri - spectra of the linear density field , which are defined through @xmath17 as we see in the above expression , in the ipt formalism , the bispectrum of the biased objects is expressed in terms of the primordial polyspectra as well as multi - point propagators .
hence , compared with the peak - background split picture , it is straightforward to take into account the effect of various types of the primordial non - gaussianity whose statistical properties are encapsulated in the primordial polyspectra .
note that the linear density field is related to the primordial curvature perturbations @xmath18 through the function @xmath19 : @xmath20 where @xmath21 , @xmath22 , @xmath23 and @xmath24 are the transfer function , the linear growth factor , the hubble parameter at present epoch , and the matter density parameter , respectively . here
@xmath25 denotes an arbitrary redshift at the matter - dominated era . with the relation ( [ eq : delta_l_phi ] ) , the linear power spectrum is given by @xmath26 with @xmath27 in fig .
[ fig : diagram_bis.eps ] , diagrammatic representation of each term in eq .
( [ eq : bispec_ipt ] ) is shown . a double solid line connected with a grey circle , and a crossed circle glued to multiple single solid lines
respectively indicate the multi - point propagator of biased objects @xmath28 , and the correlator of the initial linear density field . as we have mentioned , with the multi - point propagator , all of the loop contributions which are attached to each external vertex are resummed , and the corresponding diagram does not have self - loops .
the contributions up to the one - loop order in ipt are shown .
a double solid line connected with a grey circle , and a crossed circle glued to multiple single solid lines represent the multi - point propagator of biased objects @xmath28 , and the correlator of the initial linear density field , respectively . ] .
the contributions up to the one - loop order in ipt are shown .
a double solid line connected with a grey circle , and a crossed circle glued to multiple single solid lines represent the multi - point propagator of biased objects @xmath28 , and the correlator of the initial linear density field , respectively . ] .
the contributions up to the one - loop order in ipt are shown .
a double solid line connected with a grey circle , and a crossed circle glued to multiple single solid lines represent the multi - point propagator of biased objects @xmath28 , and the correlator of the initial linear density field , respectively . ] the multi - point propagator @xmath8 is defined as a fully non - perturbative quantity that contains all the important ingredients to describe the non - linear gravitational evolution and galaxy / halo bias properties .
it is therefore difficult to evaluate it rigorously , however , for the large scales of our interest , perturbative treatment can work well , and we obtain the simplified expressions @xcite .
in particular , taking the large - scale limit which means that the scale of interest ( @xmath29 ) is much larger than the typical scale of the formation of the collapsed object ( @xmath30 with @xmath31 being a variable of integration in eq .
( [ eq : ourform ] ) ) , we have @xmath32 where @xmath33 is the second - order kernel of standard perturbation theory , and we have ignored the contribution of @xmath34 to @xmath35 .
the expression of the kernel @xmath33 is given by @xmath36 note that due to the symmetric property of @xmath33 , we have @xmath37 in eq .
( [ eq : gamma_x_k0limit ] ) , the quantity @xmath38 is a renormalized bias function defined in lagrangian space , given by @xmath39 with @xmath40 being the number density field of biased objects defined in lagrangian space .
so far , the derivation of the formulae above is quite general , and eq .
( [ eq : ourform ] ) , together with the large - scale limit of multipoint propagators ( [ eq : gamma_x_k0limit ] ) with renormalized bias function ( [ eq : c_n ] ) , can be applied to any biased objects in order to study the scale - dependence of the bispectrum . for a quantitative investigation of the large - scale behaviors of the bispectrum , in what follows
, we specifically consider the halo clustering , whose statistical properties are well - studied analytically and numerically . in the case of halos , the analytic expressions for the renormalized bias function @xmath38 has been already derived @xcite , and we can quantitatively predict the amplitude of the halo bispectrum , which will be compared with @xmath2-body simulations . since the galaxies mainly form inside the halos , for our interest of the large - scale limit , we expect that qualitatively similar features found in the case of halo bispectrum hold in the case of galaxy bispectrum .
adopting a simple model of non - local halo bias proposed by ref .
@xcite , the renormalized bias function for halos with mass @xmath41 is given by @xmath42,\end{aligned}\ ] ] where @xmath43 is the so - called critical density of the spherical collapse model , @xmath44 is the window function smoothed with the mass scale @xmath45 , and @xmath46 is the variance of density fluctuations on the mass scale @xmath41 . here
, a function @xmath47 is defined by @xmath48 with @xmath49 being the @xmath50-th order scale - independent lagrangian bias parameter which is constructed from the universal mass function as @xmath51 throughout the paper , we adopt the fitting formula by sheth and tormen @xcite for the halo mass function @xmath52 , and explicitly compute the halo bispectrum : @xmath53 \sqrt{q } \nu e^{-q\nu^2/2},\end{aligned}\ ] ] where @xmath54 , @xmath55 , @xmath56 and the normalization factor @xmath57^{-1}$ ] . in appendix
[ sec : appendix ] , we discuss the influence of the different choice of the mass functions on the final result of halo bispectrum .
note finally that in the large scale limit where @xmath58 , the window function and its derivative asymptotically approach @xmath59 and @xmath60 , and hence the renormalized bias function does not have significant scale - dependence .
let us now study in detail the formula for bispectrum of the biased objects given by eq .
( [ eq : ourform ] ) , especially focusing on the case of halos . in
what follows , we separately consider the three types of primordial non - gaussianity ; local- , equilateral- , and orthogonal - type characterized by the specific shape of the bispectrum @xmath15 and/or trispectrum @xmath16 . in the primordial local - type non - gaussianity , the bispectrum and trispectrum of linear matter density field are respectively characterized by @xmath61,\end{aligned}\ ] ] and @xmath62 \cr\cr & & \qquad\qquad\qquad\qquad\qquad\qquad\qquad + { 25 \over 9 } \tau_{\rm nl } \left [ p_\phi(k_1)p_\phi(k_2)p_\phi(k_{13 } ) + 11~{\rm perms . }
\right ] \bigr\},\end{aligned}\ ] ] with @xmath63 . here , the constant parameters , @xmath0 , @xmath1 and @xmath3 , are called non - linearity parameters .
if one considers the case with single - sourced primordial curvature perturbations characterized by @xmath64 , the non - linearity parameter @xmath3 is related to the leading - order one @xmath0 through @xmath65 , which is nothing but the consistency relation .
note that this relation does not hold in general . in cases with multi - sourced curvature perturbations
, we obtain the inequality , @xmath66 @xcite .
the consistency relation or inequality can be checked with the measurement of both the power spectrum of biased object and cross spectrum between the biased object and the matter density field @xcite ( see also refs.@xcite ) .
below , in evaluating the halo bispectrum , we simply assume that the consistency relation holds , @xmath65 , and present the results . substituting the expressions of @xmath14 , @xmath15 and @xmath16 into eq .
( [ eq : ourform ] ) , each term of the bispectrum of the biased objects for the primordial local - type non - gaussianity is evaluated as follows . apart from the first term in eq .
( [ eq : ourform ] ) , the contribution @xmath67 becomes @xmath68.\end{aligned}\ ] ] the third term in eq .
( [ eq : ourform ] ) , @xmath69 , is separately evaluated as : @xmath70 with @xmath71 + 2~{\rm perms . }
\biggr\ } , \cr\cr b_{\tau_{\rm nl } } & = & { 25 \over 9}\tau_{\rm nl } \biggl\ { \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(1)}({{\bf k}}_2){\cal m}(k_1){\cal m}(k_2 ) \left [ p_\phi(k_1 ) + p_\phi(k_2 ) \right ] p_\phi(k_3 ) \cr\cr & & \qquad\qquad \qquad \times \int { d^3 p \over ( 2 \pi)^3 } \gamma_x^{(2 ) } ( { { \bf p } } , { { \bf k}}_3 - { { \bf p } } ) { \cal m}(p){\cal m}(|{{\bf k}}_3 -{{\bf p}}| ) p_\phi(p ) + 2~{\rm perms . } \biggr\}\cr\cr & & + { 25 \over 9 } \tau_{\rm nl } \biggl\ { \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(1)}({{\bf k}}_2){\cal m}(k_1){\cal m}(k_2 ) p_\phi(k_1 ) p_\phi(k_2 ) \cr\cr & & \qquad\qquad\qquad \times \int { d^3 p \over ( 2 \pi)^3 } \gamma_x^{(2 ) } ( { { \bf p } } , { { \bf k}}_3 - { { \bf p } } ) { \cal m}(p){\cal m}(|{{\bf k}}_3 -{{\bf p}}|)p_\phi(|{{\bf k}}_1 + { { \bf p}}| ) + 2~{\rm perms.}\biggr\ } \cr\cr & & + { 25 \over 9 } \tau_{\rm nl } \biggl\ { \biggl [ \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(1)}({{\bf k}}_2){\cal m}(k_1){\cal m}(k_2 ) p_\phi(k_1 ) \int { d^3 p \over ( 2 \pi)^3 } \gamma_x^{(2 ) } ( { { \bf p } } , { { \bf k}}_3 - { { \bf p } } ) \cr\cr & & \qquad \times { \cal m}(p){\cal m}(|{{\bf k}}_3 -{{\bf p}}| ) p_\phi(p ) p_\phi(|{{\bf k}}_3 - { { \bf p}}|)\left [ 1 + { p_\phi(|{{\bf k}}_2 + { { \bf p}}| ) \over 2p_\phi(k_1)}\right ] + ( { { \bf k}}_1 \leftrightarrow { { \bf k}}_2 ) \biggr ] + 2~{\rm perms . } \biggr\}. \end{aligned}\ ] ] the remaining one - loop contributions that are linearly proportional to @xmath0 are @xmath72 , @xmath73 , and @xmath74 , which can be respectively recast as @xmath75 + 2~{\rm perms . } \biggr\ } , \cr\cr b_{\rm bis}^{{\rm loop},2 } & = & 2 f_{\rm nl } \biggl\ { \gamma_x^{(1)}({{\bf k}}_1 ) \biggl [ \gamma_x^{(2)}({{\bf k}}_1,{{\bf k}}_2 ) p_{{\rm l}}(k_1 ) { p_{{\rm l}}(k_2 ) \over { \cal m } ( k_2 ) } \cr\cr & & \times \int { d^3 p \over ( 2 \pi)^3 } \gamma_x^{(2)}({{\bf p } } , { { \bf k}}_2 - { { \bf p } } ) { \cal m}(p){\cal m}(|{{\bf k}}_2 - { { \bf p}}|)p_\phi(p ) \left ( 1 + { p_\phi(|{{\bf k}}_2 - { { \bf p}}| ) \over 2p_\phi(k_2)}\right ) + ( { { \bf k}}_2 \leftrightarrow { { \bf k}}_3 ) \biggr ] + 2~{\rm perms . }
\biggr\ } , \cr\cr b_{\rm bis}^{{\rm loop},3 } & = & f_{\rm nl } \biggl\ { \gamma_x^{(1)}({{\bf k}}_1){\cal m}(k_1)p_\phi(k_1 ) \biggl [ \int { d^3 p \over ( 2 \pi)^3 } \gamma_x^{(2)}({{\bf p } } , { { \bf k}}_2 - { { \bf p } } ) \gamma_x^{(2 ) } ( - { { \bf p } } , { { \bf k}}_3 + { { \bf p } } ) p_{{\rm l}}(p ) { \cal m}(|{{\bf k}}_2 - { { \bf p}}|){\cal m}(|{{\bf k}}_3 + { { \bf p}}| ) \cr\cr & & \qquad\times \left ( p_\phi(|{{\bf k}}_2 - { { \bf p}}| ) + p_\phi(|{{\bf k}}_3 + { { \bf p}}| ) + { p_\phi(|{{\bf k}}_3 + { { \bf p}}|)p_\phi(|{{\bf k}}_2 - { { \bf p}}| ) \over p_\phi(k_1)}\right ) + ( { { \bf k}}_2 \leftrightarrow { { \bf k}}_3 ) \biggr ] + 2~{\rm perms . } \biggr\}. \end{aligned}\ ] ] on small scales , the halo / galaxy bispectrum is generically dominated by the non - linearity of gravitational evolution , and the contribution coming from the higher - order loops becomes non - negligible . in this respect ,
similar to the power spectrum case , large - scales are the only window where the effect of the primordial non - gaussianity would be significant , giving rise to a detectable signature on the bispectrum .
let us then consider the large - scale limit in which all of the wave numbers @xmath76 , @xmath77 , and @xmath78 are much smaller than the typical scale of the biased object ( halo ) . in this limit , we obtain the approximate expressions for @xmath79 and @xmath80 : @xmath81 , \cr\cr b_{\tau_{\rm nl } } & \approx & { 25 \over 9 } \tau_{\rm nl } \bigl\ { \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(1)}({{\bf k}}_2){\cal m}(k_1){\cal m}(k_2)\cr\cr & & \qquad \times \left [ p_\phi(k_1 ) p_\phi(k_2 ) + p_\phi(k_2)p_\phi(k_3 ) + p_\phi(k_3)p_\phi(k_1 ) \right ] \int { d^3 p \over ( 2\pi)^3 } \gamma_x^{(2)}({{\bf p } } , - { { \bf p } } ) p_{{\rm l } } ( p ) + 2~{\rm perms . }
\bigr\ } \cr\cr & & + { 25 \over 9 } \tau_{\rm nl } \bigl\ { \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(1)}({{\bf k}}_2){\cal m}(k_1){\cal m}(k_2 ) \left [ p_\phi(k_1 ) + p_\phi(k_2 ) \right ] \int { d^3 p \over ( 2\pi)^3 } \gamma_x^{(2)}({{\bf p } } , - { { \bf p } } ) p_{{\rm l } } ( p ) p_\phi ( p ) + 2~{\rm perms . }
\bigr\}. \end{aligned}\ ] ] also , the rest of the one - loop contributions is approximately described as @xmath82 ^ 3 , \cr\cr b_{\rm grav}^{{\rm loop},2 } & \approx & { 1 \over 2 } \left [ \gamma_x^{(1)}({{\bf k}}_1 )
p_{{\rm l}}(k_1 ) \int { d^3 p \over ( 2\pi)^3 } \gamma_x^{(2)}({{\bf p } } , -{{\bf p}})\gamma_x^{(3)}(-{{\bf k}}_1 , -{{\bf p } } , -{{\bf k}}_2 + { { \bf p } } ) p_{{\rm l } } ( p ) ^2 + ~5~{\rm perms.}\right ] , \cr\cr b_{\rm bis}^{{\rm loop},1 } & \approx & 2 f_{\rm nl } \left [ \gamma_x^{(1)}({{\bf k}}_1 ) \gamma_x^{(1)}({{\bf k}}_2 ) p_{{\rm l}}(k_1 ) { p_{{\rm l}}(k_2 ) \over { \cal m}(k_2 ) } \int { d^3 p \over ( 2\pi)^3 } \gamma_x^{(3)}(-{{\bf k}}_1 , -{{\bf p } } , -{{\bf k}}_2 + { { \bf p } } ) p_{{\rm l } } ( p ) + 5~{\rm perms . } \right ] , \cr\cr b_{\rm bis}^{{\rm loop},2 } & \approx & 2 f_{\rm nl } \left [ \gamma_x^{(1)}({{\bf k}}_1 ) \gamma_x^{(2)}({{\bf k}}_1,{{\bf k}}_2 ) p_{{\rm l}}(k_1 ) { p_{{\rm l}}(k_2 ) \over { \cal m}(k_2 ) } \int { d^3 p \over ( 2\pi)^3 } \gamma_x^{(2)}({{\bf p } } , - { { \bf p } } ) p_{{\rm l } } ( p ) + 5~{\rm perms . }
\right ] , \cr\cr b_{\rm bis}^{{\rm loop},3 } & \approx & 4 f_{\rm nl } \left\ { \gamma_x^{(1)}({{\bf k}}_1 ) { p_{{\rm l}}(k_1 ) \over { \cal m}(k_1 ) } \int { d^3 p \over ( 2\pi)^3 } \left [ \gamma_x^{(2)}({{\bf p } } , - { { \bf p } } ) p_{{\rm l } } ( p ) \right]^2 + 2~{\rm perms . }
\right\ } , \label{eq : lsoneloopbis } \end{aligned}\ ] ] where we have used the fact that @xmath83 . , @xmath67 , @xmath79 and @xmath80 as functions of @xmath84 and @xmath85 in momentum space for @xmath86mpc@xmath87.,width=566 ] as examples of the shape of each contribution in @xmath88-space , in fig .
[ fig : shape.eps ] , we plot @xmath89 , @xmath67 , @xmath79 and @xmath80 for fixed @xmath86mpc@xmath87 . here ,
the redshift and the mass scale of halos are set to @xmath90 and @xmath91 , respectively , and we assume @xmath92 and @xmath93 because of the triangle condition . in the following discussion , we fix the redshift and the mass scale of halos to be above values . in discussion
[ subsec : halo_mass_z ] , we investigate the redshift and the mass dependences of the halo bispectrum .
as shown in this figure , we find that the contributions of the primordial non - gaussianity become dominant and have large amplitudes in the squeezed limit , while the leading - order effect of gravitational non - linearity appears in an equilateral shape .
to clarify the scale - dependence of their contributions , let us introduce the isosceles configuration given by @xmath94 .
a large @xmath95 corresponds to the squeezed shape . then , for small @xmath88 and large @xmath95 , the dominant scale - dependence of each contribution is simply given in terms of @xmath88 and @xmath95 as @xmath96 and @xmath97 where we have assumed that the power spectrum of the primordial fluctuations is scale - invariant , that is , @xmath98 . from the above equations , we find that the contributions from the non - linearity of the gravitational evolution have positive powers of @xmath88 and they decrease as @xmath88 decreases .
we also see that the contributions from the primordial bispectrum have similar @xmath88 and @xmath95-dependences except for @xmath74 .
as we will see later , @xmath74 is suppressed on large scales compared with other @xmath99 contributions , and the total contribution from the primordial bispectrum can be simply scaled as @xmath100 .
such scale - dependent behaviors on large scales have been discussed in previous works @xcite , and our result is consistent with their results . on the other hand ,
the contributions from the primordial trispectrum characterized by @xmath1 and @xmath3 ( i.e. , @xmath79 and @xmath80 ) have the @xmath101-dependence , and hence we expect that on large scales they become dominant in the halo bispectrum .
moreover , compared with contributions from primordial bispectrum ( i.e. , @xmath67 , @xmath72 , @xmath73 , and @xmath74 ) and the term @xmath79 , the term @xmath80 has a larger power of @xmath95 , and hence in the squeezed limit ( @xmath102 ) , this can dominate the halo bispectrum on large scales . in left panel of fig .
[ fig : bispectrum1.eps ] , we plot the contributions @xmath89 ( black thin dashed line ) , @xmath67 ( red thick line ) , @xmath79 ( blue thick dashed line ) , and @xmath80 ( green thick dotted line ) , as functions of the wavenumber @xmath88 with @xmath103 .
the non - linearity parameters are given by @xmath104 , @xmath105 and @xmath65 .
the asymptotic @xmath88-dependence expressed in eq .
( [ eq : asymptotic ] ) is found to be realized at @xmath106mpc@xmath87 . with the currently allowed values of the non - linearity parameters ,
the contributions of the primordial non - gaussianitiy can dominate the halo bispectrum at @xmath107mpc@xmath87 . on the other hand , at @xmath108mpc@xmath87 ,
the scale - dependence of each contribution gradually changes and deviates from the one in eq .
( [ eq : asymptotic ] ) .
this is simply due to the behavior of the transfer function @xmath21 , and thus the approximation based on the large - scale limit would become invalid at @xmath109mpc@xmath87 .
( black thin dashed line ) , @xmath67 ( red thick line ) , @xmath79 ( blue thick dashed line ) , and @xmath80 ( green thick dotted line ) , as functions of the wavenumber @xmath88 with @xmath103 .
right panel : one - loop contributions @xmath110 ( black thin dotted line ) , @xmath111 ( black thin dot - dashed line ) , @xmath72 ( red thick dashed line ) , @xmath73 ( red thick dotted line ) , @xmath74 ( red thick dot - dashed line ) , and @xmath67 ( red thick line ) , as functions of the wavenumber @xmath88 with @xmath103 . here
, we fix the redshift and the mass scale of halos to @xmath90 and @xmath91 , respectively .
the non - linearity parameters are given by @xmath104 , @xmath105 and @xmath65.,width=566 ] right panel of fig .
[ fig : bispectrum1.eps ] shows the one - loop contributions , @xmath110 ( black thin dotted line ) , @xmath111 ( black thin dot - dashed line ) , @xmath72 ( red thick dashed line ) , @xmath73 ( red thick dotted line ) , @xmath74 ( red thick dot - dashed line ) , and @xmath67 ( red thick line ) .
the results are again plotted as function of the wavenumber @xmath88 with @xmath103 .
the one - loop contributions , @xmath72 and @xmath73 , dominate the tree - level contribution , @xmath67 . this fact has been also addressed by jeong and komatsu ( 2009 ) @xcite , who computed the halo bispectrum on the basis of the peak formalism , taking account of the contribution from the primordial trispectrum .
the reason why the contributions @xmath72 and @xmath73 become larger than the other one - loop contributions may be explained from the diagrams shown in fig .
[ fig : diagram_bis.eps ] .
although all the diagrams in ipt are irreducible , the diagram of @xmath74 graphically looks similar to @xmath67 , and connecting two of the three grey circles with linear power spectrum gives @xmath74 . in similar way ,
@xmath110 and @xmath111 can be constructed from the tree diagram @xmath89 by adding a power spectrum .
we may call them _ decomposable _ diagrams . on the other hand ,
the diagrams @xmath72 and @xmath73 can not be constructed from the tree diagrams by simply adding a power spectrum .
we may call such kind of contributions _ un - decomposable _ diagrams .
the un - decomposable diagrams in nature involve higher - order correlators of the initial linear density field , and these correlators form a specific type of loops by glueing some of the legs ( indicated by solid lines ) to a multi - point propagator . in the large - scale limit , their loop integral can be dominated by the contributions of the squeezed limit of the higher - order correlators .
since the local - type non - gaussianity is known to produce a large primordial bispectrum in the squeezed limit , the un - decomposable diagram can potentially give a significantly large contribution to the halo bispectrum .
indeed , the expressions of @xmath72 and @xmath73 in the large - scale limit , given in eq .
( [ eq : lsoneloopbis ] ) , are mostly dominated by the squeezed limit of @xmath112 , and the effect of a strong coupling between short- and long - modes results in both the scale - dependence proportional to @xmath113 and the integral retaining the short - mode contribution that produces a very large amplitude . finally , in fig .
[ fig : bispectrum2.eps ] , we compare the total contribution from the gravitational non - linearity with those from primordial non - gaussianity .
plotted results are the contributions @xmath114 ( black thin dashed line ) , @xmath115 ( red thick line ) , @xmath79 ( blue thick line ) , and @xmath80 ( green thick line ) , as functions of @xmath95 , fixing the wavenumber to @xmath116 .
as we saw before , the contributions from the primordial non - gaussianity become large as increasing @xmath95 , and eventually dominate the halo bispectrum .
in particular , in the squeezed limit ( @xmath102 ) , the term @xmath80 is found to exceed other non - gaussian contributions .
( black thin dashed line ) , @xmath115 ( red thick line ) , @xmath79 ( blue thick line ) , and @xmath80 ( green thick line ) , as functions of @xmath95 with @xmath117 . here
, we fix that the redshift @xmath90 and the mass scale of halos @xmath91 .
the non - linearity parameters are given by @xmath104 , @xmath105 and @xmath65.,width=302 ] let us next consider the equilateral - type non - gaussianity , in which the bispectrum of linear density field @xmath15 is given by @xmath118 .
\label{eq : eqbi}\end{aligned}\ ] ] here , we do not consider the contribution from the `` equilateral''-trispectrum , because its scale dependence strongly depends on the models of generating primordial non - gaussianity .
also , its exact form is much complicated compared to that of the local - type non - gaussianity .
in fact , there are several works on the estimator for the trispectrum for the cmb temperature fluctuations in the models producing the equilateral - type bispectrum @xcite .
we leave the discussion on the contribution from the equilateral - trispectrum to future work .
the tree - level contribution from the primordial bispectrum , @xmath67 , becomes @xmath119.\end{aligned}\ ] ] on the other hand , the one - loop contributions , @xmath72 and @xmath73 include the primordial bispectrum of the specific configuration , @xmath120 . in the large - scale limit @xmath121 ,
the leading - order behavior becomes @xmath122.\end{aligned}\ ] ] here @xmath123 and we assume the power - law spectrum , @xmath124 .
also , @xmath74 has the primordial bispectrum of the configuration , @xmath125 , which can be reduced to @xmath126.\end{aligned}\ ] ] then , we obtain the scale - dependence for the halo bispectrum induced from the equilateral primordial non - gaussianity . in the large scale limit
, we have @xmath127 here we simply assume the scale - invariant primordial power spectrum , that is , @xmath128 . because of the positive powers of @xmath88 , the one - loop contributions from the equilateral - type bispectrum are all suppressed on the large scales , compared to the tree - level contribution .
as function of @xmath129 and @xmath85 for @xmath130 ( left panel ) , and @xmath131 as a function of @xmath95 with fixing @xmath132 ( right panel ) for equilateral - type primordial non - gaussianity . we take @xmath133.,width=566 ] left panel of fig .
[ fig : eq_alpha.eps ] shows the shape of @xmath131 in the momentum space , fixing wavenumber @xmath76 to @xmath134 , and we assume @xmath92 and @xmath93 because of the triangle condition .
right panel of fig .
[ fig : eq_alpha.eps ] shows the @xmath95-dependence of @xmath131 , for isosceles configuration as @xmath135 , and we fix the wavenumber @xmath88 to @xmath136 and take @xmath133 , which is almost the @xmath137-@xmath138 upper bound obtained by planck collaboration @xcite .
the amplitude of @xmath67 has a peak around @xmath139 , indicating that the halo bispectrum has equilateral shape in @xmath140 space . since the contribution from the non - linearity of the gravitational evolution @xmath89 also has a peak around @xmath139 ( see fig .
[ fig : shape.eps ] ) , it seems difficult to distinguish between @xmath67 and @xmath89 in the case of equilateral - type non - gaussianity .
however , the @xmath88-dependence of @xmath67 and @xmath89 has very distinct feature , as shown in fig .
[ fig : eq_k.eps ] . here ,
we plot @xmath89 ( black dashed line ) and @xmath67 ( red thick line ) as a function of @xmath88 , fixing @xmath95 to unity and taking @xmath133 . for the non - linearity parameter consistent with the current observational limit , we find that @xmath67 dominates @xmath89 at @xmath141 .
note that the detectability of the equilateral primordial no - gaussianity from galaxy survey has been discussed in ref .
@xcite , which also found the similar scale - dependent behavior .
the authors of ref .
@xcite computed the galaxy bispectrum based on the local bias ansatz . at the tree level , the difference between our formula and the expression in ref .
@xcite appears in the scale dependence of the term , @xmath142 included in @xmath143 .
this basically comes from the bias prescription which we adopted here @xcite , and reflects the non - local nature of the halo bias , but it does not produce much difference in the halo bispectrum on large scales .
( black dashed line ) , @xmath67 ( red thick line ) , @xmath72 ( red dashed thick line ) , @xmath73 ( red dotted thick line ) , and @xmath74 ( red dotted - dashed thick line ) as a function of @xmath88 with fixing @xmath144 for equilateral - type primordial non - gaussianity .
we take @xmath133 . , width=302 ] in fig .
[ fig : eq_k.eps ] , we also plot the one - loop contributions for the equilateral - type ; @xmath89 ( black dashed line ) , @xmath67 ( red thick line ) , @xmath72 ( red dashed thick line ) , @xmath73 ( red dotted thick line ) , and @xmath74 ( red dotted - dashed thick line ) .
the one - loop contributions are all suppressed on large scales as mentioned before . in ref .
@xcite , the one - loop corrections have been also discussed .
the author computed the one - loop contributions with the eulerian local bias prescription , and mentioned that one of the loop contributions is not negligible even for the equilateral - type non - gaussianity . in the diagrammatic picture , this corresponds to the loop attached to an external vertex . in our formalism on the basis of ipt , however , such loops are already included in the multi - point propagator as a result of the resummation .
hence such higher - order loop corrections do not appear in our formula . as the third type of primordial non - gaussianity , we consider the orthogonal - type non - gaussianity .
the bispectrum of the orthogonal type is defined as @xmath145 .
\label{eq : eqbi}\end{aligned}\ ] ] then we obtain @xmath146.\end{aligned}\ ] ] as function of @xmath129 and @xmath85 for @xmath147 , right panel ; @xmath131 as a function of @xmath95 with fixing @xmath148 , for orthogonal - type primordial non - gaussianity .
we take @xmath149.,width=566 ] in fig .
[ fig : ort_alpha.eps ] , shape of the bispectrum @xmath131 ( left ) and its @xmath95-dependence ( right ) are shown . for the parameter @xmath150
, we adopt @xmath151 , consistent with the planck results at @xmath137-@xmath138 level @xcite .
right panel shows as increasing @xmath95 , @xmath67 starts to decrease and has a negative peak around @xmath139 , and then turns to increase with positive value toward the squeezed limit ( @xmath152 ) .
this is in marked contrast to the behaviors in other types of the non - gaussianity , and may be a very important clue to separately detect the orthogonal - type non - gaussianity from the halo bispectrum .
( black dashed line ) , @xmath67 ( red thick line ) , @xmath72 ( red dashed thick line ) , and @xmath73 ( red dotted thick line ) as a function of @xmath88 with fixing @xmath103 for orthogonal - type primordial non - gaussianity . we take @xmath149 .
, width=302 ] fig .
[ fig : ort_loop.eps ] shows the scale - dependence of the halo bispectrum for the orthogonal type non - gaussianity .
in addition to the tree - level contributions , @xmath89 ( black dashed line ) and @xmath67 ( red thick line ) , we also plot the one - loop contributions , @xmath72 ( red dashed thick line ) and @xmath73 ( red dotted thick line ) as functions of @xmath88 , fixing @xmath95 to @xmath153 and taking @xmath149 . here
, we do not show the contribution @xmath74 , because it turns out to be very small .
as in the case of equilateral - type non - gaussianity , the one - loop contributions are much suppressed , and tree - level contributions @xmath67 becomes dominant on very large - scales . finally , let us compare the overall trend of the total halo bispectrum at the one - loop order between three - types of primordial non - gaussianity .
[ fig : bis_total.eps ] shows the total halo bispectrum , @xmath154 , as function of wavenumber , fixing @xmath95 to @xmath153 ( left ) and @xmath155 ( right ) .
while upper panels present the results for the local - type non - gaussianity with two different parameter set , i.e. , @xmath156 and @xmath157 , bottom panels plot the halo bispectra for the equilateral- ( green ) and orthogonal- types ( magenta dashed ) .
here , we assume that the consistency relation , @xmath158 , strictly holds for local - type non - gaussianity . in each panel ,
black solid line indicates the result in the absence of the primordial non - gaussianity , and considers only the contributions from the gravitational non - linearity . for the non - linearity parameters consistent with current observations , the local - type non - gaussianity gives the largest signal of the halo bispectrum in the squeezed case at large - scale .
a notable point is that even with @xmath144 , a strong enhancement of the amplitude of bispectrum can be observed , especially in the case of @xmath159 .
the scale - dependence of the halo bispectrum also appears in the cases of the equilateral- and orthogonal - type non - gaussianities ( see bottom panel ) .
although the effect is rather moderate , at large - scales , the contribution from primordial non - gaussianity can exceed that from the gravitational non - linearity .
further , as we have seen in sec .
[ sec : orthogonal ] , additionally interesting feature can be seen for the orthogonal - type non - gaussianity . with a negative non - linearity parameter @xmath160
, the sign of the halo bispectrum eventually flips at large scales for the configuration with @xmath161 .
this would be unique and important feature to distinguish the non - gaussian signal from other types of non - gaussianity .
, for local - type with @xmath103 ( left ) and @xmath144 ( right ) .
black thin line is for the case without any primordial non - gaussianity , red thick line is for the case with @xmath104 , @xmath158 and @xmath162 , and blue thick dashed line is for the case with @xmath163 and @xmath105 .
lower panels : total halo bispectrum , @xmath154 , for equilateral- ( green thick line ) and orthogonal - types ( magenta thick dashed line ) with @xmath103 ( left ) and @xmath144 ( right ) .
we fix the non - linearity parameters to @xmath133 and @xmath164 . in all panels ,
we fix the redshift and the mass of halos to @xmath90 and @xmath91 .
, width=566 ]
in the formalism of ipt , the contribution from the primordial non - gaussianity on the halo bispectrum can be efficiently estimated from the diagrams including the linear ( primordial ) polyspectra . increasing the order of perturbations
, we can systematically calculate the contributions from higher - order polyspectra .
naively , we expect that higher - loop contributions are generally suppressed on very large scales . as we saw in sec .
[ sec : result_local ] , however , one - loop contributions from the local - type non - gaussianity are found to be non - negligible on large scales , and can eventually dominate the tree - level contributions .
this partly implies that the perturbative treatment may not work well in the case of local - type non - gaussianity , and the contribution from the two - loop order would be also dominant . in this section ,
focusing on the local - type non - gaussianity , we study the impact of two - loop contribution on the halo bispectrum . as we found in sec . [
sec : result_local ] , non - negligible contributions coming from the one - loop corrections are described by the _ un - decomposable _ diagrams , which can not be simply constructed from tree - level diagrams by adding a primordial power spectrum ( shown in fig . [
fig : diagram_bis.eps ] ) . in this respect , at two - loop order
, possible non - negligible contributions may come from the diagrams linearly proportional to the primordial trispectrum or quadratically proportional to the primordial bispectrum .
the other contributions linearly proportional to the primordial bispectrum would be certainly suppressed , since they are described as decomposable diagrams .
let us first consider the two - loop un - decomposable contributions linearly proportional to the primordial trispectrum .
the corresponding diagrams are shown in fig .
[ fig : lineargtau.eps ] , each of which are analytically expressed as @xmath165 , \cr\cr b_{\rm tris}^{2{\rm loop},2 } & = & { 1 \over 6}\biggl [ \gamma_x^{(1)}({{\bf k}}_1 ) \gamma_x^{(2)}(-{{\bf k}}_1 , - { { \bf k}}_3)p_{{\rm l}}(k_1 ) \cr\cr & & \quad\times \int { d^3 p_1 d^3 p_2 \over ( 2\pi)^6 } \gamma_x^{(3 ) } ( { { \bf p}}_1 , { { \bf p}}_2 , { { \bf k}}_3 - { { \bf p}}_1 -{{\bf p}}_2 ) t_{{\rm l}}(-{{\bf k}}_3 , { { \bf p}}_1 , { { \bf p}}_2 , { { \bf k}}_3 - { { \bf p}}_1 -{{\bf p}}_2 ) + 5~{\rm perms . }
\biggr ] , \cr\cr b_{\rm tris}^{2{\rm loop},3 } & = & { 1 \over 4 } \biggl [ \gamma_x^{(1)}({{\bf k}}_1 ) p_{{\rm l}}(k_1 ) \int { d^3 p_1 d^3 p_2 \over ( 2 \pi)^6 } \gamma_x^{(2)}({{\bf p}}_1 , { { \bf k}}_2 - { { \bf p}}_1 ) \cr\cr & & \qquad\qquad\qquad \times \gamma_x^{(3)}({{\bf k}}_1 , { { \bf p}}_2 , - { { \bf k}}_2 - { { \bf p}}_2 ) t_{{\rm l}}({{\bf p}}_1 , { { \bf k}}_2 - { { \bf p}}_1 , { { \bf p}}_2 , -{{\bf k}}_2 - { { \bf p}}_2 ) + 5~{\rm perms . } \biggr].\end{aligned}\ ] ] in a manner similar to what we did in sec .
[ sec : result_local ] , we take the large - scale limit , @xmath166 . then , we have @xmath167 { \cal m}(p_1){\cal m}(p_2){\cal m}(|{{\bf p}}_1 + { { \bf p}}_2| ) p_\phi(p_1)p_\phi(p_2 ) + 5~{\rm perms . } , \cr\cr b_{\rm tris}^{2{\rm loop},2 } & \approx & \left(3 g_{\rm nl } + { 25 \over 9}\tau_{\rm nl}\right ) \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(2)}(-{{\bf k}}_1 , -{{\bf k}}_3 ) p_{{\rm l}}(k_1){p_{{\rm l}}(k_3 ) \over { \cal m}(k_3 ) } \cr\cr & & \qquad \times \int { d^3 p_1 d^3 p_2 \over ( 2\pi)^6}c_3^{{\rm l}}({{\bf p}}_1 , { { \bf p}}_2 , - { { \bf p}}_1-{{\bf p}}_2){\cal m}(p_1){\cal m}(p_2){\cal m}(|{{\bf p}}_1 + { { \bf p}}_2| ) p_\phi(p_1)p_\phi(p_2 ) + 5~{\rm perms . } , \cr\cr b_{\rm tris}^{2{\rm loop},3 } & \approx & 3 g_{\rm nl }
\gamma_x^{(1)}({{\bf k}}_1)p_{{\rm l}}(k_1 ) \biggl\ { \int { d^3 p_1 \over ( 2\pi)^3 } c_2^{{\rm l}}({{\bf p}}_1 , -{{\bf p}}_1)p_{{\rm l}}(p_1 ) \cr\cr & & \qquad\qquad \times \int { d^3 p_2 \over ( 2\pi)^3 } \left [ -{{{\bf k}}_1 \cdot { { \bf k}}_3 \over k_1 ^ 2}c_2^{{\rm l}}({{\bf p}}_2 , -{{\bf p}}_2 ) + c_3^{{\rm l}}(-{{\bf k}}_1 , { { \bf p}}_2 , - { { \bf p}}_2 ) \right ] p_{{\rm l}}(p_2 ) p_\phi(p_2 ) \cr\cr & & + \int { d^3 p_1 \over ( 2\pi)^3 } c_2^{{\rm l}}({{\bf p}}_1 , -{{\bf p}}_1)p_{{\rm
l}}(p_1)p_\phi(p_1 ) \cr\cr & & \qquad\qquad \times \int { d^3 p_2 \over ( 2\pi)^3 } \left [ -{{{\bf k}}_1 \cdot { { \bf k}}_3 \over k_1
^ 2}c_2^{{\rm l}}({{\bf p}}_2 , -{{\bf p}}_2 ) + c_3^{{\rm l}}(-{{\bf k}}_1 , { { \bf p}}_2 , - { { \bf p}}_2 ) \right ] p_{{\rm l}}(p_2 ) \biggr\ } + 5~{\rm perms . }
\cr\cr & & + { 25 \over 9}\tau_{\rm nl}\gamma_x^{(1)}({{\bf k}}_1 ) p_{{\rm l}}(k_1){p_{{\rm l}}(k_2 ) \over { \cal m}(k_2)^2 } \int { d^3 p_1 \over ( 2\pi)^3 } c_2^{{\rm l}}({{\bf p}}_1 , -{{\bf p}}_1)p_{{\rm l}}(p_1 ) \cr\cr & & \qquad\qquad \times \int { d^3 p_2 \over ( 2\pi)^3 } \left [ -{{{\bf k}}_1 \cdot { { \bf k}}_3 \over k_1 ^ 2}c_2^{{\rm l}}({{\bf p}}_2 , -{{\bf p}}_2 ) + c_3^{{\rm l}}(-{{\bf k}}_1 , { { \bf p}}_2 , - { { \bf p}}_2 ) \right ] p_{{\rm l}}(p_2 ) + 5~{\rm perms.}. \label{eq : bgt2loop2}\end{aligned}\ ] ] here , we used the fact that @xmath168 from the above expressions , we find @xmath169 ( blue dashed thick line ) , @xmath170 ( blue dotted thick line ) , @xmath1 contribution in @xmath171 ( blue dot - dashed thick line ) , and @xmath3 contribution in @xmath171 ( green dot - dashed thick line ) , as functions of @xmath88 . we also plot @xmath89 ( black dashed line ) , @xmath79 ( blue thick line ) , and @xmath80 ( green thick line ) as references .
we take @xmath105 and @xmath172 and fix the redshift , the mass of halos and the squeezing parameter to @xmath90 , @xmath91 , and @xmath173 .
, width=302 ] in fig .
[ fig : tri_two_loop.eps ] , the three two - loop contributions , @xmath174 ( blue dashed thick line ) , @xmath170 ( blue dotted thick line ) , and @xmath171 , are plotted as functions of @xmath88 , fixing @xmath95 to @xmath153 . for the contribution , @xmath171 ,
we divide it into two pieces , and separately show the contributions proportional to @xmath1 ( blue dot - dashed thick line ) and @xmath3 ( green dot - dashed thick line ) .
then , all the un - decomposable two - loop contributions proportional to @xmath1 turn out to be sub - dominant on large scales , while the term proportional to @xmath3 ( labeled as @xmath3 part of @xmath171 ) gives a significantly large contribution , which can dominate over the one - loop contribution , @xmath80 .
this implies that we need to develop at least the two - loop order calculations if one wants to evaluate the contribution of the halo bispectrum linearly proportional to the non - linearity parameter @xmath3 . as for the case of @xmath1 ,
the one - loop order calculation seems sufficient on large scales .
consider next the un - decomposable two - loop contributions quadratically proportional to the primordial bispectrum , which are diagrammatically shown in fig .
[ fig : diagram_ff.eps ] . for local - type non - gaussianity
, the inequality , @xmath66 , generally holds . in the case of equality which we consider here ,
this implies that the diagrams of the @xmath175 order are expected to give a significant contribution comparable to the @xmath3 contribution , and potentially become dominant on large scales . in the large - scale limit
, we approximately have @xmath176 ^ 2 + 2~{\rm perms . } , \cr\cr b_{\rm bis - bis}^{2{\rm loop},2 } & \approx & 4f_{\rm nl}^2 \gamma_x^{(1)}({{\bf k}}_1)\gamma_x^{(1)}({{\bf k}}_2 ) \int { d^3p_1d^3p_2 \over ( 2\pi)^6 } c_4^{{\rm l}}({{\bf p}}_1 , - { { \bf p}}_1,{{\bf p}}_2 , - { { \bf p}}_2 ) p_{{\rm l}}(p_1)p_{{\rm l}}(p_2 ) { p_{{\rm l}}(k_1 ) p_{{\rm l}}(k_2 ) \over { \cal m}(k_1){\cal m}(k_2 ) } + 2~{\rm perms.},\cr\cr b_{\rm bis - bis}^{2{\rm loop},3 } & \approx & 4f_{\rm nl}^2 \gamma_x^{(1)}({{\bf k}}_1 ) { p_{{\rm l}}(k_1 ) p_{{\rm l}}(k_2 ) \over { \cal m}(k_1){\cal m}(k_2 ) } \int { d^3p_1 \over ( 2\pi)^3 } c_2^{{\rm l}}({{\bf p}}_1 , - { { \bf p}}_1 ) p_{{\rm l}}(p_1)\cr\cr & & \times \int { d^3 p_2 \over ( 2\pi)^3 } \left [ - { { { \bf k}}_2 \cdot { { \bf k}}_3 \over k_2 ^ 2 } c_2^{{\rm l}}(-{{\bf p}}_2 , { { \bf p}}_2 ) + c_3^{{\rm l}}(-{{\bf k}}_2 , - { { \bf p}}_2 , { { \bf k}}_1 + { { \bf p}}_2 ) \right ] p_{{\rm l}}(p_2 ) + 5~{\rm perms.}~.\end{aligned}\ ] ] then , the asymptotic behavior of these contributions becomes @xmath177 ( green thick line ) , @xmath178 ( magenta dashed line ) , @xmath179 ( magenta dotted line ) , and @xmath180 ( magenta dot - dashed line ) as a function of @xmath88 with fixing @xmath103 .
we take @xmath181 and @xmath182.,width=302 ] in fig .
[ fig : fnlfnl.eps ] , the two - loop contributions given above are plotted as functions of @xmath88 , fixing @xmath95 to @xmath153 ; @xmath178 ( magenta dashed ) , @xmath179 ( magenta dotted ) , and @xmath180 ( magenta dot - dashed ) . for comparison
, we also plot the one- and two - loop contributions linearly proportional to @xmath3 , i.e. , @xmath183 ( green thick ) , assuming @xmath181 and @xmath182 .
then , all the two - loop contributions quadratically proportional to the primordial bispectrum can become sub - dominant , and are well below the @xmath3-contribution .
the asymptotic behaviors given in eqs .
( [ eq : asymptotic ] ) , ( [ eq : asympttwotri ] ) , and ( [ eq : asymptbisbis ] ) indicate that all the contributions shown in fig . [
fig : fnlfnl.eps ] scale as @xmath101 , but the terms @xmath80 and @xmath171 have a larger power of @xmath95 , and are proportional to @xmath184 , which results in a significantly larger amplitude than that of @xmath185 by two orders of magnitude . to sum up
, it is not always the case that the higher loop contributions are suppressed compared with the lower loop contributions . in particular
, the un - decomposable loop diagrams in the context of ipt would produce dominant contributions in the large - scale halo bispectrum , and hence we need to take into account all of the un - decomposable diagrams in order to precisely evaluate the effect of primordial non - gaussianity on the halo bispectrum . in doing this , the diagrammatic approach based on the ipt helps us to systematically collect these dominant contributions .
in this section , we discuss several points which have not been yet clarified in previous sections .
first , in sec . [ subsec : halo_mass_z ] , we examine the dependence of redshift and halo mass on the halo bispectrum in the case of local - type non - gaussianity . in sec .
[ subsec : asymptotic_scale - dependence ] , to infer the asymptotic behavior of the bispectrum , we give a simple estimate of the scale - dependent behaviors using the diagrammatic approach . finally , in sec .
[ subsec : comparison ] , the results of our analysis are compared with those of recent other works .
so far , we have set the redshift and the mass scale of halos to @xmath90 and @xmath91 as representative values relevant for observations . here ,
focusing on the local - type non - gaussianity and based on the formulae for bispectrum at one - loop order , we discuss the redshift and the mass dependences of the halo bispectrum .
the redshift dependence of the halo bispectrum is shown in fig .
[ fig : fgtnlz.eps ] .
( black ) , @xmath67 ( red thick ) , @xmath72 ( red thick dashed ) , and @xmath73 ( red thick dotted ) , bottom - @xmath186 ( black ) , @xmath79 ( blue thick ) , @xmath80 ( green thick ) , and @xmath171 ( green thick dashed ) with changing the mass of halo as @xmath187 ( left ) , @xmath90 ( central ) , and the right for @xmath188 ( right).,width=642 ] in this figure , we plot each contribution of the halo bispectrum , @xmath186 ( black ) , @xmath67 ( red thick ) , @xmath72 ( red thick dashed ) , and @xmath73 ( red thick dotted ) in top panels , and @xmath186 ( black ) , @xmath79 ( blue thick ) , @xmath80 ( green thick ) , and @xmath171 ( green thick dashed ) in bottom panels . from left to right , we change the redshift as @xmath189 , @xmath155 , and @xmath190 . for these redshifts
, we have checked that the other contributions are also suppressed . from this figure
, we find that the scale - dependence of the halo bispectrum does not change with the redshift , but the most dominant contribution changes .
the redshift parameter has an affect on not only the evolution of the density perturbations but also @xmath191 .
basically , the density perturbations grow as the redshift parameter decreases . on the other hand ,
once the mass is fixed , the halo at the higher redshift becomes more highly biased object and it means larger value of @xmath191 .
as shown in fig .
[ fig : fgtnlz.eps ] , the halo bispectrum becomes larger at the higher redshift and hence we find that the halo bispectrum seems to have the stronger dependence of the redshift parameter through @xmath191 than the growth function @xmath22 . as for the most dominant contribution , we find for the more highly biased object the higher order loop contribution becomes important .
basically , the higher order loop contribution depends on the higher order scale - independent lagrangian bias parameter @xmath191 and such higher order @xmath191 tends to become larger for the more highly biased object .
hence , the higher order loop contribution becomes important for the higher redshift , as can be seen in fig .
[ fig : fgtnlz.eps ] .
the mass dependence of the halo bispectrum is shown in fig .
[ fig : fgtnlm.eps ] .
( black ) , @xmath67 ( red thick ) , @xmath72 ( red thick dashed ) , and @xmath73 ( red thick dotted ) , bottom - @xmath186 ( black ) , @xmath79 ( blue thick ) , @xmath80 ( green thick ) , and @xmath171 ( green thick dashed ) with changing the mass of halo as @xmath192 ( left ) , @xmath193 ( central ) , and @xmath194 ( right).,width=642 ] in this figure , we plot each contribution of the halo bispectrum , @xmath186 ( black ) , @xmath67 ( red thick ) , @xmath72 ( red thick dashed ) , and @xmath73 ( red thick dotted ) in top panels , and @xmath186 ( black ) , @xmath79 ( blue thick ) , @xmath80 ( green thick ) , and @xmath171 ( green thick dashed ) in bottom panels . from left to right , we change the mass of halo as @xmath192 , @xmath195 , and @xmath196 . as in the discussion about the redshift dependence , we have checked that the other contributions are suppressed , and we also find that the scale - dependence of the halo bispectrum does not change with changing the mass of halo , but the dominant contribution changes .
the dependence of the mass of halos is reflected in the scale - independent lagrangian bias parameter @xmath191 , and it becomes larger for the larger mass . as we discussed before , higher order @xmath191 tends to become larger for the larger mass and hence the higher order loop contribution becomes more important for the larger mass of the halos , as can be seen in fig .
[ fig : fgtnlm.eps ] . as the result , although which loop contribution dominates in the halo bispectrum on large scales depends on the bias parameter @xmath191 , that is , the redshift and the mass of observed halos , we stress that it should be essential to consider both contributions of one and two - loops coming from the primordial trispectrum .
the asymptotic scale - dependence is important to find dominant contributions of the halo bispectrum on large scales . in the discussion above ,
starting with the analytic expression of each diagram , we took the large - scale limit , and have finally derived the formulae for asymptotic scale - dependence .
here , we will argue that using the diagrammatic approach , these formulae can be recovered more easily and systematically .
first consider the simple case , and look at the contribution , @xmath89 , shown in fig .
[ fig : diagram_bis.eps ] .
this diagram can be graphically divided into two pieces , consisting of linear power spectrum of the biased object , @xmath197 .
thus , we may write it as @xmath198 . for scale - invariant primordial fluctuations ,
the linear power spectrum @xmath197 is proportional to @xmath88 on large scales , and we immediately obtain @xmath199 , as a dominant contribution on large scales ( @xmath200 ) in the squeezed limit ( @xmath201 ) .
similarly , we can also divide the contribution of the diagrams @xmath72 and @xmath73 in fig .
[ fig : diagram_bis.eps ] into two pieces . in this case , a part of the diagram is @xmath197 , but another piece corresponds to the one - loop power spectrum of the biased object , induced by the primordial bispectrum , @xmath202 . it is known in the literature that the local - type primordial bispectrum generates a strong scale - dependence in the bias parameter , which scales as @xmath101 . in ipt , the main contribution of this indeed comes from the one - loop power spectrum , and we have @xmath203 .
hence , we reproduce the asymptotic scale - dependence of @xmath72 and @xmath73 , and we obtain @xmath204 as a dominant contribution on large scales in the squeezed limit . this systematic approach based on the diagrammatic picture can also apply to the other contributions , and we easily find their asymptotic scale - dependences .
finally , let us compare the results of our analysis with those of the recent works . in ref .
@xcite , adopting the taylor - expansion with local bias ansatz , the authors derived the formula for the halo bispectrum based on the matarrese - lucchin - bonometto ( mlb ) formalism @xcite .
mlb formalism is nothing but the peak formalism , and we confirmed that taking the high - peak limit , our formula at the one - loop order reproduces the result in ref .
@xcite . in the present paper ,
we further investigated the two - loop contribution , and found a new dominant contribution from the primordial trispectrum , which has not been considered in ref .
@xcite . as another analytic approach , ref .
@xcite has presented a formula for the halo bispectrum based on the peak - background split picture .
the authors particularly examined the case of the single - component non - gaussianity characterized by the curvature perturbation , @xmath205 .
following ref .
@xcite , and using the cancellation properties of the highest - order parameters in press - schechter mass function , we have checked that our formula is consistent with eq .
( 5.1 ) in ref .
@xcite in the large - scale limit .
note that the bispectrum formula in ref .
@xcite additionally includes contributions proportional to @xmath206 or @xmath207 , for which we did not consider here . as shown in the figures in ref .
@xcite , however , these contributions can become significant only at very much large scales ( @xmath208 ) in the squeezed isosceles configuration ( @xmath173 ) . ref .
@xcite investigated the halo bispectrum in cosmological @xmath2-body simulations , and found that the results of their simulations are rather consistent with those predicted by ref .
@xcite , and the dominant contribution of bispectrum scales as @xmath209 in the squeezed limit .
hence , the contributions proportional to @xmath206 or @xmath207 might not be relevant for real observations . although we do not discuss in detail
, we note here that these contributions can be also constructed in the formalism of the ipt .
such contributions come from the primordial higher - order poly - spectra as follows . for the single - sourced case
, the primordial higher - order spectra can be simply parameterized by the non - linearity parameter @xmath0 .
for example , the primordial 5- and 6-point spectra respectively have the dependence of @xmath206 and @xmath207 .
including such primordial higher - order poly - spectra and considering corresponding undecomposable diagrams , we can recover the same contributions as found in ref .
we also find a crude relationship between the contributions from @xmath3 and that from @xmath206 and @xmath207 as @xmath210 and @xmath211 , where we respectively denote the contributions from @xmath206 and @xmath207 as @xmath212 and @xmath213 . from this relation , in order for @xmath212 and @xmath213 to dominate over @xmath80 , we need @xmath214 for @xmath104 .
thus , we stress here that for the accessible scale of observations and with the currently allowed values of non - linearity parameters , one- and two - loop contributions coming from the primordial trispectrum are important for the halo bispectrum . in fig .
[ fig : table1.eps ] , we summarize the consistency between the three analytic formalism for the halo bispectrum in the presence of the primordial non - gaussianity .
in this paper , on the basis of the ipt formalism , we have systematically investigated the bispectrum of the biased objects induced by the primordial non - gaussianity . basically , our formulae in the case of local - type non - gaussianity are consistent with those of the previous works .
notable point in the present paper is that we studied the halo bispectrum in the case of not only the local - type , but also the equilateral- and orthogonal - type primordial non - gaussianity .
further , in case of the local - type , we include the effect of the primordial trispectrum characterized by the two non - linearity parameters , @xmath1 and @xmath3 .
for the local - type non - gaussianity , the two - loop correction coming from the terms proportional to @xmath3 , denoted by @xmath171 , is found to produce a non - negligible contribution , whose amplitude can become comparable to the dominant one - loop corrections proportional to @xmath209 , denoted by @xmath80 . as discussed in the previous section , such non - negligible loop contributions are represented by un - decomposable loop diagrams in the context of ipt , and taking account of such un - decomposable diagrams is rather crucial to precisely investigate the dominant contributions from the primordial non - gaussianity .
we stress here that the diagrammatic approach based on the ipt helps us to systematically collect these dominant contributions .
the significance of higher - order contributions in the case of local - type non - gaussianity is our important findings , and it implies that a much more great impact on the convergence of perturbative expansion is expected in the presence of higher - order primordial polyspectra .
this should deserve further investigation , and we need a more clear and physical understanding of our results . for the current observational
upper - bound of the non - linearity parameters , the contributions from the primordial trispectrum would dominate the halo bispectrum on large scales . taking the typical redshift and the mass of halos in surveys to be @xmath90 and @xmath215 respectively , the halo bispectrum with @xmath105 is comparable to the one induced by @xmath216 with @xmath217 . in ref .
@xcite , the authors mentioned that the future galaxy surveys would have a potential to detect @xmath218 , and this is expected to give @xmath219 .
the current observational limit on @xmath1 is about @xmath220 and hence the bispectrum of the biased object would be a powerful tool to obtain tighter constraint on @xmath1 . for future idealistic wide - field surveys at @xmath221 , in which we have a survey volume of @xmath222 and the observable number density of the halos
is @xmath223 with typical mass @xmath215 , we expect that the detectability is improved to @xmath224 .
it is therefore very interesting to pursue a precise forecast for @xmath1 from the halo / galaxy bispectrum data obtained by the future surveys .
for equilateral- and orthogonal - types of non - gaussianity , we find that the one - loop corrections are all suppressed and the tree - level calculation is sufficient to investigate the dominant contribution of halo bispectrum . following the result obtained by ref .
@xcite , future galaxy surveys can have a potential to obtain a constraint on @xmath225 down to @xmath226 , and it is comparable to the constraint obtained from the cmb observations .
we expect that such future surveys also give a tight constraint on @xmath150 . in this paper
, we did not consider the contribution from the trispectrum in the cases of equilateral- and orthogonal - types non - gaussianity , because their scale dependence is strongly dependent of the model generating the primordial non - gaussianity . also , the exact form of primordial trispectrum is much more complicated compared to that of the local - type non - gaussianity .
we leave the discussion on the contribution from the equilateral or orthogonal trispectrum to future work . finally , while the recent cosmological @xmath2-body simulations suggest that the measured halo bispectrum on large scales is consistent with the previous analytic formulae ref .
@xcite , we believe that our formalism is capable of giving much more precise prediction which quantitatively reproduces the simulation results for various types of the primordial non - gaussianity even on smaller scales . along the line of this
, the detailed comparison with @xmath2-body simulation is our important next subject .
this work was supported by the grant - in - aid for jsps research under grant no .
24 - 2775 ( sy ) , grant - in - aid for scientific research ( c ) , 24540267 , 2012 ( tm ) and also 24540257 ( at ) .
to complete this work , discussions during the yitp workshop yitp - t-13 - 06 on `` the cmb and theory of the primordial universe '' were useful , and we thank yitp .
throughout this paper , we apply the sheth - tormen fitting formula as a mass function in the derivation of the scale - independent bias parameter @xmath191 .
let us discuss the dependence of the halo bispectrum on the choice of the mass function .
for the press - schechter ( ps ) mass function given by @xmath227 we have the lagrangian bias parameter as @xmath228 for the high peak formalism ( hp ) corresponding the large @xmath54 limit , we have @xmath229 the sheth - tormen fitting formula ( st ) is given by @xmath53 \sqrt{q } \nu e^{-q\nu^2/2 } , \label{eq : stmf}\end{aligned}\ ] ] and the mice mass function ( mice ) @xcite is given by @xmath230 e^{-c \nu^2 / \delta_c^2 } , \label{eq : micemf}\end{aligned}\ ] ] with @xmath231 , @xmath232 , @xmath233 , and @xmath234 .
[ fig : fgtnlmf.eps ] , ( black ) , @xmath67 ( red thick ) , @xmath72 ( red thick dashed ) , and @xmath73 ( red thick dotted ) , bottom - @xmath186 ( black ) , @xmath79 ( blue thick ) , @xmath80 ( green thick ) , and @xmath235 ( green thick dashed ) with tthe scale - independent bias parameters given by eq .
( [ eq : hpmf ] ) ( hp ) , and the mass function given by ( [ eq : psmf ] ) ( ps ) , ( [ eq : stmf ] ) ( st ) and ( [ eq : micemf ] ) ( mice ) from left to right.,width=642 ] we plot each contribution of the halo bispectrum , @xmath186 ( black ) , @xmath67 ( red thick ) , @xmath72 ( red thick dashed ) , and @xmath73 ( red thick dotted ) in top panels , and @xmath186 ( black ) , @xmath79 ( blue thick ) , @xmath80 ( green thick ) , and @xmath235 ( green thick dashed ) in bottom panels , with the scale - independent bias parameters given by eq .
( [ eq : hpmf ] ) ( hp ) , and the mass function given by ( [ eq : psmf ] ) ( ps ) , ( [ eq : stmf ] ) ( st ) and ( [ eq : micemf ] ) ( mice ) from left to right . for the halo with mass @xmath91 at @xmath90
, we have @xmath236 . in this figure
, we find that the amplitude of the halo bispectrum is also strongly dependent on the choice of the mass function .
a. slosar , c. hirata , u. seljak , s. ho and n. padmanabhan , jcap * 0808 * , 031 ( 2008 ) [ arxiv:0805.3580 [ astro - ph ] ] . t. giannantonio , a. j. ross , w. j. percival , r. crittenden , d. bacher , m. kilbinger , r. nichol and j. weller , arxiv:1303.1349 [ astro-ph.co ] . n. dalal , o. dore , d. huterer and a. shirokov , phys . rev .
d * 77 * , 123514 ( 2008 ) [ arxiv:0710.4560 [ astro - ph ] ] .
s. matarrese and l. verde , astrophys .
j. * 677 * , l77 ( 2008 ) [ arxiv:0801.4826 [ astro - ph ] ] .
d. jeong and e. komatsu , astrophys .
j. * 703 * , 1230 ( 2009 ) [ arxiv:0904.0497 [ astro-ph.co ] ] .
t. nishimichi , a. taruya , k. koyama and c. sabiu , jcap * 1007 * , 002 ( 2010 ) [ arxiv:0911.4768 [ astro-ph.co ] ] .
t. abbott _ et al . _
[ dark energy survey collaboration ] , astro - ph/0510346 .
d. schlegel _ et al . _
[ bigboss experiment collaboration ] , arxiv:1106.1706 [ astro-ph.im ] .
a. abell _ et al .
_ [ lsst science and lsst project collaborations ] , arxiv:0912.0201 [ astro-ph.im ] .
r. laureijs _ et al . _
[ euclid collaboration ] , arxiv:1110.3193 [ astro-ph.co ] .
r. ellis _ et al . _
[ pfs team collaboration ] , arxiv:1206.0737 [ astro-ph.co ] .
t. matsubara , phys .
d * 83 * , 083518 ( 2011 ) [ arxiv:1102.4619 [ astro-ph.co ] ] .
t. matsubara , phys .
d * 86 * , 063518 ( 2012 ) [ arxiv:1206.0562 [ astro-ph.co ] ] .
s. yokoyama and t. matsubara , phys .
d * 87 * , 023525 ( 2013 ) [ arxiv:1210.2495 [ astro-ph.co ] ] .
m. sato and t. matsubara , arxiv:1304.4228 [ astro-ph.co ] .
t. matsubara , arxiv:1304.4226 [ astro-ph.co ] .
t. suyama and m. yamaguchi , phys .
d * 77 * , 023505 ( 2008 ) [ arxiv:0709.2545 [ astro - ph ] ]
. t. suyama , t. takahashi , m. yamaguchi and s. yokoyama , jcap * 1012 * , 030 ( 2010 ) [ arxiv:1009.1979 [ astro-ph.co ] ] .
n. s. sugiyama , e. komatsu and t. futamase , phys .
* 106 * , 251301 ( 2011 ) [ arxiv:1101.3636 [ gr - qc ] ] .
j. bramante and j. kumar , jcap * 1109 * , 036 ( 2011 ) [ arxiv:1107.5362 [ astro-ph.co ] ] .
n. s. sugiyama , jcap * 1205 * , 032 ( 2012 ) [ arxiv:1201.4048 [ gr - qc ] ] .
j. -o .
gong and s. yokoyama , arxiv:1106.4404 [ astro-ph.co ] . t. nishimichi , jcap * 1208 * , 037 ( 2012 ) [ arxiv:1204.3490 [ astro-ph.co ] ] .
m. biagetti , v. desjacques and a. riotto , arxiv:1208.1616 [ astro-ph.co ] . | we derive a formula for the halo / galaxy bispectrum on the basis of the integrated perturbation theory ( ipt ) .
in addition to the gravity - induced non - gaussianity , we consider the non - gaussianity of the primordial curvature perturbations , and investigate in detail the effect of such primordial non - gaussianity on the large - scale halo / galaxy bispectrum . in ipt ,
the effects of primordial non - gaussianity are wholly encapsulated in the linear ( primordial ) polyspectra , and we systematically calculate the contributions to the large - scale behaviors arising from the three types of primordial bispectrum ( local- , equilateral- , and orthogonal - types ) , and primordial trispectrum of the local - type non - gaussianity .
we find that the equilateral- and orthogonal - type non - gaussianities show distinct scale - dependent behaviors which can dominate the gravity - induced non - gaussianity at very large scales .
for the local - type non - gaussianity , higher - order loop corrections are found to give a significantly large contribution to the halo / galaxy bispectrum of the squeezed shape , and eventually dominate over the other contributions on large scales . a diagrammatic approach based on
the ipt helps us to systematically investigate an impact of such higher - order contributions to the large - scale halo / galaxy bispectrum . |
observations of type ia supernovae ( snia ) @xcite , cosmic microwave background ( cmb ) @xcite and large scale structure ( lss ) @xcite all indicate the existence of mysterious dark energy ( de ) driving the current accelerating expansion of the universe .
the most obvious theoretical candidate of dark energy is the cosmological constant @xmath7 , which can fit the observations well , but is plagued with the fine - tuning problem and the coincidence problem @xcite .
numerous other dynamical de models have also been proposed in the literature , such as quintessence @xcite , phantom @xcite , @xmath8-essence @xcite , tachyon @xcite , hessence @xcite , chaplygin gas @xcite , yang - mills condensate @xcite , ect . actually , the de problem may be in essence an issue of quantum gravity @xcite .
however , by far , a complete theory of quantum gravity has not been established , so it seems that we have to consider the effects of gravity in some effective field theory in which some fundamental principles of quantum gravity should be taken into account .
it is commonly believed that the holographic principle @xcite is just a fundamental principle of quantum gravity .
based on the effective quantum field theory , cohen et al .
@xcite pointed out that the quantum zero - point energy of a system with size @xmath9 should not exceed the mass of a black hole with the same size , i.e. @xmath10 , where @xmath11 is the quantum zero - point energy density , and @xmath12 is the reduced planck mass .
this observation relates the ultraviolet ( uv ) cutoff of a system to its infrared ( ir ) cutoff .
when we take the whole universe into account , the vacuum energy related to this holographic principle can be viewed as dark energy ( its energy density is denoted as @xmath13 hereafter ) .
the largest ir cutoff @xmath9 is chosen by saturating the inequality , so that we get the holographic dark energy density @xmath14 where @xmath15 is a numerical constant characterizing all of the uncertainties of the theory , and its value can only be determined by observations . if we take @xmath9 as the size of the current universe , for instance the hubble radius @xmath16 , then the dark energy density will be close to the observational result . however , hsu @xcite pointed out this yields a wrong equation of state ( eos ) for de .
li @xcite suggested to choose the future event horizon of the universe as the ir cutoff .
this is the original holographic dark energy model .
subsequently , cai proposed @xcite that the age of the universe can be chosen as the ir cutoff , and this model is called agegraphic dark energy model . a new version of this model replacing the age of the universe by the conformal age of the universe was also discussed @xcite , in order to avoid some internal inconsistencies in the original model . furthermore , gao et al .
@xcite proposed to consider the average radius of the ricci scalar curvature as the ir cutoff , and this model is called holographic ricci dark energy model .
( for convenience , hereafter we will call them hde , ade , and rde , respectively . )
although these three models have been studied widely @xcite , so far no comparison of them has been made .
it would be very interesting to constrain the holographic dark energy models by using the latest observational data , and then make a comparison for them by using the proper model selection criterion .
this will be done in this work .
this paper is organized as follows : in section 2 , we briefly review the holographic de models . in section 3 ,
we present the method of data analysis , as well as the model selection criterion . in section 4
, we show the results of the cosmological constraints , and we also use the bayesian evidence ( be ) to make a comparison .
section 5 is a short summary . in this work ,
we assume today s scale factor @xmath17 , so the redshift @xmath18 satisfies @xmath19 ; the subscript `` 0 '' always indicates the present value of the corresponding quantity , and the unit with @xmath20 is used .
in this section , we shall briefly review the holographic de models . for
a spatially flat ( the assumption of flatness is motivated by the inflation scenario ) friedmann - robertson - walker ( frw ) universe with matter component @xmath21 and dark energy component @xmath13 , the friedmann equation reads @xmath22 or equivalently , @xmath23 where @xmath24 is the hubble parameter , @xmath25 is the present fractional matter density , and @xmath26 is the fractional dark energy density . in this work , for simplicity , we shall not discuss the issue of interaction between dark matter and dark energy , so we have @xmath27 @xmath28 where the over dot denotes the derivative with respect to the cosmic time @xmath29 , and @xmath30 is the eos of de . for this model ,
the ir cutoff is chosen as the future event horizon of the universe , @xmath31 taking derivative for eq .
( [ de ] ) with respect to @xmath32 and making use of eq .
( [ rh ] ) , we get @xmath33 combining eqs .
( [ e ] ) and ( [ derivative ] ) , we obtain the eos for hde , @xmath34 directly taking derivative for @xmath35 , and using eq .
( [ rh ] ) , we get @xmath36 where @xmath37 . from eqs .
( [ fri ] ) , ( [ m ] ) , ( [ e ] ) , and ( [ whde ] ) , we have @xmath38 for this case .
so , we have the equation of motion , a differential equation , for @xmath39 , @xmath40 since @xmath41 , we get @xmath42 solving numerically eq .
( [ density 2 ] ) and substituting the corresponding results into eq .
( [ ez ] ) , the key function @xmath43 can be obtained .
it should be mentioned that there are two model parameters , @xmath25 and @xmath15 , in the hde model .
since there are some internal inconsistencies in the original ade model , we will discuss the new version of ade model which suggests to choose the conformal age of the universe @xmath44 as the ir cutoff , so the density of ade is @xmath45 to distinguish from the hde model , a new constant parameter @xmath46 is used to replace the old parameter @xmath15 .
taking derivative for eq .
( [ rhoq ] ) with respect to @xmath47 and making use of eq .
( [ etaq ] ) , we get @xmath48 this means that the eos of ade is @xmath49 taking derivative for @xmath50 , and considering eq .
( [ etaq ] ) , we obtain @xmath51 in this case , we have @xmath52 hence , we get the equation of motion for @xmath53 , @xmath54 and this equation can be rewritten as @xmath55 as in @xcite , we choose the initial condition , @xmath56 , at @xmath57 , then eq .
( [ keyq ] ) can be numerically solved . substituting the results of eq .
( [ keyq ] ) into eq .
( [ ez ] ) , the key function @xmath43 can be obtained .
notice that once @xmath46 is given , @xmath58 can be natural obtained by solving eq.([keyq ] ) , so the ade model is a single - parameter model . for a spatially flat frw universe ,
the ricci scalar is @xmath59 as suggested by gao et al .
@xcite , the energy density of de is @xmath60 where @xmath61 is a positive numerical constant to be determined by observations . comparing to eq .
( [ de ] ) , it is seen that if we identify the ir cutoff @xmath9 with @xmath62 , we have @xmath63 . as pointed out by cai et al .
@xcite , the rde can be viewed as originated from taking the causal connection scale as the ir cutoff in the holographic setting . now
the friedmann equation can be written as @xmath64 and this equation can be further rewritten as @xmath65 where @xmath66 . solving this equation , and using the initial condition @xmath67 , we have @xmath68 there are also two model parameters , @xmath25 and @xmath61 , in rde model .
in the following , we constrain the model parameters of these three de models by using the latest observational data including the union+cfa3 sample of 397 snia , the shift parameter of the cmb given by the five - year wilkinson microwave anisotropy probe ( wmap5 ) observations , and the baryon acoustic oscillation ( bao ) measurement from the sloan digital sky survey ( sdss ) .
first , we consider the latest 397 union+cfa3 snia data , the distance modulus @xmath69 , compiled in table 1 of @xcite .
this dataset add the cfa3 sample to the 307 union sample @xcite .
the theoretical distance modulus is defined as @xmath70 where @xmath71 with @xmath72 the hubble constant @xmath73 in units of 100 km / s / mpc , and @xmath74 is the hubble - free luminosity distance @xmath75 ( here @xmath76 is the physical luminosity distance ) in a spatially flat frw universe , and here @xmath77 denotes the model parameters .
the @xmath78 for the snia data is @xmath79 ^ 2\over \sigma_i^2},\label{ochisn}\ ] ] where @xmath80 and @xmath81 are the observed value and the corresponding 1@xmath82 error of distance modulus for each supernova , respectively .
the parameter @xmath83 is a nuisance parameter but it is independent of the data and the dataset . following @xcite , the minimization with respect to @xmath83 can be made trivially by expanding the @xmath78 of eq .
( [ ochisn ] ) with respect to @xmath83 as @xmath84 where @xmath85 ^ 2\over \sigma_i^2},\ ] ] @xmath86 @xmath87 evidently , eq . ( [ ochisn ] ) has a minimum for @xmath88 at @xmath89 since @xmath90 , instead minimizing @xmath91 one can minimize @xmath92 which is independent of the nuisance parameter @xmath83 .
next , we consider the constraints from cmb and lss observations . for the cmb data , we use the cmb shift parameter @xmath93
; for the lss data , we use the parameter @xmath94 of the bao measurement .
it is widely believed that both @xmath93 and @xmath94 are nearly model - independent and contain essential information of the full cmb and lss bao data .
the shift parameter @xmath93 is given by @xcite @xmath95 where the redshift of recombination @xmath96 @xcite .
the shift parameter @xmath93 relates the angular diameter distance to the last scattering surface , the comoving size of the sound horizon at @xmath97 and the angular scale of the first acoustic peak in cmb power spectrum of temperature fluctuations @xcite .
the measured value of @xmath93 has been updated to be @xmath98 from wmap5 @xcite .
the parameter @xmath94 from the measurement of the bao peak in the distribution of sdss luminous red galaxies is defined as @xcite @xmath99^{2/3},\ ] ] where @xmath100 .
the sdss bao measurement @xcite gives @xmath101 , where the scalar spectral index is taken to be @xmath102 as measured by wmap5 @xcite .
notice that both @xmath93 and @xmath94 are independent of @xmath73 , thus these quantities can provide robust constraint as complement to snia data on the holographic dark energy models .
the total @xmath78 is given by @xmath103 where @xmath104 is given by eq .
( [ tchi2sn ] ) , and the latter two terms are defined as @xmath105 and @xmath106 where the corresponding @xmath107 errors are @xmath108 and @xmath109 , respectively . as usual , assuming the measurement errors are gaussian , the likelihood function @xmath110 the model parameters yielding a minimal @xmath111 and a maximal @xmath112 will be favored by the observations . for comparing different models , a statistical variable must be chosen .
the @xmath113 is the simplest one which is widely used .
however , for models with different number of parameters , the comparison using @xmath78 may not be fair , as one would expect that a model with more parameters tends to have a lower @xmath78 . in this work , we will use the be as a model selection criterion .
the be of a model @xmath114 takes the form @xmath115 where @xmath116 is the likelihood function given by the model @xmath114 and parameters @xmath117 ( @xmath118 denotes the data ) , and @xmath119 is the priors of parameters .
it is clear that a model favored by the observations should give a large be . since be is the average of the likelihood of a model over its prior of the parameter space and automatically includes the penalties of the number of parameters and data , it is more reasonable and unambiguous than the @xmath120 in model selection @xcite .
the logarithm of be can be used as a guide for model comparison @xcite , and we choose the @xmath121cdm as the reference model : @xmath122 .
we will show the results in this section . in fig.[fig1 ] , we plot the contours of @xmath123 and @xmath124 confidence levels in the @xmath125 plane , for the hde model .
the best - fit model parameters are @xmath126 and @xmath127 , corresponding to @xmath128 . for @xmath123 confidence level , @xmath129 , and @xmath130 ; for @xmath124 confidence level , @xmath131 , and @xmath132 .
probability contours at @xmath123 and @xmath124 confidence levels in @xmath125 plane , for the hde model .
the round point denotes the best - fit values , @xmath126 and @xmath127 , corresponding to @xmath128.,width=302 ] the @xmath111 and the corresponding likelihood @xmath112 of the ade model.,width=302 ] probability contours at @xmath123 and @xmath124 confidence levels in @xmath133 plane , for the rde model .
the round point denotes the best - fit values , @xmath134 and @xmath135 , corresponding to @xmath136.,width=302 ] comparison of the observed @xmath137 , as square dots , with the predictions from the holographic de models .
, width=302 ] in fig.[fig2 ] , we plot the @xmath111 and the corresponding likelihood @xmath112 of the ade model .
the best - fit model parameter is @xmath138 , giving @xmath139 . for @xmath123 confidence level , @xmath140 ; for @xmath124 confidence level , @xmath141 . in fig.[fig3 ]
, we plot the contours of @xmath123 and @xmath124 confidence levels in the @xmath142 plane , for the rde model .
the best - fit model parameters are @xmath134 and @xmath135 , corresponding to @xmath136 .
for @xmath123 confidence level , @xmath143 , and @xmath144 ; for @xmath124 confidence level , @xmath145 , and @xmath146 .
therefore , among these three holographic de models , the hde model can give the smallest @xmath3 . as a comparison
, we also fit the @xmath7cdm model to the same observational data , and find that the minimal @xmath147 for the best - fit parameter @xmath148 .
note that both hde model and rde model are two - parameter model , while ade model is a single - parameter model .
so , it is not suitable to choose the @xmath113 as a model selection criterion . in table
[ be ] , we list the results of @xmath149 for three holographic de models .
it is seen that although hde model performs a little poorer than @xmath7cdm model , it performs better than ade model and rde model .
therefore , among these three holographic de models , hde is more favored by the observational data .
[ be ] .the results of @xmath149 for three holographic de models . [ cols="^,^,^,^",options="header " , ] in fig.[fig4 ] , we further compare the observed expansion rate @xmath137 @xcite with that predicted by these three holographic de models ( for each model , the best - fit values determined by snia+cmb+bao analysis are taken ) . notice that the area surrounded by two dashed lines shows the @xmath150 confidence interval @xcite , and a de model would be disfavored by the observation if it gives a curve of @xmath137 that falls outside this area .
it is seen that among these three holographic de models , only the curve of @xmath137 predicted by the hde model falls inside this confidence interval .
this result verifies our conclusion , from another perspective .
in this work , we perform the best - fit analysis on three holographic de models , by using the latest observational data including the union+cfa3 sample of 397 snia , the shift parameter of the cmb given by the wmap5 observations , and the bao measurement from the sdss . the analysis shows that for hde , @xmath0 ; for rde , @xmath1 ; for ade , @xmath2 . among these three de models , the hde model can give the smallest @xmath3 , which is even smaller than that given by the @xmath7cdm model .
moreover , we use the be as a model selection criterion to make a comparison .
although hde model performs a little poorer than @xmath7cdm model , it performs better than ade model and rde model .
therefore , among these three de models , hde is more favored by the observational data . finally , adopting the best - fit values determined by snia+cmb+bao analysis , we compare the observed expansion rate @xmath137 with that predicted by these three holographic de models .
it is found that only the curve of @xmath137 predicted by hde model can fall inside the @xmath150 confidence interval .
this verifies our conclusion from another perspective .
it would be interesting to give tighter constrains on de models by adding other cosmological observations , such as gamma - ray bursts @xcite , chandra x - ray observation @xcite , and some old high - redshift objects @xcite .
this will be studied in a future work .
we would like to thank hao wei , yan gong , chang - jun gao , and feng - quan wu , for helpful discussions .
this work is supported by the natural science foundation of china .
riess et al . ,
* 116 * , 1009 ( 1998 ) ; s. perlmutter et al .
, apj * 517 * , 565 ( 1999 ) ; j. l. tonry et al . ,
apj * 594 * , 1 ( 2003 ) ; r.a .
knop et al . , apj * 598 * , 102 ( 2003 ) ; a.g .
riess et al . , apj * 607 * , 665 ( 2004 ) .
bennet et al . , apjs .
148 , 1 ( 2003 ) ; d.n .
spergel et al . ,
apjs , 148 , 175 ( 2003 ) ; d.n .
spergel et al . , apjs * 170 * , 377 ( 2007 ) ; l. page et al . , apjs * 170 * , 335 ( 2007 ) ; g. hinshaw et al . , apjs * 170 * , 263 ( 2007 ) .
s. weinberg , rev .
* 61 * , 1 ( 1989 ) ; arxiv : astro - ph/0005265 ; v. sahni and a.a .
starobinsky , int . j. mod
d * 9 * , 373 ( 2000 ) ; s.m .
carroll , living rev.rel . * 4 * , 1 ( 2001 ) ; p.j.e .
peebles and b. ratra , rev .
phys . * 75 * , 559 ( 2003 ) ; t. padmanabhan , phys . rept . * 380 * , 235 ( 2003 ) ; e. j. copeland , m. sami and s. tsujikawa , int . j. mod .
d * 15 * , 1753 ( 2006 ) .
b. ratra and p.j.e .
peebles , phys .
rev . d*37 * , 3406 ( 1988 ) ; p.j.e .
peebles and b.ratra , apj * 325 * , l17 ( 1988 ) ; c. wetterich , nucl .
b*302 * , 668 ( 1988 ) ; i. zlatev , l. wang and p.j .
steinhardt , phys .
82 * , 896 ( 1999 ) . c. armendariz - picon ,
t. damour and v. mukhanov , phys .
b * 458 * , 209 ( 1999 ) ; c. armendariz - picon , v. mukhanov and p.j .
steinhardt , phys .
rev . d*63 * , 103510 ( 2001 ) ; t. chiba , t. okabe and m. yamaguchi , phys .
rev . d*62 * , 023511 ( 2000 ) .
kamenshchik , u. moschella and v. pasquier , phys .
b * 511 * 265 ( 2001 ) ; m.c .
bento , o. bertolami and a.a .
sen , phys .
d * 66 * 043507 ( 2002 ) ; x. zhang , f.q . wu and j. zhang , jcap * 0601 * 003 ( 2006 ) .
huang and m. li , jcap * 0408 * , 013 ( 2004 ) ; q.g .
huang and m. li , jcap * 0503 * , 001 ( 2005 ) ; q.g .
huang and y.g .
gong , jcap * 0408 * , 006 ( 2004 ) ; x. zhang and f.q .
wu , phys .
d * 72 * , 043524 ( 2005 ) ; x. zhang , int .
d * 14 * , 1597 ( 2005 ) ; z. chang , f.q .
wu and x. zhang , phys .
b * 633 * , 14 ( 2006 ) ; b. wang , e. abdalla and r.k .
su , phys . lett .
b * 611 * , 21 ( 2005 ) ; b. wang , c.y .
lin and e. abdalla , phys .
b * 637 * , 357 ( 2006 ) ; x. zhang , phys .
b * 648 * , 1 ( 2007 ) ; m.r .
setare , j. zhang and x. zhang , jcap * 0703 * , 007 ( 2007 ) ; j. zhang , x. zhang and h. liu , phys .
b * 659 * , 26 ( 2008 ) ; b. chen , m. li and y. wang , nucl .
b * 774 * , 256 ( 2007 ) ; j. zhang , x. zhang and h.y .
liu , eur .
j. c * 52 * , 693 ( 2007 ) ; x. zhang and f.q .
wu , phys .
d * 76 * , 023502 ( 2007 ) ; c.j . feng , phys .
b * 633 * , 367 ( 2008 ) ; y.z .
ma , y. gong and x.l .
chen , eur .
j. c * 60 * , 303 ( 2009 ) ; y.z .
ma , y. gong and x.l .
chen , arxiv:0901.1215 ; m. li , c. lin and y. wang , jcap * 0805 * , 023 ( 2008 ) ; m. li , x.d .
li , c. lin and y. wang , commun .
* 51 * , 181 ( 2009 ) ; m. jamil , m.u .
farooq and m.a .
rashid , eur .
j. c * 61 * , 471 ( 2009 ) .
h. wei and r.g .
cai , phys .
b * 655 * , 1 ( 2007 ) ; h. wei and r.g .
cai , phys .
b * 663 * , 1 ( 2008 ) ; i.p .
neupane , phys .
d * 76 * , 123006 ( 2007 ) ; m. maziashvili , phys . lett .
b * 666 * , 364 ( 2008 ) ; j. zhang , x. zhang and h. liu , eur .
j. c * 54 * , 303 ( 2008 ) ; j.p .
ma and y. ling , phys .
b * 663 * , 152 ( 2008 ) ; h. wei and r.g .
cai , eur .
j. c * 59 * , 99 ( 2009 ) ; j. cui , l. zhang , j. zhang and x. zhang , arxiv:0902.0716 .
feng , phys .
b * 670 * , 231 ( 2008 ) ; c.j .
feng , phys .
b * 672 * , 94 ( 2009 ) ; l.n .
granda and a. oliveros , phys .
b * 669 * , 275 ( 2008 ) ; c.j .
feng , arxiv:0812.2067 ; k.y .
kim , h.w .
lee and y.s .
myung , arxiv:0812.4098 ; x. zhang , arxiv:0901.2262 ; c.j .
feng and x. zhang , arxiv:0904.0045 ; s. nojiri and s.d.odintsov , gen .
. grav . * 38 * , 1285 ( 2006 ) .
p. mukherjee , d.r .
parkinson and a.r .
liddle , apj . * 638 * l51 ( 2006 ) ; p. mukherjee , d.r .
parkinson , p.s .
corasaniti , a.r .
liddle and m. kunz , mnras . *
369 * , 1725 ( 2006 ) ; m. kunz , r. trotta and d. r. parkinson , phys .
d * 74 * , 023503 ( 2006 ) ; a. r. liddle , mnras . * 377 * , l74 ( 2007 ) ; r. trotta , mnras . *
378 * , 72 ( 2007 ) . | the holographic principle of quantum gravity theory has been applied to the dark energy ( de ) problem , and so far three holographic de models have been proposed : the original holographic dark energy ( hde ) model , the agegraphic dark energy ( ade ) model , and the holographic ricci dark energy ( rde ) model . in this work ,
we perform the best - fit analysis on these three models , by using the latest observational data including the union+cfa3 sample of 397 type ia supernovae ( snia ) , the shift parameter of the cosmic microwave background ( cmb ) given by the five - year wilkinson microwave anisotropy probe ( wmap5 ) observations , and the baryon acoustic oscillation ( bao ) measurement from the sloan digital sky survey ( sdss ) .
the analysis shows that for hde , @xmath0 ; for rde , @xmath1 ; for ade , @xmath2 . among these models
, hde model can give the smallest @xmath3 .
besides , we also use the bayesian evidence ( be ) as a model selection criterion to make a comparison .
it is found that for hde , ade , and rde , @xmath4 , @xmath5 , and @xmath6 , respectively .
so , it seems that the hde model is more favored by the observational data . |
in recent days diabetic retinopathy ( dr ) is one of the severe eye diseases causing blindness . with early stage detection and treatment
the patient can be saved from losing sight .
automatic computer aided diagnosis system will reduce burden on specialists . also for monitoring and checking the progress of disease efficiently , automatic system will perform much better than human in terms of manual evaluation time . since comparison and evaluation of images manually is a time consuming task and images are subject to various distortions . for accurate analysis of progress of diabetic retinopathy , detection of exudate
is mandatory .
exudates are primary clinical symptoms of diabetic retinopathy
. two types of exudates namely soft exudate and hard exudate appear .
hard exudates are visible in non - proliferative diabetic retinopathy and soft exudates ( cotton wool spots ) in proliferative diabetic retinopathy .
hard exudates represent the accumulation of lipid in or under the retina secondary to vascular leakage and visible as discrete yellowish deposits in color fundus images .
cotton - wool spots are nerve fibre layer infarcts and they are visible as pale white rather than yellow . also exudates are variable in sizes and shapes . a typical pathological retinal image is depicted in fig . 1 to show features like the optic disc , the macula , the blood vessels and exudates .
+ for identifying the stage of dr , classification of soft and hard exudates is foremost important to distinguish them for other retinal pathological features like drusen , heamorrhages , microaneurysms etc .
fundus images are prone to artifacts related to defocus , motion blur , fingerprints etc .
this artifacts are prone to interpretation as pathological features because of their similarity with exudates and drusen
. one important defect of fundus image is luminosity and contrast variation , which is improperly addressed in many exudate detection methods .
several methods have been presented for detection of exudates in colour fundus photograph using image processing and machine learning algorithms .
sanchez et al .
@xcite use a mixture model to separate exudate from background .
an edge detection based method was used to remove other outliers .
giancardo et al .
@xcite use a kirsch s edge method to assign score for exudate candidate to a pre - processed .
they have used background estimation and image normalization for pre - processing . for classification and detection of drusen , exudates and cotton wool spots ,
a pixel wise classification algorithm is presented by niemeijer et al .
zhang et al .
@xcite use mathematical morphological and contextual features for candidate extraction followed by random forest based classification method for exudate detection .
fuzzy c - means clustering based method was used for segmentation of features and neural network classifier for exudate detection by osareh et al .
rocha et al .
@xcite have addressed the problem of detecting bright and red lession by using one novel algorithm , they made use of surf features with machine learning method .
but they failed to properly address the problems of luminance variation and artifacts . in this work
, we propose a novel exudate detection method using gaussian scale space based interest map ( gimap ) and mathematical morphology .
this approach is robust to artefacts and illuminance variation .
secondly a disease severity prediction method is developed by using information of exudate location with respect to the macula region and the optic disc .
in addition to that we propose classification system using svm for hard and soft exudate .
distinguishing exudate as hard and soft important for severity prediction also to identify type of diabetic retinopathy , whether it is non - proliferative or proliferative .
section 2 presents our method . experimental setup and the results
are discussed in section 3 and finally conclusions are are drawn in section 4 .
exudate are the bright lessons found in retinal image , caused due to diabetic retinopathy , a most common disorder of eye with patient having diabetes .
it is also a main reason of blindness . for detection of the optic disc and fovea , the method described by niemeijer et al .
@xcite was used .
steps involve in this method is shown in fig .
2 . for correct detection and classification of hard and soft exudates
it is important to reduce noise while preserving edges . in this method
anisotropic diffusion filtering @xcite was used for reducing noise in the images and preserving edges .
anisotropic diffusion filters have been proved to be succesful in edge preserving smooting and denoising of medical images .
hard and soft exudates possesses discriminative edge structures , preserving this edge while reducing noise is also one of important step towards classification .
step involves in filtering is as follows @xcite .
@xmath0 @xmath1 @xmath2 where i is the input image , ( r , g , b ) is color channel of it , k is diffusion constant and @xmath3 is noise standard deviation over which algorithm will be iterated to find solution .
the diffusion function @xmath4 is a monotonically decreasing function of the image gradient magnitude .
gaussian scale space based interest map can capture local image structure and scale invariant image features . since exudate varies on the basis of size and hence different exudate will response to different scales of gaussian . for identifying all exudate interest map using gaussian scale space was constructed @xcite .
first step in the construction of gimap is the computation of 1st derivative of gaussian as filters at several scales and smoothed the derivative using gaussian filter and take absolute values of derivatives .
in addition to that laplacian of gaussian for each scale was computed .
because of colour difference between hard and soft exudates , they respond variably to colour channels . now taking the maximum response of all the colour channels final interest map for the scale
is constructed .
this process is repeated over different scales . for each scale
we have two filter 1st derivative gaussian and laplacian of gaussian , take maximum of absolute values from both filter ouputs in the interest map . since our concerned features may appear in variable scales so interest map of each individual map is combined using maximum operation to form a decision making interest map . scale used in these works are @xmath5{2},k\sqrt[]{2 } ... k^{n}\sqrt[]{2})$ ] .
the decision making interest map also contain some outliers such blood vessels , vein , haemorrhage etc .
filtering process is described by following equations @xcite .
@xmath6 @xmath7 @xmath8 for computation of laplacian of gaussian we have used difference of gaussian approximation . if @xmath9 be a filtered image , @xmath10 $ ] is its three color channel at a scale @xmath11 using filter @xmath12 , then the interest map for this scale is obtained by selecting maximum response over color channel as follows . @xmath13 and the decision making interest map is obtained taking maximum over filter @xmath12 and then scale @xmath11 as follows .
@xmath14 the decision map obtained from above procedure need to be improved by reducing outliers present .
first step is to enhance the features by using grayscale morphological operations . here
closing operation with a disk structuring element of size 2 and 3 was used .
@xmath15 @xmath16 if @xmath17 and @xmath18 are obtained after closing operation , then resulting enhanced image is obatined by following operation . @xmath19 where @xmath20 are pixels position in both images . in the second step
convert the interest map to binary map . for
binarization of the interest map sauvola s @xcite local adaptive thresholding technique is used .
local threshholding is efficient in this particular situation beacuse pixels values of hard and soft exudates vary significantly and to get both of those in final map we need to binarize using local windows .
let @xmath21 is interest map image , take a window size of @xmath22 , @xmath23 and @xmath24 be the mean and standard deviation of window cantered at @xmath25 , then threshold @xmath26 will be defined as follows .
@xmath27\ ] ] where @xmath28 is maximum of standard deviation of all windows and @xmath29 $ ] is a parameter . along with exudate other features such as blood vessels and veins
is also detected due to gaussian scale space .
morphological opening operation and connected component analysis was used for removing those unwanted features .
blood vessels and veins are characterized by thin long geometrical structures .
a rectangle structural element having width and length larger than the thickness of those blood vessels is used for removing those structures .
remaining vessels and outliers left after this operation will be removed by using concept of convexity and connected component analysis . for each connected region we will compute convex hull of the region and define the following @xmath30 eq . ( 13 ) term . @xmath31
the @xmath30 term computes probability of a connected region being exudate or not .
it is the ratio of the area of the region convex hull and its area .
this term will give approximate idea of the connected region .
if the value of @xmath32 then the region will be discarded , the connected region will be considered as vessel structures or flash artifacts . where @xmath33 is convex hull of the region @xmath34 and @xmath35 denotes the pixel positions . if convexity of a connected region is less than 0.2 then the area will be discarded .
also we want to ensure that artifacts like flares gets removed from gimap . from observation
it has been noticed that flares are circular shape object .
none of the features like hard and soft exudate have shapes like that of flares , approximate circular .
compactness ( @xmath36 ) is defined below eq . ( 14 ) of every connected region of the gimap should be less than @xmath37 .
@xmath38 where @xmath39 and @xmath40 are area and perimeter of the region @xmath34 respectively . in the final step
we will use colour and luminance features to select exudates and remove outliers , in general we did nt observe outliers like flares , flash etc after adapting previos mentioned steps .
but still there are possibility fo including drusens .
hard and soft exudates are characterized by its yellowish and whitish color significantly different from that of haemorhages , microanerysms and other related features . at the end , a support vector machine classifier is used to get final labels of hard and soft exudates .
two types of exudates namely hard and soft are associated with different diabetic retinopathy symptoms .
we will use a support vector machine classifier to differentiate between soft and hard exudates and outliers .
svm is one of most effective classifier , with its good generalization ability .
selection and extraction of feature vector is one of the challenging task for efficient classification .
since hard and soft exudates are varied in color magnitude we will use mean r , g , b values as feature for each detected regions .
in addition , to ensure luminance invariance in classification process , images will be converted to lab color space and values of a , b channels will be used as features .
both hard and soft exudates are varied in terms of shape and edge structures .
hard exudates are chracterizes by uneven edge and soft exudates are by smooth edge and circular structures .
firstly we will design features vector for classification by using following data .
tabel 1 shows the features used in this method for svm . *
20c feature & description + 1 & area of connected region + 2 - 4 & mean r , g , b intensity values + 5 - 7 & standard deviation of r , g , b intensity values + 8 - 9 & mean a , b intensty values in lab color space + 10 - 11 & standard deviation of a , b intensity values + 12 & ecentricity of the region + 13 & extent of the region + 14 - 15 & major and minor axis length + 16 & convexity + 17 & gradient values of edge pixels of the region + 18 & compactness of the region + 19 & energy of the region + 20 - 22 & color contrast with neighbour regions + features vector constructed from above ideas is 22 dimensional . for svm classification , we have three class specifically hard exudate , soft exudate and outliers . system is trained with hard , soft exudates and non exudates pixels data including artifacts from expert labeled images . as per ophthalmologists , exudate ( or other retinal pathology such as microaneurysms , heamorrhages , drusen etc . )
location with respect to the fovea and the optic disc determine severity of the diseases .
if exudates location are close to those features and causes defect to those features then patients may lose vision .
as per the norm of the international council of ophthalmology diabetic retinopathy @xcite divided in five groups specifically : none , mild , moderate , severe , and proliferate .
we will define some rule to detect severity by using the fovea and the optic disc location as centre of two set of concentric circles .
radius of each concentric circles with the fovea location as centre is multiple of 80 pixels .
the radius selection for each circle based on the average value of the optic disc width and height .
also for each concentric circles with the optic disc centre location as circle centre is multiple of 55 pixels .
this setting valid for images with width 1500 pixels and height 1152 pixels .
if exudates reside in vey inner circle then we label it as @xmath41 , other consecutive circle will be known as @xmath42 , where n is the index number of the circle , bigger n means circle is of bigger radius .
for each circle area of exudate(no of pixels ) will be calculated . finally by using area and
location information diseases severity will be predicted .
if location of exudate in @xmath41 circle then on the basis of its areas it will included in proliferate or severe groups . here
area of @xmath43 circle means area of @xmath44 circle .
3 depicts the scenarios described above .
@xmath45 where @xmath46 denotes area of exudate region , @xmath47 its distance from either the optic disc center or the fovea and @xmath48 $ ] are two constant function of distance .
@xmath49 @xmath50 @xmath51 @xmath52 where @xmath53 is a exudate pixel , @xmath54 denotes @xmath55 circle and @xmath56 its corresponding area .
for analysing the accuracy of our method on different set of images , we have used publicly available diaretdb1 @xcite and e - ophthaex @xcite dataset . images with various difficulty from these databases have been chosen for testing our algorithm accuracy , this dataset include image with pathological features such as haemorrhages , exudates and microaneurysms .
this dataset provided by experts in ophthalmology with proper pixel wise annotation of features location in images .
this three dataset includes 400 retinal images with variety of pathological symptoms . for implemetation matlab platform in a windows 8.1 machine with intel i7 processor
was used . for detection of exudate gaussian scale space
was constructed using 10 different scale . at first each images
is resized by a factor of @xmath57 using cubic interpolation method for reducing computational time . for each scale first derivative of gaussian and laplacian of gaussian
is computed for analysing structure of exudates present in retinal images . by using those values a decision making interest map
is formed .
different types of outliers were removed using morphological connected component analysis and opening operation . for accuracy analysis of exudates detection
we will compute true positive ( tp ) a number of exudates pixels correctly detected , false positive ( fp ) a number of non - exudate pixels which are detected wrongly as exudate pixels , false negative ( fn ) number of exudate pixels that were not detected and true negative ( tn ) a number of no exudates pixels which were correctly identified as non - exudate pixels .
also sensitivity and specificity at pixel level is computed .
thus the global sensivity se and the global specificty sp and accuracy ac for each image is defined as follows .
@xmath58 a detailed result of accuracy obtained is illustrated in table 1 .
.result of exudate detection [ cols="<,<,<,<,<",options="header " , ]
a novel method for computer aided diagnosis of retinal image for exudate detection and analysis is proposed .
we have got considerable accuracy over two different dataset comprised of several varieties of images specifically illuminance changes , with other pathological features etc . also our machine learning based classifier svm works very well for exudate classification with designed features .
this system can be used for automated processing of pathological images related to diabetic retinopathy , also will be very effective for mass screening . in near future , we will incorporate microaneurysms and haemorrhages detection to the system to enhance its credibility to evaluate the degree of diabetic retinopathy snchez , c. , garca , m. , mayo , a. , lpez , m. , hornero , r. , 2009 . retinal image analysis based on mixture models to detect hard exudates .
image anal .
13 ( 4 ) , 650658 .
giancardo , l. , meriaudeau , f. , karnowski , t. , li , y. , garg , s. , tobin , k. , chaum , e. , 2012 .
exudate - based diabetic macular edema detection in fundus images using publicly available datasets
image anal . 16 ( 1 ) , 216226 .
niemeijer , m. , van ginneken , b. , russel , s. , suttorp - schulten , m. , abrmoff , m.d . , `` automated detection and differentiation of drusen , exudates , and cottonwool spots in digital color fundus photographs for diabetic retinopathy diagnosis '' , investigative ophthalmology and visual science 48 , 2007,22602267 .
zhang et al .
`` exudate detection in color retinal images for mass screening of diabetic retinopathy '' , med .
image anal .
, 18 ( 7 ) ( 2014 ) , pp .
10261043 osareh , a. shadgar , b. ; markham , r. a computational - intelligence - based approach for detection of exudates in diabetic retinopathy images , information technology in biomedicine , ieee transactions on,2009 , 535 - 545 .
meindert niemeijer , michael d. abrmoff , bram van ginneken , `` fast detection of the optic disc and fovea in color fundus photographs '' medical image analysis , 2009 .
p perona , j malik , `` scale - space and edge detection using anisotropic diffusion '' , pattern analysis and machine intelligence,1990 t. lindeberg , `` scale space theory in computer vision , '' in kluwer , 1994 .
rachid deriche .
recursively implementating the gaussian and its derivatives .
[ research report ] rr-1893 , 1993 , pp.24 .
< inria-00074778 > j. sauvola and m. pietikainen , `` adaptive document image binarization , '' pattern recognition 33(2 ) , pp .
225236 , 2000 kauppi , t. , kalesnykiene , v. , kamarainen , j .- k . , lensu , l. , sorri , i. , raninen a. , voutilainen r. , uusitalo , h. , klviinen , h. , pietil , j. , diaretdb1 diabetic retinopathy database and evaluation protocol , in proc of the 11th conf . on medical image understanding and analysis ( aberystwyth , wales , 2007 ) http://www.adcis.net/en/downloadthirdparty/messidor.html chang , chih - chung and lin , chih - jen , libsvm : a library for support vector machines , acm transactions on intelligent systems and technology,2011 .
http://www.icoph.org/taskforce-documents/diabetic-retinopathy-guidelines.html rocha , a. , carvalho , t. , jelinek , h. f. , goldenstein , s. , wainer , j. , `` points of interest and visual dictionaries for automatic retinal lesion detection '' , biomedical engineering , ieee transactions on , 59(8 ) , 2244 - 2253 , 2012 . c. sinthanayothin
, j. f. boyce , t. h. williamson , h. k. cook , e. mensah , s. lal , d. usher , `` automated detection of diabetic retinopathy on digital fundus images '' , diabet .
19(2 ) ( 2002 ) 105 .
t. walter , j. c. klein , p. massin , a. erginay , `` a contribution of image processing to the diagnosis of diabetic retinopathy detection of exudates in color fundus images of the human retina '' , ieee trans .
imaging 21(10 ) ( 2002 ) 1236 . | in the context of computer aided diagnosis system for diabetic retinopathy , we present a novel method for detection of exudates and their classification for disease severity prediction .
the method is based on gaussian scale space based interest map and mathematical morphology .
iit makes use of support vector machine for classification and location information of the optic disc and the macula region for severity prediction .
it can efficiently handle luminance variation and it is suitable for varied sized exudates .
the method has been probed in publicly available diaretdb1v2 and e - ophthaex databases . for exudate detection
the proposed method achieved a sensitivity of 96.54% and prediction of 98.35% in diaretdb1v2 database .
exudate , diabetic rationopathy , image processing |
the large hadron collider ( lhc ) at cern , will collide heavyions at unprecedented high energies , exceeding by a factor 30 the energy available at rhic @xcite .
the main objective of alice ( a large ion collider experiment ) @xcite , is to study matter under extreme conditions of energy density to gain a better understanding of the fundamental properties of the strong interaction . in particular
, alice will explore the quark - gluon plasma ( qgp ) , the state of deconfined matter predicted by qcd @xcite .
the medium formed in heavy - ion collisions can be best probed by hard scattered partons produced in 2@xmath42 qcd processes at the leading order ( lo ) including in the final state a hard direct photon ( compton scattering : q + g @xmath5+q and quark annihilation : q+@xmath6+g ) .
on one hand , the 4-momentum of the scattered parton is modified while traversing the medium , and on the other hand , the scattered photon does not interact , thus providing a reference for the 4-momentum of the partner parton .
hence , from the modification experienced by the hard scattered partons , measured though photon tagged jets , the medium properties can be inferred . in particular , since these hard scattering processes sample the entire collision volume , the final state hadronic observables provide a real tomographic probe of the medium @xcite .
several algorithms @xcite have been developed to identify @xmath7-jet events in p
p and pb pb collisions , demonstrating the feasibility of such measurements with the alice detectors .
however , the jet identification remains challenging in the heavy - ion environment in particular for the energies @xmath8 gev where @xmath7-jet events are measurable in alice with sufficient statistics .
an equivalent approach is to measure direct - photon hadrons correlation @xcite . in the following ,
we have first established the intrinsic properties ( @xmath9 ) of @xmath7-jet events expected in pp collisions at lhc energies .
then we discuss the nucleus - nucleus ( aa ) collision case , in particular , we explore the possibility to select @xmath7-jet events as a function of their localization in the medium to validate the tomographic approach .
at leading order perturbative qcd , a pair of hard - scattered partons emerges exactly back - to - back in the center of mass of the partonic system . due to the finite size of the proton
, however , it was found that each of the colliding parton carries initial transverse momentum with respect to the colliding axis , originally described as `` intrinsic @xmath9 '' . beyond the leading order , initial and final state radiations ( isr / fsr ) will generate additional transverse momentum .
therefore , the resulting total transverse momentum of the outgoing parton pair causes an acoplanar and a momentum imbalance , @xmath10 @xcite .
it is measured as the net transverse momentum of the outgoing parton - pair @xmath11 .
it is anticipated that medium effects will generate additional transverse momentum resulting in a broadening of @xmath9 .
this transverse momentum broadening can be directly related to the transport parameter @xmath12 , which describes the transverse momentum transferred from the medium to the traversing parton @xcite . using the pythia event generator @xcite ,
we have established the collision energy dependence of @xmath10 from @xmath7-jet events , by taking available data from different experiments measurements @xcite and extrapolate to the lhc energies .
the dependence is @xmath13 with @xmath14 and @xmath15 . to study the dependence of @xmath10 with the transverse momentum of the hard scattering , we have generated @xmath7-jet and jet - jet events with pythia generator in different @xmath1 bins with collision energy 14 tev , within @xmath9 setting predicted above and isr / fsr on .
the averaged @xmath16 versus the transverse momentum , shows a weakly linear dependence .
the tomography measurement can be performed by selecting @xmath7-h pairs with different values of the parameter @xmath17 = -@xmath18 .
this criteria can effectively control hadron emission from different regions of the medium and therefore extract the corresponding jet modification parameters @xcite . to simulate the medium induced energy loss , we used the monte - carlo model qpythia @xcite , which combines an energy loss mechanism @xcite and a realistic description of the collision geometry @xcite .
the hijing @xcite generator was used to simulate the underlying events of heavy - ion collisions and pythia to simulate pp collisions .
three samples of @xmath7-jet events were generated with photon energy larger than 20 gev .
a first sample of pp collisions at 5.5 tev generated with pythia provides the baseline .
the second sample consists in similar events modified by qpythia merged with central collision events .
the last sample is obtained by merging the pythia events and peripheral collision events .
tagging events with a direct photon well identified @xcite by the alice calorimeters and measuring the distribution of hadrons emitted oppositely to the photon as a function of @xmath17 , allows us to determinate the jet fragmentation function @xcite .
the underlying event is subtracted by correlating the isolated photon with charged hadrons emitted on the same side as the photon , in the azimuthal range @xmath19 . to quantify the medium modification , @xmath20
is calculated ( fig .
[ fig : iaa ] ) @xmath21 as the ratio of @xmath7-hadrons correlation distribution measured in aa and pp collisions .
the expected enhancement at low @xmath17 and suppression at high @xmath17 for central collision is observed , whereas , @xmath20 is equal to 1 for peripheral collisions , where quenching effects are absent . to illustrate the selectivity of the tomographic measurement , the length , l
, the jet travels inside the medium is calculated .
[ fig : qlch ] indicates that most high @xmath1 leading particles are preferentially produced at the surface ( small l ) , while low @xmath1 leading particles are produced inside the whole volume ( large l ) , which demonstrates the l dependence of of the @xmath7 tagged charged hadron production for 2 different @xmath17 regions .
we have then studied the l dependence of the medium modification factor @xmath22 ( fig . [
fig : iaal ] ) by selecting different @xmath17 regions . for large @xmath17 particles ,
an obvious suppression is observed , and the suppression is stronger with increasing the medium length . for small @xmath17 , the opposite behavior is obtained as an enhancement ( @xmath23 1 ) .
this result implies that @xmath7-hadrons correlation could be used to probe volume versus surface emission by selecting @xmath7-jet events with different @xmath17 values .
however such l dependence will be challenging to measure in the experiments .
@xmath7+jet studies are widely recognized as a powerfull tool to characterize qgp .
the `` @xmath7+jet tomography '' study will enable us to extract jet quenching parameters in different regions of the dense medium via measurement of the nuclear modification factor of @xmath7-hadrons correlation . | @xmath0jet events provide a tomographic measurement of the medium formed in heavy ion collisions at lhc energies .
tagging events with a well identified high @xmath1 direct photon and measuring the correlation distribution of hadrons emitted oppositely to the photon , allows us to determine , with a good approximation , both the jet fragmentation function and the back - to - back azimuthal misalignement of the direct photon and the jet .
comparing these two observables measured in @xmath2 collisions with the ones measured in @xmath3 collisions will reveal the modifications of the jet structure induced by the medium formed in @xmath3 collisions and consequently will infer the medium properties .
direct photon , qgp , jet structure , tomography , path length 24.85.+p , 25.75.bh , 25.75.cj , 25.75.nq 2 |
these lecture notes were written for the american mathematical society ( ams ) short course on quantum computation held 17 - 18 january 2000 in conjunction with the annual meeting of the ams in washington , dc in january 2000 .
the objective of this lecture is to discuss quantum entanglement from the perspective of the theory of lie groups .
more specifically , the ultimate objective of this paper is to quantify quantum entanglement in terms of lie group invariants , and to make this material accessible to a larger audience than is currently the case .
these notes depend extensively on the material presented in lecture i @xcite .
it is assumed that the reader is familiar with the material on density operators and quantum entanglement given in the ams short course lecture i , i.e. , with sections 5 and 7 of @xcite . of necessity , the scope of this paper
is eventually restricted to the study of qubit quantum systems , and to a specific problem called the _ restricted fundamental problem in quantum entanglement _ ( _ rfpqe _ ) .
references to the broader scope of quantum entanglement are given toward the end of the paper .
+ figure 1 . quantum entanglement lab ? ? ? at first sight , a physics research lab dedicated to the pursuit of quantum entanglement might look something like the drawing found in figure 1 , i.e. , like an indecipherable , incoherent jumble of wires , fiber optic cable , lasers , bean splitters , lenses . perhaps some large magnets for nmr equipment , or some supercooling equipment for rf squids are tossed in for good measure . whatever .. it is indeed a most impressive collection of adult `` toys . ''
however , to a mathematician , such a lab appears very much like a well orchestrated collection of intriguing mathematical `` toys , '' just beckoning with new tantalizing mathematical challenges . in the hope of piquing your curiosity to read on , we give the following brief preview of what is to come : the rfpqe reduces to the mathematical problem of determining the orbits of the big adjoint action of the group of local unitary transformations @xmath0 on the lie algebra @xmath1 of the unitary group @xmath2 , as expressed by the following formula : @xmath3 where `` @xmath4 '' denotes the big adjoint operator , and where the remaining symbols are defined in the table below .
@xmath5{|c||c|}\hline $ \mathbb{l}\left ( 2^{n}\right ) = { \displaystyle\bigotimes\limits_{1}^{n } } \mathbb{su}\left ( 2\right ) $ & local unitary group\\\hline $ \ell\left ( 2^{n}\right ) = \overset{n}{\underset{1}{\boxplus}}su\left ( 2\right ) $ & lie algebra of $ \mathbb{l}\left ( 2^{n}\right ) $ \\\hline $ \mathbb{u}\left ( 2^{n}\right ) $ & unitary group\\\hline $ u\left ( 2^{n}\right ) $ & lie algebra of $ \mathbb{u}\left ( 2^{n}\right ) $ \\\hline \end{tabular}\ ] ] we attack this problem by lifting the above big adjoint action to the * induced infinitesimal action * @xmath6 which , for a @xmath7 qubit density operator @xmath8 , is explicitly given by @xmath9{c}\omega\left ( v\right ) \left ( i\rho\right ) = { \displaystyle\sum\limits_{q_{1},q_{2}=0}^{3 } } \left ( a^{(1)}\cdot x_{\ast q_{1}q_{2}}\times\frac{\partial}{\partial x_{\ast q_{1}q_{2}}}+a^{(2)}\cdot x_{q_{1}\ast q_{2}}\times\frac{\partial}{\partial x_{q_{1}\ast q_{2}}}+a^{(3)}\cdot x_{q_{1}\ast q_{2}}\times\frac{\partial } { \partial x_{q_{1}q_{2}\ast}}\right ) \end{array } $ } \ ] ] where @xmath10 and @xmath11 are given by @xmath12{ccl}v & = & a^{(1)}\cdot\xi_{\ast00}+a^{(2)}\cdot\xi_{0\ast0}+a^{(3)}\cdot \xi_{00\ast}\\ & & \\ i\rho & = & { \displaystyle\sum\limits_{r_{1},r_{2},r_{3}=0}^{3 } } x_{r_{1}r_{2}r_{3}}\xi_{r_{1}r_{2}r_{3}}\end{array } \right.\ ] ] and where @xmath13 denotes the lie algebra of vector fields on @xmath14 . the induced infinitesimal action can then be used to quantify and to classify quantum entanglement through the construction of a complete set of quantum entanglement invariants . in the pages to follow , we make every effort to make the above sneak preview more transparent and understandable .
our goal is to present the underlying intuitions without getting lost in an obscure haze of technicalities .
however , presenting this topic is much like tiptoeing through a mine field .
one false move , and everything explodes into a dense jungle and clutter of technicalities . we leave it to the reader to determine how successful this endeavor is . finally , we close this introduction with a brief historical perspective . over the past twentieth century , the scientific community s view of quantum entanglement has dramatically changed . it continues to do so even today .
initially , quantum entanglement was viewed as an unnecessary and unwanted wart on quantum mechanics .
einstein , podolsky , and rosen@xcite tried to surgically remove it .
bell@xcite,@xcite showed that such surgery can not be performed without destroying the very life of physical reality . today
, quantum entanglement is viewed as a useful resource within quantum mechanics .
it is now viewed as a commodity to be utilized and traded , much as would be a commodity on the stock exchange .
quantum entanglement appears to be one of the physical phenomena at the central core of quantum computation .
many believe that it is quantum entanglement that somehow enables us to harness the vast parallelism of quantum superposition .
but what is quantum entanglement ? how do we measure , quantify , classify quantum entanglement ? when is the quantum entanglement of two quantum systems the same ? different ? when is the quantum entanglement of one quantum system greater than that of another ?
it is anticipated that answers to the above questions will have a profound impact on the development of quantum computation . finding answers to these questions
is challenging , intriguing , and indeed very habit forming .
our tangled tale of quantum entanglement begins with alice and bob s first encounter with quantum entanglement .
alice and bob , who happen to be good friends ( as attested , time and time again , by the open literature on quantum computation ) , meet one day . a discussion ensues .
the topic , of course , is quantum entanglement . fortunately or unfortunately , depending on how one looks at it , their discussion explodes into a heated argument . after a lengthy debate , they agree that the only way to resolve their conflict is to purchase the real mccoy , i.e. , a pair of entangled qubits .
so they rush to the nearest toys for aging children store to see what they can find . almost immediately upon entering the store
, they happen to spy , on one of the store shelves , an elaborately decorated box labelled : @xmath15{c}$\mathbb{q}$\textbf{.}$\mathbb{e}$\textbf{. , inc.}\\\hline\hline $ \text{\textbf{two entangled qubits}}$\\ $ \mathcal{q}_{ab}$\\ $ \text{consisting of qubits}$\\ $ \mathcal{q}_{a}\text { and } \mathcal{q}_{b}$\end{tabular } } \ ] ] on the back of the box is the content label , required by federal law , which reads : @xmath5{|ccccc|}\hline & & $ \begin{array } [ c]{c}\text{{\scriptsize u.s .
certified}}\\ \text{\textbf{contents}}\\ \quad\fbox{$\begin{array } [ c]{c}\text{\textsf{epr pair}}\end{array } $ } ^{(\ast)}\end{array } $ & & \\\hline\hline \textbf{q sys } & \multicolumn{1}{|c}{$\begin{array } [ c]{c}\text{\textbf{hilb.}}\\ \text{\textbf{sp.}}\end{array } $ } & \multicolumn{1}{|c}{\textbf{state } } & \multicolumn{1}{|c}{$\begin{array } [ c]{c}\text{\textbf{unitary}}\\ \text{\textbf{transf.}}\end{array } $ } & \multicolumn{1}{|c|}{$\begin{array } [ c]{c}\text{\textbf{state}}\\ \text{\textbf{space}}\end{array } $ } \\\hline\hline $ \mathcal{q}_{ab}$ & \multicolumn{1}{|c}{$\mathcal{h}_{ab}$ } & \multicolumn{1}{|c}{$\rho_{ab}=\overset{}{\underset{}{\left ( \begin{array } [ c]{rrrr}\frac{1}{2 } & 0 & 0 & -\frac{1}{2}\\ 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 0\\ -\frac{1}{2 } & 0 & 0 & \frac{1}{2}\end{array } \right ) } } $ } & \multicolumn{1}{|c}{$\mathbb{u}(2^{2})_{ab}$ } & \multicolumn{1}{|c|}{$u(2^{2})_{ab}$}\\\hline $ \mathcal{q}_{a}$ & \multicolumn{1}{|c}{$\mathcal{h}_{a}$ } & \multicolumn{1}{|c}{$\rho_{a}=\overset{}{\underset{}{\left ( \begin{array } [ c]{cc}\frac{1}{2 } & 0\\ 0 & \frac{1}{2}\end{array } \right ) } } $ } & \multicolumn{1}{|c}{$\mathbb{u}(2)_{a}$ } & \multicolumn{1}{|c|}{$u(2)_{a}$}\\\hline $ \mathcal{q}_{b}$ & \multicolumn{1}{|c}{$\mathcal{h}_{b}$ } & \multicolumn{1}{|c}{$\rho_{b}=\overset{}{\underset{}{\left ( \begin{array } [ c]{cc}\frac{1}{2 } & 0\\ 0 & \frac{1}{2}\end{array } \right ) } } $ } & \multicolumn{1}{|c}{$\mathbb{u}(2)_{b}$ } & \multicolumn{1}{|c|}{$u(2)_{b}$}\\\hline & & \begin{tabular } [ c]{l}{\tiny ( * ) caveat emptor : \ \ not legally responsible}\\ \quad{\tiny \qquad\qquad\qquad\qquad\ for the effects of decoherence.}\end{tabular } & & \\\hline \end{tabular}\ ] ] alice and bob hurriedly purchase the two qubit quantum system @xmath16 . outside the store ,
they rip open the box .
alice grabs the qubit labelled @xmath17 .
bob then takes the remaining qubit @xmath18 .
alice and bob then immediately depart for their separate destinations .
alice flies to queensland , australia to continue with her ph.d .
studies at the university of queensland .
she arrives just in time to attend the first class lecture on quantum mechanics .
bob , on the other hand , flies to vancouver , british columbia to continue with his ph.d .
studies at the university of british columbia .
he just barely arrives in time to hear the first lecture in a course on differential geometry and lie groups .
soon after her quantum mechanics lecture , alice begins to have second thoughts about their joint purchase of two entangled qubits .
she quickly reaches for her cellphone , calls bob , and nervously fires off in rapid succession three questions : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` did we get our money s worth of quantum entanglement ? ''
`` how much quantum entanglement did we actually purchase ? ''
`` are we the victims of a modern day quantum entanglement scam ? '' _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ after the phone conversation , bob is indeed deeply concerned . in desperation
, he calls the u.s .
quantum entanglement protection agency , which refers him to the u.s . national institute of quantum entanglement standards and technology ( ni@xmath19st ) in gaithersburg , maryland .
after a long conversation , a representative of ni@xmath19st agrees to send alice and bob , free of charge , the ni@xmath19st quantum entanglement standards kit .
on hanging up , the ni@xmath19st representative takes the ni@xmath19st standard entangled two qubit quantum system @xmath20 off the shelf , places @xmath21 together with a user s manual into a box marked `` alice . ''
he / she also places the remaining qubit @xmath22 together with a user s manual into a second box labeled `` bob , '' and then sends the two boxes by overnight mail to alice and bob respectively .
the very next day ( in different time zones , of course ) alice and bob each receive their respective packages , take out their respective qubits , and read the enclosed user s manuals .
the ni@xmath19st user s manual reads as follows : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath23.@xmath24 .
yardstick 1*. * an epr pair @xmath16 possess the same quantum entanglement as the ni@xmath19st standard epr pair @xmath20 if it is possible for you , alice and bob , to use your own local reversible operations ( either individually or collectively ) to transform @xmath16 and @xmath20 into one another .
if this is possible , then @xmath16 and @xmath20 are of the * same entanglement type * , written @xmath25 @xmath23.@xmath24 .
yardstick 2*. * an epr pair @xmath16 * possesses more quantum entanglement than * the ni@xmath19st standard epr pair @xmath26 if it is possible for you , alice and bob , ( either individually or collectively ) to apply your own reversible and irreversible operations to your respective qubits to transform @xmath16 into @xmath20 . in this case
, we write @xmath27 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ caveat**. * * quantum entanglement may be irrevocably lost if quantum entanglement yardstick 2 is applied . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in summary , the above story about alice and bob has raised the following questions : * * question : * what type of entanglement do alice and bob collectively possess ? * * question : * is the quantum entanglement of @xmath16 the same as the quantum entanglement of @xmath20 ? * * question : * is the quantum entanglement of @xmath16 greater than the quantum entanglement of @xmath20 ?
before we continue with our story of alice and bob , now is a good opportunity to restate the definition of quantum entanglement found in @xcite . readers not familiar with this definition or related concepts should refer to sections 5 and 7 of @xcite .
let @xmath28 , @xmath29 , @xmath30 , @xmath31 be quantum systems with underlying hilbert spaces @xmath32 , @xmath33 , @xmath30 , @xmath34 , respectively . and
let @xmath35 denote the global quantum system consisting of all the quantum systems @xmath28 , @xmath29 , @xmath30 , @xmath31 , where @xmath36 denotes the underlying hilbert space of @xmath35 . finally let the density operator @xmath8 on the hilbert space @xmath37 denote the state of the global quantum system @xmath35 .
then @xmath35 is said to be * entangled * with respect to the hilbert space decomposition @xmath38 if it can not be written in the form @xmath39 for some positive integer @xmath40 , where the @xmath41 s are positive real numbers such that @xmath42 and where each @xmath43 is a density operator on the hilbert space @xmath44 .
if @xmath8 is a pure state , then @xmath35 is * entangled * if @xmath8 can not be written in the form @xmath45 where @xmath46 is a density operator on the hilbert space @xmath44 .
although the story of alice and bob was told with two qubits , the same story could have been told instead with three people , alice , bob , cathy , and three qubits . or for that matter , it could have equally been told for @xmath47 people with @xmath47 qubits . from now on
, we will consider the more general story of @xmath47 people and @xmath47 qubits .
what alice , bob , cathy , et al were trying to understand can be stated most succinctly as the fundamental problem of quantum entanglement , namely : * fundamental problem of quantum entanglement * ( * fpqe * ) .
let @xmath8 and @xmath48 be density operators representing two different states of a quantum system @xmath35 .
is it possible to move @xmath35 from state @xmath8 to state @xmath48 by applying only * local moves * ?
@xmath9{c}\text{but what is meant by the phrase `` \textbf{local move '' } ? } \end{array } $ } \ ] ] we define the * standard local moves * as : the * standard local moves * are : * local unitary transformations of the form @xmath49 for example , for bipartite quantum systems , unitary transformations of the form @xmath50 , @xmath51 , @xmath52 * measurement of local observables of the form @xmath53 for example , for bipartite quantum systems , measurement of local observables of the form @xmath54 , @xmath55 , @xmath56 we also define the * extended local moves * as the * extended local moves * are : * extended local unitary transformations of the form @xmath57 where @xmath32 , @xmath58 , @xmath30 , @xmath34 , @xmath59 are distinct non - overhapping hilbert spaces * measurement of extended local observables of the form @xmath60 where @xmath32 , @xmath58 , @xmath30 , @xmath34 , @xmath59 are distinct non - overlapping hilbert spaces moves based on unitary transformation are called * reversible*. those based on measurement are called * irreversible . * the horodeckis@xcite , @xcite , @xcite , jonathan@xcite , @xcite , linden@xcite , @xcite , nielsen@xcite , @xcite , @xcite , @xcite , plenio@xcite , @xcite , popescu@xcite , @xcite , @xcite , @xcite have made some progress in understanding the fpqe in terms of all four of the above local moves . for the rest of the talk , we restrict our discussion to reversible standard local moves .
before continuing , it should be mentioned that physics and mathematics approach quantum mechanics from two slightly different but equivalent viewpoints . to avoid possible confusion , we describe below the minor terminology differences that arise from these two slightly different perspectives .
physics describes the state of a quantum system in terms of a traceless hermitian operator @xmath8 , called the density operator .
observables are hermitian operators @xmath61 .
quantum states change via unitary transformations @xmath62 according to the rubric @xmath63 on the other hand , mathematics describes the state of a quantum system in terms of a skew hermitian operator @xmath64 , also called the density operator .
observables are skew hermitian operators @xmath65 .
quantum dynamics are defined via the rule @xmath66 where @xmath62 is a unitary operator lying in the lie group of unitary transformations @xmath67 , and where @xmath4 denotes the big adjoint operator .
please note that both density operators @xmath64 and the observables @xmath65 lie in the lie algebra @xmath68 of the unitary group @xmath67 .
these minor , but nonetheless annoying differences are summarized in the table below .
@xmath69{||c||}\hline\hline\begin{tabular } [ c]{l||l}\textbf{physics}\hspace{1.5 in } & \hspace{1.5in}\textbf{math}\end{tabular } \\\hline\hline \fbox{\begin{tabular } [ c]{c}hilbert space $ \mathcal{h}$\\ $ dim(\mathcal{h})=n$\\ \multicolumn{1}{l}{\begin{tabular } [ c]{c}unitary group\\ lie group \end{tabular }
$ \mathbb{u}(n)$}\end{tabular } } \\\hline\hline\begin{tabular } [ c]{||l||l||}\hline\hline \quad\quad\quad\begin{tabular } [ c]{l}$ ] .
[ c]llobservables : & @xmath61 + density ops : & @xmath8 } @xmath70{l}$n\times n$ hermitian ops\\ \multicolumn{1}{c}{$a^{\dagger}=\overline{a}^{t}=a$}\end{tabular } } \end{tabular } \quad\quad\quad & \begin{tabular } [ c]{l}$ ] .
[ c]llobservables : & @xmath65 + density ops : & @xmath64 } @xmath71{c}$]nn@xmath72u(n)@xmath73 ( ia ) ^=^t =- ia@xmath74{l}dynamics via $ u\in\mathbb{u}(n)$\\ \\ \multicolumn{1}{c}{$\left| \psi\right\rangle \longmapsto u\left| \psi\right\rangle $ } \\
\multicolumn{1}{c}{$\rho\longmapsto u\rho u^{\dagger}$}\end{tabular } } & \multicolumn{1}{||c||}{\begin{tabular } [ c]{c}dynamics via $ u\in\mathbb{u}(n)$\\ \\
$ \left| \psi\right\rangle \longmapsto u\left| \psi\right\rangle $ \\ $ i\rho\longmapsto ad_{u}\left ( i\rho\right ) $ \\ where $ ad_{u}(i\rho)=u(i\rho)u^{-1}$\\ \multicolumn{1}{l}{is the big adjoint rep.}\end{tabular } } \\\hline\hline \end{tabular } \\\hline\hline \end{tabular } $ ] we will use the two different terminologies and conventions interchangeably .
which terminology we are using should be clear from context . from [ physical density operator]@xcite we know that an element @xmath64 of the lie algebra @xmath75 is a physical density operator if and only if @xmath8 is positive semi - definite and of trace 1 .
thus , the set @xmath76 of physical density operators is a convex subset of the lie algebra @xmath75 .
for the sake of clarity of exposition and for the purpose of avoiding minor technicalities , from on we consider only qubit quantum systems , i.e. , quantum systems consisting of qubits .
the reader , if he / she so wishes , should be able to easily rephrase the results of this paper to more general quantum systems .
moreover , from this point on , we limit the scope of this talk to the study of quantum entanglement from the perspective of the standard local unitary transformations , i.e. , from the perspective of standard reversible local moves as defined in section 5 of this paper . to emphasize this point
, we define the group of local unitary transformations @xmath77 as follows : the * group of local unitary transformations * @xmath77 is the subgroup of @xmath78 defined by @xmath79 where @xmath80 denotes the special unitary group .
henceforth , the phrase * `` local move '' * will mean an element of the group @xmath77 of * local unitary transformations * .. * * convention . * * from this point on , * * @xmath9{c}\text{\textbf{local moves } } = \mathbb{l}(2^{n } ) \end{array } $ } \ ] ] thus , for the rest of this paper we consider only the * restricted fundamental problem of quantum entanglement * ( * rfpqe * ) , which is defined as follows : * restricted fundamental problem of quantum entanglement * ( * rfpqe * ) .
@xmath64 _ _ and _ _ @xmath81 _ _ be density operators lying in the lie algebra _ _ @xmath82 . _
does there exist a local move _
@xmath62 , i.e. , a @xmath83 such that @xmath84 we will need the following definition : two elements @xmath64 and @xmath81 in @xmath85 are said to be * locally equivalent ( or , of the same entanglement type * ) , written @xmath86 provided there exists a @xmath83 such that @xmath87 the equivalence class @xmath88 _ { e}=\left\ { i\rho^{\prime}\mid i\rho\underset { loc}{\thicksim}i\rho^{\prime}\right\}\ ] ] is called an * entanglement class * ( or , an * orbit * of the big adjoint action of @xmath77 on the lie algebra @xmath82 ) .
finally , let @xmath89 denote the * set of entanglement classes*. + *
quantum entanglement classes .
* the entanglement classes of the lie algebra @xmath90 are just the * orbits * of the big adjoint action of @xmath77 on @xmath90 . two states are entangled in the same way if and only if they lie in the same entanglement class , i.e. , in the same orbit .
local unitary transformations can not entangle quantum systems with respect to the above tensor product decomposition . however , * global unitary transformations * ( i.e. , unitary transformations lying in @xmath91 ) are those unitary transformations which can and often do produce interactions which entangle quantum systems .
@xmath9{c}\text{\textbf{but what is quantum entanglement?}}\end{array } $ } \ ] ]
we are now in a position to state clearly the main objectives of this paper .
namely , in regard to the * restricted fundamental problem of quantum entanglement ( rfpqe ) * , our objectives are twofold : * given a density operator @xmath64 , devise a means of determining the dimension of its entanglement class @xmath92 _ { e}$ ] . we will accomplish this by determining the dimension of the tangent plane @xmath93 _ { e}$ ] to the manifold @xmath94 _ { e}$ ] at the point @xmath64 .
* given two states @xmath64 and @xmath81 , devise a means of determining whether they belong to the same or different entanglement class .
we will accomplish this by constructing a * complete set of quantum entanglement invariants * , i.e. , invariants that completely specify all the orbits ( i.e. , all the entanglement classes ) . in this sense , we have completely quantified quantum entanglement . in other words , find a finite set @xmath95 of real valued functions on @xmath90 which distinguish all entanglement classes , i.e. , @xmath96
this section is meant to play the role of a litmus test for the reader .
if the reader feels reasonably comfortable with the concepts listed below , then it is suggested that the reader proceed to the next section of this paper .
if not , it is strongly suggested that the reader read appendices a and b of this paper before proceeding to the next section .
let @xmath97 be a lie group , and let @xmath98 denote its lie algebra
. the reader should be familiar with the exponential map @xmath99 which for matrix lie groups is given by the power series @xmath100 the reader should be familiar with the lie bracket @xmath101 : \mathfrak{g}\times\mathfrak{g}\longrightarrow \mathfrak{g}\text { , } \ ] ] which for matrix lie groups is given by the commutator @xmath102 = ab - ba\ ] ] the lie algebra @xmath98 of the lie group @xmath97 can be viewed in each of the following mathematically equivalent ways : * as @xmath103 , i.e. , as the tangent space to the lie group @xmath97 at the identity @xmath104 . * as @xmath105 , i.e. , as the lie algebra of right invariant smooth vector fields on the lie group @xmath97 . * as @xmath106 ,
i.e. , as the lie algebra of all derivations ( i.e. , directional derivatives ) on the algebra @xmath107 of all smooth real valued functions on @xmath97 . in summary ,
@xmath108 if you feel comfortable with the above three litmus tests , then please proceed to the next section .
let @xmath109 denote the algebra of smooth ( @xmath110 ) real valued functions on the lie algebra @xmath14 , i.e. , @xmath111 a function @xmath112 is called a * ( quantum ) entanglement invariant * if @xmath113 is invariant under the big adjoint action of @xmath0 , i.e. , if @xmath114 for all @xmath115 , and for all @xmath64 in @xmath14 .
the collection of all ( quantum ) entanglement invariants forms an algebra , which we denote by @xmath116 a subset @xmath117 of @xmath118 is called a * complete set of entanglement invariants * if @xmath119 let @xmath120 be the subalgebra of @xmath109 of all functions @xmath112 which are polynomial functions , i.e. , of all functions @xmath113 for which @xmath121 is a polynomial function of the entries in @xmath122 . we define the * algebra of polynomial entanglement invariants as * @xmath123 @xmath124 is a finitely generated algebra . a minimal set of generators of @xmath125 is called a * basic set of entanglement invariants*. it is important to note that a basic set of entanglement invariants is not always a complete set of entanglement invariants . we will show that this is the case for entanglement invariants for two qubit quantum systems .
to understand and work with the big adjoint action [ section lie alg ] @xmath127 we will need to lift this action to the corresponding infinitesimal action of the lie algebra @xmath126 of @xmath128 .
the lie algebra @xmath126 will play a crucial role in our achieving objectives 1 and 2 as stated in the previous section .
the lie algebra @xmath126 is the ( real ) lie algebra given by the following kronecker sum @xmath129 where @xmath130 denotes the lie algebra of the special unitary group @xmath131 , and where the * kronecker sum * ` @xmath132 ' of two matrices ( or operators ) @xmath133 and @xmath134 is defined by @xmath135 with ` @xmath136 ' denoting the identity matrix ( or operator ) . a basis of the ( real ) lie algebra @xmath14 is given by @xmath137 where @xmath138 and where @xmath139{cc}0 & 1\\ 1 & 0 \end{array } \right ) \text { , } \sigma_{2}=\left ( \begin{array } [ c]{cr}0 & -i\\ i & 0 \end{array } \right ) \text { , } \sigma_{3}=\left ( \begin{array } [ c]{cr}1 & 0\\ 0 & -1 \end{array } \right ) \text { , } \ ] ] denote the pauli spin matrices , and where @xmath140{cc}1 & 0\\ 0 & 1 \end{array } \right)\ ] ] denotes the @xmath141 identity matrix .
it follows that a basis of the lie algebra @xmath142 as a subalgebra of @xmath14 is : @xmath143 for example , * @xmath144 is a basis of @xmath145 * @xmath146 is a basis of @xmath147 * @xmath148 is a basis of @xmath149 thus , we have the following proposition : the lie algebra @xmath126 of the lie group @xmath0 of local unitary transformations is of dimension @xmath150 , i.e. , @xmath151
we now show how quantum entanglement invariants can be found by lifting the big adjoint action to the lie algebra @xmath126 , where the problem becomes a linear one .
[ section act def ] the big adjoint action @xmath127 induces an * infinitesimal * * action * @xmath152 as follows .
let @xmath153 .
we define the vector field @xmath154 on @xmath14 by constructing a tangent vector @xmath155 for each @xmath156 .
let @xmath157 be the smooth curve in @xmath82 defined by @xmath158 then @xmath157 is a curve which passes through @xmath64 at time @xmath159 .
we define @xmath160 as the tangent vector to @xmath157 at @xmath159 .
+ * figure 3 . induced infinitesimal action . * but what is the meaning of the infinitesimal action @xmath152 that we have just defined ?
each @xmath161 is a direction in @xmath14 from @xmath64 that we can move without leaving the quantum entanglement class @xmath92 _ { e}$ ] .
movement in all directions not in @xmath162 will force us to immediately leave @xmath94 _ { e}$ ] .
+ * figure 4 . moving in a direction that stays within the entanglement class . *
as the reader might expect , the infinitesimal action @xmath163 can naturally be expressed in terms of the small adjoint operator @xmath164 .
in particular , we have : @xmath165 for all @xmath122 in the lie algebra @xmath126 [ ad ] let @xmath122 be an arbitrary element of the lie algebra @xmath166 , and let @xmath64 be an arbitrary element of the lie algebra @xmath14 . by definition , @xmath167 is the tangent vector at @xmath159 to the curve @xmath157 in @xmath82 given by @xmath158 hence , @xmath168{ccl}\omega\left ( v\right ) \left ( i\rho\right ) & = & \frac{d}{dt}\left .
ad_{\exp\left ( tv\right ) } \left ( i\rho\right ) \right| _ { t=0}\\ & & \\ & = & \frac{d}{dt}\left .
\exp\left ( ad_{tv}\right ) \left ( i\rho\right ) \right| _ { t=0}\\ & & \\ & = & \frac{d}{dt}\left .
\exp\left ( t\cdot ad_{v}\right ) \left ( i\rho\right ) \right| _ { t=0}\\ & & \\ & = & \frac{d}{dt}\left .
\left ( 1+t\cdot ad_{v}+o(t^{2})\right ) \left ( i\rho\right ) \right| _ { t=0}\\ & & \\ & = & ad_{v}\left ( i\rho\right ) \end{array}\ ] ] the above formula will prove to be useful when we actually calculate the entanglement invariants of some examples given in later sections .
as stated in the appendices , the lie algebra @xmath169 of all smooth vector fields on @xmath14 can be identified with the lie algebra of derivations @xmath170 .
the significance of the infinitesimal action @xmath171 is best expressed in terms of the following theorem : let [ theorem ad ] @xmath172 be a basis for the lie algebra @xmath126 .
then a smooth real valued function @xmath173 is an entanglement invariant if and only if it satisfies the following system of partial differential equations @xmath12{ccc}\omega\left
( v_{1}\right ) f & = & 0\\ & & \\ \omega\left ( v_{2}\right ) f & = & 0\\ & & \\ \vdots & \vdots & \vdots\\ & & \\ \omega\left ( v_{3n}\right ) f & = & 0 \end{array } \right.\ ] ] the intuition underlying the above theorem is that @xmath174 , @xmath175 , @xmath30 , @xmath176 are the linearly independent directions we can move without leaving the entanglement class we are currently in .
hence , if @xmath113 is an entanglement invariant , then its rate of change ( i.e. , its directional derivative ) in each of the directions @xmath177 , @xmath175 , @xmath30 , @xmath178 must be zero , and vice versa .
this theorem provides us with a means of determining a complete set of entanglement invariants .
all that we need to do is to solve the above system of partial differential equations .
we now show how the infinitesimal action @xmath152 can be used to achieve the two objectives listed in section 8 of this paper .
* given an arbitrary density operator @xmath64 , devise a means of determining the dimension of its entanglement class @xmath94 _ { e}$ ] .
objective 1 is achieved as follows : we begin by noting that @xmath179 is the same as the tangent space @xmath180 to @xmath82 at the point @xmath64 , and that @xmath181 is the same as the tangent space @xmath182 _ { e}\right ) $ ] to the entanglement class @xmath92 _ { e}$ ] at @xmath64 .
hence , the dimension of @xmath92 _ { e}$ ] is same as the dimension as its tangent space at @xmath64 , i.e. , @xmath183 _ { e}\right ) = dim\left ( t_{i\rho}\left ( \left [ i\rho\right ] _ { e}\right ) \right ) = dim\left ( \left . \operatorname{im}\left ( \omega\right ) \right| _ { i\rho}\right)\ ] ] the task of finding the dimension of the entanglement class @xmath94 _ { e}$ ] reduces to that of computing the dimension of the vector space @xmath181
. we will give examples of this dimension calculation in the next two sections .
we next use the infinitesimal action to achieve : * given two states @xmath64 and @xmath81 , devise a means of determining whether they belong to the same or different entanglement class . as follows :
we begin by noting that @xmath184 can be identified with the lie algebra @xmath185 of derivations on @xmath82 .
next we recall that @xmath186 consists of all directions in @xmath82 that we can move without leaving an entanglement class that we are in .
if @xmath187 is an entanglement invariant , then @xmath113 will not change if we move in any direction within @xmath188 . as a result
we have the following theorem : let @xmath189 be a vector space basis of the ( real ) lie algebra @xmath126 .
then @xmath190 for all @xmath191 , where @xmath192 is interpreted as a differential operator in @xmath193 .
in other words , the task of finding entanglement invariants reduces to that of solving a system of linear partial differential equations
. we will give examples of this calculation in the examples found in the next two sections of this paper .
we now make use of the methods developed in the previous section to study the entanglement classes associated with @xmath194 qubits .
this is a trivial but nonetheless instructive case .
as we shall see , there is no entanglement in this case .
but there are many entanglement classes ! for this example , the local unitary group @xmath195 is the same as the special unitary group @xmath196 .
the corresponding lie algebra @xmath197 is the same as the lie algebra @xmath198 .
each density operator @xmath64 lies in the lie algebra @xmath199 . as an immediate consequence of proposition [ ad ] of section [ section act def ] , the infinitesimal action @xmath200 is simply the small adjoint action , i.e. , @xmath201 for all @xmath202 .
we can now use the bases . ]
@xmath168{c}\left\ { \xi_{1}=-\frac{1}{2}\sigma_{1},\ \xi_{2}=-\frac{1}{2}\sigma_{2},\ \xi_{3}=-\frac{1}{2}\sigma_{3}\right\ } \\
\text { \ and \ } \\ \left\ { \xi_{0}=-\frac{1}{2}\sigma_{0},\ \xi_{1}=-\frac{1}{2}\sigma_{1},\ \xi_{2}=-\frac{1}{2}\sigma_{2},\ \xi_{3}=-\frac{1}{2}\sigma_{3}\right\ } \end{array}\ ] ] of the respective lie algebras @xmath197 and @xmath203 to find a more useful expression for @xmath154 .
each element @xmath202 can be uniquely expressed in the form @xmath204 where @xmath205 and @xmath206 .
thus , @xmath207 where @xmath208 moreover , each element @xmath209 can be uniquely written in terms of the basis of @xmath203 as @xmath210 where @xmath211 and @xmath212 . in terms
of the basis of @xmath203 , @xmath213{ll}\left ( \begin{array } [ c]{cc}0 & 0\\ 0 & l_{j}\end{array } \right ) = 0\oplus l_{j } & \text{if } j=1,2,3\\ & \\ \left ( \begin{array } [ c]{cc}0 & 0\\ 0 & 0 \end{array } \right ) = 0 & \text{if } j=0 \end{array } \right . \text { , } \ ] ] where @xmath214{rrr}0 & 0 & 0\\ 0 & 0 & -1\\ 0 & 1 & 0 \end{array } \right ) \text { , } l_{2}=\left ( \begin{array } [ c]{rrr}0 & 0 & 1\\ 0 & 0 & 0\\ -1 & 0 & 0 \end{array } \right ) \text { , } l_{3}=\left ( \begin{array } [ c]{rrr}0 & -1 & 0\\ 1 & 0 & 0\\ 0 & 0 & 0 \end{array } \right)\ ] ] is the basis{cclllll}ad_{\xi_{j}}\left ( \xi_{k}\right ) & = & ad_{-i\sigma_{j}/2}\left ( -i\sigma_{k}/2\right ) & = & \left [ -i\sigma_{j}/2,-i\sigma_{k}/2\right ] & = & -\frac{1}{4}\left [ \sigma_{j},\sigma_{k}\right ] \\ & = & -\frac{1}{2}i\epsilon_{jkp}\sigma_{p } & = & \epsilon_{jkp}\xi_{p } & & \end{array}\ ] ] where @xmath215 .
] of the lie algebra @xmath216 of the special orthogonal group @xmath217 given in appendix b. let @xmath218 denote the basis of @xmath219 induced by the chart @xmath168{ccc}\mathbf{u}\left ( 2\right ) & \overset{\pi}{\longrightarrow } & \mathbb{r}^{4}\\ & & \\ i\rho=\sum_{j=0}^{3}x_{j}\xi_{j } & \longmapsto & \left ( x_{0},x_{1},x_{2},x_{3}\right ) = \left ( x_{0},x\right ) \end{array } \text { .}\ ] ] in other words , for each @xmath191 , @xmath220 denotes the vector field on @xmath203 defined at each point @xmath64 as the tangent vector to the curve @xmath221 at @xmath159 .
then , @xmath168{ccl}\omega\left ( v\right ) \left ( i\rho\right ) & = & \left ( x_{0},x\right ) \cdot\left ( 0\oplus a\cdot l\right ) \cdot\left ( \begin{array } [ c]{c}\partial/\partial x_{0}\\ \partial/\partial x_{1}\\ \partial/\partial x_{2}\\ \partial/\partial x_{3}\end{array } \right ) \\ & & \\ & = & x\cdot\left ( a\cdot l\right ) \cdot\bigtriangledown\\ & & \\ & = & a\cdot x\times\bigtriangledown\text { , } \end{array } \text { , } \ ] ] where ` @xmath222 ' denotes the vector cross product , and where @xmath223{c}\partial/\partial x_{1}\\ \partial/\partial x_{2}\\ \partial/\partial x_{3}\end{array } \right ) \text { .}\ ] ] we can now achieve objective 1 .
* objective 1 .
* _ given an arbitrary density operator _
@xmath64 _ in _
@xmath203 , _ find the dimension of an arbitrary entanglement class _
@xmath92 _ { e}$ ] .
from the above discussion , it follows that the image @xmath186 of the infinitesimal action @xmath163 is spanned by the three vector fields @xmath12{ccc}\omega\left ( \xi_{1}\right ) & = & x_{3}\frac{\partial}{\partial x_{2}}-x_{2}\frac{\partial}{\partial x_{3}}\\ & & \\ \omega\left ( \xi_{2}\right ) & = & x_{1}\frac{\partial}{\partial x_{3}}-x_{3}\frac{\partial}{\partial x_{1}}\\ & & \\ \omega\left ( \xi_{3}\right ) & = & x_{2}\frac{\partial}{\partial x_{1}}-x_{1}\frac{\partial}{\partial x_{2}}\end{array } \right .
\text { , } \ ] ] defined on @xmath224 .
in particular , the tangent space @xmath182 _ { e}\right ) = \left .
\left ( \operatorname{im}\omega\right ) \right| _ { i\rho}$ ] of the entanglement class @xmath92 _ { e}$ ] at the point @xmath64 is spanned by @xmath225 as can be easily verified by the reader , the above three vectors span a two dimensional space if @xmath226 and a zero dimensional vector space if @xmath227 .
since @xmath228 is the tangent space @xmath182 _ { e}\right ) $ ] of @xmath92 _ { e}$ ] at the point @xmath64 , and since the dimension of @xmath92 _ { e}$ ] is the same as the dimension of its tangent space @xmath182 _ { e}\right ) $ ] at @xmath64 , it follows that the dimension of the entanglement class @xmath92 _ { e}$ ] is given by : @xmath229 _ { e}=\left\ { \begin{array } [ c]{ccc}2 & \text{if } & \left| x\right| \neq0\\ & & \\ 0 & \text{if } & \left| x\right| \neq0 \end{array } \right.\ ] ] we are now ready to achieve objective 2 : * objective 2 . *
_ given two states _
@xmath64 _ and _ @xmath81 , _ devise a means of determining whether they belong to the same or different entanglement class_. we achieve this objective by determining a complete set of entanglement invariants qubits , the complete set of entanglement invariants consists of only one invariant . ] for one qubit quantum systems , i.e. , by determining a set of entanglement invariants @xmath230 such that @xmath231 .
we begin by recalling that the lie algebra @xmath232 of vector fields on @xmath233 can be identified with the lie algebra @xmath234 of all derivations on the smooth real valued functions on @xmath235 .
thus , the elements of @xmath236 can be viewed as directional derivatives , directional derivatives in those directions in which we can move and still remain in the same entanglement class . from theorem [ theorem ad ]
, it immediately follows that a real valued function @xmath237 is an entanglement invariant if and only it is a solution of the system of partial differential equations ( pdes ) : @xmath12{c}\omega\left ( \xi_{1}\right ) f=0\\ \\
\omega\left ( \xi_{2}\right ) f=0\\ \\ \omega\left ( \xi_{3}\right ) f=0 \end{array } \right.\ ] ] since from above we know that @xmath238 , we can write the above system of pdes more explicitly as : @xmath12{c}x_{3}\frac{\partial f}{\partial x_{2}}-x_{2}\frac{\partial f}{\partial x_{3}}=0\\ \\ x_{1}\frac{\partial f}{\partial x_{3}}-x_{3}\frac{\partial f}{\partial x_{1}}=0\\ \\
x_{2}\frac{\partial f}{\partial x_{1}}-x_{1}\frac{\partial f}{\partial x_{2}}=0 \end{array } \right .
\text { , } \ ] ] where , as before , @xmath239 . from theorem [ theorem ad ] , we know that a complete set of quantum entanglement invariants for one qubit systems is the same as a complete functionally independent set of solutions of the above system of pdes . thus , solving the above system of pdes by standard methods found in the theory of differential equations , we find that @xmath240 is a complete set of entanglement invariants .
a functionally equivalent complete set of entanglement invariants is @xmath241 which is also a basic set of entanglement invariants .
fortunately , in this simplest case , a complete set of entanglement invariants and a basic set of entanglement invariants are one and the same
. this will not be the case for quantum systems of more than one qubit . as a result of the previous calculation
, we have a complete set of entanglement invariants , namely @xmath242 we have completely classified all the entanglement classes for 1 qubit quantum systems . for in this case , @xmath88 _ { e}=\left [ i\rho^{\prime}\right ] \longleftrightarrow f(i\rho)=f(i\rho^{\prime})\text { .}\ ] ] as a consequence of this result , the induced foliation of the space @xmath243 of all physical density operators lying in in the lie algebra @xmath199 can be visualized in terms of the 3-ball or radius 1 in @xmath244 , called the * bloch * `` * * sphere**. '' recall from remark on page that @xmath245 is a convex subset of the of the lie algebra @xmath246 . in this special case of @xmath194 qubit ,
it is a straight forward exercise to show that @xmath247 thus , the convex subset @xmath243 of @xmath199 of all physical density operators @xmath64 in @xmath199 can naturally be identified with the 3-ball of radius one via the one - to - one correspondence @xmath248 as illustrated in figure 5 .
+ * figure 5 . the bloch `` sphere '' * it follows that each entanglement class @xmath92 _ { e}$ ] is simply a sphere of radius @xmath249 .
the sphere of radius one is the entanglement class of all pure ensembles .
all other spheres represent entanglement classes of mixed ensembles . the `` sphere '' of radius @xmath250 ( i.e. , the origin ) represents the entanglement class of the maximally mixed ensemble .
iand so we can conclude that the space of entanglement classes lying in the space formed by identifying the elements of the convex set @xmath243 via the action of the local transformation group @xmath195 , namely @xmath251 is simply a closed line segment . in terms of this picture
, it is easy to visualize the tangent space @xmath252 _ { e}\right ) $ ] to @xmath92 _ { e}$ ] at @xmath64 .
moreover , it is easy to visualize the normal bundle of @xmath94 _ { e}$ ] .
for the normal vector field is simply @xmath253 _ { e}}\ ] ] unfortunately , for quantum systems of more than one qubit , such a visualization is by no means as easy .
as one might expect , the entanglement of two qubit quantum systems is much more complex than that of one qubit quantum systems .
in fact , with each additional qubit , the entanglement becomes exponentially more complex than before .
perhaps this is a strong hint as to where the power of quantum computation is coming from ? for this example , the local unitary group @xmath255 is the lie group @xmath256 .
the corresponding lie algebra @xmath257 is kronecker sum of two matrices ( operators ) @xmath133 and @xmath134 is defined as @xmath258 where @xmath136 denotes the identity matrix ( operator ) . ]
@xmath259 .
each density operator @xmath64 lies in the lie algebra @xmath260 . as an immediate consequence of proposition [ ad ] given in section [ section act def ] ,
the infinitesimal action @xmath261 is simply the small adjoint action , i.e. , @xmath201 for all @xmath262 .
we can now use the bases . ]
@xmath168{c}\left\ { \xi_{10},\xi_{20},\xi_{30},\xi_{01},\xi_{02},\xi_{03},\right\ } \\ \text { \ and \ } \\ \left\ { \xi_{ij}\mid i , j=0,1,2,3\right\ } \end{array}\ ] ] of the respective lie algebras @xmath263 and @xmath260 to find a more useful expression for @xmath154 , where @xmath264 .
each element @xmath262 can be uniquely expressed in the form @xmath265 where @xmath266 and @xmath267 lie in @xmath244 , where @xmath268 , and where @xmath269 is the @xmath270 identity matrix .
thus , @xmath168{ccl}\omega\left ( v\right ) & = & \omega\left ( \sum_{j=1}^{3}\left ( a_{j}\xi_{j0}+b_{j}\xi_{0j}\right ) \right ) \\ & & \\ & = & \omega\left ( a\cdot\xi\boxplus b\cdot\xi\right ) \\ & & \\ & = & ad_{a\cdot\xi\boxplus b\cdot\xi}\\ & & \\ & = & ad_{\left ( a\cdot\xi\right ) \otimes i_{4}}+ad_{i_{4}\otimes\left ( b\cdot\xi\right ) } \\ & & \\ & = & i_{4}\otimes\left ( a\cdot ad_{\xi}\right ) + \left ( b\cdot ad_{\xi } \right ) \otimes i_{4}\end{array}\ ] ] where @xmath208 but as in example 1 , @xmath213{ll}\left ( \begin{array } [ c]{cc}0 & 0\\ 0 & l_{j}\end{array } \right ) = 0\oplus l_{j } & \text{if } j=1,2,3\\ & \\ \left ( \begin{array } [ c]{cc}0 & 0\\ 0 & 0 \end{array } \right ) = 0 & \text{if } j=0 \end{array } \right . \text { , } \ ] ] where @xmath214{rrr}0 & 0 & 0\\ 0 & 0 & -1\\ 0 & 1 & 0 \end{array } \right ) \text { , } l_{2}=\left ( \begin{array } [ c]{rrr}0 & 0 & 1\\ 0 & 0 & 0\\ -1 & 0 & 0 \end{array } \right ) \text { , } l_{3}=\left ( \begin{array } [ c]{rrr}0 & -1 & 0\\ 1 & 0 & 0\\ 0 & 0 & 0 \end{array } \right)\ ] ] is the basis of the lie algebra @xmath216 of the special orthogonal group @xmath217 given in appendix b on page .
let @xmath271 denote the basis of @xmath272 induced by the chart @xmath168{ccc}\mathbf{u}\left ( 2^{2}\right ) & \overset{\pi}{\longrightarrow } & \mathbb{r}^{16}\\ & & \\ i\rho=\sum_{i , j=0}^{3}x_{ij}\xi_{ij } & \longmapsto & \left ( x_{00},x_{0\ast } , x_{10},x_{1\ast},x_{20},x_{2\ast},x_{30},x_{3\ast}\right ) \end{array}\ ] ] where @xmath168{l}\left ( x_{00},x_{0\ast},x_{10},x_{1\ast},x_{20},x_{2\ast},x_{30},x_{3\ast } \right ) \\ \\
= \left ( x_{00},\quad x_{01},x_{02},x_{03},\quad x_{10},\quad x_{11},x_{12},x_{13},\quad x_{20},\quad x_{21},x_{22},x_{23},\quad x_{30},\quad x_{31},x_{32},x_{33}\right ) \end{array}\ ] ] in other words , for each pair @xmath273 , @xmath274 denotes the vector field on @xmath260 defined at each point @xmath64 as the tangent vector to the curve @xmath275 at @xmath159 . in terms of the above chart ,
@xmath276 can be written as @xmath277 \cdot\left ( \begin{array } [ c]{c}\partial/\partial x_{00}\\ \partial/\partial x_{0\ast}\\ \partial/\partial x_{10}\\ \partial/\partial x_{1\ast}\\ \partial/\partial x_{20}\\ \partial/\partial x_{2\ast}\\ \partial/\partial x_{30}\\ \partial/\partial x_{3\ast}\end{array } \right ) \text { , } \ ] ] which simplifies to @xmath278 where ` @xmath222 ' denotes the vector cross product .
we can now achieve objective 1 .
* objective 1 .
* _ given an arbitrary density operator _
@xmath64 _ in _
@xmath203 , _ find the dimension of an arbitrary entanglement class _
_ { e}$ ] . from the above discussion
, it follows that the image @xmath186 of the infinitesimal action @xmath163 is spanned by the six vector fields @xmath168{c}\left\ { \begin{array } [ c]{c}\omega\left ( \xi_{01}\right ) = \sum\limits_{q=0}^{3}\left ( x_{q2}\frac{\partial}{\partial x_{q3}}-x_{q3}\frac{\partial}{\partial x_{q2}}\right ) \\ \\ \omega\left ( \xi_{02}\right ) = \sum\limits_{q=0}^{3}\left ( x_{q3}\frac{\partial}{\partial x_{q1}}-x_{q1}\frac{\partial}{\partial x_{q3}}\right ) \\
\\ \omega\left ( \xi_{03}\right ) = \sum\limits_{q=0}^{3}\left ( x_{q1}\frac{\partial}{\partial x_{q2}}-x_{q2}\frac{\partial}{\partial x_{q1}}\right ) \end{array } \right . \\ \\ \left\ { \begin{array } [ c]{c}\omega\left ( \xi_{10}\right ) = \sum\limits_{q=0}^{3}\left ( x_{2q}\frac{\partial}{\partial x_{3q}}-x_{3q}\frac{\partial}{\partial x_{2q}}\right ) \\ \\ \omega\left ( \xi_{20}\right ) = \sum\limits_{q=0}^{3}\left ( x_{3q}\frac{\partial}{\partial x_{1q}}-x_{1q}\frac{\partial}{\partial x_{3q}}\right ) \\ \\ \omega\left ( \xi_{30}\right ) = \sum\limits_{q=0}^{3}\left ( x_{1q}\frac{\partial}{\partial x_{2q}}-x_{2q}\frac{\partial}{\partial x_{1q}}\right )
\end{array } \right .
\end{array}\ ] ] in particular , the tangent space @xmath182 _ { e}\right ) = \left . \left ( \operatorname{im}\omega\right ) \right| _ { i\rho}$ ] to the entanglement class @xmath92 _ { e}$ ] at the point @xmath64 is spanned by @xmath279 we leave it as an exercise for the reader to verify that the above six vector fields are linearly independent almost every where .
thus , it follows that almost all entanglement classes are of dimension six . however , there are notable exceptions .
consider the bell state , @xmath280 .
the corresponding density operator @xmath64 is @xmath281{rrrr}1 & 0 & 0 & -1\\ 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 0\\ -1 & 0 & 0 & 1 \end{array } \right ) = \left ( -\frac{1}{2}\right ) \xi_{00}+\left ( \frac{1}{2}\right ) \xi_{11}+\left ( -\frac{1}{2}\right ) \xi_{22}+\left ( -\frac{1}{2}\right ) \xi_{33}\text { , } \ ] ] where @xmath282 hence , @xmath283{rcl}-\frac{1}{2 } & \text{if } & j = k=0,2,3\\ & & \\ \frac{1}{2 } & \text{if } & j = k=1\\ & & \\ 0 & \text{if } & j\neq k \end{array } \right.\ ] ] thus , in this case @xmath284 is spanned by @xmath168{c}\left\ { \begin{array } [ c]{lll}\left .
\omega\left ( \xi_{01}\right ) \right| _ { i\rho } & = & \frac{1}{2}\left ( \frac{\partial}{\partial x_{23}}-\frac{\partial}{\partial x_{32}}\right ) \\ & & \\ \left .
\omega\left ( \xi_{02}\right ) \right| _ { i\rho } & = & \frac{1}{2}\left ( \frac{\partial}{\partial x_{31}}+\frac{\partial}{\partial x_{13}}\right ) \\ & & \\ \left .
\omega\left ( \xi_{03}\right ) \right| _ { i\rho } & = & \frac{1}{2}\left ( -\frac{\partial}{\partial x_{12}}-\frac{\partial}{\partial x_{21}}\right ) \end{array } \right .
\\ \\ \left\ { \begin{array } [ c]{lll}\left . \omega\left ( \xi_{10}\right ) \right| _ { i\rho } & = & \frac{1}{2}\left ( \frac{\partial}{\partial x_{32}}-\frac{\partial}{\partial x_{23}}\right ) \\ & & \\ \left . \omega\left ( \xi_{20}\right ) \right| _ { i\rho } & = & \frac{1}{2}\left ( \frac{\partial}{\partial x_{13}}+\frac{\partial}{\partial x_{31}}\right ) \\ & & \\ \left . \omega\left ( \xi_{30}\right ) \right| _ { i\rho } & = & \frac{1}{2}\left ( -\frac{\partial}{\partial x_{21}}-\frac{\partial}{\partial x_{12}}\right ) \end{array } \right . \end{array}\ ] ] hence , @xmath285 _ { e}=dim\left [ \left
. \left ( \operatorname{im}\omega\right ) \right| _ { i\rho_{bell}}\right ] = 3\ ] ] this only confirms the conventional wisdom that the entanglement class of the bell states is truly exceptional .
we are now ready for objective 2 : * objective 2 .
* _ given two states _
@xmath64 _ and _ @xmath81 , _ devise a means of determining whether they belong to the same or to different entanglement classes_. the complete functionally independent set of solutions to the above system of pdes ( hence , a complete set of entanglement invariants ) was found by linden and popescu in@xcite .
these invariants are as described below . for further details please refer to @xcite .
let @xmath64 be an arbitrary element of the lie algebra @xmath286 .
then in terms of the earlier described chart @xmath287 , @xmath288 where @xmath289 denotes the basis of @xmath286 described earlier . we will change our notation slightly .
let @xmath290 denote the @xmath291 matrix @xmath292 and let @xmath293 and @xmath294 denote the vectors @xmath12{ccc}x_{0\ast } & = & \left ( x_{01},x_{02},x_{02}\right ) \\ & & \\ x_{\ast0 } & = & \left ( x_{10},x_{20},x_{30}\right ) \end{array } \right.\ ] ] finally , let @xmath295 denote the matrix @xmath296 where the superscript ` @xmath297 ' denotes the transpose .
then the nine algebraically independent polynomial functions listed in the table @xmath5{||c||c||c||}\hline\hline
$ tr\left ( z\right ) $ & $ \overset{}{\underset{}{tr\left ( z^{2}\right ) } } $ & $ \det\left ( x_{\ast\ast}\right ) $ \\\hline\hline $ x_{0\ast}x_{0\ast}^{t}$ & $ \overset{}{\underset{}{x_{0\ast}zx_{0\ast}^{t}}}$ & $ x_{0\ast}z^{2}x_{0\ast}^{t}$\\\hline\hline $ x_{0\ast}x_{\ast\ast}x_{\ast0}^{t}$ & $ \overset{}{\underset{}{x_{0\ast } zx_{\ast\ast}x_{\ast0}^{t}}}$ & $ x_{0\ast}z^{2}x_{\ast\ast}x_{\ast0}^{t}$\\\hline\hline \end{tabular}\ ] ] form a basic set of entanglement invariants . _ but the above nine entanglement invariants do not form a complete set of entanglement invariants _ ! a tenth polynomial function @xmath298 is needed to form a complete system of entanglement invariants .
although this tenth entanglement invariant is algebraically dependent on the above nine entanglement invariants , it is still needed to determine the sign of the components of @xmath64 .
for @xmath47 qubits ( @xmath299 ) , the same methods lead to the following formula for the infinitesimal action @xmath9{c}\omega\left ( v\right ) \left ( i\rho\right ) = { \displaystyle\sum\limits_{q_{1},q_{2}\cdots q_{n-1}=0}^{3 } } { \displaystyle\sum\limits_{k=1}^{n } } a^{(k)}\cdot x_{q_{1}q_{2}\cdots q_{k-1}\ast q_{k+1}\cdots q_{n-1}}\times\frac{\partial}{\partial x_{q_{1}q_{2}\cdots q_{k-1}\ast q_{k+1}\cdots q_{n-1}}}\end{array } $ } \ ] ] where @xmath153 and @xmath300 are given by @xmath12{ccl}v & = & { \displaystyle\sum\limits_{k=1}^{n } } a^{(k)}\cdot\xi\underset{\ast\text { in } k\text{-th position}}{_{\underbrace { 00\cdots0\ast0\cdots0}}}\\ & & \\ i\rho & = & { \displaystyle\sum\limits_{r_{1},r_{2},\cdots , r_{n}=0}^{3 } } x_{r_{1}r_{2}\cdots r_{n}}\xi_{r_{1}r_{2}\cdots r_{n}}\end{array } \right.\ ] ] we will leave the solution to the corresponding system of pdes to future papers .
there is much more that could be said about quantum entanglement .
this paper presents only a small part of the big picture .
but hopefully this paper will provide the reader with some insight into this rapidly growing research field .
since this paper was written , research in quantum entanglement has literally had an explosive expansion , and even now continues to do so .
we refer the reader to the references at the end of this paper , which represent only a few of the many papers in this rapidly expanding field .
a topological space @xmath301 is an @xmath302**-dimensional manifold * * if it is locally homeomorphic to @xmath303 , i.e. , if there exists an open cover @xmath304 of @xmath301 such that for each @xmath305 , there is associated a homeomorphism @xmath306 which maps @xmath307 onto an open subset of @xmath303 .
we call @xmath308 a * chart on * @xmath301 , and @xmath309 an * atlas * on @xmath301 .
an atlas is said to be * smooth * ( @xmath110 ) , if whenever @xmath310 is defined , is a smooth ( @xmath110 ) map of @xmath311 into @xmath312 . a * smooth * ( @xmath110 ) *
manifold * is a topological manifold with a smooth atlas .
+ * figure 7 .
an atlas is smooth if every * @xmath313 * is smooth when defined . *
let @xmath314 and @xmath315 be smooth manifolds .
then a map @xmath316 is said to be smooth if for every @xmath317 there exist charts @xmath318 of @xmath314 and @xmath319 of @xmath315 containing @xmath320 and @xmath321 respectively such that @xmath322 is smooth .
let x be an element of a smooth manifold @xmath314 , and let @xmath323 and @xmath324 be smooth curves in @xmath314 which pass through @xmath320 , i.e. , such that there exists @xmath325 for which @xmath326 then @xmath327 and @xmath328 are said to be * tangentially equivalent * at @xmath320 , written @xmath329 if they are tangent at the point @xmath320 , i.e. , if there is a chart @xmath318 on @xmath314 containing @xmath320 such that @xmath330 it can easily be shown that the relation @xmath331 is independent of the chart selected .
a * * tangent vector * * @xmath332 ( also written simply as @xmath122 ) to @xmath314 at @xmath320 is a tangential equivalence class at @xmath320 .
the tangent space of @xmath333 at @xmath320 , denoted by @xmath334 , is the set of tangent vectors to @xmath314 at @xmath320 .
@xmath335 can be shown to be an @xmath47-dimensional vector space .
let @xmath336 and let @xmath287 be the map @xmath337 if @xmath338 is a chart on @xmath314 , then @xmath339 can be shown to be a chart on @xmath340 . in this way
, @xmath340 becomes a smooth manifold and @xmath287 becomes a smooth map .
@xmath340 together with the map @xmath287 is called the * tangent bundle * of @xmath314 . a * vector field * @xmath122 on a smooth manifold @xmath301 is a smooth map @xmath341
let @xmath342 be the set of all vector fields on the smooth manifold @xmath301 .
this is easily seen to be a vector space where , for example , the sum @xmath343 of two vector fields is defined by @xmath344 for all @xmath345 .
we will now consider the charts of the tangent bundle @xmath340 in a more explicit way .
let @xmath306 be a chart on the smooth manifold @xmath301 , and let @xmath346 be an arbitrary point in @xmath347 .
thus , @xmath348 for each @xmath191 ( @xmath349 ) consider the smooth curve @xmath350 in @xmath347 which passes through the point @xmath346 at time @xmath159 .
then for each such @xmath191 , let [ vec basis ] @xmath351 denote the tangent vector to the curve @xmath352 at @xmath346 .
it can be shown that @xmath353 is a vector space basis of the tangent space @xmath354 . moreover , since this construction is respect to an arbitrary point @xmath346 in @xmath307 , it can be shown that we have actually constructed for each @xmath191 a smooth vector field @xmath355 in fact , it can be shown that @xmath356 is a basis of @xmath357 , and hence a local basis of @xmath358 .
we can now express each chart @xmath359 explicitly as : @xmath360 where @xmath361 s on the left denote functions of @xmath317 , and where @xmath362 s on the right denote functions of @xmath363 .
let @xmath314 and @xmath315 be smooth manifolds , and let @xmath364 be a smooth map , and let@xmath365be an arbitrary point of @xmath314 .
we define a vector space morphism @xmath366 as follows : for each @xmath367 , there is a representative smooth curve @xmath368 in @xmath314 which passes through the point @xmath346 and which has @xmath122 as its tangent vector at the point @xmath346 .
it follows that @xmath369 is a smooth curve in @xmath315 passing through the point @xmath370 .
we define @xmath371 as the tangent vector to @xmath369 at the point @xmath370 .
it is then a simple exercise to show that @xmath372 is a vector space morphism .
since @xmath346 was an arbitrary point of @xmath314 , this leads to the definition of a smooth map @xmath373 , called the differential of @xmath113 , such that the following diagram is commutative : @xmath168{cccc}tm & \overset{df}{\longrightarrow } & tn & \\ \downarrow & & \downarrow & \\ m & \overset{f}{\longrightarrow } & n & \text{.}\end{array}\ ] ] in local coordinates , @xmath374 maps the tangent vector @xmath375 to the tangent vector @xmath376 thus , the matrix expression of the linear transformation @xmath374 is just the jacobian matrix @xmath377 let @xmath314 be a smooth manifold , and let @xmath378 be a smooth vector field on @xmath314 .
a curve @xmath379 in @xmath314 is said to be an * integral curve * of @xmath122 if @xmath380 is the tangent vector to @xmath379 for each @xmath381 for which @xmath379 is defined . in terms of local coordinates ,
an integral curve @xmath379 of a smooth vector field @xmath382 is a solution to the system of ordinary differential equations @xmath383 since @xmath122 is smooth , its coefficients @xmath384 are smooth functions .
consequently , it follows from the standard existence and uniqueness theorems for systems of ordinary differential equations that there exists a unique solution for each set of initial conditions .
thus , for each @xmath320 in @xmath314 , there exists a unique maximal integral curve @xmath385 passing through @xmath320 at time @xmath159 , and call @xmath386 the * flow * generated by the vector field @xmath122 .
we call @xmath122 the * infinitesimal generator * of the flow .
it can be easily shown that @xmath387 hence , we are justified in adopting the following suggestive notation : @xmath388 for the flow @xmath385 . in terms of our new notation , the properties of the flow can be expressed as * @xmath389 * @xmath390 * @xmath391 .
we now show how vector fields can be viewed as partial differential operators .
a * derivation * @xmath392 on an algebra @xmath393 is a map @xmath394 such that * ( linearity ) @xmath395 * ( leibnitz rule ) @xmath396 a * lie algebra * @xmath397 is a vector space together with a binary operation @xmath101 : \mathbb{a}\times\mathbb{a\rightarrow a}\text { , } \ ] ] called a * lie bracket * for @xmath397 , such that * ( bilinearity ) @xmath168{ccc}\left [ \lambda_{1}a_{1}+\lambda_{2}a_{2},\ b\right ] & = & \lambda_{1 } \left [ a_{1},b\right ] + \lambda_{2}\left [ a_{2},b\right ] \\ & & \\ \left [ a,\ \lambda_{1}b_{1}+\lambda_{2}b_{2}\right ] & = & \lambda_{1 } \left [ a , b_{1}\right ] + \lambda_{2}\left [ a , b_{2}\right ] \end{array}\ ] ] * ( skew - symmetry ) @xmath398 = -\left [ b , a\right]\ ] ] * ( jacobi identity ) @xmath399 \right ] + \left [ c,\left [ a , b\right ] \right ] + \left [ b,\left [ c , a\right ] \right ] = 0\ ] ] the set of derivations @xmath400 on an algebra @xmath393 is a lie algebra with lie bracket given by : @xmath401 = d_{1}\circ d_{2}-d_{2}\circ d_{1}\ ] ] let @xmath402 denote the * algebra of real valued functions * on the smooth manifold @xmath314 .
then , it follows that @xmath403 is a lie algebra .
we will now show how to identify the elements of @xmath404 with derivations in @xmath403 , and thereby show that @xmath404 is more than a vector space .
it is actually a lie algebra .
each smooth vector field @xmath122 on @xmath314 can be thought of as a directional derivative in the direction @xmath122 as follows : let @xmath378 and let @xmath405 .
define @xmath406 as : @xmath407 thus , we have : @xmath404 is a lie algebra of derivations on the algebra @xmath402 .
it is enlightening , to view the above in terms of local coordinates . from this perspective , @xmath408 thus , if we use the chain rule and the fact that @xmath409 we have @xmath410 hence , @xmath411 acts as a first order partial differential operator , thereby justifying the notation .
so viewing @xmath122 as a first order partial differential operator , we can write @xmath412 and , in particular , @xmath413 where @xmath414 now denotes ( locally ) @xmath415 evaluated at @xmath320 .
* @xmath424 the * real general linear group * of all automorphisms of the vector space @xmath421 .
this can be identified with the group of all nonsingular @xmath425 matrices over the reals . * @xmath426 the * real orthogonal group * is the group of all automorphisms which preserve the inner product @xmath427 .
this can be identified with the group of orthogonal matrices , i.e. , matrices @xmath133 of the form @xmath428 where the superscript @xmath429 denotes the matrix transpose . * @xmath430 the * real special linear group * is the group of all real @xmath425 matrices of determinant @xmath431 @xmath432 is the group of all rigid motions in hyperbolic @xmath47-space . * @xmath433 the * special orthogonal group * is the group of all orthogonal real @xmath434 matrices of determinant @xmath435 . * * * * this group can be identified with the group of all rotations in @xmath436 about a fixed point such as the origin .
* @xmath439 the * complex general linear group * of all automorphisms of the vector space @xmath437 .
this can be identified with the group of all nonsingular @xmath425 matrices over the complexes . *
@xmath440 the * complex special linear group * is the group of all complex @xmath425 matrices of determinant @xmath431 * @xmath441 the * unitary group * is the group of all @xmath434 unitary matrices over the complex numbers @xmath438 , i.e. , all @xmath434 complex matrices @xmath133 such that @xmath442 where @xmath443 denotes the conjugate transpose . *
@xmath444 the special unitary group is the group of all unitary matrices of determinant 1 .
let @xmath97 be a lie group . for each element @xmath445
, we define the * right multiplication map * , written @xmath446 , as @xmath447 the map @xmath446 is an autodiffeomorphism of @xmath97 . we let @xmath448 denote the corresponding differential of this diffeomorphism
. let @xmath105 denote the set of right invariant vector fields on @xmath97 .
then @xmath451 as a subset of the lie algebra @xmath452 inherits the structure of a lie algebra .
we call @xmath105 the * lie algebra * of the lie group @xmath97 .
let @xmath104 denote the identity element of the lie group @xmath97 . since a right invariant vector field @xmath453
is completely determined by its restriction to the tangent space @xmath103 via @xmath454 we can , and do , identify the lie algebra @xmath451 with the tangent space @xmath103 , i.e. , @xmath455 the tangent bundle @xmath456 of a lie group @xmath97 is trivial . for it can be shown that @xmath456 is bundle isomorphic to @xmath457 .
however , there is some additional and useful structure induced on the lie algebra @xmath458 by the lie group structure of @xmath97 , i.e. , the exponential map .
it can be shown that the exponential map @xmath463 is a local diffeomorphism .
it also follows that , for each @xmath464 , @xmath465 is a one parameter subgroup of @xmath417 .
all one parameter subgroups are of this form .
[ [ example - the - lie - algebra - mathbfuleft - nright - of - the - unitary - group - mathbbuleft - nright- . ] ] example : the lie algebra @xmath75 of the unitary group @xmath67 .
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the lie algebra @xmath75 is the tangent space @xmath468 to @xmath67 at the @xmath469 identity matrix @xmath104 .
hence , @xmath75 consists of all tangent vectors @xmath470 of all curves @xmath471 in @xmath67 which pass through @xmath104 at @xmath159 , i.e. , which satisfy @xmath472 .
let @xmath314 be an arbitrary skew @xmath466 hermitian matrix .
then @xmath477 is a curve in @xmath67 which passes through @xmath104 at @xmath159 for which @xmath478 .
hence , @xmath75 is the lie algebra of all @xmath466 skew hermitian matrices over @xmath438 .
let @xmath139{cc}0 & 1\\ 1 & 0 \end{array } \right ) \text { , } \sigma_{2}=\left ( \begin{array } [ c]{cc}0 & -i\\ i & 0 \end{array } \right ) \text { , } \sigma_{3}=\left ( \begin{array } [ c]{cc}1 & 0\\ 0 & -1 \end{array } \right)\ ] ] denote the pauli spin matrices , and let @xmath140{cc}1 & 0\\ 0 & 1 \end{array } \right)\ ] ] denote the @xmath141 identity matrix .
then the following is a basis of the lie algebra @xmath14 @xmath479 where @xmath480 please note that , although @xmath75 is a lie algebra of complex matrices , it is nonetheless a real lie algebra .
thus , the above basis @xmath481 of @xmath82 is a basis of @xmath14 over the reals @xmath422 .
but the matrices in @xmath14 are still matrices of complex numbers !
[ [ example - the - lie - algebra - mathbfsuleft - nright - of - the - special - unitary - group - mathbbsuleft - nright- . ] ] example : the lie algebra @xmath482 of the special unitary group @xmath483 .
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the lie algebra @xmath482 for the special unitary group is the same as the lie algebra of all @xmath466 traceless skew hermitian matrices , i.e. , of all @xmath466 skew hermitian matrices @xmath314 such that @xmath484 .
a basis of the lie algebra @xmath485 is @xmath486 [ [ example - the - lie - algebra - mathbfsoleft-3right - of - the - special - unitary - group - mathbbsoleft-3right- . ]
] example : the lie algebra @xmath216 of the special unitary group @xmath217 .
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ finally , we should mention that the lie algebra @xmath487 of the special orthogonal group @xmath217 is the lie algebra of all @xmath291 skew symmetric matrices over the reals @xmath422 .
the following three matrices form a basis for [ so3 basis ] @xmath488 @xmath214{rrr}0 & 0 & 0\\ 0 & 0 & -1\\ 0 & 1 & 0 \end{array } \right ) \text { , } l_{2}=\left ( \begin{array } [ c]{rrr}0 & 0 & 1\\ 0 & 0 & 0\\ -1 & 0 & 0 \end{array } \right ) \text { , } l_{3}=\left ( \begin{array } [ c]{rrr}0 & -1 & 0\\ 1 & 0 & 0\\ 0 & 0 & 0 \end{array } \right ) \text { \ .}\ ] ] let @xmath314 be a smooth manifold , and let @xmath97 be a lie group acting on @xmath314 .
then the action @xmath495 induces an * infinitesimal action * @xmath496 where @xmath497 is the tangent vector to the curve @xmath498 in @xmath314 at @xmath320 , i.e. , @xmath499 for each element @xmath445 , consider the inner automorphism : @xmath500 and let @xmath501 denote the corresponding differential
. we can now define the * big adjoint representation * @xmath502 by @xmath503 where @xmath104 denotes the identity of @xmath417 , and where @xmath504 denotes the group of automorphisms of the lie algebra @xmath98 .
we can now in turn define the * little adjoint representation * @xmath505 of the lie algebra @xmath98 by @xmath506 \text { , } \ ] ] where @xmath507 $ ] denotes the lie bracket , and where @xmath508 denotes the ring of endomorphisms of the lie algebra @xmath98 .
as the story goes , @xmath509 is actually the lie algebra of the lie group @xmath504 , and we have the following commutative diagram @xmath168{ccc}\mathfrak{g } & \overset{ad}{\longrightarrow } & end\left ( \mathfrak{g}\right ) \\
\exp\downarrow\qquad & & \quad\downarrow\exp\\ g & \overset{ad}{\longrightarrow } & aut(\mathfrak{g } ) \end{array}\ ] ] which relates the big and little adjoints .
little adjoint @xmath164 is actually the differential @xmath510 restricted to the identity @xmath104 of the big adjoint @xmath4 .
let @xmath97 be the special unitary group @xmath80 .
let @xmath511 denote its lie algebra .
then @xmath512 is the special orthogonal group @xmath513 and @xmath509 is the lie algebra @xmath514 of @xmath513 .
thus , we have the familiar commutative diagram @xmath168{ccc}su(2 ) & \overset{ad}{\longrightarrow } & so(3)\\ \exp\downarrow\qquad & & \quad\downarrow\exp\\ \mathbb{su}(2 ) & \overset{ad}{\longrightarrow } & \mathbb{so}(3 ) \end{array}\ ] ] used in quantum mechanics and in quantum computation .
bennett , charles h. , david p. divincenzo , christopher a. fuchs , tal mor , eric rains , peter w. shor , john a. smolin , and william k. wootters , * quantum nonlocality without entanglement , quant - ph/9804053 .
* cerf , nicholas j. and chris adami , `` * * quantum information theory of entanglement and measurement * * , '' in * proceedings of physics and computation , physcomp96 * , edited by j. leao t. toffoli , pp 65 - 71 .
see also quant - ph/9605039 .
einstein , a. , b. podosky , and n. rosen , * can quantum mechanical description of physical reality be considered complete ? * , phys .
rev . * 47 * , 777 ( 1935 ) ; d. bohm , `` quantum theory , '' prentice - hall , englewood cliffs , nj ( 1951 ) .
lomonaco , samuel j. , jr . , * a rosetta stone for quantum mechanics with an introduction to quantum computation : lecture notes for the ams short course on quantum computation , washington , dc , january 2000 * , to appear in the ams psamp series .
( quant - ph/0007045 ) lomonaco , samuel j. , jr . , * the shor / simon algorithm from the perspective of group representation theory * , to appear in `` * * quantum computation and information * * , '' ams contemporary mathematics series ( 2001 ) . | these lecture notes give an overview from the perspective of lie group theory of some of the recent advances in the rapidly expanding research area of quantum entanglement .
this paper is a written version of the last of eight one hour lectures given in the american mathematical society ( ams ) short course on quantum computation held in conjunction with the annual meeting of the ams in washington , dc , usa in january 2000 .
more information about the ams short course can be found at the website : http://www.csee.umbc.edu/lomonaco/ams/announce.html |
the measurement of linear polarization of radiation from active galactic nuclei ( agn ) presents additional information about the structure of radiation sources and absorbing and scattering regions ( the presence of jets , polar outflows , accretion discs near the agn centres , the existence of dust toroidal clouds and other clouds that scatter light ) .
the polarimetric results are presented in many papers ( see , for example , impey et al .
1991 , goodrich & miller 1994 , smith et al .
2002 , 2004 , 2005 ; marin et al .
2012a , b ; marin & goosmann 2013 ; marin 2014 ; afanasiev et al .
2011 , 2014 ) . in these papers
there are detailed references to the observational data and various models to explain these data .
the important parameter in any explanation is the angle of inclination @xmath1 of the line of sight @xmath2 relative to the normal @xmath3 to equatorial plane of the system related to agn .
note that the @xmath1 - value is determined by various methods which very frequently produce different results ( see the recent review of marin 2016 ) .
the inferences of the innermost structures of agns have been obtained indirectly by rowan - robinson ( 1977 ) suggested that agns are surrounded by dusty tori .
he also gives the suggestion that seyfert-2 galaxies are seen close to edge - on so that the active nuclei are obscured by the tori .
support for this picture came from statistical investigation ( see keel 1980 ) which demonstrated that seyfert-1 galaxies are preferentially seen face - on .
the further importance of the orientation effects was considered in lawrence & elvis ( 1982 ) and de zotti & gaskell ( 1985 ) . since then
, the dusty - torus model has become the standard unified model .
when the optical polarization in continuum in agns was detected ( see dibai & shakhovskoi 1966 and walker 1966 ) , it was initially accepted that this was the optical polarization of synchrotron radiation , since it had high intrinsic polarization .
later ( angel et al .
1976 ) the polarization of balmer lines was discovered , which was similar to the continuum .
after that it was clear that the scattering is responsible for both types of polarization .
a key point here was the lack of variability in polarization , which argues against a non - thermal synchrotron origin ( see becklin et al .
1973 ; knacke & capps ( 1974 ) for ir radiation and antonucci & miller 1985 ; bailey et al . 1988 ; young et al .
( 1995 ) for spectropolarimetry ) .
according to optical classification ( see antonucci 1984 , 1993 ; antonucci & miller 1985 ; lawrence 1987 , 1993 ) , there are two types of seyfert agns which also differ from each other by luminosities and spectra .
while spectra of seyfert-2 galaxies have a single set of relatively narrow emission lines , the spectra of seyfert-1 galaxies have additional much broader components of hydrogen , helium and other lines ( see di serego alighieri et al .
1994 , lal et al .
one of the main reasons for the difference between agn-1 and agn-2 is the different angle of observation .
seyfert-1 corresponds to observation closer to the normal @xmath3 to the equatorial plane .
it is the interesting feature that seyfert-1 agns have mostly relatively small linear polarization @xmath4 and the inclination angle @xmath5 . on the contrary , for seyfert-2
the polarization is larger ( @xmath6 ) and the inclination angle @xmath7 is higher ( see marin 2014 ) .
it is also important that the positional angle for seyfert-1 corresponds to wave electric field oscillations * e * nearly parallel to the symmetry axis ( more accurately parallel to the axis of the radio wave emission ) , and for seyfert-2 the oscillations are parallel to the equatorial plane . in the review of marin ( 2014 ) various aspects
are discussed : the different methods to estimate the inclination angle @xmath1 and a number of models to explain the observed polarimetric data .
particularly , the 3-component model of agn is discussed in detail . according to this model
there are 3 components - a conical polar outflow of scattering particles with the half - opening angle @xmath8 , a vast dust torus in the equatorial plane and an optically thin ( @xmath9 ) equatorial flow ( see also smith et al .
2002 ; goosmann & gaskell 2007 ) .
according to this model , the scattering of non - polarized radiation from the agn centre on electrons in a cone produces the wave electric field oscillations parallel to the equatorial plane .
on the contrary , the scattering on electrons in the optically thin equatorial flow produces the wave electric field oscillations in the the line of sight - the axis of symmetry plane , that corresponds to seyfert-1 observations . the degree of this polarization is proportional to @xmath10 and is not large .
the second reason why seyfert-1 agns have modest polarization degrees is the competition between these two scattering mechanisms with the mutually perpendicular position angles .
it is clear from a physical point of view that in axially symmetric models there can arise positional angles either perpendicular to the symmetry axis or parallel to it .
for this reason the competition can not explain the numerous cases of seyfert agns with intermediate position angles . such cases may be explained in models without axial symmetry ( with the existence of separate scattering clouds , ( see goosmann & matt 2011 and marin et al .
2015 ) , or in axially symmetric models by taking into account the faraday rotation effect ( see , for example , silantev 2002 ; buliga et al .
2014 ) . of course
, both reasons may exist independently .
the presence of magnetic field in jets and accretion discs is apparent in many objects ( see chakrabarti et al . 1994 ; pariev & colgate 2007 ) .
there are many types of jets in agns ( see , for example , blandford 2003 ) both relativistic and non - relativistic ones . up to now various models of jets
appear in literature ( for example , levinson & globus 2016 ) .
the faraday rotation of linearly polarized radiation in jets is studied in detail ( gomez et al .
the core of jet sometimes is assumed to be optically thick ( blandford & konigle 1979 ; the review of blandford 1990 ) .
it is known that the base of jets ( hot corona ) is the source of soft x - ray radiation ( see haardt & maraschi 1993 ; haardt & matt 1993 ) .
far above this region the temperature diminishes and the optical radiation appears .
most frequently the adiabatic expansion of the medium is considered ( blandford 2000 , 2003 ) .
recall that in adiabatic expansion the relation @xmath11 takes place ( landau & lifshitz 1980 ) . here
@xmath12 is the temperature in volume @xmath13 , and @xmath14 is temperature in volume @xmath15 .
the parameter @xmath16 is the ratio of thermal capacities at constant pressure and volume , respectively . for monoatomic gas @xmath17 and for diatomic gas @xmath18 .
clear that expanding jet ( mostly due to particles acceleration with the distance from the centre which results in the increase of volume @xmath15 ) gives rise to the inequality @xmath19 and consequently @xmath20 .
it is interesting that there are papers ( for example , see wang et al .
2005 ) where the existence of thermal matter in relativistic jets is discussed .
it should be noted that there are numerous seyfert agns with intermediate values of polarization degrees and position angles ( see , for example , table 3 in antonucci 1984 ; table 4 in smith et al .
note , that the aperture effects may cause distortions in both the measurement of polarization degree and the position angles of radiation in agns . in many papers ( see , for example ,
lawrence 1991 ; simpson 2005 ) there exist the modifications of a standard model ( the dependence of molecular and dust torus on luminosity of the central source , the particular distributions of dust grains in torus , and so on ) .
usually one considers optically thick accretion discs ( see shakura & sunyaev 1973 ; pariev & colgate 2007 ) .
it appears that both the optically thin and thick accretion discs really occur in agns .
our objective is to discuss the mechanism for the origin of negative polarization by reflection of non - polarized radiation from a jet in optically thick accretion discs .
the integral radiation from the surface of optically thick part of a jet has small degree of polarization ( see next chapter ) .
the thermal radiation from an optically thin jet is non - polarized .
for this reason we consider only reflection of non - polarized radiation in an accretion disc . for the case of optically thin disc
the problem is simple and corresponds to taking into account the single scattered radiation .
when the incident radiation goes from the central nucleus of a disc the polarization degree of scattered radiation is described by formula ( 26 ) .
this formula corresponds to `` negative '' polarization .
the case of optically thick disc is more difficult for consideration .
so , the optically thick accretion disc is illuminated by unpolarized radiation from a jet and we calculate the reflected radiation from the total surface of an accretion disc going into the line of sight @xmath2 . we demonstrate that this radiation has `` negative '' linear polarization for all inclination angles @xmath1 between the line of sight @xmath2 and the normal @xmath3 to the accretion disc plane . besides of reflected radiation , a telescope detects the radiation directly going from the source .
first we calculate this direct radiation from the optically thick jet ( see ch.2 ) . for calculating radiation fluxes @xmath21 and
stokes parameter @xmath22 ( here @xmath23 is the angle between the line of sight @xmath2 and direction of jet z , see fig.1 ) we use the well known chandrasekhar s ( 1950 ) results for the intensity @xmath24 and the stokes parameter @xmath25 in the milne problem for atmospheres without absorption @xmath26 . here
@xmath27 and @xmath28 are cross- sections of absorption and scattering , respectively . for calculating fluxes @xmath29 and @xmath30 of radiation scattered in the optically thick accretion disc the results of the paper by silantev & gnedin ( 2008 ) are used .
the calculations demonstrate that direct radiation from an optically thick jet for @xmath31 has the polarization with electric field oscillations perpendicular to the axis of a jet , and for @xmath32 oscillations occur in the plane ( @xmath33 ) , where @xmath2 is direction to an observer , and @xmath34-axis directed along the jet . but this negative polarization is small @xmath35 .
thus , polarization from the optically thick region of a jet for seyfert-1 is positive , and for seyfert-2 is negative , just opposite to observations ( see , for example , smith et al .
2002 ) .
the radiation reflected after multiple scatterings from an optically thick accretion disc is negative for all angles of emission . for inclination angles @xmath36
this polarization is greater than that for the radiation escaping from an optically thick jet ( the milne problem , see ch.3 ) .
in conclusion we underline that we do not present a new model to explain the seyfert -1 and 2 agns features in full .
our objective is to call attention to a new mechanism of negative polarization formation in seyfert-1 agns in optically thick accretion discs .
furthermore , we call attention to the explanation of intermediate position angles in these agns using the faraday rotation effect . we hope that our calculations are applicable to all agns , not only to seyfert agns .
let us introduce the necessary notations : the elemental radiating area on the surface of a jet is proportional to the element of the azimuthal angle @xmath37 : @xmath38 .
the axis z is directed along the jet , the direction to the telescope @xmath2 lies in the plane ( xz ) .
the normal to the radiating elemental area is @xmath3 .
the angle @xmath23 is the angle between the z - axis and the vector @xmath2 .
the vector @xmath2 is a linear combination of the coordinate unit vectors @xmath39 and @xmath40 ( see fig.1 ) : @xmath41 the normal @xmath3 can be represented as @xmath42 the cosine of angle between @xmath2 and @xmath3 is as follows : @xmath43 chandrasekhar s ( 1950 ) intensity @xmath24 [ erg/@xmath44s sr ] and the stokes parameter @xmath25 of the outgoing radiation can be approximated by the expressions : @xmath45 @xmath46 @xmath47 here @xmath48 , where @xmath49 is the total flux of outgoing radiation ; the value @xmath50 is the first moment of the function in brackets . in our approximation @xmath51 .
the definition of various moments is given in eq.(21 ) .
the ratio @xmath52 determines the degree of linear polarization .
the minus sign in @xmath25 implies that the wave electric vector @xmath53 oscillates perpendicular to the plane @xmath54 .
the stokes parameter @xmath55 .
note that expressions ( 4 ) describe the milne problem in conservative atmosphere with @xmath56 .
let us also remind the reader that the milne problem describes the situation when the sources of non - polarized radiation are located far from the surface of an atmosphere .
a telescope detects the radiation fluxes from areas with the azimuthal angles @xmath57 in the range @xmath58 : @xmath59 @xmath60 @xmath61 here @xmath62 is the distance to the observed object .
expressions for @xmath63 and @xmath64 take into account the transition of stokes parameters into the telescope s coordinate system ( xyz ) . using the expressions ( 4 ) we obtain : @xmath65 @xmath66 @xmath67 @xmath68 the value 0.5@xmath69(erg / s ) is the luminosity of observed part of a jet in the direction @xmath2 . integrating @xmath69 over all directions of the angle @xmath70 ( from @xmath71 up to @xmath72 ) and taking into account that @xmath73 , we obtain the total luminosity @xmath74 of radiating area along all directions : @xmath75 using this expression , the formula for @xmath69 can be written in the form : @xmath76 note that the positive sign in @xmath63 corresponds to wave electric field oscillations in the plane @xmath77 ( see fig.1 )
this is the case of negative polarization in seyfert-1 agns .
table 1 and fig.2 present the degrees of polarization @xmath78 in percents for a number of angles @xmath23 .
the second line of this table presents the polarization in absorbing atmosphere with @xmath79 .
here we approximated the values @xmath24 and @xmath25 using the paper by silantev ( 1980 ) .
note that the calculation of angular distribution and polarization degree is connected with exact solution of characteristic equation and the solution of the system of nonlinear equation for h - functions .
this is difficult problem , especially for large values of absorption parameter @xmath80 .
some particular results are presented some later , before eq.(24 ) .
[ % ] for absorption degrees q = 0 and 0.1 ] .polarization of radiation from an optically thick jet @xmath81 [ % ] for absorption degrees q = 0 and 0.1 [ cols="<,<,<,<,<,<,<,<",options="header " , ]
the calculations in our paper demonstrate that the negative polarization of radiation from seyfert-1 agns , corresponding to wave electric field oscillations in the plane @xmath54 ( the line of sight @xmath2 - the normal to accretion disc @xmath3 ) , can be explained not only in the model with an optically thin accretion disc ( or equatorial flow ) but also when the accretion disc is optically thick .
our calculations show that the reflection of non - polarized radiation from the central jet in an optically thick accretion disc always produces negative polarization .
the polarization of radiation emerging from the optically thick accretion disc ( the milne problem ) has positional angle perpendicular to plane @xmath54 .
the total polarization of seyfer-1 agn depends on the relative values of luminosities of radiation from the jet and the accretion disc .
we have shown that the observed polarization of seyfert -1 agn is valid even though the luminosity of disc is considerably higher than that of the jet .
allowing for faraday depolarization in the accretion disc decreases the perpendicular polarization of radiation in milne problem .
the cases with intermediate positional angles ( at an angle to the plane ( @xmath82 ) ) may be explained by presence of faraday rotations in the accretion disc . for axially symmetric model
this occurs if there is a magnetic component perpendicular to the disc .
we present the calculation of polarization degrees for a number of seyfert-1 agns .
the results of our calculations are close to the observed values of polarization .
this research was supported by the basic research program p-7 of presidium of ras and the program of the department of physical sciences of ras no.2 . 99 afanasiev v. l. , borisov n. v. , gnedin yu .
n. et al . , 2011 , astron.lett . , 37 , 302 afanasiev v. l. , borisov n. v. , gnedin yu .
n. et al . , 2014 , astron.rep . , 58 , 725 angel j.r.p . ,
stockman h.s . , woolf n.j . ,
beaver e.a . ,
martin p.g . , 1976 ,
apj , 206 , l5 antonucci r.r.j . , 1984 , apj , 278 , 499 antonucci r r.j . ,
miller j.s . , 1985 , apj , 297 , 621 antonucci r. r. j. , 1993 , ann .
rev . of a&a , 278 , 473 bailey j. a. , axon d. j. , hough j. h. , ward m. j. , mclean i. , heathcote s. r. , 1988 , mnras , 234 , 899 becklin e. e. , neugebauer g. , willner s. p. , matthews k. , 1973 , apj , 186l , 69b berriman g. , 1989 , apj , 345 , 713 blandford r.d .
, konigl a. , 1979 , apj , 232 , 34 blandford r. d. , 1990 , lecture notes of saas - fee courses , geneva observatory , xii , 280pp . ,
springer verlag , heidelberg blandford r.d .
, 2000 , rspta , 358 , 811 blandford r. d. , 2003 , active galactic nuclei : from central engine to host galaxy , aps conference series , vol.290 , p.267 ( eds . s. collin , f. combes , i. shlosman ) brenneman l. , 2013 , acta polytechn.suppl . , 53 , 652 buliga s.d . , gnedin yu .
n. , natsvlishvili t. m. , piotrovich m. yu . , silantev n. a. , 2014 , astron .
40 , 185 chakrabarti s. k. , rosner r. , vainshtein s. i. , 1994 , nature , 368 , 434 chandrasekhar s. , 1950 , radiative transfer , clarendon press , oxford chelouche d. , 2013 , apj , 772 , 9 decarli r. , labita m. , treves a. , falomo r. , 2008 , mnras , 387 , 1237 de zotti g. & gaskell c. m. , 1985 , a & a , 147 , 1 dibai e.a . & shakhovskoi n. m.,1966 , astronomicheskij tsirkular , 375 , 1 di serego alighieri s.,cimatti a. , fosbury r.a .
1994 , apj .
431 , 123 dolginov a. z. , gnedin yu . n. , silantev n. a. , 1995 , propagation and polarization of radiation in cosmic media , gordon & breach , new york dutan i. , biermann p.l . , 2008 , in proceedings of the relativistic astrophysics and cosmology , einstein legacy , eso astrophysics symposia , p. 431 dutan i. , 2010 , arxiv:1001.5434 gomez j.l .
, lobanov a. p. , bruni g. et al .
, 2016 , apj .
, 817 , 96 goodrich r. w. , miller j. s. , 1994 , apj , 434 , 82 goosmann r. w. & gaskell c.m . , 2007 , a & a , 465 , 129 greene j.e . ,
hood c. e. , barth a. j. et al .
, 2010 , apj , 723 , 409 grinin v. p. & domke h. , 1971 , astrophysics , 7 , 120 haardt f. & maraschi l. , 1993 , apj .
, 413 , 507 haardt f. & matt g. , 1993 , mnras , 261 , 346 harko t. , kavasc z. , lobo f. s. n. , 2009 , phys . rev .
d , 79 , 064001 keel w. c. , 1980 , aj , 85 , 198 ho l. c. , darling j. , greene j. e. , 2008 , apss , 177 , 103 impey c. d. , lawrence c. r. , tapia s.,1991 , apj , 375 , 46 knacke r. f. , capps r. w. , 1974 , apj , 192l , 19k kollatschny w. , zetzl m. , dietrich m. , 2006 , a & a , 454 , 459 kollatschny w. & zetzl m. , 2013 , a & a , 558 , 26 krolik j. h. , 2007 , arxiv:0709.1489 krolik j. h. , hawley j. f. , hirose s. , 2007 , rev .
aa , 27 , 1 labita m. , treves a. , falomo r. , ustenghi m. , 2006 , mnras . , 373 , 551 lal d.v . ,
shastri p. , gabuzda d. c. , 2011 , apj , 731 , 68 landau l.d .
& lifshitz e.m.,1980 , statistical physics , part 1 , ( pergamon press ) landau l.d . & lifshitz e.m.,1984 , electrodynamics of continuous media ( pergamon press ) lawrence a. & elvis m. , 1982 , apj , 256 , 410 lawrence a. , 1987 , pasp , 99 , 309 lawrence a. , 1991 , mnras , 252 , 586 lawrence a. , 1993 , the nearest active galaxies , 3 levinson a. , globus n. , 2016 , mnras , 458 , 2269 marin f. , goosmann r. w. et al . , 2012a , a & a , 548 , 121 marin f. , goosmann r. w. et al . , 2012b , mnras , 426 , l101 , 121 marin f. , goosmann r. w. , 2013 , mnras , 436 , 2522 marin f. , 2014 , mnras , 441 , 551 marin f. , goosmann r. w. , gaskell c. m. , 2015,a & a , 577 , 66 marin f. , 2016 , arxiv : 1605.02904v ( sent to mnras ) martin p. g. , thompson i. b. , maza i. , angel j. r. p. , 1983
, apj , 266 , 470 mclure r.j . &
jarvis m. j. 2002 , mnras , 337 , 109 novikov i. d. & thorne k. s. , in black holes , eds . c. de witt -morette and c. de witt ( gordon and breach , n.y . 1973 ) pariev v. i. & colgate s. a. , 2007 , apj , 658 , 114 patrick a. r. , reeves j. n. , porquet d. et al . , 2012 , mnras , 426 , 2522 piotrovich m. yu . , gnedin yu .
n. , silantev n. a. et al .
2015 , mnras , 454 , 1157 poindexter s. , morgan n. , kochanek c. , 2008 , apj , 673 , 34 rowan - robinson m. , 1977 , apj , 213 , 635 shakura n. i. , sunyaev r. a. , 1973 . , a & a , 24 , 337 silantev n. a. , 1980 , sva , 24 , 341 silantev n. a. , 2002 , a & a , 383 , 326 silantev n. a. , 2007 , astron.rep . , 84 , 67 silantev n. a. , gnedin yu . n. , 2008 , a & a , 481 , 217 silantev n. a. , piotrovich m. yu . , gnedin yu . n. , natsvlishvili t. m. , 2009 , a & a , 507 , 171 silantev n. a. , piotrovich m. yu . , gnedin yu .
n. , natsvlishvili t. m. , buliga s. d. , 2013 , astrophys.bull .
68 , 14 simpson c. , 2005 , mnras , 360 , 565 smith j. e. , young s. , robinson a. , et al . , 2002 , mnras , 335 , 773 smith j. e. , robinson a. , alexander d. m. et al . , 2004 , mnras , 350 , 140 smith j. e. , robinson a. , young s. et al . , 2005 , mnras , 359 , 846 walker m. f. , 1966 , apj , 71 , 184 walton d. j. , nardini e. , fabian a. c. , gallo l. c. , reis r. c. , 2013 , mnras , 428 , 2901 wang j .- m . , staubert r. , courvoisier t. j .- l . , 2005 , gbha conf . , 399 ( arxiv : astro - ph/0407123 v1 ) young s. , hough j. h. , axon d. j. , 1995 , mnras , 272 , 513 zhang t .- z . & wu x .- b . , 2002 , chin.astron.astrophys . , 2 , 487 | in most of seyfert-1 active galactic nucei ( agn ) the optical linear continuum polarization degree is usually small ( less than @xmath0 ) and the polarization position angle is nearly parallel to the agn radio - axis .
however , there are many types-1 agns with unexplained intermediate values for both positional angles and polarization degrees .
our explanation of polarization degree and positional angle of seyfert-1 agns focuses on the reflection of non - polarized radiation from sub - parsec jets in optically thick accretion discs .
the presence of a magnetic field surrounding the scattering media will induce faraday rotation of the polarization plane that may explain the intermediate values of positional angles if there is a magnetic field component normal to the accretion disc .
the faraday rotation depolarization effect in disc diminishes the competition between polarization of the reflected radiation with the parallel component of polarization and the perpendicular polarization from internal radiation of disc ( the milne problem ) in favor of polarization of reflected radiation .
this effect allows us to explain the observed polarization of seyfert-1 agn radiation even though the jet optical luminosity is much lower than the luminosity of disc .
we present the calculation of polarization degrees for a number of seyfert-1 agns .
[ firstpage ] accretion discs , polarization , faraday rotation , active galaxies . |
one of the methods of extensive air shower ( eas ) detection is recording fluorescence light emitted by nitrogen molecules in the air along the shower path . for very high energies of the primary particle , enough fluorescence light is produced so that the shower can be recorded from a distance of many kilometers by an appropriate optical detector system @xcite .
as the amount of fluorescence light is closely correlated to the ionization energy deposit in air , it provides a calorimetric measure of the primary energy .
the field of view of a fluorescence detector ( fd ) telescope is divided into many pixels .
for example , in case of the pierre auger observatory ( pao ) @xcite each pixel views @xmath0 of the sky and records the received light in 100 ns time intervals .
a shower passing through the telescope field of view triggers some pixels , which form together a `` shower track '' .
the lateral width of this track depends on shower geometry but can well be larger than the pixel size . for a precise energy determination one needs to collect the available signal as completely as possible , i.e. from all detector pixels which receive light from the shower . on the other hand , adding signals from many pixels implies adding the background noise as well .
therefore it is important to include in the analysis only a small number of pixels which contain the true shower signal . in this paper , monte
carlo simulations of the shower image are presented .
based on the spatial energy deposit of shower particles as calculated by corsika @xcite it is shown that the lateral shower spread can be well parameterized as a function of the shower age parameter only .
the derived parameterization can be used for reconstruction of shower profiles .
this is illustrated by applying the new parametrization to the reconstruction of the primary shower energy for several simulated events in a fluorescence detector .
the plan of the paper is the following : the definition of the shower width and algorithm of fluorescence light production based on the corsika simulation of energy deposit density are described in section 2 . in section 3 an analytical parametrization is derived and its implementation in energy reconstruction procedure is discussed .
conclusions are given in section 4 .
photons which constitute an instantaneous image of the shower originate from a range of shower development stages @xcite , namely from the surface @xmath1 shown in figure 1 .
these simultaneous photons are defined as those which arrive at the fd during a short time window @xmath2 . during this @xmath2 ( corresponding to a small change of the shower position in the sky by @xmath3 ) the shower
front moves downward along the shower axis by a small distance @xmath4 , where @xmath5 is the distance from fd to the volume @xmath6 and @xmath7 is the angle between the shower axis and the direction towards fd .
this means that the small element of surface @xmath1 corresponds to a small volume @xmath6 .
the number of photons which arrive to the fd from volume @xmath6 can be calculated as : @xmath8 where @xmath9 is the distribution of light emitted , @xmath10 is the projection of the surface @xmath11 onto a surface perpendicular to direction of the shower axis , @xmath12 is the light collecting area of the detector , @xmath13 is the light transmission factor , @xmath14 is the normalized fluorescence wavelength spectrum .
these photons form an instantaneous image of the shower which can be described by the angular distribution of light recorded by the fd @xmath15 where @xmath16 is the small angle between the direction to the center of the image spot and the direction to volume @xmath6 , @xmath17 is the azimuth angle .
the size of shower image is defined as the minimum angular diameter @xmath18 of the image spot containing a certain fraction @xmath19 of the total light recorded by the fd .
a shower viewed from a large distance has , to a very good approximation , a circular image , independent of the direction of shower axis @xcite .
the intensity distribution of light in this image is proportional to the lateral distribution of the emitted fluorescence light in the shower at the viewed stage of evolution .
therefore the fraction of light received @xmath19 can be obtained from the corresponding fraction @xmath20 of light emitted around the shower axis @xmath21 where @xmath22 is the ( normalized ) lateral distribution of fluorescence light emitted .
here we have neglected the fact that photons from the side of the shower front facing the detector have been emitted at a time later than photons from the farther side of the shower . as will be shown later ,
relevant lateral distances are in the range 150 to 300 m and hence this is also the distance scale for the maximum time differences of emission @xcite . as there is no significant change of the lateral distribution expected of a shower traversing @xmath23300 m of air , the effect of different emission is negligible in our calculation . in the following
we shall consider showers close to the fd .
for these showers , the optical image size is mainly determined by the geometric size of the shower disk .
light absorption and multiple scattering cause only a minor , negligible modification of the shower image .
the main task is therefore to derive @xmath22 , which is also referred to as the shape function , since the brightness distribution of the shower image depends on the shape of @xmath22 . at the first approximation ,
the @xmath22 is proportional to the number of particles in the shower at a given lateral distance , assuming a constant fluorescence yield per particle in the shower . for an electromagnetic shower
this number of charged particles is given by the nishimura - kamata - greisen ( nkg ) function @xcite .
in this case the @xmath20 can be determined analytically , as has been shown in ref .
. however , @xmath20 derived in this way does not describe the light distribution well in case of hadronic showers .
this is due to the fact that the number of particles in a hadronic shower does not follow the nkg distribution well .
although the assumption that the amount of emitted light is proportional to the number of particles is adequate for a determination of the fluorescence signal in many cases , it is here not suited for the following reasons .
the fluorescence yield is proportional to the ionization energy deposited by the shower rather than to the total number of charged particles .
furthermore , the simulated number of particles in a monte carlo calculation depends on the threshold energies chosen by the user @xcite , above which particles are simulated .
particles falling below the threshold energy are discarded .
a better approximation for the fluorescence yield can be obtained by using the energy deposit @xmath24 as a function of atmospheric slant depth interval @xmath25 together with a density- and temperature - dependent fluorescence yield @xmath26 @xcite . in this approximation
the distribution of photons emitted around the shower axis is proportional to the lateral distribution of energy deposit , @xmath27 , where @xmath28 is the vertical depth interval and @xmath29 is the shower zenith angle . the distribution of energy deposit @xmath30 is calculated with the corsika shower simulation program as the sum of the energy released by charged particles with energies above the simulation threshold plus the releasable energy fraction of particles discarded due to the energy cut . more specifically , the following approximation is used @xcite : @xmath31 where @xmath32 is the ionization energy deposit of all charged particles traversing the depth interval @xmath33 .
@xmath34 denotes the energy of particles of type @xmath35 falling below the simulation threshold within this interval . in the following
we will study the lateral distribution of energy deposit density in air showers , as it is directly proportional to the number of expected fluorescence photons .
using corsika a two - dimensional energy deposit distribution around the shower axis is stored in histograms during the simulation process for 20 different vertical atmospheric depths .
each of the 20 horizontal layers has a thickness of @xmath36 g/@xmath37 and corresponds to a certain atmospheric depth : the first one to @xmath38 g/@xmath37 and the last one to @xmath39 g/@xmath37 .
linear interpolation between the observation levels is performed in order to get the lateral distribution at a given vertical depth @xmath40 located between two corsika observation levels @xmath41 and @xmath42 .
the fraction of energy deposit @xmath20 is calculated by numerically integrating the histograms up to the lateral distance @xmath43 .
the shower simulations are performed with the hadronic interaction models gheisha @xcite ( for interactions below 80 gev ) and qgsjet 01 @xcite .
electromagnetic interactions are treated by a customized version of the egs4 @xcite code . to reduce computing time ,
a thinning algorithm @xcite is selected within corsika .
the thinning level of @xmath44 has been chosen with the so - called optimum weight limitation @xcite .
this ensures that the artificial fluctuations in the longitudinal shower profiles introduced by the thinning method are sufficiently small for this analysis @xcite .
the kinetic energy thresholds for explicitly tracking particles were set to : 100 , 100 , 0.25 , 0.25 mev for hadrons , muons , electrons and photons , respectively .
the knowledge of the @xmath20 function can be used to calculate the true signal ( light ) from shower , which may be divided among several neighboring detector pixels .
below we propose a universal parameterization of @xmath20 based on corsika simulations and show how the true signal can be estimated with this parameterization . in the following
we study the dependence of the lateral energy deposit density on various variables .
a natural transverse scale length in air showers , which proves to be useful for obtaining a universal parameterization of the lateral distribution , is given by the molire radius @xcite @xmath45 where @xmath46 mev is the scale energy , @xmath47 mev the critical energy and @xmath48 g/@xmath49 the radiation length in air .
the local molire radius at a given atmospheric depth of shower development ( at altitude @xmath50 ) can be obtained by dividing eq .
( [ eq - mol1 ] ) by the air density , @xmath51 , and is approximately given by @xmath52 .
it is also well known that the distribution of particles in a shower at a given depth depends on the history of the changes of @xmath53 along the shower path rather than on the local @xmath53 value at this depth . to take this into account
, the @xmath53 value is calculated at @xmath54 cascade units ( radiation length @xmath55 ) above the considered depth @xcite .
using the value of the molire radius calculated based on the atmospheric profile ( the us standard atmosphere @xcite ) for vertical depth @xmath56 , the fraction of energy deposit density @xmath57 versus the distance in molire units @xmath58 is found .
the knowledge of @xmath57 gives a possibility to study variation of the shape of energy deposit density due to properties of the atmosphere .
the variation of the density of the atmosphere along the path of a shower affects the molire radius and consequently also the radial particle distribution . to characterize the development stage of a shower
, we introduce the shower age parameter @xmath59 where @xmath60 is the atmospheric depth of shower maximum extracted from simulated data was determined by fitting a gaisser - hillas type function @xcite to the corsika longitudinal profile of energy deposit . ] . with this definition ,
a shower reaches its maximum at @xmath61 .
figure [ fig339int ] presents the integrals of energy deposit density @xmath20 and @xmath57 for a vertical proton shower with primary energy @xmath62 eev , obtained at different atmospheric depths .
it is seen in figure [ fig339int]a that the shape of this integral distribution varies considerably only at depths smaller than 360 g/@xmath37 . at larger depths , and
in particular around the shower maximum , the variation of integral energy deposit profile is not significant .
however , this variation is larger when one plots this integral versus distance measured in molire units , as shown in figure [ fig339int]b . figure [ fig - en]a shows the dependence of the integral of energy deposit density @xmath20 on energy and primary particle .
it is seen that the integral profile only slightly depends on energy and primary particle .
the differences are even smaller if we plot the fraction of energy deposit density versus distance in molire units , as shown in figure [ fig - en]b .
the same shape of the @xmath63 ) profile for different primaries and energies means that variations of the @xmath20 profile are mainly due to the atmospheric effect i.e. dependence of the molire radius on altitude . for the same shower geometry ,
there are different altitudes of the maxima of proton and iron showers , and in consequence different values of molire radius @xmath53 .
since @xmath53 determines the lateral spread of particles in the shower , the shape function @xmath20 becomes broader for iron showers ( higher altitude of shower maximum than for proton shower ) .
figure [ fig - inc ] presents the dependence of the integral of energy deposit profile on zenith angle .
we note that corsika energy deposit lateral profiles are obtained for horizontal planes at the given observation level , so if one compares energy deposit densities between vertical and inclined showers , a projection of densities from horizontal plane to the plane normal to the shower axis is performed for inclined showers .
the corrected profile for a shower inclined at @xmath64 is shown in figure [ fig - inc]a by the solid line .
it is seen that this profile differs from the profile obtained for a vertical shower .
this means that the shape of @xmath20 depends on the zenith angle .
this dependence can be understood if one takes into account the influence of the atmospheric effect on the energy deposit profile @xmath20 .
for a homogeneous atmosphere , the shape function for inclined and vertical showers must be the same for the same development stage because the molire radius does not change with altitude . in case of an inhomogeneous atmosphere
, differences of the shape function between vertical and inclined shower should be proportional to the differences of the molire radius ( i.e density of air ) .
thus if changes of the shape function are due only to the atmospheric effect , then @xmath57 profile should be the same for vertical and inclined showers .
figure [ fig - inc]b confirms this assumption . the analysis of figs .
2b , 3b and 4b leads to the following conclusion : _ the lateral shape of the energy deposit density versus distance from shower axis measured in molire units is independent of the primary energy , primary particle type and zenith angle .
it depends , to a good approximation , only on the shower age_. figure [ fig9int ] confirms this conclusion , too . in this figure we present the integral of the energy deposit density for different age parameters for 10 individual proton and 5 individual iron showers with different zenith angles ( @xmath65 ) and energy 10 eev .
it is seen that the shower - to - shower fluctuations are strongly reduced for a given age when we correct @xmath20 profiles for the atmospheric effect i.e. plot @xmath57 .
also , there are no differences in the shape of @xmath57 for showers with different zenith angles and primary particle type .
this means that it is possible to find a universal function which describes the shape of energy deposit density as a function of shower age only .
following our earlier work @xcite we will use the function @xmath66 where the parameters @xmath67 and @xmath68 are assumed to be functions of shower age .
fits of this functional form to the integral of energy deposit density were performed for the data from figures [ fig9int]b , d , f and are shown in figure [ fit - age ] .
the values of the parameters @xmath67 and @xmath68 for different shower ages are presented in figure [ fig - par ] .
the age dependence of @xmath67 and @xmath68 parameters is well described by @xmath69 @xmath70 thus , eqs .
( [ eq - fit ] ) , ( [ eq - fita ] ) and ( [ eq - fitb ] ) give us a universal function which describes the fraction of energy deposit density within a specified distance from the shower axis for different energies , zenith angles and primary particles .
moreover , eq . ( [ eq - fit ] ) can be used to simulate the size of the shower image not only at shower maximum like in ref .
@xcite , but also for any shower development stage .
inverting eq .
( [ eq - fit ] ) and taking into account the distance from the detector to the shower ( @xmath71 ) we can find the angular size of the image @xmath16 that corresponds to a certain fraction of the total fluorescence light signal : @xmath72 as the shape of the lateral distribution of energy deposit can be well described by eq .
( [ eq - fit ] ) , it may be used to take into account the knowledge on the shower width in the procedure of shower reconstruction in the fluorescence detector .
one of the first steps in shower energy reconstruction is the calculation of the light profile at the aperture of the detector , based on the signal recorded by the detector pixels .
this signal is converted to the number of equivalent photons at the detector diaphragm .
for example , one procedure to determine such a profile is described in @xcite .
this algorithm uses as input the reconstructed geometry to locate the shower image on the fd telescope camera in 100 ns intervals .
next , the signal ( charge ) and noise from pixels lying within a predetermined angular distance @xmath73 from the instantaneous position of the image spot center are collected to find the radius @xmath74 that maximizes the signal - to - noise ratio * over the whole * shower track .
finally , the charge in each 100 ns time interval ( time slot ) , @xmath75 , within that radius @xmath74 is found and converted to the number of photons using calibration constants .
this procedure works well for distant showers , when the light collected within the radius @xmath76 corresponds to about 100% of the true signal , but some differences between the signal within @xmath74 and the true signal may exist for nearby showers . in the following we investigate this problem and estimate a correction to the described reconstruction algorithm .
the necessary shower reconstructions were performed using flores - eye @xcite and fdsim @xcite programs .
first , fdsim was used to generate events based on the gaisser - hillas parameterization .
next , the geometrical and energy reconstruction were performed using flores - eye .
the reconstructed geometries for some events are listed in table [ tab1 ] . using these geometries
, we find the collected signal @xmath75 within angular distance @xmath76 at each time interval depends on geometry , but for events listed in table [ tab1 ] it equals about @xmath77 . ] .
then , for the given @xmath74 the effective radius around shower axis @xmath78 and the fraction of light @xmath79 based on the function @xmath20 was calculated .
the fraction @xmath80 for events listed in table [ tab1 ] is shown in figure [ fig - frac ] .
it is seen that for event1 @xmath79 changes from 89% for a distance - to - shower @xmath81 km to 87% for @xmath82 km .
for other events , the collected fraction of the signal within @xmath74 increases with increasing distance - to - shower and equals on average about 91% , 94% and 99% for event2 , event3 , event4 , respectively .
in other words , some portion of the signal , which falls beyond @xmath74 is missing in the shower reconstruction procedure . to take into account this lost portion of the signal ,
the signal @xmath75 is rescaled according to formula @xmath83 . in this way
, one takes into account the shape of lateral distribution of energy deposit and obtains the new integrated charge @xmath84 for each time slot .
thus the part of the signal which was contained in neighboring pixels outside @xmath74 is accounted for .
it should be pointed out that in general , any reconstruction procedure has to take into account the pixellation of the detector .
independent of the specific approach , the correction procedure developed in this work can be applied . in the following ,
we demonstrate the influence of this correction on the light profile and on energy determination . in figure [ fig - prof2]a
it is shown that our correction leads to considerable differences between the @xmath75 ( dashed line ) and @xmath84 profile ( solid line ) for a nearby shower ( event 1 ) . in case of distant showers ( like event4 )
the profile is almost unchanged ( see figure [ fig - prof2]b ) .
it should also be noted that changes in the detector - to - shower distance are accounted for in this approach .
a differential
correction is applied ( i.e. for each time slot ) , which also leads to a better reconstruction of the longitudinal shape ( and thus @xmath60 ) of the shower . accepting only a fraction of the signal
contained within @xmath74 directly influences the reconstructed primary energy of the shower . in table
[ tab2 ] we present the influence of the @xmath75 correction on the gaisser - hillas fit to the reconstructed number of particles in the showers .
it is seen that this correction changes both the number of particles at the shower maximum and the position of the shower maximum .
these changes lead to different estimates of primary energy . in the last column
the relative differences @xmath85 are listed .
one sees that @xmath86 is always positive and decreases from 14% for a distance to shower maximum of @xmath87=6.5 km to 2% for @xmath87=23 km .
in this work , the distribution of light in the shower optical image is analyzed , based on the lateral distribution of energy deposited by the shower , as derived from corsika simulations .
the lateral distribution of energy deposited is parameterized with a functional form inspired by the nkg distribution .
the angular distribution of photons arriving simultaneously at the detector ( i.e. the intensity distribution of light in the instantaneous image of the shower ) is obtained .
the shape of this distribution can be approximated by a universal function that depends on the shower age only .
this universal function is used to derive a correction to the shower energy due to the fraction of light falling into detector pixels located far from the center of the shower image . in the usual procedure of shower reconstruction , signal - to - noise ratio is optimized , so that pixels lying far from the center of the shower image are not included in the analysis .
the percentage of shower signal in those outlying pixels was determined in this paper based on the lateral distribution of light in the shower image .
the signal recorded by the fluorescence detector in the accepted central pixels is rescaled , so that a corrected light profile of the shower is obtained . for events
examined , this correction increases the estimated shower energy by 214% , depending on the detector - to - shower distance
. _ acknowledgements .
_ this work was partially supported by the polish committee for scientific research under grants no .
pbz kbn 054/p03/2001 and 2p03b 11024 and in germany by the daad under grant no .
mr is supported by the alexander von humboldt foundation .
.characteristics of events used for the comparisons in this paper . the shower zenith angle @xmath29 , azimuth angle @xmath17 , core position @xmath90 , @xmath91 are measured relative to fd detector . [ tab1 ] [ cols="^,^,^,^,^ " , ] | the light intensity distribution in a shower image and its implications to the primary energy reconstructed by the fluorescence technique are studied .
based on detailed corsika energy deposit simulations , a universal analytical formula is derived for the lateral distribution of light in the shower image and a correction factor is obtained to account for the fraction of shower light falling into outlying pixels in the detector .
the expected light profiles and the corresponding correction of the primary shower energy are illustrated for several typical event geometries .
this correction of the shower energy can exceed 10% , depending on shower geometry . |
the purpose of this paper is to show how different properties of a data set affect the results from using a method developed in a previous paper ( speights & westpfahl 2011 , hereafter paper i ) for solving the pattern speed ( @xmath0 ) equations of tremaine & weinberg ( 1984 , hereafter tw84 ) .
this is important for choosing data in future applications of the solution method , with the long - term goal of determining whether spiral patterns are rigidly rotating waves , shearing material arms , or perhaps some combination of both .
the pattern speed is traditionally determined from a guided interpretation of observations according to the density wave theory ( lin & shu 1964 , 1966 ; lin et al .
1969 ) . in density wave theory
@xmath0 is a free parameter that is measurable by assuming that it is constant ( i.e. , independent of radius ) , and that resonant radii are identifiable from photometric and kinematic features of the galaxy .
resonance is assumed to occur when @xmath0 = @xmath1 , the material speed ( corotation resonance ) , and when @xmath0 = @xmath1 @xmath2 @xmath3/@xmath4 , where @xmath3 is the epicycle frequency of a stellar orbit and @xmath4 is the mode of the wave corresponding to the number of arms ( lindblad resonance ) . the assumption that @xmath0 is constant is appealing if one assumes that spiral patterns persist for many dynamical ( orbital ) timescales because this avoids the winding dilemma for material arms . the tw84 equations ,
however , are independent of the detailed assumptions of spiral structure theories , and solutions can be used to test those assumptions .
these are integrated forms of the continuity equation that relate the pattern speed to the observable kinematic and density properties of a pattern tracer .
the derivation in tw84 assumes that the disk is flat , that the pattern speed is well defined , and that the pattern tracer obeys the continuity equation . in paper
i the tw84 equations are solved for a pattern speed that is allowed to vary with radius .
the solution method is then applied to the h1.5pt i spiral pattern of the barred , grand - design galaxy ngc 1365 .
a statistical analysis of the results rules out the hypothesis of a constant @xmath0 for the spiral arms .
the pattern speed behaves approximately as 1/@xmath5 , and is very similar to the material speed .
there are also no clear indications of unique corotation or lindblad resonances for the spiral arms .
it is noted in paper i that the winding dilemma is avoidable for ngc 1365 if the spiral pattern is a transient feature .
the results for ngc 1365 are inconsistent with traditional methods for measuring spiral arm pattern speeds .
more direct measurements that allow for a radially varying @xmath0 are therefore needed in order to estimate how common shearing patterns are .
there is an abundant amount of archived data for doing so , and a better understanding of how different properties of a data set affect the results will help with selecting useful data .
this paper reports the results from applying the solution method of paper i to the grand - design galaxy ngc 3031 , the irregular galaxy ngc 2366 , and the dwarf irregular galaxy ddo 154 .
four different data sets of ngc 3031 show how angular resolution , spectral resolution , and sensitivities to structures on different scales affect the results .
the pattern speed is found for ngc 2366 and ddo 154 to show how the results are affected in the absence of a coherent or well - organized pattern .
the results for the irregular galaxies also provide a counterexample to the results for ngc 3031 .
the rest of this paper is organized as follows . in section 2 the solution method developed in paper
i is briefly reviewed , and a new procedure is introduced for selecting models that are an alternative to a constant @xmath0 . in section 3 the method of analyzing the results is explained . in section 4
the data are described . in section 5
the results are presented and analyzed . in section 6
the results are discussed . in section 7
is a summary .
when @xmath0 is allowed to vary with radius in the tw84 equations one of the results is an integral equation , @xmath6 that is solvable for simple functional forms of @xmath0 . in equation
( 1 ) , @xmath7 is the intensity of a pattern tracer and @xmath8 is the line of sight velocity of the tracer ( corrected for the systemic velocity , @xmath9 ) times the cosecant of the disk inclination , @xmath10 .
the @xmath11 and @xmath12 coordinates are measured relative to the kinematic minor and major axis respectively . for an illustration showing how a galaxy disk is oriented in the coordinate system of equation ( 1 ) ,
see figure ( 1 ) in paper i. the two forms of @xmath0 that are used in paper i for solving equation ( 1 ) are a polynomial in @xmath5 = @xmath13 , @xmath14 and a polynomial in 1/@xmath5 , @xmath15 where that @xmath16 is an integer and @xmath17 are unknown coefficients that are determined in a solution . both equations ( 2 ) and
( 3 ) can represent a constant @xmath0 when @xmath16 = 0 , and higher order terms can be used to determine whether the pattern is rigid or shearing . determining the order at which point it is appropriate to truncate the summations in equations ( 2 ) and ( 3 ) is the primary goal of the analysis in paper i. the advantage of using such forms for @xmath0 is that the unknown coefficients can come out of the integral on the left hand side of equation ( 1 ) . using equation ( 2 ) , equation ( 1 )
becomes , @xmath18 or similarly using equation ( 3 ) , @xmath19 applying either equation ( 4 ) or ( 5 ) independently ( i.e. , at least a beam size apart in the case of radio synthesis maps ) to different parts of the galaxy transforms the integral equation into a matrix equation of the form @xmath20 = @xmath21 .
that the calculations are independent is important because this is assumed in the methods used for estimating and propagating uncertainties , and in the statistics used for analyzing the results .
the matrix equation is easily solved using the normal equations , @xmath22 = @xmath23 , for the linear least - squares problem .
methods for the analysis of the results for linear least - squares problems are well established , and this is another advantage of using such forms for @xmath0 . the results for ngc 1365 demonstrate that only a few terms are needed in the above polynomials in order to determine the radial behavior of @xmath0 . including many terms
can unnecessarily complicate the analysis , and by @xmath24 = 3 the matrix @xmath25 typically becomes ill - conditioned for numerical inversion .
when @xmath25 is ill - conditioned , the solution is unstable , and any inferences based on such solutions may be unreliable . by summing equations ( 2 ) or ( 3 ) to order @xmath16
= 3 one can infer whether there is any sign of a discontinuity in @xmath0 such as a step function .
this is useful in paper i because the bar in ngc 1365 may have a separate @xmath0 than that of the spiral arms ( sellwood & sparke 1988 ; sellwood 1993 and references therein ) , and it needed to be shown whether the pattern speed of the bar is detected in the results . since the pattern of ngc 3031 shows no obvious discontinuities or signs of a bar ( see figures 1 through 4 ) , and in order to simplify the analysis for all three galaxies , only two terms are used in this paper as an alternative to a constant @xmath0 model .
the alternative models have the form , @xmath26 but without the restriction to integer values of @xmath16 . relaxing the restriction to integer values of @xmath16
allows for different amounts of curvature in @xmath0 .
the value of @xmath16 is determined by comparing the goodness of fit for solutions of equation ( 1 ) using equation ( 6 ) and a range of different values for @xmath16 . for each solution , the distribution of the residuals , normalized by the propagated uncertainties ,
is compared with a standard normal distribution .
if one assumes that the uncertainties are well known , the distribution of the normalized residuals will be a standard normal distribution when the true form of @xmath0 is used to find a solution .
therefore , the value of @xmath16 that results in a distribution of the normalized residuals that is the most similar to a standard normal distribution is used to chose an alternative model for subsequent analysis . a kolmogorov - smirnov ( @xmath27 )
test is used to quantify how similar the distribution of the normalized residuals is to a standard normal distribution .
the @xmath27 test consists of finding the largest distance between two cumulative distribution functions ( cdf ) , @xmath28 = max@xmath29cdf(@xmath30 ) @xmath31 cdf(@xmath32)@xmath29 , where that in this case @xmath30 is the distribution of the normalized residuals and @xmath32 is a standard normal distribution .
then , the probability , @xmath33 , is found for obtaining a value of @xmath28 should the null hypothesis be true that @xmath30 is a standard normal distribution ( kendall & stewart 1973 , chapter 30 ; lupton 1993 , chapter 19 ) . if @xmath33 is sufficiently small , then there is convincing evidence that @xmath30 @xmath34 @xmath32 , and one can infer that the model in question is not the true model .
the following convention is adopted for interpreting values of @xmath33 : @xmath33 @xmath35 @xmath36 is convincing evidence that @xmath30 @xmath34 @xmath32 ; @xmath33 @xmath37 @xmath38 is suggestive of @xmath30 @xmath34 @xmath32 , but inconclusive ; and larger values can not rule out the null hypothesis that @xmath30 is a standard normal distribution . in this paper a lower limit of 0.01% is chosen as a cutoff for reporting small probabilities . as a compliment to
the analysis explained in the next section , the @xmath27 test is also applied to the results from using a constant @xmath0 model .
for the purpose of determining @xmath16 , a larger value of @xmath33 is considered an indication that the distribution of the normalized residuals is more similar to a standard normal distribution . in section 5
it is shown that two peaks consistently occur in plots of @xmath33 vs @xmath16 , one each for negative and positive values of @xmath16 .
the results are analyzed for each case , where the model using a negative value of @xmath16 = @xmath39 is referred to as alternative model 1 ( am1 ) , and the model using a positive value of @xmath16 = @xmath40 is referred to as alternative model 2 ( am2 ) .
the goal of the analysis is to determine whether the pattern is rigid or shearing .
this is based on a comparison of reduced chi - square ( @xmath41 for @xmath42 degrees of freedom ) for a solution using a constant @xmath0 model with the values obtained for am1 and am2 .
unless otherwise noted in section 5 , the validity of using @xmath41 as an indicator of the goodness of fit is based on the results of a lilliefors test for quantifying how similar @xmath30 is to a normal distribution . the lilliefors test follows the same procedure as the @xmath27 test , with the difference that @xmath28 = max@xmath29cdf(@xmath30 ) @xmath31 cdf(@xmath43)@xmath29 , where that @xmath43 is a normal distribution with a mean and a standard deviation that is the same as that for @xmath30 ( mandansky 1988 , chapter 1 ) .
if the probability , @xmath44 , of obtaining @xmath28 is sufficiently small , then @xmath30 may not have a normal distribution , and the value of @xmath41 may not be a reliable indicator of the goodness of fit .
the same convention that is adopted for interpreting values of @xmath33 is adopted for interpreting values of @xmath44 .
if a solution using am1 or am2 has a value of @xmath41 that is an improvement to the value of @xmath41 for the constant @xmath0 model , then fisher s @xmath45 test of an additional term is used to determine if the improvement is statistically significant .
the @xmath45 test is a natural choice for quantifying the significance of the improvement in @xmath41 because the constant @xmath0 model is a reduced version of am1 and am2 .
the @xmath45 test consists of calculating @xmath45 @xmath46 @xmath47/@xmath48 and finding the probability , @xmath49 , for obtaining such a value of @xmath45 should the null hypothesis be true that the extra term , @xmath50 , is zero ( bevington & robinson 2002 , chapter 10 ) .
if @xmath51 is sufficiently small , then there is convincing evidence that @xmath52 @xmath34 0 , and one can infer that the pattern is shearing . if such evidence is found then the result for the constant @xmath0 model is the mean of @xmath0 . the following convention is adopted for interpreting values of @xmath49 : @xmath49 @xmath35 @xmath36 is convincing evidence that @xmath52 @xmath34 0 ; @xmath49 @xmath37 @xmath38 is suggestive of @xmath52 @xmath34 0 , but inconclusive ; and larger values are an indication that the improvement is not statistically significant .
the uncertainties for calculating @xmath41 , confidence intervals for the coefficients , and confidence bands for @xmath0 , are estimated by propagating the uncertainties per pixel in the maps of @xmath7 and @xmath8 , as well as the uncertainties in @xmath11 , @xmath12 , @xmath10 , and @xmath53 through the calculations of both * g * and * d*. the method of estimating the uncertainties in the maps of @xmath7 and @xmath8 are adopted from the reasoning given in paper i. this takes into account the rms noise , the channel width , and the number of channels a typical detection occurs .
the total uncertainty for the @xmath54th independent calculation of equation ( 1 ) is @xmath55 @xmath46 @xmath56 @xmath57 @xmath58 ( bevington & robinson 2002 , chapter 6 ) .
the confidence intervals and confidence bands are not estimated using the covariance matrix because of the uncertainties in * g*. instead , these are estimated using monte carlo methods . an incorrect value for the position angle of the disk ( pa ) can potentially cause large errors in solutions of the tw84 equations ( debattista 2003 ; debattista & williams 2004 ) .
the pa does not explicitly appear in equation ( 1 ) , and consequently the uncertainty in the pa is not accounted for in the calculation of the propagated uncertainties . how an incorrect pa could affect the conclusions in this paper is investigated in section 5 by changing the pa and applying the solution method and analysis .
this is done for a range of -2@xmath59 @xmath60 @xmath61pa @xmath60 2@xmath59 where that @xmath59 is the 1@xmath62 uncertainty for the adopted pa . in paper
i it is explained that this method is preferred to monte carlo methods because monte carlo methods typically produce non - gaussian results for the distributions of the coefficients . in order to minimize the uncertainties in the analysis , instrumental units of [ @xmath0 ]
= km s@xmath64 arcsec@xmath64 are used for solving equation ( 1 ) and analyzing the results .
this is preferred to units commonly used for galactic disks , [ @xmath0 ] = km s@xmath64 kpc@xmath64 , because the conversion from arcseconds to kpc adds a large amount of uncertainty to @xmath11 and @xmath12 owing to the uncertainty in the distance to the galaxy .
furthermore , units commonly used for galactic disks are not necessary for the purpose of this paper .
radio synthesis intensity and velocity maps of h1.5pt i are used for applying the solution method . for a discussion of the advantages of using h1.5pt i as a pattern tracer the interested reader
is referred to paper i. although tw84 use a form of the continuity equation that assumes mass conservation , the tw84 equations are derivable form the general form of the continuity equation that includes a source function for the creation and destruction of material ( westpfahl 1998 ) . if the source function is constant within dynamical timescales , then it does not affect the calculation of the pattern speed ( paper i ) .
for the purpose of this paper it is assumed that the source function is approximately constant within dynamical timescales .
the disk parameters that are adopted for measuring @xmath0 are provided in table 1 .
the values for the kinematic centers are adopted from trachternach et al .
the values for @xmath10 , pa , and @xmath53 are adopted from de blok et al .
( 2008 ) for ngc 3031 , and from oh et al .
( 2011 ) for ngc 2366 and ddo 154 . llccccc[ht ! ] + galaxy & & & & & + & & r. a. ( j2000 ) & dec .
( j2000 ) & @xmath10 & pa & @xmath9 + & & ( @xmath65 @xmath66 @xmath67 ) & ( @xmath68 @xmath69 @xmath70 ) & ( @xmath68 ) & ( @xmath68 ) & ( km s@xmath64 ) + + ngc 3031 & & 09 55 33.5 @xmath2 0.6 & + 69 03 52.0 @xmath2 3.9 & 59 @xmath2 2.6 & 330 @xmath2 1.3 & 0.3pt @xmath3139 @xmath2 2.8 + ngc 2366 & & 07 28 53.9 @xmath2 0.7 & + 69 12 37.4 @xmath2 7.8 & 63 @xmath2 1.6 & 4.0pt 39 @xmath2 4.0 & 104 @xmath2 2.0 + ddo 154 & & 12 54 05.9 @xmath2 0.2 & + 27 09 09.9 @xmath2 3.4 & 66 @xmath2 2.5 & 229 @xmath2 7.8 & 375 @xmath2 1.5 figures 1 through 6 show the maps of @xmath7 and @xmath8 for the four different data sets of ngc 3031 , and the data sets of ngc 2366 and ddo154 , respectively .
the original maps were modified into their current form using the astronomical image processing system developed and maintained by the national radio astronomy observatory ( nrao ) .
these are rotated and shifted from their original orientation so that their kinematic major axes are along the @xmath11 axis , and the origins of the @xmath11-@xmath12 coordinates are the kinematic centers that are adopted for each galaxy . included in the figures
are optical images from the digitized sky survey .
these are based on the photographic data of the national geographic society - palomar observatory sky survey ( ngs - poss ) , and are provided for comparison with the h1.5pt i maps .
the optical images are from the 48 inch samuel oschin telescope on palomar mountain and a 103ae ( 6450 ) emulsion .
table 2 provides a summary of the properties of each data set .
the data are described in greater detail in the following subsections .
lllccc[ht ! ] + galaxy & & notes & channel seperation & beam fwhm & rms + & & & ( km s@xmath64 ) & ( @xmath70 ) & ( mjy beam@xmath64 ) + + ngc 3031 & & aw96a & 2.6 & 9.0 @xmath71 9.0 & -4.6pt 0.6 + ngc 3031 & & aw96b & 2.6 & 60.0 @xmath71 60.0 & 1.23 + ngc 3031 & & h84 & -4.5pt 10.3 & 9.0 @xmath71 9.0 & -4.6pt 2.5 + ngc 3031 & & yp99 & 6.6 & 135.3 @xmath71 124.8 & -4.6pt 6.8 + ngc 2366 & & & 2.6 & 13.1 @xmath71 11.9 & -4.6pt 0.5 + ddo 154 & & & 2.6 & 14.1 @xmath71 12.6 & -4.6pt 0.5 the four different data sets of ngc 3031 are from three different observations ( adler & westpfahl 1996 , hereafter aw96 ; hine 1984 , hereafter h84 ; yun & purton 1999 , private communication , hereafter yp99 ) .
they have different combinations of angular resolution , spectral resolution , and sensitivities to structures on different scales .
two of the four data sets , hereafter referred to as aw96a and aw96b , are made from the observations of aw96 .
they used the very large array ( vla ) of the nrao in 1992 and 1993 with all 27 antennas , 128 channels , one intermediate frequency ( if ) , and on - line hanning smoothing .
data from two pointings of the vla in the b , c , and d configurations with a channel width and separation of 12.2 khz ( 2.6 km s@xmath64 ) are combined to make a mosaic .
the maps published by aw96 do not have the continuum emission of the point sources subtracted .
for the aw96a data set this emission is subtracted from a calibrated and mapped data cube provided by adler ( 2010 , private communication ) . the continuum subtracted data cube was used to make new intensity and velocity maps .
for this data set the rms signal in a line - free channel is 0.6 mjy beam@xmath64 , and the fwhm of the synthesized beam is a 9.0@xmath70 circle .
the aw96b data set is from the same calibrated @xmath72-@xmath73 data , also provided by adler ( 2010 , private communication ) .
the @xmath72-@xmath73 data for this data set is tapered using a gaussian weighting function so that the synthesized beam is a 60.0@xmath70 circle . for the aw96b data set the rms signal in a line - free channel is 1.23 mjy beam@xmath64 .
the details of the observing and data reduction are described by aw96 .
the data of h84 are made from observations using the vla in 1982 .
owing to limitations of the telescope and correlator at that time only 18 antennas and 32 spectral channels could be used with one if .
data from two pointings of the vla in the b and d configurations with a channel separation of 48.8 khz ( 10.3 km s@xmath64 ) are combined by h84 to make a mosaic .
the rms signal in a line - free channel is 2.5 mjy beam@xmath64 .
the fwhm of the synthesized beam is a 9.0@xmath70 circle .
this data set includes the same range of baselines as does the aw96 data sets , but with relatively fewer intermediate ones due to the exclusion of the c configuration .
the details of the observing and data reduction are described by h84 .
the data of yp99 are made from observations using the synthesis telescope of the dominion radio astrophysical observatory in 1993 .
the calibrated and mapped data cube is kindly made available in advance of publication .
the channel separation is 15.63 khz ( 3.30 km s@xmath64 ) , the channel width is 24.99 khz ( 5.28 km s@xmath64 ) .
pairs of channels are averaged to improve the signal - to - noise ratio , giving final channel separations of 31.25 khz or 6.60 km s@xmath64 .
the rms signal in a line - free channel of this width is 6.8 mjy beam@xmath64 .
the fwhm of the synthesized beam is 135.3@xmath71 124.8@xmath70 with a beam pa of -4.4@xmath74 .
regions of statistically significant signal in this cube are identified by first convolving each channel with a circular gaussian of diameter 200@xmath70 .
regions in the convolved channel with signal of absolute value less than 2.5 mjy beam@xmath64 are judged to contain no significant detection , and these regions are blanked in the original , unconvolved channel .
the blanked channels are used to make the data maps .
the different data sets emphasize different attributes of ngc 3031 .
the yp99 data set excels at showing the faint outer structure , but its large synthesized beam can not show fine detail .
the aw96a data set excels at showing fine detail , but aw96 show that it is missing flux , particularly on large scales .
the aw96b data set does not show the fine detail of the aw96a data set due to the @xmath72-@xmath73 tapering .
the h84 data set is in some ways a compromise between the aw96a and yp99 data sets .
it does not have the sensitivity to the small structures of aw96a , but it is missing little or no flux from ngc 3031 itself .
more fine detail is shown in figure 1 for the data set from aw96a than in figure 3 for the data set from h84 despite having the same sized synthesized beam .
this is because of the much higher sensitivity of the aw96a data set .. the data sets for ngc 2366 and ddo 154 are from the things h1.5pt i survey ( walter et al .
both data sets were made from b , c , and d observations using the vla .
the observations of ngc 2366 are made in 2003 and 2004 , and of ddo 154 in 2004 and 2005 .
both data sets have a channel width of 2.6 km s@xmath64 and an rms signal in a line - free channel of 0.5 mjy beam@xmath64 .
the fwhm of the synthesized beam is 13.1@xmath70 @xmath71 11.9@xmath70 with a beam pa of -1.6@xmath74 for ngc 2366 , and 14.1@xmath70 @xmath71 12.6@xmath70 with a beam pa of -34.0@xmath74 for ddo 154 .
the details of the observing and data reduction are described by walter et al .
the calculations for solving equation ( 1 ) and analyzing the results are carried out in matlab .
differentiation is performed using the savitzky - golay method ( savitzky & golay 1964 ) .
integration is performed by summing over a row of pixels in a map of the integrand . the @xmath27 test and the lilliefors test
are performed using matlab s statistics toolbox .
the rest of the beginning of this section explains how the results are presented .
the details of the results and their analysis are discussed in the subsections that follow .
plots of @xmath33 vs @xmath16 are shown in figure 7 .
note that two peaks consistently occur in the plots .
it is assumed that @xmath0 does not have a large amount of curvature so that @xmath39 is not @xmath75 1 and @xmath40 is not @xmath76 1 . out of a range of -6 @xmath60 @xmath16 @xmath60 6 for
which solutions are found , the peaks occur within -2 @xmath60 @xmath16 @xmath60 2 , so only this range is shown in the figure .
the values of @xmath33 are not shown in the figure for when @xmath16 = 0 because the determinant of @xmath25 is zero for a model of @xmath0 = @xmath77 @xmath57 @xmath52 .
the values of @xmath16 , @xmath28 , and @xmath33 for the peaks in figure 7 are summarized in table 3 .
values of @xmath28 and @xmath33 for the constant @xmath0 model are also shown in table 3 .
the details of the results for the @xmath27 test are discussed in paragraph 2 of sections 5.1 , 5.2 , and 5.3 .
the results for the coefficients are shown in table 4 .
the uncertainties for the coefficients are reported as the half - widths ( hw ) of their 95% confidence intervals .
the coefficients are used to calculate plots of @xmath0 for comparison with @xmath1 and possible locations for lindblad resonance .
the plots of @xmath0 calculated from the coefficients are discussed in paragraph 3 of sections 5.1 , 5.2 , and 5.3 .
the results for the @xmath45 test are shown in table 5 .
an asterisk next to a value of @xmath78 indicates that the normalized residuals either failed a liliefors test for normality , or that the results of the test are inconclusive .
the 1@xmath62 uncertainties for @xmath41 are estimated from the variance , 2@xmath42 , of a @xmath79 distribution .
these show how similar the values of @xmath78 are for a given pair of am1 and am2 .
the details of the results for the @xmath45 test are discussed in paragraph 4 of sections 5.1 , 5.2 , and 5.3 .
results are presented for the entire spiral pattern , which consists of the region @xmath80
@xmath60 600@xmath70 in the h1.5pt i maps .
this provides 42 independent calculations of equation ( 1 ) for the aw96a and h84 data sets , 10 independent calculations for the aw96b data set , and 4 independent calculations for the yp99 data set . as a check ,
solutions are also found for three other regions : @xmath80 @xmath60 300@xmath70 , @xmath31600@xmath70 @xmath60 @xmath12 @xmath60 0@xmath70 , and 0@xmath70 @xmath60 @xmath12 @xmath60 600@xmath70 .
the results for the three other regions are consistent with those for the entire spiral pattern , so they are not presented separately . in figure 7
there are prominent peaks in the values of @xmath33 for all four data sets .
the values of @xmath33 are generally quite large at their peaks , with the exception of the peaks at @xmath40 for the aw96b and yp99 data sets .
the aw96a , aw96b , and h84 data sets show small @xmath33 near @xmath16 = 0 , but this is not the case for the yp99 data set .
for all four data sets the values of @xmath33 for the constant @xmath0 model are small enough to rule out the null hypothesis that the normalized residuals have a standard normal distribution .
this is convincing evidence that the constant @xmath0 model is not the true model .
this conclusion , however , is only reliable for the aw96a , aw96b , and h84 data sets because the @xmath27 test becomes unreliable when there are only a few calculations of the normalized residuals .
this may explain the discrepancy between the values of @xmath33 near @xmath16 = 0 in figure 7 , and the small value of @xmath33 for the constant @xmath0 model in table 3 for the yp99 data set .
llcccccccccc[b ! ] + galaxy & notes & & & & & + & & @xmath28 & @xmath33 & & @xmath39 & @xmath28 & @xmath33 & & @xmath40 & @xmath28 & @xmath33 + & & @xmath71 10@xmath64 & ( % ) & & & @xmath71 10@xmath64 & ( % ) & & & @xmath71 10@xmath64 & ( % ) + + ngc 3031 & aw96a & 3.64 & @xmath81 0.01 & & -0.89 & 1.18 & 56.4 & & 0.80 & 1.46 & 30.5 + ngc 3031 & aw96b & 6.00 & 10.0pt 0.06 & & -0.80 & 1.52 & 94.9 & & 0.36 & 3.44 & 14.7 + ngc 3031 & h84 & 4.15 & @xmath81 0.01 & & -0.61 & 0.94 & 82.1 & & 0.84 & 0.84 & 90.6 + ngc 3031 & yp99 & 7.50 & 10.0pt 0.78 & & -1.42 & 2.47 & 90.9 & & 0.31 & 4.99 & 18.9 + ngc 2366 & @xmath82 & 3.33 & 10.0pt 0.45 & & -0.11 & 1.09 & 88.4 & & 0.12 & 1.01 & 93.2 + ngc 2366 & @xmath83 & 6.14 & @xmath81 0.01 & & -0.06 & 1.58 & 82.3 & & 0.08 & 1.55 & 84.2 + ngc 2366 & @xmath84 & 4.02 & 10.0pt 2.09 & & -0.20 & 1.38 & 93.5 & & 0.25 & 1.45 & 91.2 + ngc 2366 & @xmath85 & 3.32 & 10.0pt 8.89 & & -0.31 & 2.09 & 55.2 & & 0.41 & 1.52 & 88.0 + ddo 154 & & 1.79 & 1.5pt 37.9 & & -1.01 & 1.62 & 50.4 & & 1.82 & 2.11 & 20.6 llccccccc + galaxy & notes & constant @xmath0 & & & & + & & @xmath77 @xmath2 hw@xmath86 & & -10pt @xmath77 @xmath2 hw@xmath86 & -5pt @xmath52 @xmath2 hw@xmath87 & & -10pt @xmath77 @xmath2 hw@xmath86 & -5pt @xmath52 @xmath2 hw@xmath87 + & & ( km s@xmath64 arcsec@xmath64 ) & & -10pt ( km s@xmath64 arcsec@xmath64 ) & -5pt ( km s@xmath64 ) & & -10pt ( km s@xmath64 arcsec@xmath64 ) & -5pt ( km s@xmath64 arcsec@xmath88 ) + + ngc 3031 & aw96a & ( 4.02 @xmath2 0.51 ) @xmath71 10@xmath64 & & -10pt ( -1.20 @xmath2 0.53 ) @xmath71 10@xmath64 & -1pt(1.50 @xmath2 0.15 ) @xmath71 10@xmath89 & & -41pt 0.95 @xmath2 0.09 & -0.5pt ( -0.32 -0.3pt @xmath2 0.54 ) @xmath71 10@xmath88 + ngc 3031 & aw96b & ( 2.12 @xmath2 1.20 ) @xmath71 10@xmath64 & & -10pt ( -1.04 @xmath2 0.64 ) @xmath71 10@xmath64 & -1pt(0.73 @xmath2 0.13 ) @xmath71 10@xmath89 & & -41pt 1.13 @xmath2 0.24 & -0.5pt ( -7.19 -0.3pt @xmath2 1.84 ) @xmath71 10@xmath88 + ngc 3031 & h84 & ( 3.84 @xmath2 0.60 ) @xmath71 10@xmath64 & & -10pt ( -4.57 @xmath2 1.70 ) @xmath71 10@xmath64 & -1pt ( 0.49 @xmath2 0.10 ) @xmath71 10@xmath89 & & -41pt 1.07 @xmath2 0.16 & -0.5pt ( -0.25 -0.3pt @xmath2 0.06 ) @xmath71 10@xmath88 + ngc 3031 & yp99 & ( 3.11 @xmath2 3.11 ) @xmath71 10@xmath64 & & -10pt ( -0.55 @xmath2 4.20 ) @xmath71 10@xmath64 & -4.6pt ( 25.21 -0.0pt @xmath2 -0.4pt 26.49 ) @xmath71 10@xmath89 & & -41pt 1.35 @xmath2 0.84 & -4.8pt ( -14.7 @xmath2 11.8 ) 4.1pt @xmath71 10@xmath88 + ngc 2366 & @xmath82 & ( 2.32 @xmath2 0.73 ) @xmath71 10@xmath64 & & -41pt -1.48 @xmath2 0.82 & -26pt 3.00 @xmath2 1.44 & & -41pt 1.71 @xmath2 0.72 & -0.5pt ( -7.76 -0.3pt @xmath2 3.76 ) @xmath71 10@xmath64 + ngc 2366 & @xmath83 & ( 2.66 @xmath2 1.06 ) @xmath71 10@xmath64 & & -41pt -2.63 @xmath2 2.96 & -26pt 3.93
@xmath2 4.02 & & -41pt 2.51 @xmath2 1.84 & -4.8pt ( -15.3 @xmath2 12.5 ) 4.1pt @xmath71 10@xmath64 + ngc 2366 & @xmath84 & ( 2.13 @xmath2 1.00 ) @xmath71 10@xmath64 & & -41pt -1.09 @xmath2 0.69 & -26pt 3.60
@xmath2 1.89 & & -41pt 1.21 @xmath2 0.53 & -0.5pt ( -2.66 -0.3pt @xmath2 1.42 ) @xmath71 10@xmath64 + ngc 2366 & @xmath85 & ( 2.63 @xmath2 1.24 ) @xmath71 10@xmath64 & & -41pt -0.21 @xmath2 0.44 & -26pt 2.29 @xmath2 2.09 & & -41pt 0.61 @xmath2 0.35 & -0.5pt ( -0.39 -0.3pt @xmath2 0.38 ) @xmath71 10@xmath64 + ddo 154 & & ( 1.76 @xmath2 0.33 ) @xmath71 10@xmath64 & & -10pt ( -1.22 @xmath2 0.29 ) @xmath71 10@xmath64 & -1pt ( 1.26 @xmath2 0.46 ) @xmath71 10@xmath90 & & -10pt ( 2.15 @xmath2 0.50 ) @xmath71 10@xmath64 & ( -1.33 -0.5pt @xmath2 4.3pt 1.36 ) @xmath71 10@xmath91 llccccccccc[h ! ] + galaxy & notes & constant @xmath0 & & & & + & & @xmath92 @xmath2 @xmath62 & & -6pt @xmath92 @xmath2 @xmath62 & @xmath45 & @xmath93 & & -8pt @xmath92 @xmath2 @xmath62 & @xmath45 & @xmath93 + & & & & & & ( % ) & & & & ( % ) + + ngc 3031 & aw96a & 25.8 5.5pt @xmath2 0.22 & & 1.16
@xmath2 0.22 & -8.0pt 873 & @xmath81 0.01 & & 1.06 @xmath2 0.22 & -3.5pt 955 & @xmath81 0.01 + ngc 3031 & aw96b & -4.4pt 125 12.3pt @xmath2 0.47 & & 4.28 @xmath2 0.50 & -8.0pt 255 & @xmath81 0.01 & & 2.12 @xmath2 0.50 & -3.5pt 523 & @xmath81 0.01 + ngc 3031 & h84 & 24.0 5.5pt @xmath2 0.22 & & 1.29 @xmath2 0.22 & -8.0pt 874 & @xmath81 0.01 & & 1.09 @xmath2 0.22 & -7.5pt 1036 & @xmath81 0.01 + ngc 3031 & yp99 & 73.0 5.5pt @xmath2 0.82 & & 3.40 @xmath2 1.00 & 2.9pt 52.8 & 9.5pt 1.84 & & 2.19 @xmath2 1.00 & 7.5pt 97.9 & 9.5pt 1.01 + ngc 2366 & @xmath82 & 4.3pt 9.54 1.3pt @xmath2 0.28 & & 0.90 @xmath2 0.29 & -8.0pt 241 & @xmath81 0.01 & & 0.87 @xmath2 0.29 & -3.5pt 250 & @xmath81 0.01 + ngc 2366 & @xmath83 & 8.5pt 6.97 1.4pt @xmath2 0.39@xmath94 & & 0.65 @xmath2 0.41 & -8.0pt 127 & @xmath81 0.01 & & 0.83 @xmath2 0.41 & 7.5pt 97.6 & @xmath81 0.01 + ngc 2366 & @xmath84 & 13.8 5.5pt @xmath2 0.41 & & 1.02 @xmath2 0.43 & -8.0pt 151 & @xmath81 0.01 & & 1.09 @xmath2 0.43 & -3.5pt 141 & @xmath81 0.01 + ngc 2366 & @xmath85 & 4.3pt 5.55 1.3pt @xmath2 0.41 & & 4.5pt 3.52 @xmath2 0.43@xmath94 & 11.5pt 7.94 & 9.5pt 1.67 & & 3.07 @xmath2 0.43 & 7.5pt 10.7 & 9.5pt 0.75 + ddo 154 & & 4.3pt 0.83 1.3pt @xmath2 0.29 & & 0.65 @xmath2 0.30 & 11.5pt 6.71 & 9.5pt 1.67 & & 4.3pt 0.39 @xmath2 0.30@xmath94 & 7.5pt 26.2 & @xmath81 0.01 plots of @xmath0 calculated from the coefficients for ngc 3031 are shown in figure 8 .
included in the plots is the material speed , @xmath1 , and possible locations for 2-arm lindblad resonance .
the material speed is calculated from the rotation curve by de blok et al .
the figure shows consistent plots among the four data sets for each of the different forms of @xmath0 .
the plots for am1 show the same radial behavior as @xmath1 , with smaller values of @xmath0 in the inner half of the disk for the aw96b data set , and outer half of the disk for the yp99 data set .
the solutions using @xmath39 more closely resemble the concave up behavior of @xmath1 in the inner half of the disk whereas solutions using @xmath40 more closely resemble the linear behavior of @xmath1 in the outer half of the disk .
the plots of am1 and am2 for the aw96a , aw96b , and h84 data sets do not show a unique point of corotation resonance . for the same plots the existence of an outer 2-arm lindblad resonance is inconclusive because the radius where @xmath0 intersects possible locations for such a resonance also show corotation .
for both the aw96a and h84 data sets , the plots of am1 and am2 do not show an inner 2-arm lindblad resonance .
the existence of an inner 2-arm lindblad resonance is not ruled out by am2 for the aw96b data set due to the slightly larger confidence bands . the existence or absence of these resonances in plots of am1 and am2
are not as well constrained for the yp99 data set due to the larger confidence bands .
for all four data sets , and both am1 and am2 , the values of @xmath49 are small enough to rule out the null hypothesis that @xmath52 = 0 .
this is convincing evidence that the pattern is shearing .
note that this conclusion is consistent with the results of the @xmath27 test applied to the constant @xmath0 model .
the results for the constant @xmath0 models are therefore estimates of the mean of @xmath0 . for each data set it is not clear whether am1 or am2 provides the better solution , and thus the better estimate of @xmath0 .
their values of @xmath41 are quite similar , especially given their estimated 1@xmath62 uncertainties . despite this ambiguity ,
it is possible to infer from the the above conclusions and the plots of am1 and am2 that there are no clear signs of 2-arm lindblad resonances or a unique location of corotation resonance for the spiral pattern of ngc 3031 .
it is also possible to infer that the speed of the spiral pattern more closely resembles the speed of the material than the speed of a rigidly rotating pattern .
figure 9 shows how an incorrect pa could affect the conclusions for the ngc 3031 data sets . in the figure ( and similar figures for ngc 2366 and ddo 154 )
are values of @xmath16 , @xmath28 , @xmath78 , @xmath45 , and the mean of @xmath0 that result from using a range of pa within @xmath2 2@xmath59 of the assumed pa .
the deviations for the mean of @xmath0 are shown because they are indicative of the amount of deviation in am1 and am2 .
the values of @xmath16 and @xmath78 shown in the figure for am1 and am2 do not vary as much for the aw96a and h84 data sets as they do for the aw96b and yp99 data sets . for all four
data sets the values of @xmath28 are too small to rule out the null hypothesis that the normalized residuals for am1 and am2 have a standard normal distribution , except for the h84 data set at @xmath61pa @xmath37 -1.5@xmath68 .
this means that the am1 and am2 for the h84 data sets may not be the true model if the assumed pa is incorrect by @xmath37 1.5@xmath68 .
determining whether the pattern is shearing , however , does not require that am1 or am2 are the true model .
a @xmath61pa as large as @xmath2 2@xmath59 would not change the results of the @xmath45 test for the aw96a , aw96b and h84 data sets . although some of the am1 @xmath45 values around @xmath61pa @xmath35 1@xmath68 for aw96b appear close to the value at which @xmath49 = 5@xmath95 in the figure , none of the @xmath45 values are actually small enough that @xmath49 @xmath96 1@xmath95 .
there are 6 points for the yp99 data set where that the @xmath45 test is inconclusive .
these occur at @xmath61pa = -0.94@xmath68 for am1 and -0.73@xmath68 @xmath60 @xmath61pa @xmath60 -0.31@xmath68 for am2 . the only point in figure 9 where that the @xmath45 test can not rule out the null hypothesis that @xmath52 = 0 occurs at @xmath61pa = -0.83@xmath68 for the yp99 data set and am1
. two points are worth emphasizing regarding how an incorrect pa could affect the conclusions for ngc 3031 : 1 . )
1.4pc -10.5pt all of the values of @xmath45 shown for three out of four data sets are large enough to conclusively rule out the null hypothesis that @xmath52 = 0 .
for all four data sets , 98.5@xmath95 of the @xmath45 values are large enough to conclusively rule out the null hypothesis . 2 .
) 1.4pc -10.5pt most of the mean @xmath0 results ( 87.7@xmath95 ) are within the 95@xmath95 confidence bands for the mean of @xmath0 obtained using the adopted pa .
the most deviant points occur for the aw96b data set , and these are larger than this data set s mean value for @xmath0 .
such an increase in the @xmath0 results for the aw96b data set would only bring the results for this data set into better agreement with the results for the other data sets
. therefore , the affect of an incorrect pa on the conclusions for ngc 3031 are insignificant .
the affect of an incorrect pa on the conclusions for ngc 1365 were also found to be insignificant in paper i. the robustness of the conclusions to an incorrect pa , in both cases , is attributable to the small uncertainties in the measured values of the pa , and the high quality of the h1.5pt i data .
unlike ngc 3031 , different regions of ngc 2366 produce different results .
results are presented for four different regions : @xmath80 @xmath60 400@xmath70 ,
@xmath80 @xmath60 200@xmath70 , @xmath31400@xmath70 @xmath60 @xmath12 @xmath60 0@xmath70 , and 0@xmath70 @xmath60 @xmath12 @xmath60 400@xmath70 .
these are hereafter referred to as regions @xmath82 , @xmath83 , @xmath84 , and @xmath85 , respectively .
they provide 26 , 14 , 13 , and 13 independent calculations of equation ( 1 ) , respectively . in figure 7
there are prominent peaks in the values of @xmath33 for regions @xmath82 , @xmath83 , and @xmath84 .
the peaks for region @xmath85 are not as prominent .
similar to the plots for ngc 3031 , the plots for all four regions of ngc 2366 show small @xmath33 near @xmath16 = 0 . for regions @xmath82 , @xmath83 , and @xmath84 ,
the values of @xmath33 for the constant @xmath0 model are small enough to rule out the null hypothesis that the normalized residuals have a standard normal distribution .
this is convincing evidence that the constant @xmath0 model is not the true model for these regions . for region @xmath85 ,
@xmath33 for the constant @xmath0 model is too large to rule out the null hypothesis that the normalized residuals have a standard normal distribution .
nothing can be inferred from the @xmath27 test about the constant @xmath0 model in region @xmath85 .
plots of @xmath0 calculated from the coefficients for ngc 2366 are shown in figure 10 .
the material speed in the figure is calculated from the rotation curve by oh et al .
( 2008 ; hereafter o08 ) .
possible locations for 2-arm lindblad resonances are included in the plots ( and also for ddo 154 ) to create the same sense of scale as the plots for ngc 3031 .
all of the regions show consistent plots for the constant @xmath0 model .
all of the regions also show corotation in the inner half of the disk for am1 and am2 .
note that for each region , their plots for am1 and am2 are similar although their values of @xmath16 have opposite sign .
there are clear differences among the different regions for the plots in the outer half of the disk .
these are the most prominent for regions @xmath84 and @xmath85 . in region
@xmath84 the plots for am1 and am2 approach @xmath0 = 0 at @xmath37 400@xmath97 , which is @xmath37 0.16 km
s@xmath64 arcsec@xmath64 less than @xmath1 . in region @xmath85 the results for am1 and am2 are similar to @xmath1 for the entire range in @xmath5 shown .
for all four regions , and both am1 and am2 , the values of @xmath49 are small enough to rule out the null hypothesis that @xmath52 = 0 . for regions
@xmath82 , @xmath83 , and @xmath84 , the results of the @xmath45 test are consistent with the results of the @xmath27 test applied to the constant @xmath0 model .
the results of the @xmath45 test for region @xmath83 , however , are unreliable because the distribution of the normalized residuals for the constant @xmath0 model failed a lilliefors test for normality ( @xmath44 = 1.79% ) .
the @xmath45 test result for am1 in region @xmath85 may also be unreliable because the results of a lllliefors test applied to its normalized residuals is inconclusive ( @xmath44 = 5.35% ) . from the results for the @xmath45 test ,
the results of the @xmath27 test for the constant @xmath0 model , and the plots of @xmath0 , it is possible to infer that the clumps of h1.5pt i in ngc 2366 are corotating with the material in the inner half of the disk , and that there are clumps of h1.5pt i in the outer half of the disk in region @xmath84 that are rotating more slowly than the rest of the material . the results for ngc 2366 show that the solution method can detect different values of @xmath0 in different regions of a galaxy .
this demonstrates the usefulness of the solution method for determining if there is a single , global pattern speed .
the difference for regions @xmath84 and @xmath85 may be attributed to the non - circular motions in the velocity field that are present in region @xmath84 .
these are easily identified in figure 2 by the kinks in the contours of constant @xmath8 around @xmath12 @xmath37 200@xmath97 - 400@xmath97 .
the non - circular motions were removed from the velocity field when o08 derived the rotation curve used in this paper for calculating @xmath1 .
the difference between the rotation curve of o08 , and one obtained without the non - circular motions removed ( see figure 13 of o08 ) , is not large enough to account for the difference shown for @xmath1 and @xmath0 in panels ( h ) and ( i ) of figure 10 .
the solution method is clearly detecting something other than @xmath0 @xmath98 @xmath1 .
figure 11 shows how an incorrect pa could affect the conclusions for the different regions of ngc 2366 .
for all four regions , the values of @xmath16 do not change significantly , except for am1 at @xmath61pa @xmath35 -4@xmath68 .
almost all of the values of @xmath28 that are shown are too small to rule out the null hypothesis that the normalized residuals for am1 and am2 have a standard normal distribution , consistent with the values for these models in table 3 .
the values of @xmath78 for regions @xmath82 , @xmath83 , and @xmath84 , are fairly consistent , especially given their uncertainties .
the values of @xmath45 are all quite large for @xmath82 and @xmath84 .
the smallest values of @xmath45 occur for region @xmath85 .
note that for all four regions , the values shown for the mean of @xmath0 are within the 95% confidence bands for the value obtained using the adopted pa .
it is reasonable to conclude that the accuracy of the adopted pa is not likely to significantly affect any of the conclusions for ngc 2366 .
similar to ngc 3031 , different regions : @xmath80 @xmath60 400@xmath70 , @xmath80 @xmath60 200@xmath70 , @xmath31400@xmath70 @xmath60 @xmath12 @xmath60 0@xmath70 , and 0@xmath70 @xmath60 @xmath12 @xmath60 400@xmath70 , produce consistent results .
therefore , only the results for most of the disk , @xmath80 @xmath60 400@xmath70 , are presented .
this region provides 24 independent calculations of equation ( 1 ) .
the peaks in the distribution of @xmath33 shown in figure 7 are not as narrow as they are in most of the other plots .
similar to most of the other plots , however , @xmath33 is small near @xmath16 = 0 .
note that the value of @xmath33 for the constant @xmath0 model is the largest of the values of @xmath33 shown in table 3 for the constant @xmath0 model .
it is too large to rule out the null hypothesis that the distribution of the normalized residuals is a standard normal distribution for the constant @xmath0 model .
plots of @xmath0 calculated from the coefficients are shown in figure 12 .
the material speed in the figure is calculated from the rotation curve by oh et al .
all three models of @xmath0 are quite similar in the outer half of the disk .
the plots for the constant @xmath0 model and am2 are consistent with each other to within their 95% confidence bands . the plot for am1 differs from the other two by showing corotation in the central part of the galaxy .
none of the plots show corotation at @xmath37 100@xmath99 .
the values of @xmath49 in table 5 for am1 and am2 are small enough to rule out the null hypothesis that @xmath52 = 0 .
this result for am2 may be unreliable because the results of a lllliefors test applied to its normalized residuals is inconclusive ( @xmath44 = 4.62% ) .
also note that the solution for this model has a value of @xmath78 that is smaller than 1 by more than 2@xmath62 . from the results for the @xmath45 test , the results of the @xmath27 test for the constant @xmath0 model , and the plots of @xmath0 , the clumps of h1.5pt i in ddo 154 are corotating in the central and outer parts of the disk .
some of the clumps of h1.5pt i at @xmath37 100@xmath99 are rotating more slowly than the material .
the result for the constant @xmath0 model is the mean of @xmath0 .
figure 13 shows how an incorrect pa could affect the conclusions for ddo 154 .
the results for ddo 154 are much more dependent on the value of the adopted pa than are the results for ngc 3031 or ngc 2366 .
the values of @xmath40 vary by as much as a factor of 6 .
the values of @xmath78 vary by almost as much , with the smallest values occurring around @xmath61pa @xmath37 0@xmath68 .
there are values of @xmath61pa where that the results of the @xmath45 test are inconclusive , or that @xmath45 is too small to rule out the null hypothesis that @xmath52 = 0 .
the results for the constant @xmath0 model deviates quite significantly beyond the 95% confidence intervals for the value obtained using the adopted pa .
it must be concluded that the accuracy of the adopted pa affects some of the conclusions for ddo 154 , particularly whether the solution using a constant @xmath0 model , or one of the alternative models , provides a better solution for equation ( 1 ) .
the previous section shows the results from applying the solution method to data sets with different angular resolutions , spectral resolutions , and sensitivities to structures on different scales .
results are also shown for galaxies without coherent or well - organized patterns . knowing how these properties of a data set affect the results is important for choosing data in future applications of the solution method .
of all the differences in the data sets , the angular resolution had the most noticeable affect on the results by setting the number of independent calculations of equation ( 1 ) .
the most extreme cases can be seen in figure 8 for the ngc 3031 data sets .
the results for the yp99 data set demonstrate that a constant @xmath0 , or the mean if there is evidence for a radial dependence , can be found with only 4 independent calculations of equation ( 1 ) , but more calculations would be desirable in order to reduce the uncertainty . in order to determine whether there is a single , global pattern speed ( i.e.
, @xmath0 is the same in different regions of a galaxy ) , a number of independent calculations similar to what was used for aw96b and ngc 2366 should be sufficient , given similar uncertainties .
there is no noticeable affect on the results due to spectral resolution .
consistent results were obtained for the aw96a and h84 data sets despite a factor of @xmath37 4 difference in the channel separation . also consider that results with small uncertainties were obtained in paper i for ngc 1365 using maps made from a 20.84 km s@xmath64 channel separation .
the large channel separation for the ngc 1365 data may have been compensated for by having 62 independent calculations of equation ( 1 ) for the whole disk .
the results for the three different data sets of ngc 3031 are remarkably consistent despite their different sensitivities to structures on different scales .
the results for the aw96a and h84 data sets are almost identical even though they are more sensitive to small and large scales , respectively , and the aw96a data set is missing some flux .
the results for the data sets of ngc 3031 with the largest angular resolution , aw96b and yp99 , are slightly less than the results for aw96a and h84 .
the smaller number of independent calculations of equation ( 1 ) for the aw96b and yp99 data sets , and consequently larger uncertainties , may be the cause for this small difference .
the results for ngc 2366 and ddo 154 demonstrate that the solution method can produce meaningful results that are simple to interpretin the absence of a coherent or well - organized pattern .
this shows that the solution method is not limited to grand - design patterns .
it could be used to find @xmath0 for multiple armed and flocculent spiral galaxies as well . for these types of spiral galaxies
the solution method would be measuring @xmath0 for different arms , or clumps of higher density material .
although this may violate the assumption of a well - defined pattern speed by tw84 ( i.e. , the surface brightness is assumed to be constant in a frame rotating at an angular speed @xmath0 ) , the tw84 equations are generalizable to a frame that is rotating at a variable angular speed @xmath0(@xmath5 ) ( engstrm 1994 ) . whether the results for galaxies without a coherent or well - organized pattern have any physical meaning
can be determined using a variety of statistical tools such as @xmath41 and the lilliefor s test used in this paper . a very large or small value of @xmath41 , or a small enough value of @xmath100 for the normalized residuals , could mean that equation ( 1 ) is a poor description of the dynamics of a galaxy , or part of a galaxy , for example .
this paper shows that having a sufficient number of independent calculations is important for applying the solution method .
this limits the use of h1.5pt i radio synthesis data to nearby galaxies .
for example , the resolution of the c - array of the vla for h1.5pt i observing , which has a typical fwhm of @xmath37 13@xmath70 for the synthesized beam , would be adequate for many nearby spiral galaxies .
radio synthesis data for tracers of molecular gas , which have a higher frequency , and thus a smaller synthesized beam ( typically as small as sub arcsecond to a few arcseconds at @xmath37 100 ghz ) , would be useful for applying the method to more distant galaxies that have a smaller angular size .
the resolution of optical data , such as fabry - perot h@xmath101 data , is typically @xmath37 1@xmath70 - 2@xmath70 due to atmospheric seeing .
such data would also expand the applicability of the solution method to galaxies with a smaller angular size .
there is convincing evidence in the form of the @xmath45 test that the spiral pattern in ngc 3031 is shearing .
in fact , the values of @xmath45 are quite large .
this confirms the discovery of shear in the pattern by westpfahl ( 1998 ) .
the pattern speed is very similar to the material speed , and there are no clear indications of unique corotation or lindblad resonances .
the same conclusions were found for the spiral pattern of ngc 1365 .
note that the results of the @xmath45 test are complimented by the results of the @xmath27 test for the constant @xmath0 model , from which it can be inferred that the constant @xmath0 model is not the true model .
it is important to point out that these findings for both ngc 3031 and ngc 1365 are based on calculations that are independent of the assumptions made in spiral structure theories .
the results for ngc 3031 are similar to the sheared gravitational instabilities of goldreich & lynden - bell ( 1965 ) and the swing amplifier of toomre ( 1981 ) .
close passage of the companion galaxies ngc 3034 and ngc 3077 could have caused the swing amplification in ngc 3031 .
several models agree that these galaxies passed to within 20 kpc of ngc 3031 about 600 myr ago ( van der hulst 1977 ; cottrell 1976 , 1977 ; gottesman & weliachew 1975 , 1977 ; and brouillet et al .
simulations show that such interactions can produce grand design patterns that shear over time ( s. oh 2008 ; dobbs et al .
2010 ; struck et al .
this implies that the spiral pattern in ngc 3031 is a transient feature that is driven by interactions with its companions , and in this way the winding dilemma is avoided . the results in this paper and in paper i are part of a growing body of evidence that some grand - design spiral patterns are shearing , transient features .
some of this evidence is coming from simulations , which have yet to produce persisting patterns ( e.g. , berman et al .
1978 ; sellwood & carlberg 1984 ; sorensen 1985 ; bottema 2003 ; dobbs & bonnell 2008 ; s. oh et al . 2008
; dobbs et al . 2010 ; struck et al .
2011 ; wada et al . 2011
; sellwood 2011 for a review ) . of the more recent simulations , wada et al .
( 2011 ) showed that the shearing , short lived ( @xmath37 100 myr ) , recurring spiral arms in their simulations of a live disk have a pattern speed that is similar to the material speed .
simulations are useful tools for forming hypothesis about spiral structure , but a true understanding must ultimately come from observations . in paper
i it was pointed out that the original solution method of tw84 produced results for at least 7 nearby spiral galaxies that are consistent with a shearing pattern .
foyle et al . ( 2011 ) also found evidence against long - lived , rigidly rotating patterns in spiral galaxies .
they showed that the angular offsets between star formation tracers in 11 nearby spiral galaxies , including ngc 3031 , are not systemically ordered as predicted from a long - lived , rigidly rotating pattern .
it is difficult to reconcile the findings in this paper for ngc 3031 with the density wave theory .
all versions of the theory use as a foundation the premise that the pattern is rotating approximately rigidly ( bertin & lin 1996 , chapter 2 ; bertin 2000 , chapter 15 ; binney & tremaine 2008 , chapter 6 ) .
one possibility is that @xmath0 is a step function , as advocated for by meidt et al .
( 2008 ) and meidt et al .
the solution method used in this paper would not be able to detect such step functions because continuity is assumed for am1 and am2 .
the forms used in paper i , however , are capable of detecting such step functions , and did not find evidence for this form of @xmath0 in the spiral pattern of ngc 1365 .
the same forms of @xmath0 used in paper i were also tried for the ngc 3031 data sets , and there is no sign of a step function .
most importantly , note that there are no discontinuities in the spiral pattern of ngc 3031 as one would expect if @xmath0 is indeed a step function .
in this paper the method derived in paper i for solving the pattern speed equations of tw84 is applied to ngc 3031 , ngc 2366 , and ddo 154 in order to show how different properties of a data set affect the results .
the conclusions are as follows : -10.5pt four different data sets of ngc 3031 produce consistent results despite differences in angular resolution , spectral resolution , and sensitivities to structures on different scales .
they all produce convincing evidence for shear in the pattern .
the pattern speed is more similar to the material speed than it is to the speed of a rigidly rotating pattern .
there are no unique points of corotation or 2-arm lindblad resonances .
-10.5pt the angular resolution had the most noticeable affect on the results by setting the number of independent calculations of equation ( 1 ) .
this is important to consider when selecting data for future 0.1pt applications of the solution method , especially for determining the radial behavior of @xmath0 , and whether there is a single , global pattern speed .
lower angular resolution data sets give few independent rows of data , so are less useful for determining @xmath0 as a function of radius . -10.5pt
the results for ngc 2366 and ddo 154 show that meaningful results that are simple to interpret can be obtained even in the absence of a coherent or well organized pattern .
the results for these galaxies also demonstrate that the solution method is not merely measuring the material speed , and thus serve as a counterexample to the results for ngc 3031 .
the authors acknowledge the helpful comments of the referee that improved this paper .
the authors also acknowledge erwin de blok and se - heon oh for providing the rotation curve data and the uncertainties of the orientation parameters for ngc 3031 , ngc 2366 , and ddo 154 .
renee andrae , tim shulze - hartung , and peter melchoir provided some of the inspiration for the procedure used in this paper for finding models that are an alternative to a constant @xmath0 .
this research made use of the national radio astronomy observatory which is a facility of the national science foundation operated under cooperative agreement by associated universities , inc ; and the nasa / ipac extragalactic database which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
adler , d. s. & westpfahl , d. j. 1996 , aj , 111 , 735 ( aw96 ) berman , r. h. , brownrigg , d. r. k. , & hockney , r. w. 1978 , mnras , 185 , 861 bertin , g. 2000 , dynamics of galaxies ( cambridge : cambridge university press ) bertin , g. , & lin , c. c. 1996 , spiral structure in galaxies : a density wave theory ( cambridge , ma : mit press ) bevington , p. r. , & robinson , d. k. 2002 , data reduction and error analysis for the physical sciences ( 3rd ed . ; new york : mcgraw - hill ) binney , j. & tremaine , s. 2008 , galactic dynamics , 2nd ed .
( princeton , nj : princeton univ . press ) de blok , w. j. g. , walter , f. , brinks , e. , trachternach , c. , oh , s. -h .
, & kennicut , r. c. 2008 , aj , 136 , 2648 bottema , r. 2003 , mnras , 344 , 358 brouillet , n. , baudry , a. , combes , f. , kaufman , m. , & bash , f. 1991 , a&a , 242 , 35 cottrell , g. a. 1976 , mnras , 174 , 455 cottrell , g. a. 1977 , mnras , 178 , 577 debattista , v. p. 2003 , mnras , 342 , 1194 debattista , v. p. , & williams , t. b. 2004 , apj , 605 , 714 dobbs , c. l. , & bonnell , i. a. 2008 , mnras , 385 , 1893 dobbs , c. l. , theis , c. , pringle , j. e. , & bate , m. r. 2010 , mnras , 403 , 625 foyle , k. , rix , h. -w . , dobbs , c. l. , leroy , a. k. , & walter , f. 2011 , 735 , 101 goldreich , p. , & lynden - bell , d. 1965 , mnras , 130 , 125 gottesman , s. t. , & weliachew , l. 1975 , apj , 195 , 23 gottesman , s. t. , & weliachew , l. 1977 , apj , 211 , 47 hine , b. p. a. 1984 , a high resolution neutral hydrogen study of m81 , m.a .
thesis , university of texas ( h84 ) van der hulst , j. m. 1977 , ph.d .
thesis , univ . of groningen
kendall , m. g. , & stuart , a. 1973 , the advanced theory of statistics : inference and relationship ( 3rd ed . ; new york : hafner ) lin , c. c. , & shu , f. h. 1964 , apj , 140 , 646 lin , c. c. , & shu , f. h. 1966 , proc .
usa , 55 , 229 lin , c. c. , yuan , c. , & shu , f. h. 1969 , 155 , 721 lupton , r. 1993 , statistics in theory and practice ( princeton : princeton univ .
press ) madansky , a. 1988 , prescriptions for working statisticians ( new york : springer verlag ) meidt , s. , rand , r. j. , merrifield , m. , shetty , r. , & vogel , s. n. 2008 , apj , 688 , 224 meidt , s. , rand , & r. j. , merrifield , m. 2009 , apj , 702 , 290 oh , s. h. , kim , w. -t . , & lee , h. m. 2008 , apj , 683 , 94 oh , s. -h . , de blok , w. j. g. , walter , f. , brinks , e. , & kennicut , r. c. 2008 , aj , 136 , 2761 ( o08 ) oh , s. -h . , de blok , w. j. g. , brinks , e. , walter , f. , & kennicut , r. c. 2011 , aj , 141 , 93 rand , r. j. , & wallin , j. f. 2004 , apj , 614 , 142 savitzky , a. , & golay , m. j. e. 1964 , anal .
, 36 , 1627 sellwood , j. a. 1993 , pasp , 105 , 648 sellwood , j. a. 2011 , mnras , 410 , 1637 sellwood , j. a. , & carlberg r , g. 1984 , apj , 282 , 61 sellwodd , j. a. , & sparke , l. s. 1988 , mnras , 231 , 25p sorensen , s. -a .
1985 , mnras , 212 , 723 speights , j. c. , & westpfahl , d. j. 2011 , apj , 734 , 70 ( paper i ) struck , c. , dobbs , c. l. , & hwang , j .- s .
2011 , mnras , 414 , 2498 toomre , a. 1981 , in the structure and evolution of normal galaxies , ed .
s. m. fall & d. lynden - bell ( cambridge : cambridge university press ) , 111 trachternach , c. , de blok , w. j. g. , walter , f. , brinks , e. , & kennicut , r. c. 2008 , aj , 136 , 2720 tremaine , s. , & weinberg , m. d. 1984 , apjl , 282 , l5 ( tw84 ) wada , k. , baba , j. , & saitoh , t. r. 2011 , apj , 735 , 1 walter , f. , brinks , e. , de blok , w. j. g. , bigiel , f. , kennicutt , r. c. , thornley , m. d. , & leroy , a. 2008 , aj , 136 , 2563 westpfahl , d. j. 1998 , apjs , 115 , 203 ( w98 ) yun , m. s. & purton , c. 1999 , private communication ( yp99 ) | the pattern speeds of ngc 3031 , ngc 2366 , and ddo 154 are measured using a solution of the tremaine - weinberg equations derived in a previous paper .
four different data sets of ngc 3031 produce consistent results despite differences in angular resolution , spectral resolution , and sensitivities to structures on different scales .
the results for ngc 3031 show that the pattern speed is more similar to the material speed than it is to the speed of a rigidly rotating pattern , and that there are no clear indications of unique corotation or lindblad resonances .
unlike ngc 3031 , the results for ngc 2366 and ddo 154 show clear departures from the material speed .
the results for ngc 2366 and ddo 154 also show that the solution method can produce meaningful results that are simple to interpret even if there is not a coherent or well - defined pattern in the data .
the angular resolution of a data set has the greatest affect on the results , especially for determining the radial behavior of the pattern speed , and whether there is a single , global pattern speed . |
a square identity of size @xmath1 $ ] is an identity @xmath2 where @xmath3 and @xmath4 are independent indeterminates and each @xmath5 is a bilinear form in @xmath6 and @xmath7 with integer coefficients .
the problem to determine all triples of positive integers @xmath8 for which there exists a square identity of size @xmath1 $ ] was formulated by hurwitz @xcite and remains widely open .
such triples are called _
admissible_. an admissible triple @xmath1 $ ] is _ optimal _ if @xmath9 and @xmath10 can not be increased and @xmath11 can not be decreased .
this problem has various applications and interpretations in number theory , linear algebra , geometry and topology , see @xcite,@xcite for a complete overview .
the simplest example is called brahmagupta s two - square identity @xmath12 that can be easily proved algebraically , but can also be understood as equality between the products of the norms of two complex numbers and the norm of their product : @xmath13 the similar equalities in the algebras of quaternions @xmath14 and of octonions @xmath15 coincide with euler s four - square identity ( 1748 ) and degen s eight - square identity ( 1818 ) , respectively .
we refer to @xcite for the history of these identities and their meaning . in 1898
, hurwitz proved his celebrated theorem @xcite stating that @xmath11-square identities exist only for @xmath16 and @xmath17 .
actually , hurwitz proved a stronger `` non - existence '' result in the case where @xmath5 are bilinear forms with complex coefficients . in 1918
( published in 1922 ) , hurwitz @xcite generalized his theorem and established the largest integer @xmath18 for which there exists a @xmath19$]-identity with complex coefficients . in 1922 , radon @xcite independently solved a similar problem in the case of real coefficients .
the result is the same in both cases : write @xmath11 in a form @xmath20 , then @xmath21 the function @xmath22 is called the _ hurwitz - radon function _
( this function also appears in topology @xcite and linear algebra @xcite ) . note that , hurwitz s constructive proof leads to @xmath19$]-identities with integer coefficients . according to @xcite ,
this even implies the existence of such identities with coefficients in @xmath23 .
a nice feature of the hurwitz - radon identities is their relation to clifford algebras .
every square identity of size @xmath24 $ ] defines a representation of the real clifford algebra @xmath25 by @xmath26-real matrices . in this representation
the matrices corresponding to the generators of @xmath25 are orthogonal and skew - symmetric matrices with integer coefficients .
conversely , every such representation of @xmath25 defines an @xmath24$]-identity . in his original work @xcite ,
hurwitz implicitly constructed clifford algebras and their representations .
however , this was formulated in an explicit and conceptual way much later ( see @xcite for more details ) . besides the hurwitz - radon formula , a number of solutions to the hurwitz problem are known .
we refer to @xcite,@xcite,@xcite,@xcite,@xcite,@xcite , for concrete examples of admissible triples @xmath1 $ ] and various methods to prove their optimality
. the complete list of the admissible triples with @xmath27 is given in @xcite .
in @xcite and @xcite , yuzvinsky introduced a new approach to construct square identities , see also @xcite .
his method is based on _ monomial pairings _ on the group algebra over @xmath0 .
in particular , existence of infinite series of identities of size @xmath28,\ ] ] where the function @xmath29 is given by the binomial coefficients @xmath30 was established .
the above values of @xmath1 $ ] are called yuzvinsky - lam - smith formulas ( see @xcite ) .
another series of triples : @xmath31\ ] ] can be found in @xcite . to the best of our knowledge , the hurwitz - radon formula , the expressions ( [ yuzeq ] ) and ( [ angel ] ) as well as their direct consequences , are the only known infinite series of solutions to the hurwitz problem . in the present paper
, we obtain infinite families of admissible triples @xmath1 $ ] similar to .
our method is similar to the yuzvinsky method , but the monomial pairings we consider are different .
we use a series of algebras constructed in the recent work @xcite .
these algebras generalize the classical algebra of octonions and clifford algebras at the same time . as an application of these algebras ,
explicit expressions for the hurwitz - radon identities were provided in @xcite . here
we extend the method to obtain more general identities . as well as
, the triples that we obtain are close to the hurwitz - radon triple .
whereas is obtained from @xmath32 $ ] by increasing the first entry , @xmath33 , we will decrease the third entry , @xmath34 .
let us also mention that our formulas include as the simplest case .
let us also mention that optimality of @xmath1 $ ] is often a difficult problem .
for instance , it is not known whether ( [ yuzeq ] ) is optimal .
we show that our formulas generate several optimal ( and several best known ) values @xmath1 $ ] .
however , unfortunately we do not have any result on optimality of our formulas .
the paper is organized as follows . in section [ mainrs ]
, we formulate the main results . in section [ twis ]
, we introduce the general notion of twisted group algebra over the abelian group @xmath0 . in section [ neshalgsec ]
, we recall the definition of the algebras @xmath35 and introduce the notion of hurwitzian subset of @xmath0 .
proofs of the main results are given in section [ proofsec ] .
the first family of admissible triples that we obtain depends on two parameters , @xmath36 and @xmath37 .
[ firstmain ] ( i ) for every @xmath38 , there exist square identities of size @xmath1 $ ] with @xmath39 s&=&2^n-2\left({k-1\choose2}+\ell+1\right)n + 4\,{k\choose3}+2\,k\,\ell,\\[12pt ] n&=&2^n-2\left({k-1\choose2}+\ell+1\right ) , \end{array } \right.\ ] ] for all @xmath40 .
\(ii ) if @xmath41 , then there are also identities of size @xmath42 $ ] , where @xmath8 are as in . for small values of @xmath36 and @xmath37 , formula gives the following triples : @xmath43 , \qquad \left[2n,\,2^n-6n+10,\,2^n-6 \right ] , \qquad \left[2n,\,2^n-8n+16,\,2^n-8 \right].\ ] ] we conjecture that the values ( [ step ] ) are optimal if @xmath38 is sufficiently large .
let us give a few examples of concrete numeric values .
\(a ) for @xmath44 , we obtain the following admissible triples : @xmath45 $ ] , @xmath46 $ ] ; the latter is optimal ( as proved in @xcite ) .
\(b ) for @xmath47 , we obtain @xmath48,\;[12,44,60]$ ] that correspond to the best known values , ( cf , @xcite , pp.292293 ) .
\(c ) for @xmath49 , we obtain the values @xmath50,\;[14,96,122],\;[14,104,124]$ ] from ( [ step ] ) .
furthermore , part ( ii ) of the theorem leads to @xmath51 $ ] , @xmath52 $ ] , @xmath53 $ ] .
\(d ) for @xmath54 , we obtain @xmath55 $ ] , @xmath56 $ ] .
the second family of identities is parametrized by an integer @xmath57 .
the values of @xmath10 and @xmath11 are given in terms of sums of binomial coefficients .
[ secmain ] ( i ) if @xmath58 , then there exist square identities of size @xmath1 $ ] , where @xmath59 where @xmath60 and @xmath61 .
\(ii ) if @xmath62 , then there exist square identities of size @xmath63 $ ] , where @xmath64 and @xmath11 are as in . let us give the first and simplest example . in the case
@xmath65 , the formula ( [ stepdva ] ) reads @xmath66\ ; , \qquad \big [ 2n+2\ ; , \ ; 2{n\choose m}\ ; , \ ;
2{{n+1}\choose m}\big],\ ] ] for @xmath58 and @xmath62 , respectively .
we obtain the values @xmath67\;,\ ; [ 16,70,112]\ ; , [ 18,252,420]$ ] .
the second one strengthens the known triple @xmath68 $ ] , see @xcite , p.294 .
in this section , we recall the definition of twisted group algebras over the abelian group @xmath0 . these algebras include clifford algebras and the algebra of octonions .
we then explain yuzvinsky s idea @xcite of construction of square identities out of a twisted group algebra .
consider the group algebra @xmath70 $ ] over the abelian group @xmath0 .
denote by @xmath71 the natural basis vectors indexed by elements of the group @xmath72 .
one then has : @xmath73=\bigoplus_{x\in { ( { \mathbb{z}}/2{\mathbb{z}})^n}}{\mathbb{r}}\ , u_x.\ ] ] given a function @xmath74 , satisfying @xmath75 for all @xmath76 , one defines a twisted product on @xmath70 $ ] @xmath77 for all @xmath78 .
the defined algebra is called a _ twisted group algebra _ over @xmath69 and denote by @xmath79,f)$ ] .
the basis vector @xmath80 is a unit in @xmath79,f)$ ] .
recall that the clifford algebra @xmath81 is the associative algebra with @xmath38 generators @xmath82 and relations @xmath83 write elements of @xmath0 in the form @xmath84 , where @xmath85 , and choose the function @xmath86 it turns out that the twisted group algebra @xmath79,f_{{\mathit{c}\ell}})$ ] is isomorphic to the clifford algebra @xmath81 , see @xcite for the details .
for instance , if @xmath87 , then this is nothing but the algebra of quaternions @xmath14 .
the algebra @xmath79,f)$ ] is associative if and only if @xmath88 is a 2-cocycle on the group @xmath0 , i.e. , @xmath89 for all @xmath90 .
this is obviously the case for the bilinear function ( [ cliffunet ] ) .
the following example realizes the classical non - associative algebra of octonions @xmath15 as a twisted group algebra with a cubic twisting function @xmath88 ( that is not a 2-cocycle ) .
this result was found in @xcite .
[ exo ] the twisted group algebra @xmath91,f)$ ] defined by the function @xmath92 is isomorphic to the algebra of octonions @xmath15 . in this subsection , we explain the yuzvinsky original idea ( see @xcite and also @xcite ) .
let us consider a twisted group algebra @xmath79,f)$ ] .
we define the euclidean norm : @xmath93 for @xmath94 , where @xmath95 .
recall that a composition algebra is a normed algebra such that , for any two elements @xmath96 and @xmath97 , @xmath98 it is well - known ( and follows for instance from hurwitz s theorem ) that there are no composition algebras of dimension greater than @xmath17 .
therefore , if @xmath99 , then the condition ( [ normprod ] ) can not be satisfied for arbitrary elements @xmath96 and @xmath97 of the twisted group algebra @xmath79,f)$ ] . in order to find square identities , we will look for subsets @xmath100 of @xmath0 , such that the condition ( [ normprod ] ) is satisfied for @xmath101 given two subsets @xmath102 , we say that @xmath103 is a _ multiplicative pair _ if the condition ( [ normprod ] ) holds for any two elements of the form @xmath104 note that this definition depends on the twisting function @xmath88
. moreover , existence of multiplicative pairs is an important property of @xmath88 .
recall that the _ sumset _ , @xmath105 , of subsets @xmath106 and @xmath107 in @xmath69 is the set formed by pairwise sums of the elements of @xmath106 and @xmath107 , i.e. @xmath108 [ sumprop ] if @xmath103 is a multiplicative pair , then there exists a square identity of size @xmath109=[{\textup{card}}{}(a),\,{\textup{card}}{}(b),\,{\textup{card}}{}(a+b)],\ ] ] which can be written explicitly as @xmath110 where @xmath111 this is an immediate consequence of the property . the following important statement is proved in @xcite .
it provides a criterion for subsets @xmath106 and @xmath107 to be a multiplicative pair and , therefore , it also proves the existence of corresponding square identities .
the proof is elementary and we give it here for the sake of completeness . [ propmult ] let @xmath106 and @xmath107 be two subsets of @xmath0 .
the following two conditions are equivalent 1 .
the pair @xmath103 is multiplicative .
2 . for all @xmath112 and @xmath113 such that @xmath114 , one has @xmath115 the product of the euclidean norms is obviously @xmath116 on the other hand , one obtains : @xmath117
it follows that the condition ( [ normprod ] ) is satisfied if and only if the terms @xmath118 and @xmath119 cancel whenever @xmath120 and @xmath121 , in other words , the corresponding signs are opposite . hence the result .
our next goal is to find a twisted group algebra @xmath79,f)$ ] that admits multiplicative pairs of large cardinality .
in this section , we apply the results of section [ twis ] to a particular series of twisted group algebras .
we recall the definition and the main properties of the algebras @xmath35 generalizing the algebra of octonions .
this section presents a brief account , the details can be found in @xcite . the algebra @xmath35 is the twisted group algebra defined by the function @xmath122 for all @xmath123 and @xmath124 , elements of @xmath0 .
note that @xmath125 is exactly the algebra of octonions @xmath15 , cf .
example [ exo ] .
the main property of the algebra @xmath35 is that its structure is determined ( up to isomorphism ) by a single function in one variable @xmath126 that we call the _ generating function_. given a twisted group algebra @xmath79,f)$ ] , we say that this algebra has a generating function if there exists a function @xmath127 such that 1 . for all @xmath128 ,
the following equation is satisfied : @xmath129 2 . for all @xmath130 , @xmath131 & = \a(x+y+z)+\a(x+y)+\a(x+z)+\a(y+z)+\a(x)+\a(y)+\a(z ) .
\end{array}\ ] ] the above conditions ( i ) and ( ii ) have cohomological meaning .
in particular the expression on the right - hand side of ( i ) and the left - hand side of ( ii ) are the differentials @xmath132 and @xmath133 , respectively .
the idea of existence of a generating function goes back to eilenberg and maclane ( see @xcite ) ; this idea is also crucial for the theory of error - correcting codes and code loops ( see @xcite ) . one can immediately check the following statement .
@xcite [ propalpha ] ( i ) the function @xmath134 for @xmath76 , is a generating function for the algebra @xmath35 .
\(ii ) for @xmath135 and @xmath114 , one has @xmath136
\(iii ) the function @xmath137 depends only on the weight of the element @xmath138 , i.e. , on the integer @xmath139 note that equation is equivalent to ( [ fcond ] ) .
the algebra @xmath35 is the unique twisted group algebra over @xmath0 that admits a generating function @xmath140 which is invariant under the action of the group of permutations @xmath141 on @xmath0 , i.e. , the action permuting the coordinates of @xmath138 . by ( iii ) of proposition
[ propalpha ] , the value @xmath137 only depends on the weight @xmath142 .
it turns out that @xmath140 is 4-periodic .
@xmath143 1\ ; , & \text { otherwise . } \end{array } \right.\ ] ] we use the function @xmath140 to give a combinatorial characterization of multiplicative pairs for the algebra @xmath35 .
a pair @xmath103 of subsets of @xmath0 is a multiplicative pair if and only if the following condition is satisfied .
if @xmath144 and @xmath145 , then @xmath146 .
the condition @xmath147 means @xmath148 for some elements @xmath149 such that @xmath150 .
similarly , @xmath151 , where @xmath152 . by and ,
multiplicativity of a pair @xmath103 is equivalent to @xmath153 for all @xmath149 and @xmath154 such that @xmath155 . since @xmath140 is 0 only for elements of weight multiple of 4 , the lemma now follows .
a subset @xmath156 is called a _
hurwitzian set _
if @xmath157 forms a multiplicative pair with @xmath0 , and @xmath158 is the greatest possible with this condition .
let us fix the following notation for some particular elements of @xmath0 : @xmath159 \overline{e_0}&:= & ( 1,1,\ldots,1),\\[4pt ] e_i&:=&(0,\ldots,0,1,0,\ldots,0 ) , \text { where 1 occurs at the $ i$-th position,}\\[4pt ] \overline{e_i}&:=&(1,\ldots,1,0,1,\ldots,1 ) , \text { where 0 occurs at the $ i$-th position}. \end{array}\ ] ] the following table provides a possible choice of hurwitzian sets and the size of the identities obtained by applying proposition [ sumprop ] , depending on the class of @xmath160 .
@xmath161 \hline n\equiv 0 \mod 4 & h=\{e_i , e_1+e_i\}&[2n , 2^n , 2^n]\\[6pt ] \hline n\equiv 1 \mod 4 & h=\ { e_i , \overline{e_i}\ } & [ 2n , 2^n , 2^n ] \\[6pt ] \hline n\equiv 2 \mod 4 & h=\{e_i , e_1+e_i\}&[2n , 2^n , 2^n]\\[6pt ] \hline n\equiv 3 \mod 4 & h=\{e_0 , \overline{e_0},\ , e_i , \overline{e_i}\}&[2n+2 , 2^n , 2^n]\\[6pt ] \hline \end{array}\ ] ] where @xmath162 .
indeed , it is easy to check that @xmath163 is never a multiple of 4 , when @xmath164 . therefore the condition ( [ genfuncfor ] ) is always satisfied . for @xmath165 or @xmath166 ,
the triples in the table are the optimal triples given by hurwitz - radon theorem .
when @xmath167 , we do not achieve the optimal @xmath168 $ ] .
this is related to the fact that for @xmath167 , the algebra @xmath35 is not simple .
in this section , we prove theorems [ firstmain ] and [ secmain ] .
the proofs are based on proposition [ sumprop ] .
the idea is to start with a multiplicative pair @xmath169 , where @xmath170 , and reduce the size of @xmath171 by removing elements from @xmath107 .
we first give some examples .
\1 ) for @xmath172 , the sumsets are as follows : @xmath173 for @xmath174 and @xmath175 , respectively .
this leads to the formula .
furthermore , in the case @xmath62 , one obtains an identity of size @xmath176.\ ] ] \2 ) assume @xmath38 is odd and let @xmath177 the sumset is then of the form : @xmath178 this leads to identities of size @xmath179,\ ] ] which corresponds to the second `` border case '' in theorem [ firstmain ] , namely to formula ( [ step ] ) with @xmath180 .
\3 ) again , let @xmath38 be odd and write @xmath181
. set @xmath182 then @xmath183 , see figure [ firstf ] .
this leads to identities of size @xmath184\;,\qquad \left [ 2n+2\ ; , \ ; 2{n\choose m}\ ; , \ ; 2{{n+1}\choose m } \right],\ ] ] if @xmath58 and @xmath62 , respectively .
this is theorem [ secmain ] , in the particular case @xmath65 .
set @xmath188 and @xmath189 . in order to eliminate two elements @xmath190 and
@xmath191 from @xmath171 , we have to remove from @xmath107 the following set : @xmath192 notice however , that @xmath193 and @xmath194 coincide and their intersection with @xmath157 is @xmath195 .
thus the cardinality of the removed set is @xmath196 .
finally , we fix @xmath197 with @xmath198 and let @xmath107 be the following set @xmath199 where @xmath200 is the lexicographical order .
it can be easily proved by induction that the removed set is of cardinality @xmath201 we write @xmath181 .
fix @xmath37 such that @xmath208 and set @xmath209 the sumset is as follows ( see figure [ 5and6 ] ) @xmath210 by proposition [ sumprop ] , this leads to the formulas of theorem [ secmain ] .
[ [ the - twisting - function - f_mathbbo - and - representations - of - the - clifford - algebras ] ] the twisting function @xmath211 and representations of the clifford algebras ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ for @xmath76 , we define a @xmath212-matrix @xmath213 whose coefficients are indexed by elements @xmath214 and are given by @xmath215 0 , & \text{otherwise}. \end{array } \right.\ ] ] note that for @xmath216 one obtains @xmath217 .
this leads to representations of clifford algebras in @xmath221 $ ] . indeed ,
if @xmath106 is a subset of @xmath9 elements of @xmath0 such that @xmath222 and @xmath223 , for all @xmath224 , then the matrices @xmath213 , @xmath225 , represent the real clifford algebra @xmath25 in the algebra @xmath221 $ ] .
we describe explicitly the irreducible representations of some of the simple real clifford algebras .
we give a possible choice of sets @xmath106 .
the condition on the set @xmath106 can be expressed using the generating function @xmath140 .
we use the notation of section [ hursect ] . 1 .
if @xmath62 , an irreducible representation of @xmath226 $ ] is given by @xmath227 2 . if @xmath228 , an irreducible real representation of @xmath229 $ ] is given by @xmath230 3 .
if @xmath231 , an irreducible real representation of @xmath232 $ ] is given by @xmath233 * acknowledgments*. the second author has benefited from the award of a _ leibniz fellowship _ at the mathematisches forschungsinstitut oberwolfach ( mfo ) .
the third author is also grateful to mfo for hospitality .
we would like to thank c. duval for helpful comments .
we are also deeply grateful to the anonymous referee for a number of helpful remarks and suggestions . | the hurwitz problem of composition of quadratic forms , or of `` sum of squares identity '' is tackled with the help of a particular class of @xmath0-graded non - associative algebras generalizing the octonions .
this method provides explicit formulas for the classical hurwitz - radon identities and leads to new solutions in a neighborhood of the hurwitz - radon identities . |
agent - based models ( abms ) are an attempt to understand how macroscopic regularities may emerge through processes of self - organization in systems of interacting agents .
one of the main purposes of this modeling strategy is to shed light on the fundamental principles of self - organized complexity in adaptive multi - level systems in order to gain an insight into the microscopic conditions and mechanisms responsible for the temporal and spatial patterns observed at aggregate levels .
therefore , abms are sometimes considered as a methodology to provide a > > theoretical bridge < < ( @xcite:148 ) between micro and macro theories ( see also @xcite ) . while used as a tool in economics , sociology , ecology and other disciplines , abms are often criticized for being tractable only by simulation .
this paper addresses this issue by applying markov chain and information - theoretic tools to a particular abm with the specific objective to better understand the transition from the most informative agent level to the levels at which the system behavior is typically observed .
a well posed mathematical basis for linking a micro - description of a model to a macro - description may help the understanding of many of the observed properties and therefore provide information about the transition from the interaction between individual entities to the complex macroscopic behaviors observed at the global level . for this purpose
, the paper draws upon a recently introduced markov chain framework for aggregation in agent - based and related computational models ( see @xcite and also @xcite ) .
the starting point is a microscopic markov chain description of the dynamical process in complete correspondence with the dynamical behavior of the agent model , which is obtained by considering the set of all possible agent configurations @xmath0 as the state space of a huge markov chain an idea borrowed from @xcite .
namely , if we consider an abm in which @xmath1 agents can be in @xmath2 different states this leads to a markov chain with @xmath3 states .
moreover , in models with sequential update by which one agent is chosen to update its state at a time , transitions are only allowed between system configurations that differ with respect to a single agent .
such an explicit micro formulation enables the application of the theory of markov chain aggregation namely , lumpability @xcite in order to reduce the state space of the micro chain and relate microscopic descriptions to a macroscopic formulation of interest @xcite .
namely , when performing simulations of an abm we are actually not interested in all the dynamical details , but rather in the behavior of certain macro - level properties that inform us about the global state of the system . in opinion dynamics , and in binary opinion models in particular , the typical level of observation is the number of agents in the different opinion states or respectively the average opinion ( due to the analogy to spin systems often called > > magnetization < < ) .
the explicit formulation of abms as markov chains enables the development of a mathematical framework to link a micro chain corresponding to an abm to such a macro - level description of interest .
more precisely , from the markov chain perspective , the transition from the micro to the macro level is a projection of the micro chain with state space @xmath0 onto a new state space @xmath4 by means of a ( projection ) map @xmath5 from @xmath6 to @xmath4 .
the meaning of the projection @xmath5 is to lump sets of micro configurations in @xmath6 into an aggregate set according to the macro property of interest .
such a situation naturally arises if the abm is observed not at the micro level of @xmath0 , but rather in terms of a measure @xmath7 on @xmath0 by which all configuration in @xmath0 that give rise to the same measurement are mapped into the same macro state , say @xmath8 .
an illustration of such a projection is provided in fig .
[ fig : projectiongeneral ] . ) is observed ( @xmath7 ) at a higher level and this observation defines another macro level process ( @xmath9 ) .
the micro process is a markov chain with transition matrix @xmath10 .
the macro process is a markov chain ( with @xmath11 ) only in the case of lumpability . ]
two things may happen by projecting the microscopic markov chain onto a coarser partition .
first , the macro process is still a markov chain which is the case of lumpability @xcite .
then markov chain tools can be used to compute the dynamical quantities of interest and a precise understanding of the model behavior is possible .
as shown in @xcite this depends essentially on the symmetries implemented in the model .
secondly , markovianity may be lost after the projection which means that memory effects are introduced at the macroscopic level .
noteworthy , in abms as well as more generally in markov chains , this situation is the rule rather than an exception @xcite .
while the first part of this paper derives exact markov chain description for the complete and a perfect two community graph , and the second part of this paper is devoted to the study of the non - markovian case . for these purposes
, the explicit construction of an abm s microscopic transition kernel is a necessary starting point because it helps , on the one hand , to establish the conditions for which the macro - level process remains markovian ( i.e. , lumpability ) , and enables , on the other hand , the use of information - theoretic measures for the `` closedness '' of an aggregate description .
a series of measures among them conditional past - future mutual information @xcite and micro - to - macro information flow @xcite have been developed to quantify the deviations from an idealized description which models the dynamics of the system by the state variables associated to a coarser level .
see also @xcite for the mathematical analysis of the relation between different levels of description in complex multi - level systems . in our context , this allows for a quantification of the memory effects that are introduced by a global aggregation over the agent population without sensitivity to micro- or mesoscopic structures which presents a first step to study how microscopic heterogeneity in abms may lead to macroscopic complexity when the aggregation procedure defines a non - markovian macro process . to my knowledge , this paper is the first to apply these concepts to an abm . in this paper the contrarian voter model ( cvm )
is used as a first simple scenario .
the cvm is a binary opinion model where @xmath1 agents placed on a network can adopt two different opinions : @xmath12 and @xmath13 . in the pure voter model @xcite
two agents are chosen at random and they align in the interaction because one of them imitates the other . as in other binary models of opinion dynamics
( see @xcite for two well - known variants and @xcite for an overview ) this mechanism of local alignment leads to a system which converges to a final profile of global conformity ( consensus ) in which all agent share the same opinion .
contrarian behavior , then , relates to the presence of individuals that do not seek conformity under all circumstances or to the existence of certain situations in which agents would not desire to adopt the behavior or attitude of their interaction partner . in the cvm
studied here contrarian behavior is included by introducing a small probability @xmath14 with which agents to not imitate their interaction partner , but adopt precisely the opposite opinion .
there are several different ways to include nonconformity behavior into a binary opinion model and the approach adopted here is probably the most simple one . in our choice
, we basically follow ref .
@xcite which , based on the concept of contrarian investment strategies in finance @xcite , is the first study to introduce contrarian behavior into a model of opinion dynamics ( namely , into galam s majority model ) .
while the majority model without contrarians is characterized by a relatively fast convergence to complete consensus , the introduction of only a small rate of contrarian choices leads to the coexistence of the two opinions with a clear majority - minority splitting .
noteworthy , as the contrarian rate increases further , the model exhibits a phase transition to a disordered phase in which no opinion dominates in the population .
similar observations have been made for the sznajd model @xcite .
more recently the literature often distinguishes between two types of nonconformity : ( i. ) anti - conformity or contrarian behavior and ( ii . ) independent or inflexible agents @xcite .
see @xcite for opinion models that include independent or inflexible agents .
the importance of a distinction between individuals that generally oppose the group norm or act independently of it is , from the socio - psychological perspective , relatively obvious .
the fact that these two behaviors may also give rise to qualitatively different dynamical properties , however , has been established only recently @xcite . here
we stick to contrarians .
the voter model with contrarians presented in @xcite is probably the one that relates most to the model used here .
the main difference is that a fixed number of agents always acts in a contrarian way whereas in the present model all agents take contrarian choices with a small probability @xmath14 . in that setting
, @xcite could not observe the phase transition from majority - minority splitting to disorder , but rather a change from a uniform to a gaussian equilibrium distribution .
while this difference in comparison with the sznajd and galam models has been attributed to the linearity of the cvm in @xcite , this study shows that there is in fact an order - disorder phase transition in the cvm as well .
however , the ordered phase can be observed only below a very small contrarian rate of @xmath15 at which the equilibrium distribution is uniform ( see sec .
[ cha:5.hmstatdyn ] ) , in accordance with @xcite . in the setting of @xcite with a fixed number of contrarian agents , however , this value is already reached , on average , with only a single contrarian , independent of the population size .
notice , finally , that the complete graph plays an exceptional role for the analytical treatment of nonconformity models commended on above . from the markovian point of view
, this is due to the fact that for the complete graph binary opinion models are lumpable that is reducible without loss of information to a macroscopic description in terms of the average opinion or > > magnetization < < @xcite .
this paper analyses the complete graph as well , but it goes beyond it by studying the cvm on a perfect two community graph . in that case , a loss - less macro description is obtained by taking into account separately the average opinion in the two sub - graphs , that is , by a refinement of the level of observation . alongside with the analysis of the respective model dynamics , these two cases allow to address some very interesting questions concerning the relation between the two coarse - grainings .
for instance , it is possible to illustrate why lumpability ( in its strong as well as in its weak form ) fails for the two community graph .
moreover , looking at the two - community scenario from the global perspective of total > > magnetization
< < allows for the exact computation of memory effects that emerge at the macroscopic level .
the sequel of the paper is organized as follows .
[ cha:5.cvm ] introduces the cvm and derives the corresponding microscopic markov chain .
[ cha:5.hmand2com ] deals with the model on the complete and the two - community graph with a particular focus on the stationary dynamics of the model .
the model dynamics are studies in terms of the contrarian rate @xmath14 and the coupling @xmath16 between the two - communities . after the discussion of the two stylized topologies , sec .
[ cha:5.networkdynamics ] shows the effect of various paradigmatic networks on the macroscopic stationary behavior . in sec .
[ cha:5.micromesomacro ] we return to the two - community cvm and find with it an analytical scenario to study the discrepancy between a mean - field model ( homogeneous mixing ) and the model on a more complex ( though still very simple ) topology .
it shows that memory effects are introduced at the macro level when we aggregate over agent attributes without sensitivity to the microscopic details .
the cvm is a binary opinion model where @xmath1 agents can adopt two different opinions : @xmath12 and @xmath13 .
the model is an extension of the voter model ( vm ) in order to include a form of contrarian behavior . at each step , an agent ( @xmath17 ) is chosen at random along with one of its neighbors ( @xmath18 ) .
usually ( with probability @xmath19 ) , @xmath17 imitates @xmath18 ( vm rule ) , but there is also a small probability @xmath14 that agent @xmath17 will do the opposite ( contrarian extension ) . more specifically ,
if @xmath17 holds opinion @xmath13 and meets an agent @xmath18 in @xmath12 , @xmath17 will change to @xmath12 with probability @xmath19 , and will maintain its current state @xmath13 with probability @xmath14 . likewise ,
if @xmath17 and @xmath18 are in the same state , @xmath17 will flip to the opposite state with probability @xmath14 . from the micro - level perspective , the cvm implements an update function of the form @xmath20 . for further convenience
, we denote the current state of an agent @xmath17 by @xmath21 and the updated state at time @xmath22 as @xmath23 .
the map @xmath24 can then be written as @xmath25 where @xmath26 denotes the opposite attribute of @xmath27 . in each iteration ,
two agents @xmath28 are chosen along with a random variable @xmath29 that decides whether the voter ( @xmath30 ) or the contrarian rule ( @xmath31 ) is performed .
the probability for that is @xmath32 .
notice that the update rule is equal for all agents and independent from the agent choice .
therefore the probability that an agent pair @xmath33 is chosen to perform the contrarian rule can be written as @xmath34 . respectively
, we have @xmath35 for the vm rule .
moreover , the respective probabilities @xmath36 are equal for each step of the iteration process . consequently , the cvm iteration may be seen as a ( time - homogeneous ) random choice among deterministic options , and is therefore a markov chain ( see @xcite and @xcite for all details ) . due to the sequential update scheme only one agent ( namely , agent @xmath17 ) may change its state at each time step .
this means that a non - zero transition probabilities exist only between agent configurations @xmath37 that differ in at most one element ( agent ) .
we call those configuration adjacent and write @xmath38 if @xmath39 and @xmath40 . considering that @xmath31 ( for contrarian rule ) is chosen with probability @xmath14 and @xmath30 ( vm rule ) with @xmath41 , and that this choice is independent of the agent choice , the micro - level transition probability @xmath42 between two adjacent configurations @xmath43 is given by @xmath44 notice that the configuration space of the cvm ( @xmath0 ) is the set of all bit - strings of length @xmath1 , and that therefore , the micro - level process for the cvm corresponds to a random walk on the hypercube ( as for the vm , see @xcite )
. notice also that the cvm leads to a _ regular _ chain ( as opposed to an absorbing random walk for the original vm ) because whenever @xmath45 , there is a non - zero probability that the process leaves the consensus states @xmath46 and @xmath47 .
( [ eq : phatvmcontrarian ] ) tells us that this probability is precisely @xmath14 .
therefore , the system does not converge to a fixed configuration and the long - term behavior of the model can be characterized by its stationary distribution .
the most natural level of observation in binary state dynamics is to consider the temporal evolution of the attribute densities , or respectively , the number of agents in the two different states .
while a mean - field description would typically formulate the macro dynamics as a differential equation describing the evolution of attribute densities , the markov chain approach operates with a discrete description ( in time as well as in space ) in which all possible levels of absolute attribute frequencies and transitions between them are taken into account .
let us denote as @xmath48 the number of @xmath12-agents in the population ( @xmath49 ) and refer to this level of observation as global or _
full aggregation_. in terms of markov chain aggregation @xcite , a macro description in terms of @xmath48 is achieved by a projection of the micro - level process on the hypercube onto a new process with the state space defined by the partition @xmath50 , where @xmath51 .
notice that in hypercube terminology that level of observation corresponds to the hamming weight of a configuration @xmath52 and each @xmath53 collects all micro configurations with the same hamming weight . .
the resulting macro process is , in general , a non - markovian process on the line . ]
one important observation in @xcite has been that homogeneous mixing is a prerequisite for lumpability with respect to @xmath54 , and that microscopic heterogeneities ( be it in the agents or in their connections ) translate into dynamical irregularities that prevent lumpability .
this means that full aggregation over the agent attributes ( @xmath55 ) leads in general to a non - markovian macro process .
we illustrate this process in fig .
[ fig : fullaggregation ] .
still , the process obtained by the projection from micro to macro is characterized by the fact that from an atom @xmath53 the only possible transitions are the loop to @xmath53 , or a transition to neighboring atoms @xmath56 and @xmath57 because only one agent changes at a time . however , the micro level transition rates ( [ eq : phatvmcontrarian ] ) depend essentially on the connectivity structure between the agents , and therefore , the transition probabilities at the macro level ( denoted as @xmath58 in fig .
[ fig : fullaggregation ] ) are not uniquely defined ( except for the case of homogeneous mixing ) .
that is , for two configurations in the same macro state @xmath59 the probability to go to another macro state ( e.g. , @xmath57 ) may be very different which violates the lumpability conditions of thm .
6.3.2 in @xcite .
before we use the cvm to address questions related to non - lumpable projections , we discuss the model behavior using markov chain tools for two stylized situations .
this section analyses the behavior of the cvm for homogeneous mixing and the two - community graph . as shown in @xcite at the example of the vm , the symmetries in the interaction network
can be used to define a markovian coarse - graining . as both interaction topologies are characterized by a large symmetry group , markov chain aggregation considerably
reduces the size of the micro chains such that the important entities of interest ( e.g. , stationary distribution ) can be computed on the basis of the respective macroscopic transition matrices .
the case of homogeneous mixing is particularly simple .
let us consider that the model is implemented on the complete graph without loops where the probability to choose a pair @xmath33 of agents becomes @xmath60 whenever @xmath61 and @xmath62 .
consequently , since all agents interact with all the others with equal probability , the respective transition rates depend only on the numbers @xmath48 and @xmath63 of agents in the two states : @xmath64 \vspace{6pt}\\ & = ( n - k ) \left [ ( 1-p ) k \omega + p ( n - k)\omega \right]\vspace{6pt}\\ & = ( 1-p ) \frac{(n - k ) k}{n(n-1 ) } + p \frac{(n - k)(n - k-1)}{n(n-1)}. \end{array } \label{eq : pmacrovmcontrarian01}\end{aligned}\ ] ] similarly , we obtain for @xmath65 @xmath66 and finally , @xmath67 [ cols="^,^ " , ]
let us consider that for the cylinders of length 3 .
as noted above , the grammar @xmath68 of the system is determined by the fact that @xmath69 and @xmath70 .
therefore , as illustrated in fig .
[ fig : macro3cylinder ] , for any @xmath71 with @xmath72 there are nine possible paths @xmath73 $ ] and for @xmath74 and @xmath75 there are respectively four paths . in order to compute the probability of a certain macro path , say @xmath76 $ ] @xmath14 for past and @xmath77 for future , we have to sum over all meso level paths that contribute to the given macro path .
let us denote a meso level path as @xmath78 $ ] with @xmath79 , @xmath80 and @xmath81 .
its probability is given by @xmath82 ) = \tilde{\pi}_{m_p , l_p } \tilde{p}(\tilde{x}_{m_p , l_p},\tilde{x}_{m , l } ) \tilde{p}(\tilde{x}_{m , l},\tilde{x}_{m_f , l_f } ) .
\label{eq : mu.meso}\ ] ] the l.h.s . in fig .
[ fig : meso3cylinder ] illustrates the possible paths for one @xmath83 with @xmath84 .
notice that for a given macro state @xmath53 there are @xmath85 meso states if @xmath86 and respectively @xmath87 meso states for @xmath88 ( these numbers are for the case @xmath89 with @xmath90 ) . in the case sequences of length three are considered , the situation is still quite clear .
for instance , a macro path @xmath91 $ ] can be realized in four different ways for each or @xmath92 . ]
@xmath83 with @xmath84 : @xmath93\\ \nonumber [ ( m-1 \ l),(m \ l),(m \
l+1)]\\ \nonumber [ ( m \ l-1),(m \ l),(m+1 \ l)]\\ \nonumber [ ( m \ l-1),(m \ l),(m \ l+1)]\end{aligned}\ ] ] the same reasoning can be applied to derive the probabilities for cylinders of length four even though the situation becomes slightly more complicated , as illustrated on the r.h.s . of fig .
[ fig : meso3cylinder ] . on the basis of the probabilities of blocks of length three and four
respectively , the computation of the markovianity measures @xmath94 and @xmath95 is straightforward . all that is needed is to compute the respective block entropies . fig .
[ fig : in.all ] shows @xmath96 ( dashed curves ) and @xmath97 ( solid curves ) as a function of the coupling between the two communities @xmath16 for a system of @xmath98 agents ( @xmath99 ) .
the different curves represent various different contrarian rates @xmath14 from @xmath100 to @xmath101 .
notice the log - linear scaling of the figure .
( dashed curves ) and @xmath97 ( solid curves ) as a function of the coupling between the two communities @xmath16 for a system of @xmath98 agents .
the different curves represent various different contrarian rates @xmath14 from @xmath101 to @xmath100 , see legend . ]
what becomes clear in fig .
[ fig : in.all ] , first of all , is that the deviation from markovianity is most significant for small inter - community couplings .
this means , in the reading of @xcite , that the information provided by pasts of length @xmath102 about the future state ( beyond that given by the present ) is larger than zero for small @xmath16 . in general and
not surprisingly , @xmath103 which means that both the first and the second outcome before the present provide a considerable amount of information .
in fact , the numbers indicate that the first and the second step into the past contribute in almost the same way .
noteworthy , the two measures @xmath96 and @xmath97 behave in the same way from the qualitative point of view which suggests that the computationally less expensive @xmath96 can be well - suited for the general markovianity test .
the inset in fig .
[ fig : in.all ] shows the situation for values around @xmath104 ( homogeneous mixing ) as well as @xmath105 .
as we would expect by the strong lumpability of homogeneous mixing , @xmath96 and @xmath97 are zero in the case @xmath106 .
also if the inter - community coupling becomes larger than the coupling within communities ( a situation that resembles a bipartite graph ) @xmath96 and @xmath97 are very small , indicating that a markovian macro description ( i.e. , ideal aggregation , cf .
@xcite , p. 140 and
@xcite , pp .
61 63 ) describes well these situations .
finally , we notice in fig .
[ fig : in.all ] that the measures do not generally increase monotonically with a decreasing ratio @xmath16 which is most obvious for the example with a very small @xmath107 ( green curves ) .
this is somewhat unexpected and it indicates the existence of certain parameter constellations at which macroscopic complexity ( for this is how non - markovianity may be read ) is maximized . to obtain a better understanding of this behavior , the measures @xmath96 and @xmath97 are plot in fig .
[ fig : in.p ] as a function of the contrarian rate @xmath14 .
notice again the log - linear scaling of the plot . and @xmath97 as a function of the contrarian rate @xmath14 for various coupling ratios @xmath16 and a system of @xmath99 . ]
it becomes clear that there is a strong and non - trivial dependence of the markovianity measures on the contrarian rate @xmath14 .
namely , @xmath96 and @xmath97 are very small if @xmath14 is relatively large but they are also relatively small if @xmath14 becomes very small .
there is a parameter regime in between in which deviations from markovianity become most significant .
notice that in the inset of fig .
[ fig : in.p ] the same curves are shown on a double - logarithmic scale .
this shows , first , that @xmath96 and @xmath97 for very small @xmath14 are still significantly larger compared to the case of relatively large @xmath14 ( say @xmath108 ) .
secondly , we observe that @xmath96 and @xmath97 actually vanish for @xmath109 . indeed , it is possibly to show that the two - community cvm with @xmath109 satisfies lumpability conditions independent of the topological parameter @xmath16 ( cf .
@xcite , pp.116/17 ) .
finally , a detailed picture of the dependence of @xmath110 on the contrarian rate is provided in fig .
[ fig : in.p.fine ] .
the plot compares the cases @xmath111 and @xmath112 in order to show that the peaks in the @xmath110 depend also on @xmath16 . for the interpretation of this behavior , notice that the @xmath14 at which deviations from markovianity become largest , lie precisely in the parameter interval in which switching times between the two complete consensus states become minimal .
compare fig .
[ fig : cvm.mt.0ton ] in sec .
[ cha:5.majminswitching ] . on the contrarian rate .
blue curves correspond to @xmath111 and red curves to @xmath112 . in the first case
the peak is at around @xmath113 , in the latter at @xmath114 . ] all in all , this analysis shows that global aggregation over an agent population without sensitivity to micro- or mesoscopic structures leads to memory effects at the macroscopic level .
this paper has provided an analysis of the cvm on the complete and the two - community graph . based on the previous work on markov chain aggregation for abms @xcite ,
higher - level markov chain descriptions have been derived and allow a detailed understanding of the two cases .
a large contrarian rate @xmath14 leads to a process which fluctuates around the states with approximately the same number of black and white agents , the fifty - fifty situation @xmath115 being the most probable observation .
this is true for homogeneous mixing as well as for the two - community model .
however , if @xmath14 is small , a significant difference between the two topologies emerges as the coupling between the two communities becomes weaker . on the complete graph the population is almost uniform for long periods of time , but due to the random perturbations introduced by the contrarian rule there are rare transitions between the two consensus profiles . on the community graph , an effect of local alignment is observed in addition to that , because the system is likely to approach a meta - stable state of intra - community consensus but inter - community polarization .
a order - disorder phase transition as the contrarian rate increases has been observed on the complete graph in several previous contrarian opinion models ( e.g. , @xcite ) . for the cvm , in the transition from consensus switching to disorder
there is a phase in which the process leads uniform stationary distribution in which all opinion frequency levels @xmath51 are observed with equal probability ( @xmath116 ) .
the contrarian rate @xmath14 at which this happens is @xmath15 and depends inversely on the system size such that a model with a single contrarian agent fails to enter the ordered regime .
this confirms and explains the behavior observed in @xcite for a model with a fixed number of contrarian agents .
a particular focus of this paper has been on the effect of inhomogeneities in the interaction topology on the stationary behavior . in this regard ,
the two - community cvm served as a suitable scenario to assess the macroscopic effects introduced by a slight microscopic heterogeneity .
namely , homogeneous mixing compatible with the usual way of aggregation over all agents leads to a random walk on the line with @xmath117 states whereas the two - community model leads to a random walk on a 2d lattice with @xmath118 states .
as the latter is a proper refinement of the former this gives us means to study the relation between the two coarse - grainings in a markov chain setting . in this regard , this paper has made visible the reasons for which lumpability fails , and it has also provided a first analysis of the macroscopic memory effects that are introduced by heterogeneous interaction structures . in this regard , the paper demonstrates that information - theoretic measures are a promising tool to study the relationship between different levels of description in abms .
there are various issues that deserve further discussion .
for instance , is the emergence of memory in the transition from the micro to the macro level a useful characterization for the complexity of a certain system ?
namely , the theory of markov chain aggregation makes explicit statements about when a micro process is _ compressible _ to a certain macro - level description .
this links non - lumpability to computational incompressibility , one of the key concepts in dynamical emergence ( * ? ? ?
* ; * ? ? ?
* among others ) .
this point shall be discussed in a forthcoming paper .
finally , i would like to mention the possibility of applying the arguments developed in the second part this paper to the case of models with absorbing states as , for instance , the pure vm ( @xmath119 ) . in that case
, the quasi - stationary distribution ( see @xcite ) takes the role of @xmath120 or respectively @xmath121 in the computation of cylinder measures .
one interesting issue to be addressed in this regard is to reconsider the question of weak lumpability for the vm .
finally , to understand how microscopic heterogeneity and macroscopic complexity are related , numerical experiments with different network topologies are another promising way to continue the analysis started in this paper .
banisch , s. , lima , r. , and arajo , t. , aggregation and emergence in agent - based models : a markov chain approach , in _ proceedings of the european conference on complex systems 2012 _ , eds .
gilbert , t. , kirkilionis , m. , and nicolis , g. , springer proceedings in complexity ( springer international publishing , 2013 ) , isbn 978 - 3 - 319 - 00394 - 8 , pp .
37 .
izquierdo , l. r. , izquierdo , s. s. , galn , j. m. , and santos , j. i. , techniques to understand computer simulations : markov chain analysis , _ journal of artificial societies and social simulation _ * 12 * ( 2009 ) 6 .
jacobi , m. n. and grnerup , o. , a spectral method for aggregating variables in linear dynamical systems with application to cellular automata renormalization , _ advances in complex systems _ * 12 * ( 2009 ) 131155 .
lazarsfeld , p. and merton , r. k. , friendship as a social process : a substantive and methodological analysis , in _ freedom and control in modern society _ , eds .
berger , m. , abel , t. , and page , c. h. ( new york : van nostrand , 1954 ) , pp . | an analytical treatment of a simple opinion model with contrarian behavior is presented .
the focus is on the stationary dynamics of the model and in particular on the effect of inhomogeneities in the interaction topology on the stationary behavior .
we start from a micro - level markov chain description of the model .
markov chain aggregation is then used to derive a macro chain for the complete graph as well as a meso - level description for the two - community graph composed of two ( weakly ) coupled sub - communities . in both cases ,
a detailed understanding of the model behavior is possible using markov chain tools .
more importantly , however , this setting provides an analytical scenario to study the discrepancy between the homogeneous mixing case and the model on a slightly more complex topology .
we show that memory effects are introduced at the macro level when we aggregate over agent attributes without sensitivity to the microscopic details and quantify these effects using concepts from information theory . in this way
, the method facilitates the analysis of the relation between microscopic processes and a their aggregation to a macroscopic level of description and informs about the complexity of a system introduced by heterogeneous interaction relations .
max planck institute for mathematics in the sciences ( germany ) + mpi - mis , inselstrasse 22 , d-04103 leipzig , germany + email : [email protected] |
the ads / cft correspondence @xcite , a typical example of which is a conjecture that type iib superstring on @xmath6 corresponds to @xmath0 super yang - mills ( sym ) , has been intensively investigated for a decade . however , it has not been completely proven yet , partially because it is a strong / weak duality with respect to the coupling constants .
it is , therefore , relevant to give a nonperturbative definition of @xmath0 sym that enables us to study its strong coupling regime .
the lattice gauge theory is a promising candidate for such a nonperturbative definition .
however , supersymmetric gauge theories on the lattice are generally difficult to construct , although there have been remarkable developments on this subject @xcite . to give a nonperturbative definition of @xmath0 sym will not only bring enormous progress in the study of the ads / cft correspondence , but will also yield some insights into the problem of nonperturbative formulation of supersymmetric gauge theories .
it was shown in @xcite that a gauge theory in the planar limit is equivalent to the matrix model ( the reduced model ) obtained by dimensionally reducing it to zero dimension if the @xmath7 symmetry is unbroken , where @xmath8 stands for the dimensionality of space - time .
this is the so - called large @xmath9 reduction .
the global gauge symmetry of the matrix model is naturally interpreted as the local gauge symmetry of the original gauge theory .
thus , as an alternative to the lattice gauge theory , the matrix model may serve as a nonperturbative definition of the planar gauge theory with the gauge symmetry manifestly kept .
the @xmath7 symmetry is , however , spontaneously broken except for @xmath10 , so that the above equivalence does not hold generically .
there have been two improvements of the reduced model in which the @xmath7 symmetry breaking is prevented so that the equivalence holds : one is the quenched reduced model @xcite , and the other is the twisted reduced model @xcite .
these improved models work well for nonsupersymmetric planar gauge theories .
it seems quite difficult to preserve supersymmetry manifestly in the twisted reduced model on the flat space and in the quenched reduced model , while the gauge symmetry is respected in both models .
the compactification in matrix models developed in @xcite shares the same idea with the reduced model and will be called the matrix t - duality in this paper . while it is not restricted to the planar limit
, it requires the size of matrices to be infinite from the beginning for the orbifolding condition to be imposed , so that it can not be used to define any supersymmetric gauge theory nonperturbatively as it stands .
it was argued in @xcite that by imposing an orbifolding condition on the reduced model of a supersymmetric gauge theory , one can obtain its lattice theory in which part of the supersymmetries are manifestly preserved so that the fine - tuning of only a few parameters is required . this construction can be regarded as a finite - size matrix analog of the matrix t - duality .
however , it has a problem of flat directions which is analogous to the problem of the @xmath7 symmetry breaking . to overcome this problem , for instance , one needs to introduce a mass term for the scalar field , which leads to no preservation of supersymmetries . in @xcite , takayama and
three of the present authors found the relationships among the @xmath11 symmetric theories , which include @xmath0 sym on @xmath12 , 2 + 1 sym on @xmath13 @xcite and the plane wave matrix model ( pwmm ) @xcite .
the last theory is obtained by consistently truncating the kaluza - klein modes of @xmath0 sym on @xmath1 @xcite and so are the former two theories @xcite .
in particular , 2 + 1 sym on @xmath13 and pwmm can be regarded as dimensional reductions of @xmath0 sym on @xmath1 .
these theories possess common features : mass gap , discrete spectrum and many discrete vacua . from the gravity duals of those vacua proposed by lin and maldacena @xcite , the following relations among these theories
are suggested : a ) the theory around each vacuum of 2 + 1 sym on @xmath13 is equivalent to the theory around a certain vacuum of pwmm , and b ) the theory around each vacuum of @xmath0 sym on @xmath12 is equivalent to the theory around a certain vacuum of 2 + 1 sym on @xmath13 with the orbifolding ( periodicity ) condition imposed . in @xcite ,
the relations a ) and b ) were shown directly on the gauge theory side .
the results in @xcite not only serve as a nontrivial check of the gauge / gravity correspondence for the @xmath11 theories , but they are also interesting from the point of view of the reduced model as follows . while there have been many works on realizing the gauge theories on the fuzzy sphere @xcite using matrix models @xcite and on the monopoles on the fuzzy sphere @xcite , the relation a ) shows that the continuum limit of the concentric fuzzy spheres with different radii corresponds to multiple monopoles .
note that realizing the gauge theories on the fuzzy sphere using the matrix models can be viewed as an extension of the twisted reduced model to curved space .
the relation b ) can be regarded as an extension of the matrix t - duality to that on a nontrivial @xmath14 bundle , @xmath15 , whose base space is @xmath16 .
indeed , the matrix t - duality was later extended to that on general @xmath14 bundles in @xcite and on general @xmath17 bundles in @xcite . combining the relations
a ) and b ) leads to the relation c ) , that the theory around each vacuum of @xmath0 sym on @xmath12 is equivalent to the theory around a certain vacuum of pwmm with the orbifolding condition imposed .
in particular , for @xmath18 , @xmath0 sym on @xmath1 is realized in pwmm .
the possibility of defining @xmath0 sym in terms of pwmm nonperturbatively is suggested in @xcite .
the relationships shown in @xcite are classical in the following sense : in the relation a ) , we show the equivalence at tree level and do not care about possible uv / ir mixing at higher orders , although the gravity duals suggest that any uv / ir mixing does not occur . in the relation
b ) , the size of matrices must be infinite from the beginning as in the original matrix t - duality . in this paper
, we propose a nonperturbative definition of @xmath0 sym on @xmath1 which is equivalently mapped to @xmath0 sym on @xmath19 at the conformal point and possesses the superconformal symmetry , the @xmath20 symmetry .
we restrict ourselves to the planar limit . by referring to the relation c ) in @xcite
, we regularize @xmath0 sym on @xmath1 nonperturbatively by using pwmm .
our analysis in this paper is quantum mechanical .
the restriction to the planar limit enables us not to impose the orbifolding condition and to consider finite - size matrices such that the size of matrices plays the role of the ultraviolet cutoff .
thus we use an extension of the reduced model to curved space rather than the matrix t - duality to relate @xmath21 sym on @xmath13 to @xmath0 sym on @xmath1 . because pwmm is a massive theory ,
there is no flat direction and the quenching prescription is not needed .
our regularization manifestly preserves the gauge symmetry and the @xmath11 symmetry , a subgroup of the @xmath20 symmetry .
in particular , sixteen supersymmetries among thirty - two supersymmetries are respected in our regularization .
the restriction to the planar limit and sixteen supersymmetries are probably sufficient to suppress the uv / ir mixing which may break the relation between @xmath21 sym on @xmath13 and pwmm quantum mechanically .
they also stabilize the vacua of pwmm completely .
indeed , the gravity duals of these theories suggest @xmath21 sym on @xmath13 is obtained from pwmm quantum mechanically in the continuum limit at least in the planar limit .
the full @xmath20 symmetry should be restored in the continuum limit . by performing the 1-loop analysis and comparing the results with those in continuum @xmath0 sym
, we provide some evidences that our regularization of @xmath0 sym indeed works , although our final goal is to analyze @xmath0 sym nonperturbatively by using our formulation .
our theory still has the continuum time direction , which we need to cope with in order to put our theory on computer .
for instance , we should be able to apply the method in @xcite to our case .
we comment on an interesting paper @xcite , the authors of which constructed the @xmath22 background in the iib matrix model with the myers term using the same procedure as @xcite .
they calculated the free energy of the theory around the background up to the 2-loop order to find the stability of the background .
note also that the authors of @xcite discussed practicality of @xmath0 sym on the lattice recently .
this paper is organized as follows . in section 2 , we study the large @xmath9 reduction on a finite volume . as an example , we consider the @xmath23 matrix quantum mechanics .
we examine how the theory on @xmath24 is obtained from the matrix model that is its dimensional reduction to zero dimension , emphasizing the difference between the large @xmath9 reductions for the theories on @xmath25 and @xmath24 . in section 3 , we review the relationships among @xmath0 sym on @xmath1 , 2 + 1 sym on @xmath13 and pwmm shown in @xcite . based on these relationships and the result in section 2 , we give a nonperturbative definition of @xmath0 sym on @xmath1 using pwmm . in section 4 , we perform the 1-loop calculation in our theory to give some evidences that our regularization of @xmath0 sym indeed works .
section 5 is devoted to conclusion and discussion . in appendices ,
some details are gathered .
in this section , we study the large @xmath9 reduction on a finite volume , focusing on how different it is from that on an infinite volume .
let us consider a matrix quantum mechanics , whose action is given by @xmath26 where @xmath27 is an @xmath28 hermitian matrix .
we take the t hooft limit : @xmath29 .
first , we consider the case in which the theory is defined on @xmath25 , namely @xmath30 .
the prescription of the large @xmath9 reduction is to make the following replacement @xcite : @xmath31 where @xmath32 in the right - hand side of the first equation is no longer dependent on @xmath33 and @xmath34 is an ultraviolet cutoff .
@xmath35 is a constant @xmath28 matrix given by @xmath36 with @xmath37 .
we take the limit in which @xmath38 note that @xmath39 is an infrared cutoff . the action ( [ original model ] ) is reduced to @xmath40 ( a ) [ g - afig ] ( b ) [ g - bfig ] in order to illustrate the large @xmath9 reduction , we see that the free energy of the original model ( [ original model ] ) agrees with that of the reduced model ( [ reduced model ] ) at the two - loop level .
there are two diagrams at the two - loop level for the free energy .
1-(a ) shows the planar diagram while fig .
1-(b ) shows the nonplanar one .
we evaluate the planar diagram in fig .
1-(a ) for the original model : @xmath41 the nonplanar diagram in fig .
1-(b ) for the original model is suppressed by the order of @xmath42 compared to the planar one in fig .
1-(a ) . on the other hand
, we evaluate the planar diagram in fig .
1-(a ) for the reduced model : @xmath43 by using the relation valid in the limit ( [ limit for reduced model ] ) , @xmath44 one can easily verify that @xmath45 .
indeed , one can prove that @xmath46 holds at all orders .
we further evaluate the nonplanar diagram in fig .
1-(b ) for the reduced model : @xmath47 note that there is no correspondence for the nonplanar diagram between the original and reduced models .
the nonplanar contribution ( [ nonplanar diagram for the reduced model ] ) is suppressed by the factor @xmath48 in the limit ( [ limit for reduced model ] ) , relative to the planar contribution ( [ planar diagram for the reduced model ] ) .
all the nonplanar contributions are indeed suppressed relative to the planar contributions in the reduced model .
the reduced model therefore reproduces the t hooft ( planar ) limit of the original model .
next , we compactify the @xmath33-direction to @xmath24 with the radius @xmath25 .
we evaluate the planar diagram in fig .
1-(a ) for the original model : @xmath49 where we take the @xmath50 limit with @xmath51 fixed .
note that the nonplanar diagram in fig .
1-(b ) for the original model is still suppressed by @xmath42 relative to the planar diagram in fig .
correspondingly , we consider the reduced model @xmath52 with @xmath53 .
this naive reduced model turns out not to reproduce the original model on @xmath24 .
the contribution of the planar diagram in fig .
1-(a ) to the free energy for this reduced model is @xmath54 while that of the nonplanar diagram in fig .
1-(b ) is @xmath55 ( [ nonplanar diagram for the reduced model for s^1 ] ) is not suppressed relative to ( [ planar diagram for the reduced model for s^1 ] ) , because the infrared cutoff @xmath56 is finite in this case .
thus the correspondence between the original and reduced models fails in this case . in the following ,
we modify the reduced model ( [ reduced model ] ) to recover the correspondence .
the action of the modified model takes the same form as ( [ reduced model for s^1 ] ) while @xmath32 is a @xmath57 matrix , @xmath58 run from 1 to @xmath59 and @xmath60 is given by the @xmath61-component of the matrix @xmath62 here @xmath63 is a positive even integer .
we take the limit in which @xmath64 .
@xmath63 turns out to play the role of the ultraviolet cutoff for the momentum . in the modified model , the contribution of the planar diagram in fig .
1-(a ) to the free energy is @xmath65 then , we see that @xmath66 .
indeed , it is easily verified that @xmath67 holds at all orders . on the other hand , the contribution of the nonplanar diagram in fig .
1-(b ) to the free energy for the modified model is @xmath68 this is suppressed by @xmath42 relative to ( [ planar diagram for modified model ] ) .
all the nonplanar contributions are indeed suppressed relative to the planar contributions in the modified model . hence , the modified model reproduces the t hooft ( planar ) limit of the original model on @xmath24 . for @xmath8-dimensional pure yang - mills ( ym ) , the reduction analogous to ( [ replacement ] ) leads to @xmath69 ^ 2 { \nonumber \\}&\rightarrow s_{ym , r}=-\frac{2\pi}{\lambda}\frac{1}{4g^2}\mbox{tr}[p_{\mu}+a_{\mu},p_{\nu}+a_{\nu}]^2 , \label{reduced model of ym}\end{aligned}\ ] ] where @xmath70 is the @xmath8-dimensional analogue of ( [ p ] ) .
it is known that the reduced model ( [ reduced model of ym ] ) does not reproduce the original ym , because the diagonal elements of @xmath71 are zero - dimensional massless fields and instable enough to absorb @xmath70 .
this is interpreted as the counterpart of the @xmath7 symmetry breaking in the reduced model of the lattice gauge theory .
usually , in order to overcome this problem , the eigenvalues of @xmath72 in ( [ reduced model of ym ] ) are fixed to @xmath70 @xcite .
this is a quenching prescription .
while the gauge symmetry is respected in this prescription , supersymmetry is not . in the case
we are concerned with in this paper , we want to respect both symmetries simultaneously .
we see how this problem is overcome in the next section .
in this section , we review the relationships among @xmath0 sym on @xmath1 , 2 + 1 sym on @xmath13 and pwmm shown in @xcite , and we propose a nonperturbative definition of @xmath0 sym on @xmath1 , based on these relationships and the result in the previous section . the action of @xmath0 sym on @xmath1 takes the form . ]
@xmath73[\phi^{ab},\phi^{cd}]+i\psi_a^{\dagger}d_t\psi^a + i\psi_a^{\dagger}\sigma^id_i\psi^a { \nonumber \\}&\hspace{3.7cm}\left.+\psi_a^{\dagger}\sigma^2[\phi^{ab},(\psi_b^{\dagger})^t ]
-\psi^{at}\sigma^2[\phi_{ab},\psi^b ] \right ) .
\label{actiontwo - component}\end{aligned}\ ] ] here @xmath74 are the local lorentz indices and run from 0 to 3 . 0 " corresponds to the time , @xmath75 .
@xmath76 are indices of the fundamental representation of @xmath77 and run from 1 to 4 .
@xmath78 and @xmath79 .
the radius of @xmath22 is @xmath80 .
this theory possesses the superconformal symmetry , the @xmath20 symmetry .
the action of 2 + 1 sym on @xmath13 takes the form @xmath81[\phi^{ab},\phi^{cd } ] \nonumber \\ & \hspace{2.5 cm } + i\psi_{a}^\dagger d_t\psi^a -\psi_a^\dagger\vec{\sigma}\cdot \vec{\cal d}\psi^a -\frac{3\mu}{4}\psi_a^\dagger\psi^a + \psi_a^\dagger\sigma^2[\phi^{ab},(\psi_b^\dagger)^t ] -\psi^{at}\sigma^2[\phi_{ab},\psi^b]\biggr ) , \label{s^2su(4)form}\end{aligned}\ ] ] where @xmath82 with @xmath83 , @xmath84 and @xmath85 , @xmath86 $ ] and the radius of @xmath16 is @xmath87 .
the action of pwmm takes the form @xmath88)^2 + \frac{1}{2}d_t \phi_{ab}d_t \phi^{ab}-\frac{\mu^2}{8}\phi_{ab}\phi^{ab } \right .
\nonumber\\ & \qquad\qquad + \frac{1}{2}[x_i,\phi_{ab}][x_i,\phi^{ab } ] + \frac{1}{4}[\phi_{ab},\phi_{cd}][\phi^{ab},\phi^{cd } ] + i\psi_a^\dagger d_t\psi^a \nonumber\\ & \qquad\qquad\left . -\frac{3\mu}{4}\psi_a^\dagger\psi^a + \psi_a^\dagger \sigma^i[x_i,\psi^a ] + \psi_a^\dagger \sigma^2[\phi^{ab},({\psi}_b^\dagger)^t ] -(\psi^a)^t \sigma^2 [ \phi_{ab},\psi^b ] \right ) .
\label{action_of_pwmm}\end{aligned}\ ] ] both 2 + 1 sym on @xmath13 and pwmm possess the @xmath11 symmetry , which is a subgroup of the @xmath20 symmetry and has sixteen supercharges . in the reminder of this section , for simplicity ,
we ignore the time component of the gauge field @xmath89 and the matter degrees of freedom , @xmath90 and @xmath91 .
it is easy to include these degrees of freedom in the arguments .
all the statements in the following are also valid with these degrees of freedom .
first , we summarize some useful facts about @xmath22 and @xmath16 ( see also @xcite ) .
we regard @xmath22 as the @xmath17 group manifold .
we parameterize an element of @xmath17 in terms of the euler angles as @xmath92 where @xmath93 , @xmath94 , @xmath95 .
the periodicity with respect to these angle variables is expressed as @xmath96 the isometry of @xmath22 is @xmath97 , and these two @xmath17 s act on @xmath98 from left and right , respectively . note that the superconformal group @xmath20 includes the @xmath99 group as a subgroup .
we construct the right - invariant 1-forms , @xmath100 where the radius of @xmath22 is @xmath80 .
they are explicitly given by @xmath101 and satisfy the maurer - cartan equation @xmath102 the metric is constructed from @xmath103 as @xmath104 the killing vectors dual to @xmath103 are given by @xmath105 where @xmath106 and @xmath107 are inverse of @xmath108 .
the explicit form of the killing vectors are @xmath109 because of the maurer - cartan equation ( [ maure - cartan ] ) , the killing vectors satisfy the su(2 ) algebra , @xmath110=i\epsilon_{ijk}{\cal{l}}_k$ ] .
one can also regard @xmath22 as a @xmath14 bundle over @xmath111 .
@xmath16 is parametrized by @xmath112 and @xmath113 and covered with two local patches : the patch i defined by @xmath114 and the patch ii defined by @xmath115 . in the following expressions , the upper sign is taken in the patch i while the lower sign in the patch ii .
the element of @xmath17 in ( [ euler angles ] ) is decomposed as @xmath116 @xmath117 represents an element of @xmath16 , while @xmath118 represents the fiber @xmath14 .
the fiber direction is parametrized by @xmath119 .
note that @xmath117 has no @xmath113-dependence for @xmath120 .
the zweibein of @xmath16 is given by the @xmath121 components of the left - invariant 1-form , @xmath122 @xcite .
it takes the form @xmath123 this zweibein gives the standard metric of @xmath16 with the radius @xmath87 : @xmath124 making a replacement @xmath125 in ( [ killing vector ] ) leads to the angular momentum operator in the presence of a monopole with magnetic charge @xmath126 at the origin @xcite : @xmath127 where @xmath126 is quantized as @xmath128 , because @xmath129 is a periodic variable with the period @xmath130 .
these operators act on the local sections on @xmath16 and satisfy the @xmath17 algebra @xmath131=i\epsilon_{ijk}l_k^{(q ) } $ ] .
note that when @xmath132 , these operators are reduced to the ordinary angular momentum operators ( [ angular momentum operator ] ) on @xmath16 ( or @xmath133 ) , which generate the isometry group of @xmath16 , @xmath17 . the @xmath17 acting on @xmath98 from left survives as the isometry of @xmath16 .
note that in 2 + 1 sym on @xmath13 the isometry of @xmath16 is included in the @xmath11 symmetry as a subgroup .
we dimensionally reduce the higher dimensional theories to the lower dimensional theories @xcite .
we start with @xmath0 sym on @xmath1 : @xmath134 we put @xmath135 ( note that we have ignored @xmath89 ) .
then , the curvature 2-form is given by @xmath136)\right)e^i \wedge e^j .
\label{curvature 2-form}\end{aligned}\ ] ] by using ( [ curvature 2-form ] ) , we rewrite ( [ ym on s^3 ] ) as @xmath137)\right)^2 \right ) .
\label{ym on s^3 2}\end{aligned}\ ] ] by dropping the @xmath129-derivatives in ( [ ym on s^3 2 ] ) , we obtain 2 + 1 sym on @xmath13 : @xmath138)\right)^2
\right ) , \label{ym - higgs on s^2}\end{aligned}\ ] ] where @xmath139 .
thus we obtain 2 + 1 sym on @xmath13 from @xmath0 sym on @xmath1 by dimensionally reducing the fiber direction of @xmath22 viewed as a @xmath14 bundle over @xmath16 .
one of two @xmath17 s that are the isometry of @xmath22 survives as the isometry of @xmath16 .
correspondingly , the superconformal symmetry , @xmath20 , reduces to the @xmath11 symmetry .
it is convenient for us to rewrite ( [ ym - higgs on s^2 ] ) using the gauge field and a higgs field on @xmath16 .
we decompose @xmath140 into the components tangential and horizontal to @xmath16 @xcite : @xmath141 where @xmath142 and @xmath143 are the gauge field on @xmath16 and @xmath144 is the higgs field on @xmath16 . substituting ( [ y vector ] ) into ( [ ym - higgs on s^2 ] )
leads to @xmath145 where @xmath146 run from 1 to 2 . dropping all the derivatives in ( [ ym - higgs on s^2 ] ) , we obtain @xmath147\right)^2 \right ) , \label{matrix model}\end{aligned}\ ] ]
where @xmath148 .
thus pwmm is obtained from 2 + 1 sym on @xmath13 by a dimensional reduction . in this reduction ,
the @xmath11 symmetry is preserved . while @xmath0 sym on @xmath1 possesses the unique vacuum , 2 + 1 sym on @xmath13 and pwmm
possess many nontrivial vacua @xcite .
let us see how those vacua are described .
first , the vacuum configurations of ( [ ym - higgs on s^2 2 ] ) with the gauge group @xmath149 are determined by @xmath150 in the gauge in which @xmath144 is diagonal , ( [ eom for ym - higgs on s^2 2 ] ) is solved as @xmath151 where the gauge field takes the configurations of dirac s monopoles , so that @xmath152 must be half - integers due to dirac s quantization condition .
note also that @xmath153 .
thus the vacua of 2 + 1 sym on @xmath13 are classified by the monopole charges @xmath152 and their degeneracies @xmath154 .
the vacua preserve the @xmath11 symmetry .
( [ eom for ym - higgs on s^2 2 ] ) is rewritten in terms of the notation in ( [ ym - higgs on s^2 ] ) as @xmath155=0,\end{aligned}\ ] ] which is equivalent to @xmath156=i\mu \epsilon_{ijk}(\mu l_k^{(0)}-x_k),\end{aligned}\ ] ] and ( [ vacuum of ym - higgs on s^2 2 ] ) is rewritten as @xmath157 where @xmath158 .
next , the vacuum configurations of ( [ matrix model ] ) with the gauge group @xmath159 are determined by @xmath160=-i\mu\epsilon_{ijk}x_k . \label{eom for matrix model}\end{aligned}\ ] ] ( [ eom for matrix model ] ) is solved as @xmath161 where @xmath162 are the representation matrices of the @xmath17 generators which are in general reducible , and are decomposed into irreducible representations : @xmath163{-35 } { $ \cdots \;\;\ ; \overbrace{\rotatebox[origin = c]{35}{$l_{i}^{[j_{s-1}]}$ } \ ; \cdots \ ; \rotatebox[origin = c]{35}{$l_{i}^{[j_{s-1}]}$}}^{\rotatebox{35}{$n_{s-1}$ } } \;\;\ ; \overbrace{\rotatebox[origin = c]{35}{$l_i^{[j_{s}]}$ } \ ; \cdots \ ; \rotatebox[origin = c]{35}{$l_i^{[j_{s}]}$}}^{\rotatebox{35}{$n_{s}$ } } \;\;\ ; \overbrace{\rotatebox[origin = c]{35}{$l_{i}^{[j_{s+1}]}$ } \ ; \cdots \ ; \rotatebox[origin = c]{35}{$l_{i}^{[j_{s+1}]}$}}^{\rotatebox{35}{$n_{s+1}$ } } \;\;\;\cdots $ } \end{pmatrix } , \label{matrix background}\end{aligned}\ ] ] where @xmath164}$ ] are the spin @xmath165 representation matrices of @xmath17 and @xmath166 .
the vacua of the matrix model are classified by the @xmath17 representations @xmath167 $ ] and their degeneracies @xmath154 .
( [ matrix background ] ) represents concentric fuzzy spheres with different radii .
the vacua preserve the @xmath11 symmetry . in what follows
, we obtain the higher dimensional theories from the lower dimensional theories .
first , we recall the relationship between the theory around ( [ monopole background ] ) of @xmath21 sym on @xmath13 and the theory around ( [ vacuum of matrix model ] ) of pwmm , which was shown in @xcite for the trivial vacuum of 2 + 1 sym on @xmath13 and in @xcite for generic vacua .
we introduce an ultraviolet cutoff @xmath168 and put @xmath169 then , the theory around ( [ monopole background ] ) is equivalent to the theory around ( [ vacuum of matrix model ] ) in the limit in which @xmath170 with @xmath171 and @xmath172 fixed .
the equivalence is proved as follows .
we decompose the fields into the background corresponding to ( [ monopole background ] ) and the fluctuation as @xmath173 , where @xmath174 label the ( off - diagonal ) blocks .
note that @xmath175 is an @xmath176 matrix . then
, ( [ ym - higgs on s^2 ] ) is expanded around ( [ monopole background ] ) as @xmath177^{(s , t)}\right ) \nonumber\\ & \left .
\qquad\qquad\qquad\quad\qquad \times\left(\mu x_i^{(t , s)}+i\mu\epsilon_{ilm}l^{(q_{ts})}_lx_m^{(t , s ) } -\frac{i}{2}\epsilon_{ilm}[x_l , x_m]^{(t , s)}\right)\right ] , \label{ym - higgs on s^2 expanded around vacuum}\end{aligned}\ ] ] where @xmath178 we make a harmonic expansion of ( [ ym - higgs on s^2 expanded around vacuum ] ) by expanding the fluctuation in terms of the monopole vector spherical harmonics @xmath179 : @xmath180 where @xmath181 stands for the polarization , @xmath182 and @xmath183 .
the properties of the monopole spherical harmonics are analyzed and summarized in @xcite and references therein . substituting ( [ mode expansion of x ] ) into ( [ ym - higgs on s^2 expanded around vacuum ] ) yields @xmath184 , \label{mode expansion of ym - higgs on s^2}\end{aligned}\ ] ] where @xmath185 is defined by @xmath186 and we have used the equality @xmath187 similarly , decomposing the matrices into the background given by ( [ vacuum of matrix model ] ) and the fluctuation as @xmath188 leads to the theory around ( [ vacuum of matrix model ] ) : @xmath189^{(s , t)}\right ) \right.\nonumber\\ & \left .
\qquad\qquad\qquad\qquad\qquad \times\left(\mu x_i^{(t , s)}+i\mu\epsilon_{ilm}l_l\circ x_m^{(t , s ) } -\frac{i}{2}\epsilon_{ilm}[x_l , x_m]^{(t , s)}\right)\right ] , \label{matrix model expanded around vacuum}\end{aligned}\ ] ] where @xmath190 is defined by @xmath191}x_j^{(s , t ) } -x_j^{(s , t)}l_i^{[j_t]}.\end{aligned}\ ] ] the gauge symmetry of the above theory is expressed as @xmath192^{(s ,
t)}. \label{gauge symmetry}\end{aligned}\ ] ] we make a harmonic expansion of ( [ matrix model expanded around vacuum ] ) by expanding the fluctuation in terms of the fuzzy vector spherical harmonics @xmath193 defined in appendix a as @xmath194 @xmath195 in ( [ mode expansion of x 2 ] ) is an @xmath176 matrix .
since @xmath196 , @xmath168 plays the role of the ultraviolet cutoff .
note also that @xmath197 . substituting ( [ mode expansion of x 2 ] ) into ( [ matrix model expanded around vacuum ] ) yields @xmath198 , \label{mode expansion of matrix model}\end{aligned}\ ] ] where @xmath199 is defined in ( [ ce ] ) and we have used ( [ equalities for vector and spinor fuzzy spherical harmonics ] ) . in the @xmath170 limit , the ultraviolet cutoff goes to infinity and @xmath200 because the 6-j symbol behaves asymptotically for @xmath201 as @xcite @xmath202 namely , this limit corresponds to the commutative ( continuum ) limit of the fuzzy spheres .
hence , in the @xmath170 limit with @xmath203 and @xmath204 fixed , ( [ mode expansion of matrix model ] ) agrees with ( [ mode expansion of ym - higgs on s^2 ] ) .
we have proven our statement .
this equivalence is classical in the following sense .
the asymptotic formula ( [ asymptotic form of 6j ] ) holds for @xmath205 .
namely , the reduction ( [ reduction of ce ] ) is valid for @xmath206 .
thus the equivalence is true at tree level .
the loop effect may cause a deviation between the two theories quantum mechanically , since in the loop @xmath207 can be @xmath208 .
part of this deviation should be attributed to the uv / ir mixing .
suppose we restrict ourselves to the planar limit , in which @xmath209 with @xmath210 fixed .
then , this restriction and sixteen supersymmetries are probably sufficient to suppress the uv / ir mixing , namely the noncommutativity in the continuum limit . furthermore ,
as we discuss later , they completely stabilize the vacua of pwmm .
thus , the equivalence should also hold at the quantum level . indeed , the gravity duals of 2 + 1 sym on @xmath13 and pwmm proposed in @xcite support this conjecture @xcite . in the next section , we give an evidence that the uv / ir mixing does not exist .
next , we recall that the theory around a certain vacuum of 2 + 1 @xmath211 sym on @xmath13 with the orbifolding ( periodicity ) condition imposed is equivalent to @xmath212 @xmath0 sym on @xmath1 , which was shown in @xcite ( see also @xcite ) .
this is an extension of the matrix t - duality to that on a nontrivial fiber bundle , @xmath22 as a @xmath14 bundle over @xmath16 .
the vacuum of 2 + 1 sym on @xmath13 we take is given by ( [ monopole background ] ) in the @xmath170 limit with @xmath213 running from @xmath214 to @xmath215 , @xmath216 , @xmath217 and @xmath218 .
we decompose the fields on @xmath16 into the background and the fluctuation @xmath219 and impose the periodicity ( orbifolding ) condition on the fluctuation @xmath220 the fluctuations are gauge - transformed from the patch i to the patch ii as @xcite @xmath221 we make the fourier transformation for the fluctuations on each patch to construct the gauge field on the total space from : @xmath222 we see from ( [ gauge transformation of fluctuations ] ) that the left - hand side of ( [ fourier transformation ] ) is indeed independent of the patches . using ( [ fourier transformation ] ) , we obtain @xmath223 and so on .
then , we see that ( [ ym - higgs on s^2 expanded around vacuum ] ) equals @xmath224 ( [ ym on s^3 ] ) .
we divide an overall factor @xmath225 to extract a single period and obtain @xmath212 @xmath0 sym on @xmath1 . of course
, we can verify this equivalence by seeing the agreement of the harmonic expansions of the two theories .
we expand @xmath226 in terms of the vector spherical harmonics on @xmath22 , @xmath227 ( see @xcite ) : @xmath228 where @xmath229 run over all non - negative integers and half - integers . @xmath182 and @xmath183 are the spins for the two @xmath17 s of the isometry of @xmath22 .
note that the @xmath17 whose spin is @xmath230 is broken in ( [ ym - higgs on s^2 ] ) . by using the equality @xmath231
we can easily show that the harmonic expansion of ( [ ym on s^3 2 ] ) agrees with ( [ mode expansion of ym - higgs on s^2])@xmath232 in the present set - up with the correspondence @xmath233 namely , @xmath234 is identified with @xmath235 . combining the above two equivalences
, we see that the theory around ( [ vacuum of matrix model ] ) of pwmm in the @xmath236 limit , where @xmath213 runs from @xmath214 to @xmath215 , @xmath237 and @xmath217 , is equivalent to @xmath212 @xmath0 sym on @xmath1 if @xmath238 is fixed to @xmath239 , the periodicity condition is imposed on the fluctuation and the overall factor @xmath240 is divided .
the relationship between @xmath0 sym on @xmath1 and 2 + 1 sym on @xmath13 is again classical for the following reason , and so is the relationship between @xmath0 sym on @xmath1 and pwmm . in order to construct a well - defined quantum theory ,
we need to introduce an ultraviolet cutoff to the momentum of the fiber direction , which corresponds to @xmath241 in ( [ fourier transformation ] ) .
namely , we should consider finite - size matrices by making @xmath242 run from @xmath243 to @xmath244 with @xmath63 an ultraviolet cutoff .
in this situation , however , the periodicity condition ( [ periodicity condition ] ) is not compatible with the gauge invariance . in order to resolve this problem , referring to the result for the modified reduced model in the previous section , we discard the periodicity condition and take the limit in which @xmath245 . in this case
, the @xmath24 in the previous section corresponds to the fiber direction of @xmath22 viewed as a @xmath14 bundle over @xmath16 .
our theory is a one - dimensional massive theory , so the instability discussed in the last part of the previous section is suppressed .
moreover , the @xmath11 symmetry preserved by the vacuum ( [ vacuum of matrix model ] ) completely stabilizes the vacuum .
indeed , the result in @xcite ensures the perturbative stability , and it is easily seen from the result in @xcite that the nonperturbative instability via the tunneling to other vacua of pwmm caused by the instantons is suppressed in the @xmath50 limit .
thus , we do not need any quenching , and we can respect the gauge symmetry and the @xmath11 symmetry , namely half of supersymmetries of @xmath0 sym on @xmath1 , simultaneously . indeed , ( [ gauge symmetry ] ) should correspond to the gauge symmetry of @xmath0 sym on @xmath1 .
the noncommutativity probably vanishes in the continuum limit as mentioned before . to summarize
, we propose a nonperturbative definition of @xmath0 sym on @xmath1 as follows .
we consider the theory around ( [ vacuum of matrix model ] ) of pwmm with @xmath246 we take the limit in which @xmath247 then , we obtain the t hooft ( planar ) limit of @xmath0 sym on @xmath1 . the condition @xmath248 can be relaxed to @xmath249 that should be required to obtain the continuum spheres . for simplicity of the analysis
, we adopt the stronger condition @xmath250 in this paper .
the result should not depend on how to take the limit .
our formulation preserves the gauge symmetry and the @xmath11 symmetry .
it is , in particular , remarkable that it preserves sixteen supersymmetries .
we need to check the restoration of the superconformal symmetry @xmath20 to verify that our formulation does work well .
the restoration of the superconformal symmetry should imply that no uv / ir mixing occur . in the next section ,
we give some evidences for the restoration of the superconformal symmetry .
in this section , we perform a perturbative expansion of the theory around ( [ vacuum of matrix model ] ) of pwmm . in the beginning
, we do not assume ( [ background ] ) or ( [ limit ] ) .
we make a replacement @xmath251 in ( [ action_of_pwmm ] ) .
we adopt the feynman - type gauge and add the following gauge fixing and fadeev - popov terms to the action : @xmath252)^2 + i\bar{c}\partial_td_tc+\mu\bar{c}[l_i , i\mu[l_i , c]-i[x_i , c]]\right ) .
\label{gauge fixing and f .-
terms}\end{aligned}\ ] ] the resultant gauge - fixed action is written down in ( [ gauge fixed action ] ) in appendix b. the mode expansion of the fields is given in ( [ mode_expansion_in_pwmm ] ) , which of course includes ( [ mode expansion of x 2 ] ) . the harmonic expansion of the gauge - fixed action is given in ( [ harmonic_expansion_of_s_free ] ) , ( [ harmonic_expansion_of_s_gauge_int ] ) and ( [ harmonic_expansion_of_s_matter_int ] ) which are a counterpart of ( [ mode expansion of matrix model ] ) .
one can read off the propagators from ( [ harmonic_expansion_of_s_free ] ) as in ( [ propagators ] ) and the vertices from ( [ harmonic_expansion_of_s_gauge_int ] ) and ( [ harmonic_expansion_of_s_matter_int ] ) . .
the wavy line represents the propagator of @xmath89 .
the dotted line represents the propagator of the ghost .
the solid line represents the propagator of @xmath90 .
the dashed line represents the propagator of @xmath91.,height=188 ] first , we calculate the 1-loop contribution to the tadpoles .
the only possibly nonzero contribution is the truncated 1-point function for @xmath253 , where @xmath58 run from 1 to @xmath154 and @xmath254 is dual to @xmath75 in the fourier transformation .
this quantity takes the form @xmath255 there are five 1-loop diagrams for this 1-point function as shown in fig .
2 . note that all these diagrams are planar ones .
the diagrams @xmath256 and @xmath257 completely cancel each other .
below we list the value of @xmath258 for each of the remaining diagrams .
@xmath259 where @xmath260 .
the @xmath261-@xmath165 symbol in the above expressions can be written explicitly : @xmath262 by using ( [ explicit 6j ] ) and @xmath263 , we sum up the contributions to the tadpole : @xmath264 where @xmath265 the first term in ( [ f(r ) ] ) comes from @xmath266 and @xmath267 while the second term from @xmath268 . the asymptotic behavior of @xmath269 for large @xmath25 , @xmath270 tells us that in the @xmath236 limit @xmath271 we find no @xmath168-dependent divergences .
we see that @xmath272 this means that the vev of @xmath273 that corresponds to the one - point function for the overall @xmath14 field on @xmath13 indeed vanishes .
this is consistent with the fact that it is a free field decoupled from the other fields when in the theory around ( [ monopole background ] ) one takes the feynman - like gauge , to which the gauge corresponding to ( [ gauge fixing and f .-
p . terms ] ) reduces naively in the @xmath236 limit .
one can easily verify from ( [ f(r ) ] ) that if there is no supersymmetry , the vev of the one - point function for the overall @xmath14 field does not vanish .
note that this happens even with the restriction to the planar limit since all the tadpole diagrams in fig .
2 are planar . in @xcite ,
the same phenomenon was observed in a bosonic gauge theory on the fuzzy sphere in the continuum limit and interpreted as the uv / ir mixing . on the other hand , by shifting @xmath274 in ( [ vacuum of matrix model ] ) , one can always cancel the vev of the one point function for the overall u(1 ) field and might obtain the commutative gauge theory .
however , in any case , we can not follow this prescription because it breaks supersymmetry . here
we have obtained an evidence that in our case the uv / ir mixing is avoided , that is , the noncommutativity vanishes in the continuum limit , in a way compatible with supersymmetry .
if we consider the theory around ( [ vacuum of matrix model ] ) with ( [ background ] ) and ( [ limit ] ) that would realize @xmath275 sym on @xmath1 , we find no @xmath63-dependent divergences in ( [ tadpole ] ) .
furthermore , ( [ tadpole ] ) vanishes for fixed @xmath213 in the @xmath276 due to the summation over @xmath75 . in this case , the gauge corresponding to ( [ gauge fixing and f .-
p . terms ] ) reduces naively in the limit ( [ limit ] ) to the feynman gauge in @xmath0 sym on @xmath1 .
the isometry of @xmath22 , @xmath97 , is manifest in @xmath0 sym on @xmath1 with the feynman gauge , and all the tadpoles vanish due to this isometry .
the symmetry corresponding to one of the above two @xmath17 s already exists a priori in our theory , while the other does not .
vanishing of ( [ tadpole ] ) is a signal for the restoration of the @xmath99 symmetry in the continuum limit ( [ limit ] ) .
if this restoration and the vanishing of the noncommutativity is indeed the case , we obtain a commutative gauge theory with sixteen supersymmetries on @xmath1 .
this theory should be nothing but @xmath0 sym on @xmath1 unless we perform any extra fine - tunings .
thus we have found an evidence that the superconformal symmetry is restored and our formalism does work well . .
the curly line represents the propagator of @xmath140 .
the wavy line represents the propagator of @xmath89 .
the solid line represents the propagator of @xmath90 .
the dashed line represents the propagator of @xmath91.,title="fig:",width=113 ] @xmath277 [ fig : a1 ] .
the curly line represents the propagator of @xmath140 .
the wavy line represents the propagator of @xmath89 .
the solid line represents the propagator of @xmath90 .
the dashed line represents the propagator of @xmath91.,title="fig:",width=113 ] @xmath278 [ fig : f2 ] .
the curly line represents the propagator of @xmath140 .
the wavy line represents the propagator of @xmath89 .
the solid line represents the propagator of @xmath90 .
the dashed line represents the propagator of @xmath91.,title="fig:",width=113 ] @xmath279 [ fig : a6 ] next , we calculate the fermion self - energy in the theory around ( [ vacuum of matrix model ] ) with ( [ background ] ) and ( [ limit ] ) at the one - loop level , and compare the result in @xmath0 sym on @xmath1 .
the fermion self - energy is given by the truncated two - point function @xmath280 , and this takes the form @xmath281 the diagrams which contribute to the fermion self - energy at the one - loop order are shown in fig .
we list the value of @xmath282 for each diagram in appendix c. we set the external indices to specific values to calculate the leading contribution in the continuum limit ( [ limit ] ) : @xmath283 , @xmath284 and @xmath285 .
the divergent part of @xmath286 for each diagram is evaluated as @xmath287 { \nonumber \\}&\sim -\frac{4ig^2}{\mu^2}n\left(-l-\frac{1}{4}\right)\ln t -\frac{4ig^2}{\mu^2}n\left(-\frac{l}{2}+\frac{5}{8}\right)\ln t , { \nonumber \\}[3 mm ] & ( f - b)_{\kappa=\kappa'=1,j=0}{\nonumber \\}&=\frac{ig^2}{\mu^2}n \sum_{u=-\frac{t}{2}}^{\frac{t}{2 } } \sum_{r_1=|\frac{u}{2}|}^{n_0 - 1+\frac{u}{2 } } { \nonumber \\}&\quad \times \biggl [ \sqrt{\frac{r_1 + 1}{r_1 } } \cdot \frac{1}{l-\{r_1+\sqrt{r_1(r_1 + 1)}+\frac{3}{4}\ } } + \sqrt{\frac{r_1}{r_1 + 1}}\cdot \frac{1}{l+\{r_1+\sqrt{r_1(r_1 + 1)}+\frac{1}{4}\ } } \biggr ] { \nonumber \\}&\sim -\frac{4ig^2}{\mu^2}n\left(\frac{l}{2}+\frac{3}{8}\right)\ln t , { \nonumber \\}[3 mm ] & ( f - c)_{\kappa=\kappa'=1,j=0}{\nonumber \\}&=-\frac{12ig^2}{\mu^2}n \sum_{u=-\frac{t}{2}}^{\frac{t}{2 } } \sum_{r_1=|\frac{u}{2}|}^{n_0 - 1+\frac{u}{2 } } \biggl [ \frac{r_1 + 1}{2r_1 + 1}\cdot \frac{1}{l+(2r_1+\frac{5}{4 } ) } + \frac{r_1}{2r_1 + 1}\cdot \frac{1}{l-(2r_1+\frac{3}{4 } ) } \biggr]{\nonumber \\}&\sim -\frac{12ig^2}{\mu^2}n\left(-l+\frac{3}{4}\right)\ln t , \label{r=0 case}\end{aligned}\ ] ] where @xmath288 .
note that we find no @xmath168-dependent divergences in each diagram .
this is consistent with the fact that ( 2 + 1)-dimensional gauge theory is super renormalizable .
we find that the divergence in @xmath63 is logarithmic in each diagram .
this is again consistent with the fact that the fermion self - energy has only the logarithmic divergence in four dimensions . in @xmath0 sym on @xmath1 ,
the fermion self - energy is given by the two point function @xmath289 , where @xmath229 and @xmath290 are the spins for the two @xmath17 s of the isometry of @xmath22 .
in the feynman gauge to which the gauge corresponding to ( [ gauge fixing and f .- p . terms ] ) reduces naively in the limit ( [ limit ] ) , it takes the form @xmath291 by using the technique in @xcite , we evaluate each diagram in fig
. 3 in the feynman gauge . as in @xcite
, we introduce naive cutoffs for the angular momenta and evaluate the divergent part of @xmath292 for each diagram as @xmath293,{\nonumber \\}&(f - b)_{s^3}{\nonumber \\}&=\frac{4ig^2}{\mu^2 } n \times \bigl(-\frac{1}{2}\bigr ) \left\{l+\left(j+\frac{3}{4}\right)\right\}\ln ( 2\lambda),{\nonumber \\}&(f - c)_{s^3}{\nonumber \\}&=\frac{4ig^2}{\mu^2 } n \times 3 \times \left\{l-\left(j+\frac{3}{4}\right)\right\ } \ln ( 2\lambda ) .
\label{fs in s3}\end{aligned}\ ] ] the cutoffs for the angular momenta break the gauge symmetry and supersymmetry .
nevertheless , the coefficient of the logarithmic divergent part for each diagram has a universal meaning .
we find that ( [ r=0 case ] ) completely agrees with the continuum case ( [ fs in s3 ] ) under the identification @xmath294 .
this fact provides an evidence that in the continuum limit ( [ limit ] ) our theory reproduces @xmath0 sym on @xmath1 .
furthermore , we put @xmath295 , @xmath296 , @xmath297 and @xmath298 . for @xmath299 , the fermion self - energy in @xmath0 sym on @xmath1 in the feynman gauge vanishes due to the @xmath99 symmetry , while it does not vanish a priori in our theory because the @xmath99 symmetry is not manifest in our theory .
however , it turns out that there is no divergence in @xmath300 for each diagram and it vanishes due to the summation over the blocks ( the summation over @xmath301 in ( [ r=0 case ] ) ) .
this is another evidence for the restoration of the @xmath99 symmetry , which implies the restoration of the superconformal symmetry if the noncommutativity vanishes .
in this paper , we proposed a nonperturbative definition of the t hooft limit of @xmath0 sym . we realized @xmath0 sym on @xmath1 as the theory around a vacuum of pwmm .
the size of matrices plays a role of the ultraviolet cutoff .
our formulation preserves the gauge symmetry and the @xmath11 symmetry .
@xmath11 is a subgroup of @xmath20 which is the superconformal symmetry possessed by @xmath0 sym on @xmath1 . in particular ,
sixteen supersymmetries among thirty - two supersymmetries in @xmath0 sym on @xmath1 are preserved in our formulation .
we calculated the tadpoles and the fermion self - energy at the one - loop order .
the results give some evidences that the uv / ir mixing does not exist and the @xmath20 symmetry is restored in the continuum limit so that our formulation does work well .
we should collect more evidences for the restoration of the superconformal symmetry .
higher - loop calculations are needed .
of course , the restoration should eventually be confirmed nonperturbatively .
the numerical simulation for our theory can be performed based on the method in @xcite .
unfortunately , the size of matrices available at present seems too small for the continuum limit ( [ limit ] ) .
it is now possible to perform the numerical simulation for the theory around ( [ vacuum of matrix model ] ) , for instance , with @xmath213 taking only 1 and @xmath302 and to take the continuum limit that would realize the t hooft limit of 2 + 1 sym on @xmath13 .
then , we can compare the results of the numerical simulation with those predicted by the gravity dual @xcite to check whether the uv / ir mixing is avoided .
anyway , we believe that the numerical simulation for our theory will be possible in the near future .
it is also desirable to develop an analytical ( approximation ) method that enables us to analyze our theory at strong coupling . by using the result in @xcite
, we can easily construct as a physical observable the wilson loop in our theory that corresponds to the ordinary wilson loop in @xmath0 sym on @xmath1 .
we can also consider the bps wilson loop @xcite by including the matter degrees of freedom in the loop .
it is important to calculate the vev of these wilson loops in our theory analytically and numerically in the strong coupling regime and compare the results with the predictions of the gravity dual @xcite .
we also hope to find the integrable structure of @xmath0 sym at strong coupling by analyzing our theory .
we would like to thank s. iso , h. kawai , y. kitazawa , j. nishimura , h. suzuki and k. yoshida for discussions .
the work of g.i . and s.s .
is supported in part by the jsps research fellowship for young scientists .
the work of a.t . is supported in part by grant - in - aid for scientific research ( no .
19540294 ) from the ministry of education , culture , sports , science and technology .
in this appendix , we summarize the properties of the fuzzy spherical harmonics analyzed and summarized in @xcite ( see also @xcite ) . let us consider @xmath303 rectangular complex matrices .
such matrices are generally expressed as @xmath304 we can define linear maps @xmath190 , which map the set of @xmath303 rectangular complex matrices to itself , by their operation on the basis : @xmath305}|jr\rangle\langle j'r'|- @xmath164}$ ] are the spin @xmath165 representation matrices of the @xmath17 generators .
@xmath306 satisfy the @xmath17 algebra @xmath307 = i\epsilon_{ijk}l_k \circ$ ] .
we change the basis of the rectangular matrices from the above basis @xmath308 to the new basis which is called the fuzzy spherical harmonics : @xmath309 where @xmath168 is a positive constant , which is taken to be an integer as an ultraviolet cutoff in section 3 . for a fixed @xmath229
, the fuzzy spherical harmonics also form the basis of the spin @xmath229 irreducible representation of @xmath17 which is generated by @xmath306 @xmath310 the hermitian conjugates of the fuzzy spherical harmonics are evaluated as @xmath311 the fuzzy spherical harmonics satisfy the orthonormality condition under the following normalized trace : @xmath312 where @xmath313 stands for the trace over @xmath314 matrices .
the trace of three fuzzy spherical harmonics is given by @xmath315 where the last factor of the last line in the above equation is the @xmath261-@xmath165 symbol .
we also introduce the vector fuzzy spherical harmonics @xmath316 and the spinor fuzzy spherical harmonics @xmath317 , where @xmath181 takes -1,0,1 and @xmath318 takes -1 and 1 .
they are defined in terms of the scalar spherical harmonics as @xmath319 where @xmath320 and @xmath321 .
the unitary matrix @xmath322 is given by @xmath323 the vector fuzzy spherical harmonics and the spinor fuzzy spherical harmonics satisfy @xmath324 their hermitian conjugate are @xmath325 and they satisfy the following orthonormal relations : @xmath326 we can evaluate the trace of the three fuzzy spherical harmonics , including the vector harmonics and/or the spinor harmonics , as follows : @xmath327 @xmath328 @xmath329 @xmath330
in this appendix , we make a harmonic expansion of the theory around ( [ vacuum of matrix model ] ) of pwmm .
this harmonic expansion enables us to perform the perturbative calculation of the theory in section 4 .
first , we make a replacement @xmath331 in ( [ action_of_pwmm ] ) and add the gauge fixing and the fadeev - popov terms ( [ gauge fixing and f .-
p . terms ] ) .
the resultant action is @xmath332 where @xmath333 ^ 2 -\frac{1}{2}(\partial_t a_t)^2 \right .
\nonumber\\ & \qquad\qquad\left .
-\frac{\mu^2}{2}(x_i+i\epsilon_{ijk}[l_j , x_k])^2 + \frac{\mu^2}{2}[l_i , x_i]^2 + i\bar{c}\partial_t^2c + i\mu^2 \bar{c}[l_i,[l_i , c ] ] \right ) , \label{s_gauge_free}\displaybreak[0 ] \\ s^{gauge}_{pw , int } & = \frac{1}{g_{pw}^2\mu^2}\int dt \ ; \mbox{tr } \biggl ( -i(\partial_t x_i)[a_t , x_i]-\mu[a_t , x_i][l_i , a_t ] -\frac{1}{2}[a_t , x_i]^2 { \nonumber \\}&\qquad\qquad + i\mu\epsilon_{ijk}(x_i+i\epsilon_{ilm}[l_l , x_m])x_j x_k + \frac{1}{2}\epsilon_{ijk}\epsilon_{ilm}x_j x_k x_l x_m { \nonumber \\}&\qquad \qquad -i\mu [ l_i,\bar{c}][c , x_i ] -\partial_t\bar{c}[a_t , c ] \biggr ) , \label{s_gauge_int}\displaybreak[0]\\ s^{matter}_{pw , free } & = \frac{1}{g_{pw}^2\mu^2}\int dt \ ; \mbox{tr } \left ( \frac{1}{2}\partial_t \phi_{ab}\partial_t \phi^{ab } -\frac{\mu^2}{8}\phi_{ab}\phi^{ab } + \frac{\mu^2}{2}[l_i,\phi_{ab}][l_i,\phi^{ab } ] \right . \nonumber\\ & \qquad\qquad\left . -i\psi_a^\dagger \partial_t
\psi^a -\mu\psi_a^\dagger ( \frac{3}{4}\psi^a+\sigma^i [ l_i,\psi^a]\big ) \right ) , \label{s_matter_free}\displaybreak[0]\\ s^{matter}_{pw , int } & = \frac{1}{g_{pw}^2\mu^2}\int dt \ ; \mbox{tr } \left ( -i(\partial_t \phi_{ab})[a_t,\phi^{ab } ] -\frac{1}{2}[a_t,\phi_{ab}][a_t,\phi^{ab } ] -\mu[l_i,\phi_{ab}][x_i,\phi^{ab } ] \right .
\qquad\qquad + \frac{1}{2}[x_i,\phi_{ab}][x_i,\phi^{ab } ] + \frac{1}{4}[\phi_{ab},\phi_{cd}][\phi^{ab},\phi^{cd } ] + \psi_a^\dagger [ a_t,\psi^a ] \nonumber\\ & \qquad\qquad\left .
+ \psi_a^\dagger \sigma^i [ x_i,\psi^a ] + \psi_a^\dagger \sigma^2[\phi^{ab},({\psi}_b^\dagger)^t ] -(\psi^a)^t \sigma^2 [ \phi_{ab},\psi^b ] \right).\label{s_matter_int}\end{aligned}\ ] ] we make a mode expansion of the @xmath174 blocks of the fields in terms of the fuzzy spherical harmonics defined in appendix a : @xmath334\\ c^{(s , t ) } & = \sum_{j=|j_s - j_t|}^{j_s+j_t}\sum_{m =- j}^{j}c^{(s , t)}_{jm } \otimes\hat{y}_{jm(j_sj_t ) } , \qquad \bar{c}^{(s , t ) } = \sum_{j=|j_s - j_t|}^{j_s+j_t}\sum_{m =- j}^{j}\bar{c}^{(s , t)}_{jm}\otimes \hat{y}_{jm(j_sj_t ) } , \nonumber\displaybreak[0]\\ \psi^{a(s , t ) } & = \sum_{\kappa=\pm1}\sum_{\tilde{u}=|j_s - j_t|}^{j_s+j_t}\sum_{m =- u}^{u } \psi_{jm\kappa}^{a(s , t)}\otimes\hat{y}_{jm(j_sj_t)}^\kappa \nonumber\\ & = \sum_{j=|j_s - j_t|}^{j_s+j_t}\sum_{m =- j-\frac{1}{2}}^{j+\frac{1}{2 } } \psi_{jm1}^{a(s , t)}\otimes\hat{y}_{jm(j_sj_t)}^1 + \sum_{j=|j_s - j_t|-\frac{1}{2}}^{j_s+j_t-\frac{1}{2}}\sum_{m =- j}^{j
} \psi_{jm-1}^{a(s , t)}\otimes\hat{y}_{jm(j_sj_t)}^{-1 } , \nonumber\displaybreak[0]\\ \psi_a^{(t , s)\dagger } & = \sum_{\kappa=\pm1}\sum_{\tilde{u}=|j_s - j_t|}^{j_s+j_t}\sum_{m =- u}^{u } \psi_{a , jm\kappa}^{(t , s)\dagger}\otimes\hat{y}_{jm(j_tj_s)}^{\kappa\dagger } \nonumber\\ & = \sum_{j=|j_s - j_t|}^{j_s+j_t}\sum_{m =- j-\frac{1}{2}}^{j+\frac{1}{2 } } \psi_{a , jm1}^{(t , s)\dagger}\otimes\hat{y}_{jm(j_tj_s)}^{1\dagger } + \sum_{j=|j_s - j_t|-\frac{1}{2}}^{j_s+j_t-\frac{1}{2}}\sum_{m =- j}^{j } \psi_{a , jm-1}^{(t , s)\dagger}\otimes\hat{y}_{jm(j_tj_s)}^{-1\dagger } , \nonumber\displaybreak[0]\\ x_i^{(s , t ) } & = \sum_{\rho=-1}^{1}\sum_{\tilde{q}=|j_s - j_t|}^{j_s+j_t}\sum_{m =- q}^{q } x_{jm\rho}^{(s , t)}\otimes\hat{y}{}_{jm(j_sj_t)i}^\rho \nonumber\\ & = \sum_{j=|j_s - j_t|}^{j_s+j_t}\sum_{m =- j-1}^{j+1 } x_{jm1}^{(s , t)}\otimes\hat{y}{}_{jm(j_sj_t)i}^1 + \sum_{j=|j_s - j_t|}^{j_s+j_t}\sum_{m =- j}^{j } x_{jm0}^{(s , t)}\otimes\hat{y}{}_{jm(j_sj_t)i}^0 \nonumber\\ & \;\;\;\;\;+\sum_{j=|j_s - j_t|-1}^{j_s+j_t-1}\sum_{m =- j}^{j } x_{jm-1}^{(s , t)}\otimes\hat{y}{}_{jm(j_sj_t)i}^{-1}. \label{mode_expansion_in_pwmm}\end{aligned}\ ] ] note that the modes in the right - hand sides of the equations in ( [ mode_expansion_in_pwmm ] ) are @xmath176 matrices . by using the properties of the fuzzy spherical harmonics summarized in appendix a , we rewrite the free part of ( [ gauge fixed action ] ) in terms of the modes as follows : @xmath335 where @xmath336 we can read off the propagators for the fourier transforms of the fields from ( [ harmonic_expansion_of_s_free ] ) as @xmath337{\nonumber \\}&\langle b_{jm}^{(s , t)}(p)_{ij } b_{j'm'}^{(s',t')}(p')_{kl}\rangle = ( -1)^{m-(j_s - j_t)+1}\delta_{jj'}\delta_{m\,-m ' } \delta_{st'}\delta_{ts'}\delta_{il}\delta_{jk } 2\pi\delta(p+p')\frac{ig^2}{p^2-\mu^2 j(j+1 ) } , \displaybreak[0]{\nonumber \\}&\langle c_{jm}^{(s , t)}(p)_{ij } \bar{c}_{j'm'}^{(s',t')}(p')_{kl}\rangle = ( -1)^{m-(j_s - j_t)}\delta_{jj'}\delta_{m\,-m ' } \delta_{st'}\delta_{ts'}\delta_{il}\delta_{jk } 2\pi\delta(p+p')\frac{g^2}{-p^2+\mu^2 j(j+1 ) } , \displaybreak[0]{\nonumber \\}&\langle \phi_{ab , jm}^{(s , t)}(p)_{ij } \phi_{a'b',j'm'}^{(s',t')}(p')_{kl } \rangle = \frac{1}{2}\epsilon_{aba'b ' } ( -1)^{m-(j_s - j_t)}\delta_{jj'}\delta_{m\,-m ' } \delta_{st'}\delta_{ts'}\delta_{il}\delta_{jk } 2\pi\delta(p+p')\frac{ig^2}{p^2-{\omega_j^\phi}^2 } , \displaybreak[0]{\nonumber \\}&\langle \psi_{jm\kappa}^{a(s , t)}(p)_{ij } \psi_{a',j'm'\kappa'}^{(s',t')\dagger}(p')_{kl}\rangle = \delta_{jj'}\delta_{mm'}\delta_{\kappa\kappa'}\delta^a_{a ' } \delta_{ss'}\delta_{tt'}\delta_{il}\delta_{jk } 2\pi\delta(p - p')\frac{ig^2(p+\kappa\omega_j^\psi)}{p^2-{\omega_j^\psi}^2}. \label{propagators}\end{aligned}\ ] ] the gauge part of the interaction terms in ( [ gauge fixed action ] ) is rewritten as @xmath338\\ & + \mu ( \sqrt{j_2(j_2 + 1)}\hat{\mathcal{d}}_{j_1m_1(j_sj_t)\;j_2m_2(j_tj_u)0\;j_3m_3(j_uj_s)\rho_3 } -\sqrt{j_1(j_1 + 1)}\hat{\mathcal{d}}_{j_2m_2(j_tj_u)\;j_3m_3(j_uj_s)\rho_3\;j_1m_1(j_sj_t)0 } ) \nonumber\\ & \quad\times b_{j_1m_1}^{(s , t ) } b_{j_2m_2}^{(t , u ) } x_{j_3m_3\rho_3}^{(u , s ) } \nonumber\displaybreak[0]\\ & + ( -1)^{m-(j_s - j_u)+1 } \nonumber\\ & \quad\times(\hat{\mathcal{d}}_{j_1m_1(j_sj_t)\;j_2m_2(j_tj_u)\rho_2\;j - m(j_uj_s)\rho } \hat{\mathcal{d}}_{j_4m_4(j_vj_s)\;jm(j_sj_t)\rho\;j_3m_3(j_uj_v)\rho_3 } b_{j_1m_1}^{(s , t ) } x_{j_2m_2\rho_2}^{(t , u ) } x_{j_3m_3\rho_3}^{(u , v ) } b_{j_4m_4}^{(v , s ) } \nonumber\\ & \quad-\hat{\mathcal{d}}_{j_1m_1(j_sj_t)\;j_2m_2(j_tj_u)\rho_2\;j - m(j_uj_s)\rho } \hat{\mathcal{d}}_{j_3m_3(j_uj_v)\;j_4m_4(j_vj_s)\rho_4\;jm(j_sj_u)\rho } b_{j_1m_1}^{(s , t ) } x_{j_2m_2\rho_2}^{(t , u ) } b_{j_3m_3}^{(u , v ) } x_{j_4m_4\rho_4}^{(v , s ) } ) \nonumber\displaybreak[0]\\ & + i\mu\rho_1(\rho_1 + 1 ) \hat{\mathcal{e}}_{j_1m_1(j_sj_t)\rho_1\;j_2m_2(j_tj_u)\rho_2\;j_3m_3(j_uj_s)\rho_3 } x_{j_1m_1\rho_1}^{(s , t ) } x_{j_2m_2\rho_2}^{(t , u ) } x_{j_3m_3\rho_3}^{(u , s ) } \nonumber\displaybreak[0]\\ & + \frac{1}{2}(-1)^{m-(j_s - j_u)+1 } \hat{\mathcal{e}}_{j - m(j_uj_s)\rho\;j_1m_1(j_sj_t)\rho_1\;j_2m_2(j_tj_u)\rho_2 } \hat{\mathcal{e}}_{jm(j_sj_u)\rho\;j_3m_3(j_uj_v)\rho_3\;j_4m_4(j_vj_s)\rho_4 } \nonumber\\ & \quad\times x_{j_1m_1\rho_1}^{(s , t ) } x_{j_2m_2\rho_2}^{(t , u ) } x_{j_3m_3\rho_3}^{(u , v ) } x_{j_4m_4\rho_4}^{(v , s ) } \nonumber\displaybreak[0]\\ & -i\mu\sqrt{j_3(j_3 + 1)}\ ; ( \hat{\mathcal{d}}_{j_2m_2(j_sj_t)\;j_3m_3(j_tj_u)0\;j_1m_1(j_uj_s)\rho_1 } c_{j_2m_2}^{(s , t ) } \bar{c}_{j_3m_3}^{(t , u ) } x_{j_1m_1\rho_1}^{(u , s ) } \nonumber\\ & \quad-\hat{\mathcal{d}}_{j_2m_2(j_sj_t)\;j_1m_1(j_tj_u)\rho_1\;j_3m_3(j_uj_s)0 } c_{j_2m_2}^{(s , t ) } x_{j_1m_1\rho_1}^{(t , u ) } \bar{c}_{j_3m_3}^{(u , s ) } ) \nonumber\displaybreak[0]\\ & + \hat{\mathcal{c}}_{j_1m_1(j_sj_t)\;j_2m_2(j_tj_u)\;j_3m_3(j_uj_s ) } b_{j_1m_1}^{(s , t ) } ( \partial_t \bar{c}_{j_2m_2}^{(t , u ) } c_{j_3m_3}^{(u , s ) } + c_{j_2m_2}^{(t , u)}\partial_t \bar{c}_{j_3m_3}^{(u , s ) } ) \big ] .
\label{harmonic_expansion_of_s_gauge_int}\end{aligned}\ ] ] the matter part of the interaction terms in ( [ gauge fixed action ] ) is rewritten as @xmath339\\ & + \frac{1}{2}\epsilon^{abcd } \hat{\mathcal{c}}^{jm(j_sj_u)}_{j_1m_1(j_sj_t)\;j_2m_2(j_tj_u ) } \hat{\mathcal{c}}_{jm(j_sj_u)\;j_3m_3(j_uj_v)\;j_4m_4(j_vj_s ) } \nonumber\\ & \quad\times ( b_{j_1m_1}^{(s , t ) } b_{j_2m_2}^{(t , u ) } \phi_{ab , j_3m_3}^{(u , v ) } \phi_{cd , j_4m_4}^{(v , s ) } -b_{j_1m_1}^{(s , t ) } \phi_{ab , j_2m_2}^{(t , u ) } b_{j_3m_3}^{(u , v ) } \phi_{cd , j_4m_4}^{(v , s ) } ) \nonumber\displaybreak[0]\\ & -\frac{\mu}{2}\epsilon^{abcd } ( \sqrt{j_2(j_2 + 1 ) } \hat{\mathcal{d}}_{j_1m_1(j_sj_t)\;j_2m_2(j_tj_u)0\;j_3m_3(j_uj_s)\rho_3 } \nonumber\\ & \quad-\sqrt{j_1(j_1 + 1 ) } \hat{\mathcal{d}}_{j_2m_2(j_tj_u)\;j_3m_3(j_uj_s)\rho_3\;j_1m_1(j_sj_t)0 } ) \phi_{ab , j_1m_1}^{(s , t ) } \phi_{cd , j_2m_2}^{(t , u ) } x_{j_3m_3\rho_3}^{(u , s ) } \nonumber\displaybreak[0]\\ & + \frac{1}{2}\epsilon^{abcd}(-1)^{m-(j_s - j_u)+1 } \nonumber\\ & \quad\times ( \hat{\mathcal{d}}_{j_4m_4(j_vj_s)\;jm(j_sj_u)\rho\;j_3m_3(j_uj_v)\rho_3 } \hat{\mathcal{d}}_{j_2m_2(j_tj_u)\;j - m(j_uj_s)\rho\;j_1m_1(j_sj_t)\rho_1 } x_{j_1m_1\rho_1}^{(s , t ) } \phi_{ab , j_2m_2}^{(t , u ) } x_{j_3m_3\rho_3}^{(u , v ) } \phi_{cd , j_4m_4}^{(v , s ) } \nonumber\\ & \quad -\hat{\mathcal{d}}_{j_4m_4(j_uj_v)\;j_3m_3(j_vj_s)\rho_3\;jm(j_sj_u)\rho } \hat{\mathcal{d}}_{j_2m_2(j_tj_u)\;j - m(j_uj_s)\rho\;j_1m_1(j_sj_t)\rho_1 } x_{j_1m_1\rho_1}^{(s , t ) } \phi_{ab , j_2m_2}^{(t , u ) } \phi_{cd , j_4m_4}^{(u , v ) } x_{j_3m_3\rho_3}^{(v , s ) } ) \nonumber\displaybreak[0]\\ & + \frac{1}{8}\epsilon^{abef}\epsilon^{cdgh } \hat{\mathcal{c}}^{jm(j_sj_u)}_{j_1m_1(j_sj_t)\;j_2m_2(j_tj_u ) } \hat{\mathcal{c}}_{jm(j_sj_u)\;j_3m_3(j_uj_v)\;j_4m_4(j_vj_s ) } \nonumber\\ & \quad\times ( \phi_{ab , j_1m_1}^{(s , t ) } \phi_{cd , j_2m_2}^{(t , u ) } \phi_{ef , j_3m_3}^{(u , v ) } \phi_{gh , j_4m_4}^{(v , s ) } -\phi_{ab , j_1m_1}^{(s , t ) } \phi_{ef , j_2m_2}^{(t , u ) } \phi_{cd , j_3m_3}^{(u , v ) } \phi_{gh , j_4m_4}^{(v , s ) } ) \nonumber\displaybreak[0]\\ & + \left((-1)^{m_3-(j_s - j_u)+\frac{\kappa_1-\kappa_2}{2 } } \hat{\mathcal{f}}^{j_2-m_2(j_tj_u)\kappa_2}_{j_1-m_1(j_tj_s)\kappa_1\;j_3m_3(j_sj_u ) } \psi_{a , j_1m_1\kappa_1}^{(s , t)\dagger } b_{j_3m_3}^{(s , u ) } \psi_{j_2m_2\kappa_2}^{a(u , t ) } \right.\nonumber\\ & \quad\left .
-\hat{\mathcal{f}}^{j_1m_1(j_sj_t)\kappa_1}_{j_2m_2(j_sj_u)\kappa_2\;j_3m_3(j_uj_t ) } \psi_{a , j_1m_1\kappa_1}^{(s , t)\dagger } \psi_{j_2m_2\kappa_2}^{a(s , u ) } b_{j_3m_3}^{(u , t ) } \right)\nonumber\displaybreak[0]\\ & -\left((-1)^{m_3-(j_s - j_u)+\frac{\kappa_1-\kappa_2}{2 } } \hat{\mathcal{g}}^{j_2-m_2(j_tj_u)\kappa_2}_{j_1-m_1(j_tj_s)\kappa_1\;j_3m_3(j_sj_u)\rho_3 } \psi_{a , j_1m_1\kappa_1}^{(s , t)\dagger } x_{j_3m_3\rho_3}^{(s , u ) } \psi_{j_2m_2\kappa_2}^{a(u , t ) } \right.\nonumber\\ & \quad\left .
+ \hat{\mathcal{g}}^{j_1m_1(j_sj_t)\kappa_1}_{j_2m_2(j_sj_u)\kappa_2\;j_3m_3(j_uj_t)\rho_3 } \psi_{a , j_1m_1\kappa_1}^{(s , t)\dagger } \psi_{j_2m_2\kappa_2}^{a(s , u ) } x_{j_3m_3\rho_3}^{(u , t ) } \right)\nonumber\displaybreak[0]\\ & -\frac{i}{2}\epsilon^{abcd } \left ( ( -1)^{m_1-(j_s - j_t)-\frac{\kappa_1}{2 } } \hat{\mathcal{f}}^{j_2m_2(j_tj_u)\kappa_2}_{j_1-m_1(j_tj_s)\kappa_1\;j_3m_3(j_sj_u ) } \psi_{a , j_1m_1\kappa_1}^{(s , t)\dagger } \phi_{cd , j_3m_3}^{(s , u ) } \psi_{b , j_2m_2\kappa_2}^{(t , u)\dagger } \right.\nonumber\\ & \quad\left .
+ ( -1)^{m_2-(j_u - j_s)-\frac{\kappa_2}{2 } } \hat{\mathcal{f}}^{j_1m_1(j_sj_t)\kappa_1}_{j_2-m_2(j_sj_u)\kappa_2\;j_3m_3(j_uj_t ) } \psi_{a , j_1m_1\kappa_1}^{(s , t)\dagger } \psi_{b , j_2m_2\kappa_2}^{(u , s)\dagger } \phi_{cd , j_3m_3}^{(u , t ) } \right)\nonumber\displaybreak[0]\\ & -i\left((-1)^{m_2-(j_u - j_s)-\frac{\kappa_2}{2 } } \hat{\mathcal{f}}^{j_2-m_2(j_sj_u)\kappa_2}_{j_1m_1(j_sj_t)\kappa_1\;j_3m_3(j_tj_u ) } \psi_{j_1m_1\kappa_1}^{a,(s , t ) } \phi_{ab , j_3m_3}^{(t , u ) } \psi_{j_2m_2\kappa_2}^{b(u , s ) } \right.\nonumber\\ & \quad\left .
+ ( -1)^{m_1-(j_s - j_t)-\frac{\kappa_1}{2 } } \hat{\mathcal{f}}^{j_1-m_1(j_tj_s)\kappa_1}_{j_2m_2(j_tj_u)\kappa_2\;j_3m_3(j_uj_s ) } \psi_{j_1m_1\kappa_1}^{a(s , t ) } \psi_{j_2m_2\kappa_2}^{b(t , u ) } \phi_{ab , j_3m_3}^{(u , s ) } \right ) \bigg ] .
\label{harmonic_expansion_of_s_matter_int}\end{aligned}\ ] ]
in this appendix , we list the value of @xmath340 for each diagram of the fermion self - energy in fig . 3 : @xmath341{\nonumber \\}&\qquad\qquad \times \biggl [ { { \scriptsize \begin{bmatrix } \tilde{u } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_s & \hspace{-2 mm } j_t \end{bmatrix } } } { { \scriptsize \begin{bmatrix } \tilde{u ' } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_s & \hspace{-2 mm } j_t \end{bmatrix } } } + ( -1)^{1-\frac{\kappa+\kappa'}{2 } } { { \scriptsize \begin{bmatrix } \tilde{u } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_t & \hspace{-2 mm } j_s \end{bmatrix } } } { { \scriptsize \begin{bmatrix } \tilde{u ' } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\ j_u & \hspace{-2 mm } j_t & \hspace{-2 mm } j_s \end{bmatrix } } } \biggr],\\[3 mm ] & ( f - b){\nonumber \\}&=\frac{ig^2}{\mu^2}nn_0 \sqrt{(2\tilde{u}+1)(2\tilde{u'}+1 ) } \delta_{uu ' } \sum_{ur_1r_2 } \frac{(2r_2 + 1)^2}{\sqrt{r_2(r_2 + 1)}}{\nonumber \\}&\times\biggl [ \frac{(2r_1 + 2)(2r_1 + 1)}{l-(r_1+\sqrt{r_2(r_2 + 1)}+\frac{3}{4 } ) } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u } & \hspace{-2 mm } \frac{1}{2 } \\
r_1+{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 } \\
r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u ' } & \hspace{-2 mm } \frac{1}{2 } \\ r_1+{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 } \\
r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } { \nonumber \\}&+\frac{2r_1(2r_1 + 1)}{l+(r_1+\sqrt{r_2(r_2 + 1)}+\frac{1}{4 } ) } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u } & \hspace{-2 mm } \frac{1}{2 } \\ r_1-{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 } \\
r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u ' } & \hspace{-2 mm } \frac{1}{2 } \\ r_1-{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 }
\\ r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } \biggr]{\nonumber \\}&\qquad \qquad \times \biggl [ { { \scriptsize \begin{bmatrix } \tilde{u } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_s & \hspace{-2 mm } j_t \end{bmatrix } } } { { \scriptsize \begin{bmatrix } \tilde{u ' } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_s & \hspace{-2 mm } j_t \end{bmatrix } } } + ( -1)^{1-\frac{\kappa+\kappa'}{2 } } { { \scriptsize \begin{bmatrix } \tilde{u } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_t & \hspace{-2 mm } j_s \end{bmatrix } } } { { \scriptsize \begin{bmatrix } \tilde{u ' } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_t & \hspace{-2 mm } j_s \end{bmatrix } } } \biggr],\\[3 mm ] & ( f - c){\nonumber \\}&=-\frac{12ig^2}{\mu^2}nn_0 \sqrt{(2\tilde{u}+1)(2\tilde{u'}+1 ) } \delta_{uu ' } \sum_{ur_1r_2 } ( 2r_2 + 1 ) { \nonumber \\}&\times\biggl [ ( 2r_1 + 2)(2r_1 + 1)\frac{1}{l+(r_1+r_2+\frac{5}{4 } ) } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u } & \hspace{-2 mm } \frac{1}{2 }
\\ r_1+{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 } \\ r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u ' } & \hspace{-2 mm } \frac{1}{2 } \\ r_1+{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 } \\
r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } { \nonumber \\}&+2r_1(2r_1 + 1)\frac{1}{l-(r_1+r_2+\frac{3}{4 } ) } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u } & \hspace{-2 mm } \frac{1}{2 } \\
r_1-{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 } \\
r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } { { \scriptsize \begin{bmatrix } u & \hspace{-2 mm } \tilde{u ' } & \hspace{-2 mm } \frac{1}{2 } \\ r_1-{{\tiny \mbox{$\frac{1}{2}$}}}&\hspace{-2 mm } r_1 & \hspace{-2 mm } \frac{1}{2 }
\\ r_2 & \hspace{-2 mm } r_2 & \hspace{-2 mm } 0 \end{bmatrix } } } \biggr]{\nonumber \\}&\qquad \qquad \times \biggl [ { { \scriptsize \begin{bmatrix } \tilde{u } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_s & \hspace{-2 mm } j_t \end{bmatrix } } } { { \scriptsize \begin{bmatrix } \tilde{u ' } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_s & \hspace{-2 mm } j_t \end{bmatrix } } } + ( -1)^{1-\frac{\kappa+\kappa'}{2 } } { { \scriptsize \begin{bmatrix } \tilde{u } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_t & \hspace{-2 mm } j_s \end{bmatrix } } } { { \scriptsize \begin{bmatrix } \tilde{u ' } & \hspace{-2 mm } r_1 & \hspace{-2 mm } r_2 \\
j_u & \hspace{-2 mm } j_t & \hspace{-2 mm } j_s \end{bmatrix } } } \biggr],\end{aligned}\ ] ] where @xmath288 .
d. b. kaplan , e. katz and m. unsal , jhep * 0305 * ( 2003 ) 037 [ arxiv : hep - lat/0206019 ] .
k. itoh , m. kato , h. sawanaka , h. so and n. ukita , jhep * 0302 * ( 2003 ) 033 [ arxiv : hep - lat/0210049 ] .
s. catterall , jhep * 0305 * ( 2003 ) 038 [ arxiv : hep - lat/0301028 ] .
f. sugino , jhep * 0401 * ( 2004 ) 015 [ arxiv : hep - lat/0311021 ] .
a. dadda , i. kanamori , n. kawamoto and k. nagata , phys .
b * 633 * ( 2006 ) 645 [ arxiv : hep - lat/0507029 ] .
t. eguchi and h. kawai , phys .
* 48 * ( 1982 ) 1063 .
m. teper and h. vairinhos , phys .
b * 652 * ( 2007 ) 359 [ arxiv : hep - th/0612097 ] .
t. azeyanagi , m. hanada , t. hirata and t. ishikawa , jhep * 0801 * ( 2008 ) 025 [ arxiv:0711.1925 [ hep - lat ] ] .
w. bietenholz , a. bigarini , j. nishimura , y. susaki , a. torrielli and j. volkholz , pos * lattice2007 * ( 2007 ) 049 [ arxiv:0708.1857 [ hep - lat ] ] . w. i. taylor , phys .
b * 394 * ( 1997 ) 283 [ arxiv : hep - th/9611042 ] .
g. ishiki , s. shimasaki , y. takayama and a. tsuchiya , jhep * 0611 * ( 2006 ) 089 [ arxiv : hep - th/0610038 ] .
j. maldacena , m. m. sheikh - jabbari and m. van raamsdonk , jhep * 0301 * ( 2003 ) 038 [ arxiv : hep - th/0211139 ] . u. carow - watamura , h. steinacker and s. watamura , j. geom .
* 54 * ( 2005 ) 373 [ arxiv : hep - th/0404130 ] .
t. ishii , g. ishiki , s. shimasaki and a. tsuchiya , jhep * 0705 * ( 2007 ) 014 [ arxiv : hep - th/0703021 ] . t. ishii , g. ishiki , s. shimasaki and a. tsuchiya , phys . rev .
d * 77 * ( 2008 ) 126015 [ arxiv:0802.2782 [ hep - th ] ] .
m. hanada , j. nishimura and s. takeuchi , phys .
* 99 * ( 2007 )
161602 [ arxiv:0706.1647 [ hep - lat ] ] . k. n. anagnostopoulos , m. hanada , j. nishimura and s. takeuchi , phys .
* 100 * ( 2008 ) 021601 [ arxiv:0707.4454 [ hep - th ] ] .
s. catterall and t. wiseman , arxiv:0803.4273 [ hep - th ] .
h. kaneko , y. kitazawa and k. matsumoto , phys .
d * 76 * ( 2007 ) 084024 [ arxiv:0706.1708 [ hep - th ] ] .
j. w. elliott , j. giedt and g. d. moore , arxiv:0806.0013 [ hep - lat ] .
g. ishiki , y. takayama and a. tsuchiya , jhep * 0610 * ( 2006 ) 007 [ arxiv : hep - th/0605163 ] .
t. t. wu and c. n. yang , nucl .
b * 107 * ( 1976 ) 365 .
t. ishii , g. ishiki , k. ohta , s. shimasaki and a. tsuchiya , prog .
* 119 * ( 2008 ) 863 [ arxiv:0711.4235 [ hep - th ] ] .
d. varshalovich , a. moskalev and v. khersonskii , _ quantum theory of angular momentum _
( world scientific , singapore , 1988 ) . c. s. chu , j. madore and h. steinacker , jhep * 0108 * ( 2001 ) 038 [ arxiv : hep - th/0106205 ] .
m. panero , jhep * 0705 * ( 2007 ) 082 [ arxiv : hep - th/0608202 ] . h. ling , a. r. mohazab , h. h. shieh , g. van anders and m. van raamsdonk , jhep * 0610 * ( 2006 ) 018 [ arxiv : hep - th/0606014 ] .
j. m. maldacena , phys .
* 80 * ( 1998 ) 4859 [ arxiv : hep - th/9803002 ] . s. j. rey and j. t. yee , eur .
j. c * 22 * ( 2001 ) 379 [ arxiv : hep - th/9803001 ] .
l. f. alday and j. maldacena , jhep * 0711 * ( 2007 ) 068 [ arxiv:0710.1060 [ hep - th ] ] .
b. de wit , j. hoppe and h. nicolai , nucl .
b * 305 * ( 1988 ) 545 .
j. hoppe , int .
j. mod .
phys . a * 4 * ( 1989 ) 5235 .
k. dasgupta , m. m. sheikh - jabbari and m. van raamsdonk , jhep * 0205 * ( 2002 ) 056 [ arxiv : hep - th/0205185 ] . | we propose a nonperturbative definition of @xmath0 super yang - mills ( sym ) .
we realize @xmath0 sym on @xmath1 as the theory around a vacuum of the plane wave matrix model .
our regularization preserves sixteen supersymmetries and the gauge symmetry .
we perform the 1-loop calculation to give evidences that the superconformal symmetry is restored in the continuum limit .
[ cols= " < " , ] + @xmath0 super yang - mills from the plane wave matrix model takaaki ishii@xmath2 , goro ishiki@xmath3 , shinji shimasaki@xmath2 asato tsuchiya@xmath4 + @xmath2 _ department of physics , graduate school of science _ + _ osaka university , toyonaka , osaka 560 - 0043 , japan _ + @xmath5 _ institute of particle and nuclear studies _
+ _ high energy accelerator research organization ( kek ) _ + _ 1 - 1 oho , tsukuba , ibaraki 305 - 0801 , japan
_ + @xmath4 _ department of physics , shizuoka university _ + _ 836 ohya , suruga - ku , shizuoka 422 - 8529 , japan _ |
the interesting and anomalous properties of the manganites showing a colossal magnetoresistance ( cmr ) compounds are thought to originate from the competition between charge , lattice and spin degrees of freedom .
the fundamental mechanism of cmr is reasonably understood within a framework of the double exchange model , which is based on the hopping effect of @xmath12@xmath13 electrons between mn@xmath14 and mn@xmath15 ions [ 1 - 2 ] .
in such a case the ferromagnetic interaction arises from the strong on - site coupling between the charge carriers ( @xmath12@xmath13-state ) and the local spins ( @xmath16@xmath17-state ) . however , there are many experimental reports that can not be explained by this framework at present . since the discovery of the larger magnitude of cmr in so - called 327 bilayer manganese oxide : la@xmath18sr@xmath19mn@xmath4o@xmath5 , a considerable research effort has been devoted to the understanding of the properties of these compounds [ 3 - 4 ] .
furthermore , the substitution of pr@xmath14 for la ions at the concentration of @xmath6=0.6 in ( la@xmath0pr@xmath1)@xmath2sr@xmath3mn@xmath4o@xmath5 ( hereafter denoted as lpsmo ) realizes the system with a record value of the cmr in this bilayer manganese oxide family , which amounts to a huge decrease of the @xmath20-axis resistivity by a factor of one million in the presence of a magnetic field [ 5 - 6 ] .
the first - order transition is accompanied by a remarkable negative magnetostriction , indicating the presence of important interplay among spin , carrier and orbital degrees of freedom [ 7 - 8 ] .
the host compound , la@xmath2sr@xmath3mn@xmath4o@xmath5 exhibits a paramagnetic to ferromagnetic transition at @xmath21@xmath22@xmath23125 k , accompanied by a semiconductor to metal transition [ 4 ] .
the magnitude of cmr effect reaches 98@xmath24 near @xmath21@xmath22 .
besides , this 327 bilayer manganese oxide exhibits a remarkable range of magnetic behavior , strongly depending on the concentration of sr [ 9 - 14 ] .
the substitution of the smaller pr@xmath14 ions for the larger la@xmath14 ions causes a local lattice deformation : jahn - teller distortion ( jt ) . in this system , mno@xmath25 octahedral distortions are interpreted in terms of the occupation ratio between @xmath12@xmath13 and @xmath16@xmath17 level , and ordering of @xmath12@xmath13 orbitals and their relationship to magnetic ground state [ 15 ] . the mno@xmath25 octahedra shrinks along the @xmath20-axis in the ground state with increasing @xmath26 , implying an increase of the @xmath27(@xmath26@xmath28-@xmath29@xmath28 ) orbital polarization [ 5 - 8 ] . on the other hand , the mno@xmath25 octahedra elongate along the @xmath20-axis in the ground state with increasing a pr concentration @xmath6 at the sr concentration of 0.4 , implying an increase of the @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) orbital polarization .
furthermore , an increase of the pr concentration causes a suppression of the planar ferromagnetic interaction , leading to a lowering of @xmath21@xmath22 .
the ferromagnetic phase seems to vanish around @xmath6=0.5 .
this concentration area seems to be nearby the magnetically frustrated point ( fp ) in which ferro and antiferromagnetic interactions compete considerably with one another ( although this origin is not clear yet ) .
since several vital interactions compete with each other , that is , the resultant instability of magnetic , lattice and band properties , it is expected that a unique phase or phenomena may be induced by external variables such as pressure and magnetic fields .
the ultrasonic measurement is a strong probe to examine the characteristic features of the 3@xmath27-electronic states of the mn ions .
the anomalies of the elastic constants is expected to be due to the coupling between magnetic moments or the conduction electrons and elastic strains associated with sound wave [ 19 ] . in this paper
we present our investigations on the elastic and magnetic properties of single crystal lpsmo with the concentration of z=0.6 .
the preliminary reports have been published in ref .
[ 20 - 21 ] .
a single crystal of lpsmo for @xmath6=0.6 was grown by the floating - zone ( fz ) method using a mirror furnace .
the specimen used in this study measures 5@xmath314 mm@xmath28 in the @xmath32-plane and 1 mm along the @xmath20-axis .
the sound velocity ( @xmath33 ) was measured by an ultrasonic apparatus based on a phase comparison method in magnetic fields up to 12 t generated by a superconducting magnet .
plates of linbo@xmath34 were used for the piezoelectric transducers .
the transducers were glued on to the parallel planes of the sample by an elastic polymer thiokol .
the absolute value of the elastic constant @xmath35=@xmath36@xmath33@xmath28 by using the density @xmath36 of the crystal could not be estimated explicitly because the observed echo trains were not enough clear to do that .
the high - field magnetization ( @xmath11 ) measurement was carried out by using the vibrating sample magnetometer ( vsm ) method in fields up to 12 t and at temperatures down to 4.2 k. the detailed magnetization measurement was performed by superconducting quantum interference device ( squid ) magnetometer from 1.8 k to 300 k in magnetic fields up to 5 t. @xmath37 , @xmath35@xmath38 , @xmath35@xmath39 and @xmath35@xmath40 of lpsmo for @xmath6=0.6 in the absence of magnetic fields .
the dotted vertical line indicates a transition point determined by the resistivity measurement . ]
@xmath38 of lpsmo for @xmath6=0.6 at selected temperatures .
the vertical arrows indicate a transition point .
the horizontal ones indicate a direction of the applied fields . ]
@xmath37 of lpsmo for @xmath6=0.6 at selected temperatures .
the vertical arrows indicate a transition point .
the horizontal ones indicate a direction of the applied fields . ]
@xmath39 of lpsmo for @xmath6=0.6 at selected temperatures .
the vertical arrows indicate a transition point .
the horizontal ones indicate a direction of the applied fields . ]
@xmath40 of lpsmo for @xmath6=0.6 at selected temperatures .
the vertical arrows indicate a transition point .
the horizontal ones indicate a direction of the applied fields . ]
in this section we report the elastic properties of lpsmo for @xmath6=0.6 .
figure 1 shows the temperature dependence of the longitudinal elastic constant @xmath35@xmath37 , @xmath35@xmath38 and the transverse elastic constant @xmath35@xmath39 , @xmath35@xmath40 in the absence of a magnetic field .
plotted is the relative elastic constant change . to avoid overlap and complication of curves ,
their offsets are shifted .
the same modification is utilized in all following figures in this section .
@xmath35@xmath37 and @xmath35@xmath38 are the elastic constant determined by the longitudinal sound wave propagated along @xmath41-axis and @xmath20-axis , respectively , and @xmath35@xmath39 and @xmath35@xmath40 are those determined by the transverse one propagated along @xmath41-axis and @xmath20-axis , respectively .
they both increase monotonically with decreasing a temperature .
a slight anomaly seems to be observed around @xmath21@xmath42 of 40 k in all elastic constants , where the temperature dependence of the resistivity exhibits a distinct shoulder structure[5 - 6 ] .
a pronounced elastic anomaly , however , was induced by applying magnetic fields .
figure 2 shows the field dependence of @xmath35@xmath38 at selected temperatures . at 4.2
k @xmath35@xmath38 decreases slightly with increasing a field . the sharp decrease was observed at 6 t , and then @xmath35@xmath38 increases slightly with the further increase of fields .
whereas , when the field decreases , @xmath35@xmath38 decreases slightly with decreasing a field . on the contrary ,
no anomaly was observed at 6 t with decreasing a field . that is to say ,
the transition induced by a field is accompanied by a large hysteresis , indicating the transition to be first - order transition nature .
the trace of @xmath35@xmath38 is guided by the dotted arrows in fig .
2 . the transition field shifts to the lower temperature with increasing a field below around 40 k. besides
, the hysteresis turns to be smaller with increasing a field .
whereas , the transition field shifts to the higher temperature with increasing a temperature above around 40 k. furthermore , @xmath35@xmath38 increases abruptly above the transition field in the higher fields above around 40 k. it is noted that no hysteresis accompanied by the transition was observed above 40 k. @xmath35@xmath38 slightly increases with increasing a field and no anomaly was recognized at 240 k. figure 3 shows the field dependence of @xmath35@xmath37 at selected temperatures .
the general behavior is similar to that of @xmath35@xmath38 .
however , the elastic anomaly observed at 120 k and 160 k is opposite to that observed for @xmath35@xmath38 .
the @xmath35@xmath37 decreases rapidly at the transition field .
figures 4 and 5 show the field dependence of @xmath35@xmath39 and @xmath35@xmath40 at selected temperatures , respectively . it is noted that @xmath35@xmath39 exhibits a sharp jump at the transition field at 4.2 k , in contrast to that of @xmath35@xmath37 and @xmath35@xmath38 .
it is accompanied by a distinct hysteresis as well . on the other hand
, @xmath35@xmath40 exhibits a slight drop at the transition field at 1.6 k. the general behavior around the transition field and the elastic constants are qualitatively similar to those of the longitudinal @xmath35@xmath37 and @xmath35@xmath38 . a distinct transition accompanied by a large hysteresis
was observed at around 5 t in the field dependence of elastic constants at 4.2 k. the transition shifts to lower temperatures and the accompanied hysteresis becomes smaller with increasing temperature up to about 50 k. on the other hand , the transition shifts to higher temperatures with the further increase in temperature .
this transition is not accompanied by a hysteresis within the experimental error .
inset shows the behavior around 40 k in a field of 0.05 t on an expanded scale .
the solid arrows indicate a transition point .
the dotted arrows indicate a direction of the applied fields . ]
the arrows indicate a transition point . ]
the arrows indicate a transition point .
inset shows the behavior around 40 k in a field of 0.05 t on an expanded scale . ]
-axis at 3.7 k ( a ) and 10 k. ( b ) ] -axis above 20 k. ] -axis above 20 k. ] -@xmath21 ) magnetic phase diagram determined by the present results .
the squares , triangles and circles denote the transition points determined by the @xmath11-@xmath7 curves , @xmath11-@xmath21 curves and elastic constant measurement , respectively .
the solid and dotted lines are a guide to the eye .
the solid and dotted lines denote the first- and second- nature phase transition , respectively ] in this section we report the results of the magnetization measurement of lpsmo for @xmath6=0.6 .
figure 6 shows the temperature ( @xmath21 ) dependence of the magnetization in selected fields along @xmath20-axis @xmath11@xmath22 .
@xmath11@xmath22-@xmath21 curve exhibits a slight increase without a distinct anomaly with decreasing temperature in a field of 0.05 t. inset shows the behavior around 40 k on an expanded scale .
a clear anomaly is recognized , although it is fairly small .
a distinct cusp , however , appears in the zero - field - cooled ( zfc ) magnetization curve at around 40 k in a field of 1 t. on the other hand , the field - cooled ( fc ) magnetization curve exhibits a gradual increase below about 100 k and reaches a constant value in the further decrease in temperature , without any subsequent drop below 40 k. it is noted that the fc magnetization curve is quite different from zfc one below around 40 k. an increase of magnetization below 40 k becomes steeper and a hysteresis becomes larger with increasing a field .
furthermore , the anomaly at around 40 k under a field of 1 t split into two anomalies .
the lower transition temperature shifts to lower temperatures up to 4 t , accompanied by a distinct hysteresis and disappears around 5 t. while , the higher transition temperature shifts to higher temperature .
it is noteworthy that the higher transition is not accompanied by a hysteresis , indicating a second - order transition nature .
this transition becomes steeper up to 6 t , and then smeared with the further increase of field as shown in fig .
these transitions correspond to those observed in the elastic constants , consistently .
8 shows the temperature ( @xmath21 ) dependence of the magnetization in selected fields along @xmath41-axis @xmath11@xmath43 .
inset shows the behavior around 40 k on an expanded scale .
similar to those of @xmath11@xmath22-@xmath21 curve , a clear anomaly is recognized , although it is fairly small .
a distinct cusp , however , appears in the zero - field - cooled ( zfc ) magnetization curve at around 40 k in a field of 1 t. the similar behaviors such as hysteresis , a split of transition and a steep increase of magnetization were observed under the higher fields . here
, let us compare the fc magnetization curve in a field of 0.05 t between along @xmath20-axis @xmath11@xmath22 and in @xmath41@xmath44-plane @xmath11@xmath43 as shown in inset of figs . 6 and 8 , respectively .
one can see that a magnetic moment lies in @xmath41@xmath44-plane by their anisotoropic behavior .
next , we show the high - field magnetization process , @xmath11-@xmath7 curve .
the virgin @xmath11@xmath22-@xmath7 curve at 3.7 k exhibits a steep increase around 5 t as shown in fig . 9 ( a ) . a small shoulder was also observed around 2 t only in the virgin process with respect to an increase of fields .
a clear rectangular hysteresis loop was observed in the field range between around 1 t and -1 t. the saturated magnetization reaches 3.5 @xmath45@xmath46/mn .
this hysteresis loop becomes smaller with increasing temperatures as shown in fig .
figure 10 shows the high - field magnetization measurement above 10 k. a metamagnetic transition observed in the virgin @xmath11@xmath22-@xmath7 curve shifts to lower fields with increasing temperatures up to 50 k. while , the transition shifts to higher fields and becomes gentler with the further increase of temperature .
it is noted that the transition is accompanied by no hysteresis above 50 k. no metamagnetic transition was observed at 200 k. figure 11 shows the @xmath11@xmath43-@xmath7 curve at selected temperatures .
similarly , a clear step increase was observed in @xmath11@xmath43-@xmath7 curve as well as @xmath11@xmath22-@xmath7 one .
the transition field shifts to higher fields and the accompanied hysteresis becomes smaller with increasing a temperatures .
the saturated magnetization reaches 2.5 @xmath45@xmath46/mn , which is almost 70@xmath24 of @xmath11@xmath22 , indicating that the @xmath20-axis is easy axis of magnetization .
the magnetic ( @xmath7-@xmath21 ) phase diagram was made by the measured elastic constants and magnetizations for the field along @xmath20-axis , i.e. easy axis of magnetization .
figure 12 shows the ( @xmath7-@xmath21 ) phase diagram under the field along @xmath20-axis .
there are mainly three different regions .
the phase i and ii are considered to be paramagnetic and induced ferromagnetic phase , respectively .
it is deduced from that almost full magnetic moment of 3.5 @xmath45@xmath46/mn is induced in the phase ii , where the system exhibits a metallic behavior in the resistivity .
the phase iii is the field - induced ferromagnetic phase .
there are two sub - regions in the phase iii .
the boundary is defined well only by the anomaly in the virgin @xmath11-@xmath7 and @xmath35-@xmath7 curves .
this boundary may be due to the domain rotation .
it is noted that the boundary between i and iii under zero field is quite obscure .
however , it becomes distinct in a magnetic field , indicating that the phase iii appears only in a magnetic field .
the transition is accompanied by a remarkable hysteresis in both magnetization and elastic constants at the boundary between the phase iii and ii , being indicative of a first - order transition nature . on the contrary ,
no hysteresis is accompanied at the boundary between the phase i and ii , being indicative of a second - order transition nature .
the obtained qualitative features of elastic constants are summarized as follows .
the elastic constant decreases crossing the phase boundary from the phase iii to ii accompanied by a remarkable hysteresis only in the virgin process with increasing a field . whereas , it decreases gradually with increasing a field in the phase i , and increases gradually after crossing the boundary from the phase i to ii without any hysteresis .
the temperature of around @xmath21@xmath42 of 40 k is likely to be the critical temperature to distinguish these phases , although @xmath21@xmath42 is quite vague .
@xmath28@xmath8@xmath9 ( a ) below the critical temperature of 40 k and ( b ) above that , respectively .
the dotted arrows indicate a direction of the applied fields .
40 k inc . and
denotes a result at 40 k in the increase and decrease process with respect to a magnetic field , respectively . ]
first , we discuss the obtained elastic anomalies in the field dependence of elastic constants .
let us compare them with the magnetization curves as shown in fig . 13 . a sharp softening accompanied by a prominent hysteresis
was observed at metamagnetic transition field @xmath7@xmath22 at @xmath21= 3.7 k and 30 k. on the other hand , a softening without any hysteresis was observed when crossing the transition field at the temperature of 80 k. the critical temperature is roughly estimated to be 40 k. from this result , one can see the correlative relation between them .
it is likely that the elastic anomalies accompanied with a metamagnetic transition are responsible for electron - phonon effect .
in such a case the elastic softening is related to the magnetic susceptibility , @xmath8@xmath9=@xmath10@xmath11/@xmath10@xmath7[22 - 24 ] . a strong coupling between elastic strains associated by sound waves and relevant conduction bands is usually described in terms of the magnetic gr@xmath47neisen parameter @xmath48@xmath46 = -@xmath10ln@xmath7@xmath22/@xmath10@xmath49@xmath50 , where @xmath49@xmath50 is the elastic strain with @xmath48 symmetry .
then , the elastic softening is given by @xmath51 figure 14 ( a ) and ( b ) show the field dependence of -@xmath8@xmath9@xmath7@xmath28 below the critical temperature @xmath21@xmath42 of 40 k and above that , respectively .
as compared to the measured field dependence of elastic constants , especially longitudinal ones shown in figs .
2 and 3 , a good agreement with -@xmath8@xmath9@xmath7@xmath28 is achieved by adopting the single free parameter @xmath48@xmath46 at various temperatures , described by eq .
( 1 ) . especially in the case above 40 k , we can obtain a good agreement , although some of them in the high field region and their quantitative aspects do not agree well between them .
we emphasize that this agreement gives evidence for a strong spin - lattice coupling in lpsmo below @xmath21@xmath42 of 40 k. the expected upturn of elastic constant by -@xmath8@xmath9@xmath7@xmath28 above @xmath7@xmath22 below 40 k does nt appear in the measured elastic constants , which is expected by this calculated model .
one scenario to explain
the discrepancy is that the change of carrier number probably causes the difference as discussed below .
next , let us discuss the coupling between the elastic strains and the conduction electrons .
according to the deformation potential coupling theory , the coupling can be described as @xmath52 where , @xmath53@xmath54 denotes a single particle electron state and @xmath27@xmath50 the deformation potential coupling constant with @xmath48 symmetry [ 25 - 27 ] . the suffix @xmath48 denotes the irreducible representation in the tetragonal symmetry of the present system [ 19 ] . the free energy for conduction electrons @xmath55@xmath56
may be expressed as follows , @xmath57 where @xmath53@xmath58 and @xmath59 denote the fermi energy and the number of the conduction electrons in the band , respectively . from the second derivative of the free energy , one obtains the expression of a relative change in the elastic constant .
@xmath60 where , @xmath36(@xmath49 ) and @xmath61@xmath54 denote the density of state and fermi - dirac distribution function , respectively . for simplicity
the coupling constant @xmath27@xmath50 is taken to be independent of @xmath62 ( rigid band model ) , in a two band model in which we assume them to be @xmath27(@xmath26@xmath28-@xmath29@xmath28 ) and @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) .
then , the above formula leads to the `` band - jahn - teller effect '' one .
it is expected that an increase of the density of states , i.e. , the carrier number contributes greatly to @xmath63@xmath35 .
thus , a decrease of the resistivity as a function of magnetic fields leads to a decrease of the elastic constant as described by eq .
we conjecture that this masks partly the expected upturn just above @xmath7@xmath22 .
unfortunately , we can not discuss these changes quantitatively at present since the absolute values of the elastic constants were not estimated .
nevertheless , we suggest that the elastic anomaly and the absence of the upturn of elastic constant above @xmath7@xmath22 below @xmath21@xmath42 are to be explained by the coupling between elastic strain and @xmath8@xmath9 , and between elastic strain and carriers , respectively .
secondly we would like to discuss the phase iii .
the obtained features may suggest this phase to be fairly metastable phase .
figs . 2 , 3 and 4
give typical examples to describe this situation .
actually , the oscillatory phenomena is observed for @xmath35@xmath38 at 30 k , @xmath35@xmath37 at 30 k and 50 k and @xmath35@xmath39 at 10 and 30 k below @xmath7@xmath22 .
the oscillations disappear above @xmath7@xmath22 and for higher temperatures .
we conjecture that these oscillations may be related to the vicinity of magnetic instability , such as a spin - glass state , since the transition is accompanied by a prominent hysteresis in both results of ultrasonic and magnetization measurements .
however , it is somehow unusual , because a slight anomaly is observed in the absence of a magnetic field .
a spin glass - like behavior shows up only in a magnetic field . the critical temperature in the absence of a field is estimated to be around @xmath21@xmath42 of 40 k by interpolation of the boundary between the phase i and iii .
again , this temperature corresponds to that at which a bend is observed in the resistivity curve .
a slight anomaly was observed in both the temperature dependence of elastic constant and magnetization curve at this temperature .
( la@xmath0nd@xmath1)@xmath2sr@xmath3mn@xmath4o@xmath5 and ( la@xmath64gd@xmath65)@xmath66sr@xmath67mn@xmath4o@xmath5 would be helpful to discuss this phase , which was reported previously by moritomo @xmath68 @xmath69 [ 28 ] and dho @xmath68 @xmath69 [ 29 ] , respectively .
the similar phase shows up in ( la@xmath0nd@xmath1)@xmath2sr@xmath3mn@xmath4o@xmath5 ( @xmath6=0.4 ) at low temperature . in comparison
, this phase is presumably due to metastable region where the parainsulator and ferrometal state can coexist .
that is to say , ferromagnetic clusters can be grown by applying a magnetic field .
this phase was understood within the framework of spin degree of freedom so far .
our previous results of lpsmo for @xmath6=0.6 indicated that magnetostriction ( ms ) along the @xmath20-axis and in the @xmath41@xmath44 plane showed a sudden decrease and increase at the transition field , respectively [ 7 - 8 ] .
furthermore , it should be noted that the transition is accompanied by a prominent hysteresis in ms .
this implies that an orbital degree of freedom plays a crucial role in this ordered phase iii as well .
the hysteresis region is observed in the elastic constants as a function of field becomes broader with decreasing a temperature in the phase iii .
classical thermodynamics indicates that the first - order phase transition of the metastable state can occur under the condition that the potential barrier between two states is comparable with temperature [ 28 ] .
thus it becomes difficult to transit between the metastable states with decreasing a temperature .
this energy difference may correspond to @xmath21@xmath42 of 40 k. the results of elastic constants and ms as a function of fields imply that the two states : paramagnetic and ferromagnetic states have the each different state with respect to the @xmath12@xmath13 state .
as was seen in lpsmo ( @xmath6=0 ) the @xmath12@xmath13 state is dominated by the @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) in the ferromagnetic phase , whereas an amount of @xmath27(@xmath26@xmath28-@xmath29@xmath28 ) population mixes into @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) component in the paramagnetic phase .
recently , this behavior was confirmed microscopically by the neutron scattering measurement [ 30 - 31 ] .
we speculate that this situation may provide the spatial disorder of the orbital state in addition to the magnetic disorder .
finally , we would like to comment on nature of phase ii and iii .
a distinct hysteresis was observed in phase iii .
there is another boundary in phase iii , distinguished by iii and iii. it seemed reasonable to think that the boundary between phase iii and iii may be due to the rotation of a magnetic domain wall .
we expect that the orbital glass state with respect to @xmath27(@xmath26@xmath28-@xmath29@xmath28 ) and @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) or / and orbital phase clusters : para- and ordered - cluster with @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) component could occur in the phase iii because the two orbital states can coexist at each mn sites in the phase iii in the above - mentioned way .
that is to say , it seems that the ferromagnetic cluster arises because the two levels are close to each other in energy , and the occupation ratio strongly affects on the magnetic property .
this situation lets us conjecture that the occupation ratio of @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) are different at each mn sites . as a result
two kinds of the nearly - degenerate orbital : @xmath27(@xmath26@xmath28-@xmath29@xmath28 ) or @xmath27(3@xmath6@xmath28-@xmath30@xmath28 ) could be arranged randomly at mn sites , and cause the resultant spin - glass behavior .
furthermore , this metastable state could vanish in phase ii , implying a uniform magnetic and a resultant uniform orbital state could be induced in phase ii .
the experimental fact that no hysteresis is observed in phase ii supports this scenario .
a rapid decrease of a relaxation time @xmath70 of the magnetization and magnetostriction in phase ii also supports this uniform state with respect to magnetic and orbital properties , @xmath71.@xmath12 .
, disappearance of the glass state : @xmath72 @xmath73 in phase ii.[5 ] this is left to future studies . in order to confirm this point ,
the frequency dependence of sound wave on elastic constants is currently in progress .
we performed magnetization and ultrasonic measurements on lpsmo for @xmath6=0.6 single crystal as a function of temperature and magnetic field .
a pronounced elastic anomaly was observed across the boundary of the each magnetic phases .
we gave a useful phenomenological description of these results in terms of the coupling between elastic strain and @xmath8@xmath9=@xmath10@xmath11/@xmath10@xmath7 .
the microscopic experimental and theoretical descriptions are still lacking .
in addition , we found a new metamagnetic - like transition above @xmath21@xmath42 of 40 k in the magnetization measurement accompanying no hysteresis , indicative of second - transition nature .
( @xmath7-@xmath21 ) magnetic phase diagram was determined , and it was found that the field - induced magnetic phase appears .
this field - induced phase has a remarkable hysteresis , being indicative of the strong coupling among the spin , orbital and lattice of mn ion .
it is noteworthy that a remarkable hysteresis in the elastic constants and also in magnetostrictions across the phase boundary of iii was observed .
a comparison between the field dependence of elastic constants and magnetization gives their close relation combined by a single parameter @xmath48@xmath46 , indicating a strong spin - lattice coupling in lpsmo .
there still remains the fundamental question why the phase boundary around @xmath21@xmath42 in the absence of field is not distinguished . for a further and qualitative discussion ,
it is necessary to make clear the magnetic structure in the ordered phase iii and iii , and a microscopic description in the ordered states .
we are grateful to y. tobinai , h. uematsu and m. nakamura for their help in the ultrasonic measurement and the operation of the cryogenic apparatus .
the measurements have been performed in the cryogenic division of the center for instrumental analysis , iwate university .
this work was partly supported by a grant - in - aid for science research from the minister of education , culture , sports , science , and technology of japan .
9 c. zener , phys . rev . *
82 * , 403 ( 1951 ) .
p. w. anderson and h. hasegawa , phys .
rev . * 100 * , 675 ( 1955 ) .
y. moritomo , a. asamitsu , h. kuwahara , and y. tokura , nature ( london ) * 380 * , 141 ( 1996 ) . t. kimura , y. tomioka , a. asamitsu , and y. tokura , phys . rev
81 * , 5920 ( 1998 ) . m. apostu , r. suryanarayanan , a. revcolevschi , h. ogasawara , m. matsukawa , m. yoshizawa , and n. kobayashi , phys .
b * 64 * , 012407 ( 2001 ) .
i. gordon , pwagner , v. v. moshchalkov , y. bruynseraede , m. apostu , r. suryanarayanan a. revcolevschi , phys .
b * 64 * , 92408 ( 2001 ) .
m. matsukawa , h. ogasawara , r. sato , m. yoshizawa , r. suryanarayanan , g. dhalenne , a. revcolevschi , k. ito , phys .
b * 62 * , 1 ( 2000 ) .
h. ogasawara , m. matsukawa , s. hatakeyama , m. yoshizawa , m. apostu , r. suryanarayanan , g. dhalenne , a. revcolevschi , k. ito and n. kobayashi , j. phys .
69 * , 1274 ( 2000 ) .
m. kubota , h. fujioka , k. hirota , k. ohyama , y. moritomo , h. yoshizawa and y. endoh , j. phys .
. jpn . * 69 * , 1606 ( 2000 ) .
m. kubota , y. oohara , h. yoshizawa , h. fujioka , k. shimizu , k. hirota , y. moritomo and y. endoh , j. phys .
. jpn . * 69 * , 1986 ( 2000 ) .
k. hirota , y. moritomo , h. hujioka , m. kubota , h. yoshizawa , and y. endoh , jpn .
* 67 * , 3380 ( 1998 ) .
t. kimura , y. tomioka , h. kuwahara , a. asamitsu , m. tamura , and y. tokura , science * 274 * , 1698 ( 1996 ) . t. kimura , a. asamitsu , y. tomioka , and y. tokura , phys .
lett . * 79 * , 3720 ( 1997 ) .
d. n. argyriou , j. f. mitchell , p. g. radaelli , h. n. bordallo , d. e. cox , m. medarde , and j. d. jorgensen , phys .
b * 59 * , 8695 ( 1999 ) .
a. koizumi , s. miyaki , y. kakutani , h. koizumi , n. hiraoka , k. makoshi , n. sakai , k. hirota , y. murakami , phys .
* 86 * , 5589 ( 2001 ) .
s. s. saxena , p. agarwal , k. ahllan , f. m. grosche , r. k. w. haselwlmmer , m. j. stelner , e. pugh , i. r. walker , s. r. jullan , p. monthoux , g. g. lonzarich , a. huxley , i. shelkin , d. bralthwalte and j. flouquet , nature ( london ) * 406 * , 587 ( 2000 ) .
c. pfleidere , m. uhlarz , s. m. hayden , r. vollmer , h. v. lohneysen , n. r. bernhoeft , and g. g. lonzarich , nature ( london ) * 412 * , 58 ( 2001 ) .
h. hegger , c. petrovic , e. g. moshopoulou , m. f. hundley , j. l. sarrao , z. fisk , and j. d. thompson , phys .
lett . * 84 * , 4986 ( 2000 ) .
a. imaduddin , y. nakanishi , k. shimomura , h. kanazawa , m. nakamura , n. yoshimoto , m. yoshizawa , j. phys .
. jpn . * 71 * , 1965 ( 2002 ) .
y. nakanishi , k .
shimomura , m. yoshizawa , m. apostu , r. suryanarayanan and a. revcolevischi , acta physica polonica b * 34 * , 831 ( 2003 ) .
y. nakanishi , k. shimomura , s. nakamura , m. matsukawa , m. yoshizawa , m. apostu , r. suryanarayanan and a. revcolevischi , physica b * 329 - 333 * , 817 ( 2003 ) .
i. kouroudis , d. weber , m. yoshizawa , b. l@xmath47thi , l. puech , p. haen , j. flouquet , g. bruls , u. welp , j. j. m. franse , a. menovsky , e. bucher , j. hufnagl , phys .
lett . * 58 * , 820 ( 1987 ) .
thalmeier , j. magn .
magn . mater . * 76&77 * , 299 ( 1988 )
. j. r. feller , j. b. ketterson , d. g. hinks , d. dasgupta , bimal k. sarma , phys .
b * 62 * , 11538 ( 2000 ) .
b. l@xmath47thi , j. magn .
52 * , 70 ( 1985 ) .
b. l@xmath47thi and m. yoshizawa , j. magn .
magn . mater . *
63&64 * , 274 ( 1987 ) . s. nakamura , t. goto , y. isikawa , s. sakatsume , m. kasaya , j. phys .
. jpn . * 60 * , 2305 ( 1991 ) .
y. moritomo , y. maruyama , t. akimoto , a. nakamura , phys .
b * 56 * , r7057 ( 1997 ) .
joonghoe dho , w. s. kim , n. h. hur , phys .
b * 65 * , 024404 ( 2001 ) .
f. wang , a. gukasov , f. moussa , m. hennion , m. apostu , r. suryanarayanan , and a. revcolevschi , phys .
* 91 * 047204 ( 2003 ) .
f. moussa , m. hennion , f. wang , a. gukasov , r. suryanarayanan , m. apostu , and a. revcolevschi , phys .
lett . * 93 * 107202 ( 2004 ) . | the elastic and magnetic properties of a single crystal of 327 bilayer manganese oxide ( la@xmath0pr@xmath1)@xmath2sr@xmath3mn@xmath4o@xmath5 with @xmath6=0.6 have been investigated by ultrasonic and high - field magnetization measurements .
a distinct elastic anomaly was observed at low temperatures and in high magnetic fields when crossing the phase boundary . a pronounced elastic softening as a function of magnetic field ( @xmath7 ) appears across the boundary of the low - temperature magnetic phase below around 40 k , accompanied by a distinct hysteresis . in the high field region , however , these elastic constants exhibit an increase with increasing a magnetic field in the high - field region .
the high - field magnetization measurement characterizing the magnetic state clarified a strong connection between elastic strains and magnetic moments .
the new metamagnetic transition was found in the high - field region above 40 k , accompanied with no hysteresis .
the present results of elastic constants are discussed in terms of a coupling effect between elastic strains associated with sound waves and relevant magnetic moments , and relevant conduction bands .
they are explained reasonably in terms of a strong coupling between elastic strains and magnetic susceptibility @xmath8@xmath9=@xmath10@xmath11/@xmath10@xmath7 . |
the large reservoirs of dust observed in some high redshift galaxies ( e.g. , ) have been hypothesised to originate from dust produced by supernovae from massive stars .
some theoretical studies ( e.g. , ) have supported a high efficiency of dust production ( 0.1 - 1.0m@xmath0 ) in core - collapse supernovae ( ccsne ) which would suffice to account for the dust mass budget observed in dusty high - redshift sources @xcite . however , the dust reservoirs ( @xmath8 10@xmath9m@xmath0 ) that were detected at mid - ir wavelengths during the first 1000 days in a number of ccsne remained several orders of magnitude below these theoretical predictions @xcite . with the recent advent of far - infrared ( fir ) and
submillimetre ( submm ) observing facilities ( e.g. , _ herschel _ , alma ) , the ability to also detect the emission from colder dust in ccsn remnants opened up and resulted in the detection of dust masses on the order of 0.1 - 1.0m@xmath0 in some nearby supernova remnants ( sn 1987a , crab nebula , cassiopeia a ) .
some supernova remnants show evidence for dust formation in the supernova ejecta once the ejecta material has sufficiently cooled after expansion to allow grain growth to take place ( e.g. , @xcite ) .
recent work by @xcite , @xcite and @xcite suggest that the dust mass in ccsn ejecta grows in time possibly due to accretion of material onto and coagulation of grain species .
of particular interest for studies of the mechanisms responsible for dust formation is the galactic supernova remnant cassiopeia a ( hereafter , casa ) , which shows evidence for dust in the shocked outer supernova ejecta as well as in the inner , un - shocked regions of the remnant . in this paper
, we study _
spitzer _ and _ herschel _ infrared and submm maps of the supernova remnant casa on spatially resolved scales of 0.6pc to constrain the mass and position of formed dust species .
casa ( fig.[casa_composite ] ) is the remnant of a supernova explosion of a massive progenitor about 330 years ago @xcite .
based on spectra of optical light echoes , casa was , more specifically , characterised as a hydrogen - poor type iib ccsn @xcite . due to the relatively young age of the remnant ,
the mass of swept - up material is small compared to the mass in the supernova ejecta , which makes it still possible to separate supernova dust from any swept - up circumstellar material . early _
iras_/_iso _ studies detected 10@xmath10 - 10@xmath9m@xmath0 of warm ( @xmath11 @xmath1 50 - 100
k ) dust ( e.g. , ) . based on scuba observations at submm wavelengths ,
the presence of a cold ( @xmath11 @xmath1 15 - 20k ) dust reservoir of 2 - 4 m@xmath0 was inferred from the level of excess emission after subtraction of the non - thermal synchrotron emission component @xcite .
this large dust mass was , however , questioned and much of the excess submm emission was attributed to foreground interstellar dust @xcite .
@xcite interpreted the high level of polarisation at 850@xmath3 m as due to the alignment of 1m@xmath0 of dust with the magnetic field in the snr .
several analyses of 3.6 - 160@xmath3 m _ spitzer _ data of casa found warm dust masses ( 3@xmath1210@xmath13m@xmath0 , @xcite ; 0.02 - 0.054m@xmath0 , @xcite ; @xmath1 0.04m@xmath0 , @xcite ) significantly lower compared to the submm - derived cold dust masses . by including the _ spitzer _
irs spectra in the dust sed modelling , @xcite and @xcite showed that the composition of warm dust grains in casa could be studied in more detail . while the spectral characteristics of most of the dust in the bright ejecta knots and x - ray emitting shocked ejecta ( associated with bright [ ar ] and [ ar ] line features ) were found to be consistent with a magnesium silicate composition ( with varying relative abundance ratios of mg and si ) , a smooth spectral component associated with [ ne ] emitting regions does not show any silicate features and was best reproduced by a al@xmath14o@xmath15 ( or carbonaceous ) dust composition .
the largest dust mass component was associated with an inner cold dust reservoir @xmath160.1m@xmath0 with unidentified dust composition @xcite .
@xcite inferred a dust mass of 0.06m@xmath0 , with an average @xmath11 @xmath1 33k , from _ akari _ and _ blast _ observations covering the 50 to 500@xmath3 m wavelength range , but their large beam size ( 1.3@xmath2 , 1.6@xmath2 , 1.9@xmath2 at 250 , 350 and 500@xmath3 m , respectively ) hampered a clear separation of the interstellar and supernova dust material .
the higher angular resolution of _ herschel _ ( 18.2@xmath17 , 24.9@xmath17 , 36.3@xmath17 at 250 , 350 and 500@xmath3 m , respectively ) enabled to carry out a global fit to the insterstellar and supernova far - infrared dust emission .
they derived an sn dust mass of 0.075 m@xmath18 , emitting at t@xmath135 k , but were unable to determine whether any cooler dust was present in cas a due to the difficulty to distinguish between ism and sn dust emission at longer wavelengths ( 160 - 350@xmath3 m ) .
the above results based on _ spitzer _ , _ blast _ and _ herschel _ data were consistent with the dust evolution models of @xcite who predicted 0.08m@xmath0 of new grain material in casa at a temperature of @xmath140 k , of which 0.072m@xmath0 was predicted to reside in the inner remnant regions unaffected by the reverse shock .
m ( red to green colours ) and the _ hubble space telescope _ ( hst ) in the f625w , f775w and f850lp bands ( white to purple colours ) .
, width=321 ] in this paper , we combine _ spitzer _ , _ herschel _ , _ wise _ and _ planck _ photometric data from mid - infrared ( mir ) to millimetre ( mm ) wavelengths and _ spitzer _ and _ herschel _ spectroscopic observations .
we make a detailed study of the supernova dust emission on spatially resolved scales , which allows us to more accurately separate the intrinsic supernova dust emission from the non - thermal emission component and from the continuum emission by cold interstellar dust material . in section [ sect_data ]
, we present an overview of the observational datasets used for this analysis .
section [ modelcomp ] outlines the modelling technique for the various emission components ( synchrotron radiation , ism and sn dust emission ) . in section [ dustsed ] ,
the multi - wavelength sed is modelled on resolved scales in order to derive the distribution of temperatures and masses of newly formed dust grains in casa .
section [ discussion ] presents the sn dust masses and their uncertainties resulting from the sed modelling and discusses them in light of previous results .
our main conclusions are presented in section [ conclusions ] . in the appendix
, we provide an overview of various methods used for the analysis presented in this paper .
appendix [ planck_flux ] compares the different _ planck _
measurements for casa .
appendix [ sect_line ] explains how we modelled the contribution of line emission to the different broadband images .
appendix [ sect_rf ] discusses the photo - dissociation region ( pdr ) modelling technique that was used to constrain the interstellar radiation field ( isrf ) along the sightline of casa . in appendix
[ sect_globalfit ] , we present the results of a global sed fitting analysis .
appendix [ sect_modelverification ] verifies the applicability of our models by directly comparing the models to observations ( [ discuss_comparison ] ) and discusses the effect of small variations in the isrf on the sn dust masses ( [ discuss_is ] ) . in appendix
[ sec_modelprediction ] , we apply the model results to predict the relative contribution of line emission , synchrotron radiation , ism and sn dust emission at ir - submm wavelengths ( [ sect_casa_aftercorr ] ) , to present a model image at 850@xmath3 m ( [ discuss_850mu ] ) , to estimate the interstellar and sn visual extinction ( [ avmodel ] ) .
_ herschel _ observations of casa were obtained using guaranteed time ( gt ) contributed by the spire consortium specialist astronomy group 6 ( sag-6 ) and the german pacs consortium , as part of the mess gt program ( pi : m. groenewegen , ) .
table [ casa_herscheldata ] gives an overview of the different _ herschel _ datasets , observation identification numbers ( obsids ) , observing dates and integration times for the various _ herschel _ instruments .
& & [ y - m - d ] & [ @xmath19@xmath20@xmath21 & [ @xmath22 @xmath2 @xmath17 ] & [ @xmath24 + + casa & 1342188204 & 2009 - 12 - 17 & 23:23:22.72 & 58:48:53.38 & 1889 + casa & 1342188205 & 2009 - 12 - 17 & 23:23:23.11 & 58:48:53.01 & 1889 + casa & 1342188206 & 2009 - 12 - 17 & 23:23:19.01 & 58:48:51.25 & 1889 + casa & 1342188207 & 2009 - 12 - 17 & 23:23:20.23 & 58:48:56.86 & 1889 + + casa & 1342183681 & 2009 - 09 - 12 & 23:23:19.48 & 58:49:23.66 & 5005 + casa & 1342188182 & 2009 - 12 - 17 & 23:23:21.94 & 58:49:59.49 & 5010 + + casa sp1 & 1342212249 & 2011 - 01 - 01 & 23:23:28.61 & 58:48:59.17 & 2267 + casa
sp1 & 1342212250 & 2011 - 01 - 01 & 23:23:28.20 & 58:49:05.10 & 1139 + casa
sp2 & 1342212253 & 2011 - 01 - 01 & 23:23:24.94 & 58:51:26.98 & 2267 + casa sp2 & 1342212254 & 2011 - 01 - 01 & 23:23:24.50 & 58:51:33.31 & 1139 + casa
sp3 & 1342212257 & 2011 - 01 - 01 & 23:23:12.76 & 58:49:12.26 & 2267 + casa
sp3 & 1342212258 & 2011 - 01 - 01 & 23:23:13.19 & 58:49:18.37 & 1139 + casa
sp4 & 1342212245 & 2011 - 01 - 01 & 23:23:32.82 & 58:47:48.39 & 2267 + casa
sp4 & 1342212246 & 2011 - 01 - 01 & 23:23:32.41 & 58:47:54.44 & 1139 + casa
sp5 & 1342212251 & 2011 - 01 - 01 & 23:23:27.40 & 58:47:23.04 & 2267 + casa
sp5 & 1342212252 & 2011 - 01 - 01 & 23:23:26.99 & 58:47:28.89 & 1139 + casa
sp6 & 1342212243 & 2011 - 01 - 01 & 23:23:40.49 & 58:48:52.93 & 2267 + casa
sp6 & 1342212244 & 2011 - 01 - 01 & 23:23:39.61 & 58:49:05.82 & 1139 + casa
sp7 & 1342212247 & 2011 - 01 - 01 & 23:23:30.45 & 58:50:10.21 & 2267 + casa sp7 & 1342212248 & 2011 - 01 - 01 & 23:23:30.03 & 58:50:16.42 & 1139 + casa
sp8 & 1342212255 & 2011 - 01 - 01 & 23:23:16.84 & 58:47:41.01 & 2267 + casa
sp8 & 1342212256 & 2011 - 01 - 01 & 23:23:16.43 & 58:47:46.89 & 1139 + casa
sp9 & 1342212259 & 2011 - 01 - 01 & 23:23:12.87 & 58:48:15.45 & 2267 + casa
sp9 & 1342212260 & 2011 - 01 - 01 & 23:23:12.46 & 58:48:21.36 & 1139 + + casa - centre & 1342202265 & 2010 - 08 - 08 & 23:23:29 & 58:48:54 & 3476 + casa - north & 1342204034 & 2010 - 08 - 23 & 23:23:25 & 58:50:55 & 3476 + casa - north - west & 1342204033 & 2010 - 08 - 23 & 23:23:14 & 58:49:08 & 3476 + the photodetector array camera and spectrometer ( pacs , ) observed casa on december 17 , 2009 and january 1 , 2011 , respectively .
the pacs photometry data were obtained in parallel scan - map mode with two orthogonal scans of length 22@xmath2 observed at the nominal scan speed of 20 @xmath17 s@xmath25 in the blue+red and green+red filters .
the total on - source integration in the blue ( 70@xmath3 m ) and green ( 100@xmath3 m ) filters was 2376s , while the integration time in the red filter ( 160@xmath3 m ) was 4752s .
the fwhm of the pacs beam corresponded to 5.6@xmath17 , 6.8@xmath17 and 11.4@xmath17 at 70 , 100 and 160@xmath3 m , respectively ( see pacs observers manual ) .
the pacs photometry data have been reduced with the latest ` hipe ` v14.0.0 @xcite using the standard script which allows us to combine the scan and cross scans for a single field into one map .
the script takes the level 1 data from the _ herschel _ science archive ( hsa ) , masks glitches , subtracts the baselines for separate scan legs , applies a drift correction and finally merges all scan and cross scan pairs to a final output map with default pixel sizes of 1.6@xmath17 for the blue and green filters , and 3.2@xmath17 for the red filter .
to correct for the shape of the spectrum , we apply colour corrections to the pacs maps ( see the pacs calibration document picc - me - tn-038 ) . with a dominant contribution of warm ( @xmath11 @xmath1 80k )
supernova dust emission at pacs 70@xmath3 m ( see table [ table_fluxfraction ] ) , we apply a colour correction factor of 0.989 . the pacs160@xmath3 m emission is shown later to be dominated by emission from ism dust irradiated by a radiation field @xmath260.6@xmath27 corresponds to the average fuv interstellar radiation normalised to the units of the @xcite field , i.e. @xmath27=1.6@xmath1210@xmath13 erg s@xmath25 cm@xmath9 @xcite .
a normalisation to the @xcite field ( indicated as @xmath29 ) is frequently used and is related to the @xcite field as @xmath27 = 1.7@xmath30 . ] .
the colour correction for a blackbody with temperature @xmath31 = 17.6k is 0.967 at 160@xmath3 m . for the pacs 100@xmath3 m band
, we find an equal contribution by dust emission from ism and sn dust .
the colour correction factor for the pacs 100@xmath3 m image is , therefore , calculated as the average ( 1.038 ) of the colour correction factors for blackbodies with temperatures of @xmath31 = 17.6k ( 1.069 ) and @xmath31 = 100k ( 1.007 ) .
the latter correction factors , and any factors mentioned in the remainder of this paper are multiplicative factors .
the pacs maps are assumed to have a calibration uncertainty of 5@xmath5 @xcite .
pacs spectroscopy data in pacs - ifu mode ( fov @xmath1 47@xmath17 ) were obtained at nine different positions in casa , mainly targeting the shocked and the central regions of the remnant ( see figure [ casa_aor_pacs_ifu ] ) .
each position was observed in the wavelength ranges 51 - 72@xmath3 m and 102 - 146@xmath3 m ( range mode sed b2a + short r1 ) and in the ranges 70 - 105@xmath3 m and 140 - 220@xmath3 m ( range mode sed b2b + long r1 ) .
the pacs spectra were reduced to level 2 using the standard pacs chopped large range scan and sed pipeline in ` hipe ` v14.0.0 using the pacs@xmath32cal@xmath3232@xmath320 calibration file .
the pacs ifu line measurements are assumed to have a calibration uncertainty of 13@xmath5 and 16@xmath5 short- and longwards of 150@xmath3 m , respectively , based on a combination of the absolute calibration uncertainty of 12@xmath5 and the relative uncertainty due to spaxel variations of 5@xmath5 ( @xmath8150@xmath3 m ) and 10@xmath5 ( @xmath33150@xmath3 m ) .
m image of cas a , overlaid with the on - source regions targeted with the pacs ifu spectrometer in the chopping mode .
the off - source chopped positions were offset by 6@xmath2 at a pa of @xmath1240 degrees.,width=321 ] the spectral and photometric imaging receiver ( spire , ) observed casa on august 8th and 23rd , 2010 , and on september 12th and december 17th , 2009 , respectively .
the spire observations consisted of two orthogonal scans observed at the nominal scan speed of 20@xmath17 s@xmath25 simultaneously at each wavelength ( 250 , 350 and 500@xmath3 m ) covering a 32@xmath2 @xmath12 32@xmath2 region centred on cas a with an integration time of 2876s for each field .
the fwhm of the spire beam corresponded to 18.2@xmath17 , 24.9@xmath17 and 36.3@xmath17 at 250 , 350 and 500@xmath3 m , respectively ( see spire observers manual ) .
the data were processed using hipe version v14.0.0 using the standard pipeline for the spire large map mode with extended source calibration . from level 0.5 to level 1 , an electrical crosstalk , temperature drift and bolometer time response correction is applied and a wavelet deglitching algorithm is run for all building blocks . to process the level 1 building blocks
, we use a script to combine data from different scans . on the combined data set
, we ran the destriper ( instead of the baseline subtraction ) to obtain an optimum fit between all timelines .
the _ planck _
hfi maps at 857 and 545ghz ( 350 and 550@xmath3 m ) were , furthermore , used to determine the absolute scaling of the spire maps with extended emission and 0.9@xmath5 lower at 250@xmath3 m and 350@xmath3 m and 0.5@xmath5 higher at 500@xmath3 m compared to the original spire flux calibration ( without including the _ planck _ maps to determine the absolute calibration of spire images ) . ] .
we applied colour correction factors to the spire 250@xmath3 m ( 0.9875 ) , 350@xmath3 m ( 0.9873 ) , and 500@xmath3 m ( 0.9675 ) maps , appropriate for a spectrum with @xmath34 @xmath35 @xmath36 .
the calibration uncertainties for the spire images are assumed to be 4@xmath5 , resulting from the quadratic sum of the 4@xmath5 absolute calibration error from the assumed models used for neptune ( spire observers manualom.html ] ) and the random uncertainty of 1.5@xmath5 on the repetitive measurements of neptune @xcite .
the spire fts spectra were obtained in sparse spatial sampling and high - spectral resolution mode , covering the 194 - 671@xmath3 m wavelength range .
three different regions ( centre , north , north - west ) were targeted ( see figure [ casa_aor_spire_fts ] ) with the two arrays of the bolometer detectors , each with 24 repetitions .
the 35 detectors of the ssw ( spire short wavelength ) array covered the 194 - 313@xmath3 m range , while the slw ( spire long wavelength ) array of 19 detectors covered the 303 - 671@xmath3 m wavelength range .
the ssw and slw detectors have an average fwhm of 19@xmath17 and 34@xmath17 , respectively @xcite .
the spire fts data were reduced in ` hipe ` v14.0.0 , with version spire@xmath32cal@xmath3214@xmath323 of the calibration files including the latest corrections that match the fts extended calibration with the spire photometer . we used the standard pipeline in ` hipe ` for the reduction of single pointing spire spectrometer observations , with extended source calibration and without apodisation .
the standard pipeline included a first and second order deglitching procedure , non - linearity and phase corrections , baseline subtraction , and corrections for the telescope and instrument emission .
the spectral lines in the spire fts data were fitted with the spire spectrometer line fitting algorithm in ` hipe ` using a sinc function to model the instrumental line shape @xcite .
in addition to the formal uncertainties from line fitting , we add a 10@xmath5 calibration uncertainty @xcite to the line flux uncertainties .
m image of cas a , overlaid with the three spire fts pointings with individual positions for the ssw and slw detectors indicated as white and orange circles , respectively .
, width=321 ] the infrared array camera ( irac , @xcite ) , multi - band imaging photometer ( mips , @xcite ) and infrared spectrograph ( irs , @xcite ) on board the _ spitzer _ space telescope @xcite have all targeted casa .
the mips data at 24 , 70 and 160@xmath3 m were observed as part of the _ spitzer _ early release observation program ( i d 718 , pi : g. rieke ) on november 30 , 2003 .
casa was mapped with mips at medium scan speed over a total area of 12.7@xmath2 @xmath12 30@xmath2 .
more details about the observing strategy and a detailed analysis of the mips data for casa are presented in @xcite .
we used the irac data for casa observed as part of the _ spitzer _ program _ the evolution of dust in cassiopeia a _ ( i d 3310 , pi : l. rudnick ) on january 18 , 2005 ( see @xcite for a description of the observations ) .
we retrieved the irac 3.6 and 8@xmath3 m , and mips 24@xmath3 m final data products ( with 10.4s , 10.4s and 3.67s integration time per pixel , respectively ) from the _ spitzer _ heritage archive .
extended source correction factors of 0.91 and 0.74 have been applied to the irac3.6@xmath3 m and irac8@xmath3 m images , respectively , following the recommendations of the irac instrument handbook .
the irac3.6@xmath3 m map ( dominated by synchrotron emission from casa ) did not require a correction for the shape of the spectrum ( assumed to be a power law spectrum @xmath34
@xmath35 @xmath37 with @xmath38 anywhere between @xmath39 and @xmath40 , see section [ sect_synchr ] ) .
we did correct the irac 8@xmath3 m emission ( arising primarily from hot sn dust in casa ) with a colour correction factor of 0.9818 assuming a blackbody spectrum with a temperature of 400k .
the irac3.6@xmath3 m is contaminated by many field stars , which prevents the determination of the fainter synchrotron emission at those wavelengths .
we therefore manually selected sixty - four bright targets in the field of casa and replaced the emission within an aperture with radius @xmath41 ( selected to encompass the star s emission ) with random background noise .
the mean value of the background and background variation has been measured within an annulus with inner radius @xmath41 and outer radius @xmath42 across all stars . ] .
the flux calibration uncertainties in the irac3.6@xmath3 m and 8@xmath3 m , and mips24@xmath3 m maps are assumed to be 10@xmath5 ( recommended by the irac instrument handbook ) and 4@xmath5 @xcite , respectively .
we also used the _ spitzer _
irs spectra processed by @xcite .
more details about the observations and analysis can be retrieved from their paper .
the _ spitzer _
irs observations were taken as part of the same _ spitzer _ program on january 13 , 2005 during a total observing time of 11.5 hours .
nearly the entire remnant was mapped with the irs short - low ( sl : 5 - 15@xmath3 m ) and long - low ( ll : 15 - 40@xmath3 m ) filters using 16@xmath12360 and 4@xmath1291 pointings with spectra obtained every 5@xmath17 and 10@xmath17 , respectively . the spectra were reduced using the cubism package @xcite with the standard irs pipeline ( version s12 ) , including background subtraction and corrections for extended source emission .
we derived a global irs spectrum for casa by extracting the emission within an aperture with radius r = 165@xmath17 .
the sl and ll spectra were scaled to have similar emission in overlapping wavebands .
the global spectrum was multiplied by a scaling factor of 1.12 to match the global photometry of the continuum emission in the irac8@xmath3 m , wise12@xmath3 m and 22@xmath3 m and mips24@xmath3 m wavebands . based on this global spectrum
, we derived the continuum emission for casa at 17 and 32@xmath3 m to constrain the sed fitting procedure in section [ dustsed ] .
the total flux uncertainties ( about 12@xmath5 ) for the irs spectra and the continuum fluxes were derived by combining the uncertainties in the scaling factors to match the emission in the broadband filters ( 11@xmath5 ) and the absolute calibration uncertainty ( 5@xmath5 ) . the wide - field infrared survey explorer ( _ wise _ , @xcite ) observed casa in four photometric bands at 3.4 , 4.6 , 11.6 and 22.0@xmath3 m with a resolution of fwhm = 6.1@xmath17 , 6.4@xmath17 , 6.5@xmath17 , and 12@xmath17 , respectively . we retrieved the _ wise _ 11.6 and 22.0@xmath3 m maps for casa from the nasa / ipac infrared science archive .
the _ wise _ images ( in units of dn ) were converted to vega magnitudes using the photometric zero point magnitudes indicated in the image headers ( with the magzp keyword ) . to convert the _ wise _ images to flux densities
, we applied the zero magnitude flux densities reported in the explanatory supplement to the neowise data release products .
three different types of correction factors needed to be applied to extended sources ( jarrett et al .
the first aperture correction factor corrected for the psf profile fitting that was used for the _ wise _ absolute photometric calibration , which required corrections of -0.030 mag and 0.029 mag for the wise 11.6 and 22.0@xmath3 m bands .
a second correction factor is the colour correction , which accounts for the shape of the spectrum of casa .
since the emission at mid - infrared wavelengths is mostly dominated by warm sn dust , we applied the correction factors applicable for blackbodies with temperatures of 200k and 100k to the _ wise _
11.6@xmath3 m ( 1.0006 ) and 22@xmath3 m ( 1.0032 ) bands .
the dust temperatures that dominate the emission at those wavelengths have been inferred from mid - infrared studies of casa @xcite .
the last correction factor ( 0.92 ) , which accounts for the calibration discrepancy between _ wise _ photometric standard blue stars and red galaxies , only needed to be applied to the _ wise _
22@xmath3 m band .
we assume calibration uncertainties of 4.5@xmath5 and 5.7@xmath5 @xcite for the _ wise _ 12 and 22@xmath3 m filters .
_ planck _ observed the entire sky in nine submm and mm wavebands during the mission lifetime ( 2009 - 2013 ) .
the _ planck _
satellite had two instruments , with the high frequency instrument ( hfi , ) operating at 857 , 545 , 353 , 217 , 143 and 100 ghz ( or wavelengths of 350 , 550 , 850 @xmath3 m and 1.38 , 2.1 and 3 mm , respectively ) and the low frequency instrument ( lfi , ) covering the 70 , 40 and 33 ghz frequencies ( or wavelengths of 4.3 , 6.8 and 10 mm ) .
for this analysis , we have used the customised _ planck _
flux measurements from derived from aperture photometry within an aperture optimised for the angular size of casa and the beam size at every _
planck _ frequency ( see table [ table_planckflux ] , last row ) . in appendix
[ planck_flux ] , we give an overview of the various _ planck _
measurements available for casa and discuss our choice for these flux measurements . the background in each of the maps has been determined by measuring the backgrounds in a sufficiently large number of apertures with radius = 4@xmath12fwhm at each waveband . the number of background apertures depended on the size of the background region available , and was typically about 20 .
the irac and mips images have a small field - of - view , while the _ herschel _ images are dominated by emission from interstellar material in the perseus arm which makes it hard to find emission - free regions .
we selected regions off the remnant with the final background level chosen to be the mean value from all background apertures .
we have chosen to subtract these backgrounds from the images before convolution of all images to the 500@xmath3 m resolution . to compare the emission of casa across all wavebands in an unbiased way , we convolved all ir / submm images to the same resolution , which was chosen to be the resolution of the spire 500@xmath3 m image ( fwhm = 36.3@xmath17 ) .
we used the convolution kernels from @xcite for the convolution of these images to the spire500@xmath3 m resolution .
all convolved images were rebinned to the pixel grid of the spire 500@xmath3 m map with pixel size of 14@xmath17 ( or 0.23 pc at the adopted distance of 3.4kpc for cas a , @xcite ) .
we have furthermore verified that a convolution with kernels produced based on the more recent _ herschel _ pacs and spire psf models derived from dedicated vesta and mars observations and published by would not affect the results published in this work .
we produced such convolution kernels for the specific _ herschel _ observations of casa that account for the position angle of the z - axis of the telescope during the observations of the source and target image , respectively , using the ` pypher ` ( python - based psf homogenization kernels production ) software package . with variations in the global fluxes of casa of less than 1@xmath5 ( except at 250@xmath3 m with a 2@xmath5 offset ) between the two sets of convolved images , we conclude that the type of kernels used for the convolution of the _ herschel _ maps of casa will not affect the determination of sn dust masses in casa .
for the sed modelling procedure described in section [ dustsed ] , we need to know the uncertainty on the flux in every pixel .
these uncertainties determine how well the model fits the observations , and play an important role in setting the uncertainties on the contributions from different emission components .
the main uncertainties are due to errors on the determination of the background levels , calibration uncertainties and uncertainties in the synchrotron subtraction .
the background uncertainties are driven by both large scale background variations and pixel - by - pixel noise .
the two independent background errors are computed as the standard deviation of the mean background values derived for different background regions and the mean of the standard deviation of the pixel - by - pixel variation in different background regions , respectively .
the uncertainty associated with the subtraction of the synchrotron emission is calculated based on the uncertainties associated with the spectral index and normalisation factor determined from the _ planck _ data ( see section [ sect_synchr ] ) .
table [ table_fluxfraction ] ( first column ) provides an overview of the global photometric measurements for casa .
several other works have studied the infrared emission from casa , which makes it possible to compare our global photometry to other published fluxes . used the same set of _ herschel _ observations , but their data were reduced with an earlier version of the ` hipe ` data reduction pipeline .
it is , therefore , of interest to compare the flux measurements reported by to this work for the same aperture .
while the pacs70@xmath3 m photometry is consistent within the error bars ( @xmath43=179@xmath4411 jy here and @xmath43=169@xmath4417 jy from ) , the other pacs photometric measurements published by ( @xmath45=192@xmath4419 jy , @xmath46=166@xmath4417 jy ) are 17.7@xmath5 and 29.7@xmath5 lower compared to the values derived here . in the 160@xmath3 m channel we might expect an opposite trend due to the optical field distortion that was not applied in ` hipe ` 12 and all earlier versions ( which caused an overestimate of the flux by 6 - 7@xmath5 ) .
we believe the flux difference can mainly be attributed to the different data reduction techniques that were used to process the data . while the first ` hipe ` versions only allowed one to reduce data with the standard map making tool ` photproject ` , the latest version of the data reduction was performed with ` scanamorphos ` . while ` photproject ` applied a high - pass filtering technique to remove the 1/f noise ( mainly due to data with low spatial frequencies or large scale emission features in the maps ) , ` scanamorphos ` does not make any specific assumptions to model the low frequency noise and takes into account the redundancy of observations . `
scanamorphos ` has been shown to give a better estimate of the background level in pacs maps compared to the ` photproject ` map - making algorithm which can remove or underestimate the emission of extended structures .
given that the field around casa is populated with extended emission originating from ism dust , we believe that the discrepancy between fluxes is largely due to different estimates of the background levels . in the spire wavebands ,
our integrated fluxes are also higher by 8.3@xmath5 , 17.8@xmath5 and 12.2@xmath5 compared to the values published by ( @xmath47=168@xmath4417 jy , @xmath48=92@xmath4410 jy , @xmath49=52@xmath447 jy ) , but close to being consistent within the error bars . based on the spire beam areas that were assumed during the early days of _ herschel _ observations ( 501 , 944 and 1924 arcsec@xmath50 at 250 , 350 and 500@xmath3 m , respectively , ) compared to the more recent estimates ( 464 , 822 and 1768 arcsec@xmath50 )
, we would expect the latest flux densities to be lower .
the opposite trend suggests that variations in the background level determination play a prominent role in explaining the different photometric measurements .
other than by _ herschel _ , casa has been observed by several other space and ground - based facilities at similar wavelengths . based on _ spitzer _ mips data
, @xcite reported a flux density @xmath51=107@xmath4422 jy .
observations by _ akari _ yielded flux densities of @xmath34=71@xmath4420jy , 105@xmath4421jy and 92@xmath4418jy at 65 , 90 and 140@xmath3 m @xcite , respectively .
@xcite also remeasured the _ _
iso__170@xmath3 m flux ( @xmath52=101@xmath4420jy ) , and reported _ blast _ 250 , 350 and 500@xmath3 m photometry ( @xmath53=76@xmath4416jy , @xmath54=49@xmath4410jy , @xmath55=42@xmath448jy ) .
with scuba , @xcite measured flux densities of @[email protected] and @[email protected] . while the scuba450@xmath3 m flux is consistent with the spire500@xmath3 m photometry ( see table [ table_fluxfraction ] ) , the mips , _ akari _ , _ iso _ and _ blast _ measurements are all significantly lower ( up to a factor of 2.5 ) compared to the _ herschel _ pacs and spire fluxes derived in this paper . came to a similar conclusion and attributed the flux discrepancies to the better resolution in the _ herschel _ maps which allows a more accurate background subtraction . due to the highly structured emission of the ism material in the surroundings of casa
, it can become difficult to determine the background level in an image with poor spatial resolution , due to ism dust confusion . while the high resolution of _ herschel _ images allows us to better resolve the ism emission from the true background in the ir / submm images , resulting in the recovery of higher flux densities for casa ,
the main difference of our work compared to previous studies , analysing the same set of _ herschel _ observations , is the detailed modelling on spatially resolved scales of the different emission components contributing along the sight line towards casa . in support of this claim
, we repeated the modelling based on the _ herschel _ fluxes published by and retrieved a total sn dust mass higher by a factor of three compared to the global analysis of the _ herschel _ data of casa by ( see section [ comparebefore.sec ] ) . [ cols="<,^,^,^,^,^,^,^,^,^ " , ] +
we use the global fluxes ( see table [ table_fluxfraction ] ) , after subtraction of synchrotron radiation and line contributions , to determine the total amount of emission from ism dust projected on casa and sn dust in casa .
similar to the spatially resolved sed fitting procedure in section [ dustsed ] , the global emission spectrum is fitted with a four - component sed model with an ism dust component and hot , warm and cold sn dust components .
the ism model is based on the themis dust model from illuminated by a radiation field of @xmath58=0.6@xmath27 scaled to the dust mass that is derived from the spire500@xmath3 m flux and the relations in eq .
[ eq1_label ] . the hot sn dust component is assumed to have a fixed dust composition of silicates with a low mg / si ratio of 0.7 ( i.e. , mg@xmath59sio@xmath60 ) with dust temperatures in the range 100 - 200k .
the sed modelling procedure is repeated for the dust species listed in table [ table_dustspecies ] under the assumption that it dominates the dust composition of the warm+cold sn dust components .
we assume dust temperatures between 40 - 100k and 10 - 40k for the warm and cold sn dust components .
table [ table_sedfit_global ] gives an overview of the best fitting parameters derived from the global sed fitting .
& & & & & + & @xmath11 & @xmath61 & @xmath11 & @xmath61 & @xmath11 & @xmath61 & @xmath61 & @xmath62 & @xmath11 & @xmath61(lower ) + & [ k ] & [ 10@xmath13 m@xmath0 ] & [ k ] & [ 10@xmath9 m@xmath0 ] & [ k ] & [ m@xmath0 ] & [ m@xmath0 ] & & [ k ] & [ m@xmath0 ] + + mgsio@xmath15 & 120 & [email protected] & 66 & [email protected] & 24 & [email protected] & [email protected] & 0.18 & 30 & [email protected] + mg@xmath59sio@xmath60 & 190 & [email protected] & 56 & [email protected] & 25 & [email protected] & [email protected] & 0.81 & 26 & [email protected] + mg@xmath64sio@xmath65 & 130 & [email protected] & 66 & [email protected] & 26 & [email protected] & [email protected] & 0.55 & 30 & [email protected] + al@xmath14o@xmath15 ( porous ) & 115 & [email protected] & 72 & [email protected] & 24 & [email protected] & [email protected] & 0.19 & 33 & [email protected] + caal@xmath66o@xmath67 & 105 & [email protected] & 74 & [email protected] & 16 & [email protected] & [email protected] & 1.34 & 20 & [email protected] + am .
carbon `` ac1 '' & 115 & [email protected] & 85 & [email protected] & 27 & [email protected] & [email protected] & 0.50 & 40 & [email protected] + a - c ( @xmath68=0.1ev ) & 115 & [email protected] & 77 & [email protected] & 25 & [email protected] & [email protected] & 0.31 & 37 & [email protected] + + mgsio@xmath15 & * 120 * & * [email protected] * & * 68 * & * [email protected] * & * 28 * & * [email protected] * & * [email protected] * & * 0.20 * & * 38 * & * [email protected] * + mg@xmath59sio@xmath60 & 200 & [email protected] & 55 & [email protected] & 21 & [email protected] & [email protected] & 0.58 & 29 & [email protected] + mg@xmath64sio@xmath65 & * 130 * & * [email protected] * & * 66 * & * [email protected] * & * 29 * & * [email protected] * & * [email protected] * & * 0.47 * & * 38 * & * [email protected] * + al@xmath14o@xmath15 ( porous ) & 110 & [email protected] & 72 & [email protected] & 27 & [email protected] & [email protected] & 0.22 & 40 & [email protected] + caal@xmath66o@xmath67 & 100 & [email protected] & 74 & [email protected] & 17 & [email protected] & [email protected] & 1.26 & 24 & [email protected] + am .
carbon `` ac1 '' & 115 & [email protected] & 84 & [email protected] & 29 & [email protected] & [email protected] & 0.29 & 79 & [email protected] + a - c ( @xmath68=0.1ev ) & 115 & [email protected] & 77 & [email protected] & 27 & [email protected] & [email protected] & 0.24 & 40 & [email protected] + + mgsio@xmath15 & 115 & [email protected] & 68 & [email protected] & 39 & [email protected] & [email protected] & 0.34 & 63 & [email protected] + mg@xmath59sio@xmath60 & 185 & [email protected] & 55 & [email protected] & 30 & [email protected] & [email protected] & 0.61 & 56 & [email protected] + mg@xmath64sio@xmath65 & 125 & [email protected] & 63 & [email protected] & 33 & [email protected] & [email protected] & 0.45 & 64 & [email protected] + al@xmath14o@xmath15 ( porous ) & 110 & [email protected] & 70 & [email protected] & 12 & [email protected] & [email protected] & 0.28 & 74 & [email protected] + caal@xmath66o@xmath67 & 100 & [email protected] & 75 & [email protected] & 17 & [email protected] & [email protected] & 1.24 & 75 & [email protected] + am .
carbon `` ac1 '' & 115 & [email protected] & 82 & [email protected] & 23 & [email protected] & [email protected] & 0.28 & 86 & [email protected] + a - c ( @xmath68=0.1ev ) & 115 & [email protected] & 75 & [email protected] & 22 & [email protected] & [email protected] & 0.31 & 77 & [email protected] + similarly to the resolved sed fitting procedure , the sed models with a mg@xmath59sio@xmath60 , caal@xmath66o@xmath67 and al@xmath14o@xmath15 dust composition for the warm and cold sn dust components result in sn dust masses that are unrealistically high given the constraints on metal production based on nucleosynthesis models for supernova type ii and type iib events for a 30m@xmath0 and 18m@xmath0 progenitor , respectively .
these grains can therefore be ruled out as the dominant dust species that form in casa . for carbonaceous grains ,
the best fit would imply a dust mass of 0.3 - 0.5m@xmath0 with a lower limit of 0.01 - 0.03m@xmath0 of carbon grains .
besides 0.3 - 0.5m@xmath0 of carbon locked up in grains being considered unrealistic due to the lack of sufficient carbon production by the progenitor of casa , the ejecta of casa have been shown to be predominantly oxygen - rich .
assuming that the newly formed grains are silicates ( mgsio@xmath15 or mg@xmath64sio@xmath65 ) , we would need 0.4 to 0.6m@xmath0 of these oxygen - rich grains , with a lower limit of 0.1m@xmath0 , at a temperature of @xmath11@xmath130k to reproduce the global emission spectrum of casa .
figure [ ima_casa_globalsed ] shows the best fitting global sed model assuming an ism dust sed model with radiation fields @xmath58=0.3@xmath27 , 0.6@xmath27 and 1.0@xmath27 . for these sed fits ,
we have assumed mg@xmath59sio@xmath60 dust for the hot sn dust component and mgsio@xmath15 dust for the warm and cold sn dust components . even though the irac8@xmath3 m and wise12@xmath3 m data points were not used to constrain the multi - component ism+sn model
, the best fitting model sed seems to reproduce the mir emission in those wavebands within the error bars ( with the exception of the wise12@xmath3 m datapoint for the g=0.3@xmath27 model ) .
the good agreement between model and observations seems to indicate that a combination of a warm sn dust component and the emission features originating from small hydrocarbons ( or pahs ) in the diffuse ism are able to account for the mir emission observed towards casa .
the scuba850@xmath3 m data point is also omitted from the fitting procedure .
we will discuss the scuba850@xmath3 m observations in appendix [ discuss_850mu ] .
the sn dust masses derived from the spatially resolved sed fits ( see section [ dustsed ] ) are in overall agreement with the dust masses derived from the global sed fits within the limits of observational uncertainties . since the variation in the sn dust emission spectrum and ism dust contribution within an aperture of radius 165@xmath17 can introduce possible uncertainties in the global sed parameters ( because local variations in ism dust contribution and sn dust mass and temperature are averaged out in the global spectrum that includes all the emission within a r = 165@xmath17 region ) , we attach more weight to the results of the spatially resolved sed fitting analysis .
m photometry after subtraction of the contribution from line emission and synchrotron radiation , assuming a galactic ism dust model illuminated by a radiation field g=0.3@xmath27 ( top ) , 0.6@xmath27 ( middle ) and 1.0@xmath27 ( bottom ) .
the legend in the top panel explains the line formatting and colours associated to the various sn and ism dust components.,title="fig:",width=330 ] + m photometry after subtraction of the contribution from line emission and synchrotron radiation , assuming a galactic ism dust model illuminated by a radiation field g=0.3@xmath27 ( top ) , 0.6@xmath27 ( middle ) and 1.0@xmath27 ( bottom ) .
the legend in the top panel explains the line formatting and colours associated to the various sn and ism dust components.,title="fig:",width=330 ] + m photometry after subtraction of the contribution from line emission and synchrotron radiation , assuming a galactic ism dust model illuminated by a radiation field g=0.3@xmath27 ( top ) , 0.6@xmath27 ( middle ) and 1.0@xmath27 ( bottom ) .
the legend in the top panel explains the line formatting and colours associated to the various sn and ism dust components.,title="fig:",width=330 ]
here , we compare the modelled global dust sed and spatially resolved dust emission maps to observations to verify whether the multi - component ism+sn model is capable of reproducing the observed ir - submm seds . the global relative offsets between observed and modelled fluxes ( i.e. , @xmath69-@xmath70/@xmath69 ) at various wavelengths are shown in figure [ ima_casa_globalsed_res ] for models with the three different ism dust models ( g = 0.3 , 0.6 and 1.0@xmath71 ) and various sn dust species .
all dust sed models show small residuals , independent of the assumed dust composition or strength of the illuminating isrf , and reproduce the mips24@xmath3 m , irs32@xmath3 m continuum , pacs70 , 100 and 160@xmath3 m and spire250 , 350 and 500@xmath3 m datapoints within the error bars .
while the scuba850@xmath3 m is not used to constrain the sed fitting procedure , all models are at the low end of the observed scuba 850@xmath3 m data flux ( see also appendix [ discuss_850mu ] ) . , for the global sed models .
the top , middle and bottom panels show the model deviations for ism dust models with g=0.3@xmath71 , 0.6@xmath71 and 1.0@xmath71 , respectively .
the model offsets are shown for different sn dust compositions ( mgsio@xmath15 in black , amorphous carbon in green and mg@xmath64sio@xmath65 in red ) used for the warm and cold components .
the rightmost datapoint corresponds to the scuba850@xmath3 m flux reported by @xcite .
its offset from the best fitting models is discussed in section [ discuss_850mu].,title="fig:",width=340 ] + , for the global sed models .
the top , middle and bottom panels show the model deviations for ism dust models with g=0.3@xmath71 , 0.6@xmath71 and 1.0@xmath71 , respectively .
the model offsets are shown for different sn dust compositions ( mgsio@xmath15 in black , amorphous carbon in green and mg@xmath64sio@xmath65 in red ) used for the warm and cold components .
the rightmost datapoint corresponds to the scuba850@xmath3 m flux reported by @xcite .
its offset from the best fitting models is discussed in section [ discuss_850mu].,title="fig:",width=340 ] + , for the global sed models .
the top , middle and bottom panels show the model deviations for ism dust models with g=0.3@xmath71 , 0.6@xmath71 and 1.0@xmath71 , respectively .
the model offsets are shown for different sn dust compositions ( mgsio@xmath15 in black , amorphous carbon in green and mg@xmath64sio@xmath65 in red ) used for the warm and cold components .
the rightmost datapoint corresponds to the scuba850@xmath3 m flux reported by @xcite .
its offset from the best fitting models is discussed in section [ discuss_850mu].,title="fig:",width=340 ] to verify whether the multi - component model reproduces the observed spectrum in every pixel , in figure [ ima_casa_resolvedsed_resmap ] we compare the emission predicted by the sed model with observations . to highlight any model deviations from the observations ,
the residual images ( i.e. , observed flux - modelled flux / observed flux ) and histogram of residuals are shown in the third and fourth columns of figure [ ima_casa_resolvedsed_resmap ] , respectively .
the residuals between the model and observations show random offsets centred around 0 at every wavelength rather than a systematic offset , which suggests that the multi - component sn+ism dust model is adequate to fit the ir / submm sed . at 500@xmath3 m ,
several pixels show an offset between observations and model up to -50@xmath5 .
given that the offset arises in regions where the ism contribution is low , and the model overestimates the ism dust emission in those regions , we do not believe that this discrepancy will affect the determination of the sn dust mass in casa . in this section
we analyse the effects of variations in the strength of the isrf illuminating the ism dust along the sight line of casa on the determination of the sn dust mass . for the multi - component
sed fitting procedures on resolved and global scales ( see section [ dustsed ] and appendix [ sect_globalfit ] , respectively ) , we have assumed a constant radiation field strength of @xmath58=0.6g@xmath72 .
based on local variations in @xmath58 in the areas surrounding casa ( see fig .
[ ima_casa_interstellar ] , left panel ) it is however likely that regional variations occur , which might have an effect on the sn dust temperatures and dust masses derived for casa .
the three sets of sed models that were run for @xmath58 = 0.3@xmath27 , 0.6@xmath27 and 1.0@xmath27 give an idea of the influence of the variation in radiation field illuminating the ism material on the sn dust mass and temperature estimates .
we have not considered the results of ism models with a stronger isrf ( i.e. , @[email protected]@xmath27 ) due to an overestimation of ir - submm fluxes for casa upon scaling these ism dust model seds to the spire500@xmath3 m flux ( after subtraction of the synchrotron radiation ) .
the best fitting sn dust masses and temperatures derived for these three sets of sed models can be examined in tables [ table_sedfit_resolved ] and [ table_sedfit_global ] for the spatially resolved and global sed fits , respectively . comparing the sed fitting results for different isrfs
, a general trend emerges that ism models with lower dust temperatures ( corresponding to lower @xmath58 ) yield larger sn dust masses , which trivially follows from the reduced ir / submm emission of ism dust and is compensated for with an increase in the sn dust emission at those wavelengths .
figure [ ima_casa_resolvedsed_maps_g0 ] demonstrates how the derived sn dust temperatures , and consequently the sn dust masses , are sensitive to the radiation field that is assumed to illuminate the ism dust .
the overall distribution of sn dust temperatures within casa is similar for the three models regardless of the isrf scaling factor . but with cold sn dust temperatures ranging from 20k to 40k ( middle panels ) , the total dust mass can differ by factors of a few .
more specifically , we derive dust masses of 1.4m@xmath0 , 0.5m@xmath0 and 0.2m@xmath0 from the resolved sed maps assuming that the warm and cold sn dust is composed of mgsio@xmath15 grains with ism dust illuminated by a radiation field g=0.3g@xmath72 , 0.6g@xmath72 and 1.0g@xmath72 , respectively . similarly , the resolved sed fitting procedure for a mg@xmath64sio@xmath65 dust composition results in dust masses of 0.9m@xmath0 , 0.3m@xmath0 and 0.1m@xmath0 for different isrfs , respectively .
the sn dust masses derived for a g=0.3g@xmath72 ism model exceed the maximum dust masses ( 0.4 - 1.4m@xmath0 and 0.3 - 0.9m@xmath0 for mgsio@xmath15 and mg@xmath64sio@xmath65 dust ) derived from the metal production predicted by nucleosynthesis models ( under the assumption that all the metals are locked in dust grains ) , and are therefore considered unrealistic .
given the uncertainties on the nucleosynthesis models @xcite and the mass of the progenitor , we argue that the condensation of 0.3 - 0.5m@xmath0 ( for a 0.6g@xmath72 ism model ) or 0.1 - 0.2m@xmath0 ( for a 1.0g@xmath72 ism model ) of dust in casa is reasonable . based on a total ejecta mass estimate of 2 - 4m@xmath0 derived both from x - ray observations of casa and kinetic arguments to determine the mass and explosion energy from optical observations @xcite , these dust masses imply that the condensation of elements has been extremely efficient .
given that the progenitor has been predicted to have had a mass of 23 - 30 m@xmath0 , the total ejecta mass could be larger than 4 m@xmath0 , with the missing gas mass in the form of clumped neutral gas , which might also harbour the detected cold dust in the ejecta .
alternatively , most of the initial stellar mass might have been lost by the progenitor prior to explosion with estimated values of the swept - up mass of @xmath18m@xmath0 up to possibly 20m@xmath0 .
we have shown that the ism dust model with @xmath58 = 0.6@xmath27 most closely resembles the isrf along the sight line of casa ( based on sed fitting in the surrounding regions of casa , and a pdr analysis of submm emission lines ; see figures [ ima_casa_interstellar ] and [ ima_pdr_results ] , respectively ) .
we therefore attach more weight to the sn dust masses of 0.5m@xmath0 ( mgsio@xmath15 ) , 0.3m@xmath0 ( mg@xmath64sio@xmath65 ) and 0.5 - 0.6m@xmath0 ( carbonaceous grains ) derived from sed models with @xmath58 = 0.6@xmath27 ism dust models . due to possible variations in the isrf along the line of sight
, we can not rule out that the local radiation field near casa is closer to the solar isrf which would imply sn dust masses of 0.1 - 0.2m@xmath0 .
being conservative , we can conclude that the mass of dust produced in casa is between 0.1 and 0.6m@xmath0 and likely consists of a combination of mostly silicate - type ( mgsio@xmath15 or mg@xmath64sio@xmath65 ) grains and some form of carbonaceous dust .
based on the models constructed for each of the emission components , we can determine the relative contribution of various contaminants at every wavelength .
figure [ ima_casa_schematic ] gives a schematic overview of the relative contributions at ir and submm wavelengths .
based on our models for line emission , synchrotron radiation and ism dust emission , we can furthermore construct residual images with the remaining sn dust emission .
the images in figure [ ima_casa_comp ] ( first column ) provide an overview of the total emission in the mips24@xmath3 m , pacs70 , 100 and 160@xmath3 m , and spire250 , 350 and 500@xmath3 m wavebands .
the second column shows the same maps after subtraction of the contributions from line emission and synchrotron radiation , while the third , fourth and fifth columns show the emission intrinsic to the supernova remnant derived from modelling the residual sed with a multi - component sn+ism dust model for isrf scaling factors of 0.3@xmath27 , 0.6@xmath27 and 1.0@xmath27 , respectively . at mid - infrared wavelengths , the largest contributor to the broadband emission is from warm dust in the supernova remnant , with contributions of 92@xmath5 , 97@xmath5 , 75@xmath5 and 94@xmath5 in the irs17@xmath3 m , _ _
wise__22@xmath3 m , mips24@xmath3 m and irs32@xmath3 m wavebands , respectively . in the pacs70@xmath3 m waveband
, the emission is still dominated by warm sn dust emission ( 84@xmath5 ) , while the pacs100@xmath3 m waveband has a non - negligible contribution from ism dust emission ( 34@xmath5 ) in addition to a large sn dust contribution ( 54@xmath5 ) . with the emission of ism dust peaking just longwards of 100@xmath3 m ( see figure [ ima_sed_jones ] ) ,
the contribution from sn dust emission drops significantly in the pacs160@xmath3 m and spire wavebands . for ism dust illuminated by a radiation field with g=0.6@xmath27 ,
the emission of dust intrinsic to casa contributes 30@xmath5 , 15@xmath5 , 10@xmath5 and 4@xmath5 of the total emission in the pacs160@xmath3 m and spire250 , 350 and 500@xmath3 m maps , respectively . in the pacs100 and 160@xmath3 m maps
, we can identify an emission peak in the inner regions of the remnant in addition to the three peaks of sn dust emission in the reverse shock regions . in the spire wavebands ,
this central peak becomes more prominent compared to the reverse shock region .
this sn dust component in the inner regions was already identified by and @xcite .
the same emission pattern identified in the pacs wavebands ( i.e. , one central peak and three secondary peaks in the reverse shock ) is prominently present in the sn dust emission maps derived for an isrf field g=0.6@xmath27 .
we argue that this supports the assumption that the g=0.6@xmath27 ism model best represents the ism conditions towards casa .
performed a similar decomposition of the different emission components to the _ herschel _ pacs and spire wavebands , using an earlier version of the ` hipe ` pipeline to reduce the _
herschel _ data . although their assumptions when modelling the synchrotron , line and interstellar dust emission were differentm image to subtract the ism dust contribution at 70@xmath3 m , 100@xmath3 m and submm wavelengths . ] , the contributions of the different emission sources at various wavelengths are similar . we find a contribution from synchrotron radiation of 3.6@xmath5 , 3.5@xmath5 , 4.7@xmath5 , 7.9@xmath5 , 16.3@xmath5 , and 39.7@xmath5 in the pacs70 , 100 , 160@xmath3 m and spire250 , 350 and 500@xmath3 m bands , while reported contributions of 3.7@xmath5 , 4.2@xmath5 , 6.8@xmath5 , 9.2@xmath5 , 21.1@xmath5 , and 47.9@xmath5 in those wavebands , respectively . while our adopted spectral index ( -0.644 )
is even shallower compared to their value of -0.70 , their somewhat higher synchrotron contribution most likely arises from their extrapolation of the emission at 6 cm .
the normalisation of our synchrotron component is based on recent _ planck _ measurements , which are closer in wavelength to the _ herschel _ data points .
the ism dust contributions of 10.7@xmath5 , 39.6@xmath5 , 74.1@xmath5 , 83.9@xmath5 , 75.0@xmath5 , and 52.9@xmath5 in those wavebands estimated by are consistent within 5 - 10@xmath5 to our estimated ism contributions for a radiation field with scaling factor g=0.6@xmath27 ( 12.2@xmath5 , 34.4@xmath5 , 70.2@xmath5 , 76.4@xmath5 , 77.6@xmath5 , and 59.3@xmath5 at 70 , 100 , 160 , 250 , 350 and 500@xmath3 m , respectively ) .
the contributions estimated by for the sn dust emission at 70@xmath3 m ( 85.8@xmath5 ) , 100@xmath3 m ( 47.9@xmath5 ) and 160@xmath3 m ( 19.3@xmath5 ) , and at 250@xmath3 m ( 6.9@xmath5 ) , 350@xmath3 m ( 3.9@xmath5 ) and 500@xmath3 m ( 0@xmath5 ) are within the range of percentages derived here from our spatially resolved decomposition analysis ( 83.6 , 53.9 , 29.6 , 14.9 , 9.7 , 4.4@xmath5 , respectively ) .
our supernova dust contributions are consistently higher by 4 to 10@xmath5 in the pacs100 and 160@xmath3 m and spire250 , 350 and 500@xmath3 m wavebands , due to our lower estimates for the synchrotron and ism dust emission at submm wavelengths . by extrapolating the best fitting multi - component sed model to longer wavelengths and convolving the model with the scuba 850@xmath3 m filter response curve , we are able to produce a modelled image of the ism and sn dust emission at 850@xmath3 m . to this image , we add the modelled synchrotron radiation ( using the coefficients derived in section [ sect_synchr ] ) to simulate a predicted total 850@xmath3 m image based on the models derived in this paper .
figure [ fig_simu_850mu ] presents the modelled synchrotron radiation , ism and sn dust emission at 850@xmath3 m ( top row ) .
the bottom left panel shows the sum of all model contributions at 850@xmath3 m , and can directly be compared to the observed scuba850@xmath3 m image ( bottom right panel ) . given that our model resolution of 36@xmath17 is coarser compared to the resolution of the scuba 850@xmath3 m image ( 15@xmath17 ) , our model is not sensitive to all the structures detected in the scuba 850@xmath3 m image .
the overall emission features in our simulated scuba 850@xmath3 m map compare well to the observed scuba 850@xmath3 m map , with both showing a clear emission peak in the west and secondary peaks in the north - west and south - east .
although the modelled and observed scuba 850@xmath3 m images show a great resemblance , our global model flux at 850@xmath3 m ( measured within an aperture with radius of 165@xmath17 ) , @xmath73 = [email protected] jym emission is derived from the combined uncertainty on the synchrotron model flux and the errors on the relative contribution of ism and sn dust emission at 850@xmath3 m . ]
is lower compared to the scuba 850@xmath3 m flux measurement @xmath74 = [email protected] jy @xcite , but still consistent within the large error bars due to the uncertainty on the exact synchrotron contribution at 850@xmath3 m .
our estimated contribution of synchrotron radiation ( [email protected] jy ) is similar to the values used by @xcite ( 34.9 jy ) and @xcite ( 30.7 jy ) .
our current model predicts that the ism dust emission contributes about [email protected] , with only [email protected] originating from sn dust .
the latter ism and sn dust contributions are incompatible with the lower limit of 6.0jy for the sn dust emission from casa ( and thus upper limit for the 850@xmath3 m ism dust flux of 14.1jy ) derived by @xcite based on submm polarimetry .
the high level of polarisation ( 30@xmath5 ) of sn dust found by @xcite seems hard to explain in the framework of our model with a sn dust flux of only 0.4jy . if the polarised emission would instead come from ism dust , it surpasses the average intrinsic dust polarisation of 12@xmath5 measured by @xcite .
the alignment of dust grains in the ism surrounding casa might be stronger than usually found in the diffuse ism , which might provide a possible explanation for the high degree of polarisation .
the extrapolation of our model is consistent with the conclusion from @xcite that most of the emission at 850@xmath3 m arises from ism dust emission after subtraction of the synchrotron component . our current model for synchrotron+ism+sn dust emission at 850@xmath3 m
does show a 9.5jy deficit compared to the scuba 850@xmath3 m flux ( [email protected] , @xcite ) and _ planck _
850@xmath3 m flux ( 52@xmath447jy , ) .
given that we already derive a large sn dust mass ( 0.4 - 0.7m@xmath0 ) making a 0.4jy flux contribution at 850@xmath3 m , it seems highly unlikely that the model sn dust flux could be increased without invoking unrealistically high sn dust masses .
we furthermore reject a scenario in which the offset at 850@xmath3 m is caused by model deviations from the galactic ism dust emission , given that the small relative offsets between the best fitting ism dust models and the observed pacs 100 and 160@xmath3 m and spire 250 , 350 and 500@xmath3 m data points are within the observational uncertainties .
we thus argue that the missing emission at 850@xmath3 m most likely arises from synchrotron radiation . using a power law index of -0.54 determined based only on _ planck _ data ( see section [ sect_synchr ] ) would be sufficient to increase the synchrotron radiation at 850@xmath3 m to 37.6jy .
the updated model for a mgsio@xmath15 warm+cold sn dust composition and ism model with @xmath58=0.6@xmath27 predicts 850@xmath3 m contributions of 7.9jy and 0.5jy from ism and sn dust emission , which would result in a total modelled flux at 850@xmath3 m of 46jy , consistent with the observed scuba850@xmath3 m ( [email protected] ) and _ planck _
850@xmath3 m ( 52@xmath447jy ) fluxes .
a small change in spectral index thus enables one to account for the 850@xmath3 m deficit without affecting the main results of the paper .
based on the maps of the total sn dust mass and ism dust mass obtained from the spatially resolved sed fitting in section [ dustsed ] , we can infer the spatial variation in optical extinction towards casa . for a foreground dust screen model , the @xmath75 band attenuation
is calculated based on the average extinction to gas column - density ratio in the solar neighbourhood , @xmath7/@xmath76 = 6.53@xmath1210@xmath77 mag cm@xmath50 , and a dust - to - hydrogen ratio @xmath79/(@xmath76@xmath80 ) = 0.0056 ( derived from the ism dust model ) , which results in @xmath81 figure [ ima_casa_av ] presents the @xmath7 maps for interstellar and sn dust extinction .
the position of the point - like x - ray source identified by @xcite is indicated with a black cross in the right panel .
the interstellar visual extinction varies from @xmath7 = 3mag in the north - west of casa , where the extinction is lowest , to peaks of @xmath7 = 15mag in the south - east and west , with an average @xmath7 between 6 and 8mag .
these values are consistent with the visual extinctions derived by @xcite ( @xmath7 @xmath1 4 - 6mag ) and @xcite ( @xmath7 up to 8mag ) based on optical and near - infrared spectra , and h@xmath14co absorption , respectively .
while the h@xmath14co - derived extinction maps have a resolution of 6@xmath17 ( i.e. , six times better compared to the spire500@xmath3 m resolution with fwhm @xmath1 36@xmath17 ) , our interstellar extinction map shows the same peaks in @xmath7 towards the west , south - east and near the centre of casa .
we are therefore confident that the ism dust emission was modelled accurately , and the remaining dust emission can be attributed to sn dust emission .
based on our @xmath7 maps for the ism and sn dust , we can estimate the probability to detect a possible binary companion which was argued to be responsible for the stripping of hydrogen gas from the progenitor of casa during the evolved stellar phase @xcite .
the high levels of visual extinction ( @xmath7 @xmath1 8 - 10mag ) towards the central regions make it very difficult to detect the supernova remnant or the presence of a binary star at ultraviolet and optical wavelengths but near - ir imaging could be more effective .
figure [ ima_casa_av_all ] presents a contour map of the visual extinction derived for the entire field of casa observed with _ herschel _
, where the results of the multi - component sed fitting along the line of sight of casa ( see section [ sect_isdust ] ) have been combined with the sed modelling of the ism emission in the field of casa ( see appendix [ sect_isdust ] ) .
the regions without any @xmath7 contours correspond to regions with insufficient signal - to - noise detections in the _ herschel _ images , which prevented us from modelling the ism dust emission spectrum . across the mapped area of 0.5@xmath22 , we identify regions with visual extinctions ranging from @xmath7 = 2 to 20mag ( the lower values are limited by the sensitivity of our _ herschel _ observations ) .
the latter @xmath7 variations correspond to column density variations up to a factor of 18 , showing that the amount of material that is positioned along different lines of sight can vary significantly on scales of a few pc .
we observe a peak in @xmath7 across the perseus arm which is positioned just south - east of casa . , due to ism dust for the entire field observed with _
pacs and spire instruments , as derived from the sed modelling of the ism dust emission presented in section [ sect_isdust ] and the multi - component sed modelling presented in section [ dustsed ] .
the contour levels start at @xmath7=2.5 increasing in steps of 2.5 up to @xmath7=20.,width=321 ] | theoretical models predict that core - collapse supernovae ( ccsne ) can be efficient dust producers ( 0.1 - 1.0 m@xmath0 ) , potentially accounting for most of the dust production in the early universe .
observational evidence for this dust production efficiency is however currently limited to only a few ccsn remnants ( e.g. , sn 1987a , crab nebula ) . in this paper , we revisit the dust mass produced in cassiopeiaa ( casa ) , a @xmath1330-year old o - rich galactic supernova remnant ( snr ) embedded in a dense interstellar foreground and background .
we present the first spatially resolved analysis of casa based on _ spitzer _ and _ herschel _ infrared and submillimetre data at a common resolution of @xmath10.6@xmath2 for this 5@xmath2 diameter remnant following a careful removal of contaminating line emission and synchrotron radiation .
we fit the dust continuum from 17 to 500@xmath3 m with a four - component interstellar medium ( ism ) and supernova ( sn ) dust model .
we find a concentration of cold dust in the unshocked ejecta of casa and derive a mass of 0.3 - 0.5m@xmath0 of silicate grains freshly produced in the snr , with a lower limit of @xmath40.1 - 0.2m@xmath0 . for a mixture of 50@xmath5 of silicate - type grains and 50@xmath5 of carbonaceous grains , we derive a total sn dust mass between 0.4m@xmath0 and 0.6m@xmath0 .
these dust masses estimates are higher than from most previous studies of casa and support the scenario of supernova dominated dust production at high redshifts .
we furthermore derive an interstellar extinction map for the field around casa which towards casa gives average values of a@xmath6=6 - 8mag , up to a maximum of @xmath7 = 15mag .
[ firstpage ] ism : supernova remnants supernovae : individual : cassiopeiaa ism : dust infrared : ism |
the last decade has seen a revolution in cosmology with the emergence of a standard model for the universe dominated by a strange substance known as dark energy " , with an effective negative pressure . since the first evidence of dark energy in 1998 @xcite , there has been substantial observational and theoretical research aimed at understanding the true nature of this phenomenon . in recent years , many authors have started exploring the possibility that dark energy , and the observed acceleration of the expansion of the universe , could be the consequence of an incomplete theory of gravity on cosmological scales and may require modifications to einstein s theory of general relativity .
one of the most direct methods of testing our assumed theory of gravity is to perform consistency checks between the geometrical expansion history of the universe , as measured by cosmological probes like type ia supernovae and the baryon acoustic oscillations , and the evolution of density inhomogeneities in the universe @xcite .
the growth of structures in the universe can be measured using a variety of techniques , but a popular and direct method involves measuring the coherent peculiar velocities of galaxies on large scales ( i.e. , the motion of galaxies after the cosmological expansion has been removed ) caused by their infall into large scale overdensities like clusters and superclusters of galaxies @xcite . traditionally , the local peculiar velocity field , or `` bulk flow '' of galaxies , has been estimated using samples of galaxies where the peculiar velocity of each galaxy has been determined using secondary distance indicators ( e.g. the tully fisher relationship @xcite or fundamental plane @xcite ) .
alternatively , the coherent motion , or flow , of galaxies can be statistically estimated from their effect on the clustering measurements of large redshift surveys , or through the measurement of redshift space distortions .
historically , measurements of the local galaxy peculiar velocity field have been plagued by systematic uncertainties @xcite , and small sample sizes , which have made their cosmological constraints uncompetitive compared to other probes of the universe .
however , a recent re analysis of these earlier peculiar velocity galaxy surveys @xcite has now provided a consistent observational picture from these data , and finds significant evidence for a larger than expected bulk flow in the local universe ( within 100 @xmath10 ) .
this new analysis is also consistent with new measurements of the local bulk flow using the kinetic sunyaev - zeldovich effect @xcite of massive clusters of galaxies out to at least 800 mpc @xcite , which leads to the intriguing situation that these local measurements appear to be significantly greater in amplitude , and on larger scales , than expected in a concordance , wmap7normalised @xmath4cdm cosmological model .
such discrepancies with @xmath4cdm may give support to theories of modified gravity @xcite .
given the importance of these local large scale bulk flow measurements , we investigated in a previous paper @xcite the likelihood of such large coherent flows using an alternative methodology based upon measurements of the redshift space distortions seen in the clustering of a sample of galaxy clusters selected from the sloan digital sky survey ( sdss ) @xcite .
we detected no statistical evidence for the coherent flows of the same size , and scale , as discussed above , but due to the sample size the statistical errors were large ( see @xcite for details ) .
therefore , in this paper , we re - visit our earlier work using a larger cosmological data set and recent theoretical improvements in the modelling of redshift space distortions .
the goal of this paper is to statistically study the likelihood of the local bulk flow measurements within a large cosmological volume of the universe , well beyond these nearby measurements . in section
ii of this paper , we outline in greater detail the fitting techniques used in @xcite to formulate the line of sight smearing effect due to the coherent bulk flow of galaxies . in section iii
, we explain our data analysis of a sample of sdss luminous red galaxies ( lrgs ; @xcite ) and our measurement of their two dimensional redshift space correlation function . in section iv
, we present our results from comparing our lrg measurements to our model for redshift space distortions , and compare our results to cosmological predictions and previous observations .
we conclude in section v.
the aim of this work is to statistically measure the coherent motion of galaxies on large scales . that is , the linear velocity of galaxies excluding both the cosmological hubble expansion and any smaller scale non linear components .
the coherent motions of galaxies introduce redshift space distortions an anisotropic feature into the measured clustering statistics . as originally proposed in @xcite
, a distant observer should expect a multiplicative enhancement of the overdensity field along the line of sight , compared to the transverse direction , due to such coherent peculiar motion of galaxies .
this kaiser effect " , as it is now know , can be seen as a squashing " or flattening of the two dimensional two point correlation function ( @xmath11 ) , where the correlation function is decomposed into two vectors ; one parallel to the line of sight ( @xmath12 ) and the other perpendicular to the line of sight ( @xmath13 ) .
information about the coherent velocities of galaxies can then be extracted from the two dimensional correlation function via careful theoretical modelling of these redshift space distortion effects . in this paper , we utilise several theoretical improvements to the original redshift space distortion work given in @xcite . throughout this paper
, we refer to the original work of @xcite as the `` kaiser limit '' , which formulated the coherent motions of galaxies as an additional dopper shift to the cosmological redshift .
the kaiser limit " is suitable to test theoretical models in the linear region of galaxy perturbations ( on large scales ) , but over the last decade there have been significant improvements to the modelling of these effects .
below , we repeat some of the formalism presented in our first paper @xcite to aid the reader and provide continuity with additional work presented in following subsections .
the observed power spectrum of density fluctuations in redshift space , @xmath14 , can be written as , [ eq : ps ] _
ob(k,)&= & + & & g(k,;_v ) , where @xmath15 and @xmath16 denote galaxy density and velocity fields respectively ( where @xmath17 and @xmath18 is the divergence of the velocity fields ) . in eq .
[ eq : ps ] , the function @xmath19 denotes the additional suppression effect due to coherent motions which affects the galaxy galaxy , velocity velocity and galaxy velocity power spectra equally , and is a function of separation ( @xmath20 ) , angle between the galaxies ( denoted by @xmath21 ) and has a given smoothing scale , or velocity dispersion ( @xmath6 ) , which is the parameter at the heart of our analysis and this paper .
as discussed in our previous paper , we can decompose these power spectra into a scale dependent ( @xmath22 ) and scale independent part ( @xmath23 ) given by , [ eq : pall ] p_(k , a)&=&d_(k)g_^2(a ) , + p__g_g(k , a)&=&d_m(k)g_b^2(a ) , + p_(k , a)&=&d_m(k)g_^2(a ) , where the subscript @xmath24 denotes the curvature perturbation in the newtonian gauge , @xmath25 the growth function @xmath26 is defined as @xmath27 where @xmath28 is the standard linear bias parameter between the density of galaxies and the underlying dark matter , @xmath29 . as we follow the positive sign conversion , @xmath30 is the growth function of @xmath31 .
the shape factor of the perturbed metric power spectra @xmath32 is defined as d_(k)=^2__0(k)t^2_(k ) , which is a dimensionless metric power spectra at @xmath33 ( the matter radiation equilibrium epoch ) , and @xmath34 is the initial fluctuations in the comoving gauge and @xmath35 is the transfer function normalised at @xmath36 . the primordial shape @xmath34 depends on @xmath37 ( the slope of the primordial power spectrum ) , as @xmath38 , where @xmath39 is the amplitude of the initial comoving fluctuations at the pivot scale , @xmath40 @xmath41 .
the intermediate shape factor @xmath35 depends on @xmath42 ( @xmath43 ) .
the shape factor for the matter fluctuations , @xmath44 , which is important for both the galaxy galaxy and velocity velocity power spectra in eq .
[ eq : pall ] above , is given by the conversion from @xmath32 of , d_m(k)d_(k ) , where , assuming @xmath45 , we can write @xmath46 . in the kaiser limit
, we would thus assume @xmath47 , where @xmath48 is the linear power spectrum , and no smearing of the large scale power spectra due to the random motions of galaxies within individual dark matter halos , or the `` finger of god '' ( fog ) effect seen on small scales in redshift surveys .
therefore , the observed compression of @xmath11 along the line of sight due to coherent infall of galaxies around large
scale structures can be written in configuration space as , [ eq : xis ] _
s(,)(a)&= & ( g_b^2+g_bg_+g_^2)_0(r ) p_0 ( ) + & -&(g_bg_+g_^2)_2(r ) p_2 ( ) + & + & g_^2_4(r)p_4 ( ) , where @xmath49 is the legendre polynomial and the spherical harmonic moment @xmath50 is given by , _
l(r)=d_m(k)j_l(kr ) , where @xmath51 is a spherical bessel function . the coherent evolution parts in eq .
[ eq : pall ] are not generally parameterised by known standard cosmological parameters .
we thus normalise these growth factors at @xmath33 such that , g_(a_eq)&=&1 , + g__m(a_eq)&=&a_eqg_(a_eq ) , + g_(a_eq)&=&. then , we treat @xmath26 and @xmath30 as free parameters to be measured by redshift space distortions with cmb priors used to determine the shape factor @xmath44 at the last scattering surface .
the shape factor @xmath44 depends mainly on the following set of cosmological parameters , @xmath52 and @xmath53 .
while the first two parameters ( @xmath42 and @xmath37 ) are well constrained by the cmb at the last scattering surface ( regardless of theoretical model used to describe the late time acceleration of the expansion of the universe ) , the last two parameters ( @xmath54 and @xmath53 ) can not be determined by the cmb alone .
the primordial amplitude @xmath54 has been measured to an accuracy of less than @xmath55 from wmap7 alone , however this constraint becomes weaker due to the unknown reionization history of the universe ( tighter constraints are achieved by assuming a stepwise parameterisation of the reionization history ) . for the matter content of the universe , the cmb measures @xmath42 , not @xmath53 , which depends on the hubble constant .
therefore , given our cmb priors , @xmath44 can be expressed as , d_m(k ) = ( ) ^2 d^*_m(k : a_s^*,_m^ * ) , where @xmath56 and @xmath57 are specific reference values .
we use the best fit cosmological parameters of the wmap7normalised @xmath4cdm model in this paper , @xmath58 and @xmath59 .
then the power spectra can be re written as , p__g_g(k , a)&=&(g_b)^2 d^*_m(k : a_s^*,_m^ * ) , + p_(k , a)&=&(g_)^2 d^*_m(k : a_s^*,_m^ * ) . here
, we define new growth functions as , g_b^*(a)&&g_b(a ) + g_^*(a)&&g_(a ) .
then the power spectra become p__g_g(k , a)&=&g_b^*2 d^*_m(k : a_s^*,_m^ * ) , + p_(k , a)&=&g_^*2 d^*_m(k : a_s^*,_m^ * ) , in which the power spectra are split into two parts , i.e. , @xmath60 and @xmath61 are determined through our redshift
space distortion measurements , and @xmath62 is given by the cmb priors and the reference values of @xmath56 and @xmath57 .
thus , we re - write the correlation function in eq .
[ eq : xis ] in terms of our new fitting parameters @xmath60 and @xmath61 , [ eq : xis * ] _ s(,)(a)&= & ( g_b^*2+g_b^*g_^*+g_^*2)_0^*(r ) p_0 ( ) + & -&(g_b^*g_^*+g_^*2)_2^*(r ) p_2 ( ) + & + & g_^*2_4^*(r)p_4 ( ) , where the spherical harmonic moment @xmath63 is given by , [ eq : jl ] _ l^*(r)=d_m^*(k : a_s^*,_m^*)j_l(kr ) .
we measure @xmath61 with a normalised @xmath64 , not @xmath30 paired with @xmath44 .
however , as @xmath65 , @xmath66 does not depend on any normalisation , i.e. , @xmath66 is invariant under the transformation between the @xmath67 ( asterisk ) variables and those without above .
thus , we present measured coherent motions not in terms of @xmath61 , but in terms of velocity dispersion derived from @xmath66 .
hereafter , we drop _ @xmath68 _ symbol , i.e. _ ` unasteriked ' _ quantities imply _ ` asteriked ' _ quantities from now . it has been pointed out in @xcite that the description of @xmath11 on large scales ( in linear theory ) will be modified due to dispersion effects in the @xmath12 direction .
this effect was recently investigated using @xmath69-body simulations and was found to be important to include @xcite .
therefore , in this paper , we have modified the kaiser limit using a gaussian velocity dispersion term for @xmath19 in eq .
1 , which we will refer to as the modified kaiser effect " throughout .
this modified model has been shown to work reasonable well on mock catalogues of a @xmath4cdm universe ( see @xcite ) . in detail , we can write @xmath19 as [ eq : vdispersion ] g(k,,_v)=e^-k^2 ^ 2_v^2 , where @xmath6 is the 1d velocity dispersion on large scales , and is the parameter of interest in this paper .
we can write @xmath6 as , _
v^2=p^lin_(k , z ) dk , [ eq : sigma_v ] where we see that @xmath6 has units of @xmath70 as @xmath71 and @xmath20 have units of @xmath72 and @xmath70 respectively .
this is somewhat confusing given @xmath6 is called a velocity dispersion " in previous literature , but we retain this terminology here to remain consistent with this literature .
however , we note that we can convert the units of @xmath6 to km / s using @xmath73 , but this requires a precise measurement of the expansion history of the universe ( at the redshifts of interest ) , in addition to the determined @xmath71 from our method .
we therefore modify the kaiser effect using @xmath19 defined in eq .
[ eq : vdispersion ] , but keeping @xmath74 .
then the integration of @xmath75 in eq .
[ eq : xis ] should be re expressed , as @xmath19 changes with varying @xmath76 altering @xmath6 in eq .
[ eq : sigma_v ] to , [ eq : modxis ] _ l^*(r)= d_m^*(k)e^-(k_k_v)^2_l(_k ) .
here @xmath77 is the cosine of angle between @xmath78 and the pairwise orientation . for illustrative purposes , in the left hand panel of fig .
[ fig : sigma_pi ] , we present our modelling of @xmath11 using the standard kaiser limit ( unfilled black contours ) compared to the modified kaiser effect ( filled blue contours ) discussed above . in these examples , we have used a galaxy bias of @xmath79 and @xmath80 at @xmath2 ( upper panel ) and 0.38 ( lower panel ) respectively ( these values of @xmath28 were chosen to correspond to our estimated bias values from table i to illustrate our point ) .
the additional suppression in the modified kaiser effect can be clearly seen on large scales ( in the linear regime ) at approximately the 10 to 20@xmath81 level .
the normalized distribution of 105,831 sdss lrg galaxies over plotted with the random catalogue as a smooth red line ( 1,664,948 points ) .
the lrgs are split into two redshift bins at the median redshift of the whole distribution ( @xmath82 ) as shown by the blue , dashed line .
the low redshift bin spans the range @xmath83 , while the higher redshift bin covers @xmath84 .
, scaledwidth=47.0% ] in this paper , we have analysed lrgs taken from data release seven ( dr7 ) of the sdss @xcite .
in particular , our lrg sample contains approximately 100,000 galaxies in the redshift range @xmath85 with magnitudes @xmath86 the 3-dimensional size of the survey is best described by the effective volume @xcite @xmath87 where @xmath88 is the comoving number density at every point @xmath89 and @xmath90 is the power amplitude at wavelength @xmath20 .
the sdss dr7 lrg sample used in this analysis has @xmath91 .
the lrgs are more biased than main sample galaxies and thus have a higher clustering amplitude which boosts the signal - to - noise on their measurements .
it is this large volume , and the fact that lrgs are highly biased tracers of the underlying mass distribution , which makes them an excellent population for studying the statistics of large scale structure . in detail
, we use the publicly available lrg data from @xcite , that include random catalogues and galaxy weights that are essential for accommodating the angular mask in such large scale galaxy surveys .
we have converted the observed coordinates of lrgs ( ra , dec , redshift ) into comoving cartesian distance coordinates ( @xmath92 , @xmath93 , @xmath94 ) using the wmap7 mean cosmological parameters based upon a @xmath4cdm model .
therefore , @xmath95 in eq .
[ eq : modxis ] can be re expressed using a mapping from the fiducial wmap7 normalised @xmath4cdm model of @xmath96 ( where @xmath97 ) .
we note that by assuming a cosmological model for the conversion of redshifts and angles to distances could introduce an additional anisotropic signal in the correlation function i.e. the `` alcock paczynski '' effect @xcite .
the size of this effect is smaller than the kaiser effect on the scales of interest herein , e.g. , if we vary our background cosmology , within reasonable limits allowed by the wmap analysis , we see no significant difference in our correlation functions . however , we note that a more self consistent approach will be required in the future ( see @xcite ) . in fig .
[ fig : zdist ] , we show the co - moving number density of lrgs in our whole sample ( compared to the random catalogue ) and we split this sample into two redshift bins as indicated ( bin 1 : @xmath98 and bin 2 : @xmath99 ) .
we do not split the sample further as we wish to keep the thickness of the redshift shells ( @xmath100 ) larger than the scales of interest in our analysis , thus preventing any aliasing in the redshift direction .
we estimate the correlation function using the landy - szalay " estimator @xcite , @xmath101 where @xmath102 is the number of galaxy galaxy pairs , @xmath103 the number of galaxy - random pairs , and @xmath104 is the number of random
random pairs , all separated by a distance @xmath105 and @xmath106 .
all pairs are weighted using the minimum variance weighting scheme of @xcite .
each galaxy is assigned a weight according to , @xmath107 where @xmath108 is the comoving space density at redshift @xmath109 and @xmath110 , as in @xcite .
the pairs are also corrected for fiber collisions , where two or more galaxies separated by less than 55 arcseconds can not be observed simultaneously .
this is achieved using a secondary galaxy weight , proportional to the number of unobserved galaxies within the fiber radius .
the pair counting was achieved using the parallel kd - tree code , ` ntropy ` @xcite . in the right hand panel of fig .
[ fig : sigma_pi ] , we present our measured @xmath11 . the perpendicular component , @xmath13 , is binned logarithmically into ten bins from @xmath111 to @xmath112 , and the parallel component , @xmath12 , is also binned logarithmically into ten bins , but from @xmath111 to @xmath113 .
we use the jack knife method @xcite to estimate statistical errors on @xmath11 , which involved dividing the survey into @xmath69 sub - sections with approximate equal area ( and thus volume ) and then computing the mean and variance of @xmath11 from these @xmath69 measurements of the correlation function with the @xmath114 region removed each time ( where @xmath115 ) . in our analysis , we use @xmath116 jack
knifed samples which were created using an adaption of the ` healpix ` , equal
area , pixelisation code@xcite .
knife measurements provide an estimate of the covariance matrix for all bins in our correlation function .
however , this matrix can be noisy , given the number of jack
knife samples used , although previous work @xcite has shown that the number of jack
knife samples we use should provide a stable estimate of the covariance matrix . in order to reduce the statistical noise in our covariance matrix further , we follow the same procedure as described in @xcite and `` clean '' our covariance matrix before inversion @xcite . in figure
[ fig : eigen ] , we show the distribution of eigenvalues for the decomposition of our two covariance matrices and truncate them when the eigenvalues decrease below @xmath117 , which avoids any singularity in the inversion . this threshold in @xmath118 was empirically determined and , reassuringly , gives reduced @xmath119 values close to unity ( see below ) .
= 3.2truein = 3.2truein the shape of the power spectra of density fluctuations is determined before the epoch of matter radiation equality . under the paradigm of inflationary theory ,
initial fluctuations are stretched outside the horizon at different epochs which generates the tilt in the power spectrum .
the predicted initial tilting is parameterised as a spectral index ( @xmath37 ) which is just the shape dependence due to the initial conditions . when the initial fluctuations reach the coherent evolution epoch after matter - radiation equality ,
they experience a scale - dependent shift from the moment they re - enter the horizon to the equality epoch .
gravitational instability is governed by the interplay between radiative pressure resistance and gravitational infall .
the different duration of modes during this period result in a secondary shape dependence on the power spectrum .
this shape dependence is determined by the ratio between matter and radiation energy densities and sets the location of the matter - radiation equality in the time coordinate .
therefore , the shape
dependent part of @xmath95 is well determined , in a model independent way , when both @xmath37 and @xmath42 are determined at the last scattering surface .
we thus marginalise our constraints over the growth parameters with ( @xmath37,@xmath42 ) using wmap7 priors on @xmath37 and @xmath42 , @xmath120 and @xmath121 . on small scales , there are non linear effects that are not easily modelled via analytical methods , e.g. , the internal velocities of galaxies within clusters which cause the famous fingers of god " effect seen in galaxy redshift surveys @xcite .
we investigate here the effect of such non - linearities and outline our strategy for mitigating them in our analysis .
first , the contribution of velocity velocity correlation is decomposed from the observed correlation function using the anisotropic feature of redshift distorted galaxy maps of which information is most extractable at the quadrupole moment . since the leading order of the quadrupole moment is the density velocity correlation , what we measure is the cross - correlation between the density and velocity fields not the velocity velocity component .
the velocity velocity correlation is the leading order at the hexadecapole moment , which is too weak to provide tight constraints on the coherent velocity flows . however , both fields ( density and velocity ) are perfectly correlated in the linear regime at least at scales @xmath122 .
deviations from a perfect cross
correlation coefficient @xmath123 ( @xmath124 ) are not caused by cosmological effects but rather by non linear physics . if we analyse our data only in the linear regime ( where @xmath123 ) , then @xmath26 and @xmath30 can be simultaneously determined from just the density density and density velocity components .
therefore , we impose a cut off scale on @xmath125 at the scale where @xmath126 does not significantly depart from unity .
this corresponds to @xmath122 , or @xmath127 , as seen in n
body simulations @xcite .
we use @xmath128 , instead of @xmath129 , given by the peak contribution of convolution in eq .
[ eq : jl ] due to the bessel function in the integrand .
secondly , we also discard more bins along the @xmath12 direction as they are more contaminated by the finger of god " effect . in our model
, we use a linear power spectrum instead of a non - linear power spectrum as in @xcite .
this should be fine for the accuracy required in this paper as the most important correction is due to the dispersion effect , which we have included here , and also the effect of non - linearities in the power spectra is only a few percent at @xmath130 @xcite . to estimate the impact of the non
linearity on the two point correction function in redshift space , we use the halofit model @xcite for the non - linear power spectrum with a cdm transfer function @xcite .
we apply the same non linear power spectrum for the velocity power spectrum , although this obviously overestimates the velocity power ( see @xcite for a non linear fitting function of the velocity power spectrum ) .
however , we use this non linear model just for discarding small scale data which may be significantly affected by non linearities , and our rejection will be conservative .
we find that the effect of non - linearity is over 20@xmath81 at the scale of @xmath131 in the redshift range of our data , and discard these measurements from our analysis .
shown in the right panel of fig .
[ fig : sigma_pi ] , measured @xmath11 at bins of @xmath132 is consistent with theoretical predictions ( compared with the left panel of fig .
[ fig : sigma_pi ] ) , but @xmath11 at bins of @xmath133 does not match well with the theoretical predictions from the modified kaiser effect .
in this paper , we thus impose a cut off of @xmath132 . in summary ,
we impose two cut - offs in scale to control the contamination of small - scale non - linearities on our larger scale correlation function measurements .
we achieve this by not including bins with scales @xmath134 and @xmath133 in our fitting procedure . [ cols="^,^,^,^,^,^,^ " , ]
using the data and method outlined in sections ii and iii , we measure @xmath26 and @xmath30 simultaneously from the sdss dr7 lrg correlation functions assuming the appropriate wmap7 priors as discussed above . as illustrated in fig .
[ fig : sigma_pi ] , there is excellent visual consistency on large scales between our predicted @xmath11 and the observed functions . in detail
, we obtain a reduced @xmath119 of 0.89 and 0.83 for the @xmath2 and 0.38 samples for 12 and 9 degrees of freedom .
we have 16 and 13 eigenmodes after truncation , minus four fitting parameter respectively . in fig .
[ fig : contour ] , we present constraints on @xmath26 and @xmath6 ( @xmath6 is converted from measured @xmath30 using eq .
[ eq : sigma_v ] ) . in the kaiser limit ,
variations of the coherent growth function of galaxy density fields amplifies mainly the monopole moment of @xmath11 , and variations of the coherent motions affects mostly the anisotropy of @xmath11 .
measured @xmath26 and @xmath30 using the kaiser limit are presented as unfilled black contours of fig .
[ fig : contour ] .
the velocity dispersion in the correlation function @xmath11 of the modified kaiser effect ( eq . [ eq : modxis ] ) induces an additional suppression to the @xmath11 of the conventional kaiser limit ( eq . [ eq : xis ] ) .
the variation of coherent motion leads to monopole suppression as well as anisotropic amplification .
as coherent motions increase , the corresponding @xmath26 becomes smaller due to the increasing suppression by the velocity dispersion effect .
shown as filled blue contours of fig .
[ fig : contour ] , the best fit value of @xmath26 is shifted to lower values and the contours are elongated to smaller @xmath26 at higher @xmath6 .
while the measured difference of @xmath6 between the kaiser limit and modified kaiser effect is less than @xmath135 , the monopole shift of @xmath26 due to velocity dispersion with the modified kaiser effect causes measured differences of @xmath26 from @xmath136 to @xmath137 presented in table [ tab : vp ] .
we present herein a measurement of the coherent motions using @xmath6 , converted from our measured value of @xmath30 ( shown in table [ tab : vp ] ) and shape factor @xmath138 given by our wmap7 priors . at @xmath2 , @xmath6
is measured to be @xmath139 ( one sigma errors ) , and at @xmath140 , we find @xmath141 . these measurements are less dependent on the cosmological model for the late time acceleration of the expansion of the universe ( e.g. dark energy ) as our cmb priors are fixed at a much earlier epoch in the universe , e.g. we use value of @xmath56 and @xmath57 constrained at the surface of last scattering .
our measurements are also independent of galaxy bias which is a major advantage compared to other methods to parameterise the motions of galaxies from redshift space distortions .
it is true that our measured @xmath6 correlates with @xmath26 , but we do not need to know how to separate @xmath28 and @xmath142 for determining the coherent motions .
if we assume @xmath143 is well measured , then we can convert the units of @xmath6 using @xmath144 ( assuming @xmath4cdm ) to give @xmath7 km / s and @xmath8 km / s at the two redshifts listed in table [ tab : vp ] . although we would not recommend using @xmath144 for cosmological constraints ( without a reliable @xmath145 measurement to an accuracy of a few percent ) , this conversion is more intuitive when discussing coherent motions and allows us to compare our values with other peculiar velocity measurements , e.g. , the bulk flow " measurements discussed in section i. traditionally , coherent motions are estimated using the @xmath146 parameter @xcite , _
ob(k,)=(1+^2)p__g_g(k ) .
this parameter is equivalent to @xmath147 in our notation , and measured to be @xmath148 and @xmath149 at @xmath2 and 0.38 respectively . as we have measured , @xmath26 is significantly sensitive to the different redshift space distortion models , and @xmath146 is found to be @xmath150 and @xmath151 for the kaiser limit .
thus , when measuring @xmath146 from the sdss dr7 lrg samples for use as cosmological constraints , it should be stated clearly which model for redshift - space distortions is assumed .
if we assume a @xmath4cdm model , then the galaxy bias can be estimated from our measured @xmath26 values .
we present values for @xmath28 in table [ tab : vp ] , which are @xmath152 and @xmath153 at @xmath2 and 0.38 respectively .
it is interesting to compare our measurements with @xmath154 as measured by @xcite using the higher order correlation function of lrgs from the sdss .
this consistency supports our modified kaiser formulation , as the estimated @xmath155 using the original kaiser limit is different by more than one sigma . in fig .
[ fig : vphistory ] , we present our measurements of @xmath6 , the coherent motions of galaxies in redshift space , versus redshift and compared to theoretical predictions . as can be seen , our measurements are fully consistent with the wmap7normalised @xmath4cdm model .
this represents a clean " test of these cosmological models free from contamination by non
linear physics on small scales and uncertainties of the galaxy bias determination . these data probe the growth history of fluctuations in the universe and are thus complementary to the geometrical probes of the universe . in fig [ fig : vphistory ] , we also present predicted curves for the coherent motion of galaxies for a variety of dark energy models with varying constant equation of state from @xmath156 to @xmath157 ( dotted curves ) .
all other cosmological parameters were fixed to be the same as the wmap7 best fit @xmath4cdm values , except @xmath158 . using these curves
, we can approximately estimate that we have a constraint of @xmath159 , around a mean value of @xmath160 , from our @xmath6 measurements .
this is consistent and complementary to similar constraints from geometrical observations of the universe @xcite .
finally , we also provide in fig [ fig : vphistory ] a prediction for the dvali - gabadadze - porrati ( dgp ) self - accelerating braneworld scenario @xcite .
as can be seen , this cosmological model provides a poor description of our observations ( excluded at confidence of @xmath5 for both redshift bins ) , which support other observational constraints on this model from the cosmic microwave background ( cmb ) anisotropy , supernovae and hubble constant data @xcite .
= 3.2truein = 3.2truein
we present in this paper new measurements of the coherent motions of galaxies on large - scales derived from a new methodology for analysing the redshift space distortions seen in the observed two dimensional two point correlation function of luminous red galaxies from the sdss dr7 sample .
our new methodology is based on measuring the scale independent growth functions , @xmath26 ( galaxy density ) and @xmath30 ( velocity density ) , which do not depend upon the physics of the late universe ( e.g. dark energy ) .
they do depend on our knowledge of the early universe , but this can be constrained using information from the cosmic microwave background .
we have determined values of @xmath30 from the redshift
space distortions seen in the sdss dr7 lrg data and , converting these values into the 1d velocity dispersion @xmath6 , we find @xmath161 at a mean redshift of @xmath2 and @xmath141 at @xmath140
. these values for @xmath6 are fully consistent with a wmap7normalized @xmath4cdm model with @xmath162 as illustrated in fig [ fig : vphistory ] .
our observations are however , inconsistent with a dgp model for the universe to high statistical significance ( @xmath163 ) .
our results provide a competitive , and complementary , constraint on these cosmological models compared to the usual geometric probes of the universe . we have converted our measured values of @xmath6 into velocity units ( as opposed to lengths presented above ) and find @xmath7 km / s and @xmath8 km / s at a mean redshift of @xmath2 and @xmath3 respectively , assuming a @xmath4cdm universe . as expected , these coherent motions ( or velocity dispersions ) are fully consistent with expectations from a @xmath4cdm universe .
these estimates are however , inconsistent with local measurements of the peculiar velocity field ( or bulk flows " ) which have recently been measured to be greater than these velocities and expectations from @xmath4cdm @xcite .
if the amplitude of these local , observed bulk flows were converted ( using @xmath144 and assuming a flat @xmath4cdm model ) to the redshift range studied here ( @xmath164 ) , then we might expect to see larger coherent motions .
it is difficult to perform a direct comparison of these different velocity measurements because of the different methods and redshift intervals used .
for example , the lower redshift measurements of @xcite or @xcite only probe the velocity field out to 150@xmath165mpc using their composite sample , while most of the data is within a sphere of radius of @xmath166mpc depending on the weighting scheme used ( see fig 2 of @xcite ) .
in contrast , our statistical estimate of coherent motions on the scales up to 60@xmath165mpc are derived from within a large volume of the universe at higher redshift , e.g. , 0.5@xmath167gpc@xmath168 ( @xmath83 ) and 1.1@xmath167gpc@xmath168 ( @xmath84 ) respectively , assuming a flat wmap7 cosmology .
therefore , our measurements have averaged over many hundreds of subregions of approximately the same size as the volume used by @xcite ( assuming a subregion of radius of 60@xmath165mpc ) .
furthermore , there are potential differences in the possible directions of the velocity measurements being compared .
the direction of the local velocity measurements of @xcite point towards @xmath169 degrees of right ascension ( ra ) and declination respectively on the sky , while our sdss dr7 data is centred at @xmath170 ( equatorial coordinates ) and concentrated in the northern hemisphere , i.e. our patch is not along the direction relevant to the patches used in @xcite or @xcite .
a direct comparison of the directions of the various velocity measurements is hard as again our sdss measurement is a statistical average over many subregions of space and thus has no directional information . that said , our statistical measurement of coherent velocities within such sized volumes of the universe is smaller , to high statistical significance , than that measured locally around our galaxy .
a more interesting comparison would be with the measurements of @xcite who find a bulk flow " of x ray clusters , with respect to the cmb , of @xmath171 km / s out to @xmath172mpc ( or 560@xmath173mpc for comparison herein ) in approximately the same direction as discussed above for the local measurements @xcite .
again , it is hard to make a direction comparison in terms of the bulk velocity amplitude and direction as our statistical measurements are derived from higher redshift @xcite , they note that their bulk flow velocities peak at @xmath174 with the possibility of higher redshift clusters providing little to their dipole measurements .
this is therefore , below our low redshift lrg sample data , and averaged over a larger volume , than used by @xcite .
one possible explanation for the differences we are seeing is that our galaxy is located in an unusual part of the universe , e.g. , in a highly over , or underdense region of the universe .
again , our measurements are obtained at @xmath175 , beyond the redshift limits of all these local measurements @xcite .
moreover , we can not exclude the presence of a large constant large - scale dark flow " across the volume surveyed by our dr7 data i.e. , a velocity dipole with no variation across @xmath176gpc@xmath168 of the universe .
such a constant flow would be undetectable by our method as the correlation function in redshift space is distorted by the divergence of the peculiar velocity field , while measuring the bulk flows via the sz effect can include a possible global flow .
however , this explanation would require the dark flow " of @xcite to extend out beyond the sdss lrg sample ( @xmath177 ) to leave it undetectable , i.e. , over a 1000@xmath165 mpcs in scale .
this can be tested using the sdss - iii baryon oscillation spectroscopic survey ( boss ) , which will provide redshift - space distortion measurements both at higher redshift and over larger volumes of the universe @xcite . in the future
we plan to extend our measurements to higher redshifts , in order to extend the history of coherent motions ( e.g. fig .
[ fig : vphistory ] ) .
as shown in our paper , these motions can be measured in an independent way , free of some of the problems associated with other measures of the growth history of the universe , and the assumed cosmological model for the late
time universe , e.g. , we are able to test not only the conventional dark energy model , but also interacting " and clustered " dark energy models , not to mention the general class of modified gravity theories . unlike other approaches , our measurements are free from any possible violation of the consistent equations .
we thank the referee for helpful comments that improved the quality of this paper .
y - ss , cgs and rn are supported through the stfc , and i.k .
acknowledges support by jsps research fellowship and wpi initiative , mext , japan .
we thank eyal kazin for significant assistance in using his lrg data sample , as well as many fruitful discussions , and atsushi taruya for helpful comments in formulating eq .
we also thank ofer lahav for fruitful discussions .
the authors acknowledge the use of the ucl legion high performance computing facility , and associated support services , in the completion of this work .
we thank korea institute for advanced study for providing computing resources ( kias linux cluster system ) for this work . | we present new measurements of the coherent motion of galaxies based on observations of the large - scale redshift
space distortions seen in the two dimensional two point correlation function of luminous red galaxies in data release seven of the sloan digital sky survey .
we have developed a new methodology for estimating these coherent motions , which is less dependent on the details of galaxy bias and of the cosmological model to explain the late time acceleration of the expansion of the universe .
we measure a one dimensional velocity dispersion of galaxies on large scales of @xmath0 and @xmath1 at a mean redshift of @xmath2 and @xmath3 respectively .
these values are fully consistent with predictions for a wmap7normalised @xmath4cdm universe and inconsistent at confidence of @xmath5 with a dvali - gabadadze - porrati ( dgp ) model for the universe .
we can convert the units of these @xmath6 measurements to @xmath7 km / s and @xmath8 km / s respectively ( assuming a @xmath4cdm universe ) , which are lower that expected based on recent low redshift ( @xmath9 ) measurements of the peculiar velocity field ( or bulk flows " ) .
it is difficult to directly compare these measurements as they cover different redshift ranges and different areas of the sky .
however , one possible cosmological explanation for this discrepancy is that our galaxy is located in unusually over , or under , dense region of the universe . |
there were many trials to solve the model called ( in quantum optics ) rabi hamiltonian exactly , until braak @xcite recently succeeded to do so .
the same model is known under several pseudonyms , e. g. single - mode spin - boson system , jaynes - cummings model without rotating wave approximation and others ; a brief survey can be found in @xcite .
let us introduce a more general form of rabi hamiltonians @xcite , describing the interaction between a bosonic mode with energy @xmath2 and a two - level atom with level spacing @xmath3 @xmath4,\ ] ] where @xmath5 ; @xmath6 is the interaction constant , @xmath7 , @xmath8 are the pauli matrices , @xmath9 and @xmath10 are boson creation and annihilation operators , respectively . for the most studied @xmath11 case braak @xcite
recently presented an exact algebraic solution ; we are going to solve the @xmath12 model in full analogy .
before braak s general solution , some special points , called also the juddian ones , were known to be exactly solvable .
this means that for some constraints on the hamiltonian parameters , one can find both an eigenfunction and its eigenenergy .
it does nt give the complete spectrum , just one excited eigenstate .
this was shown at first for the standard @xmath11 rabi hamiltonian @xcite .
later the same picture , of course with different constraints and eigenvalues was shown for @xmath12 , i. e. the two - photon rabi hamiltonian @xcite .
these special points manifest themselves as cross - sections of general solutions and they will serve as a check of our results . it is also known that the parameters of the two - photon rabi model are restricted by @xmath13 , otherwise the eigenfunctions are not normalizable @xcite .
the first step of the solution is going over to bargmann space @xcite , introducing complex variable @xmath14 , where the bosonic operators simplify to [ e2 ] b , b^z . the eigenfunctions have to be analytic in the whole complex plane .
let us suppose that the solution of the stationary schrdinger equation is a vector composed of two analytic functions @xmath15 .
we can insert eq.([e2 ] ) into the hamiltonian ( [ e1 ] ) , case @xmath12 , and using the well - known @xmath16 pauli matrices , we obtain a coupled system of second - order differential equations @xmath17 where @xmath18 means the derivative with respect to @xmath14 and @xmath19 is the energy from the hamiltonian s spectrum . for better symmetrization it is useful to go over to a linear combination of the considered functions , namely @xmath20 and @xmath21 .
thus we get the set @xmath22 the next step is to find important symmetries present in the set ( [ e4 ] ) before they become less clear after some transformation followed by a series expansion . here
we prefer a somewhat different way than braak @xcite , but a very obvious one .
we will perform two transformations of the variable @xmath14 .
the first one is @xmath23 .
one can easily see , that it leaves the set ( [ e4 ] ) unchanged .
simple inspection shows , that this implies common parity of both @xmath24 and @xmath25 .
plainly speaking , they must be either both even @xmath26 or alternatively , simultaneously odd @xmath27 the second applied transformation is @xmath28 .
the set ( [ e4 ] ) becomes @xmath29 now the functions @xmath24 and @xmath25 evidently swapped their places , thus up to some common multiplicative constant : @xmath30 possible values of the parameter @xmath31 can be found by inserting @xmath32 instead of @xmath14 and we get @xmath33 and analogously @xmath34 . hence @xmath35 from eq.([e5 ] ) or @xmath36 from eq.([e6 ] ) .
the symmetry parameter can acquire four values , @xmath37 . in the standard @xmath11 rabi model such parameter had only two values @xmath1 and
the transformation @xmath38 was sufficient .
finally let us introduce four functions @xmath39 whose roots with respect to @xmath19 will be used to fulfill eq.([e8 ] ) : @xmath40 in the last two cases the lower of eqs.([e8 ] ) was multiplied by @xmath41 in order to keep the @xmath42 functions real .
they share the common parity of their @xmath43 functions .
the upper of eqs.([e8 ] ) is then no more independent .
the complete discrete spectra will be given by all roots of all @xmath39 functions , again in full analogy with the standard rabi model @xcite .
of course , the roots are meant with respect to @xmath19 and they are independent of any chosen @xmath14 . having categorized the symmetries , we return to the set ( [ e4 ] ) and perform the transformation @xmath44 .
we get @xmath45 where we used the parameter @xmath46 to simplify the first equation by removing the term @xmath47 , thus specifying its value [ e11 ] = .
in fact there should be @xmath48 in front of the square root , but the plus sign would make @xmath46 divergent in the limit @xmath49 , which is physically not reasonable .
notice that @xmath46 remains real under the above mentioned restriction @xmath50 .
there is also a further analogy with the standard rabi model , where the special case @xmath51 is exactly solved with help of the coherent state @xmath52 , here @xmath53 is the lowest bosonic state @xcite . in our notation , braak performs the transformation @xmath54 , recall @xmath55 in ( [ e2 ] ) .
the @xmath12 rabi hamiltonian solution at @xmath51 involves the squeezed vacuum @xcite term @xmath56 .
we will return to this case later . in the next step
we expand the functions @xmath57 : @xmath58 to keep the solutions analytic we expect @xmath59 for @xmath60 @xcite . inserting these expressions into eqs.([e10 ] ) we get the iteration scheme @xmath61q_n \nonumber\\ + \frac{\omega_0}{2}k_n= 0 \nonumber \\
-2g(n+2)(n+1)k_{n+2 } + [ ( \omega+8g\kappa)n+4g\kappa - e]k_n \nonumber\\ - 4\omega\kappa k_{n-2}+\frac{\omega_0}{2}q_n = 0.\ \ \ \ \end{aligned}\ ] ] notice that the indexes differ by 0 , 2 or 4 , thus only the coefficients @xmath62 @xmath63 with common parity will be non - zero , either for even @xmath64 , eq.([e5 ] ) , or odd @xmath65 , eq.([e6 ] ) . the prefactor @xmath66 does nt spoil the parity .
the starting point of our iteration scheme is either at @xmath67 or @xmath68 .
let us first have a look on the case @xmath67 and the symmetry parameter @xmath69 .
upper eq.([e8 ] ) @xmath70 at @xmath71 implies @xmath72 , which can itself be a function of hamiltonian parameters , serving as normalization constant for the eigenfunctions .
but as we are interested only in the roots of @xmath73 , this constant becomes an unimportant multiplication factor and we can choose [ e14 ] q_0=1,k_0=1 .
the second case is @xmath74 .
we have @xmath75 and again at @xmath71 [ e15 ] q_0=1,k_0=-1 .
the third case with @xmath76 is only a bit more complicated .
the symmetry condition from eq.([e8 ] ) is now @xmath77 and its value at @xmath71 is rather trivial @xmath78 .
but we will compare the first derivatives w. r. t. @xmath14 , i. e. @xmath79 at @xmath71 and we get @xmath80 , thus now our starting point is [ e16 ] q_1=1,k_1=1 . at last for @xmath81
we get [ e17 ] q_1=1,k_1=-1 . concluding this part
we can see that the hilbert space of eigenfunctions splits into four disjunct subspaces .
the corresponding eigenvalues can be found separately as roots of four @xmath82 functions , eq.([e9 ] ) .
we substitute @xmath83 , the coefficients of expanded @xmath57 functions are found applying the iteration scheme ( [ e13 ] ) subsequently with four starting points , eqs.([e14]-[e17 ] ) .
the coefficients not defined by eq.([e13 ] ) are zero because of parity demands .
@xmath84 as well .
before proceeding to numerical calculations , we ought to mention several special cases , where the exact solution was already known .
they will serve as a check of our general solution .
if the interaction constant is zero , the system separates into independent two - state atom with energy levels @xmath86 and a phonon with the modes @xmath87 , where @xmath88 thus the overall energy is @xmath89 .
it is instructive to see how these values split into four groups as roots of four @xmath42 functions .
therefore we will solve this simple case explicitly .
we return to the original functions @xmath90 , as the set of equations ( [ e3 ] ) decouples for @xmath85 and the two independent solutions are found easily @xmath91 where we denoted the exponents by @xmath92 and @xmath93 , as the parity conditions ( [ e5 ] ) and ( [ e6 ] ) are common for @xmath43 and @xmath90 functions .
they force @xmath92 and @xmath93 to be integer and the analyticity demands make them non - negative . the energies @xmath94 and @xmath95 should be common , but they are generally different , thus the overall solutions will have either @xmath96 or @xmath97 .
let us first analyze the case @xmath98 with the energy @xmath94 .
the linear combinations @xmath99 and @xmath100 .
the solutions for the functions @xmath101 finally yield @xmath102 the second solution @xmath103 with @xmath104 and eigenenergies @xmath95 gives @xmath105 we will later analyze mainly the cases when @xmath3 and @xmath2 are comparable and the eigenenergies as roots of @xmath42 functions reorganize as follows : @xmath106 we can see that real values of the symmetry parameter @xmath31 are connected with even number of phonon excitations @xmath107 whereas the imaginary @xmath31 is coupled to odd @xmath107 . the global ground state is always given by the lowest root of @xmath108 , which is true also for non - zero @xmath6 .
if the gap @xmath3 between atomic levels disappears , all eigenenergies become twice degenerate .
they are exactly known and now the complete spectrum is given by @xcite [ e22 ] e=-+(n+)n=0,1,2, where another dimensionless quantity was introduced = = 1- .
let us reproduce this result .
we return to the @xmath43 functions , because now the system of equations ( [ e4 ] ) decouples for @xmath51 and the general solutions are @xmath110 .
\nonumber\\ & \left[c_{11}h_{-n-1}\left(\sqrt{\frac{\omega\omega}{4 g}}z\right)+ c_{21}\ { } _ 1f_1\left(\frac{n+1}{2},\frac{1}{2},\frac{\omega\omega}{4 g}z^2\right)\right ] \nonumber \\ & \phi_2=\exp\left({\kappa z^2}\right ) . \nonumber\\ & \left[c_{12}h_n\left(\sqrt{\frac{\omega\omega}{4 g}}z\right)+
c_{22}\ { } _ 1f_1\left(-\frac{n}{2},\frac{1}{2},\frac{\omega\omega}{4 g}z^2\right)\right],\end{aligned}\ ] ] exploiting the well known hermitian polynomials @xmath111 and hypergeometric function @xmath112 . further we introduced the quantity @xmath113 .
it is nothing else but eq.([e22 ] ) reversed .
thus if the parity demands force @xmath114 to be non - negative integer again , the spectrum is reproduced .
let us show it in detail at least for the simpler case of @xmath114 even .
we make use of a formula relating hermitian polynomials and hypergeometric function , valid for @xmath115 = i^n e^-q^2 z^2 _ 1f_1 ( , , q^2 z^2 ) , whereas for any @xmath114 , including non - integer one , the kummer transformation @xcite gives [ e26 ] _
1f_1(-,,-y)=e^-y _ 1f_1(,,y ) .
first of all we set @xmath116 , because the hermitian polynomial with negative index has no parity , whereas the rest of the solutions ( [ e24 ] ) are even functions for @xmath114 even .
we define @xmath117 , @xmath118 and require @xmath119 using eqs .
( [ e23 ] ) - ( [ e26 ] ) .
we get [ e27 ] ( -2)^n/2(n-1)!!c_12+c_22c_21 = 0 . for @xmath114 non - integer ( or odd integer ) , the term with the hermitian polynomial would become complex , but our functions @xmath120 and @xmath121 are real - hence we set also @xmath122 .
we already mentioned that one parameter can be chosen , say the integration constant @xmath123 .
thus @xmath124 and the two @xmath42 functions are identical , @xmath125 .
one can compare the series expansion of these @xmath126 solutions with those given by the scheme ( [ e13 ] ) , starting points ( [ e14 ] ) or ( [ e15 ] ) and see that in fact we managed to perform the complete sum of eqs .
( [ e12 ] ) . having exact formulas for some @xmath43 and @xmath39 functions at disposal allows us to make several observations . in the on - line supplement to ref .
@xcite , braak reports some problems with the radius of convergence @xmath127 of his series in @xmath14 .
@xmath127 seemed to be finite in some cases , though analyticity in the whole complex plane of @xmath14 is required . to keep the series analogous to our eqs.([e12 ] ) convergent , as a necessary condition , the ratio @xmath128 had to go to zero for @xmath129 .
if it was non - zero , @xmath127 became finite .
for the special case @xmath51 we can calculate this ratio explicitly .
the coefficients @xmath130 with @xmath131 become rather simple [ e28 ] q_2k=(e-_2k-2)(e-_2k-4 ) ( e-_0 ) , where @xmath132 are eigenenergies from eq.([e22 ] ) .
en route we see that the energy @xmath133 terminates @xmath134 to a ( hermite ) polynomial .
the ratio [ e29 ] q_2k+2/q_2k-0,kas required .
the equations for @xmath135 are not so simple , thus we resort to the exact solution of @xmath25 and find [ e30 ] k_2k+2/k_2k0,k . for @xmath114 odd and
especially for non - zero @xmath3 we performed at least numerical study of @xmath63 , @xmath130 from the scheme ( [ e13 ] ) and found that their ratios are also proportional to @xmath136 in leading term .
hence we experience no problems with analyticity of the expanded functions including even very large @xmath14 .
the next remark concerns the practical numerical calculations of the roots @xmath137 , independent of @xmath14 .
it turns out that such calculations are numerically more stable for large @xmath14 , which is allowed by the previous notion .
one can exploit the large-@xmath14 asymptotic for the @xmath138 contribution to @xmath121 given by @xcite _ 1f_1(a , b , y)=e^y y^a - by0 , which is analytic for non - negative integer power @xmath139 , i. e. @xmath140 as expected .
@xmath120 gives the same after kummer transformation ( [ e26 ] ) .
further we return to the complete spectrum in eq.([e22 ] ) .
for @xmath141 the quantity @xmath142 and the energy becomes infinitely many times degenerate .
this point is physically unsound , though well defined in the sense of a limit .
concluding , the eigenvalue of eq.([e22 ] ) for @xmath67 coincides with mutually equal lowest roots of @xmath108 and @xmath143 , for @xmath68 it is the lowest root of both @xmath144 and @xmath145 , for @xmath146 the second lowest root of both @xmath108 and @xmath143 , etc .
let us now recall the result of emary and bishop @xcite , who found a set of isolated solutions for our model .
we are not going to rederive it , but the basic fact is that under some constraint on hamiltonian s parameters and energy @xmath19 , at least one of the original eigenfunctions @xmath90 becomes a product of some exponential function and of a polynomial .
the main statement is that there exist exactly known eigenstates with eigenenergies @xmath148 if the following conditions are fulfilled @xmath149 etc .
we will proceed so that we choose the values of @xmath3 and @xmath2 , than we gradually change @xmath6 .
eq.([e32 ] ) yields value(s ) of @xmath6 and followingly the energy is got from eq.([e31 ] ) .
these solutions will manifest themselves as crossing points of appropriate roots of @xmath108 and @xmath143 in the case of @xmath107 even , or crossings of @xmath144 and @xmath145 roots in the case of @xmath107 odd .
this is again analogous with the standard @xmath11 rabi model , where the juddian points appeared as crossings of the roots of @xmath150 and @xmath151 @xcite .
recall that contrary to eq .
( [ e22 ] ) , this is only one known excited state for chosen set of hamiltonian parameters , not the complete spectrum .
[ t ] eigenenergies as roots of @xmath82 functions for @xmath51 and @xmath152 .
full lines are exact values.,title="fig : " ]
the resulting energies as roots of @xmath42 functions should not depend on @xmath14 ; this would be true after summing infinite number of summands in eq.([e12 ] ) . in numerical calculations
we truncate the sum and hope that the higher powers are not significant .
we can not use very small @xmath14 , because of numerical instability .
somewhat surprisingly , we can use as large values as @xmath153 or even @xmath154 without real change of the roots .
this fact was already noted for exactly solvable cases , nevertheless one even does nt have to use compromise medium values despite of the truncated expansions . using the symbolic program mathematica
, we could sum up to @xmath155 term with @xmath156 for even solutions or @xmath157 for the odd ones . if the values of a root with smaller @xmath158 converged to the same value , we accept it .
it turns out that for smaller values of @xmath159 the convergence is excellent , but it becomes poorer as we approach the maximal possible value , i. e. if @xmath160 . the interval of @xmath6 in our plots is limited to @xmath161 .
it is known that although the eigenfunctions differ after changing the sign of @xmath6 , the eigenenergies remain the same , i. e. there is a mirror symmetry @xmath162 .
let us test our calculations at first on the exactly solved case @xmath51 .
we choose @xmath152 .
parts of parabolas with common top form the exact spectrum from eq .
( [ e22 ] ) , see full lines in fig .
it is clear , that in the vicinity of the infinitely many times degenerate point with @xmath163 , the roots of @xmath82 functions become very dense and the functions themselves are quickly oscillating .
that is the reason , why even for the orders as large as @xmath156 the values of appropriate roots did not converge completely and we have to resort to some fitting procedure , yielding better guess of the saturation value for @xmath164 .
the roots are denoted so that the lowest one is @xmath165 , the second lowest one @xmath166 , etc .
we can see that the calculated eigenenergies fit the exact values almost perfectly , except for some deviation at @xmath6 close to @xmath167 and for the higher root , in this case @xmath168 .
[ t ] functions for @xmath169 and @xmath170 . ] in fig .
2 we present three lowest roots of @xmath171 , full symbols , and two lowest roots of @xmath172 , smaller empty symbols . the values @xmath169 and @xmath170 are chosen so that the spectrum at @xmath85 becomes equidistant , see eq.([e21 ] ) .
the large empty circles are exact solutions of eqs.([e32 ] ) and ( [ e31 ] ) .
the value of @xmath107 is written nearby .
we can see almost perfect match with the crossings of appropriate lines .
the bottom four lines with @xmath173 and @xmath174 do not cross ; lines with @xmath175 cross twice .
3 shows the same roots denoted by the same symbols as fig.2 , just for @xmath176 and @xmath152 . besides the well fitted exact crossing points , there are other crossings of lines with different @xmath107 , which are not exactly known .
a similar figure with in fact the same hamiltonian parameters was already published @xcite , where the authors plot also numerical results from larger matrices diagonalization .
there is a couple of simpler analytic results , that can be got from our approach .
so we can find e. g. the small-@xmath6 expansion of the ( global ) ground state energy @xmath177 : e_0 - -+o(g^4 ) , g,_0 which can be compared with similar result for the @xmath11 rabi model : @xmath178 @xcite .
[ t ] functions for @xmath176 and @xmath152 .
the description of symbols is the same as in figure 2 . ]
we have found the complete spectrum of the two - photon rabi hamiltonian as roots of four analytic functions @xmath121 and @xmath172 , in the whole parametric space .
these functions are given by the recurrence scheme ( [ e13 ] ) with four starting points ( [ e14 ] ) - ( [ e17 ] ) .
the unnormalized eigenfunctions in bargmann space can be found as well , using @xmath179/2 $ ] and @xmath180/2 $ ] . in his viewpoint : the dialogue between quantum light and matter
@xcite , e. solano states that braak @xcite managed to enlarge the class of exactly solvable models and that he added the standard rabi model on the short list of exactly solvable quantum systems .
we believe that this paper adds also the two - photon rabi hamiltonian on the same list .
this list can almost surely be extended further by using braak s approach on other related models and yet another task for future is deeper understanding of the criteria of its applicability .
d. braak , phys .
lett * 107 * , 100401 ( 2011 ) v. v. albert , g. d. scholes and p. brumer , phys .
a * 84 * , 042110 ( 2011 ) i. i. rabi , phys . rev . * 51 * , 652 ( 1937 ) b. r. judd , j. phys .
c * 12 * , 1685 ( 1979 ) c. emary and r. f. bishop , j. phys .
a * 35 * , 8231 ( 2002 ) s. n. dolya , j. math . phys . * 50 * , 033512 ( 2009 ) v. bargmann , comm .
pure appl . math .
* 14 * , 197 ( 1961 ) r. f. bishop , n. j. davidson , r. m. quick and d. m. van der walt , phys .
a * 254 * , 215 ( 1999 ) i. travenec and l. samaj , phys .
a * 375 * , 4104 ( 2011 ) c. brif , ann
. phys . * 251 * , 180 ( 1996 ) a. d. polyanin and v. f. zaitsev , exact solutions for ordinary differential equations , chapman & hall / crc , n. y. ( 2003 ) e. solano , physics * 4 * , ( 2011 ) | exact spectrum of the two - photon rabi hamiltonian is found , proceeding in full analogy with the solution of standard ( one - photon ) rabi hamiltonian , published by braak in phys .
rev .
lett .
107 , 100401 ( 2011 ) .
the hamiltonian is rewritten as a set of two differential equations .
symmetries that get hidden after further treatment are found .
one can plainly see , how the hilbert space splits into four disjunct subspaces , categorized by four values of the symmetry parameter @xmath0 .
there were only two values @xmath1 for the standard rabi model .
four analytic functions are introduced by a recurrence scheme for the coefficients of their series expansion .
all their roots yield the complete spectrum of the hamiltonian .
eigenstates in bargmann space are also at disposal . |
it is known that in conventional inflationary models vector fluctuations of the metric are not amplified by the pumping action of the gravitational field @xcite . on the contrary
both scalar and tensor modes of the geometry are copiously generated in the transition from a quasi - de sitter stage of expansion to a radiation - dominated evolution .
vector modes of the geometry may have interesting phenomenological implications .
for instance , as argued long ago by wassermann @xcite ( see also @xcite ) the existence of large - scale vorticity ( sometimes connected with a large - scale magnetic field ) may have relevant implications for the origin of global rotation of spiral galaxies .
the possible existence of primordial vorticity has been the subject of various investigations .
harrison @xcite ( see also @xcite for a complementary perspective ) noticed that if vorticity is present over sufficiently large length scales in a plasma of ions and electrons , magnetic fields can be induced over scales comparable with the typical scale of the vorticity .
the dynamical origin of primordial vorticity has been , since then , a subject of debate . in @xcite
it was suggested that cosmic strings with small scale structure may be responsible for the dynamical origin of vorticity .
the dynamical friction between cosmic strings and matter may also be responsible for further sources of vorticity @xcite . in spite of the fact that the considerations quoted so far may indeed produce large - scale vorticity , it is useful to take another approach and consider the possibility that , thanks to the pumping action of the gravitational field , the vector fluctuations of the geometry will be amplified , providing , at some later epoch , the initial condition for the evolution of the rotational fluctuations of the fluid . the motivation of the present investigation is to discuss the evolution of vector modes of the geometry in the case of a bouncing background .
as is known , the ten - dimensional degrees of freedom of a four - dimensional metric can be decomposed , with respect to three - dimensional rotations , into four scalar modes , two tensor modes ( described by a divergenceless and traceless rank - two symmetric tensor in three dimensions ) , and four vector degrees of freedom ( described by two divergenceless vectors in three dimensions ) . out of these ten degrees of freedom
, four can be fixed by a choice of coordinate system . while the tensor modes are invariant under infinitesimal gauge transformations , the scalar and vector modes are not . while this work was in progress , battefeld and brandenberger @xcite argued that it is not unlikely that vector modes of the geometry may be amplified in the case of four - dimensional curvature bounces as the ones provided by cyclic / ekpyrotic scenarios @xcite or in the context of the pre - big bang scenario @xcite .
the authors of ref .
@xcite worked in a four - dimensional conformally flat frw model and discussed the dynamics in the context of an einsteinian theory of gravity .
they found a growing - vector zero mode and correctly pointed out that the specific fate of such a mode is rather sensitive to the detailed dynamics of the bounce .
there are , in principle , different bouncing dynamics that can be analysed .
the simplest case is the one of a four - dimensional background without fluid sources but with a scalar degree of freedom ( the dilaton ) . in this case , during the contraction the vector zero mode increases , but as will be argued in specific examples , its fate depends upon the dynamics of the curvature bounce and , in all the analytical examples reported in the present paper , it matches with a decaying mode after the bounce .
a similar situation also arises , also in four dimensions but in the presence of fluid sources and together with a dilatonic background . in the multidimensional case
the situation becomes increasingly interesting .
there more scalar and vector modes arise and their quantum mechanical fluctuations can be amplified . for the sake of completeness , even if the scalar modes of the geometry are not the subject of the present investigation , we want to mention that curvature bounces have been the subject of various debates concerning the fate of the scalar fluctuations of the geometry ( see @xcite for a partial list of references ) .
the plan of the present paper will be the following . in section 2
the evolution of vector modes of the geometry will be analysed in the context of four - dimensional curvature bounces . in the string frame description
, it will be shown that the vector modes present prior to the curvature bounce match continuously with a decreasing vector mode after the bounce .
specific analytical examples will be presented both in the absence and in the presence of fluid sources . in section 3
the specific case of multidimensional curvature bounces will be discussed .
the evolution of the vector fluctuations of a five - dimensional curvature bounce will be derived in fully gauge - invariant terms . in section 4
the spectrum of the vector modes will be derived and it will be shown that vector modes can be amplified in the five - dimensional model of curvature bounce .
section 5 contains the concluding remarks with a critical summary of the main suggestions stemming from the present study .
finally , the appendix contains useful technical results needed for the derivation of the evolution equations of vector modes in more than four dimensions .
consider , to begin with , the low - energy gravi - dilaton action in four space - time dimensions and in the string frame description @xcite : @xmath0 . \label{locac}\ ] ] the potential @xmath1 is , for the moment , a local function of the dilaton field @xmath2 and @xmath3 is the string length . a linear combination of the equations of motion derived from eq .
( [ locac ] ) leads to the well known form of the low - energy beta functions @xmath4 in a conformally flat frw background and assuming the dilaton field is homogeneous @xmath5,\,\,\,\,\,\,\,\,\,\varphi= \varphi(\eta),\ ] ] the @xmath6 component of eq .
( [ beta ] ) leads to the following relation @xmath7 in eq .
( [ b0 ] ) the prime denotes a derivation with respect to the conformal time coordinate @xmath8 while @xmath9 and @xmath10 is the time derivative of the so - called shifted dilaton , which is invariant under scale - factor duality . in this framework
, the vector modes of the geometry are parametrized by two vectors , @xmath11 and @xmath12 , which are divergenceless , i.e. @xmath13 : @xmath14 from eq .
( [ fluc ] ) , the components ricci tensor can be computed to first order in the fluctuations of the metric : @xmath15 , \nonumber\\ & & \delta r_{ij } = - \partial_{(i } q_{j ) } ' - \partial_{(i } w_{j ) } '' - 2 ( { \cal h } ' + 2 { \cal h}^2 ) \partial_{(i } w_{j ) } - 2 { \cal h } [ \partial_{(i } w_{j ) } ' + \partial_{(i } q_{j ) } ] .
\label{ricci1 } \end{aligned}\ ] ] finally , using eq .
( [ ricci1 ] ) , the first - order fluctuation of eq . ( [ beta ] ) leads to the following pair of differential equations : @xmath16 q_{i } - \frac{1}{2 } \nabla^2 ( q_{i } + w_{i } ' ) = 0 , \label{int0}\\ & & - \partial_{(i } q_{j ) } ' - \partial_{(i } w_{j ) } ''
+ ( \overline{\varphi } ' + { \cal h } ) \partial_{(i } q_{j ) } + \partial_{(i } w_{j ) } ' \nonumber\\ & & - 2 \biggl[{\cal h } ' - { \cal h}^2 - { \cal h } \overline{\varphi } ' + \frac { a^2}{2 } \frac{\partial v}{\partial \varphi}\biggr ] \partial_{(i } w_{j ) } = 0 .
\label{int1}\end{aligned}\ ] ] recalling the specific form of eq .
( [ b0 ] ) , it can be appreciated that various terms cancel in eqs .
( [ int0 ] ) and ( [ int1 ] ) .
defining now the new variable @xmath17 eqs .
( [ int0 ] ) and ( [ int1 ] ) can be written as @xmath18 the variable @xmath19 is invariant under infinitesimal coordinate transformations , preserving the vector nature of the fluctuation . indeed , under infinitesimal diffeomorphisms @xmath11 and @xmath12 transform as @xmath20 where @xmath21 is the gauge function .
hence , as anticipated , @xmath19 is invariant under infinitesimal coordinate transformations .
this observation also implies that the condition @xmath22 fixes the gauge freedom completely and is analogous , in this sense , to the conformally newtonian gauge often employed in the analysis of scalar perturbations .
this gauge choice would have led to the evolution equations for @xmath11 , coinciding , in such a gauge , with the evolution of the gauge - invariant vector @xmath19 .
we will go back to the gauge issue during the discussion of vector fluctuations in the multidimensional case . from eqs .
( [ v01 ] ) and ( [ v02 ] ) we can deduce that only the zero mode is dynamical . in this case
its evolution can be written as @xmath23 where @xmath24 is an integration constant . during a conventional pre - big bang phase @xcite , as @xmath25 , the consistent solution of eq .
( [ b0 ] ) and of the other equations of the background can be written , in the absence of dilaton potential , as @xcite @xmath26 where @xmath27 is the cosmic time coordinate obeying the usual differential relation @xmath28 . inserting eq .
( [ solb1 ] ) into eq .
( [ sol0 ] ) @xmath29 this growing mode diverges in the limit @xmath30 .
this is essentially one of the arguments put forward in ref .
@xcite , with the difference that , in the present case , the discussion has been conducted , from the very beginning , in the string frame description .
the aim will now be to show that in a simple model of low - curvature transition , the growing vector mode will match to a decreasing solution in the post - bounce epoch .
as was shown in @xcite a smooth transition at low curvatures can be achieved if the dilaton potential is a local function of @xmath31 .
the generally covariant action describing this model is given , in @xmath32 space - time dimensions by @xmath33 + s_{\rm m } , \label{action1}\ ] ] where @xmath34 represents the contribution of effective ( fluid ) sources and where @xmath35 the variation of the action ( [ action1 ] ) with respect to @xmath36 and @xmath37 leads to the following equations : @xmath38 , \nonumber\\ & & - \frac{e^{-\varphi}}{2 } \sqrt { g^{\alpha\beta } \partial_{\alpha}\varphi \partial_{\beta } \varphi } \gamma_{\mu\nu } { \cal i}_{1 } = \lambda_{s}^{d -2 } t_{\mu\nu } , \label{eq1}\\ & & r + 2 g^{\alpha\beta } \nabla_{\alpha } \nabla_{\beta } \varphi - g^{\alpha\beta } \partial_{\alpha } \varphi\partial_{\beta } \varphi + v - \frac{\partial v}{\partial\overline{\varphi } } \nonumber\\ & & + e^{-\varphi } \frac{\hat{\nabla}^2 \varphi}{\sqrt{g^{\alpha\beta } \partial_{\alpha}\varphi\partial_{\beta}\varphi } } { \cal i}_1 - e^{-\varphi } v ' { \cal i}_{2 } = 0 , \label{eq2}\end{aligned}\ ] ] where ) as well as in eqs .
( [ def ] ) and ( [ comb ] ) the prime denotes a derivation with respect to the argument of the given functional and _ not _ the derivative with respect to the conformal time coordinate .
the two notations can not be confused since the latter is only employed when dealing with the explicit form of the equations on a given background . ]
@xmath39 and @xmath40 is the induced metric . in eqs .
( [ eq1 ] ) and ( [ eq2 ] ) the following integrals @xmath41 have also been defined . by combining eqs .
( [ eq1 ] ) and ( [ eq2 ] ) in order to eliminate the ricci scalar we obtain the following equation @xmath42 + \frac{1}{2 } e^{-\varphi } \biggl [ g_{\mu\nu } \frac{\hat{\nabla}^2 \varphi}{\sqrt{g^{\alpha\beta } \partial_{\alpha}\varphi \partial_{\beta } \varphi } } \nonumber\\ & & - \gamma_{\mu\nu } \sqrt{g^{\alpha\beta } \partial_{\alpha}\varphi\partial_{\beta}\varphi}\biggr ] { \cal i}_{1 } = \lambda_{s}^{d-2 } t_{\mu\nu}. \label{comb}\end{aligned}\ ] ] if @xmath43 and @xmath44 are set to zero and if @xmath45 is replaced by @xmath46 , eq .
( [ beta ] ) can be formally recovered from eq .
( [ comb ] ) . in the case of a homogeneous dilaton and for a conformally flat metric of frw type eqs .
( [ eq1 ] ) and ( [ comb ] ) lead , in @xmath47 and in units @xmath48 , to the following system of equations : @xmath49 where the stress tensor of the fluid sources @xmath50 has been assumed and where the overdot denotes a derivation with respect to the cosmic time coordinate . in eqs .
( [ b1])([b3 ] ) the energy and pressure densities have been rescaled as @xmath51 in the case @xmath52 , eqs .
( [ b1])([b4 ] ) admit the solution @xcite @xmath53^{1/\sqrt{3 } } , \label{scalefactor1}\\ & & \overline{\varphi } = - \frac{1}{2 } \log { ( 1 + \tau^2 ) } + \varphi_{0 } , \label{dilaton1}\end{aligned}\ ] ] where @xmath54 the scale @xmath55 fixes the width of the curvature bounce
. will also prove useful in the numerical discussion of the spectrum of vector fluctuations in multidimensional models . ]
another solution , which is effective in illustrating some aspects of the present analysis , may be obtained if the scale factor and the energy density of the fluid are both _
constant_. in this toy example , eqs .
( [ b1])([b4 ] ) may be integrated easily and the result is @xmath56 the evolution equations of the vector modes can be rederived by perturbing to first order eq .
( [ eq1 ] ) . in the case without sources , the result coincides in @xmath47 , with eqs .
( [ v01 ] ) and ( [ v02 ] ) . in the case with sources
( [ v01 ] ) and ( [ v02 ] ) become @xmath57 where @xmath58 is the fluctuation of the velocity field and where we used the explicit perturbed form of the energy
momentum tensor @xmath59 equations ( [ pot1])([dilaton1 ] ) can be used in order to obtain the explicit solution for the gauge - invariant vector mode , which is @xmath60 where , as before , @xmath61 is an integration constant carrying vector indices .
this shows that for @xmath62 , @xmath63 , as obtained previously .
however , for @xmath64 the dangerous growing mode turns into a decaying mode , i.e. @xmath65 .
furthermore , no divergence is present for @xmath66 . in the case of the solution given in eq .
( [ solf ] ) , eqs .
( [ source1 ] ) and ( [ source2 ] ) lead to @xmath67 where @xmath68 denotes the fourier mode .
= 9 cm the time evolution of the gauge - invariant vector mode is illustrated in fig .
[ figure1 ] for the two results obtained in eqs .
( [ vf1 ] ) and ( [ vf2 ] ) .
the analysis reported so far shows that in some simple models of curvature bounce the growing mode in the vector sector matches with a decaying mode after the bounce .
the analysis of gauge - invariant fluctuations in multidimensional and/or anisotropic cosmological models has been discussed , through the years , in different contexts .
one of the first attempts to adapt the original bardeen @xcite formalism to a multidimensional framework was discussed in @xcite with the aim of performing a general discussion of the main features of cosmological fluctuations in kaluza
klein models @xcite . of related interests are some works discussing the theory of cosmological fluctuations in bianchi models @xcite . in @xcite ,
a perturbative framework for the gauge - invariant discussion of multidimensional pre - big bang models was introduced .
the problem of gauge - invariant fluctuations in models with extradimensions received new momentum in the context of brane models , and different approaches ( see , for instance @xcite and references therein ) were developed in order to study the possible localization of modes of various spin of both gravitational and non - gravitational origin .
in @xcite vector modes of the geometry were analysed in the case of a six - dimensional abelian vortex leading to the localization of gravitational interaction ( see also @xcite for a more general discussion of codimension - two braneworlds and of their vector excitations ) .
it is useful to notice , preliminarily , that the action given in eq .
( [ action1 ] ) also admits classical solutions in the presence of internal ( contracting ) dimensions . in this section ,
the vector fluctuations of multidimensional cosmological models will be studied .
the line element will be written , in this case , as @xmath69 - b^2(t ) g_{ab } dy^{a } dy^{b } , \label{multidim}\ ] ] where @xmath36 is the @xmath32-dimensional metric ; @xmath70 and @xmath71 are the metric tensors of two maximally symmetric euclidian manifolds .
the indices @xmath72 run over the full @xmath73-dimensional space - time ; the @xmath74-expanding dimensions are typically three and the @xmath75-contracting dimensions ( labelled by capital latin letters ) will be , for a while , taken to be generic but the case @xmath76 will be of particular interest in the following part of the investigation . from eq .
( [ multidim ] ) in the particular case @xmath77 and @xmath78 , eqs .
( [ eq1 ] ) and ( [ eq2 ] ) can be written in explicit terms by introducing the expansion rates @xmath79 and @xmath80 ( the dot denotes a derivation with respect to the cosmic time coordinate ) : @xmath81 equations ( [ m1])([m3 ] ) come , respectively , from the @xmath82 , @xmath6 , @xmath83 components of eq .
( [ comb ] ) , while eq .
( [ m4 ] ) can be derived from eq .
( [ eq2 ] ) . in the case of @xmath84 , eqs .
( [ m1])([m4 ] ) are solved by the following ansatz @xmath85 } , \,\,\,\,\,\,\,\,\,\ , \tau = \frac{t}{t_0}. \label{solint2}\end{aligned}\ ] ] the solution given in eqs .
( [ solint1 ] ) and ( [ solint2 ] ) describes the expansion of the @xmath74 dimensions and the contraction ( at a dual rate ) of the @xmath75 internal dimensions .
recalling the relation between the cosmic time coordinate @xmath86 and the conformal time coordinate @xmath8 , i.e. @xmath28 , eqs .
( [ m1])([m4 ] ) can be written as @xmath87 where , in full analogy with the four - dimensional case , the rates of expansion in conformal time have been defined : @xmath88 a particularly interesting case is the one where only one extradimension is present ( i.e. @xmath89 ) . in this case , setting @xmath76 and @xmath90 in eq .
( [ solint1 ] ) the scale factors are @xmath91 from eq .
( [ explscale ] ) , it can be easily deduced that the scale factor @xmath92 smoothly interpolates between @xmath93 for @xmath94 and @xmath95 for @xmath96 .
moreover , @xmath97 will contract like @xmath98 for @xmath99 and will contract like @xmath100 for @xmath101 .
both the scale factors and the curvature invariants are regular in the origin and for @xmath102 . finally , a pleasant feature of the case @xmath90 and @xmath76 is that in the limit @xmath96 the usual radiation - dominated evolution is correctly reproduced . in fig .
[ figure2 ] the time evolution of the scale factors @xmath103 and @xmath104 and of the associated rates , i.e. @xmath105 and @xmath106 , is illustrated .
this model describes the decoupling of internal and external dimensions .
= 9 cm in the following part of the present section , the vector modes arising in the model of eq .
( [ explscale ] ) will be specifically analysed .
the equations of motion will be derived in a gauge - independent way .
the fluctuations of a five - dimensional line element of the type of eq .
( [ multidim ] ) are , in general , described by @xmath107 degrees of freedom which can be formally decomposed in terms of their transformation properties with respect to three - dimensional rotations as @xmath108 where @xmath109 according to ( [ tens ] ) @xmath110 carries two degrees of freedom since it is symmetric in the two indices .
equation ( [ vec ] ) tells us that @xmath12 , @xmath11 and @xmath111 carry overall six degrees of freedom .
the remaining degrees of freedom are seven scalars , which will not be specifically discussed since the attention of the present analysis is mainly centred on the vector modes .
the vectors @xmath12 , @xmath11 and @xmath111 obey a set of differential equations , which we ought to derive now .
notice , preliminarily , that in terms of the divergenceless gauge parameter @xmath112 the vectors @xmath12 , @xmath11 and @xmath111 transform , for infinitesimal diffeomorphisms preserving the vector nature of the fluctuation , as @xmath113 as in the four - dimensional case , gauge - invariant fluctuations can be defined , namely @xmath114 notice that ( [ gione ] ) corresponds to the gauge - invariant fluctuation of the four - dimensional case . in order to derive the coupled evolution of @xmath19 and @xmath115 , eq .
( [ eq1 ] ) shall be linearized to first order in @xmath11 , @xmath12 and @xmath116 . by inserting the definitions ( [ gione ] ) and ( [ gitwo ] ) into the obtained equations
, the system will only contain the relevant gauge - invariant combinations , i.e. @xmath19 and @xmath115 . from eq .
( [ eq1 ] ) , the general form of the various components of the perturbed equation can be written , in a compact form , as @xmath117 - \frac{{\varphi}'}{2 } [ \partial_{i } { w^{j } } ' + \partial^{j } w_{i } ' ] = 0 , \label{p1}\\ & & a^2 \delta { \cal g}_{i}^{y } + \frac{a } { 2 b } \varphi ' [ h_{i } ' + ( { \cal f } - { \cal h } ) h_{i } ] - \frac{\varphi'}{2 } \frac{a^2}{b^2 } \partial_{y } q_{i } = 0 , \label{p2}\\ & & a^2 \delta { \cal g}_{0}^{i } - q^{i } ( \varphi '' - 2{\cal h}\varphi ' ) + \frac{a^2}{2 } \frac{\partial v}{\partial \overline{\varphi } } q^{i } = 0 .
\label{p3}\end{aligned}\ ] ] in order to derive eqs .
( [ p1])([p3 ] ) the perturbed connections reported in the appendix ( see eq .
( [ christoffel ] ) ) have been inserted in the linearized form of eq .
( [ eq1 ] ) , bearing in mind that the fluctuation of the induced metric ( [ induced ] ) implies @xmath118 from eqs .
( [ p1])([p3 ] ) , using eqs .
( [ r0iup])([riyup ] ) reported in the appendix , the following expressions are obtained : @xmath119 - \frac{a}{2 b } \partial_{y}\biggl [ \partial_{i}h^{j } + \partial^{j } h_{i}\biggr ] \nonumber\\ & & + \frac{1}{2 } \biggl [ ( \partial_{i}w^{j } + \partial^{j } w_{i } ) '' + ( \overline { \varphi } ' + { \cal h } ) ( \partial_{i } w^{j } + \partial^{j}w_{i } ) ' \biggr ] \nonumber\\ & & - \frac{a^2}{2 b^2 } \partial_{y}\biggl [ \partial_{i}w^{j } + \partial^{j}w_{i}\biggr ] = 0 , \label{p1ex}\\ & & q_{i } \biggl[{\overline{\varphi}'}^2 + 3 { \cal h}^2 + { \cal f}^2 - 2 \overline{\varphi } '' + 2 { \cal h } \overline{\varphi } ' - va^2 + \frac{\partial v}{\partial \overline{\varphi } } a^2 \biggr ] \nonumber\\ & & + \frac{1}{2 } \nabla^2 q_{i } + \frac{1}{2 } \nabla^2 w_{i } ' + \frac{a^2}{2 b^2 } \nabla^2_{y } q_{i } - \frac{a}{2 b } \partial_{y } \biggl [ h_{i } ' + ( { \cal f } - { \cal h } ) h_{i } \biggr ] = 0 , \label{p2ex}\\ & & \frac{1}{2 } \partial_{y } \biggl [ q_{i } ' + ( { \cal h } - 2{\cal f } - \overline{\varphi } ' ) q_{i}\biggr ] + \frac{1}{2 } \partial_{y } \nabla^2 w_{i } \nonumber\\ & & - \frac{b}{2 a } \biggl [ h_{i } '' - ( \overline{\varphi } ' + { \cal h } ) h_{i } ' - \nabla^2 h_{i } - ( { \cal h } - { \cal f})^2 h_{i } - \nabla^2 h_{i } \biggr ] = 0 . \label{p3ex}\end{aligned}\ ] ] recalling the definitions of the gauge - invariant fluctuations given in eqs .
( [ gione ] ) and ( [ gitwo ] ) , eq .
( [ p1ex ] ) becomes @xmath120 + \frac{1}{2 } \biggl[(\partial_{i } v^{j } + \partial^{j}v_{i } ) ' - ( { \overline{\varphi } } ' + { \cal h } ) ( \partial_{i } v^{j } + \partial^{j}v_{i})\biggr ] = 0 , \label{ij}\ ] ] by virtue of eq .
( [ m4a ] ) , the first term of eq .
( [ p2ex ] ) vanishes . the use of the gauge - invariant fluctuations @xmath19 and @xmath115 into eq .
( [ p2ex ] ) then leads to @xmath121 + \frac{a^2}{2 b^2 } \nabla^2_{y } v_{i } + \frac{1}{2 } \nabla^2 v_{i}=0 . \label{0i}\ ] ] finally , inserting eqs .
( [ gione ] ) and ( [ gitwo ] ) into eq .
( [ p3ex ] ) , the last gauge - invariant evolution equation will be : the @xmath122 equation @xmath123 - \frac{b}{2a } [ z_{i } '' - ( { \overline{\varphi } } ' + { \cal h } ) z_{i } ' - ( { \cal h } -{\cal f})^2 z_{i } - \nabla^2 z_{i } ] = 0 .
\label{iy}\ ] ] equations ( [ ij])([iy ] ) can also be derived with gauge - dependent methods , as can be expected from general considerations .
for instance , fixing the gauge in such away that @xmath22 the set of equations ( [ ij])([iy ] ) can be exactly reproduced , as was checked by explicit calculation . by formally taking the limit @xmath124 and by setting to zero the derivatives with respect to the internal coordinate @xmath125 , the four - dimensional equations derived in eqs .
( [ v01 ] ) and ( [ v02 ] ) are formally recovered .
for the purposes of this investigation it is interesting to consider the spectrum of the zero modes with respect to the internal dimensions .
this is achieved by setting to zero all the terms containing one ( or more ) derivatives with respect to @xmath125 .
thus , from eqs .
( [ 0i])([iy ] ) the following system : @xmath126 is obtained .
the evolution of @xmath111 and @xmath11 is , in this approximation , decoupled .
furthermore , it can be noticed that the evolution of @xmath11 is the one already discussed in the four - dimensional case , with the minor difference that , in the present case , the background evolution differs from the four - dimensional case . from eq .
( [ ii ] ) we then have @xmath127 which also implies that @xmath128 for @xmath129 . hence , even if the time dependence is different from the four - dimensional case , it is still true that the @xmath11 decreases at large ( positive ) times even if the solution is growing just before the bounce .
let us now concentrate on the evolution of @xmath111 whose equation can be written , in fourier space and for a single vector polarization as @xmath130 u = 0 , \label{u}\\ & & { \cal u}(\eta ) = ( { \cal h } - { \cal f})^2 - \biggl ( \frac{e^{- \overline{\varphi}/2}}{\sqrt{a}}\biggr ) '' \sqrt{a } e^{\overline{\varphi}/2 } , \label{potu}\end{aligned}\ ] ] where @xmath131 . this equation will be numerically solved in a moment for different values of @xmath68 .
however , in order to cross - check the numerical results it is useful to anticipate the analytical estimate . for this purpose
a useful observation is that , for times that are sufficiently far from the bounce , the asymptotic solutions are an excellent approximation for the background evolution . with this logic in mind , for @xmath132 the solution given by eqs .
( [ solint1 ] ) and ( [ solint2 ] ) can be approximated , in conformal time , by @xmath133 for @xmath134 , the solution ( [ solint1 ] ) and ( [ solint2 ] ) behaves , in conformal time , as @xmath135 therefore , in the two opposite limits , eq .
( [ u ] ) is given , respectively , by @xmath136 u_{i}= 0 , \,\,\,\,\,\,\,\,\,\
, \eta < - \eta_{1 } , \label{nu}\\ & & u_{i } '' + \biggl[k^2 - \frac{4 \mu^2 -1}{4\eta^2}\biggr ] u_{i}= 0 , \,\,\,\,\,\,\,\,\,\
, \eta > \eta_{1 } , \label{mu}\end{aligned}\ ] ] where @xmath137 is a time scale comparable with the size of the bounce . inserting eqs .
( [ tminus ] ) and ( [ tplus ] ) into eq .
( [ potu ] ) the indices @xmath138 and @xmath139 are determined to be @xmath140 and @xmath141 .
equations ( [ nu ] ) and ( [ mu ] ) can be solved exactly in the two limits .
now , imposing quantum - mechanical initial conditions for @xmath142 , @xmath143 the solution for @xmath144 can be written in terms of the mixing coefficients @xmath145 @xmath146 , ~\eta > \eta_{0 } , \label{uplus}\ ] ] where @xmath147 are the usual hankel functions of argument @xmath148 and index @xmath149 @xcite .
notice that the phases in eqs .
( [ uminus ] ) and ( [ uplus ] ) have been carefully selected in such a way as to recover mode functions with canonical quantum mechanical normalization in the limits @xmath150 .
the two solutions ( [ uminus ] ) and ( [ uplus ] ) can be continuously matched across the bounce , i.e. imposing @xmath151 and similarly for the first derivative .
the mixing coefficients turn out to be @xmath152 , \label{cpp}\\ & & c_{-}(k)= i \frac{\pi}{4 } x_1 e^{i\pi ( \nu -\mu)/2 } \biggl [ - \frac{\nu + \mu + 1}{x_1 } h_{\mu}^{(2)}(x_1 ) h_{\nu}^{(1)}(x_1 ) \nonumber\\ & & + h_{\mu}^{(2)}(x_1 ) h_{\nu+ 1}^{(1)}(x_1 ) + h_{\mu + 1}^{(2)}(x_1 ) h_{\nu}^{(1)}(x_1 ) \biggr ] , \label{cpm}\end{aligned}\ ] ] satisfying the exact wronskian normalization condition @xmath153 . in the small argument limit , i.e. @xmath154 the leading term in eq .
( [ cpm ] ) leads to @xmath155 which implies , after inserting the specific values of @xmath138 and @xmath139 that the number of produced vector quanta is given , for each polarization , by @xmath156 implying , in turn , that the energy density per logarithmic frequency interval , i.e. @xmath157 , increases at large distance - scales ( small @xmath68 ) as @xmath158 .
the same estimate will now be obtained through a totally different procedure , where the evolution equation for @xmath111 is solved directly using the exact background solutions . since the background solutions have a simple form in cosmic time , it is more practical to solve the evolution equation of the fluctuations directly in cosmic time
. this will not affect the determination of the mixing coefficients , which are independent of the specific parametrization of the time coordinate .
equation ( [ iii ] ) becomes , in cosmic time , @xmath159 h_{i } = 0 , \label{cost1}\ ] ] where @xmath160 is the physical momentum . by defining now the rescaled variable @xmath161 , eq .
( [ cost1 ] ) becomes : @xmath162v_{i}= 0 .
\label{cost2}\ ] ] inserting eqs .
( [ solint1 ] ) and ( [ solint2 ] ) into eq .
( [ cost2 ] ) , the following simple equation for a single vector polarization can be obtained : @xmath163 v=0 , \label{cost3}\ ] ] where @xmath164 is the usual rescaled cosmic time and @xmath165 .
notice that the variable @xmath166 defined in eq .
( [ cost3 ] ) is simply related to the variable @xmath167 defined in eq .
( [ u ] ) : @xmath168 since @xmath169 , the quantum - mechanical initial conditions will be , in terms of @xmath166 @xmath170 for @xmath171 . for @xmath96 the solution
can again be written in terms of the mixing coefficients @xmath172 as @xmath173.\ ] ] equation ( [ cost3 ] ) was solved for different values of @xmath68 ( or @xmath174 , the dimensionless comoving momentum ) .
quantum - mechanical initial conditions are given when the given mode is inside the horizon at the initial time of integration . from the numerical solution for the real and imaginary parts of @xmath166 and @xmath175 ,
i.e. @xmath176 it is then easy to obtain the mixing coefficients after the relevant mode re - entered during radiation , i.e. @xmath177 , \label{sum}\\ & & |c_{+}(k)|^2 - |c_{-}(k)|^2 = 2 ( q g - f p ) .
\label{diff}\end{aligned}\ ] ] the results of this calculation are reported in fig .
[ figure3 ] .
= 9 cm recall that , according to eq .
( [ diff ] ) , @xmath178 since @xmath179 is proportional to the wronskian of the solution . from fig .
[ figure3 ] it is apparent that for an increase of one order of magnitude in @xmath180 , @xmath181 increases by @xmath182 orders of magnitude , exactly as predicted in eq .
( [ pred ] ) .
in the present paper the evolution of vector modes of the geometry has been discussed in the case of bouncing cosmologies both in four and in higher dimensions .
the main suggestions following from a series of detailed examples can be summarized as follows : * in the four - dimensional case , as also noted in @xcite , a growing vector mode is present in the regime preceding the curvature bounce ; * however , in the simple analytical models described in the present paper , the growing mode matches with a decaying mode after the curvature bounce , so that no further dangerous divergences are introduced ; * in higher dimensions , more vector modes are expected and their spectrum was computed in a simple five - dimensional model of curvature bounce , where the compactification of the internal dimension is achieved dynamically ; * the typical spectra of vector modes obtained in the specific ( but fully consistent ) example are red after the modes reenter the horizon .
the most problematic aspect of this analysis is represented by the possible occurrence of red spectra , which may be rather dangerous in models based on curvature bounces .
the three divergenceless vectors appearing in five dimensions are defined as @xmath183 and @xmath184 notice that eqs .
( [ v1 ] ) and ( [ v2 ] ) are general in the sense that no specific gauge choice has been imposed . from eqs .
( [ v1 ] ) and ( [ v2 ] ) @xmath185 , \nonumber\\ & & \delta \gamma_{i0}^{y } = \frac{a}{2 b } [ h_{i } ' + ( { \cal f } - { \cal h } ) h_{i } ] - \frac{a^2}{2 b^2 } \partial_{y } q_{i } \nonumber\\ & & \delta \gamma_{0 y}^{i } = \frac{1}{2 } \partial_{y } q^{i } + \frac{b}{2 a } [ { h^{i } } ' + ( { \cal h } - { \cal f } ) h^{i } ] , \nonumber\\ & & \delta \gamma_{i j}^{y } = \frac{a}{2 b } ( \partial_{i } h_{j } + \partial_{j } h_{i } ) + \frac{a^2}{2 b^2 } ( \partial_{i } w_{j } + \partial_{j } w_{i } ) , \nonumber\\ & & \delta \gamma_{i j}^{k } = - { \cal h } q^{k } \delta_{i j } + \frac{1}{2 } \partial^{k}(\partial_{i } w_{j } + \partial_{j } w_{i } ) - \frac{1}{2}\partial_{j } ( \partial^{k } w_{i } + \partial_{i } w^{k } ) - \frac{1}{2 } \partial_{i}(\partial_{j}w^{k } + \partial^{k}w_{j } ) , \nonumber\\ & & \delta \gamma_{i y}^{j } = \frac{b}{2 a } ( \partial_{i } h^{j } - \partial^{j } h_{i } ) - \frac{1}{2 } \partial_{y } ( \partial^{j}w_{i } + \partial_{i}w^{j } ) .
\label{christoffel}\end{aligned}\ ] ] from the expressions of the christoffel connections , the first - order fluctuations of the ricci tensors can be determined : @xmath186 q^{i } - \frac{a^2}{2 b^2 } \nabla^2_{y } q_{i } - \frac{1}{2 } \nabla^2 q_{i } - \frac{1}{2 } \nabla^2 w_{i } ' \nonumber\\ & & + \frac{a}{2 b } \partial_{y } [ h_{i } ' + ( { \cal f } - { \cal h } ) h_{i } ] , \label{r0i}\\ & & \delta r_{i j } = \frac{a}{2 b } \partial_{y } [ \partial_{i } h_{j } + \partial_{j } h_{i } ] - \frac{1}{2 } \biggl\ { ( \partial_{i } q_{j } + \partial_{j } q_{i } ) ' + ( 2 { \cal h } + { \cal f } ) ( \partial_{i } q_{j } + \partial_{j } q_{i})\biggr\ } \nonumber\\ & & - \frac{1}{2 } ( \partial_{i}w_{j } '' + \partial_{j}w_{i } '' ) - \frac{1}{2}(2 { \cal h } + { \cal f } ) ( \partial_{i}w_{j } ' + \partial_{j}w_{i } ' ) \nonumber\\ & & + \frac{a^2}{2 b^2 } \partial_{y } ( \partial_{i } w_{j } + \partial_{j}w_{i } ) - \frac{1}{2 } ( \partial_{i}w_{j } + \partial_{j}w_{i } ) ( 2 { \cal h } ' + 4 { \cal h}^2 + 2 { \cal h } { \cal f } ) , \label{rij}\\ & & \delta r_{i y } = - \frac{1}{2 } \partial_{y } [ q_{i } ' + ( 4 { \cal h } - { \cal f } ) q_{i } ] - \frac{1}{2}\partial_{y } \nabla^2 w_{i } \nonumber\\ & & + \frac{b}{2 a } \biggl\ { h_{i } '' + ( 2 { \cal h } + { \cal f } ) h_{i } ' + ( { \cal h } ' + { \cal f } ' + { \cal h}^2 + 5 { \cal h } { \cal f } ) h_{i } - \nabla^2 h_{i}\biggr\}. \label{riy}\end{aligned}\ ] ] in eqs .
( [ r0i])([riy ] ) , @xmath187 denotes the laplacian of the internal dimensions .
the perturbed ricci tensors with mixed indices are also useful for a swifter calculation of the perturbed equations : @xmath188 + \frac{1}{2 b^2 } \nabla^2_{y } q^{i } + \frac{1}{2 a^2 } \nabla^2 q^{i } \nonumber\\ & & + \frac{1}{2a^2 } \nabla^2 w_{i } - \frac{1}{2 a b } \partial_{y } [ { h^{i } } ' + ( { \cal f } - { \cal h } ) h^{i } ] , \label{r0iup}\\ & & \delta r_{i}^{j } = - \frac{1}{2 a b } \partial_{y } [ \partial_{i } h^{j } -\partial^{j } h_{i } ] + \frac{1}{2 a^2 } \biggl\ { ( \partial_{i } q^{j } + \partial^{j } q_{i } ) ' + ( 2 { \cal h } + { \cal f } ) ( \partial_{i } q^{j } + \partial^{j } q_{i})\biggr\ } \nonumber\\ & & + \frac{1}{2 a^2 } \biggl\ { ( \partial_{i } w^{j } + \partial^{j } w_{i } ) '' + ( 2 { \cal h } + { \cal f } ) ( \partial_{i } w^{j } + \partial^{j } w_{i})'\biggr\ } - \frac{1}{2b^2}\partial_{y } [ \partial_{i}w^{j } + \partial^{j}w_{i } ] , \label{rijup}\\ & & \delta r_{i}^{y } = \frac{1}{2 b^2 } \partial_{y } [ q_{i } ' + ( 4 { \cal h } - { \cal f } ) q_{i } ] + \frac{1}{2b^2 } \partial_{y } \nabla^2 w_{i } \nonumber\\ & & - \frac{1}{2 a b } \biggl\ { h_{i } '' + ( 2 { \cal h } + { \cal f } ) h_{i } ' + ( { \cal f } ' - { \cal h } ' - 3 { \cal h}^2 + 3 { \cal h}{\cal f})h_{i } - \nabla^2 h_{i } \biggr\}. \label{riyup}\end{aligned}\ ] ] clearly , since vector fluctuations are parametrized by divergence - less vectors in three dimensions , their contribution to the first order fluctuation of the ricci scalar vanishes . hence , recalling the definition of the einstein tensor @xmath189 , @xmath190 j. khouri , b. a. ovrut , p. j. steinhardt , n. turok , phys
d * 66 * , 046005 ( 2002 ) ; + j. khoury , b. a. ovrut , n. seiberg , p. j. steinhardt , and n. turok , phys .
d * 65 * , 086007 ( 2002 ) ; + n. turok and p. j. seinhardt , e - print archive [ hep - th/0403020 ] .
m. gasperini , m. giovannini , and g. veneziano , phys .
b * 569 * , 113 ( 2003 ) ; + m. gasperini , m. giovannini , and g. veneziano , e - print archive [ hep - th/0401112 ] ( to appear in nucl .
b ) ; + m. giovannini , e - print archive : hep - th/0406098 .
d. h. lyth , phys .
b * 526 * , 172 ( 2002 ) ; _ ibid_. * 524 * , 1 ( 2002 ) ; + r. brandenberger and f. finelli , jhep * 11 * 054 ( 2001 ) ; + s. tsujikawa , r. brandenberger , and f. finelli , phys .
d * 66 * , 083513 ( 2002 ) ; + s. gratton , j. khoury , p. j. steinhardt and n. turok , phys .
d * 69 * , 103505 ( 2004 ) ; + c. cartier , e - print archive : hep - th/0401036 ; + l. e. allen and d. wands , e - print archive : astro - ph/0404441 .
d. sahdev , phys . rev .
d * 30 * , 2495 ( 1984 ) ; _ ibid_. * 39 * , 3155 ( 1989 ) ; + r. abbott , s. m. barr and s. d. ellis , phys .
rev d * 30 * , 720 ( 1984 ) ; + e. w. kolb , d. lindley and d. seckel , phys .
d * 30 * , 1205 ( 1984 ) . | it is argued that in the case of a smooth transition across a ( dilaton - driven ) curvature bounce the growing mode of the vector fluctuations matches continuously with a decaying mode at later times .
analytical examples of this observation are given both in the presence and in the absence of fluid sources .
in the case of multidimensional bouncing models the situation is different , since the system of differential equations describing the vector modes of the geometry has a richer structure .
the amplification of the vector modes of the geometry is specifically investigated in a regular five - dimensional bouncing curvature model where scale factors of the external and internal manifolds evolve at a dual rate .
vector fluctuations , in this case , can be copiously produced and are continuous across the bounce .
the relevance of these results is critically illustrated .
8.5 in -30pt 0.2 cm cern - ph - th/2004 - 128 _ department of physics , theory division , cern , 1211 geneva 23 , switzerland _ |
since snell discovered an outflow from a protostar in 1980 @xcite , over 300 outflows have been found in many star - forming regions @xcite .
recently , outflows from brown dwarf @xcite and o - type stars @xcite were discovered .
these observations indicate that outflow is ubiquitous in the star formation process . observations also show that outflows have two or more distinct velocity components ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
typically , an outflow is composed of a low - velocity component ( lvc ) with 10 - 50@xmath7 and a high - velocity component ( hvc ) with @xmath6@xmath10 @xcite .
outflows have a collimation factor ( i.e. , length / width or major / minor radius of the outflow ) of @xmath11 to @xmath12 @xcite , where @xmath13 and @xmath14 are the height and radius of the outflow , respectively .
there is a clear trend toward higher collimation at higher outflow velocity @xcite . since outflows have various morphological and kinematical properties ,
they can not be explained by a single - class model .
the flows that originated from the protostar are typically classified into two types : molecular outflow ( hereafter , outflow ) observed mainly with co molecules , and optical jet ( hereafter jet ) observed by optical emission @xcite .
outflows observed by co line emission exhibit a wide opening angle ( e.g. , * ? ? ?
* ) and slow velocity ( @xmath15 ; * ? ? ?
* ; * ? ? ?
jets observed by optical emission , however , exhibit good collimation ( e.g. , * ? ? ?
* ) and high velocity ( @xmath16 ; * ? ? ?
observations indicate that around each protostar , high - speed jets with a narrow opening angle are enclosed by a low - velocity outflow with a wide opening angle ( e.g. , * ? ? ?
* ) . what mechanism drives the outflow and the jet in the star formation process ?
during star formation , neither radiation nor thermal pressure from the protostar can supply sufficient kinetic energy and momentum for driving the jet and the outflow @xcite .
these flows ( outflows and jets ) are considered to be driven by release of the gravitational energy mediated by the lorentz and centrifugal forces ( cf .
* ; * ? ? ?
* ) . however , we have not understood what drives these flows during star formation .
it is difficult to directly observe the driving points of the outflow and the jet , because they are embedded in the dense cloud at several au from the protostar .
furthermore , we can not analytically investigate the driving mechanisms of the jet and the outflow in detail , because they have a large spatial scale and density contrast , and these flows are influenced by closely related lorentz , centrifugal , thermal pressure gradient , and gravitational forces .
therefore , numerical simulation is the most effective tool for investigating the driving mechanisms of the jet and the outflow .
@xcite showed that centrifugally driven outflow from the disk is possible if the poloidal component of the magnetic field makes an angle of less than 60@xmath17 ( i.e. , the disk wind model ) .
@xcite proposed that this models as the mechanism for the protostellar jet .
subsequently many numerical simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , which showed that disk could drive the jet confirmed the disk - wind mechanism around the protostar ( see , * ? ? ?
* ; * ? ? ?
@xcite proposed the x - wind model , in which the jet is driven in close proximity to the protostar .
both the disk wind and the x - wind are driven magnetocenrifugally from open magnetic field lines anchored on rapidly rotating circumstellar disks .
the difference between these models lies in where the field lines are anchored : near the radius of magnetospherical truncation on the disk for x - wind and over a wider range of disk radii for disk wind @xcite .
the jet speeds derived from both x - wind and disk wind theories correspond to the kepler speed at their driving point on the circumstellar disk .
thus , these models can not explain the two velocity components of jet and outflow .
many people believe that outflow ( i.e. , slow speed and wide opening angle flow ) is entrained by the jet ( i.e. , high speed and narrow opening angle flow ) , and that the jet is driven by either the disk wind or the x - wind ( e.g. , * ? ? ?
however , almost all simulations , except for a few studies , calculate the evolution of the circumstellar disk ( i.e. , jet driving ) after the protostar grows up sufficiently , and provide the idealized boundary , density profile of the disk , distribution of the magnetic field and velocity , accretion rate , etc .
after the protostar is formed , we can not judge whether these configurations are realized .
therefore to acquire adequate information about the protostar and to understand the driving mechanism of the jet and the outflows , we should calculate cloud evolution from the molecular cloud that forms the protostar .
simulations of cloud evolution from the molecular cloud core to stellar core formation show that the low - velocity flow is driven from the adiabatic core formed before protostar formation , and that the high - velocity flow is driven from another core . assuming spherical symmetry , the evolution from molecular cloud to stellar core has been investigated by many authors ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and they determined the density , velocity , and temperature structures of the collapsing cloud and stellar core .
they have shown that the adiabatic core ( or the first core ) with a radius of @xmath61au forms when the central density reaches @xmath18 , because the gas becomes optically thick and the thermal pressure increases . @xcite and
@xcite calculated the stellar core formation from the molecular cloud core using sph simulations .
however , the outflow and the jet do not appear in their calculations because they ignored the magnetic effect . the evolution of the magnetized cloud until the formation of the adiabatic core ( hereafter , the first core ) was investigated by @xcite , @xcite , @xcite , @xcite , and @xcite .
the outflow appears after the first core formation in @xcite , @xcite , and @xcite .
@xcite , and @xcite investigated the evolution of the magnetized clouds up to protostar formation .
they calculated the cloud evolution from the molecular cloud ( @xmath19 ) to protostar formation ( @xmath20 ) . in their calculations
, two distinct flows appeared : low - velocity flow ( @xmath21kms@xmath22 ) driven from the first core , and high - velocity flow ( @xmath23 kms@xmath22 ) driven from the second core ( i.e. , the protostar ) .
they expected that low - velocity flow would be observed as molecular outflow and high - velocity flow as an optical jet .
they adopted ideal mhd approximation , which is valid in the low - density gas region ( @xmath24 ) ; however , it is not valid in the high - density gas region ( @xmath25 ) .
@xcite found significant magnetic flux loss for @xmath26 by ohmic dissipation .
therefore , @xcite and @xcite overestimated the magnetic flux of the cloud , particularly in the high - density gas region .
there are two predictions concerning driving outflows and jets : one is that the outflow is entrained by the jet , and another is that the outflow and jet are driven from different objects .
these predictions are mutually conflicting . in this paper , we calculate cloud evolution from the molecular cloud core ( @xmath27 , @xmath28au ) to stellar core formation ( @xmath29 , @xmath30 ) using three - dimensional resistive mhd nested grid method , study the formation process of jets and outflows , and show the driving mechanisms of these flows . the numerical method of our computations and our model framework are given in 2 , and the numerical results are presented in 3 .
we discuss the driving mechanism of the outflow and the jet in 4 , and summarize our conclusions in 5 .
our initial settings are almost the same as those of @xcite .
to study cloud evolution , we use the three - dimensional resistive mhd nested grid code .
we solve the resistive mhd equations including self - gravity : @xmath31 where @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 , and @xmath37 denote the density , velocity , pressure , magnetic flux density , resistivity , and gravitational potential , respectively .
the resistivity ( @xmath36 ) in equation ( [ eq : reg ] ) is a function of density and temperature @xcite . for resistivity ( @xmath36 )
, we use the value adopted in @xcite . to mimic the temperature evolution determined by @xcite
, we adopt the piece - wise polytropic equation of state as @xmath38 where @xmath39ms@xmath22 , @xmath40 ( @xmath41 ) , @xmath42 ( @xmath43 ) , and @xmath44 ( @xmath45 ) . for convenience ,
we define the protostar formation epoch " as that at which the central density ( @xmath46 ) reaches @xmath47 .
we also call the period for which @xmath48 the isothermal phase , " the period for which @xmath49 the adiabatic phase , " the period for which @xmath50 the second collapse phase , " and the period for which @xmath51 the protostellar phase . " in this paper , we adopt a spherical cloud with critical bonnor - ebert @xcite density profile having @xmath52 ( @xmath53 ) of the central ( number ) density as the initial condition .
the critical radius for a bonnor ebert sphere @xmath54^{-1/2}$ ] corresponds to @xmath55au for our settings .
initially , the cloud rotates rigidly ( @xmath56 ) around the @xmath57-axis and has a uniform magnetic field ( @xmath58@xmath59 g ) parallel to the @xmath57-axis ( or rotation axis ) . to promote contraction
, we increase the density by 70% from the critical bonnor - ebert sphere .
the initial model is characterized by a single one non - dimensional parameter @xmath60 .
this parameter is related to the cloud s rotation rate , and is defined using a central density @xmath61 as @xmath62 the model parameters @xmath60 , magnetic field @xmath63 , angular velocity @xmath56 , total mass inside the critical radius @xmath64 , and the ratio of the thermal @xmath65 , rotational @xmath66 , and magnetic @xmath67 energies to the gravitational energy , , @xmath68 , @xmath69 , and @xmath70 , the relative factors against the gravitational energy are defined as @xmath71 , @xmath72 , and @xmath73 . ]
are summarized in table [ table : init ] .
we adopt the nested grid method ( for details , see * ? ? ?
* ; * ? ? ?
* ) to obtain high spatial resolution near the center .
each level of a rectangular grid has the same number of cells ( @xmath74 ) , although the cell width @xmath75 depends on the grid level @xmath76 .
the cell width is reduced by a factor of 1/2 as the grid level increases by 1 ( @xmath77 ) .
we assume mirror symmetry with respect to @xmath57 = 0 .
the highest level of a grid changes dynamically .
we begin our calculations with four grid levels ( @xmath78 , 2 , 3 , 4 ) .
the box size of the initial finest grid @xmath79 is chosen to be @xmath80 , where @xmath81 denotes the radius of the critical bonnor ebert sphere .
the coarsest grid @xmath78 , therefore , has a box size of @xmath82 .
a boundary condition is imposed at @xmath83 , such that the magnetic field and ambient gas rotate at an angular velocity of @xmath56 ( for details , see @xcite ) .
a new finer grid is generated whenever the minimum local @xmath84 becomes smaller than @xmath85 .
the maximum level of grids is restricted to @xmath86 .
since the density is highest in the finest grid , the generation of a new grid ensures the jeans condition of @xcite with a margin of safety factor of two .
we adopted the hyperbolic divergence @xmath87 cleaning method of @xcite .
assuming spherical symmetry , many authors investigated the evolution from molecular cloud to stellar core using radiation hydrodynamic calculations ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we briefly summarize the evolution of the collapsing cloud core , according to their calculations .
the molecular gas obeys the isothermal equation of state with a temperature of @xmath610k until @xmath88 ( isothermal phase ) , then the cloud collapses almost adiabatically for @xmath89 ( adiabatic phase ) . a quasi - static core ( i.e. , the first core ) with @xmath6@xmath90 forms in the adiabatic phase .
subsequently dissociation of molecular hydrogen when the density exceeds @xmath91 triggers further gravitational collapse ( i.e. , second collapse ) .
finally , the second core ( i.e. , protostar ) with @xmath6@xmath92 forms at @xmath93 . the protostar increases its mass for the subsequent gas accretion phase . in this study
, we calculated the cloud evolution from the molecular cloud core with @xmath94 to form a protostar with @xmath95 .
@xcite investigated cloud evolution in the isothermal and early adiabatic phases .
the evolution of the magnetic field and angular velocity in the collapsing cloud for @xmath96 was shown by @xcite .
therefore , this paper focuses mainly on the cloud evolution after the first core formation ( @xmath97 ) because we are interested in the driving mechanism of the outflow and the jet .
we parameterized angular velocity , magnetic field strength , and the ratio of thermal to gravitational energy of the initial cloud for 80 models .
cloud evolutions are typically classified into three types when the outflow or jet appears in the collapsing cloud .
we also found that cloud evolution depends on the ratio of the angular velocity to the magnetic field ( @xmath98 ) , but does not depend on the ratio of thermal to gravitational energy , especially after the first core formation ( @xmath97 ) .
observations indicate that molecular cloud cores have large magnetic energies and small rotational energies @xcite . in these types of clouds ,
cloud evolution is sensitive to the initial angular velocity , as shown in @xcite . in the following ,
we show only three typical models in which the initial clouds have the same magnetic field strengths and ratios of thermal to gravitational energy , but different angular velocities ( slow , moderate , and rapid rotation rate ) .
figure [ fig:1 ] shows the cloud evolution for model sr after the first core formation .
model sr has a parameter of @xmath99 , which means that the initial cloud has a small rotation energy ; the ratio of the rotational to gravitational energy is @xmath100 ( see table [ table : init ] ) .
table [ table : init ] shows that , in model sr , the rotational energy is much smaller than the thermal and magnetic energy in its initial state . the density ( false colors ) and velocity ( arrows ) distribution around the first core are plotted in the upper and middle panels of figure [ fig:1 ] . the plasma beta and magnetic fields around the first core
are plotted in the lower panels of figure [ fig:1 ] .
we plotted shocked regions as dotted lines in the upper and middle panels in figure [ fig:1 ] . since the first core is surrounded by the shock layer @xcite
, dotted lines indicate the first core . for model sr , the first core forms when the central density reaches @xmath101 .
the first core at its formation epoch has a mass of @xmath102 and a radius of @xmath103au , which is equivalent to or slightly larger than that expected by one - dimensional radiation hydrodynamic calculations @xcite .
@xcite showed that a larger first core forms in a rapidly rotating cloud because the centrifugal force suppresses the cloud collapse and the shock occurs in the earlier adiabatic phase .
however , the effect of centrifugal force at the first core formation epoch is small in model sr because a slowly rotating cloud is adopted as the initial state .
the magnetic effect slightly increases the mass and size of the first core , as shown in @xcite .
figures [ fig:1]_a _ and _ b _ show the density and velocity distribution on @xmath104 and @xmath105 cut planes at @xmath106 .
these panels indicate that the first core has a nearly spherical shape . to evaluate the core shape , we define the oblateness as @xmath107 , where @xmath108 , @xmath109 , and @xmath110 are the major axis , minor axis , and @xmath111-axis , respectively , derived from the moment of inertia for the high - density gas of @xmath112 according to matsumoto & hanawa ( 1999 ) .
figure [ fig:2]_a _ shows the evolution of the oblateness as a solid line .
the oblateness increases as the cloud collapses in the isothermal phase ( @xmath113 ) , and has a peak of @xmath114 ( i.e. , the ratio of the radial to vertical scale is about 10 ) at @xmath115 .
therefore , the central region has a disk - like structure at this epoch .
since the cloud rotates slowly , the disk is formed mainly by the lorentz force .
note that the magnetic energy is comparable to both the thermal and gravitational energy at the initial state ( see table [ table : init ] ) .
the oblateness oscillates around @xmath114 for @xmath116 , and then begins to decrease ( fig .
[ fig:1]_a _ ) . this decrease is caused by an increase in thermal energy .
when the gas density reaches @xmath18 , the cloud collapses adiabatically and thermal pressure increases .
the oblateness becomes @xmath117 at the first core formation epoch ( @xmath101 ) , which indicates that the first core has a nearly spherical shape , as shown in figure [ fig:1]_a _ and _
b_. figure [ fig:1]_c _ shows the plasma beta ( @xmath118 ) around the first core on the @xmath105 cut plane .
this panel shows that the magnetic energy inside the first core is extremely small ( @xmath119 ) , while that outside the first core is comparable to the thermal energy ( @xmath120 ) .
as shown in @xcite , the magnetic field is largely removed from the cloud core by ohmic dissipation for @xmath121 .
figure [ fig:2]_b _ shows the evolution of the magnetic flux density at the center of the cloud ( @xmath122 ) for model sr ( solid line ) .
the growth rate of the magnetic flux density becomes small for @xmath121 .
the solid line in figure [ fig:2]_c _ shows the magnetic flux density normalized by the square root of the gas density at the center of the cloud ( @xmath123 ) for model sr . in the ideal mhd regime ,
the evolution of @xmath123 depends on the geometry of the collapse .
this value , @xmath123 , remains constant after a thin disk forms because the magnetic field increases in proportional to the square root of the density when the disk - like cloud collapses ( @xmath124 ; @xcite ) . on the other hand
, @xmath123 increases in proportion to @xmath125 when the cloud collapses spherically ( for details , see * ? ? ? * ) .
thus , a rapid drop of @xmath123 for @xmath121 indicates the removal of the magnetic field from the center of the cloud by ohmic dissipation .
@xcite showed that the magnetic field is largely removed by ohmic dissipation for @xmath126 . since the first core forms at @xmath127 in this model , the first core is composed of the gas from which the magnetic field is already removed
therefore , the first core has a large plasma beta ( or weak magnetic field strength ) , as shown in figure [ fig:1]_c_. figures [ fig:1]_d _ and _ e _ show the density and velocity distributions around the first core at @xmath128 ( second collapse phase ) .
the cloud can collapse again for @xmath129 owing to dissociation of molecular hydrogen ( i.e. , second collapse ; @xcite ) .
the arrows in figure [ fig:1]_d _ and _ e _ indicate that the cloud collapses intensely around its center .
the size of first core at this epoch is slightly smaller than that at the epoch of figure [ fig:1]_a _ and _ b_. however , the first core has a nearly spherical shape ( @xmath117 at @xmath130 in fig . [ fig:2]_a _ ) because the anisotropic lorentz and centrifugal forces are weaker than the isotropic forces of the gravitational and thermal pressure gradient forces in this phase .
the lorentz force becomes weak due to magnetic dissipation , and an initially small rotation rate produces a weaker centrifugal force .
figure [ fig:2]_d _ shows the angular velocity normalized by the free - fall timescale at the center of the cloud , whose value corresponds to the ratio of the rotational to gravitational energy @xcite .
when this value reaches @xmath131 , the rotational energy becomes comparable to the gravitational energy , and cloud rotation begins to affect cloud evolution @xcite . in model sr ,
the effect of cloud rotation is significantly small in the second collapse phase ( @xmath132 ) because @xmath133 .
the magnetic field is again important for cloud evolution after the second collapse phase because ohmic dissipation becomes less effective , and the magnetic field can be amplified .
figure [ fig:2]_c _ shows that @xmath134 has a minimum at @xmath135 , and then continues to increase for @xmath136 .
however , the plasma beta inside the first core reaches only @xmath137 at @xmath93 ( fig.[fig:1]_f _ ) , and therefore , the magnetic energy is much smaller than the thermal energy , even in the second collapse phase . figures [ fig:1]_g _ and _ h _ show the density and velocity distributions at @xmath138 ( the protostar formation epoch ) . in our calculations ,
the first core ( i.e. , shock layer ) does not disappear until the calculation ends .
the second core ( i.e. , protostar ; @xcite ) forms inside the first core in figures [ fig:1]_g _ and _ h_. since both the magnetic field and angular velocity continue to increase in the second collapse phase , the first core is slightly flattened both by lorentz and centrifugal forces at the protostar formation epoch , as shown in figure [ fig:1]_h_. however , the protostar has an oblateness @xmath139 ( fig .
[ fig:2]_a _ ) at its formation epoch ( @xmath140 ) , and therefore , has a nearly spherical shape . the magnetic field and rotation period reach @xmath141 and @xmath142days at the protostar formation epoch .
the plasma beta around the protostar ( or inside the first core ) becomes @xmath143 ( fig .
[ fig:1]_i _ ) . after the protostar formation ( @xmath9 )
, the plasma beta becomes large ( @xmath144 ) because the thermal energy becomes large [ see eq . ( [ eq : eos ] ) ] .
the angular velocity normalized by the free - fall timescale , however , reaches @xmath145 at the protostar formation epoch ( fig .
[ fig:2]_d _ ) ; then , the cloud begins to rotate rapidly , and the rotation significantly affects protostar evolution .
figure [ fig:3 ] shows cloud evolution after the protostar formation epoch ( @xmath51 ) .
figures [ fig:3]_a__c _ show the protostar ( i.e. , second core ) surrounded by the shock layer as a dotted line .
figure [ fig:3]_a _ shows the density and velocity distribution around the protostar 170days after the protostar formation epoch .
the protostar has a mass of @xmath146 and a radius of @xmath147 at its formation epoch ( @xmath93 ) .
the arrows in figure [ fig:3]_a _ show that the gas around the protostar accretes spherically onto the protostar .
figure [ fig:3]_b _ shows the protostar 207days after its formation epoch . figures [ fig:3]_a _ and _ b _ show that the protostar becomes flat with time because the magnetic field and cloud rotation are amplified and affect protostar evolution , as shown in figures [ fig:2]_c _ and _ d_. two shocks ( outer and inner ) are seen in figure [ fig:3]_b_. the outer shocks are located at @xmath148au , while the inner shocks corresponding to the protostar are located at @xmath149au .
as shown in figure [ fig:3]_c _ , the accretion speed inside the outer shock ( @xmath150 in the region of @xmath151au ) is slower than that outside the outer shock ( @xmath152 in the region of @xmath153au ) .
this structure of the shock and distribution of the density and velocity , shown in figure [ fig:3]_b _ , is similar to the magnetic bubble " in @xcite .
he showed that the magnetic pressure is amplified by the twisted magnetic field lines , and a bubble - like structure is formed in the weakly magnetized collapsing cloud . in model sr
, the jet appears 215days after the protostar formation epoch .
figure [ fig:3]_c _ shows that the gas flows out from the central region along the vertical axis .
the red lines in figures [ fig:3]_b_-_f _ indicate the border between the inflow and the outflow . inside the red lines ,
the gas flows out from the central region ( @xmath154 for @xmath155 ) , and outside the redlines , it flows into the central region ( @xmath156 for @xmath155 ) .
the jet has a maximum speed of @xmath157 at the same epoch as in figure [ fig:3]_c_. the jet has an elongated structure and height - to - radius ratio of @xmath158 in figure [ fig:3]_c_. the height - to - radius ratio increases with time .
figures [ fig:3]_d _ and _ e _ show the jet extending in the vertical direction keeping almost the same horizontal scale ( @xmath159au in fig .
[ fig:3]_d _ and @xmath160au in fig .
[ fig:3]_e _ ) .
the jet extends up to @xmath161au with a maximum speed of @xmath162 at @xmath163 days ( fig .
[ fig:3]_d _ ) , while it extends up to @xmath164au with @xmath165 at @xmath166 days ( fig .
[ fig:3]_e _ ) , where @xmath167 is the elapsed time from the protostar formation epoch ( @xmath140 ) .
figure [ fig:3]_f _ shows the structure of the jet at the end of the calculation .
note the butterfly - like density distribution ( white - density contour ) , which is caused by the strong mass ejection from the protostar .
the jet has a well - collimated structure at this epoch .
the height - to - radius ratio of the jet reaches @xmath168 . in figure [ fig:3]_f _ , the jet penetrates the first core , shown as a thick broken line , and extends up to the outside of the first core . at the end of the calculation , the jet reaches @xmath169au with a maximum speed of @xmath170 .
figure [ fig:4 ] shows the shape of the jet ( purple iso - velocity surface ) and magnetic field lines ( stream lines ) at the same epoch as figure [ fig:3]_f _ but with a different grid size ( left : @xmath171 , right : @xmath172 ) . figure [ fig:4 ] left panel confirms that the jet has a well - collimated structure showing that the jet is strongly coiled by the magnetic field lines , with the bow shocks clearly visible near the upper and lower boundaries .
figure [ fig:3 ] ( right panel ) provides a close - up view of the left panel , indicating that the jet is driven from the protostar , which is shown by a red iso - density surface .
we can see ripping density contours on the projected wall caused by strong mass ejection @xcite . at the end of the calculation
, the protostar in model sr has a mass of @xmath173 and a radius of @xmath174 , while the mass flowing out from the protostar is @xmath175 . at the same epoch ,
the protostar and jet have the angular momenta of @xmath176 and @xmath177 , respectively .
therefore , the specific angular momenta of the protostar and jet are @xmath178 and @xmath179 , respectively .
this indicates that the jet largely removes the angular momentum from the protostar .
@xcite showed that , at its formation epoch , the protostar has a short rotation period compared with the observation .
the jet may be one of the mechanisms for the angular momentum transfer after protostar formation .
figure [ fig:5 ] shows the cloud evolution for model mr after first core formation .
model mr has a parameter of @xmath180 .
the ratio of the rotational to gravitational energy for model mr is @xmath181 , which is 100 times larger than that for model sr at its initial state ( see table [ table : init ] ) . in model
mr , the first core forms at @xmath101 , and has a mass of @xmath182 and a radius of @xmath183au .
figures [ fig:5]_a _ and [ fig:1]_a _ show that the first core of model mr is larger than that of model sr , because the centrifugal force for model mr is stronger . the angular velocity , normalized by free - fall timescale at the center of the cloud , saturates @xmath184 for @xmath185 ( fig .
[ fig:2]_d _ ) .
this indicates that cloud rotation significantly affects cloud evolution after the central density reaches @xmath101 .
due to the rotation ( or the centrifugal force ) , the first core for model sr , has a more oblate structure than that for model sr , as shown in figure [ fig:5]_b_. the evolution of the oblateness for model mr is almost the same as that for model sr for @xmath186 . on the other hand , the oblateness increases after the central density reaches @xmath135 , while it maintains @xmath117 for @xmath185 in model sr .
figure [ fig:5]_c _ shows the plasma beta around the first core at @xmath187 .
the plasma beta for model mr is as large as that for model sr ( @xmath188 ) at the first core formation epoch ( fig .
[ fig:5]_c _ ) .
thus , the magnetic field slightly affects cloud evolution inside the first core at this epoch .
figures [ fig:5]_d _ and _ e _ show the density and velocity distribution at @xmath189 ( the second collapse phase ) . the arrows in figure [ fig:5]_d _ show that the cloud collapses while rotating : however , it collapses only in the radial direction for model sr ( fig . [ fig:1 ] ) .
the azimuthal component of the velocity ( @xmath190 ) inside the first core is comparable to or larger than the radial one ( @xmath191 ) in model mr , while the azimuthal component of the velocity is much smaller than the radial one in model sr .
since the gas cloud inside the first core rotates rapidly , the central region increases its oblateness as the cloud collapses ( fig .
[ fig:2]_a _ ) .
therefore , the central region has a more oblate or disk - like structure , similar to the first core shown in @xcite , in which they investigated the equilibrium state of the rotating first cores . in general , since the angular momentum is transferred by magnetic braking in the magnetized clouds @xcite , the cloud rotation slows with time . in this model
, however , the cloud rotation hardly slows because the magnetic energy around the first core is too small to effectively work the magnetic braking .
the small energy of the magnetic field is realized by ohmic dissipation .
therefore , the cloud around the first core has a large plasma beta ( @xmath192 ; fig .
[ fig:5]_c _ ) . figure [ fig:2]_c _ compares the magnetic fields normalized by the density at the center of the cloud ( @xmath123 ) for model mr ( broken line ) , which are smaller than those for model sr ( solid line ) because for model mr , the cloud collapses , maintaining the disk - like structure , while that in model sr collapses spherically .
the growth rate of the magnetic field depends on the geometry of the collapse and is much larger for a spherical collapse than for a disk - like collapse @xcite . as a result
, the protostar has a magnetic field of @xmath193 for model mr which is five times smaller than that for model sr at its formation epoch ( @xmath194 ) .
figures [ fig:5]_g _ and _ h _ show the density and velocity distribution at @xmath195 ( the protostar formation epoch ) . in each panel , the protostar ( i.e. , the second core ) is represented by a red - dotted line . in model
mr , the protostar forms at @xmath196 , and has a mass of @xmath197 and a radius of @xmath198 . as shown in figure [ fig:5]_h _ , at the protostar formation epoch , the first core sags downward in the center in a concave structure similar to model c in @xcite .
this concave structure is considered to be caused by rapid rotation and an increase in the density of the first core ( for details , see * ? ? ?
as shown in figure [ fig:5]_h _ , since the gas density falls below @xmath199 in the regions just above and below the first core at the center of the cloud ( @xmath200au ) , ohmic dissipation is less effective in these regions .
note that ohmic dissipation is effective for @xmath121 @xcite .
therefore , the magnetic field is strong around the protostar ( @xmath201 ) .
since the central region becomes increasingly thinner and has an increasingly concave structure over time , the magnetic field around the protostar barely dissipates through ohmic dissipation .
therefore in model mr , the torsional alfvn wave generated by the rotation of the protostar can be transferred outside without dissipation . in model sr , however , the magnetic field dissipates and the plasma beta maintains a high value ( @xmath202 ) around the protostar , because the cloud collapses spherically with slow rotation , and ohmic dissipation is effective around the protostar .
figure [ fig:5]_h _ shows that the outflows are driven near the first core .
the region where the outflow appears in figure [ fig:5]_h _ corresponds to the low - plasma beta region ( @xmath203 ) shown in figure [ fig:5]_i_. figure [ fig:6 ] shows the evolution of outflow for model mr .
figures [ fig:6]_a__c _ show the density and velocity distribution around the first core on the @xmath105 plane .
the gas accretes onto the first core ( white dotted lines ) in the whole region outside the first core ( fig .
[ fig:6]_a _ ) , while the gas in the regions above and below the first core ( inside the red lines ) outflows from the central region ( fig .
[ fig:6]_b _ ) . in the ideal mhd regime
, the outflow in the collapsing cloud appears in close proximity to the first core @xcite .
however , as shown in figure [ fig:6]_b _ , the outflow appears far from the first core in the non - ideal mhd regime because the magnetic field near the first core is not well coupled with the neutral gas due to ohmic dissipation .
thus , the torsional alfvn wave caused by the rotation of the first core is not effectively transferred .
however , the torsional alfvn wave generated far from the first core can maintain itself , because the gas density is low and ohmic dissipation is less effective in these regions .
as the cloud collapses , the outflow can be driven even near the first core , as shown in figure [ fig:6]_c _ , because in these regions , the gas density gradually decreases and ohmic dissipation becomes less effective over time , as shown in figures [ fig:5]_b _ , _ e _ , and _
finally , the outflow is anchored by the first core , as seen in the ideal mhd calculations . at the end of the calculation
, the outflow has a maximum speed of @xmath204 , which is comparable to the results of ideal mhd calculations @xcite .
therefore , the outflow in our calculation ( the non - ideal mhd calculation ) differs from that of the ideal mhd calculation in the early phase , while the outflow features are almost the same ( e.g. , the shape and speed of the outflow ) for both the non - ideal and ideal mhd calculations in the later phase . figure [ fig:7 ] shows the evolution of a jet driven from the protostar . in this figure , the first core and protostar ( i.e. , the second core )
are shown as dotted white and red lines , respectively .
the protostar has a disk - like structure , while the first core has a concave shape ( fig .
[ fig:7]_a _ ) .
the gas flows out from the regions above and below the protostar ( fig .
[ fig:7]_b _ ) . figure [ fig:7]_b _ shows horn - like structures at @xmath205au , which are caused by strong mass ejection .
this panel shows that only the gas located in the regions just above and below the protostars flows out from the protostar , while a large fraction of the gas accretes onto the protostar . at the end of the calculation , a strong jet with a maximum velocity of @xmath206 is driven from the protostar , as shown in figure [ fig:7]_c_. figures [ fig:8]_a__c _ ( upper panels ) show the density ( color and contours ) and velocity distributions ( arrows ) around the protostar for model mr .
the velocity of @xmath57-component ( @xmath207 ; color and contours ) and magnetic field ( arrows ) corresponding to each upper panel of figure [ fig:8 ] are plotted in figures [ fig:8]_e__f _ ( lower panels ) . the first core and protostar ( i.e. , second core )
are plotted as dotted white and red lines , respectively .
the outflow extends from the region just above the first core to the outside of the @xmath208 grid boundary ( fig .
[ fig:8]_a _ ) . in model
mr , the outflow extends to 7au at the end of the calculation .
there are two velocity peaks in figure [ fig:8]_d_. the upper one ( peak 1 ) has a velocity of @xmath209 at @xmath210au , 5au ) , while the lower ( peak 2 ) has a velocity of @xmath211 at @xmath212au , 1au ) .
figures [ fig:8]_b _ and _ e _ provide close - up views of figures [ fig:8]_a _ and _ d_. in figure [ fig:8]_e _ , strong flows ( peak 3 ) appear near the surface of the first core [ ( @xmath213 ( @xmath214 au , 0.3@xmath215au ) ] .
figures [ fig:8]_c _ and _ f _ show close - up views of figures [ fig:8]_b _ and _ e_. a strong jet is driven from the region above the protostar in figure [ fig:8]_c _ , corresponding to peak 4 in figure [ fig:8]_f_. this peak ( peak 4 ) has a velocity of @xmath216 at @xmath217au , 0.08au ) .
the lower panels of figure [ fig:8 ] show four velocity peaks ( peak 1peak 4 ) .
these peaks imply that speeds ( strength ) of the outflow and jet change with time , because they represent a history of the outflow and jet .
we confirmed that peaks 1 and 2 appear before protostar formation , and peaks 3 and 4 appear after protostar formation .
we also confirmed that each peak is related to the oscillation of each core .
for example , peaks 1 and 2 appear every time the first core oscillates .
it is , therefore , considered that the outer peaks ( peaks 1 and 2 ) originate from the first core , while the inner peaks ( peaks 3 and 4 ) originate from the protostar .
thus , in our definition , the outer peaks correspond to the outflow , while the inner peaks correspond to the jet .
figure [ fig:9 ] shows the structure of the outflow and jet , and the configuration of the magnetic field lines .
it also shows the shapes of the first core ( left panel ; the projected density contours on the wall ) and the protostar ( right panel ; the red isosurface ) .
the purple and blue surfaces in figure [ fig:9 ] indicate the iso - velocity surface of @xmath218 and @xmath219 , respectively .
the flow inside the purple iso - velocity surface has a velocity of @xmath220 ( low - velocity component ; lvc ) , while the flow inside the blue iso - velocity surface has a velocity of @xmath221 ( high - velocity component ; hvc ) .
the hvc is enclosed by the lvc .
the lvc flow is mainly driven from the first core , while the hvc flow is mainly driven from the protostar .
the lvc and hvc are strongly coiled by the magnetic field lines anchored to the first core and protostar , respectively . in model
mr , the protostar has a mass of @xmath222 and a radius of @xmath174 , while at the end of the calculation , the mass flowing out from the protostar is @xmath223 . at the same epoch ,
the protostar and outflow gas ( outflow and jet ) have angular momenta of @xmath224 and @xmath225 , respectively .
therefore , the specific angular momenta of the protostar and outflow gas ( i.e. , the outflow and jet ) are @xmath226 and @xmath227 , respectively .
figure [ fig:10 ] shows the cloud evolution for model fr after the first core formation .
model fr has a parameter of @xmath228 .
the ratio of the rotational to gravitational energy for model fr is @xmath229 , which is @xmath230 times larger than that for the initial stage of model sr(table [ table : init ] ) .
the first core forms at @xmath231 , and has a mass of @xmath232 and a radius of @xmath233au at its formation epoch .
the mass and size of the first core for model fr are larger than those for models sr and mr because the cloud is rotating rapidly .
the cloud in model fr has @xmath145 in the initial stage , and maintains this value for @xmath199 , as shown in figure [ fig:2]_d _ , which indicates that the centrifugal force significantly affects the formation of the first core . owing to rapid rotation
, the cloud has a disk - like structure near the center , as shown in figure [ fig:10]_a_. the oblateness of model fr increases rapidly , as shown by the dotted line in figure [ fig:2]_a _ , and saturates @xmath234 for @xmath235 . in this model , the first core has a disk - like shape at its formation epoch .
then , the disk - like first core deforms to a ring , and the ring fragments into several pieces at @xmath236 .
@xcite showed that fragmentation occurs in the rapidly rotating cloud .
since we focus on the mechanism of the outflow and jet , we stopped the calculation when fragmentation occurred .
we will discuss the fragmentation process of the first and second cores in a subsequent paper .
figure [ fig:10]_b _ shows the density and velocity distribution around the center of the cloud 330.3yr after the first core formation .
this panel shows that outflow is driven from the center of the cloud .
the outflow extends up to 120au with a maximum speed of @xmath237 at this epoch .
figure [ fig:10]_c _ shows the structure of the outflow 476.1yr after the first core formation . at the end of the calculation , the outflow extends up to 200au with a maximum speed of @xmath238 .
figures [ fig:10]_b _ , [ fig:3 ] , and [ fig:8 ] show that the opening angle of the outflow for model fr is larger than that of the jet for models sr and mr .
the height - to - radius ratio of the outflow is @xmath239 ( fig .
[ fig:10]_b _ ) to @xmath240 ( fig .
[ fig:10]_c _ ) in model fr , while that of the jet is @xmath168 in model sr . in model fr ,
the outflow maintains its shape ( i.e. , @xmath241 ) regardless of time , while the jet in model sr becomes slender along the vertical axis with time ( i.e. , @xmath242 increases with time ) .
figure [ fig:11 ] shows the density and velocity distribution on the @xmath105 ( upper panel ) and @xmath104 ( lower panel ) planes at the same epoch as figure [ fig:10]_c _ , but with different grid scales .
figure [ fig:11]_c _ is 16 times magnification of figure [ fig:11]a and shows the ripping density contours caused by the strong mass ejection from the center of the cloud . since the outflow interrupts gas accretion in the regions above and below the central core , the gas accretes onto the central region only via a thin disk ( fig . [ fig:11]_a _ ) .
the upper panel in figure [ fig:11]_b _ shows that the outflow is driven from the first core ( dotted line ) . at this epoch
, the first core , which has a disk - like shape at the formation epoch , has deformed to a ring - like shape ( fig .
[ fig:11]_b _ lower panel ) .
as shown in figure [ fig:11]_c _ , the density of the first core exceeds @xmath8 , while the density of the ambient gas outside the first core is @xmath199 .
since ohmic dissipation is effective for @xmath121 @xcite , the dissipation of the magnetic field becomes effective inside the first core , while it is barely dissipated by ohmic dissipation outside the first core .
therefore , the magnetic field lines can be twisted just outside the first core , and the torsional alfvn wave transfers to the outside , as shown in the ideal mhd calculations @xcite .
figure [ fig:12 ] shows the shape of the outflow ( blue iso - velocity surface ) and the configuration of the magnetic field lines ( streamlines ) at the same epoch as figure [ fig:11 ] .
this figure shows that the magnetic field lines are strongly twisted inside the outflow , while they are loosely twisted outside the outflow . due to rapid rotation , the first core forms at a lower density for model fr ( @xmath243 ) than for models sr and mr ( @xmath101 ) ; therefore , ohmic dissipation is less effective outside the first core in model fr .
the magnetic field lines anchored to the rapidly rotating first core are effectively twisted , and the lorentz force can drive the strong outflow near the first core .
we stopped the calculation for model fr at @xmath135 , because fragmentation occurred .
thus , we can not determine further cloud evolution for model fr .
we expect that each fragment has a small spin angular momentum because of redistribution of the angular momentum ( spin and orbital angular momenta ) for fragmentation @xcite , and the magnetic field is dissipated by ohmic dissipation in each fragment .
thus , we expect that each fragment traces a similar evolution to models sr and mr . in model fr
, the first core has a mass of @xmath244 and a radius of @xmath245au , while the mass flowing out from the first core is @xmath246 at the end of the calculation . at the same epoch ,
the first core and outflow have the angular momenta of @xmath247 and @xmath248 , respectively .
therefore , the specific angular momenta of the first core and outflow are @xmath249 , and @xmath250 , respectively .
thus , in model fr , the outflow largely removes the angular momentum from the first core . in the previous sections
, we showed the evolutions of three different clouds , while we investigated the cloud evolutions of 80 models in total .
as shown in @xcite , cloud evolution depends only on the initial angular velocity in the magnetic - force dominant clouds @xmath251 , where @xmath56 , @xmath63 , and @xmath252 are the initial angular velocity , magnetic flux density , and sound speed , respectively .
our choice of these strongly magnetized clouds is supported by observations @xcite . from our calculations
, we found the emergence conditions for the outflow and the jet fall into three categories , depending on the initial rotation rate @xmath60 : 1 .
@xmath253 , both outflow and jet can be driven from each core , 2 .
@xmath254 , only a jet can be driven from the protostar , and 3 .
@xmath255 , neither outflow nor jet can be driven from each core .
the outflow always appears when condition ( 1 ) is satisfied .
however , the jet does not appear in some models even when condition ( 1 ) or ( 2 ) is satisfied .
thus , condition ( 1 ) is the necessary and sufficient condition for driving outflow , while condition ( 1 ) and ( 2 ) is the necessary condition for driving jet .
for example , although 24 models satisfy condition ( 2 ) , the jet appeared only in nine models .
although we followed the evolution of the accreting protostar up to @xmath6200 cpu hours for each model , we did not observe any sign of the jet in those nine models .
after the protostar is formed , the magnetic field becomes strong in the regions above the poles of the protostar ; thus , the plasma beta becomes @xmath256 in these regions .
therefore , it is difficult to calculate for a long period after protostar formation because the time step becomes significantly short owing to the increased alfvn speed
. therefore , we could not determine the reason the jet does not appear in some models .
we could not determine if the jet would never appear in those models or if it might appear much later . to clarify the long - term evolutions
, we need to develop a calculation method with implicit time integration .
in this paper , for convenience , we called the flows driven from the first core the outflow , and the flows driven from the protostar ( i.e. , the second core ) the jet . in model sr , no outflow appears because the initial cloud rotates slowly , and the rotation rate is not sufficiently amplified at the first core formation epoch . on the other hand , we can not observe the jet driven from the protostar in model fr because we stopped calculation once it showed fragmentation . in model mr , both the outflow and the jet appear . in this section , to investigate the driving mechanism of the outflow and the jet , we focus on the flows that appeared in models sr ( the jet ) and fr ( the outflow ) .
we do not discuss the flows in model mr because it is difficult to separate the jet from the outflow as shown in figures [ fig:6][fig:9 ] . at the end of the calculation ,
the jet has a well - collimated structure with high speed ( @xmath257 ) and a large height - to - radius ratio of @xmath258 .
the @xmath259 ratio of the jet increases with time , as shown in 3.1.2 and 3.2.2 .
the outflow , however , has a wide opening angle with slow speed ( @xmath260 ) , and expands outwardly , maintaining the height - to - radius ratio of @xmath261 .
the difference of the speeds between the jet and outflow can be understood from the difference in the gravitational potential . in model sr
, the protostar has a mass of @xmath262 and is a radius of @xmath263 at the end of the calculation .
the kepler speed corresponding to the mass and radius is @xmath264 , which is comparable to the speed of the jet in model sr ( @xmath257 ) . on the other hand , at the end of the calculation for model fr
, the first core has a mass of @xmath265 and a radius of @xmath266au .
thus , the kepler speed of the first core is @xmath267 , which is comparable to the speed of the outflow ( @xmath260 ) . as a result ,
both the jet and the outflow have flow speeds similar to the kepler speed of their respective cores .
the difference in the degree of collimation between the outflow and the jet ( the well - collimated structure of the jet and the wide opening angle of the outflow ) can be understood from the difference in their driving mechanisms and the configuration of the magnetic field lines around their drivers ( i.e. , the fist core and protostar ) .
figure [ fig:13 ] shows the shapes of the jet ( model sr ; left panels ) and outflow ( model fr ; outflows ) with thick red lines indicating the border between the accretion and the jet or outflow .
this shows that the collimation of the jet ( left panel ) is stronger than that of the outflow ( right panel ) .
first , we discuss the collimation and driving mechanism of the jet .
the upper panel of figure [ fig:13 ] show the ratio of the toroidal to poloidal components of the magnetic field ( @xmath268 ) at each mesh point around the outflow ( model sr ; left ) and the jet ( model mr ; right ) . in figure
[ fig:13]_a _ , the toroidal field dominates the poloidal field inside the jet ( @xmath269 ) , while the toroidal field barely exists outside the jet ( @xmath270 ) . in model sr , the magnetic field is dissipated by ohmic dissipation , and decoupled from the neutral gas inside the first core ( or outside the second core ) , as shown in figure [ fig:1]_i_. thus , the magnetic field lines are relaxed by the magnetic tension force and are almost straight ( @xmath271 in the cylindrical coordinates ) , because magnetic field lines move freely , irrespective of the neutral gas .
even after the magnetic field is well - coupled with the neutral gas for @xmath136 , the straight magnetic field lines are distributed around the protostar , which reflects past ohmic dissipation . therefore , the vertical component of the magnetic field ( @xmath272 ) is dominant outside the protostar .
the lower panels in figure [ fig:13 ] show the plasma beta around the jet ( left ; model mr ) and the outflow ( right ; model fr ) .
figure [ fig:13]_b _ shows that the magnetic field is weak ( @xmath273 ) outside the jet ( outside the red line ) , while it is very strong ( @xmath274 ) inside the jet ( inside the red line ) , compared to the thermal pressure .
the sharing motion between the protostar and ambient gas amplifies the magnetic field inside the jet , since the formed protostar rapidly rotates because the magnetic braking is less effective , explained in 3.1 . in general , when the magnetic field is weak around the driver , it is considered that the magnetic pressure gradient force dominates the magnetocentrifugal force in terms of driving the flow @xcite . to investigate the driving mechanism of the jet and the outflow , we calculate the lorentz ( @xmath275 ) , centrifugal ( @xmath276 ) , and thermal pressure gradient ( @xmath277 ) forces parallel to the poloidal component of the magnetic field ( @xmath278 ) . when the self - gravity term in equation ( [ eq : eos ] ) is ignored , the equation of motion can be rewritten as @xmath279 .
\label{eq : eom2}\ ] ] we consider the dot product between each term in the left - hand side in the equation ( [ eq : eom2 ] ) and the poloidal component of the magnetic field ( @xmath280 ) .
thus , each force parallel to the @xmath280 is written as @xmath281 where @xmath282 . in the left panels of figure [ fig:14 ] , we plot the ratio of the lorentz to the centrifugal force ( @xmath283 ; fig .
[ fig:14]_a _ ) , and the ratio of the lorentz to the thermal pressure gradient force ( @xmath284 ; fig .
[ fig:14]_b _ ) in the outflow region at each mesh point , where the outflow region is defined as meshes with @xmath285 for @xmath286 . the right panels ( fig .
[ fig:14]_c _ and _ d _ ) show the plots for the same ratios of @xmath283 , and @xmath284 in the accreting region ( @xmath287 for @xmath286 ) . in figure
[ fig:14 ] , red dots and lines are data from the region around the jet in model sr , while black dots and lines are data from the region around the outflow in model fr . in figure
[ fig:14 ] , left panels show that , inside the jet ( red dots and line ) , the lorentz force is equivalent to the centrifugal force ( @xmath288 ; fig .
[ fig:14]_a _ ) , while the lorentz force is about @xmath289 times stronger than the thermal pressure gradient force ( @xmath290 ; fig .
[ fig:14]_b _ ) . in the accreting region around the jet ( red dots and lines in fig .
[ fig:14]c and d ) , the lorentz force is much weaker than both the centrifugal ( @xmath291 ) and thermal pressure gradient forces ( @xmath292 ) .
thus , the lorentz force is not effective in the accreting regions . in model sr ,
a rapidly rotating protostar forms because the magnetic field around the center of the cloud is weak and magnetic braking is less effective .
thus , the magnetic field lines anchored to the protostar make a strong toroidal field for the rapid rotation of the protostar .
the toroidal field caused by the protostar transmits in the vertical direction as the torsional alfvn wave , generating a large magnetic pressure gradient along the vertical axis .
the jet is considered to be driven by this strong magnetic pressure gradient force . in general , the jet has good collimation when the magnetic pressure gradient force is dominant ( @xmath293 ) for driving the flow , because the magnetic field lines are pinched by the toroidal field , as shown by @xcite .
furthermore , the jet in model sr is guided by the straight configuration of the magnetic field lines ( @xmath294 ) outside the jet region .
thus , the well - collimated jet is caused by both the driving mechanism of the magnetic pressure gradient and the straight configuration of the magnetic field lines around the protostar .
figure [ fig:13 ] shows the ratio of the toroidal to poloidal components of the magnetic field ( @xmath268 ; fig .
[ fig:13]_c _ ) and the plasma beta ( @xmath295 ; fig .
[ fig:13]_d _ ) around the outflow for model fr .
the toroidal field is @xmath296 times stronger than the poloidal field ( @xmath297 ) inside the outflow ( inside the red line ) .
however , this ratio is smaller than that of the jet ( @xmath269 ; model sr ) because the rotation rate in model fr , which causes the toroidal field , is small around the first core , due to effective magnetic braking . outside the outflow ,
however , the poloidal field is slightly larger than the toroidal field ( @xmath298 ) . figure [ fig:13]_c _ shows that the magnetic energy is larger than the thermal energy inside the outflow ( @xmath274 ) , while the magnetic energy is comparable to or slightly weaker than the thermal energy outside the outflow ( @xmath299 ) .
the gas around the first core has a density range of @xmath300 in model sr ( fig .
[ fig:10 ] ) .
since ohmic dissipation is ineffective in this region , the magnetic energy becomes comparable to the thermal energy as demonstrated by @xcite .
after the first core formation , the toroidal field can be amplified because the rotational timescale is shorter than the collapsing timescale inside the first core , and the magnetic field lines begin to twist @xcite .
therefore , the magnetic energy inside the outflow ( or near the first core ) becomes larger than that outside the outflow ( or far from the first core ) .
as shown in figures [ fig:14]_a _ and _ b _ , the centrifugal force is dominant over the lorentz and thermal pressure gradient forces inside the outflow ( @xmath301 , and @xmath302 1 ; black dots and lines ) , while it is comparable to the lorentz force inside the jet ( red dots and lines ) . outside the outflow ,
both the centrifugal and thermal pressure gradient forces are stronger than the lorentz force ( @xmath303 ) , since @xmath304 ( fig . [ fig:14]_c _ ) and @xmath305 ( fig .
[ fig:14]_d _ ) .
these results indicate that the outflow in model sr is mainly driven by the magnetocentrifugal mechanism @xcite , which causes the loosely pinched magnetic field lines inside the outflow .
in addition , before the outflow appears , the magnetic field lines around the first core expand sideways toward the rotation axis from the center of the cloud , whose shape looks like a capital letter u or v , as shown by @xcite .
thus , the collimation of the magnetic field lines becomes worse with distance from the center , as shown in figure 8_a _ of @xcite .
since the outflow traces this configuration of the magnetic field lines , it does not have good collimation and the outflow driven from the first core has a wide opening angle because of the magnetocentrifugal driving mechanism and the configuration of the magnetic field lines .
the former realizes the loosely pinched magnetic field lines , and the latter swells the outflow . in this paper , we investigated cloud evolution with same magnetic field strengths ( @xmath306 g ) but different rotation rates [ @xmath307s@xmath22 ( model sr ) , @xmath308s@xmath22 ( model mr ) , and @xmath309s@xmath22 ( model fr ) ] .
as shown in previous section , the speeds of the jets and outflows depend on the mass and size of their drivers .
the mass and size of the first and second cores at their formation epochs depend on the initial rotation rate . therefore , the speeds of the jet and outflow depend on the initial rotation rate of the cloud core . for example , in a rapidly rotating clouds , the first core forms at lower density with larger size .
when the first core has a large size and low density , the outflow is slow because the gravitational potential is shallow .
on the other hand , a slowly rotating cloud produces a first core with smaller size and higher density . in this case
, there is high - speed outflow , owing to the deep gravitational potential of the first core .
thus , the outflow has higher speed in a more slowly ( rapidly ) rotating cloud .
in addition , a cloud with an extremely low rotation rate produces no outflow because the first core has such a small angular momentum at its formation epoch that the magnetic field lines are barely twisted .
the speed of the jet , however , depends on the initial rotation rate only slightly because the angular momentum around the center of the cloud converges to a certain value by the time the protostar ( i.e. , second core ) is formed @xcite .
thus , the variation caused by the rotation of initial cloud is not a factor affecting the time of the protostar formation epoch .
note that the properties of the jet and outflow might possibly change much later because we followed the evolution of these flows for only a short period .
does the difference of magnetic field strength in the initial cloud affect the properties of the jet and outflow ?
@xcite showed that the evolution of the molecular cloud depends only on the ratio of the angular velocity to the magnetic flux density of the initial cloud ( @xmath310 ) .
initial differences in the distribution of the density , velocity , and magnetic field hardly affect cloud evolution , especially once the central region becomes adiabatic ( @xmath97 ) .
cloud evolution is controlled by the magnetic field when @xmath311 , while it is controlled by the rotation when @xmath312 @xcite .
although we can choose arbitrary parameter of @xmath313 and @xmath314 , observations indicate that the magnetic energy is much larger than the rotation energy in many molecular clouds @xcite , and clouds have @xmath312 ( i.e. , the magnetic - force dominant clouds in @xcite ) . in magnetic - force dominant clouds
, the magnetic field converges to a specific value in the isothermal phase ( the magnetic flux - spin relation ; @xcite ) .
thus , cloud evolution does not depend much on the initial magnetic field strength . as a result
, outflows and jets with similar properties may appear in these clouds when they have the same angular velocities at their initial state . in this paper
, we have shown that the outflow is driven from the first core .
after the first core formation epoch , the collapse timescale becomes longer than the rotational timescale , the magnetic field lines are strongly twisted for the rotating first core , and outflow appears as shown in 3.2.2 .
observations have shown that outflows have lifetimes of @xmath315yr @xcite . however , since we followed cloud evolution only through the very early phase , we could not determine whether outflow continues to be driven for a long time .
the spherical symmetric calculation indicates that the first core disappears @xmath316 yr after its formation epoch . in model
fr , we followed cloud evolution for @xmath317yr after the first core is formed . in this model ,
the first core remains until the end of the calculation .
the outflow that appears just after the first core formation epoch continues to be driven from the first core .
the rotating first cores seem to have lifetimes of several thousand years or more ( e.g. , * ? ? ?
thus , outflows driven from these cores may maintain themselves for several thousand years .
note that it is expected that a jet will continue to be driven from the protostar once it appears because the protostar does not disappear .
after the first core disappeared , we could not determine whether the outflow continues to be driven .
@xcite suggested that after the first core disappears , the remnant of the first core forms a disk - like structure or torus .
outflow may continue to be driven from these objects even after the disappearance of the first core .
however , if the outflow does not continue after the first core disappears , then outflow , as shown in this paper , might be a transient phenomena , in which the outflow is driven only for @xmath316yr . in this case
, the flow from the first core , which we call outflow in this paper , might be a special phenomenon that would be observed only in limited numbers of young stellar objects . to understand the duration of the outflow driven from the first core , further long - term calculations , possibly with an implicit code , are necessary . at present
we conjecture that the outflow in our calculations will be driven for @xmath318 years .
observation shows that the outflows have wide opening angles and low flow speeds , while jets have good collimation and high flow speeds .
outflow has been considered to be entrained by the jet driven from a circumstellar disk around the protostar . in this paper , we calculated the cloud evolution from the molecular cloud core to protostar formation . as a result of calculations
, we found that two distinct flows ( outflow and jet ) are driven from different objects ( the first , and second cores ) , and the features of outflow and jet ( flow speed and collimation factor ) were naturally reproduced .
our results show that the flow appearing around the first core has a wide opening angle and slow speed , while the jet appearing around the protostar has a well - collimated structure and high speed , as shown in figure [ fig:15 ] .
the speed difference is caused by the difference of the depth in the gravitational potential .
the flow speed ( i.e. , the speed of the outflow and jet ) corresponds to the kepler speed of each object ( the first and the second cores ) .
because the first core has a shallow gravitational potential , its flow ( outflow ) is slower .
the flow ( jet ) driven from the protostar , which has a deeper gravitational potential , has a high speed . in our calculations ,
the outflow and the jet have speeds of @xmath319 and @xmath257 , respectively .
these speeds are slower than those of observations . typically , observed molecular outflow and optical jet have speeds of @xmath320 , and @xmath321 , respectively .
however , since the first and second cores ( protostar ) have mass of @xmath322 and @xmath323 , respectively , at the end of the calculations , each core increases its mass in the subsequent gas accretion phase .
the kepler speed increases with the square root of the mass of the central object at a fixed radius .
when the mass of each core increases by 100 times , the kepler speed increases 10 times .
thus , the speed of the outflow and jet may increase by 10 times , and reach @xmath324 and @xmath325 , respectively , which correspond to typical observed values .
the outflow has a wide opening angle , while the jet has a well - collimated structure .
this is caused both by the configuration of the magnetic field lines around the drivers and their driving mechanisms .
the magnetic field lines around the first core have an hourglass configuration because they converge to the cloud center as the cloud collapses , and ohmic dissipation is ineffective before the first core formation .
the flow appearing near the first core is mainly driven by the magnetocentrifugal wind mechanism ( disk wind ) .
the centrifugal force is dominant in the outflow , whereas near the protostar , the magnetic field lines are a straight , and the magnetic pressure gradient mechanism is more effective for driving the jet .
the magnetic field lines are stretched by the magnetic tension force near the protostar because the magnetic field is decoupled from the neutral gas . however , the magnetic field lines are strongly twisted in the region in close proximity to the protostar , where the magnetic field is coupled with the neutral gas again .
thus , the strong toroidal field generated around the protostar can drive the jet , which is guided by the straight configuration of the magnetic field .
our calculations do not completely reject the well - known concept that the observed molecular outflow is entrained by the jet , because we calculate the formation of the jet and outflow only in the early star - formation phase .
further long - term calculations are needed to understand the mechanism of the jet and outflow in more detail .
we have greatly benefited from the discussion with t. nakano and k. saigo .
we also thank t. hanawa for making a contribution to the nested grid code .
numerical calculations were carried out with a fujitsu vpp5000 at the astronomical data analysis center , the national astronomical observatory of japan .
this work is supported by the grant - in - aid for the 21st century coe `` center for diversity and universality in physics '' from the ministry of education , culture , sports , science and technology ( mext ) of japan , and partially supported by the grants - in - aid from mext ( 15740118 , 16077202 , 18740113 , 18740104 ) .
arce , h. g. , shepherd , d. , gueth , f. , lee , c .- f .
, bachiller , r. , rosen , a. , & beuther , h. , 2005 , in protostars and planets v , ed .
b. reipurth , d. jewitt , & k. keil ( tucson : univ .
arizona press ) , in press bouvier , j. , alencar , s. h. p. , harries , t. j. , johns - krull , c. m. , & romanova , m. m. 2006 , in protostars and planets v , ed .
b. reipurth , d. jewitt , & k. keil ( tucson : univ .
arizona press ) , in press ( astro - ph/0603498 ) johns - krull , c. m. , valenti , j. a. , piskunov , n. e. , saar , s. h. , & hatzes , a. p. 2001 , in asp conf .
248 , magnetic fields across the hertzsprung - russell diagram , ed .
g. mathys , s. k. solanki , & d. t. wickramasinghe ( san francisco : asp ) , 527 plane ( _ a _ , _ d _ , and _ g _ ) .
middle panels : density ( color - scale ) and velocity distribution ( arrows ) on the cross - section in the @xmath105 plane ( _ b _ , _ e _ , and _
lower panels : plasma beta ( color - scale and contours ) and magnetic field ( arrows ) on the cross - section in the @xmath105 plane ( _ c _ , _ f _ , and _
. panels from left to right are snapshots at the stages of @xmath326 ( @xmath327 , panels _ a _ , _ b _ , and _
c _ ) , @xmath328 ( @xmath327 ; panels _ d _ , _ e _ , and _
f _ ) , and @xmath329 ( @xmath327 ; panels _ g _ , _ h _ , and _ i _ ) , respectively , where @xmath76 denotes the level of subgrid .
white - dotted lines in upper and middle panels show the shocked regions , which indicate the first core . the elapsed time ,
density at the center of the cloud and arrow scale are denoted in each panel .
, width=604 ] ) within @xmath330 , ( b ) thick lines : magnetic flux density ( @xmath122 ; left axis ) .
thin lines : rotation period ( @xmath34 ; right axis ) , ( c ) magnetic field normalized by the square root of the density ( @xmath123 ) , and ( d ) : angular velocity normalized by the free - fall timescale [ @xmath331 at the center of the cloud against the central number density ( @xmath4 ) for models sr , mr , and fr .
, width=604 ] plane are plotted in each panel .
panels ( a ) through ( f ) are snapshots at the stages of ( a ) @xmath332days ( @xmath333 ) , ( b ) 207days ( @xmath333 ) , ( c ) 221days ( @xmath334 ) , ( d ) 229days ( @xmath335 ) , ( e ) 236days ( @xmath336 ) , and ( f ) 265days ( @xmath171 ) , where @xmath167 denotes the elapsed time after protostar formation ( @xmath140 ) .
red thick lines mean the border between the jet and the accretion flow ( i.e. , contour of @xmath207 = 0 ) .
white dotted lines in panels _
a_-_c _ indicate the second core ( i.e. , protostar ) , while white dotted line in panel _
f _ indicate the first core .
the elapsed time after protostar formation ( @xmath167 ) , the elapsed time from the initial ( @xmath337 ) , gas density at the center of the cloud ( @xmath4 ) , arrow scale , and grid scale are denoted in each panel .
the level of the subgrid is shown in the upper left corner .
, width=566 ] _ f _ , but with different scales of @xmath338au ( left ) , and @xmath339au ( right ) , where @xmath340 means the grid scale .
the structure of high - density region ( @xmath341 ; red iso - density surface ) , and magnetic field lines ( black - and - white streamlines ) are plotted in each panel .
the structure of the jet is shown by purple iso - velocity surface in which the gas is outflowing from the center .
the density contours ( false color and contour lines ) , velocity vectors ( thin arrows ) on the mid - plane of @xmath342 , @xmath105 , and @xmath3430 are , respectively , projected in each wall surface . ,
width=604 ] but for model mr .
panels from left to right are snapshots at the stages of @xmath344 ( @xmath327 , panels _ a _ , _ b _ , and _ c _ ) , @xmath345 ( @xmath327 ; panels _ d _ , _ e _ , and _
f _ ) , and @xmath346 ( @xmath327 ; panels _ g _ , _ h _ , and _ i _ ) , respectively .
the white - dotted lines in upper and middle panels mean the first core , while the red - dotted lines in panels _
g _ and _ h _ indicate the protostar .
, width=604 ] plane are plotted in each panel .
panels ( a ) through ( c ) are snapshots at the stages of ( a ) @xmath347 ( @xmath348 ) , ( b ) @xmath349 ( @xmath348 ) , and ( c ) @xmath350 ( @xmath348 ) .
white - dotted lines indicate the first core .
red - solid lines indicate the border between the outflow and accretion flow ( i.e. , contour of @xmath207 = 0 ) .
the elapsed time from the initial ( @xmath337 ) , density at the center of the cloud ( @xmath4 ) , arrow scale are denoted in each panel .
, width=604 ] plane are plotted in each panel .
panels ( a ) through ( c ) are snapshots at the stages of ( a ) @xmath351 ( @xmath348 ) , ( b ) @xmath352 ( @xmath348 ) , and ( c ) @xmath350 ( @xmath348 ) .
the first core and second core ( or protostar ) are shown by white- and red - dotted lines , respectively .
red - thick lines indicate the border between the jet and accretion flow ( i.e. , contour of @xmath207 = 0 ) .
the elapsed time from the initial ( @xmath337 ) , density at the center of the cloud ( @xmath4 ) , arrow scale are denoted in each panel .
, width=604 ] plane for model mr at the same epoch as fig .
[ fig:6 ] _ c_. lower panels : the vertical component of the velocity ( @xmath207 ; color - scale and contours ) and magnetic field ( arrows ) with the same scale as each upper panel .
the characters of ` peak ' indicate the position where @xmath207 has a local peak . the first core and second core ( or protostar )
are shown by the white- and red - dotted lines , respectively .
red - thick lines indicate the border between the outflow and accretion flow ( i.e. , contour of @xmath207 = 0 ) . , width=604 ] _ c _ and fig . [ fig:7]_c _ , but with different scales of @xmath353au ( left ) , and @xmath354au ( right ) .
the structure of high - density region ( @xmath341 ; red iso - density surface ) , and magnetic field lines ( black - and - white streamlines ) are plotted in each panel .
the structures of the jet ( @xmath355 ) and outflow ( @xmath356 ) are shown by purple and blue iso - velocity surfaces , respectively .
the density contours ( false color and contour lines ) , velocity vectors ( thin arrows ) on the mid - plane of @xmath342 , @xmath105 , and @xmath3430 are , respectively , projected in each wall surface . , width=604 ] plane are plotted .
panels ( a ) through ( c ) are snapshots at the stages of ( a ) @xmath357 ( @xmath358 ) , ( b ) @xmath359 ( @xmath360 ) , and ( c ) @xmath361 ( @xmath362 ) .
red - thick lines indicate the border between the outflow and accretion flow ( i.e. , contour of @xmath207 = 0 ) .
the elapsed time after the first core formation ( @xmath167 ) , elapsed time from the initial ( @xmath337 ) , density at the center of the cloud ( @xmath4 ) , and arrow scale are denoted in each panel .
, width=604 ] plane ( upper panels ) and @xmath104 plane ( lower panels ) at the same epoch as fig .
[ fig:10]_c _ , but with different grid levels of @xmath363 ( left panels ) , @xmath364 ( middle panels ) , and @xmath208 ( right panels ) , respectively .
white - dotted lines indicate the first core .
red - thick lines indicate the border between the outflow and accretion flow ( i.e. , contour of @xmath207 = 0 ) .
, width=604 ] ) at the same epoch as fig .
[ fig:10]_c_. the structure of high - density region ( @xmath365 ; red iso - density surface ) , and magnetic field lines ( black - and - white streamlines ) are plotted .
the structure of the outflow is shown by the blue iso - velocity surface inside which the gas is outflowing from the center .
the density contours ( false color and contour lines ) , velocity vectors ( thin arrows ) on the mid - plane of @xmath342 , @xmath105 , and @xmath3430 are , respectively , projected in each wall surface . , width=340 ] , color and contours ) and velocity vector ( arrows ) are plotted in @xmath286 region for models with initially slow rotation model sr ( left panels ) , and rapidly rotation fr ( right panel ) .
the plasma beta ( @xmath295 ; color and contours ) and magnetic field ( arrows ) are plotted in @xmath366 region for the same models .
red thick lines indicate the border between the outflow and accretion flow [ i.e. , the shapes of the jet ( left panel ) and outflow ( right panel ) ] . ,
width=604 ] ) at each mesh point in the region with @xmath367 ( i.e. , the outflow and jet region ) .
b _ : the ratio of the lorentz to the thermal pressure gradient forces ( @xmath284 ) in the region with @xmath367 ( i.e. , the outflow and jet region ) .
c _ : the ratio of the lorentz to the centrifugal forces ( @xmath283 ) in the region with @xmath368 ( i.e. , the inflow region ) . and panel _
d _ : the ratio of the lorentz to the centrifugal forces ( @xmath283 ) in the region with @xmath368 ( i.e. , the inflow region ) .
the horizontal axis means the number of the cell .
the dots are the data from models fr ( black dots ) , and sr ( red dots ) , respectively .
the thick lines are average of adjacent 100 cells for models fr ( black lines ) , and sr ( red lines ) , respectively .
the horizontal - dotted lines indicate the value at which each force equals .
, width=529 ] | the driving mechanism of jets and outflows in star formation process is studied using three - dimensional resistive mhd nested grid simulations . starting with a bonnor - ebert isothermal cloud rotating in a uniform magnetic field , we calculated cloud evolution from the molecular cloud core ( @xmath0 , @xmath1au ) to the stellar core ( @xmath2 , @xmath3 ) , where @xmath4 and @xmath5 denote the central density and radius of each object , respectively . in the collapsing cloud core
, we found two distinct flows : low - velocity flows ( @xmath65@xmath7 ) with a wide opening angle , driven from the adiabatic core when the central density exceeds @xmath8 , and high - velocity flows ( @xmath630@xmath7 ) with good collimation , driven from the protostar when the central density exceeds @xmath9 .
high - velocity flows are enclosed by low - velocity flows after protostar formation .
the difference in the degree of collimation between the two flows is caused by the strength of the magnetic field and configuration of the magnetic field lines .
the magnetic field around an adiabatic core is strong and has an hourglass configuration ; therefore , flows from the adiabatic core are driven mainly by the magnetocentrifugal mechanism and guided by the hourglass - like field lines .
in contrast , the magnetic field around the protostar is weak and has a straight configuration owing to ohmic dissipation in the high - density gas region .
therefore , flows from the protostar are driven mainly by the magnetic pressure gradient force and guided by straight field lines
. differing depth of the gravitational potential between the adiabatic core and the protostar cause the difference of the flow speed .
low - velocity flows correspond to the observed molecular outflows , while high - velocity flows correspond to the observed optical jets .
we suggest that the outflow and the jet are driven by different cores , rather than that the outflow being entrained by the jet . |
the wfcam science archive (; hambly et al .
2007 , collins et al . 2006 ) holds the image and catalogue data products generated by the wide field camera ( wfcam ) on ukirt ( united kingdom infrared telescope ) .
the data comprise pipeline processed multi - extension fits files ( multiframes ) containing pixel / image and catalogue data for four detectors at one pointing .
the latter contains all detections of stacked multiframes .
the data are pipeline processed at the cambridge astronomical survey unit ( casu ) and transferred to edinburgh where the wide field astronomy unit ( wfau ) processes it for ingestion into the database . since the release database contains advanced products which can take a lot of cpu time to produce , it is preferable to carry out the ingest procedure as fast as possible .
another aspect is that the pixel / image and catalogue data need to be ingested completely before further processing is done .
another constraint is the uniqueness of multiframes and detections .
each multiframe has to have a unique identifier across the whole database and each detection must be unique for a given survey .
to ingest the large amount of data into the wsa as the first stage in building a release database , a set of curation usecases ( cus ) have been designed . they are coded in a python / c++ environment , where c++ is used where high performance is needed and python to facilitate an easy to use object - oriented environment .
table [ tab : p4.1_tab1 ] shows an overview of the ingest cus and the average volume of data to be processed per observing night .
lll 1 & data transfer from casu to wfau & @xmath0 gbyte + 2 & creation of compressed images ( jpegs ) & @xmath1 jpegs + 3 & & @xmath2 files + 4 & & @xmath3 detections + in the following we will concentrate on cus 2 to 4 , as the data transfer ( cu1 ) is described in detail in bryant et al .
after transfer the data are split into daily chunks and then processed on multiple computers .
+ the following dependencies need to be observed to avoid duplicate entries or missing data .
each fits file gets assigned a unique multiframe i d associating data across the database with its source .
also each object in a catalogue gets assigned a unique object i d associating data across the database with its original detection . to update the database with the paths to compressed images ( cu2 ) , the general fits file metadata has to be ingested beforehand ( cu3 ) . and
finally catalogue data can only be processed ( cu4 ) if the corresponding image metadata is available ( cu3 ) .
the normal procedure of executing the cus for small amounts of data would comprise of a run of each cu followed by an ingest as shown in figure [ fig : p4.1_fig1 ] .
this might then be followed again by the whole procedure for the next batch of data .
as the creation of compressed images ( cu2 ) itself can be done independently , it can be decoupled from image and catalogue data processing . since
this procedure is linear , unique ids for pixel files and catalogue detections are applied during cu3 and cu4 , respectively .
+ to improve cpu usage and speed up the process for a whole cycle the following enhancements were applied : 1 .
files get multiframe ids depending on the largest mulitframe i d in the database directly after transfer ( cu1 ) .
the database is accordingly updated .
the cu4 object i d is turned into a temporary negative , only per day unique i d . this way we avoid overlaps with already in the database existing ( positive )
object ids .
it can be translated into a global unique i d directly after ingest with a simple addition to the last maximal object i d .
the ingest process is completely de - coupled from extraction and processing .
the first two steps allow us to process metadata simultaneously on different computers .
the last one allows the ingest to be run on available ingestable data while new data is processed . since ingest can take as much time as processing , this can nearly halve the run time of the full cycle .
figure [ fig : p4.1_fig2 ] shows the flow chart for the final design .
a data daemon has been created that checks the available computers , their load , and the tasks that need to be executed as well as tasks already running . at the moment
it suggests the distribution of these tasks to enhance usage of cpus on the pixel servers .
the database operator then takes the final decision in running the tasks .
in addition to the data daemon the ingester daemon can be run , checking automatically for ingestable data and ingesting them into the database .
this maximises cpu usage on the database server .
+ since these daemons can be run automatically , we are investigating the best way to run individual tasks remotely from a master computer .
a number of python - related solutions exist and are described in the .
running each of the cus 2 to 4 for 30 nights of data on five computers at the same time and ingesting the data as soon as it was available improved the total run times as follows : | one of the major challenges providing large databases like the wfcam science archive ( wsa ) is to minimize ingest times for pixel / image metadata and catalogue data . in this article
we describe how the pipeline processed data are ingested into the database as the first stage in building a release database which will be succeeded by advanced processing ( source merging , seaming , detection quality flagging etc . ) . to accomplish the ingestion procedure as fast as possible
we use a mixed python / c++ environment and run the required tasks in a simple parallel modus operandi where the data are split into daily chunks and then processed on different computers .
the created data files can be ingested into the database immediately as they are available . this flexible way of handling the data allows the most usage of the available cpus as the comparison with sequential processing shows . |
an intriguing alignment of gamma - hadron families ( i.e. , the outgoing high energy secondary particles from a single collision in the atmosphere ) along a straight line in a target ( transverse ) plane has been observed with ( lead and carbon ) @xmath6-ray emulsion chambers ( @xmath6rec s ) in the pamir mountains @xcite .- ray films .
this provides a few interaction lengths for hadrons and a quasicalorimeter determination of the particle s energy .
the c - chambers contain a 60 cm carbon layer covered on both sides by lead plates sandwiched with @xmath6-ray films .
the carbon block provides a large cross section for hadron interaction , while the lead blocks are of minimal thickness allowing determination of particle energies .
the total area of the chambers is few tens of square meters .
electron - photon cascades initiated by high energy hadrons and gamma - rays inside the @xmath6rec s produced dark spots whose sizes are proportional to the cascade energy deposited on the @xmath6-ray film . ]
these families can be reconstructed by measuring the coordinates and the incident direction of each particle in the film emulsion .
this allows determination of the total energy in gamma - rays and the total energy of hadrons release to gamma - rays .
recall that most of the hadrons in the family are pions and the average fraction of energy transferred by pions to the electromagnetic component is @xmath7 .
all families in the experiment are classified by the value of the total energy observed in gamma - rays , @xmath8 .
the centers of the main energy fluxes deposited on the @xmath6-ray film ( a.k.a .
`` subcores '' ) include halos of electromagnetic origin , gamma - ray clusters , single gamma - rays of high energy , and high energy hadrons .
the criterion for alignment is given by the asymmetry parameter @xmath9 where @xmath10 is the number of subcores and @xmath11 is the angle between vectors issuing from the @xmath12-th subcore to the @xmath13-th and @xmath14-th subcores @xcite .
the parameter @xmath15 decreases from @xmath16 ( corresponding to @xmath10 subcores disposed along a straight line ) to @xmath17 ( corresponding to the isotropic case ) .
events are referred to as aligned if the @xmath10 most energetic subcores satisfy @xmath18 .
a common choice is @xmath19 and @xmath20 .
the data have been collected at an altitude of 4400 m a.s.l . , _
i.e. _ , at a depth of 594 g/@xmath21 in the atmosphere . for low energy showers , @xmath22 ,
the fraction of aligned events coincides with background expectation from fluctuations in cosmic ray cascade developments .
however , for @xmath23 , the alignment phenomenon appears to be statistically significant @xcite .
namely , the fraction ( @xmath24 ) of aligned events is @xmath25 ( 6 out of 14 ) in the pb-@xmath6rec catalogue , and @xmath26 ( 13 out of 59 ) in the c-@xmath6rec catalogue .
the predominant part of the gamma - hadron families is produced by hadrons with energy @xmath27 , corresponding to interactions with a center - of - mass energy @xmath28 tev .
data analyses suggest that the production of most aligned groups occurs low above the chamber @xcite .
thus , it is not completely surprising that the kascade collaboration has found no evidence of this intricate phenomenon at sea level ( @xmath29 g/@xmath21 ) @xcite .
interestingly , the fraction of events with alignment registered in fe-@xmath6rec s at mt .
kanbala ( in china ) is also unexpectedly large @xcite . for gamma - hadron families with energy @xmath30
tev the fraction of aligned events is @xmath31 ( 3 out of 6 ) . in addition ,
two events with @xmath32 tev have been observed in stratospheric experiments @xcite .
both events are highly aligned : _ ( i ) _ the so - called strana superfamily , detected by an emulsion chamber on board a russian stratospheric balloon , has @xmath33 ; _ ( ii ) _ the jf2af2 superfamily , detected by an emulsion chamber during a high - altitude flight of the supersonic aircraft concord , has @xmath34 .
it is worth noting that stratospheric experiments record the alignment of particles , whereas mountain - based facilities register the alignment of the main fluxes of energy originated by these particles on a target plane .
the strong collinearity of shower cores has been interpreted as a tendency for coplanar scattering and quasiscaling spectrum of secondary particles in the fragmentation region @xcite . if the aligned phenomenon observed in cosmic ray showers
is not a statistical fluctuation , then events with unusual topology may be produced at the large hadron collider ( lhc ) . in this paper
we carry out a systematic study of @xmath35 scattering processes to establish the sensitivity of lhc experiments to such planar - shape topology .
the analysis technique described herein constitute an entirely general approach to search for planar scattering at the lhc .
firstly , we generate standard model ( sm ) qcd events that contain 4 light jets and 2 @xmath36-quark jets and 2 light jets utilizing the alpgen monte carlo code @xcite .
next , the event shape variable is classified according to the standard aplanarity parameter , @xmath37 @xcite .
namely , we define the signal region as @xmath38 and the control region with @xmath39 , varying the cut among these three choices : @xmath40 after that , we extract the ratio of bi - planar to planar events ( @xmath41 ) and calculate the @xmath41 uncertainty based on poisson statistics for a given luminosity @xmath42 . finally , we calculate the required luminosity to obtain @xmath2 , @xmath43 , and 95% c.l .
away from various values of @xmath41 .
though the search technique is agnostic regarding the hypothetical physics underpinning the planar configuration of events , an observation could have bearing on the recently proposed idea that spatial dimensions collapse at short distance , shutting off one - by - one with rising energy @xcite .
the layout of the paper is as follows . in sec .
[ iv ] we study the lhc sensitivity to coplanar particle escape .
we quantify signal and background rates of 4 jet events and show that future measurements of the aplanarity distribution of multi - jet events can provide a potent method for exposing a dimensional reduction of phase space . in sec .
[ ii ] we associate the dimensional reduction in momentum space to a reduction in spatial dimensions .
we assume that space at its fundamental level is an anisotropic lattice @xcite .
this idea that spatial dimensions effectively reduce with increasing energy directly constrasts with field / string theories in continuous spacetime dimensions , where dimensionality increases with a rise in energy . in sec .
[ iii ] we present a phenomenological analysis of the drell - yan scattering processes @xcite for a 3d @xmath1 2d crossover and we show that tevatron data is insufficient to constrain the model for a dimensional reduction above 1 tev .
we summarize our conclusions in sec .
in this section we estimate the lhc sensitivity to coplanar events and corresponding discovery reach at @xmath44 using @xmath45 scattering processes .
such scattering processes involve multiple virtual particles .
we define @xmath0 as the energy - scale of the onset of new physics .
we assume that when the momentum transfer @xmath46 ( @xmath47 ) in each of the propagators is comparable with @xmath0 , a growing fraction of the jets are produced in one plane in their center - of - mass frame .
this coplanarity is drastically different from the usual topology of 4d scattering , where the four outgoing partons are in general acoplanar . for simplicity ,
we consider three values of the coplanarity fraction at @xmath0 . we investigate the fractions 30% , 50% , and 100% , utilizing simulated qcd events ( with alpgen @xcite ) containing 4 jets ( either 4 light jets or 2 @xmath36-quark jets and 2 light jets ) .
while a fraction of the simulated events is planar , the entire sample is generally bi - planar . given the limited number of events containing four high transverse momentum ( @xmath48 ) jets , the probe of truly planar events is statistics limited .
therefore , we look to compare the sm qcd prediction of bi - planar jets with a purely planar sample by determining how many events are required for the observation of _ aplanar _ events to be significant . for a given propagator scale , @xmath0 , we accept events that have 4 jets that pass the following acceptance cuts : @xmath49 where @xmath50 is the @xmath14 jet pseudorapidity , and @xmath51 is the separation in the azimuthal angle ( @xmath52 ) - pseudorapidity ( @xmath53 ) plane among jets : @xmath54 in addition , to reject events that do not have a hard momentum scale , @xmath55 , we require that the invariant mass of any two pairs of jets satisfy @xmath56 and that for any two pairs of jets , the transverse momentum of a jet relative to the boost axis of the jet pair is @xmath57 finally , we model detector resolution effects by smearing the final state jet energy according to : @xmath58 the event shape variable , aplanarity , can be calculated based on the constructed momentum tensor @xmath59 where @xmath13 runs over the 4 jets in each event , with all momenta measured in the c.m .
the aplanarity is defined as @xcite @xmath60 where @xmath61 is the smallest normalized eigenvalue of the momentum tensor , giving @xmath37 a maximum value of @xmath62 .
therefore , planar or collinear events possess @xmath63 values , while more 3d events approach the maximum value , @xmath64 .
now , we define events as exhibiting a planar topology by having the aplanarity below some cut value , @xmath38 . in fig . [ fig:1].a , we show @xmath37 for a variety of jet @xmath48 cuts for the lhc with 100 fb@xmath4 of integrated luminosity .
to facilitate distinguishing planar events from the usual sm qcd bi - planar events , we define the variable @xmath65 the ratio of 3d bi - planar events to planar events using the observable @xmath66 as a discriminator for planar and bi - planar events .
for pure planar events , this ratio should vanish , while for 3d sm events , it is generally nonzero . for a given @xmath48 cut
, we can identify at what luminosity it is statistically different from a ratio the model predicts , i.e. @xmath67 .
we vary the value of @xmath68 and find that @xmath69 gives the maximum sensitivity for @xmath70 , denoted by the vertical line in fig . [ fig:1].a . in fig .
[ fig:1].(b - d ) , we show the reach the lhc may have for a given luminosity for planar events , recast into the propagator scale , @xmath0 for values of @xmath71 . with 10(100 ) fb@xmath4 , the lhc may discover propagators of scale @xmath72 tev with @xmath70 .
the reach is degraded as @xmath71 increases such that at @xmath73 , @xmath2 discovery at 100 fb@xmath4 may be possible only for @xmath74 tev .
this strong dependence on @xmath71 is due to the 3d region in fig .
[ fig:1]a having fewer events than the planar region .
in addition to the aplanarity , one can construct other variables to test for planar events .
for instance , two planes may be defined as going through each jet pair .
the subsequent azimuthal angle , @xmath75 , between these two planes should be zero for all combinations of jet pairings for planar events , and nonzero otherwise . when using this alternate variable and placing a cut as done for the @xmath37 analysis above ( see fig .
[ fig:2].a ) , we arrive at a very similar reach for @xmath76 , shown in fig .
[ fig:2].b ; other @xmath77 cases are shown in fig .
3(c , d ) . the weaker dependence of the reach on the nonzero value of @xmath78 than in the @xmath37 case is due to the smaller @xmath79 value in the sm .
thus , making it easier to statistically distinguish between the bi - planar sm and planar events .
and @xmath80 are not necessarily equivalent . ]
an approach for production of highly coplanar multi - jet events , which exploits semi - hard qcd processes with a high transverse momentum transfer , has been discussed in @xcite .
this model is susscesful in explaining coplanar 3-jet events in cosmic ray data .
the leading particle provide one jet and the collinear singularity of qcd correlates the second and third jets , therefore producing a roughly coplanar 3-jet events in the lab system .
the enhanced amplitude due to collinear gluon emission may explain , for example , the two jets `` ridge '' phenomenon recently observed at the lhc in @xmath81 collisions @xcite and previouly observed at rhic in heavy ion collisions@xcite .
however , qcd collinearity can not explain @xmath82 jet events in cosmic ray physics , which corresponds to @xmath83 jet events in the center - of - mass system ( which is also the lab system ) at the lhc .
in the remainder of this paper we will discuss a model where multijet events are coplanar for all values of @xmath10 .
motivated by condensed matter systems , some us recently proposed spacetime may be an ordered lattice structure that becomes anisotropic at very small distances @xcite .
the proposed set up , shown in fig
. [ lattice ] , resembles that of dimensional crossover in layered strongly correlated metals @xcite .
these materials have an insulating character in the direction perpendicular to the layers at high temperatures but become metal - like at low temperatures , whereas transport parallel to the layers remains metallic over the whole temperature range .
the analogy which we adopt is to replace the temperature variable in the materials system with short - distance `` virtuality '' in the parton scattering processes .
we further assume that the lattice orientation is randomized on a scale sufficiently small to avoid any preferred direction in space on the macroscopic scale . on the small scale , there will be a preferred direction given by the local lattice orientation .
therefore , hard scattering processes can resolve the lattice spacings that separate `` conducting space '' from the `` crystal world '' of insulating space . on the other hand `` macroscopic ''
objects like beam protons effectively see a spacetime continuum .
it is of interest to explore the consequences of this extreme viewpoint ( and let experiment be the arbiter ) .
we are not aware of any data that would rule out this conjecture .
we emphasize that this conjecture is radically different from the increase of dimensions at small spatial scales that are postulated in string theories and in many modern field theories .
in contrast with our conjecture the effective number of dimensions decreases as partons probe smaller scales . with the reduction in spatial dimension ,
phase space is reduced , the cross - sections are reduced , and multi - jet final states are necessarily coplanar ( for @xmath84 ) . with an increase in spatial dimension ,
phase space is increased , cross - sections increase , and multi - jet final states fill the three spatial dimensions but also lose energy and multiplicity into the extra dimensions . as demonstrated in sec .
[ iv ] for the coplanarity , striking collider signals would be observed at the lhc if @xmath85 tev .
many related aspects of beyond the sm field theory will occur at high - energy : the transition of the renormalizable sm to a super - renormalizable field theory , modification in the evolution properties of parton distribution functions , running coupling constants , running anomalous dimensions of operators , and so on .
a study of these new effects requires a more specific model .
such a study is beyond the scope of this paper .
if the lhc provides indications for the correctness of our conjecture , then it becomes neccessary and even imperative to explore the new physics in detail .
in the next section we address an obvious probe of new physics , namely drell - yan production of lepton pairs . above @xmath86 , the reduction of phase space
is expected to reduce the drell - yan cross - section , we show that present tevatron data does not impact on the conjecture , but that future lhc data will provide a new reach in @xmath0 .
in this section we assess the effect on drell - yan cross sections at tevatron and at lhc of the restriction to coplanarity of scattering above the lattice energy @xmath87 1 tev . in order to effectuate this , we will need to hypothesize an algorithm for a smooth continuation between @xmath88 to @xmath89 . in a standard manner
, we define the cross section as the probability transition rate divided by the flux , and obtain @xmath90 where @xmath91 is the feynman amplitude invariant in @xmath92 dimensions ; @xmath93 and @xmath94 are the initial and final center - of - mass momenta @xmath95 ; and @xmath96 is the solid angle aperture in the center of mass .
thus @xmath97 whereas for @xmath98 @xmath99 ( note that since @xmath100 the `` cross section '' @xmath101 in two spatial dimensions is a straight line with dimension length , as expected . ) to assess the impact on drell - yan , it is of critical importance to modify @xmath101 in order that it gain entry as a @xmath102 in the parton model calculation of @xmath103 .
this entails a change in dimension for the cross sections , and thus it _ can not _ occur in the probability transition rate , which has the same dimension @xmath104 for any @xmath92 .
thus it must reside in the incoming flux @xcite . assuming that this modification is purely of geometric origin , without energy dependence , we insert a factor of @xmath0 to adjust dimensions and obtain the following prescription for the transition to @xmath98 : @xmath105 at this point we must deal with the matrix element @xmath106 .
there are essentially two choices : _
( i ) _ one can assume total ignorance of the dynamics , ( seeing as it may involve the interaction of particles with a space time lattice ) , and simply work on dimensional grounds .
this approach is decidedly unsatisfactory , since it would permit arbitrary powers of @xmath107 vitiating any hope of making contact with quantitative experimental findings ; or _
( ii ) _ we can assume that @xmath106 can be calculated from a 3 dimensional version of qed . in case _
( ii ) _ , the operational lagrangian is @xmath108 where @xmath109 is the field strength , @xmath110 is the coupling , and @xmath111 is a four - component spinor with the corresponding four - dimensional representation of the dirac algebra @xmath112 with @xmath113 the pauli matrices @xcite .
then the calculation of drell - yan proceeds the same manner as in @xmath114 , with the result @xmath115 where the factor @xmath116 comes from the dirac spinor normalization and the @xmath117 from the virtual photon ( or @xmath118 ) propagator .
an engineering dimensional analysis of ( [ eq : lag ] ) shows @xmath119 so that indeed @xmath120 . would be energy ( and dimension ) dependent , receiving contributions from radiative corrections @xcite . ] in this case , eq.([eq : fourthree ] ) reads @xmath121 for @xmath122 more explicitly , the working algorithm for examining effects of the transition to coplanarity would be the following : @xmath123 where @xmath124 is the drell - yan parton - parton cross section as calculated in the sm .
for convinience we use as an interpolating function @xmath125 where @xmath126 parameterizes the the probability to remain confined to the two dimensional space . the quantity @xmath0
may also be regarded as characteristic of a potential barrier for phase transitions from 3 to 2 spatial dimensions , with @xmath126 representing the probability of tunneling between these states .
this is perhaps reminiscent of the the factor @xmath127 the probability per unit time per unit volume for creating an @xmath128 pair in a constant electric field @xmath129 @xcite . to obtain the total cross section for the process
@xmath103 we have to convolute ( [ eq : working ] ) with the parton distribution functions .
we do not at present have a controlled calculation coming from a well - defined formalism for the parton distribution with particle momenta oblique to the lattice layer . a crude approximation that conserves 4-momentum among the particles
is to assume the beam axis is aligned with the lattice layer .
this hadron alignment in turn means that the partons in the infinite momentum frame are aligned with the lattice . in reality , it is only a projection of the parton momentum that is aligned with the lattice .
each of the two partons oblique to the lattice will presumably undergo an inelastic scattering with the lattice plane in advance of their mutual hard scattering .
this new mechanism of parton energy loss is somewhat analogous to an initial state radiation off the partons beyond the physical radiation included in factorization theorems .
a question that immediately arises is whether there are constraints on this model from tevatron data .
the d0 collaboration reported the most recent study of the dielectron invariant mass spectrum analyzing 5.4 fb@xmath4 of data collected at @xmath130 tev @xcite .
there are four events observed in the energy bin @xmath131 .
since there is negligible background from other sm processes , the 68.27%cl spread in the poisson signal mean is ( @xmath132 ) events @xcite . at this point
it is worth recalling that the drell - yan @xmath133 measures lepton pair spectra at parton collision energy @xmath134 therefore , from eq .
( [ interpolation ] ) it is straightforward to verify that a dimensional reduction at @xmath135 tev ( which predicts @xmath136 events in the @xmath137 energy bin ) is consistent with tevatron data at the @xmath138 level .
a similar analysis follows from the invariant mass spectrum of dimuon data reported by the cdf collaboration @xcite .
interestingly , the @xmath139 tev region will be tested by the early lhc run at @xmath140 tev .
in the first part of this paper we have presented a complete model independent study to search for planar events in 4 jet final states .
the only free parameter of our analysis is the characteristic energy scale for the onset of coplanarity , @xmath0 . for the extreme scenario in which all @xmath3 scattering processes become coplanar above @xmath0
, we have shown that with an integrated luminosity of 10(100 ) fb@xmath4 the lhc experiments have the potential to discover correlations between jets if @xmath141 tev . in the second part of this paper
we have discussed some aspects of the recently proposed @xcite latticized spatial dimensions , with different characteristic lattice spacings in each dimension .
we have studied specific collider signatures that will probe this idea . in particular , we have shown that the predicted energy spectrum of drell - yan scattering is significantly modified in this model . remarkably , the anisotropic crystal world yields planar events when the energies of hard scatterings exceed the scale at which space transitions from 3d @xmath1 2d .
therefore , four jet events at the lhc will exhibit striking planar alignment ( if the parton - parton momentum transfer @xmath46 exceeds the energy scale @xmath0 of the lattice ) .
jets with this strong azimuthal anisotropy may have been already observed by the pamir collaboration : the effect know as alignment , which can not be explained by conventional physics .
we would like to thank malcolm fairbairn for valuable discussions .
this work is partially supported by the us national science foundation , under grants no .
phy-0757598 , phy-0757959 , and phy-0914893 , us department of energy , under grants no .
de - fg02 - 91er40688 , de - ac02 - 06ch11357 , de - fg02 - 91er40684 , and de - fg05 - 85er40226 , eu marie curie network universenet ( hprn - ct-2006 - 035863 ) , and the uwm research growth initiative .
l. t. baradzei _
[ pamir collaboration ] , bull .
* 50n11 * , 46 ( 1986 ) [ izv .
nauk ser .
* 50 * , 2125 ( 1986 ) ] ; mgu-89 - 67 - 144 ( 1989 ) ; bull .
* 55n4 * , 24 ( 1991 ) [ izv .
nauk ser .
fiz . * 55 * , 650 ( 1991 ) ] ; bull .
phys . * 57 * , 612 ( 1993 ) [ izv .
nauk ser .
* 57n4 * , 40 ( 1993 ) ] .
i. p. ivanenko , v. v. kopenkin , a. k. managadze and i. v. rakobolskaya , jetp lett .
* 56 * , 188 ( 1992 ) ; v. v. kopenkin , a. k. managadze , i. v. rakobolskaya and t. m. roganova , phys .
d * 52 * , 2766 ( 1995 ) [ arxiv : hep - ph/9408247 ] .
j. n. capdevielle , j. phys .
g * 14 * , 503 ( 1988 ) ; a. k. managadze _ et al .
. nucl .
lett . * 112 * , 19 ( 2002 ) ; v. i. galkin _ et al .
_ , bull .
* 66 * , 1697 ( 2002 ) [ izv
. ross .
* 66 * ( 2002 ) 1544 ] ; a. k. managadze and v. i. osedlo , bull .
* 73 * , 615 ( 2009 ) [ izv .
nauk ser .
fiz . * 73 * , 653 ( 2009 ) ] . a. de roeck , i. p. lokhtin , a. k. managadze , l. i. sarycheva , and a. m. snigirev , in _ proceedings of the 13th international conference on elastic and diffractive scattering _ ( edts .
m. deile , d. denterria , and a. de roeck , 2010 ) p.308 ; m. deile _ et al .
_ , arxiv:1002.3527 [ hep - ph ] ; r. mukhamedshin , eur . phys
. j. c * 60 * , 345 ( 2009 ) .
see also ref .
@xcite .
r. jackiw and s. templeton , phys .
d * 23 * , 2291 ( 1981 ) ; r. d. pisarski , phys . rev .
d * 29 * , 2423 ( 1984 ) ; t. w. appelquist , m. j. bowick , d. karabali and l. c. r. wijewardhana , phys .
d * 33 * , 3704 ( 1986 ) . | alignment of the main energy fluxes along a straight line in a target plane has been observed in families of cosmic ray particles detected in the pamir mountains .
the fraction of events with alignment is statistically significant for families with superhigh energies and large numbers of hadrons .
this can be interpreted as evidence for coplanar hard - scattering of secondary hadrons produced in the early stages of the atmospheric cascade development .
this phenomenon can be described within the recently proposed `` crystal world , '' with latticized and anisotropic spatial dimensions .
planar events are expected to dominate particle collisions at a hard - scattering energy exceeding the scale @xmath0 at which space transitions from 3d @xmath1 2d .
we study specific collider signatures that will test this hypothesis .
we show that the energy - spectrum of drell - yan scattering is significantly modified in this framework . at the lhc ,
two jet and three jet events are necessarily planar , but four jet events can test the hypothesis . accordingly
, we study in a model - independent way the @xmath2 discovery reach of the atlas and cms experiments for identifying four jets coplanarities . for the extreme scenario in which all @xmath3 scattering processes become coplanar above @xmath0
, we show that with an integrated luminosity of 10(100 ) fb@xmath4 the lhc experiments have the potential to discover correlations between jets if @xmath5 tev . |
the neutral current measurements at sno @xcite have confirmed the oscillations of solar neutrinos . after the first evidence of antineutrino disappearance in a beam of electron antineutrinos reported by kamland experiment @xcite ,
all other explanations of the solar neutrino problem are either completely discarded or may just be subdominant effects . after these pioneering experiments , there is hardly any scope for doubt about the physical reality of neutrino mass and their , consequent , oscillations . once neutrino mass is observed , neutrino magnetic moments are an inevitable consequence in the standard model and beyond .
kamland is the first experiment to explore the neutrino parameter space relevant to solar neutrinos with a beam of terrestrial neutrinos .
kamland has , convincingly , demonstrated the existence of neutrino oscillations with parameters confined to the lma region .
apart from the measurement of reactor antineutrino flux , kamland is sesitive to any boron antineutrinos produced from solar boron neutrinos as a result of spin flavor oscillations .
the flux of solar antineutrinos is expected to be considerable , in case , the neutrinos possess majorana character and , in addition , the neutrino magnetic moment is high enough .
thus , the observation of a small solar antineutrino flux is likely to establish the majorana character of the neutrino and the existence of physics beyond the standard model .
an important signature of rsfp is an observable flux of electron antineutrinos from the sun if neutrinos , in addition to transition magnetic moments , have a sizable flavor mixing .
the flux of solar electronic antineutrinos can , in principle , be detected at borexino and sno .
if both transition magnetic moments and flavor mixing of massive neutrinos are present , the combined action of rsfp and the flavor oscillations may lead to an observable flux of electron neutrinos .
the existing bounds on the solar antineutrino flux from kamland are fairly stringent @xcite .
it may be worthwhile to emphasize again that the observation of solar electronic antineutrino flux will be a unique signature for a majorana neutrino transition magnetic moment and the resonant spin - flavor precession inside the sun .
therefore , the results of the electronic antineutrino flux measurements by borexino will be eagerly awaited since the sensitivity of borexino to the electronic antineutrinos is expected to be about 37 times larger than that to the electronic neutrinos .
the detection of solar antineutrinos is one of the major goals of several existing and forthcoming solar neutrino experiments .
the kamland bounds on the boron antineutrino flux are available now @xcite and the beryllium antineutrinos will be observable in borexino .
sno has already indicated that the solar electron neutrinos are being converted into active neutrinos by observing the neutral current processes @xcite .
therefore , only a small fraction of electron neutrinos is allowed to be converted into the muon antineutrinos via the rsfp mechanism within the solar interior .
these muon neutrinos can , then , oscillate into the electron antineutrinos while travelling from the solar surface to the earth via vacuum transitions with a probability of about half .
the bounds on the antineutrino flux have been used to obtain the bounds on @xmath0 for the case of high energy boron neutrinos @xcite .
after kamland results and earlier sno neutral current measurements , there is hardly any doubt about the fact that the neutrino oscillations with lma parameters explain the solar neutrino deficit and that the electron neutrinos are being , dominantly , converted into muon neutrinos .
only a small fraction is allowed to be converted into antineutrinos via rsfp mechanism . in this work ,
we consider the solar antineutrino production in the rsfp scenario@xcite in coexistence with the lma msw scenario and reexamine the earlier results for boron neutrinos .
we also extend our analysis to include the low and intermediate energy neutrinos which have not been considered in the earlier analyses @xcite . in order to obtain bounds on low and intermediate energy antineutrino fluxes ,
we compare the model independent values of the survival probabilities with their values in the pure lma case assuming that the difference between the two is due to the rsfp driven @xmath1 transitions .
it is noted that beryllium neutrino flux is smaller than the lma expectations .
this could explain the low ar - production rate in the homestake experiment sought to be explained by smirnov @xcite by invoking the existence of a light sterile neutrino . in this work
, we surmise that the missing beryllium neutrinos are being converted into muon antineutrinos via rsfp transitions within the sun .
the muon antineutrinos , thus produced , undergo vacuum oscillations into the electron antineutrinos in their flight to the earth .
this antineutrino flux can be detected by borexino .
thus , the hypothesis advanced in the present work can be directly confirmed or discarded by borexino in contrast to the hypothesis put forth by smirnov @xcite where the missing beryllium neutrinos are assumed to be oscillating into sterile neutrinos . in this scenario ,
the neutrino transition magnetic moments and solar magnetic fields required to account for the @xmath1 conversion of missing beryllium neutrino flux are well within their known limits . in order to avoid an anomalously large rsfp conversion of boron neutrinos , the magnetic field has to undergo a decrease towards the edge of the core .
a suitably decreasing magnetic field profile also explains the absence of the rise up of neutrino energy spectrum at superkamiokande .
this rise up can be as large as 10 percent in the pure lma scenario @xcite .
in our model of antineutrino production , the @xmath2 flavor transitions are described by the lma scenario while the @xmath3 transitions are described by the rsfp scenario where both types of transitions are assumed to be occurring independently of each other .
this approximation is well justified for the lma oscillation parameters for which the rsfp resonance width is too small to interfere with the lma flavor transitions @xcite .
the rsfp resonance condition @xcite @xmath4 implies that for the neutrinos of energy e , the resonance will occur at the point @xmath5 where @xmath6 , and we have considered the standard model density profile @xcite . here @xmath7 and the energy e is in the units of mev .
boron neutrinos have average energy 6.7 mev .
the production region of the boron neutrinos extends up to x=0.15 @xcite .
thus , in the course of their propagation , the boron neutrinos will get the resonance density at x=0.24 .
the resonance point for the beryllium neutrinos @xmath8 will be at x=0.06 which is near the point of maximal production @xcite .
thus , the production and the resonance regions may coincide for the @xmath9be ( 0.861 mev ) neutrinos for certain values of lma parameters .
this makes the lma transition probability for beryllium neutrinos about half .
however , the adiabatic rsfp transition probability for beryllium neutrinos is very small for values of @xmath0 within the range constrained by various laboratory and astrophysical bounds .
therefore , the non - adiabatic effects in the resonance region will be important .
as the energies of the low energy pp neutrinos are below 0.45 mev , they will not get the resonance densities anywhere .
the production region for the neutrinos is an extended one . for the pp
neutrinos the maximal production occurs at the point x=0.1 whereas for the boron and beryllium neutrinos , the maximal production occurs at about x=0.05 @xcitethe mixing angle in matter in the presence of a magnetic field is given by @xcite @xmath10 where @xmath11 the rsfp transition probability is given by @xmath12 where @xmath13 is to be evaluated at the production point whereas @xmath14 is to be evaluated at the edge and is approximately one .
the crossing probability given by @xmath15 is to be evaluated at the respective resonances @xcite . for lma parameters
, it simplifies to @xmath16 where @xmath17 is in the units of @xmath18 and e is in the units of mev . for beryllium neutrinos
, we have @xmath19 . for the boron and pp neutrinos
, we have @xmath20 at the respective production points for their respective average energies .
the bounds on @xmath21 are available from the various sources @xcite .
the reactor bounds on the neutrino magnetic moment are @xmath22 .
the bounds on @xmath23 @xcite are less certain and little is known about its magnitude and spatial variation , especially in the deep solar interior .
we can safely assume that @xmath24 in the solar core .
for @xmath25 , the sfp adiabatic transition probability is 0.64 for pp neutrinos .
a considerable transformation into antineutrinos is , thus , allowed , even in the adiabatic approximation .
hence , such a large field is directly ruled out by virtue of the reactor bounds on the neutrino transition moments .
this lowers considerably the chandershekhar bound on b @xcite .
we , therefore , take @xmath26 as an upper bound , which implies @xmath27 .
there will be considerably large adiabatic transitions for this value of @xmath17 .
the values of cos2@xmath28 for pp , beryllium and boron neutrinos are 0.96 , -0.12 and -0.78 respectively .
thus , the transition probability is 0.02 for pp neutrinos in the adiabatic approximation which is justified for pp neutrinos because they do not undergo resonance . for
be and b neutrinos , the crossing probability is not zero and will be the only effect for smaller values of @xmath0 . for pp neutrinos , the transition probability ( adiabatic ) will approach zero as @xmath29 decreases .
if @xmath30 is one order of magnitude below the above upper bound , _
i.e. _ @xmath31e.g.@xmath32 , the transition probability is @xmath33 for the pp neutrinos . for the beryllium and boron neutrinos ,
@xmath34 so that the adiabatic transition probability will be zero for them .
therefore , the transition of pp neutrinos into antineutrinos is not expected at all and the transformation of beryllium and boron neutrinos is highly non - adiabatic ( adiabatic transition probability being zero ) .
hence , for the intermediate and high energy neutrinos , we have @xmath35 for @xmath36 .
the bounds on antineutrino flux produced from the @xmath9be neutrinos will be provided by borexino . from these bounds
, one can constrain @xmath0 within the core .
conversely , assuming suitable values for @xmath0 , one can predict the beryllium electron antineutrino probability and , hence , the beryllium electron antineutrino flux .
the probability for the electron antineutrinos to appear at the earth is given by @xmath37 where , @xmath38 .
alternative models of antineutrino production studied in the literature @xcite either consider a highly chaotic magnetic field in the radiation zone of the sun or use perturbation calculations [ valid only for small @xmath0 ] to obtain the rsfp transition probability . in perturbative calculations ,
the transition probability depends only on the magnetic field in the production region . in both categories of models ,
the transition probabilities for beryllium and boron neutrinos can not be much different .
thus , these models can not explain a large transition probability for beryllium neutrinos into antineutrinos while still keeping the boron antineutrino production at a phenomenologically acceptable level , in contrast to the model considered in the present work .
the model independent survival probabilities for the low , intermediate and high energy neutrinos have been calculated from the solar neutrino data by barger _
et al _ @xcite .
the survival probability for high energy neutrinos can be obtained from sno nc and cc measurements alone the value for which in the pure lma scenario is @xmath39 .
the agreement between the two values is the first step towards the solution of the solar neutrino problem .
the experimental values , however , do not agree with the survival probabilities in the pure lma scenario for low and intermediate energy neutrinos .
the survival probability for the beryllium neutrinos is expected to be nearly half in the pure lma scenario compared to the value of 0.3 obtained from the model independent analysis of the solar neutrino data @xcite poinitng towards a significant portion of this component to be missing @xcite .
the two values do not agree even if the uncertainties in the experiments and the ssm fluxes are taken into consideration .
similarly , the lma survival probability for pp neutrinos is the same as the vacuum survival probability @xmath40 whereas the experimental value is about 0.8 @xcite .
thus , the observed pp neutrino flux is larger than the lma expectation .
therefore , there is no scope for antineutrino production in the low energy range
. the missing beryllium neutrino flux can account for the low ar - production rate at homestake sought to be explained by smirnov @xcite by invoking mixing with a sterile component .
it is clear that a portion of the beryllium neutrinos is undergoing other subdominant transitions in addition to the dominant lma flavor oscillations .
these transitions have to be effective only at the intermediate energies .
an attractive candidate for these are the rsfp transitions , though , the transitions into a light sterile neutrino can also explain this anomaly @xcite .
the attractiveness of the rsfp transitions of missing beryllium neutrinos into antineutrinos lies in the possibility of their direct detection at borexino . if the event rate at borexino is found to be smaller than @xmath41 ( pure lma value )
, it would confirm that a portion of the beryllium neutrinos are missing .
in addition , if borexino , also , observes an antineutrino flux of about 0.1 of the ssm beryllium neutrino flux , it would be confirmed that the missing beryllium neutrinos are undergoing rsfp transitions .
otherwise , the possiblity of oscillation into the sterile neutrinos has to be taken seriously .
similarly , the rsfp transitions of boron neutrinos will result in values of the fluxes smaller than the lma expectations . in the pure lma scenario ,
the flux of boron neutrinos is expected to increase slightly as the energy decreases .
this rise up in the neutrino energy spectrum can be as large as 10 percent @xcite .
the possibility of rsfp transitions can result in a decrease in the neutrino flux which can explain the lack of the rise up of the boron neutrino spectrum at low energies . from equations ( 5 ) , ( 6 ) , and ( 7 )
, it follows that to account for the missing beryllium neutrinos via rsfp tarnsitions , we require @xmath42 deep inside the core ( @xmath43 ) . to explain the upper limit on the boron antineutrino appearance probability ( about 0.0034 ) obtained at kamland @xcite
, we must have @xmath44 towards the edge of the core ( @xmath45 ) .
the recent direct bounds on the solar magnetic field at the bottom of radiation zone are : @xmath46 @xcite which are independent of the values of the neutrino magnetic moment . if these bounds are to be taken seriously , we must take @xmath47 to have the required value of @xmath30 at @xmath45 .
thus , a magnetic field of the order of about 10000 t is required deep within the core . from the equations ( 5),(6 ) and
( 7 ) , and the rise up in the pure lma survival probability , one can calculate the solar magnetic field profile .
if p@xmath48(e ) is the energy dependent survival probability in the lma scenario , we must have @xmath49 in order to have no rise up in the boron neutrino spectrum at superkamiokande where the effects of flux - averaging over the production region have been neglected .
the pure lma survival probability @xcite as a function of energy is given by @xmath50 where @xmath51 the lma survival probability @xmath52 has been ploted in fig.1.as a function of energy e ( in the units of mev ) .
the rise up in the survival probability for boron neutrinos at low energies is about 4 percent . also , for the average value of energy ( 6.7 mev ) , @xmath53 while @xmath54 .
this implies an electron antineutrino appearance probability of 0.0031 in fairly close agreement with the kamland bound @xcite . the magnetic field profile in the resonance region of boron neutrinos extending from 0.2 to 0.3
can be obtained , directly , from equation ( 9 ) . in fig.2 .
, the magnetic field b ( in the units of tesla ) has been plotted against the distance from the center of the sun in the units of the solar radius , assuming @xmath55 .
to conclude , we considered the solar antineutrino production via rsfp transitions in coexistence with the dominant lma msw transitions to explain the kamland bounds on solar antineutrinos .
the bounds on the antineutrino flux have been translated into the bounds on solar magnetic fields .
recent studies @xcite on solar antineutrinos are limited to boron antineutrinos for which direct bounds are available from kamland . however , we have incuded low and intermediate energy neutrinos in this analysis .
indirect bounds on low and intermediate energy antineutrino fluxes have been obtained by comparing the model independent values of the survival probabilities obtained from the solar neutrino data with their values in the pure lma case .
it is noted that beryllium neutrino flux is smaller than the lma expectations implying that the beryllium neutrinos are also undergoing transitions other than the flavor oscillations at a subdominant level . a natural candidate for these transitions
are the spin - flavor transitions which can result in lower event rate at borexino as compared to the pure lma expectations .
whether this is actually happening can , directly , be verified by borexino .
the neutrino transition moments and solar magnetic fields required to account for the missing beryllium neutrino flux are well within the known limits on these quantities .
moreover , to avoid an anomalously large rsfp conversion of boron neutrinos , the magnetic field has to undergo a decrease towards the edge of the core .
in addition , a suitably decreasing magnetic field profile can also explain the absence of the rise up of neutrino energy spectrum at superkamiokande .
one of the authors ( sk ) gratefully acknowledges the financial support provided by the council for scientific and industrial research(csir ) , government of india . c. s. lim and w. j. marciano , _ phys .
rev . _ * d37*,1368 ( 1988 ) ; e. kh .
akhmedov , sov .
phy . jetp68,690 ( 1989 ) ; b. c. chauhan , s. dev and u. c. pandey , _ phys .
_ * d * * 59 * , 083002 ( 1999 ) ; e. k. akhmedov and j. pulido , _ phys .
_ b * 485 * , 178 ( 2000 ) ; e. k. akhmedov and j. pulido , _ phys .
lett._*b * * 529 * , 193 ( 2002 ) ; e. k. akhmedov and j. pulido , phys .
b * 553 * , 7 ( 2003 ) . * * * * | even after the confirmation of the large mixing angle ( lma ) solution of the solar neutrino problem , the scope for resonant spin- flavor precession ( rsfp ) transitions as a subdominant effect still exists . in this work
, we have considered suitably suppressed rsfp transitions in addition to the dominant lma flavor transitions and translated the bounds on the antineutrino flux into the bounds on the product of neutrino magnetic moment and solar magnetic field .
the low and intermediate energy neutrinos have been included in the analysis by obtaining indirect bounds on the corresponding antineutrino fluxes for these components .
it is assumed that the missing beryllium neutrinos are being converted into muon antineutrinos via rsfp transitions in the sun .
= 0.1 in = -0.3 in = 0.7truecm = 0.1truecm = -0.4truecm = -0.4truecm * neutrino oscillations , solar antineutrinos and the solar magnetic fields * + s. dev and sanjeev kumar + department of physics , himachal pradesh university , shimla-5 , india . 0.6truecm |
with the enormous progress of research on colossal magnetoresistance ( cmr ) manganites@xcite over the last decade it has been realized that besides charge and spin also _ lattice degrees of freedom _ @xcite play a significant role in the formation of the cmr and the metal - insulator transition ( mit ) . special attention focused on the optimal doped manganites , characterized by the largest cmr effects in particular , la@xmath0ca@xmath1mno@xmath2 ( lcmo ) .
as the temperature range in which the cmr effect and the mit are observed is close to room temperature , these compounds have become potential candidates for technical applications .
the phonon excitation spectrum has been studied in detail for undoped lamno@xmath2 @xcite which is , below approximately 750 k , in a jahn - teller ( jt ) distorted orthorhombic ( pnma ) structure . a theoretical analysis of the complete set of phonon modes in the doped compounds is still missing .
lcmo has an orthorhombic symmetry whereas the sr - doped compound , la@xmath0sr@xmath1mno@xmath2 ( lsmo ) , is rhombohedrally distorted .
the infrared - active phonons of undoped rhombohedral ( r@xmath3c ) lamno@xmath2 were identified @xcite however these zero - doping investigations can not make up for experiments at finite doping .
in fact , the local mn - o geometry depends strongly on the size of the ions and doping concentration @xcite which accounts for the structural differences between lcmo and lsmo .
experimental studies of two single crystals , undoped and with @xmath4% sr - doping , were performed by paolone _
et al._@xcite they discussed the assignment of all infrared active phonon modes in lamno@xmath2 at low temperature .
however , to the best of our knowledge , there are no reports on phonon modes in the metallic phase of lsmo . for bulk samples of lcmo , three sets of broad phonon bands
are observed which reveal external- , bending- and stretching - type lattice vibrations , as classified in the cubic symmetry .
@xcite apart from providing a fingerprint of the lattice structure , phononic spectra can indicate the relevance of electronic correlations .
the dominant effect of mobile electrons on the lattice vibrations is generically a suppression of phononic resonances due to screening . for metals
, this may render the phonons unobservable in far - infrared ( fir ) optical spectroscopy .
however , since the polycrystalline samples are poor metals , and even more so the investigated thin films , phonon resonances are still clearly resolved even in the conducting low - temperature ferromagnetic phase . in lcmo
a significant shift of the stretching mode near the mit was found @xcite and associated with a reduced screening through charge carriers close to the mit .
although the observed shifts of phonons in lcmo are evidently related to electronic correlations since they come along with the mit it is not obvious that they are explained exclusively by a reduced screening .
in particular , not all phonon modes display the shift and , furthermore , in lsmo we observe phononic shifts which are tied to the ferromagnetic - to - paramagnetic ( fm - pm ) transition , not the mit ( see sec . [
sec : comp ] ) . in lcmo
the fm - pm transition and the mit coincide whereas in lsmo they are well separated .
a comparative study of lcmo and lsmo should address and illuminate the fundamental question if the phononic excitations couple to spin degrees of freedom . since the phonon resonances are well pronounced in the optical spectroscopy measurements of the metallic manganite films they provide a unique opportunity to analyze the consequences of a coupling of lattice degrees of freedom ( dof ) with spin and charge .
correlation of lattice dof with charge dof has to be in an weak ( lsmo ) to strong coupling range ( lcmo ) as , for example , the polaronic excitations in the mid - infrared ( mir ) evidence a coupling in this range.@xcite also neutron diffraction measurements for lcmo revealed an anomalous volume thermal expansion at the paramagnetic - to - ferromagnetic mit,@xcite which has to originate from a sufficiently strong coupling .
we will elaborate further on the consequences of a strong phonon - electron coupling for the phononic resonances in this paper and we will identify the characteristics in the fir spectra signifying cooperative dynamics not only of lattice and electronic charge but also of lattice and spin degrees of freedom . the paper is organized as follows : sec .
[ sec : expdetails ] deals with the experimental details , sample characterization and data analysis . in sec .
[ sec : tempr ] , we present reflectivity measurements of the infrared active phonons in thin films of lsmo and lcmo at temperatures ranging from the insulating paramagnetic ( pm ) to the metallic ferromagnetic ( fm ) phase . in sec .
[ sec : ht ] we give a comprehensive analysis of phonons from the respective optical data .
the spectra confirm the anticipated mode structure for the respective symmetry of the lattice and prove the high quality of the thin films . in sec .
[ sec : comp ] we elaborate on the close relation between lattice and spin degrees of freedom by studying the temperature dependent shift of phonon modes across the ferromagnetic transition . in sec .
[ sec : fano ] , we discuss fano - like modifications of the line shape of distinct phonon modes in lcmo which testify a significant coupling of the lattice excitations to the electronic continuum in the ca - doped manganite films .
the interference of nearly degenerate modes with electronic excitations may result in spectral weight transfer and level repulsion which we discuss within a phenomenological approach for coupled phononic and electronic modes .
a summary of our findings and conclusions are presented in the last section , sec .
[ sec : conclusion ] .
the films for this investigation have been prepared using a standard pulsed laser deposition technique.@xcite lcmo was grown to a thickness of 200 nm onto ndgao@xmath2 single crystalline substrates .
lsmo # 1 with a thickness of @xmath5 nm and lsmo # 2 with @xmath6 nm were deposited on ( laalo@xmath2)@xmath7(sr@xmath8altao@xmath9)@xmath10 .
x - ray diffraction of lcmo revealed an orthorhombic structure with a preferred growth along the [ 110 ] axis while lsmo was confirmed to be in a rhombohedral structure with preferred growth along the [ 100 ] axis .
the infrared reflectivities of the sample and the pure substrate were measured using a combination of fourier transform spectrometers , bruker ifs 113v and ifs 66 v / s , to cover the frequency range from 50 to 40000 @xmath11 .
temperature dependent measurements from 6 to 295 k were performed with a he - cryostat .
higher temperatures were measured in a home made oven in which the sample was exposed to a continuous flow of heated nitrogen gas . to obtain the optical conductivity @xmath12 a kramers - kronig ( kk ) analysis
was performed which included the results from submillimeter spectroscopy between 5 and 30 @xmath11 . applying the fresnel optical formulas for the complex reflectance coefficient of the substrate - film system , the optical conductivity for the films was calculated.@xcite magnetization measurements
were carried out between 2 and 400 k by using a commercial quantum design squid magnetometer .
the dc resistance measurements were performed by a standard four - probe method .
the temperature dependence of the ac resistancee was done in a mach - zehnder interferometer arrangement , which allows the measurements of both , transmission and phase shift . applying the fresnel formulas , the ac resistance was determined directly without any approximations .
upper panel ( lsmo ) : temperature dependence of the magnetization at @xmath13 oe ( left scale ) and dc - resistivity ( right scale ) .
lower panel ( lcmo ) : magnetization at @xmath14 oe ( left scale ) ; dc- and ac - resistivity at @xmath15 ghz ( right scale ) . ] frequently , manganite films exhibit fm transition temperatures @xmath16 lower than those reported for the bulk materials with the same nominal composition .
the differences can be attributed to grain boundaries , strain , microstructure depending on the details of the growth process and on the lattice parameters of the substrate.@xcite they all result in a variation of the strength of the lattice distortion , which determines the structure and , accordingly , the electrical and magnetic behavior .
it is well known , that differences between bulk samples and thin films can occur , depending on the quality of the films .
however , we observe that the overall optical characteristics of the films are in good agreement with bulk properties . .metal - insulator ( @xmath17 ) , magnetic ( @xmath18 ) and electric ( @xmath19 ) transition temperatures of the lcmo and lsmo films . [ cols="<,<,^,^,^,^,^,<,<,<,<",options="header " , ] all line shapes of the first phonon mode in fig .
[ ca - fano ] exhibit a dip on the high - energy side , a consequence of the negative sign of the coupling parameters .
the coupling constant is proportional to the effective mass . in manganites
the carriers have hole character.@xcite the associated inverse sign of the effective mass may be responsible for a negative value of @xmath20 and @xmath21 .
it turns out that neither a coupling to the drude nor to the incoherent contribution can induce the nearly perfect antisymmetric shape of the 114 @xmath11 peak - dip structure .
only a sufficiently strong coupling to the polaronic excitation supports this antisymmetric fano shape .
the increase of coupling strength @xmath22 with temperature is significant .
it results from the fact that the line shape keeps its antisymmetric form up to the highest temperature ( 295 k ) while the electronic background is reduced by a factor 8 from 160 k to 295 k. since the line shape of the phonon is controlled by a product of the coupling constant and the strength of the polaronic background ( @xmath23 ) , the coupling constant has to increase correspondingly to keep the observed shape .
the line shapes of the two modes at about 170 @xmath11 do not deviate strongly from lorentzians which is consistent with rather small couplings @xmath24 and @xmath25 ( cf . tab .
[ doublemode ] ) .
however , in order to reproduce the `` hump '' in between the phonon at 114 @xmath11 and the nearly degenerate external modes at about 170 @xmath11 one needs a sufficiently strong electron - phonon coupling @xmath26 to the electronic resonance at about 300 @xmath11 .
these electron - phonon couplings , @xmath26 and @xmath27 , affect the position and the spectral weight of the resonances at @xmath28 and @xmath29 considerably .
the three coupled modes ( @xmath28 , @xmath29 , @xmath30 ) constitute a `` triad '' with a bonding ( lowest ) , non - bonding ( intermediate ) and anti - bonding state . the bonding and anti - bonding modes always experience level repulsion so that the bare frequencies @xmath28 and @xmath29 can be both at 171 @xmath11 ( degenerate modes , cf . tab .
[ doublemode ] ) but the two resonances are still well separated at 295 k ( at 167 @xmath11 and at 171 @xmath11 , cf . fig .
[ ca - fano ] and tab .
[ tabmoden - ca ] ) . @xcite the level repulsion is accompanied by a transfer of spectral weight : according to tab .
[ doublemode ] and fig .
[ ca - fano ] , the lowest mode of the triad takes nearly all its weight from the electronic mode at 295 k. this observation may explain why the two - mode structure at 170 @xmath11 in lcmo is still intense at or above the mit whereas in lsmo the mode at about 170 @xmath11 is weak ( cf . fig .
[ external ] , right panel ) . in fig .
[ ca_fano_vgl_t295 ] we display the `` bare modes '' , with parameters of tabs . [ tab - ca - fano ] and [ doublemode ] , jointly with the interaction - dominated mode spectrum in order to summarize visually the effects of the strong electron - coupling on the external mode group in lcmo .
fits of the low frequency phonon group with and without electron - phonon ( e - ph ) coupling for data at room temperature . ]
in contrast to lcmo , the phonon spectra for lsmo exhibit no asymmetric shape and can be described by lorentzian oscillators . in the latter compound
the electron - phonon coupling is smaller , as is confirmed by shape , position and temperature dependence of the polaronic mid - infrared ( mir ) resonance.@xcite
manganite thin films provide an excellent opportunity to study the phononic excitations of lsmo and lcmo in their metallic phases .
the films allow for a well - resolved , distinct phonon spectrum in infrared spectroscopy , even for temperatures well below the mit .
in lsmo we identified each of the eight infrared - active phonons expected for the r@xmath3c symmetry and in lcmo we were able to resolve 18 out of 25 phonons expected for the pnma - symmetry .
the phonon spectra reveal a number of anomalies which result from a finite electron - lattice coupling and which confirm the importance of cooperative effects of microscopic degrees of freedom when approaching the ferromagnetic - to - paramagnetic and the metal - insulator transition .
first and most apparent , we observe that the fir spectra change substantially up to the mit ( see figs . [ ca - r - temp ] and [ sr - r - temp ] ) but then , above the mit , temperature dependent modifications are negligible .
this overall @xmath31-dependence of the spectra up to the mit is related to a strongly reduced screening of the phononic excitations , due to localization or diffusive motion of electronic charges .
however also a coupling of lattice and spin degrees of freedom is evidenced by the phononic excitation spectrum : temperature dependent shifts of several phonon modes , which vary on the same temperature scale as the magnetization , confirm the correlation of these degrees of freedom ( see figs . [ ca - min ] [ casquid ] ) . in lcmo the pronounced asymmetric line shapes , observed for several phonons ( see figs .
[ ca - phononen ] and [ ca - fano ] ) , are a manifestation of the strong electron - phonon coupling .
we analyzed the fano lineshapes of the external group within a phenomenological approach .
@xcite it assigns independent modes to the various spectral contributions of the electronic excitations and couples them linearly to the phononic oscillators . in fir and mir spectra of lcmo , three electronic continua
are typically identified : a `` continuum of incoherent electronic excitations '' ( cf . refs . ) , observed as a background to the phonons in the fir range , polaronic excitations in the mir and possibly a rather insignificant drude contribution on the low - frequency side of the fir spectrum .
the lineshape of the lowest - frequency external mode , which is remarkably antisymmetric at all temperatures , can be reproduced only with a strong coupling to the polaronic excitations .
in contrast , the fir continuum , which is parameterized by a strongly damped oscillator spectral function , is responsible for frequency shifts ( level repulsion ) and transfer of spectral weight from the electronic excitations into distinct phononic modes ( see fig . [ ca_fano_vgl_t295 ] ) . in the presented scheme , we find that the mode splitting of the 170 @xmath11 resonance and its considerable spectral weight at high temperature ( above the mit ) is induced by the electronic continuum .
this observation may explain the discrepancy with lsmo where these two external modes are degenerate at high temperature since lsmo is characterized by a weaker electron - phonon coupling where fano - like lineshapes are entirely missing .
a microscopic evaluation is mandatory in order to assign unambiguously these anomalous characteristics of the phononic spectra to the interference with specific electronic excitations .
the preliminary identification within the phenomenological approach has provided ample motivation for such an effort .
1 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , , , * * , ( ) . , , * * , ( ) . , , , , , , , , * * , ( ) . , * * , ( ) .
, , , , , , , * * , ( ) . , , , , * * , ( ) . , , . , , , , , * * , ( ) .
, , , , , * * , ( ) . , , , , * * , ( ) . , , , , , , , * * , ( ) .
, , , , * * , ( ) .
, , , , , * * , ( ) . , , , , , , , , , * * , ( ) . , , , , , * * , ( ) . , , . , , . , , , , , * * , ( ) . , , , * * , ( ) . , * * , ( ) . , , , , ,
* * , ( ) . , , * * , ( ) . , ph.d .
thesis , ( ) .
, , , , , , , * * , ( ) . ,
* * , ( ) . , , * * , ( ) . moreover , a full evaluation of screening would have to include the diffusive nature of the charge transport which results in a retarded screening dynamics with a less prominent signature .
, , , , * * , ( ) . near the magnetic transition an irregular shift at 609 @xmath11 was attributed to spin - lattice interaction due to the observed variations of the lattice parameters and sublattice magnetization .
, , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , * * , ( ) .
the origin of the mode degeneracy in lsmo and its relation to the evolution of the spin degrees of freedom will be investigated in a subsequent publication ( hartinger _ et al .
_ , unpublished ) . , * * , ( ) . , , , , , , , , * * , ( ) , , , * * , ( ) . , , , , , , * * , ( ) .
, , * * , ( ) . , , * * , ( ) . , , * * , ( ) . , , , , , * * , ( ) . , , , , , , , * * , ( ) . in the original fano model a parameter @xmath32 determines the degree of the line asymmetry of the peak , which is inversely related to the strength of interaction.@xcite , , , * * , ( ) . , , , , , , * * , ( ) .
, , * * , ( ) . , , , , , , ,
* * , ( ) .
although the scheme with a pure triad of bonding , non - bonding and anti - bonding modes is oversimplified , it visualizes sufficiently the expected modifications of mode positions and spectral weight with increasing coupling .
certainly , the higher - energy external mode at about 172 @xmath11 has not a pure non - bonding character and it also experiences a small level shift and modifications of its spectral shape through its interaction with the other modes . | we present optical measurements of phononic excitations in la@xmath0ca@xmath1mno@xmath2 ( lcmo ) and la@xmath0sr@xmath1mno@xmath2 ( lsmo ) thin films covering the full temperature range from the metallic ferromagnetic to the insulating paramagnetic phase .
all eight phonons expected for the r@xmath3c symmetry in lsmo and 17 out of the expected 25 phonons for the pnma symmetry in lcmo have been determined .
close to the ferromagnetic - to - paramagnetic transition both compounds reveal an anomalous behavior but with different characteristics .
anomalies in the phononic spectra are a manifestation of the coupling of lattice degrees of freedom ( dof ) to electronic dof .
specifically , the low - frequency external group proves to be an indicator for lattice modifications induced by electronic correlations .
the enhanced electron - phonon coupling in lcmo is responsible for fano - like interference effects of distinct phonon modes with electronic continuum excitations : we observe asymmetric phonon line shapes , mode splitting and spectral weight transfer between modes . |
the intrinsic high - energy emission from extragalactic sources is attenuated by the pair - production process ( @xcite , @xcite , @xcite ) .
photons in the tev range ( @xmath4 ) interact with low
energy photons from the extragalactic infrared background ( @xmath5 ) and produce electron positron pairs ( @xmath6 ) .
the total absorption depends on the history of the reaction rate along the line of sight , and therefore , on the distance travelled and on the local density of the low - energy target photon field , both a function of redshift @xmath7
. this process will affect the high energy tev end of the observed blazar spectra . since the optical depth depends strongly on the incident tev
photon energy the process leads to a steepening of the observed spectrum with respect to the intrinsic ( unabsorbed ) one .
thus , the observed spectra from distant blazars contain information on both the ir background radiation field history and the intrinsic properties of the source . from the viewpoint of blazar studies
the latter is just an undesirable extrinsic effect that must be corrected for .
in contrast , from the opposite perspective , it allows the measurement of the ir background . in both cases
these two independent processes have to be disentangled by either modeling both simultaneously ( e.g. @xcite , @xcite ) , or measuring one independently ( see @xcite for a review ) .
unfortunately it proved very difficult to constrain the ir background independently , and different groups tried to constrain models for the ir background from tev observations _ assuming _ that the blazar spectrum is a well - behaved power
law ( e.g. @xcite , @xcite ) .
the best perspective for such studies is offered by high
redshift blazars , since the absorption will be strong .
recently , @xcite presented a study along these lines , based on the tev spectra of high redshift blazars recently measured by the h.e.s.s .
they de reddened the observed tev data of 1es 1101 - 232 with different models of the ir backgrounds , and concluded that the resulting intrinsic spectrum is _ always _ very hard .
the index of the intrinsic spectrum should be equal to @xmath8 if the level of the ir background is close to the absolute lower limit obtained from direct integration of the light from resolved galaxies @xcite .
larger ir background levels should correspond to intrinsic tev spectra harder than @xmath8 . note that this spectral index is already harder than the spectral index observed in other nearby tev sources and that the difference will be more important for larger background levels .
@xcite also commented on the theoretical difficulties to explain spectra harder than @xmath8 , which would imply quite unexpected physics .
these facts led @xcite to prefer the solution of minimal ir background . in this paper
we show instead that spectra harder than @xmath8 can be obtained in
normal " one zone homogeneous synchrotron self
compton ( ssc ) models , as long as there is a deficit of low energy electrons with respect to the extrapolation from higher energies .
this can be achieved by assuming an emitting particle distribution which is a double power law , with a flat slope ( @xmath9 ) of the first , low energy portion , or assuming a simple power law , but with a relatively high low energy cut off . in these cases
the limits on how hard a tev spectrum can be is @xmath10 , while well accounting for a much softer x ray spectrum .
we discuss this result with the aim of finding limits on the ir background and also in the context of explaining the other tev sources ( with a softer intrinsic spectrum ) with the same model .
the simplest possible scenario able to explain the x ray and tev emission of a blazar assumes a homogeneous , spherical source filled by tangled magnetic field and relativistic electrons . since in tev blazars
we must have electrons of very large energies , the scattering process occurs partly in the klein nishina regime . in this regime
the scattering process becomes inefficient , and for typical electron distributions the ssc process can be well described by approximating the scattering cross section with a step function , equal to the thomson cross section for incident photon energies @xmath11 , and zero otherwise .
this means that electrons of increasingly large energies scatter an increasingly smaller fraction of synchrotron seed photons , steepening the resulting ssc spectrum with respect to the canonical " value of @xmath12 .
note that this steepening introduced by the klein
nishina effects is virtually inescapable , since the scattering occurs completely in the thomson limit only for extremely large values of the beaming factor : @xmath13 ( see e.g. @xcite ) . in the next section
we will find what is the high energy spectral index corresponding to a given particle energy distribution , assuming usual " parameters for our tev source , namely : source radius @xmath14 [ cm ] , magnetic field intensity @xmath15 [ g ] , doppler factor @xmath16 and approximating the particle energy distribution by a broken power law with the index @xmath17 [ @xmath18 below the break energy ( @xmath19 ) and @xmath20 above the break [ @xmath21 , where @xmath22 is the density of the particles with @xmath23 .
if the minimal @xmath24 is greater than unity ( @xmath25 ) the density of the particles with minimum energy ( @xmath26 ) is given by @xmath27 .
a flat slope of the electron distribution ( @xmath28 ) may help to explain the intrinsic tev spectrum of 1es 1101232 .
assuming the lowest possible absorption level , @xcite have estimated the intrinsic spectral index @xmath29 for this particular source .
we verified that it is roughly possible to reproduce such a spectrum if we assume that @xmath30 ( note that in the thomson limit @xmath8 would correspond to @xmath31 . ) . in this case , however , we do not obtain a power law intrinsic tev spectrum .
the spectrum is curved , with @xmath32 changing from @xmath33 in the gev range to @xmath34 at the peak of the emission at a few tev .
since @xmath30 can explain the spectrum derived for the minimal ir background case , we then ask if a smaller value of @xmath17 , producing an harder spectrum , can allow for a larger level of the ir background . in fig .
[ fig_n1 t ] we show different synchrotron and inverse compton spectra for different @xmath17 indices .
the spectral index of the first part of the thin synchrotron emission ( @xmath35 ) depends directly on @xmath17 [ @xmath36 down to the limiting value @xmath37 . below this limit
the index remains constant @xmath38 .
this is due the fact that , below the limit , the dominant part of the low frequency emission is generated by electrons with @xmath39 .
therefore , the spectrum in the discussed range is equivalent to the spectrum produced by a monoenergetic population of electrons , which can be approximated by a power law with @xmath10 below the maximum of the emission ( e.g. @xcite ) .
this means that even in the extreme case of the truncated , single power law energy distribution ( @xmath40 ) there is always a `` source '' of low energy photons due to the low frequency emission ( @xmath41 ) of the high energy electrons ( @xmath39 ) .
the spectral index of the inverse compton emission in the thompson regime is equivalent to the index of the target radiation field ( e.g. @xcite ) . since @xmath35 is limited to 1/3 also the index of the inverse
compton emission below the peak is limited to this value . moreover ,
this index does not depend on @xmath20 .
the index of the inverse
compton emission above the peak is given by @xmath42 , where @xmath43 .
this relationship can be easily derived from the approximation of the scattering derived by @xcite . to directly check the applicability of the ssc scenario that assumes @xmath44
we have applied our computations to the emission of 1es 1101232 observed by @xmath45sax @xcite and h.e.s.s .
@xcite experiments .
the tev data ( here and in all the other cases in the following ) have been unabsorbed according to the `` best - fit '' model proposed recently by @xcite .
the predictions of this model are also consistent with those of the model proposed by @xcite .
the model provides a significantly stronger absorption level than the minimal possible absorption suggested by @xcite .
however , our best fit for the expected intrinsic spectrum , in a framework of the above mentioned ssc scenario , has been obtained for the limiting case ( @xmath40 ) .
this indicates that the absorption is close to the highest possible value that can be accommodated by this scenario .
if the absorption is stronger , as suggested by some models ( e.g. @xcite ) , then our ssc scenario can not explain the intrinsic tev emission of 1es 1101232 .
the detailed values of the physical parameters used in our computations are discussed in the last section of this work , where we apply this scenario also to other sources .
the truncated particle energy distribution corresponds to the limit @xmath46 of a broken power law . for clarity
, we will refer in the following to the break energy as @xmath47 .
for such a electron distribution , the shape of the tev emission depends strongly on @xmath47 , and we will therefore investigate this is some detail . as we have shown in the previous subsection a particle distribution with a low energy cut - off provides the best explanation for the emission of 1es 1101232 .
we here show the effects of changing the value of the low energy cut off , which we will call @xmath47 .
the result of our calculations is shown in fig .
[ fig_gmt ] .
we have applied this test for the @xmath45sax @xcite and cat @xcite observations of mrk 501 from 16@xmath48 april , 1997 . for this test
, we fix the slope of the particle distribution to @xmath49 , which provides the best fit for the used x
ray observations of mrk 501 . in the first approach
we assume the minimal possible value of @xmath50 .
this means that the model is equivalent to the standard ssc scenario and is providing a relatively broad peak for the inverse
compton emission . nevertheless , this approach is able to roughly explain the unabsorbed tev spectrum .
however , the increase of the @xmath47 value by about three orders of magnitude provides a slightly narrower peak and therefore better explanation for the spectrum .
a value of @xmath47 greater by one more order of magnitude is the limiting value if we want to explain the entire x
ray spectrum .
however , this already high value of @xmath51 is drastically modifying the tev spectrum , where the limiting slope ( @xmath52 ) of the emission becomes well visible ( blue line in fig .
2 ) . finally ,
monoenergetic electron distribution can not explain the x ray nor the tev emission .
note that for values of @xmath53 close to @xmath47 , the second part of the tev spectrum , above the peak , is curved and it is not a power law .
we have already applied our specific ssc scenario to 1es 1101232 and mrk 501 . in this section
we apply this model to two other tev sources , mrk 421 and pks 2155304 , and compare the results .
the rxte and hegra observations of mrk 421 @xcite can be fitted assuming that the x ray emission is done by the exponential synchrotron tail of the highest energy particles at @xmath54 .
therefore , the index of the particle energy spectrum and @xmath47 are not constrained by the data .
the same approach can explain the rxte and h.e.s.s .
observations of pks 2155304 @xcite .
note that it is also possible to explain these observations , as well as the observations of mrk 501 , assuming @xmath55 .
we performed our computations to show that the assumption of having @xmath56 ( required to explain the emission for 1es 1101232 ) , works very well also for other sources . in fig .
[ fig_lum ] we show the results of our modeling and give the values of the input parameters .
note that for mrk 501 we provide also the spectral fits for the observations made on april 7@xmath48 , 1997 .
moreover , in the case of mrk 421 , mrk 501 and pks 2155304 the x - ray and tev observations were made almost simultaneously , whereas for 1es 1101232 we use the x ray spectrum from january 1997 and the tev data collected over march - june 2004 - 2005 .
note that the value of the magnetic field used in our calculations is significantly smaller that the value estimated from the equipartition between the magnetic field energy density and the electron energy density .
we have shown that the standard ssc model can explain the specific observations of the relatively distant tev blazar 1es 1101232 . assuming a truncated power law energy distribution , with @xmath56
, we can well explain the intrinsic tev spectrum even assuming a stronger absorption than the upper limit suggested by @xcite .
this assumption appears in very good agreement with the recent estimates of the ir intergalactic background ( @xcite , @xcite ) .
the limit on how hard an ssc spectrum can be is @xmath10 .
this limit in turn translates in a limit on the amount of absorption suffered by tev photons , and hence on the level of the ir background . concerning the reason of having @xmath56
, we can propose two answers .
the first is the injection of an already truncated power law spectrum from a shock region .
if , in addition , the cooling process is inefficient for the the particles at lower energies ( close to @xmath47 ) , the particle distribution will not develop a low energy tail in a reasonable " time , namely a dynamical time , over which we can assume that the physical conditions ( size , densities , magnetic field ) are constant ( since the source has bulk motion , after a dynamical time it will be located in a larger part of the jet , and not contribute any longer to the observed flux ) .
note that in our particular simulations also particles with @xmath57 are not affected by the cooling in one crossing time ( @xmath58 ) . in a second , more complex
, scenario all particles are accelerated efficiently at the shock front almost up to some maximal energy .
then they escape downstream of the shock and cool .
however , the cooling process may be compensated ( at least in part ) by some stochastic turbulent acceleration @xcite .
the competition between re - acceleration ( heating ) and cooling may single out an equilibrium energy where the two processes balance .
this equilibrium energy may correspond to the @xmath47 assumed in our simple ssc scenario .
the changes of @xmath47 and @xmath53 may have various impacts for the source variability , especially for the correlation between the x ray and tev emission .
as an example , when @xmath59 and @xmath60 the change of @xmath47 may not affect the x ray emission while having a strong impact for the tev radiation .
such a scenario could explain the orphan tev flare in 1es 1959 + 650 @xcite .
on the contrary , a change of @xmath53 would produce a significant x ray variability without any change in the tev flux below the tev peak .
we are grateful to a. djannati - atai , h. krawczynski and b. giebels for the data obtained by the cat , hegra and h.e.s.s . experiments .
we acknowledge the ec funding under contract hprcn - ct-2002 - 00321 ( enigma network ) .
jg thanks for the hospitality at the osservatorio di brera . | recent gamma ray observations of the blazar 1es 1101232 ( redshift @xmath0 ) reveal that the unabsorbed tev spectrum is hard , with spectral index @xmath1 [ @xmath2 . we show that simple one zone synchrotron self
compton model can explain such hard spectra if we assume a power law energy distribution of the emitting electrons with a relatively high minimum energy
. in this case the intrinsic tev spectrum can be as hard as @xmath3 , while the predicted x ray spectrum can still be much softer .
the observations of 1es 1101232 can therefore be reconciled with relatively high intensities of the infrared background , even if some extreme background levels can indeed be excluded .
we show that the other tev sources ( mrk 421 , mrk 501 & pks 2155304 ) can be interpreted in the same framework , with a somewhat larger minimum energy .
[ firstpage ] radiative transfer
bl lacertae objects : individual : mrk 421 , mrk 501 , pks 2155304 , 1es 1101232 |
we introduce new mathematical techniques for analyzing complex spatiotemporal nonlinear dynamics and demonstrate their efficacy in problems from two different paradigms in hydrodynamics .
our approach employs methods from algebraic topology ; earlier efforts have shown that computing the homology of topological spaces associated to scalar or vector fields generated by complex systems can provide new insights into dynamics @xcite .
we extend prior work by using a relatively new tool called persistent homology @xcite .
complex spatiotemporal systems often exhibit complicated pattern evolution .
the patterns are given by scalar or vector fields representing the state of the system under study .
persistent homology can be viewed as a map @xmath0 that assigns to every field a collection of points in @xmath1 , called a _ persistence diagram_. for a given scalar field @xmath2 , the points in the persistence diagram @xmath3 encode geometric features of the sub - level sets @xmath4 for all values of @xmath5 .
a feature encoded by the point @xmath6 appears in @xmath7 for the first time and disappears in @xmath8 .
therefore , @xmath9 and @xmath10 are called birth and death coordinates of this feature .
the lifespan @xmath11 indicates the prominence of the feature . in particular
, features with long lifespans are considered important and features with short lifespans are often associated with noise .
thus , the persistence diagram is a highly simplified representation of the field generating the pattern .
the space of all persistence diagrams , @xmath12 , can be endowed with a variety of metrics under which @xmath0 is a continuous function .
this has several important implications that we exploit in this paper .
first , continuity implies that small changes in the field pattern , e.g. bounded errors associated with measurements or numerical approximations , lead to small changes in the persistence diagrams .
second , by using different metrics , we can vary our focus of interest between larger and smaller changes in the persistence diagrams .
moreover , by comparing different metrics , we can infer if the changes in a pattern affect geometric features with longer or shorter life spans . finally , since
, applying the map @xmath0 to a time series of patterns produces a time series in @xmath12 , the distance between the consecutive data points in @xmath12 can be used to quantify the average rate at which the geometry of the patterns is changing .
as mentioned above , the dynamics of spatiotemporal systems are characterized by the time - evolution of the patterns corresponding to the fields generated by the system . however , capturing these vector fields , either experimentally or numerically , results in multi - scale high dimensional data sets . in order to efficiently analyze these data sets ,
a dimension reduction must be performed .
we use persistent homology to perform nonlinear dimension reduction from a time series of patterns to a time series of persistence diagrams .
we show that this reduction can cope with redundancies introduced by symmetries ( both discrete and continuous ) present in the system .
in particular , this approach directly quotients out symmetries and , thereby , permits easy identification of solutions that lie on a group orbit .
separately , we also apply persistent homology to extract information about dynamical structures in the reduced data . characterizing dynamics in the space of persistence diagrams
can not be done using conventional methods ( e.g. , time delay embeddings ) , since choosing a coordinate system in @xmath12 is currently an open problem . however , since @xmath12 is a metric space , the geometry of the point cloud @xmath13 , generated by the time series of the reduced data , is encoded by a scalar field which assigns to each point in @xmath12 its distance to @xmath13 .
we show how persistent homology may be applied to describe dynamics by characterizing the geometry of @xmath13 .
an outline of the paper is as follows . in section [ sec : systems ] we present a brief overview of the two fluid flows examined in this paper : ( 1 ) kolmogorov flow and ( 2 ) rayleigh - bnard convection .
we note here , for emphasis , that while persistent homology can be applied to vector fields , it will be sufficient for this paper to focus on scalar fields drawn from these systems ( specifically , one component of the vorticity field for kolmogorov flow , and the temperature field for rayleigh - bnard convection ) . in section [ sec : ph * ]
we discuss key issues related to the application of persistent homology . by now , the mathematical theory of persistent homology is well developed .
therefore , our main emphasis is on the computational aspect of passing from the data to the persistence diagrams .
section [ sec : ph ] describes the correspondence between the geometric features of a scalar field and the points in its corresponding persistence diagram .
section [ sec : spaceofpd ] discusses the structure of the space @xmath12 and the properties of the associated metrics . in sections [ sec : distances ] and [ sec : analyzingpointcloud ] we discuss how these metrics can be used to analyze dynamics .
first , we interpret distance between the persistence diagrams representing the consecutive data points in the time series as a rate at which geometry of the corresponding scalar fields is changing .
second , we motivate and explain the procedure for extracting the geometric structure of the point cloud in @xmath12 .
we close the paper by applying the developed techniques to the following problems . in section [ sec : fixedpoints ] , we identify distinct classes of symmetry - related equilibria for kolomogorov flow . in section [ sec : periodicorbit ] , we show that a relative periodic orbit for kolmogorov flow collapses to a closed loop in @xmath12 .
finally , in section [ sec : rbc ] , we deal with identifying recurrent dynamics that occur on different time scales in our study of rayleigh - bnard convection flow .
for the study of turbulence in two dimensions , kolmogorov proposed a model flow where the two - dimensional ( 2d ) velocity field @xmath14 is given by @xmath15 ( with @xmath16 and @xmath17 ) , where @xmath18 is the pressure field , @xmath19 is the kinematic viscosity , @xmath20 is fluid density , and @xmath21 is the forcing that drives the flow @xcite .
laboratory experiments in electromagnetically - driven shallow layers of electrolyte can exhibit flow dynamics that are well - described by equations ( [ eq : q2dns ] ) with appropriate choices of @xmath22 and @xmath23 to capture three - dimensional effects , which are commonly present in experiments @xcite . in this paper
, we refer to all models described by equations ( [ eq : q2dns ] ) ( including experimentally - realistic versions ) as kolmogorov flows .
it is convenient to use the vorticity - stream function formulation @xcite to study kolmogorov flow analytically and numerically .
equations ( [ eq : q2dns ] ) , written in terms of the z - component of the vorticity field @xmath24 , a scalar field , take the form @xmath25 for the current study , we choose @xmath26 , @xmath27 m@xmath28/s , @xmath29 s@xmath30 , @xmath31 kg / m@xmath32 , and @xmath33 m. we express the strength of the forcing in terms of a non - dimensional parameter , the reynolds number @xmath34 .
equation ( [ eq : q2dvor_nd ] ) is solved numerically by using a pseudo - spectral method @xcite , assuming periodic boundary conditions in both @xmath35 and @xmath36 directions , i.e. , @xmath37 , where @xmath38 m and @xmath39 m are the dimensions of the domain in the @xmath35 and @xmath36 directions , respectively .
it is important to note that equation ( [ eq : q2dvor_nd ] ) , with periodic boundary conditions , is invariant under any combination of three distinct coordinate transformations : ( 1 ) a translation along @xmath35 : @xmath40 , @xmath41 $ ] ; ( 2 ) a rotation by @xmath42 : @xmath43 ; and ( 3 ) a reflection and a shift : @xmath44 . because of these symmetries , each particular solution to equation ( [ eq : q2dvor_nd ] ) generates a set of solutions which are dynamically equivalent .
physically , invariance under continuous translation leads to the existence of relative equilibria ( req ) and relative periodic orbit ( rpo ) solutions , in addition to equilibria ( eq ) and periodic orbit ( po ) solutions .
for @xmath45 , the flow is characterized by a steady rpo ; figure [ fig : projections](a ) shows a projection , plotted using the three dominant fourier modes of this rpo .
the rpo has a period 34.78 seconds and a drift speed @xmath46 m / s .
the tunnel - like structure is a result of the periodic motion superposed over the slow drift along the @xmath35-direction . for larger
forcing ( @xmath47 ) , the flow becomes weakly turbulent , as can be seen from the fourier projections in figure [ fig : projections](b ) .
the turbulent dynamics in this regime are of great interest as the flow explores a region of the state space which contains weakly " unstable eq , po , req , and rpo solutions .
recent theoretical advances have shown that the identification of these solutions could aid the understanding of weakly turbulent dynamics @xcite .
for instance , if the turbulent trajectory is close to an eq solution ( @xmath48 ) , which is characterized by @xmath49 , we would expect the instantaneous rate of change of @xmath50 to be relatively small , i.e. , @xmath51 .
similarly , a close pass to a po solution would mean @xmath52 , where @xmath53 is the period of the po that is guiding the dynamics of the turbulent trajectory .
the turbulent trajectory depicted in figure [ fig : projections](b ) passes close to unstable eq and req solutions which are indicated by the red dots .
rayleigh - bnard convection is a canonical pattern forming system that has been used to gain many new fundamental insights into the spatiotemporal dynamics of systems that are driven far - from - equilibrium @xcite .
rayleigh - bnard convection is the buoyancy driven fluid flow that occurs when a shallow layer of fluid is heated uniformly from below in a gravitational field .
the dynamics are governed by the boussinesq equations , @xmath54 where @xmath55 is a vector field of the fluid velocity , @xmath56 is the pressure field , and @xmath57 is the temperature field . in our notation ,
the origin of the cartesian coordinates @xmath58 at the center of the domain are at the lower heated plate where @xmath59 is a unit vector opposing gravity .
equations ( [ eq : rbc1])-([eq : rbc3 ] ) represent the conservation of momentum , energy , and mass , respectively .
the equations have been nondimensionalized using the vertical diffusion time of heat as the time scale , the layer depth as the length scale , and the constant temperature difference between the lower and upper plates as the temperature scale . in our work ,
we consider rayleigh - bnard convection in a shallow domain with a cylindrical cross - section .
the no - slip fluid boundary condition @xmath60 is applied to all material surfaces .
the lower and upper plates are held at a constant temperature where @xmath61 and @xmath62 , respectively .
the lateral sidewalls of the cylindrical container are assumed to be perfectly conducting , which yields @xmath63 .
the dynamics can be described using three non - dimensional parameters .
the rayleigh number @xmath64 represents the ratio of buoyancy to viscous forces . at the critical value @xmath65 , an infinite layer of fluid undergoes a bifurcation to straight and parallel convection rolls . for increasing values of the rayleigh number @xmath66 , the dynamics become periodic , chaotic , and eventually turbulent .
the prandtl number @xmath67 is the ratio of the momentum and thermal diffusivities . for typical experiments using compressed gasses , @xmath68 .
lastly , the aspect ratio of the cylindrical domain @xmath69 is the ratio of the domain s radius to its depth .
we numerically integrate eqs .
( [ eq : rbc1])-([eq : rbc3 ] ) using a highly parallel spectral element algorithm that has been tailored for the study of convection ( c.f .
figure [ fig : klf](b ) shows a typical pattern from a numerical simulation of rayleigh - bnard convection . in this simulation , @xmath70 , @xmath71 , and
the aspect ratio of the domain is @xmath72 .
the numerical simulation is initiated from a field of small random perturbations to the temperature field and is integrated for long times .
figure [ fig : klf](b ) illustrates the fluid temperature field at the horizontal mid - plane ( @xmath73 ) , where light is warm rising fluid and dark is cool falling fluid . this image is a snap shot in time of a time - dependent pattern where the dynamics are nearly periodic in time .
the pattern shown does not include the region near the sidewall .
specifically , a distance of one - layer depth from the lateral sidewall is not shown ( this distance is approximately the width of a convection roll ) .
this is done to remove the complex fluid flow that occurs in the small region near the sidewalls to allow our diagnostics to focus upon the bulk patterns and dynamics ( c.f .
the aim of this paper is to introduce an approach for analyzing the dynamics of the pattern evolution in spatiotemporal systems .
this is done in two steps .
first , we perform nonlinear data reduction , and then we extract information about the dynamical structures from this reduced data . we formulate both of these tasks in terms of analyzing the structure of the sub - level sets of a scalar function @xmath74 , where @xmath75 is a topological space .
tools from algebraic topology , homology in particular , are used to capture and quantify the geometry of the sub - level sets .
recall that given any topological space @xmath76 , homology assigns to @xmath76 a sequence of vector spaces @xmath77 , @xmath78 .
the dimension of @xmath77 is called the @xmath79-th _ betti number _ and is denoted by @xmath80 .
betti numbers provide geometric information about @xmath13 : @xmath81 is the number of connected components , or pieces , of @xmath76 ; @xmath82 indicates the number of loops or tunnels in @xmath76 ; and @xmath83 is the number of cavities .
our goal is to understand structure of the sub - level sets @xmath84 for all values of @xmath85 . as we vary @xmath5 , the number of components , loops , and cavities in @xmath86 changes , implying that @xmath87 , @xmath88 , also changes .
( see section [ sec : ph ] for examples . )
what is remarkable is that , under very weak conditions , we can choose bases for the vector spaces @xmath89 over all values of @xmath5 such that , given a basis element of @xmath89 , we can identify a unique value @xmath9 at which this basis element appears and a unique value @xmath10 at which this basis element disappears .
we refer to @xmath9 as the _ birth value _ , @xmath10 as the _ death value _ , and the pair @xmath90 as a _ persistence point _ corresponding to the chosen basis element of @xmath89 .
the difference @xmath91 is called the _ life span _ of the persistence point .
longer life spans are associated with geometric features that persist through larger variations of @xmath5 , and persistence diagrams are a codification of this information . given a scalar field @xmath92 , the set of associated _ persistence diagrams _ are denoted by @xmath93 , where @xmath94 consists of all persistence points corresponding to the @xmath79-th level of homology ( keeping track of multiple copies of a single point ) , along with infinitely many points at each point along the diagonal @xmath95 .
the reason for the inclusion of the diagonal is made clear in definition [ defn : bottleneck ] , when we define metrics on the space of persistence diagrams .
for the systems introduced in section [ sec : systems ] , we first use persistent homology as a nonlinear data reduction method .
for kolmogorov flow we study the scalar field @xmath96 , the @xmath97-component of the vorticity field , while for rayleigh - bnard convection we study the scalar field @xmath98 , the temperature field at the mid - plane .
it is important to note that the domains for these two scalar fields are different . for kolmogorov flow ,
the domain @xmath75 is a torus since we are using periodic boundary conditions , while for rayleigh - bnard convection , @xmath75 is a disk . for the disk ,
we need only to concern ourselves with the vector spaces @xmath99 for @xmath100 .
however , for the torus , the vector spaces @xmath101 also need to be considered , since the torus encloses a three - dimensional cavity . in section [ sec : ph ] , we explain how the persistence diagrams @xmath3 capture important information about the patterns given by the scalar fields @xmath50 and @xmath53 .
the set of all persistence diagrams @xmath0 is a metric space , denoted by @xmath12 ( see section [ sec : spaceofpd ] ) .
since we are studying the evolution of kolmogorov flow and rayleigh - bnard convection , we have time series of the vorticity @xmath102 and temperature @xmath103 fields , and , therefore , we have time series of persistence diagrams @xmath104 and @xmath105 .
we view each of these time series as a point cloud @xmath106 . to extract information about dynamical structures present in the time series , we use persistent homology a second time to quantify the geometry associated with this point cloud .
this is achieved by introducing a new scalar function @xmath107 that gives the distance from any point in @xmath12 to the point cloud @xmath13 and is defined by @xmath108 where @xmath109 is an appropriate metric on the space of persistence diagrams .
the associated sub - level sets @xmath86 are once again given by .
to carry out the steps mentioned above requires the ability to compute the persistence diagrams @xmath3 . to do this
, we need to calculate @xmath89 , which requires a discrete representation of @xmath86 called a _
complex_. in the context of nonlinear data reduction , we make use of a _ cubical _ complex . when analyzing the geometry of the point cloud , we approximate @xmath86 using a _ vietoris - rips _ complex , which is a special form of a simplicial complex .
this is a classical subject and thus there are a variety of references providing precise definitions of complexes , e.g. @xcite for vietoris - rips complexes and @xcite for cubical complexes , discussions of issues related to approximations @xcite , and how one proceeds from a complex to computing persistent homology @xcite .
the homological computations in this paper were performed using the perseus software @xcite .
the numerical data for the vorticity and the temperature fields is presented in the form of piecewise - constant functions defined on a rectangular lattice .
for kolmogorov flow , values of @xmath50 are reported in double precision .
recall that the vector spaces @xmath99 can only change for @xmath110 , where @xmath111 is the finite set of values that @xmath50 attains on the given lattice .
each of the sets @xmath112 is a cubical complex , and we use the perseus software to compute the corresponding persistence diagrams using only the values @xmath110 . numerical simulations for rayleigh - bnard convection are carried out with high precision as well .
however , keeping in mind our goal to compare the numerical simulations with experimental data , we convert the temperature field to an @xmath113-bit temperature field @xmath114 ( an integer - valued function with values between @xmath115 and @xmath116 ) , which can be obtained experimentally . consequences of this rescaling are examined in section [ sec : distances ] .
the purpose of this section is to provide intuition and interpretation of the information that persistence diagrams present .
as indicated in the previous section , we are interested in the diagrams @xmath117 , @xmath118 , of the vorticity field for kolmogorov flow , and the diagrams @xmath119 , @xmath120 of the temperature field for rayleigh - bnard convection , shown in figure [ fig : klf ] .
+ + + we begin by discussing @xmath121 , shown in figure [ fig : pds_kolmogorov](e ) , computed from a single time snapshot of the vorticity field for the kolmogorov flow .
the minimum value of the vorticity field is @xmath122 , and therefore , @xmath123 for all @xmath124 . at @xmath125 ,
two components appear , indicated by the two persistence points with birth value @xmath126 .
the death value of one of these two persistence points is @xmath127 , and so the two components merge at this value , resulting in a single component .
this explains the persistence point @xmath128 .
the reason the other persistence point is denoted by @xmath129 , with @xmath130 , is because when features merge , a choice must be made about which topological feature ( in this case , a connected component ) dies .
having a consistent choice of basis over all values of @xmath5 requires that the homology generator associated with the geometric feature that has the larger birth value die first . if the birth values are the same , then it does not matter which topological feature with this birth value is chosen to be the one that persists .
in particular , this implies that the generator associated with one of these two initial components can never die .
figure [ fig : pds_kolmogorov](a ) indicates the subset of @xmath75 corresponding to @xmath131 .
we remind the reader that the domain @xmath75 for kolmogorov flow is a torus , since the left ( top ) and right ( bottom ) boundaries are identified . therefore , @xmath131 consists of eight distinct connected components instead of nine .
the existence of these eight connected components can also be extracted from @xmath121 , shown in figure [ fig : pds_kolmogorov](e ) .
observe that these connected components correspond to connected regions with birth value @xmath132 and death value @xmath133 . in figure
[ fig : pds_kolmogorov](e ) , this corresponds to the eight points in the rectangular region @xmath134 .
figure [ fig : pds_kolmogorov](b ) indicates that @xmath135 consists of four connected horizontal bands , which agrees with the number of persistence points in the rectangular region @xmath136 of @xmath121 .
each stripe is created as two distinct components present in figure [ fig : pds_kolmogorov](a ) grow and merge , causing one component to die each time .
the deaths of these components are captured by the points in the rectangle @xmath137 which are not in the rectangle @xmath138 , since these are components that are born before @xmath139 but die before @xmath140 .
three horizontal stripes merge together before @xmath141 , as indicated by two points inside the rectangle @xmath137 that are not in the rectangle @xmath142 .
the two remaining connected components merge together soon thereafter , and for all greater threshold values , there is only one connected component . to finish our analysis of @xmath121 , we turn our attention to the persistence points close to the diagonal .
these have very short life spans , which suggests that these features may be numerical artifacts . in our example , these points represent the narrow horizontal bands formed in between two connected components before they merge into a single band ( see video 1 available in the supplementary materials ) .
these narrow bands are formed by small oscillations of the vorticity field at the places where the field is almost constant .
we now turn our attention to the @xmath143 persistence diagram , which characterizes loops in @xmath112 .
appendix [ sec : homologyoftorus ] provides a detailed discussion of independent loops on a torus . from @xmath143 , we see that the first loop appears at threshold @xmath144 .
it corresponds to one of the four horizontal bands shown in figure [ fig : pds_kolmogorov](b ) .
each horizontal band generates a single independent loop , corroborated by the existence of four persistence points in the rectangle @xmath138 of @xmath143 .
we note that the full torus has two loops captured by homology .
this is expressed in @xmath143 by the two persistence points with @xmath130 .
observe that the first loop that appears at @xmath144 is equivalent to one of the toral loops , thus it can not be killed by any other loop , and hence is captured by the persistence point @xmath145 .
the other three loops present at @xmath140 correspond to the same toral loop and thus must die .
in fact , they do so by @xmath146 .
note that the birth values @xmath9 of these persistence points are close to the death values @xmath10 of the persistence points in @xmath147 of @xmath121 .
this implies that shortly after the components merge , they form horizontal bands across the entire domain .
new loops are also created as the bands start merging .
if two horizontal bands are connected by @xmath148 links , then the number of loops generated by this object ( two bands plus the links ) is @xmath149 .
thus , the first additional loop appears when a second link is created ( see appendix [ sec : homologyoftorus ] ) . in our example
, this happens near the threshold @xmath150 . in figure
[ fig : pds_kolmogorov](c ) , there are four distinct links between the two horizontal bands at the top of the figure .
the small punctures visible in figure [ fig : pds_kolmogorov](c ) are filled in almost immediately , and the four links merge into two distinct links .
the points in @xmath143 that are close to the diagonal capture this behavior .
the other two links are present for a wider range of thresholds , and the loop they generate is represented by one of the persistence points in @xmath143 with birth coordinate slightly smaller than @xmath151 .
the horizontal band at the top and the horizontal band at the bottom are linked in a similar fashion .
this explains the presence of another point with birth coordinate slightly smaller than @xmath151 . at @xmath152 , a connection from the top to the bottom boundary
is created .
this loop is homologically equivalent to the second of the two independent loops of the torus , and hence is identified by the persistence point @xmath153 . as the threshold passes the value @xmath154 , the punctures shown in figure [ fig : pds_kolmogorov](d ) start disappearing and the corresponding loops start dying .
again , there are @xmath155 independent loops for @xmath156 punctures .
since the maximum value of @xmath50 is @xmath157 , the sub - level set is the whole torus for any threshold above this , i.e. @xmath158 for all @xmath159 . in this case , there are no more punctures , and the rectangle @xmath160 contains only two persistence points . finally , we address the @xmath161 persistence diagram , not shown for brevity .
this diagram contains a single persistence point at @xmath162 .
the birth coordinate , @xmath9 , corresponds to the minimum value of @xmath5 for which @xmath163 , the whole torus . since @xmath163 for all @xmath164 , this point never dies .
we now discuss the persistence diagrams for the temperature field @xmath114 shown in figure [ fig : pds_panam ] for rayleigh - bnard convection .
again , beginning with @xmath165 , the points with short life spans correspond the large number of small connected components that make up @xmath166 , as shown in figure [ fig : pds_panam](a ) .
points with long life spans represent the well - defined connected components shown in figure [ fig : pds_panam](b ) . from the persistence diagram , we can see that these components merge almost simultaneously at two threshold values , @xmath167 and @xmath168 .
turning to @xmath169 , we note that the domain of the temperature field is a disk , so the independent loops correspond to punctures inside of the disk .
the diagram @xmath169 indicates that there are no loops with long life spans , and the loops that do appear do so roughly at the same threshold values at which the dominant components merge .
these features are due to the small fluctuations of the temperature field close to the critical values at which different rolls merge together .
this is consistent with their short life spans .
as explained in the previous section , a persistence diagram codifies , in a reasonably compact form , considerable information about the geometry of a scalar function . as suggested by the examples , we use persistence diagrams to provide a reduced description of the state of the dynamical system of interest at any given point in time .
therefore , to analyze the dynamics , we need to be able to compare one collection of persistence diagrams @xmath0 ( corresponding to a snapshot of the flow pattern at an instant of time ) to another collection of diagrams @xmath170 ( from another flow snapshot ) .
there are a variety of metrics that can be imposed on persistence diagrams @xcite .
the metrics used in this paper rely on pairing the points @xmath171 in a one - to - one correspondence ( bijection ) with the points in @xmath172 .
according to the definition , every persistence diagram contains an infinite number of copies of the diagonal .
hence , there are many different bijections @xmath173 between @xmath174 and @xmath172 . roughly speaking
, the distance between @xmath0 and @xmath170 is defined using the bijections that `` minimize the shift '' in the mapping of the points @xmath175 from @xmath174 to @xmath176 in @xmath172 .
this notion is made more precise in the following definition .
[ defn : bottleneck ] _ let @xmath177 and @xmath178 be two collections of persistence diagrams . the _ bottleneck distance _ between @xmath0 and @xmath170 is defined to be @xmath179 where @xmath180 and @xmath173 ranges over all bijections between persistence points .
similarly , the _ degree-@xmath175 wasserstein distance _ is defined as @xmath181^{1/p}. \label{eqn : dwdef}\ ] ] _ roughly speaking , a function @xmath182 is _ tame _ if , for every @xmath183 , the vector space @xmath184))$ ] is finite dimensional for every @xmath79 , and there are only finitely - many thresholds at which the vector spaces change ( for a precise definition see @xcite ) . for our purposes , it suffices to remark that if @xmath92 is a piecewise - constant function on a finite complex , then @xmath92 is tame .
in particular , the numerically - computed vorticity field @xmath50 and @xmath113-bit temperature field @xmath114 are tame functions . for the remainder of this paper
, we use @xmath185 to denote the set of persistence diagrams corresponding to @xmath186 and @xmath12 to denote the set of all persistence diagrams .
let @xmath187 denote the set of tame functions @xmath188 equipped with the @xmath189 norm .
a fundamental result @xcite is that , using the wasserstein or bottleneck metrics , @xmath190 is a lipschitz - continuous function .
in particular , if @xmath191 , then @xmath192 these results on lipschitz continuity have two important implications for this work , both stemming from the fact that our analysis is based on numerical simulations .
assume for the moment that @xmath188 denotes the exact solution at a given time to either kolmogorov flow or the boussinesq equations .
ideally , we want to understand @xmath3 .
our computations of persistent homology are based on @xmath193 , a cubical complex defined in terms of the numerically - reported values @xmath194 , where @xmath195 represents the associated piecewise - constant function .
if the numerical approximation @xmath195 satisfies @xmath196 , then by we have a bound on the bottleneck distance between the actual persistence diagram @xmath3 and the computed persistence diagram @xmath197 , so that @xmath198 .
figure [ fig : bottleneck ] provides a schematic justification of this claim .
as indicated in the introduction , persistent homology is invariant under certain continuous deformations of the domain . to be more precise , if @xmath199 is a homeomorphism and @xmath188 , then @xmath200 .
of particular relevance to this paper is a function @xmath201 which arises as a symmetric action on the domain . in this paper
, we work with piecewise - constant numerical approximations of the actual functions of interest , and we can not assume that this equality holds .
however , if @xmath92 is given and @xmath202 , where @xmath201 is as above , and we have an @xmath189 bound @xmath203 on the difference between the approximation and the true function , then by , @xmath204 in summary , under the assumption of bounded noise or errors from numerical simulations ( or experimental data ) , we have explicit control of the errors of the distances in @xmath12 .
the goal of this section is twofold : one , to provide intuition about the information contained in the different metrics , and two , to suggest how viewing a time series in @xmath12 can provide insight into the underlying dynamics .
we begin by remarking that the bottleneck distance @xmath205 measures only the single largest difference between the persistence diagrams and ignores the rest .
the wasserstein distance @xmath206 includes all differences between the diagrams .
thus , it is always true that @xmath207 the sensitivity of the wasserstein metric to small differences ( possibly due to noise ) can be modulated by the choice of the value of @xmath175 , i.e. if @xmath208 , then one expects @xmath206 to be less sensitive to small changes than @xmath209 . in this paper , we restrict ourselves to the bottleneck distance @xmath205 and the wasserstein distances @xmath206 for @xmath210 .
the most obvious use of these metrics is to identify or distinguish patterns . as an example , we consider patterns along an orbit from the kolmogorov flow . as indicated in section [ intro_kolmogorov ] , this particular trajectory arises from a periodic orbit with a slow drift along an orbit of continuous symmetry .
in particular , we consider the three time points indicated in figure [ fig : projections](a ) : two that appear to differ by the continuous symmetry , and a third that lies on the ` opposite ' side of the periodic orbit
. plots of the associated vorticity fields at these points ( see figure [ fig:3states ] ) agree with this characterization of the time points .
we want to identify this information through the associated persistence diagrams @xmath211 , @xmath212 , and @xmath213 , shown in figure [ fig:3statesdiagram ] .
indeed , the plots of @xmath214 and @xmath215 are difficult to distinguish , but @xmath216 is clearly distinct . to quantify this difference , we make use of the distances between the persistence diagrams using @xmath205 , @xmath217 , and @xmath218 .
these values are recorded in table [ table : threenetworksfulldistance ] . not surprisingly , the distances between @xmath211 and @xmath212 are much smaller than the distances between @xmath211 and @xmath213 .
we want to use these distances , as opposed to the detailed information in the persistence diagrams , to obtain rough information about how the pattern at figure [ fig:3states](a ) differs from the pattern at figure [ fig:3states](c ) .
persistence diagrams @xmath219 and @xmath213 for the vorticity fields shown in figure [ fig:3states ] .
the points in @xmath211 and @xmath212 are almost identical because the corresponding vorticity fields are similar .
the points in @xmath213 are more spread out and do not shadow the points in @xmath211 so well .
the same is true for the @xmath220 persistence diagrams which are not shown .
so @xmath221 , for @xmath222 , as indicated by table [ table : threenetworksfulldistance ] . ,
title="fig:",width=163 ] persistence diagrams @xmath219 and @xmath213 for the vorticity fields shown in figure [ fig:3states ] .
the points in @xmath211 and @xmath212 are almost identical because the corresponding vorticity fields are similar .
the points in @xmath213 are more spread out and do not shadow the points in @xmath211 so well .
the same is true for the @xmath220 persistence diagrams which are not shown .
so @xmath221 , for @xmath222 , as indicated by table [ table : threenetworksfulldistance ] .
, title="fig:",width=163 ] .distances between selected persistence diagrams ( rounded to @xmath223 decimal places ) shown in figure [ fig:3statesdiagram ] , corresponding to the vorticity fields given by figure [ fig:3states ] . [ cols="^,^,^,^",options="header " , ]
as mentioned in section [ sec::largedata ] , characterizing the geometry of a continuous trajectory becomes a challenge in the setting of dynamics with multiple time scales . to demonstrate this
, we consider the numerical simulation of rayleigh - bnard convection , where from multiple perspectives it appears that the trajectory is close to a periodic orbit and that the rate of change in the patterns of the temperature field is far from constant .
this can be clearly seen visually ( see video 6 , 7 , or 8 in the supplementary materials ) . moreover , both the speed plot , figure [ fig : consecdist](b ) , and the distance matrix , figure [ fig : distancematrixperiodicorbit](b ) , suggest recurrent dynamics . however , we note that the rate of change , especially using the bottleneck distance , is typically small except for short periods of time at which the speed spikes .
the distance matrix has a distinct checkerboard pattern , with the edges corresponding to the spikes , again indicating a rapid and large change in location in the space of persistence diagrams .
the maximum bottleneck distance between the consecutive sampling points is @xmath224 ( figure [ fig : consecdist](b ) caption ) , while the diameter of the point cloud is only @xmath225 ( figure [ fig : distancematrixperiodicorbit](b ) ) .
therefore , we expect that significant portions of the trajectory are missing .
indeed , figure [ fig : persistencediagramsrbc500](a ) shows that there are several persistence points in @xmath226 with a ( finite ) death coordinate larger than ten .
thus , at a length scale of @xmath227 ( which is forty times larger than the noise threshold ) , the sample of the trajectory is broken into several pieces .
the largest gap between different pieces of the trajectory is @xmath228 , as indicated by the persistence point with coordinates @xmath229 .
this means that the sampling rate is far from adequate .
the diagram @xmath230 in figure [ fig : persistencediagramsrbc500](b ) contains a single dominant point at @xmath231 with life span @xmath232 .
however , unlike in our analysis of the kolmogorov flow in the previous section , we can not argue that this point corresponds to a single dominant loop along which the data is organized because of the gaps in the sampling of the orbit . as mentioned in section [ sec::largedata ]
, the missing parts of the orbit could introduce loops of similar size corresponding to secondary structures .
these structures might occur due to the fact that the loop corresponding to the underlying almost - periodic dynamics might be twisted , pinched , or bent in @xmath12 . in order to obtain information about secondary structures
, we require a faster sampling rate . we increased the sampling rate considerably and collected approximately @xmath233 equally - spaced snapshots of the temperature field over four - and - a - half periods and compute the associated persistence diagrams , producing a point cloud @xmath234 .
the maximal distances between the consecutive frames for the increased sampling rate drop to @xmath235 , @xmath236 and @xmath237 .
the new value of @xmath205 is much closer to our estimate of the numerical error and it is more than 24 times smaller than the diameter of the point cloud generated from the slower sampling .
since the point cloud could only increase in diameter through increasing the sample rate , we consider this sampling rate to be satisfactory .
our next step is to use the ideas introduced in section [ sec::largedata ] to reduce the size of the sample and to complete our analysis .
first we construct a @xmath238-dense , @xmath238-sparse subsample @xmath239 of the point cloud @xmath240 .
the smallest value of @xmath238 for which we were able to compute the persistence diagrams @xmath241 , using @xmath242 gb of memory , is @xmath243 .
this value is only slightly larger than the largest distance between the consecutive states and , since the diameter of the subsampled point cloud is @xmath244 , the relationship between the length scale of the smallest detectable feature and the length scale of the diameter of the point cloud is still sufficient to resolve the geometry of the dynamics .
the resulting persistence diagrams @xmath241 are shown in figure [ fig : persistencediagramsrbc ] .
as shown by @xmath245 , figure [ fig : persistencediagramsrbc](a ) , the point cloud merges to a single connected component at @xmath246 .
this indicates that the sample of the trajectory is not broken into different pieces separated from each other .
since the maximum consecutive distance between any two points in @xmath240 is @xmath247 , the loop along which the data is organized should be present for @xmath248 .
however , after subsampling , it is possible that the loop will not be born until @xmath249 . looking at the diagram @xmath250 in figure [ fig : persistencediagramsrbc](b ) , we see that it contains a dominant point at @xmath251 , and so the loop was indeed born before @xmath249 .
this is the loop along which the point cloud is organized .
now , there is another point , @xmath252 , with life span @xmath253 .
this point corresponds to a secondary structure of the orbit .
indeed , it can be seen from the distance matrix for the @xmath238-sparse , @xmath238-dense subsample ( not shown for brevity ) that the part of the orbit corresponding to the fast dynamics ( missing for the slow sampling rate ) revisits very similar states before continuing along the main loop .
however , the development of more sensitive tools is required to fully understand these secondary features .
we now turn our attention to the differences between the persistence diagrams of the original point cloud @xmath240 and its subsample @xmath254 .
theorem [ thm::subsample ] implies that @xmath255 , and so there exists a bijection between the points in @xmath256 and @xmath241 such that the distance between matched points is less than @xmath257 . according to remark [ rem::subsample ] , for the dominant point @xmath258
, there is exactly one corresponding point in @xmath259 .
this point is the unique point in @xmath259 that lies inside of the shaded box touching the point @xmath251 , see figure [ fig : persistencediagramsrbc](b ) .
the same is true for the other dominant point .
moreover , there are no points in @xmath259 outside of the shaded regions .
points in @xmath259 that do not correspond to the off - diagonal points in @xmath250 can appear only @xmath260 far away from the diagonal .
we have shown how persistent homology can be used to identify equilibria and study periodic dynamics , and how this method is particularly natural when solutions must be identified that lie on a group orbit .
we study two regimes in kolmogorov flow : chaotic dynamics due to the appearance of unstable fixed points , and a periodic flow that exhibits drift in a direction of continuous symmetry .
we also study an almost - periodic orbit from rayleigh - bnard convection .
we solve for the unstable equilibria in the first case and sample the periodic orbits in the other two cases , and use persistent homology to project these solutions to the space of persistence diagrams .
we provide theoretical results that show this projection is stable with respect to numerical errors and discuss how the projection naturally identifies symmetry - related solutions .
we give three different metrics on the space of persistence diagrams that can be used to study pattern evolution on large versus small spatial scales , and how these metrics can be used to estimate numerical error in the space of persistence diagrams .
we develop an intuition for studying dynamics in the space of persistence diagrams by looking at point clouds in two - dimensional euclidean space , and discuss methods for determining if a continuous trajectory has been sampled densely enough to resolve the underlying dynamics , as well as mathematical methods used to address issues associated with computing on large sample sets .
we demonstrate the efficacy of these methods on kolmogorov flow and rayleigh - bnard convection , comparing our methods to traditional fourier methods where appropriate .
our results show that the geometry of the dynamics are recovered in each case . for rayleigh - bnard convection in particular
, we show that the dynamics are recovered even after truncating the simulated data to an 8-bit temperature field , and so this approach is suitable for studying data collected experimentally , rather than numerically .
also for this flow , we recover more subtle aspects of the geometry in the space of persistence diagrams . in summary
, we have shown that this method is both robust to noise and sensitive to more complicated dynamics , and that it is appropriate for studying dynamics on datasets obtained experimentally .
our ongoing research will further refine these tools .
the work of mk , rl , and km has been partially supported by nsf grants nsf - dms-0835621 , 0915019 , 1125174 , 1248071 , and contracts from afosr and darpa .
the work of jrf , bs and mfs has been partially supported by nsf grants dms-1125302 , cmmi-1234436 .
topologically , a torus @xmath261 is a closed surface defined as the product of two circles .
it can be also described as a quotient of the cartesian plane under the identifications @xmath262 .
the homology groups of @xmath53 are given by @xmath263 intuitively , this means that @xmath53 has a single connected component ( @xmath264 ) , two independent loops ( @xmath265 ) , and a single cavity ( @xmath266 ) . for a more detailed treatment of the following material , see see hatcher , ch 0 for a reference to homotopies of maps , and hatcher ch 1 for a reference on identifying independent loops in a space , or the notion of the fundamental group . in this section ,
we explain the notion of independent loops of subsets of a torus . by a loop ,
we mean a continuous path @xmath267 \rightarrow t$ ] such that @xmath268 .
we will also be using the notion of a homotopy of loops , which can be thought of as deforming one loop continuously to another .
more precisely , a homotopy of loops from a loop @xmath269 to a loop @xmath270 is a continuous function @xmath271\times[0,1 ] \rightarrow t$ ] such that @xmath272 and @xmath273 for all @xmath274 $ ] , and @xmath275 for all @xmath276 $ ] .
define @xmath277 , which runs the loop @xmath173 backwards in time , and define @xmath278 , which traverses the loop @xmath173 @xmath148 times . finally , given two loops @xmath279 and
@xmath280 , we can form their sum @xmath281 by taking a path @xmath282 \rightarrow t$ ] such that @xmath283 and @xmath284 and form the loop @xmath285\\ \delta(4 t - 1 ) & : t \in [ 1/4,1/2]\\ \alpha_2(4t-2 ) & : t \in [ 1/2 , 3/4]\\ -\delta(4t-3 ) & : t \in [ 3/4 , 1 ] .
\end{array } \right.\ ] ] algebraically , this can be written as @xmath286 .
we say that a loop @xmath173 is independent of a collection of loops @xmath287 if there does not exist a homotopy of loops from @xmath173 to a linear combination of the loops @xmath287 .
figure [ fig : holeintorus ] shows eight subsets @xmath288 of a torus .
note that @xmath289 for @xmath290 , and the sets can be considered as sub - level sets of some scalar function @xmath92 .
we will now examine each set and identify the independent loops in each .
the set @xmath292 can not be contracted to a point .
it forms a band that wraps around the torus .
there are many different loops ( wrapping once around the torus from left to right in the picture ) inside of this band .
however , we can choose a single loop @xmath279 that represents all of them ; every other loop can be either continuously deformed to a linear combination of the loop @xmath279 , or contracted to a point .
similarly , the set @xmath293 contains two independent loops .
the set @xmath294 is formed by linking the horizontal bands present in @xmath293 . the loops @xmath279 and @xmath280 are still independent in @xmath294 ( one can not be deformed to the other inside @xmath294 ) .
it might seem that there is a new independent loop , @xmath173 .
however , this is not case because @xmath173 can be deformed ( inside of @xmath294 ) to the union of the black lines corresponding to @xmath295 and @xmath238 . after this deformation
, the loop traverses @xmath238 twice : the right part of the deformed loop traverses @xmath238 from the top to the bottom , and the left part in the opposite direction .
algebraically , the deformed loop can be expressed as @xmath296 .
this shows that @xmath173 can be deformed to a linear combination of the loops @xmath279 and @xmath280 .
thus , @xmath173 is not a new independent loop .
the set @xmath297 , obtained from @xmath294 by adding another link between the horizontal bands , contains a new independent loop , @xmath298 , consisting of the edges @xmath299 and @xmath300 ( @xmath301 ) .
this means that the loop @xmath298 can not be deformed inside of @xmath297 to a linear combination of the loops @xmath279 and @xmath280 . again , the loop @xmath173 is not independent from @xmath279 , @xmath280 , and @xmath298 because it can be perturbed to @xmath302 , which is a linear combination of the loops @xmath295 , and @xmath298 .
therefore , there are three independent loops in this case . adding another link between the horizontal bands creates another independent loop .
hence , the number of independent loops for two bands with @xmath148 links is @xmath303 .
alternatively , we can view the set @xmath294 as a single band with one puncture , and @xmath297 as a single band with two punctures .
the number of independent loops is @xmath303 , where @xmath148 is the number of punctures , and the extra loop is generated by the band . due to the identification @xmath304
, the set @xmath305 contains another link between the horizontal bands .
this band creates another puncture . in figure
[ fig : holeintorus](f ) , this puncture seems to have four distinct components ( white blocks in the corners ) .
however , under the boundary identification , they correspond to a single component .
therefore , there are four independent loops .
the independent loops start disappearing as the punctures are filled in .
the set @xmath306 contains a single puncture , and according to the previous argument , there are two independent loops , @xmath23 and @xmath22 . in this case , the loop @xmath173 can be deformed to a point inside of the set @xmath306 .
finally , the set @xmath307 contains two independent loops corresponding to the two copies of @xmath308 that generate the torus .
k. krishan , h. kurtuldu , m. f. schatz , m. gameiro , k. mischaikow , and s. madruga , `` homology and symmetry breaking in rayleigh - bnard convection : experiments and simulations , '' _ phys . fluids _ , vol . 19 , nov .
2007 .
h. kurtuldu , k. mischaikow , and m. f. schatz , `` extensive scaling from computational homology and karhunen - love decomposition analysis of rayleigh - bnard convection experiments , '' _ phys .
_ , vol .
107 , no . 3 , 2011 . m. paul , k .- h . chiam , m. cross , p. fischer , and h. greenside , `` pattern formation and dynamics in rayleigh
bnard convection : numerical simulations of experimentally realistic geometries , '' _ physica d _ , vol .
184 , no . 1 ,
pp . 114126 , 2003 . | we use persistent homology to build a quantitative understanding of large complex systems that are driven far - from - equilibrium ; in particular , we analyze image time series of flow field patterns from numerical simulations of two important problems in fluid dynamics : kolmogorov flow and rayleigh - bnard convection . for each image
we compute a persistence diagram to yield a reduced description of the flow field ; by applying different metrics to the space of persistence diagrams , we relate characteristic features in persistence diagrams to the geometry of the corresponding flow patterns .
we also examine the dynamics of the flow patterns by a second application of persistent homology to the time series of persistence diagrams .
we demonstrate that persistent homology provides an effective method both for quotienting out symmetries in families of solutions and for identifying multiscale recurrent dynamics .
our approach is quite general and it is anticipated to be applicable to a broad range of open problems exhibiting complex spatio - temporal behavior . |
there has been recently an increasing interest for study of systems with a finite chiral density , namely @xmath10 .
such chirality imbalance can be obtained dynamically because of the adler - bell - jackiw anomaly @xcite when fermions interact with nontrivial gauge field configurations characterized by a topological index named the winding number , @xmath11 . in the context of quantum chromodynamics ( qcd ) such nontrivial gauge field configurations at finite temperature in minkowski space are named sphalerons , whose production rate has been estimated to be quite large @xcite . the large number of sphaleron transitions in high temperature suggests the possibility that net chirality might be abundant ( locally ) in the quark - gluon plasma phase of qcd ; when one couples this thermal qcd bath with an external strong magnetic field , @xmath1 , produced in the early stages of heavy ion collisions , the coexistence of @xmath12 and @xmath13 might lead to a charge separation phenonemon named the chiral magnetic effect ( cme ) @xcite which has been observed experimentally in zirconium pentatelluride @xcite . beside cme other interesting effects related to anomaly and chiral imbalance can be found in @xcite . in order to describe systems with finite chirality in thermodynamical equilibrium
, it is customary to introduce a chiral chemical potential , @xmath5 , conjugated to the @xmath4 @xcite .
the chiral chemical potential describes a system in which chiral density is in thermodynamical equilibrium ; however because of anomaly as well as of chirality changing processes due to finite quark condensate , @xmath4 is not a strictly conserved quantity hence the meaning of @xmath5 is not so clear ; however naming @xmath9 the typical time scale in which chirality changing processes take place , one might assume that @xmath14 describes a system in thermodynamical equilibrium with a fixed value of @xmath4 on a time scale much larger than @xmath9 , the latter representing the time scale needed for @xmath4 to equilibrate .
in this article we study chiral phase transition and chiral density production in the context of quark matter in a background static and homogeneous parallel electric , @xmath0 , and magnetic , @xmath1 , fields .
one of our goal is to investigate the effect of the background fields on chiral symmetry breaking at zero temperature , and on the critical temperature for chiral symmetry restoration , @xmath15 .
this part of the study embraces previous studies about chiral symmetry breaking / restoration in the background of external fields @xcite , completing them by adding the computation of the critical temperature versus the strengh of @xmath2 and @xmath3 .
we find that the effect of the electric field is to lower the critical temperature , in agreement with the scenario of inverse catalysis depicted in @xcite where however the zero temperature case has only been considered ; the inverse catalysis scenario does not change considerably when the magnetic field is added , as long as the magnetic field is not very large compared to the electric one .
this finding is in agreement with a previous study at zero temperature @xcite where the role of the second electromagnetic invariant , @xmath16 , has been recognized as inhibitor of chiral symmetry breaking .
we are also interested to study the effect of chiral density on the thermodynamics of the system .
the model studied here has the advantage that a chiral density is obtained dynamically without the need to introduce , _ a priori _ , a chiral chemical potential . as a matter of fact
chirality can be produced combining @xmath0 which produces pairs via the schwinger mechanism , and @xmath1 which alignes particles spin along its direction .
the mechanism producing chirality is very simple : we assume for sake of simplicity a very large @xmath1 , so that only the lowest landau level ( lll ) is occupied ; moreover we assume the system made only of one flavor of quarks , namely @xmath17 quarks , and we focus on a single @xmath18 created by the schwinger effect .
the @xmath17 quark must have its spin aligned along @xmath1 because it sits in the lll , and its momentum will be initially rather parallel or antiparallel to @xmath1 , so the initial helicity can be either positive or negative . on the other hand the effect of @xmath19 is to accelerate @xmath17 along the direction of @xmath1 so after some time @xmath17 quark will have positive helicity .
an analogous discussion can be done for the @xmath20 .
therefore as a consequence of the schwinger effect , lll and @xmath19 each time a pair is created , there is an increase of a factor two of the net chiral density of the system .
the dynamical evolution of @xmath4 produced by this mechanism can be computed explicitly @xcite and it has been shown to be the one expected from the adler - bell - jackiw anomaly : this is not surprising because @xmath21 meaning that axial current is not conserved at the quantum level and @xmath4 should evolve according to the anomaly equation . if @xmath4 evolution was governed only by the anomaly , however , there would be no chance for reaching a thermodynamical equilibrium because @xmath4 would grow indefinitely ( assuming the fields as external fields and neglecting any backreaction from the fermion currents ) .
but in the thermal bath there are also chirality flipping processes related to the existence of the chiral condensate as well as of the finite current quark mass : we introduce a relaxation time for chirality , namely @xmath9 , giving the time scale necessary for the equilibration of @xmath4 . then it is possible to show that for times @xmath22 chiral density equilibrates to @xmath23 , the actual value depending on quark electric charge , fields magnitude and temperature . because @xmath4 equilibrates it is possible to introduce the chiral chemical potential , @xmath5 , conjugated to @xmath23 at equilibrium . differently from previous calculations with chirality imbalance ,
in the present study we compute the value of @xmath5 self - consistently by coupling the gap equation to the number equation , even if we limit ourselves to the approximation of small fields and small @xmath5 , namely working at the leading order in @xmath24 and @xmath6 , @xmath7 . as a consequence ,
@xmath5 will depend on temperature as well as on external fields , and on the relaxation time .
we focus on the effects of the external fields on the chiral phase transition , with emphasis on the role of chirality production in the critical region . because of the small fields approximation involved in the solution of the gap as well as the number equations , we are aware that our picture about thermodynamics might change in case of large fields . in this study
we compute the effect of the dynamically produced @xmath4 on @xmath15 .
as mentioned above , the @xmath25 term tends to lower the critical temperature ; on the other hand the chiral chemical potential has the effect to increase @xmath15 @xcite .
therefore , it is interesting to compute the response of @xmath15 to the simultaneous presence of @xmath5 and fields , to check if the inverse catalysis scenario obtained at @xmath26 still persists at @xmath27 .
we can anticipate our results , namely that chiral density does not affect drastically the thermodynamics at the phase transition , confirming the inverse catalysis induced by the fields , as long as the average chiral chemical potential in the crossover region turns out to be small with respect to temperature . in section v we present a detailed study of this effect , showing concrete numbers and among other things how changing the field strenghts and/or the relaxation time magnitude affects the inverse catalysis .
in fact we have found and report about situations in which we can measure a net effect of the chiral chemical potential on the constituent quark mass and on critical temperature , even if we take these results with a grain of salt as the value of @xmath5 at equilibrium turns out to be of the order of the critical temperature , hence potentially validating our quantitative predictions .
the relaxation time for chirality adds the greatest theoretical uncertainty to our calculations : in absence of a specific calculation of @xmath9 it is possible to give only a rough estimate based on dimensional analysis as well as on physical reasons ; we chose @xmath28 where @xmath29 is the constituent quark mass which is computed self - consistently within the model : it depends on temperature and fields , and by construction it brings informations about the chiral condensate at zero as well as finite temperature .
because of this uncertainty on @xmath9 we feel it is not so important , in this explorative study , to present the most complete calculation possible taking into account the full propagators with the full @xmath5 dependence : we suspect in fact that even within the most accurate calculation possible , the new effects of the chiral density on the phase transition might be cancelled by changing @xmath9 which still would remain unknown .
we therefore prefer to limit ourselves to a simple weak fields and small @xmath5 approximation to explore the effects the chiral density will have on the phase diagram , leaving a more complete calculation to a future study . the plan of the article is as follows . in section
ii we briefly review the model we use for our calculations . in section
iii we present few selected results at zero temperature which show the interplay between the electric and magnetic fields on chiral symmetry breaking . in section
iv we discuss some result at finite temperature , with emphasis on the chiral phase transition without taking into account chirality production . in section
v we compute chirality at equilibrium and the related chiral chemical potential , and study the effect of this chirality on the critical temperature . finally in section
vi we draw our conclusions .
in this article we are interested to study quark matter in a background of an electric - magnetic fluc tube made of parallel electric , @xmath0 , and magnetic , @xmath1 , fields .
we assume the fields are constant in time and homogeneous in space ; moreover we assume they develop along the @xmath30direction . in this section
we describe the model we use for our calculations .
more specifically , we use a nambu - jona - lasinio ( njl ) model @xcite ( see @xcite for reviews ) with a local interaction kernel , in which we introduce the coupling of quarks with the external alectric and magnetic fields .
the set up of the gap equation has been presented in great detail in @xcite which we follow , therefore we will skip all the technical details and report here only the few equations we need to specify the interactions used in the calculations . the euclidean lagrangian density is given by @xmath31,\ ] ] with @xmath32 being a quark field with dirac , color and flavor indices , @xmath33 is the current quark mass and @xmath34 denotes a vector of pauli matrices on flavor space .
the interaction with the background fields is embedded in the covariant derivative @xmath35 , where @xmath36 denotes the set of euclidean dirac matrices and @xmath37 is the quark electric charge matrix in flavor space . in this work we use the gauge @xmath38 . introducing the auxiliary field @xmath39 and using a mean field approximation ,
the thermodynamic potential can be written as @xmath40 where the constituent quark mass is @xmath41 , @xmath42 and @xmath43 corresponds to the euclidean quantization volume .
the constituent quark mass differs from @xmath33 because of spontaneous chiral symmetry breaking , the latter being related to a nonvanishing chiral condensate , @xmath44 .
even if it would be more appropriate to discuss chiral symmetry restoration via the quark condensate , because it has its counterpart in qcd , in this article we will refer to @xmath29 for simplicity , keeping in mind that whenever we discuss about the chiral phase transition in terms of @xmath29 the decrease of the latter is related to the decreasing chiral condensate . in this model
the main task is to compute self - consistently @xmath29 at finite temperature and in presence of the external fields .
this is achieved by requiring the physical value of @xmath29 minimizes the thermodynamic potential , and this in turn implies that @xmath29 satisfies the gap equation , @xmath45 , namely @xmath46 where @xmath47 corresponds to the full fermion propagator in the electric and magnetic field background .
the computation of the propagator has been already given in detail in @xcite therefore here we merely quote the final result for the gap equation , that is @xmath48 where we have introduced the functions @xmath49 with @xmath50 being the third elliptic theta function , and @xmath51 in eq .
eq : gkk we have added and subtracted the zero field contribution on the right hand side which is the only one to diverge , and we have regularized it by cutting the integration at @xmath52 ; on the other hand we have not added a cutoff on the field dependent part as it is not divergent . for the parameters choice
we use the standard parameter set for a proper time regularization @xcite , namely @xmath53 mev and @xmath54 .
the presence of the @xmath55 in eq .
eq : fs implies the existence of an infinite set of poles on the integration in @xmath56 in eq .
eq : gkk ; these poles appear in @xmath57 as well .
following the original treatment by schwinger @xcite these poles are moved to the complex plane by adding a small imaginary part which allows to perform the @xmath58integration in principal value ; this leads to an imaginary part of the free energy , which is a sign of the vacuum instability induced by the static electric field @xcite and leads to particle pair creation . we will consider the effect of this vacuum instability in section [ eq : cd ] because it can be directly connected to chiral density production in case of parallel @xmath0 and @xmath1
dynamical quark mass with electric and/or magnetic field strength at zero temperature .
maroon dot - dashed line corresponds to the case of a pure magnetic field ; green dashed line to a pure electric field ; finally solid orange line corresponds to the case @xmath59 .
left panel corresponds to @xmath60 mev , while right panel corresponds to the @xmath61.,width=264 ] in this section we present few results at zero temperature . in fig .
[ fig:1 ] we plot the constituent quark mass as a function of the external field strength at @xmath62 for several cases : maroon dot - dashed line corresponds to the case of a pure magnetic field ; green dashed line to a pure electric field ; finally solid orange line corresponds to the case @xmath59 .
left panel corresponds to @xmath60 mev which is the value of the current quark mass necessary to have @xmath63 mev ; right panel corresponds to the chiral limit @xmath61 . for the case of a pure magnetic field
we find the magnetic catalysis of chiral symmetry breaking ; on the other hand the electric field has the opposite effect leading to an inverse magnetic catalysis @xcite . in this pure electric field case
there exists a critical electric field at which chiral symmetry is restored in the chiral limit : we find the transition to be of the second order . in the case of massive quarks
the phase transition is changed into a smooth crossover characterized by a smooth but net change in the slope of the condensate , resulting in smaller value of the condensate itself , as it happens for the chiral phase transition at finite temperature . in this case
it is not possible to define in a rigorous way a critical field , but it is still possible to identify a range of electric fields in which @xmath29 has its highest change with @xmath2 , and identify this range with the pseudo - critical region . _ upper panel . _
dynamical quark mass versus magnetic field strength at zero temperature , for several values of the background electric field .
maroon dot - dashed line corresponds to the case of a pure magnetic field ; indigo dotted line to @xmath64 ; dashed magenta line to @xmath65 ; dot - dot - dashed brown line to @xmath66 ; finally orange line to @xmath59 . for comparison
we have also shown data for @xmath67 borrowed from fig.[fig:1 ] ; in this case on @xmath68axis we show @xmath69 in units of @xmath70 .
_ lower panel .
_ comparison of the perturbative solution eq .
eq : l2 with the full one.,title="fig:",width=264 ] + _ upper panel .
_ dynamical quark mass versus magnetic field strength at zero temperature , for several values of the background electric field .
maroon dot - dashed line corresponds to the case of a pure magnetic field ; indigo dotted line to @xmath64 ; dashed magenta line to @xmath65 ; dot - dot - dashed brown line to @xmath66 ; finally orange line to @xmath59 . for comparison
we have also shown data for @xmath67 borrowed from fig.[fig:1 ] ; in this case on @xmath68axis we show @xmath69 in units of @xmath70 . _
lower panel .
_ comparison of the perturbative solution eq .
eq : l2 with the full one.,title="fig:",width=264 ] it is interesting to study what happens when @xmath2 and @xmath3 act together : naturally one would expect a competition among the effects of the magnetic ( catalysis ) and electric ( inverse catalysis ) fields . in fig .
[ fig:1 ] we have shown the case @xmath59 in which it is clear that , regardless we work in the chiral or in the physical current quark mass limit , the magnetic field has some catalysis effect increasing the value of @xmath29 ( i.e. chiral condensate ) and shifts the critical ( or pseudo - critical ) value of the electric field slightly upwards compared to the case @xmath67 . in fig .
[ fig:1b ] we show @xmath29 as a function of @xmath71 for several choices of @xmath2 , starting from @xmath72 up to @xmath73 . already for @xmath65 we find a sign of competition among direct and inverse catalysis , which manifests in a non - monotonic behaviour of @xmath29 versus @xmath71 .
we can also read the results of fig .
[ fig:1b ] in the opposite way : given a background of an electric field @xmath74 , even introducing a magnetic field of the same magnitude of @xmath74 does not result in a considerable change of spontaneous chiral symmetry breaking , compare green dashed and orange solid lines in fig .
[ fig:1 ] ; for @xmath3 as large as @xmath75 we find that qualitatively the behaviour of the chiral condensate versus field strength is like the one at @xmath67 , even if for very small values of the field strength we still find the mass increases ; the net effect of the magnetic field is to shift the critical value of the electric field to larger values because of catalysis . in order to measure a catalysis effect
one has to introduce a larger magnetic field , for example @xmath76 in fig .
[ fig:1b ] .
the inverse catalysis effect induced by the electric field and the second electromagnetic invariant , @xmath77 , are in agreement with previous studies at zero temperature @xcite .
the behaviour of @xmath29 for small values of the fields can be easily understood quantitatively by the gap equation at @xmath62 and @xmath61 .
we can find an analytical solution for the gap equation eq : gkk for small fields by writing @xmath78 where @xmath79 corresponds to the solution of the gap equation for @xmath80 .
moreover for small values of the fields we can keep only the order @xmath81 in the field dependent term in eq .
eq : gkk . taking into account that @xmath79 is the solution of the gap equation at @xmath80 we find @xmath82 where @xmath83 with @xmath84 , @xmath85
; moreover @xmath86 denotes the exponential integral function , @xmath87 .
the field dependence in the above equation resembles that occurring in the euler - heisenberg lagrangian @xcite as it should , since the latter can be obtained by integrating the gap equation over @xmath29 . from eq .
eq : l2 we notice that for @xmath67 , @xmath88 neglecting higher order contributions ; the curvature of @xmath89 versus @xmath69 does not change as long as @xmath90 . for @xmath91 one has to take into account the contribution @xmath92 which still shows @xmath93 leading to a decreasing @xmath29 .
finally for @xmath94 the catalysis sets in , at least for small values of the fields , eventually leading to @xmath95 for @xmath72 . in the lower panel of fig.[fig:1b ]
we have compared the perturbative solution in eq .
eq : l2 with the full one , for two cases .
we find a fair agreement among the two for @xmath96 .
_ ( upper panel)_. dynamical quark mass versus temperature for @xmath97 ( red solid line ) , @xmath98 ( cyan dot - dashed line ) and @xmath99 ( dashed green line ) . _
( middle panel)_. @xmath100 versus temperature used to identify the pseudo - critical temperature for the chiral crossover .
line convention is the same of the upper panel . _
( lower panel)_. dynamical quark mass versus temperature for @xmath97 ( red solid line ) , @xmath98 ( cyan dot - dashed line ) and @xmath99 ( dashed green line ) in the chiral limit.,title="fig:",width=264 ] + _ ( upper panel)_. dynamical quark mass versus temperature for @xmath97 ( red solid line ) , @xmath98 ( cyan dot - dashed line ) and @xmath99 ( dashed green line ) . _ ( middle panel)_. @xmath100 versus temperature used to identify the pseudo - critical temperature for the chiral crossover .
line convention is the same of the upper panel . _
( lower panel)_. dynamical quark mass versus temperature for @xmath97 ( red solid line ) , @xmath98 ( cyan dot - dashed line ) and @xmath99 ( dashed green line ) in the chiral limit.,title="fig:",width=264 ] + _ ( upper panel)_. dynamical quark mass versus temperature for @xmath97 ( red solid line ) , @xmath98 ( cyan dot - dashed line ) and @xmath99 ( dashed green line ) . _
( middle panel)_. @xmath100 versus temperature used to identify the pseudo - critical temperature for the chiral crossover .
line convention is the same of the upper panel . _
( lower panel)_. dynamical quark mass versus temperature for @xmath97 ( red solid line ) , @xmath98 ( cyan dot - dashed line ) and @xmath99 ( dashed green line ) in the chiral limit.,title="fig:",width=264 ] in this section we discuss our results about chiral symmetry restoration at finite temperature . in the upper panel of fig .
[ fig:2 ] we plot @xmath29 versus @xmath101 for @xmath102 and @xmath103 , and compare it with the result at @xmath97 .
the general trend of data shown in the figure is in agreement with the scenario depicted at @xmath62 discussed above .
in particular the inverse catalysis due to the electric field implies the lowering of the critical temperature ; on the other hand the catalysis due to the magnetic field at @xmath62 is still present at @xmath104 leading to the increase of the pseudo - critical temperature at @xmath103 .
it has been discussed that the magnetic catalysis of chiral symmetry breaking within the njl model at finite temperature is due to the fact the njl interaction kernel does not take into account the effects of screening as well as of coupling lowering which instead occur in qcd and are important for inverse catalysis @xcite . although it would be possible to insert by hand a @xmath105dependence of the njl coupling in order to reproduce the inverse magnetic catalysis @xcite
, we prefer to not do this in our study because it would hide the effect of the electric field ; we will add this important ingredient in our upcoming works . in the middle panel of fig .
[ fig:2 ] we plot @xmath100 : we identify its maximum with the crossover temperature . for @xmath97
we find @xmath106 mev .
we notice that the electric field not only makes the pseudo - critical temperature lower than the one in the case @xmath97 , but it also smooths the crossover because the variation of the quark mass with temperature is smaller in magnitude than in the case with no fields . finally , in the lower panel of fig . [ fig:2 ]
we plot @xmath29 versus temperature for the ideal case of quarks with vanishing current mass : as expected , the effect of the external fields is qualitatively the same we have found in the realistic case of quarks with finite current mass . dynamical quark mass versus temperature for several values of @xmath2 and @xmath3 .
thin lines correspond @xmath67 , while with thick lines we denote the results for @xmath59 .
blue solid line corresponds to @xmath107 , orange dotted line to @xmath108 , and green dashed line to @xmath109 .
color convention for thick lines follows that we have used for thin lines.,width=264 ] in fig .
[ fig:2a ] we plot @xmath29 versus @xmath101 for several values of @xmath2 and @xmath3 : thin lines correspond @xmath67 , while with thick lines we denote the results for @xmath110 .
blue solid line corresponds to @xmath111 , orange dotted line to @xmath108 , and green dashed line to @xmath109 .
increasing the electric field strength results in a lowering of the critical temperature , and the effect of @xmath103 is just to increase a bit the quark mass and shift the critical temperature towards slightly higher values .
the results collected in figures [ fig:2 ] and [ fig:2a ] show that even when @xmath112 the effect of the fields on the critical temperature does not cancel and the electric field gives the more important contribution , leading to an inverse catalysis .
in fact one would need a larger value of @xmath3 to observe an increase of the critical temperature .
this can be understood easily : close to the second order phase transition ( we work now at @xmath61 which allows an analytical treatment ) we can make an expansion of the thermodynamic potential in powers of @xmath29 , namely @xmath113 where the coefficient @xmath114 at @xmath115 ; @xmath116 is negative in the chirally broken phase and vanishes at the phase transition .
the coefficient @xmath116 can be easily computed taking the derivative of the gap equation eq .
eq : gkk , and expanding for small values of the fields .
it is then possible to write @xmath117 where @xmath118 denotes a field - independent term and @xmath119 corresponds to a term @xmath120 .
the field - independent term is not interesting because it just determines the critical temperature when the external fields are set to zero . on the other hand
the field - dependent contributions are more relevant for the discussion ; a straightforward calculation leads to @xmath121 where we have used the shorthand notation @xmath122}. \label{eq : ops2}\end{aligned}\ ] ] the term on the right hand side in the first line of eq .
eq : ops shows that the correction to @xmath116 due to the pure magnetic field is negative , hence shifting the phase transition to larger temperatures . on the other hand the term proportional to @xmath123 is positive , and gets a further positive contribution from the second line of eq .
eq : ops : indeed the latter is proportional to the derivative of the @xmath50 function , which is a decreasing function of its second argument . as a consequence
the coefficient proportional to @xmath123 is positive and because of the additional contribution at finite temperature , one needs a value of @xmath124 in order to change the sign of @xmath119 and turning the inverse catalysis into a direct one .
this explains why for @xmath73 we still find an inverse catalysis of chiral symmetry breaking at finite temperature .
critical temperature for chiral symmetry restoration versus electric field strength , measured in units of @xmath70 , for several values of the external magnetic field.,width=264 ] in fig .
[ fig:2c ] we plot @xmath15 versus @xmath69 ( measured in units of @xmath70 ) for several values of the external magnetic field : black squares correspond to @xmath67 , red diamonds to @xmath125 and green triangles to @xmath126 .
this figure summarizes one of the main finding of our work , namely that the electric field leads to a lowering of the critical temperature for chiral symmetry restoration , and the presence of the parallel magnetic field does not change this result unless @xmath127 .
the electric - magnetic background considered in this article is dynamically unstable because of the schwinger pair production @xcite .
this is due to the presence of poles in the thermodynamic potential , @xmath57 , which in turn make the quantum corrections to the electromagnetic lagrangian complex , with the imaginary part related to the vacuum persistency probability .
because of the quantum anomaly , the schwinger mechanism eventually leads to a nonzero chiral density , @xmath4 ; we assume for the moment the background magnetic field @xmath3 is very large so that it is reasonable to assume that only the lowest landau level ( lll ) is occupied , to simplify the discussion ; moreover we assume the system made only of one flavor of quarks , namely @xmath17 quarks .
let us focus on one single @xmath18 created by the schwinger effect .
the @xmath17 quark must have its spin aligned along @xmath1 because of the lll approximation ; because of dimensional reduction in the lll @xmath17 momentum will be initially rather parallel or antiparallel to @xmath1 , so the initial chirality can be either positive or negative .
on the other hand the effect of @xmath19 is to accelerate @xmath17 along the direction of @xmath1 so after some time @xmath17 quark will have positive chirality .
an analogous discussion can be done for the @xmath20 .
therefore as a consequence of the schwinger effect , lll and @xmath19 each time a pair is created , there is an increase of a factor two of the net chiral density of the system , @xmath128 .
obviously higher landau levels do not contribute to @xmath4 because particle spin can be either parallel or antiparallel to @xmath1 leading to a cancellation of @xmath4 .
hence chirality is produced dynamically in the background field configuration studied here .
this makes the study very interesting because if chiral density relaxes to an equilibrium value , it might affect the equilibrium properties of quark matter .
the time evolution of @xmath4 in case of massive particles in the background with constant and homogeneous fields has been derived for the first time by warringa in @xcite , where he has shown it can be directly obtained from the schwinger production rate for the case of @xmath129 , namely @xcite @xmath130 indeed only the lowest landau level ( lll ) gives a contribution to @xmath4 , and this lll contribution can be easily extracted from the above equation by taking the @xmath131 limit because in such a limit it is reasonable to assume that only the lll is occupied ; because each pair in the lll changes the chiral density of a factor of 2 we have from the above equation in the @xmath131 limit : @xmath132 in agreement with @xcite .
if evolution of @xmath4 was given only by the above equation then the system would never be able to reach thermodynamical equilibrium ( assuming the fields as external fields neglecting any backreaction from the fermion currents ) .
however eq . eq : al1 is just half of the story : because of finite quark mass there are chirality changing processes which should lead to equilibration of @xmath4 .
in order to take into account of these processes we add a relaxation term on the right hand side of the above equation , @xmath133 where @xmath9 corresponds to the relaxation time of chirality changing processes .
for @xmath134 the solution of eq .
eq : al2 relaxes to teh equilibrium value @xmath135 the equilibrium value of chiral density depends on the value of @xmath9 . it is reasonable to assume both by virtue of dimensional considerations and by naive physical arguments that @xmath136 where @xmath29 corresponds to the constituent quark mass : for large values of @xmath29 chirality changing processes will be very fast hence reducing drastically the relaxation time and the net chirality produced at equilibrium ; on the other hand for small @xmath29 the system will be less efficient in changing chirality which implies a larger relaxation time and a larger chirality produced at equilibrium .
thus we assume @xmath137 the above equation implicitly contains effects of the chiral condensate in the chirally broken phase via the larger value of @xmath29 in this phase .
needless to say the parameter @xmath138 adds the largest uncertainty in our calculations : because @xmath23 depens linearly on @xmath9 , a change of an order of magnitude in @xmath138 will produce the same change in @xmath23 .
we will study how changing @xmath138 might affect our results .
equation eq : al3 shows that on a time scale larger than the relaxation time an equilibrium value of @xmath4 , namely @xmath23 , is produced . because of the different charges of @xmath17 and @xmath139 quarks the equilibrium value of @xmath4 for the two flavors to be different : at equilibrium in fact we find @xmath140 the actual value depending on @xmath2 and on temperature via @xmath29 .
the existence of an equilibrium value for the chiral density means it is possible to introduce a chemical potential for the chiral charge , namely the chiral chemical potential @xmath5 , conjugated to @xmath141 . a self - consistent computation of @xmath5 given the value of @xmath23 in eq .
eq : al3 requires a canonical ensemble calculation in which the gap equation for the quark mass is solved self - consistently with the number equation , namely @xmath142 with @xmath5 introduced in the quark propagator with @xmath143 .
this full calculation is well beyond the purpose of the present article and is left to a future study . here
we limit ourselves to consider this problem only in the limit of small @xmath5 as well as small fields , in which we can use the njl model with @xmath97 but @xmath14 to compute the relation between @xmath5 and @xmath23 , as well as to take into account self - consistently the effect of @xmath5 in the gap equation .
the cheap procedure we use here to solve self - consistently the problem should be accurate up to the lowest nontrivial order in @xmath5 and fields , that is @xmath144 . the njl thermodynamic potential at @xmath97 and @xmath14 can be written as @xcite @xmath145 with @xmath146 and @xmath147 denotes the chiral chemical potential for the flavor @xmath148 : we allow for a flavor dependence of @xmath5 because the equilibrium value of @xmath4 depends on the flavor itself . at lowest order in @xmath5 the correction to the thermodynamic potential
can be written as @xmath149 which allows to write the @xmath150dependent correction to the gap equation , namely @xmath151 moreover the relation among @xmath152 and @xmath5 is given by @xmath153 and the number equation eq .
eq : ne can be written as @xmath154 we have verified that in the chiral limit @xmath115 the above equation gives @xmath155 in agreement with @xcite ; in the case @xmath156 the relation between @xmath4 and @xmath5 is more complicated and we have to compute it by performing numerically the integration in eq .
eq : pa2 . taking into account eq .
eq : pa1 the gap equation eq .
eq : gkk becomes @xmath157 and @xmath5 has to be computed self - consistently according to the number equation , eq .
we notice that although an explicit dependence of @xmath5 on @xmath2 and @xmath3 is not present in the above equations , the gap equation eq .
eq : pa2 are coupled because of the dependence of @xmath23 on the @xmath29 . _
( upper panel ) .
_ @xmath29 versus temperature .
maroon solid line corresponds to @xmath97 and @xmath158 .
indigo dashed line corresponds to @xmath97 and @xmath159 mev for @xmath17 and @xmath139 quarks .
orange dotted line corresponds to the case @xmath160 and @xmath158 : these data are the same we have shown in fig .
[ fig:2a ] .
green dot - dashed line corresponds to @xmath160 , with both @xmath29 and @xmath5 computed self - consistently by the condition @xmath161 with @xmath4 given by eq .
eq : pa2 and the gap equation eq .
eq : pa1 . _
( lower panel ) .
_ self - consistent @xmath5 ( green lines ) for @xmath17 ( thick dot - dashed line ) and @xmath139 ( thin dot - dashed ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel . for comparison
we have also shown the chiral chemical potential obtained by @xmath29 computed with @xmath158 shown in the upper panel.,title="fig:",width=264 ] + _ ( upper panel ) .
_ @xmath29 versus temperature .
maroon solid line corresponds to @xmath97 and @xmath158 .
indigo dashed line corresponds to @xmath97 and @xmath159 mev for @xmath17 and @xmath139 quarks .
orange dotted line corresponds to the case @xmath160 and @xmath158 : these data are the same we have shown in fig . [ fig:2a ] .
green dot - dashed line corresponds to @xmath160 , with both @xmath29 and @xmath5 computed self - consistently by the condition @xmath161 with @xmath4 given by eq .
eq : pa2 and the gap equation eq .
eq : pa1 . _
( lower panel ) .
_ self - consistent @xmath5 ( green lines ) for @xmath17 ( thick dot - dashed line ) and @xmath139 ( thin dot - dashed ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel . for comparison we have also shown the chiral chemical potential obtained by @xmath29 computed with @xmath158 shown in the upper panel.,title="fig:",width=264 ] in fig .
[ fig : i ] we plot @xmath29 versus temperature for several cases : maroon solid line corresponds to @xmath97 and @xmath158 .
indigo dashed line corresponds to @xmath97 and a common value @xmath159 mev for @xmath17 and @xmath139 quarks : we plot these data to show that the njl model we use in the calculation is capable to capture the catalysis of chiral symmetry breaking at finite @xmath5 since both @xmath29 and @xmath15 are shifted towards higher values in comparison with the case @xmath158 .
orange dotted line corresponds to the case @xmath160 and @xmath158 : these data are the same we have shown in fig .
[ fig:2a ] .
finally green dot - dashed line corresponds to @xmath160 , with both @xmath29 and @xmath5 computed self - consistently by solving the number equation eq .
eq : ne2 and the gap equation eq .
eq : pa1 simultaneously .
although @xmath99 sounds large for the approximation to be reliable , we present this result first because it magnifies the effect of the self - consistent @xmath5 which would be otherwise not easy to see . in the lower panel of fig .
[ fig : i ] we plot the chiral chemical potential for @xmath17 ( thick dot - dashed line ) and @xmath139 ( thin dot - dashed ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel . for comparison
we have also shown the chiral chemical potential obtained by @xmath29 computed with @xmath158 shown in the upper panel . comparing the results obtained with @xmath158 and self - consistent @xmath5 we notice a slight backreaction of the equilibrium chiral density on the quark condensate , which reflects in a small change of @xmath29 ; moreover the catalysis induced by @xmath5 is observed thanks to a slight shift of the inflection point of @xmath29 towards larger temperature .
however still the combined effect of @xmath4 at equilibrium and @xmath143 is to lower @xmath15 with respect to the case @xmath97 . _
( upper panel ) .
_ @xmath29 versus temperature .
maroon solid line corresponds to @xmath97 and @xmath158 .
indigo dashed line corresponds to @xmath97 and @xmath159 mev for @xmath17 and @xmath139 quarks .
green lines correspond to @xmath160 , orange lines to @xmath160 .
open symbols denote calculations at @xmath158 while thick lines are for results with @xmath5 computed self - consistently by the number equation and the gap equation . _
( middle panel)_. zoom of the upper panel in the temperature range of the chiral crossover . _ ( lower panel ) .
_ self - consistent @xmath5 for @xmath17 ( thick lines ) and @xmath139 ( thin lines ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel.,title="fig:",width=264 ] + _ ( upper panel ) . _
@xmath29 versus temperature .
maroon solid line corresponds to @xmath97 and @xmath158 .
indigo dashed line corresponds to @xmath97 and @xmath159 mev for @xmath17 and @xmath139 quarks .
green lines correspond to @xmath160 , orange lines to @xmath160 .
open symbols denote calculations at @xmath158 while thick lines are for results with @xmath5 computed self - consistently by the number equation and the gap equation .
_ ( middle panel)_. zoom of the upper panel in the temperature range of the chiral crossover . _
( lower panel ) . _ self - consistent @xmath5 for @xmath17 ( thick lines ) and @xmath139 ( thin lines ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel.,title="fig:",width=264 ] + _ ( upper panel ) .
_ @xmath29 versus temperature .
maroon solid line corresponds to @xmath97 and @xmath158 .
indigo dashed line corresponds to @xmath97 and @xmath159 mev for @xmath17 and @xmath139 quarks .
green lines correspond to @xmath160 , orange lines to @xmath160 .
open symbols denote calculations at @xmath158 while thick lines are for results with @xmath5 computed self - consistently by the number equation and the gap equation . _
( middle panel)_. zoom of the upper panel in the temperature range of the chiral crossover .
_ ( lower panel ) .
_ self - consistent @xmath5 for @xmath17 ( thick lines ) and @xmath139 ( thin lines ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel.,title="fig:",width=264 ] we have verified that this scenario is in qualitative agreement with the one obtained for smaller values of @xmath2 and @xmath3 , in which case our approximation should be quantitatively more reliable . in the upper panel fig .
[ fig : ii ] we plot @xmath29 versus temperature for @xmath160 ( orange lines ) and @xmath162 ( green lines ) , as well as for the case @xmath97 which we use as a benchmark ( solid maroon line ) . in the lower panel of the fig .
[ fig : ii ] we plot the chiral chemical potentials for @xmath17 and @xmath139 quarks at equilibrium computed self - consistently .
we find no qualitative difference between the cases of small and large fields . for completeness
we report the average value of @xmath4 in the crossover region , namely in the temperature range @xmath163 mev , which can be obtained directly by using eq .
we find it runs in the range @xmath164 @xmath165 in the case of @xmath162 , and @xmath166 @xmath165 in the case of @xmath160 .
the reason why @xmath29 is poorly affected by @xmath5 for small values of the fields is the different relative change of critical temperature induced by @xmath5 on the one hand , and the electric field on the other hand . in the case @xmath167 in fig .
[ fig : ii ] , the average values of @xmath5 are less than 10 mev in the crossover region .
taking @xmath158 , the effect of @xmath2 and @xmath3 is to lower the critical temperature of about the @xmath168 ; on the other hand taking @xmath97 and @xmath169 mev the shift of @xmath15 is practically zero . even increasing by hand the value of mu5 of a factor of 10 ,
the increase of @xmath15 due to mu5 is practically negligible compared to the lowering induced by the fields . _
( upper panel ) .
_ @xmath29 versus temperature .
maroon solid line corresponds to @xmath97 and @xmath158 .
green line corresponds to @xmath162 with @xmath170 .
data represented by indigo stars denote @xmath29 computed for @xmath162 with @xmath170 .
orange data correspond to @xmath171 .
finally turquoise plus correspond to a calculation at fixed value of @xmath172 mev . _
( lower panel ) .
_ self - consistent @xmath5 for @xmath17 ( thick lines ) and @xmath139 ( thin lines ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel.,title="fig:",width=264 ] + _ ( upper panel ) .
_ @xmath29 versus temperature .
maroon solid line corresponds to @xmath97 and @xmath158 .
green line corresponds to @xmath162 with @xmath170 .
data represented by indigo stars denote @xmath29 computed for @xmath162 with @xmath170 .
orange data correspond to @xmath171 .
finally turquoise plus correspond to a calculation at fixed value of @xmath172 mev . _
( lower panel ) .
_ self - consistent @xmath5 for @xmath17 ( thick lines ) and @xmath139 ( thin lines ) quarks versus temperature , corresponding to @xmath29 shown in the upper panel.,title="fig:",width=264 ] the results shown in fig .
[ fig : ii ] have been obtained for @xmath173 in eq .
we have checked the stability of the results in the case @xmath174 by increasing the relaxation time of an order of magnitude : we collect the results of this check in fig .
[ fig : ii2 ] . in the upper panel of fig .
[ fig : ii2 ] we plot @xmath29 versus temperature in the pseudo - critical region .
maroon and green lines represent the same quantities of fig .
[ fig : ii ] ; indigo stars correspond to @xmath29 computed with @xmath175 in eq .
eq : al3 , and turquoise plus denote the solution of the gap equation for a fixed value of @xmath172 mev .
we find that for large temperatures the effect of the larger relaxation appears as a tiny shift of @xmath29 towards larger values ; this can be understood because the values of @xmath176 , @xmath177 in this case are larger ot those found with @xmath173 , see lower panel of fig .
[ fig : ii2 ] .
however still the average value of the chiral chemical potentials is quite small in the pseudo - critical region . for comparison
we have shown the results of a computation at fixed value of @xmath172 mev in the figure : this value of chemical potential approximately corresponds to the average value @xmath178 computed self - consistently in the case @xmath175 at @xmath179 mev , see indigo stars in the lower panel of fig . [
fig : ii2 ] .
we find a fair agreement among the calculations with fixed and self - consistent @xmath5 , showing that the values of @xmath29 we obtain in the self - consistent calculation are indeed those expected .
we notice that in the case @xmath180 , shown in fig .
[ fig : ii2 ] by orange data , we measure a slightly larger increase of @xmath29 due to @xmath14 in the crossover region .
this result clearly shows how a large @xmath5 would affect the thermodynamics balancing the effect of the external fields ; the concrete value of the average @xmath5 we have in this case however runs in the range @xmath181 mev , so the result should not be trusted quantitatively . _
( upper panel ) .
_ @xmath29 versus temperature for @xmath182 . as benchmarks
we plot by red solid line @xmath29 for the zero field case , and by .
circles @xmath29 with @xmath183 .
diamonds correspond to @xmath170 , triangles to @xmath184.,width=264 ] we have performed the stability check against variations of @xmath138 or the case @xmath160 discussed above .
we plot in fig.[fig : ii2a ] the result of this check for the cases of @xmath170 ( diamonds ) and @xmath185 ( triangles ) .
we have found that taking @xmath186 affects @xmath29 considerably , hence showing a net effect of chiral density on the phase transition .
however we take this result not too seriously because doubling @xmath138 would roughly correspond to double @xmath5 , which is already quite large in the pseudo - critical region as it is shown in fig .
[ fig : ii ] hence making the use of our approximation questionable .
our conclusion is that as long as the values of @xmath2 and @xmath5 are not too large , our approximate solution to the self - consistent problem is fairly good , while for larger values of the background field it has to be taken with a grain of salt .
@xmath29 versus temperature for @xmath174 and several values of @xmath71.,width=264 ] in fig .
[ fig : iiab ] we plot @xmath29 versus temperature for @xmath174 and several values of @xmath71 .
the computations have been performed by taking into account the dynamically generated @xmath4 for @xmath17 and @xmath139 quarks .
it is interesting that even if the magnetic field acts as a catalyzer of chiral symmetry breaking at small temperatures , the presence of the electric field helps an inverse catalysis in the critical region : for example in the case @xmath187 we find that @xmath29 close to the critical temperature still sits on the zero field result ; increasing the magnitude of @xmath2 just moves the critical temperature to a lower value .
this is in agreement with the analytical discussion of section iv .
in this article we have studied spontaneous chiral symmetry breaking for quark matter in the background of static , homogeneous and parallel electric field @xmath0 and magnetic field @xmath1 .
we have used a nambu - jona - lasinio model with a local kernel interaction to compute the relevant quantities to describe chiral symmetry breaking at finite temperature for a wide range of @xmath2 and @xmath3 .
part of our study has been devoted to a mean field calculation of the response of the chiral condensate to the external fields , both at zero and at nonzero temperature .
we have derived both numerically and analytically the magnetic catalysis and the electric inverse catalysis at zero temperature ; we have also studied the behaviour of the quark condensate at finite temperature , finding a competition between the magnetic and electric fields which affects the critical temperature .
we have not considered a by - hand modification of the njl coupling constant in order to reproduce inverse magnetic catalysis for small @xmath3 at finite temperature , because this would have masked the genuine response of the model to an electric field , but we will certainly consider this necessary modification to the interation term in the future .
our result in this direction is that the critical temperature for chiral symmetry restoration , @xmath15 , is lowered by the simultaneous presence of the parallel electric and magnetic fields .
we have then focused on the effect of equilibration of chiral density , @xmath4 , produced dynamically by axial anomaly on the critical temperature .
chiral density is produced thanks to schwinger tunneling and spin alignment in the magnetic field .
the equilibration of @xmath4 happens as a consequence of chirality flipping processes in the thermal bath ; we have introduced the relaxation time for chirality , namely @xmath9 , giving the time scale necessary for the equilibration of @xmath4 . in absence of a specific calculation of @xmath9 it
is possible to give only an ansatz ; we chose @xmath28 where @xmath29 is the constituent quark mass . because this dynamical system reaches a thermodynamical equilibrium state for @xmath22 , with a specified value of @xmath188 depending on the actual values of the field and of the temperature , it is possible to introduce the chiral chemical potential , @xmath5 , conjugated to @xmath23 at equilibrium .
the value of @xmath5 has been computed by coupling the gap equation to the number equation , at the leading order in @xmath189 , @xmath190 and @xmath24 .
because of the different electric charges of @xmath17 and @xmath139 quarks at equilibrium @xmath191 and the ratio of the two is about @xmath192 in the critical region ; we have therefore introduced two chemical potentials , @xmath176 and @xmath177 conjugated respectively to @xmath193 and @xmath194 .
we have found that the equilibrated chiral density does not change drastically the thermodynamics as long as @xmath5 at equilibrium is not too large ; namely , the inverse catalysis effect induced by the background fields is not spoiled by the presence of the @xmath5 background .
the weak effect of @xmath5 on the shift of @xmath15 in presence of the background fields can be understood because the change of @xmath15 induced by @xmath5 itself are smaller than the ones induced by the background fields .
for example in the case @xmath158 , the effect of the background fields is to lower the critical temperature of about the @xmath168 ; on the other hand taking @xmath97 and @xmath169 mev which corresponds to the average value of the chemical potential we find in the crossover region , the shift of @xmath15 is practically zero . this conclusion might be no longer valid in the case of large @xmath5 . for larger values of the fields
we have found that @xmath29 is effectively pushed towards larger values in the critical region by @xmath14 . a firm conclusion about this finding
can be achieved however only by solving the problem beyond the perturbative analysis used in our study .
we would like to remark that the results presented here have to be considered only explorative : the study of this problem beyond the the weak field and small @xmath5 approximation will be the topic of upcoming research .
moreover , the theoretical calculation of the equilibrium value of @xmath5 has an uncertainty because of the lack of information about the relaxation time for chirality flipping processes , @xmath9 in eq .
eq : c , whose computation will be the theme of near future research .
d. e. kharzeev , l. d. mclerran and h. j. warringa , nucl .
a * 803 * , 227 ( 2008 ) .
k. fukushima , d. e. kharzeev and h. j. warringa , phys .
d * 78 * , 074033 ( 2008 ) .
q. li _ et al .
_ , arxiv:1412.6543 [ cond-mat.str-el ]
. d. t. son and p. surowka , phys .
lett . * 103 * , 191601 ( 2009 ) .
n. banerjee , j. bhattacharya , s. bhattacharyya , s. dutta , r. loganayagam and p. surowka , jhep * 1101 * , 094 ( 2011 ) . k. landsteiner , e. megias and f. pena - benitez , phys .
lett . *
107 * , 021601 ( 2011 ) .
d. t. son and a. r. zhitnitsky , phys .
d * 70 * , 074018 ( 2004 ) .
m. a. metlitski and a. r. zhitnitsky , phys .
d * 72 * , 045011 ( 2005 ) .
d. e. kharzeev and h. u. yee , phys .
d * 83 * , 085007 ( 2011 ) .
m. n. chernodub , jhep * 1601 * , 100 ( 2016 ) . m. n. chernodub and m. zubkov , arxiv:1508.03114 [ cond-mat.mes-hall ] .
m. n. chernodub , a. cortijo , a. g. grushin , k. landsteiner and m. a. h. vozmediano , phys .
b * 89 * , no . 8 ,
081407 ( 2014 ) . v. braguta , m. n. chernodub , k. landsteiner , m. i. polikarpov and m. v. ulybyshev , phys . rev .
d * 88 * , 071501 ( 2013 ) .
a. v. sadofyev and m. v. isachenkov , phys .
b * 697 * , 404 ( 2011 ) .
a. v. sadofyev , v. i. shevchenko and v. i. zakharov , phys .
d * 83 * , 105025 ( 2011 ) .
z. v. khaidukov , v. p. kirilin , a. v. sadofyev and v. i. zakharov , arxiv:1307.0138 [ hep - th ] . v. p. kirilin , a. v. sadofyev and v. i. zakharov , arxiv:1312.0895 [ hep - th ] .
a. avdoshkin , v. p. kirilin , a. v. sadofyev and v. i. zakharov , phys .
b * 755 * , 1 ( 2016 ) .
r. gatto and m. ruggieri , phys .
d * 85 * , 054013 ( 2012 ) .
k. fukushima , m. ruggieri and r. gatto , phys .
d * 81 * , 114031 ( 2010 ) .
m. n. chernodub and a. s. nedelin , phys .
d * 83 * , 105008 ( 2011 ) .
m. ruggieri , phys .
d * 84 * , 014011 ( 2011 ) .
l. yu , h. liu and m. huang , arxiv:1511.03073 [ hep - ph ] .
l. yu , j. van doorsselaere and m. huang , phys . rev .
d * 91 * , no .
7 , 074011 ( 2015 ) .
m. frasca , arxiv:1602.04654 [ hep - ph ] . v. v. braguta , e .-
ilgenfritz , a. y. kotov , b. petersson and s. a. skinderev , arxiv:1512.05873 [ hep - lat ] . v. v. braguta , v. a. goy , e .-
ilgenfritz , a. y. kotov , a. v. molochkov , m. muller - preussker and b. petersson , jhep * 1506 * , 094 ( 2015 ) .
v. v. braguta and a. y. kotov , arxiv:1601.04957 [ hep - th ] .
m. hanada and n. yamamoto , pos lattice * 2011 * , 221 ( 2011 ) [ arxiv:1111.3391 [ hep - lat ] ] .
s. s. xu , z. f. cui , b. wang , y. m. shi , y. c. yang and h. s. zong , phys .
d * 91 * , no .
5 , 056003 ( 2015 ) .
a. y. babansky , e. v. gorbar and g. v. shchepanyuk , phys .
b * 419 * , 272 ( 1998 ) .
s. p. klevansky and r. h. lemmer , phys .
d * 39 * , 3478 ( 1989 ) . h. suganuma and t. tatsumi , annals phys . * 208 * , 470 ( 1991 ) .
k. g. klimenko , theor .
phys . * 90 * , 1 ( 1992 ) [ teor . mat .
* 90 * , 3 ( 1992 ) ] .
i. v. krive and s. a. naftulin , phys .
d * 46 * , 2737 ( 1992 ) .
v. p. gusynin , v. a. miransky and i. a. shovkovy , phys .
* 73 * , 3499 ( 1994 ) , erratum : [ phys .
lett . * 76 * , 1005 ( 1996 ) ] .
v. p. gusynin , v. a. miransky and i. a. shovkovy , phys .
b * 349 * , 477 ( 1995 ) .
g. cao and x. g. huang , phys .
b * 757 * , 1 ( 2016 ) .
m. delia , m. mariti and f. negro , phys .
lett . * 110 * , no . 8 , 082002 ( 2013 ) .
y. nambu and g. jona - lasinio , phys . rev . *
122 * , 345 ( 1961 ) .
y. nambu and g. jona - lasinio , phys .
124 * , 246 ( 1961 ) .
s. p. klevansky , rev .
phys . * 64 * , 649 ( 1992 ) .
t. hatsuda and t. kunihiro , phys .
rept . *
247 * , 221 ( 1994 ) .
a. i. nikishov , zh . eksp .
fiz . * 57 * , 1210 ( 1969 ) .
t. d. cohen and d. a. mcgady , phys .
d * 78 * , 036008 ( 2008 ) .
f. v. bunkin and i. i. tugov , dokl .
nauk ser .
fiz . * 187 * , 541 ( 1969 ) .
g. v. dunne , in * shifman , m. ( ed . )
et al . : from fields to strings , vol . 1 * 445 - 522 [ hep - th/0406216 ] . | in this article we study spontaneous chiral symmetry breaking for quark matter in the background of static and homogeneous parallel electric field @xmath0 and magnetic field @xmath1 .
we use a nambu - jona - lasinio model with a local kernel interaction to compute the relevant quantities to describe chiral symmetry breaking at finite temperature for a wide range of @xmath2 and @xmath3 .
we study the effect of this background on inverse catalysis of chiral symmetry breaking for @xmath2 and @xmath3 of the same order of magnitude .
we then focus on the effect of equilibration of chiral density , @xmath4 , produced dynamically by axial anomaly on the critical temperature .
the equilibration of @xmath4 , a consequence of chirality flipping processes in the thermal bath , allows for the introduction of the chiral chemical potential , @xmath5 , which is computed self - consistently as a function of temperature and field strength by coupling the number equation to the gap equation , and solving the two within an expansion in @xmath6 , @xmath7 and @xmath8 .
we find that even if chirality is produced and equilibrates within a relaxation time @xmath9 , it does not change drastically the thermodynamics , with particular reference to the inverse catalysis induced by the external fields , as long as the average @xmath5 at equilibrium is not too large . |
elementary excitations such as electrons and phonons can be probed by the inelastic scattering of light using the raman spectroscopy technique . in graphene - related systems , studying the shape of the raman spectra can give us a deep understanding of the electron energy dispersion , @xcite phonon energy dispersion , @xcite lifetime of excitations , @xcite the kohn anomaly effect , @xcite and structure characterization .
@xcite in particular , the asymmetric breit - wigner - fano ( bwf ) lineshape , historically observed in the raman spectra of graphite interacalation compounds ( gics ) @xcite and metallic nanotubes ( m - swnts ) , @xcite probes interference between the continuum spectra with discrete spectra .
@xcite recently , the bwf asymmetry has been observed by yoon _ et al . _
@xcite in monolayer graphene indicating a common origin of the bwf lineshape of the graphite - related systems ( i.e. gics , m - swnts , and monolayer graphene ) that arise due to the presence of the dirac cone or linear energy band structure .
the bwf lineshape is defined by the following formula @xmath1,\end{aligned}\ ] ] where @xmath2 .
here @xmath3 , @xmath4 , @xmath5 , @xmath6 , and @xmath7 are the raman shift , the spectral peak position , the asymmetric factor , the spectral width , and the maximum intensity of the bwf spectra , respectively . the right hand side of eq .
tells us that the bwf lineshape respectively consists of a constant continuum spectrum , a discrete lorentzian spectrum , and an interference effect between both spectra .
when @xmath8 , eq .
gives a lorentzian lineshape which represents a discrete phonon spectrum .
the interference term gives rise to an asymmetric lineshape for positive and negative values of @xmath9 , in which the asymmetry is proportional to a dimensionless parameter @xmath5 , mimicking the ratio between the probability amplitude of the continuum spectra to that of the discrete spectra .
@xcite in the raman spectroscopy studies of graphite - related systems , @xcite continuum spectra come from the electronic excitations and are usually observed only in metallic systems .
the bwf lineshapes in graphene have been found in various kinds of phenomena such as scanning tunneling microscopy , @xcite optical conductivity , @xcite photoabsorption spectroscopy , @xcite and infrared spectroscopy @xcite revealing that electron - hole pair excitations in the vicinity of the dirac cone play an important role in the continuum spectra .
the asymmetric bwf lineshapes in graphite - related systems are normally found in the raman shift around @xmath10 , known as the @xmath11 modes , which correspond to two zone - center ( @xmath12 ) phonon modes , namely the in - plane tangential optic ( ito ) and longitudinal optic modes . in graphene , the bwf asymmetry of the @xmath11 band is observed using the gate - modulated raman spectroscopy .
@xcite the asymmetric factor ( @xmath5 ) has a value around @xmath13 or one - order of magnitude smaller than those found in m - swnts ( @xmath14 ) @xcite and in gics ( @xmath15 ) .
@xcite the absolute value of the bwf asymmetric factor greatly decreases as we change the fermi energy ( @xmath16 ) to be further from the dirac point by applying a positive or a negative bias with respect to the charge neutrality point .
@xcite these results give a clue that the asymmetric factor strongly depends on the electronic density of states ( dos ) near the dirac cone . in this work ,
we show that the origin of the bwf spectra in graphene comes from the continuous single particle electron - hole pair spectra , interfering with the discrete phonon spectra .
hereafter , we refer to the single particle electron - hole pair spectra as the electronic raman spectra ( ers ) .
@xcite in the previous work for m - swnts , we discuss that the ers spectra originate from the second - order coulomb interaction with non - zero momentum transfer @xmath17 , due to the symmetry of the a and b sublattice wavefunctions which gives rise to the absence of the direct coulomb interaction at the zone center @xmath12 .
@xcite unlike the previous calculation for m - swnts which utilized exciton wavefunctions , @xcite in this calculation we use electron wavefunctions from the tight binding ( tb ) method because our calculation regime ( @xmath18 ev ) is far from the saddle point energy dispersion ( @xmath19 ev ) , and thus the exciton effects are negligible .
@xcite the use of electron wavefunctions give considerable contributions of the intervalley scattering to the raman intensity which was neglected in the previous study.@xcite after calculating the raman amplitudes of the ers and the phonon spectra , we found that the interference between the ers and the phonon spectra gives a drastic change in the constructive - destructive interference near the phonon spectra , giving an asymmetry to the phonon lineshape when fitted to the bwf lineshape . by considering the second - order raman process , we systematically reproduce the @xmath16 dependence of the raman spectra of graphene that was observed in experiment .
@xcite our paper is organized as follows . in sec .
[ sec : theor ] we describe our calculation of the electron - electron interaction using the tb method and considering up to second - order raman processes . in sec .
[ sec : ers ] , we discuss the calculated ers spectra as a function of @xmath16 and compare the asymmetric bwf factor @xmath5 obtained from our calculation with that from the experiment . finally , in sec .
[ sec : summ ] we give a summary of this work .
the possible ers processes are described in fig .
[ fig1 ] , which consist of either intravalley ( a ) or intervalley ( e ) interaction , either intravalley ( a ) or intervalley ( e ) scattering and either zero momentum transfer ( @xmath20 first - order ) or non - zero momentum transfer ( @xmath21 second - order ) processes .
@xcite when a photon with the laser excitation energy @xmath22 is introduced to the graphene sample , the photon excites an electron from an initial state @xmath23 to an intermediate state @xmath24 with an energy matched to @xmath22 ( incident resonance ) .
this photoexcited electron ( pe ) is then scattered to another intermediate state @xmath25 by the dirac electrons ( des ) on the dirac cone and the electron finally recombines with a hole by emitting a scattered photon energy @xmath26 as shown in fig . [ fig1 ] .
the coulomb interaction between the pe and the des causes the pe to reduce its energy and changes pe s momentum while the des are being excited .
the number of des to be excited for each process depends on the number of the scattering order . in the first - order process , only one de is excited and this process requires a zero momentum transfer ( @xmath27 ) since the pe momentum ( @xmath28 ) should be the same as its hole momentum in order to emit a scattered photon with energy @xmath26 by the electron - hole recombination process . in the second - order process , on the other hand , the pe is scattered twice ( @xmath29 ) and ( @xmath30 ) and the pe excites two des with relative non - zero electron - hole momenta @xmath31 and @xmath32 . due to the degeneracy of the dirac cone at the @xmath33 and @xmath34 points of the graphene brillouin zone ( bz ) ,
both the first - order processes and the second - order processes may occur in the intravalley ( a ) interactions or in the intervalley ( e ) interactions . in the a interactions ,
the des are excited on the same dirac cone as the pe , while in the e interactions , the des are excited on the other dirac cone . in the case of the e interaction , the initial and final states of the pe and des
can be in the same ( different ) valley which is defined by intravalley ( intervalley ) scattering labelled by a small letter `` a '' ( `` e '' ) .
the e scattering is not possible in the a interaction because @xmath35 and @xmath31 scattering are pointing to two different directions at the high symmetry points of graphene ; one is pointing to the @xmath36 direction while the other is pointing to the @xmath37 direction .
thus the ae interaction does not conserve energy during the scattering processes . combining all possible a and e interactions with the a and e scatterings we have : an aa [ fig .
[ fig1](a ) ] and an ea [ fig .
[ fig1](b ) ] in the first - order processes ; and an aa [ fig .
[ fig1](c ) ] , an ea [ fig . [ fig1](d ) ] , and an ee [ fig .
[ fig1](e ) ] in the second - order processes .
the bwf asymmetry comes from the interference effect mentioned above because both the ers and the phonon spectra have the same initial and final states for a single pe .
based on this standpoint , we define the raman intensity : @xmath38 ^ 2,\ ] ] where @xmath39 , in which @xmath40 and @xmath41 are , respectively , the @xmath42-th phonon scattering amplitude and the ers scattering amplitude . the phonon scattering amplitude in the resonance raman spectra
is given by @xcite @xmath43}\nonumber\\ & \times \frac{{\cal m}_{{\rm el-}\nu}^{n^\prime , n } } { \left[\delta e_{n^\prime i}-\hbar\omega_{g}-i(\gamma+\gamma_{\nu})\right]}\nonumber\\ & \times \frac{{\cal m}_{\rm el - op}^{f , n^\prime } } { \left[e_{\mathrm{l}}-\hbar\omega_{g}-\hbar \omega_{\rm s } -i\gamma_{\nu}\right ] } \bigg ] , \label{eq : amplitude}\end{aligned}\ ] ] where for the phonon modes we only consider the first - order process @xmath42 = ito or lo modes , and @xmath44 .
here we use a broadening factor @xmath45 ev , which is related to the inverse of the life time of the photoexcited carriers . on the other hand ,
@xmath46 is related to the life time of the electron - phonon interaction .
@xcite the values of @xmath46 and @xmath4 are considered as follows . in the gate - modulated raman spectra
, we expect phonon frequency softening and spectral broadening as we shift the fermi energy from @xmath47 to @xmath48 .
this effect is due to the kohn anomaly , i.e. renormalization of the phonon energy by electron - hole pair excitation in the @xmath11 mode raman spectra of graphene .
@xcite we do not consider the kohn anomaly effect explicitly in this calculation , but we can fit the peak position @xmath49 for @xmath50 ev and @xmath51 for @xmath52 ev .
the inverse of the phonon life time is also fitted by @xmath53 for @xmath54 ev and @xmath55 for @xmath56 ev so as to reproduce the experimental results .
@xcite it is important to note that the kohn anomaly does not give the asymmetric bwf of the @xmath11 band spectra because the kohn anomaly is not an inteference phenomenon ; only the interference effect between the @xmath11 band and the ers does however show a bwf lineshape .
the electron - photon ( @xmath57 ) and electron - phonon ( @xmath58 ) matrix elements for a transition between states @xmath59 are adopted from previous works within the tb method .
@xcite we approximate the intermediate states ( virtual states ) to become a real state with @xmath60 , which is a good approximation for the resonance condition .
@xcite the ers amplitude @xmath41 is the summation of the amplitude from the first - order @xmath61 and second - order @xmath62 processes . the amplitude of the first - order ers process is given by @xmath63}\nonumber\\ & \times \frac{k_{n^\prime , l^\prime , n , l}(0 ) } { \left[\delta e_{n^\prime i}-\hbar\omega_{\rm e}-i(\gamma+\gamma_{\rm e})\right]}\nonumber\\ & \times \frac{{\cal m}_{\rm el - op}^{f , n^\prime } } { \left[e_{\mathrm{l}}-\hbar\omega_{\rm e}-\hbar \omega_\mathrm{s } -i\gamma_{\rm e}\right ] } \bigg ] , \label{eq : ampli1}\end{aligned}\ ] ] where @xmath64 and @xmath65 mev are , respectively , the energy of the excited de electron and the inverse life time of the electron - electron interaction .
the electron - electron interaction @xmath66 defines the scattering of the pe [ de ] from an initial state @xmath67 $ ] to a final state @xmath68 $ ] which consists of direct @xmath69 and exchange @xmath70 interaction terms , @xmath71 for a spin singlet state .
we do not consider spin triplet states for simplicity due to the fact that the exchange interaction is sufficiently small .
@xcite the direct @xmath72 and exchange @xmath73 coulomb interactions between two electrons in the tb approximation is given by @xmath74 , \label{eq : kdfin}\\ k_{1,2,3,4}^{\rm x}({\bf q})= & \sum_{ss^\prime={\rm a , b}}c_{s}^1 c_{s^\prime}^2 c_{s^\prime}^{*3 } c_{s}^{*4 } \re \left[v_{ss^\prime } \left(\mathbf{\kvec^\prime-\kvec-\qvec}\right ) \right ] , \label{eq : kxfin}\end{aligned}\ ] ] where @xmath75=[\kvec\cond,\kvec^\prime\val,(\kvec-\qvec)\cond , ( \kvec^\prime+\qvec)\cond]$ ] in the case of ers in undoped - graphene ( @xmath48 ) [ fig .
[ fig2](a ) ] . in the electron doped ( @xmath76 ) and
the hole doped ( @xmath77 ) cases , we add possible intraband transitions @xmath78=[\kvec^\prime\cond,(\kvec^\prime+\qvec)\cond]$ ] and @xmath78=[\kvec^\prime\val,(\kvec^\prime+\qvec)\val]$ ] , respectively , as long as state @xmath79 is occupied and state @xmath80 is unoccupied .
@xmath81 is a tight binding coefficient for an atomic site @xmath82 and a state @xmath83 .
@xcite the fourier transform of the coulomb potential @xmath84 is defined by @xmath85 where @xmath86 is the effective coulomb potential for the @xmath87 electron system modeled by the ohno potential @xcite @xmath88 in which @xmath89 is the on - site coulomb potential for two @xmath87 electrons at the same site @xmath90 , defined by @xmath91 the amplitude of the second - order ers process is given by @xmath92 } \nonumber\\&\times \frac{k^{\rm d}_{n^\prime , m^\prime , n , m}({\bf q } ) } { \left [ \delta e_{n^\prime i}-\hbar\omega_{1 } - i(\gamma+\gamma_{\mathrm{e } } ) \right ] } \nonumber\\ & \times \frac{k^{\rm d}_{n^{\prime\prime},l^\prime , n^\prime , l}(-{\bf q } ) } { \left[\delta e_{n^{\prime\prime } i}-\hbar\omega_{1}-\hbar\omega_{2}-i(\gamma+ 2\gamma_{\mathrm{e}})\right]}\nonumber\\ & \times \frac{{\cal m}_{\rm el - op}^{f , n^\prime } } { \left[e_{\mathrm{l}}-\hbar\omega_{1}-\hbar\omega_{2 } - \hbar\omega_{\mathrm{s}}-2i\gamma_{\mathrm{e}}\right ] } \bigg],\end{aligned}\ ] ] where we also consider the same virtual state approximation as in eq . . here ,
@xmath93 and @xmath94 are the energies of the des emitted for the electron - electron interaction in the second - order ers process .
( color online ) ( a ) ilustration of the direct coulomb intravalley ( a ) interaction and intervalley ( e ) interaction .
( b ) the averaged absolute value of the direct coulomb interaction matrix element @xmath95 as a function of momentum transfer @xmath32 for the intravalley ( a ) interaction and intervalley ( e ) interaction .
the intervalley ( e ) scattering is not shown in this figure.,width=302 ] ( color online ) ( a ) calculated results of the total raman intensity in eq . for @xmath96 ( solid line ) compared with the lorentzian @xmath11 mode intensity by taking the square of @xmath97 in eq .
( dashed line ) .
the @xmath11 mode constituents , i.e. ito and lo , are indicated by a dotted line and a dot - dashed line , respectively . an asymmetric lineshape ( solid line )
appears due to the interference effect of the @xmath11 mode with the ers .
the inset shows calculated results of the first - order ( dashed line ) and the second - order ( solid line ) ers spectra indicating that the second - order processes have an intensity value six - orders of magnitude greater than that of the first - order processes .
( b ) calculated results of the total raman intensity for @xmath98 ( solid line ) and @xmath99 ( dashed line ) . the bwf asymmetric factor @xmath5 decreases by increasing the absolute value of @xmath100 away from the dirac point because the ers intensity also decreases by increasing @xmath100 ( inset).,width=302 ] since the electron - electron interaction depends on initial states @xmath101 of pe and de and also on a momentum transfer ( @xmath32 ) , we consider the absolute average value of the matrix elements over the initial states in order to visualize the strength of the electron - electron interaction in a simple manner .
the direct coulomb interaction can occur in either the a or e interaction as shown in fig .
[ fig2](a ) . figure [ fig2](b )
depicts the absolute average value of @xmath102 over the initial states @xmath101 @xmath103 where @xmath104 a and e. the e scattering is not shown in fig .
[ fig2](b ) for convenient comparison between the a and e interaction , since the a interaction does not have the e scattering . @xmath105 only depends on @xmath32 after taking the summation over the initial states @xmath101 because the final states @xmath106 depend on @xmath101 by momentum conservation in eq . .
as shown in fig .
[ fig2](b ) , for both the a and e interactions , @xmath95 disappears at @xmath12 , indicated by a small dot at @xmath107 , due to the symmetry of the a and b sublattice wavefunctions in the graphene unit cell which cancel in the summation of @xmath108 in eq . .
@xcite the absence of the direct coulomb interaction suggests that the ers should come from the second order @xmath109 electron - electron interaction , similar to what we found in m - swnts .
@xcite the first - order ers can only be possible by means of the exchange coulomb interaction .
although we take into account the exchange coulomb interaction , the raman intensity from the first - order process is still six - orders of magnitude smaller than that of the second order process [ see inset of fig . [ fig3](a ) ] .
therefore , we can neglect the first - order processes for both the a and e interactions . in fig .
[ fig3 ] we present the raman intensity calculation @xmath110 of eq . .
the solid curve in fig . [ fig3](a ) shows the total raman intensity after considering the interference of the @xmath11 mode spectra with the ers spectra , while the dashed line shows the lorentzian @xmath11 phonon spectra by taking the square of its probability amplitudes @xmath111 [ eq . ] .
the @xmath11 mode constituents , i.e. the ito and lo modes , are indicated by a dotted line and a dot - dashed line , respectively .
it is clear from fig .
[ fig3](a ) that the calculated raman spectra shows asymmetry around the peak position at @xmath112 . by fitting the calculated result to eq .
, we obtain the fitted values of @xmath5 , which have the same negative sign as the experimental data .
@xcite for a negative @xmath5 , when @xmath113 is smaller ( greater ) than @xmath4 , @xmath114 is greater ( smaller ) than @xmath115 , indicating that the interference between the @xmath11 mode and the ers spectra is constructive ( destructive ) below ( above ) the resonance condition @xmath116 . by decreasing ( increasing ) @xmath16 further from the dirac cone ,
transitions from ( to ) the unoccupied ( occupied ) states are suppressed due to the pauli principle .
thus we expect that the asymmetric factor @xmath5 decreases as we change the @xmath16 from the dirac point @xmath117 ev to @xmath118 ev as shown in fig .
[ fig3](b ) .
the solid line is the intensity of the spectrum with @xmath119 when @xmath117 ev , while the dashed line is the corresponding curve with @xmath120 when @xmath118 ev .
the raman intensity and peak position at @xmath121 ev are larger than that at @xmath117 ev due to the kohn anomaly effect .
@xcite unlike the ers spectra in m - swnts which are lorentzian functions , @xcite the ers intensity in graphene is a linear function of @xmath113 [ inset of fig .
[ fig3](b ) ] .
the positive gradient of the ers intensity is due to the greater scattering path available to excite des in the second - order processes as @xmath113 increases .
the ers intensity will increase monotonically and will get saturated at @xmath122 .
the absence of the ers peak intensity in graphene is related to the absence of van - hove singularities within the @xmath11 mode energy @xmath123 ev to @xmath124 ev .
the absence of the ers peak also becomes the reason why the @xmath5 values of the @xmath11 mode in graphene are one - order of magnitude smaller compared to that in m - swnts .
the ers intensity is about two - orders of magnitude smaller than that of the @xmath11 mode , and by increasing the @xmath16 the ers intensity decreases only less than @xmath125% ; nevertheless the change of the @xmath5 is significant [ fig .
[ fig3](b ) ] . thus
, this bwf feature is very sensitive to the presence or absence of the continuum spectra . in figures [ fig4](a ) and
( b ) , we respectively show our calculated result and corresponding experimental results ( ref . ) of the @xmath11 band raman intensity as a function of raman shift , which is plotted for various values of @xmath16 in the range @xmath126 ev . in the original version ,
@xcite fig .
[ fig4](b ) was given as a function of gate voltages @xmath127 .
for our purpose of comparing the calculated results and experimental results , here we convert @xmath127 to @xmath16 using the relation @xmath128 where the fermi velocity @xmath129 , the constant voltage adjusted to the dirac point @xmath130 , and the capacitance @xmath131 for the @xmath132 dielectric medium with a thickness @xmath133 .
@xcite at the charge neutrality point @xmath117 ev , the @xmath11 band spectrum is broadened and its frequency is softened due to the kohn anomaly effect .
comparison of the bwf asymmetric factor @xmath5 between the theory ( square ) and experiment ( circle ) shows a reasonable agreement in fig .
[ fig4](c ) except for @xmath134 ev , when the experimental results deviate from the calculated results .
we suppose that the deviation is related to the difficulties of observing the bwf asymmetry at @xmath135 ev in the experiment because the continuum ers intensity is about two or three - orders of magnitude smaller compared to the @xmath11 band intensity .
such weak ers spectra might couple strongly with the background spectra in the experiment which make it difficult to observe .
the calculated asymmetric factor @xmath5 has a `` v''-shaped curve structure as a function of @xmath16 with the dip position at @xmath136 ev .
the decrease of @xmath5 is related to the decrease of the ers intensity due to the suppression of electron - hole pair excitations on the dirac cone upon doping .
the present agreement also reconfirms that plasmons do not contribute to the continuum spectra .
the reason is as follows . when @xmath137 , collective excitations ( plasmons ) are expected to be generated , and consequently the ers spectra should be enchanced .
@xcite however , what we obtain in the present study is that the ers spectra are in fact suppresed if we increase @xmath100 .
therefore , we rule out the contribution of plasmons in the ers spectra and we conclude that only single - particle electron - hole pair excitations are important .
we show that the origin of the bwf spectra in graphene comes from the continuum single particle electron - hole pair ers spectra interfering with the discrete @xmath11 band phonon spectra . after calculating the raman amplitudes of the ers and the phonon spectra
, we found that the interference effect between the ers and the phonon spectra gives a drastic change in the constructive - destructive interference near the phonon resonance condition , leading to an asymmetry of the phonon lineshape when fitted to the bwf lineshape . by considering the second - order raman process
, we are able to reproduce the @xmath16 dependence of the raman spectra systematically .
we expect that the asymmetric bwf feature appears generally in the phonon raman spectra of all dirac cone systems .
the authors are grateful to duhee yoon for sharing information regarding his gate - modulated raman spectroscopy experiment ( ref . ) .
e.h . is supported by a mext scholarship .
acknowledges a jsps research fellowship for young scientists no .
r.s . acknowledges mext grants no .
25286005 and no .
m.s.d . acknowledges nsf - dmr grant no .
10 - 04147 . | excitation of electron - hole pairs in the vicinity of the dirac cone by the coulomb interaction gives rise to an asymmetric breit - wigner - fano lineshape in the phonon raman spectra in graphene .
this asymmetric lineshape appears due to the interference effect between the phonon spectra and the electron - hole pair excitation spectra .
the calculated breit - wigner - fano asymmetric factor @xmath0 as a function of the fermi energy shows a `` v''-shaped curve with a minimum value at the charge neutrality point and gives good agreement with the experimental result . |
interlayer exchange coupling ( iec ) was discovered in late 80-ties in fe / cr / fe trilayers @xcite . since then it has been observed in a variety of multilayer structures composed of alternating magnetic and nonmagnetic layers .
these studies concentrated on the coupling between ferromagnetic , metallic layers , although separated by both metallic and insulating spacers .
thus , the ferromagnetic character of the magnetic layers and the fact that in these structures the fermi level was situated in the region of high density of electronic states inhered to the theoretical models , which were designed to explain the origins of the iec phenomena ( see @xcite and the references therein ) . surprisingly enough , the iec was also discovered in all - semiconductor superlattices ( sls ) .
moreover , the semiconductor sls in which it was first observed were the mnte / cdte @xcite , mnte / znte @xcite and eute / pbte @xcite , all with antiferromagnetic layers .
recently , such coupling was also identified in semiconductor multilayer structures with ferromagnetic eus , @xcite and ga(mn)as layers @xcite . while the iec in the trilayer ga(mn)as / ga(al)as / ga(mn)as with the high concentration of free carriers , @xcite , can be , at least qualitatively , explained in terms of the models tailored for metallic systems , the other semiconductor structures exhibiting iec call for a different approach .
several attempts to explain iec in all - semiconductor structures have already been reported in the literature .
two models , in which the interlayer coupling is mediated by carriers localized on shallow impurities in the spacer region , were proposed for ii - vi sls,@xcite .
these models do not apply to iv - vi structures , with the pbte and pbs spacers , since in lead chalcogenides localized shallow impurity states were never detected @xcite . for these sls , mechanisms of interlayer spin - spin interactions mediated by valence - band electrons were suggested .
the calculations of the difference between total electronic energies obtained for two different ( b1 and b2 in fig .
[ a - fm ] ) spin configurations of the sl , performed within a frame of a very simple 1d tight - binding model , put first in evidence the significant role of the valence band electrons in iec in all - semiconductor magnetic / nonmagnetic layer structures @xcite .
this role was further demonstrated for eute / pbte / eute trilayers by wilczyski and wirkowicz in @xcite , where a 3d tight binding model , still oversimplifying the band structure , was used . a different approach to the magnetic interlayer interactions mediated by valence electrons
have been chosen by dugaev et al , @xcite .
these authors studied the blombergen - rowland mechanism within the effective mass approximation and obtained a ferromagnetic coupling between two magnetic impurities situated at the opposite interfaces of a narrow - gap iv - vi semiconductor spacer .
as the experimentally observed iec in eus / pbs sls is antiferromagnetic , @xcite , this means that the blombergen - rowland interactions are not dominating iec in these sls .
still , because of the low concentration of free carriers and the absence of shallow impurity states in pbs , the iec in these sls is most likely mediated by valence electrons . in this situation , the total energy calculations , which do not focus on a particular interaction mechanism but account globally for the spin - dependent structure of valence bands , seem to constitute the most appropriate approach .
the calculations of this type , reported in @xcite and @xcite , were clearly oversimplified and performed for a different spin structure than that of eus / pbs ( 001 ) sl .
in this paper we present the results of refined 3d total energy calculations , which take into account the crystal and the band structures of the sls component materials . the tight - binding model with its assumptions and the results for three different spin structures , corresponding to all experimentally studied iv - vi semiconductor magnetic / nonmagnetic sls ,
are presented in section iii . in section
ii the magnetic and band structures of the constituent materials are described and the sl geometry is specified .
the comparison with experimental data and the conclusions are presented in the last section .
all the components of the eux / pbx sls , where x = te or s , crystallize in the rock - salt structure .
bulk pbs and pbte are narrow gap nonmagnetic semiconductors with very similar band structures .
they both have a direct energy gap between the p - anion valence band and the p - cation conduction band at the point l of the bz @xcite .
it is well known that in both these lead chalcogenides the spin - orbit terms are important for the detailed description of the energy bands in the vicinity of the l point,@xcite .
one notices , however , that the spin - orbit corrections affect predominantly the conduction bands originating from the states of the heavy pb atoms . bulk eus is a classical heisenberg ferromagnet with the curie temperature 16.6 k @xcite .
bulk eute exhibits type ii afm structure with the nel temperature 9.6 k @xcite . in eute the spins of eu ions
are ferromagnetically ordered in ( 111)-type planes , which in turn are coupled antiferromagnetically to one another .
all the eu chalcogenides are semi - insulating , large gap semiconductors .
the results of the nonrelativistic apw calculations of the eus spin - polarized band structure @xcite show the narrow f(@xmath0 ) bands situated in the energy gap between the valence band , formed essentially of anion p states , and the conduction band , built mostly of cation d states .
the valence band maximum is situated at the center of the brillouin zone ( bz ) and the conduction band minimum at the point x. the spin splitting of the valence band results predominantly from the spin dependent mixing of p anion and f cation states , whereas that of the conduction band is mostly due to f - d and s - d on - site direct exchange .
much less is known about the eute band structure .
the optical experiments performed at t=300 k indicate @xcite , that in the paramagnetic phase of eute the f - d gap is somewhat larger than in eus , in agreement with the general trends visible in the experimental data @xcite and the results of apw calculations for other europium chalcogenides @xcite .
these trends seem not to be followed in the recent @xcite calculations of eute band structure , focused predominantly on the conduction bands .
two types of eus / pbs sls were experimentally studied , one grown on kcl substrate along [ 001 ] and the other on baf@xmath1 along the [ 111 ] crystallographic axis .
the measurements show that in both cases the ferromagnetically ordered eu spins within each magnetic layer lie in the planes perpendicular to the growth axis @xcite . in the ( 001 ) structures
each atomic monolayer consists of both anions and cations , with the monolayers @xmath2 apart , where @xmath3 is the cubic lattice constant .
the schematic view of two such monolayers is presented in fig .
the distances between the cation and its four in - plane nearest neighbors ( anions ) and four in - plane next nearest neighbors ( cations ) are also shown in the figure .
the spin structure of the magnetic ( 001 ) sl , which corresponds to the observed antiferromagnetic interlayer coupling is shown schematically in fig .
[ a - fm ] , a2 . in the case of ( 111 )
eus / pbs sls iec has not been yet observed .
these sls have the same crystallographic structure as the experimentally studied eute / pbte sls grown along the [ 111 ] axis . with this growth direction , the subsequent ( 111 ) atomic monolayers
are built either of anions or of cations , in alternation .
the distance between cation and anion monolayers is @xmath4 . a schematic view of three successive cation layers is presented in fig . [ 111 ] - the analogous anion sublattice is shifted by @xmath5 along the [ 111 ] direction and is not shown in the figure for the clarity reasons . in eute
/ pbte sls the neutronographic measurements show that the afm type ii structure is preserved in each eute layer , but the fm spin sheets form exclusively on the ( 111 ) planes parallel to the layers .
this is in contrast to bulk eute , where there are four symmetry - equivalent type ii afm arrangements in which the ferromagnetic spin sheets form on the \{111 } , \{@xmath611 } , \{@xmath71 } , or \{@xmath8 } plane families .
moreover , for the nonmagnetic spacers thin enough , the satellites observed in neutronographic spectra clearly indicate the existence of some long range order proving that the spins in consecutive magnetic layers are not randomly oriented , but tend to align along the same direction in a correlated way @xcite .
although for the antiferromagnetic layers the notions of afm and fm iec are not applicable , two types of co - linear correlated spin orientations in successive layers are still possible : identical ( in - phase ) and reversed ( out - of - phase ) , as shown in fig . [ a - fm]b
. both types of iec were observed in the experiment .
the solid gray circles represent anions , the white circles cations small circles are for ions in upper layer , the big ones for ions in the other .
the distances r@xmath9 from a cation to the four in - plane nn anions are shown by solid lines , by dotted lines the distances r@xmath10 to the in - plane nnn cations , width=321 ] above the black cation plane , are shown by open squares .
the distances to all twelve nnn are marked by dotted lines , width=321 ] all the ( eux)@xmath11/(pbx)@xmath12 sls ( where m and n denote the number of eux and pbx layers , respectively ) have relatively small lattice mismatch . this mismatch , as well as the strains resulting from it , will be ignored in the following , though the strains were shown to affect the magnetic properties of the eux layers , and their transition temperatures to the paramagnetic state,@xcite . to discuss the spin coupling between the magnetic layers
one has to consider a sl magnetic elementary cell containing at least two such layers . in sls
grown along [ 001 ] crystallographic direction the situation is simple , as the stacking sequence is abab - type .
such stacking does not enlarge the size of the magnetic elementary cell , whatever @xmath13 value .
in contrast , the stacking sequence abcabc - type ( compare fig . [ 111 ] ) for both anions and cations , in sls grown along [ 111 ] axis does enlarge the elementary cell when @xmath13 is not a multiple of 3 . thus , to limit the size of elementary cells , we consider only the ( 111 ) sls with @xmath14 . in the case of ( 001 ) sls the primitive lattice vectors , which define our elementary cells are : @xmath15 $ ] ; @xmath16 $ ] ; @xmath17 $ ] . for ( 111 )
sls , when the z - axis is chosen along the growth direction , we have @xmath18$];@xmath19 $ ] ; @xmath20 $ ] ( the plane views of these cells are sketched in fig . [ 001 ] and fig .
[ 111 ] ) . for both types of sl ,
our magnetic elementary cells contain , therefore , @xmath21 anions , @xmath22 magnetic cations and @xmath23 nonmagnetic cations . in order to determine iec in the above structures we compare the total valence - electron energies in two magnetic sls : with the in - phase and out - of - phase spin ordering . for the n and
m values typical for the experimentally studied sls the elementary cells contain several tens of atoms . in view of this complexity
, we decide to use the simplest calculation scheme still leading to fairly realistic band structure , namely , an empirical tight - binding method .
even though the one - electron methods are not designed for the total - energy calculations , the small spin - dependent changes in the total energy should be described adequately within this approach .
to construct the empirical tight - binding hamiltonian matrix one has to select the set of atomic orbitals for every type of involved ions and to specify the range of the ion - ion interactions .
this selection is always a compromise between the best description of the band structure and the minimization of the hamiltonian matrix dimensions and of the number of parameters used . in the following
we assume that the proper description of sl band structure is reached , when the hamiltonian reproduces in the @xmath24 and @xmath25 limits the known band structures of the bulk constituent magnetic and nonmagnetic materials , respectively .
this criterion determines in principle the selection of the ionic orbitals and gives the values of the parameters , all but those characterizing the interaction between magnetic and nonmagnetic cations . to calculate the band structure of the lead chalcogenides ( pbs and pbte ) we took into account s and p orbitals for both anions and cations , which lead to 8x8 hamiltonian matrix .
we allowed for the s
s , s p and p
p anion - cation nearest - neighbor ( nn ) interactions and the anion anion and cation cation p
p next - nearest neighbor ( nnn ) interactions .
it turned out that the band structure can be reproduced much better when we include also , by second
order perturbation , the interactions of p - orbitals with the three nn d - orbitals belonging to the @xmath26 representation .
the values of the parameters describing all these interactions and the values of the on - site orbital energies were determined by a @xmath27 minimization procedure , in which the band structure was fitted to the energies in the high symmetry points of the bz , taken from @xcite and @xcite .
the obtained energy bands for pbs and pbte along the symmetry axes of the bz , are presented in fig .
[ pbtes ] . in the other limit , for europium chalcogenides ( eus and eute ) , to describe the cations we take explicitly the one s- and five d - orbitals , whereas the anions are described as before by s- and p - orbitals .
the nn interaction involving the anion p - orbitals and cation s- and d - orbitals as well as the nnn cation - cation d d and anion - anion p
p interaction were included in the 10x10 hamiltonian .
the s - anion s - cation , s - anion d - cation interactions , turned out to be less important and were neglected . instead
, we included , again by second order perturbation , the hybridization of anion p(@xmath0)-orbitals with the cation f@xmath28-orbitals - this was necessary for reproducing the spin splittings of the valence bands in the ferromagnetic eus ( we neglected the hybridization with the energetically distant f@xmath29 band ) . to reproduce the spin splittings in the eus conduction bands the on - site exchange constants @xmath30 and @xmath31 had to be introduced .
the band structure of eus , presented in fig . [ eutes ] was obtained with the parameters fitted to the results of the apw spin
polarized calculations reported in @xcite .
the presented in fig .
[ eutes ] band structure of eute was obtained with the parameter values extrapolated from the values for eus and euse by exploiting the chemical trends in europium chalcogenides .
the elementary cell of the antiferromagnetic eute has a twice larger volume and completely different shape than the one of the ferromagnetic eus to facilitate the comparison with eus , we present the band structure of eute in the paramagnetic phase . in the above
, the number of independent fitting parameters was partially reduced according to the harrison relations @xcite , e.g. , instead of two nnn interatomic matrix elements @xmath32 and @xmath33 , we used @xmath34 .
it has to be also noted , that in all calculations we neglected the spin
orbit terms , known to be important in lead chalcogenides .
these terms would increase the number of model parameters and double the matrix dimensions and would , therefore , pose a problem in the case of the sls .
fortunately , we are mainly interested in valence bands , for which the spin
orbit is much less important than for the conduction bands . in the magnetic elementary cell of the ( eux)@xmath11/(pbx)@xmath12 sl
there are @xmath35 nonequivalent ions .
seven orbitals ( s , p , d ) have to be taken into account for each anion and each nonmagnetic cation and thirteen ( s , d , and f ) for each magnetic cation . in principle , the sl tight - binding hamiltonian is therefore a @xmath36x@xmath36 matrix , which has to be completed by the nonmagnetic cation
magnetic cation interactions .
the constants describing these interaction can not be inferred from the @xmath25 and @xmath24 limits , they were estimated from harrison s formula for interatomic matrix elements .
the matrix was then perturbationally reduced to the @xmath37x@xmath37 matrix . to determine the small difference between the large total energies of the valence electrons in the two , in - phase and out - of - phase , spin configurations
, we did not calculate these energies separately .
instead , the difference between the two energies of the valence electrons was calculated at a given k - point , after the numerical diagonalization of the two corresponding hamiltonian matrices and summation of the occupied states energies .
the results , obtained for a mesh of k - points ( 16000 points in the case of the tetragonal bz of the ( 100 ) sls , and 17280 in the case of the hexagonal bz of the ( 111 ) sls ) , were then used in the triple simpson procedure for integrating over the entire bz .
the calculations were performed for ( eus)@xmath38/(pbs)@xmath39 sls grown in both [ 001 ] and [ 111 ] directions and for ( eute)@xmath38/(pbte)@xmath39 sls grown along the [ 111 ] axis .
we denote by @xmath40 the absolute value of the energy difference between the in - phase and out - of - phase spin configuration per unit surface of the layer .
@xmath40 can be regarded as a measure of the strength of the interlayer spin coupling in the sls for ferromagnetic structures it can be expressed in terms of the constant @xmath41 @xcite , commonly used to characterize the iec in metallic magnetic / nonmagnetic trilayers by the relation : @xmath42 ( here the factor 4 , instead of 2 , accounts for the fact , that in sl each magnetic layer is coupled to two neighboring layers ) . the sign of the calculated energy difference determines the spin configuration in consecutive magnetic layers . in the ferromagnetic ,
eus - based , sls the out - of - phase spin configuration in consecutive magnetic layers is energetically preferred , so that iec in these structures has an antiferromagnetic character , in agreement with the experiment .
for the antiferromagnetic , eute - based sls the situation is more complicated : for odd number @xmath43 of spin planes in the magnetic layer the out - of - phase configuration has the lower energy , whereas for even @xmath43 it is the in - phase configuration , which is energetically favored .
thus , one can notice that in all studied sls the valence electron mediated iec prefers the spin configuration with the opposite directions of spins at the two interfaces bordering the spacer .
many various @xmath43 and @xmath44 values were selected to study the range of the interlayer coupling and the iec dependence on the thickness of the magnetic layer .
it turned out , that in all sls for fixed spacer thickness @xmath44 the strength of iec is almost independent on the magnetic layer thickness @xmath43 .
this seems to prove that in the considered here sls , composed of two semiconductors with very different energy gaps , the valence electron mediated iec is essentially a surface effect .
the dependence of the strength of the interlayer coupling on the spacer thickness @xmath44 for all three studied types of sls is presented in fig .
we recall that for the [ 111 ] growth direction , we calculated iec only for @xmath14 , so that for these sls in the figure the points for different @xmath44 values do not correspond to the same @xmath43 .
as can be seen in fig .
[ j1 ] , the strength of the coupling in all three cases decreases with the spacer thickness approximately exponentially .
the strongest and the least rapidly decreasing iec was obtained in the case of the fm eus / pbs ( 001 ) sls .
the comparison of the results obtained with the same set of model parameters for the two different types of eus / pbs sls , ( i.e. , the grown along [ 001 ] and [ 111 ] axis ) indicates that the valence electron mediated iec depends strongly on the lattice geometry .
for example , for the pbs spacer @xmath45 the obtained coupling between eus ( 111 ) magnetic layers is about five times weaker than that between eus ( 001 ) layers .
moreover , in the ( 111 ) sls the strength of the coupling decreases more quickly with @xmath44 .
the small regular deviations from the smooth dependence of @xmath41 vs. @xmath44 , which can be best seen for every second @xmath44 in the ( 001 ) case , but also for every third @xmath44 in the ( 111 ) results , reflect the periodic effects of stacking .
( left y - axis ) as a function of the spacer thickness for ferromagnetic eus / pbs ( 001 ) ( squares ) and ( 111 ) ( triangles ) sls . for the antiferromagnetic eute / pbte sls ( circles )
the absolute value of the energy difference @xmath46@xmath47 was divided by 4 ( at the right y - axis ) for comparison with the fm case , width=321 ] the iec calculated for the afm eute / pbte ( 111 ) sl turned out to be stronger than that for fm eus / pbs ( 111 ) sl ( see fig . [ j1 ] ) .
the difference in the band parameters by itself does not explain this result the calculations performed for afm and fm ( 111 ) sls with identical sets of band parameters have shown that the coupling between the afm layers is approximately two times stronger .
this indicates that the valence electron mediated iec is not a purely surface effect . in view of the present results for iec in antiferromagnetic eute / pbte sls
, we note that the simple 1d model in @xcite overestimated the strength of iec for large spacer thicknesses now the decrease of iec with @xmath44 is much more rapid ; the character of the interlayer coupling was however described properly .
the iec was observed in ferromagnetic eus / pbs ( 001 ) sls by magnetic @xcite and neutron diffraction and reflectivity @xcite methods . for the samples with the thin enough spacers , i.e. , 4.5@xmath48 ( 1.5 monolayers , probably a mixture of @xmath45 and @xmath49 ) and 10@xmath48 ( ca @xmath50 )
, the antiferromagnetic interlayer coupling was observed in the magnetic moment measurements ; the magnetic methods did not reveal any iec in the samples with larger spacer thicknesses . in the neutronographic experiments
the afm iec was confirmed in the above two samples , but it was also observed in the sample with 23@xmath48 pbs layers .
further measurements in the external magnetic fields , parallel to the layers , allowed to estimate the experimental strength of the coupling constant @xmath41 , from the magnetic field @xmath51 erasing the afm neutron reflectivity peak @xcite .
this was possible for the two samples with the thinner spacers , but not for the sample with the thicker spacer . for the latter ,
the field - induced changes in the neutronographic spectra were irreversible , what suggests that in this case the iec was weaker than the magnetic anisotropies .
the estimated experimental values of @xmath41 are : 0.063 , 0.031 and 0.019 ( in mj / m@xmath52 ) , for n=1 , 2 and 3 , respectively , @xcite .
the corresponding theoretical values obtained from our model are : 0.77 , 0.33 , 0.18 mj / m@xmath52 .
thus , one can conclude , that the model of valence electron mediated iec describes properly the sign and the rate of the decrease of the coupling with the spacer thickness , but overestimates the strength of the coupling .
the fact that the theoretical results obtained for crystallographicaly perfect sls lead to exchange constants order of magnitude larger than those observed for the real multilayer structures , is probably due to the interface diffusion , which in the case of metallic structures was shown to reduce significantly the strength of the iec @xcite .
unfortunately , for the afm type of sls there are no experiments , which provide direct information about the strength of the coupling .
the evidences of the existence of the coupling between the afm eute layers come from the satellite structure of the neutron diffraction spectra , seen in a variety of eute / pbte sls consisting of several hundreds of periods,@xcite . the detailed analysis of the shapes of the satellite lines in the neutronographic spectra indicates that in these sls the eute layers are not entirely coupled , but only partially correlated , the less the thicker are the pbte spacer layers of the sl @xcite . under the strong assumption that the structures are morphologically perfect , with the same @xmath43 and @xmath44 values throughout the entire sl , this degree of correlation can be quantitatively determined . under the same assumption
, the analysis of the satellites positions allows to distinguish which spin configuration , the in - phase or out - of - phase , is dominating .
the observed spectra for the sls with nominally even @xmath43 and even @xmath44 reveal the preference for the in - phase spin configurations , whereas those for the sls with odd @xmath43 and even @xmath44 exhibit the preference for the out - of - phase configuration , both in agreement with the predictions of the present model . for the case of even @xmath43 and odd @xmath44 there are no available data . finally , for the samples with @xmath43 and @xmath44 both odd , the neutron diffraction spectra seem to indicate that the in - phase configuration is preferred , contrary to the theoretical predictions . however , the in - phase spin configuration for sls with odd number of spin planes in each antiferromagnetic layer should exhibit ferrimagnetic properties , i.e. , lead to a significant net magnetic moment of ferrimagnetic domains .
no such magnetic moments were detected in these samples @xcite .
these somewhat confusing results seem to indicate that both chemical and magnetic structures of the studied sls are not perfect enough .
new technological and experimental efforts to observe the iec in eute / pbte sls with smaller number of sl periods , i.e. , in sls with better controlled @xmath43 and @xmath44 values , are undertaken . in conclusion , we have shown , within a 3d tight binding model , that the valence electron mediated interlayer exchange coupling explains the afm coupling between the fm layers observed in eus / pbs ( 001 ) sls with narrow pbs spacers .
the strength of the calculated coupling depends strongly on the lattice geometry and decreases approximately exponentially with the spacer thickness @xmath44 . for a given type of sl
it is almost independent on the number @xmath43 of the spin planes within each magnetic layer .
these features distinguish the considered mechanism from another mechanism of afm coupling between the fm layers , namely the dipolar coupling possible in multilayer structures with tiny magnetic domains @xcite .
the valence electron mediated interlayer coupling is , up to now , the only effective mechanism capable to explain the origin of the interlayer correlations observed in the antiferromagnetic eute / pbte sls with no localized impurity states .
the current , not complete understanding of the experimental data for afm sls , does not allow , however , to draw definite conclusions about the comparison between the details of the experimental and theoretical results .
p.grnberg , r.schreiber , y. pang , m.b .
brodsky , and h. sower , phys .
rev . lett.*57 * 2442 ( 1986 ) .
p. bruno , phys .
b * 52 * , 411 ( 1995 ) ; e - print : cond - mat/9905022 , and references cited therein . v. nunez , t.m .
giebultowicz , w. faschinger , g. bauer , h. sitter and j.k .
furdyna , j. magn . mag . mat.*140 - 144 * , 633 ( 1995 )
. j. j. rhyne , j. lin , j. k. furdyna and t. m. giebultowicz , j. mag . mag . mat .
* 177 - 181 * , 1195 ( 1998 ) .
giebultowicz , v. nunez , g. springholz , g. bauer , j. chen , m.s .
dresselhaus and j.k .
furdyna , j. magn . mag . mat.*140 - 144 * , 635 ( 1995 ) .
h. kpa , j. kutner - pielaszek , a. twardowski , a.yu .
sipatov , c.f .
majkrzak , t. story , r.r .
gaazka , t.m .
giebultowicz , proc .
icm , receife , brasil 2000 , j. mag . mag . mat .
, in print .
k. ha , c.h.w .
swste , h.j.m .
swagten , a.a .
smits , w.j.m .
de jonge , t. story , w.dobrowolski , m. arciszewska , r.r .
gaazka , a.yu .
sipatov , international colloquium on magnetic thin films and surfaces , natal , brasil 2000 , extended abstracts .
n. akiba f. matsukura , a. shen , y. ohno , h. ohno , a. oiwa , s. katsumoto and y. iye , appl.phys.lett.*73*,2122 ( 1998 ) ; d. chiba , n. akiba , y. ohno and h. ohno , appl.phys.lett.*77*,1873 ( 2000 ) .
p.shevchenko , l.wierkowski and j. oitmaa , j. mag . mag
* 177 - 181 * , 1168 ( 1998 ) .
rusin , phys.rev.b * 58 * , 2107,(1998-ii ) h. heinrich in lectures notes in physics , vol .
* 133 * ( springer - verlag , heidelberg , 1980 ) p. 407 .
j. blinowski and p. kacman , acta phys .
a * 92 * , 719 ( 1997 ) .
m. wilczyski and r. wirkowicz , phys.stat.sol.(b ) * 212 * , 165 ( 1999 ) .
dugaev , v.i .
litvinov , w. dobrowolski , and t. story , solid state commun . * 110 * , 351 ( 1999 ) .
su - huai wei and a.zunger , phys . rev.b
* 55 * , 13605 ( 1997 ) .
g. guentherodt , phys .
matter * 18*,37 ( 1974 ) .
s. mathi jaya and w. nolting , j. phys . : condens .
matter * 9 * , 10439 ( 1997 ) .
a. stachow - wojcik , t. story , w.dobrowolski , m. arciszewska , r.r .
gaazka , m.w .
kreijveld , c.h.w .
swste , h.j.m .
swagten , w.j.m .
de jonge , a. twardowski , a.yu .
sipatov , phys.rev.b*60 * , 15220 ( 1999 ) .
i.n . goncharenko and i. mirebeau , phys .
lett . * 80 * , 1082 ( 1998 ) .
harrison `` electronic structure and the properties of solids '' ( w.h.freeman and co. , san francisco , 1980 ) .
parkin , phys .
lett.*67 * 3598 ( 1991 ) .
h. kpa , j. kutner - pielaszek , j. blinowski , a. twardowski , c.f .
majkrzak , t. story , p. kacman , r.r .
gaazka , k. ha , h.j.m .
swagten , w.j.m .
de jonge , a.yu .
sipatov , t.m .
giebultowicz , submitted to prl .
h. kpa , j. blinowski , p. kacman , g. springholz , g. bauer , c.f .
majkrzak , k.i .
goldman and t.m .
giebultowicz , proc .
xxiv icps , jerusalem 1998 ( cd , world scientific , singapore 1999 ) s. holl , f. schinagl , t.m .
giebultowicz and h. krenn , j. mag . mag
* 220 * , 293 ( 2000 ) .
borchers , p.m. gehring , r.w .
erwin , j.f .
ankner , c.f .
majkrzak , t.l .
hylton , k.r .
coffey , m.a .
parker , j.k .
howard , phys .
b * 54 * , 9870 ( 1996 ) . | the interlayer exchange coupling mediated by valence band electrons in all - semiconductor iv - vi magnetic / nonmagnetic superlattices is studied theoretically .
a 3d tight - binding model , accounting for the band and magnetic structure of the constituent superlattice components is used to calculate the spin - dependent part of the total electronic energy .
the antiferromagnetic coupling between ferromagnetic layers in eus / pbs superlattices is obtained , in agreement with the experimental evidences .
the results obtained for the coupling between antiferromagnetic layers in eute / pbte superlattices are also presented . |
by gravitationally lensing distant background sources , galaxy clusters act as natural magnifying lenses in the sky , providing a unique window into the early universe .
several high - redshift galaxies have been discovered through the magnification power of galaxy clusters ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the current highest - redshift lensed galaxy is at @xmath20 in the field of abell 1689 @xcite , but among the highest - redshift _ multiply - lensed _ galaxies known are a galaxy at @xmath21 in the field of abell 383 ( comprising two images ; @xcite ) , a @xmath22 galaxy in the field of abell 2218 ( comprising three images ; @xcite ) , and a @xmath23 candidate in the `` bullet cluster '' ( comprising two images ; @xcite ) .
in addition , the lens model can be used to map the lensed images back into the source plane , while the high magnification enables their spatially resolved internal structural properties to be measured .
such measurements are not possible without the aid of the lensing power of the cluster ( e.g. * ? ? ?
* ) .
the lyman - alpha forest produces a sharp drop in flux below @xmath24 .
the expansion of the universe moves this spectral break to longer wavelengths , allowing high - redshift galaxies to be identified as `` dropouts '' in the observed - frame optical and near - ir ( see * ? ? ?
* ; * ? ? ?
* ) . as these young
, high - redshift galaxies are usually actively forming stars , their rest - frame uv spectra should be relatively blue .
this combination of properties allows the redshift of the galaxy to be estimated accurately , even without spectroscopy : as low as @xmath25 uncertainty on the redshift is obtained here , for 95% confidence levels .
the study of high - redshift galaxies enables important constraints to be placed on galaxy evolution and structure formation .
particularly , observing high - redshift galaxies provides direct measurements of the high-@xmath26 luminosity - function , early sfr , and the epoch of the intergalactic medium reionization ( see * ? ? ? * and references therein ) .
we report the discovery of one of the highest - redshift ( @xmath0 ) multiply - lensed galaxies known to date , lensed into four images by the galaxy cluster macs j0329.6 - 0211 ( @xmath26=0.45 ; macs0329 hereafter ) .
macs0329 is an x - ray selected system found by the massive cluster survey , macs ( @xcite ) .
@xcite classified macsj0329 as having evidence for substructure in its x - ray surface brightness , although @xcite classified it as relaxed .
we found no record of previous strong lensing ( sl ) analysis of this cluster , and only 1-band shallow _
hst_/wfpc2 previous imaging .
the 16 _ hst _ bands chosen for the cluster lensing and supernova survey with hubble ( clash ; @xcite ) enable us to conduct the first sl study of this cluster , and to obtain accurate photometric redshifts for the multiply - lensed background galaxies .
we use these data in conjunction with a parametric sl modeling method ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , to find several multiple image families across the central field of macs0329 ( fig .
[ curves0329 ] ) so that its mass distribution and inner profile can be well - constrained , allowing us to deduce the source - plane properties of the high - redshift galaxy presented here .
the paper is organized as follows : in 2 we describe the observations , and in 3 we detail the sl model . in
4 we report the results of the sl analysis and the physical properties of the high - redshift galaxy based on detailed sed modeling .
the results and conclusions are summarized in 5 .
we adopt a concordance @xmath27cdm cosmology with @xmath28 , @xmath29 , and @xmath30 . with these parameters
, one arcsecond corresponds to a physical scale of @xmath31 kpc at the cluster redshift ( @xmath32 ; @xcite ) , and @xmath33 kpc at @xmath34 .
macs0329 was observed as part of clash with _ hst _ from 2011 august to 2011 october .
this is one of 25 clusters to be observed to a total depth of 20 _ hst _ orbits in 16 filters with the wide field camera 3 ( wfc3 ) uvis and ir cameras , and the advanced camera for surveys ( acs ) wfc .
the images are reduced and mosaiced ( @xmath35 pixel@xmath18 ) using standard techniques implemented in the mosaicdrizzle pipeline @xcite .
hst photometry for the arcs is obtained within the apertures shown in fig .
[ dropout ] , and photometric redshifts are estimated for all the galaxies in the field using the full 16-band uvis / acs / wfc3-ir
aperture - matched ( e.g. , fig [ dropout ] ) photometry via both the bpz @xcite and lephare @xcite programs .
see @xcite for additional details .
we supplement these _ hst _ observations with infrared array camera ( irac ) imaging obtained in 2011 october as part of the _ spitzer _ cycle 8 warm mission ( pi : r. bouwens ; see also @xcite ) .
irac observations in channels 1 and 2 ( 3.6 and 4.5@xmath36 m ) have a total integration time of @xmath37s , and clearly detect three of the four images of the background source we report here ( 1.1 through 1.3 ) , with a formal @xmath38 detection for arc 1.4 .
because of crowding in the spitzer images , we adopt a circular @xmath39 diameter aperture for the irac photometry .
background levels are estimated by sampling non - crowded sections of the irac mosaic while avoiding nearby bright sources .
the accuracy of this ( manual ) photometry is verified by subtracting possible contaminating sources with galfit , especially for arc 1.2 which has a bright star nearby .
we use a flux aperture correction factor of @xmath40 , based on a curve of growth analysis of unresolved sources in the post basic calibrated data ( pbcd ) mosaic .
the relevant multi - wavelength photometry is given in table [ systems ] .
the lens modeling method used here ( e.g. , * ? ? ? * ; * ? ? ? * ) begins with the assumption that mass approximately traces light .
we model the distribution of cluster mass by assigning a power - law mass profile to each red - sequence cluster galaxy , scaled by its ( relative ) brightness .
the sum of all galaxy contributions represents the lumpy galaxy component , which is then smoothed using 2d spline interpolation to obtain a smooth - component representing the dark matter ( dm ) distribution .
the polynomial degree of smoothing and the index of the power - law are the most important free parameters determining the mass profile .
a worthwhile improvement in fitting the location of the lensed images is generally found by introducing an external shear describing the overall matter ellipticity .
the direction of the shear and its amplitude are free parameters , allowing for some flexibility in the relation between the distribution of dm and the distribution of galaxies .
the weight of the lumpy component relative to the dm , and the overall normalization , bring the total number of free parameters in our modeling to six ( see * ? ? ?
the best fit is assessed by a @xmath41 minimization in the image plane , which is preferred because it generally does not produce solutions that are biased towards shallow mass profiles ( and high magnifications ) , as is the case with source - plane minimization .
the positions and morphologies of the multiply - imaged arcs are accurately reproduced by the best - fit mass model in their measured photometric redshifts ( for a morphological comparison example see fig .
[ rep1to4 ] ) , and the model is successively refined as additional sets of multiple images are incorporated .
i d & ( j2000.0)&(j2000.0 ) & & & + 1.1 & 03:29:40.18 & -02:11:45.60 & & & + 1.2 & 03:29:40.06 & -02:11:51.72 & & & + 1.3 & 03:29:41.28 & -02:12:05.04 & & & + 1.4 & 03:29:43.16 & -02:11:17.30 & & & + + arc & f775w & f814w & f850lp & f105w & f110w & f125w & f140w & f160w & @xmath42 & @xmath43 + 1.1 & @xmath44 & @xmath45 & @xmath46 & @xmath47 & @xmath48 & @xmath49 & @xmath50 & @xmath51 & @xmath52 & @xmath53 + 1.2 & @xmath54 ( 2@xmath55 ) & @xmath56 & @xmath57 & @xmath58 & @xmath59 & @xmath60 & @xmath61 & @xmath62 & @xmath63 & @xmath64 + 1.3 & @xmath65 & @xmath66 & @xmath67 & @xmath68 & @xmath69 & @xmath70 & @xmath71 & @xmath72 & @xmath730.6 & @xmath740.3 + 1.4 & @xmath75 ( 2@xmath55 ) & @xmath76 ( 2@xmath55 ) & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 ( 2@xmath55 ) +
we uncovered in _ hst_/clash and _ spitzer_/irac imaging of macs0329 one of the highest - redshift multiply - lensed galaxies known to date .
we use the independent photometric - redshift distributions of its four lensed images to obtain a source redshift ( and @xmath85 limits , see table 1 ) of @xmath0 . using our mass model and the extensive imaging and resulting photometric redshifts ,
we physically matched 15 additional new multiple images and candidates of five background , lower-@xmath26 sources ( see figure [ curves0329 ] ) , which are used in turn to refine the mass model .
explicitly , in addition to the four images of the high-@xmath26 galaxy , we use as additional constraints the seven multiple - images of systems 2 and 3 , at photometric redshifts ( and @xmath85 limits ) of @xmath86 and @xmath87 , respectively , and verify that all other systems and candidates are plausible in the context of the resulting mass model . the magnification by the cluster is coupled to the surface - density gradient in each point ( i.e. , the mass profile ) , so multiple - systems with redshift information are important to assess the true magnification across the cluster field . for the four images of the @xmath88 galaxy , we obtain magnifications ( and @xmath89 uncertainties ) of @xmath6 , @xmath7 , @xmath8 , and @xmath9 , respectively , so that in total the galaxy is magnified by a factor of @xmath90 . galaxy .
the upper panel shows the observed and rest - frame sed and the maximum likelihood model fit .
the filled red points indicate bands in which the object is well detected , while the open blue triangles indicate @xmath85 upper limits ; the open squares show the photometry of the best - fitting model convolved with the wfc3 , acs , and irac filter response functions .
the open red circle indicates a spurious , marginally significant detection of arc 1.2 in the f606w band ; however , we emphasize that none of the other images are detected in this band .
we also show an inset with the photometric redshift distribution , @xmath91 , obtained by combining the independent photometric redshift measurements of the four arcs , yielding a source redshift of @xmath92 with bpz , and @xmath93 with lpz ( @xmath1 confidence levels ) .
the lower panels show the posterior distributions on the demagnified stellar mass , sfr weighted age , stellar metallicity ( relative to @xmath94 ) , rest @xmath95-band dust attenuation , demagnified sfr , and specific sfr , ssfr@xmath96sfr/@xmath97 .
the vertical dashed lines show the median of each posterior distribution .
we summarize the results in
[ results].,width=309 ] we leverage the magnification boost of the most magnified image , arc 1.2 , to construct a high - resolution image of the background galaxy , which occupies a source - plane area of @xmath98 kpc@xmath13 ( see figure [ curves0329 ] ) .
to further constrain the physical properties of this galaxy , we model the observed sed of arc 1.2 using the bayesian sed - fitting code isedfit ( see @xcite ) coupled to the flexible stellar population synthesis models of @xcite .
we adopt the @xcite initial mass function from @xmath99 , assume the @xcite dust attenuation law , and adopt uniform priors ( based on monte carlo draws ) on the stellar metallicity ( @xmath100 ) , @xmath95-band attenuation ( @xmath101 mag ) , and galaxy age ( @xmath102 gyr ) . for reference , the age of the universe at @xmath34 is @xmath103 gyr .
we parameterize the star formation history @xmath104 as an exponentially declining function of time , @xmath105 , given by @xmath106 , where @xmath107 is the characteristic time for star formation .
we draw @xmath107 from a uniform distribution between @xmath108 gyr , which spans the range from passively evolving to continuous star formation . in figure [ sf ]
we show the observed and rest - frame sed of arc 1.2 , the maximum likelihood model fit , and the posterior distributions on all the model parameters . adopting the median of each posterior distribution as our best estimate of the properties of the galaxy
, we find a demagnified stellar mass of @xmath14 , low dust content ( @xmath16 mag ) , a demagnified sfr of @xmath17 yr@xmath18 , and a sfr - weighted age of @xmath109 myr .
these results imply a specific sfr , ssfr@xmath96sfr/@xmath97 , of @xmath19 gyr@xmath18 , corresponding to a mass - doubling time of just @xmath110 myr ( assuming a 50% return fraction ) .
the stellar metallicity and @xmath107 parameter ( not shown ) are not particularly well constrained , although solutions with subsolar metallicity and @xmath111 gyr are generally favored ; the median of the posterior distributions imply subsolar metallicity , @xmath15 , and @xmath112 gyr .
we verified that performing our sed modeling on the other ( less magnified ) images yields overall similar results .
the demagnified sed of arc 1.1 is nearly identical to that of arc 1.2 , and therefore the posterior distributions of the physical quantities we derive are very similar to those shown in figure [ sf ] ; the median quantities all agree to well within 0.1 dex ( @xmath113 ) .
the demagnified f125w magnitude of arc 1.3(1.4 ) , on the other hand , is 0.8 mag brighter(fainter ) than arcs 1.1 and 1.2 , although this discrepancy is well within the statistical uncertainties on the magnifications .
nevertheless , at face value the fainter two arcs imply an 0.3 - 0.4 dex larger stellar mass and sfr , 0.2 - 0.6 mag more dust attenuation , and a very similar age , 200 myr .
however , we emphasize that the posterior distributions on these quantities overlap significantly with those of arcs 1.1 and 1.2 , and therefore these differences do not affect our conclusions about the nature of this object .
note also that the center of emission in the 3.6@xmath36 m image is slightly different than that of the 4.5@xmath36 m image ( figure [ dropout ] ) . repeating the photometry on the two different centers
, we find this offset may introduce uncertainty of @xmath1140.3 mag in the irac photometry .
the effect of possible contamination from bright neighbors was re - examined for both emission - centers , and was found to be typically 0.4 mag .
however , we importantly verified that these higher uncertainties have only a negligible ( @xmath115 dex ) effect on the resulting physical properties , as the fit is governed by the hst photometry .
source , relative to lower-@xmath26 systems such as systems 2 and 3 , independently strengthening its detection as a high - redshift source.,width=302 ] finally , we use our lens model to constrain the physical characteristics of macs0329 . for the @xmath116 source ,
the critical curves enclose a relatively large area , with an effective einstein radius of @xmath117 ( @xmath118195 kpc at @xmath2 ) , and a projected mass of @xmath119 ( figure [ curves0329 ] ) . for the lower redshift of system 2 , @xmath120 ,
the einstein radius is @xmath121 , and the critical curve encloses a projected mass of @xmath122 . for completeness
, we measure the ( total ) mass profile slope , @xmath123 ( in the range @xmath124 , or @xmath125 kpc ; about twice the einstein radius for @xmath126 ) , typical of _ relaxed _ and well - concentrated lensing clusters ( e.g. * ? ? ?
* ; * ? ? ?
* ) , and in agreement with the fairly circularly - symmetric x - ray contours centered on the bcg ( see * ? ? ?
the discovery of a high - redshift galaxy in the field of macs0329 adds to several known high - reshift galaxies lensed by galaxy clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
here we summarize the properties of this unique source .
\(1 ) it is one of the highest - redshift multiply lensed objects known to date , and lensed into four separate images .
the angular separation between arcs 1.2 and 1.4 is @xmath127 , considerably larger than previously reported cases @xcite .
\(2 ) the source is one of the brightest at @xmath128 : its @xmath129 magnitude is 24.0 ab , making it a viable candidate for follow - up spectroscopy .
\(3 ) the galaxy is consistent with being a dwarf galaxy .
its intrinsic ( delensed ) magnitude of @xmath130 ab makes it a sub-@xmath131 galaxy at this redshift .
it occupies a source - plane area of @xmath1142.2 kpc@xmath13 , similar to previously deduced sizes of high-@xmath26 lensed galaxies ( e.g. * ? ? ?
* ) . due to the hierarchical growth of structure
, galaxies are expected to be small at high redshifts , with dwarf galaxies constituting the building material of larger structures .
our source - plane reconstruction shows at least three ( possibly star - forming ) knots , consistent with several other reports of high - redshift galaxies with multiple components @xcite , possibly as the result of merging .
in addition , we measure an overall half - light radius of @xmath132 , consistent with that found in @xcite and @xcite .
\(4 ) the sed fits to the multiband photometry of the source suggest a demagnified stellar mass of @xmath14 , a sfr - weighted age of @xmath109 myr , subsolar metallicity ( @xmath15 ) , low dust content ( @xmath16 mag ) , and a demagnified sfr of @xmath17 yr@xmath18 .
the specific sfr of @xmath19 gyr@xmath18 , which is slightly higher than that found by other recent studies @xcite , implies a mass - doubling time of just @xmath110 myr and therefore vigorous ongoing star formation considering its low mass .
\(5 ) the uv continuum is blue , with a uv - slope @xmath133 , consistent with measurements of other faint @xmath134 galaxies and suggests that these sources are largely dust free @xcite .
the discovery of the galaxy presented here shows once more the novelty and tremendous potential of galaxy clusters for observationally accessing the faint early universe .
the authors thank saurabh jha for useful discussions , and the anonymous referee for valuable comments .
the clash multi - cycle treasury program ( go-12065 ) is based on observations made with the nasa / esa hubble space telescope .
the space telescope science institute is operated by the association of universities for research in astronomy , inc . under nasa contract
nas 5 - 26555 .
this work is based in part on observations made with the _ spitzer _ space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
support for this work was provided by nasa through an award issued by jpl / caltech .
we thank the hospitality of institut fr theoretische astrophysik , universitt heidelberg , where part of this work took place , and the baden - werttemberg stiftung .
, a. m. , fruchter , a. s. , hook , r. n. , & hack , w. 2002 , in the 2002 hst calibration workshop : hubble after the installation of the acs and the nicmos cooling system , ed .
s. arribas , a. koekemoer , & b. whitmore , 337 | we report the discovery of a @xmath0 ( @xmath1 confidence level ) dwarf galaxy , lensed into four images by the galaxy cluster macs j0329.6 - 0211 ( @xmath2 ) .
the galaxy is observed as a high - redshift dropout in _
hst_/acs / wfc3 clash and _
spitzer_/irac imaging .
its redshift is securely determined due to a clear detection of the lyman - break in the 18-band photometry , making this galaxy one of the highest - redshift multiply - lensed objects known to date with an observed magnitude of f125w@xmath3 ab @xmath4 for its most magnified image .
we also present the first strong - lensing analysis of this cluster uncovering 15 additional multiply - imaged candidates of five lower - redshift sources spanning the range @xmath5 .
the mass model independently supports the high photometric redshift and reveals magnifications of @xmath6 , @xmath7 , @xmath8 , and @xmath9 , respectively , for the four images of the high - redshift galaxy . by delensing the most magnified image
we construct an image of the source with a physical resolution of @xmath10 pc when the universe was @xmath11 gyr old , where the @xmath12 galaxy occupies a source - plane area of approximately 2.2 kpc@xmath13 .
modeling the observed spectral energy distribution ( sed ) using population synthesis models , we find a demagnified stellar mass of @xmath14 , subsolar metallicity ( @xmath15 ) , low dust content ( @xmath16 mag ) , a demagnified sfr of @xmath17 yr@xmath18 , and a specific sfr of @xmath19 gyr@xmath18 , all consistent with the properties of local dwarf galaxies . |
let @xmath0 be a smooth stationary solution of the einstein
maxwell system @xmath1 with @xmath2 a 4-manifold , @xmath3 a lorentzian metric with signature @xmath4 , and @xmath5 a 2-form on @xmath2 ; here @xmath6 is the electromagnetic energy - momentum tensor , and @xmath7 is the cosmological constant , which we will take to be _ positive _ and fixed . if @xmath0 is close to a non - degenerate reissner
nordstrm de sitter ( rnds ) black hole in a neighborhood of the domain of outer communications ( more precisely , close to the future development of rnds initial data on a spacelike hypersurface intersecting and extending a bit past the future event and cosmological horizons ) , we show that @xmath0 is isometric to a slowly rotating kerr
de sitter ( knds ) black hole .
we deduce this from the full non - linear stability result proved by the author in @xcite . to state our theorem
, we introduce @xmath8 with @xmath9 specified below , and @xmath10 denote the parameters of a knds black hole by @xmath11 , where @xmath12 is its mass , @xmath13 its angular momentum per unit mass , and @xmath14 are its electric and magnetic charge .
let @xmath15 denote the parameters of a fixed non - degenerate rnds black hole , see ( * ? ? ?
* definition 3.1 ) , which thus has an event horizon at @xmath16 and a cosmological horizon at @xmath17 .
we choose @xmath18 and @xmath19 such that @xmath20 for a fixed small @xmath21 , and such that moreover the cauchy horizon ( which exists if @xmath22 ) of the rnds black hole lies in @xmath23 .
we only consider parameters @xmath24 which are close to @xmath25 .
the knds family , recalled below , is then a smooth family @xmath26 of solutions of on the fixed manifold @xmath2 ; the vector field @xmath27 is killing in the sense that @xmath28 and @xmath29 for all @xmath24 .
we assume that the event , resp .
cosmological , horizon of @xmath30 is located at @xmath31 , resp .
@xmath32 , with @xmath33 , and that the cauchy horizon of @xmath30 is located in @xmath23 ; this holds for @xmath24 sufficiently close to @xmath25 .
lastly , if @xmath34 is any smooth map , so @xmath35 with @xmath36 real - valued and @xmath37 taking values in @xmath38 , we define its _ stationary extension _ to be the map @xmath39 , @xmath40 [ thmintrouniq ] suppose @xmath41 is a smooth solution of the einstein
maxwell system on @xmath2 which is stationary with respect to @xmath27 , i.e. @xmath42 .
suppose moreover that @xmath41 is close to @xmath43 in the topology of @xmath44 for some large fixed @xmath45 .
then there exist parameters @xmath24 close to @xmath25 and a smooth hypersurface @xmath46 , together with a smooth map @xmath47 @xmath48-close in @xmath49 to the map @xmath50 such that for the stationary extension @xmath51 of @xmath52 , we have @xmath53 see figure [ figintrouniq ] . here , we define @xmath44 using the embedding @xmath54 by means of polar coordinates .
if one merely assumes @xmath41 to have fixed but high regularity , the conclusion remains valid , only with @xmath52 having finite ( but high ) regularity only .
the conclusion can not be strengthened : if @xmath34 , @xmath55 small , is any smooth map , close to the time @xmath56 translation map , with stationary extension @xmath51 , then @xmath57 satisfies the assumptions of the theorem ( after slightly shrinking the size of the interval @xmath58 $ ] and thus of @xmath2 ) , and the conclusion holds for @xmath52 and @xmath59 . , drawn in a penrose diagrammatic fashion together with a level set of @xmath60 .
_ right : _ penrose diagram of a knds solution @xmath26 , with event horizon @xmath61 , cosmological horizon @xmath62 , and future infinity @xmath63 .
_ red : _ the stationary extension @xmath51 of the map @xmath64 pulls the knds solution back to the given one.,title="fig : " ] in boyer
lindquist coordinates , a knds solution @xmath26 with magnetic charge @xmath65 and @xmath66 takes the form @xmath67 where we use polar coordinates @xmath68 and @xmath69 on @xmath70 , and @xmath71 the radii @xmath72 are the two largest roots of @xmath73 .
the 4-potential is @xmath74 and the electromagnetic tensor is correspondingly given by @xmath75 .
these expressions for @xmath30 and @xmath76 are singular at the horizons @xmath77 , but a simple change of coordinates removes these singularities , giving an extension of @xmath26 to @xmath2 ; see @xcite for details .
knds solutions with non - zero magnetic charge can be constructed from magnetically uncharged solutions using the electric - magnetic duality , as described in @xcite .
theorem [ thmintrouniq ] is a rigidity theorem for stationary black holes in the einstein
maxwell theory in the _ smooth _ category and _ without symmetry assumptions_. in the case of @xmath78 , rigidity of the kerr
newman family was proved by wong in his thesis @xcite conditional on a technical assumption , see ( * ? ? ?
* item ( t ) in 4.4 ) using a tensorial characterization of kerr
newman black holes @xcite .
these results built on the work on rigidity theorems for kerr spacetimes , in the context of the einstein vacuum equations and under smallness assumptions on the angular momentum or the mars
simons tensor , by alexakis
klainerman @xcite , which in turn relied on earlier work by ionescu
klainerman @xcite .
see also @xcite .
these works make certain geometric assumptions in particular on the structure of the future and past event horizons .
while we do not explicitly make such assumptions in the statement of theorem [ thmintrouniq ] , they are in fact automatically satisfied , as a careful analysis of the solution produced by the main theorem of @xcite shows ; this will be discussed elsewhere .
earlier uniqueness results for the kerr spacetime by hawking and ellis @xcite , carter @xcite , and robinson @xcite , as well as later work by mazur @xcite on the rigidity of kerr
newman black holes , relied on the restrictive assumption of real analyticity of the metric and the electromagnetic field .
we refer the reader to the reviews @xcite for further results and references .
proving uniqueness for the _ linearized _ einstein(maxwell ) equations amounts to performing a mode stability analysis at @xmath79 frequency , which was done by vishveshwara @xcite , moncrief @xcite , and kodama @xcite in various contexts . via the non - linear stability theorem ,
a key ingredient of our proof of theorem [ thmintrouniq ] is the analysis of ( generalized ) @xmath79 frequency modes for linear perturbations of non - degenerate rnds black holes in the spirit of the work by kodama
ishibashi @xcite on perturbations of schwarzschild de sitter and rnds spacetimes ; we do however need the full strength of @xcite in the argument presented here .
using the kerr de sitter ( kds ) stability result with vasy @xcite , our arguments also prove the uniqueness of slowly rotating kds black holes . furthermore ,
as pointed out in ( * ? ? ?
* remark 1.2 ) , non - linear stability for perturbations of slowly rotating kds or knds black holes also holds for the einstein(maxwell ) equations coupled to a massless or massive scalar field .
correspondingly , perturbative uniqueness results , analogous to theorem [ thmintrouniq ] , hold for such black holes .
( when coupling to a massive scalar field , the restriction to small angular momenta in theorem [ thmintrouniq ] is likely necessary with the smallness depending on the mass of the scalar field , as the construction of time - periodic perturbations of kerr black holes by chodosh and shlapentokh - rothman @xcite suggests . )
we close this introduction by recalling the non - linear stability result for knds black holes @xcite : we take @xmath80 as the cauchy surface for the formulation of the initial value problem of the einstein maxwell system .
we denote initial data on @xmath81 by @xmath82 , where @xmath83 is a riemannian metric on @xmath81 , @xmath84 a symmetric 2-tensor , and @xmath85 and @xmath86 are 1-forms , assumed to be smooth . a metric @xmath3 and a 2-form @xmath5 on @xmath2 , satisfying the einstein
maxwell system , are said to induce these data provided @xmath87 , with @xmath88 the inclusion , and @xmath84 are the metric and second fundamental form of @xmath81 induced by @xmath3 , respectively , and if for the future unit normal @xmath89 to @xmath81 , we have @xmath90 and @xmath91 .
the _ constraint equations _ are necessary and sufficient conditions for a local solution of this initial value problem to exist ; see ( * ? ? ?
* equations ( 2.14)(2.15 ) ) for their form in the present context .
[ thmintrostab ] ( see ( * ? ? ?
* theorem 9.2 and remark 9.3 ) . )
fix the parameters @xmath25 of a non - degenerate rnds spacetime .
then there exists @xmath92 such that the following holds : suppose @xmath82 are smooth initial data on @xmath81 , satisfying the constraint equations , which are close , in the topology of @xmath93 , to the initial data induced by the non - degenerate rnds solution @xmath43 . then there exist a smooth global forward solution @xmath41 of the einstein
maxwell system , attaining the given initial data on @xmath81 , and knds parameters @xmath11 close to @xmath25 , such that @xmath94 where @xmath95 and @xmath96 are exponentially decaying .
moreover , @xmath24 , @xmath97 and @xmath98 depend continuously on @xmath82 . here , @xmath99 denotes bounded functions ( or sections of a vector bundle ) which are bounded together will all their coordinate derivatives , using the embedding @xmath54 .
the continuous dependence is not explicitly stated in the reference , but is automatic for the nash
moser iteration scheme employed there , see also ( * ? ? ?
* theorem 11.2 ) .
since we do not track the number of derivatives on the data which need to be controlled in order to control a fixed @xmath44 norm of @xmath100 , we work with @xmath44 spaces of high but unspecified regularity @xmath101 in theorem [ thmintrouniq ] and in our arguments below ; in principle however one could specify a numerical value for @xmath45 by a careful study of the proof of the nash
moser iteration scheme . in order to prove theorem [ thmintrouniq ]
, we use this non - linear stability result as follows : if in theorem [ thmintrostab ] we take as data the initial data of a stationary solution @xmath41 which are close to non - degenerate rnds data , then the solution @xmath102 given by theorem [ thmintrostab ] asymptotes to a stationary slowly rotating knds solution @xmath26 ; however , @xmath102 itself is stationary since it is the same as @xmath41 , up to diffeomorphism equivalence and hence must in fact be equal to @xmath26 , again up to diffeomorphism equivalence .
we implement this idea in [ secext ] and [ secpf ] .
i am grateful to andrs vasy and maciej zworski for useful discussions , and to the miller institute at the university of california , berkeley , for support .
given a solution @xmath41 of equation on a globally hyperbolic spacetime @xmath103 with cauchy surface @xmath104 , we show : [ propext ] let @xmath105 be an open neighborhood of @xmath81 , and suppose that @xmath106 is a smooth vector field such that @xmath107 and @xmath108 in @xmath105
. then there exists a unique global extension @xmath109 of @xmath110 such that @xmath111 and @xmath112 .
let @xmath113 be any vector field , and define the 2-tensor , resp .
2-form @xmath114 denote by @xmath115 the time @xmath116 flow of @xmath113 . due to the diffeomorphism invariance of the einstein
maxwell system , we have @xmath117 we then recall from ( * ? ? ?
* equation ( 2.4 ) ) the formula @xmath118 , where @xmath119 is the trace reversal , @xmath120 the symmetric gradient , and @xmath121 is a zeroth order operator . aiming to derive a wave equation for @xmath122
, we note that @xmath123 thus , if @xmath113 satisfies the equation @xmath124 then @xmath125 , hence @xmath122 satisfies the wave equation @xmath126 ( conversely , if @xmath113 is a killing extension of @xmath110 , then @xmath127 , hence @xmath113 satisfies equation , which proves the uniqueness of the extension @xmath113 . )
next , we observe that @xmath128 , hence @xmath129 , and @xmath130 hence , we have the wave equation @xmath131 for any choice of @xmath113 .
note that while up to second derivatives of @xmath122 appear in this equation , @xmath132 in appears only undifferentiated .
thus , energy estimates do apply to the coupled system of wave equations : one estimates @xmath122 in @xmath133 and @xmath132 in @xmath134 .
( see the discussion around ( * ? ? ?
* equation ( 8.42 ) ) . ) to prove the existence of @xmath113 , we solve the wave equation with @xmath135 near @xmath81 ; since we have @xmath127 and @xmath136 near @xmath81 , and since @xmath122 and @xmath132 solve the coupled system of linear wave equations , we conclude that @xmath137 and @xmath138 on @xmath2 .
let @xmath139 denote the initial data induced on @xmath81 by the given smooth stationary solution @xmath41 . by assumption , @xmath140 is close ( in a high regularity @xmath44 norm ) to the data @xmath141 induced on @xmath81 by the rnds solution @xmath43 .
let @xmath102 denote the forward solution of the einstein maxwell system , with initial data @xmath140 on @xmath81 , given by theorem [ thmintrostab ] , thus there exist black hole parameters @xmath24 with @xmath142 by uniqueness of solutions to the initial value problem , @xmath102 and @xmath41 represent the same solution .
more precisely : [ lemmapfsame ] fix @xmath101 , then there exists @xmath143 such that the following holds : if @xmath140 is sufficiently close to @xmath141 in @xmath144 , then there exist a neighborhood @xmath105 of @xmath145 in @xmath2 and a diffeomorphism @xmath146 with @xmath147 and such that @xmath148 and @xmath149 .
embedding @xmath54 by means of @xmath150 , the smooth function @xmath151 is small in @xmath152 .
denote by @xmath89 and @xmath159 the future unit normal vector fields on @xmath81 with respect to @xmath3 and @xmath160 , respectively . for small @xmath161 ,
the normal exponential map @xmath162 ( defined with respect to the metric @xmath3 ) is a diffeomorphism onto its image ; pulling @xmath3 back by this map , we have @xmath163 and at @xmath81 @xmath164 we can perform the same construction for the metric @xmath160 and the map @xmath165 .
define the diffeomorphism @xmath166 by @xmath167 , @xmath168 ; this is well - defined in a neighborhood of @xmath169 .
since @xmath3 and @xmath160 induce the same initial data pointwise on @xmath81 , @xmath3 and @xmath170 , defined near @xmath169 , have the same 1-jet on @xmath169 . by the same token ,
@xmath5 and @xmath171 agree on @xmath169 ; in fact , since they satisfy maxwell s equations with respect to the metrics @xmath3 and @xmath172 , respectively , their 1-jets agree .
define now the map @xmath173 to be the solution of the ( semilinear ) wave map equation @xmath174 with trivial initial data , i.e. @xmath175 and @xmath176 .
( see ( * ? ? ?
* remark 2.1 ) for the notation . ) on a neighborhood of @xmath177 , @xmath173 is a diffeomorphism onto its image , which we denote by @xmath105 .
let @xmath178 and @xmath179 , then the identity map @xmath180 induces a wave map @xmath181 ; of course , @xmath182 also induces a wave map @xmath183 . by the local ( pointwise ! ) uniqueness for the einstein maxwell system in this wave map gauge ,
i.e. with @xmath3 as a background metric ( see @xcite ) , we conclude that @xmath184 near @xmath145 , as desired .
since the killing vector field @xmath113 on @xmath102 is in general _ not _ equal to @xmath27 , the global existence of @xmath113 as a killing field is not automatic at this point ; but using proposition [ propext ] for @xmath102 , i.e. solving the wave equation @xmath187 we can extend @xmath113 from a neighborhood of @xmath145 to a killing vector field satisfying @xmath188 . here
, we assume that the future domain of dependence of @xmath145 with respect to @xmath160 contains a neighborhood @xmath189 of the black hole exterior region , which holds for sufficiently small perturbations of the rnds data @xmath141 ; the vector field @xmath113 is then defined in @xmath189 .
[ lemmapfkillingasymp ] the vector field @xmath113 takes the form @xmath190 where @xmath191 is a constant , @xmath192 is a rotational killing field , and @xmath193 .
moreover , @xmath194 is small , @xmath195 is a small rotation , and @xmath196 has small norm in @xmath197 .
assume first that @xmath160 is stationary , or more generally a smooth b - metric on the compactification of @xmath189 defined by letting @xmath198 , and adding @xmath199 as a boundary at future infinity ; equivalently , assume that the coefficients of @xmath160 ( and thus of @xmath97 in ) , as a metric on @xmath200 , extend to smooth functions of @xmath201 down to @xmath199 .
since equation agrees with the tensor wave operator on 1-forms up to sub - subprincipal terms , the main result of @xcite applies , showing that @xmath113 has an asymptotic expansion @xmath202 where @xmath203 , the @xmath204 are pairwise distinct resonances , @xmath205 , @xmath206 , @xmath207 , and @xmath193 ; here @xmath92 is a fixed small number .
since @xmath160 is stationary up to a remainder which is exponentially decaying in @xmath60 , the killing equation @xmath208 implies that for each @xmath209 , the generalized mode @xmath210 satisfies the killing equation @xmath211 .
however , the only killing vector fields of @xmath30 are @xmath27 and @xmath212 if the black hole with parameters @xmath24 has non - zero angular momentum , and all rotational vector fields on @xmath70 for non - rotating black holes .
this forces @xmath213 and @xmath214 , hence we conclude that holds for some @xmath215 .
observe now that if the initial data were the ones induced by the rnds solution @xmath43 , then we would have @xmath216 , the local diffeomorphism @xmath217 constructed in lemma [ lemmapfsame ] would be the identity , and the killing vector field @xmath113 would be @xmath218 , so @xmath219 , @xmath220 , and @xmath221 in .
the smallness assertions then follow from the continuous dependence of the resonance expansion on the coefficients of the operator in equation and on the initial data , see @xcite for details . in general
, we only have the regularity stated in for @xmath160 , i.e. conormality instead of smoothness on the compactification of @xmath189 . in this case , one first notes that energy estimates immediately give @xmath222 for some @xmath223 , and in fact @xmath224 by propagation of b - singularities , see ( * ? ? ?
* lemma 5.2 ) .
one then has a partial expansion of the form , with @xmath225 and a remainder @xmath226 , see the proof of ( * ? ? ?
* theorem 5.6 ) .
as long as @xmath227 , the above arguments show that the partial expansion must vanish identically , implying an improved decay rate for @xmath228 .
iterating this argument , one obtains @xmath224 for all @xmath229 , and in the next iterative step , one obtains the partial expansion , as desired .
rescaling @xmath113 and redefining @xmath195 and @xmath196 , we thus have @xmath230 denote by @xmath115 the time @xmath116 flow of @xmath113 , and let @xmath231 be the time @xmath116 flow of @xmath232 , so @xmath233 . 1 . [ itpflongtimeexists ] the flow @xmath115 maps @xmath236 for all @xmath237 .
[ itpflongtimeconv ] let @xmath238 , and write @xmath239 for @xmath240 , @xmath237 .
then the limit @xmath241 exists and defines a smooth function satisfying @xmath242 .
moreover , the convergence is exponentially fast , that is , @xmath243 .
the estimates @xmath244 , @xmath245 , with @xmath246 small for @xmath140 close to @xmath141 , imply that for the @xmath247-coordinate @xmath248 , we have @xmath249 this proves ( [ itpflongtimeexists ] ) .
next , note that @xmath250 is well - defined for all @xmath237 by part ( [ itpflongtimeexists ] ) .
estimates similar to the previous one show that , for @xmath251 small enough , @xmath252 and moreover @xmath253 for all @xmath254 . decomposing @xmath255 , with @xmath256 a spatial vector field , and writing @xmath257 for the time and spatial components , respectively
, @xmath258 satisfies the ode @xmath259 with initial condition @xmath260 . in view of and the decay properties of @xmath196 , this gives @xmath261 for all @xmath84 , and hence the conclusion . by definition of @xmath113
, we have @xmath262 in @xmath263 for all @xmath116 .
hence , taking the limit as @xmath264 in the topology of @xmath265 , we have @xmath266 where for the second equality we used that the pull - back by @xmath231 preserves @xmath26 .
the proof of theorem [ thmintrouniq ] is complete . | we show that a stationary solution of the einstein maxwell equations which is close to a non - degenerate reissner
nordstrm
de sitter solution is in fact equal to a slowly rotating kerr
newman
de sitter solution .
the proof uses the non - linear stability of the kerr newman
de sitter family of black holes with small angular momenta , recently established by the author , together with an extension argument for killing vector fields .
our black hole uniqueness result only requires the solution to have high but finite regularity ; in particular , we do not make any analyticity assumptions . |
due to their remarkable mechanical properties upon loading followed by heating [ `` shape memory effect '' ( sme ) ] and vice versa [ `` superelasticity '' ( se ) ] , shape memory alloys ( smas ) are widely applied in these days in diverse macroscopic engineering devices @xcite .
microscopically , these effects result from the martensitic phase transformation between crystal structures of different symmetries , and from the formation of differently oriented variants of the low - temperature ground state phase ( `` martensite '' ) .
the question , whether and how this behavior takes place in systems with very small lengths scales , where a strong influence of the free surfaces comes into play , is of increasing interest in research and technology , for example in order to construct functional nanoscale actuators @xcite . +
a lot of effort has been done in studying the martensitic and reverse ( `` austenitic '' ) transformation in nanosized systems ( an overview is given in @xcite ) , where for example new types of domain configurations emerge upon cooling of nickel - titanium nanocrystals below strongly size dependent transition temperatures @xcite .
experimental studies of se and sme of niti @xcite and cualni nanopillars @xcite show , that these effects take place at least down to length scales of several hundred nanometers , but , in case of cualni , there is a considerable size dependence of the stress - strain curves if compared to bulk systems @xcite .
+ besides experiments , computer simulations provide a tool for investigating sme and se at an atomic scale . by applying loading , unloading and heating conditions to a martensitic nial system under periodic boundary conditions ,
total deformation recovery was obtained in a molecular dynamics ( md ) simulation @xcite .
strong zigzag behavior of the stress - strain curve upon loading due to abrupt stress releases was shown to be smoothed by differently oriented martensitic grains as starting configuration @xcite , which fits the realistic situation in a better way .
concerning nanosized systems , the austenite - martensite transformation upon loading , which is the origin of se , has been studied by md by straining a niti nanocrystal with free surfaces , leading to stress decrease if martensite nucleates heterogeneously in the sample @xcite . by applying md , several characteristics of structural phase transformations in niti model systems
have been investigated in the past by suzuki and coworkers @xcite with a pair potential , as well as by ishida and hiwatari @xcite , who used the modified embedded atom method .
a new type of sme in fcc metal nanowires , triggered by an extremely high surface - to - volume fraction and the one - dimensional character of the system , could be identified by md @xcite . here
, a structure conserving change of the lattice orientation propagates through the wire in order to accommodate the external load along the wire axis , leading to strains of over 50% .
heating above a critical temperature changes the configuration back to the initial lowest energy orientation due to strong surface stresses , whereupon the strain is totally released .
+ by performing md , a related kind to this nanoscale shape memory behavior is studied in the present work in a niti model system with different ni content .
the term `` model system '' is used , since the applied interatomic potential leads to a lowest energy martensitic ground state structure , which has strong similarities to the experimentally observed monoclinic b19@xmath0 , without being exactly the same .
nevertheless , the simulated nanocrystals contain of differently oriented variants in the martensite , what can be explained in general by minimization of the total energy , composed of surface , bulk and domain wall parts .
an external load is applied to the system , leading to a reorientation of the variants at a critical stress , accompanied by a plateau in the stress - strain curves .
a permanent strain remains after relaxing the load , but it is released upon heating above a strongly concentration dependent temperature .
since these transition temperatures , at which the original domain configuration is recovered , lie below the austenitic start temperatures , this kind of sme at the nanoscale is conspicuously different to the mechanism in conventional bulk smas .
in the the present study , a semiempirical interatomic potential of the finnis - sinclair type @xcite is used .
it has its origin in the tight binding method in second moment approximation of the electronic density of states @xcite .
the parameters for describing a niti alloy , i.e. values for ni - ni , ti - ti , and ni - ti interaction , have been determined by lai and liu @xcite .
the authors of the present paper have shown in a recent work @xcite , that the application of a cut - off function proposed by baskes _
@xcite with a width of @xmath1=@xmath2 results in a stable monoclinic b19@xmath0 structure in a bulk system at low temperatures , with lattice parameters comparing well with experimental @xcite and _ ab initio _ @xcite values .
in addition , an entropically stabilized nearly cubic b2 like structure is obtained upon heating . if the system is cooled out of this high - temperature phase , a martensitic ground state structure denoted @xmath3 emerges .
this structure is closely related to b19@xmath0 , since it shows nearly the same peak structure in the radial distribution function , but the angular distribution of the nearest neighbor lengths is different , resulting in a ground state energy , which is about 0.04% lower than that of b19@xmath0 @xcite . in that work ,
periodic boundary conditions were applied in order to simulate bulk behavior .
the present paper deals with a nano structure , where angular distributions of the nearest neighbor lengths in @xmath3 are observed , which differ slightly from the one described in @xcite . at free surfaces , these structures , denoted a and b , are energetically more favourable .
fig.[f1 ] shows all these structures resulting from the model potential , together with an order parameter @xmath4 , which is calculated out of the nearest neighbor environment of each atom @xcite . due to the shear in the monoclinic b19@xmath0 structure , two nearest neighbor lengths are elongated about 13.5% compared to the other six , denoted by a dashed connection between the atoms . in the b2 like high temperature state , the lengths are all nearly the same , differing only about 3% in another distribution . as mentioned above , in the @xmath3 configuration ,
the elongations of nearest neighbor lengths are as long as in b19@xmath0 , but they are arranged in another way .
the order parameter @xmath4 is calculated out of the nearest neighbor distances , relative to an underlying b19@xmath0 reference structure with determined shear direction [ 100 ] .
this leads to different values of @xmath4 in @xmath3 , depending on the direction of the elongated lengths relative to [ 100 ] ( [ 100 ] a , or [ 010 ] b ) .
these are just differently oriented variants of the same structure . and @xmath3 ( a , b ) , 3% in b2 ] .
@xmath4 is an order parameter , with which the structures can be resolved on an atomic scale , and can be visualized by a color coding . ] for symmetry reasons , one could think of a @xmath3 version c as well , where the elongated neighbor lengths point into the [ 001 ] direction .
in contrast to the versions a and b , we could not identify its existence in the martensitic state of the simulated nanoparticle . as explained below in more detail
, the formation of domain walls between different orientations of the same structure costs a certain amount of energy . at small system sizes ,
this loss in energy is higher than the energy gain due to a possible free surface with a third orientation c , leading to only two observed variants in our system .
since the martensitic ground state of the used model potential is the @xmath3 structure , which is not observed to exist experimentally in niti , the system is denoted as a niti model system in the following . in this sense ,
the present study deals with the general behavior of a nano shape memory alloy consisting of differently oriented variants , where load and temperature is applied , and explains it from a fundamental point of view , rather than reproducing exact niti behavior .
in addition , it is investigated , how the behavior , which is mainly triggered by a huge surface - to - volume fraction , changes , if the concentration of a perfectly ordered bi - atomic alloy is altered slightly .
the md simulations are performed with a velocity - verlet algorithm @xcite ( timestep @xmath5 s ) , and the temperature is controlled by a nos - hoover thermostat @xcite .
in the following , a niti nanoparticle with dimensions 20x8x8 nm@xmath6 , consisting of 90000 atoms , is considered , with an ordered alloy structure and a concentration of nickel varying between 50% and 52% . in the martensitic state , which can be obtained by cooling a b2 system below the martensitic transition temperature , the particle consists of differently oriented @xmath3 domains ( twins ) , in a configuration as shown in fig.[f2](a ) .
this type of domain structure is determined by the geometry of the nanoparticle , since it results from an interplay between surface energy contributions and the energy of twin boundaries . at the free y - z - surface ,
the a orientation of @xmath3 is energetically more favorable , while at the free x - z - surface it is the b version .
this can be explained easily , since there is an angle of 90 degrees between the y - z- and the x - z - surfaces , and the @xmath3 orientation b is obtained by rotating the version a by the same angle .
so , both surfaces have the same atomic configuration .
this is only possible , if there are twin boundaries in the system between variants a and b. since the formation of twin boundaries costs an amount of energy @xmath7 , a multi variant martensitic structure in the nanoparticle only forms , if the energy gain due to the surface reorientation is bigger than @xmath7 .
then , the total energy of the multi variant configuration is lower than that of a single oriented @xmath3 domain .
the dimension of the variants depends on the relation between the surface areas , as was observed in the simulation of 40x5x4 nm@xmath6 and 30x4x4 nm@xmath6 systems , where much smaller @xmath3 ( a ) domains were formed in the martensitic transition .
this kind of geometry dependent variant configuration , triggered by the interplay between surface and twin boundary energies , might be a general attribute of nanoscale martensitic transforming systems . in martensitic start configuration ( a ) , during loading [ ( b)-(d ) ] , unloading ( e ) and heating [ ( f)-(h ) ] process . ]
it is experimentally well known and one of the main reasons for the occurrence of the shape memory effect , that the orientation of martensitic domains changes according to an externally applied load , resulting in a remanent deformation @xcite . in order to observe and study this behavior in the nanoparticle by simulation
, a tensile load was applied in x - direction , as indicated by the arrows in fig.[f2](a ) .
the outermost layer of the y - z - plane in both directions was moved about @xmath5 m at every timestep , leading to a loading rate of 1 m / s .
the atoms of these surfaces were prevented from moving back to the equilibrium configuration by setting their velocities and accelerations to zero at every timestep .
this lead to a linearly increasing strain of the nanoparticle .
the corresponding stress was calculated as the negative internal pressure in the system .
fig.[f3 ] shows the stress - strain curves for five different ni concentrations at @xmath8=1 k. concentrations varying from 50% were achieved by choosing a corresponding amount of ni atoms randomly and replacing them by ti atoms .
the curves represent mean values over five different of those random distributions .
= @xmath9 k for 5 different ni concentrations .
marked points ( a)-(e ) correspond to the visualizations in fig.[f2 ] . ]
there is a strong stress increase at small strains , corresponding to elastic deformation .
the slopes are slightly bigger at concentrations lying closer to the perfect configuration with 50% ni , indicating a stiffer elastic behavior in this case .
at a critical strain of about 1.8% , a stress plateau is reached until about 7% strain .
the higher the ni content , the smoother is the stress in this range , with less abrupt increases and releases .
this behavior , as well as the lower stiffnesses at higher ni contents can be explained by a growing lattice instability , what was figured out to be the reason for the strong decrease of transition temperatures with ni content in a previous work @xcite .
figs.[f2](b)-(d ) show the development of the microstructure during the loading process .
it is seen , that at the critical stress , a reorientation of the b to the a version of the @xmath3 structure begins , what is proceeded mainly by twin boundary movement , until a nearly single domain structure remains . at higher strains ,
the stress increases linear due to elastic deformation again . in the next step of the simulation , the external load
was instantaneously removed , leading to a relaxation of the system and vanishing stress .
after the relaxation process , which was performed over 50000 timesteps ( 50 ps ) , the single domain structure stayed stable [ fig.[f2](e ) ] , accompanied by a remanent strain between about 5.6% ( 50% ni ) and 5% ( 52% ni ) . as known for macroscopic smas , the recovery of the shape , which they had before loading , takes place , if the systems are heated over a certain temperature , at which each martensitic variant adopts the same configuration of a higher symmetric structure ( austenite ) , independent of the initial orientation . in order to analyze ,
whether and how this effect occurs in the nanoparticle studied in this work , the systems were heated with a rate of 1 k every 1000 timesteps ( 1 k / ps ) .
fig.[f4 ] shows the strain behavior during this heating process up to 400 k for the different ni concentrations .
it is common for all systems , that the strain stays almost constant at low temperatures . as depicted in fig.[f2](f ) , the martensitic structure containing a single orientation remains stable within this temperature range . at a certain temperature , which depends strongly on concentration , a drop of the strain begins , leading to a nearly complete disappearance . the microstructural evolution during this process is shown in fig.[f2](g ) .
instead of a shape recovery due to the transformation to austenite , the initial multi variant configuration is adopted by the system .
a subsequent cooling would end up in the starting configuration again .
an explanation of this behavior is as follows : both , the single , as well as the multi variant state of the nanoparticle are free energy minima , with the latter being the global minimum , i.e. the ground state of the system .
but there is a free energy barrier between these states , corresponding to the energy needed to form twin boundaries and to reorientate parts of the structure .
the height of this barrier depends on ni content , since , as mentioned above , higher ni concentrations lead to more instable lattice structures , in which twin boundaries can more easily be formed . at a certain temperature , the energy barrier can be overcome by the system , leading to a recovery of the multi variant structure .
further heating beyond this temperature initiates the phase transformation from @xmath3 martensite to b2 austenite , ending up in a system with 2.5% strain , independent of ni concentration [ fig.[f2](h ) ] .
cooling below the martensitic transition temperature would trigger the b2@xmath10@xmath3 phase change , transforming the system back to the initial configuration [ fig.[f2](a ) ] , what is the conventional sme . .
in conclusion , a md study of the sme of a niti model system at the nanoscale is presented .
a tensile load was applied to a nanoparticle with a martensitic structure containing variants of different orientation . due to the load , the boundaries between these twins move and the system reaches a single variant configuration with a remaining non - zero strain after unloading . during a subsequent heating process , a shape recovery different to the conventional mechanism of the sme occurs .
it is not the phase transition to austenite which makes the strain vanish , but a transformation back to the ground state configuration with multiple twin orientations , without a change of the crystal structure .
the height of the energy barrier , which has to be overcome for this process , decreases significantly , if the perfectly ordered bi - atomic alloy composition is changed slightly .
since the distribution of variant orientations in the martensite depends mainly on surface energy contributions , this kind of sme occurs due to the huge surface - to - volume fraction in nanoscale systems and not in macroscopic smas .
+ 28 natexlab#1#1[2]#2 ( ed . ) , , , .
, , , , ( ) . , , , , ( ) .
, , , , ( ) . , , , ( ) .
, , , ( ) . , , , ( ) .
, , , ( ) . , , , ( ) .
, , , ( ) . , , ( ) .
, , , , , , , , ( ) .
, , , ( ) .
, , ( ) . , , ( ) .
, , , ( ) . , , , , , , ( ) .
, , , ( ) . , , ( ) .
, , , , ( ) . , ( ) .
, , , ( ) . | the shape memory behavior of a niti nanoparticle is analyzed by molecular dynamics simulations . after a detailed description of the equilibrium structures of the used model potential , the multi variant martensitic ground state , which depends on the geometry of the particle ,
is discussed .
tensile load is applied , changing the variant configuration to a single domain state with a remanent strain after unloading .
heating the particle leads to a shape memory effect without a phase transition to the austenite , but by variant reorientation and twin boundary formation at a certain temperature .
these processes are described by stress - strain and strain - temperature curves , together with a visualization of the microstructure of the nanoparticle .
results are presented for five different ni concentrations in the vicinity of 50% , showing for example , that small deviations from this ideal composition can influence the critical temperature for shape recovery significantly . |
theoretical studies on black holes have attracted substantial attention since the advent of general relativity .
it is well known that black holes can not be directly observed but they can be detected through their action on their neighborhoods .
the behavior of the matter and fields surrounding a black hole not only tells us about its presence but also helps us determine its parameters .
a real black hole can never be fully described by its basic parameters ( mass , charge and angular momentum ) and is always in the perturbed state ( @xcite ) . for several decades
, many efforts are made in investigating the evolution of external field perturbation around black holes .
these perturbations lead to damped oscillations called quasinormal modes(@xcite ) .
different kinds of perturbation in the geometry of a black hole can excite certain combination of its characteristic frequencies of the normal modes for which the frequencies are no longer purely real showing that the system is losing energy .
the real part of qn frequency represents the ring down frequency and the imaginary part , the decay time .
the resonant frequencies of the response of external perturbations called quasinormal frequencies are one of the most essential characteristics of black holes .
it is widely believed that the qnms carry a unique characteristic `` fingerprint '' which would lead to the direct identification of the black hole existence .
quasinormal modes associated with perturbations of different fields were considered in a lot of works ( @xcite ) .
we know that the vacuum undergoes quantum fluctuations .
this phenomenon is the appearance of energetic particles out of nothing , as allowed by the uncertainty principle .
vacuum fluctuations have observable consequences like casimir force between two plates in vacuum . due to quantum fluctuations
, the evolution of slices in the black hole geometry will lead to the creation of particle pairs which facilitates black holes radiation ( @xcite ) .
it was shown that quantum vacuum fluctuations modify the geometry of the schwarzschild black hole .
@xcite obtained new expressions of the schwarzschild metric when the back reaction of the spacetime due to quantum fluctuations is taken into account .
recently , wontae and yongwan ( @xcite ) investigated phase transition of the quantum - corrected schwarzschild black hole and concluded that there appear a type of grass - perry - yaffe phase transition due to the quantum vacuum fluctuations and this held even for the very small size black hole .
more recently , we investigate qnms of scalar perturbation around a quantum - corrected schwarzschild black hole and show that the scalar field damps more slowly and oscillates more slowly due to the quantum fluctuations ( @xcite ) . in this paper , the qnms of a quantum - corrected schwarzschild black hole is investigated in order to highlight the behavior of a gravitational and dirac perturbations when the vacuum fluctuations are taken into account .
the paper is organized as follows . in section [ sec:1 ] ,
the quantum - corrected schwarzschild black hole is presented . in section [ sec:01 ]
, we derive the wave equation of a gravitational perturbation in the quantum - corrected schwarzschild black hole background . in section [ sec:2 ] , the wave equation for dirac perturbation of the black hole
is derived . in section [ sec:3 ]
, we evaluate the qn frequencies of the gravitational and dirac perturbations by using the third order wkb approximation method .
the last section is devoted to a summary and conclusion .
according to the work of kazakov and solodukhin on quantum deformation of the schwarzschild solution ( @xcite ) , the background metric of the schwarzschild black hole is defined by @xmath1 where @xmath2 for an empty space , @xmath3 .
thus we obtain the schwarzschild metric @xmath4 where @xmath5 is the black hole mass . taking into account the quantum fluctuation of vacuum , the quantity @xmath6 transforms to ( @xcite ) @xmath7 where @xmath8
, @xmath9 is the newton constant and @xmath10 is a scale parameter .
the background metric of the quantum - corrected schwarzschild black hole can then be read as @xmath11 with @xmath12 and @xmath13 .
the quantum correction parameter @xmath0 has the dimensionality of length @xmath14 $ ] . using this form of the metric
, the scalar curvature of the spacetime is transformed to @xmath15\ ] ] as we can see , the scalar curvature does not depend on the mass of the gravitating body but depends on the quantum correction parameter . for large @xmath16 , the metric becomes @xmath17 which looks like the metric of a charged body with mass @xmath5 and imaginary charge @xmath18 .
thus , we can say that the quantum correction parameter @xmath0 acts as an imaginary charge added to the spacetime metric .
its coincides with the reissner - nordstrm metric but with an imaginary electric charge .
it is therefore interesting for us to evaluate the qnms of the schwarzschild black hole when the quantum fluctuations modify it .
for the gravitational perturbation , we will deal with the regge - wheeler gauge ( @xcite ) . in this gauge
, the gravitational perturbation is regarded as perturbation to the background metric .
we regard the perturbed background metric @xmath19 as the sum of the unperturbed metric @xmath20 and the perturbation in it @xmath21 @xmath22 where @xmath23 is the perturbed metric .
the perturbations @xmath24 are supposed to be very small compared with @xmath25 .
the @xmath24 can be calculated from @xmath25 , and @xmath26 from @xmath27 .
the quantity @xmath28 can be expressed in the form(@xcite ) @xmath29 where @xmath30 similarly , we will use the unperturbed christoffel symbols in computing covariant derivatives of perturbation quantities . from eq .
( [ 8 ] ) it follows that @xmath31 the perturbed christoffel symbols are given by @xmath32 where @xmath33 the last equation reveals that the variation of the christoffel symbols , @xmath34 , forms a tensor , even though the christoffel symbols themselves do not .
adopting the @xcite matching , the final canonical form for an odd wave is then @xmath35 this form represents the perturbation of a spherically symmetric black hole in regge - wheeler gauge .
the radial wave equation is the final equation which gives us the behavior of the perturbation s oscillation .
it can be reduced to a simple linear differential equation . in vacuum
, the perturbed field equations simply reduce to ( @xcite ) @xmath36 using that and the background metric , we get from @xmath37 , @xmath38=0 \end{array}\ ] ] and from @xmath39 , @xmath40 defining @xmath41 and @xmath42 and eliminating @xmath43 , we then get @xmath44 where @xmath45 equation ( [ 19 ] ) is a second order linear differential equation which represents a perturbation equation .
the potential @xmath46 depends explicitly on the radius @xmath47 .
its behavior is represented on fig.[fig:1 ] . through this figures
, we can see that the potential decreases with increasing @xmath0 .
moreover , the potential takes constant values when @xmath48 evolves to infinity .
that allows us using the wkb approximation method provided by iyer and will ( @xcite ) to evaluate the quasinormal frequencies .
for the general background spacetime , the massless dirac equation is @xmath49\psi=0,\ ] ] where @xmath50 is the dirac matrix , @xmath51 is the inverse of the tetrad @xmath52 by the metric @xmath53 , @xmath54 is the minkowski metric given by @xmath55 , @xmath56 is the spin connection defined as @xmath57e^\nu_ae_{b\nu;\mu}$ ] and @xmath58 . in order to separate the dirac equation , we must introduce the tetrad @xmath59 using this expression , the dirac equation becomes @xmath60 in order to simplify the above equation , we can define the perturbation @xmath61 as @xmath62 then , eq . ( [ dirac3 ] ) can be rewritten in its simplified form as @xmath63 the pauli matrices @xmath64 are given by @xmath65 we introduce the tortoise coordinate @xmath66 and the ansatz for the dirac spinor @xmath67 where @xmath68 @xmath69 @xmath70 represent the ordinary spherical harmonics . since @xmath71 equation ( [ dirac3 ] ) can be written in the simplified matrix form @xmath72 the cases ( + ) and ( - ) in the functions can be put together after some matching and the equation can be decoupled as ( @xcite ) @xmath73 with@xmath74 where @xmath75 is the total angular momentum number . the behavior of the potential is plotted on fig.[dp ] . for @xmath76,title="fig:",scaledwidth=45.0% ]
+ these equations can be solved using the wkb approximation method to find the quasinormal modes .
the wave equation ( [ 19 ] ) can be rewritten as : @xmath77 where @xmath78 . for a black hole ,
the qn frequencies correspond to solution of perturbation equation which satisfy the boundary conditions appropriate for purely ingoing waves at the horizon and purely outgoing waves at infinity .
incoming and outgoing waves correspond to the radial solution proportional to @xmath79 and @xmath80 , respectively .
only a discrete set of complex frequencies satisfies these conditions .
to evaluate the qn frequencies , we applied here the third order wkb approximation method derived by schutz , will ( @xcite)and iyer ( @xcite ) to the above equation and these qn frequencies are given by ( @xcite ) @xmath81-i(n+\frac{1}{2})(-2v''_{0})^{1/2}[1+\tilde{\omega}],\ ] ] where @xmath82 @xmath83 , and @xmath84 . using equation ( [ 22 ] )
, we calculated numerically the qn frequencies of the gravitational perturbation for @xmath85 without and with quantum correction of the black hole .
the results are shown in tables [ tab:1 ] and [ tab:2 ] , where @xmath86 is the harmonic angular index , @xmath87 is the overtone number , @xmath88 is the complex qn frequency and @xmath0 is the quantum correction parameter . [
cols="<,<,<,<,<,<",options="header " , ] as observed for the gravitational perturbation , we can also remark that the real parts and the absolute values of the imaginary parts of the quasinormal frequencies decrease when increasing the quantum correction parameter .
the particularity of dirac perturbation here is that the imaginary parts of the ( @xmath89 ) frequencies do not depend on the angular harmonic index ( @xmath86 ) .
the relationship between the real and imaginary parts of quasinormal frequencies of the gravitational as well as dirac perturbations in the background of the black hole is plotted in fig.[fig:2 ] .
the behaviors of the real and imaginary parts of the quasinormal frequencies when varying the quantum - correction parameter , @xmath0 , are also plotted ( see figs .
[ fig:02 ] and [ fig:002 ] ) .
.,scaledwidth=45.0% ] .,scaledwidth=45.0% ] .,scaledwidth=45.0% ]
in summary , qnms of a gravitational and dirac perturbations around a schwarzschild black hole were evaluated using the third order wkb approximation . for the gravitational perturbation ,
the results of table [ tab:1 ] are obtained without considering any kind of correction while those of table [ tab:2 ] are obtained when considering the quantum - corrected schwarzschild black hole metric .
dirac quasinormal frequencies are summarized in tables [ tab:3 ] and [ tab:4 ] for the cases of the schwarzschild black hole with and without quantum correction , respectively . through the above tables
, we can remark that the absolute values of the imaginary parts of the quasinormal frequencies of the quantum - corrected black hole are smaller compared to those without quantum correction , for fixed set of @xmath86 and @xmath87 .
moreover , we can remark that these values decrease when increasing @xmath0 .
that can be clearly seen through figs .
[ fig:2 ] and [ fig:002 ] . through fig .
[ fig:2 ] , we an also remark that the real parts of dirac quasinormal frequencies are higher compared to gravitational ones .
that implies that dirac qnms oscillate more rapidly than gravitational ones . concerning the real parts
, they are also decreasing when increasing @xmath0 .
thus , we can conclude that , due to the quantum fluctuations , the quasinormal modes of gravitational and dirac perturbations of the schwarzschild black hole damp more slowly and oscillate more slowly .
cardoso v. , lemos , j.p.s . : * 63 * , 124015 ( 2001 ) cardoso , v. , konoplya , r.a . , &
lemos , j.p.s . , * 68 * , 044024 ( 2003 ) castello - branco , k.h.c . , konoplya , r.a . ,
zhidenko , a. : * 71 * , 047502 ( 2005 ) chandrasekhar , s. , detweiler , s.l .
london a * 344 * , 441 ( 1975 ) cho , h.t . * 68 * , 024003 ( 2003 ) dreyer , o. : * 90 * , 081301 ( 2003 ) eisenhart , l.p . : _
riemannian geometry _ 1926 ( princeton : princeton university ) hawking , s.w . , ellis , g.f.r .
1973 , the large scale structure of space - time , cambridge university press , cambridge hod , s. : * 81 * , 4293 ( 1998 ) hod , s. , piran , t. * 58 * , 024017 ( 1998 ) iyer , s. , will , c.m .
* 35 * , 3621 ( 1987 ) jing , j. , pan , q. : phys .
b * 660 * , 13 ( 2008 ) jing , j.l .
d * 69 * , 084009 ( 2004 ) . j.l .
: jing , phys .
d * 70 * , 065004 ( 2004 ) ; phys .
d * 71 * , 124011 ( 2005 ) ; phys .
d * 71 * , 124006 ( 2005 ) kazakov , d.i . ,
solodukhin , s.n .
b * 429 * , 153 ( 1994 ) konoplya , r.a . * 66 * 084007 ( 2002 ) .
konoplya , r.a . , abdalla , e. : * 71 * , 084015 ( 2005 ) konoplya , r.a . ,
zhidenko , a. : rev .
phys . * 83 * , 793 ( 2011 ) leaver , e.w .
a * 402 * , 285 ( 1985 ) ; phys .
d * 34 * , 384 ( 1986 ) mahamat , s. , bouetou , b.t . , kofane , t.c . : chin .
* 26 * , 109802 ( 2009 ) ; astrophys .
space sci . * 333 * , 449 ( 2011 ) mahamat , s. , bouetou , b.t .
, kofane , t.c .
: astrophys .
space sci .
* 350 * , 721 ( 2014 ) maldacena , j. : adv .
* 2 * , 231 ( 1998 ) mathur , s.d . : class .
quantum grav .
* 26 * 224001 ( 2009 ) mellor , f. , moss , i.g . :
d * 41 * , 403 ( 1990 ) moss , i.g . , norman , j.p . : class .
quantum grav .
* 19 * , 2323 ( 2002 ) natario , j. , schiappa , r. : hep - th/0411267 .
nollert , h.p . : class .
quantum grav .
* 16 * , r159 ( 1999 ) rama , s.k .
, sathiapalan : mod phys .
a * 14 * , 2635 ( 1999 ) regge , t. , wheeler , j.a .
: phys . rev . *
108 * , 1063 ( 1957 ) schutz , b.f .
, & will , c.m . , 291 , l33 ( 1985 )
setare , m.r . : class .
. grav . * 21 * , 1453 ( 2003 ) ; phys .
d * 69 * , 044016 ( 2004 ) .
starinets , a.o .
d * 66 * , 124013 ( 2002 ) wang , c .- y .
, zhang , y. , gui , y .- x . ,
l , j .- b .
. phys . * 53 * , 882 ( 2010 ) witten , e. : adv .
* 2 * , 253 ( 1998 ) .
wontae , k. , yongwan , k. : phys .
b * 718 * , 687 ( 2012 ) zhang yu , gui , y.x .
, fei yu , feng lin li gen . rel
. grav . * 39 * , 1003 ( 2007 ) zerilli , f.j .
: phys . rev .
* 24 * , 737 ( 1970 ) a. zhidenko , class .
quant . grav .
21 273 ( 2004 ) . | in this work , quasinormal modes ( qnms ) of the schwarzschild black hole are investigated by taking into account the quantum fluctuations .
gravitational and dirac perturbations were considered for this case .
the regge - wheeler gauge and the dirac equation were used to derive the perturbation equations of the gravitational and dirac fields respectively and the third order wentzel - kramers - brillouin ( wkb ) approximation method is used for the computing of the quasinormal frequencies .
the results show that due to the quantum fluctuations in the background of the schwarzschild black hole , the qnms of the black hole damp more slowly when increasing the quantum correction factor ( @xmath0 ) , and oscillate more slowly . |
many mathematical models of real - life processes pose challenges during numerical computations , due to their large size and complexity .
model order reduction ( mor ) techniques are methods that reduce the computational complexity of numerical simulations , an overview of mor methods is provided in @xcite .
mor techniques such as balanced truncation ( bt ) and singular perturbation approximation ( spa ) are methods which have been introduced in @xcite and @xcite , respectively , for linear deterministic systems @xmath0 here @xmath1 is asymptotically stable , @xmath2 , @xmath3 and @xmath4 , @xmath5 , @xmath6 are state , output and input of the system , respectively . from the gramians @xmath7 and @xmath8 which solve dual lyapunov equations @xmath9 a balancing transformation is found , which is used to project the state space of size @xmath10 to a much smaller dimensional state space ( see , e.g. @xcite ) .
recently , the theory for bt and spa has been extended to stochastic linear systems of the form @xmath11 where @xmath12 , @xmath13 and @xmath14 as above , and @xmath15 and @xmath16 ( @xmath17 ) are uncorrelated scalar square integrable lvy processes with mean zero ( often @xmath18 and the special case of wiener processes are considered , see , for example , @xcite ) . in this case
bt and spa require the solution of more general lyapunov equations of the form @xmath19 where @xmath20 $ ] for general lvy processes .
note that @xmath21 , @xmath22 for the case of a wiener process @xcite . we refer to @xcite for a detailed theoretical and numerical treatment of balancing related mor for ( [ linsysafterdisintr ] ) . in this paper
we are going to study balancing related mor for systems of the form @xmath23 where @xmath24 and @xmath25 is the @xmath26 column of @xmath27 .
the processes @xmath28 are the components of a square integrable mean zero lvy process @xmath29 that takes values in @xmath30 . consequently , these components are not necessarily uncorrelated . for a general theoretical treatment of sdes with lvy noise
we refer to @xcite .
the setting in ( [ linsysafterdisintr2 ] ) is of particular interest in many applications .
if one is interested in a large number of different realisations of the output @xmath31 ( e.g. to compute moments of the form @xmath32 $ ] ) , then one needs to solve the sde in ( [ linsysafterdisintr2 ] ) a large number of times . for a state space of high dimension
this is computationally expensive .
reduction of the state space dimension decreases the computational complexity when sampling the solution to ( [ linsysafterdisintr2 ] ) , as the sde can then be solved in much smaller dimensions . hence the computational costs are reduced dramatically .
the linear system ( [ linsysafterdisintr2 ] ) is a problem where the control is noise . in this case
the standard theory for balancing related mor applied to a deterministic system no longer applies .
balanced truncation has been applied to linear systems with white noise before .
the discrete time setting was discussed in @xcite . for the continuous time
setting , dissipative hamiltonian systems with wiener noise were treated in @xcite , but no error bounds were provided . in this paper
we consider both bt and spa model order reduction .
as far as we are aware , no theory and in particular error bounds for balancing related mor have been developed for continuous time sdes with lvy noise . using theory for linear stochastic differential equations with additive lvy noise we provide a stochastic concept of reachability .
this concept motivates a new formulation of the reachability gramian .
we prove bounds for the error between the full and reduced system which provide criteria for truncating , e.g. criteria for a suitable size of the reduced system .
we analyse both bt and spa and apply the theory directly to an application arising from a second order damped wave equation .
we now consider a particular example which explains why the above setting is of practical interest .
[ [ motivational - example ] ] motivational example + + + + + + + + + + + + + + + + + + + + in @xcite the lateral time - dependent displacement @xmath33 of an electricity cable impacted by wind was modeled by the following one - dimensional symbolic second order spde with lvy noise : @xmath34 for @xmath35 $ ] , @xmath36 $ ] and @xmath37 , with boundary and initial conditions @xmath38 for small @xmath39 , the output equation @xmath40 is approximately the position of the middle of the cable . in @xcite , it is shown that transforming this spde in into a first order spde and then discretising it in space , leads to a system of the form ( [ linsysafterdisintr ] ) where @xmath41 .
one drawback of the approach above is , that , when the electricity cable is in steady state , the wind has no impact . a more realistic scenario , which models the wind as some form of stochastic input , is the following symbolic equation @xmath42 for @xmath35 $ ] , @xmath36 $ ] and @xmath37 , boundary and initial conditions as in ( [ eq : bcic ] ) , and @xmath16 the components of a square integrable mean zero lvy process @xmath29 that takes values in @xmath30 .
in this paper , we consider a framework which covers this model .
moreover we modify the output in ( [ introbspout ] ) and let @xmath43 so that both the position and velocity of the middle of the string are observed . transformation and discretisation of
this spde leads to a system of the form ( [ linsysafterdisintr2 ] ) where @xmath12 is an asymptotically stable matrix , i.e. @xmath44 .
this paper is set up as follows .
section [ sec : balancing ] provides the theoretical tools for balancing linear sdes with additive lvy noise .
we explain the theoretical concepts of reachability and observability in this setting and show how this motivates mor using bt and spa .
moreover we provide theoretical error bounds for both methods . in section
[ sec : wave ] we show how a wave equation driven by lvy noise can be transformed into a first order equation and then reduced to a system of the form ( [ linsysafterdisintr2 ] ) by using a spectral galerkin method . numerical results which support our theory are provided in section [ sec : numerics ] .
in @xcite balancing related mor was considered for deterministic systems of the form @xmath45 where @xmath46 was assumed to be asymptotically stable , i.e. @xmath44 , @xmath2 , @xmath3 and @xmath47)$ ] for all @xmath48 was a deterministic control .
we now turn our attention to a stochastic system @xmath49 which , in section [ subsecspeltralgal ] , represents a spatially discretised version of an spde .
the matrices @xmath12 , @xmath13 and @xmath14 are as above and the @xmath30-valued process @xmath50 is a square integrable lvy processes with mean zero .
one might interpret system ( [ stochsysnew ] ) as system ( [ detsysant ] ) with @xmath51 but the noise @xmath52 is no control in the classical sense .
first of all , stochastic controls were not admissible in the deterministic setting and secondly the classical derivative does not exist .
so , if we want to study balancing related mor for the `` particular control '' @xmath53 , we need to make sense of this setting which we do by the ito - type sde in ( [ stochsysnew ] ) . in the deterministic case reachability and observability concepts are introduced to characterise the importance of states .
difficult to reach states ( states which require large energy to reach them ) and difficult to observe states ( states which only produce little observation energy ) are seen to be unimportant in the systems dynamics . in balancing related mor
the idea is to create a system , where the dominant reachable and observable states are the same .
those are then truncated to obtain a reduced order model ( rom ) . applying balancing related mor to ( [ stochsysnew ] )
requires a few modifications compared to the classical deterministic framework .
we introduce a stochastic reachability concept in section [ reachsec ] which also leads to a different reachability gramian compared to the deterministic case . for the observation concept
we follow the deterministic approach .
we then describe the procedure of balancing for systems with additive noise in section [ subsecprocedure ] which is similar to the deterministic case .
afterwards , we will discuss two particular techniques which are bt and spa . since it is not a priori clear whether these approaches for system ( [ stochsysnew ] ) perform as well as for deterministic systems , we contribute an error bound for both bt and spa in section [ btandspa ] .
these error bounds enable us to point out the cases , where bt and spa work well , and they can be used to find a suitable rom dimension . with suitable reachability and observability concepts we want to analyze which states in system ( [ stochsysnew ] ) are unimportant and hence can be neglected .
[ [ reachability ] ] reachability + + + + + + + + + + + + we begin with a stochastic reachability concept , where the particular choice of @xmath50 is taken into account .
starting from zero ( @xmath54 ) in @xmath55 we investigate how much the noise can control the state away from zero .
we define what is meant be reachability in the stochastic case , where @xmath56 , @xmath57 , denotes the solution to ( [ stochstatenew ] ) with initial condition @xmath58 and noise process @xmath50 .
a state @xmath59 is not reachable from zero on the time interval @xmath60 $ ] , @xmath48 , if it is contained in an open set @xmath61 with @xmath62\right\}=0 , \end{aligned}\ ] ] else @xmath63 is reachable .
the system is called completely reachable if @xmath64\right\}>0 \end{aligned}\ ] ] for every open set @xmath65 .
we refer to @xcite , where weak controllability was analyzed for equations with wiener noise .
weak controllability turns out to be similar to condition ( [ completereach ] ) . to characterise the degree of reachability of a state ,
we introduce finite time reachability gramians @xmath66 $ ] which are the covariance matrices of @xmath67 at fixed times @xmath57 .
before we study the meaning of these gramians , we show that @xmath68 is the solution of a matrix differential equation .
the matrix - valued function @xmath68 , @xmath57 , is the solution to @xmath69 where @xmath70 $ ] is the covariance matrix of @xmath50 at time @xmath71 .
we replace @xmath67 by @xmath72 to shorten the notation in the proof . using ito s formula in corollary [ iotprodformelmatpro ]
, we obtain the following for @xmath73 , @xmath57 : @xmath74_t\right)_{i , j=1 , \ldots , n},\end{aligned}\ ] ] where @xmath75 is the @xmath76-th unit vector and we used @xmath54 . inserting the stochastic differential of @xmath72 yields @xmath77
since the ito integrals have mean zero , we have @xmath78=\int_0^t \mathbb e\left[x(s ) x^t(s)\right ] a^t ds+\int_0^t a \mathbb e\left[x(s ) x^t(s)\right ] ds+\left(\mathbb e[e_i^t x , e_j^t x]_t\right)_{i , j=1 , \ldots , n},\end{aligned}\ ] ] where we replaced @xmath79 by @xmath80 .
this does not impact the integrals since a cdlg process has at most countably many jumps on a finite time interval ( see ( * ? ? ?
* theorem 2.7.1 ) ) .
applying corollary [ iotprodformelmatpro ] again , the stochastic differential of @xmath81 is given by : @xmath82_t\right)_{i , j=1 , \ldots , n},\end{aligned}\ ] ] taking the expected value , we have @xmath83b^t=\mathbb e\left([e_i^t bm , e_j^tbm]_t\right)_{i , j=1 , \ldots , n}$ ] . in ( * ? ? ?
* theorem 4.44 ) it was shown that the covariance function is linear in @xmath84 , i.e. @xmath85=\mathcal q_m t$ ] . since the @xmath76th component @xmath86 has the same jumps and the same martingale part as @xmath87 , we know by ( [ decomqucov ] ) that @xmath88_t=[e_i^t x , e_j^t x]_t$ ] for @xmath89 . summarizing the results , we have @xmath78=\int_0^t \mathbb e\left[x(s ) x^t(s)\right ] a^t ds+\int_0^t a \mathbb e\left[x(s ) x^t(s)\right ] ds+b\mathcal q_m b^t t,\end{aligned}\ ] ] which concludes the proof . to find a representation for @xmath68 we need the following straightforward result .
[ prodeinfachh ] let @xmath90 and @xmath91 , then @xmath92 the product rule yields @xmath93 and integrating gives the result .
setting @xmath94 and @xmath95 in proposition [ prodeinfachh ] , we see that @xmath96 solves the differential equation ( [ matrixdiffgl ] ) .
since the solution to ( [ matrixdiffgl ] ) is unique , we have @xmath97 consequently , @xmath98 is an increasing function . if @xmath99 , then we obtain the reachability gramian of the deterministic setting ( [ detsysant ] ) , see @xcite .
this is also the case if @xmath50 is a standard wiener process .
the finite reachability gramian @xmath68 provides information about the reachability of a state which we see from the following identity : @xmath100}\mathbb e \left\langle x(t , 0 , m ) , x \right\rangle^2_{\mathbb r^n}=x^t p(t)x\quad \text{for } \quad x\in\mathbb r^n.\end{aligned}\ ] ] consequently , we know that @xmath101 , @xmath102 $ ] , @xmath103 a.s . if and only if @xmath104 meaning that @xmath67 is orthogonal to @xmath105 . since @xmath106 is symmetric positive semidefinite , we have @xmath107 and hence @xmath108\right\}=1,\end{aligned}\ ] ] we observe from ( [ impreachset ] ) that all the states that are not in @xmath109 are not reachable and thus they do not contribute to the system dynamics . as a first step to reduce the system dimension it is necessary to remove all the states that are not in @xmath109 .
we will see in the next proposition , that the finite reachability gramians can be replaced by the infinite gramian @xmath110 since their images coincide .
this ( infinite ) gramian exists due to the asymptotic stability of @xmath12 .
it is easier to work with @xmath7 since it can be computed as the unique solution to @xmath111 @xmath7 satisfies ( [ lyapeqreach ] ) since @xmath68 satisfies ( [ matrixdiffgl ] ) and @xmath112 if @xmath113 due the asymptotic stability of @xmath12 . for the case
@xmath114 this gramian was discussed in ( * ? ? ?
* section 4.3 ) in the context of balancing for deterministic systems ( [ detsysant ] ) .
[ refsameimbla ] the images of the finite reachability gramians @xmath68 , @xmath115 , and the infinite reachability gramian @xmath7 are the same , that is , @xmath116 since @xmath7 and @xmath68 are symmetric positive semidefinite , it is enough to show that their kernels are equal .
let @xmath117 .
this implies @xmath118 , since @xmath119 is increasing .
hence @xmath120 . on the other hand , if @xmath120 , we have @xmath121 consequently , @xmath122 for all @xmath123 $ ] .
since the entries of @xmath124 are analytic functions , the scalar function @xmath125 is analytic , such that @xmath126 on @xmath127 .
thus , @xmath128 and the result follows .
let us now assume that we already removed all the unreachable states from ( [ stochstatenew ] ) .
so , ( [ completereach ] ) holds which implies that @xmath129 .
we choose an orthonormal basis of @xmath130 , consisting of eigenvectors @xmath131 of @xmath7 , and the following representation holds : @xmath132 we investigate how much the noise influences @xmath67 in the direction of @xmath133 .
if a state remains close to zero , it barely contributes to the system dynamics .
those states can be identified with the help of the positive eigenvalues @xmath134 of @xmath7 . using ( [ supkleinpt ] ) and
the fact that @xmath106 is increasing , we obtain @xmath135}\mathbb e \left\langle x(t , 0 , m ) , p_k\right\rangle^2_{\mathbb r^n}=p_k^t p(t ) p_k\leq p_k^t p p_k=\lambda_k.\end{aligned}\ ] ] hence , if @xmath136 is small , then the the corresponding coefficient @xmath137 in ( [ eq : expansion ] ) is small ( in the @xmath138 sense ) .
this means that the noise hardly steers the state in the direction of @xmath133 .
consequently , the states that are difficult to reach are contained in the space spanned by the eigenvectors corresponding to the small eigenvalues of @xmath7 .
we continue by reasoning why using the modified reachability gramian @xmath7 is better than using the reachability gramian @xmath139 ( @xmath114 ) of the deterministic system ( [ detsysant ] ) .
[ propdworse ] the following properties hold for the ( modified ) reachability gramians @xmath7 and @xmath140 : * in general , we have @xmath141 * if @xmath142 ( positive definite ) , then @xmath143 . *
if @xmath144 , then @xmath145 .
let @xmath146 , then @xmath147 which is equivalent to @xmath148 on @xmath149 and implies @xmath150 on @xmath149 .
equivalently , we have @xmath151 and since @xmath152 if and only if @xmath153 , we have @xmath154 .
consequently , we obtain @xmath155 due to @xmath156 and @xmath157 .
if @xmath142 , then @xmath158 on @xmath149 implies @xmath148 on @xmath149 . in this case , all the above statements are equivalent .
therefore @xmath159 and hence @xmath160 . to prove ( c ) , assume @xmath161
pre- and postmultiplying ( [ lyapeqreach ] ) with @xmath162 and @xmath163 , respectively , yields @xmath164 this implies @xmath165 but if @xmath166 , then there is a @xmath117 such that @xmath167 .
we set @xmath168 , @xmath57 , and observe that @xmath169 is an analytic function that is not constantly zero since @xmath170 .
consequently , @xmath169 has only countably many zeros such that @xmath171 is a purely positive function up to lebesgue zero sets . hence ,
@xmath172 such that @xmath173 .
having @xmath174 implies ( c ) . by ( [ impreachset ] ) and proposition [ propdworse ] ( a )
, we obtain @xmath175\right\}=1.\end{aligned}\ ] ] one could now think of using @xmath140 instead of @xmath7 but from ( [ pdnaja ] ) not all unreachable states can be identified especially if case ( c ) in proposition [ propdworse ] holds . hence ,
if we were to use @xmath140 , we would underestimate the set of unreachable states .
even if we assume that the system is already completely reachable ( e.g. ( [ completereach ] ) holds ) , inequality ( [ interlambad ] ) can not be obtained with @xmath140 .
this means that we can not identify the difficult to reach states with the help of the eigenvalues of @xmath140 . in section [ btandspa ]
we will see that @xmath7 , rather than @xmath140 , enters the error bound for the rom .
finally we note that the reachability gramian @xmath140 of system ( [ detsysant ] ) does not depend on the input @xmath176 .
if a `` noisy control '' is used , this does not apply , since @xmath7 depends on @xmath177 and hence on the lvy process @xmath50 .
[ [ observability ] ] observability + + + + + + + + + + + + + we conclude this section by introducing a deterministic observability concept for the output equation @xmath178 corresponding to ( [ stochstatenew ] ) with @xmath179 .
we recall known facts from ( * ? ? ?
* subsection 4.2.2 ) to characterise the importance of certain initial states in the system dynamics since we are in a situation without noise .
we assume to have an unknown initial state @xmath58 in the following observation problem and aim to reconstruct @xmath180 from the observation @xmath181 on the entire time interval @xmath182 .
an initial state @xmath180 is not observable if @xmath183 on @xmath182 , i.e. it can not be reconstructed by the observation .
otherwise , @xmath180 is called observable .
a system a called completely observable if every initial state is observable . in order to determine the observability of a state
, we consider the energy that is caused by the observations of @xmath180 : @xmath184 where we used that @xmath185 and set @xmath186 .
the observability gramian @xmath8 exists due to the asymptotic stability of @xmath12 and is the unique solution to @xmath187 the above relation is obtained by replacing @xmath12 and @xmath188 in ( [ infinitreachgram ] ) and ( [ lyapeqreach ] ) by @xmath189 and @xmath190 , respectively . from ( [ observenergy ] )
we see that @xmath180 is unobservable if and only if @xmath191 .
hence , the system is completely observable if and only if @xmath192 . besides the unobservable states we aim to remove the difficult to observe states from the system in order to obtain an accurate rom . the difficult to observe states
are those producing only little observation energy , i.e. the corresponding observations @xmath181 are close to zero in the @xmath193 sense . using ( [ observenergy ] ) again , the difficult to observe states are contained in the eigenspaces spanned by the eigenvectors of @xmath8 corresponding to the small eigenvalues . before considering balanced truncation ( bt ) and singular perturbation approximation ( spa ) we summarise the general theory for balancing and how to find a balancing transformation . states that are difficult to reach have large components in the span of the eigenvectors corresponding to small eigenvalues of the reachability gramian @xmath7 , cf .
( [ interlambad ] ) .
similarly , states that are difficult to observe are the ones that have large components in the span of eigenvectors corresponding to small eigenvalues of the observability gramian @xmath8 , see ( [ observenergy ] ) . hence in order to produce accurate roms one eliminates states that are both difficult to reach and difficult to observe . to this end
we need to find a basis in which the dominant reachable and observable states are the same , which is done by a simultaneous transformation of the gramians .
let @xmath194 be a nonsingular matrix .
transforming the states using @xmath195 the system ( [ linsysafterdisintr2 ] ) becomes @xmath196 where @xmath197 , @xmath198 , @xmath199 .
the input - output map remains the same , only the state , input and output matrices are transformed .
@xmath7 and @xmath8 , the reachability and observability gramians of the original systems which satisfy ( [ lyapeqreach ] ) and ( [ observlyap ] ) can be transformed into reachability and observability gramians of the transformed system @xmath200 and @xmath201 ( by multiplying ( [ lyapeqreach ] ) with @xmath202 from the left and @xmath203 from the right and ( [ observlyap ] ) with @xmath204 from the left and @xmath205 from the right ) .
the hankel singular values ( hsvs ) of @xmath206 , where @xmath207 for the original and transformed system are the same .
the above transformation is a balancing transformation if the transformed gramians are equal to each other and diagonal .
such a transformation always exists if @xmath208 and can be obtained by choosing @xmath209 where @xmath210 are the hsvs .
@xmath211 , @xmath212 , @xmath213 and @xmath214 are computed as follows .
let @xmath215 , @xmath216 be square root factorisations of @xmath7 and @xmath8 , then an svd of @xmath217 gives the required matrices . with this transformation @xmath218 .
below , let @xmath202 be the balancing transformation as stated above , then we partition the coefficients of the balanced realisation as follows:@xmath219},\;\;\ ; t{b } = { \left[\begin{array}{c}}{b}_1\\ { b}_2{\end{array}\right]},\;\;\ ; { ct^{-1 } } = { \left[\begin{array}{cc}}{c}_1 & { c}_2{\end{array}\right ] } , \end{aligned}\ ] ] where @xmath220 etc . furthermore , setting @xmath221}$ ] , where @xmath222 , we obtain the transformed partitioned system @xmath223}&={\left[\begin{array}{cc}}{a}_{11}&{a}_{12}\\ { a}_{21}&{a}_{22}{\end{array}\right]}{\left[\begin{array}{cc}}{x}_1(t)\\ { x}_2(t){\end{array}\right]}dt+ { \left[\begin{array}{c}}{b}_1\\ { b}_2{\end{array}\right]}dm(t),\label{balrelpart}\\ y(t)&= { \left[\begin{array}{cc}}{c}_1 & { c}_2{\end{array}\right]}{\left[\begin{array}{cc}}{x}_1(t)\\ { x}_2(t){\end{array}\right]}.\label{balrelpartout}\end{aligned}\ ] ] in this system , the difficult to reach and observe states are represented by @xmath224 , which correspond to the smallest hsvs @xmath225 , but of course @xmath226 has to be chosen such that the neglected hsvs are small ( @xmath227 ) .
we discuss two methods ( bt and spa ) to neglect @xmath224 leading to a reduced system of the form @xmath228 where @xmath229 , @xmath230 and @xmath231 ( @xmath232 ) .
[ [ balanced - truncation ] ] balanced truncation + + + + + + + + + + + + + + + + + + + for bt the second row in ( [ balrelpart ] ) is truncated and the remaining @xmath224 components in the first row and in ( [ balrelpartout ] ) are set to zero .
this leads to reduced coefficients @xmath233 which is similar to the deterministic case .
the next lemma states that bt preserves asymptotic stability , which is known from the deterministic case , see ( * ? ? ?
* theorem 7.9 ) .
[ btpresstab ] let the gramians @xmath7 and @xmath8 be positive definite and @xmath234 , then @xmath235 for @xmath236 , i.e. @xmath237 and @xmath238 are asymptotically stable . the above lemma is vital for the error bound analysis in section [ btandspa ] .
[ [ singular - perturbation - approximation ] ] singular perturbation approximation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + instead of setting @xmath239 , one assumes @xmath240 .
this idea originates from the deterministic case , where it can be observed that @xmath224 are the fast variables meaning that they are in a steady state after a short time . in our framework , the classical derivative of @xmath224 does not exist but we proceed with setting @xmath241 in ( [ balrelpart ] ) .
this yields an algebraic constraint @xmath242}{\left[\begin{array}{cc}}{x}_1(s)\\ { x}_2(s){\end{array}\right]}ds+{b}_2 m(t)=:r(t),\end{aligned}\ ] ] where we assumed zero initial conditions .
applying ito s product formula ( [ profriot ] ) to every summand of @xmath243 ( @xmath244 is the @xmath76th component of @xmath245 ) yields @xmath246_t .
\end{aligned}\ ] ] inserting the differential of @xmath245 and exploiting that the expectation of the ito integrals is zero , gives @xmath247=\mathbb e \left[\int_0^t
a^t(s ) r(s)ds\right]+\mathbb e\left[\int_0^t r^t(s ) a(s ) ds\right ] + \mathbb e \sum_{i=1}^{n - r}[r_i , r_i]_t , \end{aligned}\ ] ] where we set @xmath248 .
setting @xmath249 gives @xmath250_t$ ] .
since @xmath244 and @xmath251 have the same martingale parts and the same jumps , their compensator processes coincide by ( [ decomqucov ] ) and hence @xmath252_t=\mathbb e \sum_{i=1}^{n - r}[e_i^t b_2 m , e_i^t b_2 m]_t .
\end{aligned}\ ] ] applying ito s product formula to @xmath253 and taking the expectation , we have @xmath254_t=\mathbb e ( b_2 m)^t b_2 m \end{aligned}\ ] ] which implies @xmath255 @xmath103-a.s . using this simplification in ( [ algbraicconst ] ) yields @xmath256 which is well - defined by lemma [ btpresstab ] and which we use in the first row of ( [ balrelpart ] ) and in ( [ balrelpartout ] ) .
this leads to reduced order coefficients @xmath257 this reduced model is different to the deterministic case , that requires @xmath258 with an additional term in the output equation which does not depend on the state , see ( * ? ? ?
* section 2 ) . in the deterministic case ,
the rom is balanced @xcite , which is not true here due to the modification .
like in the deterministic case , the observability gramian is given by @xmath259 .
this property is obtained by multiplying ( [ partobeq ] ) with @xmath260 ( @xmath261 is the matrix of the balanced system ) from the left and with @xmath262 from the right and then evaluating the left upper block of the resulting equation , see also ( [ bzwzw ] ) .
the following example shows that the reduced order reachability is @xmath263 is not equal to @xmath264 in general which is different from the deterministic case .
let @xmath50 be a standard wiener process , then @xmath114 and set @xmath265 @xmath12 is asymptotically stable and the system is balanced since @xmath266 .
we fix the reduced order dimension to @xmath267 and compute the reduced order coefficients by spa in ( [ coefficientspa ] ) .
we know that @xmath268 but the reachability gramian is up to the digits shown @xmath269 which we computed numerically .
this implies that the hsvs are not a subset of the original ones anymore . here , they are @xmath270 and @xmath271 .
we conclude this section by a stability result from @xcite .
[ spapresstab ] let the gramians @xmath7 and @xmath8 be positive definite and @xmath234 , then @xmath272 for @xmath273 with @xmath274
. the inverses in lemma [ spapresstab ] exist because of the asymptotic stability of the matrices @xmath237 and @xmath238 , see lemma [ btpresstab ] .
before we specify the error bounds for bt and spa , we provide a general error bound comparing the outputs of ( [ stochsysnew ] ) and ( [ generalreducedsys ] ) with asymptotically stable matrices @xmath12 , @xmath275 and initial conditions @xmath54 , @xmath276 . these outputs are then given by ornstein - uhlenbeck processes @xmath277 as mentioned in @xcite , see also @xcite
. using these representations and cauchy s inequality as well as ito s isometry ( see @xcite ) , we obtain @xmath278 where @xmath279 is the covariance matrix of @xmath50 . substitution and taking limits yields @xmath280 using the definition of the frobenius norm and the linearity of the trace , we have @xmath281}\mathbb e \left\|y(t)- y_r(t)\right\|_{\mathbb r^p}\leq \left({\operatorname{tr}}\left(c p c^t\right)+{\operatorname{tr}}\left(c_r p_r c_r^t\right)-2{\operatorname{tr}}\left(c p_g c_r^t\right)\right)^{\frac{1}{2}},\end{aligned}\ ] ] where @xmath282 is the reachability gramian of the original system satisfying ( [ lyapeqreach ] ) , @xmath283 the one of the reduced system satisfying @xmath284 and @xmath285 is the solution to @xmath286 equation ( [ gemischtegram ] ) is a consequence of proposition [ prodeinfachh ] with @xmath287 and @xmath288 and the fact that @xmath289 if @xmath290 due to the asymptotic stability of @xmath12 and @xmath275 .
the matrices @xmath291 and @xmath292 in the bound in ( [ computeerrorbound ] ) are all well - defined because @xmath12 and @xmath275 are asymptotically stable . the error bound in ( [ computeerrorbound ] ) holds for both bt and spa since both approaches preserve asymptotic stability , see lemmas [ btpresstab ] and [ spapresstab ] . for both bt and spa the representation in ( [ computeerrorbound ] )
can be used for practical computations of the error bound .
the gramian @xmath7 is already available since it is required in the balancing procedure .
the reduced model gramian @xmath263 is computationally cheap because it is low dimensional assuming that we fix a small rom dimension .
the same is true for @xmath292 since it has only a few columns which makes the solution to ( [ gemischtegram ] ) easily accessible .
since the error bound ( [ computeerrorbound ] ) is computationally cheap , it can be computed for several rom dimensions and hence be used to find a suitable @xmath226 . in the next two theorems
we specify the general error bound in ( [ computeerrorbound ] ) for both bt and spa and represent it in terms of the truncated hsvs @xmath225 of the system . using the balanced realisation ( [ balancedrels ] ) of the original system with @xmath293 and its corresponding partition , we have @xmath294}\hspace{-0.1cm}{\left[\begin{array}{cc}}\sigma_1\hspace{-0.2cm}{}&\\ & \sigma_2{\end{array}\right]}\hspace{-0.05cm}+\hspace{-0.05 cm } { \left[\begin{array}{cc}}\sigma_1\hspace{-0.2cm}{}&\\ & \sigma_2{\end{array}\right]}\hspace{-0.1 cm } { \left[\begin{array}{cc}}a^t_{11}&a^t_{21}\\ a^t_{12 } & a^t_{22}{\end{array}\right]}\hspace{-0.05cm}= & \hspace{-0.05 cm } -{\left[\begin{array}{cc}}\tilde { b}_1\tilde b_1^t&\tilde { b}_1\tilde b_2^t \\ \tilde { b}_2\tilde b_1^t & \tilde { b}_2\tilde b_2^t{\end{array}\right]}\\ \label{partobeq } { \left[\begin{array}{cc}}{a}^t_{11}&{a}^t_{21}\\ { a}^t_{12}&{a}^t_{22}{\end{array}\right]}\hspace{-0.1cm}{\left[\begin{array}{cc}}\sigma_1\hspace{-0.2cm}{}&\\ & \sigma_2{\end{array}\right]}\hspace{-0.05cm}+\hspace{-0.05 cm } { \left[\begin{array}{cc}}\sigma_1\hspace{-0.2cm}{}&\\ & \sigma_2{\end{array}\right]}\hspace{-0.1 cm } { \left[\begin{array}{cc}}a_{11}&a_{12}\\ a_{21 } & a_{22}{\end{array}\right]}\hspace{-0.05cm}= & \hspace{-0.05 cm } -{\left[\begin{array}{cc}}c_1^tc_1&c_1^tc_2 \\ c^t_2c_1 & c^t_2c_2{\end{array}\right]}\end{aligned}\ ] ] where @xmath295}={\left[\begin{array}{c}}{b}_1\mathcal{q}_m^{\frac{1}{2}}\\ b_2 \mathcal{q}_m^{\frac{1}{2 } } { \end{array}\right]}$ ] , @xmath296 and @xmath297 . [ bterrorbound ]
let @xmath298 be the output of the reduced order system obtained by bt , then under the assumptions of lemma [ btpresstab ] , we have@xmath299}\mathbb e \left\|y(t)- y_{bt}(t)\right\|_{\mathbb r^p}\leq \left({\operatorname{tr}}(\sigma_2 ( b_2 \mathcal{q}_m b_2^t+2 p_{g , 2 } a_{21}^t))\right)^{\frac{1}{2}},\end{aligned}\ ] ] where @xmath300 are the last @xmath301 rows of @xmath302 with @xmath202 being the balancing transformation . evaluating the left and right upper block of ( [ partobeq ] ) yields@xmath303 from ( [ computeerrorbound ] ) the error bound has the form @xmath304 since @xmath305 . using the balancing transformation @xmath202 and the partition of @xmath306 in ( [ balancedrels ] )
, we obtain @xmath307 .
now , the left upper block of ( [ partreacheq ] ) is @xmath308 such that @xmath309 . using the partitions of @xmath310 and @xmath311}$ ]
, we obtain @xmath312 .
inserting these results into ( [ eq : epserror ] ) gives @xmath313 using @xmath314 and substituting ( [ c1c2eqbt2 ] ) yields @xmath315 multiplying ( [ gemischtegram ] ) with the balancing transformation from the left and using the partitions of @xmath316 and @xmath317 from ( [ balancedrels ] ) yields @xmath318}{\left[\begin{array}{c } } p_{g , 1 } \\
p_{g , 2 } { \end{array}\right]}+{\left[\begin{array}{c } } p_{g , 1 } \\
p_{g , 2 } { \end{array}\right]}a_{11}^t = -{\left[\begin{array}{c}}{b}_1\\ { b}_2{\end{array}\right]}\mathcal q_m b_1^t .
\end{aligned}\ ] ] with the first row of this equation , @xmath319 we have@xmath320 and substituting ( [ firstc1c2bt2jk ] ) , we obtain @xmath321 so that @xmath322 inserting this result into ( [ insertforebbt2fh ] ) gives@xmath323 with ( [ firstc1c2bt2jk ] ) and ( [ blaaa ] ) , and the properties of the trace function we obtain @xmath324 similarly @xmath325 can be shown using the right lower blocks of ( [ partreacheq ] ) and ( [ partobeq ] ) .
hence , @xmath326 which gives the result .
[ spaerrorbound ] let @xmath327 be the output of the reduced order system obtained by spa , then under the assumptions of lemma [ spapresstab ] , we have@xmath299}\mathbb e \left\|y(t)- y_{spa}(t)\right\|_{\mathbb r^p}\leq \left({\operatorname{tr}}(\sigma_2 ( b_2 \mathcal{q}_m b_2^t-2(a_{22}p_{g , 2}+a_{21 } p_{g , 1})(a_{22}^{-1 } a_{21})^t))\right)^{\frac{1}{2}},\end{aligned}\ ] ] where @xmath328 are the first @xmath226 and @xmath300 the last @xmath301 rows of @xmath329 with @xmath202 being the balancing transformation .
let @xmath330}$ ] , then , since @xmath331 are invertible by lemma [ btpresstab ] , its inverse is given in block form @xmath332},\end{aligned}\ ] ] where @xmath333 .
if we multiply ( [ partobeq ] ) with @xmath260 from the left hand side and select the left and right upper block of this equation , we obtain @xmath334 where @xmath335 and thus @xmath336 furthermore , multiplying ( [ partobeq ] ) with @xmath260 from the left and with @xmath262 from the right , the resulting left upper block of the equation is @xmath337 and thus@xmath338 we define @xmath339 which is the error bound for spa . from the proof of theorem [ bterrorbound ] we know that the following holds@xmath340 by ( [ bzwzw ] ) and the definition of the reachability equation of the rom , we have @xmath341 this leads to @xmath342 we multiply ( [ gemischtegram ] ) with the balancing transformation @xmath202 from the left ( here @xmath343 ) and use the partitions of @xmath316 , @xmath317 from ( [ balancedrels ] ) and the partition of @xmath311}$ ] .
thus , @xmath344}{\left[\begin{array}{c } } p_{g , 1 } \\
p_{g , 2 } { \end{array}\right]}+{\left[\begin{array}{c } } p_{g , 1 } \\
p_{g , 2 } { \end{array}\right]}\bar a^t = -{\left[\begin{array}{c}}{b}_1\\ { b}_2{\end{array}\right]}\mathcal q_m b_1^t . \end{aligned}\ ] ] we obtain @xmath345 using the partition of @xmath310 in ( [ balancedrels ] ) . with ( [ barcczwei ] )
we obtain @xmath346 inserting the upper block of ( [ partmixedeqspa ] ) leads to @xmath347 using ( [ barcceins ] ) and the properties of the trace function we have @xmath348 consequently , @xmath349 holds and hence @xmath350 which provides the required result .
the error bound representations in theorems [ bterrorbound ] and [ spaerrorbound ] depend on the @xmath301 smallest hsvs @xmath225 .
if the corresponding truncated components are unimportant , i.e. they are difficult to reach and observe , then the values @xmath225 are small and consequently the error bound is small . hence , the rom is of good quality . the error bounds in theorems [ bterrorbound ] and [ spaerrorbound ] can be used to find a suitable reduced order dimension @xmath226 .
small hsvs @xmath225 for fixed @xmath226 would guarantee a small error .
note that , if @xmath114 , as for example in the standard wiener case , then the error bound in theorem [ bterrorbound ] coincides with the @xmath351-error bound in the deterministic case when using a normalised control , see ( * ? ? ? * lemma 7.13 ) .
the next section provides a particular sde to which we will apply the theory developed in this section .
in this section , we deal with a setting that covers the spde with its output in ( [ inrobspeq])-([introbspoutnew ] ) , a damped wave equation with additive noise which can formally be interpreted as @xmath352 with @xmath353 ( @xmath236 ) , @xmath37 and the @xmath354th component of the control @xmath355 , @xmath356 . here , @xmath357 are the components of an @xmath358-valued lvy processes @xmath50 that is square integrable and has mean zero .
this is in contrast to the setting in @xcite where ( [ nichtlinsde ] ) with multiplicative lvy noise was considered , e.g. @xmath359 linear bounded operators and @xmath176 an @xmath360-dimensional stochastic control , @xmath361 and @xmath362 uncorrelated scalar lvy processes .
for the stability analysis of the uncontrolled equation ( [ nichtlinsde ] ) with wiener noise ( @xmath363 ) we refer to @xcite .
since lvy noise is no feasible control in the framework in @xcite , this setting requires further analysis .
we transform damped wave equation with additive noise into a first order spde and define the corresponding solution in section [ setting ] , following the approach in @xcite . in section [ subsecspeltralgal ] ,
we explain how the resulting first order spde can be approximated by a spectral galerkin scheme .
we refer to @xcite , where similar techniques were applied .
let @xmath364 be square integrable lvy processes with zero mean that takes values in @xmath30 .
moreover , @xmath50 is defined on a complete probability space @xmath365 , is right - continuous and that @xmath366 contains all @xmath103 null sets . ]
it is adapted to the filtration @xmath367 and its increments @xmath368 are independent of @xmath369 for @xmath370 .
let @xmath371 be a self adjoint and positive definite operator on a separable hilbert space @xmath372 and let @xmath373 be an orthonormal basis of eigenvectors of @xmath374 for @xmath372 , @xmath375 where @xmath376 are the corresponding eigenvalues .
we denote the well - defined square root of @xmath374 by @xmath377 . @xmath378
equipped with the inner product @xmath379 represents a separable hilbert space as well .
the ( symbolic ) second order spde we consider is given by @xmath380 with initial conditions @xmath381 , @xmath382 , @xmath37 and output @xmath383 we assume @xmath384 and @xmath385 . using the separable hilbert space @xmath386 with the inner product @xmath387 we transform this second order system into a first order system following the approach in @xcite .
the system ( [ nichtlinsde2])-([eq : output ] ) can be expressed as : @xmath388 where @xmath389\;\;\;\text{and}\;\;\;\mathcal b_k=\left[\begin{matrix } 0 \\ -\tilde { \mathcal b}_k \end{matrix}\right]\in h.\end{aligned}\ ] ] so far , we only worked with symbolic equation , since the classical derivative of a lvy process does not exist in general .
the next lemma from @xcite provides a stability result and is vital to define a cdlg mild solution of ( [ firstordertransf ] ) .
[ lemwaveconexsemigr ] for every @xmath37 the linear operator @xmath390 with domain @xmath391 generates an exponential stable contraction semigroup @xmath392 with @xmath393 we use this result to define the solution to ( [ firstordertransf ] ) . an @xmath367-adapted cdlg process @xmath394 is called mild solution to ( [ firstordertransf ] ) if for all @xmath57 @xmath395 we refer to @xcite for the definition of the ito integral in ( [ mildsolalternative ] ) .
for the finite dimensional case , the definition of an ito integral with respect to lvy processes can be found in @xcite and ito integrals with respect to martingales are defined in @xcite .
if we set @xmath396)$ ] , @xmath397)$ ] , @xmath398 , @xmath399)$ ] and the output operator @xmath400 with @xmath401 @xmath402)$ ] in system ( [ nichtlinsde2])-([eq : output ] ) then we obtain the system in ( [ inrobspeq])-([introbspoutnew ] ) , which is therefore well - defined in the mild sense ( [ mildsolalternative ] ) .
we study a spectral galerkin scheme to approximate the mild solution of ( [ firstordertransf ] ) with output ( [ firstordertransfout ] ) , similar to the approach in @xcite ( mainly for spdes with wiener noise ) .
this approximation is based on a particular choice of an orthonormal basis @xmath403 of @xmath404 , given by @xmath405 where @xmath373 and @xmath406 are defined in ( [ eigvaleq ] ) , see @xcite . in ( [ inrobspeq])-([introbspoutnew ] ) , we have @xmath407 on @xmath408 $ ] . in this case @xmath409 and @xmath410 for @xmath411 . to approximate the @xmath404-valued process @xmath412 in ( [ firstordertransf ] )
, we construct a sequence @xmath413 of finite dimensional adapted cdlg processes with values in @xmath414 , defined by @xmath415 where we set * @xmath416 for all @xmath417 , * @xmath418 for @xmath419 , * @xmath420 . for the mild solution to ( [ approxsde ] ) ,
let @xmath421 be a @xmath422-semigroup on @xmath423 given by @xmath424 for all @xmath425 .
it is generated by @xmath426 such that the mild solution of equation ( [ approxsde ] ) is @xmath427 since @xmath426 is bounded , the @xmath422-semigroup on @xmath423 is represented by @xmath428 , @xmath57 .
we formulate the main result of this section , which uses ideas from @xcite and is proved in appendix [ sec : app2 ] .
[ th : mildsol ] the mild solution @xmath429 of equation ( [ approxsde ] ) approximates the mild solution @xmath412 of equation ( [ firstordertransf ] ) , i.e. @xmath430 for @xmath431 and @xmath57 .
this implies the convergence of the corresponding outputs @xmath432 . in the following , we make use of the property that the mild and the strong solution of ( [ approxsde ] )
coincide , since we are in finite dimensions .
we write the output @xmath433 of the galerkin system as an expression depending on the fourier coefficients of the galerkin solution @xmath429 .
the coefficients of @xmath433 are @xmath434 for @xmath435 , where @xmath436 is the @xmath437-th unit vector in @xmath438 .
we set @xmath439 and obtain @xmath440 .
the components @xmath441 of @xmath72 satisfy @xmath442 using the fourier series representation of @xmath429 , we obtain @xmath443 hence , the vector of fourier coefficients @xmath63 is given by @xmath444 where @xmath445 with @xmath446 ( @xmath447 ) , and @xmath448 the eigenvalues of @xmath449 , and @xmath450 for @xmath451 .
we will often make use of the compact form of the sde in ( [ galerkwavenowspec ] ) which is @xmath452 where @xmath29 and @xmath453 $ ] . applying the spectral galerkin method to the system ( [ inrobspeq])-([introbspoutnew ] ) the matrices of the semi - discretised system ( [ galerkwavenowspec ] )
are given by @xmath454 with @xmath455 , @xmath456 with @xmath457)}},\ ] ] and @xmath458 $ ] with @xmath459 & 0 \end{matrix}\right)^t,\\ \mathcal c h_{2 \ell}&=\left(\begin{matrix } 0 & \frac{1}{\sqrt{2\pi}\ell\epsilon}\left [ \cos \left(\ell \left(\frac{\pi}{2 } -\epsilon \right)\right)-\cos\left(\ell\left(\frac { \pi } { 2 } + \epsilon \right)\right)\right]\end{matrix}\right)^t,\end{aligned}\ ] ] where we assume @xmath10 to be even and @xmath460 .
in figure [ fig : wavetime ] we plot the numerical solution to the stochastic damped wave equation for @xmath36 $ ] and in the time interval @xmath461 $ ] where we set @xmath462 , @xmath463 and @xmath464 ( e.g. @xmath465 stochastic inputs ) .
the weighting functions for the two inputs are @xmath466 and @xmath467 .
the noise processes are @xmath468 and @xmath469 is a compound poisson process , where @xmath470 is a poisson process with parameter equal to @xmath71 , @xmath471 are independent uniformly distributed jumps and @xmath472 is a standard wiener process . @xmath473 and @xmath474 are independent .
the plot in figure [ fig : wavetime ] shows a particular realisation of the solution to ( [ inrobspeq ] ) at @xmath475 specific times @xmath476 .
we see that the string moves up and down as expected due to the nonzero ( stochastic ) input .
we observe that the third snapshot is taken after a jump occured in stochastic process . the corresponding output ,
namely both the position and the velocity in the middle of the string , is shown in figure [ fig : position_velocity ] . in the plot for the velocity
the noise generated by the lvy process can be seen .
the trajectory of the velocity is impacted by lvy noise with jumps , where the velocity ( e.g. the impact by wind ) is randomly increased or reduced . )
-([introbspoutnew ] ) in the phase plane . ]
the trajectory for the position of the cable in figure [ fig : position_velocity ] is smoother as it is the integral of the velocity .
finally , figure [ fig : paths ] shows the velocity versus the position of the string , for the same sample path , in a phase portrait .
the four jumps are clearly visible .
we consider the spectral galerkin discretisation of the second order damped wave equation which we discussed in detail in section [ sec : wave ] , and in particular , the example in ( [ inrobspeq])-([introbspoutnew ] ) with two stochastic inputs and two outputs , namely position and velocity of the middle of the string .
we set @xmath477 and choose the weighting functions @xmath478 and the noise processes @xmath28 ( @xmath236 ) as in figure [ fig : wavetime ] .
we fix the state dimension to @xmath479 and reduce the galerkin solution by bt and spa . for computing the trajectories of the sde we use the euler - maruyama method ( see , e.g. @xcite ) . and velocity @xmath480 with @xmath481 . ]
figures [ fig : btr6 ] and [ fig : btr24 ] show the logarithmic errors for the position @xmath482 and the velocity @xmath480 of the middle of the string , if mor by bt is applied to the wave equation with stochastic inputs when reduced models of dimension @xmath475 and @xmath483 , respectively , are computed . and velocity @xmath480 with @xmath484 . ]
the first two plots in each of the figures show the logarithmic mean error for both the position and the velocity .
one observation is that the position is generally more accurate than the velocity ( about one order of magnitude here ) , since the trajectories are smoother .
moreover , comparing the expected values of the errors of the reduced model of dimension @xmath481 ( first two plots in figure [ fig : btr6 ] ) with the one of dimension @xmath484 ( first two plots in figure [ fig : btr24 ] ) it can be seen that the latter ones are more accurate ( an improvement of about one order of magnitude ) as one would expect .
the last two plots in figures [ fig : btr6 ] and [ fig : btr24 ] show the logarithmic errors for position and velocity for one particular trajectory , which is the same as the one for the sample we considered in section [ sec : wave ] . and velocity @xmath480 with @xmath481 . ]
figures [ fig : spar6 ] and [ fig : spar24 ] show the logarithmic errors for the position @xmath482 and the velocity @xmath480 of the middle of the string , if mor by spa is applied to the wave equation with stochastic inputs when reduced models of dimension @xmath475 and @xmath483 , respectively , are computed .
again , the first two plots show the mean errors while the last two plots show the errors in particular trajectories . and velocity @xmath480 with @xmath484 . ]
we observe that the error in the position is smaller than the error in the velocity , and , the error is smaller if a larger dimension of the reduced order model is used .
finally , we compare the error bounds for bt ( see theorem [ bterrorbound ] ) and spa ( see theorem [ spaerrorbound ] ) with the worst case mean errors , that is @xmath485 } \mathbb e\left\|y(t)-y_r(t)\right\|_{\mathbb{r}^p } \end{aligned}\ ] ] for both methods in table [ tab : errorbd ] , where @xmath486 is the full output of the original model and @xmath487 the rom output .
.error and error bounds for both bt and spa and several dimensions of the reduced order model ( rom ) . [ cols="^,^,^,^,^,^ " , ] first , as expected both mean errors and error bounds are getting smaller the larger the size of the rom .
moreover , both error bounds are rather tight and close to the actual error of the rom , e.g. the bounds , which are worst case bounds also provide a good prediction of the true time domain error . we also note that bt performs better than spa , both in actually computed mean errors as well as in terms of the error bounds .
we have presented theory for balancing related model order reduction ( mor ) applied to linear stochastic differential equations ( sdes ) with additive lvy noise .
in particular we extended the concepts of reachability and observability to stochastic systems and formulated a new reachability gramian .
we then showed how balancing related mor which is well known for deterministic systems can be extended to sdes with additive lvy noise , e.g. leads to the solution of a lyapunov equation ( with a slightly different right hand side ) .
we proved a general error bound for reduced ( asymptotically stable ) systems in this setting and then gave specific bounds for balanced truncation ( bt ) and singular perturbation approximation ( spa ) which depended on the neglected ( small ) hankel singular values of the linear system .
we finally applied our theory to a second order damped wave equation , discretised using a spectral galerkin method , and controlled by lvy noise .
the numerical results showed that mor can be applied successfully and that errors for both bt and spa are small , and the error bounds tight .
let all stochastic processes appearing in this section be defined on a filtered probability space @xmath488 shall be right continuous and complete . ] .
we denote the set of all cdlg square integrable @xmath489-valued martingales with respect to @xmath367 by @xmath490 .
let @xmath491 be scalar semimartingales .
we set @xmath492 with @xmath493 for @xmath236 .
then the ito product formula @xmath494_t\end{aligned}\ ] ] for @xmath57 holds , see @xcite or @xcite for the special case of lvy - type integrals . by (
* theorem 4.52 ) , the compensator process @xmath495 $ ] is given by @xmath496_t=\left\langle m_1^c , m_2^c\right\rangle_t+\sum_{0\leq s\leq t } \delta z_1(s ) \delta z_2(s)\end{aligned}\ ] ] for @xmath57 , where @xmath497 , @xmath498 are the continuous martingale parts of @xmath499 and @xmath500 ( cf .
* theorem 4.18 ) ) .
the process @xmath501 is a uniquely defined angle bracket process that ensures that @xmath502 is an @xmath367- martingale , see ( * ? ? ?
* proposition 17.2 ) . as a simple consequence of ( [ profriot ] )
, we have : [ iotprodformelmatpro ] let @xmath211 be an @xmath503-valued and @xmath212 be an @xmath130-valued semimartingale , then we have@xmath504_t\right)_{{i=1 , \ldots , d}\atop { j=1 , \ldots , n}}\end{aligned}\ ] ] for all @xmath57 .
considering the stochastic differential of the @xmath505-th component of the matrix - valued process @xmath506 , @xmath57 , and using ( [ profriot ] ) gives the result , see also @xcite .
using @xmath507 for @xmath508 , we obtain @xmath509 ito s isometry ( see e.g. @xcite ) yields that the right hand side can be bounded by @xmath510 $ ] . since @xmath511 is a contraction semigroup , we have@xmath512 by the representation @xmath513 ( @xmath425 ) and lebesgue s theorem , the bound in ( [ inconcon2 ] ) tends to zero for @xmath431 .
for @xmath514 we get @xmath515 which tends to zero for @xmath431 and hence @xmath516\rightarrow 0\end{aligned}\ ] ] for @xmath431 by lebesgue s theorem . | when solving linear stochastic differential equations numerically , usually a high order spatial discretisation is used .
balanced truncation ( bt ) and singular perturbation approximation ( spa ) are well - known projection techniques in the deterministic framework which reduce the order of a control system and hence reduce computational complexity .
this work considers both methods when the control is replaced by a noise term .
we provide theoretical tools such as stochastic concepts for reachability and observability , which are necessary for balancing related model order reduction of linear stochastic differential equations with additive lvy noise .
moreover , we derive error bounds for both bt and spa and provide numerical results for a specific example which support the theory . [ [ keywords ] ] keywords : + + + + + + + + + model order reduction , balanced truncation , singular perturbation approximation , stochastic systems , lvy process , gramians , lyapunov equations .
[ [ ams - subject - classifications ] ] ams subject classifications + + + + + + + + + + + + + + + + + + + + + + + + + + + primary 93a15 , 93b40 , 93e03 , 93e30 , 60j75 .
secondary 93a30 , 15a24 . |
we consider the quadratic optimization problem over the @xmath0 unit ball @xmath3 which is known as an @xmath0-norm trust - region subproblem in nonlinear programming @xcite and @xmath0 grothendieck problem in combinatorial optimization @xcite .
applications of ( qpl1(q ) ) can be also found in compressed sensing where @xmath4 is introduced to approximate @xmath5 , the number of nonzero elements of @xmath6 .
if @xmath7 is negative or positive semidefinite , ( qpl1(q ) ) is trivial to solve , see @xcite .
generally , ( qpl1(q ) ) is np - hard , even when the off - diagonal elements of @xmath7 are all nonnegative , see @xcite . in the same paper , hsia showed that ( qpl1(q ) ) admits an exact nonconvex semidefinite programming ( sdp ) relaxation , which was firstly proposed as an open problem by pinar and teboulle @xcite . very recently
, different sdp relaxations for ( qpl1(q ) ) have been studied in @xcite .
the tightest one is the following doubly nonnegative ( dnn ) relaxation due to bomze et al .
@xcite : @xmath8 where @xmath9 is the vector with all elements equal to @xmath10 , @xmath11 is the set of @xmath12 symmetric matrices , @xmath13 means that @xmath14 is componentwise nonnegative , @xmath15 stands for that @xmath14 is positive semidefinite , @xmath16 is the standard inner product of @xmath17 and @xmath18 , and @xmath19.\ ] ] notice that the set of extreme points of @xmath20 is @xmath21 , where @xmath22 is the @xmath23-th column of the identity matrix @xmath24 .
define @xmath25=[i~-i ] \in \re^{n\times 2n}.\ ] ] then we have @xmath26 consequently , ( qpl1(q ) ) can be equivalently transformed to the following standard quadratic program ( qps ) @xcite : @xmath27 now we can see that @xmath28 exactly corresponds to the well - known doubly nonnegative relaxation of ( qps ) @xcite .
moreover , as mentioned in @xcite , @xmath28 can be also derived by applying the lifting procedure @xcite to the following homogeneous reformulation of ( qps ) : @xmath29 a natural extension of ( qpl1(q ) ) is @xmath30 it is a relaxation of the sparse principal component analysis ( spca ) problem @xcite obtained by replacing the original constraint @xmath31 with ( [ l1:1 ] ) due to the following fact : @xmath32 a well - known sdp relaxation for ( qpl2l1(q ) ) is due to daspremont et al .
@xcite : @xmath33 recently , xia @xcite extended the doubly nonnegative relaxation approach from ( qpl1(q ) ) to ( qpl2l1(q ) ) and obtained the following sdp relaxation : @xmath34 it was proved in @xcite that @xmath35 , where @xmath36 denote the optimal value of problem @xmath37 . unfortunately , this equivalence result is incorrect though it is true that @xmath38 .
a first counterexample will be given in this paper ( see example 2 below ) to show it is possible @xmath39 .
the other extension of ( qpl1(q ) ) is @xmath40 where @xmath41 and @xmath42 .
( qplp ) is known as a special case of the @xmath1 grothendieck problem if the diagonal entries of @xmath7 vanish . according to the survey @xcite
, there is no approximation and hardness results for the @xmath1 grothendieck problem with @xmath43 .
though @xmath44 has an exact nonconvex sdp relaxation similar to that of @xmath45 , the computational complexity of @xmath44 is still unknown @xcite . since the @xmath1 unit balls ( @xmath43 ) are included in the @xmath46 unit ball , a trivial bound for @xmath44 is @xmath47 where @xmath48 is the largest eigenvalue of @xmath7 . as mentioned by nesterov in the sdp handbook @xcite ,
no practical sdp bounds of @xmath44 are in sight for @xmath43 .
recently , bomze @xcite used the hlder inequality @xmath49 to propose the following sdp bound @xmath50 in general , @xmath51 dominates @xmath52 when @xmath53 close to @xmath10 , though lacking a proof . in this paper , based on a new variable - splitting reformulation for the @xmath0-constrained set
, we establish a new sdp relaxation for ( qpl1(q ) ) , which is proved to dominate @xmath28 .
we use a small example to show the improvement could be strict . then we extend the new approach to ( qpl2l1(q ) ) and obtain two new sdp relaxations .
we can not prove the first new sdp bound dominates @xmath54 , though it was demonstrated by examples .
however , under a mild assumption , the second new sdp bound dominates @xmath54 . finally , motivated by the model ( qpl2l1(q ) )
, we establish a new sdp bound for ( qplp(q ) ) and show it is in general tighter than @xmath55 .
the paper is organized as follows . in section 1
, we propose a new variable - splitting reformulation for the @xmath0-constrained set and then a new sdp relaxation for ( qpl1(q ) ) .
we show it improves the state - of - the - art sdp - based bound . in section 2 ,
we extend the new sdp approach to ( qpl2l1(q ) ) and study the obtained two new sdp relaxations . in section 3 , we establish a new sdp relaxation for ( qplp(q ) ) , which improves the existing upper bounds .
conclusions are made in section 4 .
in this section , we establish a new sdp relaxation for ( qpl1(q ) ) based on a new variable - splitting reformulation for the @xmath0-constrained set . for any @xmath56 ,
let @xmath57 then we have @xmath58 now we obtain a new variable - splitting reformulation of the @xmath0-constrained set : @xmath59 it follows that @xmath60 applying the lifting procedure @xcite , we obtain the following new doubly nonnegative relaxation of ( qpl1(q ) ) @xmath61 we first compare the qualities of @xmath62 and @xmath63 .
[ thm:1 ] @xmath64 .
according to the definitions , we have @xmath65 and @xmath66 .
it is sufficient to prove the first inequality .
since @xmath67 is a feasible solution of @xmath68 , we have @xmath69 suppose @xmath70 .
let @xmath71 be an optimal solution of ( @xmath72 ) .
since @xmath73 , we have @xmath74 and therefore @xmath75 consequently , @xmath76 .
similarly , we can show @xmath77 .
now we assume @xmath78 .
there is a vector @xmath79 such that @xmath80 and @xmath81 .
that is , @xmath82 .
it follows that @xmath83 .
let @xmath71 be an optimal solution of ( @xmath72 )
. then @xmath84 .
moreover , since @xmath85 , we have @xmath86 .
we conclude that @xmath87 if this is not true , then @xmath88 .
define @xmath89 it is trivial to see that @xmath90 is also feasible to ( @xmath72 ) .
moreover , we have @xmath91 which contradicts the fact that @xmath71 is a maximizer of ( @xmath72 ) .
according to the equality ( [ eq1 ] ) , @xmath71 is also a feasible solution of ( @xmath92 ) .
consequently , @xmath93 .
the proof is complete .
the following small example illustrates that @xmath94 could strictly improve @xmath95 .
[ exam1 ] consider the following instance of dimension @xmath96 @xmath97\ ] ] we modeled this instance by cvx 1.2 ( @xcite ) and solved it by sedumi ( @xcite ) within cvx .
then we obtained that @xmath98 finally , we show that there are some cases for which @xmath68 has no improvement .
this `` negative '' result is also interesting in the sense that in case we solve @xmath99 , we can fix @xmath100 ( @xmath101 ) at zeros in advance . [ thm:2 ] suppose @xmath102 . @xmath103 .
let @xmath71 be an optimal solution of @xmath99 .
suppose there is an index @xmath104 such that @xmath105 .
let @xmath106 and define a symmetric matrix @xmath107 where @xmath108 and all other elements are zeros
. then @xmath109 it follows that @xmath110 then , @xmath111 is also an optimal solution of @xmath99 .
repeat the above procedure until we obtain an optimal solution of @xmath99 , denoted by @xmath112 , satisfying @xmath113 for @xmath101 .
notice that @xmath112 is a feasible solution of @xmath114 .
therefore , we have @xmath115 . combining this inequality with theorem [ thm:1 ]
, we can complete the proof .
in this section , we extend the above new reformulation approach to ( qpl2l1(q ) ) and obtain two new semidefinite programming relaxations .
similar to the reformulation ( [ x - y:1])-([x - y:4 ] ) , we have @xmath116 it follows that @xmath117 introducing @xmath118 , we obtain the following new sdp relaxation for ( qpl2l1(q ) ) : @xmath119 according to the definition , we trivially have : @xmath120 . [ prop ] @xmath121 .
both @xmath122 and @xmath123 share the same relaxation : @xmath124 let @xmath125 .
we have @xmath126 therefore , @xmath127 can be further relaxed to @xmath128 let @xmath129 be the eigenvalue decomposition of @xmath7 , where @xmath130 and @xmath131 are column - orthogonal . since @xmath132 we can further relax @xmath133 to the following linear programming problem : @xmath134 now it is trivial to verify that @xmath135 the proof is complete .
[ cor ] suppose @xmath136 , then we have @xmath137 we are unable to prove @xmath138 , though we failed to have found an example such that @xmath139 .
moreover , the following example shows that it is possible @xmath140 . as a by - product , we observe @xmath39 from the example , which means that the result @xmath35 ( theorem 3.2 @xcite ) is incorrect . notice that it is true that @xmath38 .
[ exam2 ] consider the same instance of example [ exam1 ] and let @xmath141 .
we modeled this instance by cvx 1.2 ( @xcite ) and solved it by sedumi ( @xcite ) within cvx .
we obtained that @xmath142 thus , in order to theoretically improve @xmath143 , we consider @xmath144 it is trivial to see that @xmath145 however , @xmath146 may be not an upper bound of @xmath147 , which is indicated by the following example .
[ exam3 ] consider the same instance of example [ exam1 ] and let @xmath148 .
we modeled this instance by cvx 1.2 ( @xcite ) and solved it by sedumi ( @xcite ) within cvx .
we obtained that @xmath149 so , we have to identify when @xmath146 is an upper bound of @xmath147 . [ thm:4 ]
suppose @xmath150 we have @xmath151 .
we first notice that the maximum eigenvalue problem @xmath152 is a homogeneous trust - region subproblem and hence has no local - non - global maximizer @xcite . therefore , suppose there is an optimal solution of @xmath147 , denoted by @xmath153 , satisfying @xmath154 , then @xmath153 also globally solves ( e ) , i.e. , @xmath155 consequently , the assumption ( [ as:2 ] ) implies that @xmath156 taking the transformation ( [ xy:1])-([xy:4 ] ) and then applying the lifting approach @xcite , we obtain the sdp relaxation @xmath157 .
the proof is complete .
the assumption ( [ as:2 ] ) is generally not easy to verify .
however , when @xmath7 has a unique maximum eigenvalue , ( [ as:2 ] ) holds if and only if @xmath158 , where @xmath79 is the @xmath46-normalized eigenvector corresponding to the maximum eigenvalue of @xmath7 .
moreover , according to corollary [ cor ] and proposition [ prop ] , the assumption ( [ as:2 ] ) can be replaced by the following easy - to - check sufficient condition @xmath159
in this section , we first propose a new sdp relaxation for ( qplp(q ) ) and then show it improves both @xmath160 ( [ b2 ] ) and @xmath161 ( [ bom ] ) . motivated by the hlder inequality ( [ hol ] ) and the model ( qpl2l1(q ) ) , we obtain the following new relaxation for ( qplp(q ) ) : @xmath162 taking the transformation ( [ xy:1])-([xy:4 ] ) and then applying the lifting approach @xcite , we obtain the following sdp relaxation for @xmath163 , which is very similar to @xmath123 : @xmath164 @xmath165 according to the definitions , the second inequality is trivial . it is sufficient to prove the first inequality .
we first show @xmath166 .
let @xmath167 . since @xmath168 @xmath169 has the following relaxation : @xmath170 let @xmath129 be the eigenvalue decomposition of @xmath7 , where @xmath130 and @xmath131 are column - orthogonal . according to ( [ x:1])-([x:3 ] )
, we can further relax @xmath171 to the following linear programming problem : @xmath172 it is not difficult to verify that @xmath173 now we prove @xmath174 .
notice that @xmath175 where the last inequality follows from theorem [ thm:1 ] .
the proof is complete .
we randomly generated a symmetric matrix @xmath7 of order @xmath176 using the following matlab scripts : .... rand('state',0 ) ; q = rand(n , n ) ; q = ( q+q')/2 ; .... and then compared the qualities of the three upper bounds , @xmath177 , @xmath51 and @xmath52 .
the results were plotted in figure 1 , where the lower bound of qplp(q ) is computed as follows .
solve @xmath169 and obtain the optimal solution @xmath71 .
let @xmath178 be the unit eigenvectors corresponding to the maximum eigenvalues of @xmath179 and @xmath7 , respectively .
then @xmath180 and @xmath181 are two feasible solutions of ( qplp(q ) ) and @xmath182 gives a lower bound of @xmath183 . from figure 1
, we can see that for @xmath43 , though @xmath52 and @xmath51 can not dominate each other , both are strictly improved by @xmath177 .
[ fig ] , @xmath161 and @xmath177 in dependence of @xmath53.,scaledwidth=80.0% ]
the sdp relaxation has been known to generate high quality bounds for nonconvex quadratic optimization problems . in this paper ,
based on a new variable - splitting characterization of the @xmath0 unit ball , we establish a new semidefinite programming ( sdp ) relaxation for the quadratic optimization problem over the @xmath0 unit ball ( qpl1 ) .
we show the new developed sdp bound dominates the state - of - the - art sdp - based upper bound for ( qpl1 ) .
there is an example to show the improvement could be strict .
then we extend the new reformulation approach to the relaxation problem of the sparse principal component analysis ( qpl2l1 ) and obtain two sdp formulations .
examples demonstrate that the first sdp bound is in general tighter than the dnn relaxation for ( qpl2l1 ) .
but we are unable to prove it . under a mild assumption ,
the second sdp bound dominates the dnn relaxation .
finally , we extend our approach to the nonconvex quadratic optimization problem over the @xmath1 ( @xmath2 ) unit ball ( qplp ) and show the new sdp bound dominates two upper bounds in recent literature .
y. nesterov , global quadratic optimization via conic relaxation , in handbook of semidefinite programming , h. wolkowicz , r. saigal and l. vandenberghe , eds . , kluwer academic publishers , boston , 363387 , 2000 | in this paper , by improving the variable - splitting approach , we propose a new semidefinite programming ( sdp ) relaxation for the nonconvex quadratic optimization problem over the @xmath0 unit ball ( qpl1 ) .
it dominates the state - of - the - art sdp - based bound for ( qpl1 ) . as extensions ,
we apply the new approach to the relaxation problem of the sparse principal component analysis and the nonconvex quadratic optimization problem over the @xmath1 ( @xmath2 ) unit ball and then show the dominance of the new relaxation .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
the observation of the new baryon @xmath6 with positive strangeness and minimal quark content @xmath7 @xcite has motivated intense theoretical investigations to clarify its quantum numbers and to understand the under - structures @xcite . through the pentaquark state @xmath6 can be signed to the top of the antidecuplet @xmath8 with isospin @xmath9 , the spin and parity have not been experimentally determined yet and no consensus has ever been reached on the theoretical side @xcite . the discovery has opened a new field of strong interaction and provides a new opportunity for a deeper understanding of low energy qcd especially when multiquarks are involved .
the magnetic moments of the pentaquark states are fundamental parameters as their masses , which have copious information about the underlying quark structures i.e. different substructures result in different predictions , can be used to distinguish the preferred configurations from various theoretical models and deepen our understanding of the underlying dynamics .
furthermore , the magnetic moment of the pentaquark state @xmath6 is an important ingredient in studying the cross sections of the photo- or electro - production , which can be used to determine the fundamental quantum number of the pentaquark state @xmath6 , such as spin and parity @xcite , and may be extracted from experiments eventually in the near future .
there have been several works on the magnetic moments of the pentaquark state @xmath6 @xcite , in this article , we take the point of view that the @xmath10 baryon is diquark - diquark - antiquark ( @xmath1[ud]\bar{s}$ ] ) state with the quantum numbers @xmath11 , @xmath9 , @xmath12 , and study its magnetic moment in the framework of the light - cone qcd sum rules approach , which carries out the operator product expansion near the light - cone @xmath13 instead of the short distance @xmath14 while the nonperturbative matrix elements are parameterized by the light - cone distribution amplitudes which classified according to their twists instead of the vacuum condensates @xcite .
the article is arranged as follows : we derive the light - cone qcd sum rules for the magnetic moment of the pentaquark state @xmath6 in section ii ; in section iii , numerical results ; section iv is reserved for conclusion .
we can study the magnetic moments of the baryons using two - point correlation functions in an external electromagnetic field , with vacuum susceptibilities introduced as parameters for nonperturbative propagation in the external field , i.e. the qcd sum rules in the external field @xcite . as nonperturbative vacuum properties
, the susceptibilities can be introduced for both small and large momentum transfers in the external fields .
the alternative way is the light - cone qcd sum rules approach , which was firstly used to calculate the magnetic moments of nucleons in ref.@xcite . for more discussions about the magnetic moments of the baryons in the framework of the light - cone qcd sum rules approach
, one can consult ref .
@xcite . in the following ,
we write down the two - point correlation function @xmath15 in the framework of the light - cone qcd sum rules approach @xcite , @xmath16 \left[u_c^t(x ) c \gamma_5 d_e(x)\right]c\bar{s}^t_e(x ) - ( u\leftrightarrow d)\right\ } , \nonumber \\
\eta_2(x)&=&{1\over\sqrt{2}}\epsilon^{abc}\left\{\left[u_a^t(x ) c d_b(x)\right ] \left[u_c^t(x ) c d_e(x ) \right ] c\bar{s}^t_e(x ) - ( u\leftrightarrow d)\right\ } .
\nonumber\end{aligned}\ ] ] here the @xmath17 represents the external electromagnetic field with the vector potential @xmath18 , the @xmath19 is the photon polarization vector and the field strength @xmath20 .
the @xmath21 , @xmath22 , @xmath23 and @xmath24 are color indexes , the @xmath25 is the charge conjugation operator , and the @xmath26 is an arbitrary parameter .
the constituents @xmath27 and @xmath28 represent the scalar @xmath29 and the pseudoscalar @xmath30 @xmath31 diquarks respectively .
they both belong to the antitriplet @xmath32 representation of the color @xmath33 group and can cluster together with diquark - diquark - antiquark type structure to give the correct spin and parity for the pentaquark @xmath6 @xmath34 .
the scalar diquarks correspond to the @xmath35 states of @xmath36 quark system .
the one gluon exchange force and the instanton induced force can lead to significant attractions between the quarks in the @xmath37 channels @xcite .
the pseudoscalar diquarks do not have nonrelativistic limit , can be taken as the @xmath38 states .
according to the basic assumption of current - hadron duality in the qcd sum rules approach @xcite , we insert a complete series of intermediate states satisfying the unitarity principle with the same quantum numbers as the current operator @xmath39 into the correlation function in eq.(1 ) to obtain the hadronic representation . after isolating the pole terms of the lowest pentaquark states , we get the following result , @xmath40\frac{{\not p}+{\not q}+m_{\theta^+}}{(p+q)^2-m_{\theta^+}^2 } + \cdots \nonumber \\ & = & - \frac{f_0 ^ 2\left[f_1(q^2)+ f_2(q^2)\right]}{(p^2-m_{\theta^+}^2)((p+q)^2-m_{\theta^+}^2 ) } \not\!p \not\!\varepsilon ( \not\!p+\not\!q ) + \cdots \nonumber\\ & = & \pi ( p , q)i\varepsilon_{\mu\nu\alpha\beta}\gamma_5\gamma^\mu\varepsilon^\nu q^\alpha p^\beta+\cdots,\end{aligned}\ ] ] where @xmath41 and the @xmath42 and @xmath43 are electromagnetic form factors . here
we write down only the double - pole term which corresponds to the magnetic moment of the pentaquark state @xmath6 explicitly , and choose the tensor structure @xmath44 .
the contributions concerning the excited and continuum states are suppressed after double borel transformation , and not shown explicitly for simplicity . from the electromagnetic form factors @xmath42 and @xmath45 , we can obtain the magnetic moment of the pentaquark state @xmath6 , @xmath46 the calculation of the operator product expansion near the light - cone @xmath13 at the level of quark and gluon degrees of freedom is straightforward and tedious , here technical details are neglected for simplicity .
the photon can couple to the quark lines perturbatively and nonperturbatively , which results in two classes of diagrams . in the first class of diagrams , the photon couples to the quark lines perturbatively through the standard qed , @xmath47 the second class of diagrams involve the nonperturbative interactions of photons with the quark lines , which are parameterized by the photon light - cone distribution amplitudes in stead of the vacuum susceptibilities . in this article , the following two - particle photon light - cone distribution amplitudes are useful @xcite , @xmath48\ } \right . \nonumber\\ & & + \left .
\{qx ( \varepsilon_\alpha x_\beta - \varepsilon_\beta x_\alpha ) + \varepsilon x ( x_\alpha q_\beta - x_\beta q_\alpha)\ } g_2(u)\right\ } \ , , \\ \langle \gamma ( q)|\bar q(x ) \gamma_\mu\gamma_5q(0)|0\rangle & = & \frac{f}{4 } e_q \epsilon_{\mu\nu\rho\sigma } \varepsilon^\nu q^\rho x^\sigma
\int_0 ^ 1 due^{iuqx } \psi(u ) \ , ,
\\ \langle \gamma ( q)|\bar q(x ) \gamma_\mu q(0)|0\rangle & = & f^{(v ) } e_q \varepsilon_\mu \int_0 ^ 1 due^{i uqx } \psi^{(v)}(u ) \
, , \end{aligned}\ ] ] where the @xmath49 is the magnetic susceptibility of the quark condensate , its values with different theoretical approaches are different from each other , for a short review , one can see ref.@xcite .
the @xmath50 is the quark charge , @xmath51 and @xmath52 are the twist2 photon light - cone distribution amplitudes , while @xmath53 and @xmath54 are the twist4 light - cone distribution amplitudes .
the twist3 photon light - cone distribution amplitudes are neglected due to their small contributions .
once the analytical results are obtained , then we can express the correlation functions at the level of quark - gluon degrees of freedom into the following form through dispersion relation , @xmath55}{s - up^2-(1-u)(p+q)^2}+ b(p , q ) \right\}+\cdots\ , , \end{aligned}\ ] ] where @xmath56}{\pi}&= & \frac{(5t^2 + 2t+5)s^4}{2^{12}5!4!\pi^8}-\frac{(5t^2 + 2t+5)s^3f \psi(u)}{2^{13}5!\pi^6}+\frac{(7t^2 - 2t-5)s\langle \bar{q}q\rangle^2}{2^{7}3 ^ 2\pi^4}\nonumber \\ & & -\frac{(7t^2 - 2t-5)\langle \bar{q}q\rangle^2f
\psi(u)}{2^{6}3\pi^2}+\frac{(5t^2 + 2t+5)s^2}{2^{15}4!\pi^6}\langle \frac{\alpha_s gg}{\pi } \rangle \nonumber \\ & & -\frac{(5t^2 + 2t+5)sf \psi(u)}{2^{14}3\pi^4}\langle \frac{\alpha_s gg}{\pi } \rangle , \nonumber \\
b(p , q)&=&-\frac{(5t^2 + 2t+5)\langle \bar{q}q\rangle^4}{2^{3}3 ^ 3 \left(-up^2-(1-u)(p+q)^2\right)^2}. \nonumber\end{aligned}\ ] ] from eq.(10 ) , we can see that due to the special interpolating current ( eq.(2 ) ) , only the @xmath57 quark has contributions to the magnetic moment of the pentaquark @xmath6 , which is significantly different from the results obtained in refs.@xcite , where all the @xmath58 , @xmath59 and @xmath57 quarks have contributions . in refs.@xcite ,
the interpolating current @xmath60 is used , @xmath61 then we make double borel transformation with respect to the variables @xmath62 and @xmath63 in eq.(10 ) , @xmath64^n } = ( m^2)^{2-n } e^{-{m^2\over m^2 } } \delta ( u - u_0 ) , \nonumber\ ] ] here @xmath65 is the borel parameter and @xmath66 . in this way
the single - pole terms can be eliminated . for more discussions about double borel transformation
, one can consult ref.@xcite .
finally we obtain the sum rule , @xmath67 @xmath68 here we have used the functions @xmath69 to subtract contributions come from the excited states and continuum states . as the initial and final states are the same pentaquark state @xmath6 .
it is natural to take the values @xmath70 and @xmath71 . if we replace the final @xmath17 state with the vacuum state in eq.(1 ) , we can obtain the sum rules for the coupling constant @xmath72 @xcite , @xmath73 @xmath74 from above equations , we can obtain @xmath75
in this article , we take the values of the parameter @xmath26 for the interpolating current in eq.(2 ) to be @xmath76 , which can give stable mass i.e. @xmath77 with respect to the variations of the borel mass @xmath78 in the considered interval @xmath79 in ref.@xcite .
the other parameters are taken as @xmath80 , @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 and @xmath87 @xcite .
here we have neglected the uncertainties about the vacuum condensates , small variations of those condensates will not lead to larger changes about the numerical values .
the threshold parameter @xmath88 is chosen to vary between @xmath89 to avoid possible contaminations from higher resonances and continuum states .
for @xmath90 , we obtain the values @xmath91 where the @xmath92 is the nucleon magneton . from the table 1
, we can see that although the numerical values for the magnetic moment of the pentaquark state @xmath6 vary with theoretical approaches , they are small in general ; our numerical results are consistent with most of the existing values of theoretical estimations in magnitude , however , with negative sign . in previously work @xcite , we observe that the sum rules in the external electromagnetic field with different interpolating currents can lead to very different predictions for the magnetic moment , although they can both give satisfactory masses for the pentaquark state @xmath6 .
the present results show that the same interpolating current with different approaches i.e. the qcd sum rules in the external fields and the light - cone qcd sum rules , can result in different predictions in magnitude with the same sign @xcite .
when the experimental measurement of the magnetic moment of the pentaquark state @xmath6 is possible in the near future , we might be able to test the theoretical predictions of the magnetic moment and select the preferred quark configurations . in the region @xmath79 , the sum rules for @xmath93 as functions of the borel parameter @xmath78
is plotted in fig.1 for @xmath94 as an example .
.[tablechi]the values of @xmath95 ( in unit of @xmath92 ) [ cols="^,^ " , ]
in summary , we have calculated the magnetic moment of the pentaquark state @xmath6 as diquark - diquark - antiquark ( @xmath1[ud]\bar{s}$ ] ) state in the framework of the light - cone qcd sum rules approach .
the numerical results are consistent with most of the existing values of theoretical estimations in magnitude , however , with negative sign , @xmath96 . in previously work @xcite , we observe that the sum rules in the external electromagnetic field with different interpolating currents can lead to very different predictions for the magnetic moment , although they can both give satisfactory masses for the pentaquark state @xmath6 .
the present results show that the same interpolating current with different approaches i.e. the qcd sum rules in the external fields and the light - cone qcd sum rules , can result in different predictions in magnitude with the same sign @xcite .
the magnetic moments of the baryons are fundamental parameters as their masses , which have copious information about the underlying quark structures , different substructures can lead to very different results .
the width of the pentaquark state @xmath6 is so narrow , the small magnetic moment may be extracted from photo - production experiments eventually in the near future , which may be used to distinguish the preferred configurations from various theoretical models , obtain more insight into the relevant degrees of freedom and deepen our understanding about the underlying dynamics that determines the properties of the exotic pentaquark states .
this work is supported by national natural science foundation , grant number 10405009 , and key program foundation of ncepu .
the authors are indebted to dr .
j.he ( ihep ) and dr .
x.b.huang ( pku ) for numerous help , without them , the work would not be finished . 99 leps collaboration , t. nakano _ et al _ , phys .
* 91 * ( 2003 ) 012002 ; diana collaboration , v. v. barmin _
et al _ , phys .
* 66 * ( 2003 ) 1715 ; clas collaboration , s. stepanyan _
et al_. , phys .
* 91 * ( 2003 ) 252001 ; clas collaboration , v. kubarovsky _
et al _ , phys .
* 92 * ( 2004 ) 032001 ; saphir collaboration , j. barth _ et al _ , phys . lett .
* b572 * ( 2003 ) 127 ; a. e. asratyan , a. g. dolgolenko and m. a. kubantsev , phys . atom . nucl . *
67 * ( 2004 ) 682 ; hermes collaboration , a. airapetian _
et al _ , phys
* b585 * ( 2004 ) 213 ; svd collaboration , a. aleev _
et al _ , hep - ex/0401024 ; zeus collaboration , s. v. chekanov , hep - ex/0404007 ; clas collaboration , h. g. juengst , nucl - ex/0312019 .
n. ishii , t. doi , h. iida , m.oka , f. okiharu , h. suganuma , phys .
* d71 * ( 2005 ) 034001 ; n. mathur , f.x .
lee , a. alexandru , c. bennhold , y. chen , s. j. dong , t. draper , i. horvath , k. f. liu , s. tamhankar and j.b .
zhang , phys . rev .
* d70 * ( 2004 ) 074508 .
p. z. huang , w. z. deng , x. l. chen and s. l. zhu , phys .
* d69 * ( 2004 ) 074004 .
q. zhao , phys . rev .
* d69 * ( 2004 ) 053009 ; erratum - ibid . *
d70 * ( 2004 ) 039901 .
h. c. kim and m. praszalowicz , phys .
b585 * ( 2004 ) 99 .
s. i. nam , a. hosaka and h. c. kim , phys . lett .
* b579 * ( 2004 ) 43 .
y. r. liu , p. z. huang , w. z. deng , x. l. chen and s. l. zhu , phys . rev .
* c69 * ( 2004 ) 035205 .
t. inoue , v. e. lyubovitskij , t. gutsche and a. faessler , hep - ph/0408057 .
r. bijker , m. m. giannini and e. santopinto , phys .
b595 * ( 2004 ) 260 .
k. goeke , h. c. kim , m. praszalowicz and g. s. yang , hep - ph/0411195 .
d. k. hong , y. j. sohn and i. zahed , phys . lett . *
b596 * ( 2004 ) 191 .
p. jimenez delgado , hep - ph/0409128 .
z. g. wang , w. m. yang and s. l. wan , hep - ph/0501278 .
z. g. wang , s. l. wan and w. m. yang , hep - ph/0503007 .
i. i. balitsky , v. m. braun and a. v. kolesnichenko , sov
* 44 * ( 1986 ) 1028 ; nucl . phys .
* b312 * ( 1989 ) 509 ; v. l. chernyak and i. r. zhitnitsky , nucl . phys . *
b345 * ( 1990 ) 137 ; v. l. chernyak and a. r. zhitnitsky , phys . rep .
* 112 * ( 1984 ) 173 .
v. m. braun and i. e. filyanov , z. phys . * c44 * ( 1989 ) 157 .
aliev , i. kanik and m. savci , phys .
* d68 * ( 2003 ) 056002 ; t.m .
aliev , a. ozpineci and m. savci .
phys . rev .
* d66 * ( 2002 ) 016002 , erratum - ibid .
* d67 * ( 2003 ) 039901 ; phys . rev .
* d65 * ( 2002 ) 096004 ; phys . rev . * d65 * ( 2002 ) 056008 ; phys . lett .
* b516 * ( 2001 ) 299 ; nucl . phys .
* a678 * ( 2000 ) 443 ; t. m. aliev and a. ozpineci .
rev . * d62 * ( 2000 ) 053012 .
a. de rujula , h. georgi and s. l. glashow , phys . rev .
* d12 * ( 1975 ) 147 ; t. degrand , r. l. jaffe , k. johnson and j. e. kiskis , phys .
d12 * ( 1975 ) 2060 ; g. t hooft , phys . rev .
* d14 * ( 1976 ) 3432 [ erratum - ibid .
* d18 * ( 1978 ) 2199 ] ; e. v. shuryak , nucl . phys .
* b203 * ( 1982 ) 93 ; t. schafer and e. v. shuryak , rev .
* 70 * ( 1998 ) 323 ; e. shuryak and i. zahed , phys
* b 589 * ( 2004 ) 21 .
i. i. balitsky , v. m. braun , and a. v. kolesnichenko , nucl .
b312 * ( 1989 ) 509 .
v. m. braun and i. e. filyanov , z. phys .
* c48 * ( 1990 ) 239 ; a. ali and v. m. braun , phys .
* b359 * ( 1995 ) 223 ; g. eilam , i. halperin and r. r. mendel , phys . lett .
* b361 * ( 1995 ) 137 ; p. ball , v. m. braun and n. kivel , nucl .
b649 * ( 2003 ) 263 .
z. g. wang , j. phys . * g28 * ( 2002 ) 3007 . v.
m. belyaev , v. m. braun , a. khodjamirian and r. rckl , phys . rev . *
d51 * ( 1995 ) 6177 ; h. kim , s. h. lee and m. oka , prog .
* 109 * ( 2003 ) 371 ; z. g. wang , w. m. yang and s. l. wan , eur .
j. * c37 * ( 2004 ) 223 . | in this article , we study the magnetic moment of the pentaquark state @xmath0 as diquark - diquark - antiquark ( @xmath1[ud]\bar{s}$ ] ) state in the framework of the light - cone qcd sum rules approach .
the numerical results indicate the magnetic moment of the pentaquark state @xmath0 is about @xmath2 . + zhi - gang wang@xmath3 , shao - long wan@xmath4 and wei - min yang@xmath5 + @xmath3 department of physics , north china electric power university , baoding 071003 , p. r. china + @xmath4 department of modern physics , university of science and technology of china , hefei 230026 , p. r. china + pacs : 12.38.aw , 12.38.lg , 12.39.ba , 12.39.-x * key words : * light - cone qcd sum rules , magnetic moment , pentaquark |
accumulating results of theoretical investigations indicate that the gravitational field exists in different phases .
first indications supporting such an idea came from considerations of three - dimensional euclidean quantum gravity @xcite . by means of monte carlo simulations it was possible to explore the configuration of the gravitational field under various conditions . for four - dimensional euclidean gravity ,
gravity exhibits two phases : the _ crumpled _ phase and the _ branched polymer _ phase @xcite .
this result has since been generalized to the case of four - dimensional gravity with an imposed causality condition , formulation known as causal dynamical triangulations ( cdt ) .
the causality condition turned out to be essential for the correct phase structure of gravity , leading to emergence of the four - dimensional spacetime @xcite .
generation of such extended phase in the euclidean approach introducing a non - trivial path integral measure remains an interesting possibility @xcite .
furthermore , in similarity to a phase structure of the lifshitz scalar @xcite , the critical surface of cdt has been divided into three regions separated by the first and second order transition lines @xcite .
interestingly , a theory describing gravity at a triple point ( lifshitz point ) of the phase diagram has been constructed , and shown to be power - counting renormalizable @xcite .
further evidence for the non - trivial phase structure of gravity comes from quantum graphity @xcite .
this approach utilizes the idea of _ geometrogenesis _ - a transition between geometric and non - geometric phases of gravity .
the basic question one can ask , assuming the existence of the different phases of gravity , is where the other phases can be found ?
a natural place to search for them are high curvature regions such as interiors of the black holes and the early universe . because of a horizon , a possibility of empirical verification of a phase change inside of black holes is out of reach .
more promising is a search for signatures of the gravitational phase transitions which took place in the early universe .
so far , there has been very little attention devoted to this issue in the literature .
most studies of the phase transitions in the early universe were dedicated to the matter sector , rather than gravity @xcite . among the few studies on the gravitational phase transitions in the early universe , the work of refs .
@xcite is especially noteworthy . in ref .
@xcite a specific model of geometrogenesis , through a second order phase transition , has been proposed .
it was shown that by assuming the holographic principle to be fulfilled in the high temperature phase , it is possible to generate a power spectrum of primordial perturbations that is in agreement with observations . in ref .
@xcite the cosmological relevance of second order phase transitions is discussed .
arguments supporting generation of `` inflationary '' power spectrum from critical behavior of the gravitational field have been presented . in what follows
we attract attention to the fact that a second order gravitational phase transition has recently been observed within causal dynamical triangulations @xcite .
the transition takes place exactly between the phases of the form discussed in refs .
therefore , cdt gives a concrete realization of the scenario of geometrogenesis .
we also show that gravitational phase transition may be associated with the deformation of general covariance , recently observed in the context of loop quantum cosmology ( lqc ) . in both cases ,
the phase transition is of second order , suggesting a critical nature of the emergence of classical spacetime in the early universe .
analysis performed within four - dimensional cdt with a positive cosmological constant indicates the presence of three different phases of the gravitational field , called a , b and c @xcite .
the phases are separated by the first ( a - c ) and second order ( b - c ) transition lines presumably intersecting at the triple point .
the order of the a - b phase transition has not been determined so far . at large scales ,
phase c forms the four - dimensional de sitter space @xcite .
the phase is , however , not fully classical since it exhibits dimensional reduction to two dimensions at short scales @xcite .
this can be shown by investigating properties of the spectral dimension , defined via a diffusion process .
nevertheless , the phase c can be associated with the usual " phase of gravity .
the two remaining phases , are fundamentally different from this phase .
phase a is characterized by a vanishing interaction between adjacent time slices .
phase b , resembling the _ crumpled _ phase in euclidean gravity , is characterized by a large ( tending to infinity in the @xmath0volume limit ) haussdorf and spectral dimension .
phase b shares features of the high temperature phase postulated in quantum graphity .
moreover , this phase is separated with the low energetic phase c by the second order phase transition .
this is in one - to - one correspondence to the quantum graphity case .
based on this observation , we hypothesize the following : _ hypothesis 1 . in the early universe , there is a second order phase transition from the high temperature phase b to the low temperature phase c. the transition is associated with a change of the connectivity structure between the elementary chunks of space . _
the change of connectivity can be inferred from the considerations of the spectral dimensions of the phases b and c. in order to see it explicitly let us consider a toy model of the universe composed of the @xmath1 chunks of space
. they will be represented by the nodes of a graph .
a structure of adjacency is represented by the links . in phase c , which is a geometric phase
, the degree of vertices is low . in our toy model
it is equal @xmath2 and the resulting space is represented by the ring graph ( see the left panel in fig .
[ graphs ] ) .
the spectral dimension of this graph can be found by determining spectrum ( eigenvalues @xmath3 ) of the laplace operator @xmath4 , where @xmath5 is an adjacency matrix and @xmath6 is a degree matrix . by using the expression for the trace of the heat kernel
one can find that @xmath7 where @xmath8 is a diffusion time .
@xmath9{ringgraph10.eps } & \includegraphics[width=5 cm , angle = 0]{completegraph10.eps } \end{array}$ ] in fig .
[ spectdim ] we plot function ( [ spectraldim ] ) for the ring graph , for which the eigenvalues are @xmath10 . at intermediate diffusion times
the spectral dimension is equal to one , as expected classically .
the short time behavior , corresponding to dimensional reduction observed in the four - dimensional case , is due to the discrete nature of the network .
furthermore , at large diffusion times the spectral dimension is again falling to zero due to the compactness of space ( not visible in fig .
[ spectdim ] ) .
let us now model the high temperature non - geometric phase b by assuming , for computational simplicity , a maximal degree of nodes .
the resulting complete graph is shown in fig .
[ graphs ] .
the assumption of completeness allows one to determine spectrum of the laplace operator analytically , and enables the simplification of eq .
[ spectraldim ] to @xmath11 we plot this function in fig .
[ spectdim ] , comparing it with the low temperature case .
as we see , the spectral dimension is now peaked at small diffusion times .
the maximal value of the spectral dimension grows with the number of nodes as @xmath12 , where @xmath13 is the lamber function .
this behavior is in qualitative agreement with the numerical computations performed in four - dimensional cdt @xcite .
recent developments in lqc indicate that the hypersurface deformation algebra ( hda ) is deformed due to the quantum gravitational effects @xcite .
this means that the general covariance is quantum deformed , but not broken . in condensed matter physics ,
a change of symmetry is often associated with the occurrence of a phase transition . by extrapolating this observation to the sector of gravitational interactions
, we make the following hypothesis : _ hypothesis 2 .
gravitational phase transitions are associated with deformations of the hypersurface deformation algebra . _ in order to support this hypothesis we present an example based on holonomy corrections in lqc . in this case , hda is deformed such that @xmath14 , where @xmath15 are scalar constraints and @xmath6 is a diffeomorphism constraint , and the remaining brackets are unchanged @xcite .
the @xmath16 is a deformation factor , @xmath17 denotes energy density of matter and @xmath18 is a maximal energy density expected to be of the order of the planck energy density . at low energy densities , the classical lorentzian hda with @xmath19
is recovered , while at @xmath20 , @xmath21 corresponding to euclidean space . approaching the planck epoch
is therefore associated with the signature change from lorentzian spacetime to euclidean four - dimensional space @xcite .
interestingly , at @xmath22 the hda reduces to the ultralocal form ( @xmath23 ) describing a state of _ silence _ @xcite .
this state shares properties of phase a in cdt , giving a first indication for the relationship between the phase of gravity and deformation of hda .
further evidence comes from a simple model of the signature change as a spontaneous symmetry breaking ( sbb ) associated with a second order phase transition .
the symmetry breaking is from @xmath24 to @xmath25 at the level of an effective homogeneous vector field @xmath26 .
this translates to a symmetry change from @xmath24 to @xmath27 , experienced by the field living on a geometry described by the metric @xmath28 .
let us assume that the free energy for the model with a massless scalar field @xmath29 is @xmath30}_{v(\phi^{\mu},\rho ) } , \nonumber\end{aligned}\ ] ] where @xmath31 and @xmath32 is a constant .
due to the symmetry breaking kinetic factor , the expression for the free energy is not explicitly @xmath24 invariant . treating this term as a perturbation ( which is valid for sufficiently small values of @xmath29 ) ,
the equilibrium is obtained by minimizing value of the potential @xmath33 . at energy densities
@xmath34 the vacuum state maintains the @xmath24 symmetry , leading to @xmath35 .
the metric elements are therefore @xmath36 and @xmath37 , representing the four - dimensional euclidean space .
this region is however forbidden due to the constraint @xmath38 present in lqc . in turn , below the critical energy @xmath39 , the minimum of the potential is located at @xmath40 in some spontaneously chosen direction ( see fig .
[ ssb ] ) . without loss of generality ,
let us assume that the ssb takes place in direction @xmath41 , for which @xmath42 leading to the effective speed of light @xmath43 . as a consequence ,
the equation of motion for the scalar field @xmath29 takes the form @xmath44 manifesting @xmath24 symmetry at the critical point ( @xmath21 ) , and @xmath27 symmetry in the low temperature limit ( @xmath19 ) .
the form of equation ( [ eqmv ] ) agrees with the one derived from the holonomy deformations of the hda @xcite . in lqc ,
energy densities above @xmath18 can not be reached .
therefore , the evolution starts at the critical point located at @xmath18 .
an interesting possibility is that the system has been maintained at the critical point before the energy density started to drop .
this may not require a fine - tuning if the dynamics of the system exhibited self organized criticality ( soc ) @xcite , which is observed in various complex systems .
interestingly , this concept has already been applied to quantum gravity , however , in order to describe classical configuration of space @xcite .
furthermore , it is possible that the region @xmath45 is described by a non - geometric phase of gravity , which properties can not be captured within the presented model .
we have shown that cdt offers a concrete realization of geometrogenesis , having a second order gravitational phase transition between the non - geometric and geometric phase .
we have explained how measurements of the spectral dimension are related in this case with connectivity of the chunks of space .
the critical nature of the emergence of classical spacetime in the early universe may give a first possibility of testing cdt .
however , this would require more detailed investigations of the properties of the second order phase transition in cdt . in lqc , which is an alternative to cdt ,
a gravitational second order phase transition may explain a signature change in the planck epoch . in the presented toy model , the universe originates just at the critical point .
interestingly , the latest results in cdt indicate the existence of a new phase between phases b and c @xcite , characterized by a `` bifurcation '' of the kinetic term .
this behavior resembles the signature change observed in lqc .
j. ambjorn , d. v. boulatov , a. krzywicki and s. varsted , phys .
b * 276 * ( 1992 ) 432 .
j. ambjorn and j. jurkiewicz , phys .
b * 278 * ( 1992 ) 42 .
j. ambjorn , j. jurkiewicz and r. loll , phys .
* 93 * ( 2004 ) 131301 [ hep - th/0404156 ] .
j. laiho and d. coumbe , phys . rev .
* 107 * ( 2011 ) 161301 [ arxiv:1104.5505 [ hep - lat ] ] .
r. m. hornreich , m. luban and s. shtrikman , phys .
* 35 * ( 1975 ) 1678 .
j. ambjorn , s. jordan , j. jurkiewicz and r. loll , phys .
d * 85 * ( 2012 ) 124044 [ arxiv:1205.1229 [ hep - th ] ] .
p. horava , phys .
d * 79 * ( 2009 ) 084008 [ arxiv:0901.3775 [ hep - th ] ] .
t. konopka , f. markopoulou and l. smolin , hep - th/0611197 .
t. w. b. kibble , j. phys . a * 9 * ( 1976 ) 1387 . w. h. zurek , nature * 317 * ( 1985 ) 505 .
j. magueijo , l. smolin and c. r. contaldi , class .
* 24 * ( 2007 ) 3691 [ astro - ph/0611695 ] .
o. dreyer , arxiv:1307.6169 [ gr - qc ] .
j. ambjorn , s. jordan , j. jurkiewicz and r. loll , phys .
lett . * 107 * ( 2011 ) 211303 [ arxiv:1108.3932 [ hep - th ] ] . j. ambjorn , a. grlich , j. jurkiewicz and r. loll , phys .
* 100 * ( 2008 ) 091304 [ arxiv:0712.2485 [ hep - th ] ] .
j. ambjorn , j. jurkiewicz and r. loll , phys .
* 95 * ( 2005 ) 171301 [ hep - th/0505113 ] .
a. grlich , private communication .
m. bojowald and g. m. paily , phys .
d * 86 * ( 2012 ) 104018 [ arxiv:1112.1899 [ gr - qc ] ] .
t. cailleteau , j. mielczarek , a. barrau and j. grain , class .
* 29 * ( 2012 ) 095010 [ arxiv:1111.3535 [ gr - qc ] ] .
j. mielczarek , arxiv:1207.4657 [ gr - qc ] .
j. mielczarek , aip conf .
* 1514 * ( 2012 ) 81 [ arxiv:1212.3527 [ gr - qc ] ] .
p. bak , c. tang and k. wiesenfeld , phys .
* 59 * ( 1987 ) 381 .
m. h. ansari and l. smolin , class .
* 25 * ( 2008 ) 095016 [ hep - th/0412307 ] .
j. ambjorn , j. gizbert - studnicki , a. grlich and j. jurkiewicz , arxiv:1403.5940 [ hep - th ] . | this essay addresses the issue of gravitational phase transitions in the early universe .
we suggest that a second order phase transition observed in the causal dynamical triangulations approach to quantum gravity may have a cosmological relevance .
the phase transition interpolates between a non - geometric crumpled phase of gravity , and an extended phase with classical properties .
transitions of this kind have been postulated earlier in the context of geometrogenesis in quantum graphity .
we show that critical behavior may also be associated with a signature change event in loop quantum cosmology . in both cases ,
classical spacetime originates at the critical point associated with a second order phase transition . |
many of network service provider s pain points can be traced back to a fundamental issue : the lack of network visibility .
for example , the network congestion collapse can be avoided in many cases if we know exactly when and where the congestion is happening or even better , if we can precisely predict it well before any impact is made ; sophisticated network attacks can be prevented through stateful and distributed network behavior analysis ; in order to monetize the traffic and provide application - centric service , user flows and their interaction with networks need to be tracked and understood .
all these pose vast needs on generalized network data , either passing through networks or generated by network devices .
not surprisingly , people start to model the network visibility as a big data problem
. traditional algorithms can still apply , but more advanced big data analytics such as machine learning can provide unlimited opportunities to mine values from network data . if we can retrieve any data of interest in real time with direct helps from the data source ( i.e. , the network data plane ) , then most problems can be solved by the comprehensive data analytics at the application plane . therefore , our value proposition is to _ build a unified and general - purpose network data analytics platform with integrated data plane support to provide the omni network visibility_. this is in contrast to the ad - hoc solutions which only deal with one single problem a time with a special means to acquire the relevant data .
sdn appears to be an ideal architecture to support the omni network visibility .
the logically centralized controller is at the unique vantage point to see any data in networks .
however , so far the sdn controller has limited view of network data and states , because the data plane is incapable of providing enough data to sustain all application requirements .
therefore , we need to make the data plane programmable so any data can be collected if needed .
more importantly , the programmability must be open to the upper layer applications so service provider can directly take advantage of this flexibility for application layer data collection .
we further argue that the data plane should also support on - demand and real - time programming to meet the dynamic data needs from various runtime applications ( detailed in sec .
[ motiv ] ) .
this interactive programming capability is in contrast to the conventional programming with static languages .
with such a data plane , we can add network probes anywhere and anytime through a standard interface .
these passive probes will not alter the forwarding behavior but add the monitor points which are responsible to collect and preprocess data for applications .
we make two major contributions in this paper .
first , we devise the dynamic network probe ( dnp ) as a flexible and dynamic means for sdn data plane data collection .
second , we show the possibility to build a universal network data analytics platform in which network devices play an integrated role .
the first contribution forms the foundation for the second one .
data collected from data plane provide input to the sdn control loop .
what data to collect is determined by the purpose of the data - consuming application .
the most basic application is routing and forwarding decision for which the network topology and link states need to be collected .
other applications , such as traffic engineering , network security , network health monitoring , trouble shooting , and fault diagnosis , require different type of data . the data are either normal traffic packets that are filtered , sampled , or digested , or metadata generated by network devices to convey network states and status . in either case , data collection is meant to be passive and should not change the network forwarding behavior .
the sdn controller analyzes the collected data and then makes decisions accordingly to actively change the network behavior .
forwarding information base ( fib ) and access control list ( acl ) table updates are the most notable examples , as well as traffic manager ( tm ) parameters .
a few other possible changes are less obvious . for example , some applications keep refining the ways to collect data based on previous observations .
the control loop algorithm continuously interacts with the data plane and modify the data source and content .
we can also imagine that in some other applications , new network and packet processing functions can be enabled to alter the forwarding behavior at runtime .
this paper concerns with the passive data collection only .
although the technology discussed in this paper can also be applied to actively modify the network behavior , we leave that topic to future work .
ideally , we want to gain the full visibility to know any states anytime anywhere in the entire network data plane . in reality , this is extremely difficult if not impossible .
theoretically , any network state can be inferred if all the traffic through it can be seen , so a simple option is to mirror all the raw traffic to servers where data analytical engine is running .
however , this brute - force method requires to double the device port count and the traffic bandwidth , and poses enormous computing and storage cost . as a tradeoff
, test access port ( tap ) or switch port analyzer ( span ) is used to selectively mirror only a portion of the overall traffic .
network packet broker ( npb ) is deployed along with tap or span to process and distribute the raw data to various data analytical tools .
there are some other ad - hoc solutions ( e.g. , sflow @xcite and everflow @xcite ) which can provide sampled and digested packet data and some traffic statistics .
meanwhile , network devices also generate various log files to record miscellaneous events in the system .
when aggregating all these solutions together , we can gain a relatively comprehensive view of the network .
however , the main problem here is the lack of a unified platform to deal with the general data collection problem .
moreover , each ad - hoc solution inevitably loses information due to data plane resource limitation which makes the data analytical results suboptimal , so does the follow - up data plane control based on the results .
we also note that application s requests on data are often dynamic and realtime .
when third party applications run on top of the network operator s controller , their data needs are diversified and unpredictable .
even a single application may need to constantly adjust what data to collect ( e.g. , an elephant flow detector continues to narrow down the flow granularities and gather their statistics ) .
trying to design an omnipotent system to support all possible runtime data requests is inviable because the resources required are prohibitive ( e.g. , even a simple counter per flow is impossible in practice ) .
an alternative is to reprogram or reconfigure the data plane device whenever an unsupported data request appears .
this is possible thanks to the recently available programmable chips @xcite and the trend to open the programmability to service providers @xcite .
unfortunately , the static programming approach can not meet the realtime requirements due to the latency incurred by the programming and compiling process .
the reprogramming process also risks breaking the normal operation of network devices .
then a viable solution left to us is : whenever applications request data which is unavailable in the data plane , the data plane can be configured in real time to return the requested data .
that is , we do not attempt to make the network data plane provide all data all the time . instead , we only need to ensure that _ any application can acquire any necessary data instantly whenever it actually asks for it .
_ this data - on - demand model can support effectively `` omni '' network visibility , but it is still unthinkable with the current stiff and black - boxed data plane .
so we first introduce the recent technology advance to ground the feasibility of our vision in sec .
[ tech ] and [ pof ] .
driven by sdn , network data plane is evolving to become open programmable @xcite .
this means the network operators are in control of customizing the network device s function and forwarding behavior .
several ongoing trends in industry , as shown in figure [ fig_opdp ] , are validating this idea .
these trends are shaping new forms of network devices and inspiring innovative ways to use them .
the first trend is led by the ocp networking project @xcite , which advocates the decoupling of the network operating system and the network device hardware .
a common switch abstract interface ( sai ) allows applications to run on heterogeneous substrate devices .
however , such devices are built with fixed function asics , which provide limited flexibility for application customization .
the second trend is built upon the first one yet makes a big leap . chip and device vendors are working on opening the programmability of the npu , cpu , and fpga - based network devices to network operators .
most recently , programmable asics has been proved feasible @xcite .
high level language such as p4 @xcite is developed to make the network device programming easy and fast .
now a network device can be incarnated into different functioning boxes depending on the program installed . however , such programming process is considered static .
even a minor modification to the existing application requires to recompile the updated source code and reinstall the application .
this incurs long deployment latency and may also temporarily break the normal data plane operation .
hence , although this generation of trend is still in its infancy , we have seen some of its limitations .
open programmable data plane should be stretched further to support runtime interactive programming in order to extend its scope of usability .
dynamic application requirements can not be foreseen at design time , and runtime data plane modifications are required to be done in real time ( for agile control loop ) and on demand ( to meet data plane resource constraints ) . meanwhile , the data plane devices are capable of doing more complex things such as stateful processing without always resorting to controller for state tracking .
this allows network devices to offload a significant portion of the data processing task and only hand off the preprocessed data to the data - requesting applications .
we introduced the pof programming model to support this trend as shown in figure [ fig_pof1 ] @xcite .
we still use static programing with high level languages to define the main data plane processing and forwarding function .
but at runtime , whenever an application requires to make some modification to the data plane , we deploy the incremental modification directly through the runtime control channel . the key to make this dynamic and interactive programming work is to maintain a unified interface to devices for both configuration and runtime control , because both programming paths share the same data plane abstraction and use the same back - end adapting and mapping method .
the data plane processing and forwarding function is abstracted as the standard intermediate representation ( ir ) .
ir contains high - level and abstract objectives that structurize the data plane components and flow logic @xcite .
these objectives are not deviated too much from the high level language such as p4 @xcite .
this means for incremental changes , it is possible to provide apis directly at ir level .
user can therefore interactively apply data plane changes and avoid the full source - code compiling process . of course
, the network devices need to be flexible enough to support the interactive programming .
we have implemented an npu - based hardware prototype to support it .
the virtual network device running on cpu / gpu can certainly support it easily .
asics and fpgas are also possible to support it but do need some architectural innovations which will be briefly discussed in section [ conclude ] .
the forwarding application in data plane can be modeled in different ways ( e.g. , forces @xcite ) , but so far the most popular abstract model is the match - action table pipeline@xcite . in this model ,
the streaming packets coming from some physical or logical input ports enter a pipeline . at each pipeline stage
, some data extracted form the packet or the metadata is used as key to search a table .
the matching entry triggers some associated action .
the action may lead the packet to another pipeline stage or an output port .
a single network device may contain multiple such pipelines , segregated by logical ports ( e.g. , controller , switch fabric , or black box modules ) .
this simple yet expressive model is sufficient to describe arbitrary data plane forwarding applications .
we make several enhancements to the basic model in order to support the runtime interactive programming .
first , the actions are no longer statically associated with flow table entries as in openflow @xcite and p4 @xcite . as the chief programming point
, actions contain primitive packet processing instructions and control instructions .
each custom action can be dynamically loaded into network devices .
each flow entry has a pointer field which holds a pointer to an action , and a parameter field which holds the parameters used by the associated action[multiblock footnote omitted ] .
this mechanism allows users to change flow actions at runtime .
it also allows each action to have different number of instructions , as long as the performance and action storage can sustain the size . with this mechanism , users are free to download new actions to change a flow s behavior on the run .
even new pipeline stages ( i.e. , new tables ) can be inserted into the pipeline dynamically without interrupting the data path .
second , some data plane resources are maintained in a dynamically shared pool .
such resources include meters , counters , timers , and global registers and tables .
the rationale is multifold : ( 1 ) network devices can not afford to statically allocated counters or meters to all the flows and flow tables . for example , a two - million - entry ip forwarding table alone can consume 64 mb memory for counters . since it is impossible to know a priori how to reasonably allocate these limited resources at design time ,
counters and meters must be shared dynamically ; ( 2 ) it is easy to see if a counter or meter can be allocated to multiple flows , the aggregated flow statistics and measurement can be realized with ease .
likewise , multiple counters and meters can also be allocated to a single flow so various aspects of the flow can be counted and metered with conditional instructions ; ( 3 ) a global register or table is essentially a persistent storage . when it is assigned to a flow
, the flow packet can assess and update its value , therefore some stateful processing can be realized .
similarly , a register can be assigned to multiple flows to enable inter - flow information sharing , and a flow can be assigned with multiple registers to hold more stateful data . at last , for better stateful processing support , actions are given the privilege to write tables ( i.e. , insert , modify , or delete table entries ) .
this can effectively turn a table into a state table while the search key is no longer a flow signature but a state signature .
this idea was originated in @xcite and further developed in @xcite .
figure [ fig_dpmodel ] summarizes the major data plane features that support interactive programming .
such data plane is so flexible that the inattentive use of interactive programming may accidentally break the device operation or even paralyze the entire network .
however , we will see at least in one scenario we are immune from the negative effects yet can still enjoy the benefits of interactive programming .
network probes are passive monitors which are installed at specific forwarding data path locations to collect specific data .
dnps are _ dynamically _ deployed and revoked probes by applications at runtime .
the customizable dnps can collect simple statistics or conduct more complex data preprocessing .
since dnps may require actively modifying the existing data path pipeline beyond simple flow entry manipulation , these operations need to be done through interactive programming process . when a dnp is revoked , the involved shared resources are automatically recycled and returned back to the global resource pool .
dnps can be deployed at various data path locations including port , queue , buffer , table , and table entry . when the data plane programmability is extended to cover other components ( e.g. , cpu load , fan speed , gps coordinations , etc . ) , dnps can be deployed to collect corresponding data as well .
a few data plane objectives can be composed to form probes .
these objectives are counter , meter , timer , timestamp , and register . combining these with the packet filter through flow table entry configuration
, one can easily monitor and catch arbitrary states on the data plane .
the simplest probe is just a counter .
the counter can be configured to count bytes or packets and the counting can be conditional .
the more complex probes are modeled as finite state machines ( fsm ) which are configured to capture specific events[multiblock footnote omitted ] .
we have shown how stateful processing is supported in section [ pof ] .
fsms essentially preprocess the raw stream data and only report the necessary data to controller - side applications .
these complex dnps help reduce the device to controller bandwidth consumption and also alleviate the controller s processing load .
applications can use pull mode or push mode to access probes and collect data .
the normal counter probes are often accessed via pull mode .
applications decide what time and how often the counter value is read . on the other hand ,
the complex fsm probes are usually accessed in push mode .
when the target event is triggered , a report is generated and pushed to the application .
below is the pseudo code of a push - mode packet counter : whenever a counter is incremented by a predefined amount , a packet is generated to report the event to the subscribing application . ....
cntr[i ] + + ; if(cntr[i ] = = threshold ) { gen_pkt(to_app , flow_id , now ) ; cntr[i ] = 0 ; } .... to install this probe , we first identify the target action . if the action @xmath0 exists , we generate a new action @xmath1 by inserting this piece of code to the code of old action @xmath0 .
then we download this new action @xmath1 to the target device .
next we issue a command to switch the flow pointer from @xmath0 to @xmath1 . now
the probe in @xmath1 takes effect and @xmath0 can be deleted if no longer needed .
sometimes the target action does not exist which means a new flow entry needs to be installed along with the action .
a solid example is that the application wants to gather the statistics of a new flow @xmath2 that does not exist .
we need to choose a target table @xmath3 first and then analyze the normal processing action @xmath0 to this flow at the table @xmath3 .
we augment @xmath0 with the probe code to get @xmath1 and download @xmath1 to the target device .
next we issue commands to insert @xmath2 to @xmath3 and associate @xmath2 to @xmath1 .
note that if the flow @xmath2 overlaps with another flow @xmath4 in table @xmath3 , the action of @xmath4 may also need to be modified .
timer is a special global resource .
a timer can be configured to link to some action . when the time is up , the corresponding action is executed .
for example , to get notification when a port load exceeds some threshold , we can set a timer with a fixed time - out interval , and link the timer to an action which reads the counter and generates the report packet if the condition is triggered .
this way , the application avoids the need to keep pulling statistics from the data plane .
the pseudo code of the action is as follows : .... if(cntr[i ] > = threshold ) { gen_pkt(to_app , port_id , cntr[i ] , now ) ; } cntr[i ] = 0 ; .... with the use of global registers and state tables , more complex fsm probes can be implemented .
for example , to monitor the half - open tcp connections , for each syn request , we store the flow signature to a state table
. then for each ack packet , the state table is checked and the matched entry is removed .
the state table can be periodically pulled to acquire the list of half - open connections .
the application can also choose to only retrieve the counter of half - open connections .
when the counter exceeds some threshold , further measure can be taken to examine if a syn flood attack is going on .
the pseudo code is shown below . ....
if(tcp ) { if(syn ) { write flow_sig to stb ; cntr[i ] + + ; } else if(ack ) { remove flow_sig from stb ; cntr[i ] -- ; } } .... registers can be considered mini state tables which are good to track a single flow and a few state transitions .
for example , to get the duration of a particular flow , when the flow is established , the state and the timestamp are recorded in a register ; when the flow is teared down , the flow duration can be calculated with the old timestamp and the new timestamp .
in another example , we want to monitor a queue by setting a low water mark and a high water mark for the fill level . every time when an enqueue or a dequeue event happens ,
the queue depth is compared with the marks and a report packet is generated when a mark is crossed .
some probes are essentially packet filters which are used to filter out a portion of the traffic and mirrored the traffic to the application or some other target port for further processing .
there are two ways to implement a packet filter : use a flow table that matches on the filtering criteria and specify the associated action ; or directly make decision in the action .
an example of the former case is to filter all packets with a particular source ip address .
an example of the latter case is to filter all tcp fin packets at edge .
although we can always use a flow table to filter traffic , sometimes it is more efficient and convenient to directly work on the action . as being programmed by the application ,
the filtered traffic can be further processed before being sent .
two most common processes are digest and sample , both aiming to reduce the quantity of raw data .
the digest process prunes the unnecessary data from the original packet and only pack the useful information in the digest packet .
the sample process picks a subset of filtered traffic to send based on some predefined sampling criteria .
the two processes can be used jointly to maximize the data reduction effect .
an application may need to install multiple dnps in one device or across multiple devices to finish one data analytical task .
for example , to measure the latency of any link in a network .
we install a dnp on the source node to generate probe packets with timestamp .
we install another dnp at the sink node to capture the probe packets and report both the source timestamp and the sink timestamp to the application for link latency calculation .
the probe packets are also dropped by the sink dnp .
the source dnp can be configured to generate probe packets at any rate .
it can also generate just one probe packet per application request .
the pseudo code of the sink dnp action is as follows : .... if(is_probe_packet ) { gen_pkt(to_app , old_time , now ) ; drop(this ) ; } .... using the similar idea , we can deploy dnps to measure the end - to - end flow latency or trace exact flow paths
. the information can be piggybacked on packets of normal traffic or on generated probe packets .
since potentially we may have many such tasks but each of such tasks may not be constantly needed and each consumes some network resources , making them dynamic is no doubt more efficient . in summary
, dnp is a versatile tool to prepare and generate just - in - time data for data analytical applications .
in the past , network data analytics is considered a separate function from networks .
they consume raw data extracted from networks through ad hoc protocols and interfaces . with the open programmable data plane
, we expect a paradigm shift that makes the data plane be an active component of the data analytical solution .
the programmable in - network computing is efficient and flexible to offload the data preprocessing through interactive data plane programming . a universal network data analytical platform built on top of this
enables a tight and agile sdn control loop .
while dnp is a passive data plane data collection mechanism , we need to provide a declarative interface for applications to use the target - specific dnps for data analytics .
a proposed dynamic networking data analytical system is illustrated in figure [ fig_platform ] .
an application translates its data requirements into some dynamic transactional queries .
the queries are then compiled into a set of dnps targeting a subset of data plane devices and the instructions for data post - processing after data are collected from the data plane . after the dnps are deployed , each dnp conducts in - network data preprocessing and feeds the preprocessed data to the collector .
the collector finishes the data post - processing and presents the results to the data - requesting application
. a query can be either continuous or one - shot .
the continuous query may require the application to continuously refine the existing dnps or deploy new dnps .
when an application revokes its queries , the idle dnp resource is released .
since one dnp may be subscribed by multiple applications , the runtime system needs to keep track of the active dnps . as our future work
, we aim to build a dnp - based network data analytics platform which can address multiple business requirements of network service providers , including qoe measurement , security enforcement , customer care , and network optimization .
we have built a hardware - based dnp prototype based on huawei ne40e device .
the device line card is equipped with a 200gbps npu . the pof interface protocol , essentially an extension of openflow 1.4 , is used as the device programming interface @xcite .
the preliminary results are promising .
we can deploy an arbitrary counter probe and start to collect results in less than @xmath5-ms without interrupting the normal service and impacting the forwarding performance .
in contrast , to achieve the same effect with the static programming approach , we need to edit and recompile the source code of the entire design , delete the old device configuration , and download the new configuration .
the compiling process consumes about @xmath6 second and the normal service is interrupted by about @xmath6 second for new program download .
dnp reduces the deployment latency by 40 times .
note that this is only for a small forwarding application with 3 flow tables and 10 instruction blocks . for a larger forwarding application ,
the advantage of dnp is even more prominent , because deploying a dnp has a fixed cost but compiling and downloading a larger design consume more time .
we also evaluated the dnp s performance impact to the normal forwarding throughput on our platform by keeping inserting more flow counters and observing the achievable forwarding throughput .
the results are shown in figure [ fig_eval ] .
the factors that affect the performance include the additional instructions and memory accesses incurred by the dnps .
our analysis shows the extra memory accesses is the dominant factor .
one counter operation needs one read and one write to an sram block reserved for counters .
our platfom s counter memory can sustain at most 425 m accesses per second . the memory bandwidth and latency
interact with the limited number of cores and threads , which eventually drags down the throughput .
our evaluation shows that our platform can support line speed forwarding with up to 14k flow counters .
when more complex dnps are applied , we expect the performance impact will be noticeable with fewer dnps .
the results also serve as an evidence why dynamic rather than static probes are needed .
interested readers may use the open source software provided at pof website @xcite to conduct the similar experiments on a software target .
many technique challenges need to be addressed to realize dnp and the universal network data analytics platform on general sdn data plane .
we list a few here and also provide our initial thoughts on potential solutions .
\a ) allowing applications to modify the data plane has security and safety risks ( e.g. , dos attack ) .
the counter measure is to supply a standard and safe api to segregate applications from the runtime system and provide applications limited accessibility to the data plane .
each api can be easily compiled and mapped to standard dnps . an sql - like query language which adapts to the stream processing system
might be feasible for the applications .
\b ) when multiple correlated dnps are deployed across multiple network devices or function blocks , or when multiple applications request the same dnps , the deployment consistency needs to be guaranteed for correctness .
this requires a robust runtime compiling and management system which keeps track of the subscription to dnps and controls the dnp execution time and order .
\c ) the performance impact of dnps must be evaluated before deployment to avoid unintentionally reducing the forwarding throughput .
fortunately , the resource consumption and performance impact of standard dnps can be accurately profiled in advance .
a device is usually over provisioned and is capable of absorbing extra functions up to a limit .
moreover , programmable data plane allows users to tailor their forwarding application to the bare bones so more resources can be reserved for probes . the runtime system needs to evaluate the resulting throughput performance before committing a dnp .
if it is unacceptable , either some old dnps need to be revoked or the new request must be denied .
\d ) while dnp is relatively easy to be implemented in software - based platform ( e.g. , npu and cpu ) , it is harder in asic - based programmable chips .
architectural and algorithmic innovations are needed to support a more flexible pipeline which allows new pipeline stage , new tables , and new custom actions to be inserted at runtime through hitless in - service updates .
an architecture with shared memory and flexible processor cores might be viable to meet these requirements . alternatively , dnps can be implemented using an `` out - of - band '' fashion .
that is , some reserved pipeline stages are dedicated for dnps .
relevant data and/or packets are configured to be passed to these stages for processing and counting .
mainly working on fixed function network devices , sflow @xcite , planck @xcite , flexam @xcite , and everflow @xcite are all packet - based mirroring and sampling techniques , through setting flow filters or relying on particular sampling policies .
such techniques can only provide partial data plane visibility with information loss .
netsight @xcite aggressively collects the complete packet history in order to achieve full network visibility with the significant storage and computing cost at control plane . some other work ( e.g. , tpp @xcite , int @xcite , and flowradar @xcite ) builds on programmable network devices ,
therefore non - packet data can be retrieved from data plane through programming means .
tpp revives the idea of active network @xcite but keeps it simple .
it allows packets to carry a tiny program in header so the program instructions can be executed by network devices and data collected along the forwarding path .
int realizes the similar idea by using p4 @xcite as the high level programming language .
flowradar can maintain full flow statistics with succinct data structure .
this group of techniques falls into the static programming category in which the custom functions are predefined at design time . to collect data that were not supported by the current design ,
one has to reprogram the data plane .
some techniques are introduced to enhance the data plane data processing capability .
opensketch @xcite proposes a set of data plane primitives to facilitate the network measurement .
openstate @xcite devises a stateful data plane programming abstraction which can be used to implement fsms .
insp @xcite provides a generic data plane packet generation api which can be used to preprocess and encapsulate collected data from data plane .
dpt @xcite suggests to augment a timestamp header to all packets for various data analytical applications .
interactive control on what data to collect based on network dynamics is discussed in some work ( e.g. , tpp @xcite , dream @xcite , and mozart @xcite ) . however , due to the lack of interactive programming capability , new data can only be collected through selecting existing probes or configuring new flow table entries .
gigascope @xcite and path query @xcite provide sql - like languages for applications to initiate interactive queries for network states and data .
frenetic queries manipulate flow entries with the assumption that each flow entry has its counters and rely on the controller - side runtime system to aggreagte the data collected @xcite .
the high - level and network - wide queries can be compiled into dnps to collect preprocessed data direclty through data plane programming .
while the network data plane is naturally a data streaming system , a data stream management system ( dsms ) @xcite would rely on dnps to realize in - network sampling and sketching techniques .
based on ebpf , io - visor @xcite is a kernel i / o and networking infrastructure which can be used to implement virtual switch .
it supports dynamic tracing by installing virtual probes converted from high level programs at runtime .
this is inline with the idea of dnp .
dnp is enabled by the most recent technology advances : open programmable data plane and interactive data plane programming .
it takes advantage of the data plane processing capability and provides real - time and on - demand network visibility at low cost and high performance .
we believe dnp is a solid stepping stone to a network data analytical platform which can help service providers to gain deeper insight and mine more values from their networks . | effective sdn control relies on the network data collecting capability as well as the quality and timeliness of the data .
as open programmable data plane is becoming a reality , we further enhance it with the support of runtime interactive programming in order to cope with application dynamics , optimize data plane resource allocation , and reduce control - plane processing pressure .
based on the latest technologies , we propose the dynamic network probes ( dnp ) as a means to support real - time and on - demand network visibility .
dnps serve as an important building block of an integrated networking data analytics platform which involves the network data plane as an active component for in - network computing . in this paper , we show the types of dnps and their role in the big picture .
we have implemented an np - based hardware prototype to demonstrate the feasibility and efficiency of dnps .
we lay out the research challenges and our future work to realize the omni network visibility based on dnps . |
it is known that electrostatic forces determine the properties of various systems : physical as well as chemical or biological . in particular , the coulomb interactions are of great importance when dealing with ionic fluids i.e. , fluids consisting of dissociated cations and ions . in most cases
the coulomb interaction is the dominant interaction and due to its long - range character can substantially affect the critical properties and the phase behavior of ionic systems .
thus , the investigations concerning these issues are of great fundamental interest and practical importance . over the last ten years
, both the phase diagrams and the critical behavior of ionic solutions have been intensively studied using both experimental and theoretical methods .
these studies were stimulated by controversial experimental results , demonstrating the three types of the critical behavior in electrolytes solutions : ( i ) classical ( or mean - field ) and ( ii ) ising - like behavior as well as ( iii ) crossover between the two @xcite . in accordance with these peculiarities
, ionic solutions were conventionally divided into two classes , namely : `` solvophobic '' systems with ising - like critical behavior in which coulomb forces are not supposed to play a major role ( the solvent is generally characterized by high dielectric constant ) and `` coulombic '' systems in which the phase separation is primarily driven by coulomb interactions ( the solvent is characterized by low dielectric constant )
. hence the criticality of the coulombic systems became a challenge for theory and experiment .
a theoretical model which demonstrates the phase separation driven exclusively by coulombic forces is a restricted primitive model ( rpm ) @xcite . in this model
the ionic fluid is described as an electroneutral binary mixture of charged hard spheres of equal diameter immersed in a structureless dielectric continuum .
early studies @xcite established that the model has a gas - liquid phase transition . a reasonable theoretical description of the critical point in the rpm was accomplished at a mean - field ( mf ) level using integral equation methods @xcite and debye - hckel theory @xcite . due to controversial experimental findings
, the critical behavior of the rpm has been under active debates @xcite and strong evidence for an ising universal class has been found by recent simulations @xcite and theoretical @xcite studies . in spite of significant progress in this field ,
the criticality of ionic systems are far from being completely understood .
the investigation of more complex models is very important in understanding the nature of critical behavior of real ionic fluids demonstrating both the charge and size asymmetry as well as other complexities such as short - range attraction .
a description of a crossover region when the critical point is approached is of particular interest for such models .
based on the experimental findings one can suggest that in ionic fluids the temperature interval of crossover regime , characterized by the ginzburg temperature , is much smaller than observed in nonionic systems @xcite .
in particular , a sharp crossover was reported for the systems @xmath2 @xcite ( see also @xcite ) .
the analysis of experimental data for various ionic solutions confirmed that such systems generally exhibit crossover or , at least a tendency to crossover from the ising behavior asymptotically close to the critical point , to the mean - field behavior upon increasing distance from the critical point @xcite .
moreover , the systematic experimental investigations of the ionic systems such as tetra-@xmath3-butylammonium picrate , @xmath4 , ( for tetra-@xmath3-butylammonium picrate we will follow the notations from @xcite ) in long chain @xmath3-alkanols with dielectric constant ranging from @xmath5 for @xmath6-tetradecanol to @xmath7 for @xmath8-propanol suggest an increasing tendency for crossover to the mean - field behavior when the coulomb contribution becomes essential @xcite .
they also indicate that the `` coulomb limit '' reduced temperature of the rpm @xmath9 is valid for the almost non - polar long chain alkanols @xcite .
it has been stressed @xcite that for solutions of @xmath4 in @xmath6-alkanols , the upper critical solution points are found to increase linearly with the chain length of the alcohols ( that corresponds to the decrease of dielectric constant of the solvent ) .
the experimental data for the critical points and the dielectric permittivities for solutions of @xmath4 in @xmath6-alkanols are given in table 1 @xcite .
.the experimental parameters of the critical points ( critical temperature @xmath10 , critical mass fraction @xmath11 ) and the corresponding dielectric constants @xmath12 for solutions of @xmath4 in @xmath6-alkanols @xcite . [ cols="^,^,^,^",options="header " , ]
in this paper we study the reduced ginzburg temperature as a function of the interplay between the short- and long - range interactions .
the ionic fluid is modelled as a charge asymmetric continuous system that includes additional short - range attractions . the model without coulomb interactions exhibits a gas - liquid critical point belonging to the ising class of criticality .
we derive an effective glw hamiltonian for the model whose coefficients have the form of an expansion in powers of the ionicity . using these coefficients
we calculate a ginzburg temperature depending on the ionicity . to this end
we introduce a specific model which consists of charged hard spheres of the same diameter interacting through the additional square - well potentials . to study the effect of the interplay between short- and long - range interactions we change , besides the ionicity , the range of the square - well potential .
as a result , we obtain the similar tendency for the reduced ginzburg temperature as in @xcite when the region of the short - range attraction increases i.e. , its nonmonotonic character but with different numerical characteristics
. however , our results demonstrate a much faster decrease of the ginzburg temperature when the ionicity increases . we found a good qualitative and sufficient quantitative agreement with the experimental findings for @xmath4 in @xmath3-alkanols .
this confirms the experimental observations that an interplay between the solvophobic and coulomb interactions alters the temperature region of the crossover regime i.e. , the increase of the ionicity that can be related to the decrease of dielectric constant leads to the decrease of the crossover region .
we suggest that the quantitative discrepancy of the results for @xmath0 obtained in @xcite and in this work could be due to the fact , besides the difference in the symmetry of the two models , that the chemical potential ( or density ) dependence of the hamiltonian coefficients was taken into account explicitly in our case .
it should be noted that in the approximation considered in this paper only the critical chemical potential depends explicitly on the charge magnitude . in order to obtain the charge dependence of the other quantities terms of order higher than @xmath13 should be taken into account into the effective hamiltonian .
finally , we emphasize that the functional representation ( [ 2.5])-([action ] ) allows to consider more complicated models in particular models including charge and size asymmetry .
@xmath14 ^{2}{\widetilde g}_{n-2}(k_{1},k_{2},\ldots,\vert{\mathbf k}_{n-2}+{\mathbf k}_{n-1}+{\mathbf k}_{n}\vert ) \right .
\nonumber \\ & & \left.+\left(q_{\alpha}^{4}c_{\alpha } -3\left[q_{\alpha}^{2}c_{\alpha}\right ] ^{2}\right){\widetilde g}_{n-3}(k_{1 } , k_{2},\ldots,\vert{\mathbf k}_{n-3}+\ldots+{\mathbf k}_{n}\vert)\right\rbrace , % \label{3.14f}\end{aligned}\ ] ] [ [ the - nth - particle - structure - factors - of - a - one - component - hard - sphere - systems - in - the - percus - yevick - approximation ] ] the nth - particle structure factors of a one component hard sphere systems in the percus - yevick approximation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ let us write the ornstein - zernike equation in the fourier space @xmath18 where @xmath19 is the fourier transform of the ornstein - zernike direct correlation function @xcite we have for @xmath19 in the percus - yevick approximation @xcite @xmath20 where @xmath21 from ( [ s-2 ] ) and ( [ c-2 ] ) we get for @xmath22 @xmath23 taking into account ( [ phi - sr - k ] ) we have @xmath24 . as a result
, @xmath17 is as follows @xmath25 where @xmath26 is given by ( [ t-01 ] ) .
[ [ explicit - expressions - for - the - integrals - used - in - equations-2.21a-2.21e ] ] explicit expressions for the integrals used in equations ( [ 2.21a])-([2.21e ] ) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ j.m.h .
levelt sengers and j. a. given , mol . phys . *
80 * , 899 ( 1993 ) .
pitzer , j. phys . chem .
* 99 * , 13070 ( 1995 ) .
k. gutkowski , m.a .
anisimov and j.v .
sengers , j. chem .
phys . * 114 * 3133 ( 2001 ) .
w. schrer and h. weigrtner , pure appl . chem . *
76 * , 19 ( 2004 ) .
w. schrer , in _ ionic soft matter : modern trends and applications _ , edited by d. henderson et al ( dordrecht : nato asi series ii , springer , 2005 ) , p. 143 .
fisher , j. stat . phys . * 75 * , 1 ( 1994 ) . g. stell , j. stat .
phys . * 78 * , 197 ( 1995 ) .
f.h . stillinger and r. lovett , j. chem . phys .
* 48 * , 3858 ( 1968 ) .
vorontsov - veliaminov , a.m. elyashevich , l.a .
morgenshtern and v.p .
chasovshikh , teplofiz .
vysokikh temp.*8 * , 277 ( 1970 ) .
g. stell , k.c .
wu and b. larsen , phys .
lett . * 37 * , 1369 ( 1976 ) .
y. zhou , s. yeh and g. stell , j. chem .
phys.*102*,5785 ( 1995 ) .
y. levin and m.e .
fisher , physica a * 225 * , 164 ( 1996 ) .
fisher and b.p .
lee , phys .
lett . * 77 * , 3361 ( 1996 ) .
w. schrer and v.c .
weiss , j. chem .
phys . * 109 * , 8504 ( 1998 ) .
leote de carvalho and r. evans , j. phys .
: condensed matter * 7 * , 575 ( 1995 ) .
caillol , d. levesque and j .- j .
weis , j. chem .
phys . * 107 * , 1565 ( 1997 ) .
j. valleau and g. torrie , j. chem .
phys . * 108 * , 5169 ( 1998 ) .
p.j . camp and g.n .
patey , j. chem . phys . * 114 * , 399 ( 2001 ) .
e. luijten , m.e .
fisher and a.z .
panagiotopoulos , j. chem . phys . * 114 * , 5468 ( 2001 ) .
caillol , d. levesque and j .- j .
weis , j. chem . phys . * 116 * , 10794 ( 2002 ) .
patsahan , condens .
matter phys.*7 * , 35 ( 2004 ) .
patsahan , i. mryglod and j .- m .
caillol , j. phys . : condens .
matter * 17 * , l251 ( 2005 ) .
e. luijten , m.e .
fisher and a.z .
panagiotopoulos , phys .
lett . * 88 * , 185701 ( 2002 ) .
kim and m.e .
fisher , phys . rev .
lett . * 92 * , 185703 ( 2004 ) .
a. ciach and g. stell , j. mol .
liq . * 87 * , 253 ( 2000 ) .
a. ciach and g. stell , int .
b * 21 * , 3309 ( 2005 ) .
a. ciach , phys .
e * 73 * , 066110 ( 2006 ) .
o. patsahan and i. mryglod , j. phys . : condens .
matter * 16 * , l235 ( 2004 ) .
p.chieux , m.j .
sienko , j. chem . phys . * 53 * , 566 ( 1970 ) . t.narayanan and k.s .
pitzer , j. phys . chem . * 98 * , 9170 ( 1994 ) .
t.narayanan and k.s .
pitzer , j. chem . phys . * 102 * , 8118 ( 1995 ) . m.a .
anisimov , j.jacob , a. kumar , v.a .
agayan and j.v .
sengers , phys .
lett . * 85 * , 2336 ( 2000 ) .
k. gutkowskii , m.a .
anisimov and j.v .
sengers , j. chem . phys .
* 114 * , 3133 ( 2001 ) .
m. kleemeier , s. wiegand , w. schrer and h. weigrtner , j. chem . phys . * 110 * , 3085 ( 1999 ) .
a.g . moreira , m.m . telo de gama and m.e .
fisher , j. chem . phys . * 110 * , 10058 ( 1999 ) .
mcquarrie , _ statistical mechanics _ ( harper - collins , new york , 1976 ) .
j.d . weeks , d. chandler and h.c .
andersen , j. chem . phys . * 54 * , 5237 ( 1971 ) .
g. stell , in _ phase transitionss and critical phenomena _
, * 5b * , edited by c. domb and m.s .
green ( academic press , new york , 1975 ) .
hansen and i.r .
mcdonald , _ theory of simple liquids _
( academic press , 1986 ) .
n.w . ashcroft and j. lekner , phys .
b * 83 * , 5237 ( 1966 ) . | according to extensive experimental findings , the ginzburg temperature @xmath0 for ionic fluids differs substantially from that of nonionic fluids [ schrer w. , weigrtner h. 2004 _ pure appl .
chem . _ * 76 * 19 ] .
a theoretical investigation of this outcome is proposed here by a mean field analysis of the interplay of short and long range interactions on the value of @xmath0 .
we consider a quite general continuous charge - asymmetric model made of charged hard spheres with additional short - range interactions ( without electrostatic interactions the model belongs to the same universality class as the @xmath1 ising model ) .
the effective landau - ginzburg hamiltonian of the full system near its gas - liquid critical point is derived from which the ginzburg temperature is calculated as a function of the ionicity .
the results obtained in this way for @xmath0 are in good qualitative and sufficient quantitative agreement with available experimental data . |
during last two decades the experimental studies of phase transitions in finite and even in small systems are inspiring the high interest to their rigorous theoretical treatment @xcite .
one of the main reasons for such an interest is that the nuclear systems which are experimentally studied at intermediate @xcite and at high @xcite collision energies have no thermodynamic limit due to the presence of the long range coulomb interaction and , hence , in a strict thermodynamical sense the phase transition in such systems is not defined .
therefore , the practical need to formulate the reliable experimental signals of phase transformations in the systems consisting of a few hundreds or thousands particles lead researchers to development of the non - traditional statistical methods which may be suited to small systems @xcite
. one of such directions of research , the concept of bimodality , is based on the old t. l. hill idea @xcite that in finite system the interface between two pure phases costs " additional free energy and , hence , their coexistence is suppressed .
the practical conclusion coming out of this idea is that the resulting distribution of the order parameter should demonstrate a bimodal behavior and , hence , each maximum or peak of the bimodal distribution has to be associated with the pure phase @xcite .
although in @xcite the authors claimed to establish the one - to - one correspondence between the bimodal structure of specially constructed partition of some measurable quantity , known on average , and the properties of the lee - yang zeros @xcite of this partition in the complex plane of fugacity of this measurable quantity , there appeared certain doubts @xcite about so strict relation between the bimodal - like mass / charge distributions of nuclear fragments and the liquid - gas nuclear phase transition in finite systems . in particular , the doubts follow from the experimental analysis of high momentum decay channels in nucleus - nucleus collisions @xcite and from the exact analytical solutions of the simplified statistical multifragmentation model @xcite found for finite @xcite and infinite @xcite volumes .
two explicit counterexamples suggest that the bimodal mass distribution of nuclear fragments appears in a finite volume analog of gaseous phase @xcite with positive surface tension and that it can also occur in the thermodynamic limit at supercritical temperatures @xcite , if at such temperatures there exists negative surface tension . despite the existing counterexamples @xcite , up to now it is unclear , whether the statistical systems without phase transition can , in principle , generate the bimodal size distributions of their constituents for finite volumes . in order to demonstrate that this is , indeed
, the case and the statistical systems in finite volume can generate the bimodal size distributions , here we analytically calculate the size distribution of geometrical spin clusters of the simplest statistical model which has no phase transition , the one dimensional ising model @xcite .
another principal purpose of this work is to further develop connections between the spin models and cluster models .
the cluster models are successfully applied to study phase transitions in simple liquids @xcite , in nuclear matter @xcite and in the quark - gluon plasma phenomenology @xcite .
their basic assumption is that the properties of studied system consisting from the elementary degrees of freedom ( molecules , nucleons , partons , etc . )
can be successfully described in terms of physical clusters ( vapor droplets , nucleus , hadrons , etc . )
formed by any natural number of the elementary degrees of freedom . since the cluster models proved to be a successful tool to investigate the phase transition mechanisms both in finite and in infinite systems
, it seems that the development of equivalent cluster formulation for the well - known spin models and the determination of the properties of their physical clusters is a theoretical task of high priority .
it is necessary to mention that some results on such a connection were already reported in @xcite , in which a parameterization of the fisher liquid droplet model @xcite was successfully applied to the numerical description of cluster multiplicities of two- and three - dimensional ising model .
however , below we implement this program to the one dimensional ising model and directly calculate the size distribution of the geometrical spin clusters from the corresponding partition function .
the work is organized as follows . in the next section
we explain the necessary mathematical aspects of calculating the spin cluster size distribution . then , in sect .
3 the derived expression is analyzed in details and the physical origin of bimodality in the present model is discussed .
the results of our numerical simulations are also presented in this section .
finally , our conclusions and perspectives of further research are summarized in sect .
here we use the geometrical definition of clusters which seems to be the most natural for lattice models . within this framework the monomers , the dimers , the trimers and so on are built from the neighboring spins of the same direction .
this approach is useful for different lattice systems @xcite .
a high level of its generality allows one to study the su(2 ) and su(3 ) gauge lattice models in terms of the clusters composed from the polyakov loops @xcite .
let s consider the one dimensional ising model in which spins are arranged in a linear lattice of size @xmath2 and they can be in two states only ( up or down ) .
interaction energy of two neighboring spins is proportional to their product .
hence , the neighboring pair of parallel spins contribute @xmath3 to the total energy of the system , whereas the corresponding contribution of the neighboring pair of antiparallel spins is @xmath4 . for a convenience
we choose the following boundary conditions : two edge spins have only a single neighbor which is able to interact with them .
evidently , in the present model there are only two types of clusters , i.e. the clusters composed of spins up and of spins down .
hence , for a convenience hereafter they are called the up clusters and the down clusters and are marked with the superscripts @xmath5 and @xmath6 , respectively . the cluster multiplicities are defined as their occupancy numbers @xmath7 and @xmath8 , where @xmath9 is the cluster size .
it is clear that size of the maximal cluster can not exceed @xmath2 and , hence , only the spin configurations obeying the condition @xmath10 can be realized in the considered system .
the total numbers of up and down spin clusters are , respectively , denoted as @xmath11 and @xmath12 . hereafter , it is assumed that the blind index of cluster size runs over all positive integers . in
what follows the set of all occupancy numbers @xmath13 is called as a microscopic configuration .
each microscopic configuration @xmath14 of the system defines its total energy , which receives a contribution @xmath15 from every up cluster or from every down cluster of size @xmath9 .
in addition , there exist the contacts of neighboring clusters with opposite spins and their number is @xmath16 .
evidently , any such a contact gives a contribution @xmath4 to the system energy .
therefore , using the definitions of @xmath17 and @xmath18 we can cast the total energy of the system for a given microscopic configuration @xmath14 as @xmath19 .
% \ ] ] obviously , the total number of spins @xmath20 should coincide with the lattice size @xmath2 . since up and down clusters alternate each other , then the only configurations with @xmath21 can be realized in the considered system .
hence , taking into account the fact that the total number of cluster permutations is @xmath22 we write the degeneracy factor of the microscopic configuration @xmath14 as @xmath23 where @xmath24 is the kronecker delta - symbol .
note that every configurations with @xmath25 should have a degeneracy 2 , since it can be `` constructed '' only from the configurations @xmath26 and @xmath27 by combining one of their outer clusters with any other cluster of the spin sign which coincides with the one of outer cluster .
the @xmath28 statistical moment of the down cluster occupancy number we define as @xmath29^q g_\sigma \ , e^{-\frac{h_\sigma}{t}+a\beta(n_\sigma - n ) } , % \ ] ] where @xmath30 is the temperature and @xmath31 is the lattice constant .
a summation over the configurations @xmath14 in ( [ iii ] ) should be understood as a summation over all nonnegative integers @xmath32 and @xmath33 .
the dimensionless parameter @xmath34 is introduced into ( [ iii ] ) for the convenience of analytical calculations and its value is chosen to satisfy the conditions @xmath35 and @xmath36 . however
, this parameter does not affect any observable due to the condition @xmath37 provided by the kronecker delta - symbols in eq .
( [ ii ] ) .
note , that @xmath38 yields the partition function @xmath39 , whereas a symmetry between up and down spins leads to the equality @xmath40 . using eqs .
( [ i ] ) and ( [ ii ] ) , the definition of @xmath41 along with the integral representation of the kronecker delta - symbol @xmath42 and of the factorials @xmath43 we can write @xmath44 now the summations over @xmath7 and @xmath45 in ( [ iv ] ) can be performed trivially giving the corresponding exponential functions . in what follows we are mainly interested in finding the statistical moments of @xmath46 of the zeroth and first order . for @xmath47
the summation over @xmath48 generates an exponential function multiplied by the factor @xmath49 .
then , each product over @xmath9 in ( [ iv ] ) can be carried out , since it is equivalent to a summation of decreasing geometric progressions in the exponential .
in addition , the chosen range of @xmath34 provides a convergence of the integrals over @xmath50 and @xmath51 variables due to negative real part of these exponential functions .
hence , for @xmath38 and @xmath52 one can easily find @xmath53 where the real variables of integration @xmath54 and @xmath55 are now changed to the complex ones @xmath56 and @xmath57 .
the integrand with respect to the variable @xmath58 in eq .
( [ v ] ) contains three poles .
the first one @xmath59 exists for @xmath38 only .
however , the contribution of this pole does not produce any singularity with respect to the @xmath60 variable and , hence , it can be safely neglected . the second pole existing at any value of @xmath61 corresponds to @xmath62 .
it does not contribute to @xmath63 since it is located out of the contour @xmath64 due to the chosen value of the regularization parameter @xmath34 .
thus , accounting only for the simple pole @xmath65 we obtain @xmath66 this expression clearly demonstrates that for @xmath67 and @xmath68 the integral function has a specific simple pole @xmath69 which , as we discuss below , is responsible for the non - monotonic behavior of @xmath70 .
taking @xmath38 we immediately reproduce the well - known partition function of the one dimensional ising model with free edges @xcite , i.e. @xmath71 for @xmath72 the average occupancy number of the down clusters is obtained from eq .
( [ vi ] ) for @xmath67 @xmath73 whereas @xmath74 .
note , that @xmath75 for all @xmath76 .
the direct calculation based on this expression demonstrates that @xmath77 which is an obvious consequence of the symmetry between the spins up and down , if the external magnetic field is absent . an explicit expression ( [ viii ] ) for @xmath70 allows us to find the size distribution function of clusters @xmath78 which is proportional to their average occupancy number . since the size distribution functions are the same for both kinds of clusters , we do distinguish them .
the normalized distribution @xmath78 is obtained from the condition that the total probability to find up ( or down ) cluster of any size is @xmath79 , i. e. @xmath80 .
this condition follows from the symmetry between up and down spins .
hence , we obtain @xmath81 and @xmath82 . from eq .
( [ ix ] ) one can see that @xmath78 has not only the exponential part @xmath83 , but also it includes the factor which is linear in @xmath76 .
the power part of the size distribution functions is known in several statistical models of cluster type .
traditionally it is taken into account by the fisher topological exponent @xmath84 as @xmath85 @xcite .
however , the linear @xmath76-part of the size distribution function ( [ ix ] ) disappears in the thermodynamic limit @xmath86 , since in this limit one finds @xmath87 .
hence , it is interesting to analyze the distribution @xmath78 of other lattice models in order to find possible restrictions on values of the fisher index @xmath84 .
the growing interest to a bimodality is caused by the widespread belief that it can serve as a robust signal of the first order phase transition in finite systems . as it was mention above this concept
is based on t. l. hill idea @xcite that , each peak of the bimodal distribution is associated with a pure phase .
however , this idea can not be sufficiently justified , since in finite systems the analog of mixed phase is not just a mixture of two pure phases @xcite . according to exact analytical solutions found for several cluster models in finite systems @xcite the finite volume analog of mixed phase is represented by a combination of an analog of gaseous phase , which is stable , and some even number of different metastable states .
the number of metastable states is determined by thermodynamic parameters , of course , but the whole point is that their identification with the help of a single peak ( or maximum ) of the size distribution function does not seem to be a well defined procedure @xcite .
having at hand an explicit expression for cluster size distribution we are able to answer the question what is the reason for a bimodal behavior . from eq .
( [ ix ] ) it follows that @xmath88 for finite @xmath2 and for @xmath89 ( a ferromagnetic system at zero temperature ) , whereas the limit @xmath90 ( an antiferromagnetic system at zero temperature ) generates @xmath91 for finite @xmath2 .
therefore , we conclude that at some intermediate value of the lattice constant @xmath0 the monomer dominant regime should switch to the regime of dominance of the cluster of maximal size . this switching is characterized by the comparable probabilities to find the monomers ( dimers , trimers , etc . ) and the cluster of maximal size and , hence , at some value of @xmath0 the distribution function @xmath78 becomes a non - monotonic one .
a direct inspection of eq . ( [ ix ] ) shows that for @xmath92 and finite @xmath2 the derivative @xmath93 is always negative , i.e. @xmath94 and , hence , at this interval of cluster sizes the function @xmath78 is a monotonic one .
the non - monotonic behavior of the distribution @xmath78 is caused by presence of the kronecker delta - symbol @xmath95 .
it is clear that the cluster size distribution is non - monotonic , if the inequality @xmath96 is obeyed or , equivalently , if @xmath97 this inequality can be fulfilled for @xmath98 , only .
this means that the non - monotonicity of @xmath78 in the one dimensional ising model can appear in the ferromagnetic case only .
it is necessary to stress that the value of @xmath1 does not depend on @xmath2 and @xmath4 and , hence , it is a universal constant .
thus , in the ferromagnetic case the found cluster distribution function @xmath78 is non - monotonic ( bimodal ) for any size of the lattice and any value of the spin coupling .
however , we should note that the bimodal behavior of the cluster size distribution is washed out in the thermodynamic limit @xmath86 , since in this case @xmath99 for all finite values of the lattice constant @xmath0 .
the mathematical reason of the non - monotonic behavior of cluster size distribution in the present model is clear now .
although in this model the bimodality phenomenon appears as the finite size effect , we would like to add a few words on its substantial origin . for this purpose let us consider an auxiliary infinite lattice with the cluster distribution function @xmath100 which for all @xmath76 obeying inequality @xmath101
is defined as @xmath102 where the parameter @xmath103 provides the nonnegative value of @xmath104 .
for @xmath105 the distribution ( [ x ] ) describes clusters of almost all sizes ( up to @xmath106 ) on an equal footing .
note , that in eq .
( [ x ] ) the quantity @xmath2 is treated as a finite parameter .
suppose now that we randomly choose on this infinite lattice the infinite amount of intervals of the length @xmath2 each .
it is clear that the size distribution of small clusters will be given by eq .
( [ x ] ) , whereas the clusters with @xmath107 will increase the occupancy number of the cluster of maximal size ( they will also contribute to the smaller clusters , but this fact is already taken into account in ( [ ix ] ) ) .
this can be easily seen , if one sums up all values of @xmath100 for @xmath107 @xmath108 although in the sum above the terms with @xmath109 are negative , their contribution is practically negligible due to the adopted assumption that @xmath105 .
the result of summation is nothing else , but the term in @xmath110 of eq .
( [ ix ] ) which is proportional to the kronecker delta - symbol .
( [ xi ] ) demonstrates that in a finite system the clusters of size larger than the lattice size are `` condensing '' into the largest cluster of size @xmath2 , whereas their probabilities @xmath111 are `` absorbed '' into @xmath110 . ) ( triangles ) and found numerically for @xmath112 and @xmath113 ( squares and circles ) for three values of the lattice constant @xmath0 . for @xmath114 ( upper panel ) @xmath78 monotonically decreases with the cluster size growth , while for @xmath115 ( middle panel ) and @xmath116 ( lower panel ) it has a maximum at @xmath68 . ] ) ( triangles ) and found numerically for @xmath112 and @xmath113 ( squares and circles ) for three values of the lattice constant @xmath0 . for @xmath114 ( upper panel ) @xmath78 monotonically decreases with the cluster size growth , while for @xmath115 ( middle panel ) and @xmath116 ( lower panel ) it has a maximum at @xmath68 . ] ) ( triangles ) and found numerically for @xmath112 and @xmath113 ( squares and circles ) for three values of the lattice constant @xmath0 . for @xmath114 ( upper panel )
@xmath78 monotonically decreases with the cluster size growth , while for @xmath115 ( middle panel ) and @xmath116 ( lower panel ) it has a maximum at @xmath68 . ] in order to demonstrate the occurrence of bimodality in the one dimensional ising model we present here the results of our numerical simulations for the lattice sizes @xmath112 and @xmath113 .
these simulations were made using the swendsen - wang algorithm .
for each run the first @xmath117 lattice configurations were discarded to ensure a complete thermalization , and the next @xmath118 lattice measurements were made while discarding every 5 lattice configurations between the measurements . for each measurement the total expectation of the number of clusters of all sizes was calculated , then the jackknife error analysis was used to obtain the error estimate of the mean values obtained .
our numerical study includes 14 values of the lattice constant in the range from 0 to 2.5 .
we would like to stress that the coincidence of distribution functions simulated numerically and calculated according to formula ( @xmath119 ) is perfect . in fig .
[ fig ] the cluster size distribution is shown for three values of the lattice constant @xmath0 .
for @xmath120 the cluster size distribution is monotonic for both values of @xmath2 . at the same time for the values of lattice constant above the `` critical '' one , i.e. for @xmath121 and for @xmath122 , one can clearly see the bimodal distribution in fig .
it is also seen that bimodality is enhanced with the lattice constant increasing .
we have to remind once again that the present model does not have any phase transition in the thermodynamic limit .
hence , the one dimensional ising model gives an explicit counterexample to treating bimodality as a signal of phase transition in finite systems .
in this work we found an analytical solution of the one dimensional ising model in terms of the geometrical clusters composed of the neighboring spins of the same direction .
it is clear that the present formulation is an important step towards establishing a firm connection between the lattice spin systems and the statistical models of cluster type .
the developed approach allows us to exactly calculate the size distribution of spin clusters for finite and infinite lattice sizes . using the exact formulae we showed that for finite size of the lattice the one dimensional ising model has a bimodal size distribution of clusters for @xmath123 .
we demonstrate that the bimodal size distribution of spin clusters in the present model is a finite size effect which appears due to a `` condensation '' of clusters whose size exceeds the lattice extent to the largest cluster .
such a `` condensation '' to the maximal cluster leads to the non - monotonic size distribution at sufficiently large values of the lattice constant in ferromagnetic case only .
the obtained result is valid for all sizes of the lattice , however , in the thermodynamic limit it is washed out .
it provides us with an explicit counterexample to the widely spread belief about an exclusive role of bimodality as a signal of phase transition in finite systems @xcite .
our criticism is in full coherence with the results of refs .
the developed formalism considers the simplest exactly solvable lattice model , but it would be interesting to apply it to more realistic physical systems .
+ + * acknowledgments . *
the authors appreciate the fruitful discussions with k. a. bugaev and v. k. petrov and their valuable comments .
the work of a. i. i. is supported in part by the national academy of sciences of ukraine grant of grid simulations for high energy physics .
k. a. bugaev , a. i. ivanytskyi , e. g. nikonov , v. k. petrov , a. s. sorin and g. m. zinovjev , phys .
, 75 , 6 , 707 ( 2012 ) . i. zakout , c. greiner , j. schaffner - bielich , nucl .
phys . , a 781 , 150 ( 2007 ) . | the size distribution of geometrical spin clusters is exactly found for the one dimensional ising model of finite extent . for the values of lattice constant @xmath0 above some critical value "
@xmath1 the found size distribution demonstrates the non - monotonic behavior with the peak corresponding to the size of largest available cluster . in other words , at high values of lattice constant there are two ways to fill the lattice : either to form a single largest cluster or to create many clusters of small sizes .
this feature closely resembles the well - know bimodal size distribution of clusters which is usually interpreted as a robust signal of the first order liquid - gas phase transition in finite systems .
it is remarkable that the bimodal size distribution of spin clusters appears in the one dimensional ising model of finite size , i.e. in the model which in thermodynamic limit has no phase transition at all .
keywords : phase transitions , small systems , bimodal distribution |
an indirect ( or dipolar ) exciton ( _ ix _ ) consists of a coulomb - correlated electron - hole pair with the individual particles confined in two closely spaced quantum wells ( qws ) in a double quantum well ( dqw ) structure ( cf .
[ figixsetup ] ) subjected to a transverse electric field @xmath1 . @xmath1
controls via the quantum confined stark effect both the _ ix _ energy and radiative lifetime , which can reach the ms - range.@xcite the electric field can be used to laterally confine _ ix _ gases in dqw plane using electrostatic gates @xcite as well as to control the interconversion between _ ix _ and photons .
in addition , the electrical dipoles arising from the charge separation give rise to strong repulsive exciton - exciton interactions @xcite . finally , _
_ ix__s are composite bosons and , therefore , susceptible to the formation of collective exciton phases , as evidenced by the experiments presented in refs . .
these interesting _ ix _ properties have been combined to demonstrate different device functionalities including light storage cells , switches , and transistors @xcite .
the long _ ix _ lifetimes also enable the long - range transport and coupling of _ ix _ packets .
different approaches have also been introduced for the transport of _ _ ix__s based on bare diffusion,@xcite drift induced by repulsive _
interactions,@xcite or by using spatially varying electric fields , including electrostatic ramps @xcite as well as moving electrostatic @xcite and acoustic lattices @xcite .
recently , the confinement of single _ _
ix__s has been demonstrated , @xcite thus opening the way for applications of single _ _ ix__s in quantum information .
@xcite here , the manipulation and transport of _ ix _ spins offers further exciting perspectives . _
_ ix__s have several favorable properties for spin control and manipulation .
the reduced overlap between the electron and hole wave functions reduces not only the radiative lifetime but also the exchange interaction between the exciton - bound electron spins and the short - living hole spins ( the excitonic or bir - aronov - pikus ( bap ) mechanism @xcite ) , thus enhancing the exciton spin lifetime .
@xcite the long lifetime also enables for spin control using the spin - orbit ( so ) coupling . due to the bulk inversion asymmetry ( bia )
intrinsic to the noncentrosymmetric iii - v materials ( such as gaas),@xcite moving spins feel an effective magnetic field ( the dresselhaus field,@xcite @xmath2 ) , which modifies the spin vector during precession .
in addition , the structural inversion anisotropy ( sia ) created by @xmath1 induces a tunable contribution to the so magnetic field ( the rashba field,@xcite @xmath3 ) , thus providing a powerful tool for the dynamic spin control during motion .
@xmath1 also affects spins by changing the @xmath4-factor.@xcite the so fields can , however , also lead to the dephasing of exciton spin ensembles ( the dyakonov - perel dephasing mechanism ) .
the spin dephasing time can be electrically controlled by selecting @xmath1 to modulate the total so field @xmath5.@xcite the _ ix _ spin lifetime also enhances under high particle densities . in analogy with the spin dynamics of atomic condensates,@xcite
the latter has been attributed to the formation of collective _ ix _
phases.@xcite in particular , experiments carried out in this regime have lead to the observation of spin currents in _ ix _ gases with a variety of spatially resolved polarization patterns,@xcite which have been attributed to the so magnetic field .
most of the previous _ ix _ spin studies have addressed the spin dynamics very close to the spin excitation area ( i.e. , within distances @xmath6 m ) , one exception being the self - organized extended ( i.e. , @xmath7 m ) spin patterns reported for the collective regime at high _ ix _
densities.@xcite in this work , we demonstrate that optically injected _
ix _ spins can be transported over distances exceeding 20 @xmath8 m from the injection spot .
the spins coherently precess during motion : the spatial precession period depends on the motion direction and can be controlled by the vertical field @xmath1 applied across the dqw .
this precession behavior , which does not depend on _ ix _ density , is attributed to the precession of the electron spins in the anisotropic so field . from the dependence of the spin polarization on @xmath1 , motion direction , and external magnetic field @xmath9 we have determined the _ ix _ spin lifetime as well as the absolute magnitude and orientation of the so fields in the dqw structures .
in contrast to the spin patterns in collective regime addressed in ref .
@xcite , we show that transport and manipulation of _ ix _ ensembles reported here takes place in the drift - diffusive regime and can be potentially extended down to the single spin level . in the following ,
we first describe the sample structure and the photoluminescence ( pl ) technique used to probe the transport of excitons spins . we then present experimental results demonstrating the transport and coherent precession of ix spins in the so field together with detailed studies of the spin dynamics as a function of propagation direction , bias , and external magnetic fields .
these results on spin transport and so effects are compared to the state - of - the art in the final sections of the manuscript . )
induced by a bias @xmath10 applied between the doped substrate and a semitransparent top electrode ( ste ) . _
ix _ spin transport is mapped by imaging the ix photoluminescence ( pl ) emitted along direction @xmath11 with polarization sensitivity . during motion along @xmath11 , the spins precess under the spin - orbit ( @xmath12 ) and external magnetic field ( @xmath13 ) .
( b ) energy band diagram of the gaas/(al , ga)as dqw along the vertical ( @xmath14 ) direction showing the direct ( _ dx _ ) and _ ix _ transitions under the electric field @xmath1 .
( c ) spatially and spectrally resolved images of the pl with right ( @xmath15 , upper panel ) and left ( @xmath16 , lower panel ) circular polarizations recorded under a bias @xmath17 v and a laser power of @xmath18w .
_ dx _ and _ ix _ denote the emission from direct and indirect excitons in the dqw structure , while the line at 1.515 ev arises from excitons in the gaas substrate . ]
the samples used in the experiment contain three sets of asymmetric gaas dqws grown by molecular beam epitaxy on a @xmath19-doped gaas(001 ) substrate [ cf . figs .
[ figixsetup](a)-(b ) ] .
each dqw is formed by a @xmath20 nm ( qw@xmath21 ) and a @xmath22 nm - thick qw ( qw@xmath23 ) separated by a thin ( 4 nm - thick ) @xmath24 barrier .
multiple dqws were used in order to enhance the photon yield of the pl experiments .
the different qw widths enable the selective generation of direct excitons in either qw@xmath21 or qw@xmath23 by the appropriate selection of the excitation wavelength @xcite .
the electric field @xmath25 for the formation of _ _ ix__s was induced by a bias voltage @xmath10 applied between the n - doped substrate and a thin ( 10 nm - thick ) semitransparent ti electrode ( ste ) deposited on the sample surface .
the spectroscopic experiments were carried out in a he bath cryostat at 2 k. excitons were generated using a right circularly polarized beam from a cw ti : sapphire laser resonant with the electron - heavy hole direct exciton ( dx ) transition of the wider ( qw@xmath23 ) qw .
the laser was focused onto a spot with a diameter of 2.5 @xmath8 m using a microscope objective .
the pl around the laser spot was collected by the same objective , analyzed regarding the circular polarization , and then imaged on the entrance slit of a spectrometer connected to a cooled charge - coupled - device ( ccd ) camera .
the detected ccd images contain information about the energy ( direction perpendicular to the slit ) and spatial distribution of the pl intensity ( along the slit ) .
the spatial ( coordinate @xmath26 ) dependence of the spin polarization was calculated according to @xmath27/\left[i_{\sigma^+}(r)+i_{\sigma^-}(r)\right]$ ] , where @xmath28 ( @xmath29 ) denote the pl intensity at @xmath26 with right ( left ) circular polarization .
figure [ figixsetup](c ) compares pl images with right ( @xmath15 ) and left ( @xmath16 ) polarization recorded with the spectrometer slit aligned with the sample crystallographic direction @xmath30 $ ] .
the emission patterns at 1.515 ev and 1.525 ev arise , respectively , from excitons in the gaas substrate and from the low energy tail of dxs in qw@xmath23 . in both cases ,
the emission is restricted to a region close to the excitation spot , thus indicating negligible diffusion distances for these excitations .
the line at 1.520 ev is associated with the emission of _
_ ix__s in the dqw structure , which in this case was subject to a bias @xmath17 v. the emission extends to radii exceeding 40 @xmath8 m - these long transport distances arise from the fact that the ixs are driven not only by diffusion but also by the drift forces induced by repulsive exciton - exciton dipolar interactions at the areas of high exciton density close to the excitation area .
this assertion is supported by the strong dependence of the extension of the _ ix _ pl cloud on the excitation density @xmath31 , as displayed by the dashed and dash - dotted lines in fig .
[ figixso](a ) .
these curves displayed the spatial dependence of the total pl intensity ( i.e. , @xmath32 $ ] determined from pl images for laser excitation powers of @xmath33w ( as in fig . [ figixsetup](c ) ) and 1.6 @xmath8w , respectively .
the characteristic shape of the spatial pl profiles are consistent with the observations of ref .
@xcite , which reported that the _ ix _ pl reduces in the area of hot _ ix _ close to the excitation spot .
furthermore , a closer look to fig . [ figixsetup](c ) also reveals that the _ ix _ emission energy red - shifts as one moves away from the excitation spot .
in fact , the average _ ix _ emission energy at @xmath34 m is @xmath35 mev higher than at @xmath36 m .
this energetic blue - shift is attributed to the strong repulsive _
ix_-_ix _ interaction at the high _ ix _ densities ( estimated @xcite to be of @xmath37 @xmath38 ) close to @xmath39 . the intensity of the circularly polarized _
ix _ emission in fig .
[ figixsetup](c ) is also spatially modulated with the maxima for the @xmath15 polarization corresponding to minima in the @xmath16 polarization .
the symbols in fig .
[ figixso](a ) display the corresponding profiles for the spin polarization @xmath40 .
this spatial pl modulation is attributed to the precession of the _ ix _ electron spins in the effective so magnetic field @xmath12 during their motion away from the excitation spot .
the @xmath40 oscillations extend over a range of radial distances shorter than the radius of the emission cloud .
while the amplitude of the oscillations increases with @xmath31 , their spatial period remains essentially constant .
a similar increase in spin polarization with excitation density was previously reported in ref .
@xcite and attributed to the onset of collective exciton effects .
the fact that the spatial precession period does not change with the excitation density indicates , however , that the spin precession rates under the so field does not depend on the _ ix _ density .
in addition , the larger @xmath40 amplitudes in our case is at least in part due to the shorter expansion times of the _ ix _ cloud resulting from _
ix_-_ix _ interactions under high @xmath31 , which reduces the impact of spin scattering processes .
the solid lines in fig .
[ figixso](a ) are fits of the experimental results to damped oscillations with a spatial dependence given by : @xmath41 here @xmath42 is the spin spatial precession period for propagation along @xmath43 , which is equal to the so precession period @xmath44 in the absence of external magnetic fields .
@xmath45 denotes the effective spin transport decay length .
equation ( [ eqix1 ] ) describes the spin dynamics in a non - degenerate qw at low temperatures ( i.e. , for thermal energies less than the qw confinement energy ) , when spin precession rate @xmath46 becomes proportional to the electron wave vector * k*. under this condition , the spin precession period ( for @xmath47 ) @xmath48 only depends on direction @xmath43 .
the spin decay length @xmath45 , in contrast , depends on temporal dynamics of the carrier motion as well as on the details of the spin scattering processes .
the fits carried out with a single value for @xmath49 ( i.e. , independent of @xmath26 ) displayed in eq .
[ eqix1 ] reproduces very well the measured spin polarization profiles .
, symbols ) and for the average pl intensity ( @xmath50 , dashed and dot - dashed lines ) for exciton transport along the @xmath51 $ ] ( symbols ) recorded for laser excitation densities @xmath52w ( open symbols ) and @xmath18w ( solid symbols ) . and @xmath53 $ ] directions ( open symbols ) .
( b ) corresponding profiles for transport along @xmath54 $ ] recorded for @xmath55w .
the solid lines superimposed on the symbols are fits of the experimental results to eq . [ eqix1 ] .
( c ) in - plane dependence of the dresselhaus ( @xmath56 ) and rashba ( @xmath57 ) contributions to the so effective magnetic field @xmath58 . ] the wave vector ( @xmath59 ) dependence of the so contributions @xmath60 and @xmath61 for a iii - v qw is illustrated in fig .
[ figixso](c ) . here , we use a reference frame defined by the axes @xmath51 $ ] , @xmath54 $ ] , and @xmath62 $ ] .
the associated ( temporal ) electron spin precession frequencies @xmath63 , @xmath64 can be written as a function of the wave vector @xmath59 as:@xcite @xmath65 \quad\textit{and } \label{eqixa1}\ ] ] @xmath66 \label{eqixa2}\ ] ] where @xmath67 and @xmath68 are the dresselhaus and rashba spin constants , respectively
. equations [ eqixa1 ] and [ eqixa2 ] apply for small wave vectors ( i.e. , @xmath69 .
@xmath70 is the effective extension along @xmath14 of the electronic wave function in the qw confining the electrons ( qw@xmath23 in the present case , as determined by the sign of v@xmath71 ) .
@xmath70 is determined by both the qw width and by the penetration depth into the barrier layers .
the corresponding amplitudes of the spatial precession frequencies @xmath72 can be derived from the precession distance required for a spin rotation by @xmath73 .
the following expressions are obtained:@xcite @xmath74 if @xmath75 the resulting spatial precession frequency will be given by @xmath76 , where the sign is defined by the relative orientation of the two fields .
equations [ eqixa1 ] and [ eqix2 ] predicts that the spatial precession period @xmath44 depends both on transport direction and on bias electric field @xmath1 . in order to check the in - plane anisotropy of @xmath77 , we show in fig .
[ figixso](b ) spin polarization profiles recorded for _ ix _ motion along the @xmath78 $ ] surface direction ( i.e. , orthogonal to the one in figs .
[ figixsetup](c ) and [ figixso](b ) ) .
the period of the precession oscillations is much larger than for the motion along the @xmath79 $ ] direction . the larger period , which in this case becomes comparable with the spin decay length @xmath45 , makes the amplitude of the oscillations weaker than in fig .
[ figixso](a ) .
the dependence of @xmath44 on transport direction is in agreement with the diagrams of fig .
[ figixso](c ) : while @xmath80 and @xmath81 add for motion along @xmath79 $ ] , they partially cancel each other for motion along @xmath53 $ ] , thus leading to a weaker so field .
interestingly , while a weaker @xmath77 reduces spin scattering , it does not necessarily improve the visibility of the @xmath40 oscillations since one also has to satisfy the constraint @xmath82 .
on the other hand , a stronger @xmath77 reduces the spatial precession period . in this case
, however , one needs to reduce the excitation spot to dimensions at least a factor of 2 smaller than the oscillation period in order to measure the oscillations . , symbols ) recorded for increasing electric fields @xmath83 applied along the growth direction of the dqw structure .
( b ) dependence of the spatial spin precession frequency @xmath84 on the electric field @xmath25 for transport along @xmath85 $ ] ( solid symbols ) and @xmath54 $ ] ( open symbol ) .
the solid line is a fit of the data along @xmath79 $ ] to @xmath86 ( cf .
[ eqix2 ] ) with @xmath87 ev@xmath88 , @xmath89 ev@xmath90 , and @xmath91 nm . the corresponding inverse spatial precession periods @xmath92 and @xmath93 are illustrated by the dashed lines .
the dot - dashed line shows the expected inverse precession period for motion along @xmath78 $ ] . ] the electric control of the spin precession frequency is demonstrated by the series of profiles recorded for transport along @xmath79 $ ] in fig .
[ figixbias](a ) .
the amplitude @xmath1 of electric field was determined from the applied bias @xmath10 using the procedure described in ref .
the spatial precession period @xmath44 reduces by over 15% with increasing @xmath25 , thus indicating an enhancement of the rashba field @xmath3 as well as of the total so field @xmath77 .
larger changes can be in principle be achieved by increasing the bias across the dqws .
these , however , also reduce the pl yield and may lead to field - induced carrier leakage out of the dqw structure , thus hampering an efficient spin detection.@xcite the measured spatial precession frequencies ( @xmath94 ) are plotted as a function of the transverse electric field in fig .
[ figixbias](b ) .
the solid line is a fit of the data measured along @xmath79 $ ] to the expression : @xmath86 ( cf .
eq . [ eqix2 ] ) , which gives @xmath87 ev@xmath88 , @xmath89 ev@xmath90 , and @xmath91 nm ( the signs for @xmath67 and @xmath68 will be determined below . the corresponding inverse precession periods @xmath92 and @xmath93 for propagation along this direction are illustrated by the dashed lines . for the applied bias ,
indirect excitons are created in the dqw structure with the electrons confined within the wider ( 17 nm wide ) qw .
the value for @xmath70 used in the fits is compatible with an electronic wave function with a penetration depth of 2 nm in the barriers .
as will be further discussed later , these so parameters agree well with values reported in the literature .
the open symbol in fig . [ figixbias](b ) shows the inverse precession period measured for spin propagation along @xmath95 , where the dresselhaus contribution changes in sign relative to the @xmath96 .
the measured value agrees well with the calculated inverse periods @xmath94 for this propagation along this direction ( dash - dotted line ) .
note that the change in relative signs of the rashba and dresselhaus effective magnetic fields can in principle also be attained for a fixed motion direction by simply reversing the bias polarity . for the asymmetric dqws used in the present experiments , however , a reversal in direction of @xmath1 moves the electrons to the narrow qw , thus also changing the magnitude of the dresselhaus contribution .
in addition , while the leakage current is very low for negative biases ( a few na ) , it increases considerably for high positive biases , thus indicating enhanced carrier leakage .
an external magnetic field @xmath97 applied along the dqw plane provides an additional degree of freedom to control the spin vector .
the experiments were carried out under the configuration illustrated in fig .
[ figixsetup](a ) , where one maps the spins moving along @xmath96 under an external field @xmath98 and collinear with @xmath99 .
figure [ figixb](a ) displays @xmath40 profiles along @xmath96 recorded under different external magnetic fields .
large magnetic fields ( @xmath100 mt ) mainly induce a reduction of the overall spin polarization ( hanle effect ) , from which one can extract the spin dephasing time @xmath101 . for that purpose ,
we plot in fig .
[ figixb](b ) the spin ix polarization at the excitation spot @xmath102 as a function of @xmath97 .
the lines superimposed on the data points are fits to the hanle expression : ) .
the curves are stacked for clarity .
( b ) @xmath40 close to the excitation spot as a function of @xmath13 measured for optical excitation densities @xmath33w ( squares ) and @xmath103w ( dots ) .
the dashed lines display fits to eq .
[ eqixhanle ] , from which one extracts spin lifetimes of 6 and 3 ns , respectively .
( c ) dependence of the inverse precession period ( @xmath104 , where @xmath44 and @xmath105 are the precession period around the so and external fields , respectively ) on @xmath13 ( symbols ) .
the lines are the predictions from the diffusion ( @xmath106 ) and drift ( @xmath107 ) models discussed in the text . ]
@xmath108 where @xmath109 is the degree of polarization at the generation point , @xmath110 with @xmath111 the indirect exciton lifetime . by using the conduction band g - factor @xmath112 and considering @xmath113 , one extracts from the fits spin lifetimes of 3 and 6 ns for excitation powers of @xmath114w and @xmath18w , respectively .
these lifetimes are considerably longer than the typical carrier scattering times , thus indicating that the spin dynamics observed here takes place in diffusive ( or strong scattering ) regime . weak magnetic fields , in contrast , mainly affect the spin spatial precession frequency , as indicated by the symbols in fig . [
figixb](c ) .
in contrast to the so field , which reverses with carrier momentum , the external field has a fixed direction and its action on the spin vector only depends on time .
these features can be exploited to determine the absolute orientation of @xmath99 .
when @xmath97 is oriented along @xmath115 ( upper curves for @xmath116 in fig .
[ figixb](a ) ) , the spatial spin precession rate increases ( reduces ) for spins moving along @xmath117 ( @xmath118 ) .
the opposite behavior is observed when the orientation of @xmath9 reverses ( curves for @xmath119 ) . from these results , together with the increase of the spatial precession frequency along the @xmath96 with @xmath120 ( cf .
[ figixbias](a ) , we unambiguously establish that both the rashba and dresselhaus spin orbit fields point towards @xmath95 when spins move along @xmath96 .
the dresselhaus splitting constant @xmath67 in eq .
[ eqixa1 ] is thus positive while the rashba parameter @xmath121 is negative .
the different natures of the two fields allows us to extract information about the precession dynamics during motion and their impact on the precession period @xmath49 . here
, we first note that the different @xmath49 values in fig .
[ figixb](a ) along @xmath117 and @xmath118 implies that both fields ( i.e. , @xmath99 and @xmath97 ) must act simultaneously on the spins during the motion .
this becomes clear if one considers the hypothetical situation , where the spins initially move to the final position @xmath96 within a very short time ( thereby precessing only around @xmath122 ) and then remain trapped there until recombination ( thus precessing only around @xmath97 ) . in this case , @xmath97 would simply add a fixed ( i.e. , position independent ) phase to the spatial oscillations determined by @xmath123 and @xmath49 would be the same for spin motion along both the positive and negative @xmath96 axes , in contrast to the results displayed in fig .
[ figixb](a ) .
since @xmath97 and @xmath123 are collinear in fig . [ figixb](a ) , the inverse precession period can be written as @xmath124 , where @xmath125 is the contribution associated with the external field . in order to account for the dependence of @xmath126 on @xmath97 one
needs information about the spin position @xmath96 as a function of time ( @xmath127 ) .
we consider the following two extreme situations .
first , if the motion is mainly driven by the ix - ix repulsion forces at the excitation region , the carriers are expected to move with an approximately constant velocity @xmath107 , yielding @xmath128 .
this model predicts the linear dependence of @xmath126 on field indicated by the dashed line in fig .
[ figixb](c ) .
the very small values used for @xmath129 m / s , which corresponds to only @xmath1301% of the average exciton thermal velocity at 2 k , probably arise from the fact that diffusion has been completely neglected .
while the behavior for small fields is well accounted for , the model fails to reproduce the deviations from linearity observed as the @xmath131 increases .
the second extreme situation occurs if motion in purely driven by diffusion , in which case @xmath132 , where @xmath106 is the ix diffusion coefficient .
a straight - forward calculation yields @xmath133 , leading to the behavior depicted by the solid line in fig .
[ figixb](c ) .
the calculations were carried out using an exciton diffusion coefficient @xmath134 cm@xmath90/s , which is approximately a factor of 2 larger than the one expected from ref .
@xcite for diffusion in a 15 nm wide qw . despite its simplicity , this approximation reproduces reasonably well the measured behavior for large field amplitudes .
a good fit requires , however , values for @xmath106 higher that those expected from ref .
this becomes necessary to reduce the ix transport ( and precession ) time and , in that way , compensate for the errors arising neglecting of drift effects .
a more precise description of the spin dynamics by taking into account drift and diffusion is presently under study .
the experimental results of the previous section unambiguously show the coherent precession of optically excited spins in the effective magnetic field @xmath99 induced by the so interaction .
the spin precession period depends on the in - plane motion direction - it is essentially independent of the _ ix _ density and can be controlled by external electric and magnetic fields .
the measured spin lifetimes ( of a few ns ) indicate that transport takes place in the drift - diffusive regime .
the spin dynamics is well - described by taking into account the so parameters as well as the g - factor for electrons , thus showing that it is determined by the _ ix _ electron spins .
in fact , the so parameter @xmath87 ev@xmath88 compares well with those determined from electron spin transport in @xmath136,@xcite @xmath137,@xcite and @xmath138 qws @xcite with comparable widths .
the magnitude of the rashba parameter @xmath89 ev@xmath90 , however , lies above the experimental values of @xmath139 and @xmath140 ev@xmath90 of refs . @xcite and @xcite , respectively , as well as
the calculated value of @xmath141 ev@xmath90 from ref .
while further investigations are required to clarify this difference , we speculate that it may be related to the asymmetric barriers in the dqw structure , which make the electronic wave function more susceptible to the applied electric fields .
the coherent transport and precession of electron spins in the so field in the drift - diffusive transport regime has been previously reported for bulk gaas layers @xcite as well as qw structures .
@xcite in these experiments , electron transport has been driven by an external electric ( or piezoelectric field ) : in this way , the spins could be moved over microscopic distances before dephasing ( assumed to be due to the dyakonov - perel mechanism @xcite ) sets in . to our knowledge , however , the coherent precession of spins injected in an undoped qws in the absence of an electric driving field demonstrated here has so far not been reported .
the observation of coherent spin precession in the so field requires small ratios @xmath142 ( cf .
eq . [ eqix1 ] ) . here ,
( @xmath144 is a proportionality constant and the @xmath145 s denote thermal averages over the transport path ) only depends on the strength of the so interaction .
the spin decay length @xmath146 , in contrast , depends on transport time . for purely diffusive motion in the dyakonov - perel regime , @xmath147 , where @xmath148 is the carrier scattering time and @xmath106 the diffusion coefficient , reduces with transport distance @xmath26 .
as a result , the requirement @xmath149 becomes increasingly difficult to be satisfied for long transport distances @xmath26 ( as well as for strong so fields ) . in the present experiments , the expansion forces due to _ ix_-_ix _ repulsive interactions appear to have a fundamental role , since they considerably reduce the transport time , thus increasing @xmath45 and enabling the observation of precession oscillations .
coherent spin patterns resulting from precession in the so field have been reported for macroscopically ordered _ ix _ phases.@xcite the long spin coherence lengths required for their formation have been attributed to the reduced scattering in the ordered state , which enables the exciton transport in the ballistic regime .
the results presented in the previous section show that long spin lifetimes and transport lengths , as well as coherent precession in the so field can also be achieved in the drift - diffusive regime . this behavior is followed occurs over a wide range of exciton densities , thus indicating that it is not associated with collective effects .
interestingly , the spatial extent of the precession oscillations reported here is comparable to the ones for the self - organized spin patterns.@xcite there are additional fundamental differences between the two cases . while the spin pattern observed in ref .
@xcite results from self - organization and does not rely on the initial exciton polarization , the ones presented here appear as a consequence of the coherent transport of optically injected spins .
in addition , the dynamics observed here is associated with electron spin precession , while ref .
@xcite has invoked the simultaneous evolution of hole and electron spins . finally , as in most of the reports of coherent precession of optically injected spins in the so field , the absence of hole spin contributions to the spin dynamics
is attributed to the short hole spin coherence times as well as the spatial separation of the electron and hole wave function in the dqw structure .
since we have excited and probed _ ix _ spins , an interesting question is whether the electrons and hole remain bound during transport .
although we can not present a conclusive evidence for this behavior , we mention that the _ ix _ spins have been excited with small excess energy ( approx .
10 mev ) , thus reducing the probability of exciton dissociation followed by the spatial separation of electrons and holes .
in addition , the _ ix _ spatial profiles ( cf . , e.g. , fig . [ figixsetup](c ) ) do not show the characteristic rings observed when electrons and holes diffuse independently .
we have presented experimental evidence for the coherent precession of exciton spins during transport in the effective magnetic field induced by the so interaction .
the precession dynamics is essentially independent of the ix density , but can be controlled by an external bias ( via changes in the so field ) or magnetic field .
the spin dynamics are well explained by taking into account electron spin precession , with no noticeable hole spin contribution .
the technique used by us allows for the control and discrimination of precession effects arising from the rashba and dresselhaus so contributions , as well as those induced by an external magnetic field . in this way
, we were able to directly determine the absolute value and sign of the spin splitting parameters for the rashba and dresselhaus so mechanisms .
in addition , the spin dynamics in the presence of an external magnetic field yield information about the precession dynamics during transport .
the long spin transport distances as well as the control of the spin vector open the way for the use of ix spins to encode and manipulate quantum information in opto - electronic semiconductor devices .
we thank manfred ramsteiner for discussions and comments on the manuscript .
we are indebted to s. rauwerdink , w. seidel , s. krau , and a. tahraoui for the assistance in the preparation of the samples .
we also acknowledge financial support from the german dfg ( grant sa-598/9 ) .
g. j. schinner , j. repp , e. schubert , a. k. rai , d. reuter , a. d. wieck , a. o. govorov , a. w. holleitner , and j. p. kotthaus . confinement and interaction of single indirect excitons in a voltage - controlled trap formed inside double ingaas quantum wells .
, 110:127403 , mar 2013 .
a. g. winbow , j. r. leonard , m. remeika , y. y. kuznetsova , a. a. high , a. t. hammack , l. v. butov , j. wilkes , a. a. guenther , a. l. ivanov , m. hanson , and a. c. gossard .
electrostatic conveyer for excitons .
, 106:196806 , 2011 .
e. a. burt , r. w. ghrist , c. j. myatt , m. j. holland , e. a. cornell , and c. e. wieman .
coherence , correlations , and collisions : what one learns about bose - einstein condensates from their decay .
, 79:337340 , 1997 .
a. a. high , a. t. hammack , j. r. leonard , sen yang , l. v. butov , t. ostatnick , m. vladimirova , a. v. kavokin , t. c. h. liew , k. l. campman , and a. c. gossard .
spin currents in a coherent exciton gas .
, 110:246403 , jun 2013 .
m. stern , v. garmider , e. segre , m. rappaport , v. umansky , y. levinson , and i. bar - joseph .
photoluminescence ring formation in coupled quantum wells : excitonic versus ambipolar diffusion .
, 101(25):257402 , 2008 .
p. s. eldridge , w. j. h. leyland , p. g. lagoudakis , r. t. harley , r. t. phillips , r. winkler , m. henini , and d. taylor .
rashba spin - splitting of electrons in asymmetric quantum wells .
, 82(4):045317 , jul 2010 .
h. sanada , t. sogawa , h. gotoh , k. onomitsu , m. kohda , j. nitta , and p. v. santos . acoustically induced spin - orbit interactions revealed by two - dimensional imaging of spin transport in gaas .
, 106(21):216602 , may 2011 .
h. sanada , y. kunihashi , h. gotoh , k. onomitsu , m. kohda , j. nitta , p. v. santos , and t. sogawa .
manipulation of mobile spin coherence using magnetic - field - free electron spin resonance . ,
9:280 , march 2013 . | we report on the coherent control and transport of indirect exciton ( ix ) spins in gaas double quantum well ( dqw ) nanostructures .
the spin dynamics was investigated by optically generating spins using a focused , circularly polarized light spot and by probing their spatial distribution using spatially and polarization resolved photoluminescence spectroscopy . optically injected exciton spins precess while diffusing over distances exceeding @xmath0 m from the excitation spot with a spatial precession frequency that depends on the spin transport direction as well as on the bias applied across the dqw structure .
this behavior is attributed to the spin precession in the effective magnetic field induced by the spin - orbit interaction . from the dependence of the spin dynamics on the transport direction , bias and external magnetic fields we directly determined the dresselhaus and rashba spin splitting coefficients for the structure .
the precession dynamics is essentially independent on the ix density , thus indicating that the long spin lifetimes are not associated with ix collective effects .
the latter , together with the negligible contribution of holes to the spin dynamics , are rather attributed to spatial separation of the electron and hole wave functions by the electric field , which reduces the electron - hole exchange interaction .
coherent spin precession over long transport distances as well as the control of the spin vector using electric and magnetic field open the way for the application of _ ix _ spins in the quantum information processing . |
in the fast - developing fields of spintronics , @xcite magnonics , @xcite and spin caloritronics , @xcite many intriguing effects have been observed .
one of the most interesting discoveries is the spin seebeck effect ( see ) , which is a phenomenon that a temperature bias can drive a pure spin current .
it has been identified in many kinds of materials , including magnetic metals , @xcite magnetic semiconductors , @xcite magnetic insulators , @xcite and nonmagnetic materials with spin - orbit coupling .
@xcite the real breakthrough is made by the observation of sse in magnetic insulators , @xcite which has clearly uncovered that , distinct from spin - dependent seebeck effect in metallic materials , sse possesses the unique ability to generate a pure flow of spin angular momentum by mere thermal excitations without moving charge carriers .
the thermal - generated pure spin current is carried by excitations of the magnetization , known as magnons , instead of by moving charges .
this is an advantage because charge carriers are often problematic for the thermal design of devices , of which the issue can be avoided by the sse in insulating magnets without conducting charge currents .
it allows us to construct efficient thermoelectric devices upon new principles @xcite and to realize robust , nondissipative information transmission and energy transfer @xcite in the absence of joule heating .
therefore , the sse is expected to act as a new method facilitating the functional use of `` waste '' heat and opens a new possibility of spintronics , @xcite magnonics , @xcite and spin caloritronics , @xcite and thus has ignited a new upsurge of research interest @xcite in these fields . [ 0.45 ] the magnonic spin current is easily flowing from left to right but for @xmath0 the magnonic spin current from right to left is severely suppressed and even prohibited .
, title="fig : " ] a recent work by one of the authors has uncovered the _ rectification _ and _ negative differential _
sse in a metal - insulating magnetic interface system , @xcite where the spin transfer is assisted by the interface electron - magnon inelastic scattering .
it was shown that with the interfacial electron - magnon coupling , reversing the thermal bias is able to give asymmetric spin currents , and increasing thermal bias across the interface abnormally gives an increasing spin current .
@xcite in this work , we study the nonequilibrium transport for the asymmetric and negative differential magnon tunneling driven by temperature bias . we demonstrate in magnon tunneling junctions that the rectification and negative differential sse emerge as a consequence of the magnon - magnon interaction that makes the magnonic spectrum _ temperature - dependent_. this mechanism is different from the one underlying the asymmetric and negative differential sses in metal - magnetic insulator interfaces @xcite that needs the _ energy - dependent _ electronic spectrum .
we also illustrate the concept of spin seebeck transistor based on the negative differential sse .
the rectification and negative differential electronic transports are fundamental for realizing functional electronic diodes and transistors , which are building blocks of modern electronics .
similarly , the rectification of heat flux and negative differential thermal conductance @xcite are also crucial for designing heat diodes and transistors that are as well the fundamental building blocks of phononics .
@xcite thus , we believe our results are relevant for constructing magnonic and spin caloritronic circuits with efficient spin seebeck diodes and transistors , which could play crucial roles in controlling energy and information in magnonics @xcite and spin caloritronics .
@xcite we summarize the transport theory of magnon tunneling in a spin tunneling junction in sec .
we point out in sec .
iiia that the many - body interaction is important for rectifying spin seebeck currents . in sec .
iiib , we show one example of the spin seebeck diode with magnon - magnon interaction induced temperature - dependent spectrum . in sec iiic
, we demonstrate the negative differential sse in a magnon tunneling junction with temperature - dependent excitation gaps and illustrate the functionality of a spin seebeck transistor .
we conclude in sec .
iv and discuss the possible extension of the present results to phononics of controlling phononic information and thermal energy
the magnonic spin tunneling junction consists of three parts : when putting together the left and right insulating magnetic segments , the central interface will be formed , which may result from the lattice mismatch , the vacuum , and the depleted- or nonmagnetic regime that acting as the tunneling barrier [ see fig . [ fig1 ] ] .
therefore , the total hamiltonian can be written as : @xmath1 the left and right insulating magnetic materials , in the low temperature or large spin size limit , are well described by the noninteracting magnon model @xmath2 where @xmath3 and @xmath4 are the creation and annihilation operators of the magnon with momentum @xmath5 and energy @xmath6 at the left ( right ) segments .
these field operators satisfy the commutation relations for bosons .
the left and right magnons are assumed at their own equilibriums with temperature @xmath7 and @xmath8 , respectively .
note that since we are interested in the sse where the pure spin current is induced by the mere temperature bias , we do not consider the possible chemical potential imbalance for the left and right magnons and set them both to zero . in the next section
we will discuss more details , where many - body magnon interaction is included .
for the tunneling part , the magnonic spin transfer is described by the tunneling matrix : @xmath9 which results from the exchange interaction across the interface .
because this term could be small compared to other parts of the total hamiltonian , it is usually treated by the perturbation theory .
physically , each magnon carries a unit spin angular momentum @xmath10 .
the annihilation of a magnon at one side and the successive creation of a magnon at the other side correspond to the rising of the spin component at one side but lowering the spin component at the other side .
in other words , the tunneling magnon current is equivalent to the pure spin down current if the magnetization of the material is defined as up . the tunneling magnon current operator @xmath11 through the interface is given by @xmath12 , where @xmath13 and @xmath14 are the magnon number operators of the left and right segments , respectively .
using the heisenberg s equation @xmath15=\frac{i}{\hbar}[\hat h_i , \hat n_v]$ ] , ( @xmath16 , we can express the tunneling magnon current operator as @xmath17 the magnonic spin current @xmath18 across the tunneling interface is the average value of the operator @xmath19 , which under the second - order perturbation of the tunneling matrix is calculated through @xmath20\right\rangle,\ ] ] where @xmath21 and @xmath22 are represented in the interaction picture , with @xmath23 . substituting the operator @xmath11 , we arrive at the single - magnon tunneling current @xmath24_{\omega=0}=\frac{i}{\hbar^2}\left[g_{\mathrm{tot}}^r(\omega)-g_{\mathrm{tot}}^a(\omega)\right]_{\omega=0},\ ] ] where the retarded ( advanced ) green s function reads @xmath25 with @xmath26 \rangle$ ] and @xmath27 .
although this tunneling spin current is formulated in terms of the bosonic magnon tunneling and the bosonic green s functions are defined in a different way from the electronic ones due to the different ( anti)commutation relations , we note that it shares the same expression as for the standard result of fermionic electron tunneling .
@xcite the green s functions are often calculated by evaluating equivalent matsubara functions of imaginary frequency and successively by the analytic continuation @xmath28 .
@xcite in the following , we alternatively adopt the simple way in the real time and frequency so that the calculations are accessible for the beginner . to evaluate the tunneling spin current , we use the apparent relation @xmath29 so that @xmath30_{\omega=0}$ ] , where the lesser green s function in time domain is @xmath31 and the greater green s function is @xmath32 . in the tunneling process , since the left and right segments of the junction are independent , we then can factorize the green s function as a product of the corresponding left and right ones .
for example , @xmath33 then , noticing that @xmath34 and @xmath35 e^{-i\varepsilon_{\bm p } t/\hbar}$ ] where @xmath36^{-1}$ ] is the corresponding bose - einstein distribution of the magnon population at the @xmath37 segment with temperature @xmath38 , one immediately arrives at @xmath39\delta(\omega+\frac{\varepsilon_{\bm k}-\varepsilon_{\bm p}}{\hbar}).\ ] ] with the same procedure , one also has @xmath40n_r(\varepsilon_{\bm
p})\delta(\omega+\frac{\varepsilon_{\bm k}-\varepsilon_{\bm p}}{\hbar}).\ ] ] therefore , we finally obtain the tunneling spin current : @xmath41 .
\label{eq : is1}\ ] ] for the noninteracting magnon picture within the left and right segments , we have the spectral functions for both sides as @xmath42 and @xmath43 , so that the magnonic tunneling spin current can be expressed as @xmath44 .
\label{eq : is2}\ ] ] note the density of states ( dos ) is related to the spectral functions as @xmath45 and @xmath46 , when approximating the tunneling matrix element as a constant @xmath47 , @xcite one will get @xmath48 .
\label{eq : is3}\ ] ] this formula shows that the tunneling spin current is determined by the magnon population difference and the overlap between magnon dos of the left and right segments . in spite of the fact that the spin current is formulated in terms of the bosonic magnon tunneling and the bosonic green s functions are different from the electronic ones , the expression eq .
( [ eq : is3 ] ) is still reminiscent of that for the electron tunneling .
@xcite the only differences are that here the integral of magnon energy is from @xmath49 from @xmath50 instead of the energy range @xmath51 for electrons ; the magnon population is characterized by the bose - einstein distribution instead of the fermi - dirac distribution for electrons . by the substitutions @xmath52 and @xmath53 , eq . ( [ eq : is3 ] ) can be also obtained directly from eq .
( [ eq : is1 ] ) .
in the above section , we have summarized the theoretical derivations of the tunneling spin current carried by magnons .
it shows clearly that the temperature bias only manifests in the distribution difference of magnons [ see eqs .
( [ eq : is1 ] , [ eq : is2 ] , [ eq : is3 ] ) ] . as such , when reversing the thermal bias @xmath54 , we merely get the reversed spin current without changing magnitude , @xmath55 , i.e. , the spin seebeck diode is absent . in order to obtain the rectification of spin current controlled by the temperature bias ,
it is necessary for both sides of the tunneling junction to have different responses to temperature change .
therefore , we need to go beyond the noninteracting magnon picture so that the temperature - dependent dos becomes possible . in many - body systems ,
the ( nonlinear , higher order than quadratic ) interaction generally changes the energy spectrum . in terms of the language of green s functions
, the many - magnon interaction manifests as the so - called self energy @xmath56 that is generally temperature - dependent .
following the dyson equation , @xcite the retarded ( advanced ) green s function at temperature @xmath57 including magnon - magnon ( or other many - body ) interactions has the form @xmath58\mp i \text{im}[\sigma(\bm k , \varepsilon , t)]}. \label{eq : magnong}\ ] ] accordingly , the spectral function including the many - magnon interaction obtains the temperature - dependence as well , through the self - energy in green s functions : @xmath59=i\left[g^r - g^a\right ] \nonumber \\ & & = \frac{-2\text{im}[\sigma(\bm k , \varepsilon , t)]}{(\varepsilon-\varepsilon_{\bm k}-\text{re}[\sigma(\bm k , \varepsilon , t)])^2 + ( \text{im}[\sigma(\bm k , \varepsilon , t)])^2}.\end{aligned}\ ] ] taking the limit of @xmath60 as in the mean - field approach , we retrieve the delta - function - type spectral function @xmath61 as for the noninteracting case .
the temperature - dependence now enters in the renormalized effective free quasiparticle energy @xmath62 $ ] .
thereupon , the magnon dos becomes also temperature - dependent @xmath63 .
as such , eq . ( [ eq : is3 ] )
is modified by the many - body interaction , as @xmath64 .
\label{eq : is4}\ ] ] generally , the left and right magnetic materials of the tunneling junction can be different , i.e. , their magnon dos have different responses for temperature change . after reversing the thermal bias , i.e. , exchanging the temperatures @xmath65
, we generally have the different dos - overlaps ( or , say , different spectral - overlaps ) @xmath66 . as a result ,
the reversed spin current will have different magnitudes under reversing thermal bias .
if the density - overlap @xmath67 is much larger or smaller than @xmath68 , we will obtain a spin seebeck diode .
that is , in one direction the temperature bias could produce a considerable spin current but in the opposite direction the temperature bias produces less spin current inefficiently , as illustrated in fig .
[ fig1 ] . even if the left and right magnetic materials are identical , in some cases the density - overlap @xmath68 may decrease as increasing the thermal bias @xmath69 .
when the decreasing of the density - overlap surpasses the increasing of @xmath70 $ ] , we will have the negative differential sse .
that is , the magnonic spin current decreases with an increasing thermal bias .
therefore , the many - body interaction induced temperature - dependence of the dos ( as well as the temperature - dependent spectral functions ) is the key to achieve the rectification and negative differential transport of spin current by controlling temperature bias . in the following ,
let us use a simple standard model to illustrate how the magnon - magnon interaction induces the temperature - dependent magnon energy @xcite that in turn leads to the possible spin seebeck diode . without loss of generality , we focus on the ferromagnets .
the results and discussions are readily generalized to the antiferromagnets .
@xcite we first look into the left insulating magnetic material , which is conventionally described by a heisenberg lattice : @xmath71 , \end{aligned}\ ] ] where @xmath72 is the raising ( lowering ) operator for the spin at site @xmath73 and @xmath74 denotes the nearest - neighbor bond with exchange coupling strength @xmath75 .
this heisenberg exchange hamiltonian can be expanded in increasing powers of the magnon operators by the holstein - primakoff transformation , @xcite which maps the spin operators into bosonic magnons through @xmath76 , @xmath77 , @xmath78 , where @xmath79 is the spin size of the left material .
clearly , the creation ( annihilation ) of a magnon @xmath80 at site @xmath73 corresponds to that spin points less ( more ) in @xmath81 component so that each magnon carries an spin angular momentum of @xmath82 ( associated with a magnetic moment ) .
thus , the tunneling magnons driven by temperature bias are responsible for the seebeck spin current .
when keeping up to the lowest order of two - body magnon - magnon interaction , we have the truncated expansion : @xmath83 where we dropped an irrelevant constant term corresponding to the ground - state energy .
since the holstein - primakoff transformation requires the constraint @xmath84 , the above expansion is justified , as in the large spin limit or low temperatures . after performing the fourier
transform into the momentum space , one gets @xmath85 where @xmath86 is the total number of sites , @xmath87 denotes the nearest - neighbor vector , and the energy spectrum has the dispersion @xmath88 . to handle the two - body magnon interaction ,
we adopt the standard mean - field approximation and keep the hartree - fock terms .
finally , after some algebra , one arrives at the quasifree magnon hamiltonian , as in ref . :
@xmath89 where all the magnon - magnon interactions are renormalized into the temperature - dependent factor @xmath90 with @xmath81 the lattice coordinate number .
it requires self - consistency to solve @xmath91 and the analytic expression is absent .
however , to the lowest order we can set the factor at the right - hand side as @xmath92 and use the long - wave limit @xmath93 for the three - dimensional ( 3d ) cube lattice with @xmath94 the lattice constant , then one can integrate eq .
( [ eq : aa ] ) and readily obtain @xmath95 with @xmath96 denoting the zeta function .
accordingly , the temperature - dependent magnon dos for the 3d cube lattice is obtained as @xmath97 ^ 3 } } \ ; , \label{eq : dos}\ ] ] with the total number of spins being normalized .
same expressions also apply for the right insulating magnetic materials with the subscript interchange @xmath98 .
[ 0.4]^{-3/2}$ ] shows the asymmetric behavior .
( b ) * asymmetric thermal - generated spin current in a spin seebeck diode .
* inset shows the rectification ratio defined as @xmath99 .
material parameter values are @xmath100 mev , @xmath101 mev , @xmath102 , which are comparable with the estimated ones from typical magnetic insulators , cf .
yig ( y@xmath103fe@xmath104o@xmath105 ) in refs . .
other parameters are @xmath106 mev , @xmath107 , @xmath108 with @xmath109 k. to meet the constraint @xmath110 , we set a magnon gap of @xmath111 mev , which however will not affect the data plotted here .
, title="fig : " ] from eq .
( [ eq : aa1 ] ) , we know that increasing temperature @xmath112 will decrease @xmath113 that in turn increases the dos @xmath114 through eq .
( [ eq : dos ] ) [ see also fig . [ fig2](a ) ] .
this is understandable because when temperature increases , more magnons are excited so that their interactions effectively soften the exchange stiffness through the decreasing factor @xmath113 , which equivalently increases the magnon population . as a consequence , the dos @xmath114 increases ( decreases ) as temperature @xmath112 increases ( decreases ) and the increasing ( decreasing ) rate is a function of @xmath115 and @xmath116
therefore , when the left and right segments are made of different magnetic materials , i.e. , @xmath117 , @xmath118 , or different spin lattice structures , their dos then generally possess different responses to the temperature change through eq .
( [ eq : dos ] ) , which makes the spin seebeck diode possible .
figure [ fig2](b ) illustrates one example of the asymmetric sse : in the positive thermal bias @xmath119 ( @xmath120 , the thermal - generated spin current is considerable while it is suppressed in the negative thermal bias @xmath121 .
this is reasonable by looking into the dos - overlap @xmath122^{-3/2}$ ] , which as shown in fig . [
fig2](a ) is asymmetric for positive and negative thermal bias as the consequence of different responses of @xmath123 and @xmath124 to the temperature change .
the behavior of the dos - overlap is consistent with that of the seebeck spin current , displayed in fig .
[ fig2](b ) .
naturally , the rectification ratio @xmath99 increases as enlarging the thermal bias [ see the inset of fig . [
fig2](b ) ] .
although the rectification ratio @xmath125 shows that it is not as good as a perfect diode at small thermal bias , the example demonstrates in principle the feasibility of the spin seebeck diode .
note here the rectification of sse emerges from the _ temperature - dependent _ magnonic spectrum due to the magnon - magnon interaction .
it applies for the pure magnonic tunneling system .
this mechanism is different from the asymmetric sse in hybrid metal - magnetic insulator systems , @xcite where the interfacial magnon - electron scattering and the _ energy - dependent _ electronic spectral density are crucial .
although we exemplify the spin seebeck rectification with only the lowest - order magnon - magnon interaction , at high temperatures more magnons will be excited and high - order magnon interactions will be manifested . as such
, the temperature - dependence of the magnon dos will become stronger , which then magnifies the response difference of the magnon dos of the two segments to temperature changes .
therefore , more and stronger ( high - order ) magnon interactions are preferable for realizing efficient spin seebeck diode . in the above example
, we have illustrated that the possible spin seebeck diode effect can result from the temperature - dependent spectra induced by the magnon - magnon interaction with however temperature - independent exchange couplings .
in fact , in real situations , the exchange interaction itself @xmath126 may be sensitive to temperature changes , which in turn also has the contribution to the spin seebeck diode effect . beside the exchange coupling
, the dipolar interaction will have a similar effect .
in particular , the dipolar interaction as well as the anisotropic exchange coupling usually opens a gap for the magnon spectrum , which thus makes the magnon gap temperature - dependent ; see refs . for experimental examples .
the temperature - dependent magnon gap not only contributes to the spin seebeck diode effect but also makes the negative differential sse possible ; that is , increasing ( decreasing ) the thermal bias gives a decreasing ( increasing ) spin current . as an example
, we consider the two materials at both sides of the tunneling interface are both 3d cube lattices with finite magnon gaps so that their dos can be obtained as @xmath127 ^ 3\cos[x(\varepsilon-\delta_v-6j_vs_v ) ] , \label{eq : dos3d}\ ] ] where @xmath128 denotes the bessel function of the first kind and @xmath129 ( @xmath130 ) is the magnon gap of the @xmath37 side , which we assume linearly increases with temperature @xmath131 , as indicated by the experiments in refs . .
we would like to iterate that although here we use a phenomenological temperature - dependence of magnon gap , the microscopic reason relies on the general mechanism of the many - body magnon interaction ( or other particle - magnon interactions ) induced temperature - dependence .
we focus on the model study at present , which in principle is sufficient to demonstrate the possible nontrivial properties .
future works should be continued from the aspect of first - principle or _ ab initio _
theory for more realistic material calculations .
[ 0.4 ] mev , @xmath102 for 3d cube lattices that are comparable with the estimated ones from typical magnetic insulators , cf .
yig in refs . , and @xmath106 mev .
the right segment has a fixed magnon gap @xmath132 mev with a fixed temperature @xmath133 k. the magnon gap for the left material is assumed as @xmath134 with @xmath135 mev / k .
the three dos profiles are calculated from eq .
( [ eq : dos3d]).,title="fig : " ] the corresponding negative differential sse is illustrated in fig .
when @xmath7 increases larger than @xmath8 , the spin seebeck current first increases as expected but then counterintuitively decreases with further increasing thermal bias , so called _ negative differential sse_. this anomalous behavior can be understood with the insets of the figure , where the corresponding magnon dos of the left and right materials are depicted through calculating eq .
( [ eq : dos3d ] ) : with increasing the thermal bias , the dos of two sides are drawn away from each other , which in turn decreases the dos - overlap in eq .
( [ eq : is4 ] ) ; when the dos - overlap decreasing surpasses the increasing of thermal bias , the spin current starts to decrease although with increasing the bias , and the negative differential sse then emerges .
this mechanism for the negative differential sse is different from the one discussed in ref .
, where both the interfacial electron - magnon coupling and the strong energy - dependent electronic dos play the crucial role .
[ 0.5 ] now , based on the negative differential sse , let us illustrate the concept of spin seebeck transistor following the spirit of phononics @xcite : similar to the electronic counterpart of a field - effect transistor ( fet ) , the spin seebeck transistor is composed of three parts : the source ( s ) , the gate ( g ) , and the drain ( d ) , as depicted in fig . [ fig4 ] , which may then connect to external spin or magnon circuits .
the temperatures at the source and drain are fixed with @xmath136 so that the seebeck - effect - generated magnonic spin current flows from the source to the drain .
a third temperature @xmath137 at the gate side is tunable to control the spin currents at the source , gate , and drain , with @xmath138 .
the amplification ability of a spin seebeck transistor can be characterized as the change rate of the gain spin current at the drain upon the change of the control current at the gate , which is expressed as @xmath139 where @xmath140 and @xmath141 are the differential spin seebeck conductance for the source and drain segments , respectively .
it is straightforward to see that only when either negative differential spin seebeck conductance @xmath142 or @xmath143 is present then @xmath144 emerges to be larger than unity . therefore
, the negative differential sse is crucial for realizing a spin seebeck transistor to amplify a spin current with a weak input signal , as the fet in modern electronics and the thermal transistor in phononics .
in summary , we have studied the nonequilibrium transport for the asymmetric and negative differential magnon tunneling driven by temperature bias . we have demonstrated that in magnon tunneling junctions , the rectification and negative differential spin seebeck effects are emerging as a consequence of the many - body magnon interaction that makes the magnonic spectrum temperature - dependent , which are then used to build spin seebeck diodes and transistors .
considering the fact that the magnon carries not only the spin angular momentum but also the energy [ see , e.g. , ref . ] , the present results further indicate the potential of a magnon tunneling junction acting as a thermal diode and/or a thermal transistor .
although in calculations we normalized the number of spin at the interface , the size effect will not influence the asymmetric and negative differential magnon tunneling behaviors , but only affects the magnitude of the spin current .
we can restore the interface size by denoting the number of spin at the tunneling interface as @xmath145 . in this way , @xmath146 and @xmath147 .
thus , the tunneling spin current will be magnified as @xmath148 . therefore , in macroscopic magnon tunneling junctions , the spin current can be sufficiently large and much easier for measurements and applications .
extension of the asymmetric and negative differential tunneling to phononics for the phonon diode and transistor @xcite is also possible , but should be done with care . although the phonon and magnon are both bosons , phonons have different physics in the real space compared to magnons . for phonons ,
the left and right side can still be described by the free bosonic gas model as in eq .
( [ eq : freeboson ] ) , but then the tunneling matrix is modified to @xmath149 because the phonon tunneling is caused by the bilinear coupling of atomic displacements in the real space .
also , the retarded ( advanced ) green s function for phonons , instead of eq .
( [ eq : magnong ] ) , has a different form @xcite : @xmath150\mp i 2\omega_{\bm k}\text{im}[\sigma(\bm k , \omega , t)]},\ ] ] which is also due to the fact that the physics of the phononic green s functions in real space are expressed in terms of the displacement - displacement correlation .
but the underlying mechanism should not change , that is , the many - body interaction induces temperature - dependent self - energies , which in turns makes the quasiparticle spectrum temperature - dependent . as a consequence , the rectification and negative differential thermal conductance will be emergent and the phonon diode and transistor become possible .
recent studies also imply that the phonon - drag plays an important role in sse .
@xcite thus taking account of the effect of nonequilibrium phonons on asymmetric and negative differential sse will be an interesting future topic .
we therefore believe that by integrating the phononics @xcite with spintronics , @xcite magnonics , @xcite and spin caloritronics , @xcite there are more opportunities to achieve the smart control of energy and information in low - dimensional nanodevices .
this work was supported by the national nuclear security administration of the us doe at lanl under contract no .
de - ac52 - 06na25396 , and the ldrd program at lanl ( j.r . ) and in part by the center for integrated nanotechnologies , a u.s .
doe office of basic energy sciences user facility ( j .- x.z . ) .
k. uchida , s. takahashi , k. harii , j. ieda , w. koshibae , k. ando , s. maekawa , and e. saitoh , nature * 455 * , 778 ( 2008 ) . c. m. jaworski , j. yang , s. mack , d. d. awschalom , j. p. heremans , and r. c. myers , nature mater .
* 9 * , 898 ( 2010 ) .
c . le breton , s. sharma , h. saito , s. yuasa , and r. jansen , nature * 475 * , 82 ( 2011 ) .
k. uchida , j. xiao , h. adachi , j. ohe , s. takahashi , j. ieda , t. ota , y. kajiwara , h. umezawa , h. kawai , g. e. w. bauer , s. maekawa , and e. saitoh , nature mater .
* 9 * , 894 ( 2010 ) .
k. uchida , h. adachi , t. ota , h. nakayama , s. maekawa , and e. saitoh , appl .
lett . * 97 * , 172505 ( 2010 ) .
j. bardeen , phys .
* 6 * , 57 ( 1961 ) .
m. h. cohen , l. m. falicov , and j. c. phillips , phys .
lett . * 8 * , 316 ( 1962 ) . j. r. schrieffer , d. j. scalapino , and j. w. wilkins , phys . rev .
* 10 * , 336 ( 1963 ) .
i. s. tupitsyn , p. c. e. stamp , and a. l. burin , phys .
lett . * 100 * , 257202 ( 2008 ) .
a. kreisel , f. sauli , l. bartosch , and p. kopietz , eur .
j. b * 71 * , 59 ( 2009 ) .
h. adachi , j. ohe , s. takahashi , and s. maekawa , phys .
b * 83 * , 094410 ( 2011 ) .
j. xiao , g. e. w. bauer , k. uchida , e. saitoh , and s. maekawa , phys .
b * 81 * , 214418 ( 2010 ) .
h adachi , k. uchida , e. saitoh , j. ohe , s. takahashi , and s. maekawa , appl .
phys lett .
* 97 * , 252506 ( 2010 ) . c. m. jaworski , j. yang , s. mack , d. d. awschalom , r. c. myers , and j. p. heremans , phys .
lett . * 106 * , 186601 ( 2011 ) .
k. uchida , h. adachi , t. an , t. ota , m. toda , b. hillebrands , s. maekawa , and e. saitoh , nature mater . * 10 * , 737 ( 2011 ) .
m. agrawal , v. i. vasyuchka , a. a. serga , a. d. karenowska , g. a. melkov , and b. hillebrands , phys .
* 111 * , 107204 ( 2013 ) .
m. schreier , a. kamra , m. weiler , j. xiao , g. e. w. bauer , r. gross , s. t. b. goennenwein , phys .
b * 88 * , 094410 ( 2013 ) . | we study the nonequilibrium transport for the asymmetric and negative differential magnon tunneling driven by temperature bias .
we demonstrate that the many - body magnon interaction that makes the magnonic spectrum temperature - dependent is the crucial factor for the emergence of rectification and negative differential spin seebeck effects in magnon tunneling junctions .
when magnonic junctions have temperature - dependent density of states , reversing the temperature bias is able to give asymmetric spin currents and increasing temperature bias could give an anomalously decreasing magnonic spin current .
we show that these properties are relevant for building spin seebeck diodes and transistors , which could play important roles in controlling information and energy in magnonics and spin caloritronics . |
fracture processes in heterogeneous media , initiated by external loading has rich physical and mathematical aspects@xcite . since past decades
, people are trying to study this fracture phenomena through different models as well as through direct experiments .
the knowledge of the strength of a disordered solid and prior knowledge of its failure properties are of extreme importance in architectural engineering , textile engineering and in construction of any mechanical structure .
this knowledge is also required to design material microstructures which can be used to construct highly reliable components . in the context of geophysics the dynamical aspects of fracture process
is useful to explain the earthquake phenomena .
some theoretical models like lattice models in various dimensions have been proposed to analyse the details of the fracture phenomena ; among these the fiber bundle model is the earliest and the simplest one .
a loaded bundle of fibers represents the various aspects of fracture process through its self - organised dynamics .
the fiber bundle model study was initiated by peirce@xcite in the context of testing the strength of cotton yarns . since
then this model has been studied from various points of view .
fiber bundles are of two classes with respect to the time dependence of fiber strength : the ` static ' bundles contain fibers whose strengths are independent of time , whereas the ` dynamic ' bundles are assumed to have time dependent elements to capture the creep rupture and fatigue behaviors . according to the load sharing rule , fiber bundles are being classified into two groups : equal load - sharing ( els ) bundles or democratic bundles and local load - sharing ( lls ) bundles . in democratic bundles
intact fibers bear the applied load equally and in local load - sharing bundles the terminal load of the failed fiber is given equally to all the intact neighbors .
the classic work of daniels@xcite on the strength of the static fiber bundles under equal load sharing ( els ) assumption initiated the probabilistic analysis of the model@xcite@xmath0@xcite .
the distribution of burst avalanches during fracture process is a marked feature of the fracture dynamics and can be observed in ultrasonic emissions during the fracture process .
it helps characterizing different physical systems along with the possibility to predict the large avalanches . from a nontrivial probabilistic analysis , hemmer and hansen@xcite
got power law distribution of avalanches for static els bundles , whereas the power law exponent observed numerically for static lls bundles differs significantly .
this observation induces the possibility of presenting loaded fiber bundles as earthquake models@xcite .
the recent mean field estimate of the ` avalanches ' in the els bundles@xcite , gives a new power law@xcite .
the phase transition@xcite and dynamic critical behavior of the fracture process in such bundles has been established through recursive formulation@xcite of the failure dynamics .
the exact solutions@xcite of the recursion relations suggest universal values of the exponents involved .
attempt has also been made@xcite to study the els ans lls bundles from a single framework introducing a ` range of interaction ' parameter which determines the load transfer rule .
coleman@xcite started working on time dependent bundles under els to obtain their life time with steady load .
later , subsequent generalization was made by phoenix@xcite where both els and lls were considered .
some recent developments@xcite@xmath1@xcite show fatigue behavior of els bundles considering fluctuation in applied load . also , introducing noise - induced failure probability of fibers , fatigue behavior is achieved@xcite in a homogeneous fiber bundle under els with steady load . in this report
, we give a brief summary of the exactly solvable@xcite static fiber bundle models under els assumptions .
this gives the mean field behavior of the failure dynamics , its critical behavior and its universality .
an analytic study of fatigue - failure in homogeneous bundle under els@xcite is also discussed . some modifications and discussions
have been added here to correlate the sequential developments .
we also show numerically that critical strength of such linear bundles under lls , vanishes in the large chain limit .
this is basically a confirmation of earlier observations@xcite . a brief magnetic mapping model of fiber bundles
is also discussed later .
* fig . 1 : * the fiber bundle consists initially of @xmath2 fibers attached in parallel to a fixed plate at the top and a movable plate from which a load @xmath3 is suspended at the bottom .
in the els model the load @xmath3 is equally shared by all the intact fibers .
.1 in the fiber bundle consists of @xmath2 fibers , each having identical spring constant @xmath4 .
the bundle supports a load @xmath5 and the breaking threshold @xmath6 of the fibers are assumed to be different for different fiber ( @xmath7 ) .
els model assumes equal load sharing , i.e. , the intact fibers share the applied load @xmath5 equally where @xmath8 is the initial applied stress ( load per fiber ) .
the strength of each of the fiber @xmath9 in the bundle is given by the stress value it can bear , and beyond which it fails .
the strength of the fibers are taken from a randomly distributed normalised density @xmath10 within the interval @xmath11 and @xmath12 such that @xmath13 the equal load sharing assumption neglects ` local ' fluctuations in stress ( and its redistribution ) and renders the model as a mean - field one .
.2 in .2 in the breaking dynamics starts when an initial stress @xmath8 ( load per fiber ) is applied on the bundle .
the fibers having strength less than @xmath8 fail instantly .
due to this rupture , total number of intact fibers decreases and rest of the ( intact ) fibers have to bear the applied load on the bundle . hence effective
stress on the fibers increases and this compels some more fibers to break .
these two sequential operations , the stress redistribution and further breaking of fibers continue till an equilibrium is reached , where either the surviving fibers are strong enough to bear the applied load on the bundle or all fibers fail .
this self organised breaking dynamics can be represented by recursion relations in discrete time steps .
let @xmath14 be the fraction of fibers in the initial bundle that survive after time step @xmath15 , where time step indicates the number of occurrence of stress redistribution .
then the redistributed load per fiber after @xmath15 time step becomes @xmath16 and after @xmath17 time steps the surviving fraction of fiber is @xmath18 where @xmath19 is the cumulative distribution of the corresponding density @xmath10 : @xmath20 now using eq .
( 2 ) and eq . ( 3 ) we can write the recursion relations which show how @xmath21 and @xmath14 evolve in discrete time : @xmath22 and @xmath23 the recursion relations ( 5 ) and ( 6 ) represent the basic dynamics of failure in equal load sharing models . at the equilibrium or steady state @xmath24 and @xmath25 .
this is a fixed point of the recursive dynamics .
( 5 ) and eq . ( 6 ) can be solved at the fixed point for some particular form of @xmath10 and these solutions near @xmath26 ( or @xmath27 ) give the detail features of the failure dynamics of the bundle .
.1 in .2 in * fig . 2 : *
the simplest model considered here assumes uniform density @xmath10 of the fiber strength distribution up to a cutoff strength .
.1 in we choose the uniform density of fiber strength distribution to solve the recursive failure dynamics of democratic bundle . here , the cumulative distribution becomes @xmath28 therefore @xmath14 follows a simple recursion relation ( following eq .
( 6 ) ) @xmath29 at the equilibrium state ( @xmath30 ) , the above relation takes a quadratic form of @xmath26 : @xmath31 the solution is @xmath32 here @xmath33 is the critical value of initial applied stress beyond which the bundle fails completely . the solution with ( @xmath34 ) sign is the stable one , whereas the one with ( @xmath35 sign gives unstable solution@xcite . the quantity @xmath36 must be real valued as it has a physical meaning : it is the fraction of the original bundle that remains intact under a fixed applied stress @xmath8 when the applied stress lies in the range @xmath37 . clearly , @xmath38 ( putting @xmath39 in eq .
therefore the stable solution can be written as @xmath40 for @xmath41 we can not get a real - valued fixed point as the dynamics never stops until @xmath42 when the bundle breaks completely . it may be noted that the quantity @xmath43 behaves like an order parameter that determines a transition from a state of partial failure ( @xmath44 ) to a state of total failure ( @xmath41)@xcite : @xmath45 .2 in * ( a ) at @xmath46 * .2 in to study the dynamics away from criticality ( @xmath47 from below ) , we replace the recursion relation ( 8) by a differential equation @xmath48 close to the fixed point we write @xmath49 + @xmath50 ( where @xmath51 ) .
this , following eq .
( 10 ) , gives@xcite @xmath52 where @xmath53 $ ] . near the critical point
we can write @xmath54 therefore the relaxation time diverges following a power - law as @xmath47 from below@xcite .
one can also consider the breakdown susceptibility @xmath55 , defined as the change of @xmath36 due to an infinitesimal increment of the applied stress @xmath8@xcite @xmath56 from equation ( 10 ) .
hence the susceptibility diverges as the applied stress @xmath8 approaches the critical value @xmath57 .
such a divergence in @xmath55 had already been observed in the numerical studies@xcite .
.2 in * ( b ) at * * @xmath58 * .2 in at the critical point ( @xmath39 ) , we observe a dynamic critical behavior in the relaxation of the failure process to the fixed point . from the recursion relation (
8) it can be shown that decay of the fraction @xmath59 of unbroken fibers that remain intact at time @xmath15 follows a simple power - law@xcite : @xmath60 starting from @xmath61 . for large @xmath15 ( @xmath62 ) , this reduces to @xmath63 ; @xmath64 ; a strict power law which is a robust characterization of the critical state . .2 in to check the universality class of the model we have taken two other types of fiber strength distributions : linearly increasing density distribution and linearly decreasing density distribution within the limit @xmath11 and @xmath12 .
we solve the recursion equations ( 5 ) and ( 6 ) in these two cases .
we show that while @xmath33 changes with different strength distributions , the critical behavior remains unchanged : @xmath65 , @xmath64 .
.1 in * ( a ) linearly increasing density of fiber strength * .2 in .1 in here , the cumulative distribution becomes @xmath66 therefore @xmath14 follows a recursion relation ( following eq .
( 6 ) ) @xmath67 at the fixed point ( @xmath30 ) , the above recursion relation can be represented by a cubic equation of @xmath26 @xmath68 solving the above equation we get the value of critical stress @xmath69@xcite which is the strength of the bundle for the above fiber strength distribution . here
, the order parameter can be defined as @xmath43 and this goes as @xmath70 the susceptibility diverges as the critical point is approached from below : @xmath71 we can also show that for any @xmath72 @xmath73 with @xmath54 .1 in and at @xmath39 .1 in @xmath74 * ( b ) linearly decreasing density of fiber strength * .1 in * fig . 4 : * the linearly decreasing density @xmath10 of the fiber strength distribution up to a cutoff strength .
.1 in in this case , the cumulative distribution becomes @xmath75 and @xmath14 follows a recursion relation ( following eq .
( 6 ) ) @xmath76 at the fixed point ( @xmath30 ) , the above recursion relation can be represented by a cubic equation of @xmath26 @xmath77 solution of the above equation suggests the value of critical stress @xmath78@xcite which is the strength of the bundle for the above fiber strength distribution .
also , the order parameter goes as @xmath79\propto ( \sigma _ { c}-\sigma ) ^{\beta } ; \beta = \frac{1}{2}\ ] ] and the susceptibility diverges with the similar power law as in eqs .
( 16 and 22 ) when the critical point is approached from below : @xmath71 here also for any @xmath72 @xmath73 where @xmath54 and at @xmath39 @xmath74 .1 in thus the democratic fiber bundles ( for different fiber strength distributions ) show phase transition with a well defined order parameter which shows similar power law variation on the way the critical point is approached .
the susceptibility and relaxation time also diverge with same power exponent for all the cases .
therefore , failure of democratic fiber bundles belong to a universality class characterized by the universal values of the associated exponents ( @xmath80 , @xmath81 , @xmath82 and @xmath83 ) .
.2 in .15 in one can now consider a slightly modified strength distribution of the democratic fiber bundle , showing nonlinear deformation characteristics@xcite .
for this , we consider an uniform density distribution of fiber strength , having a lower cutoff . until failure of any of the fibers ( due to this lower cutoff )
, the bundle shows linear elastic behavior .
as soon as the fibers start failing , the stress - strain relationship becomes nonlinear .
the dynamic critical behavior remains essentially the same and the static ( fixed point ) behavior shows elastic - plastic like deformation before rupture of the bundle .
here the fibers are elastic in nature having identical force constant @xmath4 and the random fiber strengths distributed uniformly in the interval @xmath84 $ ] with @xmath85 ; the normalised distribution of the threshold stress of the fibers thus has the form ( see fig .
5 ) : @xmath86 * * fig . 5 * * : the fiber breaking strength distribution @xmath10 considered for studying elastic - plastic type nonlinear deformation behavior of the els model . .1 in for an applied stress @xmath87 none of the fibers
break , though they are elongated by an amount @xmath83 @xmath88 .
the dynamics of breaking starts when applied stress @xmath8 becomes greater than @xmath89 .
now , for @xmath90 the fraction of unbroken fibers follows a recursion relation ( for @xmath10 as in fig .
5 ) : @xmath91 \frac{1}{1-\sigma _ { l}}=\frac{1}{1-\sigma _ { l}}\left [ 1-\frac{\sigma } { u_{t}}\right ] , \ ] ] which has stable fixed points : @xmath92 ; \sigma _ { c}=\frac{1}{4(1-\sigma _ { l})}.\ ] ] the model now has a critical point @xmath93 $ ] beyond which total failure of the bundle takes place .
the above equation also requires that @xmath94 ( to keep the fraction @xmath95 ) .
as one can easily see , the dynamics of @xmath14 for @xmath72 and also at @xmath39 remains the same as discussed in the earlier section . at each fixed point
there will be an equilibrium elongation @xmath96 and a corresponding stress @xmath97 develops in the system ( bundle ) .
this @xmath96 can be easily expressed in terms of @xmath36 .
this requires the evaluation of @xmath27 , the internal stress per fiber developed at the fixed point , corresponding to the initial ( external ) stress @xmath8 ( @xmath98 ) per fiber applied on the bundle when all the fibers were intact .
from the first part of eq .
( 36 ) , one then gets ( for @xmath90)@xmath99 consequently , @xmath100 it may be noted that the internal stress @xmath101 ( @xmath102 from ( 2 ) ) is universally equal to @xmath103 ( independent of @xmath89;from ( 36 ) ) at the failure point @xmath39 of the bundle .
this finally gives the stress - strain relation for the els model : @xmath104 * fig .
6 : * schematic stress ( @xmath105)-strain ( @xmath83 ) curve of the bundle ( shown by the solid line ) , following eq .
( 39 ) , with the fiber strength distribution ( 34 ) ( as shown in fig . 5 ) .
.1 in this stress - strain relation is schematically shown in fig .
6 , where the initial linear region has slope @xmath4 ( the force constant of each fiber ) .
this hooke s region for stress @xmath105 continues up to the strain value @xmath106 , until which no fibers break ( @xmath107 ) .
after this , nonlinearity appears due to the failure of a few of the fibers and the consequent decrease of @xmath36 ( from unity ) .
it finally drops to zero discontinuously by an amount @xmath108=\sigma _ { c } $ ] at the breaking point @xmath39 or @xmath109 for the bundle .
this indicates that the stress drop at the final failure point of the bundle is related to the extent ( @xmath89 ) of the linear region of the stress - strain curve of the same bundle .
.1 in fatigue in fiber bundle model was first studied by coleman in 1958@xcite .
thermally activated failures of fiber have recently been considered and approximate fatigue behavior has been studied@xcite .
we consider here a very simple fiber bundle model with noise - induced activated failure , for which the dynamics can be analytically solved .
let us consider a homogeneous bundle of @xmath2 fibers under load @xmath110 , each having identical failure strength @xmath111 which is the strength ( @xmath33 ) of the bundle also . without any noise ( @xmath112 ) , the model is trivial : the bundle does not fail ( failure time @xmath113 is infinity ) for stress @xmath72 and it fails immediately ( @xmath114 ) for @xmath115 .
we now assume that each such fiber has a finite probability @xmath116 ) of failure at any stress @xmath8 induced by a non - zero noise @xmath117 : @xmath118 , & 0\leq \sigma \leq \sigma _ { c}\\ 1 , & \sigma > \sigma _ { c } \end{array}\right.\ ] ] as one can see , each fiber now has got a non - vanishing probability @xmath119 to fail under a stress @xmath72 at any non - zero noise parameter @xmath117 .
@xmath119 increases as @xmath117 increases and for @xmath115 , @xmath120 .
unlike at @xmath112 , the bundle therefore fails at @xmath72 after a finite time @xmath113 .
here we assume each fiber to have unique threshold , while their breaking probability at any @xmath8 ( @xmath121 ) is due to noise - activated hopping over the barrier height ( @xmath122 ) .
this differs from the earlier model studies@xcite where the load distribution is noise induced .
.1 in * failure time * .05 in at @xmath123 and under any stress @xmath8 ( @xmath121 ) , some fibers fail due to noise and the load gets shared among the surviving fibers , which in turn enhances their stress value , inducing further failure . denoting the fraction of fibers to the initial bundle that remains intact at time @xmath15 by @xmath14
, a discrete time recursion relation@xcite can be written as @xmath124 , \ ] ] where @xmath125 is the redistributed load per fiber among the @xmath126 surviving fibers at time @xmath15 . in the continuum limit ,
we can write the above recursion relation in a differential form @xmath127 , \ ] ] the failure time @xmath113 is defined as @xmath128 when @xmath42 .
integrating eq .
( 42 ) within proper limits we get @xmath129 du\ ] ] or@xcite @xmath130 , \ ] ] for @xmath72 . for @xmath115
, starting from @xmath131 at @xmath132 , one gets @xmath133 from eq .
( 41 ) , giving @xmath114 . for small @xmath117 and as @xmath47 , @xmath134 $ ] .
this failure time @xmath113 therefore approaches infinity as @xmath135 . for @xmath72 ,
one gets finite failure time @xmath113 which decreases exponentially as @xmath8 approaches @xmath33 or as @xmath117 increases and @xmath114 for @xmath115 .
this last feature is absent in the earlier formulations@xcite .
however , all these features are very desirable and are in qualitative agreement with the recent experimental observations@xcite .
our numerical study confirms the above analytic results [ obtained using the continuum version of the recursion relation ( 41 ) ( see fig .
7 ) well . .15 in * * fig . 7**. the simulation results showing variation of average failure time @xmath113 against ( a ) stress @xmath8 and ( b ) against noise @xmath117 , for a bundle containing @xmath136 fibers .
the theoretical results are shown by dotted and dashed lines [ from eq .
the insets show the simulation results for the variation of the fraction @xmath137 of unbroken fibers with time @xmath15 for different @xmath117 values [ 1.2 ( cross ) and 1.0 ( plus ) ] in ( a ) and @xmath8 values [ 0.15 ( cross ) and 0.12 ( plus ) ] in ( b ) .
the dotted and dashed lines represent the theoretical results [ eq .
( 42 ) ] .
.2 in the strength of a bundle of fibers plays important role in the failure dynamics of the bundle when external load is applied . by a probabilistic approach
daniel@xcite pioneered the study of finding the strength distribution of a bundle in terms of strength of the constituent fibers .
daniel assumed equal sharing of applied load ( els ) .
later , this work has been expanded in the context of material science@xcite .
this type of model shows ( both analytically and numerically ) existence of a critical strength ( non zero @xmath33 ) of the bundle@xcite beyond which it collapses instantly .
the other extreme model , i , e . , the local load sharing ( lls ) model has been proved to be difficult to tackle analytically@xcite .
however some approximate asymptotic methods@xcite have been developed to tackle the problem in one dimension .
recently pacheco et .
al.@xcite introduced and solved the one sided load transfer model analytically .
it is basically a simplification of the conventional ( both sided ) lls model . considering weibull distribution of fiber strength , they obtained the system size dependence of the strength ( @xmath33 ) of the bundle as @xmath138 the subscript @xmath139 indicates the shape factor or weibull index of the fiber strength distribution .
this clearly shows @xmath140 as @xmath141 .
smith@xcite conjectured a similar logarithmic dependence for lls bundles from numerical results . here
, we have simulated the above two types of lls models : the one sided load transfer model and the conventional both sided load transfer model considering uniform distribution ( random ) of fiber strengths [ fig . 8 ] . we can not use the strictly uniform distrinution ( as in fig .
2 ) , to avoid the increasing sequence of fiber strength arrangement which is fatal in case of lls bundles though it does not matter in case of els bundles .
.2 in .2 in * * fig . 8 * * : the uniform ( random ) fiber strength distribution @xmath10 considered to study the lls models .
.2 in we observe the @xmath142 dependence of the bundle s strength ( @xmath33 ) for both the cases [ fig .
9 ] which confirms the non existence of any critical strength ( non zero @xmath33 ) of the bundle in one dimension . * * fig . 9 * * : the system size dependence of strength of the bundle ( @xmath33 ) in lls models .
the results are obtained after averaging over @xmath143configurations .
.2 in we can map the fiber bundle problem in terms of a magnetic model or specifically an ising model .
the state of the fibers represent different states of the spins .
we introduce a variable @xmath144 to indicate the status of the @xmath7-th fiber at the time step @xmath15 such that @xmath145 for intact fibers and @xmath11 for broken fibers .
the corresponding spin variable @xmath146 can then take two values @xmath147 , which can represent the up and down states of the ising spin at @xmath7-th site .
the strength threshold @xmath148 of the @xmath7-th fiber represents a random field at the @xmath7-th site ( fiber ) favoring @xmath149 state ( or @xmath150 ) and the stress @xmath151 on the @xmath7-th fiber represents a competing field at each site favoring the state @xmath152 ( or @xmath153 ) .
the value of @xmath154 or @xmath144 depends on the strength of the resultant field @xmath155 : @xmath156 ; \ ] ] @xmath157 .\ ] ] the only restriction on the competing fields is @xmath158 \sigma _ { i}(t)=f;\ ] ] @xmath3 is the applied force ( load ) on the bundle .
@xmath151 , and hence @xmath155 , evolves with time @xmath15 following the above relation .
this is a general restriction for both , equal and local load sharing cases .
the average magnetization at time @xmath15 can be defined as @xmath159 this non - interacting spin model is somewhat special as the zero - temperature dynamics of the spins here is determined by the strength of the effective feild @xmath160 at that site ( 46 ) , where one part ( @xmath151 ) of @xmath160 evolves with time following ( 48 ) .
the stable manetization @xmath161 ( @xmath162 at @xmath163 ) decreases continuously with @xmath8 until it reaches a value @xmath164 , where @xmath165 at @xmath39 , beyond which @xmath161 drops to @xmath166 discontinuously . for equal load sharing case ,
the above equation becomes @xmath167 @xmath8 being the initial applied stress ( @xmath168 ) .
for uniform fiber strength distribution we have @xmath38 ( from ( 10 ) ) . here
, @xmath169 and @xmath161 decreases continuously from unity ( at @xmath163 ) to @xmath170 at @xmath39 and then abruptly crosses over to @xmath171 at @xmath41 .
.2 in the inherent mean - field nature of the els models enables to construct recursion relations ( eqs . 5 and 6 ) which captures essentially all the intriguing features of the failure dynamics .
though we have identified @xmath172 as the order parameter ( with exponent @xmath173 ) for the continuous transition in the els models , unlike in the conventional phase transitions it does not have a real - valued existence for @xmath41 .
the ` type ' of phase transition in els models is still a controversial issue . earlier@xcite it was suggested to be a first order phase transition , because the the surviving fraction of fibers has a discontinuity at the breakdown point of the bundles .
however , as the susceptibility shows divergence ( @xmath174 ) at the breakdown point , the transition has been later identified to be of second order@xcite .
the dynamic critical behavior of the els models and the universality of the exponent values are straightforward . here , divergence of relaxation time ( @xmath113 ) at the critical point ( @xmath175 ) indicates ` critical slowing ' of the dynamics which is characteristic of conventional critical phenomena . at the critical point , one observes power law decay of the surviving fraction in time ( @xmath176@xmath64 ) .
we demonstrated the universality of the failure behavior near @xmath39 , for three different distributions : uniform ( fig .
2 ) , linearly increasing ( fig .
3 ) and linearly decreasing ( fig .
4 ) distributions of fiber strength .
the critical strengths of the bundles differ in each case : @xmath177 and @xmath178 respectively for these three distributions . however , the critical behavior of the order parameter @xmath179 , susceptibility @xmath55 , relaxation time @xmath113 and of the time decay at @xmath33 , as given by the exponents @xmath180 and @xmath83 remain unchanged : @xmath65 and @xmath64 for all three distributions .
the els model also shows realistic nonlinear deformation behavior with a shifted ( by @xmath89 , away from the origin ) uniform distribution of fiber strengths .
the stress - strain curve for the model clearly shows three different regions : elastic or linear part ( hooke s region ) when none of the fibers break ( @xmath107 ) , plastic or nonlinear part due to the successive failure of the fibers ( @xmath181 ) and then finally the stress drops suddenly ( due to the discontinuous drop in the fraction of surviving fibers from @xmath182 to zero ) at the failure point @xmath93 $ ] .
the fatigue study in a homogeneous fiber bundle suggests if the each fiber has a finite probability of failure ( due to noise @xmath117 as in ( 40 ) ) below its normal strength , then the failure time of the bundle decreases exponentially ( @xmath134 $ ] ) as @xmath8 approaches @xmath33 from below and @xmath183 for @xmath41 .
these features agree well with the experimental observations in disordered solids .
the lls bundles show ` zero ' critical strength as the bundle size goes to infinity in one dimension .
it is not clear at this stage if , in higher dimensions , lls bundles are going to have non - zero critical strength@xcite . in any case , the associated dynamics of failure of these higher dimensional bundles with variable range load transfer@xcite should be interesting .
we believe , the elegance and simplicity of the model , its common - sense appeal , the exact solubility of its critical behavior in the mean field ( els ) limit , its demonstrated universality , etc , would promote the model eventually to a level competing with the ising model of magnetic critical behavior .
we are grateful to pratip bhattacharyya for useful discussions .
one of the author ( bkc ) is grateful to the selected high - school leaving students ( in the undergraduate associateship programme of the saha institute of nuclear physics ) for their enthusiastic participation , comments and suggestions .
their involvement and appreciation encouraged the comparative view of the usefulness of the model over the ising model discussed above .
b. k. chakrabarti and l. g. benguigui , _ statistical physics of fracture and breakdown in disorder systems _ , oxford univ .
press , oxford ( 1997 ) , see also h. j. herrmann and s. roux ( eds ) , _ statistical models of disordered media _ , north holland , amsterdam ( 1990 ) ; p. bak , _ how nature works , _
oxford univ . press , oxford ( 1997 ) ;
m. sahimi , phys .
rep**. 306 * * , 213 ( 1980 ) .
f. t. peires , j. textile inst . *
17 * , t355 - 368 ( 1926 ) .
h. e. daniels , proc .
london a * 183 * 405 ( 1945 ) .
d. e. gucer and j. gurland , j. mech .
solids * 10 * , 365 ( 1962 ) ; s. l. phoenix , adv .
11 * , 153 ( 1979 ) ; r. l. smith and s. l. phoenix , j. appl . mech . * 48 * , 75 ( 1981 ) ; r. l. smith , ann
* 10 * , 137 ( 1982 ) ; l. n. mccartney and r. l. smith , j. appl . mech . * 50 * , 601 ( 1983 ) .
p. c. hemmer and a. hansen , j. appl .
mech . * 59 * 909 ( 1992 ) ; a. hansen and p. c. hemmer , phys .
a * 184 * 394 ( 1994 ) ; m. kloster , a. hansen and p. c. hemmer , phys . rev .
e * 56 * 2615 ( 1997 ) .
d. sornette , j. phys .
a * 22 * l243 ( 1989 ) ; d. sornette , j. phys .
i ( france ) * 2 * 2089 ( 1992 ) ; a. t. bernardes and j. g. moreira , phys .
b * 49 * 15035 ( 1994 )
. s. zapperi , p. ray , h. e. stanley and a. vespignani , phys .
lett . * 78 * 1408 ( 1997 ) ; y. moreno , j. b. gomez and a. f. pacheco , phys .
. lett . * 85 * 2865 ( 2000 ) .
r. da silveria , am . j. phys .
* 67 * 1177 ( 1999 ) .
s. pradhan and b. k. chakrabarti , phys .
e * 65 * , 016113 ( 2001 ) ; s. pradhan , p. bhattacharyya and b. k. chakrabarti , phys . rev .
e * 66 * , 016116 ( 2002 ) ; p. bhattacharyya , s. pradhan and b. k. chakrabarti , phys .
e * * 6**7 , 046122 ( 2003 ) ; r. c. hidalgo , f. kun and h. j. herrmann , phys . rev .
e * * 64**066122 ( 2001 ) .
b. d. coleman , j. appl . phys . * 27 * , 862 ( 1956 ) ; b. d. coleman , trans .
. rheol . * 1 * , 153 ( 1957 ) ; b. d. coleman , trans .
. rheol . * 2 * , 195 ( 1958 ) . s. l. phoenix , siam ( soc .
appl . math . )
j. appl . math . *
34 * , 227 ( 1978 ) ; s. l. phoenix , adv .
. probab . * 11 * , 153 ( 1979 ) ; w. i. newman and s. l. phoenix , phys .
e * 63 * , 021507 ( 2000 ) ; . s. roux , phys .
e * 62 * , 6164 ( 2000 ) ; r. scorretti , s. ciliberto and a. guarino , europhys .
* 55 * , 626 ( 2001 ) .
s. pradhan and b. k. chakrabarti , phys .
e * * 6**7 , 046124 ( 2003 ) .
l. golubovic and s. feng , phys .
a * 43 * , 5223 ( 1991 ) .
r. banerjee and b. k. chakrabarti , bull .
* 24 * , 161 ( 2001 ) ; a. guarino , s. ciliberto , a. garcimartin , m. zei and r. scorretti , cond - mat/0201257 ( 2002 ) .
d. g. harlow and s. l. phoenix , j. composit mater .
* 12 * , 314 ( 1978 ) ; d. g. harlow and s. l. phoenix , adv . appl . prob .
* 14 * , 68 ( 1982 ) ; r. l. smith , proc .
london * a 382 * , 179 ( 1982 ) ; r. l. smith , adv , appl . prob . * 15 * , 304 ( 1982 ) . s. l. phoenix and r. l. smith , j. solid struct . * 19 * , 479 ( 1983 ) .
j. b. gomez , d. iniguez and a. f. pacheco , phys .
71 , 380 ( 1993 ) .
r. l. smith , proc .
london * a 372 * , 539 ( 1980 ) . g. caldarelli and a. petri , phys .
lett . * 83 * , 1483 ( 1999 ) . o. e. yewande , y. moreno , f. kun , r. c. hidalgo and h. j. herrman , cond mat/0210119 ( 2002 ) . | we study the failure properties of fiber bundles when continuous rupture goes on due to the application of external load on the bundles .
we take the two extreme models : equal load sharing model ( democratic fiber bundles ) and local load sharing model .
the strength of the fibers are assumed to be distributed randomly within a finite interval .
the democratic fiber bundles show a solvable phase transition at a critical stress ( load per fiber ) .
the dynamic critical behavior is obtained analytically near the critical point and the critical exponents are found to be universal .
this model also shows elastic - plastic like nonlinear deformation behavior when the fiber strength distribution has a lower cut - off .
we solve analytically the fatigue - failure in a democratic bundle , and the behavior qualitatively agrees with the experimental observations . the strength of the local load sharing bundles is obtained numerically and compared with the existing results .
finally we map the failure phenomena of fiber bundles in terms of magnetic model ( ising model ) which may resolve the ambiguity of studying the failure properties of fiber bundles in higher dimensions .
failure properties of fiber bundle models |
determinations of the kobayashi - maskawa matrix elements [ 1 ] provide important checks on the consistency of the standard model and ways to search for new physics . in decades , several methods have been discussed to provide interior angles @xmath8 , @xmath9 and @xmath0 of the unitarity triangle . the @xmath0 measurement is still on the way since the measurement has been statistically limited even with modern electron - positron b factories .
two of those analyses to evaluate @xmath0 are reported here .
as space is limited , belle experiment and kekb accelerator are not explained here . detailed description of the belle detector is found elsewhere [ 2 ] .
the time - dependent @xmath10 analysis of the @xmath11 system provides a theoretically clean measurement of the product @xmath12 [ 3 ] , where @xmath13 is the ratio of the magnitudes of the doubly cabibbo - suppressed decay amplitude to the cabibbo - favoured decay amplitude . unlike the @xmath14 process
, @xmath2 , which is predominantly a spectator process with a @xmath15 transition , does not have contributions from @xmath16 decays to the same final state and can provide clean experimental access to @xmath17 . assuming su(3 ) flavour symmetry between @xmath18 and @xmath19 , @xmath17 is given by @xmath20 where @xmath21 is the cabibbo angle , @xmath22 and @xmath23 are the meson form factors , and the @xmath24 s stand for the corresponding branching fractions . the @xmath2 process , in addition , does not have a penguin loop contribution and hence can in principle be used to determine @xmath25 [ 4 ] .
the decay @xmath2 does not have a contribution from the @xmath26-exchange amplitude , as the quark - antiquark pair with the same flavor , required for such a diagram , is absent from the final state .
we assume the @xmath26-exchange contributions in @xmath27 to be negligible , in making the correspondence between @xmath28 and @xmath29 in the @xmath17 calculation .
the size of the w - exchange diagram can be estimated from the @xmath3 decay , which proceeds only via @xmath26 exchange . the @xmath3 branching fraction was expected to be enhanced due to rescattering effects [ 5 , 6 ] . here
we briefly report an improved measurement of the branching fractions for @xmath2 and @xmath3 with a data sample consisting of @xmath30 @xmath31 pairs .
detailed description of the analysis is found elsewhere [ 7 ] .
the signal is reconstructed in three @xmath32 modes : @xmath33 with @xmath34 , @xmath35 with @xmath36 , and @xmath37 with @xmath38 . after applying belle
standard event selection criteria , we obtained results of fits ( figure 1 ) summarized in table 1 .
the significance is defined as @xmath39 , where @xmath40 are the likelihoods for the best fit and with the signal branching fraction fixed to zero .
table 2 summarizes the systematic uncertainties involved .
the overall systematic uncertainty is obtained by adding these contributions in quadrature .
lcrllr @xmath41 mode & @xmath32 mode & @xmath42(% ) & @xmath43 & @xmath44 & @xmath45 + @xmath2 & @xmath46 & @xmath47 & @xmath48 & @xmath49 & @xmath50 + & @xmath51 & @xmath52 & @xmath53 & @xmath54 & @xmath55 + & @xmath56 & @xmath57 & @xmath58 & @xmath59 & @xmath60 + & simultaneous & @xmath61 & @xmath61 & @xmath62 & @xmath63 + @xmath3 & @xmath46 & @xmath64 & @xmath65 & @xmath66 & @xmath50 + & @xmath51 & @xmath67 & @xmath68 & @xmath69 & @xmath70 + & @xmath56 & @xmath71 & @xmath72 & @xmath73 & @xmath74 + & simultaneous & @xmath61 & @xmath61 & @xmath75 & @xmath76 + captypetable lcc source & + & @xmath77 & @xmath78 + @xmath32 branching fraction uncertainties & & + signal & 5.9 & 6.2 + peaking background & 1.5 & 1.9 + total(@xmath24 ) & 6.1 & 6.5 + tracking efficiency & 4.0 & 4.0 + photon detection efficiency & 7.0 & 7.0 + particle identification efficiency & 2.4 & 2.1 + @xmath79 efficiency & 1.1 & 1.1 + @xmath80 & 0.6 & 0.5 + @xmath81 & 1.4 & 1.4 + mc statistics & 1.4 & 1.6 + pdf shape & 3.4 & 1.5 + fit bias & 0.9 & 0.3 + total ( other ) & 9.4 & 8.8 + we obtain @xmath82 and @xmath83 with significance of @xmath84 and @xmath85 , respectively , where the systematic uncertainties on the signal yield as well as the statistical uncertainties are included in the significance evaluation . using observed value for the @xmath2 branching fraction , the latest values for @xmath86 , @xmath87 [ 8 ] , and the theoretical estimate of the ratio @xmath88 [ 9 ] , we obtain @xmath89 where the first error is statistical , the second corresponds to the experimental systematic uncertainty , and the third accounts for the theoretical uncertainty in the @xmath90 estimation .
we have assumed that the ratio @xmath91 is equal to the ratio of vector meson decay constants , @xmath92 .
the value we obtain for @xmath13 is consistent with theoretical expectation of 2% .
the observed value for the @xmath3 branching fraction is 2 orders of magnitude lower than that for the cabibbo - favoured decay @xmath93 .
among various methods to measure the angle @xmath0 using @xmath10 violation in @xmath95 decays [ 10 - 14 ] , three body final states such as @xmath96 have been suggested as promising modes for the extraction of @xmath0 [ 15 ] . assuming no @xmath10 asymmetry in neutral @xmath97 decays , the amplitude of the neutral @xmath97 decays from @xmath98 as a function of dalitz plot variables @xmath99 and @xmath100 is @xmath101 where @xmath102 is the amplitude of the @xmath103 decay , @xmath104 is the ratio of the magnitudes of the two interfering amplitudes , and @xmath105 is the strong phase difference between them .
the @xmath103 decay amplitude @xmath106 can be determined from a lage sample of flavor - tagged @xmath103 decays produced in @xmath107 continuum process .
once @xmath106 is known , a simultaneous fit of @xmath108 and @xmath109 data allows the contribution of @xmath104 , @xmath0 and @xmath105 to be separated . in this paper
, we report a preliminary result of a measurement of @xmath0 using the modes @xmath110 and @xmath111 with @xmath112 , based on a 605 @xmath113 data sample .
detailed description is found in reference [ 16 ] . after event selection(figure 2 and 3 ) , we obtained @xmath0 and other parameters by fitting the dalitz plots for two event samples , @xmath110 , @xmath111 and combined results of angle @xmath0 as @xmath4 ( table 3 ) .
.@xmath10 fit results .
the first error is statistical , the second is experimental systematic , and the third is the model uncertainty .
the error in combined result of @xmath0 is only statistical .
consult the reference for detailed results [ 16 ] .
[ cols="<,^,^,^,^",options="header " , ] decays from selected @xmath114(top ) and @xmath115(bottom ) candidates , shown separately for @xmath109(left ) and @xmath108(right ) tags . ]
decays from selected @xmath114(top ) and @xmath115(bottom ) candidates , shown separately for @xmath109(left ) and @xmath108(right ) tags . ]
we report results of measurements of the angle @xmath0 of kobayashi - maskawa triangle with @xmath1 dalitz plot analysis and related issues with @xmath2 and @xmath3 decay processes .
measurement of angle @xmath0 is on the way to verify the standard model prediction , and will be a rich physics subject for a decade with next generation @xmath116 factories and lhc . | we report results of measurement of the unitarity triangle angle @xmath0 with @xmath1 dalitz plot analysis and related issues with @xmath2 and @xmath3 decay processes .
the dalitz analysis improves accuracy of the angle @xmath0 as @xmath4 and the branching fractions are found to be @xmath5 decays are set as @xmath6 and @xmath7 . |
ly@xmath1 emitters ( laes ) have the potential to be a powerful cosmological probe .
the evolution of their number density at * @xmath7 * offers the opportunity to constrain the opacity of the intergalactic medium ( igm ) and the timing of cosmic reionization ( e.g. , * ? ? ?
their clustering properties may be able to constrain efficiently the expansion history of the universe ( e.g. , * ? ? ?
* ; * ? ? ?
furthermore , at the highest redshifts ( @xmath8 ) ly@xmath1 is the only emission line that we can detect from the ground , making it the only probe of the internal structure of these galaxies .
however , to use laes as probes of the high - redshift universe , we must first understand their physical properties .
the ly@xmath1 line is resonantly scattered by neutral hydrogen , making its flux and line profile notoriously hard to interpret ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for example , unlike rest - frame uv continuum observations , one can not merely apply a dust extinction correction to obtain an estimate of the star formation rate ( sfr ) .
thus , constraints on the ly@xmath1 escape fraction must be empirically measured .
furthermore , observational studies are required to determine which physical properties facilitate the escape of ly@xmath1 emission . due to the limited resources in the uv ,
most observational studies have focused on constraining the physical properties of @xmath9 laes , where the ly@xmath1 line moves into the optical .
detailed analysis of the host galaxies is difficult at these high redshifts , and different groups have come to drastically different conclusions about them ( e.g. , compare @xcite to @xcite ) .
furthermore , at these redshifts ly@xmath1 emission may be significantly altered by the intervening igm .
for example , based on hydro - simulations , @xcite find that only @xmath10 , @xmath11 , and @xmath12 of ly@xmath1 photons are transmitted through the igm at @xmath13 , @xmath14 , and @xmath15 , respectively . to make progress
, we need to study lower - redshift samples where laes are bright due to smaller luminosity distances and where laes can be integrated into comprehensive studies of galaxies at the same redshifts to understand how laes are drawn from the general galaxy population .
recently , a @xmath16 @xmath2 0.3 lae sample has been found by searching for emission - line objects in _ galaxy evolution explorer _ ( _ galex _ ; * ? ? ?
* ) fuv pipeline spectra .
since the _ galex _ grism pipeline only includes objects whose uv continuum magnitudes are bright enough to generate measurable continuum spectra , only a fraction of the emission - line objects are extracted .
thus , this sample s selection process is most analogous to locating laes in the high - redshift lyman break galaxy ( lbg ) population via spectroscopy ( e.g. , * ? ? ?
however , the procedure enables the selection of a substantial sample of @xmath16 @xmath2 0.3 sources , and many papers have investigated their properties ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the picture that emerged from the optical follow - up of the @xmath3 _ galex _ sample is that laes , when compared to uv - continuum selected galaxies , are relatively young , compact , metal poor , star - forming galaxies @xcite .
however , it is not clear that the properties of these low - redshift ( @xmath3 ) sources are representative of high - redshift laes . in particular , laes are considerably fainter and much less common at @xmath3 than they were in the past , with only about @xmath17 of @xmath3 uv - continuum selected galaxies having rest - frame ly@xmath1 equivalent widths ( ew@xmath18(ly@xmath1 ) ) greater than @xmath19 , compared with @xmath20 at @xmath21 @xcite .
the ew@xmath18(ly@xmath1 ) @xmath22
20 requirement is typically used to define high - redshift samples ( e.g. , * ? ? ? * ) .
the first redshift where laes are seen that are comparable in luminosity to high - redshift laes is at @xmath16 @xmath2 1 .
however , only a handful of @xmath0 laes have _ galex _ detectable continuum spectra .
thus , we need to remove the pipeline s continuum selection requirement ( nuv@xmath2322 ) in order to study a large sample of @xmath16 @xmath2 1 laes .
this has the additional virtue of producing a purely flux - limited sample that is straightforward to analyze .
ccccccc cdfs & 3@xmath2430@xmath2540@xmath26 & -27@xmath2727@xmath2843@xmath29 & 353 ks & 1.8@xmath3010@xmath31 & 40 & 24 + groth & 14@xmath3219@xmath3358@xmath34 & 52@xmath3546@xmath2854@xmath29 & 291 ks & 2.0@xmath3010@xmath31 & 41 & 19 + ngpdws & 14@xmath3236@xmath3337@xmath34 & 35@xmath3510@xmath2817@xmath29 & 165 ks & 2.3@xmath3010@xmath31 & 35 & 11 + cosmos & 10@xmath3200@xmath3329@xmath34 & + 2@xmath3512@xmath2821@xmath29 & 140 ks & 3.2@xmath3010@xmath31 & 19 & 4 + [ fields ] in ( * ? ? ?
* hereafter ,
bcw12 ) , we solved the problem of obtaining a flux - limited sample of @xmath0 laes for the _ chandra _ deep field - south ( cdfs ) . to do this
, we converted the multiple _ galex _ grism images into a three - dimensional ( two spatial axes and one wavelength axis ) data cube .
we then treated the wavelength `` slices '' as narrowband images in which we searched for emission - line galaxies . through simulations
, we showed that we could recover more than 80% of the sources with @xmath36(ly@xmath1 ) @xmath37 10@xmath31 erg @xmath38 s@xmath39 , which corresponds to @xmath40(ly@xmath1 ) @xmath37 10@xmath41erg s@xmath39 at @xmath0 . by comparing our sample to x - ray data , existing optical spectroscopy , and deep _
u_-band imaging , we determined that nearly all of our 28 new laes are real and that the uv spectroscopic redshifts based on the ly@xmath1 identifications are reliable .
we also determined the fraction of active galactic nuclei ( agns ) in the sample relative to star formers . in this paper
, we apply the data cube search method to all of the deepest _ galex _ grism fields ( cdfs-00 , 353 ks ; groth-00 , 291 ks ; ngpdws-00
, 165 ks ; cosmos-00 , 140 ks ) , which correspond to some of the most intensively studied regions in the sky .
this work provides the first large sample of @xmath42 laes ( n=60 ) that can be used to investigate the physical properties of these galaxies .
there are relatively large samples of known laes at redshifts of @xmath3 and @xmath5 .
thus , a sample of @xmath0 laes is needed to map the evolution of laes over a @xmath27 gyr gap .
this redshift regime is where the star - forming properties of galaxies change very rapidly and where the star formation begins to decline .
it is a key area for connecting to the high - redshift universe . in paper ii , we will leverage existing archival spectra and followup optical spectra to constrain their physical properties .
in particular , we will study which properties facilitate the escape of ly@xmath1 emission and how laes are drawn from the overall galaxy population .
this will help us understand how the ly@xmath1 emission properties of galaxies evolve as we move to higher redshifts and higher ly@xmath1 luminosities .
here we catalog our candidate @xmath0 lae samples in each field and give optical redshifts from both archival and newly obtained observations . with x - ray ,
uv , and optical data , we determine the false detection rate ( cases where the emission line is either not confirmed or is not ly@xmath1 ) and the agn contamination rate of our sample . with the remaining laes , we compute the lae galaxy luminosity function ( lf ) at @xmath0 and use this to investigate the evolution of the ly@xmath1 lf and the ly@xmath1 escape fraction over the redshift range from @xmath3 to @xmath5 . unless otherwise noted , we give all magnitudes in the ab magnitude system ( @xmath43log@xmath44 with @xmath45 in units of @xmath46jy ) .
we use a standard @xmath47 km s@xmath39 mpc@xmath39 , @xmath48 , and @xmath49 cosmology .
we study the four deepest nuv grism observations : cdfs-00 , groth-00 , ngpdws-00 , and cosmos-00 .
the _ galex _ fields are large ( @xmath50 deg@xmath51 ) , but in subregions of the fields , there are many objects with optical spectra ( cdfs , @xcite , @xcite ; groth , @xcite ; cosmos , @xcite , @xcite ) . in subregions , there are also multi - color observations from the _ hst _ goods @xcite , candels @xcite , cosmos @xcite , and gems @xcite programs that provide the galaxy morphologies and spectral energy distributions from the rest - frame far - uv to the mid - infrared . when corrected for the emission - line contributions using the spectra @xcite , the spectral energy densities can be used to compute the ages and extinctions of the galaxies . finally , the 4 ms _ chandra _ image of the cdfs @xcite region , along with shallower x - ray observations in the extended cdfs @xcite , cosmos @xcite , groth @xcite , and ngpdws @xcite fields , can be used to identify agns .
agns may also be identified using the uv grism spectra obtained with _
galex _ and the optical spectra , but the x - rays provide a valuable cross - check , where they are available .
( a color version of this figure is available in the online journal . ) as described in bcw12 , we converted multiple _ galex _ nuv grism images into a three - dimensional ( two spatial axes and one wavelength axis ) data cube .
each background subtracted data cube consists of thirty @xmath19 narrowband slices covering a wavelength range of 2030 to 2630 and a 50@xmath28@xmath3050@xmath28 field of view .
we used sextractor @xcite to identify all 4@xmath52 sources within the cube and then visually inspected each source and its spectrum ( 1-d and 2-d ) to eliminate objects that were artifacts . applying the data cube search produced 40 cdfs , 41 groth , 35 ngpdws , and 19 cosmos objects for a full ly@xmath1 selected candidate sample of 135 ( see table [ fields ] ) . in figure [ ex](a ) , we show the extracted 1-d spectrum for galex033100 - 273020 to illustrate the quality of our _ galex _ spectra .
bcw12 performed this routine on the deepest nuv grism field ( cdfs-00 ) . to verify the bcw12 sample of 28 lae candidates , we have independently searched the cdfs cube for emission - line sources .
we found 27 of the 28 objects that were identified by bcw12 , and we found 12 new objects .
approximately half of these new lae candidates are close to the field s @xmath53 completeness flux threshold ( see section 3.3 ) and were probably missed due to their faintness .
the other half are relatively bright and were probably mis - classified as data cube artifacts , such as the remaining edge effects from brighter objects . the bcw12 lae candidate missed by this search ( galex033124 - 275625 ) is real and has a consistent optical redshift .
this lae was eliminated in the visual inspection phase of our search due to its proximity to a uv - bright star .
we include this missed lae in all subsequent analysis . to date ,
we have confirmed the uv redshifts of 92@xmath54 ( 11 out of 12 ) of the newly discovered cdfs laes with optical follow - up spectra .
spectroscopic observations were primarily obtained with the deep imaging multi - object spectrograph ( deimos ; * ? ? ?
* ) on keck ii in a series of runs in 2012 to 2013 .
the observations were made with the zd600 line mm@xmath39 grating blazed at 7500 .
this gives a resolution of @xmath2 5 with a 1@xmath29 slit and a wavelength coverage of 5300 .
the high spectral resolution is necessary to distinguish the [ oii]@xmath553727 doublet structure .
the observations were not generally taken at parallactic angle , since the position angle was determined by the mask orientation .
each @xmath230 min exposure was broken into three subsets , with the objects stepped along the slit by 1.5@xmath29 in each direction .
the raw two - dimensional spectra were reduced and extracted using the procedure described in @xcite . in figure [ ex](b ) , we show the optical spectrum for galex033100 - 273020 to illustrate the quality of our _ _ deimos spectra .
we used the remaining space available in our mos masks to observe a control sample , a uv - continuum selected sample without detected ly@xmath1 that have the same luminosities and are expected , based on their colors , to lie in the same redshift interval as our primary lae sample . in paper ii , we will use the control sample to determine how the lae galaxies are drawn from the general galaxy population . ) .
the red histogram shows the number of lae candidates as a function of flux for all four _ galex _ fields . ]
( a color version of this figure is available in the online journal . )
we observed our southern and equatorial fields ( cosmos and cdfs ) with the robert stobie spectrograph ( rss ; * ? ? ?
* ) on salt from 2011 to 2013 .
the observations were made with the pg1300 grating with the grating angle adjusted to ensure spectral coverage of lines redward of [ nev ] .
this gives a resolution of @xmath2 4 with a 1.5@xmath29 slit and a wavelength coverage of 2000 .
we observed each target for @xmath230 min .
the raw two - dimensional spectra were reduced and extracted using the iraf packages longslit and apextract . to determine the limitations of our multi - field catalogs and to compute the lae galaxy lf ,
we have developed a simulation to determine the completeness of recovery versus flux . for each field
, we added 1000 simulated emitters uniformly within the field s data cube .
we did not model morphology or size difference , since nearly all emitters are unresolved at the spatial ( @xmath56 ) and spectral resolution ( @xmath57 ) of the _ galex _ grism data .
we then ran our standard selection procedure and found the number of recovered objects .
we independently performed the above procedure ten times , giving a total of 10,000 input sources . in table
[ fields ] , we list the flux threshold above which each field is greater than 80@xmath54 complete .
as expected , the completeness limit scales as the inverse square root of the exposure time . below this threshold ,
the completeness of our sample rapidly declines ( see figure [ compl ] ) .
luminosity bin for the ew@xmath58(ly@xmath1 ) @xmath22 20 lae data cube sample .
the histogram shows the sample wide agn fractions , while the points with error bars show the agn fractions for regions with deep x - ray data . ]
( a color version of this figure is available in the online journal . ) in tables [ cdfs]-[cosmos ] , we list all of the lae candidates in the cdfs , groth , ngpdws , and cosmos fields ordered by redshift . in column 1
, we give the _ galex _ name ; in columns 2 and 3 , the j2000 right ascension and declination based on nuv position ; in columns 4 and 5 , the nuv and fuv ab magnitudes ; in column 6 , the redshift from the _ galex _ uv spectrum ; in column 7 , the logarithm of the ly@xmath1 luminosity ; in column 8 , the rest - frame ew@xmath58(ly@xmath1 ) with 1@xmath52 errors ; in column 9 , the logarithm of the x - ray flux ; in column 10 , the uv classified agns ; in columns 11 and 12 , the j2000 right ascension and declination based on optical position ; in column 13 , the offset between the optical to uv positions ; in column 14 , the optical redshift ; and in column 15 the optically classified agns .
we measured nuv and fuv ab magnitudes from the archival _ galex _ background subtracted intensity maps @xcite .
we first determined the magnitudes within 8@xmath29 diameter apertures centered on each of the emitter positions . to correct for flux that falls outside our apertures , we measured the offset between 8@xmath29 aperture magnitudes and _ galex _ pipeline total magnitudes for all bright cataloged objects ( 20 - 23 mag range ) within our fields .
we determined the median offset for each field and applied these to the magnitudes listed in columns 4 and 5 .
all objects are bright in the nuv when compared to their fuv magnitudes . in some cases
, we measure negative fuv fluxes . for these cases
, we list the magnitude corresponding to the absolute value of the flux with a minus sign in front to indicate that the flux was negative .
we corrected our one - dimensional nuv spectra for galactic extinction assuming a @xcite reddening law with r@xmath59=3.1 .
we obtained @xmath60 values from the @xcite recalibration of the @xcite extinction map as listed in the nasa / ipac extragalactic database ( ned ) .
galactic extinction modifies the ly@xmath1 flux by @xmath211% for the cosmos laes , @xmath24% for the groth laes , and @xmath26% for the cdfs and ngpdws laes . from these extinction
corrected spectra , we measured the redshifts , the ly@xmath1 fluxes , and the line widths using a two step process .
first , we fit a 140 rest - frame region around the ly@xmath1 line with a gaussian and a sloped continuum ( see figure [ ex](a ) ) .
a downhill simplex optimization routine was used to @xmath61 fit the five free parameters ( continuum level and slope plus gaussian center , width , and area ) .
we used the results of this fitting process to eliminate the two continuum parameters and as a starting point for the second step . in the second step
, we used the idl mpfit procedures of @xcite to @xmath61 fit the remaining three gaussian parameters .
we also employed the same fitting routine but with only a flat continuum ( one continuum free parameter instead of two ) .
we find that our results are not significantly affected by this model assumption . with the best - fit redshifts and ly@xmath1 fluxes , we calculated ly@xmath1 luminosities .
when available , we used the more precise optical redshift rather than the uv redshift to calculate the ly@xmath1 luminosities .
we list the uv redshifts and ly@xmath1 luminosities in columns 6 and 7 .
the rest - frame ew@xmath58(ly@xmath1 ) measured on the spectra are quite uncertain due to the very faint uv continuum .
we obtained a more accurate rest - frame ew by dividing the measured ly@xmath1 flux by the continuum flux measured from the broadband nuv image .
it is these rest - frame ews that are given in column 8 . in section 5 ,
we investigate the @xmath0 rest - frame ew distribution .
we made a classification of whether the emitter was an agn based on the presence of either broad or high - excitation emission lines in its uv or optical spectra or on the presence of an x - ray counterpart .
candidate x - ray counterparts were identified by matching all x - ray sources within a @xmath62 radius from the data cube position .
we then manually inspected the matches to reject false counterparts .
we list the x - ray flux of each identified counterpart in column 9 .
the x - ray luminosity threshold of @xmath63 erg s@xmath39 that is usually used to define agn activity @xcite corresponds to an x - ray flux of @xmath3710@xmath64 erg @xmath38 s@xmath39 for our @xmath65 sample .
all x - ray detected sources exceed this luminosity threshold , and we hereafter consider these sources to be agns . in column 10 , we list sources classified as agns based on the presence of broad or high - excitation emission lines in their uv spectrum ( see , e.g. , @xcite ) . our candidate star - forming galaxy sample will still contain some remaining agns , so we used the optical spectra to make a final determination of whether a galaxy is , in fact , star - forming . in column 15
, we identify optical agns based on the presence of either detectable [ nev ] or broad balmer or mgii@xmath552798 emission lines in the optical spectra .
we obtained optical redshifts for 122 of our 135 candidate laes .
we found that 15 optical redshifts did not match the measured uv redshifts .
for these objects , the ly@xmath1 luminosity and the rest - frame ew@xmath58(ly@xmath1 ) fields are left blank in columns 7 and 8 .
these sources are either spurious , stars , or strong civ@xmath551549 emitters .
for example , we found three lae candidates in our cosmos-00 field to be @xmath66 civ@xmath551549 emitters .
we indicate these objects by showing their optical redshift in parentheses or by setting their optical redshift and type to ` star ' in columns 14 and 15 .
additionally , we observed 4 candidate laes with keck deimos without recovering an optical redshift .
we indicate these objects by setting their optical redshifts to ` no z ' in column 14 . for the purpose of computing the lf
, we retain the nuv brightest ` no z ' source ( galex033045 - 274506 ) , which has two nearby untargeted potential optical counterparts .
the remaining ` no z ' candidate laes have relatively faint nuv counterparts and may be spurious . given the spatial resolution of our data cubes ( @xmath56 ) , we find that a total of five ` no z ' and non - matching redshift sources have potential alternative optical counterparts .
we list these ` [ alt ] ' sources in the table [ cdfs]-[cosmos ] notes .
overall , we exclude 18 optically unconfirmed lae candidates in the following discussion and in the construction of the @xmath0 ly@xmath1 lf .
* _ _ * excluding these 18 , we are left with a sample of 117 laes .
we classified 57 of these sources as uv , x - ray , or optical agn .
this establishes an agn fraction ( @xmath36@xmath67 ) of 49@xmath688@xmath54 for our sample .
after we exclude uv identified agns , our @xmath36@xmath67 drops to 41@xmath688@xmath54 .
when performing the optical spectroscopic followup , we preferentially targeted laes that were not classified as uv or x - ray agns . with this policy
, we obtained optical redshifts for 98% ( 59 out of 60 ) of our lae galaxies but only 89@xmath54 ( 51 out of 57 ) of our agns .
we found that 9 of the 37 laes classified as x - ray agns were already classified as agns based on their uv spectra .
we observed 19 laes classified as x - ray agns with keck deimos .
only 3 of these x - ray agns were not classified as optical agns . as a cross check to our sample s agn fraction of 49@xmath688@xmath54
, we only consider laes within deep x - ray fields and re - compute the agn fraction .
there are 79 candidate laes within deep x - ray fields ; we classified 44 ( or 56@xmath6810@xmath54 ) as agns .
this is not significantly different our sample - wide agn contamination . in figure
[ fagn ] , we show that the agn fraction increases with ly@xmath1 luminosity ( similar trends have been noted by @xcite and @xcite ) .
breaking each sample into 5 luminosity bins from log @xmath40@xmath69 = 42.5 to 43.4 ( used below in the ly@xmath1 lf computation ) and requiring ew@xmath58(ly@xmath1 ) @xmath22 20 , we find that the agn fraction increases from 6% to 71% for the full _ galex _ data cube sample and from 13% to 83% for the deep x - ray subsample .
given this small but systematic difference , we derive the ly@xmath1 lf for our full lae galaxy sample and for the subsample limited to regions with deep x - ray data to determine the effect , if any , on our results . in common with other low - redshift samples ,
our agn fraction is high compared to the agn fractions quoted for higher redshift samples .
( note that many of the low - redshift samples quote agn fractions that exclude sources known to be agns based on the uv spectra ; our agn fraction includes these sources . )
however , it is essential to emphasize that the agn fraction is strongly dependent on the luminosity range . * * our laes have ly@xmath1 luminosities from 10@xmath70 to 10@xmath71 erg s@xmath39 .
as shown in @xcite , the agn fraction rapidly increases over this range and approaches 100% at ly@xmath1 luminosities above 10@xmath71erg s@xmath39 .
the high - luminosity laes are primarily agns , and it is these objects that are included in the _ galex _ pipeline extractions . only with the increased sensitivity of the data cube search do we probe faint enough to develop large samples of star - forming laes . in figure [ pipe](a ) , we show the lae redshift versus the nuv magnitudes for our lae data cube sample ( blue diamonds ) and for the pipeline lae sample found by ( * ? ? ?
* ; * ? ? ?
* black squares ) constrained to the data cubes 50@xmath28@xmath3050@xmath28 fovs and redshift range ( @xmath42 ) .
we note that there are seven pipeline lae galaxies listed within the cdfs , groth , ngpdws , and cosmos fields ( see table 2 in * ? ? ?
we excluded one pipeline lae galaxy due to its low redshift @xmath72 , and we excluded two pipeline lae galaxies because they do not fall within our data cubes 50@xmath28@xmath3050@xmath28 fovs . although not excluded for the purposes of this discussion , if we also require ew@xmath18(ly@xmath1 ) @xmath22 20 , then there are only two lae galaxies that remain in the pipeline sample . by construction ,
the _ galex _ pipeline extractions miss all laes fainter than the pipeline magnitude limit of nuv@xmath73 ( dashed line in figure [ pipe](a ) ) .
objects classified as agns in any way have their symbols outlined in red . in figure [ pipe](b ) , we show the lae redshift versus the ly@xmath1 luminosity . directly below figures [ pipe](a ) and
[ pipe](b ) , we show these figures again but with all agns removed . we find that the agn fraction reaches 100% at a ly@xmath1 luminosity of @xmath74 erg s@xmath39
this is also the ly@xmath1 luminosity where the pipeline extraction begins to miss sources . while there is some luminosity overlap between the data cube and pipeline samples
, we find that our data cube search is necessary to obtain a large lae galaxy sample .
we computed the ly@xmath1 lf of the combined cdfs , groth , ngpdws , and cosmos lae galaxy samples in the redshift range @xmath42 using the @xmath75 technique @xcite .
we only included sources that were not classified as agn in any way and that have ew@xmath18(ly@xmath1)@xmath76 .
with these criteria , the sample includes two lae galaxies , which appeared in the pipeline sample of ( * ? ? ?
* see their table 2 ) . the rest ( 58 ) come from our data cube sample and
are not found in the _ galex _ pipeline extractions .
ccc cdfs & 8.5@xmath3010@xmath77 & @xmath78 + groth & 5.2@xmath3010@xmath77 & @xmath79 + ngpdws & 10.9@xmath3010@xmath77 & @xmath80 + cosmos & 24.1@xmath3010@xmath77 & @xmath81 + [ flimit ] to compute the lf , we divided our ly@xmath1 survey into 20 samples ( 5 samples per field ) , where the sample @xmath82 covers a ly@xmath1 flux range @xmath83 and a solid angle area @xmath84 ( the masking of bright sources alters @xmath84 from one _ galex _ 50@xmath28@xmath3050@xmath28 field to another ) . for each field , we took the difference between the minimum and maximum observed ly@xmath1 flux and divided this into 5 flux bins of equal size ( see table [ flimit ] for minimum ly@xmath1 flux values and corresponding completeness limits ) .
the sampling rate @xmath85 ( the fraction of lae galaxies in the given flux range that were observed ) was estimated with our completeness simulations .
the effective area of a sample is @xmath86 .
the accessible volume of the @xmath87 lae galaxy in sample @xmath82 is @xmath88 where the comoving volume is calculated at the highest and lowest redshifts at which the @xmath87 lae galaxy remains both within sample @xmath82 s ly@xmath1 flux range @xmath83 and within the redshift range @xmath42 .
the total accessible volume of the @xmath87 lae galaxy is @xmath89 and the ly@xmath1 lf in the luminosity range @xmath90 and redshift range @xmath42 is @xmath91 in figure [ lf](a ) , we show our raw lf ( black open circles ) along with our lf corrected for incompleteness ( black solid circles ) .
error bars are @xmath921@xmath52 poisson errors . in table
[ tlf ] columns 1 , 2 , and 3 , we list the luminosity bins , the number of lae galaxies per bin , and the computed luminosity function values corrected for incompleteness .
we have tested that these results are not dependent on our method of constructing flux bins .
alternatively , for each field , we generated 5 flux bins of equal size within the range of @xmath93 to the maximum observed ly@xmath1 flux , where @xmath93 is the ly@xmath1 flux corresponding to 50% completeness ( see figure [ compl ] ) .
this method reduces the dependence on our completeness simulations but does not significantly change our results .
ccc 42.468 - 42.655 & 17 & [email protected]@xmath94 + 42.655 - 42.842 & 13 & [email protected]@xmath94 + 42.842 - 43.029 & 15 & [email protected]@xmath94 + 43.029 - 43.216 & 8 & [email protected]@xmath94 + 43.216 - 43.403 & 7 & [email protected]@xmath94 + [ tlf ] we may correct for any remaining agns in our sample by restricting our field to regions with deep x - ray data .
this removes the ngpdws field and restricts the area of the remaining 50@xmath9550@xmath28 _ galex _ fields but ensures a robust agn classification . in figure [ lf](b ) , we show our x - ray data limited lf ( black circles ) . comparing this lf to the lf computed from the full lae galaxy sample , we find that all points are consistent within 1@xmath52 error bars .
we find that our lae galaxy lf is consistent with the @xmath0 lae galaxy lf computed in bcw12 .
bcw12 based their lf on @xmath4 lae galaxies found in one of the four _ galex _ fields contained in this study .
we have increased the sample size by a factor of three .
furthermore , we have searched for agns in the bcw12 lae galaxy sample ( none found ) and removed spurious sources ( 3 bcw12 lae candidates ) with our newly obtained optical spectroscopic data .
in figure [ ew_lalum ] , we show our rest - frame ews versus redshift .
because we have a flux - limited sample , the distribution of these ews should be directly comparable to narrowband ly@xmath1 selected samples at higher redshifts . in figure [ ew_dists](a ) , we show the distribution of these ews for the full sample .
ly@xmath1 luminosity for the full sample.,width=321 ] . *
( b ) * the @xcite sample at @xmath96 . * ( c ) * the @xcite sample at @xmath97 . in each case
, the blue curve shows the exponential distribution computed using a maximum likelihood procedure .
the rest - frame scale length is 80@xmath9210 for @xmath0 , 56@xmath927 for @xmath96 , and 56@xmath927 for @xmath97.,width=321 ] . *
( b ) * the @xcite sample at @xmath96 . * ( c ) * the @xcite sample at @xmath97 . in each case
, the blue curve shows the exponential distribution computed using a maximum likelihood procedure .
the rest - frame scale length is 80@xmath9210 for @xmath0 , 56@xmath927 for @xmath96 , and 56@xmath927 for @xmath97.,width=321 ] . *
( b ) * the @xcite sample at @xmath96 . * ( c ) * the @xcite sample at @xmath97 . in each case
, the blue curve shows the exponential distribution computed using a maximum likelihood procedure .
the rest - frame scale length is 80@xmath9210 for @xmath0 , 56@xmath927 for @xmath96 , and 56@xmath927 for @xmath97.,width=321 ] ( a color version of this figure is available in the online journal . )
sample ( black square ) , the @xmath96 @xcite sample ( green triangle ) , and the @xmath98 @xcite , @xmath97 @xcite , and @xmath97 @xcite samples ( blue diamonds ; values quoted from ciardullo et al . ) vs. redshift .
the red circle shows the scale length from the @xmath99 _ galex _ lae sample of @xcite after correction for their continuum selection , as described in their section 5.4.,width=321 ] ( a color version of this figure is available in the online journal . )
the rest - frame ew distribution is normally fit with an exponential function . for the lae samples , which are truncated below @xmath19 ,
the maximum likelihood estimate of the scale length is the mean of the rest - frame ew values minus @xmath19 .
we find a scale length of @xmath100 for the full sample , where we computed the error using the parameterized bootstrap method .
separating the sample by luminosity gives a scale length of @xmath101 for sources with logarithmic ly@xmath1 luminosities above 42.8 and @xmath102 for those with lower luminosities .
this is consistent with the distribution being invariant as a function of ly@xmath1 luminosity , as has been found in the higher redshift samples @xcite .
we compare our sample at @xmath0 with a sample at @xmath96 from @xcite ( figure [ ew_dists](b ) ) and a sample at @xmath97 from @xcite ( figure [ ew_dists](c ) ) .
the nilsson et al .
sample has a logarithmic luminosity limit of 42.4 erg s@xmath39 , which is very similar to that of the present sample , while the ciardullo et al .
sample covers 42 erg s@xmath39 to just above 43 erg s@xmath39 , making it slightly fainter , on average .
the nilsson et al . and ciardullo et al .
maximum likelihood estimates of the rest - frame ew scale length are both @xmath103 .
nilsson et al . also give a least squares fit value of @xmath103 , which is consistent within the errors .
however , when @xcite allowed for biases in their sample , such as those introduced by the filter shape , then their rest - frame ew scale length rose to @xmath104 .
when ciardullo et al.analyzed an alternate sample at @xmath97 from @xcite in the same way , they found @xmath105 . from the combined samples , they found @xmath106 .
ciardullo et al .
also reanalyzed the @xmath98 sample from @xcite in a self - consistent way and found @xmath107 .
the ciardullo et al .
@xmath98 value is consistent with the @xcite @xmath96 value .
based on an anderson - darling test , they concluded that there is a significant difference between the @xmath97 and @xmath98 samples .
the @xcite analysis emphasizes that differences inherent in the methodology and fitting procedure are important .
in particular , comparisons between the present sample , where the line fluxes are computed from the spectra and the continua from the broadband magnitudes , and samples such as the ciardullo et al . sample , which are based purely on imaging data , can be tricky ( e.g. , * ? ? ? * ) .
nevertheless , the present sample appears to indicate that , if there is indeed a narrowing of the rest - frame ew distribution from @xmath108 to @xmath109 , then it has reversed by @xmath110 . in figure
[ ew_z ] , we show the rest - frame ew scale lengths measured from the maximum likelihood procedures for the present sample ( black square ) , the @xmath96 @xcite sample ( green triangle ) , and the @xmath98 @xcite , @xmath97 @xcite , and @xmath97 @xcite samples ( blue diamonds ; values quoted from ciardullo et al . ) while there may be a drop at @xmath5 , overall the scale length shows little variation with redshift .
this is in sharp contrast to the analysis of @xcite , who found a redshift evolution of @xmath111 . the zheng et al .
result appears to be due primarily to their misinterpretation of the scale length given by @xcite for their @xmath99 _ galex _ lae sample .
zheng et al .
used @xmath112 , which was what cowie et al .
gave for their continuum selected sample .
however , as cowie et al . emphasized in their section 5.4 , this needs to be corrected for missing high ew sources before it can be compared with ly@xmath1 luminosity selected samples . in figure
[ ew_z ] , we show the cowie et al
. corrected scale length of 75 ( red circle ) .
( because of the large systematic errors , we do not show an error bar . )
this is fully consistent with there being a constant value of the scale length with redshift .
luminosity vs. redshift for lae galaxies with ew@xmath18(ly@xmath1 ) @xmath22 20 .
we show pipeline _ galex _ laes from ( * ? ? ?
* red triangles and red diamonds ) , data cube _ galex _ laes from this study ( black circles ) , @xmath113 hetdex laes from ( * ? ? ?
* magenta right pointing triangles ) , @xmath114 laes from ( * ? ? ?
* green squares ) , @xmath115 laes from ( * ? ? ?
* blue left pointing triangles ) , @xmath116 laes from ( * ? ? ?
* blue squares ) , @xmath117 laes from ( * ? ? ?
* magenta diamonds ) , and @xmath118 laes from ( * ? ? ?
* green diamonds ) and ( * ? ?
* magenta triangles ) .
the red line shows the @xmath119 derived for the @xmath3 lf assuming a slope of @xmath120 .
the cyan area shows the range of @xmath119 values derived for the @xmath121 lfs . ]
( a color version of this figure is available in the online journal . )
luminosity function at @xmath42 for the lae galaxies in deep _
galex _ grism fields ( black circles from figure [ lf](a ) ) compared to other ly@xmath1 luminosity functions in close redshift proximity .
the fitted schechter functions are shown as solid curves over the the extent of their observed data points .
the black curve indicates our best fit schechter function assuming a fixed slope of @xmath120 .
the red curve indicates the @xmath122 @xcite lf .
the blue curve indicates the @xmath98 @xcite lf .
the green curve indicates the @xmath123 @xcite lf .
the cyan curve indicates the @xmath124 @xcite lf .
the magenta curve indicates the @xmath113 @xcite lf . ]
( a color version of this figure is available in the online journal . ) , 2@xmath52 , and 3@xmath52 * _ _ * contours for the @xmath0 luminosity function parameters @xmath119 and @xmath125 assuming a faint end slope of @xmath126 ( blue ) and @xmath127 ( red ) .
the blue , black , and red circles show our best fit schechter function results for the full lae galaxy sample assuming @xmath126 ( blue circle ) , @xmath120 ( black circle ) , and @xmath127 ( red circle ) .
regardless of the assumed faint end slope , log @xmath128 . ]
( a color version of this figure is available in the online journal . ) vs. redshift for the lae galaxy lfs shown in figure [ lf ] .
the black solid circle shows the product @xmath129 from our best fit @xmath0 lae galaxy lf .
the black open circle shows the product @xmath130 from our fit to the @xmath0 lae galaxy lf limited to regions with deep x - ray data .
both @xmath0 fits assume a faint end slope @xmath120 .
error bars are estimated from the published uncertainties in @xmath119 and @xmath125 . ]
( a color version of this figure is available in the online journal . )
escape fractions for all lfs presented in figure [ lf ] .
the black solid circle shows @xmath131 computed from our best fit @xmath0 lf .
the black open circle shows @xmath132 computed from our fitted lf limited to regions with deep x - ray data .
diamonds ( from low redshift to high ) indicate @xmath133 derived from @xcite , @xcite , @xcite , @xcite , and @xcite .
error bars are estimated from the published uncertainties in @xmath119 and @xmath125 .
the solid and dashed black curves are from @xcite and @xcite , respectively , and show the best fit power law to @xmath133 data .
the dotted black curve shows the best fit transition model from @xcite .
the calculated points and the selected curves have not been corrected for igm absorption , which should be negligible at @xmath0 and within the 1@xmath52 uncertainties at @xmath134 . ]
in figure [ lstar ] , we show the dramatic increase in the observed ly@xmath1 luminosities to a redshift of @xmath0 followed by relatively no luminosity evolution from @xmath0 to @xmath135 .
this large luminosity boost was previously noted by @xcite with a sample of @xmath136 laes at a redshift of @xmath0 ( see figure [ lstar ] , red diamonds ) . we have now increased the sample size of @xmath0 laes to 60 galaxies ( see figure [ lstar ] , black circles ) .
our relatively nearby @xmath0 sample will facilitate the study of laes with luminosities analogous to high - redshift laes . in paper ii , we will use the available spectroscopic and imaging data to constrain the properties of our sample and to serve as a baseline for studies of higher redshift laes . in order to quantify the evolution in @xmath119 ,
we have held the slope @xmath1 fixed and found the best @xcite function fit to our data .
we assume a slope of @xmath120 , which allows us to compare directly our results to the lae galaxy lf at @xmath3 ( @xmath120 , log @xmath137 , log @xmath138 ; * ? ? ?
we obtain log @xmath139 and log @xmath140 ( see figure [ lf](a ) ) . limiting our sample to regions with deep x - ray data , we obtain log @xmath141 and log @xmath142 ( see figure [ lf](b ) ) . both our @xmath143 values are roughly consistent with @xmath119 values found for lae galaxy lfs at @xmath21 ( log @xmath144 , @xcite ; log @xmath145 , @xcite ; log @xmath145 , @xcite ) and imply an @xmath119 increase by a factor of @xmath146 from @xmath3 to @xmath0 ( from log @xmath147 to log @xmath148 ) .
a large boost in @xmath119 is in agreement with the evolution seen in other lfs tracing star - forming galaxies .
both h@xmath1 and uv lfs are found to be dominated by luminosity evolution over the redshift range @xmath149 .
@xcite investigated h@xmath1 lfs and found that @xmath150 increases by a factor of 10 from @xmath151 to @xmath5 .
@xcite investigated uv lfs and found that @xmath152 increases by a factor of @xmath153 from @xmath151 to @xmath21 .
however , our fitted schechter parameters indicate a decrease in @xmath125 from @xmath3 to @xmath0 ( from log @xmath154 to log @xmath155 ) .
it is only beyond @xmath0 that there a significant increase in @xmath125 . referring again to other lfs tracing star - forming galaxies
, we note that , unlike @xmath119 , @xmath125 does not increase monotonically over the redshift range @xmath149 . within the uncertainties of their measurements , both ( * ? ? ?
* for @xmath156 ) and ( * ? ? ?
* for @xmath157 ) found very little evidence for @xmath125 evolution from @xmath151 to @xmath5 .
as discussed in more detail below , we find that our best fit schechter functions with higher @xmath119 values tend to have lower @xmath125 values .
given this issue and the poorly constrained faint end slope , a clearer picture of the evolution of individual schechter function parameters awaits a @xmath0 lae survey probing lower ly@xmath1 luminosities .
such a survey is not possible with currently available telescopes but would allow stronger constraints on the @xmath143 and @xmath158 values . in figure [ complf ] , we compare our @xmath0 lf computed from the full lae galaxy sample to other ly@xmath1 luminosity functions in close redshift proximity .
we show the fitted schechter functions as solid lines over the the extent of their observed data points . for the @xmath5 schechter function parameters , the log @xmath40@xmath159 values range from 42.3 to 43.2 , the log @xmath125 values range from @xmath160 to @xmath161 , and the @xmath1 values range from @xmath162 to @xmath163 ( see the blue , green , cyan , and magenta lines in figure [ complf ] ) .
due to the differing survey volumes and fitting assumptions , it is still possible that all @xmath5 data are in overall agreement @xcite . the schechter function parameters @xmath125 and @xmath119 are known to be strongly correlated , while assuming different values of @xmath1 tends to leave the product @xmath164 unchanged .
we show this for our schechter function fit in figure [ contour ] .
for these reasons , we consider the product @xmath164 and the luminosity density , which is proportional to @xmath164 , to be more reliable than the values of individual schechter parameters . in figure
[ lphi ] , we show that a coherent picture emerges by comparing our results to the @xmath164 values found in the literature . as noted by @xcite
, there is a trend ( in agreement with our uncertainty contours ) for the @xmath5 schechter function fits with higher @xmath119 values to have lower @xmath125 values .
@xmath165thus , the @xmath5 schechter functions which have @xmath119 values that differ by about an order of magnitude are found to have consistent @xmath164 values .
the product @xmath164 increases by a factor of @xmath166 from @xmath3 to @xmath0 followed by an increase by a factor of @xmath4 from @xmath0 to @xmath5 .
the observed ly@xmath1 luminosity density provides an estimate of the total amount of ly@xmath1 light emitted by galaxies per unit volume : @xmath167 where @xmath168 is a ly@xmath1 schechter function .
integrating this expression from @xmath169 to @xmath170 gives @xmath171 where @xmath172 is the gamma function .
the main source of uncertainty in this expression comes from the poorly constrained faint end slope @xmath1 . taking the extreme values for the @xmath5 schechter functions , @xmath127 and @xmath126
, we find that @xmath172 changes by a factor of 1.7 .
modulo this factor , the evolution of the product @xmath164 seen in figure [ lphi ] from @xmath3 to @xmath5 results from the change in observed ly@xmath1 luminosity density .
to first order , the intrinsic production of ly@xmath1 photons in a galaxy is proportional to the sfr @xcite .
thus , an overall increase in the ly@xmath1 luminosity density from @xmath3 to @xmath5 is expected due to the well established increase in the cosmic sfr density . however , observed ly@xmath1 luminosities will significantly deviate from intrinsic ly@xmath1 luminosities due to resonant scattering of ly@xmath1 photons by neutral hydrogen .
to quantify this , we calculate the @xmath0 volumetric ly@xmath1 escape fraction , @xmath133 , and compare its value to other measurements taken at nearby redshifts .
the volumetric ly@xmath1 escape fraction measures the fraction of ly@xmath1 photons that escape from the survey volume .
we emphasize that its value will be systematically lower than the ly@xmath1 escape fractions derived for individual laes .
for example , @xcite find a @xmath173 median ly@xmath1 escape fraction for their @xmath113 lae sample but only a @xmath174 volumetric ly@xmath1 escape fraction .
the volumetric escape fraction is defined as the observed ly@xmath1 luminosity density divided by the intrinsic ly@xmath1 luminosity density , @xmath175 where the intrinsic ly@xmath1 luminosity density is found by taking a measure of the cosmic sfr density and then converting sfr to ly@xmath1 luminosity , @xmath176 the conversion factors come from the assumption of case b recombination @xcite and the sfr to h@xmath1 @xcite relation . for consistency with @xcite
, we integrate the ly@xmath1 lfs down to a @xmath177erg s@xmath39 limit and obtain cosmic sfr density measurements from @xcite .
@xcite estimate that the choice of the luminosity integration limit may alter the computed luminosity density by at most @xmath178 ( for further discussion of this issue see * ? ? ?
@xcite derive cosmic sfr densities from extinction corrected rest - frame uv lfs .
we do not attempt to correct for igm absorption , which should be negligible for our @xmath0 sample .
both @xcite and @xcite have compiled measurements from various ly@xmath1 lf studies spanning a redshift range of @xmath179 to determine the evolution of @xmath133 . both groups fit power laws to these data and find evidence for a rapidly increasing @xmath133 , which approaches 100@xmath54 at high redshifts ( @xmath180 ) .
we find a @xmath0 @xmath133 of @xmath181 . in figure
[ fesc ] , we show that the @xmath133 computed from our best fit @xmath0 schechter function is consistent with the @xmath133 value of @xmath182 interpolated from the power law fits of @xcite and @xcite . in figure
[ fesc ] , we also show @xmath132 ( open circle ) computed from our fitted lf limited to regions with deep x - ray data .
@xcite argue that the rising trend in the volumetric ly@xmath1 escape fraction with increasing redshift is consistent with expectations due to the general decline of dust content in star - forming galaxies .
we have now constrained the @xmath133 from @xmath3 to @xmath5 and find results consistent with this inferred trend .
we presented a catalog of 135 candidate @xmath0 laes .
we obtained optical spectral data for 90% of our sample .
we found that only @xmath213% of our sample are either spurious , stars , or strong civ@xmath551549 emitters . combining these optical data with the uv spectra and x - ray imaging data
, we found that 49% of our ly@xmath1 emitters are agns .
eliminating agns and laes with ew@xmath18(ly@xmath1)@xmath2320 gives a final sample of 60 star - forming laes , which we used to compute the @xmath0 ly@xmath1 lf and the @xmath0 ly@xmath1 volumetric escape fraction ( @xmath131 ) .
we note that no improved @xmath0 ly@xmath1 lfs are possible for the foreseeable future due to the lack of uv telescopes .
our best fit lf implies a significant increase in @xmath119 between @xmath3 and @xmath0 without much change in the luminosity density .
this requires a rapid increase in @xmath125 between @xmath0 and @xmath5 .
it is clear that the intrinsic properties of the laes must be changing rapidly in the @xmath183 interval , resulting in a rapid increase in the ly@xmath1 escape fraction .
our cataloged sources offer the best opportunity to study emitters with luminosities comparable to laes found in the early universe . with the optical data in hand
, paper ii will compare our sample s physical properties to a uv - selected control sample .
, e. b. , nordsieck , k. h. , kobulnicky , h. a. , et al .
2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4841 , society of photo - optical instrumentation engineers ( spie ) conference series , ed . m. iye & a. f. m. moorwood , 14631471 , s. m. , phillips , a. c. , kibrick , r. i. , et al .
2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4841 , society of photo - optical instrumentation engineers ( spie ) conference series , ed . m. iye & a. f. m. moorwood , 16571669 , g. j. , gebhardt , k. , komatsu , e. , et al .
2008 , in astronomical society of the pacific conference series , vol .
399 , panoramic views of galaxy formation and evolution , ed .
t. kodama , t. yamada , & k. aoki , 115 ccccccccccccccc galex033111 - 281725 & 52.796374 & -28.290444 & 23.55 & 27.50 & 1.168 & 43.61 & 281@xmath9217 & & & 52.796253 & -28.290321 & 0.6 & 1.164 & agn + galex033112 - 281517 & 52.800377 & -28.254936 & 23.02 & 28.01 & 1.162 & 43.31 & 87@xmath925 & & & 52.800663 & -28.255022 & 1.0 & 1.161 & + galex033359 - 275326 & 53.499160 & -27.890792 & 22.04 & 24.13 & 1.143 & 43.53 & 63@xmath922 & & agn & 53.499096 & -27.890846 & 0.3 & 1.133 & agn + galex033301 - 273227 & 53.255196 & -27.540939 & 23.45 & -27.75 & 1.135 & 43.08 & 82@xmath927 & & & 53.255268 & -27.540880 & 0.3 & 1.136 & + galex033146 - 272942 & 52.942574 & -27.495223 & 22.73 & 25.44 & 1.137 & & & & & 52.942543 & -27.495167 & 0.3 & ( 0.629 ) & + galex033348 - 280216 & 53.453649 & -28.038008 & 23.40 & 25.53 & 1.128 & 43.09 & 82@xmath928 & & & 53.453671 & -28.038050 & 0.2 & 1.130 & + galex033230 - 280804 & 53.125683 & -28.134514 & 22.34 & 23.83 & 1.124 & 43.27 & 47@xmath923 & & agn & & & & & + galex033120 - 274917 & 52.837504 & -27.821383 & 22.82 & 32.73 & 1.116 & 42.91 & 30@xmath925 & & & 52.837444 & -27.821638 & 0.9 & 1.113 & + galex033255 - 281320 & 53.231551 & -28.222466 & 22.46 & -27.50 & 1.087 & 42.95 & 28@xmath923 & & & 53.231567 & -28.222486 & 0.1 & 1.085 & + galex033112 - 281442 & 52.803095 & -28.245108 & 22.19 & -28.20 & 1.055 & 42.85 & 19@xmath923 & & agn & 52.801987 & -28.244835 & 3.6 & 1.042 & + galex033413 - 275940 & 53.555847 & -27.994665 & 24.21 & 26.59 & 1.052 & 42.78 & 118@xmath9212 & & & 53.555477 & -27.994663 & 1.2 & 1.051 & agn + galex033212 - 274913 & 53.051599 & -27.820316 & 22.68 & 27.04 & 1.053 & 42.89 & 33@xmath924 & & & 53.051613 & -27.820168 & 0.5 & 1.048 & + galex033200 - 274319 & 53.001430 & -27.722033 & 22.40 & 24.39 & 1.042 & 43.25 & 59@xmath924 & -14.2 & & 53.001520 & -27.722071 & 0.3 & 1.041 & agn + galex033228 - 273614 & 53.121022 & -27.604317 & 23.62 & 26.66 & 1.036 & 42.77 & 59@xmath929 & & & 53.118469 & -27.604019 & 8.2 & 1.046 & + galex033113 - 274949 & 52.807630 & -27.830242 & 22.45 & 31.95 & 1.031 & 42.81 & 24@xmath923 & & & 52.807545 & -27.830231 & 0.3 & 1.026 & + galex033331 - 280625 & 53.382523 & -28.107164 & 24.41 & 27.66 & 1.029 & 42.80 & 141@xmath9214 & & & 53.382084 & -28.106623 & 2.4 & 1.026 & agn + galex033202 - 280320 & 53.010734 & -28.055702 & 23.44 & 25.22 & 1.018 & 42.94 & 81@xmath926 & -13.7 & agn & 53.010658 & -28.055834 & 0.5 & 1.015 & agn + galex033111 - 275506 & 52.798577 & -27.918436 & 23.81 & -27.52 & 1.010 & 43.12 & 168@xmath9217 & & & 52.798473 & -27.918442 & 0.3 & 1.023 & + galex033336 - 274224 & 53.402806 & -27.706674 & 23.02 & 26.37 & 0.977 & 43.22 & 119@xmath926 & -14.8 & agn & 53.402771 & -27.706703 & 0.2 & 0.976 & + galex033204 - 273725 & 53.017064 & -27.623785 & 22.78 & 24.25 & 0.977 & 43.36 & 133@xmath928 & -13.6 & agn & 53.016788 & -27.623751 & 0.9 & 0.970 & agn + galex033246 - 274154 & 53.195680 & -27.698408 & 23.17 & 26.61 & 0.976 & 43.20 & 128@xmath9210 & & & 53.196271 & -27.698547 & 2.1 & no z & + galex033342 - 274035 & 53.425620 & -27.676525 & 24.73 & 26.03 & 0.956 & 42.59 & 146@xmath9215 & & & 53.425163 & -27.676567 & 1.5 & 0.948 & + galex033335 - 273934 & 53.398183 & -27.659591 & 22.54 & 23.31 & 0.949 & & & -14.3 & & 53.398163 & -27.659760 & 0.6 & ( 0.622 ) & + galex033329 - 280127 & 53.373274 & -28.024282 & 23.16 & 25.95 & 0.942 & 42.62 & 37@xmath926 & & & 53.373272 & -28.024244 & 0.1 & 0.940 & + galex033206 - 281408 & 53.026465 & -28.235703 & 21.79 & 25.75 & 0.913 & 43.30 & 56@xmath923 & & & 53.026493 & -28.235905 & 0.7 & 0.909 & + galex033044 - 280237 & 52.686897 & -28.043707 & 23.25 & 26.58 & 0.869 & 42.82 & 116@xmath9213 & & & 52.686813 & -28.043819 & 0.5 & 0.859 & + galex033042 - 274215 & 52.679011 & -27.704257 & 23.03 & 25.18 & 0.861 & 42.57 & 38@xmath924 & & & 52.678730 & -27.704462 & 1.2 & 0.856 & + galex033409 - 280129 & 53.541285 & -28.024923 & 23.34 & 26.40 & 0.857 & 42.63 & 60@xmath926 & & & 53.540737 & -28.024443 & 2.5 & 0.854 & + galex033251 - 280809 & 53.214351 & -28.135966 & 22.70 & 26.17 & 0.838 & 42.71 & 43@xmath924 & & & 53.214489 & -28.136051 & 0.5 & 0.832 & + galex033314 - 274834 & 53.310626 & -27.809526 & 24.23 & 25.48 & 0.830 & 42.65 & 155@xmath9221 & & & 53.310975 & -27.809095 & 1.9 & 0.826 & + galex033045 - 274506 & 52.689797 & -27.751818 & 23.04 & 25.67 & 0.832 & 42.74 & 63@xmath926 & & & 52.689950 & -27.752197 & 1.4 & no z & + galex033405 - 275023 & 53.521292 & -27.839993 & 23.13 & 25.03 & 0.825 & 42.56 & 47@xmath929 & & & 53.520832 & -27.839972 & 1.5 & 0.821 & + galex033057 - 273316 & 52.739776 & -27.554542 & 22.37 & 24.09 & 0.794 & 42.73 & 38@xmath923 & & & 52.739708 & -27.554583 & 0.3 & 0.790 & + galex033124 - 275625 & 52.851698 & -27.940441 & 22.69 & 25.82 & 0.778 & 42.57 & 36@xmath924 & & & 52.851761 & -27.940475 & 0.2 & 0.773 & + galex033235 - 274059 & 53.148932 & -27.683184 & 23.67 & 25.70 & 0.745 & 42.79 & 182@xmath9215 & & & 53.149258 & -27.683273 & 1.1 & 0.735 & + galex033146 - 274846 & 52.942503 & -27.812779 & 22.98 & 25.88 & 0.742 & 42.75 & 86@xmath926 & & & 52.942345 & -27.812771 & 0.5 & 0.736 & + galex033100 - 273020 & 52.751186 & -27.505665 & 22.52 & 24.80 & 0.733 & 42.68 & 51@xmath925 & & & 52.750877 & -27.505367 & 1.5 & 0.727 & + galex033131 - 273429 & 52.880399 & -27.574774 & 22.56 & 25.77 & 0.695 & 43.28 & 280@xmath9211 & -14.1 & & 52.880260 & -27.574841 & 0.5 & 0.688 & agn + galex033321 - 275031 & 53.341285 & -27.842052 & 23.16 & 25.06 & 0.691 & 42.59 & 90@xmath929 & & agn & & & & & + galex033150 - 274116 & 52.962287 & -27.687773 & 22.28 & 23.15 & 0.673 & 42.70 & 54@xmath925 & -14.2 & agn & 52.962346 & -27.687700 & 0.3 & 0.667 & + [ cdfs ] ccccccccccccccc galex142028 + 524640 & 215.11946 & 52.777836 & 23.41 & 26.75 & 1.131 & 42.57 & 25@xmath924 & & & 215.11989 & 52.777527 & 1.5 & 1.134 & + galex141814 + 524415 & 214.56075 & 52.737531 & 23.97 & -27.83 & 1.128 & 43.10 & 144@xmath9214 & & & 214.56084 & 52.737473 & 0.3 & 1.123 & + galex141950 + 522542 & 214.96249 & 52.428532 & 24.09 & 26.34 & 1.114 & 42.97 & 124@xmath9212 & & & 214.96179 & 52.427998 & 2.5 & 1.111 & + galex142013 + 524008 & 215.05510 & 52.669118 & 25.11 & 29.17 & 1.100 & 43.05 & 380@xmath9249 & & & 215.05558 & 52.669342 & 1.2 & no z & + galex141842 + 522919 & 214.67858 & 52.488791 & 22.69 & 23.68 & 1.084 & 42.98 & 37@xmath925 & -14.3 & & 214.67863 & 52.488586 & 0.7 & 1.084 & agn + galex141824 + 522329 & 214.60416 & 52.391463 & 23.51 & 24.99 & 1.083 & 43.12 & 110@xmath9212 & -13.6 & & 214.60379 & 52.391666 & 1.1 & 1.081 & agn + galex141718 + 525521 & 214.32734 & 52.922402 & 22.47 & 24.41 & 1.067 & 43.58 & 127@xmath924 & & agn & 214.32666 & 52.922764 & 2.0 & 1.064 & agn + galex142028 + 522516 & 215.12077 & 52.421360 & 23.00 & 26.88 & 1.052 & 42.94 &
51@xmath924 & & & 215.12045 & 52.421455 & 0.8 & 1.045 & agn + galex141842 + 523735 & 214.67549 & 52.626454 & 23.07 & 25.43 & 1.038 & 43.17 & 93@xmath9214 & & & 214.67545 & 52.625916 & 1.9 & 1.035 & agn + galex141737 + 524236 & 214.40603 & 52.709992 & 22.20 & 26.33 & 1.027 & 43.53 & 99@xmath926 & -13.3 & & 214.40543 & 52.710182 & 1.5 & 1.025 & agn + galex141815 + 530452 & 214.56554 & 53.081148 & 22.36 & 23.26 & 1.028 & 43.07 & 41@xmath923 & & & 214.56442 & 53.081223 & 2.4 & 1.020 & agn + galex142015 + 525024 & 215.06357 & 52.840222 & 24.04 & -28.66 & 1.021 & 43.13 & 223@xmath9220 & & & 215.06401 & 52.840656 & 1.8 & 1.014 & + galex142144 + 523627 & 215.43607 & 52.607699 & 21.83 & 24.66 & 1.011 & 43.32 & 46@xmath923 & & & 215.43503 & 52.607677 & 2.3 & 1.006 & agn + galex142027 + 530455 & 215.11514 & 53.081989 & 23.18 & -27.68 & 1.002 & 42.82 & 54@xmath926 & -13.1 & & 215.11458 & 53.081806 & 1.4 & 0.994 & agn + galex141851 + 523600 & 214.71624 & 52.600051 & 23.54 & 26.00 & 1.000 & 43.00 & 112@xmath9214 & -14.1 & & 214.71530 & 52.599945 & 2.1 & 0.986 & agn + galex142242 + 525245 & 215.67896 & 52.879323 & 22.07 & -26.74 & 1.003 & 42.65 & 12@xmath923 & & & 215.67851 & 52.879360 & 1.0 & 0.999 & + galex142033 + 530617 & 215.14121 & 53.104797 & 23.62 & 25.29 & 0.999 & 42.80 & 75@xmath928 & -14.5 & & 215.14162 & 53.105114 & 1.5 & 0.994 & + galex141902 + 525637 & 214.76161 & 52.943789 & 22.47 & 24.55 & 0.984 & 42.56 & 21@xmath924 & & & 214.76091 & 52.943871 & 1.6 & 0.980 & + galex142207 + 525411 & 215.52938 & 52.903111 & 23.07 & 25.35 & 0.996 & 43.16 & 104@xmath925 & & & 215.52910 & 52.902931 & 0.9 & 0.994 & agn + galex142148 + 522902 & 215.45007 & 52.484033 & 23.04 & 27.40 & 0.961 & 43.23 & 131@xmath9213 & & & 215.44913 & 52.484470 & 2.6 & 0.961 & + galex142244 + 525442 & 215.68588 & 52.911824 & 23.85 & 26.89 & 0.958 & 43.26 & 300@xmath9222 & & & 215.68564 & 52.911930 & 0.6 & 0.954 & + galex141746 + 525259 & 214.44522 & 52.882997 & 24.17 & -27.28 & 0.957 & 42.58 & 87@xmath9210 & & & 214.44400 & 52.884399 & 5.7 & 0.957 & + galex141946 + 524755 & 214.94289 & 52.798752 & 22.23 & 28.91 & 0.928 & 42.21 & 6@xmath923 & & & 214.94319 & 52.798855 & 0.7 & 0.914 & + galex141935 + 524127 & 214.89632 & 52.690888 & 22.13 & 24.21 & 0.920 & 42.72 & 20@xmath923 & & & 214.89662 & 52.690529 & 1.4 & 0.917 & + galex141733 + 530403 & 214.39094 & 53.067620 & 23.21 & 24.87 & 0.918 & 43.07 & 120@xmath9210 & & & 214.39050 & 53.067913 & 1.4 & 0.915 & agn + galex141833 + 525525 & 214.64107 & 52.923739 & 23.46 & 26.32 & 0.916 & 42.74 & 72@xmath927 & & & 214.64212 & 52.924011 & 2.5 & 0.912 & + galex142201 + 530637 & 215.50453 & 53.110604 & 24.13 & 25.62 & 0.912 & 43.19 & 398@xmath9253 & -13.3 & & 215.50462 & 53.110195 & 1.5 & 0.898 & agn + galex142202 + 530823 & 215.50844 & 53.139988 & 22.85 & 24.75 & 0.872 & & & & & 215.50819 & 53.140114 & 1.3 & ( 0.572 ) & + galex142028 + 525839 & 215.11758 & 52.977690 & 23.30 & -27.77 & 0.870 & 42.81 & 83@xmath928 & -13.9 & & 215.11682 & 52.978155 & 2.3 & 0.871 & + galex141800 + 522514 & 214.50369 & 52.420519 & 22.59 & 25.99 & 0.838 & 42.89 & 60@xmath925 & & & 214.50280 & 52.419863 & 3.1 & 0.833 & + galex141722 + 524209 & 214.34177 & 52.702458 & 22.26 & 26.40 & 0.838 & 43.07 & 65@xmath925 & & & 214.34238 & 52.702278 & 1.5 & 0.838 & + galex142202 + 522931 & 215.50918 & 52.492172 & 24.88 & 28.45 & 0.812 & 42.66 & 307@xmath9252 & & & 215.51004 & 52.493038 & 3.6 & 0.807 & + galex142223 + 525642 & 215.59777 & 52.945235 & 23.55 & 26.86 & 0.770 & 42.90 & 187@xmath9221 & & & 215.59738 & 52.944958 & 1.3 & 0.763 & + galex141858 + 523209 & 214.74578 & 52.535948 & 21.76 & 22.67 & 0.762 & 43.11 & 57@xmath924 & -14.4 & & 214.74509 & 52.536110 & 1.6 & 0.758 & agn + galex142049 + 523203 & 215.20764 & 52.534438 & 23.62 & 28.61 & 0.754 & 43.03 & 273@xmath9221 & & & 215.20830 & 52.534378 & 1.5 & 0.758 & + galex142212 + 522634 & 215.55384 & 52.443058 & 23.67 & 27.45 & 0.762 & 42.47 & 79@xmath9211 & & & 215.55397 & 52.443161 & 0.5 & 0.759 & + galex142057 + 525642 & 215.24117 & 52.945171 & 22.54 & 24.98 & 0.756 & 43.00 & 101@xmath926 & & & 215.24106 & 52.944958 & 0.8 & 0.747 & + galex141746 + 525646 & 214.44476 & 52.946049 & 23.28 & 25.21 & 0.750 & 42.97 & 191@xmath9222 & & & 214.44504 & 52.946918 & 3.2 & 0.746 & agn + galex141842 + 522256 & 214.67524 & 52.382180 & 23.65 & 26.29 & 0.750 & 42.49 & 85@xmath9216 & & & 214.67545 & 52.381447 & 2.7 & 0.747 & + galex141842 + 530140 & 214.67697 & 53.027806 & 23.03 & 25.01 & 0.748 & 42.59 & 62@xmath929 & & & 214.67871 & 53.028191 & 4.0 & 0.741 & + galex142139 + 523401 & 215.41400 & 52.567205 & 23.82 & 26.24 & 0.741 & & & & agn & 215.41359 & 52.568192 & 3.8 & ( 0.692 ) & + [ groth ] ccccccccccccccc galex143557 + 350702 & 218.98764 & 35.117445 & 23.82 & -28.46 & 1.158 & 43.27 & 169@xmath9226 & & & 218.98735 & 35.116795 & 2.5 & 1.152 & + galex143539 + 351001 & 218.91250 & 35.167233 & 22.20 & 25.99 & 1.083 & 43.40 & 63@xmath926 & & & 218.91272 & 35.167511 & 1.2 & 1.082 & + galex143545 + 350039 & 218.93744 & 35.011330 & 24.17 & 26.63 & 1.063 & & & & agn & 218.93967 & 35.010777 & 5.9 & ( 0.722 ) & + galex143520 + 350414 & 218.83421 & 35.070876 & 21.55 & 22.69 & 1.055 & 43.32 & 31@xmath923 & -14.5 & & 218.83405 & 35.070351 & 2.0 & 1.052 & agn + galex143824 + 352325 & 219.60174 & 35.390787 & 21.92 & 24.92 & 1.039 & 42.92 & 18@xmath923 & & agn & 219.60257 & 35.390755 & 2.4 & 1.031 & + galex143512 + 352338 & 218.80005 & 35.394162 & 24.84 & -31.03 & 1.037 & & & & & 218.80109 & 35.392956 & 4.2 & ( 1.080 ) & + galex143726 + 351448 & 219.36103 & 35.247140 & 22.44 & 24.39 & 1.034 & 43.13 & 47@xmath926 & -15.1 & & & & & & + galex143443 + 353037 & 218.68009 & 35.510756 & 23.39 & 28.50 & 1.034 & 42.85 & 61@xmath928 & & & 218.68036 & 35.510929 & 1.0 & 1.025 & + galex143750 + 350645 & 219.45954 & 35.112962 & 22.22 & 23.70 & 1.010 & & & -15.0 & & 219.46024 & 35.113098 & 1.9 & ( 0.577 ) & + galex143449 + 350248 & 218.70428 & 35.046988 & 23.21 & 29.53 & 1.009 & 43.60 & 303@xmath929 & -15.3 & & & & & & + galex143810 + 350424 & 219.54441 & 35.073741 & 23.73 & 26.81 & 1.006 & 43.15 & 177@xmath9214 & & & 219.54455 & 35.073328 & 0.7 & no z & + galex143450 + 352520 & 218.70844 & 35.422650 & 22.85 & 26.74 & 0.967 & 43.29 & 122@xmath927 & -14.2 & & 218.70876 & 35.422508 & 1.1 & 0.964 & agn + galex143512 + 345907 & 218.79986 & 34.985750 & 24.58 & 28.79 & 0.963 & & & & & 218.80116 & 34.984894 & 3.8 & ( 0.884 ) & + galex143510 + 351602 & 218.79406 & 35.267619 & 23.42 & 26.28 & 0.941 & & & & & 218.79521 & 35.267330 & 2.5 & ( 0.478 ) & + galex143639 + 350557 & 219.16252 & 35.099672 & 22.31 & 23.43 & 0.937 & 42.87 & 32@xmath923 & & & 219.16257 & 35.099346 & 1.2 & 0.932 & + galex143702 + 350155 & 219.26169 & 35.032282 & 22.94 & 24.99 & 0.933 & 43.34 & 149@xmath9210 & -14.6 & & 219.26167 & 35.032078 & 0.7 & 0.927 & agn + galex143531 + 351415 & 218.88092 & 35.237743 & 23.09 & 27.50 & 0.934 & 42.96 & 78@xmath9211 & & & 218.88094 & 35.238354 & 2.2 & 0.930 & + galex143730 + 352016 & 219.37524 & 35.338103 & 22.51 & 24.34 & 0.907 & 42.90 & 46@xmath925 & & & 219.37563 & 35.337757 & 1.7 & 0.904 & agn + galex143521 + 350509 & 218.83940 & 35.086214 & 21.89 & 26.57 & 0.901 & & & & & 218.83961 & 35.085414 & 2.1 & star & star + galex143819 + 351542 & 219.58154 & 35.262070 & 23.01 & 26.70 & 0.889 & 42.63 & 39@xmath929 & & & 219.58224 & 35.261349 & 3.3 & 0.889 & + galex143652 + 350537 & 219.21699 & 35.093944 & 21.90 & 23.03 & 0.870 & 43.25 & 63@xmath925 & -15.0 & & 219.21677 & 35.093887 & 0.7 & 0.866 & agn + galex143729 + 350732 & 219.37419 & 35.125830 & 22.68 & 24.16 & 0.868 & 43.31 & 151@xmath9211 & -15.0 & & 219.37398 & 35.126053 & 1.0 & 0.860 & agn + galex143457 + 353213 & 218.73762 & 35.537287 & 21.75 & 23.02 & 0.865 & 42.82 & 21@xmath923 & -14.9 & & 218.73782 & 35.537319 & 0.6 & 0.857 & agn + galex143715 + 353355 & 219.31319 & 35.565674 & 22.38 & 23.84 & 0.854 & 43.27 & 127@xmath9211 & -15.1 & & 219.31288 & 35.565460 & 1.2 & 0.846 & agn + galex143459 + 352504 & 218.74643 & 35.418085 & 22.48 & -28.10 & 0.848 & & & & & 218.74582 & 35.417044 & 4.0 & star & star + galex143800 + 352206 & 219.50182 & 35.368824 & 22.24 & 24.81 & 0.829 & 42.81 & 35@xmath925 & -15.4 & & & & & & + galex143642 + 345027 & 219.17868 & 34.841281 & 24.09 & 27.73 & 0.823 & 42.91 & 254@xmath9249 & & & 219.17896 & 34.841167 & 0.9 & 0.818 & + galex143655 + 344610 & 219.23087 & 34.769857 & 24.20 & 28.61 & 0.820 & 42.88 & 269@xmath9230 & & & 219.23146 & 34.769501 & 2.2 & 0.813 & + galex143801 + 352533 & 219.50452 & 35.426310 & 21.62 & 22.75 & 0.806 & 43.09 & 42@xmath923 & -14.7 & & & & & & + galex143700 + 353132 & 219.25323 & 35.525830 & 22.50 & 24.40 & 0.791 & 42.96 & 75@xmath927 & & & 219.25333 & 35.525555 & 1.0 & 0.784 & agn + galex143515 + 353203 & 218.81441 & 35.534506 & 21.77 & 23.47 & 0.789 & & & & & 218.81459 & 35.534233 & 0.5 & ( 0.410 ) & + galex143821 + 350707 & 219.59055 & 35.118996 & 21.62 & 23.15 & 0.741 & 43.23 & 75@xmath925 & -14.8 & agn & 219.59068 & 35.119396 & 1.5 & 0.728 & + galex143829 + 352308 & 219.62144 & 35.385905 & 22.19 & 24.26 & 0.720 & 42.49 & 25@xmath924 & & & 219.62173 & 35.385715 & 1.1 & 0.710 & + galex143716 + 352324 & 219.32011 & 35.390438 & 24.61 & 27.67 & 0.716 & 42.61 & 365@xmath9265 & & & 219.31993 & 35.391083 & 2.4 & 0.699 & + galex143813 + 351121 & 219.55694 & 35.189666 & 24.20 & 27.14 & 0.710 & 42.67 & 199@xmath9253 & & & 219.55760 & 35.190166 & 2.6 & 0.717 & + [ ngpdws ] ccccccccccccccc galex095954 + 021707 & 149.97779 & 2.2852787 & 22.17 & 23.93 & 1.163 & 43.24 & 30@xmath924 & -14.7 & & & & & & + galex100145 + 023237 & 150.43904 & 2.5436927 & 23.59 & 24.84 & 1.154 & 43.35 & 158@xmath9213 & & & 150.43942 & 2.5435109 & 1.5 & 1.149 & agn + galex100016 + 015104 & 150.06776 & 1.8512265 & 21.74 & 23.40 & 1.140 & 43.63 & 57@xmath924 & -13.6 & agn & 150.06784 & 1.8513390 & 0.5 & 1.134 & agn + galex100031 + 023012 & 150.13260 & 2.5034525 & 22.72 & 25.38 & 1.103 & 43.21 & 59@xmath926 & & & 150.13235 & 2.5033109 & 1.0 & 1.099 & + galex095918 + 014933 & 149.82753 & 1.8260336 & 22.89 & 25.11 & 1.070 & 43.38 & 109@xmath9210 & & & 149.82774 & 1.8258640 & 1.0 & 1.069 & + galex095916 + 015048 & 149.81704 & 1.8467845 & 21.94 & 24.30 & 1.034 & 43.20 & 33@xmath924 & -14.0 & agn & 149.81688 & 1.8467280 & 0.6 & 1.034 & agn + galex100141 + 021029 & 150.42197 & 2.1748164 & 21.83 & 23.00 & 0.989 & 43.47 & 65@xmath925 & -13.5 & agn & 150.42221 & 2.1754150 & 2.3 & 0.982 & agn + galex100150 + 020936 & 150.45907 & 2.1598448 & 22.75 & 25.61 & 0.927 & 43.11 & 79@xmath927 & & & 150.45946 & 2.1603360 & 2.3 & 0.926 & agn + galex100124 + 021447 & 150.34970 & 2.2463910 & 22.34 & 23.46 & 0.900 & 43.23 & 77@xmath928 & -13.8 & & 150.34988 & 2.2461319 & 1.1 & 0.894 & + galex100202 + 020145 & 150.51078 & 2.0290676 & 22.95 & 24.72 & 0.894 & 43.20 & 124@xmath929 & -13.6 & & 150.51064 & 2.0292540 & 0.8 & 0.898 & agn + galex100049 + 021707 & 150.20630 & 2.2852412 & 23.88 & 24.96 & 0.877 & 42.96 & 181@xmath9230 & -13.7 & & 150.20627 & 2.2857921 & 2.0 & 0.874 & agn + galex100136 + 020653 & 150.40266 & 2.1148013 & 22.28 & 25.06 & 0.853 & 42.92 & 42@xmath925 & & & 150.40346 & 2.1155031 & 3.8 & 0.849 & + galex100133 + 015451 & 150.39033 & 1.9141639 & 21.85 & 24.96 & 0.848 & 43.02 & 37@xmath923 & & & 150.39020 & 1.9144530 & 1.1 & 0.844 & + galex100002 + 021628 & 150.00988 & 2.2746592 & 21.90 & 23.78 & 0.847 & 42.99 & 34@xmath924 & -13.5 & & 150.00922 & 2.2755051 & 3.9 & 0.850 & agn + galex100207 + 021119 & 150.53190 & 2.1886307 & 22.31 & 22.60 & 0.838 & 43.25 & 101@xmath9219 & -14.0 & & 150.53186 & 2.1889460 & 1.1 & 0.830 & + galex100113 + 022548 & 150.30739 & 2.4301342 & 22.76 & 23.40 & 0.757 & & & -13.6 & & 150.30808 & 2.4300640 & 2.5 & ( 0.374 ) & agn + galex100029 + 022129 & 150.12442 & 2.3581919 & 22.26 & 23.42 & 0.739 & 42.93 & 66@xmath929 & -13.6 & agn & 150.12370 & 2.3582580 & 2.6 & 0.728 & agn + galex095910 + 020732 & 149.79323 & 2.1258027 & 22.12 & 22.22 & 0.730 & & & -14.8 & & 149.79295 & 2.1256490 & 1.1 & ( 0.353 ) & + galex100017 + 020013 & 150.07274 & 2.0037061 & 22.69 & 23.04 & 0.729 & & & -13.6 & & 150.07301 & 2.0035551 & 1.1 & ( 0.350 ) & agn + [ cosmos ] | we construct a flux - limited sample of 135 candidate @xmath0 ly@xmath1 emitters ( laes ) from _ galaxy evolution explorer _ ( _ galex _ ) grism data using a new data cube search method .
these laes have luminosities comparable to those at high redshifts and lie within a 7 gyr gap present in existing lae samples .
we use archival and newly obtained optical spectra to verify the uv redshifts of these laes .
we use the combination of the _ galex _ uv spectra , optical spectra , and x - ray imaging data to estimate the active galactic nucleus ( agn ) fraction and its dependence on ly@xmath1 luminosity .
we remove the agns and compute the luminosity function ( lf ) from 60 @xmath0 lae galaxies .
we find that the best fit lf implies a luminosity density increase by a factor of @xmath21.5 from @xmath3 to @xmath0 and @xmath4 from @xmath0 to @xmath5 .
we find a @xmath0 volumetric ly@xmath1 escape fraction of @xmath6 . |
there have been wide interests in the study of the topological objects such as the vortices @xcite , monopoles @xcite , and instantons @xcite because of their importance for understanding some crucial aspects of the nonlinear field theories @xcite .
they are not included in the original action manifestly , but appear as the localized finite energy solutions of the nonlinear field equations .
these solutions can be classified with the topological number which comes from the topological property of the system .
since the topological objects have spatially localized , finite energy configurations and some topologically conserved charge , it is fascinating to regard them as dynamical degrees of freedom in the system .
one of such objects is the abrikosov vortex @xcite in ginzburg - landau ( gl ) theory , which appears as the magnetic flux squeezed by the cooper - pair condensate .
when the external magnetic field applied to the superconductor exceeds some critical strength , the cooper - pair is dissociated and the node of the cooper - pair forms a vortex line .
the single - valued property of the cooper - pair around vortex lines leads the quantization of the total magnetic flux , as @xmath4 , where @xmath5 represent the effective gauge coupling of the cooper - pair , the total magnetic flux and the topological number , respectively .
the topological number @xmath1 characterizes the topology of the system and we can classify vortices in terms of @xmath1 .
there exist many applications employing the concept of the abrikosov vortex and its relativistic version , the nielsen - olesen vortex @xcite .
one example is the squeezed color - electric flux between the quark and anti - quark in qcd
. nambu @xcite , t hooft @xcite and mandelstam @xcite discussed that there exist the vortices which appear as the color - electric flux tubes squeezed by the condensation of magnetic monopoles .
as a result of the color - electric flux squeezing , we can only observe colorless mesons and baryons , which are color singlet bound states of the colored quarks with the bond of the color - electric flux .
this scenario is a possible explanation for `` color confinement '' , which is the experimental fact in hadron physics @xcite .
the lattice qcd monte carlo calculations @xcite for @xmath6 @xcite and 3@xmath7 @xcite potentials strongly support the flux tube picture , and show the universality of the string tension of the flux - tube , @xmath8 0.89 gev , in both @xmath6 and 3@xmath7 cases . an another example is the cosmic string in astrophysics as a seed of the galaxy formation and baryogensis @xcite .
kibble @xcite and zurek @xcite discussed the distribution of the topological defects and their cosmological evolution after the quench of the early universe and the resulting phase transition .
they argued that the cosmic strings as topological defects would produce a characteristic signature in the cosmic microwave background .
following their scenarios , the phase transition dynamics are closely investigated in the numerical simulations using the phenomenological time - dependent gl ( tdgl ) theory @xcite in both cases with global symmetry and local gauge symmetry . in particular , in the case of the gauge symmetry , it is discussed that the initial thermal magnetic fluctuations just after the quench play important roles on the later distribution of the vortices such as the clusters of vortices @xcite .
these investigations are closely correlated with the condensed matter physics in a viewpoint to understand the phase transition dynamics , and provide the interesting subjects to link different fields of research .
now we turn to the dynamical aspects of single vortex as a `` pseudo - particle '' . for the treatment of the vortex motion ,
the vortex is usually regarded as one dimensional ( 1-d ) thin object like a string , neglecting the extension perpendicular to the vortex line .
such a treatment is expected to be sufficient when the length of the string is much larger than the length scale in the radial direction .
this condition comes from the fact that the excitation energy of string as a 1-d object is typically @xmath9 ( @xmath10 is the string length ) .
then , for the vortex with large length , the stringy excitation is more important than radial one with changing its thickness .
even if the above condition is not well - satisfied , we have only to consider extensive modes of the magnetic flux when the type - ii nature is very strong , i.e. , the ginzburg - landau parameter @xmath11 is very large ( here @xmath12 is the penetration depth of the magnetic field and @xmath13 is the coherence length of the higgs field ) .
this is because when type - ii nature is very strong , the topological defect of the higgs field along the vortex line is strongly squeezed and its extension modes perpendicular to the vortex line cost a large energy .
then , in the type - ii limit of the vortex , the main excitation is given by the stringy modes , and the subject is linking to one of the most fundamental problems in the quantum string theory@xcite , which gives not only a candidate of the grand - unification including gravity but also a new method to analyze nonperturbative qcd @xcite .
so far , there are so many interesting studies done on the quantum string . for instance
, the quantum fluctuations of the string lead to a roughening " , i.e. , a widening of the string @xcite , and several studies of quantum nielsen - olesen strings @xcite showed that quantum vibrational modes make the vortex world - sheets crumpled and dominated by branched - polymers " as a disease of quantum strings@xcite .
however , we sometimes encounter the cases where the vortex can not be regarded as a stringy object .
we are interested in such cases .
one example is the type - ii superconductors in condensed matter physics near @xmath2=1/2 , i.e. , bogomolnyi - prasad - sommerfeld ( bps ) @xcite value or critical coupling . in such cases , the squeezing of the magnetic flux due to
cooper - pair condensation is not very strong and both the cooper - pair and photon fluctuations can play important roles .
moreover , their interplay may provide interesting phenomena .
an another interesting example is the quark systems such as the meson and baryon which have the color - electric flux tube with the total length about 1 fm , which is not very large in comparison with the flux tube extension of about 0.4 fm .
in addition , according to the analysis using the dual - ginzburg - landau model @xcite , type - ii nature also seems to be not so strong , and then the radial motion of the vortex can play an important role .
studies on such excitation modes are important to understand the structure of hadrons . for the gluonic excitations , the lattice qcd studies for heavy @xmath6 @xcite and 3@xmath7 @xcite system both provide the 1st gluonic excitation energy as @xmath14 1 gev in the the total length of the flux , @xmath15=0.5 - 1.0 fm , as the typical hadronic scale . small dependence on @xmath15
provides the possibility to understand the 1st gluonic excitation as not simple stringy vibration but the radial excitation with the fixed boundary effects on the flux - tube , because the simple stringy mode is likely to behave proportionally to 1/@xmath15 .
the studies on the energy of the radial excitation with comparing it to that of lattice results will give a qualitative explanation for the gluonic excitations in the case of relatively small but not too short @xmath15 .
there seem to exist several other cases where the radial excitations play important roles .
, width=321 ] putting applications to the above examples in perspective , we will discuss the axial - symmetric radial excitation modes and their propagation along the vortex line .
hereafter , we will call this excitation mode _
peristaltic _ mode .
the schematic picture of the peristaltic mode is shown in fig.[fig : peristic ] . to discuss the peristaltic modes
, we utilize the abelian higgs action which includes the higgs and the photon fields . in terms of this action ,
the peristaltic modes are characterized by the higgs and photon field fluctuations . in particular , the time - dependent photon fluctuations , which are sometimes neglected ,
play important roles on the corporative behavior with the higgs field fluctuations .
these fluctuations are regarded as the vibration of the vortex from the viewpoint that the vortex is the dynamical degrees of freedom in the system . in this work , since we concentrate on the radial excitation modes , we simply neglect the effects of the boundary of the vortex edge , considering the vortex with the infinite length .
the bending of the vortex is also neglected .
we investigate both type - i and type - ii vortices with not only usual @xmath1=1 but also @xmath16 for the future applications to the non - equilibrium or multi - vortices systems where the giant vortices can appear as excited states . in the type - i case ,
if the system contains only finite magnetic flux in contrast to the equilibrium case usually realized in the laboratory , these flux gather and become a @xmath17 2 giant vortex because type - i vortices attract each other . on the other hand , in the type - ii case , the giant vortex is unstable and splits into the small vortices . in spite of this , when @xmath2 is not too large , it is worth considering the vibration of the giant vortex because the gradient of the vortex - vortex interaction is small for the small separation of vortices @xcite and we can expect that the fission process develops slowly .
the organization of this paper is as follows . in sec.ii
we briefly review the treatments of the static profile of single vortex .
in sec.iii we derive the effective action and the resulting equation of motion for the fluctuations around the vortex .
the effects of the static vortex background on the fluctuations are emphasized . in sec.iv
we show the numerical results for the excitation modes of the @xmath1=1 vortex with 0.1@xmath18@[email protected] .
based on the classification with @xmath2 , we closely discuss their excitation energy , correlation energy between the higgs and photon fields , spatial distribution of fluctuations , and the resulting magnetic flux behavior .
it will be shown that , in the case of @xmath20 1/2 , the interplay of the higgs and the photon fields plays an essential role for the existence of the new - type discrete pole .
we also discuss the excitation modes around @xmath21 2 vortices . in sec.v
we summarize this paper with the perspective for the future direction .
in this section , we briefly review the treatments of the static vortex in terms of abelian higgs model @xcite .
this type of action is very popular and has been used to describe cosmic string in the astrophysics @xcite , the color flux tube in qcd @xcite , and so on .
we use this lorentz invariant action to investigate the properties of the vortex as a `` pseudo - particle '' . for the finite temperature cases , one usually uses the phenomenological model with replacing the 2nd time - derivative in abelian higgs model with the 1st derivative although the microscopic foundation of this treatment is still not conclusive .
the noise term is sometimes further added .
these replacements change the equation of motion from the wave equation to the diffusive equation which respects the dissipation effects due to the thermal fluctuations .
we will discuss the finite temperature case elsewhere .
let us consider the ideal system which contains only one vortex with infinite length along the @xmath22-axis .
our purpose is to investigate the excitation mode of the vortex in such a system .
we start with the following action with natural units ( @xmath23 ) @xmath24 where @xmath25 and @xmath26 represent the higgs field and photon field respectively . here
@xmath27 are the effective gauge coupling constant , the strength of self - coupling , and the vacuum expectation value of the higgs field , respectively .
@xmath28 denote space - time coordinates and @xmath29 . through this paper , we describe the variables in physical unit with bar . for later convenience
, we rescale the fields and the coordinates to make them dimensionless quantities , @xmath30 we also rewrite the action @xmath31 the energy relation between the physical and our rescaled unit can be obtained with noting the following relation , @xmath32 here @xmath33 and @xmath34 are lagrangian in physical and rescaled unit , respectively , and they are related as @xmath35 .
similarly , we can obtain the energy and momentum in the physical unit with multiplying @xmath36 to our rescaled energy and momenta .
after these rewriting , the action reads @xmath37 here we have introduced the gl parameter @xmath38 , which is used to distinguish the type - i ( @xmath39 ) and type - ii ( @xmath40 ) superconductors . the critical case of @xmath41
is called as the bps case , which lies between type - i and type - ii . as a remarkable fact in the bps case ,
the photon mass @xmath42 and the higgs mass @xmath43 coincide as @xmath44 , because of @xmath45 and @xmath46 , i.e. , @xmath47 , which can be derived from the original action ( [ action1 ] ) . in this work
, we will consider the system which is translation invariant in the @xmath22-direction and contains single vortex with infinite length along the @xmath22-axis .
we also impose the axial - symmetry around the @xmath22-axis on the static solution . with cylindrical coordinates ( @xmath48 )
, we denote the complex field @xmath49 in the polar - decomposed form @xmath50 where @xmath1 is the topological number ( @xmath1=0 , @xmath511 , ) related to the phase of the higgs field wave function around the @xmath22-axis . for single vortex with the topological number @xmath52 , the sign of @xmath1 is not relevant because of charge conjugation symmetry of the abelian higgs model , and therefore we have only to consider @xmath1=0 , 1 ,
cases without loss of generality .
since the abelian higgs model has the u(1 ) local gauge symmetry , we have the degrees of freedom to choose the gauge . for the treatment of the vortex solution , the familiar gauge is the @xmath53 gauge . instead , we adopt the gauge which eliminates the small amplitude of the phase , @xmath54 to simplify the equations of fluctuations in the later analysis . for @xmath1=0 ,
this gauge fixing corresponds to the unitary gauge . from the rotational symmetry and translational invariance along the @xmath22-direction
, we adopt the following ansatz for the static solutions @xmath55 and the static electric and magnetic fields are obtained as @xmath56 as seen soon later in eq.([sta1 ] ) and ( [ sta2 ] ) , our gauge fixing and ansatz for the static fields reproduce the same results for the static field solutions as those in usual @xmath57 gauge . the total static energy of the system ( in the rescaled unit ) is obtained as @xmath58 with putting the time derivative terms zero , @xmath59,\end{aligned}\ ] ] where @xmath60 = & & \!\!\ ! ( \partial_z \phi)^2 + ( \partial_r \phi)^2 \nonumber \\ & & \!\!\ ! + \frac{1}{2}(\partial_z
a_\theta ) ^2
+ \frac{1}{2}(\partial_r a_\theta + \frac { a_\theta } { r } ) ^2 \nonumber \\ & & \!\!\ !
+ \frac { ( n- r a_{\theta } ) ^2 } { r^2 } \phi^2 + \kappa^2 ( \phi^2 -1)^2.\end{aligned}\ ] ] ( the full action and the energy functional are given in appendix [ appendixa ] . ) under the above circumstances , we consider the static vortex solutions @xmath61 and @xmath62 which minimize the static energy . from the variational principle @xmath63 and then the static field equations read @xmath64 since we are interested in the vortex solution , we consider the boundary condition ( b.c . ) which is appropriate to describe the vortex at the center . for @xmath65 , @xmath66 and for @xmath67 , @xmath68 these boundary conditions are obtained by the asymptotic analysis of the static field equations which satisfy the following physical requirements to describe the vortex at the center : ( i ) at @xmath69 the vacuum expectation value of higgs field takes zero .
( ii ) the static field energy @xmath70 asymptotically approaches to the vacuum energy when @xmath71 .
note that the first condition represents the existence of the vortex at @xmath69 and the second one represents the magnetic fields are completely screened in the region far from the center of the vortex .
the boundary conditions ( [ bc1 ] ) and ( [ bc2 ] ) also lead the quantization of the total magnetic flux @xmath72 , @xmath73 ( @xmath72 corresponds to @xmath74 in the physical unit . ) we can numerically solve the static field equations ( [ sta1 ] ) and ( [ sta2 ] ) with the boundary conditions ( [ bc1 ] ) and ( [ bc2 ] ) and obtain the static solutions @xmath75 and @xmath76 for the higgs field @xmath61 and the photon field @xmath77 .
fig.[profile1 ] shows the static profiles of the higgs field @xmath78 , the magnetic field @xmath79 , and the resulting static energy distribution @xmath80 in the vortex with various @xmath0 . using the distribution of the static field energy
, we can calculate the vortex mass ( with quantum number @xmath1 ) per unit length in the @xmath22-direction , i.e. , @xmath81.\end{aligned}\ ] ] note that by definition of the action , the origin of energy is zero when the system has no vortex .
the vortex mass is classified in the following way , @xmath82 these relations are analytically @xcite and numerically @xcite investigated .
fig.[fig : masssmall ] and [ fig : vortexmass ] showed the numerically solved @xmath2-dependence of @xmath83 , the vortex mass divided by the topological number @xmath1 , and we can see @xmath84 these relations indicate that the interaction between vortices is attractive for @xmath851/2 , and repulsive for @xmath861/2 .
these are confirmed by the studies of the vortex - vortex interaction analytically in large separation case @xcite and numerically in arbitrary distance case @xcite . the qualitative understanding for the potential property is as follows .
the attractive part of the potential comes from the reduction of the topological defect of the higgs field condensate .
this is because the total energy of the system decreases when the domain of the normal conducting state decreases . on the other hand ,
the repulsive part comes from the magnetic field interaction of the vortices .
then , in the case of @xmath851/2 where the coherence length of the higgs field exceeds the penetration depth of the magnetic field , the attractive part becomes larger than the repulsive part in the whole region , and the vortex - vortex potential becomes attractive . on the other hand , in the case of @xmath861/2 , the relation is reversed and the vortex - vortex potential becomes repulsive . in the case of the vortex - anti - vortex case ,
the potential becomes attractive independently from @xmath2 since the magnetic field interaction between the vortex and the anti - vortex becomes attractive in contrast to the vortex - vortex case , and there is no repulsive part in the potential .
, width=264 ] , width=264 ] although we consider the isolated vortex under the idealized situation from the beginning , it is worth mentioning under which situations single vortex appears referring to the potential properties explained above . in type - i superconductor , the multi - vortices with total topological number @xmath87 transfer to the single , giant vortex with topological number @xmath87 due to their attractive interaction . after the successive fusion processes , the multi - vortices system transfers to the normal conducting state .
for the finite size superconducting materials under the homogeneous magnetic fields and thermal equilibrium , such fusion processes lead the difficulty to observe the type - i vortices because the magnetic fluxes successively invade from the interface between the normal and superconducting phase .
nevertheless , under the non - equilibrium situations or inhomogeneous magnetic fields concentrated on the small area , it is expected that the type - i vortex or , more generally , the domain of the normal state in the superconductor can appear as a transient state in the process such as the dynamical evolution of the thermodynamic unstable state into the equilibrium state , i.e. , the nucleation or spinodal decomposition processes @xcite . from the viewpoint of the particle physics ,
an another interesting situation where type - i vortex appears is the case where magnetic monopoles are immersed in superconductors and the isolated vortex emerges between the monopole and anti - monopole @xcite . in this case
, no collapse with the other vortices occurs and type - i vortex behaves as the stable object due to the topological charge conservation . on the other hand , in type - ii superconductor , the giant vortex with topological charge @xmath87 fissions to the @xmath87 vortices with charge @xmath1=1 due to their repulsive nature @xcite .
this repulsive interaction prevents the vortices to coalescence and , under the thermal equilibrium , the @xmath1=1 vortices form some stable structure like the abrikosov triangle lattice @xcite . then , in type - ii superconductor , the @xmath1=1 case is realistic in the thermodynamic equilibrium in contrast to type - i case .
however , when the system experience a certain quenching @xcite such as the rapid change of the temperature or density , the @xmath88 vortex may be generated as the excited states .
this is the reason why we also consider the type - ii giant vortex in the following .
although they are unstable against the fission into small vortices , when @xmath2 is not too large , it may be interesting to consider the vibrations of the giant vortex because such vibrations can occur during the slow development of the fission process due to the small gradient of the vortex - vortex potential for the small separation of vortices @xcite .
moreover , the study of the fluctuation modes of the @xmath88 vortex may give some insight for the early stage of the evolution of the giant vortex into the multi - vortices .
in this section , using the results of the previous section , we analyze axial - symmetric fluctuations in the system which has only one vortex in the background . in subsection
a , using the static vortex profile , we derive the _ vortex - induced potential _ for the higgs and photon field fluctuations , and the equation of motion . in subsection b
, we closely discuss the potential feature to classify the type of the excitation modes .
we investigate the axial - symmetric fluctuations of the higgs field and the related photon fields around the static vortex .
we expand the fields with small amplitude fluctuations around the static field , @xmath89 here we only consider the axial - symmetric fluctuations and we include @xmath90-dependence at this stage .
it is important to notice that , with the ansatz ( [ ansatz ] ) for the static solution , all the other type of fluctuations such as fluctuations of @xmath91 and the @xmath92-dependent fluctuations are _ decoupled _ from the axial - symmetric fluctuations which we consider .
hence we can analyze the axial - symmetric fluctuations ( [ static ] ) independently from the other type of fluctuations and we drop off such fluctuation terms from the following lagrangian .
the complete form of the lagrangian is given in the appendix [ appendixa ] .
substituting eq.([static ] ) into eq.([action2 ] ) and using the static solution , the following expression for the axial - fluctuation part of lagrangian can be derived as @xmath93 = { \cal l } _ { { \rm axial } } ^ { { \rm cl } } + { \cal
l } _ { { \rm axial } } ^{(2 ) } + { \cal l } _ { { \rm axial } } ^{(3 ) } + { \cal
l } _ { { \rm axial } } ^{(4)}.\end{aligned}\ ] ] the first term @xmath94 represents @xmath95 $ ] .
the 1st order terms of fluctuations @xmath96 and @xmath97 do not appear because we expand around the static solution .
the 2nd , 3rd , and 4th order fluctuation terms take the following forms ; @xmath98 \varphi^2 a_\theta -2 \phi _ { { \rm cl } } \varphi a_\theta^2 , \\
{ \cal l } _ { { \rm axial } } ^{(4 ) } \!\!\ ! & = & \!\!\ !
- \kappa^2 \varphi^4 - \varphi^2 a_\theta^2 .
\end{aligned}\ ] ] here @xmath99 , @xmath100 , @xmath101 denote the _ vortex - induced potential _
, i.e. , these interactions are induced by the static vortex configuration which can be regarded as the large amplitude background external field . @xmath99 and @xmath100 play roles as the _ single particle potential _ for @xmath96 and @xmath97 respectively , and @xmath101 is the interaction which mixes the @xmath96 and @xmath97 .
hereafter , we will simply call these interactions `` potentials '' .
the concrete forms of the potentials are @xmath102 ^ 2 + 2 \kappa^2 ( 3 \phi _ { { \rm cl } } ^2 - 1 ) , \nonumber \\ & & v_{\varphi a}(r ) = - 4 \bigg [ \frac{n}{r } - a _ { { \rm cl } } \bigg ] \phi _ { { \rm cl } } , \nonumber \\ & & v_{aa}(r ) = \phi _ { { \rm cl } } ^2 . \ ] ] to investigate the fluctuation spectrum , we solve the euler - lagrange equation for @xmath103 neglecting the higher order fluctuations .
the quantum loop corrections can be calculated using the green s functions constructed of bases which diagonalize @xmath103 . however , because of the inhomogeneous vortex - background , the green s functions are not functions of the distance between arbitrary two points , i.e. , @xmath104 , and the calculations of the loop corrections are not straightforward . in addition , higher order terms of the axial - symmetric fluctuations couple with the terms of the non - axial symmetric ones .
we reserve the calculation of the loop corrections for the future work and , in this work , we safely neglect the quantum corrections constraining ourselves to the case where the perturbative expansion parameters @xmath105 and @xmath106 are sufficiently small .
( recall that the perturbative corrections are calculated with expansion of @xmath107 .
see also eq.([reaction ] ) . )
these constraints are not so strict because the higgs - photon coupling @xmath108 is usually small and we mainly consider @xmath109 cases , where @xmath106 is also sufficiently small . for later convenience to discuss the effect of the potential , we change the variables @xmath110 these changes of variables enable us to considerably simplify the equation of motion as will be seen .
the euler - lagrange equation for 2nd order lagrangian is obtained through the variation , @xmath111 then we obtain the following coupled equation @xmath112 & = & \left [ \begin{array}{cc } - \partial_r^2 + \frac { 4 ( n - a _ { { \rm cl } } ) ^2 -1 } { 4 r^2 } + 2 \kappa^2 ( 3\phi _ { { \rm cl } } ^2 -1 ) & -\frac { 2 \sqrt{2 } \phi _ { { \rm cl } } ( n - a _ { { \rm cl } } ) } { r } \\ -\frac { 2 \sqrt{2 } \phi _ { { \rm cl } } ( n - a _ { { \rm cl } } ) } { r } & - \partial_r^2 + \frac { 3 } { 4 r^2 } + 2 \phi _ { { \rm cl } } ^2 \end{array } \right ] \left [ \begin{array}{c } \tilde{\varphi } ( t , z , r ) \\ \tilde{a}_\theta ( t , z , r ) \end{array } \right ] \\
& \equiv & \left [ \begin{array}{cc } - \partial_r^2 + v_{\alpha \alpha}(r ) & v_{\alpha \beta}(r ) \\
v_{\beta \alpha}(r ) & - \partial_r^2 + v_{\beta \beta}(r ) \end{array } \right ] \left [ \begin{array}{c } \tilde{\varphi } ( t , z , r ) \\
\tilde{a}_\theta ( t , z , r ) \end{array } \right ] .\end{aligned}\ ] ] here @xmath113 ( this is the consequence of the change of the variables in eq.([variable ] ) ) .
note that the @xmath90 dependence of @xmath114 and @xmath115 is trivial because the background static vortex configuration is @xmath90-independent and @xmath90 dependence does not appear in the potential .
then the partial differential equation can be decoupled into the total differential equations for @xmath116 , respectively .
the system has the translation invariance in both @xmath22-direction and time direction , and then the solution takes the following plane wave solution , @xmath117 then the total differential equation in the radial direction for @xmath118 and @xmath119 is obtained as @xmath120 = \left [ \begin{array}{cc } - \frac{d^2}{dr^2 } + v_{\alpha \alpha}(r ) & v_{\alpha \beta}(r ) \\
v_{\beta \alpha}(r ) & - \frac{d^2}{dr^2 } + v_{\beta \beta}(r ) \end{array } \right ] \left [ \begin{array}{c } \alpha_j \\
\beta_j \end{array } \right ] , \nonumber \\ \label{psi1}\end{aligned}\ ] ] where @xmath121 .
note that the hermiticity of the matrix in the rhs in eq.([psi1 ] ) ensures the real value of @xmath122 and our later analysis shows @xmath123 , i.e. , the stability of the static profile against the fluctuations .
the root of the eigenvalue for the eq.([psi1 ] ) , @xmath124 , plays a role of the mass for the propagating mode in the @xmath22-direction .
the hermiticity of eq.([psi1 ] ) also enables us to take the functions @xmath118 and @xmath119 real , and the orthonormal condition for the @xmath118 and @xmath119 is expressed as @xmath125 \nonumber \\ & = & \delta_{jk}. \label{psi2}\end{aligned}\ ] ] here we give a brief explanation for the qualitative picture for the excitation modes .
the shape of the radial direction is given by @xmath126 and @xmath127 .
this wave function propagates in the @xmath22-direction with the frequency @xmath128 and the wave vector @xmath129
. therefore the root of the eigenvalue @xmath122 of the radial wave functions @xmath126 and @xmath127 play roles of the mass for the propagation in the @xmath22-direction .
then we call this type of new mode as the _ peristaltic _ mode of the vortex as illustrated in fig.[fig : peristic ] .
we show in fig.[potential1 ] the single particle potentials @xmath130 , @xmath131 and @xmath132 around the vortex for various values of @xmath0 .
we can deduce the considerable information of the excitation modes from the examination of these potentials . to get the qualitative picture about the wave functions of fluctuations , it is convenient to divide the region of the coordinate space @xmath133 into three regions , _ central _ ( @xmath134 ) , _ intermediate _ ( @xmath135 ) , and _ asymptotic _ ( @xmath136 ) region in the rescaled abelian higgs model . in the _ central _ region ,
when @xmath137 , the asymptotic behavior of the potential is @xmath138 from these asymptotic behavior , it is expected that the fluctuations @xmath118 and @xmath119 around the vortex center are suppressed by the strong repulsive interaction proportional @xmath139 and their distributions are shifted to the outside .
on the other hand , in the _ intermediate _ region , the strong repulsive part of @xmath130 and @xmath132 becomes weak and the attractive pocket emerges .
the mixing term @xmath131 also plays an important role to enhance the mixing between the higgs and photon field fluctuations , or @xmath140 and @xmath141 .
since the @xmath131 always takes the negative value , the excitation energy is reduced when the relative sign of the @xmath140 and @xmath141 fluctuations is positive . to avoid the confusion ,
we stress that the minus sign of @xmath131 does not mean attractive interaction .
even if @xmath131 takes the positive value , the energy is reduced when the relative sign of @xmath140 and @xmath141 fluctuations is negative .
this is also recognized from the fact that the replacement @xmath142 changes only the sign of the off - diagonal part of eq.([psi1 ] ) . in treatment
up to the second order of the fluctuations , only the absolute value of @xmath131 is relevant to the mixing of @xmath140 and @xmath141 .
all these features of the potentials @xmath143 make low energy modes of the fluctuations localized in the intermediate region .
finally we discuss the _ asymptotic _ region .
when @xmath144 , these potentials asymptotically behave as @xmath145 note that the asymptotic behaviors do not depend on the topological charge @xmath1 .
since @xmath131 goes to zero , the correlation between @xmath140 and @xmath141 vanishes and these fluctuations behave as independent modes .
then we have only to consider the one - particle problem in this region , i.e. , we treat the asymptotic forms of eq.([psi1 ] ) as @xmath146 the values @xmath147 and @xmath148 in the lhs play roles of the ( square of ) _ continuum thresholds _ of the @xmath140 and @xmath141 respectively .
these thresholds are equivalent to the masses of the fluctuations in the system with no vortex , i.e. , @xmath149 corresponds to the mass of the higgs field and @xmath150 is the photon mass generated through the anderson - higgs mechanism @xcite . then , the threshold energy for the continuum states is expressed as @xmath151 when @xmath122 is smaller ( larger ) than the thresholds , the fluctuations behave like bound ( continuum ) states .
it is also worth mentioning that @xmath130 and @xmath132 take the same value in the asymptotic region and the continuum thresholds are the same both for @xmath140 and @xmath141 for the case of @xmath2=1/2 , where bps saturation occurs .
this note is sometimes useful to roughly classify the type of the fluctuation spectrum in the low - energy region in terms of the type of superconductor , as is explained below .
it is expected that , for type - i case ( @xmath851/2 ) , the threshold for the higgs field is lower than that for the photon field , and the higgs field fluctuation dominates over the low energy physics . for type - ii case ( @xmath861/2 ) , the relation is reversed , the photon field fluctuation dominates .
we also expect that the collective properties induced by the mixing of the higgs and photon field fluctuations are most pronounced when the bps saturation occurs .
this is because the strength of both fluctuations are the same order and they affect each other . in the next section , we will see the above mechanism is crucial to emerge the characteristic discrete pole around @xmath152 .
in this section , we summarize the numerical results in both type - i and type - ii superconductors .
the following results are obtained by solving the coupled channel equations ( [ psi1 ] ) numerically with the lattice spacing 0.02 and the lattice size 40.0 in our rescaled unit .
we impose the boundary condition on the wave function @xmath118 and @xmath119 to be zero value at the box boundary . in the following ,
we consider @xmath153 case .
the energy and radial wave functions shown in this section are translated into those in the general @xmath154 case with the replacements @xmath155 and multiplying @xmath156 to the radial wave functions , respectively . in this subsection
, we show the numerical results for @xmath1=1 case in both type - i and type - ii superconductors . in fig.[fig : n1spec ] , the @xmath2-dependence of @xmath122 for @xmath1=1 case and the thresholds for @xmath140 and @xmath141 fields are plotted . shown in fig.[fig : n1specw ] are the same quantities as fig.[fig : n1spec ] , but without channel coupling , which is artificially dropped off . with smaller @xmath157 ,
the threshold for @xmath140 is shifted lower than threshold for @xmath141 , and @xmath140 can excite with smaller energy and the amplitude of @xmath140 is enhanced compared to that of @xmath141 field
. then the lower excitation modes are mainly characterized by the @xmath140 field fluctuation . under the normalization condition ( [ psi2 ] )
, there appears a large excess of the amplitude of @xmath140 field compared to that of @xmath141 field , and this mismatch between @xmath140 and @xmath141 leads the suppression of the mixing effect .
then the properties of the lower excitations asymptotically approach to those calculated without channel coupling .
the property of the discrete pole can be interpreted as the bound state of the static vortex and the fluctuating higgs field . on the other hand , with larger @xmath158 , the enhancement of the threshold energy for @xmath140 leads the suppression of the @xmath140 field excitation , and simultaneously enhances the @xmath141 field ratio in the lower excitation modes .
then the increasing of @xmath2 value leads the reduction of the mixing correlation energy which is indispensable for the appearance of the discrete pole .
eventually , at @xmath159 , the discrete pole is buried in the continuum state spectrum of the fluctuating photon field and no discrete pole appears for @xmath160 .
this indicates that such a pole is difficult to be identified in most type - ii superconductors .
finally we consider the most interesting case , @xmath161 case , i.e. , near bps saturation . in this case , the mixing effect becomes larger , and , as a result , the correlation between @xmath140 and @xmath141 considerably lowers the excitation energy .
this is clearly seen from the comparison of fig.[fig : n1spec ] with fig.[fig : n1specw ] . in table
[ tab : table1 ] , we summarize the results for the lowest excitation energy @xmath162 and their relations to the threshold energy @xmath163 , the lowest excitation energy without channel coupling , @xmath164 , and the energy differences among them . the quantity @xmath165 represents the binding energy and specifies the spatial extension of the wave functions .
the quantity @xmath166 is , roughly speaking , the correlation energy induced by the mixing effect between higgs and photon field fluctuations .
as expected , the maximum mixing correlation is realized in the bps case ( @xmath2=1/2 ) .
.[tab : table1 ] @xmath2 dependence of the various energy in the rescaled unit .
the physical value can be obtained by multiplying @xmath36 .
@xmath167 , @xmath168 , @xmath164 represent the threshold energy , the lowest excitation energy and the lowest excitation energy without the channel coupling , respectively .
the negative sign of @xmath169 indicates the bound state .
the quantity @xmath170 roughly characterizes the correlation energy of the higgs - photon field mixing .
[ cols="^,^,^,^,^,^,^,^",options="header " , ] the features explained above can be explicitly seen from the behavior of the radial wave function of the higgs and the photon fields , i.e. , @xmath171 for the argument of the radial part , we hereafter drop off the trivial factor @xmath172 appearing in the relations ( [ variable ] ) and ( [ expansion ] ) .
in fact , @xmath173 and @xmath174 denote the radial part of the higgs and the photon fluctuations in eq.([static ] ) . in fig.[fig : n1wave ] we show the wave functions @xmath173 and @xmath174 of the low - lying modes and the corresponding magnetic fluctuations @xmath175 for @xmath2= 0.1 , 0.5 , 1.0 . for @xmath2 = 0.1 and 0.5 , we plot the lowest eigenfunction with radial mass smaller than the threshold @xmath167 . on the other hand , for @xmath2 = 1.0 , there is no discrete pole because the radial mass @xmath124 is larger than the threshold @xmath176 , i.e. , @xmath177 .
then we plot the typical wave function of the low lying mode whose radial mass satisfying @xmath178 . the behavior of the fluctuation fields coincides to our previous expectations . for @xmath2=0.1 ,
the higgs field fluctuation considerably exceeds to the photon fluctuation .
although both of fluctuation fields have the smaller excitation energy than the threshold @xmath179 and this leads localization of them around the vortex core , the distribution of the higgs field fluctuation spreads broader than that of the photon field fluctuation .
this is because the lowest excitation energy @xmath168 is close to the threshold for the higgs field fluctuations , @xmath149 , and the lowest mode is characterized as the shallow bound - state between the static vortex and the higgs field fluctuation .
when we change to the @xmath2=0.5 , the mixing correlation induces the corporative behavior of the higgs and photon field fluctuations and enlarges the binding energy @xmath180 . as a result ,
the lowest excitation mode becomes the deeper bound - state , and then , the higgs field localizes closer to the core , which affects the magnetic field vibration more strongly . for @xmath2=1.0 ,
because of @xmath178 for low - lying modes , only higgs field fluctuation is localized and the photon field @xmath174 oscillates asymptotically ( @xmath181 ) as @xmath182 leading to the asymptotic behavior of the magnetic field @xmath183 like @xmath184 with @xmath185 . (
more precisely , @xmath174 and @xmath183 are expressed as the linear combinations of the @xmath186 and @xmath187 because @xmath173 and @xmath174 are real functions . ) here we comment on the total magnetic flux . in the static case , the total magnetic flux takes the quantized value @xmath188 , as seen in eq.([totalflux ] ) . on the other hand ,
the photon fluctuation generally induces a time - dependent magnetic flux , whose proper time - average vanishes due to the factor @xmath189 .
the contribution from the radial part of the magnetic field fluctuation is @xmath190^{r=\infty}_{r=0}. \label{totalflux2}\end{aligned}\ ] ] for @xmath65 , @xmath174 asymptotically approaches to zero . on the other hand , for @xmath67 , @xmath191 where @xmath192 , @xmath193 represent some constants .
then , for the discrete pole , @xmath194 becomes zero and thus the total magnetic flux is conserved in the arbitrarily cross section perpendicular to the vortex line . in contrast , for the continuum states , the total magnetic flux becomes sensitive to the situation far from the vortex core because of the long tail behavior of @xmath183 .
for the continuum state , the radial photon fluctuation @xmath174 asymptotically behaves as a two - dimensional spherical wave in the homogeneous no - vortex vacuum , and hence its asymptotic part can be regarded as a non - vortex - origin fluctuation . in other words ,
the continuum states can be regarded as the vortex plus a `` scattering wave '' .
to discuss the character change of the excitation modes in wider range of @xmath2 , we define the functions @xmath195 notice that @xmath196=1 $ ] in the normalization condition .
( see eq.([psi2 ] ) . )
we also define the function to estimate the @xmath2-dependence of the collective properties and their mixing ratio for the lowest eigenmode .
one of the good indicators to estimate the degree of the collective nature is the function @xmath197 when the oscillation is sympathetic , this integral takes the large value . in fig.[fig : n1mix ] we plot the behavior of the functions @xmath198 , @xmath199 and @xmath200 for various @xmath2 . as expected , the ration of @xmath118 , or higgs field decreases compared to that of @xmath119 , or photon fluctuation with increasing @xmath2 because the threshold for the higgs field fluctuation increases .
eventually , around @xmath159 , the discrete pole buries in the continuum , and the lowest mode becomes the continuum state of the photon field . at the same time , the correlation between the higgs and photon field fluctuations becomes very small , as seen in the lower panel in fig.[fig : n1mix ] . in this subsection
, we show the results for the peristaltic modes of the @xmath88 vortex , which is realistic in the case of @xmath202 .
we emphasize the difference of the static vortex between @xmath1=1 vortex and @xmath1=2 , 3 vortex , and then discuss their effects on the potential behavior and the resulting excitation spectrum . in this subsection , when we argue @xmath1-dependence of the classical profile and the potential , @xmath2 value is fixed to bps value , i.e. , @xmath2=1/2 .
shown in fig.[potential2 ] are the @xmath1=1 , 2 , 3 static vortex profiles and the corresponding vortex - induced potentials @xmath130 , @xmath132 , @xmath131 for the fluctuations @xmath118 and @xmath119 , respectively . here
we qualitatively explain the changes of the static vortex profiles in terms of @xmath1 . since the total penetrating flux of the magnetic field is proportional to the topological number @xmath1 , the magnetic field around the vortex core increases with increasing @xmath1 .
then the enlarged magnetic flux pushes the higgs field outward or reduce the expectation value of the higgs field .
roughly , this process spreads the normal phase region near the vortex core and leads the outward shift of the interface between normal and superconducting phase .
according to the outward shift of the surface region , where the kinetic energy of the higgs field is enhanced , the maximum of the local energy density also shifts outward , as seen in the upper panel of fig.[potential2 ] .
next we present the qualitative explanation for the single particle potential for the fluctuation fields .
first we give the explanation for the change of @xmath130 for the higgs field fluctuation .
since the area of the magnetic flux is spread with increasing @xmath1 , the intersection surface of the static magnetic and the higgs fields moves outward .
this leads the outward shift of the potential @xmath130 .
the changes of @xmath132 can be interpreted in somewhat indirect way as follows .
the fact read off from fig .
[ potential2 ] is the depth enhancement of the attractive pocket in @xmath132 with increasing @xmath1 .
this enables the photon field fluctuations to enhance and localize around the surface region more easily for larger @xmath1 cases .
the localization of the photon field fluctuation means somewhat rapid spatial change of the wave function @xmath174 , and leads a large fluctuation of the magnetic field @xmath203 .
then we can say that the depth enhancement of the attractive pocket in @xmath132 for large @xmath1 leads the large fluctuation of the magnetic field .
the large magnetic fluctuation reflects the reduction of the higgs field condensate or the weakened meissner effect .
in fact , the depth enhancement of the attractive pocket is closely related to the outward shift of the static higgs field distribution .
finally we comment on the potential @xmath131 , which induces the mixing correlation between the higgs and the photon field fluctuations . as seen in fig .
[ potential2 ] , with increasing @xmath1 , the maximum of @xmath204 shifts outward .
the physical meaning of the change in this potential is naturally explained as the outward shift of the interface between the static higgs field and magnetic field because , at the interface , the fluctuations of the higgs and photon field can frequently interplay . with increasing @xmath1 , the resulting changes in the static configurations and the vortex - induced potential lead the change of the excitation energy of the peristaltic modes .
now we see the excitation spectra of the giant vortex ( @xmath205 ) and their dependence on the @xmath2 value .
the numerical results for @xmath205 cases are shown in figs .
[ fig : nspe2 ] , [ fig : nspe3 ] , respectively . for larger @xmath1 ,
the excitation energy for the lowest mode becomes much smaller as expected from the arguments in the previous paragraphs . as a consequence of the depth enhancement of the attractive pocket in @xmath132 , the discrete pole appears up to considerably large @xmath2 value in the type - ii region .
in such large @xmath206 cases , of course , the fission processes of the giant vortex to many @xmath1=1 vortices become important . according to the analysis for the instability of the static vortex @xcite and the interaction between two vortices @xcite
, there exists no potential barrier for the fission from single type - ii giant vortex to many @xmath1=1 vortices and the giant vortex can fission through the non - axial symmetric modes .
however , in spite of this observation , we have a possibility that axial - symmetric discrete poles play important roles if the lifetime ( decay width ) of the giant vortex is very large compared to the typical time scale ( excitation energy ) of the axial - symmetric fluctuations .
the decay rate can be calculated through the analysis for the non - axial - symmetric modes and the fluctuations of @xmath207 , which are not considered in this work .
we expect that , for @xmath2 not so far beyond the bps value , @xmath208 , the lifetime is large because the force to break up the giant vortex is small . then , if the lifetime is confirmed to be large enough , we can immediately apply our results without further consideration about the correlation between the axial - symmetric and the other modes because these modes behave independently as far as the distortion is sufficiently small , as noticed in sec.iiia .
on the other hand , for @xmath209 cases , since the giant vortex is stable and the axial - symmetric modes are independent from the other modes not considered in this work , the strong bound - state - like modes more easily appear than the case for @xmath210 because of non - existence of the fission processes .
in addition , according to the negative surface tension , vortex with @xmath209 tends to reduce the surface area and then the axial - symmetric fluctuation becomes increasingly prominent with smaller @xmath2 .
therefore , we expect that the discrete pole shown in this section shares the considerable relevance to the low - energy behavior of the giant vortex with @xmath209 .
in this paper , as a new - type mode beyond the stringy one , we have studied the axial - symmetric _ peristaltic _ modes of the vortex , peculiar to the extension of the flux - tube .
we have investigated the peristaltic modes both in type - i and type - ii superconductors for @xmath2111 , 2 , 3 and @xmath212 .
we have treated single vortex in the situation where type - i vortex can be realized without collapses with the other vortices and behaves as the stable object . here
we summarize our main results about the fluctuation modes . with @xmath213 ,
the lowest excitation modes are characterized by the higgs field fluctuation binding with the static vortex . on the other hand , with @xmath214 ,
the mass of the higgs field fluctuation becomes heavier and the photon fluctuation modes are relevant . for @xmath152 ( bps cases ) , there appears the characteristic discrete pole . in this case , the higgs and the photon field fluctuations have the same masses , and then both fluctuations play important roles and their corporative behavior makes the lowest excitation softer one . this mechanism is the main origin of the characteristic discrete pole . since the energy separation between this discrete pole and the continuum threshold is large around @xmath161 , the peristaltic mode may appear most clearly in the superconductor with @xmath161 .
we have also investigated the fluctuation modes of the giant vortex with the topological number @xmath201 .
it is found that , with larger @xmath1 , the photon field fluctuation becomes more important for the lowest eigenmode because the distribution of the static higgs field is shifted outward and the photon fluctuation can excite easily around the core .
we expect that , in @xmath215 cases , the axial - symmetric excitation mode with the bound - state - like pole plays an important role in the low - energy dynamics of the vortex . on the other hand , for @xmath210 cases ,
although the equations for axial - symmetric excitation modes are decoupled from those of the other modes , it is necessarily not only to investigate axial - symmetric modes but also to consider whether these excitation modes can be realized or not during the lifetime of the giant vortex . to estimate the lifetime , we have to treat the non - axial - symmetric modes and the fluctuations of @xmath207 , which drive the fission of the single giant vortex to many @xmath1=1 vortices .
we expect that , for @xmath2 value not so far from the bps value , @xmath2=1/2 , the life time is long enough and we can explore the possibility of the appearance of the discrete pole . finally , we end with a future perspective toward the further investigation for the dynamics of the vortices ( or color - electric flux tubes ) regarding them as dynamical degrees of freedom .
although we restrict ourselves to the axial - symmetric static vortex profile and fluctuations of @xmath216 and @xmath217 , we will investigate the more general cases without using the axial - symmetric ansatz for the static profile and classify the large class of the fluctuation modes , i.e. , ( @xmath216 , @xmath218 , @xmath217 , @xmath219 ) in @xmath220 gauge , or equivalently , ( @xmath216 , @xmath221 , @xmath217 , @xmath219 ) in @xmath57 gauge , which are not fully considered in this work .
for this purposes , we will calculate the static vortex profile and excitation spectrum of the fluctuations in two - dimensional analysis in near future . for type - i case
, we expect that , in low energy , the relevant excitation mode around single vortex is axial - symmetric since the type - i vortex tends to reduce the surface area owing to the positive surface tension . also from the point of view on the kinetic energy ,
the axial - symmetric excitation is favorable . on the other hand ,
the arguments for the type - ii case are more subtle .
because type - ii vortex tends to increase the surface area according to the negative surface tension , it may be important to compare the surface energy to the kinetic energy of the distortion .
for @xmath222 giant vortices , interesting subjects are remained .
one of them is the calculation of the lifetime of single giant vortex for @xmath210 and the comparison to the excitation modes considered in this work .
the lifetime of the giant vortex is closely related to the dynamics of the multi - vortices in non - equilibrium systems . as well as the information for the vortex - vortex potential , the information for the lifetime of single giant vortex may provide the good building block to understand the interesting but complicated phenomena such as the transition of the system from the thermodynamically unstable state to the stable one .
in connection with them , we are also interested in the fluctuations between multi - vortices and their evolution through the vortex - vortex scattering @xcite . as a successive study
, we would like to investigate the results of the full two - dimensional analysis for excitation modes of the vortex in both single vortex and the multi - vortices system @xcite . in application to qcd , the dynamics of the color - electric flux - tubes , or gluonic excitations as nonperturbative modes are also important to discuss the hadrons , which are composed of quarks and color vortex between quarks . in particular , the analysis for the multi - color - flux - tube system can give an good insight to understand the color _ confinement - deconfinement _ transition in qcd , and properties of the hot and dense matter of hadrons or quarks .
it is also interesting to consider heavy ion collisions and the successive occuring expansion of the hot quark matter in close analogy with the evolution of the early universe .
the distribution of the topological defects may affect collective properties of quark gluon plasma and the resulting particle productions . in this work
, we have studied the dynamics of single vortex as one example of the dynamics of the topological objects .
the treatments of topological objects as essential degrees of freedom are quite general and important to understand the crucial aspects of the nonlinear field theory , which sometimes can not be reached with the perturbative treatments for the nonlinear terms .
we would like to extend our analysis to the other topological objects such as color flux - tubes , instantons , magnetic monopoles in gauge theories , or branes in the string theory @xcite , and their many - body dynamics .
the authors are grateful to prof .
r. ikeda and dr .
k. nawa for helpful discussions .
thanks to drs .
m. nitta and e. nakano for useful discussions and suggestions in the yitp workshop , `` fundamental problems and applications of quantum field theory '' .
was supported in part by a grant for scientific research ( no.16540236 , no.19540287 ) from the ministry of education , culture , sports , science and technology ( mext ) of japan .
this work is supported by the grant - in - aid for the 21st century coe
center for diversity and universality in physics " from the mext of japan .
* note added * : after this work is completed , we notice a similar study using a different gauge where ghosts appear @xcite .
their numerical result on the physical modes is consistent with ours .
we give the complete form of the action in the cylindrical coordinates ( @xmath48 ) .
we use the polar - decomposition of eq.([polar ] ) for the higgs field @xmath223 and fix the gauge to @xmath220 . then we obtain the action as @xmath224 where here the electric and magnetic fields are described as @xmath226 we decompose the field into the static part and fluctuation part , @xmath227 where @xmath228 . for the static part , we impose the ansatz ( [ ansatz ] ) , or equivalently , @xmath229 putting these forms into the lagrangian ( [ lag3 ] ) , we obtain @xmath230 = { \cal l}^ { { \rm cl } } + { \cal l}^{(2 ) } + { \cal l}^{(3 ) } + { \cal l}^{(4)},\end{aligned}\ ] ] where @xmath231 + \phi _ { { \rm cl } } ^2 ( a_0 ^ 2 - a_r^2 - a_\theta^2 - a_z^2 ) + \frac { 4\phi_{\rm cl } ( n - ra^ { { \rm cl } } _ \theta ) } { r } \varphi a_{\theta } , \\ & & { \cal l}^{(3 ) } = - 4\kappa^2 \phi^ { { \rm cl } } \varphi^3 + \frac { 2(n - ra^ { { \rm cl } } _ \theta ) } { r } \varphi^2 a_\theta + 2 \phi^ { { \rm cl } } \varphi ( a_0 ^ 2 - a_r^2 - a_\theta^2 - a_z^2 ) , \\ & & { \cal l}^{(4 ) } = \varphi^2 ( a_0 ^ 2 -a_r^2 -a_\theta^2 -a_z^2 ) - \kappa^2 \varphi^4 . \end{aligned}\ ] ] the lagrangian for the static part , i.e. , @xmath232 is related to static energy ( [ ene ] ) as @xmath233 .
the fluctuations of the electric and magnetic fields , i.e. , @xmath234 and @xmath235 are expressed as the eq.([elemag ] ) with the substitution @xmath236 . the linearized euler - lagrange equations for the fluctuations @xmath237 are obtained as @xmath238 \varphi \!\!\!\ !
& & = -\bigg [ \frac{(n - ra^{{\rm cl}}_\theta)^2}{r^2 } + 2\kappa^2(3\phi^{{\rm cl}}-1)^2 \bigg ] \varphi + \frac{2\phi^{{\rm cl}}(n - ra_\theta^{{\rm cl}})}{r } a_\theta , \label{varphi } \\ \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\ ! \bigg [ \frac{\partial^2}{\partial t^2 } -\frac{1}{r } \frac{\partial}{\partial r } \big(r \frac{\partial}{\partial r}\big ) -\frac{\partial^2}{\partial z^2 } \bigg ] a_\theta \!\!\!\!\!\ ! & & = -\bigg[\frac{1}{r^2}+ 2\phi_{{\rm cl}}^2 \bigg ] a_\theta - \frac{4\phi^{{\rm cl}}(n - ra^{{\rm cl}}_\theta)^2}{r^2}\varphi + \frac{1}{r } \frac{\partial}{\partial \theta } \bigg [ \frac{\partial a_0}{\partial t } -\frac{1}{r } \frac{\partial ( r a_r)}{\partial r } -\frac{\partial a_z}{\partial z } \bigg ] , \label{atheta } \\ \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\ ! \bigg [ - \frac{1}{r } \frac{\partial}{\partial r } \big(r \frac{\partial}{\partial r}\big ) -\frac{1}{r^2}\frac{\partial^2}{\partial \theta^2 } -\frac{\partial^2}{\partial z^2 } \bigg ] a_0 \!\!\!\!\!\ ! & & = -2\phi_{\rm cl}^2 a_0 - \frac{\partial}{\partial t } \bigg [ \frac{1}{r^2 } \frac{\partial a_{\theta}}{\partial \theta } + \frac{1}{r } \frac{\partial ( r a_r)}{\partial r } + \frac{\partial a_z}{\partial z}\bigg ] , \label{a0 } \\ \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\ ! \bigg [ \frac{\partial^2}{\partial t^2 } -\frac{1}{r^2 } \frac{\partial^2}{\partial \theta^2 } -\frac{\partial^2}{\partial z^2 } \bigg ] a_r \!\!\!\!\!\ !
& & = - 2\phi _ { { \rm cl } } ^2 a_r - \frac{2}{r^2 } \frac{\partial a_\theta}{\partial \theta } + \frac{\partial } { \partial r } \bigg [ \frac{\partial a_0}{\partial t } - \frac{1}{r}\frac{\partial a_\theta}{\partial \theta } - \frac{\partial a_z}{\partial z } \bigg ] , \label{ar } \\ \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\ !
\bigg [ \frac{\partial^2}{\partial t^2 } -\frac{1}{r } \frac{\partial}{\partial r } \big(r \frac{\partial}{\partial r}\big ) -\frac{1}{r^2}\frac{\partial^2}{\partial \theta^2 } \bigg ] a_z \!\!\!\!\!\ !
& & = -2\phi_{{\rm cl}}^2 a_z -\frac{\partial}{\partial z } \bigg [ \frac{\partial a_0}{\partial t } -\frac{1}{r } \frac{\partial a_\theta}{\partial \theta } -\frac{1}{r } \frac{\partial ( ra_r)}{\partial r } \bigg ] \label{az}.\end{aligned}\ ] ] it is important to notice that we can decouple the equations for the axial - symmetric fluctuations of @xmath96 and @xmath97 expanding the fluctuation modes with the legendre polynomial .
if we restrict ourselves to the @xmath239 modes analysis , we can neglect angular - derivative terms in eq.([varphi])-([az ] ) .
therefore , we can treat the fluctuations @xmath96 and @xmath97 independently from fluctuations @xmath240 , because @xmath96 and @xmath97 are related to @xmath240 only through the angular - derivative terms . as a result , to investigate the axial - symmetric fluctuations of @xmath96 and @xmath97 , we have only to treat the lagrangian ( [ fluc ] ) as mentioned in sec.iiia
. 2 a. a. abrikosov , soviet phys .
jetp * 5 * , 1174 ( 1957 ) .
p. a. m. dirac , proc .
* a133 * , 60 ( 1934 ) , g. t hooft , nucl . phys .
* b79 * , 276 ( 1974 ) , a. m. polyakov , jetp lett * 20 * , 194 ( 1974 ) .
a. a. belavin , a. m. polyakov , a. s. schwartz , and y. s. tyupkin , phys .
59b * , 85 ( 1975 ) .
r. rajaraman , _ solitons and instantons _ , ( north - holland , 1987 ) . n. manton and p. sutcliffe , _ topological solitons _ , ( cambridge university press , cambridge , 2004 ) and references therein .
h. b. nielsen and p. olesen , nucl .
b61 * , 45 ( 1973 ) .
y. nambu , phys .
rev . d*10 * , 4262 ( 1974 ) .
g. t hooft , nucl . phys . *
b190 * , 455 ( 1981 ) .
s. mandelstam , phys .
c23 * , 245 ( 1976 ) .
edited by h. suganuma , n. ishii , m. oka , h. enyo , t. hatsuda , t. kunihiro , and k. yazaki ( world scientific , singapore , 2004 ) h. j. rothe , _ lattice gauge theories : an introduction , 3rd edition _ ( world scientific , singapore , 2005 ) m. creutz , phys .
lett . * 43 * , 553 ( 1979 ) ; * 43 * , 890 ( 1979 ) ; phys .
d * 21 * , 2308 ( 1980 ) .
takahashi , h. matsufuru , y. nemoto , and h. suganuma , phys .
* 86 * , 18 ( 2001 ) ; t.t .
takahashi , h. suganuma , y. nemoto , and h. matsufuru , phys .
d * 65 * , 114509 ( 2002 ) .
h. ichie , v. bornyakov , t. streuer , and g. schierholz , nucl . phys . *
a721 * , c899 ( 2003 ) .
a. vilenkin and e. p. s. shellard , _ cosmic strings and other topological defects _ , ( cambiridge university press , cambridge , 1994 ) and references therein .
t. w. b. kibble , j. phys . a * 9 * , 1387 ( 1976 ) .
w. h. zurek , nature ( london ) * 317 * , 505 ( 1985 ) ; phys .
rep . * 276 * , 177 ( 1996 ) .
see , for example , m. tinkham , _ introduction to superconductivity _
( krieger , malabar , 1985 ) .
m. hindmarsh and a. rajantie , phys .
lett . * 85 * , 4660 ( 2000 ) .
g. j. stephens , l. m. a. bettencourt , and w. h. zurek , phys .
lett . * 88 * , 137004 ( 2002 ) .
for instance , j. polchinski , _ string theory _ , ( cambridge university press , cambridge , 1998 ) .
k. nawa , h. suganuma and t. kojo , phys .
rev . d*85 * , 086003 ( 2007 ) , and its references .
m. luscher , g. munster and p. weisz , nucl . phys . * b180 * , 1 ( 1981 ) .
p. orland , nucl .
b428 * , 221 ( 1994 ) and its references . e.t .
akhmedov , m.n .
chernodub , m.i .
polikarpov and m.a .
zubkov , phys .
rev . d*53 * , 2087 ( 1996 ) and its references .
e. b. bogomolnyi , sov .
* 24 * , 449 ( 1976 ) .
m. k. prasad and c. m. sommerfield , phys .
* 35 * , 760 ( 1975 ) .
h. suganuma , s. sasaki , and h. toki , nucl . phys .
* b435 * , 207 ( 1995 )
. h. suganuma , s. sasaki , h. toki , and h. ichie , prog .
* 120 * , 57 ( 1995 ) .
k. j. juge , j. kuti , and c. j. morningstar , nucl .
( proc . suppl . )
* b63 * , 326 ( 1998 ) ; phys .
lett . * 90 * , 161601 ( 2003 ) .
takahashi and h. suganuma , phys .
90 * , 182001 ( 2003 ) ; phys .
d * 70 * , 074506 ( 2004 ) .
l. jacobs and c. rebbi , phys .
rev . b*19 * , 4486 ( 1979 ) .
h. j. de vega and f. a. schaposnik , phys . rev .
* d14 * , 1100 ( 1976 ) .
p. tholfsen and h. meissner , phys .
rev . * 169 * , 413 ( 1968 ) .
l. kramer , phys .
b * 3 * , 3821 ( 1971 ) .
h. frahm , s. ullah and a. t. dorsey , phys .
lett . * 66 * , 3067 ( 1991 ) .
f. liu , m. mondello and n. goldenfeld , phys .
. lett . * 66 * , 3071 ( 1991 ) .
p. w. anderson , phys .
130 * , 439 ( 1963 ) .
p. w. higgs , phys .
* 13 * , 508 ( 1964 ) .
s. gustafson and i. m. sigal , commun .
* 212 * , 257 ( 2000 ) .
e. myers , c. r. rebbi , and r. strilka , phys . rev . *
d45 * , 1355 ( 1992 ) t. kojo , h. suganuma , and k. tsumura , in preparation . for recent review , m. eto , y. isozumi , m. nitta , k. ohashi , and n. sakai , j. phys .
* a39 * , r315 ( 2006 ) . m. goodband and m. hindmarsh , phys .
d*52 * , 4621 ( 1995 ) . | using the abelian higgs model , we study the radial excitations of single vortex and their propagation modes along the vortex line .
we call such beyond - stringy modes _ peristaltic _
modes of single vortex . with the profile of the static vortex
, we derive the vortex - induced potential , i.e. , single - particle potential for the higgs and the photon field fluctuations around the static vortex , and investigate the coherently propagating fluctuations which corresponds to the vibration of the vortex .
we derive , analyze and numerically solve the field equations of the higgs and the photon field fluctuations around the static vortex with various ginzburg - landau parameter @xmath0 and topological charge @xmath1 . around the bps value or critical coupling @xmath2=1/2
, there appears a significant correlation between the higgs and the photon field fluctuations mediated by the static vortex . as a result ,
for @xmath2=1/2 , we find the characteristic new - type discrete pole of the peristaltic mode corresponding to the quasi - bound - state of coherently fluctuating fields and the static vortex .
we investigate its excitation energy , correlation energy of coherent fluctuations , spatial distributions , and the resulting magnetic flux behavior in detail .
our investigation covers not only usual type - ii vortices with @xmath1=1 but also type - i and type - ii vortices with @xmath3 for the application to various general systems where the vortex - like objects behave as the essential degrees of freedom . |
the discoveries of c@xmath0 fullerene by kroto and co - workers in 1985 @xcite and carbon nanotubes ( cnts ) by iijima in 1991 @xcite , following a preliminary carbon filamentous growth work @xcite , have opened up a new scientific revolution for nanotechnology . since then
, an enormous amount of research has been performed on representative carbon nanomaterials , notably single - wall carbon nanotubes ( swcnts ) , discovered in 1993 @xcite , and graphene , first isolated in 2004 @xcite . not only attracting much attention for future applications in microscale and nanoscale optoelectronic devices , these carbon nanomaterials are stimulating much interest from a fundamental point of view , considered to be one of the most ideal systems that exhibit the quantum mechanical nature of interacting low - dimensional electrons . while initial investigations on these materials concentrated on dc characteristics , recent theoretical studies have instigated a flurry of new experimental activities to uncover unusual ac properties .
both swcnts and graphene are expected to show exotic terahertz ( thz ) dynamics that can lead to innovative optoelectronic applications @xcite .
these properties are inherently related to their unique , low - dimensional band structure , combined with many - body interactions of quantum - confined carriers . from a fundamental point of view
, dynamic ( or optical or ac ) conductivity @xmath1 , where @xmath2 is the ( angular ) frequency of the applied electric field , is expected to provide a wealth of information on quantum confinement , electron interactions , and disorder . for one - dimensional ( 1-d ) electron systems such as swcnts ,
there have been detailed theoretical calculations of @xmath1 , taking into account interactions and disorder to varying degrees ( see , e.g. , @xcite and pp .
219 - 237 of @xcite ) . for a metallic swcnt , for example , ando @xcite calculated @xmath1 within a self - consistent born approximation , which indicated that there can exist non - drude - like conductivity , depending on the range of scattering potentials .
one particularly important question for metallic swcnts is whether the optical conductivity satisfies the so - called @xmath3 scaling , i.e. , the frequency dependence appears only through a particular combination of @xmath2 and @xmath4 , i.e. , @xmath3 @xcite , where @xmath4 is the sample temperature .
this implies that the system is at a quantum critical point @xcite .
a tomonaga - luttinger liquid has the characteristic feature that it is quantum critical over a finite range of parameters due to enhanced quantum fluctuations in 1-d ( see , e.g. , @xcite ) , and thus , is expected to exhibit a @xmath5 scaling law in conductivity . a number of experimental thz / far - infrared spectroscopic studies have been performed over the last decade on swcnts of various forms @xcite , producing an array of conflicting results with contradicting interpretations .
this is partly due to the widely differing types of samples used in these studies grown by different methods ( hipco , comocat , cvd , arc discharge , and laser ablation ) and put in a variety of polymer films that are transparent in the thz range .
nanotubes in most of these samples were bundled and typically consisted of a mixture of semiconducting and metallic nanotubes with a wide distribution of diameters .
some of the samples used were partially aligned through mechanical stretching , showing some degree of anisotropy in their thz response @xcite .
one common spectral feature that many groups have detected is a broad absorption peak around 4 thz ( or 135 @xmath6 or @xmath717 mev ) .
this feature , first observed by ugawa _
et al_. @xcite , has been interpreted as interband absorption in mod-3 non - armchair nanotubes with curvature - induced gaps @xcite or absorption due to plasmon oscillations along the tube axis @xcite , but a consensus has not achieved @xcite .
the ac dynamics of dirac fermions in graphene have attracted much theoretical attention the influence of linear dispersions , two - dimensionality , and disorder has been extensively discussed by many theorists @xcite . however , the influence of electron - electron interactions on the optical conductivity of graphene is somewhat controversial .
theoretical studies using different methods have led to different conclusions as to the magnitude of many - body corrections to the drude - like intraband optical conductivity ( see , e.g. , @xcite and references cited therein ) . experimentally , while a number of studies have confirmed the so - called universal optical conductivity @xmath8 for _ interband _ transitions in a wide spectra range @xcite , successful experimental studies of the _ intraband _ conductivity have been very limited @xcite , except for cyclotron resonance measurements in magnetic fields @xcite . on the other hand , theoretical studies on _ nonlinear and non - equilibrium properties _ of graphene have emerged in recent years , urging further experiments to be done using time - resolved and/or nonlinear optical spectroscopy . for example ,
mikhailov and ziegler @xcite , using a semiclassical approximation , have shown that dirac fermions in graphene with a dispersion @xmath9(@xmath10 ) = @xmath11@xmath12@xmath10@xmath12 ( where @xmath11 is the fermi velocity ) in the presence of an applied ac electric field @xmath13(@xmath14 ) = @xmath15@xmath16@xmath17 would produce an ac current @xmath18 _ within linear response_. note that not only does the current oscillate at the frequency of the applied field @xmath2 , but it also contains all the odd harmonics .
this should be contrasted to conventional semiconductors such as gaas , which , within linear response , will only produce an ac current with frequency @xmath2 , i.e. , @xmath19 , where @xmath20 is the effective mass .
therefore , the dynamics of graphene in an ac electric field are intrinsically nonlinear , and efficient frequency multiplications for thz generation can be expected for microwave - driven graphene . another outstanding theoretical prediction is thz amplification in optically pumped graphene @xcite .
ryzhii and co - workers calculated the dynamic conductivity of a non - equilibrium 2d electron - hole system in graphene under interband optical excitation .
both interband and intraband transitions were taken into account in their model . under optical pumping with photon energy @xmath21@xmath22
, electrons and holes are photogenerated with energy @xmath23 = @xmath24/2 and emit a cascade of optical phonons with energy @xmath25 , leading to population accumulation of the bottom of the conduction band ( by electrons ) and the top of the valence band ( by holes ) , as shown in fig .
[ ryzhiijap ] .
they demonstrated that sufficiently strong optical pumping will result in population inversion , making the real part of the net ac conductivity negative , i.e. , amplification .
due to the gapless energy spectrum , this negative ac conductivity takes place in the range of thz frequencies .
this effect might be used in graphene - based coherent sources for thz radiation .
the 1-d character of confined carriers in swcnts manifests itself as strong anisotropy in electric , magnetic , and optical properties @xcite .
individual metallic swcnts have been shown to be excellent 1-d electrical conductors with very long coherence lengths @xcite , while individual semiconducting swcnts have been shown to absorb and emit light only when the light polarization is parallel to the tube axis @xcite . moreover ,
individualized swcnts , both metallic and semiconducting , have been shown to align well in an external magnetic field @xcite due to their anisotropic magnetic susceptibilities @xcite , and aligned swcnts exhibit strong linear dichroism due to their anisotropic optical properties . even in bundled samples where swcnts form aggregates through van der waals attraction ,
their anisotropic properties are expected to be preserved to a large degree , even though systematic optical spectroscopy studies have been limited @xcite due to the rarity of ensemble samples in which the swcnts are highly aligned . here
, we describe results of polarization - dependent thz transmission measurements on films of macroscopically - aligned swcnts , which demonstrate an extremely high degree of anisotropy @xcite .
strikingly , when the thz polarization is perpendicular to the nanotube alignment direction , practically no absorption is observed despite the macroscopic thickness of the film . on the other hand ,
when the polarization is parallel to the alignment direction , there is strong absorption .
the degree of polarization in terms of absorbance is 1 and the reduced linear dichroism is 3 , throughout the entire frequency range of our experiment ( 0.1 - 2.2 thz ) . using the theory of linear dichroism for an ensemble of anisotropic molecules @xcite , we show that this value of reduced dichroism ( i.e. , 3 ) is possible only when the nematic order parameter ( @xmath26 ) is 1 .
these observations are a direct result of the 1-d nature of conduction electrons in the nanotubes and at the same time demonstrate that any misalignment of nanotubes in the film must have characteristic length scales much smaller than the wavelengths used in these experiments ( 1.5 mm to 150 @xmath27 m ) . these results on aligned swcnts , as well as a recent report by kyoung _
et al_. on aligned multi - wall cnts @xcite , indeed suggest that aligned cnt films perform as ideal linear polarizers in the thz frequency range . these aligned films ( or stacks of films
@xcite ) exhibit comparable performance to commercial wire - grid technology but have added benefits of ( i ) broadband thz absorption driven by the inherent 1-d character of the cnts and ( ii ) mechanical robustness in diverse operation conditions .
namely , in comparison to wire - grid technology , the thz performance of these materials is driven not by the precise structure of the conductive wires but rather the inherent anisotropic thz absorption properties of aligned cnts .
the aligned swcnt films studied were produced by natural self - assembly of swcnts into densely packed and highly aligned macroscopic materials during synthesis . utilizing optical lithography to define the pitch between lines of catalyst ,
water - assisted chemical vapor deposition was employed to grow aligned swcnts in high aspect - ratio lines .
following growth , the lines were transferred to a sapphire or silicon substrate to produce a thz polarizer @xcite .
to optimize the device performance with respect to extinction , multiple layers were stacked until full extinction of linearly polarized thz radiation was achieved in a configuration where the thz field is parallel to the alignment .
the benefit of this approach compared to other possible approaches is the simplicity of the contact - transfer process that makes this scalable to large areas .
there are no known limitations in this approach that would inhibit scaling to full wafers or even continuous roll - to - roll growth and contact transfer techniques to yield high - throughput production of thz devices .
the thz - tds system used a ti : sapphire femtosecond laser that excited either a znte crystal or a photoconductive antenna to generate and detect coherent thz radiation .
the sample was positioned at the thz focus and was rotated about the propagation direction of the thz beam , which changed the angle , @xmath28 , between the nanotube axis and the thz polarization direction from 0@xmath29 to 90@xmath29 , as shown in fig [ swntsthzpolarizer ] .
more details about the experimental setup are described elsewhere @xcite . with an intrinsic silicon substrate , which is transparent in almost the entire infrared ( ir ) range , we were also able to perform similar polarization - dependent transmission measurements using fourier transform infrared ( ftir ) spectroscopy .
figure [ swntsthzpolarizer ] shows transmitted time - domain waveforms for four different angles ( @xmath28 = 0@xmath29 , 30@xmath29 , 45@xmath29 , and 90@xmath29 ) between the thz polarization direction and the cnt alignment direction , together with the waveform transmitted through a reference sapphire substrate with no nanotubes .
note that the 90-degree trace precisely coincides with the reference waveform , which means that there is no attenuation for this thz polarization ( perpendicular to the tubes ) . from the measured absorption anisotropy , we calculated the reduced linear dichrosim to be 3 , corresponding to a nematic order parameter of 1 @xcite .
this result suggests that any misalignment of nanotubes in the film is negligible compared to the wavelengths used in these experiments ( 1.5 mm @xmath30 150 @xmath27 m ) . in order to optimize the device performance with respect to extinction ,
multiple layers were stacked in parallel with each other , as shown in fig .
[ swntsfilmstacks ] , until full extinction of linearly polarized thz radiation was achieved in a configuration where the thz field is parallel to the alignment . to elucidate polarizer performance , we calculated the degree of polarization ( dop ) , defined as dop = ( @xmath31)/(@xmath32 ) , and the extinction ratio ( er ) , defined as er = @xmath33/@xmath34 , where @xmath33 is the transmittance for the parallel case , and @xmath34 is the transmittance for the perpendicular case . as shown in fig .
[ swntsstacksdoper](a ) , the dop value of our swcnts polarizer increases with the swcnts film thickness and reaches a value at 99.9@xmath35 throughout the whole measured frequency range for a triple - stacked film .
figure [ swntsstacksdoper](b ) indicates that the er value is increasing dramatically with the film thickness , achieving an average value of 33.4 db in the [email protected] thz range two orders of magnitude better than the thinner swcnt films .
these results demonstrate the remarkable utility of aligned swcnts for thz applications .
we fourier - transformed the time - domain data of the single highly aligned swcnt thin film shown in fig .
[ swntsthzpolarizer ] into the frequency domain and extracted the parallel and perpendicular elements of the the dynamic complex conductivity tensor @xmath36 , i.e. , @xmath37 and @xmath38 , respectively , as shown in fig .
[ condtensor ] .
highly anisotropic responses are again observed . at 90@xmath29 ( thz polarization perpendicular to the nanotube axis ) ,
the real part of the conductivity is zero throughout the entire frequency range , showing no sign of absorption .
in contrast , at 0@xmath29 ( thz polarization parallel to the nanotube axis ) the conductivity is finite and reaches @xmath7140 s@xmath39@xmath6 at 1.8 thz .
note that the real part monotonically increases with increasing frequency in this frequency range .
this is consistent with the existence of a peak at a higher frequency , as observed by many other groups on various types of swcnts , as discussed previously and shown in fig .
[ 4thz - peak ] .
the results shown here unambiguously demonstrate that such a finite - frequency peak appears only in the parallel component of the dynamic conductivity tensor .
graphene has been intensively studied since its first isolation in 2004 @xcite .
this zero - gap semiconductor consisting of a single layer of @xmath40-bonded carbon atoms arranged into a 2-d honeycomb lattice possesses a photon - like , linear energy dispersion , which is expected to lead to exceptionally nonlinear electro - dynamic properties @xcite .
since both intraband and interband transitions in graphene are expected to be sensitive to the location of the fermi energy , ac studies with tunable carrier concentration would provide significant new insights into the dynamics of 2-d dirac fermions .
the fermi level can be tuned either by doping or gating .
substitution of carbon atoms in graphene by nitrogen and boron have been attempted , but this dramatically decreases the mobility by breaking its lattice structure .
physically adsorbed molecules can also dope graphene , but this is not stable and leads to a suppression of mobility as well . so far , applying a controllable gate voltage to graphene to transfer carriers from the doped silicon substrate is still the most common and reliable way . by utilizing applied gate voltages ,
different groups have observed tunable interband optical transitions @xcite , tunable intraband far - infrared conductivity @xcite , and a systematic g - band change with gate voltage in raman spectra @xcite . here , we describe our thz and ir study of _ large - area _ ( centimeter scale ) graphene with an electrically tunable fermi level ( @xmath41 ) . in a field effect transistor consisting of graphene on a sio@xmath42/@xmath43-si substrate , the intensity of the transmitted thz electromagnetic wave was observed to change with the gate voltage .
both the drude - like intraband conductivities and the ` 2@xmath41 onset ' of the interband transitions , monitored through thz time - domain spectroscpy and fourier - transform infrared spectroscopy , respectively , were both modulated by the gate voltage .
[ gateexpconf ] the large - area graphene sample used for this spectroscopy study was grown from a solid state carbon source
poly(methyl methacrylate ) ( pmma ) @xcite .
grown single - layer graphene was transferred to a @xmath71.5 cm @xmath44 1.5 cm @xmath43-type silicon wafer ( 5 - 10 @xmath45-cm ) with a 300-nm thick sio@xmath42 layer .
gold electrodes were then deposited on corners of the 8 mm @xmath44 8 mm graphene film and on the back of the p - si substrate ( see fig . [ gateexpconf ] .
the thz wave was normal incident onto the center of graphene , and the transmitted thz wave for was detected and analyzed as a function of gate voltage .
figure [ thztdsgrgating](a ) shows gate - voltage - dependent transmitted thz waveforms , the signal peaks of which are zoomed in and illustrated in fig .
[ thztdsgrgating](b ) and ( c ) . at + 30
v , the fermi energy is at the charge neutrality point ( cnp ) , and thus , the highest thz transmission is obtained . at all other voltages above and below + 30
v , thz transmission decreases ( or absorption increases ) monotonically with the voltage change , as shown fig .
[ thztdsgrgating](b ) and ( c ) .
figure [ thztdsgrgating](d ) shows the transmitted thz beam power as a function of gate voltage ( blue circled line ) , demonstrating that + 30 v is indeed closest to the cnp and the unbiased ( 0-v ) point is on the @xmath43-doped side .
the dc resistance of the device was measured _ in situ _ and is plotted as a function of gate voltage in fig .
[ thztdsgrgating](d ) ( the red trace ) , showing agreement with the gate dependence of the transmitted thz power .
after fourier transforming the time - domain data shown in fig .
[ thztdsgrgating ] , we obtained the corresponding transmission spectrum in the frequency domain with the signal taken at + 30 v ( the cnp ) as a reference . using standard analysis methods appropriate for thin conducting films ,
the 2d thz conductance of the graphene sample was extracted as a function of frequency , as shown for three representative gate voltage values in fig .
[ graphenethz - ir](a ) .
the ac conductance is seen to decrease with increasing frequency throughout the range of 0.3 - 2.1 thz ( 10 - 70 @xmath6 ) , in contrast to the case of swcnts ( see fig .
[ condtensor ] ) .
the thz data is combined with ftir data taken at higher frequencies ( 100 - 600 @xmath6 ) .
also shown in fig .
[ graphenethz - ir](a ) are theoretical fits to the data ( colored dashed lines ) to deduce the fermi energy ( @xmath41 ) and scattering rate ( @xmath46 ) , using @xcite @xmath47 where @xmath48 is the electronic charge , @xmath21 is the reduced planck constant , @xmath49 is the boltzmann constant , and @xmath4 is the temperature (= 300 k in our experiments ) .
the larger the gate voltage is , measured from the cnp , the larger the fermi energy , and as a result , the larger the overall conductance .
the extracted values for the fermi energy and scattering rate are listed within fig .
[ graphenethz - ir](a ) .
the obtained values for @xmath46 are on the order of 2 @xmath44 10@xmath50 sec@xmath51 ( or a scattering time of @xmath750 fs ) , which agrees with our results on the same type of graphene samples obtained from high - field mid - ir cyclotron resonance measurements @xcite .
the obtained @xmath41 values are also generally consistent with the `` 2@xmath41 onset '' observed in the mir range , shown in fig .
[ graphenethz - ir](b ) . here
, the ratio of the transmission spectrum under a certain gate voltage by that under the cnp gate voltage is plotted for various gate voltages . as we can see , as the gate voltage was changed , the 2@xmath41 onset peak shifts monotonically , indicating a tuned fermi level .
these gated large - area graphene samples are thus promising for further basic studies of low - frequency phenomena with a tunable fermi energy as well as for manipulation of thz and infrared waves .
in this article , we reviewed our recent work progress on the optical conductivities of low - dimensional carriers in single - wall carbon nanotubes ( swcnts ) and graphene , two types of carbon nanomaterials with molecules structured with @xmath40-bonded carbon atoms .
polarization - dependent thz transmission measurements on highly aligned swcnt films revealed extreme anisotropy .
based on this , we synthesized polarizers made of aligned swcnts on sapphire substrates that exhibit remarkable broadband performance in the thz frequency range , with degrees of polarization of 99.9@xmath35 from 0.2 to 2 thz , and extinction ratios of up to @xmath735 db . this material yields broadband polarization and extinction features that outperform the conventional wire - grid polarizers . through a proper model ,
the dynamic complex conductivity tensor elements of these swcnt films were determined from the thz - tds data , and both the real part and the imaginary part of the parallel conductivity show a non - drude frequency dependence .
we also measured the transmission spectra of large - area graphene films grown from solid state carbon source on sio@xmath52/@xmath43-si through thz - tds and fourier transform infrared spectroscopy , and determined the optical conductivity of graphene in the thz and infrared ranges .
the fermi energy was tuned by applying a gate voltage to the device , which , in turn , modulated the transmission of thz and ir waves .
the frequency dependence of the transmission allowed us to determine the fermi energy and scattering time at each gate voltage , both from the thz ( intraband ) and ir ( interband ) spectra .
this work was supported by the national science foundation ( through grant no .
oise-0530220 and no .
oise-0968405 ) , the department of energy ( through grant no .
defg02 - 06er46308 ) , and the robert a. welch foundation ( through grant no .
we thank cary l. pint and robert h. hauge for providing us with the aligned carbon nanotube films and jun yao , zhengzong sun , zheng yan , zhong jin , and james m. tour for providing us with the large - area graphene samples . | low - dimensional carbon nanostructures , such as single - wall carbon nanotubes ( swcnts ) and graphene , offer new opportunities for terahertz science and technology .
being zero - gap systems with a linear , photon - like energy dispersion , metallic swcnts and graphene exhibit a variety of extraordinary properties .
their dc and linear electrical properties have been extensively studied in the last decade , but their unusual finite - frequency , nonlinear , and/or non - equilibrium properties are largely unexplored , although they are predicted to be useful for new terahertz device applications .
terahertz dynamic conductivity measurements allow us to probe the dynamics of such photon - like electrons , or massless dirac fermions . here , we use terahertz time - domain spectroscopy and fourier transform infrared spectroscopy to investigate terahertz conductivities of one - dimensional and two - dimensional electrons , respectively , in films of highly aligned swcnts and gated large - area graphene . in swcnts , we observe extremely anisotropic terahertz conductivities , promising for terahertz polarizer applications . in graphene , we demonstrate that terahertz and infrared properties sensitively change with the fermi energy , which can be controlled by electrical gating and thermal annealing .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
recent lattice qcd calculations predict that there is a first order phase transition from hadronic matter to a quark - gluon plasma that ends in a critical point , with a continuous phase transition on the other side of the critical point .
the relativistic heavy ion collider ( rhic ) has conducted a program to probe different regions of the qcd phase diagram in the vicinity of the possible critical point with a beam energy scan . during 2010 and 2011 ,
rhic provided au+au collisions to phenix at @xmath0 200 gev , 62.4 gev , 39 gev , 27 gev , 19.6 gev , and 7.7 gev .
analysis of the data concentrates on two strategies : looking for signs of the onset of deconfinement by comparing to results at the top rhic energy , and searching for direct signatures of a critical point . results from phenix covering transverse energy production , multipicity fluctuations , and the skewness and kurtosis of net charge distributions will be discussed in this article .
phenix has measured transverse energy ( @xmath1 ) production in au+au collisions at the following collision energies : 200 , 62.4 , 39 , 27 , 19.6 , and 7.7 gev .
these observables are closely related to the geometry of the system and are fundamental measurements necessary to understand the global properties of the collision .
this work extends the previous phenix measurements in 200 , 130 , and 19.6 gev au+au collisions @xcite .
total @xmath1 production results are summarized in figure [ fig : ebjexcite ] , which shows the excitation function of the estimated value of the bjorken energy density @xcite expressed as @xmath2 where @xmath3 is the formation time and @xmath4 is the transverse overlap area of the nuclei .
the bjorken energy density increases monotonically over the range of the rhic beam energy scan . also shown is the estimate for 200 gev u+u collisions taken during the 2012 running period .
although @xmath1 production dramatically increases at lhc energies compared to rhic energies , the shape of the distributions as a function of the number of participants , @xmath5 , is independent of the collision energy .
this is illustrated in figure [ fig : detdetarhiclhc ] , which shows an overlay of the distributions of @xmath6 normalized by the number of participant pairs for 7.7 gev , 200 gev , and 2.76 tev au+au collisions .
the 200 gev and 7.7 gev distributions have been scaled up by a factors of 2.6 and 9.7 , respectively .
the shape of the distribution as a function of @xmath5 appears to be driven by the collision geometry . , multiplied by the formation time in central au+au collisions at mid - rapidity as a function of @xmath7 .
the open circle represents the estimate for 200 gev u+u collisions.,scaledwidth=50.0% ] normalized by the number of participant pairs as a function @xmath5 .
overlayed are the distributions from 7.7 gev , 200 gev , and 2.76 tev au+au collisions .
the phenix data has been scaled up to overlay the alice data @xcite.,scaledwidth=50.0% ]
near the qcd critical point , it is expected that fluctuations in the charged particle multiplicity will increase @xcite .
phenix has extended the previous analysis of multiplicity fluctuations in 200 and 62.4 gev au+au collisions @xcite to 39 and 7.7 gev au+au collisions .
charged particle multiplicity fluctuations are measured using the scaled variance , @xmath8 , which is the standard deviation scaled by the mean of the distribution .
the scaled variance is corrected for contributions due to non - dynamic impact parameter fluctuations using the method described in @xcite .
figure [ fig : svarexcite ] shows the phenix results for central collisions compared to results from the na49 collaboration @xcite as a function of @xmath7 .
there is no indication of the presence of a critical point from the phenix results alone . .
circles represent the results from phenix.,scaledwidth=50.0% ]
the shapes of the distributions of the event - by - event net charge are expected to be sensitive to the presence of the critical point @xcite .
phenix has measured the skewness ( @xmath9 ) and the kurtosis ( @xmath10 ) of net charge distributions in au+au collisions at 200 , 62.4 , 39 , and 7.7 gev .
these values are expressed in terms that can be associated with the quark number susceptibilities , @xmath11 : @xmath12 and @xmath13 @xcite .
the skewness and kurtosis for central collisions are shown in figure [ fig : skewkurt ] as a function of @xmath7 .
the data are compared to urqmd and hijing simulation results processed through the phenix acceptance and detector response .
there is no excess above the simulation results observed in the data at these four collision energies .
presented here are some of the phenix results from the rhic beam energy scan program . from the analyses completed to date , there is no significant indication of the presence of the qcd critical point .
however , many analyses from phenix , particularly at @xmath0 27 gev and 19.6 gev , will be available soon .
adler et al . , phys .
c 71 , 034908 ( 2005 ) .
j. d. bjorken , phys .
d 27 , 140 ( 1983 ) . c. loizides et al . , arxiv:1106.6324v1 ( 2011 ) .
m. stephanov et al , phys .
d 60 , 114028 ( 1999 ) . a. adare et al , phys
c 78 , 044902 ( 2008 ) . c. alt et al , phys .
c 78 , 034914 ( 2008 ) .
r. v. gavai and s. gupta , phys .
b 696 , 459 ( 2011 ) .
f. karsch and k. redlich , phys . lett .
b 695 , 136 ( 2011 ) . | the phenix experiment at rhic has conducted a beam energy scan at several collision energies in order to search for signatures of the qcd critical point and the onset of deconfinement .
phenix has conducted measurements of transverse energy production , muliplicity fluctuations , and the skewness and kurtosis of net charge distributions .
the data analyzed to date show no significant indications of the presence of the critical point . |
protostellar jets are seen associated with low - mass protostars in the early stages of star formation . in spite of numerous studies ,
their physical properties ( e.g. , speed , episodic nature , collimation , and angular momentum ) and thus launching mechanisms are still not well understood .
they are believed to be launched from accretion disks around the protostars ( see recent reviews by e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , allowing us to probe the accretion process , which remains heretofore unresolved , as it requires us to observe directly in the inner parts at the au scale . the submillimeter array ( sma )
@xcite , with the capability to probe warm and dense molecular gas at high angular resolution , can be and has been used to study the physical properties of the jets in detail ( e.g. , * ? ? ?
* ; * ? ? ?
the hh 211 outflow is an archetypical outflow with a highly collimated jet , located in the ic 348 complex in perseus .
the distance of the ic 348 complex is assumed to be 320 pc @xcite , but it could be 250 pc @xcite .
the outflow was discovered in 2 shock emission at 2.12 @xcite , powered by a young , low - mass , and low - luminosity ( @xmath8 ) class 0 protostar with @xmath9 k @xcite .
a collimated co jet was seen surrounded by cavity walls in co ( j=2 - 1 ) ( gueth & guilloteau 1999 , hereafter @xcite ) .
recent observations with the sma in sio ( j=5 - 4 ) @xcite and ( j=8 - 7 ) @xcite also revealed a collimated sio jet consisting of a chain of spatially unresolved knots aligned with the co jet , tracing shock emission along the jet .
in this paper , we present observations of the jet in sio ( j=8 - 7 ) and co ( j=3 - 2 ) at higher - angular resolution than in @xcite , in order to better constrain the physical properties and thus the launching mechanisms of the jet .
observations toward the hh 211 jet were carried out with the sma on 2004 october 4 and 18 in the compact configuration and on 2004 september 10 in the extended configuration .
note that the observation on 2004 october 18 was only a short partial track and thus not included in our analysis .
the zenith opacities were @xmath10 and 0.13 , respectively , on 2004 october 4 and 2004 september 10 .
sio ( @xmath0 ) and co ( @xmath1 ) lines were observed simultaneously with continuum using the 345 ghz band receivers .
the receivers have two sidebands , lower and upper , covering the frequency range from 335.58 to 337.55 and from 345.59 to 347.56 ghz , respectively . combining the line - free portions of the two sidebands results in a total continuum bandwidth of @xmath2 3.7 ghz centered at @xmath2 342 ghz ( or @xmath11 880 ) for the continuum .
the baselines have projected lengths ranging from @xmath2 20 to 225 m. the primary beam has a size of @xmath2 and three pointings were used to map the jet . for the correlator , 128 spectral channels were used for each 104 mhz chunk , resulting in a velocity resolution of @xmath2 0.7 per channel .
the visibility data were calibrated with the mir package , with saturn , venus , quasars 3c84 and j0355 + 508 as passband calibrators , quasars 3c84 and j0355 + 508 as gain calibrators , and uranus ( @xmath12 jy ) as a flux calibrator .
the flux uncertainty is estimated to be @xmath2 20% .
the calibrated visibility data were imaged with the miriad package .
the dirty maps that were produced from the calibrated visibility data were cleaned using the steer clean method , producing the clean component maps .
the final maps were obtained by restoring the clean component maps with a synthesized ( gaussian ) beam fitted to the main lobe of the dirty beam . with natural weighting ,
the synthesized beam has a size of @xmath13 at a position angle ( p.a . ) of @xmath2 70 .
the rms noise level is @xmath2 0.33 in the channel maps and 6.5 in the continuum map .
the velocities of the channel maps are lsr .
the typical rms of the gain phases is @xmath2 20 , resulting in an absolute positional accuracy of one tenth of the synthesized beam , or @xmath2 .
our results are presented in comparison to the ir image ( 2 at 2.12 + continuum ) made with the vlt on 2002 january 4 @xcite , which shows clear shock interactions along the jet axis . since our observations were carried out @xmath2 3 years later than the ir image , the 2 shock knots and bow shocks might have moved down along the jet axis by @xmath2 @xmath4 from their positions in that image , assuming a typical jet velocity of 100@xmath4200 .
this movement , however , does not affect significantly our comparison and conclusions , considering that the angular resolution of our observations is @xmath2 1 along the jet axis . the systemic velocity in this region is assumed to be @xmath14 lsr , as in @xcite and @xcite .
throughout this paper , the velocity is relative to this systemic value .
continuum emission is detected at 342 ghz at the center of the 2 flow with a total ( integrated ) flux of 0.44@xmath15 jy .
it has a peak at @xmath16 , @xmath17 , with a positional uncertainty of ( fig .
[ fig : cont ] ) .
this peak position is within of that found at 43.3 ghz ( or @xmath18 mm ) with the vla at an angular resolution of @xmath2 @xcite , and is thus considered as the position of the protostar throughout this paper .
note that the continuum flux here is about twice as that found in @xcite , which was based on the observation on 2004 october 18 .
as mentioned , that observation was only a short partial track with poor @xmath19 coverage .
thus , even though their visibility amplitude versus @xmath19 distance plot is similar to ours , most of the extended emission was not recovered in their image ( see their fig .
the emission is seen extending @xmath2 to the southwest from the protostar roughly perpendicular to the jet axis , similar to that seen at 230 ghz @xcite , likely tracing the flattened envelope perpendicular to the jet axis .
the envelope , however , seems asymmetric with less emission extending to the northeast .
faint emission is also seen extending to the northwest , similar to that seen at 220 ghz @xcite , probably tracing the envelope material around the west outflow lobe . near the protostar ,
the structure is compact and not resolved , as seen in the map made using the visibility data with the @xmath19 distance greater than 100 k@xmath20 ( see fig .
[ fig : cont]b ) , with a flux of @xmath2 0.08@xmath21 jy .
this unresolved compact source is also seen in the visibility amplitude versus @xmath19 distance plot , with a similar flux ( fig .
[ fig : uvsedcont]a ) .
the size ( diameter ) of this compact source has an upper limit of @xmath2 or 130 au set by the longest baseline .
two components , an envelope and a compact source , have also been suggested in other class 0 sources , e.g. , iras 16293 - 2422 and l1448-c @xcite , and hh 212 @xcite . the spectral energy distribution ( sed ) of the continuum source ( see fig .
[ fig : uvsedcont]b ) indicates that the continuum emission at 342 ghz is mainly thermal dust emission . assuming a constant temperature @xmath22 , a frequency - independent source size @xmath23 , and a mass opacity @xmath24 @xmath25 g@xmath6 @xcite for the dust , the sed can be fitted with @xmath26 30 k , @xmath27 3.6 arcsec@xmath28 , @xmath29 , and an optical depth @xmath30 0.086 at 342 ghz . our value of @xmath31 is similar to that found by @xcite and @xcite .
the dust temperature is also consistent with the bolometric temperature , which was found to be @xmath32 k @xcite .
however , our @xmath31 is smaller than that found including fluxes from the larger - scale envelope ( with a mass of @xmath2 0.8 ) in photometric broadband observations , which is 1.3 @xcite .
@xmath33 was also found toward large - scale envelopes around other embedded ysos @xcite .
thus , excluding the larger - scale envelope tends to decrease the value of @xmath31 , suggesting that the dust grains grow bigger toward the source .
the compact source at the center may have even lower @xmath31 . to examine this , fluxes of the compact source at 342 and 220 ghz are compared with @xmath34 or @xmath35 . assuming that fluxes at long baselines are dominated by the compact source , the fluxes are @xmath2 0.08@xmath21 jy at 342 ghz and 0.04@xmath21
jy at 220 ghz ( 3 @xmath36 detection with the sma data in * ? ? ?
* with 1 @xmath36 @xmath2 0.013 and the uncertainty in flux scale @xmath2 20 % ) .
the sed of the compact source at those frequencies seems roughly consistent with @xmath35 , or an optically thick emission , as in hh 212 @xcite . in addition
, half of the emission at 43.3 ghz could be from the unresolved compact source . if the compact emission at 342 and 43.3 ghz are indeed arisen from the same region
, then the compact source has a size ( diameter ) of @xmath37 or 30 au @xcite .
therefore , the compact source probably has different origin and is likely to be a warm ( with a brightness temperature @xmath38 k ) optically thick ( accretion ) disk deeply embedded in the cold envelope .
observations at higher angular resolution are really needed to confirm this .
the total mass of the continuum emission ( including both the extended and compact components ) is estimated to be @xmath2 0.05 , assuming optically thin emission with a temperature of 30 k. since part of the emission is likely to be optically thick , the mass here is only a lower limit .
a knotty jet is seen in sio , as in @xcite , but with more knots resolved at higher angular resolution ( fig .
[ fig : jet]c ) .
it is bipolar with the blueshifted side in the southeast and the redshifted side in the northwest , but with the redshifted side brighter than the blueshifted side .
it consists of a chain of knots on each side of the source with an interknot spacing of @xmath2 @xmath4 or 600@xmath4900 au .
it extends out to @xmath2 away , with the innermost pair of knots at only @xmath2 or 535 au from the source .
most of the knots have 2 counterparts , except for those in the inner part .
the lack of 2 counterparts there was already seen by @xcite and attributed to the heavy dust extinction associated with the dense envelope as seen in @xcite and nh@xmath39 @xcite .
the eastern component and western component of the sio jet are not exactly antiparallel ( fig .
[ fig : jet]c ) .
they are misaligned slightly by @xmath2 1 , with their axes estimated to have a p.a . of 116.1@xmath40 and 297.1@xmath40 , respectively , by connecting the source to the sio knots .
since their original paths of motion are likely to be antiparallel , this misalignment suggests a presence of a jet bending . assuming both bent by the same degree , they are both bent by @xmath2 0.5 to the north with a jet ( or mean ) axis having a p.a .
of 116.6@xmath40 .
the peaks of the closest pair of the sio knots ( bk1 and rk1 ) are not exactly aligned with this jet axis , probably suggesting a slight precession of the jet . that the jet is slightly sinuous also supports this possibility .
the angle of precession is estimated to be @xmath3 1 .
the jet may also have a large - scale precession as discussed in @xcite and @xcite , with an angle of @xmath41 .
the axis of the sio jet , which is aligned with the 2 knots , is considered as the jet axis in the inner part out to 2 bow shocks bb1 and rb1 ( fig .
[ fig : jet]a ) . the jet axis in the outer part , which can be estimated by connecting the source to the tip of bow shock bb2 further out in the east , is found to be different by @xmath2 3 with a p.a . @xmath2
113.6 ( indicated by the green dashed line ) .
if we connect this axis to the west , the 2 emission rb2 could be the counterpart of bow shock bb2 but mainly seen with the southern wing .
co emission is detected not only along the jet axis but also toward outflow shells . in the following , two velocity ranges , high ( from @xmath423.7 to @xmath48.7 and from 10.3 to 35.3 ) and low ( from @xmath45.7 to @xmath41.7 and from 0.3 to 4.3 ,
respectively for the blueshifted emission and redshifted emission ) , are selected to show these components .
a knotty jet - like structure is also seen in co here at high velocity ( fig .
fig : jet]d ) , as in lower excitation line @xcite . however , this jet - like structure actually contains two components , a jet component with a similar velocity structure to the sio jet and a high - velocity shell component with a different velocity structure ( see next section ) . with the selected velocity range ,
the high - velocity co emission at @xmath2 away from the source is mainly from the jet component .
it is faint , associated with knots bk1 and rk1 seen in sio , but slightly downstream . at low velocity ,
rim - brightened shell - like structures are seen in co ( fig .
[ fig : coshell ] ) , coincident with those seen in the ir image , which are mainly from the continuum except near the bow shocks where the 2 emission dominates @xcite .
these shell - like structures were also seen in lower excitation lines of co and thought to trace the dense cavity walls of the outflow lobes produced by the 2 bow shocks located at the ends of the jet @xcite . in the eastern lobe , however , the southern and northern shells are seen associated with two different bow shocks , bb1 and bb2 , respectively .
thus , the shells likely trace the dense cavity walls recently shocked ( excited ) by the internal 2 bow shocks at different jet axes because of a large - scale jet precession ( see [ sec : jetaxis ] ) .
position - velocity ( pv ) diagrams of the sio and co emission cut along the jet axis are used to study the kinematics of the jet ( fig .
[ fig : pvjet ] ) .
note that the co emission seen at low velocity ( indicated by dashed lines ) likely traces the cavity walls along the jet axis . in the jet ,
the sio emission is localized toward the knots , with a range of velocities detached from the systemic velocity : knots bk1 , rk1 , rk2 , and rk5 are bright and seen with a velocity range of @xmath2 25 ( see also their spectra in fig . [
fig : specsio ] ) , while other knots are faint and seen with a narrower velocity range . at high velocity ,
co emission ( labeled co jet ) also arises from the jet , showing a similar velocity structure with similar mean velocity but with narrower velocity range ( see also fig . [
fig : specsio ] ) .
thus , the sio emission and the co jet emission can be used together to study the kinematics of the jet .
they are blueshifted in the east and redshifted in the west .
they are , however , more redshifted ( with a mean velocity of @xmath2 @xmath42 ) in the west than blueshifted ( with a mean velocity of @xmath2 @xmath43 ) in the east .
in addition , for each component of the jet , the velocity centroids of the individual knots are also different .
linear velocity structures ( indicated with solid lines ) are seen in co , as in lower excitation line @xcite , with the velocity magnitude increasing with the distance from the source .
these velocity structures are different from that of the sio jet .
to study their origin , we plot in figure [ fig : shell2 ] the emission of the linear velocity structure ( marked by an ellipse ) in the west , where it can be better separated from the jet emission .
the emission is seen around the jet axis but brighter in the north , with the transverse width increasing with the distance from the source , surrounded by the rim - brightened ir shell ( cavity wall ) . in the far end ,
the emission shows shell - like structures associated with the 2 knots . in the near end
, however , it is unclear but may show shell - like structures associated with the sio knots ( comparing fig .
[ fig : shell2 ] with fig .
[ fig : jet]c ) .
pv cuts across the jet axis in the west at and in the far end , where the structure is better resolved , also show a ring - like velocity structure as expected for a shell ( fig .
[ fig : pv_per ] ) .
note that , in these pv cuts , the co emission seen at low velocity is likely from the cavity wall .
therefore , the co emission associated with the linear velocity structures likely traces the ( internal ) shells or wakes driven by the ( internal ) bow shocks ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) , surrounded by the dense cavity walls . the co emission is assumed to have an excitation temperature of 100 k in the jet and 50 k in the high - velocity shells .
note that , however , the co emission , which was detected also in higher-@xmath44 transitions in the far infrared , could have a higher excitation temperature @xcite .
the 2 column density can be derived assuming a co abundance of 8.5@xmath45 @xcite and optically thin emission in local thermal equilibrium .
it is found to be @xmath46 in the jet , which has an intensity of 57@xmath4104 toward the knots ( see fig . [
fig : jet]d , integrated from -23.7 to -8.7 and 10.3 to 35.3 ) , and @xmath47 in the high - velocity shells , which have a mean intensity of 85 ( see fig .
[ fig : shell2 ] , integrated from 0.3 to 20.3 ) .
the jet is unresolved with a size ( diameter ) of @xmath3 and thus has a density of @xmath48 @xmath49 . assuming a shell thickness of @xmath2 , the high - velocity shells have a density of @xmath2 @xmath50 .
note that , however , it is not meaningful to derive the density for the low - velocity shells because their fluxes are mostly resolved out in our observations ( see also * ? ? ?
using the sio(@xmath51 ) observations from @xcite , the line ratio of sio(@xmath0)/sio(@xmath51 ) is found to be @xmath2 0.7 for the first pair of knots but decreasing to @xmath2 0.4 at the far ends of the sio jet ( fig .
[ fig : sioratio ] ) . note that this line ratio has been found to be @xmath2 1.0 for the first pair of knots but decreasing to @xmath2 0.5 at the far ends of the sio jet by @xcite at lower angular resolution .
however , with their maps , we recalculated the line ratio at lower angular resolution using the same velocity interval and intensity cutoff level and found it to be @xmath2 0.8 for the first pair of knots , and thus consistent with ours within the flux uncertainty .
the kinetic temperature of the sio emission can be assumed to be 300 k for the first pair of knots but 100 k for the knots further out @xcite .
the density can be estimated by comparing the line ratio to that in the lvg calculations with an assumption of optically thin emission @xcite .
note that , however , the emission could be optically thick @xcite .
the density is estimated to be @xmath52 for the first pair of knots but decreasing to @xmath53 at the far ends of the sio jet .
these densities , however , are a factor of @xmath2 2 lower than those derived from the line ratio of sio(@xmath51)/sio(@xmath54 ) in @xcite
. a detailed study with multiple transitions of sio will address this in a later paper ( hirano et al . in prep ) .
the sio knots are seen with a range of velocities , likely tracing the unresolved internal ( bow ) shocks ( i.e. , working surfaces ) in the jet , as in hh 212 @xcite .
it is thought that sio abundance is greatly enhanced in the shocks as a consequence of grain sputtering or grain - grain collisions releasing si - bearing material into the gas phase , which reacts rapidly with o - bearing species ( e.g. , o@xmath55 and oh ) to form sio @xcite .
the emission is consistent with its production in c - type shocks , with a velocity range of @xmath2 25 , similar to that found to produce the observed sio column densities in molecular outflows @xcite . at high velocity
, co is also seen tracing the shocks in the jet .
it has smaller velocity range than sio , tracing weaker shocks where the temperature and density are both lower .
observations at higher angular resolution are needed to study the morphological relationship between the co and sio emission by resolving their structures .
the jet inclination , @xmath56 , can be estimated from the mean velocity of the sio emission , @xmath57 , with @xmath58 , where @xmath59 is the jet velocity .
the eastern and western components of the jet may have different inclinations because of their different mean velocities . assuming a typical jet velocity of 100@xmath4200 , the eastern and western components of the jet have an inclination of @xmath48.6 to @xmath44.3 and 11.4 to 5.7 , respectively .
note that , however , it is also possible that the eastern and western components of the jet have the same inclinations but different jet velocities as in t tauri star jets @xcite .
the mass - loss rate of the jet can be estimated from the co emission of the knots .
as mentioned , the knots likely trace the internal ( bow ) shocks or the jet itself but highly compressed by shocks .
since the knots are unresolved , the compression factor is assumed to be @xmath2 3 , as found in hh 212 at similar angular resolution @xcite .
note that , however , the actual compression factor due to the shocks could be higher , if the knots are resolved .
thus , the ( two - sided ) mass - loss rate is given by @xmath60 where @xmath61 and @xmath62 are the co column density and the linear size of the synthesized beam perpendicular to the jet axis . with @xmath63 ( see [ sec : tcd ] ) , @xmath64 cm , and @xmath65 , @xmath66 yr@xmath6 , similar to that found in hh 212 .
the accretion rate can be estimated assuming that the bolometric luminosity @xmath67 is mainly from the accretion .
assuming a stellar mass of @xmath68 ( derived from an evolution model , * ? ? ?
* ) and a stellar radius of @xmath69 @xcite , then the accretion rate @xmath70 , with @xmath71 @xcite .
thus , the mass - loss rate is estimated to be @xmath72% of the accretion rate , similar to that found in hh 212 @xcite .
the jet is seen with a chain of knots in sio and co with a semiperiodic spacing of @xmath2 @xmath4 or 600@xmath4900 au .
the knots may trace the unresolved ( bow ) shocks resulting from a semiperiodic variation in the jet velocity ( see , e.g. , * ? ? ?
a temporal variation in the jet properties was also suggested in @xcite .
the period , which can be estimated by dividing the interknot spacing by the jet velocity , is found to be @xmath73 yr , assuming a jet velocity of 100@xmath4200 .
the jet is believed to be launched from an accretion disk around the source .
the periodic velocity variation may be due to ( 1 ) periodic perturbation of the accretion disk by an unresolved companion at a few au away or ( 2 ) magnetic cycle like the solar magnetic cycle , which has a period of @xmath2 22 yr @xcite .
the jet seems to have a small - scale ( @xmath3 1 ) precession with a sinuous structure , in addition to a large - scale ( @xmath2 3 ) precession suggested in @xcite and @xcite .
small - scale jet precession is also seen in other jets , e.g. , hh 34 and hh 212 , and is generally ascribed to the tidal effects of a companion star on the direction of the jet axis @xcite .
however , it may also be due to kink instability in the jet @xcite .
the jet is also seen bent in sio with its eastern and western components bent by @xmath2 0.5 to the north .
the jet was also seen bent in co but by 3 times larger @xcite due to contamination from high - velocity shells ( see [ sec : cojs ] & [ sec : kinematics ] ) .
the jet may have an additional bending into the plane of the sky because the eastern and western components of the jet may have different inclinations due to their different mean velocities ( see [ sec : dis_inc ] ) .
assuming their original paths of motion are antiparallel , the eastern and western components of the jet are both bent by @xmath2 [email protected] into the plane of the sky .
thus , the jet may have a total bending of @xmath2 [email protected] .
possible mechanisms for jet bending have been discussed in @xcite and @xcite .
one of the possibilities is due to motion of the jet source in a binary system .
the envelope , which is seen elongated toward the southwest , may result from an interaction with a companion in the southwest .
the jet is expected to be rotating , carrying away extra angular momentum from accretion disk .
the hh 211 jet , being close to the plane of the sky , is one of the best candidates to study the rotation . here ,
only the first four bright sio knots ( bk1 , bk3 , rk1 , and rk2 ) are used , because the sio knots further out are likely to be more affected by shock interactions .
a velocity gradient is seen across the first four knots ( see the solid lines in fig . [
fig : pvrotation ] ) , with @xmath2 [email protected] at @xmath2 30@xmath715 au
( i.e. , ) away from the jet axis
. the jet may have a radius ( i.e. , jet edge ) of @xmath2 30 au , similar to that of hh 212 , which is estimated to be @xmath2 45 au @xcite .
however , the beam dilution due to insufficient angular resolution could make the velocity shift smaller than it really is ( see , e.g. , * ? ? ?
* ) . this velocity gradient has the same direction as that seen in the rotating ammonia envelope ( wiseman , private communication ) , suggesting that it is from the jet rotation .
the velocity beyond @xmath71.5 seems increasing toward the jet axis , also as predicted in magneto - centrifugal wind models . however , the interpretation of the velocity gradient perpendicular to the jet axis is complicated by the presence of a velocity gradient along the jet axis ( fig . [
fig : maprotation ] ) . on the blueshifted ( east ) side , the higher - blueshifted emission ( blue contours ) is upstream of the lower - blueshifted emission ( red contours ) , while on the redshifted ( west ) side , the higher - redshifted emission ( red contours ) is upstream of the lower - redshifted emission ( blue contours ) .
this velocity gradient along the jet axis is consistent with the sio knots being formed by a velocity variation in the jet ( see [ sec : jetebp ] ) , with the faster jet material catching up with the slower jet material . without sufficient angular resolution ,
this velocity gradient can not be separated from the component perpendicular to the jet axis .
the velocity gradient seen across the knots rk1 and bk1 may even arise from the velocity gradient along the jet axis due to the beam position angle .
the jet also has a small - scale precession that may introduce a velocity asymmetry between the two jet edges ( see * ? ? ?
the increase in velocity toward the jet axis may also be due to the side - ways ejection of the shocked material .
therefore , further observations at higher angular resolution are really needed to confirm the jet rotation .
the launching radius of the jet can be estimated if the measured velocity gradient is indeed from the jet rotation . in magneto - centrifugal wind models ,
the jet can be considered as the dense part of the wind along the rotational axis . in the wind
, the specific angular momentum and energy can be assumed to be conserved along any given field line .
thus , for a given stellar mass , the specific angular momentum and poloidal velocity ( i.e. , jet velocity ) at large distance can be used to derive the angular velocity and thus the wind launching radius at the foot point of the field line in the accretion disk ( see , e.g. , * ? ? ?
* ) . the angular velocity at the foot point is found to be @xmath74 s@xmath6 or [email protected] day@xmath6 , using eq . 4 in @xcite , assuming a stellar mass of @xmath75 @xcite and a jet velocity of 100@xmath4200 .
thus , the wind launching radius is @xmath76 au or @xmath77 , with the stellar radius @xmath78 @xcite .
therefore , the jet could be launched either from near the corotation radius ( @xmath79 au ) as in the x - wind model @xcite or beyond in the inner edge ( @xmath80 au ) of a disk as in disk - wind model @xcite , depending on the jet velocity .
note that , if the jet is really launched from near the corotation radius , the rotation period of the protostar would be locked to be @xmath2 22 days .
a classical t tauri star has been found to have a typical rotation period of @xmath2 8 days ( see , e.g. , * ? ? ?
* ) , with a stellar mass of [email protected] .
the protostar here in hh 211 , with a mass of @xmath2 10 times smaller , could have a rotation period of @xmath2 3 times longer .
we have mapped the protostellar jet hh 211 in 342 ghz continuum , sio ( @xmath0 ) , and co ( @xmath1 ) emission .
thermal dust emission is seen in continuum at the center of the jet , tracing an envelope and a possible optically thick compact disk ( with a size @xmath3 130 au ) around the protostar .
a knotty jet is seen in co and sio as in 2 , but extending closer to the protostar .
it consists of a chain of knots on each side of the protostar , with an interknot spacing of @xmath2 @xmath4 or 600@xmath4900 au and the innermost pair of knots at only @xmath2 or 535 au from the protostar .
these knots likely trace unresolved internal ( bow ) shocks in the jet , with a velocity range up to @xmath2 25 .
the jet is episodic , precessing , and bending .
the two - sided mass - loss rate of the jet is estimated to be @xmath81 yr@xmath6 , about 9@xmath436% of the accretion rate .
a velocity gradient is seen consistently across two bright sio knots ( bk3 and rk2 ) perpendicular to the jet axis , with @xmath2 [email protected] at @xmath2 30@xmath715 au , suggesting a presence of a jet rotation .
the launching radius of the jet , derived from the potential jet rotation , is @xmath2 [email protected] au in the inner disk .
we thank the sma staff for their efforts in running and maintaining the array , and the anonymous referee for the insightful comments .
f . lee thanks frank h. shu , zhi - yun lee , and nagayoshi ohashi for fruitful conversations .
a. p. is grateful to robert estalella for helpful discussions .
anderson , j. m. , li , z .- y . , krasnopolsky , r. , & blandford , r. d. 2003 , , 590 , l107 avila , r. , rodrguez , l. f. , & curiel , s. 2001 , revista mexicana de astronomia y astrofisica , 37 , 201 beckwith , s. v. w. , sargent , a. i. , chini , r. s. , & guesten , r. 1990 , , 99 , 924 bouvier , j. , cabrit , s. , fernandez , m. , martin , e. l. , & matthews , j. m. 1993 , , 272 , 176 cabrit , s. , codella , c. , gueth , f. nisini , b. , gusdorf , a. , dougados , c. , & bacciotti , f. 2007 , , in press caselli , p. , hartquist , t. w. , & havnes , o. 1997 , , 322 , 296 cerqueira , a. h. , velzquez , p. f. , raga , a. c. , vasconcelos , m. j. , & de colle , f. 2006 , , 448 , 231 codella , c. , cabrit , s. , gueth , f. , cesaroni , r. , bacciotti , f. , lefloch , b. , & mccaughrean , m. j. 2007 , , 462 , l53 dent , w. r. f. , matthews , h. e. , & ward - thompson , d. 1998 , , 301 , 1049 eislffel , j. , froebrich , d. , stanke , t. , & mccaughrean , m. j. 2003 , , 595 , 259 enoch , m. l. , et al .
2006 , , 638 , 293 fendt , c. , & zinnecker , h. 1998 , , 334 , 750 frerking , m. a. , langer , w. d. , & wilson , r. w. 1982 , , 262 , 590 froebrich , d. 2005 , , 156 , 169 froebrich , d. , smith , m. d. , hodapp , k .- w .
, & eislffel , j. 2003 , , 346 , 163 giannini , t. , nisini , b. , & lorenzetti , d. 2001 , , 555 , 40 gueth , f. & guilloteau , s. 1999 , , 343 , 571 hirano , n. , liu , s .- y . , shang , h. , ho , p. t. p. , huang , h .- c . , kuan , y .- j . , mccaughrean , m. j. , & zhang , q. 2006 , , 636 , l141 ho , p. t. p. , moran , j. m. , & lo , k. y. 2004 , apj , 616 , l1 konigl , a. , & pudritz , r. e. 2000 , protostars and planets iv , 759 lada , c. j. , et al . 2006 , , 131 , 1574 lee , c .- f . , stone , j. m. , ostriker , e. c. , & mundy , l. g. 2001 , , 557 , 429 lee , c .- f .
, ho , p. t. p. , hirano , n. , beuther , h. , bourke , t. l. , zhang , q. , & shang , h. 2007 , , 659 , 499 lpez - martn , l. , cabrit , s. , & dougados , c. 2003 , , 405 , l1 mccaughrean , m. j. , rayner , j. t. , & zinnecker , h. 1994 , , 436 , l189 nisini , b. , codella , c. , giannini , t. , & richer , j. s. 2002 , , 395 , l25 palau , a. , ho , p. t. p. , zhang , q. , estalella , r. , hirano , n. , shang , h. , lee , c .- f . , bourke , t. l. , beuther , h. , & kuan , y .- j . 2006 , , 636 , l137 pesenti , n. , dougados , c. , cabrit , s. , ferreira , j. , casse , f. , garcia , p. , & obrien , d. 2004 , , 416 , l9 pudritz , r. e. , ouyed , r. , fendt , c. , & brandenburg , a. 2007 , protostars and planets v , 277 raga , a. & cabrit , s. 1993 , , 278 , 267 ray , t. , dougados , c. , bacciotti , f. , eislffel , j. , & chrysostomou , a. 2007 , protostars and planets v , 231 rebull , l. m. , et al .
2007 , arxiv astrophysics e - prints , arxiv : astro - ph/0701711 reipurth , b. , heathcote , s. , morse , j. , hartigan , p. , & bally , j. 2002 , , 123 , 362 schilke , p. , walmsley , c. m. , pineau des forets , g. , & flower , d. r. 1997 , , 321 , 293 schier , f. l. , jrgensen , j. k. , van dishoeck , e. f. , & blake , g. a. 2004 , , 418 , 185 shang , h. , li , z .- y . , & hirano , n. 2007 , protostars and planets v , 261 shu , f. h. , shang , h. , glassgold , a. e. , & lee , t. 1997 , science , 277 , 1475 shu , f.h .
, najita , j. , shang , h. , & li , z. -y .
2000 , in protostars and planets iv , ed . v. mannings , a. p. boss & s. s. russell ( tucson : university of arizona press ) , 789 stahler , s. w. , shu , f. h. , & taam , r. e. 1980 , , 241 , 637 suttner , g. , smith , m. d. , yorke , h. w. & zinnecker , h. 1997 , , 318 , 595 todo , y. , uchida , y. , sato , t. , & rosner , r. 1993 , , 403 , 164 wiseman , j. 2001 , science with the atacama large millimeter array , 235 , 179 | we have mapped the protostellar jet hh 211 in 342 ghz continuum , sio ( @xmath0 ) , and co ( @xmath1 ) emission at @xmath2 resolution with the submillimeter array ( sma ) .
thermal dust emission is seen in continuum at the center of the jet , tracing an envelope and a possible optically thick compact disk ( with a size @xmath3 130 au ) around the protostar .
a knotty jet is seen in co and sio as in 2 , but extending closer to the protostar .
it consists of a chain of knots on each side of the protostar , with an interknot spacing of @xmath2 @xmath4 or 600@xmath4900 au and the innermost pair of knots at only @xmath2 or 535 au from the protostar .
these knots likely trace unresolved internal ( bow ) shocks ( i.e. , working surfaces ) in the jet , with a velocity range up to @xmath2 25
. the two - sided mass - loss rate of the jet is estimated to be @xmath5 yr@xmath6 .
the jet is episodic , precessing , and bending .
a velocity gradient is seen consistently across two bright sio knots ( bk3 and rk2 ) perpendicular to the jet axis , with @xmath2 [email protected] at @xmath2 30@xmath715 au , suggesting a presence of a jet rotation .
the launching radius of the jet , derived from the potential jet rotation , is @xmath2 [email protected] au in the inner disk . |
hard partonic reactions in perturbative quantum chromodynamics ( pqcd ) lead ( to leading order in the strong coupling @xmath1 ) to highly virtual back - to - back parton configurations which subsequently evolve as parton showers , hadronize and are then experimentally observed as back - to - back jet events .
one of the most striking manifestations of the interaction of hard partons with a qcd medium as created in ultrarelativistic heavy - ion collisions ( urhic ) is the appearance of monojets in events in which one of the outgoing parton showers has lost such a high amount of energy into the medium that it is no longer identified by a jet finding algorithm .
such monojets have been observed in experiment indirectly via triggered two - particle correlations at rhic energies e.g. by the star collaboration @xcite , but only recently directly on the level of reconstructed jets by the atlas and cms collaborations at lhc energies @xcite .
the lhc experiments quantify the observation by focusing on the imbalance of the energy reconstructed for two jets in opposite hemispheres .
specifically , in the case of the atlas experiment , calorimeter towers are clustered within a jet cone radius of @xmath2 , a trigger condition on the energy @xmath3 of the most energetic jet of @xmath4 gev is imposed and for valid trigger the energy of the leading jet in the opposite hemisphere @xmath5 is measured with the additional condition @xmath6 gev imposed .
the asymmetry or dijet imbalance is then defined as @xmath7 where @xmath8 indicates a perfect energy balance between the two jets whereas @xmath9 indicates a true monojet event ( note that the condition @xmath6 gev prevents this situation in the atlas analysis ) .
a number of models which were able to account for the measured imbalance were subsequently proposed .
such models range from parametric estimates @xcite via schematic modelling of energy loss and jet finding @xcite to sophisticated nlo pqcd modelling of jet evolution in a medium , however without detailed medium or jet - finding modelling @xcite and complete monte carlo ( mc ) modelling of a jet embedded into a bulk medium with realistic jet finding using the martini code @xcite .
the success of these models quickly established that the measured dijet asymmetry does not pose a fundamental challenge to the radiative energy loss picture .
however , at least an equally interesting question is what the information content of the @xmath0 measurement is , i.e. what model properties are constrained by the observable . the aim of this work is to address this question by computing @xmath0 under a number of different assumptions for both the parton - medium interaction and the evolution of the bulk medium
a rigorous comparison of a jet quenching theory with e.g. the atlas measurement would in principle require the following steps : 1 ) generation of a bulk event using for instance hydrodynamics with event by event fluctuating initial conditions 2 ) generation of a hard event inside the same bulk event ( since there is an observable correlation between medium fluctuation hotspots and hard vertices @xcite , the hard and soft physics can not be modelled independently in a rigorous treatment ) and evolution of the perturbative shower inside the medium 3 ) computation of the back - reaction of the shower to the medium to account for correlated soft physics inside the jet area ( for instance elastic recoil of medium constituents ) 4 ) simulating the detector response to turn particle information into calorimeter tower information 5 ) clustering and analysis of the calorimeter - level event as in the experimental procedure .
the full set 1 ) - 5 ) corresponds to a numerically very demanding workflow , and consequently in the most ambitious theory comparison so far @xcite , only part of this has been included consistently using the bulk matter evolution code music without fluctuations and the in - medium shower evolution code martini @xcite . in the present work ,
we do not aim at a rigorous comparison with the data but rather at establishing the information content of an @xmath0 measurement , therefore the following simplifications are assumed : * the event by event fluctuations of the medium are ( for the purpose of embedding and propagating hard partons ) replaced by an initial - state averaged hydrodynamical model * no back - reaction of the medium to energy - deposition by the shower is included , all energy transferred to the medium is assumed to be outside the jet definition , i.e. lost * a simple set of cuts on hadron species and @xmath10 replaces the mapping of particles to calorimeter towers * instead of the anti @xmath11 algorithm clustering the whole event , a simple cone drawn around the primary parton direction with radius @xmath2 is used to define the jet , the ( by assumption ) smooth medium background can be subtracted by definition * medium ( and calorimeter ) fluctuations are then re - introduced as uncorrelated gaussian fluctuations of the medium background energy inside the cone area while these assumptions have been made to mimick the behaviour of the published atlas baseline simulation @xcite , the reader should be cautious to read anything more than qualitative trends into a comparison with data .
note however that qualitatively , the simplified jet definition contains the same elements as the real calorimetric jet problem , e.g. hadron - species dependence of the calorimeter response or a momentum cut introduced by the deflection of low @xmath10 charged hadrons by the magnetic field . inside this framework , we test three different scenarios for the parton - medium interaction and three different hydrodynamical models for the medium evolution . out of each model class , two of three model are known ruled out by some other set of data and are only used to test if @xmath0 is able to discriminate the elements in which these models fail .
since the various test cases are easily distinguished by other observables ( such as the single hadron suppression factor @xmath12 ) , it is important to set up the calculation in such a way that only the information provided by @xmath0 is probed . in particular ,
the dijet asymmetry is related to a conditional probability , i.e. given a trigger @xmath4 gev , what is the probability to find an away side jet with @xmath5 ?
we re specifically interested in the information content of this conditional probability , not in the probability to find a trigger jet per event ( which is related to the suppression factor @xmath13 of jets ) in the first place .
the dependence on the reduction of the trigger rate is removed as follows : the parameters of the only realistic model combination ( yajem - de in lhc hydrodynamics ) are tuned such that the single inclusive charged hadron suppression factor @xmath12 at 60 gev is described ( this implies typically partonic kinematics above 100 gev as relevant for the dijet observation ) .
the parameters of all other model combinations are then adjusted to create the same suppression of the jet @xmath14 gev trigger rate , i.e. all scenarios by definition show the same suppression of the single jet rate .
this procedure is very similar to what has been used to assess the information content of the dihadron correlation suppression factor @xmath15 in @xcite .
the modelling thus consists of the mc simulation of a hard partonic back - to - back event , the choice of a medium evolution and embedding of the hard event into the medium , the propagation of the evolving parton showers through the medium and the clustering and binning according to the jet definition . in lo
factorized pqcd , the production of two hard partons @xmath16 is described by @xmath17 where @xmath18 and @xmath19 stand for the colliding objects ( protons or nuclei ) and @xmath20 is the rapidity of parton @xmath21 .
the distribution function of a parton type @xmath22 in @xmath18 at a momentum fraction @xmath23 and a factorization scale @xmath24 is @xmath25 .
the distribution functions are different for free protons @xcite and nucleons in nuclei @xcite .
the fractional momenta of the colliding partons @xmath22 , @xmath26 are given by @xmath27 + \exp[\pm y_2 ] \right)$ ] .
expressions for the pqcd subprocesses @xmath28 as a function of the parton mandelstam variables @xmath29 and @xmath30 can be found e.g. in @xcite . to account for various effects , including higher order pqcd radiation , transverse motion of partons in the nucleon ( nuclear ) wave function and effectively also the fact that hadronization is not a collinear process , the distribution is commonly folded with an intrinsic transverse momentum @xmath11 with a gaussian distribution , thus creating a momentum imbalance between the two partons as @xmath31 .
these expressions are evaluated in a mc framework described in @xcite , resulting in an approximately back - to - back pair of hard partons with known parton types and momenta . for the modelling of the medium created in urhic at the lhc
, we use a 2 + 1d ideal hydrodynamical @xcite framework which has successfully been tested against bulk spectra and already been used as a background to compute single hadron suppression at lhc energies @xcite .
we contrast this with two hydrodynamical evolutions for rhic energies , a 3 + 1d code @xcite and a 2 + 1d code @xcite which do clearly _ not _ reproduce bulk observables at lhc correctly .
these frameworks differ substantially in the geometry of the simulated medium , for a comparison between 2 + 1d frameworks at rhic and lhc see e.g. @xcite , for a comparison between 2 + 1d and 3 + 1d at rhic see @xcite and are known to be discriminated by other high @xmath10 observables , for instance the reaction plane angle dependence of single hadron suppression @xmath32 @xcite or by hard dihadron correlations @xcite . for the actual embedding of parton showers into the medium , we assume that the distribution of vertices in the medium follows binary collision scaling as appropriate for a lo pqcd calculation .
thus , the probability density to find a vertex in the transverse plane is @xmath33 where the thickness function is given in terms of woods - saxon distributions of the the nuclear density @xmath34 as @xmath35 and @xmath36 is the standard nuclear overlap function @xmath37 for impact parameter @xmath38 .
we place the parton pair generated in the previous step at a probabilistically sampled vertex @xmath39 sampled from this distribution with a random orientation @xmath40 with respect to the reaction plane . when embedding parton showers into the medium , hydrodynamical parameters need to be related to transport coefficients which are probed by the shower . in the following ,
this is always done by assuming the relevant coefficients @xmath41 ( virtuality growth per unit pathlength ) or @xmath42 ( mean energy loss per unit pathlength ) are related to the energy density of the medium @xmath43 via @xmath44(\zeta ) = k[k_d ] \cdot 2 \cdot \epsilon(\zeta)^{3/4 } ( \cosh \rho(\zeta ) - \sinh \rho(\zeta ) \cos\psi).\ ] ] here @xmath45 is the position coordinate of a parton in the medium , @xmath46 is the transferse flow rapidity and @xmath47 the angle between parton propagation and flow direction ( the last factor contains a correction for relative motion between parton and fluid element , for instance it involves lorentz contraction ) .
the only parameters which are not fixed by the hydrodynamical medium are the adjustible constants @xmath48 ( or @xmath49 ) which dial the overall strength of the medium modification .
these are determined by a comparison to reference data or by the requirement of a constant jet rate suppression as described above .
the in - medium shower for each parton given the path through the medium determined in the previous step is computed using the mc code yajem @xcite .
yajem is a framework to compute medium - induced changes to the kinematics of the evolving parton shower where the medium is chacacterized in a phenomenological way by transport coefficients such as @xmath41 or @xmath42 .
the reliance on transport coefficients means that yajem does not resolve the medium microscopically .
it is based on the pyshow algorithm @xcite and has an option to simulate coherent medium - induced radiation as well as an incoherent drag force ( as characteristic e.g. for elastic collisions with the medium constituents ) . in the following , we use yajem to set up three scenarios : the default scenario ( yajem - de )
is characterized by medium - induced radiation with a @xmath50% component of elastic energy loss @xcite .
the minimum scale @xmath51 down to which the shower is evolved in the medium is set by formation time arguments as @xcite @xmath52 where @xmath53 is the parton energy and @xmath54 the in - medium length .
this creates a length and energy dependence of the medium interaction which is indicated by a large set of high @xmath10 observables @xcite . here
, the small elastic component is required by the low momentum strength of dihadron correlations which is overestimated without this contribution @xcite . in an alternative scenario ( yajem )
, we set the scale down to which the shower is evolved to a constant @xmath55 gev .
this changes pathlength and energy dependence and leads to disagreement with the high @xmath10 observables both at rhic @xcite and lhc @xcite . for a second alternative scenario ( yajem - e )
, we assume that the medium effect is only manifest via incoherent drag and has no induced radiation component , i.e. @xmath56 .
this changes observables like subjet fraction or jet shapes to a behaviour opposite to an induced radiation scenario @xcite while induced radiation _ widens _ a jet , a pure drag force scenario _ narrows _ the jet shape ( cf . rad vs. drag in @xcite ) .
the pathlength dependence of such a purely incoherent scenario is also incompatible with data @xcite . the yajem output in terms of hadron jets
is then clustered to compute @xmath57 , i.e. the probability distribution for a given parton @xmath22 with energy @xmath53 produced at a position @xmath58 in the transverse plane propagating at an angle @xmath40 to be registered as a jet of energy @xmath59 after passage through the medium .
@xmath59 is defined as the sum of the energy of hadrons fulfilling the following conditions : a ) at an angle of @xmath60 or less with the jet axis b ) the hadron species is one of @xmath61 c ) @xmath62 gev ( similar criteria have been suggested in the context of jet measurements by the star collaboration @xcite ) . to account for uncorrelated fluctuations of the medium background , a @xmath63 is sampled from a gaussian distribution centered around zero with width 20 gev and is added to the jet energies in the presence of a medium . in @xcite , it has been pointed out that uncorrelated background fluctuations have a potentially strong influence on the observed dijet asymmetry and hence need to be accounted for . in this way
, the near and away side jet energies @xmath3 and @xmath5 can be computed from the partonic event and compared with the experimental trigger conditions .
given a triggered event , @xmath0 is then be evaluated and the procedure is repeated with the generation of the next partonic event .
before looking at the outcome of the calculation , let us briefly review what we might reasonably expect to see by a comparison with dihadron correlations .
given the same suppression of the trigger rate , triggered dihadron correlations are sensitive to the pathlength dependence of energy loss @xcite and clearly able to rule out incoherent scenarios with linear pathlength dependence .
they also have tomographic capability , i.e. they are able to distinguish two different hydrodynamical background models @xcite . at the heart of these is a phenomenon commonly called _ surface bias _ the distribution of vertices leading to a triggered event is very different from the unbiased binary collision distribution eq .
( [ e - profile ] ) and does not peak in the medium center but towards the medium surface @xcite .
this in turn forces a longer than average path for the away side parton , and the precise balance between surface bias and pathlength dependence of energy loss is reflected in the strength of the dihadron correlation . in addition , the for low associate momentum bins , the dihadron correlation is able to trace the flow of energy to subleading particles and establish a lower bound on the rate with which energy is transferred into the bulk fluid medium @xcite . in principle , all these phenomena can also be realized in dijet correlations . even in the absence of a medium
, there is some imbalance between the jet energy seen on the near and away side .
the strength of the imbalance depends on the jet definition energy - momentum conservation ensures that the energy is balanced ( up to intrinsic @xmath11 ) when summing over the complete flow of energy and momentum resulting from the original parton evolution , and hence any imbalance is caused by part of the flow of energy and momentum going outside of the jet definition . as evident from fig .
[ f - base ] , this is not an academic discussion the result with angular cut only and without any cuts on @xmath10 or hadron species is quite different from the result for the full in - medium baseline where the whole set of cuts and fluctuations has been applied .
note again that while the jet definition has been chosen to resemble the atlas baseline simulation results , there is no reason to think that it really corresponds to the atlas detector response and clustering procedure , however the implication is that the probability distribution for losing energy out of the jet definition is similar .
the dependence of the distribution of @xmath0 on the assumed scenario for shower - medium interaction is shown in fig .
[ f - model ] . within statistical and systematic uncertainties ,
all three scenarios are indistinguishable and the width of the distibution qualitatively agrees with the atlas data . given that both the pathlength dependence and the energy - momentum flow inside the modified jet is very different in all three scenarios , this is a somewhat remarkable outcome .
the dijet asymmetry observable is neither strongly sensitive to the fact that yajem and yajem - e have a linear pathlength dependence which is incompatible with the angular dependence of single hadron suppression @xcite nor to the difference between widening and focusing of the jet shape @xcite . only for large asymmetries @xmath64
there is a hint of different model dynamics , indicating that a purely elastic energy loss scenario into the medium is somewhat less efficient in creating large asymmetries . the suppression of the trigger jet rate as compared to the vacuum case as found in these scenarios corresponds to a suppression factor of @xmath65 in the trigger momentum range ( i.e. for jets of 100 to about 150 gev ) , in line with atlas results for @xmath66 @xcite . in fig .
[ f - hydro ] the dependence on the assumed hydrodynamial modelling is shown .
within statistical uncertainties , there is no such dependence visible despite the obvious differences in geometry , evolution time and freeze - out conditions when going from lhc to rhic energies .
note however that by definition the differences in mean density between rhic and lhc are not probed here , since in all cases the suppression of the trigger rate is required to be a constant .
this implies that the quenching power scales ( @xmath48 and @xmath49 ) are adjusted to a much higher value for the same value of the energy density in the rhic hydrodynamical runs .
if the same relation between medium density and transport coefficients would be assumed in all cases , the results would be very different , but note that such an overall normalization factor is much easier probed in single hadron suppression .
the obvious similarity of the outcome leaves only the conclusion that @xmath0 is not sensitive to the medium geometry , i.e. the observable has no tomographic capability .
the dependence of @xmath0 on two other factors is illustrated by fig .
[ f - fluct ] .
if no fluctuations of the uncorrelated background medium inside the jet area are applied , then the shape of the @xmath0 distribution is found to be less broad .
similarly , if the medium transport coefficients are set to 50% of the values used in the previous investigations , the resulting shape is quite different .
however , note that both procedures alter the rate suppression of triggered events and are therefore not on par with the dependencies investigated in the previous sections .
these results demonstrate that the dijet asymmetry is sensitive to the overall strength of shower - medium interaction and that likewise that the proper treatment of the uncorrelated background fluctuations is important .
one of the experimental handles to control the amount of background fluctuations is the jet cone radius , and hence the systematics of the calculation with @xmath67 is an important crosscheck .
since the radius parameter controls the jet area and hence the total amount of background probed , the effect of uncorrelated background fluctuations should scale linearly with the radius parameter , i.e. clustering jets with @xmath68 should reduce background fluctuations by 50% of the strength seen at @xmath69 . in the following , we use this scaling assumption for the uncorrelated background .
however , while reducing the cone radius reduces background fluctuations , at the same time the probability for energy to fall outside the jet definition is increased . as a result ,
the fluctuations in clustered jet energy given a certain parton energy increase as compared to the case of @xmath60 ( in principle even in the absence of a medium , but certainly in a more pronounced way in a medium ) .
the net result of a variation in cone size depends on how well these two opposing effects cancel , since @xmath0 is in the end a measure of all jet energy fluctuations regardless of their origin .
the results using yajem - de in the lhc 2 + 1d hydrodynamical evolution are shown in fig .
[ f - radius ] .
to first approximation the result indicates that the decrease of background fluctuations with @xmath67 is compensated by the increase in fluctuations of jet energy inside the cone .
however , this cancellation is nt perfect and leads to small diffferences at low values of @xmath0 . for the same reason , also
the suppression of the trigger rate is almost unchanged between @xmath70 and @xmath68 .
for the smaller cone radius , the calculation yields @xmath71 for jets in the trigger range , i.e. from 100 gev to about 150 gev , in line with the atlas results @xcite .
note that this cancellation of background against jet energy fluctuations inside the cone can not occur in the absence of background fluctuations .
the fact that experimentally similar @xmath66 and @xmath0 is observed for both @xmath60 and @xmath68 is a strong indication that both types of fluctuations are present , strengthening the message of @xcite that uncorrelated background fluctuations need to be properly understood .
the apparent insensitivity of the dijet asymmetry observable @xmath0 to even gross changes in shower - medium interaction scenario or medium geometry is certainly at first glance more than somewhat surprising .
however , upon closer inspection , several points can be identified which may help to explain the findings .
jet definitions as used in jet finding algorithms are designed to make meaningful comparison between pqcd calculations valid at large momentum scales @xmath72 and observed hadron showers in terms of e.g. calorimetric measurements possible . for this purpose ,
jet definitions are created to systematically remove dependence to physics at low @xmath72 where pqcd starts breaking down , i.e. by construction jet definitions are relatively blind to enhanced soft gluon emission or the dynamics of hadronization .
however , the medium modifications of a shower take place precisely at scales of the medium temperature , which in itself is not far from the qcd fundamental scale , @xmath73 .
the implication is that vacuum jet finding is designed to minimize the sensitivity to physics at the medium scale , and as a result jets are fairly robust against final state interaction with the medium when compared to e.g. leading high @xmath10 hadrons .
the relative robustness against medium modification translates directly into a lack of surface bias .
this is illustrated by fig .
[ f - vertices ] which shows the probability distribution of the location of a vertex in the transverse plane given a triggered event . while the peak of such distributions is usually shifted off - center by @xmath74 fm in hadron - triggered correlations ,
no surface bias is visible in the case of dijet correlations . in other words , while the strength of medium modification and the in - medium pathlength are correlated in dijet as well as dihadron correlations , the strength of the correlation is much weaker due to the relative robustness of jets vs. single hadrons .
the presence of strong uncorrelated background fluctuations @xmath75 gev in the jet area ( with the typical scale of energy transported out of cone of the same order @xcite ) further dilutes the position - energy correlation and effectively removes any surface bias . as a result , the observable is insensitive to pathlength dependence or medium geometry .
while the suppression of single hadron production can be achieved by fairly collinear gluon emission , the suppression of a jet rate requires the emission of large angle gluons to shuffle energy out of the jet cone . applying this argument to medium - induced gluon production above a thermal momentum scale
explains the robustness mentioned above .
however , as pointed out in @xcite , this argument no longer applies to soft gluons in the thermal momentum range , as they are easily scattered to different angles .
this behaviour is fairly generic : whenever a model can produce soft gluons in the shower , these soft gluons are allowed to re - interact with the medium and there is energy and momentum exchange between shower and medium , the soft gluon angles becomes randomized no matter what dynamics is assumed in detail ( in yajem for instance this effect has been observed the distortion of low @xmath10 jet shapes long before @xmath0 was measured @xcite ) .
in essence , once a gluon in the thermal momentum range is embedded into a thermal medium and interacts with the medium , it is difficult to find any argument why the gluon should not thermalize within a very short time , and so soft gluon production generically equals energy transport out of cone .
we thus end up in a situation in which hard gluon emission is too rare to make an effect ( and not governed by shower - medium physics in any case since the momentum scales are too different ) , the dynamics of semi - soft gluon emission is largely erased by the clustering process into jets and the dynamics of soft gluons is generic and independent of specific model assumptions .
this explains why any model which is capable of suppressing the rate of triggered jets expects also a very similar dijet imbalance .
the above arguments suggest that the overall strength of the medium quenching power is the only jet - medium interaction related parameter constrained by the measured distribution of @xmath0 .
however , judging from figs .
[ f - base ] and [ f - fluct ] the outcome is very sensitive to both the precise cuts used to relate hadrons to calorimeter jets and to the assumed amount of background fluctuations ( the latter point has also been made in @xcite ) .
this leads to the surprising conclusion that in an approach which models the whole event , the correct modelling of the initial state fluctuations of the hydrodynamical medium and their evolution is more important than the modelling of shower - medium interactions .
the dependence on the way hadrons are related to calorimeter jets may then explain the somewhat different shapes of the @xmath0 as observed by various groups @xcite .
this should again be taken as a sign of caution not to read too much into the ( dis-)agreement of the results presented in this work ( and in other works ) with the data .
but the very observation that the observable is more sensitive to technical details of jet finding and treatment of fluctuations than to the physics it is supposed to measure is clearly a reason to worry .
following the reasoning presented in the previous section , it seems that observing the emission of soft gluons at the medium momentum scale is problematic in principle , as such gluons would thermalize under very general conditions .
a more meaningful question is then how the energy and momentum carried by such gluons is transported by the bulk medium , i.e. by hydrodynamical excitations .
the physics of medium modifications to a parton shower is then contained in semi - soft gluons with @xmath76 which are sufficiently hard not to be deflected to large angles by re - interactions with the medium . however , the physics at such momentum scales is best not observed via clustering into jets , as the clustering algorithms are designed to suppress sensitivity of observables to physics at these scales . as a consequence , triggerd hadron correlations have much greater tomorgraphic power .
it would thus appear that fully reconstructed jets do not make the best observables to do high @xmath10 tomography of the medium or to determine the precise nature of shower - medium interaction , and that approches to focus on this physics specifically by suitable designed @xmath77-particle correlation are more likely to answer such questions .
this work was supported by the academy researcher program of the finnish academy ( project 130472 ) .
discussions with jrn putschke , helen caines , jan rak , guangyou qin , berndt mller , jamie nagle , brian cole , matteo cacciari and bjrn schenke are gratefully acknowledged . | the appearance of monojets is among the most striking signature of jet quenching in the context of ultrarelativistic heavy - ion collisions .
experimentally , the disappearance of jets has been quantified by the atlas and cms collaborations in terms of the dijet asymmetry observable @xmath0 . while the experimental findings initially gave rise to claims that the measured @xmath0 would challenge the radiative energy loss paradigm ,
the results of a systematic investigation of @xmath0 in different models for the medium evolution and for the shower - medium interaction presented here suggest that the observed properties of @xmath0 arise fairly generically and independent of specific model assumptions for a large class of reasonable models .
this would imply that rather than posing a challenge to any particular model , the observable prompts the question what model dynamics is not compatible with the data . |
r0.5 in the ild collaboration the particle flow calorimetry is the mandatory technique which requires at least 10 mm @xmath010 mm lateral segmentation in the electromagnetic calorimeter ( ecal ) @xcite .
the scintillator - strip electromagnetic calorimeter ( scecal ) is one of the proposed concepts for such ecal . in the concept strip - shape scintillators in odd layers
are put orthogonally with respect to those even layers and forms about 30 layers of sampling calorimeter .
the tungsten layers are put between such sensitive layers as the absorber layers .
for example , a typical size of the scintillator strip in lateral dimension is 10 mm @xmath0 45 mm in order to achieve the effective 10 mm @xmath010 mm segmentation , which has been implemented in the prototype module and tested at fermi national accelerator laboratory ( fnal ) in 2008 and 2009 @xcite . to extract the best performance of such the strip segmentation detector
, reconstruction algorithm is being developed . including the development of such algorithm , status of the simulation tools for the study of the scecal performance
is reported in this manuscript .
resent studies suggest that 5 mm @xmath0 5 mm segmentation is better for ecal @xcite .
therefore , 5 mm width scintillator strips are also discussed in this study .
figure [ fig : geometry ] shows the geometry of the ecal in mokka @xcite which is the detector simulator for the ild . the scecal has the same global structure as the silicon tungsten ecal ( siecal ) @xcite which is another concept of ecal .
ecal barrel part is made from eight staves in azimuthal angle direction and a stave is made of five modules as shown in figure [ fig : module ] .
hr0.5 in the current plan
the scecal is designed having 27 common thick layers , while siecal has two different absorber thickness layers in 20 layers and 10 layers .
a pair of layers where the scintillator strips in each layer are mutually orthogonal to the ones in the other layer , constructs a detector unit called slab .
figure [ fig : slab ] shows a structure of slab .
these two scintillator layers are mounted on the both above and below sides of a `` h '' shape tungsten absorber , and they are sealed between the printed circuit boards ( pcbs ) .
besides , the copper plates cover them as the thermal radiators . with most outside shielding films ,
a salb is inserted into a column of shelf structure made of the tungsten plates which also form absorber layers as shown in figure [ fig : module ] .
the default thickness of the scintillator is 2.0 mm and tungsten absorber layers has 3.0 mm thickness .
the hits in ecal are generated as the hits on 5 mm @xmath0 5 mm square cells in mokka , and hits on the square cells then are merged into strip shapes in the digitizer of the marlin @xcite . to achieve the effective 5 mm @xmath0 5 mm segmentation with scintillator strips which have 5 mm width and some longer longitudinal length ,
the triplet method was developed by daniel jeans @xcite .
in the first step of the triplet method , the mip tracks are searched considering only cells with a hit consistent with the passage of a simple mip through the cell , and which have a limited number of nearest neighbors in the same layer .
the tpc tracks associated with these calorimeter tracks are considered as the tracks break into the calorimeter . using the remaining unassociated hits the two dimensional calorimeter clustering ( 2-d clustering )
is carried out .
the 2-d clustering is to make the clusters of neighbor strip in each layer .
when multiple energy peaks exist in a 2-d cluster , it is split into the clusters for the respective peaks .
figure [ fig:2dclustering ] shows a schematic diagram of the 2-d clustering .
these 2-d clusters are combined into `` triplets '' in the next step .
hr0.65 a triplet is made from the three clusters in successive three layers , which overlap .
figure [ fig : triplet ] left shows a schematic diagram to make a triplet , where the scintillator strips have longitudinal direction along the @xmath1 axis in @xmath2 layer and they are finely segmented along to the @xmath1 axis in @xmath3 layers .
the hit position in the longitudinal direction of scintillator strip in @xmath2 layer is calculated from the average of the energy weighted cluster position of above and below clusters . after making triplets for every successive three layers , triplets which have two clusters in common
are combined to make a `` calorimeter track '' explained in figure [ fig : triplet ] right .
this procedure is done from the inner surface of ecal .
when a combination of calorimeter tracks reduces a discrepancy between the tpc track momentum and calorimeter track energy , such calorimeter tracks are combined together . the position and deposited energy of the hits included in those calorimeter tracks are turned over to the pandorapfaprocessor which makes the particle flow approach .
figure [ fig : resolution ] shows the jet energy resolution as a function of the length of the scintillator strips performed before the triplet method is implemented .
the energy resolution gradually degrades as the length increases.the triplet method is expected to reduce those degradation .
one of the possible reason of the degradation of jet energy resolution is the two - fold ambiguity by the multi hits in a square which has four edges shaped with the longitudinal size of the scintillator strips shown in figure[fig : twofoldambiguity ] . when the tpc track - information works correctly , the fake calorimeter tracks are merged into the correct calorimeter tracks optimizing the matching between tpc track momentum and the energy of the calorimeter track .
one of the way to resolve this problem is to replace some layers with some fine segmented square cell layers .
figure [ fig : hybrid ] shows an example of such structure of ecal layers .
we will study these type of detectors in future .
the smallest unit of sensitive scintillator detector is made of a plastic scintillator strip , a pixelated photon detector ( ppd ) sit in the housing made on the one side of the scintillator strip , wave length shift ( wls ) fiber to collect the scintillation photons and a film reflector envelopes the scintillator .
an example of the structure of the detector unit was used in the tested prototype module at fnal @xcite .
figure [ fig:10x45strip ] shows the detector unit used in the prototype module . some of these materials make dead volume of detector .
these dead volume potentially degrade the performance of the detector . according to our simulation study for the prototype module tested in fnal , the effect from reflector film thickness
is small enough on both the constant term and the stochastic term of quadratic parameterization of the energy resolution .
although the effect from the thickness of ppd is also small enough , the stochastic term of the energy resolution increase if the ppd thickness increases more than current size . additionally , we have a plan to change the width of the scintillator strip to 5 mm for effective segmentation of 5 mm @xmath0 5 mm . in this case the dead volume increases .
therefore the realistic simulation including the dead volume mentioned above is required .
the scintillator size is shown in figure [ fig : newdetail ] . in order to decrease the dead volume , the smallest size of mppc produced by hamamatsu k.k .
will be used ; the size of its package is 2.0 mm @xmath0 2.5 mm @xmath0 0.9 mm . according to the study of position dependence of the sensitivity of 5 mm @xmath0 45 mm @xmath0 3 mm scintillator
, we do not need wls fiber to improve the uniformity of sensitivity for the 5 mm width scintillator strip .
although more detail study of the position dependence of sensitivity without wls fiber is ongoing , our plan is not to use the wls fiber in the next step .
ild software group has plan to implement these realistic simulation in mokka near future .
the goal of scecal group of ild is to achieve precise jet energy resolution using strip shape scintillators with the same performance as the one done with 5 mm @xmath0 5 mm square cell .
if it is achieved , the amount of photon sensors can be reduced . to achieve this
, we are developing the algorithm to extract the best performance from the scecal . to avoid the two - fold ambiguity , hybrid calorimeter of the square cell layers and the strip cell layers
are suggested .
the improvement of the geometry model in mokka is in progress to include those new configurations .
the authors would like to thank to ild group for many vital and important discussions .
the authors would also like to express thanks to the calice collaboration for providing the indispensable information of the experiment .
this work is supported in part by the creative science research grants no .
18gs0202 of the japan society for promotion of science ( jsps )
. 99 ild letter of intent , http://www.ilcild.org/ ( 2009 ) .
calice collaboration , `` _ first stage analysis of the energy response and resolution of the scintillator ecal in the beam test at fnal , 2008 _ '' , calice analysis note no .
16 ( 2010 ) .
d. grondin and d. jeans , `` _ next generation calice electromagnetic calorimeter _ '' , in _ proceedings of ilcws2010 _ ( 2010 ) .
d. jeans `` _ strip clustering in pfa _ '' international linear collider ecfa workshop + http://ecfa2008.fuw.edu.pl/ ( 2008 ) . | the scintillator - strip electromagnetic calorimeter ( scecal ) is one of the calorimeter technic for the ilc . to achieve the fine granularity from the strip - segmented layers the strips in odd layers are orthogonal with respect to those in the even layers . in order to extract the best performance from such detector concept ,
a special reconstruction method and simulation tools are being developed in ild collaboration .
this manuscript repots the status of developing of those tools . |
head - on beam beam compensation was first proposed as a four - beam e@xmath0e@xmath1e@xmath0e@xmath1 scheme for coppelia @xcite and implemented for dispositif de collisions dans ligloo ( dci ) @xcite .
the dci experience , however , fell short of expectations ; luminosities with two , three , or four beams were about the same .
the shortfall is generally attributed to coherent beam beam instabilities @xcite , and head - on beam beam compensation has not been tested again since . nevertheless , various proposals have been made , such as for the ssc @xcite , tevatron @xcite , lhc @xcite , and b - factories @xcite . in hadron colliders ,
the compensation can be achieved by colliding positively charged beams with a negatively charged low - energy electron beam , in a device usually referred to as an electron lens .
doing so avoids the coherent instabilities seen in dci , as the electron beam will not couple back to the hadron beam , except for single - pass effects ; these can be significant @xcite and may require the addition of a transverse damper in rhic .
two electron lenses were installed in the tevatron @xcite , where they were routinely used as a gap cleaner , but not for head - on beam beam compensation .
the tevatron experience is valuable for several reasons : ( i ) the reliability of the technology was demonstrated , as no store was ever lost because of the lenses @xcite ; ( ii ) the tune shift of selected bunches due to pacman effects was corrected , leading to lifetime improvements @xcite ; ( iii ) the sensitivity to positioning errors , transverse profile shape , and electron beam current fluctuations was explored @xcite ; ( iv ) experiments with a gaussian profile electron beam were performed ; and ( v ) a hollow electron beam was tested in a collimation scheme @xcite . for the design of the rhic electron lenses we have benefited greatly from the tevatron experience .
we have also drawn on the expertise gained in the construction and operation of an electron beam ion source ( ebis ) at brookhaven national laboratory ( bnl ) @xcite , which is a device similar to an electron lens but with a different purpose .
in rhic there are two head - on beam beam interactions at interaction points ip6 and ip8 ( fig .
[ fig : ov ] ) , as well as four long - range beam beam interactions with large separation ( about 10 mm ) between the beams at the other interaction points .
the luminosity is limited by the head - on effect in polarized proton operation @xcite , as can be seen in fig .
[ fig : bunch12 ] .
bunches with two collisions experience a larger proton loss throughout the store than bunches with only one collision .
the enhanced loss is particularly strong at the beginning of a store .
beam effects in other hadron colliders are reported in refs .
@xcite .
we consider the partial indirect compensation of the head - on beam beam effect with one electron lens in each ring . together with intensity and emittance upgrades @xcite , our goal is to approximately double the luminosity over what can be achieved without these upgrades .
this article gives a summary of previous studies and progress reports on head - on beam beam compensation in rhic with electron lenses @xcite , updated with the latest available information .
if a collision of a proton beam with another proton beam is followed by a collision with an electron beam , the head - on beam beam kick can in principle be reversed . for simplicity
we consider only the horizontal plane and beams with a gaussian transverse distribution .
figure [ fig : hobbc_bl ] shows the beam line layout for head - on compensation , and fig . [ fig : hobbc_ps ] shows the normalized phase space view .
, a proton experiences a beam beam kick from another proton bunch with intensity @xmath2 and root - mean - square beam size @xmath3 . at the second location , with lattice parameters
@xmath4 , another beam beam kick is generated by the electron beam with effective bunch intensity @xmath5 and root - mean - square beam size @xmath6.,width=302 ] before experiencing a beam beam kick from another proton beam at location 1 , a proton has transverse phase space coordinates @xmath7 .
the proton then receives a kick from the other proton beam @xcite , @xmath8,\ ] ] where @xmath2 is the bunch intensity of the second proton beam , @xmath9 is the relativistic factor of the proton receiving the kick , @xmath10 is the classical proton radius , and @xmath3 is the root - mean - square ( rms ) beam size of the second proton beam .
the new coordinates are then @xmath11 after transport through the linear beam line , the coordinates are @xmath12 with ( see @xcite ) @xmath13 where @xmath14 . in the electron lens
, the proton receives the kick @xmath15,\ ] ] where @xmath5 is the effective bunch intensity of the electron lens beam ( i.e. the number of electrons the proton passes in the lens ) and @xmath6 is the rms beam size of the electron lens beam . the coordinates after passing the electron lens are then @xmath16 one can now express the final coordinates @xmath17 as a function of the intensities @xmath18 and require , for exact compensation , that @xmath19 and @xmath20 i.e. that the final coordinates are the same with and without beam beam interaction and compensation . from the condition ( [ eq : c1 ] ) it follows that @xmath21 and hence @xmath22 , where @xmath23 is an integer . from the condition ( [ eq : c2 ] ) it follows that @xmath24 and @xmath25 .
therefore , if the following three conditions are met , the beam beam kicks are cancelled exactly . 1
. the ion beam and electron beam produce the same amplitude - dependent force by having the same effective charge and profile .
the phase advance between the two beam beam collisions is a multiple of @xmath26 in both transverse planes .
there are no nonlinearities between the two collisions . in practice
the above can be achieved only approximately .
deviations from condition 1 include : * an electron current that does not match the proton bunch intensity ; * a non - gaussian electron beam profile ( assuming that the proton beam transverse profile is gaussian ) ; * an electron beam size that differs from the proton beam size ; * time - dependence of the electron and proton beam parameters . deviations from condition 2 include : * a phase advance @xmath27 between the head - on collision and the electron lens ; * long bunches , i.e. @xmath28 .
deviations from condition 3 include : * lattice sextupoles and octupoles , as well as multipole error between the head - on collision and the electron lens .
tolerances were studied extensively in simulations and reported in ref .
@xcite , and bunch length effects have been investigated in refs .
the tevatron experience also provides tolerances for positioning errors , transverse shape and size mismatches , and electron current variations .
we give the tolerances for all devices below .
we plan to compensate for only one of the two head - on collisions in rhic , since a full compensation would lead to a small tune spread and could give rise to instabilities .
in designing the electron lens , we were aiming for a technically feasible implementation that would come as close as possible to the ideal compensation scheme outlined above .
in addition , a major design consideration was ease of commissioning and operation .
our goal is a commissioning that is largely parasitic to the rhic operation for physics .
the main design process can be summarized as follows . lcccc quantity & unit & + + total energy @xmath29 & gev & 100 & 255 & 255 + bunch intensity @xmath30 & @xmath31 & 2.5 & 2.5 & 3.0 + @xmath32 at ip6 , ip8 ( p p ) & m & 0.85 & 0.5 & 0.5 + @xmath32 at ip10 ( p e ) & m & 10.0 & 10.0 & 10.0 + lattice tunes @xmath33 & & + rms emittance @xmath34 , initial & mm@xmath35mrad & + rms beam size at ip6 , ip8 @xmath36 & @xmath37 m & 140 & 70 & 70 + rms beam size at ip10 @xmath36 & @xmath37 m & 485 & 310 & 310 + rms bunch length @xmath38 & m & 0.50 & 0.40 & 0.20 + hourglass factor @xmath39 , initial & & 0.88 & 0.85 & 0.93 + beam beam parameter @xmath40/ip & & 0.012 & 0.012 & 0.015 + number of beam beam ips & & + + distance of centre from ip & m & + effective length @xmath41 & m & + kinetic energy @xmath42 & kev & 7.8 & 7.8 & 9.3 + relativistic factor @xmath43 & & 0.18 & 0.18 & 0.19 + electron line density @xmath44 & @xmath45m@xmath46 & 1.0 & 1.0 & 1.2 + electrons in lens @xmath47 & @xmath31 & 2.1 & 2.1 & 2.5 + electrons encountered @xmath48 & @xmath31 & 2.5 & 2.5 & 3.0 + current @xmath49 & a & 0.85 & 0.85 & 1.10 + + * condition 1 * ( same amplitude - dependent forces from the proton beam and electron lens ) has a number of implications .
since both proton beams are round in the beam beam interactions ( @xmath50 and @xmath51 ) , we also require that @xmath52 at the electron lens location , and require matched transverse proton and electron beam profiles , i.e. that the electron beam profile is also gaussian with @xmath53 and @xmath54 .
the condition @xmath55 limits the electron lens locations to the space between the dx magnets ; in these locations the rhic lattice also has a small dispersion .
the tolerances for the main solenoid field straightness and for the relative beam alignment are easier to meet with a larger proton beam .
a larger beam is also less susceptible to coherent instabilities @xcite .
the @xmath56-function at ip10 can not be larger than 10 m at 250 gev proton energy without modifications to the buses and feedthroughs of the ir10 superconducting magnets .
such modifications are currently not under consideration because of costs , but could be implemented if coherent instabilities occur and can not be mitigated by other means . with a fully magnetized electron beam , the beam size @xmath57 in the main solenoid is given by its size at the cathode , @xmath58 , together with the solenoid fields @xmath59 at the cathode and @xmath60 in the main solenoid as @xmath61 .
for technological and cost reasons , the field @xmath60 can not be much larger than 6 t , and a strong field makes a correction of the field straightness more difficult .
the field @xmath59 has to be large enough to suppress space charge effects . with the limits in the @xmath59 and @xmath60 fields and a given beam size @xmath57 , the electron beam size and current density at the cathode follow , and they must be technically feasible . unlike the tevatron electron lenses
, we use a dc electron beam to avoid the noise possibly introduced through the high - voltage switches .
a dc beam requires the removal of ions created in the electron lens through residual gas ionization .
* condition 2 * ( phase advance of multiples of @xmath26 between p p and p
e interaction ) can be achieved through lattice modifications .
we have installed four phase - shifter power supplies for both transverse planes of both rings so that the betatron phase between ip8 and the electron lenses in ir10 can be adjusted . to have @xmath62 in both planes of both rings , it is also necessary to change the integer tunes from @xmath63 to @xmath64 in the blue ring and from @xmath63 to @xmath65 in the yellow ring to find a solution . with the new lattices ,
higher luminosities were reached in 2013 than in previous years , but the polarization was lower .
the lower polarization is still being investigated and may not have resulted from the new lattices .
other lattice options are also under study : ( i ) a solution was found for the yellow ring that maintains the integer tunes of @xmath63 and has the correct phase advances ; ( ii ) the phase advance of a multiple of @xmath26 may also be realized between ip6 and the electron lenses .
* condition 3 * ( no nonlinearities between the p p and p e interactions ) is best achieved when the p
e interaction is as close as possible to the p
p interaction . with the location in ir10 ( fig .
[ fig : ov ] ) , there is only one arc between the p
p interaction at ip8 and the p
e interaction at ir10 . in this configuration , a proton , after receiving a beam beam kick at ip8 , passes a triplet with nonlinear magnetic fields from field errors , an arc with chromaticity sextupoles and dodecapoles in the quadrupoles as dominating nonlinear field errors , and another triplet in ir10 . to avoid bunch length effects
, the parameter @xmath66 can not be too small @xcite . in simulations , a value as low as @xmath67 m was found to be acceptable @xcite .
the location of both the blue and the yellow electron lenses in ir10 , in a section common to both beams ( fig . [ fig:3 ] ) , allows local compensation of the main solenoid effect on both linear coupling and spin orientation by having the two main solenoids with opposing field orientations . at 255 gev proton energy ,
one superconducting solenoid with a 6 t field introduces coupling that leads to @xmath68 @xcite and increases all spin resonance strengths by 0.003 @xcite . in this configuration
it is also possible to ramp the magnets together during rhic stores without affecting the beam lifetime or spin orientation .
the instrumentation must allow for monitoring of the electron beam current and shape as well as the relative position and angle of the electron and proton beams in the electron lens .
two modes are foreseen : a setup mode in which the electron beam current is modulated and affects only a single bunch in rhic , and a compensation mode with a dc electron beam .
the main parameters of the electron beams are presented in table [ tab : ref ] .
a rhic electron lens consists of ( see fig . [
fig : elens ] ) an electron gun , an electron beam transport to the main solenoid , the superconducting main solenoid in which the interaction with the hadron beam occurs , an electron beam transport to the collector , an electron collector , and instrumentation . the electron gun ( see fig . [
fig : egun ] and table [ tab : egun ] ) @xcite has to provide a beam with a transverse profile that is close to gaussian .
considering the magnetic compression of the electron beam into the main solenoid centre with a maximum magnetic field of 6.0 t , a cathode radius of 4.1 mm gives a gaussian profile with 2.8 rms beam sizes .
the perveance of the gun is @xmath69 av@xmath70 .
the current density of the electron beam on its radial periphery can be changed with the control electrode voltage ( fig .
[ fig : egun ] , top ) , while the general shape of the beam profile remains gaussian .
the cathodes ( lb@xmath71 and irce ) were produced at binp in novosibirsk @xcite . with a nominal current density of 12 a/@xmath72
, irce was chosen as the cathode material for its long lifetime ( greater than 10000 h ) .
an assembled gun is shown in the bottom panel of fig .
[ fig : egun ] .
the gun has three operating modes : ( i ) dc for continuous compensation ; ( ii ) 100 hz for electron beam positioning with bpms , such that the electron current rises between the last two rhic bunches and falls in the abort gap ; ( iii ) 78 khz for single - bunch compensation , with rise and fall time as in the 100 hz mode .
the gun and collector vacuum is uhv compatible , with a design pressure of 10@xmath73 torr and a nominal pressure of 10@xmath74 torr for the interface to the rhic warm bore . for this reason ,
all of the components are bakeable to 250@xmath75c .
the gun and collector chambers will have a confined gas load by using a conductance - limiting aperture and enough installed pumping speed .
all vacuum chambers interfacing with the rhic warm bore will be made from stainless steel .
.main parameters of the thermionic electron gun . [
cols="<,^,^",options="header " , ] the instrumentation monitors the current and shape of the electron beam , the electron beam losses , and the overlap of the electron beam with the proton beam . the following items are included ( the quantities given in parentheses are for each lens ) : * dual - plane beam position monitors ( 2 ) ; * e p beam overlap monitor based on backscattered electrons ( 1 ) @xcite ; * differential current monitor ( 1 ) ; * beam loss monitor drift tubes ( 8) ; * collector temperature sensor ( 1 ) ; * profile monitor ( yag screen ) ( 1 ) ; * profile monitor ( pin - hole ) ( 1 ) ; * ion collector ( 1 ) .
the layout of the vacuum system with the drift tubes is shown in fig .
[ fig : dt ] . a total of eight drift tubes allow for changes in the electron beam energy and the removal of ions in the interaction region ; the split drift tube 4 enables the removal of backscattered electrons @xcite , which can be trapped with a low main field @xmath60 and high fringe fields .
figure [ fig : bpm ] shows the detail of a section containing a beam position monitor ( bpm ) , two drift tubes , cables , feedthroughs , and a heat sink to cool the cables , which can heat up when the proton beam deposits radio - frequency energy in the structure .
the bpms see only a signal with a pulsed beam .
the proton beams are bunched , and a fill pattern can be created so that a bunch in one beam is detected when there is a gap in the other beam . the electron beam needs to be pulsed ( at 100 hz or 80 khz ) to be visible .
the bpms are used to bring the electron and proton beams in close proximity .
the final alignment is done with the beam overlap monitor based on backscattered electrons @xcite .
alignment was found to be a critical parameter in the tevatron electron lenses , and the beams have to be aligned to within a fraction of the rms beam size , which can be as small as 310 @xmath37 m ( see table [ tab : ref ] ) .
figure [ fig : bs ] shows the beam overlap monitor .
the differential current monitor , drift tubes , ion collector , and collector temperature sensor all monitor the electron beam loss in the lens .
the yag screen and pin - hole profile monitors can only be used in a low - power mode .
the extracted ion current is monitored in a collector @xcite .
the test bench ( figs .
[ fig:10 ] and [ fig:2 ] ) uses the location and the superconducting solenoid of the bnl ebis test stand . of the rhic electron lenses ,
the following components were installed : a gun and collector , a gs1 solenoid with power supply , a movable pin - hole detector , a movable yag screen with camera , and an electron halo detector .
the test bench work is complete and the following have been demonstrated @xcite . *
the gun operated in 80 khz pulsed mode and dc mode , and reached 1 a of dc current with a current ripple of @xmath76% . * the gun perveance with a la@xmath71b cathode
was measured to be 0.93 @xmath37av@xmath77 . *
the collector temperature and pressure was measured with the 1 a dc current and found to be within expectations . *
the gaussian transverse electron beam profile was verified . *
the machine protection system was prototyped .
* part of the controls software was tested .
after completion of the test bench , the components were removed and installed in the rhic tunnel and service building .
for the ongoing rhic run-13 , the hardware of both lenses is partially installed ( fig . [ fig:3 ] ) . the blue lens has a complete electron beam transport system , although instead of the superconducting main solenoid designed for the electron lens a spare solenoid of the bnl ebis is installed .
this magnet is a 2m - long superconducting solenoid with a maximum field strength of 5 t , but it does not have an iron yoke and therefore the field lines are not straight enough for beam beam compensation ; it does , however , allow for propagation of the electrons from the gun to the collector even at field strengths as low as 1 t. the low field is necessary to minimize the effect on the proton spin , as long as the second superconducting solenoid is not yet powered .
the blue lens also has a full complement of instrumentation , with the exception of the overlap monitor based on backscattered electrons .
all drift tubes are grounded . in this configuration ,
all warm magnets can be commissioned as well as the electron beam in pulsed mode .
the two dual - plane bpms inside the superconducting solenoid , the yag screen profile monitor , and the pin - hole detector can be tested .
interaction with the proton beam is in principle possible .
the yellow lens has one of the new superconducting solenoids installed , but with a straight beam pipe that does not have bpms or drift tubes ( i.e. the vacuum system of the electron gun and collector is not connected to the proton beam vacuum system ) .
this configuration allows for commissioning of the superconducting main solenoid and all superconducting correctors , as well as all warm magnets . the yellow lens is shown in fig .
[ fig:9 ] .
the second superconducting solenoid is set up in the superconducting magnet division as a test bed for the field - straightness measurement system . as of the submission of this paper ,
the following have been achieved .
a new lattice was commissioned for both rings that has a phase advance of a multiple of @xmath26 between ip8 and the electron lens ; for this new phase shifter , power supplies were installed in both rings and both transverse planes .
a bunch - by - bunch loss monitor has become available , and bunch - by - bunch btf measurements are being tested . the derivation of the incoherent beam
beam tune spread in the presence of coherent modes from transverse btf measurements is under investigation @xcite . in the blue lens ,
a field of 1 t in the superconducting solenoid has been established .
all warm solenoids were tested at operating currents , and all gsb and csb solenoids ran concurrently with rhic polarized proton operation . in the summer of 2013 the second superconducting main solenoid will be installed , and the field straightness of both magnets will be measured in place and corrected .
after that , the installation will be completed for both lenses , including the overlap detector based on backscattered electrons . in 2014 rhic is likely to operate predominantly with heavy ions .
the beam beam effect with heavy ions is too small for compensation , but all electron beam operating modes ( pulsed and dc ) can be established , and the electron beam can interact with the ion beam . the first compensation test can be done in polarized proton operation .
partial head - on beam beam compensation is being implemented in rhic .
one of two beam beam interactions is to be compensated with two electron lenses , one for each of the two proton beams .
this allows for an increase in the bunch intensity with a new polarized proton source @xcite , with the goal of doubling the average luminosity in polarized proton operation .
the components of two electron lenses have been manufactured and partially installed .
the current installation allows for commissioning of the warm magnets , electron beam , and instrumentation in the blue lens . in the yellow lens , the new superconducting solenoid and the warm magnets
can be commissioned .
first tests with ion beams are anticipated for the next year , after which the compensation can be commissioned for polarized proton operation .
we are grateful to v. shiltsev , a. valishev , and g. stancari of fnal for many discussions on the tevatron electron lenses and for the opportunities to participate in electron lens studies at the tevatron .
we also greatly benefited from the experience of the bnl ebis team , including j. alessi , e. beebe , m. okamura , and d. raparia .
we had many fruitful conversations about beam beam and compensation problems with v. shiltsev , a. valishev , t. sen , and g. stancari of fnal ; x. buffat , r. demaria , u. dorda , w. herr , j .-
koutchouk , t. pieloni , f. schmidt , and f. zimmerman of cern ; k. ohmi of kek ; v. kamerdziev of fz jlich ; a. kabel of slac ; and p. grgen of tu darmstadt .
we are thankful to the u.s .
lhc accelerator research program ( larp ) for support of beam beam simulations .
we also acknowledge the technical and administrative support received from the bnl superconducting magnet division , as well as from all groups of the collider - accelerator department , in particular k. mirabella and g. ganetis .
derbenev , `` collective instability of compensated colliding beams , '' nuclear physics institute , siberian division , academy of sciences ussr , novosibirsk , report iyaf 70 - 72 , in russian ( 1972 ) ; slac - trans 151 in english .
u. dorda , f. zimmermann , w. fischer , and v. shiltsev , `` lhc beam beam compensation using wires and electron lenses , '' proc .
2007 particle accelerator conf . ,
albuquerque , new mexico , usa ; lhc - project - report-1023 ( 2007 ) .
v. shiltsev , y. alexahin , k. bishofberger , v. kamerdzhiev , g. kuznetsov , and x .-
zhang , `` experimental demonstration of colliding - beam - lifetime improvement by electron lenses , '' phys .
99 ( 2007 ) 244801 .
zhang , k. bishofberger , v. kamerdzhiev , v. lebedev , v. shiltsev , r. thurman - keup , and a. tollestrup , `` generation and diagnostics of uncaptured beam in the fermilab tevatron and its control by electron lenses , '' phys . rev .
beams 11 ( 2008 ) 051002 .
y. luo , m. bai , w. fischer , c. montag , and s. white , `` beam beam effects in rhic , '' proc .
52th icfa advanced beam dynamics workshop on high intensity high brightness hadron beams ( hb2012 ) , beijing , china ( 2012 ) .
white , w. fischer , and y. luo , `` coherent beam beam effects observation and mitigation at the rhic collider , '' proc .
particle accelerator conf . ,
new orleans , louisiana , usa , pp . 193195 ( 2012 ) .
y. luo , w. fischer , n. abreu , e. beebe , j. beebe - wang , c. montag , m. okamura , a. pikin , and g. robert - demolaize , `` head - on beam beam compensation with electron lenses in the relativistic heavy ion collider , '' proc .
2008 european particle accelerator conf .
, genoa , italy , pp . 16161618 ( 2008 ) .
y. luo , g. robert - demolaize , n. abreu , and w. fischer , `` multi - particle weak - strong simulation of head - on beam beam compensation in the rhic , '' proc .
2008 european particle accelerator conf .
, genoa , italy , pp . 31253127 ( 2008 ) .
abreu , w. fischer , y. luo , and g. robert - demolaize , `` the effect of head - on beam beam compensation on the stochastic boundaries and particle diffusion in rhic , '' proc .
2008 european particle accelerator conf .
, genoa , italy , pp . 25212523 ( 2008 ) .
w. fischer et al .
, `` long - range and head - on compensation studies in rhic with lessons for the lhc , '' proc .
final care - hhh workshop on scenarios for the lhc upgrade and fair ( hhh-2008 ) , chavannes - de - bogis , switzerland ; w. scandale and f. zimmermann ( eds . ) , cern-2009 - 004 ( 2009 ) ; care - conf-08 - 032-hhh ( 2008 ) . y. luo , n.p .
abreu , r. de maria , w. fischer , g. robert - demolaize , and e. mcintosh , `` weak - strong simulation of head - on beam beam compensation in the rhic , '' proc .
2009 particle accelerator conf . ,
vancouver , canada , pp .
9496 ( 2009 ) .
y. luo , w. fischer , x. gu , and a.i .
pikin , `` optimizing the electron beam parameters for head - on beam beam compensation in rhic , '' proc .
2011 particle accelerator conf . , new york , usa , pp .
7072 ( 2011 ) .
x. gu , w. fischer , r.c .
gupta , j. hock , y. luo , m. okamura , a.i .
pikin , and d. raparia , `` designing a beam transport system for rhic s electron lens , '' proc .
2011 particle accelerator conf . , new york , usa , pp .
12051207 ( 2011 ) .
x. gu , w. fischer , r.c .
gupta , j. hock , y. luo , m. okamura , a.i .
pikin , and d. raparia , `` the effects of the rhic e - lenses magnetic structure layout on the proton beam trajectory , '' proc .
2011 particle accelerator conf . , new york , usa , pp .
12021204 ( 2011 ) .
p. thieberger et al . , `` proposed electron halo detector system as one of the beam overlap diagnostic tools for the new rhic electron lens , '' proc .
2011 particle accelerator conf .
, new york , usa , pp .
489491 ( 2011 ) . c. montag , m. bai , k.a .
drees , w. fischer , a. marusic , and g. wang , `` beam experiments related to the head - on beam beam compensation project at rhic , '' proc .
2011 particle accelerator conf .
, new york , usa , pp .
22432245 ( 2011 ) .
p. thieberger et al . , `` design of a proton - electron beam overlap monitor for the new rhic electron lens , based on detecting energetic backscattered electrons , '' proc .
beam instrumentation workshop biw12 , newport news , virginia , usa ( 2012 ) .
x. gu , w. fischer , d.m .
gassner , k. hamdi , j. hock , y. luo , c. montag , m. okamura , a.i .
pikin , and p. thieberger , `` a split - electrode for clearing scattered electrons in the rhic e - lens , '' proc .
particle accelerator conf .
, new orleans , louisiana , usa , pp . 40384040 ( 2012 ) .
y. luo , w. fischer , n.p .
abreu , a. pikin , and g. robert - demolaize , `` six - dimensional weak - strong simulation of head - on beam beam compensation in the relativistic heavy ion collider , '' phys .
st accel . beams 15 ( 2012 ) 051004 .
white , w. fischer , and y. luo , `` simulations of coherent beam beam effects with head - on compensation , '' proc .
particle accelerator conf . ,
new orleans , louisiana , usa , pp . 187189 ( 2012 ) .
gassner , e. beebe , w. fischer , x. gu , k. hamdi , j. hock , c. liu , t.a .
miller , a. pikin , and p. thieberger , `` rhic electron lens test bench diagnostics , '' proc .
dipac2011 , hamburg , germany , pp . 3840 ( 2011 ) .
a. jain , `` plans for measurement of field straightness in the solenoids for the electron lens system for rhic , '' presentation at 17th international magnetic measurement workshop ( immw17 ) , barcelona , spain ( 2011 ) . | in polarized proton operation , the performance of the relativistic heavy ion collider ( rhic ) is limited by the head - on beam beam effect . to overcome this limitation ,
two electron lenses are under commissioning .
we give an overview of head - on beam beam compensation in general and in the specific design for rhic , which is based on electron lenses .
the status of installation and commissioning are presented along with plans for the future . |
observations @xcite indicate that approximately 70% of the energy density in the universe is in the form of an exotic , negative - pressure component , dubbed dark energy .
( see ref .
@xcite for a recent review ) .
taking @xmath2 to be the ratio of pressure to density for the dark energy : @xmath9 recent observations suggest that @xmath2 is close to @xmath3 . for example , if @xmath2 is assumed to be constant , then @xmath10 @xcite .
if we are interested in dynamical models for dark energy , such as those that arise from a scalar field @xmath11 , then such models can be significantly simplified in the limit that @xmath2 is close to @xmath3 .
this fact was exploited in ref .
@xcite , which examined quintessence models with a nearly flat potential , defined as a @xmath12 satisfying the slow - roll conditions : @xmath13 and @xmath14 with these assumptions , it is possible to derive a generic expression for @xmath2 as a function of the scale factor , @xmath15 , that provides an excellent approximation to this entire class of potentials .
( for other approaches to this problem , see refs .
@xcite ) . here
we extend these results to the case of phantom models , i.e. , models for which @xmath16 .
it was first noted by caldwell @xcite that observational data do not rule out the possibility that @xmath17 .
these phantom dark energy models have several interesting properties . the density of the dark energy increases with increasing scale factor , and the phantom energy density can become infinite at a finite time , a condition known as the big rip " @xcite .
further , it has been suggested that the finite lifetime for the universe which is exhibited in these models may provide an explanation for the apparent coincidence between the current values of the matter density and the dark energy density @xcite .
( see ref .
@xcite for a related , but different argument ) .
a simple way to achieve a phantom model is to take a scalar field lagrangian with a negative kinetic term .
such models have well - known problems @xcite , but they nonetheless provide an interesting set of representative phantom models , and they have been widely studied @xcite . here we assume a negative kinetic term , and then use techniques similar to those in ref .
@xcite to derive an expression for @xmath4 for phantom models satisfying conditions ( [ slow1 ] ) and ( [ slow2 ] ) .
we assume a spatially - flat universe containing only nonrelativistic matter and phantom dark energy , since radiation can be neglected in the epoch of interest .
then [ fried ] h^2 = _ t/3 , where @xmath18 , @xmath15 is the scale factor , @xmath19 is the total energy density , and we take @xmath20 throughout . in a phantom model with negative kinetic term and potential @xmath12 , the energy density and pressure of the phantom are given by _
= -(1/2)^2 + v
( ) , and p_=-(1/2)^2 - v ( ) , so that the equation of state parameter is @xmath21 this equation of motion can be rewritten as @xcite @xmath23,\ ] ] where @xmath5 is the density of the phantom field in units of the critical density ( note that @xmath5 evolves with time ) . in equation ( [ vprime / v ] ) , @xmath24 , so that @xmath25 and @xmath2 are related via @xmath26 ( note that in equation ( [ vprime / v ] ) we use the sign convention in @xcite , rather than the one used in @xcite ) . equation ( [ vprime / v ] ) is the phantom version of the quintessence equation of motion first derived in ref .
@xcite ; it differs from the quintessence equation only in the sign of @xmath27 on the right - hand side .
we are interested in the limit where @xmath2 is near @xmath3 , so following ref .
@xcite , we define @xmath28 , where we take @xmath29 to be small and positive
. this will allow us to drop terms of higher order in @xmath29 .
then equation ( [ vprime / v ] ) becomes @xmath30 where we have defined @xmath31 .
the fractional density in phantom dark energy , @xmath5 , evolves as @xmath32 combining equations ( [ betaprime ] ) and ( [ omegaprime ] ) yields @xmath33 this can be compared to the corresponding equation for @xmath34 for quintessence , where @xmath35 @xcite : @xmath36 clearly , equations ( [ exact1 ] ) and ( [ exact2 ] ) predict that @xmath37 for a phantom field and @xmath38 for a quintessence field will evolve quite differently .
now , however , we make our slow - roll assumptions . as in ref .
@xcite , we take @xmath39 and we assume @xmath40 to be a constant , @xmath41 , given by the initial value of @xmath42 . both of these results are a consequence of equations ( [ slow1 ] ) and ( [ slow2 ] ) ; see ref .
@xcite for the details
. then equation ( [ exact1 ] ) becomes @xmath43 this is identical to the equation one obtains , in the corresponding limit , for @xmath38 for quintessence @xcite .
thus , in this limit , @xmath44 for quintessence and @xmath45 for a phantom field evolve in exactly the same way . using the exact solution for equation ( [ step2 ] ) from ref .
@xcite , we obtain @xmath46 ^ 2.\ ] ] further , when @xmath27 is close to @xmath3 , we can approximate the solution to equation ( [ omegaprime ] ) as @xcite @xmath47^{-1},\ ] ] where @xmath48 is the present - day value of @xmath5 , and we take @xmath49 at the present . combining equations ( [ wom ] ) and ( [ oma ] ) , and normalizing to @xmath50 at present , we obtain : @xmath51 ^ 2\nonumber\\ \label{wpred } \times \left[\frac{1}{\sqrt{\omega_{\phi 0 } } } - \left(\frac{1}{\omega_{\phi 0 } } - 1 \right ) \tanh^{-1}\sqrt{\omega_{\phi 0}}\right]^{-2}.\end{aligned}\ ] ] this equation is identical to the corresponding result for @xmath27 for the case of quintessence @xcite .
the fact that these two different models ( phantom and quintessence ) yield an identical form for @xmath52 is not _ a priori _ obvious ; indeed , this result is valid only in the slow roll " limit considered here .
we now compare the approximation given in equation ( [ wpred ] ) to exact numerical results .
kujat et al .
@xcite showed that phantom potentials with a negative kinetic term can be divided into three broad classes , depending on the late - time asymptotic behavior of @xmath53 .
when @xmath54 at late times , @xmath55 .
these models include , for instance , negative power laws . for models in which @xmath42 asymptotically approaches a constant
, @xmath2 also approaches a constant , @xmath56 , where @xmath57 .
this class of models includes exponential potentials .
finally , when @xmath58 at late times , @xmath59 .
this class of models includes , for example , positive power law potentials .
this final class of models , for which @xmath60 , can be further subdivided into models which yield a future singularity ( a big rip " ) , and those which do not .
the exact conditions necessary to avoid a big rip involve an integral function of @xmath12 @xcite , but for power - law potentials , the condition is simpler : for @xmath61 , @xmath62 yields a big rip , while @xmath63 yields no future singularity @xcite .
we have therefore chosen to compare our analytic approximation for @xmath4 with a model from each of these four classes .
the results are displayed in fig .
1 . as in the case of quintessence , the agreement between the actual evolution and our approximation in equation ( [ wpred ] ) is remarkably good , with the error in @xmath2 less than 0.5% .
our analytic approximation is well - fit by a linear relation , @xcite @xmath64 in the regime @xmath65 , which corresponds to a redshift @xmath66 .
further , for the value of @xmath67 adopted here ( @xmath68 ) , we have @xmath8 .
this is the same relation between @xmath69 and @xmath56 noted empirically for linear quintessence potentials @xcite , predicted for generic slow - roll quintessence models @xcite , and derived analytically for linear quintessence potentials @xcite .
( note that the latter reference gives a value of @xmath70 rather than @xmath71 , but this difference is not significant , as the result depends on the assumed value of @xmath48 and on the assumed definition of @xmath69 ; the latter is ambiguous since @xmath4 is not exactly linear in @xmath15 in these models ) .
given that equation ( [ wpred ] ) provides a generic prediction for the behavior of @xmath4 in a wide class of phantom models , it is useful to compare this prediction to the observations . further , since both quintessence and phantom models produce the same generic form for @xmath4 in the limit where @xmath42 is small and nearly constant , we can combine the results for @xmath4 for quintessence models from ref .
@xcite with the results for @xmath4 for phantom models given here into a single likelihood plot . in fig .
2 , we compare this generic prediction for @xmath4 to the snia and baryon acoustic oscillation data .
the likelihoods were constructed using the 60 essence supernovae , 57 snls ( supernova legacy survey ) and 45 nearby supernovae , and the data release of 30 sne ia detected by hst and classified as the gold sample by riess et al . @xcite .
the combined dataset can be found in ref .
we also add the measurement of the baryon acoustic oscillation ( bao ) scale at @xmath72 as observed by the sloan digital sky survey @xcite .
it is important to note that fig .
2 combines two completely different models .
the portion of the figure with @xmath73 corresponds to quintessence models , while @xmath57 corresponds to phantom models .
we have simply taken advantage of the fact that both models yield the same form for @xmath4 as a function of @xmath56 and @xmath67 in the limiting case where the potential is nearly flat .
our main conclusion is that the analytic approximation previous derived for quintessence with a nearly flat potential @xcite has the same form when it is extended to phantom models with a negative kinetic term .
thus , the generic prediction for the evolution @xmath2 at low redshift ( @xmath6 ) in such models , @xmath74 , can be applied to both @xmath73 and @xmath57 .
r.j.s . was supported in part by the department of energy ( de - fg05 - 85er40226 ) .
a.a.s acknowledges the financial support from the university grants commission , govt .
of india through major research project grant ( project no : 33 - 28/2007(sr ) ) . | we examine phantom dark energy models produced by a field with a negative kinetic term and a potential that satisfies the slow roll conditions : @xmath0 ^ 2 \ll 1 $ ] and @xmath1 .
such models provide a natural mechanism to produce an equation of state parameter , @xmath2 , slightly less than @xmath3 at present . using techniques previously applied to quintessence ,
we show that in this limit , all such phantom models converge to a single expression for @xmath4 , which is a function only of the present - day values of @xmath5 and @xmath2 .
this expression is identical to the corresponding behavior of @xmath4 for quintessence models in the same limit . at redshifts
@xmath6 , this limiting behavior is well fit by the linear parametrization , @xmath7 , with @xmath8 . |
the cosmic microwave background ( cmb ) offers a picture of the early universe when it was only 400,000 years old .
the cmb photons last scattered off electrons at that time and since then they traveled free through the space .
the primordial perturbations set up during inflation are imprinted in both radiation and matter distribution .
the cmb temperature anisotropies , related with the primordial perturbations , can be used to test some assumptions of the so called standard model . in particular
, we can test the prediction of the standard , single field , slow roll inflation @xcite which states that the anisotropies are compatible with a nearly gaussian random field .
other non - standard models of inflation predict detectable levels of non - gaussianity in the anisotropies ( see e.g. * ? ? ?
primordial non - gaussianity of the local form is characterised by the non - linear coupling parameter @xmath0 @xcite : @xmath5 where @xmath6 is the primordial gravitational potential and @xmath7 is a linear random field which is gaussian distributed and has zero mean .
there are many studies based on different statistical tools to constrain the local @xmath0 parameter from the cmb anisotropies , using the data of different experiments .
we can mention the analyses using the angular bispectrum and wavelets on the cosmic background explorer ( cobe ) data @xcite , the angular bispectrum on maxima data @xcite , the angular bispectrum on wmap data @xcite , different kind of wavelet analyses on wmap data @xcite , the minkowski functionals on boomerang data @xcite , the minkowski functionals on archeops data @xcite , the minkowski functionals on wmap data @xcite among others
. we can also mention new promising techniques as for example one based on the @xmath8-point probability density distribution @xcite , and other based on needlets @xcite .
other works are based on the use of the large scale structure to constrain @xmath0 @xcite .
this work is a continuation of the wavelet - based analysis by @xcite of the wmap data .
we use high resolution wmap data maps , and compute the wavelet coefficients for 12 angular scales logarithmically spaced from 6.9 arcmin to 500 arcmin . with these wavelet coefficients
we compute all the possible third order moments involving these scales .
the article is organised as follows .
section [ method ] presents the estimators used to test gaussianity and to constrain @xmath0 , the data maps and the simulations .
section [ results ] summarises the main results of this work and the conclusions are in section [ conclusions ] .
this analysis is based on the spherical mexican hat wavelet ( smhw ) as defined in @xcite . for references about the use of the smhw to test gaussianity in the cmb see for example the review by @xcite .
we compute the wavelet coefficient maps at several scales @xmath9 logarithmically separated ( @xmath10 constant ) .
the considered scales are : @xmath11 , @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 .
we also include the unconvolved map , which will be represented by the scale @xmath22 as in @xcite . for each possible combination of three scales @xmath9 , @xmath23 , and @xmath24 ( where the indices @xmath25 , @xmath26 , and @xmath27 can be repeated ) we define a third order statistic @xmath28 where @xmath29 is the total number of pixels of the map , @xmath30 is the number of pixels available after combining the extended masks corresponding to the three scales @xmath9 , @xmath23 and @xmath24 , @xmath31 is the wavelet coefficient in the pixel @xmath32 evaluated at the scale @xmath9 , and @xmath33 is the dispersion of @xmath34 .
each map @xmath34 is masked out with the corresponding extended mask at the scale @xmath9 as in @xcite .
for a set of @xmath8 scales we have @xmath35 $ ] third order statistics such as the one defined in eq .
( [ statistic ] ) .
we have tested with simulations that these statistics have a gaussian - like distribution .
we can construct a vector @xmath36 of dimension @xmath37 @xmath38.\ ] ] with this vector we can perform two different analyses using a @xmath39 statistic : one to test gaussianity and a second one to constrain @xmath0 ( eqs . ( 7 ) and ( 8) of * ? ? ?
we use the 5-yr wmap foreground reduced data , available in the legacy archive for microwave background data analysis ( lambda ) web site .
we combine the maps of different radiometers using the inverse of the noise variance as an optimal weight @xcite .
in particular we analyse the v+w , q , v and w combined maps at a resolution of 6.9 arcmin , corresponding to a healpix @xcite @xmath40 .
we use the @xmath4175 mask and also a set of extended masks for the wavelet coefficient maps .
we use the same masks as the ones described in @xcite for a threshold of 0.01 .
this corresponds to an available fraction of the sky from 71.2% for the @xmath42 scale to 31.4% for the @xmath43 scale .
notice that larger scales have the restriction of a lower available area , which means a lower sensitivity to @xmath0 .
finally we analyse the data and compare them with gaussian and non - gaussian simulations .
the gaussian simulations are performed using the best fit power spectrum @xmath44 for wmap provided by lambda and the instrumental white noise of each wmap radiometer .
the non - gaussian simulations with the @xmath0 contribution are computed following the algorithms described in @xcite and transformed into wmap maps with the instrumental noise included .
we also estimate the unresolved point source contribution to @xmath0 for the v+w case by analysing point source simulations .
these simulations have been generated as in @xcite following the source number counts @xmath45 given by @xcite .
in this section we present the gaussianity analysis of the wmap data for the combined v+w , q , v , and w data maps .
we also constrain @xmath0 for these maps through a @xmath39 test .
we estimate the contribution of point sources to the v+w map . finally we constrain @xmath0 for northern and southern pixels separately .
.@xmath39 constructed from the 364 statistics for v+w , q , v , and w data maps .
[ stats_vw ] we also present the mean and the dispersion of the @xmath39 corresponding to 1,000 gaussian simulations , and the cumulative probability for the @xmath39 of the data obtained from the simulations . [ cols="^,^,^,^,^,^",options="header " , ]
we have tested the gaussianity and constrained the @xmath0 parameter with the 5-yr wmap data .
we use an optimal wavelet - based test .
we have considered the v+w , q , v and w combined maps at high resolution .
we have used a set of 300 realistic non - gaussian simulations and thousands of gaussian simulations for the analysis .
we have computed the wavelet coefficient maps at scales from 6.9 arcmin to 500 arcmin and computed all the possible third order moments ( eq . [ statistic ] ) using appropriate extended masks .
the data are compatible with gaussian simulations for the considered combined maps ( see table [ stats_vw ] ) .
we have imposed constraints on the non - linear coupling parameter @xmath0 by using non - gaussian simulations with @xmath0 .
the results show that @xmath0 increases when we go from the q to the v and w combined maps .
this frequency dependence also appears in the results by @xcite .
the results are compatible with zero at 95% cl for the v+w , q , and v combined maps , but not for the w map ( which is compatible at 99% cl ) .
this value can not be explained by unresolved point sources since their contribution is @xmath46 for the w map .
we have estimated the probability of having those values with simulations and the results do not show incompatibility among different channels .
we have also seen that the relatively large @xmath0 value obtained for the w band may come from systematics in one or several radiometers of this band .
we have also estimated the contribution of unresolved point sources to @xmath0 for the v+w map using a realistic model given by @xcite .
the results are @xmath47 .
taking into account this value , our best estimate for @xmath0 is @xmath48 at 95% cl .
the use of new scales and all the possible third order moments has returned better constrains to @xmath0 and lower error bars compared with the results by @xcite and previous works .
our best estimate is compatible with the values obtained by @xcite and it excludes the best - fitting @xmath0 value obtained by @xcite at the @xmath399% cl .
finally we have constrained @xmath0 for the north and south hemispheres and the results give two best - fitting values that are compatible with zero at 95% cl and also are compatible between them .
therefore we do not find any north - south asymmetry for this parameter .
we are thankful to frode hansen , michele liguori and sabino matarrese for supplying maps with primordial non - gaussianity .
the authors thank j. gonzlez - nuevo for providing the @xmath49 counts and for his useful comments on unresolved point sources .
the authors thank p. vielva , e. komatsu and m. cruz for useful discussions .
we also thank r. marco and l. cabellos for computational support .
we acknowledge partial financial support from the spanish ministerio de ciencia e innovacin project aya2007 - 68058-c03 - 02 .
a. c. thanks the spanish ministerio de ciencia e innovacin for a pre - doctoral fellowship .
the authors acknowledge the computer resources , technical expertise and assistance provided by the spanish supercomputing network ( res ) node at universidad de cantabria .
we acknowledge the use of legacy archive for microwave background data analysis ( lambda ) .
support for it is provided by the nasa office of space science .
the healpix package was used throughout the data analysis @xcite . , c. l. , hill , r. s. , hinshaw , g. , nolta , m. r. , odegard , n. , page , l. , spergel , d. n. , weiland , j. l. , wright , e. l. , halpern , m. , jarosik , n. , kogut , a. , limon , m. , meyer , s. s. , tucker , g. s. , & wollack , e. 2003 , apjs , 148 , 97 , e. , dunkley , j. , nolta , m. r. , bennett , c. l. , gold , b. , hinshaw , g. , jarosik , n. , larson , d. , limon , m. , page , l. , spergel , d. n. , halpern , m. , hill , r. s. , kogut , a. , meyer , s. s. , tucker , g. s. , weiland , j. l. , wollack , e. , & wright , e. l. 2009 , apjs , 180 , 330 , e. , kogut , a. , nolta , m. r. , bennett , c. l. , halpern , m. , hinshaw , g. , jarosik , n. , limon , m. , meyer , s. s. , page , l. , spergel , d. n. , tucker , g. s. , verde , l. , wollack , e. , & wright , e. l. 2003 , apjs , 148 , 119 , m. g. , heavens , a. , balbi , a. , borrill , j. , ferreira , p. g. , hanany , s. , jaffe , a. h. , lee , a. t. , rabii , b. , richards , p. l. , smoot , g. f. , stompor , r. , winant , c. d. , & wu , j. h. p. 2003 ,
mnras , 341 , 623 | we present constraints on the non - linear coupling parameter @xmath0 with the wilkinson microwave anisotropy probe ( wmap ) data .
we use an updated method based on the spherical mexican hat wavelet ( smhw ) which provides improved constraints on the @xmath0 parameter .
this paper is a continuation of a previous work by curto et al .
where several third order statistics based on the smhw were considered . in this paper
, we use all the possible third order statistics computed from the wavelet coefficient maps evaluated at 12 angular scales .
the scales are logarithmically distributed from 6.9 arcmin to 500 arcmin .
our analysis indicates that @xmath0 is constrained to @xmath1 at 95% confidence level ( cl ) for the combined v+w wmap map .
this value has been corrected by the presence of undetected point sources , which adds a positive contribution of @xmath2 .
our result excludes at @xmath399% cl the best - fitting value @xmath487 reported by yadav & wandelt .
we have also constrained @xmath0 for the q , v and w frequency bands separately , finding compatibility with zero at 95 % cl for the q and v bands but not for the w band .
we have performed some further tests to understand the cause of this deviation which indicate that systematics associated to the w radiometers could be responsible for this result .
finally we have performed a galactic north - south analysis for @xmath0 .
we have not found any asymmetry , i.e. the best - fitting @xmath0 for the northern pixels is compatible with the best - fitting @xmath0 for the southern pixels . |
the studies of strongly correlated electrons in transition metal oxides ( tmos ) focus traditionally on @xmath0 materials @xcite , mainly because of high - temperature superconductivity discovered in cuprates and more recently in iron - pnictides , and because of colossal magnetoresistance manganites .
the competition of different and complex types of order is ubiquitous in strongly correlated tmos mainly due to coupled spin - charge - orbital where frustrated exchange competes with the kinetic energy of charge carriers .
the best known example is spin - charge competition in cuprates , where spin , charge and superconducting orders intertwine @xcite and stripe order emerges in the normal phase as a compromise between the magnetic and kinetic energy @xcite .
remarkable evolution of the stripe order under increasing doping is observed @xcite and could be reproduced by the theory based on the extended hubbard model @xcite .
hole doping in cuprates corresponds to the removal of the spin degree of freedom .
similarly , hole doping in a simplest system with the orbital order in @xmath8 configuration removes locally orbital degrees of freedom and generates stripe phases which involve orbital polarons @xcite .
it was predicted recently that orbital domain walls in bilayer manganites should be partially charged as a result of competition between orbital - induced strain and coulomb repulsion @xcite , which opens a new route towards charge - orbital physics in tmos .
we will show below that the stripe - like order may also occur in doped spin - orbital systems .
these systems are very challenging and their doping leads to very complex and yet unexplored spin - orbital - charge phenomena @xcite .
a prerequisite to the phenomena with spin - orbital - charge coupled degrees of freedom is the understanding of undoped systems @xcite , where the low - energy physics and spin - orbital order are dictated by effective spin - orbital superexchange @xcite and compete with spin - orbital quantum fluctuations @xcite .
although ordered states occur in many cases , the most intriguing are quantum phases such as spin @xcite or orbital @xcite liquids .
recent experiments on a copper oxide ba@xmath9cusb@xmath10o@xmath11 @xcite have triggered renewed efforts in a fundamental search for a quantum spin - orbital liquid @xcite , where spin - orbital order is absent and electron spins are randomly choosing orbitals which they occupy . a signature of strong quantum effects in a spin - orbital system is a disordered state which persists down to very low temperatures .
a good example of such a disordered spin - orbital liquid state is as well fesc@xmath10s@xmath12 which does not order in spite of finite curie - weiss temperature @xmath13 k @xcite , but shows instead signatures of quantum criticality @xcite .
spin - orbital interactions may be even more challenging for instance previous attempts to find a spin - orbital liquid in the kugel - khomskii model @xcite or in linio@xmath10 @xcite turned out to be unsuccessful .
in fact , in the former case certain types of exotic spin order arise as a consequence of frustrated and entangled spin - orbital interactions @xcite , and a spin - orbital entangled resonating valence bond state was recently shown to be a quantum superposition of strped spin - singlet covering on a square lattice @xcite .
in contrast , spin and orbital superexchange have different energy scales and orbital interactions in linio@xmath10 are much stronger and dominated by frustration @xcite . hence the reasons behind the absence of magnetic long range order are more subtle @xcite . in all these cases orbital fluctuations play a prominent role and spin - orbital entanglement @xcite determines the ground state .
the role of charge carriers in spin - orbital systems is under very active investigation at present . in doped la@xmath14(sr ,
ca)@xmath15mno@xmath9 manganites several different types of magnetic order compete with one another and occur at increasing hole doping @xcite .
undoped lamno@xmath9 is an antiferromagnetic ( af ) mott insulator , with large @xmath16 spins for @xmath17 ionic configurations of mn@xmath18 ions stabilized by hund s exchange , coupled via the spin - orbital superexchange due to @xmath19 and @xmath7 electron excitations @xcite .
the orbital @xmath19 degree of freedom is removed by hole doping when mn@xmath18 ions are generated , and this requires careful modeling in the theory that takes into account both @xmath17 and @xmath2 electronic configurations of mn@xmath18 and mn@xmath20 ions @xcite .
in fact , the orbital order changes radically with increasing doping in la@xmath14(sr , ca)@xmath15mno@xmath9 systems at the magnetic phase transitions between different types of magnetic order @xcite , as weel as at la@xmath21ca@xmath22mno@xmath9/bifeo@xmath9 heterostructures , where it offers a new route to enhancing multiferroic functionality @xcite . the double exchange mechanism @xcite triggers ferromagnetic ( fm ) metallic phase at sufficient doping ; in this phase the spin and orbital degrees of freedom decouple and spin excitations are explained by the orbital liquid @xcite .
due to distinct magnetic and kinetic energy scales , even low doping may suffice for a drastic change in the magnetic order , as observed in electron - doped manganites @xcite . a rather unique example of a spin - orbital system with strongly fluctuating orbitals , as predicted in the theory @xcite and seen experimentally @xcite , are the perovskite vanadates with competing spin - orbital order @xcite . in these @xmath7 systems @xmath23 orbitals are filled by one electron and orbital order of active @xmath24 orbitals is strongly influenced by doping with ca ( sr ) ions which replace y ( la ) ones in yvo@xmath9 ( lavo@xmath9 ) . in this case
finite spin - orbit coupling modifies the spin - orbital phase diagram @xcite .
in addition , the af order switches easily from the @xmath25-type af ( @xmath25-af ) to @xmath26-type af ( @xmath26-af ) order in the presence of charge defects in y@xmath14ca@xmath15vo@xmath9 .
already at low @xmath27 doping the spin - orbital order changes and spectral weight is generated within the mott - hubbard gap @xcite .
although one might imagine that the orbital degree of freedom is thereby removed , a closer inspection shows that this is not the case as the orbitals are polarized by charge defects @xcite and readjust near them @xcite .
removing the orbital degree of freedom in vanadates would be only possible by electron doping generating instead @xmath28 ionic configurations , but such a doping by charge defects would be very different from the doping by transition metal ions of the same valence considered below . also in @xmath1 materials spin - orbital physics plays a role @xcite , as for instance in ca@xmath29sr@xmath15ruo@xmath12 systems with ru@xmath20 ions in @xmath4 configuration @xcite .
recently it has been shown that unconventional magnetism is possible for ru@xmath20 and similar ions where spin - orbit coupling plaus a role @xcite .
surprisingly , these systems are not similar to manganites but to vanadates where one finds as well ions with active @xmath7 orbitals . in the case of ruthenates
the @xmath30 ru@xmath20 ions have low @xmath31 spin as the splitting between the @xmath7 and @xmath19 levels is large .
thus the undoped ca@xmath10ruo@xmath12 is a hole analogue of a vanadate @xcite , with @xmath7 orbital degree of freedom and @xmath31 spin per site in both cases .
this gives new opportunities to investigate spin - orbital entangled states in @xmath7 system , observed recently by angle resolved photoemission @xcite .
here we focus on a novel and very different doping from all those considered above , namely on a _ substitutional _ doping by other magnetic ions in a plane built by transition metal and oxygen ions , for instance in the @xmath32 plane of a monolayer or in perovskite ruthenates or vanadates . in this study
we are interested primarily in doping of a tmo with @xmath7 orbital degrees of freedom , where doped magnetic ions have no orbital degree of freedom and realize _
orbital dilution_. in addition , we deal with the simpler case of @xmath0 doped ions where we can neglect spin - orbit interaction which should not be ignored for @xmath1 ions .
we emphasize that in contrast to manganites where holes within @xmath19 orbitals participate in transport and are responsible for the colossal magnetoresitance , such doped hole are immobile due to the ionic potential at @xmath0 sites and form defects in spin - orbital order of a mott insulator .
we encounter here a different situation from the dilution effects in the 2d @xmath19 orbital system considered so far @xcite as we deel with magnetic ions at doped sites .
it is challenging to investigate how such impurities modify locally or globally spin - orbital order of the host .
the doping which realizes this paradigm is by either mn@xmath20 or cr@xmath18 ions with large @xmath33 spins stabilized by hund s exchange , and orbital dilution occurs either in a tmo with @xmath34 ionic configuration as in the vanadium perovskites , or in @xmath1 mott insulators as in ruthenates .
it has been shown that dilute cr doping for ru reduces the temperature of the orthorhombic distortion and induces fm behavior in ca@xmath10ru@xmath14cr@xmath15o@xmath12 ( with @xmath35 ) @xcite .
it also induces surprising negative volume thermal expansion via spin - orbital order .
such defects , on one hand , can weaken the spin - orbital coupling in the host , but on the other hand , may open a new channel of interaction between the spin and orbital degree of freedom through the host - impurity exchange , see fig .
[ fig : bond_schem ] .
consequences of such doping are yet unexplored and are expected to open a new route in the research on strongly correlated oxides .
ion with no orbital degree of freedom and spin @xmath33 substitutes @xmath4 one with spin @xmath31 on a bond having specific spin and orbital character in the host ( gray arrows ) .
spins are shown by red arrows and doubly occupied @xmath7 orbitals ( doublons ) are shown by green symbols for @xmath36 and @xmath37 orbitals , respectively .
( b ) if an inactive orbital along the bond is removed by doping , the total spin exchange is af .
( c ) on the contrary , active orbitals at the host site can lead to either fm ( top ) or af ( bottom ) exchange coupling , depending on the energy levels mismatch and difference in the coulomb couplings between the impurity and the host .
we show the case when the host site is unchanged in the doping process .
[ fig : bond_schem ] ] the physical example for the present theory are the insulating phases of @xmath5 hybrid structures , where doping happens at @xmath38 transition metal sites , and the value of the spin is locally changed from @xmath31 to @xmath33 . as a demonstration of the highly nontrivial physics emerging in @xmath5 oxides , remarkable effects have already been observed , for instance , when ru ions are replaced by mn , ti , cr or other @xmath0 elements .
the role of mn doping in the srruo ruddlesden - popper series is strongly linked to the dimensionality through the number @xmath39 of ruo@xmath10 layers in the unit cell .
the mn doping of the srruo@xmath9 cubic member drives the system from the itinerant fm state to an insulating af configuration in a continuous way via a possible unconventional quantum phase transition @xcite .
doping by mn ions in sr@xmath40ru@xmath10o@xmath41 leads to a metal - to - insulator transition and af long - range order for more than @xmath42 mn concentration @xcite .
subtle orbital rearrangement can occur at the mn site , as for instance the inversion of the crystal field in the @xmath19 sector observed via x - ray absorption spectroscopy @xcite .
neutron scattering studies indicate the occurrence of an unusual @xmath43-type antiferromagnetism in doped systems ( planar order with fm zigzag chains with af order between them ) with moments aligned along the @xmath37 axis within a single bilayer @xcite .
furthermore , the more extended @xmath1 orbitals would _ a priori _ suggest a weaker correlation than in @xmath0 tmos due to a reduced ratio between the intraatomic coulomb interaction and the electron bandwidth . nevertheless , the ( effective ) @xmath44-bandwidth is reduced by the changes in the @xmath0-@xmath45-@xmath0 bond angles in distorted structures which typically arise in these materials .
this brings these systems on the verge of a metal - insulator transition @xcite , or even into the mott insulating state with spin - orbital order , see fig .
[ fig : host ] .
hence , not only @xmath1 materials share common features with @xmath0 systems , but are also richer due to their sensitivity to the lattice structure and to relativistic effects due to larger spin - orbit @xcite or other magneto - crystalline couplings .
-af spin order coexisting with @xmath25-ao orbital order in the @xmath32 plane of an undoped mott insulator with @xmath4 ionic configurations .
spins are shown by arrows while doubly occupied @xmath23 and @xmath46 orbitals ( @xmath37 and @xmath36 doublons , see text ) form a checkerboard pattern .
equivalent spin - orbital order is realized for v@xmath18 ions in @xmath47 planes of lavo@xmath9 @xcite , with orbitals standing for empty orbitals ( holes ) . ] to simplify the analysis we assume that onsite coulomb interactions are so strong that charge degrees of freedom are projected out , and only virtual charge transfer can occur between @xmath0 and @xmath1 ions via the oxygen ligands . for convenience ,
we define the orbital degree of freedom as a _ doublon _ ( double occupancy ) in the @xmath30 configuration .
the above @xmath0 doping leads then effectively to the removal of a doublon in one of @xmath7 orbitals which we label as @xmath48 ( this notation is introduced in ref .
@xcite and explained below ) and to replacing it by a @xmath49 ion . to our knowledge , this is the only example of removing the orbital degree of freedom in @xmath7 manifold realized so far and below we investigate possible consequences of this phenomenon .
another possibility of orbital dilution which awaits experimental realization would occur when a @xmath7 degree of freedom is removed by replacing a @xmath34 ion by a @xmath28 one , as for instance by cr@xmath18 doping in a vanadate here a doublon is an empty @xmath7 orbital , i.e. , filled by two holes . before presenting the details of the quantitative analysis ,
let us concentrate of the main idea of the superexchange modified by doping in a spin - orbital system .
the @xmath28 ions have singly occupied all three @xmath7 orbitals and @xmath33 spins due to hund s exchange .
while a pair of @xmath28 ions , e.g. in srmno@xmath9 , is coupled by af superexchange @xcite , the superexchange for the @xmath50 bond has a rather rich structure and may also be fm .
the spin exchange depends then on whether the orbital degree of freedom is active and participates in charge excitations along a considered bond or electrons of the doublon can not move along this bond due to the symmetry of @xmath7 orbital , as explained in fig .
[ fig : bond_schem ] .
this qualitative difference to systems without active orbital degrees of freedom is investigated in detail in sec .
[ sec : model ] .
the main outcomes of our analysis are : ( i ) the determination of the effective spin - orbital exchange hamiltonian describing the low - energy sector for the @xmath5 hybrid structure , ( ii ) establishing that a @xmath2 impurity without an orbital degree of freedom modifies the orbital order in the @xmath4 host , ( iii ) providing the detailed way how the microscopic spin - orbital order within the @xmath4 host is modified around the @xmath2 impurity , and ( iv ) suggesting possible spin - orbital patterns that arise due to periodic and finite substitution ( doping ) of @xmath1 atoms in the host by @xmath0 ones .
the emerging physical scenario is that the @xmath0 impurity acts as an orbital vacancy when the host - impurity coupling is weak and as an orbital polarizer of the bonds active @xmath7 doublon configurations when it is strongly coupled to the host .
the tendency to polarize the host orbitals around the impurity turns out to be robust and independent of spin configuration . otherwise , it is the resulting orbital arrangement around the impurity and the strength of hund s coupling at the impurity that set the character of the host - impurity magnetic exchange .
the remaining of the paper is organized as follows . in sec .
[ sec : model ] we introduce the effective model describing the spin - orbital superexchange at the @xmath5 bonds which serves to investigate the changes of spin and orbital order around individual impurities and at finite doping .
we arrive at a rather general formulation which emphasizes the impurity orbital degree of freedom , being a doublon , and present some technical details of the derivation in appendix a. the strategy we adopt is to analyze first the ground state properties of a single @xmath2 impurity surrounded by @xmath4 atoms by investigating how the spin - orbital pattern in the host may be modified at the nearest neighbor ( nn ) sites to the @xmath0 atom .
this study is performed for different spin - orbital patterns of the @xmath1 host with special emphasis on the alternating fm chains ( @xmath26-af order ) which coexist with @xmath25-type alternating orbital ( @xmath25-ao ) order , see fig .
[ fig : host ] .
we address the impurity problem within the classical approximation in sec .
[ sec : mfa ] . as explained in sec .
[ sec : two ] , there are two nonequivalent cases which depend on the precise modification of the orbital order by the @xmath0 impurity , doped either to replace a doublon in @xmath36 orbital ( sec .
[ sec : impa ] ) or the one in @xmath37 orbital ( sec . [
sec : impc ] ) .
starting from the single impurity solution we next address periodic arrangements of @xmath0 atoms at different concentrations .
we demonstrate that the spin - orbital order in the host can be radically changed by the presence of impurities , leading to striped patterns with alternating fm / af domains and islands of fully fm states . in sec .
[ sec : dopg ] we consider the modifications of spin - orbital order which arise at periodic doping with macroscopic concentration . here
we limit ourselves to two representative cases : ( i ) commensurate @xmath51 doping in sec .
[ sec : dop8 ] , and ( ii ) two doping levels @xmath52 and @xmath53 being incommensurate with underlying two - sublattice order ( fig .
[ fig : host ] ) which implies simultaneous doping at two sublattices , i.e. , at both @xmath36 and @xmath37 doublon sites , as presented in secs .
[ sec : dop5 ] and [ sec : dop9 ] . finally , in sec .
[ sec : orb ] we investigate the modifications of the classical phase diagram induced by quantum fluctuations , and in secs .
[ sec : soc ] and [ sec : jh ] we discuss representative results obtained for finite spin - orbit coupling ( calculation details of the treatment of spin - orbit interaction are presented in appendix b ) .
the paper is concluded by a general discussion of possible emerging scenarios for the @xmath2 impurities in @xmath4 host , a summary of the main results and perspective of future experimental investigations of orbital dilution in sec .
[ sec : sum ] .
in this section we consider a @xmath0 impurity in a strongly correlated @xmath1 tmo and derive the effective @xmath54 spin - orbital superexchange .
it follows from the coupling between @xmath0 and @xmath1 orbitals via oxygen @xmath45 orbitals due to the @xmath55 hybridization . in a strongly correlated system
it suffices to concentrate on a pair of atoms forming a bond @xmath56 , as the effective interactions are generated by charge excitations @xmath57 along a single bond @xcite . in the reference @xmath1 host
both atoms on the bond @xmath56 are equivalent and one considers , @xmath58 the coulomb interaction @xmath59 is local at site @xmath60 and we describe it by the degenerate hubbard model @xcite , see below .
we implement a strict rule that the hopping within the @xmath7 sector is allowed in a tmo only between two neighboring orbitals of the same symmetry which are active along the bond direction @xcite , and neglect the interorbital processes originating from the octahedral distortions such as rotation or tilting .
indeed , in ideal undistorted ( perovskite or square lattice ) geometry the orbital flavor is conserved as long as the spin - orbit coupling may be neglected .
the interorbital hopping elements are smaller by at least one order of magnitude and may be treated as corrections in cases where distortions play a role to the overall scenario established below .
the kinetic energy for a representative @xmath0-@xmath45-@xmath1 bond , i.e. , after projecting out the oxygen degrees of freedom , is given by the hopping in the host @xmath61 between sites @xmath60 and @xmath62 , @xmath63 here @xmath64 are the electron creation operators at site @xmath60 in the spin - orbital state @xmath65 .
the bond @xmath56 points along one of the two crystallographic directions , @xmath66 , in the two - dimensional ( 2d ) square lattice . without distortions ,
only two out of three @xmath7 orbitals are active along each bond @xmath67 and contribute to @xmath68 , while the third orbital lies in the plane perpendicular to the @xmath69 axis and thus the hopping via oxygen is forbidden by symmetry .
this motivates a convenient notation as follows @xcite , @xmath70 with the @xmath7 orbital inactive along a given direction @xmath71 labeled by the index @xmath69 .
we consider a 2d square lattice with transition metal ions connected via oxygen orbitals as in a ruo@xmath10 @xmath32 plane of ca@xmath10ruo@xmath12 ( srruo@xmath9 ) . in this case
@xmath72 ( @xmath73 ) orbitals are active along the @xmath74 ( @xmath36 ) axis , while @xmath75 orbitals are active along both @xmath76 axes . to derive the superexchange in a mott insulator , it is sufficient to consider a bond which connects nearest neighbor sites , @xmath77 .
below we consider a bond between an impurity site @xmath78 occupied by a @xmath0 ion and a neighboring host @xmath1 ion at site @xmath79 .
the hamiltonian for this bond can be then expressed in the following form , @xmath80 the total hamiltonian contains the kinetic energy term @xmath81 describing the electron charge transfer via oxygen orbitals , the onsite interaction terms @xmath82 for the @xmath0 @xmath83 ion at site @xmath84 , and the local potential of the @xmath1 atom , @xmath85 , which takes into account the mismatch of the energy level structure between the two ( @xmath1 and @xmath0 ) atomic species and prevents valence fluctuations when the host is doped , even in the absence of local coulomb interaction . the kinetic energy in eq . ( [ hub ] ) is given by , @xmath86 where @xmath87 is the electron creation operator at site @xmath84 in the spin - orbital state @xmath65 .
the bond @xmath67 points along one of the two crystallographic directions , @xmath66 , and again the orbital flavor is conserved @xcite .
the coulomb interaction on an atom at site @xmath84 depends on two parameters @xcite : ( i ) intraorbital coulomb repulsion @xmath88 , and ( ii ) hund s exchange @xmath89 .
the label @xmath90 stands for the ion and distinguishes between these terms at the @xmath0 and @xmath1 ion , respectively . the interaction is expressed in the form , @xmath91 the terms standing in the first line of eq .
( [ hint ] ) contribute to the magnetic instabilities in degenerate hubbard model @xcite and decide about spin order , both in an itinerant system and in a mott insulator .
the remaining terms contribute to the multiplet structure and are of importance for the correct derivation of the superexchange which follows from charge excitations , see below .
finally , we include a local potential on the @xmath1 atom which encodes the energy mismatch between the host and the impurity orbitals close to the fermi level and prevents valence fluctuations on the @xmath1 ion due to the @xmath0 doping .
this term has the following general structure , @xmath92 with @xmath93 .
the effective hamiltonian for the low energy processes is derived from @xmath94 ( [ hub ] ) by a second order expansion for charge excitations generated by @xmath81 , and treating the remaining part of @xmath94 as an unperturbed hamiltonian .
we are basically interested in virtual charge excitations in the manifold of degenerate ground states of a pair of @xmath0 and @xmath1 atoms on a bond , see fig .
[ fig:3d4d ] .
these quantum states are labeled as @xmath95 with @xmath96 and @xmath97 with @xmath98 and their number follows from the solution of the onsite quantum problem for the hamiltonian @xmath59 . for the @xmath0 atom
the relevant states can be classified according to the four components of the total spin @xmath99 for the @xmath0 impurity atom at site @xmath100 , three components of @xmath101 spin for the @xmath1 host atom at site @xmath102 and for the three different positions of the double occupied orbital ( doublon ) .
thus , the effective hamiltonian will contain spin products @xmath103 between spin operators defined as , @xmath104 for @xmath84 sites and the operator of the doublon position at site @xmath102 , @xmath105 the doublon operator identifies the orbital @xmath69 within the @xmath7 manifold of the @xmath1 ion with a double occupancy ( occupied by the doublon ) and stands in what follows for the orbital degree of freedom .
it is worth noting that the hopping ( [ eq : hop ] ) does not change the orbital flavor thus we expect that the resulting hamiltonian is diagonal in the orbital degrees of freedom with only @xmath106 operators .
bond @xmath67 as given by the local coulomb hamiltonian @xmath82 ( [ hint ] ) with @xmath84 .
the dominant exchange processes considered here are those that move one of the four electrons on the @xmath1 atom to the @xmath0 neighbor and back .
the stability of the @xmath107-@xmath108 charge configurations is provided by the local potential energy @xmath109 , see eq .
( [ eq : ion2 ] ) . ] following the standard second order perturbation expansion for spin - orbital systems @xcite , we can write the matrix elements of the low energy exchange hamiltonian , @xmath110 , for a bond @xmath111 along the @xmath69 axis as follows , @xmath112 with @xmath113 being the excitation energies for atoms at site @xmath84 with respect to the unperturbed ground state . the superexchange hamiltonian @xmath114 for a bond along @xmath69 can be expressed in a matrix form by a @xmath115 matrix , with dependence on @xmath88 , @xmath89 , and @xmath116 elements .
there are two types of charge excitations : ( i ) @xmath117 one which creates a doublon at the @xmath0 impurity , and ( ii ) @xmath118 one which adds another doublon at the @xmath1 host site in the intermediate state .
the second type of excitations involves more doubly occupied orbitals and has much larger excitation energy .
it is therefore only a small correction to the leading term ( i ) , as we discuss in appendix a. similar as in the case of doped manganites @xcite , the dominant contribution to the effective low - energy spin - orbital hamiltonian for the @xmath5 bond stems from the @xmath117 charge excitations , as they do not involve an extra double occupancy and the coulomb energy @xmath119 .
the @xmath120 charge excitations can be analyzed in a similar way as the @xmath121 ones for an @xmath56 bond in doped manganites @xcite . in both cases
the total number of doubly occupied orbitals does not change , so the main contributions come due to hund s exchange . in the present case , one more parameter plays a role , @xmath122 which stands for the mismatch potential energy ( [ eq : ion2 ] ) renormalized by the onsite coulomb interactions @xmath123 and by hund s exchange @xmath124 . on a general ground
we expect @xmath125 to be a positive quantity , since the repulsion @xmath88 should be larger for smaller @xmath0 shells than for the @xmath1 ones and @xmath88 is the largest energy scale in the problem .
let us have a closer view on this dominant contribution of the effective low - energy spin - orbital hamiltonian for the @xmath5 bond , given by eq .
( [ eq : s - ex ] ) . for the analysis performed below and the clarity of our presentation it is convenient to introduce some scaled parameters related to the interactions within the host and between the host and the impurity . for this purpose
we employ the exchange couplings @xmath126 and @xmath127 , @xmath128 which follow from the virtual charge excitations generated by the kinetic energy , see eqs .
( [ eq : hop4d ] ) and ( [ eq : hop ] ) .
we use their ratio to investigate the influence of the impurity on the spin - orbital order in the host . here
@xmath129 is the hopping amplitude between two @xmath7 orbitals at nn @xmath1 atoms , @xmath130 and @xmath131 refer to the host , and @xmath125 ( [ delta ] ) is the renormalized ionization energy of the @xmath5 bonds .
the results depend as well on hund s exchange element for the impurity and on the one at host atoms , @xmath132 note that the ratio introduced for the impurity , @xmath133 ( [ eq : etai ] ) , has here a different meaning from hund s exchange used here for the host , @xmath134 ( [ eq : etah ] ) , which can not be too large by construction , i.e. , @xmath135 . with the parametrization introduced above , the dominant term in the impurity - host hamiltonian for the impurity spin @xmath136 interacting with the neighboring host spins @xmath137 at @xmath138 , deduced from @xmath139 eq .
( [ eq : s - ex ] ) , can be written in a rather compact form as follows @xmath140 where the orbital ( doublon ) dependent spin couplings @xmath141 and the doublon energy @xmath142 depend on @xmath133 .
the evolution of the exchange couplings are shown in fig .
[ fig : jsed ] .
we note that the dominant energy scale is @xmath143 , so for a single @xmath5 bond the doublon will avoid occupying the inactive ( @xmath69 ) orbital and the spins will couple with @xmath144 which can be either af if @xmath145 or fm if @xmath146 .
thus the spins at @xmath147 will decouple according to the @xmath110 exchange . and the doublon energy @xmath142 , both given in eq .
( [ eq : h123 ] ) for increasing hund s exchange @xmath133 at the impurity .
[ fig : jsed ] ] let us conclude this section by writing the complete superexchange hamiltonian , @xmath148 where @xmath149 includes all the @xmath5 bonds around impurities , @xmath150 stands for the the effective spin - orbital hamiltonian for the @xmath1 host bonds , and @xmath151 is the spin - orbit interaction in the host .
the former term we explain below , while the latter one is defined in sec .
[ sec : soc ] , where we analyze the quantum corrections and the consequences of spin - orbit interaction .
the superexchange in the host for the bonds @xmath56 along the @xmath66 axes @xcite , @xmath152 depends on @xmath153 and @xmath154 operators acting only in the orbital space .
they are expressed in terms of the pseudospin operators defined in the orbital subspace spanned by the two orbital flavors active along a given direction @xmath69 , i.e. , @xmath155 with @xmath156 standing for the multiplet structure in charge excitations , and the orbital operators @xmath157 that for the @xmath158 axis take the form : @xmath159{cc } a_{i}^{\dagger } & b_{i}^{\dagger}\end{array}\big)\cdot\vec{\sigma}\cdot \big(\begin{array}[t]{cc}a_i^ { } & b_i^{}\end{array}\big)^{\intercal } , \\
n_{i}^{(c)}&=&a_{i}^{\dagger}a_{i}^{}+b_{i}^{\dagger}b_{i}^{}.\end{aligned}\ ] ] for the directions @xmath66 in the considered @xmath32 plane one finds equivalent expressions by cyclic permutation of the axis labels @xmath48 in the above formulas .
this problem is isomorphic with the spin - orbital superexchange in the vanadium perovskites @xcite , where a hole in the @xmath160 doublet plays an equivalent role to the doublon in the present case .
the operators @xmath161 are the doublon ( hard core boson ) creation operators in the orbital @xmath162 , respectively , and they satisfy the local constraint , @xmath163 meaning that exactly _
one _ doublon ( [ eq : dub ] ) occupies one of the three @xmath7 orbitals at each site @xmath60 .
these bosonic occupation operators coincide with the previously used doublon occupation operators @xmath164 , i.e. , @xmath165 with @xmath162 .
below we follow first the classical procedure to determine the ground states of single impurities in sec . [
sec : imp ] , and at macroscopic doping in sec [ sec : dop ] .
in this section we describe the methodology that we applied for the determination of the phase diagrams for a single impurity reported below in secs . [
sec : impa ] , and next at macroscopic doping , as presented in sec .
[ sec : dop ] .
let us consider first the case of a single @xmath0 impurity in the @xmath1 host . since the interactions in the model hamiltonian are only effective ones between nn sites , it is sufficient to study the modification of the spin - orbital order around the impurity for a given spin - orbital configuration of the host by investigating a cluster of @xmath166 sites shown in fig .
[ fig : setup ] .
we assume the @xmath26-af spin order ( fm chains coupled antiferromagnetically ) accompanied by @xmath25-ao order within the host which is the spin - orbital order occurring for the realistic parameters of a ruo@xmath10 plane @xcite , see fig .
[ fig : host ] .
such a spin - orbital pattern turns out to be the most relevant one when considering the competition between the host and the impurity as due to the ao order within the @xmath32 plane .
other possible configurations with uniform orbital order and af spin pattern , e.g. @xmath25-af order , will also be considered in the discussion throughout the manuscript .
the sites @xmath167 inside the cluster in fig .
[ fig : setup ] have active spin and orbital degrees of freedom while the impurity at site @xmath168 has only spin degree of freedom . at the remaining sites
the spin - orbital configuration is assigned , following the order in the host , and it does not change along the computation .
impurity within the @xmath1 host in an @xmath32 plane .
the impurity is at the central site @xmath168 which belongs to the @xmath37 orbital host sublattice . for the outer sites in this cluster
the spin - orbital configuration is fixed and determined by the undoped @xmath1 host ( with spins and @xmath37 orbitals shown here ) having @xmath26-af/@xmath25-ao order , see fig .
[ fig : host ] .
for the central @xmath168 site the spin state and for the host sites @xmath169 the spin - orbital configurations are determined by minimizing the energy of the cluster . ] to determine the ground state we assume that the spin - orbital degrees of freedom are treated as classical variables .
this implies that for the bonds between atoms in the host we use the hamiltonian ( [ eq : hhost ] ) and neglect quantum fluctuations , i.e. , in the spin sector we keep only the @xmath170th ( ising ) spin components and in the orbital one only the terms which are proportional to the doublon occupation numbers ( [ eq : dub ] ) and to the identity operators .
similarly , for the impurity - host bonds we use the hamiltonian eq .
( [ eq : h123 ] ) by keeping only the @xmath170th projections of spin operators . since we do neglect the fluctuation of the spin amplitude it is enough to consider only the maximal and minimal values of @xmath171 for spin @xmath33 at the impurity sites and @xmath31 at the host atoms . with these assumptions
we can construct all the possible configurations by varying the spin and orbital configurations at the sites from @xmath78 to @xmath172 in the cluster shown in fig .
[ fig : setup ] .
note that the outer ions in the cluster belong all to the same sublattice , so two distinct cases have to be considered to probe all the configurations .
since physically it is unlikely that a single impurity will change the orbital order of the host globally thus we will not compare the energies from these two cases and analyze two classes of solutions separately , see sec .
[ sec : two ] .
then , the lowest energy configuration in each class provides the optimal spin - orbital pattern for the nns around the @xmath0 impurity . in the case of degenerate classical states ,
the spin - orbital order is established by including quantum fluctuations . in the case of a periodic doping analyzed in sec .
[ sec : dop ] , we use a similar strategy in the computation . taking the most general formulation ,
we employ larger clusters having both size and shape that depend on the impurity distribution and on the spin - orbital order in the host . for this purpose ,
the most natural choice is to search for the minimum energy configuration in the elementary unit cell that can reproduce the full lattice by a suitable choice of the translation vectors .
this is computationally expensive but doable for a periodic distribution of the impurities that is commensurate to the lattice because it yields a unit cell of relatively small size for doping around @xmath173 . otherwise , for the incommensurate doping the size of the unit cell can lead to a configuration space of a dimension that impedes finding of the ground state .
this problem is computationally more demanding and to avoid the comparison of all the energy configurations , we have employed the metropolis algorithm at low temperature to achieve the optimal configuration iteratively along the convergence process .
note that this approach is fully classical , meaning that the spins of the host and impurity are treated as ising variables and the orbital fluctuations in the host s hamiltonian eq .
( [ eq : hhost ] ) are omitted .
they will be addressed in sec .
[ sec : orb ] .
the single impurity problem is the key case to start with because it shows how the short - range spin - orbital correlations are modified around the @xmath0 atom due to the host and host - impurity interactions in eq .
( [ fullh ] ) .
the analysis is performed by fixing the strength between hund s exchange and coulomb interaction within the host ( [ hint ] ) at @xmath174 , and by allowing for a variation of the ratio between the host - impurity interaction ( [ eq : h123 ] ) and the coulomb coupling at the impurity site .
the choice of @xmath174 is made here because this value is within the physically relevant range for the case of the ruthenium materials .
small variations of @xmath134 do not affect the obtained results qualitatively . as we have already discussed in the model derivation , the sign of the magnetic exchange between the impurity and the host depends on the orbital occupation of the @xmath1 doublon around the @xmath0 impurity . the main aspect that controls the resulting magnetic configuration
is then given by the character of the doublon orbitals around the impurity , depending on whether they are active or inactive along the considered @xmath5 bond .
to explore such a competition quantitatively we investigate @xmath25-ao order for the host with alternation of @xmath36 and @xmath37 doublon configurations accompanied by the @xmath26-af spin pattern , see fig .
[ fig : host ] .
note that the @xmath36 orbitals are active only along the @xmath74 axis , while the @xmath37 orbitals are active along the both axes : @xmath36 and @xmath74 @xcite .
this state has the lowest energy for the host in a wide range of parameters for hund s exchange , coulomb element and crystal - field potential @xcite . due to the specific orbital pattern of fig .
[ fig : host ] , the @xmath0 impurity can substitute one of two distinct @xmath1 sites which are considered separately below , either with @xmath36 or with @xmath37 orbital occupied by the doublon .
since the two @xmath1 atoms have nonequivalent surrounding orbitals , not always active along the @xmath5 bond , we expect that the resulting ground state will have a modified spin - orbital order . indeed ,
if the @xmath0 atom replaces the @xmath1 one with the doublon in the @xmath36 orbital , then all the @xmath1 neighboring sites have active doublons along the connecting @xmath5 bonds because they are in the @xmath37 orbitals . on the contrary
, the substitution at the @xmath1 site with @xmath37 orbital doublon configuration leads to an impurity state with its neighbors having both active and inactive doublons .
therefore , we do expect a more intricate competition for the latter case of an impurity occupying the @xmath1 site with @xmath37 orbital configuration .
indeed , this leads to frustrated host - impurity interactions , as we show in sec .
[ sec : impc ] .
we start by considering the physical situation where the @xmath0 impurity replaces a @xmath1 ion with the doublon within the @xmath36 orbital .
the ground state phase diagram and the schematic view of the spin - orbital pattern are reported in fig .
[ fig : pd_caf_a ] in terms of the ratio @xmath175 ( [ eq : etai ] ) and the strength of hund s exchange coupling @xmath133 ( [ eq : bothj ] ) at the @xmath0 site .
there are three different ground states that appear in the phase diagram .
taking into account the structure of the @xmath5 spin - orbital exchange ( [ eq : h123 ] ) we expect that , in the regime where the host - impurity interaction is greater than that in the host , the @xmath1 neighbors to the impurity tend to favor the spin - orbital configuration set by the @xmath5 exchange . in this case , since the orbitals surrounding the @xmath0 site already minimize the @xmath5 hamiltonian , we expect that the optimal spin configuration corresponds to the @xmath1 spins aligned either antiferromagnetically or ferromagnetically with respect to the impurity @xmath0 spin .
impurity in the @xmath32 plane with the @xmath26-af/@xmath25-ao order in @xmath1 host for the impurity doped at the sublattice with an @xmath36-orbital doublon .
different colors refer to local spin order around the impurity , af and fm , while fs indicates the intermediate regime of frustrated impurity spin
. bottom schematic view of spin - orbital patterns for the ground state configurations shown in the top panel .
the @xmath0 atom is at the central site , the dotted frame highlights the @xmath1 sites where the impurity induces a a spin reversal . in the fs@xmath36 phase
the question mark stands for that the frustrated impurity spin within the classical approach but frustration is released by the quantum fluctuations of the nn @xmath37 orbitals in the @xmath36 direction resulting in small af couplings along the @xmath36 axis , and spins obey the @xmath26-af order ( small arrow ) .
the labels fm@xmath36 and af@xmath36 refer to the local spin order around the @xmath0 impurity site with respect to the host these states differ by spin inversion at the @xmath0 atom site .
[ fig : pd_caf_a ] ] the neighbor spins are af to the @xmath0 spin impurity in the af@xmath36 phase , while the fm@xmath36 phase is just obtained from af@xmath36 by reversing the spin at the impurity , and having all the @xmath5 bonds fm .
it is interesting to note that due to the host - impurity interaction the @xmath26-af spin pattern of the host is modified in both the af@xmath36 and the fm@xmath36 ground states .
another intermediate configuration which emerges when the host - impurity exchange is weak in the intermediate fs@xmath36 phase where the impurity spin is undetermined and its configuration in the initial @xmath26-af phase is degenerate with the one obtained after the spin - inversion operation .
this is a singular physical situation because the impurity does not select a specific direction even if the surrounding host has a given spin - orbital configuration .
such a degeneracy is clearly verified at the critical point @xmath176 where the amplitude of the @xmath5 coupling vanishes when the doublon occupies the active orbital .
interestingly , such a degenerate configuration is also obtained at @xmath177 when the host dominates and the spin configuration at the @xmath1 sites around the impurity are basically determined by @xmath127 . in this case ,
due to the @xmath26-af spin order , always two bonds are fm and other two have af order , independently of the spin orientation at the @xmath0 impurity .
this implies that both fm or af couplings along the @xmath5 bonds perfectly balance each other which results in the degenerate fs@xmath36 phase .
it is worth pointing out that there is a quite large region of the phase diagram where the fs@xmath36 state is stabilized and the spin - orbital order of the host is not affected by doping with the possibility of having large degeneracy in the spin configuration of the impurities . on the other hand , by inspecting the @xmath37 orbitals around the impurity ( fig .
[ fig : pd_caf_a ] ) from the point of view of the full host s hamiltonian eq .
( [ eq : hhost ] ) with orbital flips included , @xmath178 , one can easily find out that the frustration of the impurity spin can be released by quantum orbital fluctuations . note that the @xmath37 orbitals around the impurity in the @xmath36 ( @xmath74 ) direction have quite different surroundings .
the ones along the @xmath36 axis are connected by two active bonds along the @xmath74 axis with orbitals @xmath36 , as in fig .
[ fig : host_flips](a ) , while the ones along @xmath74 are connected with _ only one _ active @xmath36 orbital along the same @xmath74 axis .
this means that in the perturbative expansion the orbital flips will contribute only along the @xmath74 bonds ( for the present @xmath25-ao order ) and admix the @xmath36 orbital character to @xmath37 orbitals along them , while such processes will be blocked for the bonds along the @xmath36 axis , as also for @xmath74 orbitals along the @xmath74 axis , see fig .
[ fig : host_flips](b ) .
host : ( a ) an active bond with respect to orbital flips , @xmath178 , and ( b ) an inactive bond , where orbital fluctuations are blocked by the orbital symmetry here the orbitals are static and only ising terms contribute to the ground state energy . ]
this fundamental difference can be easily included in the host - impurity bond in the mean - field manner by setting @xmath179 for the bonds along the @xmath74 axis and @xmath180 for the bonds along the @xmath36 axis .
then one can easily check that for the impurity spin pointing downwards we get the energy contribution from the spin - spin bonds which is given by @xmath181 , and for the impurity spin pointing upwards we have @xmath182 , with @xmath183 .
thus , it is clear that any admixture of the virtual orbital flips in the host s wave function polarize the impurity spin upwards so that the @xmath26-af order of the host will be restored .
let us move to the case of the @xmath0 atom replacing the doublon at @xmath37 orbital . as anticipated above
, this configuration is more intricate because the orbitals surrounding the impurity , as originated by the @xmath26-af / ao order within the host , lead to nonequivalent @xmath5 bonds .
there are two bonds with the doublon occupying an inactive orbital ( and has no hybridization with the @xmath7 orbitals at @xmath184d atom ) , and two remaining bonds with doublons in active @xmath7 orbitals .
since the @xmath5 spin - orbital exchange depends on the orbital polarization of @xmath1 sites we do expect a competition which may modify significantly the spin - orbital correlations in the host .
indeed , one observes that three configurations compete , denoted as af1@xmath37 , af2@xmath37 and fm@xmath37 , see fig .
[ fig : pd_caf_c ] . in the regime
where the host - impurity exchange dominates the system tends to minimize the energy due to the @xmath5 spin - orbital coupling and , thus , the orbitals become polarized in the active configurations compatible with the @xmath26-af/@xmath25-ao pattern and the host - impurity spin coupling is af for @xmath185 , while it is fm otherwise .
this region resembles orbital polarons in doped manganites @xcite .
also in this case , the orbital polarons arise because they minimize the double exchange energy @xcite .
impurity in the @xmath1 host with @xmath26-af/@xmath25-ao order and the impurity doped at the @xmath37 doublon sublattice .
different colors refer to local spin order around the impurity : af@xmath186 , af@xmath187 , and fm .
bottom schematic view of spin - orbital patterns for the two af ground state configurations shown in the top panel ; the fm@xmath37 phase differs from the af@xmath187 one only by spin inversion at the @xmath0 atom .
the @xmath0 atom is at the central site and has no doublon orbital , the frames highlight the spin - orbital defects caused by the impurity . as in fig .
[ fig : pd_caf_a ] , the labels af and fm refer to the impurity spin orientation with respect to the neighboring @xmath1 sites . ] on the contrary , for weak spin - orbital coupling between the impurity and the host there is an interesting cooperation between the @xmath0 and @xmath1 atoms . since the strength of the impurity - host coupling is not sufficient to polarize the orbitals at the @xmath1 sites , it is preferable to have an orbital rearrangement to the configuration with inactive orbitals on @xmath5 bonds and spin flips at @xmath1 sites . in this way the spin - orbital exchange is optimized in the host and also on the @xmath5 bonds .
the resulting state has an af coupling between the host and the impurity as it should when all the orbitals surrounding the @xmath0 atoms are inactive with respect to the bond direction .
this modification of the orbital configuration induces the change in spin orientation .
the double exchange bonds ( with inactive doublon orbitals ) along the @xmath74 axis are then blocked and the total energy is lowered , in spite of the frustrated spin - orbital exchange in the host . as a result , the af1@xmath37 state the spins surrounding the impurity are aligned and antiparallel to the spin at the @xmath0 site .
concerning the host @xmath26-af/@xmath25-ao order we note that it is modified only along the direction where the fm correlations develop and spin defects occur within the chain doped by the @xmath0 atom .
the fm order is locally disturbed by the @xmath0 defect antiferromagnetically coupled spins surrounding it .
note that this phase is driven by the orbital vacancy as the host develops more favorable orbital bonds to gain the energy in the absence of the orbital degrees of freedom at the impurity . at the same time
the impurity - host bonds do not generate too big energy losses as : ( i ) either @xmath133 is so small that the loss due to @xmath142 is compensated by the gain from the superexchange @xmath188 ( all these bonds are af ) , see fig .
[ fig : jsed ] , or ( ii ) @xmath175 is small meaning that the overall energy scale of the impurity - host exchange remains small .
interestingly , if we compare the af1@xmath37 with the af2@xmath37 ground states we observe that the disruption of the @xmath26-af/@xmath25-ao order is anisotropic and occurs either along the fm chains in the af1@xmath37 phase or perpendicular to the fm chains in the af2@xmath37 phase .
no spin frustration is found here , in contrast to the fs@xmath36 phase in the case of @xmath36 doublon doping , see fig .
[ fig : pd_caf_a ] . finally , we point out that a very similar phase diagram can be obtained assuming that the host has the fm/@xmath25-ao order with @xmath36 and @xmath74 orbitals alternating from site to site
. such configuration can be stabilized by a distortion that favors the out - of - plane orbitals . in this case
there is no difference in doping at one or the other sublattice .
the main difference is found in energy scales for the @xmath25-af/@xmath26-ao order the diagram is similar to the one of fig .
[ fig : pd_caf_c ] if we rescale @xmath126 by half , which means that the @xmath25-af order is softer than the @xmath26-af one .
note also that in the peculiar af1@xmath37 phase the impurity does not induce any changes in the host for the fm / ao ordered host . thus we can safely conclude that the observed change in the orbital order for the @xmath26-af host in the af1@xmath37 phase is due to the presence of the @xmath37 orbitals which are not directional in the @xmath32 plane . summarizing ,
we have shown the complexity of local spin - orbital order around @xmath49 impurities in a @xmath189 host .
it is remarkable that such impurity spins not only modify the spin - orbital order around them in a broad regime of parameters , but also are frequently frustrated .
this highlights the importance of quantum effects beyond the present classical approach which release frustration as we show in sec .
[ sec : orb ] .
in this section we analyze the spin - orbital patterns due to a finite concentration @xmath190 of @xmath0 impurities within the @xmath1 host with @xmath26-af/@xmath25-ao order , assuming that the @xmath0 impurities are distributed in a periodic way .
the study is performed for three representative doping distributions the first one @xmath51 is commensurate with the underlying spin - orbital order and the other two are incommensurate with respect to it , meaning that in such cases doping at both @xmath36 and @xmath37 doublon sites is imposed simultaneously . as the impurities lead to local energy gains due to @xmath5 bonds surrounding them
, we expect that the most favorable situation is when they are isolated and have maximal distances between one another . therefore , we selected the largest possible distances for the three doping levels used in our study : @xmath51 , @xmath52 , and @xmath53 .
this choice allows us to cover different regimes of competition between the spin - orbital coupling within the host and the @xmath5 coupling .
while single impurities may only change spin - orbital order locally , we use here a high enough doping to investigate possible global changes in spin - orbital order , i.e. , whether they can occur in the respective parameter regime .
the analysis is performed as for a single impurity , by assuming the classical spin and orbital variables and by determining the configuration with the lowest energy . for this analysis we set the spatial distribution of the @xmath0 atoms and we determine the spin and orbital profile that minimizes the energy .
we begin with the phase diagram obtained at @xmath51 @xmath0 doping , see fig .
[ fig:1to8 ] . in the regime of strong impurity - host coupling
the @xmath5 spin - orbital exchange determines the orbital and spin configuration of the @xmath1 atoms around the impurity .
the most favorable state is when the doublon occupies @xmath37 orbitals at the nn sites to the impurity .
the spin correlations between the impurity and the host are af ( fm ) , if the amplitude of @xmath133 is below ( above ) @xmath191 , leading to the af@xmath36 and the fm@xmath36 states , see fig .
[ fig:1to8 ] .
the af@xmath36 state has a striped - like profile with af chains alternated by fm domains ( consisting of three chains ) along the diagonal of the square lattice .
even if the coupling between the impurity and the host is af for all the bonds in the af@xmath36 state , the overall configuration has a residual magnetic moment originating by the uncompensated spins and by the cooperation between the spin - orbital exchange in the @xmath1 host and that for the @xmath5 bonds .
interestingly , at the point where the dominant @xmath5 exchange tends to zero ( i.e. , for @xmath192 ) , one finds a region of the fs@xmath36 phase which is analogous to the fs@xmath36 phase found in sec .
[ sec : impa ] for a single impurity , see fig .
[ fig : pd_caf_a ] .
again the impurity spin is frustrated in purely classical approach but this frustration is easily released by the orbital fluctuations in the host so that the @xmath26-af order of the host can be restored .
this state is stable for the amplitude of @xmath133 being close to @xmath191 .
doping @xmath51 .
different colors refer to local spin order around the impurity : af@xmath36 , af@xmath37 , fs@xmath36 , and fm@xmath36 .
bottom panel schematic view of the ground state configurations within the four 8-site unit cells ( indicated by blue dashed lines ) for the phases shown in the phase diagram .
the question marks in fs@xmath36 phase indicate frustrated impurity spins within the classical approach the spin direction ( small arrows ) is fixed only by quantum fluctuations .
the @xmath0 atoms are placed at the sites where orbitals are absent . ]
the regime of small @xmath175 ratio is qualitatively different an orbital rearrangement around the impurity takes place , with a preference to move the doublons into the inactive orbitals along the @xmath5 bonds .
such orbital configurations favor the af spin coupling at all the @xmath5 bonds which is stabilized by the @xmath193 superexchange @xcite .
this configuration is peculiar because it generally breaks inversion and does not have any plane of mirror symmetry .
it is worth pointing out that the original order in the @xmath1 host is completely modified by the small concentration of @xmath0 ions and one finds that the af coupling between the @xmath0 impurity and the @xmath1 host generally leads to patterns such as the af@xmath37 phase where fm chains alternate with af ones in the @xmath32 plane .
another relevant issue is that the cooperation between the host and impurity can lead to a fully polarized fm@xmath36 state .
this implies that doping can release the orbital frustration which was present in the host with the @xmath26-af/@xmath25-ao order . and fs@xmath194 phases in the case of @xmath52 periodic doping ( fig .
[ fig:1to5 ] ) .
frames mark the clusters which are not connected with orbitally active bonds .
[ fig : fsv ] ] next we consider doping @xmath52 with a given periodic spatial profile which concerns both doublon sublattices .
we investigate the @xmath0 spin impurities separated by the translation vectors @xmath195 and @xmath196 ( one can show that for general periodic doping @xmath190 , @xmath197 ) so there is a mismatch between the impurity periodicity and the two - sublattice @xmath25-ao order in the host .
one finds that the present case , see fig .
[ fig:1to5 ] , has similar general structure of the phase diagram to the case of @xmath51 ( fig .
[ fig:1to8 ] ) , with af correlations dominating for @xmath133 lower than @xmath191 and fm ones otherwise . due to the specific doping distribution
there are more phases appearing in the ground state phase diagram . for @xmath198
the most stable spin configuration is with the impurity coupled antiferromagnetically to the host .
this happens both in the af vacancy ( af@xmath199 ) and the af polaronic ( af@xmath194 ) ground states .
the difference between the two af states arises due to the orbital arrangement around the impurity . for weak ratio of the impurity to the host spin - orbital exchange , @xmath175 ,
the orbitals around the impurity are all inactive ones . on the contrary , in the strong impurity - host coupling regime all the orbitals
are polarized to be in active ( polaronic ) states around the impurity .
both states have been found as af1@xmath37 and af2@xmath37 phase in the single impurity problem ( fig .
[ fig : pd_caf_c ] ) .
more generally , for all phases the boundary given by an approximate hyperbolic relation @xmath200 separates the phases where the orbitals around impurities in the @xmath37-orbital sublattice are all inactive ( small @xmath133 ) from those where all the orbitals are active ( large @xmath133 ) .
the inactive orbital around the impurity stabilize always the af coupling between the impurity spin and host spins whereas the active orbitals can give either af or fm exchange depending on @xmath133 ( hence @xmath191 , see fig .
[ fig : jsed ] ) .
since the doping does not match the size of the elementary unit cell , the resulting ground states do not exhibit specific symmetries in the spin - orbital pattern .
they are generally fm due to the uncompensated magnetic moments and the impurity feels screening by the presence of the surrounding it host spins being antiparallel to the impurity spin . by increasing hund s exchange coupling at the @xmath0 ion the system develops fully fm state in a large region of the ground state diagram due to the possibility of suitable orbital polarization around the impurity . on the other hand , in the limit where the impurity - host bonds are weak , so either for @xmath192 and large enough @xmath175 so that all orbitals around the impurity are active , or just for small @xmath175 we get the fs phases where the impurity spin at the @xmath36-orbital sublattice is undetermined in the present classical approach .
this is a similar situation to the one found in the fs@xmath36 phase of a single impurity problem and at @xmath51 periodic doping , see figs .
[ fig : pd_caf_a ] and [ fig:1to8 ] , but there it was still possible to identify the favored impurity spin polarization by considering the orbital flips in the host around the impurity . however , the situation here is different as the host s order is completely altered by doping and has became isotropic , in contrast to the initial @xmath26-af order ( fig .
[ fig : host ] ) which breaks the planar symmetry between the @xmath36 and @xmath74 direction .
it was precisely this symmetry breaking that favored one impurity spin polarization over the other one . here
this mechanism is absent one can easily check that the neighborhood of the @xmath37 orbitals surrounding impurities is completely equivalent in both directions ( see fig .
[ fig : fsv ] for the view of these surroundings ) so that the orbital flip argument is no longer applicable .
this is a peculiar situation in the classical approach and we indicate frustration in spin direction by question marks in fig .
[ fig:1to5 ] . in fig .
[ fig : fsv ] we can see that both in fs vacancy ( fs@xmath199 ) and fs polaronic ( fs@xmath194 ) phase the orbitals are grouped in @xmath201 clusters and @xmath202 plaquettes , respectively , that encircle the degenerate impurity spins . for the fs@xmath199 phase we can distinguish between two kind of plaquettes with non - zero spin polarization differing by a global spin inversion . in the case of fs@xmath194 phases we observe four plaquettes with zero spin polarization arranged in two pairs related by a point reflection with respect to the impurity site .
it is worthwhile to realize that these plaquettes are completely disconnected in the orbital sector , i.e. , there are no orbitally active bonds connecting them ( see fig .
[ fig : host_flips ] for the pictorial definition of orbitally active bonds ) .
this means that quantum effects of purely orbital nature can appear only at the short range , i.e. , inside the plaquettes .
however , one can expect that if for some reason the two degenerate spins in a single elementary cell will polarize then they will also polarize in the same way in all the other cells to favor long - range quantum fluctuations in the spin sector related to the translational invariance of the system .
finally we investigate low doping @xmath53 with a given periodic spatial profile , see fig .
[ fig:1to9 ] . here
the impurities are separated by the translation vectors @xmath203 and @xmath204 .
once again there is a mismatch between the periodic distribution of impurities and the host s two - sublattice ao order , so we again call this doping incommensurate as it also imposes doping at both doublon sublattices .
the ground state diagram presents gradually increasing tendency towards fm @xmath5 bonds with increasing @xmath133 , see fig .
[ fig:1to9 ] .
these polaronic bonds polarize as well the @xmath193 bonds and one finds an almost fm order in the fm@xmath194 state . altogether , we have found the same phases as at the higher doping of @xmath52 , see fig .
[ fig:1to5 ] , i.e. , af@xmath199 and af@xmath194 at low values @xmath133 , fm@xmath199 and fm@xmath194 in the regime of high @xmath133 , separated by the regime of frustrated impurity spins which occur within the phases : fs@xmath199 , fs1@xmath194 , and fs2@xmath194 .
the difference between the two af ( fm ) states in fig .
[ fig:1to9 ] is due to the orbital arrangement around the impurity . as for the other doping levels considered so far , @xmath51 and @xmath52
, we find neutral ( inactive ) orbitals around @xmath0 impurities in the regime of low @xmath175 in af@xmath199 and fm@xmath199 phases which lead to spin defects within the 1d fm chains in the @xmath26-af spin order .
a similar behavior was reported for single impurities in the low doping regime in sec .
[ sec : imp ] .
this changes radically above the orbital transition for both types of local magnetic order , where the orbitals reorient into the active ones .
one finds that spin orientations are then the same as those of their neighboring @xmath1 atoms , with some similarities to those found at @xmath52 , see fig .
[ fig:1to5 ] .
frustrated impurity spins occur in the crossover regime between the af and fm local order around impurities .
this follows from the local configurations around them , which include two @xmath205-spins and two @xmath206-spins accompanied by @xmath37 orbitals at the nn @xmath1 sites .
this frustration is easily removed by quantum fluctuations and we suggest that this happens again in the same way as for @xmath51 doping , as indicated by small arrows in the respective fs phases shown in fig .
[ fig:1to9 ] .
so far , we analyzed the ground states of @xmath0 impurities in the @xmath32 plane of a @xmath1 system using the classical approach . here
we show that this classical picture may be used as a guideline and is only quantitatively changed by quantum fluctuations if the spin - orbit coupling is weak .
we start the analysis by considering the quantum problem in the absence of spin - orbit coupling ( at @xmath207 ) .
the orbital doublon densities , @xmath208 with @xmath162 , and total @xmath209 are conserved quantities and thus good quantum numbers for a numerical simulation . to determine the ground state configurations in the parameters space and the relevant correlation functions we diagonalize exactly the hamiltonian matrix ( [ fullh ] ) for the cluster of @xmath210 sites by means of the lanczos algorithm . in fig .
[ fig : diag](a ) we report the resulting quantum phase diagram for an 8-site cluster having one impurity and assuming periodic boundary conditions , see fig .
[ fig : diag](b ) .
this appears to be an optimal cluster configuration because it contains a number of sites and connectivities that allows us to analyze separately the interplay between the host - host and the host - impurity interactions and to simulate a physical situation when the interactions within the host dominate over those between the host and the impurity .
such a problem is a quantum analogue of the single unit cell presented in fig .
[ fig:1to8 ] for @xmath51 periodic doping . as a general feature that resembles the classical phase diagram
, we observe that there is a prevalent tendency to have af - like ( fm - like ) spin correlations between the impurity and the host sites in the region of @xmath133 below ( above ) the critical point at @xmath176 which separates these two regimes , with intermediate configurations having frustrated magnetic exchange .
as we shall discuss below it is the orbital degree of freedom that turns out to be more affected by the quantum effects .
following the notation used for the classical case , we distinguish various quantum af ( qaf ) ground states , i.e. , qaf@xmath211 ( @xmath212 ) and qaf@xmath213 ( @xmath212 ) , as well as a uniform quantum fm ( qfm ) configuration , i.e. , qfm@xmath36 , and quantum frustrated one labeled as qfs@xmath36 . in order to visualize the main spin - orbital patterns contributing to the quantum ground state it is convenient to adopt a representation with arrows for the spin and ellipsoids for the orbital sector at any given host site .
the arrows stand for the on - site spin projection @xmath214 , with the length being proportional to the amplitude .
the length scale for the arrows is the same for all the configurations .
moreover , in order to describe the orbital character of the ground state we employed a graphical representation that makes use of an ellipsoid whose semi - axes @xmath48 length are given by the average amplitude of the squared angular momentum components @xmath215 , or equivalently by the doublon occupation eq .
( [ eq : dub ] ) .
for instance , for a completely flat circle ( degenerate ellipsoid ) lying in the plane perpendicular to the @xmath69 axis only the corresponding @xmath69 orbital is occupied . on the other hand
, if the ellipsoid develops in all three directions @xmath48 it implies that more than one orbital is occupied and the distribution can be anisotropic in general .
if all the orbitals contribute equally , one finds an isotropic spherical ellipsoid . due to the symmetry of the hamiltonian , the phases shown in the phase diagram of fig .
[ fig : diag](a ) can be characterized by the quantum numbers for the @xmath170-th spin projection , @xmath209 , and the doublon orbital occupation @xmath216 ( [ doc ] ) , @xmath217 : qaf@xmath186 @xmath218 , qfs@xmath219 @xmath220 , qaf@xmath221 @xmath222 , qaf@xmath219 and qaf@xmath219 @xmath223 , qfs@xmath221 @xmath224 , and qfm@xmath36 @xmath225 . despite the irregular shape of the cluster [ fig .
[ fig : diag](b ) ] there is also symmetry between the @xmath36 and @xmath74 directions .
for this reason , the phases with @xmath226 can be equivalently described either by the set @xmath227 or @xmath228 .
the outcome of the quantum analysis indicates that the spin patterns are quite robust as the spin configurations of the phases qaf@xmath36 , qaf@xmath37 , qfs@xmath36 and qfm@xmath36 are the analogues of the classical ones . the effects of quantum fluctuations are more evident in the orbital sector where mixed orbital patterns occur if compared to the classical case .
in particular , orbital inactive states around the impurity are softened by quantum fluctuations and on some bonds we find an orbital configuration with a superposition of active and inactive states .
the unique af states where the classical inactive scenario is recovered corresponds to the qaf@xmath186 and qaf@xmath187 ones in the regime of small @xmath133 .
a small hybridization of active and inactive orbitals along both the af and fm bonds is also observed around the impurity for the qfs@xmath36 phases as one can note by the shape of the ellipsoid at host sites .
moreover , in the range of large @xmath133 where the fm state is stabilized , the orbital pattern around the impurity is again like in the classical case .
a significant orbital rearrangement is also obtained within the host .
we generally obtain an orbital pattern that is slightly modified from the pure ao configuration assumed in the classical case .
the effect is dramatically different in the regime of strong impurity - host coupling ( i.e. , for large @xmath229 ) with af exchange ( qaf@xmath219 ) with the formation of an orbital liquid around the impurity and within the host , with doublon occupation represented by an almost isotropic shaped ellipsoid .
interestingly , though with a different orbital arrangement , the qfs@xmath221 and the qfs@xmath219 states are the only ones where the @xmath26-af order of the host is recovered .
for all the other phases shown in the diagram of fig .
[ fig : diag ] the coupling between the host and the impurity is generally leading to a uniform spin polarization with fm or af coupling between the host and the impurity depending on the strength of the host - impurity coupling .
altogether , we conclude that the classical spin patterns are only quantitatively modified and are robust with respect to quantum fluctuations . in this section
we analyze the quantum effects in the spin and orbital order around the impurity in the presence of the spin - orbit coupling at the host @xmath38 sites . for the @xmath230 configuration
the strong spin - orbit regime has been considered recently by performing an expansion around the atomic limit where the angular @xmath231 and spin @xmath136 momenta form a spin - orbit singlet for the amplitude of the total angular momentum , @xmath232 ( i.e. , @xmath233 ) @xcite . the instability towards an af state starting from the @xmath233 liquid has been obtained within the spin - wave theory @xcite for the low energy excitations emerging from the spin - orbital exchange . in the analysis presented here
we proceed from the limit of zero spin - orbit to investigate how the spin and orbital order are gradually suppressed when approaching the @xmath233 spin - orbit singlet state .
this issue is addressed by solving the full quantum hamiltonian ( [ fullh ] ) exactly on a cluster of @xmath210 sites including the spin - orbital exchange for the host and that one derived for the host - impurity coupling ( [ fullh ] ) as well as the spin - orbit term , @xmath234 where the sum includes the ions of the @xmath1 host and we use the spin @xmath31 and the angular momentum @xmath235 , as in the ionic @xmath4 configurations . here
@xmath236 is the spin - orbit coupling constant at @xmath1 host ions , and the components of the orbital momentum @xmath237 are defined as follows : @xmath238 to determine the ground state and the relevant correlation functions we use again the lanczos algorithm for the cluster of @xmath210 sites .
such an approach allows us to study the competition between the spin - orbital exchange and the spin - orbit coupling on equal footing without any simplifying approximation .
moreover , the cluster calculation permits to include the impurity in the host and deal with the numerous degrees of freedom without making approximations that would constrain the interplay of the impurity - host versus host - host interactions .
finite spin - orbit coupling significantly modifies the symmetry properties of the problem . instead of the su(2 )
spin invariance one has to deal with the rotational invariance related to the total angular momentum per site @xmath232 .
though the @xmath239 term in eq .
( [ hso ] ) commutes with both total @xmath240 and @xmath241 , the full hamiltonian for the host with impurities eq .
( [ fullh ] ) has a reduced symmetry because the spin sector is now linearly coupled to the orbital which has only the cubic symmetry .
thus the remaining symmetry is a cyclic permutation of the @xmath242 axes .
moreover , @xmath241 is not a conserved quantity due to the orbital anisotropy of the spin - orbital exchange in the host and the orbital character of the impurity - host coupling .
there one has a @xmath243 symmetry associated with the parity operator ( -1)@xmath244 .
hence , the ground state can be classified as even or odd with respect to the value of @xmath241 .
this symmetry aspect can introduce a constraint on the character of the ground state and on the impurity - host coupling since the @xmath241 value for the impurity is only due to the spin projection while in the host it is due to the combination of the orbital and spin projection .
a direct consequence is that the parity constraint together with the unbalance between the spin at the host and the impurity sites leads to a nonvanishing total projection of the spin and angular momentum with respect to a symmetry axis , e.g. the @xmath245 axis .
it is worth to note that a fixed parity for the impurity spin means that it prefers to point in one direction rather than the other one which is not the case for the host s spin and angular momentum .
thus the presence on the impurity for a fixed @xmath246 will give a nonzero polarization along the quantization axis for every site of the system .
such a property holds for any single impurity with a half - integer spin .
another important consequence of the spin - orbit coupling is that it introduces local quantum fluctuations in the orbital sector even at the sites close to the impurity where the orbital pattern is disturbed .
the spin - orbit term makes the on - site problem around the impurity effectively analogous to the ising model in a transverse field for the orbital sector , with nontrivial spin - orbital entanglement @xcite extending over the impurity neighborhood . in figs .
[ fig:4conf ] and [ fig:2conf ] we report the schematic evolution of the ground state configurations for the cluster of @xmath210 sites , with one - impurity and periodic boundary conditions as a function of increasing spin - orbit coupling .
these patterns have been determined by taking into account the sign and the amplitude of the relevant spatial dependent spin and orbital correlation functions .
the arrows associated to the spin degree of freedom can lie in @xmath23 plane or out - of - plane ( along @xmath170 , chosen to be parallel to the @xmath37 axis ) to indicate the anisotropic spin pattern .
the out - of - plane arrow length is given by the on - site expectation value of @xmath171 while the in - plane arrow length is obtained by computing the square root of the second moment , i.e. , @xmath247 and @xmath248 of the @xmath190 and @xmath249 spin components corresponding to the arrows along @xmath36 and @xmath74 , respectively .
moreover , the in - plane arrow orientation for a given direction is determined by the sign of the corresponding spin - spin correlation function assuming as a reference the orientation of the impurity spin .
the ellipsoid is constructed in the same way as for the zero spin - orbit case above , with the addition of a color map that indicates the strength of the average @xmath239 ( i.e. , red , yellow , green , blue , violet correspond with a growing amplitude of the local spin - orbit correlation function ) .
the scale for the spin - orbit amplitude is set to be in the interval @xmath250 .
the selected values for the ground state evolution are given by the relation ( with @xmath251 ) , @xmath252j_{\rm host}. \label{lambda}\ ] ] the scale is set such that @xmath253 and @xmath254 .
this range of values allows us to explore the relevant physical regimes when moving from @xmath0 to @xmath1 and @xmath255 materials with corresponding @xmath236 being much smaller that @xmath127 , @xmath256 and @xmath257 , respectively . for the performed analysis the selected values of @xmath236 ( [ lambda ] )
are also representative of the most interesting regimes of the ground state as induced by the spin - orbit coupling .
let us start with the quantum af phases qaf@xmath186 , qaf@xmath187 , qfs@xmath221 , qfs@xmath219 , qaf@xmath221 , and qaf@xmath219 . as one can observe the switching on of the spin - orbit coupling ( i.e. , @xmath258 in fig .
[ fig:4conf ] ) leads to anisotropic spin patterns with unequal moments for the in - plane and out - of - plane components . from weak to strong spin - orbit coupling
, the character of the spin correlations keeps being af between the impurity and the neighboring host sites in all the spin directions .
the main change for the spin sector occurs for the planar components . for weak spin - orbit coupling the in - plane spin pattern
is generally af for the whole system in all the spatial directions ( i.e. , @xmath25-af order ) .
further increase of the spin - orbit does not modify qualitatively the character of the spin pattern for the out - of - plane components as long as we do not go to maximal values of @xmath259 where local @xmath214 moments are strongly suppressed . in this limit the dominant tendency of the system is towards formation of the spin - orbital singlets and the spin patterns shown in fig .
[ fig:4conf ] are the effect of the virtual singlet - triplet excitations @xcite . concerning the orbital sector , only for weak spin - orbit coupling around the impurity one can still observe a reminiscence of inactive orbitals as related to the orbital vacancy role at the impurity site in the af phase . such an orbital configuration is quickly modified by increasing the spin - orbit interaction and it evolves into a uniform pattern with almost degenerate orbital occupations in all the directions , and with preferential superpositions of @xmath37 and @xmath32 states associated with dominating @xmath260 and @xmath261 orbital angular components ( flattened ellipsoids along the @xmath37 direction ) .
an exception is the qaf@xmath187 phase with the orbital inactive polaron that is stable up to large spin - orbit coupling of the order of @xmath127 .
and qfs@xmath219 phases for selected increasing values of spin - orbit coupling @xmath262 , see eq .
( [ lambda ] ) .
arrows and ellipsoids indicate the spin - orbital state at a given site @xmath60 .
color map indicates the strength of the average spin - orbit , @xmath263 , i.e. , red , yellow , green , blue , violet correspond to the growing amplitude of the above local correlation function . ] when considering the quantum fm configurations qfm@xmath221 in fig .
[ fig:4conf ] , we observe similar trends in the evolution of the spin correlation functions as obtained for the af states . indeed
, the qfm@xmath36 exhibits a tendency to form fm chains with af coupling for the in - plane components at weak spin - orbit that evolve into more dominant af correlations in all the spatial directions within the host .
interestingly , the spin exchange between the impurity and the neighboring host sites shows a changeover from af to fm for the range of intermediate - to - strong spin - orbit amplitudes . a peculiar response to the spin - orbit coupling
is obtained for the qfs@xmath221 phase , see fig .
[ fig:2conf ] , which showed a frustrated spin pattern around the impurity already in the classical regime , with fm and af bonds .
it is remarkable that due to the close proximity with uniform fm and the af states , the spin - orbit interaction can lead to a dramatic rearrangement of the spin and orbital correlations for such a configuration . at weak spin - orbit coupling ( i.e. , @xmath264 ) the spin - pattern is @xmath26-af and the increased coupling ( @xmath265 ) keeps the @xmath26-af order only for the in - plane components with the exception of the impurity site
. it also modulates the spin moment distribution around the impurity along the @xmath170 direction .
further increase of @xmath236 leads to complete spin polarization along the @xmath170 direction in the host , with antiparallel orientation with respect to the impurity spin .
this pattern is guided by the proximity to the fm phase .
the in - plane components develop a mixed fm - af pattern with a strong @xmath23 anisotropy most probably related to the different bond exchange between the impurity and the host . when approaching the regime of a spin - orbit coupling that is comparable to @xmath266 , the out - of - plane spin components dominate and the only out - of - plane spin polarization is observed at the impurity site .
such a behavior is unique and occurs only in the qfs@xmath36 phases .
the cooperation between the strong spin - orbit coupling and the frustrated host - impurity spin - orbital exchange leads to an effective decoupling in the spin sector at the impurity with a resulting maximal polarization . on the other hand , as for the af states , the most favorable configuration for strong spin - orbit has af in - plane spin correlations .
the orbital pattern for the qfs@xmath36 states evolves similarly to the af cases with a suppression of the active - inactive interplay around the impurity and the setting of a uniform - like orbital configuration with unquenched angular momentum on site and predominant in - plane components .
the response of the fm state is different in this respect as the orbital active states around the impurity are hardly affected by the spin - orbit while the host sites far from the impurity the local spin - orbit coupling is more pronounced . finally , to understand the peculiar evolution of the spin configuration it is useful to consider the lowest order terms in the spin - orbital exchange that couple directly the orbital angular momentum with the spin . taking into account the expression of the spin - orbital exchange in the host
( [ hso ] ) and the expression of @xmath231 one can show that the low energy terms on a bond that get more relevant in the hamiltonian when the spin - orbit coupling makes a non - vanishing local angular momentum . as a result ,
the corresponding expressions are : @xmath267 with positive coefficients @xmath268 and @xmath269 that depend on @xmath270 and @xmath271 ( [ rr ] ) .
a definite sign for the spin exchange in the limit of vanishing spin - orbit coupling is given by the terms which go beyond eq .
( [ ls ] ) . then ,
if the ground state has isotropic fm correlations ( e.g. qfm@xmath36 ) at @xmath207 , the term @xmath272 would tend to favor af - like configurations for the in - plane orbital angular components when the spin - orbit interaction is switched on .
this opposite tendency between the @xmath170 and @xmath273 components is counteracted by the local spin - orbit coupling that prevents to have coexisting fm and af spin - orbital correlations .
such patterns would not allow to optimize the @xmath239 amplitudes .
one way out is to reduce the @xmath245 spin projection and to get planar af correlations in the spin and in the host .
a similar reasoning applies to the af states where the negative sign of the @xmath274 correlations favors fo alignment of the angular momentum components .
as for the previous case , the opposite trend of in- and out - of - plane spin - orbital components is suppressed by the spin - orbit coupling and the in - plane fo correlations for the @xmath275 components leads to fm patterns for the in - plane spin part as well . summarizing , by close inspection of figs .
[ fig:4conf ] and [ fig:2conf ] one finds an interesting evolution of the spin patterns in the quantum phases : + ( i ) for the qaf states ( fig .
[ fig:4conf ] ) , a spin canting develops at the host sites ( i.e. , the relative angle is between 0 and @xmath276 ) while the spins on impurity - host bonds are always af . the canting in the host evolves , sometime in an inhomogeneous way , to become reduced in the strong spin - orbit coupling regime where ferro - like correlations tend to dominate . in this respect , when the impurity is coupled antiferromagnetically to the host it does not follow the tendency to form spin canting . + ( ii ) in the qfm states ( fig .
[ fig:2conf ] ) , at weak spin - orbit one observes spin - canting in the host and for the host - impurity coupling that persists only in the host whereas the spin - orbit interaction is increasing . to probe the phase diagram of the system in presence of the spin - orbit coupling ( @xmath277 ) we solved the same cluster of @xmath210 sites as before along three different cuts in the phase diagram of fig .
[ fig : diag](a ) for three values of @xmath236 , i.e. , small @xmath278 , intermediate @xmath279 , and large @xmath280 .
each cut contained ten points , the cuts were parameterized as follows : ( i ) @xmath281 and @xmath282 , ( ii ) @xmath283 and @xmath282 , and ( iii ) @xmath284 and @xmath285 . in fig .
[ fig : jh](a ) we show the representative spin - orbital configurations obtained for @xmath279 along the first cut shown in fig .
[ fig : jh](b ) .
values of @xmath133 are chosen as @xmath286 with @xmath287 but not all the points are shown in fig .
[ fig : jh](a ) only the ones for which the spin - orbital configuration changes substantially .
the cut starts in the qaf@xmath187 phase , according to the phase diagram of fig .
[ fig : jh](b ) , and indeed we find a similar configuration to the one shown in fig .
[ fig:4conf ] for qaf@xmath187 phase at @xmath288 . moving up in the phase diagram from @xmath289 to @xmath290
we see that the configuration evolves smoothly to the one which we have found in the qaf@xmath221 phase at @xmath288 ( not shown in fig .
[ fig:4conf ] ) .
the evolution of spins is such that the out - of - plane moments are suppressed while in - plane ones are slightly enhanced .
the orbitals become more spherical and the local spin - orbit average , @xmath291 , becomes larger and more uniform , however for the apical site @xmath292 in the cluster [ fig .
[ fig : diag](b ) ] the trend is opposite initially large value of spin - orbit coupling drops towards the uniform value .
the points between @xmath293 and @xmath294 we skip as the evolution is smooth and the trend is clear , however the impurity out - of - plane moment begins to grow above @xmath295 , indicating proximity to the qfs@xmath221 phase . for this phase at intermediate and high @xmath236
the impurity moment is much larger than all the others ( see fig .
[ fig:2conf ] ) . and for a fixed value of spin - orbit coupling @xmath279 along a cut in the phase diagram shown in panel ( b ) , i.e. , for @xmath281 and @xmath296 .
arrows and ellipsoids indicate the spin - orbital state at a given site @xmath60 .
color map indicates the strength of the average spin - orbit , @xmath297 , i.e. , red , yellow , green , blue , violet correspond to the growing amplitude of the above correlation function . ] for @xmath298 the orbital pattern clearly shows that we are in the qfs@xmath221 phase at @xmath288 which agrees with the position of the @xmath294 point in the phase diagram , see fig .
[ fig : jh](b ) . on the other hand ,
moving to the next @xmath133 point upward along the cut eq .
( [ cut ] ) we already observe a configuration which is very typical for the qfm@xmath36 phase at intermediate @xmath236 ( here @xmath299 shown in fig .
[ fig:4conf ] but also @xmath300 , not shown ) .
this indicates that the qfs@xmath221 phase can be still distinguished at @xmath279 and its position in the phase diagram is similar as in the @xmath207 case , i.e. , as an intermediate phase between the qaf@xmath301 and qfm@xmath36 one . finally , we have found that also the two other cuts which were not shown here , i.e. , for @xmath283 and increasing @xmath133 and for @xmath284 and increasing @xmath126 confirm that the overall character of the phase diagram of fig .
[ fig : diag](a ) is preserved at this value of spin - orbit coupling , however firstly , the transitions between the phases are smooth and secondly , the subtle differences between the two qfs@xmath36 , qaf@xmath36 and qaf@xmath37 phases are no longer present .
this also refers to the smaller value of @xmath236 , i.e. , @xmath278 , but already for @xmath280 the out - of - plane moments are so strongly suppressed ( except for the impurity moment in the qfs@xmath221 phase ) and the orbital polarization is so weak ( i.e. , almost spherical ellipsoids ) that typically the only distinction between the phases can be made by looking at the in - plane spin correlations and the average spin - orbit , @xmath297 . in this limit
we conclude that the phase diagram is ( partially ) melted by large spin - orbit coupling but for lower values of @xmath236 it is still valid .
we have derived the spin - orbital superexchange model for @xmath2 impurities replacing @xmath4 ( or @xmath3 ) ions in the @xmath1 ( @xmath0 ) host in the regime of mott insulating phase .
although the impurity has no orbital degree of freedom , we have shown that it contributes to the spin - orbital physics and influences strongly the orbital order .
in fact , it tends to project out the inactive orbitals at the impurity - host bonds to maximize the energy gain from virtual charge fluctuations . in this case
the interaction along the superexchange bond can be either antiferromagnetic or ferromagnetic , depending on the ratio of hund s exchange coupling at impurity ( @xmath302 ) and host ( @xmath303 ) ions and on the mismatch @xmath125 between the @xmath0 and @xmath1 atomic energies , modified by the difference in hubbard @xmath304 s and hund s exchange @xmath305 s at both atoms .
this ratio , denoted @xmath133 ( [ eq : etai ] ) , replaces here the conventional parameter @xmath306 often found in the spin - orbital superexchange models of undoped compounds ( e.g. , in the kugel - khomskii model for kcuf@xmath9 @xcite ) where it quantifies the proximity to ferromagnetism . on the other hand ,
if the overall coupling between the host and impurity is weak in the sense of the total superexchange , @xmath126 , with respect to the host value , @xmath127 , the orbitals being next to the impurity may be forced to stay inactive which modifies the magnetic properties in such cases the impurity - host bond is always antiferromagnetic . as we have seen in the case of a single impurity , the above two mechanisms can have a nontrivial effect on the host , especially if the host itself is characterized by frustrated interactions , as it happens in the parameter regime where the @xmath26-af phase is stable .
for this reason we have focused mostly on the latter phase of the host and we have presented the phase diagrams of a single impurity configuration in the case when the impurity is doped on the sublattice where the orbitals form a checkerboard pattern with alternating @xmath37 and @xmath36 orbitals occupied by doublons .
the diagram for the @xmath37-sublattice doping shows that in some sense the impurity is never weak , because even for a very small value of @xmath175 it can release the host s frustration around the impurity site acting as an orbital vacancy . on the other hand , for the @xmath36-sublattice doping when the impurity - host coupling is weak , i.e. , either @xmath175 is weak or @xmath133 is close to @xmath191 , we have identified an interesting quantum mechanism releasing frustration of the impurity spin ( that can not be avoided in the purely classical approach ) .
it turned out that in such situations the orbital flips in the host make the impurity spin polarize in such a way that the @xmath26-af order of the host is completely restored .
the cases of the periodic doping studied in this paper show that the host s order can be completely altered already for _ rather low _ doping of @xmath51 , even if the @xmath175 is small . in this case
we can stabilize a ferrimagnetic type of phase with a four - site unit cell having magnetization @xmath307 , reduced further by quantum fluctuations .
we have established that the only parameter range where the host s order remains unchanged is when @xmath133 is close to @xmath191 and @xmath308 .
the latter value is very surprising as it means that the impurity - host coupling must be large enough to keep the host s order unchanged this is another manifestation of the orbital vacancy mechanism that we have already observed for a single impurity .
also in this case the impurity spins are fixed with the help of orbital flips in the host that lift the degeneracy which arises in the classical approach .
we would like to point out that the quantum mechanism that lifts the ground state degeneracy mentioned above and the role of quantum fluctuations are of particular interest for the periodically doped checkerboard systems with @xmath309 doping which is a challenging problem for future research .
from the point of view of generic , i.e. , non - periodic doping , the most representative cases are those of a doping which is incommensurate with the two - sublattice spin - orbital pattern . to uncover the generic rules in such cases , we have studied periodic @xmath52 and @xmath53 doping .
one finds that when the period of the impurity positions does not match the period of 2 for both the spin and orbital order of the host , interesting novel types of order emerge .
in such cases the elementary cell must be doubled in both lattice directions which clearly gives a chance of realizing more phases than in the case of commensurate doping .
our results show that indeed , the number of phases increases from 4 to 7 and the host s order is altered in each of them .
quite surprisingly , the overall character of the phase diagram remained unchanged with respect to the one for @xmath51 doping and , if we ignore the differences in configuration , it seems that only some of the phases got divided into two versions differing either by the spin bond s polarizations around impurities ( phases around @xmath191 ) , or by the character of the orbitals around the impurities ( phases with inactive orbitals in the limit of small enough product @xmath310 , versus phases with active orbitals in the opposite limit ) .
orbital polarization in this latter region resembles orbital polarons in doped manganites @xcite also here such states are stabilized by the double exchange @xcite .
a closer inspection of underlying phases reveals however a very interesting degeneracy of the impurity spins at @xmath52 that arises again from the classical approach but this time it can not be released by short - range orbital flips .
this happens because the host s order is already so strongly altered that it is no longer anisotropic ( as it was the case of the @xmath26-af phase ) and there is no way to restore the orbital anisotropy around the impurities that could lead to spin - bonds imbalance and polarize the spin . in the case of lower @xmath53 doping such an effect
is absent and the impurity spins are always polarized , as it happens for @xmath51 .
it shows that this is rather a peculiarity of the @xmath52 periodic doping . indeed
, one can easily notice that for @xmath52 every atom of the host is a nearest neighbor of some impurity .
in contrast , for @xmath51 we can find three host s atoms per unit cell which do not neighbor any impurity and for @xmath53 there are sixteen of them .
for this reason the impurity effects are amplified for @xmath52 which is not unexpected although one may find somewhat surprising that the ground state diagrams for the lowest and the highest doping considered here are very similar .
this suggests that the cooperative effects of multiple impurities are indeed not very strong in the low - doping regime , so the diagram obtained for @xmath53 can be regarded as generic for the dilute doping regime with uniform spatial profile .
for the representative case of @xmath51 doping , we have presented the consequences of quantum effects beyond the classical approach .
spin fluctuations are rather weak for the considered case of large @xmath31 and @xmath33 spins , and we have shown that orbital fluctuations on superexchange bonds are more important .
they are strongest in the regime of antiferromagnetic impurity - host coupling ( which suggests importance of entangled states @xcite ) and enhance the tendency towards frustrated impurity spin configurations but do not destroy other generic trends observed when the parameters @xmath133 and @xmath175 increase .
increasing spin - orbit coupling leads to qualitative changes in the spin - orbital order . when hund s exchange is small at the impurity sites , the antiferromagnetic bonds around it have reduced values of spin - orbit coupling term , but the magnetic moments reorient and survive in the @xmath32 planes , with some similarity to the phenomena occurring in the perovskite vanadates @xcite .
this quenches the magnetic moments at @xmath0 impurities and leads to almost uniform orbital occupancies at the host sites .
in contrast , frustration of impurity spins is removed and the impurity magnetization along the @xmath37 axis survives for large spin - orbit coupling .
we would like to emphasize that the _ orbital dilution _ considered here influences directly the orbital degrees of freedom in the host around the impurities .
the synthesis of hybrid compounds having both @xmath0 and @xmath1 transition metal ions will likely open a novel route for unconventional effects in complex materials .
there are several reasons for expecting new scenarios in mixed @xmath5 spin - orbital - lattice materials , and we pointed out only some of them . on the experimental side ,
the changes of local order could be captured using inelastic neutron scattering or resonant inelastic x - ray scattering ( rixs ) .
in fact , using rixs can also bring an additional advantage : rixs , besides being a perfect probe of both spin and orbital excitations , can also ( indirectly ) detect the nature of orbital ground state ( supposedly also including the nature of impurities in the crystal ) @xcite . unfortunately , there are no such experiments yet but we believe that they will be available soon
. short range order around impurities could be investigated by the excitation spectra at the resonant edges of the substituting atoms .
taking them both at finite energy and momentum can dive insights into the nature of the short range order around the impurity and then unveil information of the order within the host as well . even if there are no elastic superlattice extra peaks one can expect that the spin - orbital correlations will emerge in the integrated rixs spectra providing information of the impurity - host coupling and of the short range order around the impurity .
even more interesting is the case where the substituting atom forms a periodic array with small deviation from the perfect superlattice when one expects the emergence of extra elastic peaks which will clearly indicate the spin - orbital reconstruction . in our case an active orbital diluted site can not participate coherently in the host spin - orbital order but rather may to restructure the host ordering @xcite . at dilute impurity concentration
we may expect broad peaks emerging at finite momenta in the brillouin zone , indicating the formation of coherent islands with short range order around impurities .
we also note that local susceptibility can be suitably measured by making use of resonant spectroscopies ( e.g. nuclear magnetic resonance ( nmr ) , electron spin resonance ( esr ) , nuclear quadrupole resonance ( nqr ) , muon spin resonance ( @xmath311sr ) , _ etcetera _ ) that exploit the different magnetic or electric character of the atomic nuclei for the impurity and the host in the hybrid system .
finally , the random implantation of the muons in the sample can provide information of the relaxation time in different domains with unequal dopant concentration which may be nonuniform .
for the given problem the differences in the resonant response can give relevant information about the distribution of the local fields , the occurrence of local order and provide access to the dynamical response within doped domains .
the use of local spectroscopic resonance methods has been widely demonstrated to be successful when probing the nature and the evolution of the ground state in the presence of spin vacancies both for ordered and disordered magnetic configurations @xcite . in summary , this study highlights the role of spin defects which lead to orbital dilution in spin - orbital systems .
using an example of @xmath2 impurities in a @xmath4 ( or @xmath3 ) host we have shown that impurities change radically the spin - orbital order around them , independently of the parameter regime . as a general feature
we have found that doped @xmath2 ions within the host with spin - orbital order have frustrated spins and polarize the orbitals of the host when the impurity - host exchange as well as hund s exchange at the impurity are both sufficiently large .
this remarkable trend is independent of doping and is expected to lead to global changes of spin - orbital order in doped materials . while the latter effect is robust , we argue that the long - range spin fluctuations resulting from the translational invariance of the system will likely prevent the ground state from being macroscopically degenerate , so if the impurity spins in one unit cell happens to choose its polarization then the others will follow . on the contrary , in the regime of weak hund s exchange @xmath2 ions
act not only as spin defects which order antiferromagnetically with respect to their neighbors , but also induce doublons in inactive orbitals .
finally , we remark that this behavior with switching between inactive and active orbitals by an orbitally neutral impurity may lead to multiple interesting phenomena at macroscopic doping when global modifications of the spin - orbital order are expected to occur .
most of the results were obtained in the classical approximation but we have shown that modifications due to spin - orbit coupling do not change the main conclusion .
we note that this generic treatment and the general questions addressed here , such as the release of frustration for competing spin structures due to periodic impurities , are relevant to double perovskites @xcite . while the local orbital polarization should be similar , it is challenging to investigate disordered impurities , both theoretically and in experiment , to find out whether their influence on the global spin - orbital order in the host is equally strong .
we thank maria daghofer and krzysztof wohlfeld for insightful discussions . w. b. and a. m. o. kindly acknowledge support by the polish national science center ( ncn ) under project no .
2012/04/a / st3/00331 .
w. b. was also supported by the foundation for polish science ( fnp ) within the start program .
m. c. acknowledges funding from the eu fp7/2007 - 2013 under grant agreement no .
264098 mama .
here we present the details of the derivation of the low energy spin - orbital hamiltonian for the @xmath54 bonds around the impurity at site @xmath60 .
@xmath312 , which follows from the perfurbation theory , as given in eq .
( [ eq : pert_exp ] ) . here
we consider a single @xmath54 bond @xmath56 .
two contributions to the effective hamiltonian follow from charge excitations : ( i ) @xmath313 due to @xmath314 , and ( ii ) @xmath315 due to @xmath316 .
therefore the low energy hamiltonian is , @xmath317 consider first the processes which conserve the number of doubly occupied orbitals , @xmath314 . then by means of spin and orbital projectors , it is possible to express @xmath313 for @xmath78 and @xmath79 as @xmath318 with the excitation energy @xmath125 defined in eq .
( [ delta ] ) .
the resulting effective @xmath5 exchange in eq .
( [ eq : s - ex ] ) consists of three terms : ( i ) the first one does not depend on the orbital configuration of the @xmath1 atom and it can be fm or af depending on the values @xmath125 and the hund s exchange on the @xmath0 ion
. in particular , if @xmath125 is the largest or the smallest energy scale , the coupling will be either af or fm , respectively .
( ii ) the second term has an explicit dependence on the occupation of the doublon on the @xmath1 atom via the projecting operator @xmath106 .
this implies that a magnetic exchange is possible only if the doublon occupies the inactive orbital for a bond along a given direction @xmath69 .
unlike in the first term , the sign of this interaction is always positive favoring an af configuration at any strength of @xmath125 and @xmath319 .
( iii ) finally , the last term describes the effective processes which do not depend on the spin states on the @xmath0 and @xmath1 atoms .
this contribution is of pure orbital nature , as it originates from the hopping between @xmath0 and @xmath1 atoms without affecting their spin configuration , and for this reason favors the occupation of active @xmath7 orbitals along the bond by the doublon . within the same scheme
, we have derived the effective spin - orbital exchange that originates from the charge transfer processes of the type @xmath320 , @xmath321 .
the effective low - energy contribution to the hamiltonian for @xmath78 and @xmath79 reads @xmath322 by inspection of the spin structure involved in the elemental processes that generate @xmath321 , one can note that it is always af independently of the orbital configuration on the @xmath1 atom exhibiting with a larger spin - exchange and an orbital energy gain if the doublon is occupying the inactive orbital along a given bond .
we have verified that the amplitude of the exchange terms in @xmath321 is much smaller than the ones which enter in @xmath323 which justifies that one may simplify eq .
( [ eq : hj ] ) for @xmath78 and @xmath79 to @xmath324 and neglect @xmath321 terms altogether .
this approximation is used in sec .
[ sec : model ] .
the starting point to express the orbital operators appearing in the spin - orbital superexchange model ( [ fullh ] ) is the relation between quenched @xmath325 , @xmath326 , and @xmath327 orbitals at site @xmath60 and the eigenvectors @xmath328 , @xmath329 , and @xmath330 of the angular momentum operator @xmath331 .
these are known to be @xmath332 from this we can immediately get the occupation number operators for the doublon , @xmath333 and the related @xmath334 operators , @xmath335 the doublon hopping operators have a slightly different structure that reflects their noncommutivity , i.e. , @xmath336 these relations are sufficient to write the superexchange hamiltonian for the host - host and impurity - host bonds in the @xmath337 operator basis for the orbital part .
however , in practice it is more convenient to work with real operators @xmath338 rather than with the original ones , @xmath337 .
thus we write the final relations which we used for the numerical calculations in terms of these operators , @xmath339+\frac12\left(l^z_i\right)^2,\nonumber \\
d_i^{(b ) } & = & \hskip .2 cm \frac{1}{4}\left[\left(l^+_i\right)^{2 } + \left(l^-_i\right)^{2}\right]+\frac12\left(l^z_i\right)^2,\nonumber \\
d_i^{(c ) } & = & 1-\left(l^z_i\right)^{2},\end{aligned}\ ] ] for the doublon occupation numbers and going directly to the orbital @xmath340 operators we find that , @xmath341-\frac{i}{2}l^{z}_i,\end{aligned}\ ] ] for the off - diagonal part and @xmath342+\frac{3}{4}\left(l^{z}_i\right)^2 -\frac{1}{2},\nonumber \\ \tau^{z(b)}_i & = & \frac{1}{8}\left[\left(l^{+}_i\right)^{2 } + \left(l^{-}_i\right)^2\right]-\frac{3}{4}\left(l^{z}_i\right)^2 + \frac{1}{2},\nonumber \\ \tau^{z(c)}_i & = & -\frac{1}{4}\left[\left(l^{+}_i\right)^{2 } + \left(l^{-}_i\right)^{2}\right],\end{aligned}\ ] ] for the diagonal one . note that the complex phase in @xmath343 and @xmath344 is irrelevant and can be omitted here as @xmath345 is always accompanied by @xmath346 on a neighboring site .
this is a consequence of the cubic symmetry in the orbital part of the superexchange hamiltonian and it can be altered by a presence of a distortion , e.g. , octahedral rotation . for completeness we also give the backward relation between angular momentum components , @xmath347 with @xmath348 , and the orbital operators @xmath349 ; these are : @xmath350 k. yamada , c. h. lee , k. kurahashi , j. wada , s. wakimoto , s. ueki , h. kimura , y. endoh , s. hosoya , g. shirane , r. j. birgeneau , m. greven , m. a. kastner , and y. j. kim , _ doping dependence of the spatially modulated dynamical spin correlations and the superconducting - transition temperature in _
la@xmath29sr@xmath15cuo@xmath12 , phys .
b * 57 * , 6165 ( 1998 ) .
q. li , k. e. gray , s. b. wilkins , m. garcia fernandez , s. rosenkranz , h. zheng , and j. f. mitchell , _ prediction and experimental evidence for thermodynamically stable charged orbital domain walls _ , phys . rev .
x * 4 * , 031028 ( 2014 ) .
a. m. ole , g. khaliullin , p. horsch , and l. f. feiner , _ fingerprints of spin - orbital physics in cubic mott insulators : magnetic exchange interactions and optical spectral weights _ ,
b * 72 * , 214431 ( 2005 ) .
s. nakatsuji , k. kuga , k. kimura , r. satake , n. katayama , e. nishibori , h. sawa , r. ishii , m. hagiwara , f. bridges , t. u. ito , w. higemoto , y. karaki , m. halim , a. a. nugroho , j. a. rodriguez - rivera , m. a. green , and c. broholm , _ spin - orbital short - range order on a honeycomb - based lattice _
, science * 336 * , 559 ( 2012 ) .
j. a. quilliam , f. bert , e. kermarrec , c. payen , c. guillot - deudon , p. bonville , c. baines , h. luetkens , and p. mendels , _ singlet ground state of the quantum antiferromagnet _
ba@xmath9cusb@xmath10o@xmath11 , phys .
lett . * 109 * , 117203 ( 2012 ) . v. fritsch , j. hemberger , n. bttgen , e .- w .
scheidt , h .- a .
krug von nidda , a. loidl , and v. tsurkan , _ spin and orbital frustration in _
mnsc@xmath10s@xmath10 _ and _ fesc@xmath10s@xmath12 , phys .
lett . * 92 * , 116401 ( 2004 ) .
l. mittelstdt , m. schmidt , z. wang , f. mayr , v. tsurkan , p. lunkenheimer , d. ish , l. balents , j. deisenhofer , and a. loidl , _ spin - orbiton and quantum criticality in _ fesc@xmath10s@xmath12 , phys .
b * 91 * , in press ( 2015 ) .
f. vernay , k. penc , p. fazekas , and f. mila , _ orbital degeneracy as a source of frustration in _ linio@xmath10 ,
b * 70 * , 014428 ( 2004 ) .
f. reynaud , d. mertz , f. celestini , j. debierre , a. m. ghorayeb , p. simon , a. stepanov , j. voiron , and c. delmas , _ orbital frustration at the origin of the magnetic behavior in _
linio@xmath10 , phys .
86 * , 3638 ( 2001 ) .
j. geck , p. wochner , s. kiele , r. klingeler , p. reutler , a. revcolevschi , and b. bchner , _
orbital polaron lattice formation in lightly doped _
la@xmath14sr@xmath15mno@xmath9 , phys .
* 95 * , 236401 ( 2005 ) .
y. m. sheu , s. a. trugman , l. yan , j. qi , q. x. jia , a. j. taylor , and r. p. prasankumar , _ polaronic transport induced by competing interfacial magnetic order in a _ la@xmath21ca@xmath22mno@xmath9/bifeo@xmath9 _ heterostructure _ , phys . rev .
x * 4 * , 021001 ( 2014 ) .
h. sakai , s. ishiwata , d. okuyama , a. nakao , h. nakao , y. murakami , y. taguchi , and y. tokura , _ electron doping in the cubic perovskite _
srmno@xmath9 : _ isotropic metal versus chainlike ordering of jahn - teller polarons _ ,
b * 82 * , 180409 ( 2010 ) . c. ulrich , g. khaliullin , j. sirker , m. reehuis , m. ohl , s. miyasaka , y. tokura , and b. keimer , _ magnetic neutron scattering study of _ yvo@xmath9 : _ evidence for an orbital peierls state _ , phys . rev .
* 91 * , 257202 ( 2003 ) .
m. reehuis , c. ulrich , k. proke , s. mata , j. fujioka , s. miyasaka , y. tokura , and b. keimer , _ structural and magnetic phase transitions of the orthovanadates
_ @xmath351vo@xmath9 ( @xmath351=dy , ho , er ) _ as seen via neutron diffraction _
b * 83 * , 064404 ( 2011 ) .
j. fujioka , t. yasue , s. miyasaka , y. yamasaki , t. arima , h. sagayama , t. inami , k. ishii , and y. tokura , _ critical competition between two distinct orbital - spin ordered states in perovskite vanadates _
b * 82 * , 144425 ( 2010 ) .
j. fujioka , s. miyasaka , and y. tokura , _ doping variation of anisotropic charge and orbital dynamics in _
y@xmath14ca@xmath15vo@xmath9 : _ comparison with _
la@xmath14sr@xmath15vo@xmath9 , phys .
b * 77 * , 144402 ( 2008 ) .
a. avella , p. horsch , and a. m. ole , _ defect states and excitations in a mott insulator with orbital degrees of freedom : mott - hubbard gap versus optical and transport gaps in doped systems _ ,
b * 87 * , 045132 ( 2013 ) .
t. mizokawa , l. h. tjeng , g. a. sawatzky , g. ghiringhelli , o. tjernberg , n. b. brookes , h. fukazawa , s. nakatsuji , and y. maeno , _ spin - orbit coupling in the mott insulator _ ca@xmath10ruo@xmath12 , phys .
87 * , 077202 ( 2001 ) .
j. s. lee , y. s. lee , t. w. noh , s .-
oh , j. yu , s. nakatsuji , h. fukazawa , and y. maeno , _ electron and orbital correlations in _
ca@xmath29sr@xmath15ruo@xmath12 _ probed by optical spectroscopy _ , phys .
lett . * 89 * , 257402 ( 2002 ) . c. n. veenstra , z .- h .
zhu , m. raichle , b. m. ludbrook , a. nicolaou , b. slomski , g. landolt , s. kittaka , y. maeno , j. h. dil , i. s. elfimov , m. w. haverkort , and a. damascelli , _ spin - orbital entanglement and the breakdown of singlets and triplets in _ sr@xmath10ruo@xmath12 _ revealed by spin- and angle - resolved photoemission spectroscopy _ , phys .
lett . * 112 * , 127002 ( 2014 ) .
t. f. qi , o. b. korneta , s. parkin , l. e. de long , p. schlottmann , and g. cao , _ negative volume thermal expansion via orbital and magnetic orders in _ ca@xmath10ru@xmath14cr@xmath15o@xmath12 ( @xmath35 ) , phys .
lett . * 105 * , 177203 ( 2010 ) .
g. cao , s. chikara , x. n. lin , e. elhami , v. durairaj , and p. schlottmann , _ itinerant ferromagnetism to insulating antiferromagnetism : a magnetic and transport study of single crystal _
srru@xmath14mn@xmath15o@xmath9 ( @xmath353 ) , phys .
b * 71 * , 035104 ( 2005 ) .
m. a. hossain , b. bohnenbuck , y. d. chuang , m. w. haverkort , i. s. elfimov , a. tanaka , a. g. cruz gonzalez , z. hu , h .- j .
lin , c. t. chen , r. mathieu , y. tokura , y. yoshida , l. h. tjeng , z. hussain , b. keimer , g. a. sawatzky , and a. damascelli , _ mott versus slater - type metal - insulator iransition in mn - substituted _
sr@xmath9ru@xmath10o@xmath354 , phys .
b * 86 * , 041102(r ) ( 2012 ) .
m. a. hossain , z. hu , m. w. haverkort , t. burnus , c. f. chang , s. klein , j. d. denlinger , h .- j .
lin , c. t. chen , r. mathieu , y. kaneko , y. tokura , s. satow , y. yoshida , h. takagi , a. tanaka , i. s. elfimov , g. a. sawatzky , l. h. tjeng , and a. damascelli , _ crystal - field level inversion in lightly mn - doped _
sr@xmath9ru@xmath10o@xmath354 , phys .
* 101 * , 016404 ( 2008 ) .
d. mesa , f. ye , s. chi , j. a. fernandez - baca , w. tian , b. hu , r. jin , e. w. plummer , and j. zhang , _ single - bilayer e - type antiferromagnetism in mn - substituted _ sr@xmath9ru@xmath10o@xmath354 : _ neutron scattering study _ ,
b * 85 * , 180410(r ) ( 2012 ) .
s. a. j. kimber , j. a. rodgers , h. wu , c. a. murray , d. n. argyriou , a. n. fitch , d. i. khomskii , and j. p. attfield , _ metal - insulator transition and orbital order in _
pbruo@xmath9 , phys .
* 102 * , 046409 ( 2009 ) .
m. a. hossain , i. zegkinoglou , y .- d .
chuang , j. geck , b. bohnenbuck , a. g. cruz gonzalez , h .- h .
wu , c. schler - langeheine , d. g. hawthorn , j. d. denlinger , r. mathieu , y. tokura , s. satow , h. takagi , y. yoshida , z. hussain , b. keimer , g. a. sawatzky , and a. damascelli , _ electronic superlattice revealed by resonant scattering from random impurities in _
sr@xmath9ru@xmath10o@xmath354 , sci . rep . * 3 * , 2299 ( 2013 ) .
l. limot , p. mendels , g. collin , c. mondelli , b. ouladdiaf , h. mutka , n. blanchard , and m. mekata , _ susceptibility and dilution effects of the kagom bilayer geometrically frustrated network : a ga nmr study of _
srcr@xmath355ga@xmath356o@xmath357 , phys .
b * 65 * , 144447 ( 2002 ) .
j. bobroff , n. laflorencie , l. k. alexander , a. v. mahajan , b. koteswararao , and p. mendels , _ impurity - induced magnetic order in low - dimensional spin - gapped materials _
rev . lett . * 103 * , 047201 ( 2009 ) .
p. bonf , p. carretta , s. sanna , g. lamura , g. prando , a. martinelli , a. palenzona , m. tropeano , m. putti , and r. de renzi , _ magnetic properties of spin - diluted iron pnictides from @xmath311sr and nmr in _
lafe@xmath358ru@xmath15aso , phys .
b * 85 * , 054518 ( 2012 ) . a. k. paul , m. reehuis , v. ksenofontov , b. yan , a. hoser , d. m. tbbens , p. m. abdala , p. adler , m. jansen , and c. felser , _ lattice instability and competing spin structures in the double perovskite insulator _ sr@xmath10feoso@xmath359 , phys .
rev . lett . * 111 * , 167205 ( 2013 ) . | we investigate the changes in spin and orbital patterns induced by magnetic transition metal ions without an orbital degree of freedom doped in a strongly correlated insulator with spin - orbital order . in this context
we study the @xmath0 ion substitution in @xmath1 transition metal oxides in the case of @xmath2 doping at either @xmath3 or @xmath4 sites which realizes _ orbital dilution _ in a mott insulator .
although we concentrate on this doping case as it is known experimentally and more challenging than other oxides due to finite spin - orbit coupling , the conclusions are more general .
we derive the effective @xmath5 ( or @xmath6 ) superexchange in a mott insulator with different ionic valencies , underlining the emerging structure of the spin - orbital coupling between the impurity and the host sites and demonstrate that it is qualitatively different from that encountered in the host itself .
this derivation shows that the interaction between the host and the impurity depends in a crucial way on the type of doubly occupied @xmath7 orbital .
one finds that in some cases , due to the quench of the orbital degree of freedom at the @xmath0 impurity , the spin and orbital order within the host is drastically modified by doping .
the impurity acts either as a spin defect accompanied by an orbital vacancy in the spin - orbital structure when the host - impurity coupling is weak , or it favors doubly occupied active orbitals ( orbital polarons ) along the @xmath5 bond leading to antiferromagnetic or ferromagnetic spin coupling .
this competition between different magnetic couplings leads to quite different ground states .
in particular , for the case of a finite and periodic @xmath0 atom substitution , it leads to striped patterns either with alternating ferromagnetic / antiferromagnetic domains or with islands of saturated ferromagnetic order .
we find that magnetic frustration and spin degeneracy can be lifted by the quantum orbital flips of the host but they are robust in special regions of the incommensurate phase diagram .
orbital quantum fluctuations modify quantitatively spin - orbital order imposed by superexchange .
in contrast , the spin - orbit coupling can lead to anisotropic spin and orbital patterns along the symmetry directions and cause a radical modification of the order imposed by the spin - orbital superexchange .
our findings are expected to be of importance for future theoretical understanding of experimental results for @xmath1 transition metal oxides doped with @xmath2 ions .
we suggest how the local or global changes of the spin - orbital order induced by such impurities could be detected experimentally . |
planar cellular structures formed by tessellation , tiling , or subdivision of plane into contiguous and nonoverlapping cells have always generated interest among scientists in general and physicist in particular because cellular structures are ubiquitous in nature .
examples include acicular texture in martensite growth , tessellated pavement on ocean shores , agricultural land division according to ownership , grain texture in polycrystals , cell texture in biology , soap froths and so on @xcite .
for instance , voronoi lattice formed by partitioning of a plane into convex polygons or apollonian packing generated by tiling a plane into contiguous and non - overlapping disks has found widespread applications @xcite .
there are some theoretical models , both random and deterministic , developed either to directly mimic these structures or to serve as a tool on which one can study various physical problems @xcite .
however , cells in most of the existing structures do not have sides which share with more than one side of other cells . in reality
, the sides of a cell can share a part or a whole side of another cell . as a result
, the number of neighbours of a cell can be higher than the number of sides it has .
moreover , cellular structure may emerge through evolution where cells can be of different sizes and have different number of neighbours since nature favours these properties as a matter of rule rather than exception .
a lattice with such properties can be of great interest as it can mimic disordered medium on which one can study percolation or random walk like problems . in this article
, we propose a weighted planar stochastic lattice ( wpsl ) as a space - filling cellular structure where annealed coordination number disorder and size disorder are introduced in a natural way .
the definition of the model is trivially simple .
it starts with an initiator , say a square of unit area , and a generator that divides it randomly into four blocks .
the generator thereafter is sequentially applied over and over again to only one of the available blocks picked preferentially with respect to their areas .
it results in the partitioning of the square into ever smaller mutually exclusive rectangular blocks .
a snapshot of the wpsl at late stage ( figure 1 ) provides an awe - inspiring perspective on the emergence of an intriguing and rich pattern of blocks .
we intend to investigate its topological and geometrical properties in an attempt to find some order in this seemingly disordered lattice .
if the blocks of the wpsl are regarded as isolated fragments then the model can also describe the fragmentation of a @xmath10d object by random sequential nucleation of seeds from which two orthogonal cracks parallel to the sides of the parent object are grown until intercepted by existing cracks @xcite .
in reality , fragments produced in fracture of solids by propagation of interacting cracks is a formidable mathematical problem .
the model in question can , however , be considered as the minimum model which should be capable of capturing the essential features of the underlying mechanism .
the wpsl can also describe the martensite formation as we find its definition remarkably similar to the model proposed by rao _
et al . _ which is also reflected in the similarity between the figure 1 and figure 2 of @xcite .
yet another application , perhaps a little exotic , is the random search tree problem in computer science @xcite .
searching for an order in the disorder is always an attractive proposition in physics . to this end
, we invoke the concept of complex network topology to quantify the coordination number disorder and the idea of multifractality to quantify the size disorder of the blocks in wpsl .
it is interesting to note that the dual of the wpsl ( dwpsl ) obtained by replacing each block with a node at its center and common border between blocks with an edge joining the two corresponding vertices emerges as a network .
the area of the respective blocks is assigned to the corresponding nodes to characterize them as their fitness parameter .
nodes in the dwpsl , therefore , are characterized by their respective fitness parameter and the corresponding degree @xmath8 defined as the number of links a node has . for a decade , there has been a surge of interest in finding the degree distribution @xmath7 triggered by the work of a .-
barabasi and his co - workers who have revolutionized the notion of the network theory by recognizing the fact that real networks are not static rather grow by addition of new nodes establishing links preferentially , known as the preferential attachment ( pa ) rule , to the nodes that are already well connected @xcite . incorporating both
the ingredients , growth and the pa rule , barabasi and albert ( ba ) presented a simple theoretical model and showed that such network self - organizes into a power - law degree distribution @xmath5 with @xmath11 @xcite .
the phenomenal success of the ba model lies in the fact that it can capture , at least qualitatively , the key features of many real life networks .
interestingly , we find that the dwpsl has all the ingredients of the ba model and its degree distribution @xmath7 follows heavy - tailed power - law but with exponent @xmath6 revealing that the coordination number of the wpsl is scale - free in character .
in addition to characterizing the blocks of the wpsl by the coordination number @xmath8 , they can also be characterized by their respective length @xmath12 and width @xmath13 .
we then find that the dynamics of the wpsl is governed by infinitely many conservation laws , namely the quantity @xmath14 remains independent of time @xmath15 @xmath16 where blocks are labbeled by index @xmath17 .
for instance , total area is one obvious conserved quantity obtained by setting @xmath18 , sum of the cubic power of the length ( or width ) of all the existing blocks @xmath19 is a non - trivial conserved quantity obtained by setting @xmath20 ( or @xmath21 ) .
interestingly , we find that when the @xmath4th block is populated with the fraction of the measure @xmath9 then the distribution of the population in the wpsl emerges as multifractal indicating further development towards gaining deeper insight into the complex nature of the wpsl we proposed .
multifractal analysis was initially proposed to treat turbulence but later successfully applied in a wide range of exciting field of research @xcite . recently though it has received a renewed interest as it has been found that the wild fluctuations of the wave functions in the vicinity of the anderson and the quantum hall transition can be best quantified by using a multifractal analysis @xcite .
the organization of this paper is as follows . in section 2
, we give its exact algorithm of the model . in section 3 various structural topological properties of the wpsl and its dual
are discussed in order to quantify the annealed coordination number disorder . in section 4
we discuss the geometric properties of the wpsl in an attempt to quantify the annealed size disorder .
finally , section 5 gives a short summary of our results .
perhaps an exact algorithm can provide a better description of the model than the mere definition . in step one ,
the generator divides the initiator , say a square of unit area , randomly into four smaller blocks .
we then label the four newly created blocks by their respective areas @xmath22 and @xmath23 in a clockwise fashion starting from the upper left block ( see figure 2 ) . in each step
thereafter only one block is picked preferentially with respect to their respective area ( which we also refer as the fitness parameter ) and then it is divided randomly into four blocks . in general
, the @xmath24th step of the algorithm can be described as follows .
( i ) subdivide the interval @xmath25 $ ] into @xmath26 subintervals of size @xmath27 $ ] , @xmath28$],@xmath29 , @xmath30 $ ] each of which represents the blocks labelled by their areas @xmath31 respectively .
( ii ) generate a random number @xmath32 from the interval @xmath25 $ ] and find which of the @xmath33 sub - interval contains this @xmath32 .
the corresponding block it represents , say the @xmath34th block of area @xmath35 , is picked .
( iii ) calculate the length @xmath36 and the width @xmath37 of this block and keep note of the coordinate of the lower - left corner of the @xmath34th block , say it is @xmath38 .
( iv ) generate two random numbers @xmath39 and @xmath40 from @xmath41 $ ] and @xmath42 $ ] respectively and hence the point @xmath43 mimics a random point chosen in the block @xmath34 .
( v ) draw two perpendicular lines through the point @xmath43 parallel to the sides of the @xmath34th block in order to divide it into four smaller blocks .
the label @xmath35 is now redundant and hence it can be reused .
( vi ) label the four newly created blocks according to their areas @xmath35 , @xmath44 , @xmath45 and @xmath46 respectively in a clockwise fashion starting from the upper left corner . (
vii ) increase time by one unit and repeat the steps ( i ) - ( vi ) _ ad infinitum_. with exactly @xmath47 neighbours as a function of time .
, width=321,height=207 ]
we first focus our analysis on the blocks of the wpsl and their coordination numbers .
note that for square lattice , the deterministic counterpart of the wpsl , the coordination number is a constant equal to @xmath48 .
however , the coordination number in the wpsl is neither a constant nor has a typical mean value , rather the coordination number that each block assumes in the wpsl is random .
moreover , it is allowed to evolve with time and hence the coordination number disorder in the wpsl can be regarded as of annealed type . defining each step of the algorithm as one time unit and imposing periodic boundary condition in the simulation
, we find that the number of blocks @xmath49 which have coordination number @xmath8 ( or @xmath8 neighbours ) continue to grow linearly with time @xmath50 ( see figure 3 ) . on the other hand , the number of total blocks @xmath51 at time @xmath52 in the lattice also grow linearly with time @xmath53 which we can write @xmath54 in the asymptotic regime .
the ratio of the two quantities @xmath55 that describes the fraction of the total blocks which have coordination number @xmath8 is @xmath56 .
it implies that @xmath57 becomes a global property since we find it independent of time and size of the lattice .
we now take wpsl of fixed size or time and look into its structural properties . for instance
, we want to find out what fraction of the total blocks of the wpsl of a given size has coordination number @xmath8 . for this
we collect data for @xmath58 as a function of @xmath8 and obtain the coordination number distribution function @xmath59 where subscript @xmath52 indicates fixed time .
interestingly , the same wpsl can be interpreted as a network if the blocks of the lattice are regarded as nodes and the common borders between blocks as their links which is topologically identical to the dual of the wpsl ( dwpsl ) . for the dwpsl network where data points represent average of @xmath60 independent realizations .
the line have slope exactly equal to @xmath61 revealing power - law degree distribution with exponent @xmath6 .
, width=321,height=207 ] is shown using using the same data as of figure 4 . the dotted line with slope equal to @xmath62 is drawn to guide our eyes .
, width=321,height=207 ] in fact , the data for the degree distribution @xmath7 of the dwpsl network is exactly the same as the coordination number distribution function @xmath59 of the wpsl i.e. , @xmath63 .
the plot of @xmath64 vs @xmath65 in figure 4 using data obtained after ensemble average over @xmath60 independent realizations clearly suggests that fitting a straight line to data is possible .
this implies that the degree distribution decays obeying power - law @xmath66 however , note that figure 4 has heavy or fat - tail , benchmark of the scale - free network , which represents highly connected _ hub _ nodes .
the presence of messy tail - end complicates the process of fitting the data into power - law forms , estimating its exponent @xmath67 , and identifying the range over which power - law holds .
one way of reducing the noise at the tail - end of the degree distribution is to plot cumulative distribution @xmath68 .
we therefore plot @xmath69 vs @xmath65 in figure 5 using the same data of figure 4 and find that the heavy tail smooths out naturally where no data is obscured .
the straight line fit of figure 5 has a slope @xmath62 which indicates that the degree distribution ( figure 4 ) decays following power - law with exponent @xmath6 .
we find it worthwhile to mention as a passing note that the mean coordination number @xmath70 reaches asymptotically to a constant value equal to @xmath71 .
we thus find that in the large - size limit the wpsl develops some order in the sense that its annealed coordination number disorder or the degree distribution of its dual is scale - free in character .
this is in sharp contrast to the quenched coordination number disorder found in the voronoi lattice where it is almost impossible to find cells which have significantly higher or fewer neighbours than the mean value @xmath72 @xcite .
in fact , it has been shown that the degree distribution of the dual of the voronoi lattice is gaussian .
the square lattice , on the other hand , is self - dual and its degree distribution is @xmath73 .
further , it is interesting to point out that the exponent @xmath6 is significantly higher than usually found in most real - life network which is typically @xmath74 .
this suggests that in addition to the pa rule the network in question has to obey some constraints .
for instance , nodes in the wpsl are spatially embedded in euclidean space , links gained by the incoming nodes are constrained by the spatial location and the fitting parameter of the nodes .
owing to its unique dynamics this was not unexpected . perhaps , it is noteworthy to mention that the the degree distribution of the electric power grid , whose nodes like wpsl are also embedded in the spatial position , is shown to exhibit power - law but with exponent @xmath75 @xcite .
the power - law degree distribution @xmath7 has been found in many seemingly unrelated real life networks .
it implies that there must exists some common underlying mechanisms for which disparate systems behave in such a remarkably similar fashion @xcite .
barabasi and albert argued that the growth and the pa rule are the main essence behind the emergence of such power - law .
indeed , the dwpsl network too grows with time but in sharp contrast to the ba model where network grows by addition of one single node with @xmath76 edges per unit time , the dwpsl network grows by addition of a group of three nodes which are already linked by two edges .
it also differs in the way incoming nodes establish links with the existing nodes . to understand the growth mechanism of the dwpsl network
let us look into the @xmath24th step of the algorithm .
first a node , say it is labeled as @xmath35 , is picked from the @xmath26 nodes preferentially with respect to the fitness parameter of a node ( i.e. , according to their respective areas ) .
secondly , connect the node @xmath35 with two new nodes @xmath77 and @xmath78 in order to establish their links with the existing network . at the same time , at least two or more links of @xmath35 with other nodes are removed ( though the exact number depends on the number of neighbours @xmath35 already has ) in favour of linking them among the three incoming nodes in a self - organised fashion . in the process ,
the degree @xmath79 of the node @xmath35 will either decrease ( may turn into a node with marginally low degree in the case it is highly connected node ) or at best remain the same but will never increase .
it , therefore , may appear that the pa rule is not followed here .
a closer look into the dynamics , however , reveals otherwise .
it is interesting to note that an existing nodes during the process gain links only if one of its neighbour is picked not itself .
it implies that the higher the links ( or degree ) a node has , the higher its chance of gaining more links since they can be reached in a larger number of ways .
it essentially embodies the intuitive idea of pa rule .
therefore , the dwpsl network can be seen to follow preferential attachment rule but in disguise .
we again focus on the blocks of the wpsl but this time we characterize them by their length @xmath12 and width @xmath13 instead of the number of neighbours they have .
then the evolution of their distribution function @xmath80 can be described by the following kinetic equation @xcite @xmath81 incorporating the @xmath10-tuple millen transform @xmath82 in the kinetic equation yields @xmath83 iterating it to get all the derivatives of @xmath84 and then substituting them into the taylor series expansion of @xmath84 about @xmath85 one can write its solution in terms of generalized hypergeometric function @xcite @xmath86 where @xmath87 for symmetry reason and @xmath88^{{{1}\over{2}}}.\ ] ] one can immediately see that ( i ) @xmath89 is the total number of blocks @xmath90 and ( ii ) @xmath91 is the total area of all the blocks which is obviously a conserved quantity .
the behaviour of @xmath84 in the long time limit is @xmath92 it implies that the system is in fact governed by several conservation laws namely , @xmath93 are independent of time for all @xmath16 which has been confirmed by numerical simulation ( see figure 6 ) .
vs @xmath94 for @xmath95 are drawn using data collected from one realization .
, width=321,height=207 ] we now focus on the distribution function @xmath96 that describes the concentration of blocks of length @xmath12 at time @xmath52 regardless of the size of their widths @xmath13 .
then the @xmath97th moment of @xmath98 is defined as @xmath99 appreciating the fact that @xmath100 and using equation ( [ eq : aminus ] ) one can immediately write that @xmath101 note that had we focus on @xmath102 instead of @xmath98 we would have exactly the same result since their @xmath97th moments are identical , @xmath103 , due to symmetry reason .
we find that the quantity @xmath104 and hence @xmath19 or @xmath19 is a conserved quantity .
however , yet another
interesting fact is that although @xmath104 remains a constant against time in every independent realization , their exact numerical value fluctuates sample to sample ( see figure 7 ) .
it clearly indicates lack of self - averaging or wild fluctuation .
nonetheless , during each realization we can use @xmath104 as a measure to populate the @xmath4th block with the fraction of the total population @xmath105 .
the corresponding `` partition function '' then is @xmath106 whose solution can be written immediately from equation ( [ eq : qmoment ] ) to give @xmath107 vs @xmath94 for four different realizations shows that the numerical value is different at every independent realization .
, width=321,height=207 ] we find it instructive to express @xmath108 in terms of square root of the mean area @xmath109 as it gives the weighted number of squares @xmath110 needed to cover the measure which scales as @xmath111 where the mass exponent @xmath112 the non - linear nature of @xmath113 suggests that an infinite hierarchy of exponents is required to specify how the moments of the probabilities @xmath114s scales with @xmath115 .
note that @xmath116 is the hausdorff - besicovitch ( h - d ) dimension of the wpsl since the bare number @xmath117 and @xmath118 follows from the conservation laws ( or normalization of the probabilities @xmath119 ) .
, width=321,height=207 ] we now perform the legendre transformation of @xmath113 by using the lipschitz - hlder exponent @xmath120 as an independent variable to obtain the new function @xmath121 substituting it in equation ( [ weightednumber ] ) we find that @xmath122 since it is the dominant value of the integral @xmath123 obtained by the extremal conditions
. we can infer from this that the number of squares @xmath124 needed to cover the psl with @xmath125 in the range @xmath125 to @xmath126 scales as @xmath127 where @xmath128 is the number of times the subdivided regions indexed by @xmath125 @xcite .
it implies that a spectrum of spatially intertwined fractal dimensions @xmath129 are needed to characterize the measure which is always concave in character ( see figure 8) .
it implies that the size disorder of the blocks are multifractal in character since the measure @xmath130 is related to size of the blocks .
that is , the distribution of @xmath130 in wpsl can be subdivided into a union of fractal subsets each with fractal dimension @xmath131 in which the measure @xmath132 scales as @xmath133 .
note that @xmath134 is always concave in character ( see figure 8) with a single maximum at @xmath135 correspond to the dimension of the wpsl with empty blocks . on the other hand , we find that the entropy @xmath136 associated with the partition of the measure on the support ( wpsl ) by using the relation @xmath137 in the definition of @xmath138
. then a few steps of algebraic manipulation reveals that @xmath138 exhibits scaling @xmath139 where the exponent @xmath140 obtained from @xmath141 it is interesting to note that @xmath142 is related to the generalized dimension @xmath143 , also related to the rnyi entropy @xmath144 in the information theory , given by @xmath145={{\tau(q)}\over{1-q}},\ ] ] which is often used in the multifractal formalism as it can also provide insightful interpretation .
for instance , @xmath146 is the dimension of the support , @xmath147 is the rnyi information dimension and @xmath148 is known as the correlation dimension @xcite .
we have proposed and studied a weighted planar stochastic lattice ( wpsl ) which has annealed coordination number and the block size disorder . we have shown that the coordination number disorder is scale - free in character since the degree distribution of its dual ( dwpsl ) , which is topologically identical to the network obtained by considering blocks of the wpsl as nodes and the common border between blocks as links , exhibits power - law .
however , the novelty of this network is that it grows by addition of a group of already linked nodes which then establish links with the existing nodes following pa rule though in disguise in the sense that existing nodes gain links only if one of their neighbour is picked not itself . besides
, we have shown that if the blocks of the wpsl are characterized by their respective length @xmath12 and width @xmath13 then we find @xmath0 remains a constant regardless of the size of the lattice .
however , the numerical values of the conserved quantities except @xmath18 varies from sample to sample revealing absence of self - averaging - an indication of wild fluctuation .
we have shown that if the blocks are occupied with a fraction of the measure equal to cubic power of their respective length or width then its distribution on the wpsl is multifractal nature .
such multifractal lattice with scale - free coordination disorder can be of great interest as it has the potential to mimic disordered medium on which one can study various physical phenomena like percolation and random walk problems etc .
nip acknowledges support from the bose centre for advanced study and research in natural sciences .
rao m , sengupta s , and sahu h k 1995 phys .
* 75 * , 2164 atkinson h v 1998 acta metall , * 36 * 469 j. c. m. mombach j c m , vasconcellos m a z , and de almeida r m c j. 1990 phys .
d * 23 * , 600 weaire d and rivier n , 1984 contemp . phys . * 25 * , 59 okabe a , boots b , sugihara k , and chiu s n 2000 _ spatial tessellations - concepts and applications of voronoi diagrams _ ( chicester : wiley ) delaney g w , hutzler s and aste t 2008 phys .
* 101 * 120602 hassan m k and rodgers g j 1996 phys .
a * 218 * , 207 krapivsky p l and ben - naim e 1994 phys . rev .
e * 50 * 3502 krapivsky p l and ben - naim e 1996 phys .
* 76 * 3234 majumdar s n , dean d s and krapivsky p l 2005 pramana - j. phys . * 64 * , 1187 barabasi a l and r. albert r 1999 science * 286 * , 509 albert r and barabasi a l 2002 rev . mod . phys . * 74 * , 47 mandelbrot b b 1982 _ fractal geometry of nature _ ( new york : w. h. freeman ) evers f and mirlin a d 2008 rev .
* 80 * , 1355 m. e. j. newman , siam review * 45 * , 167 ( 2003 ) .
de oliveira m m , alves s g , ferreira s c , and dickman r 2008 phys . rev .
e * 78 * , 031133 luke y l 1969 _ the special functions and their approximations i _ ( new york : academic press ) feder j 1988 fractals ( new york : plenum , new york ) hentschel h g e and procaccia i 1983 physica * 8d * , 435 sporting j , weickert j 1999 ieee trans .
theory * 45 * , 1051 | we propose a weighted planar stochastic lattice ( wpsl ) formed by the random sequential partition of a plane into contiguous and non - overlapping blocks and find that it evolves following several non - trivial conservation laws , namely @xmath0 is independent of time @xmath1 , where @xmath2 and @xmath3 are the length and width of the @xmath4th block .
its dual on the other hand , obtained by replacing each block with a node at its center and common border between blocks with an edge joining the two vertices , emerges as a network with a power - law degree distribution @xmath5 where @xmath6 revealing scale - free coordination number disorder since @xmath7 also describes the fraction of blocks having @xmath8 neighbours . to quantify the size disorder ,
we show that if the @xmath4th block is populated with @xmath9 then its distribution in the wpsl exhibits multifractality . |
with the advent of the internet , many machine learning applications are faced with very large and inherently high - dimensional datasets , resulting in challenges in scaling up training algorithms and storing the data . especially in the context of search and machine translation , corpus sizes used in industrial practice
have long exceeded the main memory capacity of single machine .
for example , @xcite experimented with a dataset with potentially 16 trillion ( @xmath1 ) unique features .
@xcite discusses training sets with ( on average ) @xmath2 items and @xmath3 distinct features , requiring novel algorithmic approaches and architectures . as a consequence ,
there has been a renewed emphasis on scaling up machine learning techniques by using massively parallel architectures ; however , methods relying solely on parallelism can be expensive ( both with regards to hardware requirements and energy costs ) and often induce significant additional communication and data distribution overhead . this work approaches the challenges posed by large datasets by leveraging techniques from the area of _ similarity search _
@xcite , where similar increase in dataset sizes has made the storage and computational requirements for computing exact distances prohibitive , thus making data representations that allow compact storage and efficient approximate distance computation necessary .
the method of @xmath0-bit minwise hashing @xcite is a very recent progress for efficiently ( in both time and space ) computing _ resemblances _ among extremely high - dimensional ( e.g. , @xmath4 ) binary vectors . in this paper , we show that @xmath0-bit minwise hashing can be seamlessly integrated with linear support vector machine ( svm ) @xcite and logistic regression solvers . in @xcite , the authors addressed a critically important problem of training linear svm when the data can not fit in memory . in this paper
, our work also addresses the same problem using a very different approach . in the context of search , a standard procedure to represent documents ( e.g. , web pages ) is to use @xmath5-shingles ( i.e. , @xmath5 contiguous words ) , where @xmath5 can be as large as 5 ( or 7 ) in several studies @xcite .
this procedure can generate datasets of extremely high dimensions .
for example , suppose we only consider @xmath6 common english words .
using @xmath7 may require the size of dictionary @xmath8 to be @xmath9 .
in practice , @xmath10 often suffices , as the number of available documents may not be large enough to exhaust the dictionary . for @xmath5-shingle data , normally only abscence / presence ( 0/1 ) information is used , as it is known that word frequency distributions within documents approximately follow a power - law @xcite , meaning that most single terms occur rarely , thereby making a @xmath5-shingle unlikely to occur more than once in a document .
interestingly , even when the data are not too high - dimensional , empirical studies @xcite achieved good performance with binary - quantized data .
when the data can fit in memory , linear svm is often extremely efficient after the data are loaded into the memory .
it is however often the case that , for very large datasets , the data loading time dominates the computing time for training the svm @xcite .
a much more severe problem arises when the data can not fit in memory .
this situation can be very common in practice .
the publicly available _ webspam _ dataset needs about 24 gb disk space ( in libsvm input data format ) , which exceeds the memory capacity of many desktop pcs .
note that _ webspam _ , which contains only 350,000 documents represented by 3-shingles , is still a small dataset compared to the industry applications @xcite .
we propose a solution which leverages _ b - bit minwise hashing_. our approach assume the data vectors are binary , very high - dimensional , and relatively sparse , which is generally true of text documents represented via shingles .
we apply @xmath0-bit minwise hashing to obtain a compact representation of the original data . in order to use the technique for efficient learning
, we have to address several issues : * we need to prove that the matrices generated by @xmath0-bit minwise hashing are indeed positive definite , which will provide the solid foundation for our proposed solution . *
if we use @xmath0-bit minwise hashing to estimate the resemblance , which is nonlinear , how can we effectively convert this nonlinear problem into a linear problem ? * compared to other hashing techniques such as random projections , count - min ( cm ) sketch @xcite , or vowpal wabbit ( vw ) @xcite , does our approach exhibits advantages ?
it turns out that our proof in the next section that @xmath0-bit hashing matrices are positive definite naturally provides the construction for converting the otherwise nonlinear svm problem into linear svm .
+ @xcite proposed solving the memory bottleneck by partitioning the data into blocks , which are repeatedly loaded into memory as their approach updates the model coefficients .
however , the computational bottleneck is still at the memory because loading the data blocks for many iterations consumes a large number of disk i / os .
clearly , one should note that our method is not really a competitor of the approach in @xcite .
in fact , both approaches may work together to solve extremely large problems .
_ minwise hashing _ @xcite has been successfully applied to a very wide range of real - world problems especially in the context of search @xcite , for efficiently computing set similarities .
minwise hashing mainly works well with binary data , which can be viewed either as 0/1 vectors or as sets .
given two sets , @xmath11 , a widely used ( normalized ) measure of similarity is the _ resemblance _ @xmath12 : @xmath13 in this method , one applies a random permutation @xmath14 on @xmath15 and @xmath16 .
the collision probability is simply @xmath17 one can repeat the permutation @xmath18 times : @xmath19 , @xmath20 , ... ,
@xmath21 to estimate @xmath12 without bias , as @xmath22 the common practice of minwise hashing is to store each hashed value , e.g. , @xmath23 and @xmath24 , using 64 bits @xcite .
the storage ( and computational ) cost will be prohibitive in truly large - scale ( industry ) applications @xcite .
b - bit minwise hashing _
@xcite provides a strikingly simple solution to this ( storage and computational ) problem by storing only the lowest @xmath0 bits ( instead of 64 bits ) of each hashed value . for convenience ,
we define the minimum values under @xmath25 : @xmath26 and @xmath27 , and define @xmath28th lowest bit of @xmath29 , and @xmath30th lowest bit of @xmath31 .
@xcite[the_basic ] assume @xmath32 is large .
@xmath33^{2^b-1}}{1-\left[1-r_1\right]^{2^b}},\hspace{1 in } a_{2,b } = \frac{r_2\left[1-r_2\right]^{2^b-1}}{1-\left[1-r_2\right]^{2^b}}.\box\end{aligned}\ ] ] this ( approximate ) formula ( [ eqn_basic ] ) is remarkably accurate , even for very small @xmath32 .
some numerical comparisons with the exact probabilities are provided in appendix [ app_basic_error ] .
we can then estimate @xmath34 ( and @xmath12 ) from @xmath18 independent permutations : @xmath19 , @xmath20 , ... ,
@xmath21 , @xmath35@xmath36 ^ 2 } = \frac{1}{k}\frac{\left[c_{1,b}+(1-c_{2,b})r\right]\left[1-c_{1,b}-(1-c_{2,b})r\right]}{\left[1-c_{2,b}\right]^2}\end{aligned}\ ] ] we will show that we can apply @xmath0-bit hashing for learning without explicitly estimating @xmath12 from ( [ eqn_basic ] ) .
this section proves some theoretical properties of matrices generated by resemblance , minwise hashing , or @xmath0-bit minwise hashing , which are all positive definite matrices .
our proof not only provides a solid theoretical foundation for using @xmath0-bit hashing in learning , but also illustrates our idea behind the construction for integrating @xmath0-bit hashing with linear learning algorithms .
+ * definition * : a symmetric @xmath37 matrix @xmath38 satisfying @xmath39 , for all real vectors @xmath40 is called _ positive definite ( pd)_. note that here we do not differentiate pd from _ nonnegative definite_. [ thm_pd ] consider @xmath41 sets @xmath15 , @xmath16 , ... , @xmath42 .
apply one permutation @xmath25 to each set and define @xmath43 .
the following three matrices are all pd . 1 . the _ resemblance matrix _ @xmath44 , whose @xmath45-th entry is the resemblance between set @xmath46 and set @xmath47 : @xmath48 2 .
the _ minwise hashing matrix _
@xmath49 : @xmath50 3 . the _ b - bit minwise hashing matrix _
@xmath51 : @xmath52 , where @xmath53 is the @xmath54-th lowest bit of @xmath55 .
consequently , consider @xmath18 independent permutations and denote @xmath56 the b - bit minwise hashing matrix generated by the @xmath57-th permutation .
then the summation @xmath58 is also pd . + * proof : * a matrix @xmath59 is pd if it can be written as an inner product @xmath60 . because @xmath61 @xmath62 is the inner product of two d - dim vectors .
thus , @xmath63 is pd .
similarly , the b - bit minwise hashing matrix @xmath64 is pd because @xmath65 the resemblance matrix @xmath66 is pd because @xmath67 and @xmath62 is the @xmath45-th element of the pd matrix @xmath63 .
note that the expectation is a linear operation .
@xmath68 our proof that the @xmath0-bit minwise hashing matrix @xmath64 is pd provides us with a simple strategy to expand a nonlinear ( resemblance ) kernel into a linear ( inner product ) kernel .
after concatenating the @xmath18 vectors resulting from ( [ eqn_m_b ] ) , the new ( binary ) data vector after the expansion will be of dimension @xmath69 with exactly @xmath18 ones .
linear algorithms such as linear svm and logistic regression have become very powerful and extremely popular .
representative software packages include svm@xmath70 @xcite , pegasos @xcite , bottou s sgd svm @xcite , and liblinear @xcite .
given a dataset @xmath71 , @xmath72 , @xmath73 , the @xmath74-regularized linear svm solves the following optimization problem : @xmath75 and the @xmath74-regularized logistic regression solves a similar problem : @xmath76 here @xmath77 is an important penalty parameter .
since our purpose is to demonstrate the effectiveness of our proposed scheme using @xmath0-bit hashing , we simply provide results for a wide range of @xmath78 values and assume that the best performance is achievable if we conduct cross - validations . in our approach , we apply @xmath18 independent random permutations on each feature vector @xmath79 and store the lowest @xmath0 bits of each hashed value .
this way , we obtain a new dataset which can be stored using merely @xmath80 bits . at run - time , we expand each new data point into a @xmath69-length vector .
+ for example , suppose @xmath81 and the hashed values are originally @xmath82 , whose binary digits are @xmath83 .
consider @xmath84 .
then the binary digits are stored as @xmath85 ( which corresponds to @xmath86 in decimals ) . at run - time , we need to expand them into a vector of length @xmath87 , to be @xmath88 , which will be the new feature vector fed to a solver : @xmath89 clearly , this expansion is directly inspired by the proof that the @xmath0-bit minwise hashing matrix is pd in theorem [ thm_pd ] .
note that the total storage cost is still just @xmath80 bits and each new data vector ( of length @xmath69 ) has exactly @xmath18 1 s .
also , note that in this procedure we actually do not explicitly estimate the resemblance @xmath12 using ( [ eqn_basic ] ) .
our experimental settings follow the work in @xcite very closely . the authors of
@xcite conducted experiments on three datasets , of which the _ webspam _ dataset is public and reasonably high - dimensional ( @xmath90 , @xmath91 ) .
therefore , our experiments focus on _ webspam_. following @xcite , we randomly selected @xmath92 of samples for testing and used the remaining @xmath93 samples for training .
we chose liblinear as the tool to demonstrate the effectiveness of our algorithm .
all experiments were conducted on workstations with xeon(r ) cpu ( [email protected] ) and 48 gb ram , under windows 7 system .
thus , in our case , the original data ( about 24 gb in libsvm format ) fit in memory . in applications for which the data do not fit in memory
, we expect that @xmath0-bit hashing will be even more substantially advantageous , because the hashed data are relatively very small .
in fact , our experimental results will show that for this dataset , using @xmath94 and @xmath95 can achieve the same testing accuracy as using the original data .
the effective storage for the reduced dataset ( with 350k examples , using @xmath94 and @xmath95 ) would be merely about 70 mb .
we implemented a new resemblance kernel function and tried to use libsvm to train the _
webspam _ dataset .
we waited for * * over one week * * but libsvm still had not output any results .
fortunately , using @xmath0-bit minswise hashing to estimate the resemblance kernels , we were able to obtain some results .
for example , with @xmath96 and @xmath95 , the training time of libsvm ranged from 1938 seconds ( @xmath97 ) to 13253 seconds ( @xmath98 ) .
in particular , when @xmath99 , the test accuracies essentially matched the best test results given by liblinear on the original _ webspam _ data . therefore , there is a significant benefit of _ data reduction _ provided by @xmath0-bit minwise hashing , for training nonlinear svm .
this experiment also demonstrates that it is very important ( and fortunate ) that we are able to transform this nonlinear problem into a linear problem .
since there is an important tuning parameter @xmath78 in linear svm and logistic regression , we conducted our extensive experiments for a wide range of @xmath78 values ( from @xmath100 to @xmath101 ) with fine spacings in @xmath102 $ ] .
we mainly experimented with @xmath97 to @xmath98 , and @xmath103 , 2 , 4 , 8 , and 16 .
figures [ fig_acc ] ( average ) and [ fig_acc_std ] ( std , standard deviation ) provide the test accuracies .
figure [ fig_acc ] demonstrates that using @xmath104 and @xmath105 achieves about the same test accuracies as using the original data . since our method is randomized , we repeated every experiment 50 times .
we report both the mean and std values .
figure [ fig_acc_std ] illustrates that the stds are very small , especially with @xmath106 . in other words ,
our algorithm produces stable predictions . for this dataset ,
the best performances were usually achieved when @xmath107 . compared with the original training time ( about 100 seconds )
, we can see from figure [ fig_training ] that our method only need about @xmath108 seconds near @xmath96 ( about 3 seconds for @xmath95 ) .
note that here the training time did not include the data loading time .
loading the original data took about 12 minutes while loading the hashed data took only about 10 seconds .
of course , there is a cost for processing ( hashing ) the data , which we find is efficient , confirming prior studies @xcite .
in fact , data processing can be conducted during data collection , as is the standard practice in search . in other words , prior to conducting the learning procedure , the data may be already processed and stored by ( @xmath0-bit ) minwise hashing , which can be used for multiple tasks including learning , clustering , duplicate detection , near - neighbor search , etc . + compared with the original testing time ( about @xmath109 seconds ) , we can see from figure [ fig_testing ] that the testing time of our method is merely about 1 or 2 seconds .
note that the testing time includes both the data loading time and computing time , as designed by liblinear .
the efficiency of testing may be very important in practice , for example , when the classifier is deployed in an user - facing application ( such as search ) , while the cost of training or pre - processing ( such as hashing ) may be less critical and can often be conducted off - line .
figure [ fig_acc_logit ] presents the test accuracies and figure [ fig_training_logit ] presents the training time using logistic regression . again , with @xmath105 ( or even @xmath110 and @xmath104 , @xmath0-bit minwise hashing can achieve the same test accuracies as using the original data .
figure [ fig_acc_std_logit ] presents the standard deviations , which again verify that our algorithm produces stable predictions for logistic regression . from figure [ fig_training_logit ]
, we can see that the training time is substantially reduced , from about 1000 seconds to about @xmath111 seconds only ( unless @xmath112 and @xmath18 is large ) .
+ in summary , it appears @xmath0-bit hashing is highly effective in reducing the data size and speeding up the training ( and testing ) , for both ( nonlinear and linear ) svm and logistic regression .
we notice that when using @xmath112 , the training time can be much larger than using @xmath113 .
interestingly , we find that @xmath0-bit hashing can be combined with _
vowpal wabbit ( vw )
_ @xcite to further reduce the training time , especially when @xmath0 is large .
the two methods , random projections @xcite and vowpal wabbit ( vw ) @xcite are not limited to binary data ( although for ultra high - dimensional used in the context of search , the data are often binary ) .
the vw algorithm is also related to the count - min sketch @xcite . in this paper , we use vw particularly for the algorithm in @xcite . for convenience ,
we denote two @xmath32-dim data vectors by @xmath114 .
again , the task is to estimate the inner product @xmath115 . the general idea is to multiply the data vectors , e.g. , @xmath116 and @xmath117 , by a random matrix @xmath118 , where @xmath119 is sampled i.i.d . from the following generic distribution with @xcite @xmath120 we must have @xmath121 because @xmath122 .
this generates two @xmath18-dim vectors , @xmath123 and @xmath124 : @xmath125 the general distributions which satisfy ( [ eqn_r_ij ] ) includes the standard normal distribution ( in this case , @xmath126 ) and the `` sparse projection '' distribution specified as @xmath127 @xcite provided the following unbiased estimator @xmath128 of @xmath129 and the general variance formula : @xmath130\end{aligned}\ ] ] which means @xmath131 achieves the smallest variance .
the only elementary distribution we know that satisfies ( [ eqn_r_ij ] ) with @xmath131 is the two point distribution in @xmath132 with equal probabilities , i.e. , ( [ eqn_sparse_r ] ) with @xmath131 .
again , in this paper , `` vw '' always refers to the particular algorithm in @xcite .
vw may be viewed as a `` bias - corrected '' version of the count - min ( cm ) sketch algorithm @xcite . in the original cm algorithm , the key step is to independently and uniformly hash elements of the data vectors to buckets @xmath133 and the hashed value is the sum of the elements in the bucket .
that is @xmath134 with probability @xmath135 , where @xmath136 . for convenience ,
we introduce an indicator function : @xmath137 which allow us to write the hashed data as @xmath138 the estimate @xmath139 is ( severely ) biased for the task of estimating the inner products .
the original paper @xcite suggested a `` count - min '' step for positive data , by generating multiple independent estimates @xmath140 and taking the minimum as the final estimate . that step can not remove the bias and makes the analysis ( such as variance ) very difficult .
here we should mention that the bias of cm may not be a major issue in other tasks such as sparse recovery ( or `` heavy - hitter '' , or `` elephant detection '' , by various communities ) .
+ @xcite proposed a creative method for bias - correction , which consists of pre - multiplying ( element - wise ) the original data vectors with a random vector whose entries are sampled i.i.d . from the two - point distribution in @xmath141 with equal probabilities , which corresponds to @xmath142 in ( [ eqn_sparse_r ] ) . here
, we consider a more general situation by considering any @xmath121 . after
applying multiplication and hashing on @xmath116 and @xmath117 as in @xcite , the resultant vectors @xmath143 and @xmath144 are @xmath145 where @xmath146 is defined as in ( [ eqn_r_ij ] ) , i.e. , @xmath147 .
we have the following lemma .
[ lem_vw ] @xmath148\end{aligned}\ ] ] * proof : * see appendix [ proof_lem_vw].@xmath68 interestingly , the variance ( [ eqn_var_vw ] ) says we do need @xmath131 , otherwise the additional term @xmath149 will not vanish even as the sample size @xmath150 .
in other words , the choice of random distribution in vw is essentially the only option if we want to remove the bias by pre - multiplying the data vectors ( element - wise ) with a vector of random variables .
of course , once we let @xmath131 , the variance ( [ eqn_var_vw ] ) becomes identical to the variance of random projections ( [ eqn_var_rp ] ) .
we implemented vw ( which , in this paper , always refers to the algorithm developed in @xcite ) and tested it on the same webspam dataset .
figure [ fig_vw ] shows that @xmath0-bit minwise hashing is substantially more accurate ( at the same sample size @xmath18 ) and requires significantly less training time ( to achieve the same accuracy ) .
for example , @xmath151-bit minwise hashing with @xmath94 achieves about the same test accuracy as vw with @xmath152 .
note that we only stored the non - zeros of the hashed data generated by vw .
this empirical finding is not surprising , because the variance of @xmath0-bit hashing is usually substantially smaller than the variance of vw ( and random projections ) . in appendix
[ app_compare_vw ] , we show that , at the same storage cost , @xmath0-bit hashing usually improves vw by 10- to 100-fold , by assuming each sample of vw requires 32 bits storage .
of course , even if vw only stores each sample using 16 bits , an improvement of 5- to 50-fold would still be very substantial .
note that this comparison makes sense for the purpose of _ data reduction _
, i.e. , the sample size @xmath18 is substantially smaller than the number of non - zeros in the original ( massive ) data .
+ there is one interesting issue here . unlike random projections ( and minwise hashing )
, vw is a _ sparsity - preserving _ algorithm , meaning that in the resultant sample vector of length @xmath18 , the number of non - zeros will not exceed the number of non - zeros in the original vector .
in fact , it is easy to see that the fraction of zeros in the resultant vector would be ( at least ) @xmath153 , where @xmath40 is the number of non - zeros in the original data vector . when @xmath154 , then @xmath155 . in other words , if our goal is _ data reduction _ ( i.e. , @xmath156 ) , then the hashed data by vw are dense .
+ in this paper , we mainly focus on _ data reduction_. as discussed in the introduction , for many industry applications , the relatively sparse datasets are often massive in the absolute scale and we assume we can not store all the non - zeros .
in fact , this is the also one of the basic motivations for developing minwise hashing .
however , the case of @xmath157 can also be interesting and useful in our work .
this is because vw is an excellent tool for achieving _ compact indexing _ due to the _ sparsity - preserving _ property .
basically , we can let @xmath18 be very large ( like @xmath158 in @xcite ) .
as the original dictionary size @xmath32 is extremely large ( e.g. , @xmath4 ) , even @xmath159 will be a meaningful reduction of the indexing . of course , using a very large @xmath18 will not be useful for the purpose of _ data reduction_. +
in our algorithm , we reduce the original massive data to @xmath160 bits only , where @xmath41 is the number of data points . with ( e.g. , ) @xmath94 and @xmath95
, our technique achieves a huge _ data reduction_. in the run - time , we need to expand each data point into a binary vector of length @xmath161 with exactly @xmath18 1 s
. if @xmath0 is large like @xmath162 , the new binary vectors will be highly sparse .
in fact , in figure [ fig_training ] and figure [ fig_training_logit ] , we can see that when using @xmath112 , the training time becomes substantially larger than using @xmath113 ( especially when @xmath18 is large ) . on the other hand ,
once we have expanded the vectors , the task is merely computing inner products , for which we can actually use vw .
therefore , in the run - time , after we have generated the sparse binary vectors of length @xmath163 , we hash them using vw with sample size @xmath164 ( to differentiate from @xmath18 ) .
how large should @xmath164 be ? lemma [ lem_b - bit+vw ] may provide some insights
. + recall section [ sec_minwise ] provides the estimator , denoted by @xmath165 , of the resemblance @xmath12 , using @xmath0-bit minwise hashing . now , suppose we first apply vw hashing with size @xmath164 on the binary vector of length @xmath163 before estimating @xmath12 , which will introduce some additional randomness ( on top of @xmath0-bit hashing ) .
we denote the new estimator by @xmath166 .
lemma [ lem_b - bit+vw ] provides its theoretical variance .
[ lem_b - bit+vw ] @xmath167
^ 2}\left(1+p_b^2-\frac{p_b(1+p_b)}{k}\right),\\\notag & \hspace{0.73 in } = \frac{1}{k}\frac{p_b(1-p_b)}{\left[1-c_{2,b}\right]^2 } + \frac{1}{m}\frac{1+p_b^2}{\left[1-c_{2,b}\right]^2 } - \frac{1}{mk}\frac{p_b(1+p_b)}{\left[1-c_{2,b}\right]^2}\end{aligned}\ ] ] where @xmath168 ^ 2}$ ] is given by ( [ eqn_var_b ] ) and @xmath169 is the constant defined in theorem [ the_basic ] . + * proof * : the proof is quite straightforward , by following the conditional expectation formula : @xmath170 , and the conditional variance formula @xmath171 .
recall , originally we estimate the resemblance by @xmath172}$ ] , where @xmath173 and @xmath174 is the number of matches in the two hashed data vectors of length @xmath18 generated by @xmath0-bit hashing . @xmath175 and @xmath176 .
now , we apply vw ( of size @xmath164 and @xmath131 ) on the hashed data vectors to estimate @xmath174 ( instead of counting it exactly ) . we denote this estimates by @xmath177 and @xmath178 .
because we know the vw estimate is unbiased , we have @xmath179 using the conditional variance formula and the variance of vw ( [ eqn_var_vw ] ) ( with @xmath131 ) , we obtain @xmath180\right ) + var\left(t\right)\right]\\\notag = & \frac{1}{k^2}\left[\frac{1}{m}\left(k^2+kp_b(1-p_b)+k^2p_b^2 - 2kp_b\right ) + kp_b(1-p_b)\right]\\\notag = & \frac{p_b(1-p_b)}{k } + \frac{1}{m}\left(1+p_b^2-\frac{p_b(1+p_b)}{k}\right)\end{aligned}\ ] ] this completes the proof .
@xmath68 + compared to the original variance @xmath181 ^ 2}$ ] , the additional term @xmath182 ^ 2}$ ] in ( [ eqn_var_b - bit+vw ] ) can be relatively large if @xmath164 is not large enough . therefore , we should choose @xmath183 ( to reduce the additional variance ) and @xmath184 ( otherwise there is no need to apply this vw step ) .
if @xmath112 , then @xmath185 may be a good trade - off , because @xmath186 .
+ figure [ fig_16-bit+vw ] provides an empirical study to verify this intuition .
basically , as @xmath187 , using vw on top of 16-bit hashing achieves the same accuracies at using 16-bit hashing directly and reduces the training time quite noticeably .
[ fig_16-bit+vw ] we also experimented with combining 8-bit hashing with vw .
we found that we need @xmath185 to achieve similar accuracies , i.e. , the additional vw step did not bring more improvement ( without hurting accuracies ) in terms of training speed when @xmath95 .
this is understandable from the analysis of the variance in lemma [ lem_b - bit+vw ] .
minwise hashing has been widely used in ( search ) industry and @xmath0-bit minwise hashing requires only very minimal ( if any ) modifications ( by doing less work ) .
thus , we expect @xmath0-bit minwise hashing will be adopted in practice .
it is also well - understood in practice that we can use ( good ) hashing functions to very efficiently simulate permutations . in many real - world scenarios ,
the preprocessing step is not critical because it requires only one scan of the data , which can be conducted off - line ( or on the data - collection stage , or at the same time as n - grams are generated ) , and it is trivially parallelizable .
in fact , because @xmath0-bit minwise hashing can substantially reduce the memory consumption , it may be now affordable to store considerably more examples in the memory ( after @xmath0-bit hashing ) than before , to avoid ( or minimize ) disk ios . once the hashed data have been generated , they can be used and re - used for many tasks such as supervised learning , clustering , duplicate detections , near - neighbor search , etc .
for example , a learning task may need to re - use the same ( hashed ) dataset to perform many cross - validations and parameter tuning ( e.g. , for experimenting with many @xmath78 values in svm ) .
+ nevertheless , there might be situations in which the preprocessing time can be an issue .
for example , when a new unprocessed document ( i.e. n - grams are not yet available ) arrives and a particular application requires an immediate response from the learning algorithm , then the preprocessing cost might ( or might not ) be an issue .
firstly , generating n - grams will take some time . secondly ,
if during the session a disk io occurs , then the io cost will typically mask the cost of preprocessing for @xmath0-bit minwise hashing .
note that the preprocessing cost for the vw algorithm can be substantially lower .
thus , if the time for pre - processing is indeed a concern ( while the storage cost or test accuracies are not as much ) , one may want to consider using vw ( or _ very sparse random projections _
@xcite ) for those applications .
as data sizes continue to grow faster than the memory and computational power , machine - learning tasks in industrial practice are increasingly faced with training datasets that exceed the resources on a single server . a number of approaches have been proposed that address this by either scaling out the training process or partitioning the data , but both solutions can be expensive . in this paper , we propose a compact representation of sparse , binary datasets based on @xmath0-bit minwise hashing .
we show that the @xmath0-bit minwise hashing estimators are positive definite kernels and can be naturally integrated with learning algorithms such as svm and logistic regression , leading to dramatic improvements in training time and/or resource requirements .
we also compare @xmath0-bit minwise hashing with the vowpal wabbit ( vw ) algorithm , which has the same variances as random projections .
our theoretical and empirical comparisons illustrate that usually @xmath0-bit minwise hashing is significantly more accurate ( at the same storage ) than vw for binary data .
interestingly , @xmath0-bit minwise hashing can be combined with vw to achieve further improvements in terms of training speed when @xmath0 is large ( e.g. , @xmath188 ) .
10 dimitris achlioptas .
database - friendly random projections : with binary coins . , 66(4):671687 , 2003 .
alexandr andoni and piotr indyk .
near - optimal hashing algorithms for approximate nearest neighbor in high dimensions . in _ commun .
volume 51 , pages 117122 , 2008 .
harald baayen . , volume 18 of _ text , speech and language technology_. kulver academic publishers , 2001 .
michael bendersky and w. bruce croft .
finding text reuse on the web . in _ wsdm _ , pages 262271 , barcelona , spain , 2009 .
leon bottou .
available at http://leon.bottou.org/projects/sgd .
andrei z. broder . on the resemblance and containment of documents . in _ the compression and complexity of sequences _ , pages 2129 , positano , italy , 1997 .
andrei z. broder , steven c. glassman , mark s. manasse , and geoffrey zweig .
syntactic clustering of the web . in _
www _ , pages 1157 1166 , santa clara , ca , 1997 .
gregory buehrer and kumar chellapilla .
a scalable pattern mining approach to web graph compression with communities . in _
wsdm _ , pages 95106 , stanford , ca , 2008 .
olivier chapelle , patrick haffner , and vladimir n. vapnik .
support vector machines for histogram - based image classification .
, 10(5):10551064 , 1999 .
ludmila cherkasova , kave eshghi , charles b. morrey iii , joseph tucek , and alistair c. veitch . applying syntactic similarity algorithms for enterprise information management . in _
kdd _ , pages 10871096 , paris , france , 2009 .
flavio chierichetti , ravi kumar , silvio lattanzi , michael mitzenmacher , alessandro panconesi , and prabhakar raghavan . on compressing social networks .
in _ kdd _ , pages 219228 , paris , france , 2009 .
graham cormode and s. muthukrishnan .
an improved data stream summary : the count - min sketch and its applications .
, 55(1):5875 , 2005 .
yon dourisboure , filippo geraci , and marco pellegrini .
extraction and classification of dense implicit communities in the web graph .
, 3(2):136 , 2009 .
rong - en fan , kai - wei chang , cho - jui hsieh , xiang - rui wang , and chih - jen lin .
liblinear : a library for large linear classification .
, 9:18711874 , 2008 .
dennis fetterly , mark manasse , marc najork , and janet l. wiener .
a large - scale study of the evolution of web pages . in _ www _ , pages 669678 , budapest , hungary , 2003 .
george forman , kave eshghi , and jaap suermondt .
efficient detection of large - scale redundancy in enterprise file systems .
, 43(1):8491 , 2009 . sreenivas gollapudi and aneesh sharma . an axiomatic approach for result diversification . in _ www _ , pages 381390 , madrid , spain , 2009 .
matthias hein and olivier bousquet .
hilbertian metrics and positive definite kernels on probability measures . in _ aistats _ , pages 136143 , barbados , 2005 .
cho - jui hsieh , kai - wei chang , chih - jen lin , s. sathiya keerthi , and s. sundararajan . a dual coordinate descent method for large - scale linear svm . in _ proceedings of the 25th international conference on machine learning _ ,
icml , pages 408415 , 2008 .
yugang jiang , chongwah ngo , and jun yang .
towards optimal bag - of - features for object categorization and semantic video retrieval . in _ civr _ , pages 494501 , amsterdam , netherlands , 2007 .
nitin jindal and bing liu .
opinion spam and analysis . in _
pages 219230 , palo alto , california , usa , 2008 .
thorsten joachims . training linear svms in linear time . in _ kdd _ , pages 217226 , pittsburgh , pa , 2006 .
konstantinos kalpakis and shilang tang .
collaborative data gathering in wireless sensor networks using measurement co - occurrence .
, 31(10):19791992 , 2008 .
ping li , trevor j. hastie , and kenneth w. church . very sparse random projections . in _
kdd _ , pages 287296 , philadelphia , pa , 2006 .
ping li and arnd christian knig . theory and applications b - bit minwise hashing . in _ commun .
acm _ , to appear .
ping li and arnd christian .
b - bit minwise hashing . in _ www _ , pages 671680 , raleigh , nc , 2010 .
ping li , arnd christian , and wenhao gui .
b - bit minwise hashing for estimating three - way similarities . in _ nips _ , vancouver , bc , 2010 .
gurmeet singh manku , arvind jain , and anish das sarma . etecting near - duplicates for web - crawling . in _ www _ , banff , alberta , canada , 2007 .
marc najork , sreenivas gollapudi , and rina panigrahy .
less is more : sampling the neighborhood graph makes salsa better and faster . in _ wsdm _ , pages 242251 ,
barcelona , spain , 2009 .
shai shalev - shwartz , yoram singer , and nathan srebro .
pegasos : primal estimated sub - gradient solver for svm . in _ icml _ , pages 807814 , corvalis , oregon , 2007 .
qinfeng shi , james petterson , gideon dror , john langford , alex smola , and s.v.n . vishwanathan .
hash kernels for structured data .
, 10:26152637 , 2009 .
simon tong .
lessons learned developing a practical large scale machine learning system .
available at http://googleresearch.blogspot.com/2010/04/lessons-learned-developing-practical.html , 2008 .
tanguy urvoy , emmanuel chauveau , pascal filoche , and thomas lavergne .
tracking web spam with html style similarities .
, 2(1):128 , 2008 .
kilian weinberger , anirban dasgupta , john langford , alex smola , and josh attenberg .
feature hashing for large scale multitask learning . in _ icml _ , pages 11131120 , 2009 .
hsiang - fu yu , cho - jui hsieh , kai - wei chang , and chih - jen lin .
large linear classification when data can not fit in memory . in _ kdd _ , pages 833842 , 2010 .
note that the only assumption needed in the proof of theorem [ the_basic ] is that @xmath32 is large , which is virtually always satisfied in practice .
interestingly , ( [ eqn_basic ] ) is remarkably accurate even for very small @xmath32 .
figure [ fig_approximateerror ] shows that when @xmath189 , the absolute error caused by using ( [ eqn_basic ] ) is @xmath190 .
the exact probability , which has no closed - form , can be computed by exhaustive enumerations for small @xmath32 .
the vw algorithm @xcite provides a bias - corrected version of the count - min ( cm ) sketch algorithm @xcite .
the key step in cm is to independently and uniformly hash elements of the data vectors to @xmath191 .
that is @xmath192 with equal probabilities , where @xmath193 . for convenience ,
we introduce the following indicator function : @xmath194 thus , we can denote the cm `` samples '' after the hashing step by @xmath195 and estimate the inner product by @xmath196 , whose expectation and variance can be shown to be @xmath197\end{aligned}\ ] ] from the definition of @xmath198 , we can easily infer its moments , for example , @xmath199 the proof of the mean ( [ eqn_mean_cm ] ) is simple : @xmath200= \sum_{i=1}^d u_{1,i}u_{2,i } + \frac{1}{k}\sum_{i\neq j } u_{1,i}u_{2,j}.\end{aligned}\ ] ] the variance ( [ eqn_var_cm ] is more complicated : @xmath201 the following expansions are helpful : @xmath202 ^ 2 = & \sum_{i=1}^d a_i^2b_i^2 + \sum_{i\neq j } a_i^2b_j^2 + 2a_i^2b_ib_j+2b_i^2a_ia_j+2a_ib_ia_jb_j\\\notag + & \sum_{i\neq j\neq c } a_i^2b_jb_c+b_i^2a_ja_c + 4a_ib_ia_jb_c+\sum_{i\neq j\neq c\neq t}a_ib_ja_cb_t\\\notag \left[\sum_{i\neq j } a_i b_j\right]^2 = & \sum_{i\neq j } a_i^2b_j^2+a_ib_ia_jb_j + \sum_{i\neq j\neq c } a_i^2b_jb_c+b_i^2a_ja_c + 2a_ib_ia_jb_c+\sum_{i\neq j\neq c\neq t}a_ib_ja_cb_t\\\notag \sum_{i=1}^da_ib_i\sum_{i\neq j } a_i b_j = & \sum_{i\neq j } a_i^2b_ib_j+b_i^2a_ia_j+\sum_{i\neq j\neq c } a_i^2b_jb_c\end{aligned}\ ] ] which , combined with the moments of @xmath198 , yield @xmath203 @xmath204\\\notag = & ( k-1)\left[\frac{1}{k}\sum_{i\neq j } u_{1,i}u_{2,i}u_{1,j}u_{2,j}+\frac{2}{k^2}\sum_{i\neq j \neq c } u_{1,i}u_{2,i}u_{1,j}u_{2,c}+\frac{1}{k^3}\sum_{i\neq j\neq c\neq t } u_{1,i}u_{2,j}u_{1,c}u_{2,t}\right]\end{aligned}\ ] ] @xmath205 therefore , @xmath206\end{aligned}\ ] ] the nice approach proposed in the vw paper @xcite is to pre - element - wise - multiply the data vectors with a random vector @xmath207 before taking the hashing operation .
we denote the two resultant vectors ( samples ) by @xmath143 and @xmath144 respectively : @xmath208 where @xmath209 with equal probabilities . here
, we provide a more general scheme by sampling @xmath146 from a sub - gaussian distribution with parameter @xmath57 and @xmath210 which include normal ( i.e. , @xmath126 ) and the distribution on @xmath211 with equal probabilities ( i.e. , @xmath131 ) as special cases .
let @xmath212 .
the goal is to show @xmath213\end{aligned}\ ] ] we can use the previous results and the conditional expectation and variance formulas : @xmath214 @xmath215 , @xmath216 , @xmath217 , @xmath218 .
@xmath219 as @xmath220 , we need to compute @xmath221 thus , @xmath222 by examining the expectation ( [ eqn_mean_cm ] ) , the bias of cm can be easily removed , because @xmath223\end{aligned}\ ] ] is unbiased with variance @xmath224,\end{aligned}\ ] ] which is essentially the same as the variance of vw .
we compare vw ( and random projections ) with @xmath0-bit minwise hashing for the task of estimating inner products on binary data . with binary data ,
i.e. , @xmath225 , we have @xmath226 , @xmath227 , @xmath228 .
the variance ( [ eqn_var_vw ] ) ( by using @xmath131 ) becomes @xmath229 we can compare this variance with the variance of @xmath0-bit minwise hashing . because the variance ( [ eqn_var_b ] ) is for estimating the resemblance , we need to convert it into the variance for estimating the inner product @xmath129 using the relation : @xmath230 for @xmath0-bit minwise hashing , each sample is stored using only @xmath0 bits . for vw ( and random projections ) , we assume each sample is stored using 32 bits ( instead of 64 bits ) for two reasons : ( i ) for binary data , it would be very unlikely for the hashed value to be close to @xmath4 , even when @xmath10 ; ( ii ) unlike @xmath0-bit minwise hashing , which requires exact bit - matching in the estimation stage , random projections only need to compute the inner products for which it would suffice to store hashed values as ( double precision ) real numbers . if @xmath233 , then @xmath0-bit minwise hashing is more accurate than binary random projections .
equivalently , when @xmath233 , in order to achieve the same level of accuracy ( variance ) , @xmath0-bit minwise hashing needs smaller storage space than random projections .
there are two issues we need to elaborate on : 1 . here
, we assume the purpose of using vw is for _ data reduction_. that is , @xmath18 is small compared to the number of non - zeros ( i.e. , @xmath234 , @xmath235 ) .
we do not consider the case when @xmath18 is taken to be extremely large for the benefits of _ compact indexing _ without achieving _
data reduction_. 2 .
because we assume @xmath18 is small , we need to represent the sample with enough precision .
that is why we assume each sample of vw is stored using 32 bits .
in fact , since the ratio @xmath236 is usually very large ( e.g. , @xmath237 ) by using 32 bits for each vw sample , it will remain to be very large ( e.g. , @xmath238 ) even if we only need to store each vw sample using 16 bits . without loss of generality
, we can assume @xmath239 ( hence @xmath240 ) .
figures [ fig_gb8 ] to [ fig_gb1 ] display the ratios ( [ eqn_g_vw ] ) for @xmath241 , respectively . in order to achieve high learning accuracies , @xmath0-bit minwise hashing requires @xmath242 ( or even 8) . in each figure , we plot @xmath243 for @xmath244 and full ranges of @xmath235 and @xmath129 .
we can see that @xmath236 is much larger than one ( usually 10 to 100 ) , indicating the very substantial advantage of @xmath0-bit minwise hashing over random projections .
note that the comparisons are essentially independent of @xmath32 .
this is because in the variance of binary random projection ( [ eqn_g_vw ] ) the @xmath245 term is negligible compared to @xmath246 in binary data as @xmath32 is very large . to generate the plots
, we used @xmath247 ( although practically @xmath32 should be much larger ) .
+ * conclusion * : our theoretical analysis has illustrated the substantial improvements of @xmath0-bit minwise hashing over the vw algorithm and random projections in binary data , often by 10- to @xmath248-fold .
we feel such a large performance difference should be noted by researchers and practitioners in large - scale machine learning . | in this paper , we first demonstrate that @xmath0-bit minwise hashing , whose estimators are positive definite kernels , can be naturally integrated with learning algorithms such as svm and logistic regression . we adopt a simple scheme to transform the nonlinear ( resemblance ) kernel into linear ( inner product ) kernel ; and
hence large - scale problems can be solved extremely efficiently .
our method provides a simple effective solution to large - scale learning in massive and extremely high - dimensional datasets , especially when data do not fit in memory .
+ we then compare @xmath0-bit minwise hashing with the vowpal wabbit ( vw ) algorithm ( which is related the count - min ( cm ) sketch ) .
interestingly , vw has the same variances as random projections .
our theoretical and empirical comparisons illustrate that usually @xmath0-bit minwise hashing is significantly more accurate ( at the same storage ) than vw ( and random projections ) in binary data .
furthermore , @xmath0-bit minwise hashing can be combined with vw to achieve further improvements in terms of training speed , especially when @xmath0 is large . |
initially introduced in @xcite as a discrete approximation of langevin dynamics , the model of repeated quantum interactions has found since many applications ( quantum trajectories , stochastic control , etc . ) . in this work
we generalize this model by allowing _
random _ interactions at each time step .
our main focus is the long - time behavior of the reduced dynamics .
our viewpoint is that of quantum open systems , where a `` small '' system is in interaction with an inaccessible environment ( or an auxiliary system ) .
we are interested in the reduced dynamics of the small system , which is described by the action of quantum channels .
when repeating such interactions , under some mild conditions on the spectrum of the quantum channel , we show that the successive states of the small system converge to the invariant density matrix of the channel .
these considerations motivated us to consider random invariant states , and we introduce a new probability measure on the set of density matrices .
there exists extensive literature @xcite on what is a `` typical '' density matrix .
there are two general categories of such probability measures on @xmath3 : measures that come from metrics with statistical significance and the so - called `` induced measures '' , where density matrices are obtained as partial traces of larger , random pure states .
our construction from section [ sec : fixed ] falls into the second category , since our model involves an open system in interaction with a chain of `` auxiliary '' systems .
next , we introduce two models of random quantum channels . in the first model ,
we allow for the states of the auxiliary system to be random . in the second one ,
the unitary matrices acting on the coupled system are assumed random , distributed along the haar invariant probability on the unitary group , and independent between different interactions .
since the ( random ) state of the system fluctuates , almost sure convergence does not hold , and we state results in the ergodic sense .
the article is structured as follows .
the section [ sec : rep_int ] is devoted to presenting the model of quantum repeated interactions and its description via quantum channels .
section [ sec : spectral ] contains some general facts about the spectra of completely positive maps , as well as some related tools from matrix analysis .
next , in section [ sec : fixed ] we study our first model , where the interaction unitary is a fixed , deterministic matrix .
we prove that , under some assumptions on the spectrum of the quantum channel , the state of the system converges to the invariant state of the channel .
it is at this time that we introduce the new ensemble of random density matrices , by transporting the unitary haar measure via the application which maps a channel to its invariant state .
the final two sections are devoted to introducing two models of random quantum channels , one where the interaction unitary is constant and the auxiliary states are i.i.d .
density matrices ( sec . [
sec : random_env ] ) and another where the interaction unitaries are independent and haar distributed ( sec .
[ sec : iid_unitaries ] ) .
we introduce now some notation and recall some basic facts and terminology from quantum information theory .
we write @xmath4 for the set of self - adjoint @xmath5 complex matrices and @xmath3 for the set of _ density matrices _ ( or states ) , @xmath6 = 1\}$ ] .
since our main focus is quantum information , all hilbert spaces in this article are complex and finite dimensional .
scalar products are assumed linear in the second coordinate and , for two vectors @xmath7 we denote by @xmath8 the map @xmath9 an unit vector @xmath10 is called a _ pure state _ and it is assimilated often with the orthogonal projection on @xmath11 , @xmath12 . finally , for a matrix @xmath13
, we define its _ partial trace _ with respect to @xmath14 as the unique element @xmath15 \in { \mathcal{b}}({\mathcal{h}})$ ] which verifies @xmath16 = \operatorname{tr}[a ( x \otimes \operatorname{i}_{\mathcal{k } } ) ] , \quad \forall x \in { \mathcal{b}}({\mathcal{h}}).\ ] ] we shall also extensively use the _ haar _ ( or uniform ) measure @xmath17 on the unitary group @xmath18 ; it is the unique probability measure which is invariant by left and right multiplication by unitary elements : @xmath19
in this introductory section we give a description of the physical model we shall use in the rest of the paper : _ repeated quantum interactions_. the setting , a system interacting repeatedly with `` independent '' copies of an environment , was introduced by s. attal and y. pautrat in @xcite where it was shown that in the continuous limit ( when the time between interactions approaches zero ) , the dynamics is governed by a quantum stochastic differential equation . a different model , where after each interaction an indirect quantum measurement of the system is performed ,
was considered by the second named author in @xcite and shown to converge in the limit to the so - called stochastic schrdinger equations . here ,
we are concerned only with the discrete setting and with the limit of a large number of interactions .
the study of random quantum trajectories is postponed to a later paper .
consider a quantum system @xmath20 described by a complex hilbert space state @xmath21 .
in realistic physical models , @xmath20 is usually a quantum system with relatively few degrees of freedom and it represents the object of interest of our study ; we shall refer to it as the _ small system_. consider also another quantum system @xmath22 which interacts with the initial small system @xmath20 .
we shall call @xmath22 the _ environment _ and we denote by @xmath14 its hilbert state space . in this work
we consider finite dimensional spaces @xmath23 and @xmath24 .
we shall eventually be interested in _ repeated _ interactions between @xmath20 and independent copies of @xmath22 , but let us start with the easier task of describing a single interaction between the `` small '' system and the environment .
assume that the initial state of the system is a product state @xmath25 , where @xmath26 and @xmath27 are the respective states of the small system and the environment .
the coupled system undergoes an unitary evolution @xmath28 and @xmath29 is the global state after the interaction .
the unitary operator @xmath28 comes from a hamiltonian @xmath30 where the operators @xmath31 and @xmath32 are the free hamiltonians of the systems @xmath20 and @xmath22 respectively and @xmath33 represents the interaction hamiltonian
. we shall be interested in the situation where @xmath34 , otherwise there is no coupling and the system and the environment undergo separate dynamics . in this general case , the evolution unitary operator @xmath28 is given by @xmath35 where @xmath36 is the interaction time .
hence , the state of the coupled system @xmath37 after one interaction is given by @xmath38 since one is interested only in the dynamics of the `` small '' system @xmath20 , after taking the partial trace we obtain the final state of @xmath20 , @xmath39.\ ] ] we now move on to describe successive interactions between @xmath20 and a chain of independent copies of @xmath22 . in order to do this , consider the countable tensor product @xmath40 where @xmath41 is the @xmath42-th copy of the environment ( @xmath43 ) .
this setting can be interpreted in two different ways : globally , as an evolution on infinite dimensional countable tensor product @xmath44 , or by discarding the environment , as a discrete evolution on @xmath45 .
since we are interested only in the evolution of the `` small '' system , the latter approach is the better choice . from eq .
( [ eq : single_interaction ] ) , we obtain the recurrence relation @xmath46,\ ] ] where @xmath47 are the successive states of the system @xmath20 at times @xmath48 and @xmath42 , and @xmath49 and @xmath50 are the interaction unitary and respectively the state of the auxiliary system @xmath22 for the @xmath42-th interaction .
note that at this stage we work in a general setting , without making any assumptions on the sequences @xmath51 and @xmath52 .
we introduce now a more parsimonious description of repeated quantum interactions , via quantum channels .
recall that a linear map @xmath53 is called @xmath54-positive if the extended map @xmath55 is positive .
@xmath56 is called _ completely positive _ if it is @xmath54-positive for all @xmath57 ( in fact @xmath58 suffices ) and _ trace preserving _ if @xmath59 = \operatorname{tr}[x]$ ] for all @xmath60 . by definition ,
quantum channel _ is a trace - preserving , completely positive linear map .
the next proposition gives two very important characterizations of quantum channels .
[ prop : stinespring_kraus ] a linear map @xmath61 is a quantum channel if and only if one of the following two equivalent conditions holds . 1 . * ( stinespring dilation ) * there exists a finite dimensional hilbert space @xmath62 , a density matrix @xmath63 and an unitary operation @xmath64 such that @xmath65 , \quad \forall x \in { \mathcal{m}}_d({\mathbb{c}}).\ ] ] 2 . *
( kraus decomposition ) * there exists an integer @xmath54 and matrices @xmath66 such that @xmath67 and @xmath68 it can be shown that the dimension of the ancilla space in the stinespring dilation theorem can be chosen @xmath69 and @xmath27 can be chosen to be a rank one projector .
a similar result holds for the number of kraus operators : one can always find a decomposition with @xmath70 operators .
choi rank _ of a quantum channel @xmath56 is the least positive integer @xmath54 such that @xmath56 admits a kraus decomposition ( [ eq : kraus_decomposition ] ) with @xmath54 operators @xmath71 .
we see now that eq .
( [ eq : rec_relation ] ) can be re - written as @xmath72 where @xmath73 is the quantum channel @xmath74.\end{aligned}\ ] ] after @xmath42 such interactions , the state of the system becomes @xmath75 let us now consider a fixed channel @xmath76 and show that the stinesping and kraus form of @xmath56 are connected in a simple fashion . to this end ,
start with the stinespring form of @xmath56 and pick some orthonormal bases @xmath77 and @xmath78 of respectively @xmath79 and @xmath62 such that the state of the environment @xmath27 diagonalizes : @xmath80 next , endow the product space @xmath81 with the basis @xmath82 this particular ordering of the tensor product basis was preferred in order to have a simple expression for the partial trace operation with respect to the environment @xmath14 .
indeed , if a matrix @xmath83 is written in this basis and viewed as a @xmath84 matrix of blocks @xmath85 : @xmath86 then the computation of the partial trace with respect to @xmath87 reads @xmath88 = \operatorname{tr}_k\left[\sum_{i , j=1}^{d'}a_{ij } \otimes { | f_i \rangle \langle f_j |}\right ] = \sum_{i , j=1}^{d'}a_{ij } \cdot { \langle f_j , f_i\rangle } = a_{11 } + a_{22 } + \cdots + a_{d'd'}.\ ] ] in other words , the partial trace of @xmath89 over the environment @xmath14 is simply the trace of the block - matrix , that is the sum of the diagonal blocks of @xmath89 .
we apply now these ideas to the stinespring form of a quantum channel , @xmath90 $ ] . written as a block matrix in the basis defined in eq .
( [ eq : product_basis ] ) , the matrix @xmath91 is diagonal , with diagonal blocks given by @xmath92 .
writing @xmath64 in the same fashion and taking the partial trace , we obtain @xmath93 = \sum_{i , j=1}^{d'}b_j u_{ij } x u_{ij}^ * = \sum_{i , j=1}^{d'}(\sqrt{b_j } u_{ij } ) x ( \sqrt{b_j } u_{ij})^*,\ ] ] where @xmath94 are the blocks of the interaction unitary @xmath28 .
one recognizes a kraus decomposition for @xmath56 , where the kraus elements are rescaled versions of the blocks of the stinespring matrix @xmath28 .
moreover , if @xmath27 is a rank one projector then all the @xmath95 s are zero except one , hence the kraus decomposition we obtained has @xmath96 elements .
since we shall be interested in repeated applications of quantum channels , it is natural that spectral properties of these maps should play an important role in what follows .
one should note that most results of this section can be generalized to infinite dimensional hilbert spaces .
the next lemma gathers some basic facts about quantum channels .
since quantum channels preserve the compact convex set of density matrices @xmath3 , the first affirmation follows from the fixed point theorem of markov - kakutani @xcite . the second and the third assertions are trivial ( see @xcite for further results on @xmath97 norms of quantum channels ) , and
the last one is a consequence of 2-positivity .
[ lem : channel ] let @xmath61 a quantum channel
. then 1 .
@xmath56 has at least one invariant element , which is a density matrix ; 2 .
@xmath56 has trace operator norm of 1 ; 3 .
@xmath56 has spectral radius of 1 ; 4 .
@xmath56 satisfies the schwarz inequality @xmath98 if one looks at a channel @xmath56 as an operator in the hilbert space @xmath99 endowed with the hilbert - schmidt scalar product , then one can introduce @xmath100 , the _ dual map _ of @xmath56 .
it is defined by the relation @xmath101 = \operatorname{tr}[\psi(x ) y ] , \quad \forall x , y \in { \mathcal{m}}_d({\mathbb{c}}).\ ] ] from kraus decomposition @xmath102 , one can obtain a kraus decomposition for the dual channel , @xmath103 .
note that the trace preserving condition for @xmath56 , @xmath104 reads now @xmath105 .
hence , the dual of a quantum channel is a unital ( not necessarily trace - preserving ) completely positive linear map .
using this idea , one can see that the partial trace operation @xmath106 is the dual of the tensoring operation @xmath107 , @xmath108 .
we now introduce some particular classes of positive maps which are known to have interesting spectral properties .
let @xmath61 be a positive linear map .
@xmath56 is called _ strictly positive _ ( or positivity improving ) if @xmath109 for all @xmath110 .
@xmath56 is called _ irreducible _ if there is no projector @xmath111 such that @xmath112 for some @xmath113 .
[ eg : unitary_conj ] let @xmath114 be a fixed unitary and consider the channel @xmath115 , @xmath116 .
it is easy to check that the spectrum of @xmath56 is the set @xmath117 since @xmath56 maps pure states ( i.e. rank - one projectors ) to pure states , it neither irreducible , nor strictly positive . obviously , a strictly positive map is irreducible .
in fact , the following characterization of irreducibility is known @xcite . a positive linear map @xmath53 is irreducible if and only if the map @xmath118 is strictly positive . irreducible unital maps which satisfy the schwarz inequality have very nice peripheral spectra .
the proof of the following important result can be found in one of @xcite , in more general settings .
if @xmath100 is a unital , irreducible map on @xmath99 which satisfies the schwarz inequality , then the set of peripheral ( i.e. modulus one ) eigenvalues is a ( possibly trivial ) subgroup of the unit circle @xmath119 .
moreover , every peripheral eigenvalue is simple and the corresponding eigenspaces are spanned by unitary elements of @xmath99 .
irreducible ( and , in particular , strictly positive ) quantum channels have desirable spectral properties , hence the interest one has for these classes of maps .
as we shall see in section [ sec : fixed ] , irreducible maps are in certain sense generic . on the other hand ,
the strict positivity condition is rather restrictive and not suitable for the considerations on this work .
next , we develop these ideas , giving criteria for irreducibility and for strict positivity .
let us start by analyzing strict positivity .
subspaces of product spaces @xmath120 with high entanglement have received recently great attention . in this direction ,
applications to the additivity conjecture @xcite are the most notable ones .
the results in these papers , which rely on probability theory techniques deal with von neumann entropy .
when one looks at the rank , projective algebraic geometry comes into play .
indeed , possible states of the coupled system are modeled by the projective space @xmath121 .
this space contains the _ product states _
, @xmath122 as a subset called _ the segre variety_. the following lemma , a textbook result in algebraic geometry , is obtained by computing the dimension of the segre variety ( see @xcite ) .
[ lem : entangled_subspace ] the maximum dimension of a subspace @xmath123 which does not contain any non - zero product elements @xmath124 is @xmath125 . as a rather simple consequence of this lemma , we obtain a necessary condition for strict positivity .
let @xmath61 be strictly positive quantum map .
then the choi rank of @xmath56 is at least @xmath126 .
let @xmath127 be a minimal kraus decomposition of a strictly positive channel @xmath56 .
for all @xmath128 , @xmath129 has full rank , and thus , for all non - zero @xmath130 , @xmath131 = \sum_{i=1}^k { \left| { \langle y ,
l_i x\rangle}\right|}^2 > 0.\ ] ] hence , for all non - zero @xmath132 , there exist an @xmath133 such that @xmath134 \neq 0 $ ] , or , in other words , @xmath135 is _ not _ orthogonal to @xmath136 with respect to the hilbert - schmidt scalar product .
consider now the space @xmath137 .
obviously , @xmath138 does not contain any rank one matrices @xmath136 . under the usual isomorphism @xmath139 , product vectors
@xmath124 are identified with rank one matrices @xmath136 , so , by the lemma [ lem : entangled_subspace ] , we get @xmath140 . since @xmath138 is the intersection of @xmath54 subspaces of dimension @xmath141
, we get @xmath142 which implies @xmath143 .
we now turn to irreducible quantum maps and state some results which will be useful later , when showing that irreducibility is generic for a specific model of random quantum channels .
the following result of @xcite gives necessary and sufficient conditions for a map written in the kraus form to be irreducible .
we denote by @xmath144 the lattice of invariant subspaces of an operator @xmath145 .
[ prop : irred_lat ] consider the map @xmath53 defined by @xmath146 , with @xmath147 , @xmath148 . then @xmath56 is irreducible if and only if @xmath149 is trivial . of course
, quantum channels of choi rank one ( i.e. unitary conjugations , see also example [ eg : unitary_conj ] ) , @xmath150 , with @xmath151 can not be irreducible , since they leave invariant eigenprojectors of @xmath152 . when looking at channels with choi rank at least two , an useful criterion for deciding whether @xmath153 is trivial or not is given by the following two results .
the first proposition gives necessary and sufficient conditions for two matrices @xmath89 and @xmath154 to share a common eigenvector , and the second one generalizes this idea to arbitrary common subspaces .
two matrices @xmath155 have a common eigenvector if and only if @xmath156 \neq \{0\},\ ] ] or , equivalently , iff @xmath157^ * \cdot [ a^i , b^j ] = 0.\ ] ] in order to move on from common eigenvectors to common invariant subspaces , we consider antisymmetric tensor powers ( or wedge powers ) of matrices ( see @xcite , ch .
i ) . given @xmath158 and and an integer @xmath159 , the @xmath54-th wedge power of @xmath89 , denoted by @xmath160 , is defined as the restriction of @xmath161 to the antisymmetric tensor product @xmath162 .
more precisely , @xmath160 is a @xmath163 matrix , where @xmath164 .
its matrix elements are indexed by couples @xmath165 of strictly increasing sequences of size @xmath54 from @xmath166 : @xmath167,\ ] ] where @xmath168 $ ] is the submatrix of @xmath89 with rows indexed by @xmath169 and columns indexed by @xmath27 .
the next result of @xcite is an easy consequence of the fact that if @xmath170 are eigenvalues of @xmath89 with linear independent vectors @xmath171 , then @xmath172 is an eigenvalue of @xmath160 with corresponding eigenvector @xmath173 .
[ prop : shemesh_gen ] let @xmath174 be two complex matrices . if @xmath89 and @xmath154 have a common invariant subspace of dimension @xmath54 ( for @xmath175 ) , then their @xmath54-th wedge powers have a common eigenvector , and hence ( we put @xmath164 ) @xmath176 \neq \{0\},\ ] ] or , equivalently , @xmath177^ * \cdot [ ( a^{\wedge k})^i,(b^{\wedge k})^j ] = 0.\ ] ] the preceding conditions turn out to be sufficient under more stringent assumptions on the matrices @xmath89 and @xmath154 ( see @xcite for further details ) .
the main point of the two preceding results is that there exists an universal polynomial @xmath178 $ ] with the property that whenever two matrices @xmath179 and @xmath180 have a non - trivial common invariant subspace , @xmath181 .
this fact ( together with proposition [ prop : irred_lat ] ) will be useful later in this work , when we shall show that a generic class of quantum maps are irreducible .
in this section we consider repeated interactions with a fixed unitary matrix @xmath28 ( @xmath182 ) and fixed state of the environment @xmath27 ( @xmath183 ) . by the results of the previous section ,
the recurrence relation which governs the discrete , deterministic dynamics is @xmath184.\ ] ] iterating this formula , one obtains the state of the system after @xmath42 interactions : @xmath185 where @xmath186 was the initial state of the system .
there is one obvious situation in which the asymptotic properties of the sequence @xmath187 can be established .
indeed , from lemma [ lem : channel ] , one knows that all quantum channels have eigenvalue 1 and that all other eigenvalues have module less than 1 .
let @xmath188 be the set of all quantum channels that have 1 as a simple eigenvalue and all other eigenvalues are contained in the _ open _ unit disc . since 1 is a simple eigenvalue , @xmath56 has an unique fixed point which is ( by lemma [ lem : channel ] ) a density matrix @xmath189 . using the jordan form of @xmath56
, one can show the following result ( @xcite ) .
[ prop : conv_1_unique ] let @xmath190 be a fixed quantum channel . then , for all @xmath191 , @xmath192 where @xmath193 is the unique invariant state of @xmath56 .
the importance of the peripheral spectrum of a quantum channel is illustrated in the following example .
[ eg : periph ] consider the following channel @xmath194 @xmath195 where the pauli matrices are given by @xmath196 direct computation shows that @xmath197 , @xmath198 and @xmath199 .
hence , the peripheral spectrum of @xmath56 has 2 simple eigenvalues , @xmath200 and @xmath201 .
however , for @xmath202 , one has @xmath203 which does not converge in the limit @xmath204 .
hence , the simplicity of the eigenvalue @xmath200 does not suffice to have convergence to the invariant state .
note also that the channel @xmath56 is irreducible , since @xmath205 and @xmath206 do not have any common non - trivial invariant subspaces .
let us now show that the class @xmath207 of quantum channels which have 1 as an unique peripheral eigenvalue is generic in a certain sense . to this end
, we shall introduce a model of _ random _ quantum channel , based on the stinespring decomposition .
to start , fix the dimension of the environment @xmath96 and a state @xmath63 .
next , consider an unitary random matrix @xmath28 distributed along the ( uniform ) haar measure @xmath208 on @xmath209 . to the state @xmath27 and the evolution operator @xmath28
, we associate the quantum channel @xmath210 . in this way
, we define a model of random quantum channels by considering the image measure of the haar distribution @xmath208 on the set of quantum channels . in the recent preprint @xcite ,
the authors study a similar model of random quantum channels , focusing on the spectral properties of the random matrix defining the channel .
more precisely , we claim that if the state of the environment @xmath27 is fixed and the interaction unitary @xmath64 is chosen randomly with the uniform haar distribution @xmath208 , then , with probability one , the channel @xmath211 admits 1 as the unique eigenvalue on the unit circle . here
we need another fact from algebraic geometry , summarized in the following lemma ( for a similar result , one should have a look at proposition 2.6 of @xcite ) .
[ lem : unitary_poly ] given a polynomial @xmath212 $ ] , the set @xmath213 is either equal to the whole set @xmath18 or it has haar measure 0 .
we start by noticing that the real algebraic set @xmath18 is irreducible .
this follows from the connectedness of @xmath18 ( in the usual topology ) and from the fact that irreducible components of a linear algebraic group are disjoint ( @xcite , 7.3 ) .
the set @xmath214 is the intersection of the irreducible variety @xmath18 with the variety @xmath215 of zeros of the polynomial @xmath111 . if @xmath216 , then @xmath217 ; otherwise , the dimension of @xmath214 is strictly smaller than @xmath218 , the real dimension of @xmath18 .
since the haar measure is just the integration of an invariant differential form , it has a density in local coordinates ( @xcite , ch .
5 ) and hence @xmath219 in this case .
[ thm : haar_1_unique ] let @xmath27 be a fixed density matrix of size @xmath96 . if @xmath28 is a random unitary matrix distributed along the haar invariant probability @xmath208 on @xmath209 , then @xmath220 almost surely .
the proof goes in two steps .
first , we show that @xmath73 is almost surely irreducible and then we conclude by a simple probabilistic argument .
let us start by applying lemma [ lem : unitary_poly ] to show that a random quantum channel is almost surely irreducible .
to this end , using eq .
( [ eq : kraus_from_u ] ) , we obtain a set of kraus operators for @xmath73 which are sub - matrices of @xmath64 .
consider two such kraus operators @xmath174 ( choose @xmath221 such that @xmath222 and take @xmath223 , @xmath224 ) . using proposition [ prop : irred_lat ] , to show irreducibility it suffices to see that @xmath89 and @xmath154 do not have a non - trivial common invariant subspace .
let @xmath175 be the dimension of a potentially invariant common subspace of @xmath89 and @xmath154 .
by the criterion in proposition [ prop : shemesh_gen ] , there exists a polynomial @xmath225 in the entries of @xmath89 and @xmath154 ( and thus in the entries of @xmath28 ) such that if @xmath226 is non - zero , then @xmath89 and @xmath154 do not share a @xmath54-dimensional invariant space .
note that @xmath225 can not be identically zero : for two small enough matrices @xmath227 without common invariant subspaces , one can build a unitary matrix @xmath228 such that @xmath229 , @xmath230 . by the lemma [ lem : unitary_poly ] , @xmath208-almost all unitary matrices @xmath28
give kraus operators @xmath89 and @xmath154 that do not have any @xmath54-dimensional invariant subspaces in common .
since the intersection of finitely many full measure sets has still measure one , almost all quantum channels are irreducible .
consider now a random channel @xmath73 which we can assume irreducible .
since the peripheral spectrum of an irreducible channel is a multiplicative subgroup of the unit circle @xmath119 , it suffices to show that for all element @xmath231 of the finite set @xmath232 , with haar probability one , @xmath231 is not an eigenvalue of @xmath73 .
we use the same trick as earlier .
consider such a complex number @xmath231 and introduce the polynomial @xmath233 $ ] , where @xmath73 is seen as a matrix @xmath234 .
since @xmath235 and the identity channel @xmath236 has only unit eigenvalues , @xmath237 can not be identically zero , and the conclusion follows .
the main difficulty in the proof of the preceding result comes from the fact that the matrices @xmath89 and @xmath154 are `` correlated '' : two blocks of an unitary matrix must satisfy norm and ( maybe ) orthogonality relations .
hence the need to use sophisticated geometric algebra techniques .
proving that two independent random gaussian ( or unitary ) matrices do not share non - trivial invariant subspaces is much simpler and does not require the use of such techniques .
we now move on and apply the previous results to constructing a new family of probability distributions on the set of density matrices .
the main idea is to assign , whenever possible , to a random unitary @xmath64 its unique invariant density matrix @xmath193 . in this way ,
the haar measure @xmath208 on the unitary group @xmath209 is transported to the set of density matrices @xmath3 .
let us now make this construction more precise .
the new family of probability measures shall be indexed by an integer @xmath238 ( the dimension of the auxiliary system ) and by a non - increasing probability vector @xmath239 : @xmath240 and @xmath241 ( these are the eigenvalues of the state of the auxiliary system ) .
for such a couple @xmath242 consider a density matrix @xmath63 with eigenvalue vector @xmath243 ( the eigenvectors of @xmath27 do not matter , see lemma [ prop : asympt_measures ] ) . as it follows from proposition [ thm : haar_1_unique ] , for almost all unitaries @xmath64 , the channel @xmath73 satisfies the hypotheses of proposition [ prop : conv_1_unique ] .
hence , for almost all @xmath28 and for all density matrices @xmath191 , @xmath244 , where @xmath193 is the unique invariant state of @xmath73 .
we have defined almost everywhere an application @xmath245 we denote by @xmath246 the image measure of the haar probability @xmath208 on @xmath209 by the previous application ( notice that we dropped the integer parameter @xmath96 , since this is the dimension of the vector @xmath243 ) .
we call @xmath247 the _ asymptotic induced measure _ on the set of density matrices .
we now motivate the term `` asymptotic induced '' in the previous definition by showing how the measures @xmath247 relate to the induced random density matrices considered in @xcite .
let us recall here how these measures are constructed and how one can sample from this distribution .
the physical motivation behind the induced measures comes from the following setup .
assume that a system @xmath20 is coupled to an environment @xmath22 and that the whole is in a pure state @xmath248 .
if one has no _ a priori _ knowledge about the state @xmath249 , then it is natural to assume that @xmath249 is a random uniform element on the unit sphere of the product space @xmath250 .
the distribution of the partial trace over the environment @xmath251\ ] ] is called the _ induced measure _ and it is denoted by @xmath252 ( the parameter @xmath253 is the dimension of the environment ) .
we refer the interested reader to @xcite for more information on these measures .
since the distribution of an uniform norm - one vector @xmath249 is the equal to the distribution of @xmath254 , where @xmath255 is any fixed norm - one vector and @xmath28 is a haar unitary , @xmath252 is also the distribution of the matrix @xmath256.\ ] ] if one chooses @xmath257 , where @xmath258 and @xmath259 are the first vectors of the canonical basis of @xmath260 and respectively @xmath261 , then @xmath262 = \phi^{u , \beta_0}(\rho_0),\ ] ] with @xmath263 and @xmath264 .
hence , the induced measure @xmath252 is the distribution of the result of _ one _ application of a random channel @xmath265 on the constant matrix @xmath186 : @xmath266 . on the other hand , after a large number of identical interactions , one gets @xmath267^n(\rho_0).\ ] ] in this work we have shown that with @xmath208-probability one , @xmath193 is a well defined density matrix - valued random variable which does not depend on the value of @xmath186 . since the eigenvalue vector of @xmath268 is @xmath269 we have that @xmath270 .
now , the relation between the two families of measures is clear : the induced measure @xmath252 is the distribution of the density matrix after one interaction , whereas @xmath271 is the distribution at the limit , after a large number of interactions .
the reader may notice that this analogy is valid only in the case where @xmath272 ( pure state on the environment ) .
generalizations of the ( usual ) induced measures to other environment states are possible , but out of the scope of the present work . to further compare the asymptotic and
the one interaction induced measures , we plotted the spectra of samples of density matrices from both families in figures [ fig : asympt ] and [ fig : induced ] .
in particular , one should compare figure [ fig : asympt:22 ] with figure [ fig : induced:22 ] ( @xmath273 ) , figure [ fig : asympt:33 ] with figure [ fig : induced:33 ] ( @xmath274 ) and figure [ fig : asympt:35 ] with figure [ fig : induced:35 ] ( @xmath275 ) .
+ one particularly simple case is obtained by taking @xmath276 .
the measure @xmath247 is then trivial , being equal to the dirac mass supported on the `` chaotic state '' @xmath277 . in the next lemma
we prove some basic properties of the newly introduced measures @xmath247 .
a more thorough investigation of these measures is postponed to a later work .
[ prop : asympt_measures ] the probability measures @xmath247 have the following properties : 1 . for every probability vector @xmath243 , the measure @xmath247 is well defined , in the sense that the distribution of @xmath278^n(\rho_0)$ ] does not depend on the eigenvectors of @xmath27 , but only on the eigenvalue vector @xmath243 .
2 . for all unitary matrix @xmath279 , @xmath26 and @xmath280 have the same distribution ( we say that the measure @xmath247 is _ unitarily invariant _ ) .
there exists a probability measure @xmath281 on the probability simplex @xmath282 such that if @xmath283 is a diagonal matrix sampled from @xmath281 and @xmath215 is an independent haar unitary on @xmath18 , then @xmath284 has distribution @xmath246 .
in other words , the distribution of a random density matrix @xmath285 is determined by the distribution of its eigenvalue vector @xmath286 . to prove the first assertion
, we show that for all @xmath287 , replacing @xmath27 with @xmath288 does not change the distribution of @xmath193 . to see this
, note that by the invariance of the haar probability measure @xmath208 , the random matrices @xmath28 and @xmath289 have the same distribution .
it follows that the same holds for the random channels @xmath73 and @xmath290 and thus for their invariant states .
the second affirmation is proved in the same manner ( this time using a fixed unitary @xmath215 acting on @xmath21 ) and the third one is a trivial consequence of the second .
in the previous section we considered repeated _
identical _ quantum interactions of a system @xmath20 with a chain of identical environment systems @xmath22 .
we now introduce classical randomness in our model by considering random states on the environment @xmath22 . in this model ,
the unitary describing the interaction is a fixed deterministic matrix @xmath64 .
the @xmath42-th interaction between the small system @xmath20 and the environment @xmath291 is given by the following relation : @xmath292,\ ] ] where @xmath52 is a sequence of independent identically distributed random density matrices .
notice that , since @xmath28 is constant , we use the shorthand notation @xmath293 .
we are interested , as usual , in the limit @xmath204 . in this case however , the ( random ) channels @xmath294 do not have in general a common invariant state , so one has to look at ergodic limits .
we use here the machinery developed by l. bruneau , a. joye and m. merkli in @xcite ( see @xcite for additional results in this direction ) . for the sake of completeness ,
let us state their main result .
[ thm : bjm ] let @xmath295 be a sequence of i.i.d
. random contractions of @xmath99 with the following properties : 1 .
there exists a constant vector @xmath296 such that @xmath297 for ( almost all ) @xmath298 ; 2 .
@xmath299 .
then the ( deterministic ) matrix @xmath300 $ ] has eigenvalue 1 with multiplicity one and there exists a constant vector @xmath301 such that @xmath302},\ ] ] where @xmath303}$ ] is the rank - one spectral projector of @xmath300 $ ] corresponding to the eigenvalue 1 .
note that this result doest apply to our situation , mainly for two reasons : the order of the composition of the channels @xmath56 is reversed and the linear applications @xmath294 do not necessarily share a constant invariant state @xmath249 .
this inconvenient can be overcome by considering dual channels ( see section [ sec : spectral ] ) , or , in physicists language , by switching from the schrdinger to the heisenberg picture of quantum mechanics .
duals of quantum channels are unital , hence they have in common the invariant element @xmath304 .
another important benefit of considering duals is that the order of composition of maps is reversed . indeed ,
if one starts from a state @xmath186 , applies successively @xmath42 channels @xmath305 and finally measures an observable @xmath306 , it is easy to see that the expected outcome is @xmath307 = \operatorname{tr}[(\phi_{n-1}\circ \cdots \circ \phi_1)(\rho ) \cdot \psi_n(a ) ] = \cdots = \operatorname{tr}[\rho \cdot ( \psi_1 \circ \cdots \circ \psi_n)(a)].\ ] ] we are now in position to state and prove the analogue of theorem [ thm : bjm ] for infinite products of quantum channels , simply by replacing quantum channels with their duals .
[ thm : prod_q_channels ] let @xmath308 be a sequence of i.i.d .
random quantum channels acting on @xmath99 such that @xmath309 . then @xmath310 $ ] is a quantum channel with an unique invariant state @xmath311 and , @xmath312-almost surely , @xmath313(\rho_0 ) = \theta , \quad \forall \rho_0 \in { \mathcal{m}^{1 , + } } _ d({\mathbb{c}}).\ ] ] let us start by introducing some notation .
let , for some initial state @xmath191 , @xmath314(\rho_0),\ ] ] and consider the dual operators @xmath315 which are , as described earlier , the adjoints of @xmath316 with respect to the hilbert - schmidt scalar product on @xmath99 .
then , for a self - adjoint observable @xmath306 , one has @xmath317 = \operatorname{tr}\left [ \rho_0 \frac 1 n \sum_{n=1}^n ( \psi_1 \circ \cdots \circ \psi_n)(a ) \right].\ ] ] it is easy to see that the random operators @xmath315 satisfy the hypotheses of theorem [ thm : bjm ] on the hilbert space @xmath99 endowed with the hilbert - schmidt scalar product .
indeed , the spectrum of @xmath100 is the complex conjugate of the spectrum of @xmath56 , hence @xmath100 is a contraction ( with respect to the hilbert - schmidt norm ) .
moreover , with non - zero probability , @xmath318 is the unique invariant state of @xmath100 . from the theorem [ thm : bjm ] ,
one obtains the existence of a non - random element @xmath319 such that , @xmath312-almost surely , @xmath320 plugging this into eq .
( [ eq : mu_n_a ] ) , one gets @xmath321 = \operatorname{tr}[\rho_0 { | \operatorname{i}_d \rangle \langle \theta |}a ] = { \langle \theta , a\rangle}_{\text{hs } } \operatorname{tr}[\rho_0 \operatorname{i}_d ] = \operatorname{tr}[\theta^ * a].\ ] ] since the set of density matrices @xmath3 is ( weakly ) closed , @xmath322 and @xmath323 . the fact that @xmath324 is the _
unique _ invariant state of @xmath310 $ ] follows again from theorem [ thm : bjm ] .
when comparing the preceding theorem with the proposition [ prop : conv_1_unique ] , one notes that the hypotheses are relaxed here , asking only that the eigenvalue 1 is simple , without further constraints on the peripheral spectrum .
this is due to the fact that we are considering csaro means and fluctuations ( such as the ones seen in example [ eg : periph ] ) cancel out at the limit .
we now move on to apply the preceding general result to the setting described in the beginning of this section .
recall that the successive interactions were described by i.i.d .
random quantum channels @xmath325 , where @xmath326.\ ] ] since the previous equation is linear in @xmath27 , @xmath327 = \phi^{{\mathbb{e}}[\beta]}$ ] and one gets the following corollary .
let @xmath328 be a sequence of i.i.d .
random density matrices and consider the repeated quantum interaction scheme with constant interaction unitary @xmath28 .
assume that , with non - zero probability , the induced quantum channel @xmath329 has an unique invariant state .
then , @xmath312-almost surely , for all initial states @xmath191 , one has @xmath330(\rho_0 ) = \theta,\ ] ] where @xmath311 is the unique invariant state of the deterministic channel @xmath331}$ ] .
in particular , if @xmath332 = \operatorname{i}_{d'}/{d'}$ ] , then @xmath324 is the `` chaotic '' state @xmath333 .
we now consider a rather different framework from the one studied in sections [ sec : fixed ] and [ sec : random_env ] .
we shall assume that the interaction unitaries @xmath49 acting on the coupled system @xmath250 are random independent and identically distributed ( i.i.d . ) according to the unique invariant ( haar ) probability measure @xmath208 on the group @xmath209 .
this is a rather non - conventional model from a physical point of view , but it permits to relax hypothesis on the successive states of the environment and to obtain an ergodic - type convergence result .
as before , we start with a fixed state @xmath191 .
the @xmath42-th interaction is given by @xmath334 , where @xmath52 is a ( possibly random ) sequence of density matrices on @xmath14 and @xmath51 is a sequence of i.i.d .
haar unitaries of @xmath209 independent of the sequence @xmath52 .
note that we make no assumption on the joint distribution of the sequence @xmath52 ; in particular , the environment states can be correlated or they can have non - identical probability distributions .
the state of the system after @xmath42 interactions is given by the forward iteration of the applications @xmath335 : @xmath336 since we made no assumption on the successive states of the environment @xmath337 , the sequence @xmath187 is not a markov chain in general .
indeed , the density matrices @xmath52 were not supposed independent , hence @xmath338 ( and thus @xmath339 ) may depend not only on the present randomness , but also on past randomness , such as @xmath340 , @xmath341 , etc .
although the sequence @xmath187 lacks markovianity , it has the following important invariance property .
[ lem : u - iid - same - law ] let @xmath342 be a sequence of i.i.d .
haar unitaries independent of the family @xmath343 and consider the sequence of successive states @xmath187 defined in eq .
( [ eq : cocycle ] ) .
then the sequences @xmath187 and @xmath344 have the same distribution .
consider a i.i.d .
sequence @xmath342 of @xmath17-distributed unitaries independent from the @xmath49 s and the @xmath50 s appearing in eq .
( [ eq : cocycle ] ) . to simplify notation
, we put @xmath345 . we also introduce the following sequence of ( random ) @xmath346 unitary matrices : @xmath347 a simple calculation shows that @xmath348 it follows that , in order to conclude , it suffices to show that the family @xmath349 is i.i.d . and @xmath208-distributed ( it is obviously independent of the @xmath27 s ) .
we start by proving that , at fixed @xmath42 , @xmath350 is @xmath208-distributed .
since the families @xmath51 and @xmath342 are independent , one can consider realizations of these random variables on different probability space @xmath351 and @xmath352 .
for a positive measurable function @xmath353 , one has ( we put @xmath354 ) @xmath355 & = { \mathbb{e}}[f((v_n \otimes \operatorname{i } ) u_n ( v_{n-1}^ { * } \otimes \operatorname{i } ) ) ] = \\ & = \int f((v_n(\omega^2_{n } ) \otimes \operatorname{i } ) u_n(\omega^1_{n } ) ( v_{n-1}^{*}(\omega^2_{n-1 } ) \otimes \operatorname{i } ) ) d{\mathbb{p}}(\omega^2_{n})d{\mathbb{p}}(\omega^1_{n})d{\mathbb{p}}(\omega^2_{n-1})\\ & = \int \left ( \int f((v_n(\omega^2_{n } ) \otimes \operatorname{i } ) u_n(\omega^1_{n } ) ( v_{n-1}^{*}(\omega^2_{n-1 } ) \otimes \operatorname{i } ) ) d{\mathbb{p}}(\omega^1_{n})\right ) d{\mathbb{p}}(\omega^2_{n})d{\mathbb{p}}(\omega^2_{n-1})\\ & \stackrel{(*)}{= } \int { \mathbb{e}}[f(u_n ) ] d{\mathbb{p}}(\omega^2_{n})d{\mathbb{p}}(\omega^2_{n-1 } ) = { \mathbb{e}}[f ( u_n)],\end{aligned}\ ] ] where we used in @xmath356 the fact that the haar probability on @xmath209 is invariant by left and right multiplication with constant unitaries .
we now claim that the r.v .
@xmath350 are independent . for some positive measurable functions @xmath357 ,
one has @xmath358 & = { \mathbb{e}}\left[\prod_{k=1}^n f_k((v_k \otimes \operatorname{i } ) u_k ( v_{k-1}^ { * } \otimes \operatorname{i}))\right ] = \\ & = \int \prod_{k=1}^n f_k((v_k(\omega^2_{k } ) \otimes \operatorname{i } ) u_k(\omega^1_{k } ) ( v_{k-1}^{*}(\omega^2_{k-1 } ) \otimes \operatorname{i } ) ) \prod_{k=1}^n d{\mathbb{p}}(\omega^1_{k})d{\mathbb{p}}(\omega^2_{k})\\ & = \int \prod_{k=1}^n \left ( \int f_k((v_k(\omega^2_{k } ) \otimes \operatorname{i } ) u_k(\omega^1_{k } ) ( v_{k-1}^{*}(\omega^2_{k-1 } ) \otimes \operatorname{i } ) ) d{\mathbb{p}}(\omega^1_{k } ) \right ) \prod_{k=1}^n d{\mathbb{p}}(\omega^2_{k } ) \\ & \stackrel{(**)}{= } \int { \mathbb{e}}[f_k(u_k ) ] \prod_{k=1}^n d{\mathbb{p}}(\omega^2_{k } ) = \prod_{k=1}^n { \mathbb{e}}[f_k(u_k ) ] \stackrel{(***)}{= } \prod_{k=1}^n { \mathbb{e}}[f_k(\tilde u_k)].\end{aligned}\ ] ] again , we used in the equality @xmath359 the invariance of the @xmath360-dimensional haar measure and in @xmath361 the fact that @xmath362 and @xmath363 have the same distribution .
we conclude from the above result that although the successive states of the small system @xmath187 are random density matrices that can be correlated in a very general way , their joint probability distribution is invariant by independent unitary basis changes .
in other words , the correlations manifest only at the level of the spectrum , the matrices being independently rotated by random haar unitaries .
the ergodic convergence result in such a case is established in the following proposition .
let @xmath364 be a sequence of random density matrices ( we make no assumption whatsoever on their distribution ) and @xmath342 a sequence of i.i.d .
haar unitaries independent of @xmath364
. then , almost surely , @xmath365 since both sides of the previous equation are self - adjoint matrices , it suffices to show that for any self - adjoint operator @xmath158 we have @xmath366 = \operatorname{tr}[a]/d$ ] . using the invariance of the haar measure
, one can assume that the observable @xmath89 is diagonal @xmath367 in some fixed orthonormal basis @xmath77 of @xmath260 .
in the same basis , we write @xmath368 and @xmath369 . to simplify notation , we put @xmath370 = \frac{t_1 + \cdots + t_n}{n},\ ] ] where @xmath371 = \sum_{i_1,i_2,j=1}^d t^{(k)}_{i_1,i_2 } s_j v_{i_1,j}^{(k ) } \overline{v_{i_2,j}^{(k)}}$ ] .
using the fact that @xmath372 = \delta_{i , i ' } \delta_{j , j ' } \frac 1 d,\ ] ] one can easily check that that the random variables @xmath373 have mean @xmath374/d$ ] , finite variance ( a rough bound for @xmath375 $ ] is @xmath374 ^ 2 $ ] ) and that they are not correlated ( @xmath376 , if @xmath377 ) .
it is a classical result in probability theory that in this case the ( strong ) law of large numbers holds and thus , almost surely , @xmath378 = \frac{\operatorname{tr}[a]}{d}.\ ] ] putting the previous proposition and lemma [ lem : u - iid - same - law ] together , one obtains the main result of this section , an ergodic - mean convergence result for the sequence of states of the `` small '' system . | we consider a generalized model of repeated quantum interactions , where a system @xmath0 is interacting in a random way with a sequence of independent quantum systems @xmath1 .
two types of randomness are studied in detail .
one is provided by considering haar - distributed unitaries to describe each interaction between @xmath0 and @xmath2 .
the other involves random quantum states describing each copy @xmath2 . in the limit of a large number of interactions , we present convergence results for the asymptotic state of @xmath0 .
this is achieved by studying spectral properties of ( random ) quantum channels which guarantee the existence of unique invariant states .
finally this allows to introduce a new physically motivated ensemble of random density matrices called the _
asymptotic induced ensemble_. |
layered transition metal dichalcogenides tx@xmath1 ( t = transition metal , x = chalcogen ) are among the most studied two - dimensional electronic systems .
the charge - density - wave ( cdw ) instability and its coexistence / competition with superconductivity are central characteristics of this family.@xcite on the other hand , due to the weak van der waals interaction between x - t - x sandwich layers , a large variety of atoms and molecules can be intercalated into the interlayer vacant sites , and dramatically change the physical properties.@xcite among various intercalates , the 3__d__-transition metal intercalated 2@xmath3-type m@xmath4tx@xmath1 are particularly interesting ( m = v , cr ,
mn , fe , co , ni ; 2@xmath3 means _ two _ x - t - x sandwiches in one unit cell , and the compound has a _ hexagonal _ symmetry ) . at 1/4 or 1/3
doping , the intercalated 3__d _ _ ions form an ordered two dimensional magnetic array , and show either ferromagnetic ( fm ) or antiferromagnetic ( afm ) order at low _
t _ , depending on the intercalated species and the host compound.@xcite transport measurements commonly indicate metallic conductivity in the paramagnetic ( pm ) state and a more rapid decrease in resistivity in the spin ordered state .
the easy axis of the magnetization can be either parallel or perpendicular to the hexagonal layers for different fm intercalates.@xcite the types of magnetic ordering encountered provide an interesting venue for exploring spin - related effects , for example , the anomalous hall effect@xcite or the magneto - optical effect , topics of renewed and increased interest due to the recent development of intrinsic mechanism arising from the berry phase of the bloch state.@xcite ( color online ) the unit cell of 2@xmath3-cr@xmath0nbse@xmath1 : ( a ) a 3d view ; ( b ) a ( 010 ) projection : cr atoms occupy one third of the octahedral holes in the van der waals gap ; ( c ) a ( 001 ) projection ; ( d ) a comparison of the nbse@xmath1 unit cell ( red ) and the cr@xmath0nbse@xmath1 @xmath5 superlattice ( grey).,width=240 ] the _ t_-dependent band structure , especially its modification across the fm transition , provides essential information in understanding ferromagnetism . in the itinerant stoner model,@xcite
the energy gain in the fm state arises from the exchange interaction , and the collapse of exchange splitting leads to a vanishing magnetic moment above the curie _
t_. while in the localized model,@xcite ferromagnetism originates from the long - range ordered local moments , which are orientation disordered in the pm state .
however , practical ferromagnets usually belong to the intermediate regime between the above two extremes , requiring a unified picture which covers the whole range from itinerant to localized ferromagnetism.@xcite in this work , we present an optical spectroscopy study and first - principles calculations on cr@xmath0nbse@xmath1 to elucidate the electronic structure change across the fm transition . in comparison with pure 2@xmath3-nbse@xmath1 , two new interband transitions in the mid- and near - infrared region emerge in cr@xmath0nbse@xmath1 , which split into four in the fm state .
band structure calculation indicates a strong hybridization between cr 3__d _ _ and nb 4__d _ _ bands near e@xmath6 , and the spin up cr e@xmath2 shows a double - peak character in the density of states ( dos ) map .
then the four optical peaks in the fm state are all interband transitions to cr e@xmath2(@xmath7 ) , and the band broadening in this cr - derived state is the cause for the peak splitting in the optical response .
both 2@xmath3-nbse@xmath1 and 2@xmath3-type cr@xmath0nbse@xmath1 single crystals were grown by the vapor transport method .
the structure of cr@xmath0nbse@xmath1 is shown in fig .
1 . the _ t_-dependent resistivity was obtained by the four contact technique in a quantum design ppms .
the near - normal incident reflectance spectra were measured by a bruker ifs 66v / s spectrometer in the frequency range from 40 to 25000 @xmath8 .
an _ in situ _ gold and aluminium overcoating technique was used to get the reflectivity r(@xmath9 ) .
the real part of conductivity @xmath10 is obtained by the kramers - kronig transformation of r(@xmath9 ) .
first - principles calculations based on the local density approximation ( lda ) and local spin density approximation ( lsda ) were performed to get the electronic structure of nbse@xmath1 and cr@xmath0nbse@xmath1 , respectively .
the calculations were done with our state code.@xcite we adopt the experimental structure parameters@xcite , and an initial se position z=0.125__c _ _ for both compounds , then relax the se ions to their lowest energy positions ( z@xmath11=0.132__c _ _ for 2@xmath3-nbse@xmath1 ; z@xmath11=0.131__c _ _ for cr@xmath0nbse@xmath1 ) . in previous calculation ,
z@xmath11 for 2@xmath3-nbse@xmath1 is around 0.118__c__@xcite .
the discrepancy arises from different choices for the atomic coordination . here
we set the nb position as ( 0,0,0 ) , not ( 0,0,1/4__c__).@xcite note z@xmath11=0.132__c _ _ in nb(0,0,0 ) is in fact equivalent to z@xmath11=0.118__c _ _ in nb(0,0,1/4__c _ _ ) coordination .
as shown in fig .
2 , a metallic behavior is found for cr@xmath0nbse@xmath1 from 300 to 10 k. the resistivity drops more rapidly with decreasing _ t _ below 115 k ,
consistent with the onset temperature of fm ordering.@xcite similar transport behavior is observed in other 3__d _ _ transition metal intercalates.@xcite the inset of fig .
2 focuses on the cdw and superconducting transitions for 2@xmath3-nbse@xmath1 .
t_-dependent resistivity for cr@xmath0nbse@xmath1 .
inset : an expanded plot for the cdw ( @xmath1232 k ) and superconducting transitions ( @xmath127.2 k ) in 2@xmath3-nbse@xmath1.,width=220 ] the r(@xmath9 ) for 2@xmath3-nbse@xmath1 ( fig .
3 ( a ) ) are the same as found in a previous study,@xcite except for a new feature in the mid - infrared region : r(@xmath9 ) at 10 k is suppressed slightly below that at 50 k from 600 to 5000 @xmath8 .
this is intrinsic , since the same result was repeatedly obtained , with each study carried out on a fresh , newly cleaved surface .
the midinfrared suppression implies the formation of a partial energy gap on the fermi surface in the cdw state , which closely resembles that of 2@xmath3-tas@xmath1@xcite , another cdw - bearing member with a higher _ t@xmath13 _ ( 75 k ) and thus a more apparent gap character .
r(@xmath9 ) is greatly modified after cr intercalation ( fig . 3 ( b ) ) . the well - defined plasma edge for 2@xmath3-nbse@xmath1 decays into a overdamped one in cr@xmath0nbse@xmath1
such an overdamped shape might arise from impurity scattering due to cr disorder , but not the scattering by cr moments , since the shape of the plasma edge is almost unchanged from fm to pm states .
the most interesting features are two additional peaks around 4300 and 8500 @xmath8 at 300 k , and their splitting below 115 k. no similar feature is found in the host compound .
( color online ) reflectivity of ( a ) 2@xmath3-nbse@xmath1 and ( b ) cr@xmath0nbse@xmath1 below 20000 @xmath8 at various temperatures .
the inset figures amplify the low frequency characters.,width=240 ] the low energy interband transitions in cr@xmath0nbse@xmath1 can be more clearly resolved in the real part of the conductivity @xmath14@xmath15(@xmath9 ) , as illustrated in fig .
4 ( a ) . beside the drude component , two peaks around 4300 and 8500 @xmath8 at 300 k split into four peaks near 2600 , 4700 , 8400 and 10500 @xmath8 in the fm state . in order to extract these features from the background , we fit @xmath14@xmath15(@xmath9 ) by the drude - lorentz model.@xcite in the inset of fig .
4 ( a ) , the fitting curves of 10 and 300 k match the original data well .
subtracting the drude components and the high energy interband transitions from @xmath14@xmath15(@xmath9 ) , we then get a clear picture of the absorption peaks and their splitting in fig .
4 ( b ) : the peak @xmath16 ( 4250 @xmath8 ) at 300 k splits into @xmath17 ( 2610 @xmath8 ) and @xmath18 ( 4690 @xmath8 ) at 10 k ; the peak @xmath19 ( 8840 @xmath8 ) splits into @xmath20 ( 8400 @xmath8 ) and @xmath21 ( 10520 @xmath8 ) .
the splitting energies @xmath22 and @xmath23 are almost the same , hence those peaks should have a similar origin . the apparent spectral difference for nbse@xmath1 and cr@xmath0nbse@xmath1 indicates cr intercalation is not a simple charge transfer process . detailed information about the band structure is required .
we note that the spectral difference ( i.e. the splitting of the interband transitions ) at 100 k and 130 k in the near - infrared region is obscure .
this is understandable as those features in r(@xmath9 ) are close to the reflectivity minimum and the overall peak strengths are rather weak . in addition , both t=100 k and 130k are very close to the fm transition temperature ( 115k ) , therefore fluctuation effect is also prone to influence the spectra near the critical temperature . on the other hand , in our overcoating technique for reflectance measurement ,
the sample is not in exchange gas but in vacuum ( better than 3@xmath24 torr ) , the measurement becomes relatively surface sensitive in such high energy , a more detailed inspection of temperature variations in @xmath10 for the fm phase , which requires to keep the sample at low temperature for a long time , becomes impractical .
however , the splitting indeed turns more apparent as the temperature goes down to 50 and 10 k from our current data , which confirms the connection between the splitting in the optical conductivity and the fm ordering .
( color online ) ( a ) @xmath10 for cr@xmath0nbse@xmath1 below 12000 @xmath8 .
the inset illustrates the experimental data and the drude - lorentz fitting for 10 and 300 k below 20000 @xmath8 .
the raw data and the fitting for 300 k are shifted down by 1000 @xmath25@xmath8 for clarity .
( b ) the first four lorentz terms extracted from the fitting.,width=259 ]
conventionally , the band structure modification after intercalation is understood by the rigid - band model : the host band structure is unchanged upon intercalation , meanwhile , electrons from the guest species fill the host conduction band , thus raise the chemical potential
. however , our optical data apparently violate the rigid - band model for the emergence of new interband transitions in cr doped nbse@xmath1 and their further splitting at low _ t_. to understand these low energy interband transitions , details about the band structure for both the pure and the cr intercalated compounds are required .
we obtain the ground state electronic structure for nbse@xmath1 and cr@xmath0nbse@xmath1 by first - principles calculations .
the projected density of states ( pdos ) for 2@xmath3-nbse@xmath1 is shown in fig . 5 ( a ) and ( b ) .
the crystal structure of 2@xmath3-nbse@xmath1 is the same with that of cr@xmath0nbse@xmath1 if cr atoms were removed .
each nb is in trigonal prismatic coordination with six se neighbors , so the nb 4__d _ _ states are split into : a@xmath26 ( _ _ d__@xmath27 ) state , e@xmath2 doublet ( _ _ d__@xmath28 ; _ _ d__@xmath29 ) , and e@xmath2 doublet ( _ _ d__@xmath30 ; _ _ d__@xmath31 ) . far below the fermi energy , the total density of states are of se 4__p _ _ character .
the tail of this se band hybridizes with nb ( a@xmath26 , e@xmath2 ) complex and extends above e@xmath6 .
a high density of states at the fermi energy consists with the optical observation of a well - defined plasma edge in 2@xmath3-nbse@xmath1 .
we note that the partially filled nb conduction band is frequently referred to as `` _
_ d__@xmath27 '' band in the early studies .
however , recent calculation on 2@xmath3-nbse@xmath1@xcite show that the half - filled conduction band has _ _ d__@xmath27 symmetry around the @xmath32 point of the brillouin zone , and ( _ _ d__@xmath29 , _ _
d__@xmath28 ) symmetry around the @xmath33 point .
similar result is reproduced in fig . 5 ( a ) , that both a@xmath26 ( _ _ d__@xmath27 ) and e@xmath2 ( _ _ d__@xmath29 , d@xmath28 ) contribute to the pdos of the half - filled conduction band .
figure 5 ( c ) to ( e ) shows the pdos for cr@xmath0nbse@xmath1 in the fm ground state .
since cr is in a trigonally distorted crse@xmath34 octahedron,@xcite the cr 3__d _ _ bands are split into t@xmath35 ( which is further split into a@xmath26 and e@xmath2 ) and e@xmath2 states . in the spin up channel ,
the cr e@xmath2 state is localized and fully occupied ; while the cr a@xmath26 state and e@xmath2 doublet have weak itinerant character , and strongly hybridize with nb conduction band in the range [email protected] ev across e@xmath6 .
all the spin down cr 3__d _ _ bands are unoccupied . from fig .
5 ( g ) and ( h ) , one will find that cr@xmath0nbse@xmath1 is metallic with a fermi level crossing in both spin directions . comparing with 2@xmath3-nbse@xmath1 ,
the half - filled nb conduction band is now almost filled up in cr@xmath0nbse@xmath1 , with a large removal of dos at the fermi energy , leading to a marked reduction in the free carrier spectral weight .
although there is a strong hybridization between cr t@xmath35 ( @xmath7 ) and e@xmath2 ( @xmath7 ) states near e@xmath6 , the occupied nb states in the spin down projection is almost unaffected after intercalation , besides a shift of the chemical potential .
in comparison with the band structure calculations , we can identify the origin of the experimental interband transitions ( associated with the joint density of states ) at low _
t_. for nbse@xmath1 , the close - to - e@xmath6 interband transition from occupied se to unoccupied nb states is the origin of the first peak around 12000 @xmath8 in fig .
3 ( a ) . while the second peak at 20000 @xmath8 is the nb
_ d _ to _ d _ hopping from e@xmath6 to the lowest empty state .
such _ d_-_d _ transition is not forbidden with the help of se 4__p _ _ hybridization . for cr@xmath0nbse@xmath1 ,
all possible low - energy ( @xmath361.5 ev ) interband transitions in the fm ground state include cr - to - cr , nb - to - nb , and cr - to - nb ( also nb - to - cr ) transitions .
note the pdos scales for fig .
5 ( c)-(e ) are different that cr has apparently larger density of states than that of nb or se , so cr - cr and cr - nb transitions should dominate the low - energy interband optical peaks . as shown in fig .
5 ( g ) , cr e@xmath2(@xmath7 ) bands have rather special dispersion : they are flat near the band bottom and top but dispersive elsewhere ( which can be better resolved along a - l - h - a ) , so it has a double - peak pdos .
then the two peaks , @xmath17 and @xmath18 , in @xmath10 mainly come from the spin up cr a@xmath26@xmath37cr e@xmath2 transition , and the @xmath20 ( 1.04 ev ) and @xmath21 ( 1.30 ev ) peaks are mainly contributed by the occupied spin up nb ( a@xmath26 , e@xmath2)@xmath37cr e@xmath2 transition .
other cr - cr or cr - nb transitions exceed the energy scope in fig .
4 ( b ) , while the low energy se - nb transition in nbse@xmath1 shifts to higher energies because cr intercalation raises the chemical potential .
the predicted interband contribution in @xmath10 from the lsda calculation are shown in fig . 5 ( f ) .
the experimental data with the removal of the drude component are also presented for comparison . in general
, the theoretical result qualitatively reproduces the experimental observation of a four - peak character in the low energy optical response . since the curie - weiss pm state is a spin - disorder system , that the orientation and amplitude of neighboring spins would fluctuate with temperature
, one can not calculate the pm band from first - principles .
however , the band modification across the fm transition can be speculated from the changes in optical interband transitions . as cr e@xmath2(@xmath7 ) is the common final state for the four interband transitions in the fm ground state , then the two - to - four splitting of the optical peaks should be caused by a splitting , or more generally , a band broadening , in this common e@xmath2 state .
provided cr e@xmath2(@xmath7 ) turns narrower in the pm state , then the interband transitions @xmath38 and @xmath39 are indistinguishable .
the band width for a correlated electronic system is related to the hopping integral between neighboring sites , then the long - range ordering of cr local moment favors a broader bandwidth from pm to fm state .
the spin up cr a@xmath26 and e@xmath2 have partial itinerant character with a pdos `` tail '' at e@xmath6 ( fig . 5 ( c ) ) .
therefore the lowering ( i.e. gain ) of the kinetic energy of carriers in the fm transition would lead to the conduction band broadening.@xcite here the phase transition does not apparently change the drude response in cr@xmath0nbse@xmath1 , indicating the conducting carriers are mainly of nb 4__d _ _ character , thus are less affected by possible effective mass reduction when the cr - derived bands turn broadener in the fm phase . as shown above , the interband peak splitting is a band broadening effect in cr e@xmath2 doublet , now further discussions on other possibilities are necessary .
one possible cause for the band modification is a structure distortion , which splits the degenerated e@xmath2 into two separated states .
here we should emphasize that the relaxed se position in our calculation indicates a trigonal distorted crse@xmath34 octahedra ( along the c - axis ) in the fm ground state , but it apparently does not change the degeneracy of the e@xmath2 doublet ( d@xmath31 ; d@xmath30 ) .
in fact , the double - peak pdos for cr e@xmath2(@xmath7 ) is not two separated states , but due to special band dispersion , so the splitting of the optical peaks is not caused by possible structural distortion .
another possible case is the exchange splitting .
regardless of the detailed band structure for this specific material , the splitting in the interband transition for a ferromagnet is generally attributed to the exchange splitting : when magnetism is on the itinerant side , then the pm phase will have some set of absorption peaks , and in the fm phase each of them will split into two , reflecting the fact that the exchange splitting is not rigid . however , our band structure calculations rule out such a possibility . as shown in fig
5 ( c ) , although the spin up cr a@xmath26 and e@xmath2 have a weak itinerant character , the cr 3__d _ _ bands are mainly localized , therefore the dominating magnetic mechanism for cr@xmath0nbse@xmath1 is away from the itinerant side . in the curie - weiss pm state ,
the cr 3__d _ _ local moments do not vanish . then the only close - to - e@xmath6 band which might experience spin polarization from the pm to fm state is the nb 4__d _ _ conduction band ( fig .
5 ( d ) ) .
however , one can not find any nb - related interband transitions which satisfy : 1 ) a [email protected] ev and [email protected] ev gap for both spin projections ; and 2 ) a 0.25 ev energy difference for the common hopping in different spin channels ( as a 2000 @xmath8 splitting from optical measurements ) .
in fact , the magnetic moment for nb atom is less than 0.03@xmath40 ( while cr moment is 2.6@xmath40 ) from our calculation , hence the nb spin polarization is so weak that it can be neglected , so the splitting of the interband transitions is irrelevant to the exchange splitting
. here some complementary discussion on the pm band structure should also be added . in the curie - weiss
paramagnetic state , there is no spin polarization for the energy band , then the optical response from different spin channels have equal contribution to @xmath10 .
as shown above , we conclude that the `` spin up '' cr e@xmath2 turns narrower in the pm state . here the spin label was used to better describe this specific band when comparing with the fm case .
one should keep in mind that the two spin channels are mixed in the pm state , reflecting the fluctuation of the local spins .
however , such a band mixing will not bring new low - energy interband transitions .
note @xmath38 and @xmath39 ( i=1,2 ) are all come from cr - related bands , which are mainly localized and has a large splitting in different spin projections in the fm ground state ( fig . 5 ( c ) ) ;
furthermore , possible spin polarization in the fm state for the nb and se related bands are rather weak , then the band mixing in the pm state will not bring new peaks for the low frequency @xmath10 .
finally , it should be noted that the qualitative electronic structure around fermi level for cr@xmath0nbse@xmath1 is not sensitive to the on - site coulomb repulsion @xmath41 of cr site due to the strong cr - nb hybridization .
we have tried a lsda+@xmath41 ( @xmath41=3.0 ev for cr site ) calculation , the low energy optical conductivity is not qualitatively changed ( except the @xmath17 and @xmath18 peaks are weaker compared to lsda ) .
we report a combined optical spectroscopy study and first - principles calculations on 2@xmath3-type nbse@xmath1 and the ferromagnetic intercalated compound cr@xmath0nbse@xmath1 .
a weak mid - infrared suppression in r(@xmath9 ) of 2@xmath3-nbse@xmath1 testifies the existence of a cdw induced partial gap on the fermi surface . for cr@xmath0nbse@xmath1 , the appearance of new interband transitions and their remarkable _
evolution are direct experimental evidence for the invalidity of the rigid - band model .
the multi - peak feature in @xmath42(@xmath9 ) has been clearly extracted by a drude - lorentz fitting . based on the lsda calculation
, we found that the four optical peaks in the fm phase are interband transitions involving both the intercalated cr 3__d _ _ and the host nb 4__d _ _ states .
considering the temperature evolution for these optical interband transitions , we further conclude that their splitting from pm to fm phase is not an exchange splitting effect , but a dispersion modification ( broadening ) in the spin up cr @xmath43 doublet , which have partial itinerant character and strongly hybridize with nb 4__d _ _ states .
such an evident band broadening in a rather close - to - e@xmath6 3__d _ _ state requires further experimental investigations , which might provide new insight into understanding the ferromagnetism induced band modification across the phase transition .
this work is supported by the national science foundation of china , the knowledge innovation project of the chinese academy of sciences , and the 973 project of the ministry of science and technology of china .
the work at brookhaven national laboratory is operated for the u.s .
department of energy by brookhaven science associates ( no .
de - ac02 - 98ch10886 ) , and at princeton by the national science foundation of the usa .
there are two types of nb sites ( fig .
1 ) : nb1 has no direct cr neighbor along the c - axis , while nb2 has one . besides a weaker hybridization with cr 3__d _ _ states
, the nb2 pdos shows no intrinsic difference with that of nb1 , so we only plot the nb1 case in fig . 5 ( d ) .
note the nb2 to cr hopping does not contribute to the in - plane @xmath10 . | the electronic structure of cr@xmath0nbse@xmath1 is studied via optical spectroscopy .
we observe two low - energy interband transitions in the paramagnetic phase , which split into four peaks as the compound enters the ferromagnetic state .
the band structure calculation indicates the four peaks are interband transitions to the spin up cr e@xmath2 states .
we show that the peak splitting below the curie temperature is _ not _ due to the exchange splitting of spin up and down bands , but directly reflects a band broadening effect in cr - derived states upon the spontaneous ferromagnetic ordering . |
the hudson - parthasarathy quantum stochastic calculus @xcite provides a unified framework for modelling the dynamics of open quantum systems interacting with environment .
the governing quantum stochastic differential equations ( qsdes ) are driven by a quantum counterpart of the classical wiener process which acts on a boson fock space and models a heat bath of external fields .
an important feature of the qsdes is that their drift and dispersion terms are specified by the hamiltonian and coupling operators , which describe the energetics of the system and its interaction with the environment , and also involve a scattering matrix representing the photon exchange between the fields @xcite .
these energy operators are usually functions ( for example , polynomials or weyl quantization integrals @xcite ) of the system variables , and the form of this dependence affects tractability of the resulting quantum system .
in particular , quadratic hamiltonians and linear system - field coupling operators lead to a class of open quantum harmonic oscillators ( oqhos ) @xcite which play an important role in the linear quantum control theory @xcite .
the linear dynamics of such systems are relatively well understood and are similar to the classical linear sdes in some regards , such as the preservation of gaussian nature of quantum states @xcite driven by vacuum fields .
nevertheless , the coherent ( that is , measurement - free ) quantum analogue @xcite of the classical lqg control problem @xcite for oqhos is complicated by the physical realizability ( pr ) constraints on the state - space matrices of the fully quantum controller which come from the special structure of qsdes mentioned above .
similar constraints are present in the coherent quantum filtering ( cqf ) problem @xcite .
one of the existing approaches to the coherent quantum lqg ( cqlqg ) control and filtering problems employs the frechet derivatives of the lqg cost ( with respect to the state - space matrices subject to the pr constraints ) for obtaining the optimality conditions @xcite and for the numerical optimization @xcite .
this approach takes into account the quantum nature of the underlying problem only through the pr constraints , being `` classical '' in all other respects , which has its own advantages from the viewpoint of well - developed optimization methods .
however , a disadvantage of this approach lies in the fact that it is limited to certain parametric classes of linear controllers and filters . in particular
, the resulting optimality conditions do not provide information on whether nonlinear quantum controllers or filters can outperform the linear ones for linear quantum plants . in the present paper ,
we outline a fully quantum variational technique which allows the sensitivity of the dynamic and output variables of a nonlinear quantum stochastic system to be investigated with respect to arbitrary ( that is , not only linear - quadratic ) perturbations of the hamiltonian and coupling operators .
this approach is based on using a transverse hamiltonian , an auxiliary time - varying operator which encodes the propagation of such perturbations through the unitary system - field evolution .
this leads to an infinitesimal perturbation formula for quantum averaged performance criteria ( such as the lqg cost functional ) which is applicable to the development of optimality conditions in coherent quantum control problems for larger classes of controllers and filters .
in particular , this approach allows the sensitivity of oqhos with quadratic performance criteria to be studied with respect to general perturbations of the energy operators .
in fact , the transverse hamiltonian method has already been employed in @xcite to establish local sufficiency of linear filters in the mean square optimal cqf problem @xcite for linear plants with respect to varying the energy operators of the filter along weyl operators @xcite .
note that our approach is different from @xcite which develops a quantum hamilton - jacobi - bellman principle for a measurement - based quantum feedback control problem , where the role of state variables is played by the density operator .
it is also relevant to mention a parallel between the presented perturbation analysis and the fluctuation - dissipation theorem @xcite .
the paper is organised as follows .
section [ sec : not ] outlines notation used in the paper .
section [ sec : qsde ] specifies the class of open quantum systems in hamiltonian formulation .
section [ sec : oqho ] discusses sensitivity of oqhos to parametric perturbations within quadratic hamiltonians and linear system - field coupling operators .
section [ sec : th ] introduces the transverse hamiltonian .
the latter is used in section [ sec : sys ] for perturbation analysis of system operators .
section [ sec : func ] extends the transverse hamiltonian technique to quantum averaged performance criteria .
section [ sec : cqf ] applies this approach to a mean square optimal cqf problem .
section [ sec : conc ] makes concluding remarks .
in what follows , @xmath0:= ab - ba$ ] denotes the commutator of linear operators @xmath1 and @xmath2 on a common space . as a linear superoperator , the commutator with a fixed operator @xmath1 ,
is denoted by @xmath3 $ ] .
this extends to the commutator @xmath4-matrix @xmath5 : = xy^{\rt } - ( yx^{\rt})^{\rt } = ( [ x_j , y_k])_{1\ < j\ < n,1\ < k\ < m } $ ] for a vector @xmath6 of operators @xmath7 and a vector @xmath8 of operators @xmath9
. vectors are organized as columns unless indicated otherwise , and the transpose @xmath10 acts on matrices of operators as if their entries were scalars . in application to such matrices , @xmath11 denotes the transpose of the entry - wise operator adjoint @xmath12 . for complex matrices
, @xmath13 is the usual complex conjugate transpose @xmath14 .
the subspaces of real symmetric , real antisymmetric and complex hermitian matrices of order @xmath15 are denoted by @xmath16 , @xmath17 and @xmath18 , respectively , where @xmath19 is the imaginary unit .
the real and imaginary parts of a complex matrix are denoted by @xmath20 and @xmath21 .
these extend to matrices @xmath22 with operator - valued entries as @xmath23 and @xmath24 which consist of self - adjoint operators .
positive ( semi- ) definiteness of matrices and the corresponding partial ordering are denoted by ( @xmath25 ) @xmath26 .
also , @xmath27 and @xmath28 denote the sets of positive semi - definite real symmetric and complex hermitian matrices of order @xmath15 , respectively .
the tensor product of spaces or operators ( in particular , the kronecker product of matrices ) is denoted by @xmath29 .
the tensor product @xmath30 of operators @xmath1 and @xmath2 acting on different spaces will sometimes be abbreviated as @xmath31 .
the identity matrix of order @xmath15 is denoted by @xmath32 , while the identity operator on a space @xmath33 is denoted by @xmath34 .
the frobenius inner product of real or complex matrices is denoted by @xmath35 .
the expectation @xmath36 of a quantum variable @xmath37 over a density operator @xmath38 extends entrywise to vectors and matrices of such variables .
consider an open quantum system which interacts with a heat bath of @xmath39 external boson fields and is endowed with @xmath15 dynamic variables @xmath40 evolving in time @xmath41 .
the system variables are self - adjoint operators on a composite system - field hilbert space @xmath42 , where @xmath43 is the initial space of the system which provides a domain for @xmath44 , and @xmath45 is a boson fock space @xcite for the action of quantum wiener processes @xmath46 .
the latter are self - adjoint operators which model the external boson fields .
the energetics of the system - field interaction is specified by the system hamiltonian @xmath47 and the system - field coupling operators @xmath48 which are self - adjoint operators , representable as time invariant functions ( for example , polynomials with constant coefficients ) of the system variables @xmath40 . both @xmath49 and @xmath50 act on the initial space @xmath43 . for what follows ,
the system and field variables , and the coupling operators are assembled into vectors : @xmath51 depending on the context , a system operator @xmath52 , acting on the initial space @xmath43 , will sometimes be identified with its extension @xmath53 to the composite system - field space @xmath42 .
the system and the fields form an isolated entity which undergoes a unitary evolution @xmath54 on @xmath42 driven by the system - field interaction according to the following qsde @xcite with initial condition @xmath55 : @xmath56 which corresponds to the schrdinger picture of quantum dynamics . here and in what follows , the initial values of time - varying operators ( or vectors or matrices of operators ) at @xmath57 are marked by subscript @xmath58 ( so that @xmath59 , @xmath60 and @xmath61 ) , while the time dependence will often be omitted for brevity .
the qsde ( [ du ] ) corresponds to a particular yet important case of the identity scattering matrix , when the photon exchange between the fields and the gauge processes @xcite can be eliminated from consideration . here , @xmath62 is the quantum ito matrix of the wiener process @xmath63 : @xmath64 where the dimension @xmath39 is assumed to be even , and the matrix @xmath65 specifies the cross - commutations between the quantum wiener processes @xmath66 : @xmath67 = 2ij\rd t , \qquad j : = \bj\ox i_{m/2 } , \qquad \bj : = { \begin{bmatrix } 0 & 1\\ -1 & 0 \end{bmatrix}}.\ ] ] accordingly , the term @xmath68 in ( [ du ] ) can be interpreted as an incremental hamiltonian of the system - field interaction , while @xmath69 is associated with the ito correction term in the differential relation @xmath70 describing the preservation of co - isometry of @xmath54 . the system variables at time @xmath71 , as operators on the system - field space @xmath42 , are the images @xmath72 of their initial values under the flow @xmath73 which maps a system operator @xmath74 on the initial space @xmath43 to @xmath75 note that , in view of ( [ du ] ) , the flow @xmath73 depends on the energy operators @xmath76 and @xmath77 .
the resulting quantum adapted process @xmath52 satisfies the following hudson - parthasarathy qsde @xcite : @xmath78\rd w , \qquad \cg(\sigma ) : = i[h,\sigma ] + \cd(\sigma),\ ] ] where use is made of the hamiltonian and the coupling operators evolved by the flow @xmath73 from ( [ flow ] ) as @xmath79 ( the flow acts entrywise on vectors and matrices of operators ) .
also , @xmath80 in ( [ dsigma ] ) denotes the gorini - kossakowski - sudarshan - lindblad ( gksl ) decoherence superoperator @xcite which acts on @xmath81 as @xmath82 + [ l_j,\sigma]l_k ) \\
\nonumber & = \frac{1}{2 } \big ( l^{\rt}\omega [ \sigma , l ]
+ [ l^{\rt},\sigma]\omega l \big)\\ \label{cd } & = -[\sigma , l^{\rt}]\omega l - \frac{1}{2 } [ l^{\rt}\omega l , \sigma].\end{aligned}\ ] ] the third equality in ( [ cd ] ) is convenient for the entrywise application of the superoperator @xmath80 to vectors of operators . the superoperator @xmath83 in ( [ dsigma ] ) is referred to as the gksl generator .
the identity @xmath84 , which holds for system operators in view of ( [ flow ] ) and the unitarity of @xmath54 , allows the qsde ( [ du ] ) to be represented in the heisenberg picture using ( [ hl ] ) : @xmath85 here , use is also made of the property that for any @xmath71 , the future - pointing increments @xmath86 commute with adapted processes evaluated at the same ( or an earlier ) moment of time . in application to the vector @xmath87 of system variables ,
the flow @xmath88 is understood ( also entrywise as mentioned before ) in accordance with ( [ xu ] ) and ( [ flow ] ) , and the corresponding qsde ( [ dsigma ] ) is representable in a vector - matrix form as @xmath89,\ ] ] where the @xmath15-dimensional drift vector @xmath90 and the dispersion @xmath4-matrix @xmath91 consist of time - varying operators .
furthermore , the interaction of the input field with the system produces an @xmath39-dimensional output field @xmath92 where the system - field unitary evolution is applied to the current input field variables ( which reflects the innovation nature of the quantum wiener process and the continuous tensor product structure of the fock space @xcite ) .
the output field satisfies the qsde @xmath93 where the matrix @xmath94 is given by ( [ j ] ) , and @xmath95 is the vector of system - field coupling operators from ( [ xwl ] ) .
the common unitary evolution in ( [ x ] ) and ( [ y ] ) preserves the commutativity between the system variables and the output field variables in time @xmath41 : @xmath96 = u(t)^{\dagger}[x_0 \ox \ci_{\cf},\ci_{\ch_0}\ox w(t)^{\rt}]u(t ) = 0,\ ] ] where the entries of @xmath97 and @xmath98 commute as operators on different spaces . by a similar reasoning ,
the output field @xmath8 inherits the ccr matrix @xmath94 from the input quantum wiener process @xmath63 : @xmath99 & = u(t)^{\dagger}[\ci_{\ch}\ox w(t ) , \ci_{\ch}\ox w(t)^{\rt } ] u(t)\\ & = u(t)^{\dagger}(\ci_{\ch } \ox[w(t ) , w(t)^{\rt } ] ) u(t)\\ % \nonumber%\label{yy } & = 2it j u(t)^{\dagger}\ci_{\ch\ox \cf } u(t ) = 2itj.\end{aligned}\ ] ] this relation can also be established as a corollary of the property that @xmath100 in view of ( [ omega ] ) , ( [ j ] ) and ( [ dy ] ) .
in a particular case , when the system being considered is an oqho @xcite , its dynamic variables satisfy the heisenberg ccrs @xmath101 = 2i \theta\ ] ] on a dense domain , where the ccr matrix @xmath102 represents @xmath103 and remains unchanged .
the hamiltonian of the oqho is a quadratic polynomial and the system - field coupling operators in ( [ xwl ] ) are linear functions of the system variables : @xmath104 where @xmath105 and @xmath106 are constant matrices . in this case ,
( [ dx ] ) and ( [ dy ] ) become linear qsdes @xmath107 ( with linear drift @xmath108 and dispersion matrix @xmath109 with real - valued entries ) with constant coefficients which comprise the matrices @xmath110 , @xmath111 and @xmath112 : @xmath113 these matrices satisfy the following algebraic equations @xmath114 which describe the fulfillment of the ccrs ( [ theta ] ) and the commutativity ( [ xy ] ) at any moment of time and provide necessary and sufficient conditions of physical realizability @xcite of the linear qsdes ( [ dxdy ] ) as an oqho with the ccr matrix @xmath115 . the first equality in ( [ pr ] ) can be regarded as an algebraic lyapunov equation ( ale ) with respect to @xmath115 which has a unique solution if and only if the kronecker sum @xmath116 is a nonsingular matrix ( that is , @xmath1 has no two eigenvalues which are symmetric to each other about the origin of the complex plane ) . for general open quantum systems ( whose dynamic variables satisfy ccrs , while the hamiltonian and the coupling operators are not necessarily quadratic and linear functions of the system variables , respectively ) , the ccr preservation follows from the unitary evolution of the system variables in ( [ x ] ) .
using the standard notation of classical linear control theory @xcite , the quadruple of the state - space realization matrices of the qsdes ( [ dxdy ] ) , which specify the input - system - output map from @xmath63 to @xmath8 for the oqho , can be written as @xmath117}.\ ] ] if the matrices @xmath118 and @xmath119 , which specify the hamiltonian and the coupling operators in ( [ rn ] ) , depend smoothly on a scalar parameter @xmath120 ( while @xmath63 is independent of @xmath120 ) , then so do the matrices @xmath1 , @xmath2 , @xmath121 in ( [ abc ] ) and the system and output variables which comprise the vectors @xmath87 and @xmath122 . the derivatives @xmath123 and @xmath124 can be interpreted for any @xmath41 as adapted processes with self - adjoint operator - valued entries satisfying the qsdes @xmath125 ( the second of which is , in fact , an ode ) with zero initial conditions @xmath126 and @xmath127 since @xmath97 and @xmath128 do not depend on @xmath120 . here ,
@xmath129 are the derivatives of the matrices @xmath1 , @xmath2 , @xmath121 from ( [ abc ] ) with respect to @xmath120 . in combination with ( [ dxdy ] )
, the qsdes ( [ dxdy ] ) allow the input - system - output map from @xmath63 to @xmath130 to be represented in a form , similar to ( [ abci ] ) , as @xmath131 } % : = { \left [ \begin{array}{cc||c } a & 0 & b\\ a ' & a & b'\\ \hline \hline c & 0 & i_m\\ c ' & c & 0 \end{array } \right]}.\ ] ] the special block lower triangular structure of the state dynamics matrix in ( [ aaa ] ) is closely related to the gateaux derivative @xmath132 of the matrix exponential in the direction @xmath133 ; see , for example , @xcite . at any time @xmath71 ,
both @xmath134 and @xmath135 depend linearly on the matrices @xmath136 and @xmath137 from ( [ abc ] ) in view of the following representation of the solutions of the qsdes ( [ dxdy ] ) which regards @xmath138 as a forcing term : @xmath139 where @xmath140 is the unperturbed solution of the first qsde in ( [ dxdy ] ) which does not depend on @xmath136 or @xmath137 . in the case of linear qsdes ,
whose coefficients depend smoothly on parameters , the mean square differentiability of their solutions with respect to the parameters can be verified directly by using the closed form ( [ xlin ] ) ( under certain integrability conditions for the underlying quantum state ) . the relations ( [ abc])([y ] ) exhaust the analysis of sensitivity of the oqho to the perturbations of the matrices @xmath118 and @xmath119 under which the energy operators @xmath33 and @xmath95 ( [ rn ] ) remain in the corresponding classes of quadratic and linear functions of the system variables .
however , this analysis is not applicable to more general perturbations ( for example , higher order polynomials of the system variables ) and is restricted to linear systems .
we will now consider the response of the general open quantum system , described in section [ sec : qsde ] and governed by ( [ dx ] ) and ( [ dy ] ) , to perturbation of the system hamiltonian and the system - field coupling operators : @xmath141 here , @xmath142 and the entries of the @xmath39-dimensional vector @xmath143 are self - adjoint operators on the initial system space @xmath43 , and @xmath120 is a small real - valued parameter , so that @xmath144 and @xmath145 . in
what follows , the derivative @xmath146 will be taken at @xmath147 .
the perturbations @xmath142 and @xmath143 in ( [ km ] ) are also assumed to be functions of the system variables which are evolved by the _
unperturbed _ flow ( [ flow ] ) as @xmath148 for example , in the case of the oqho considered in section [ sec : oqho ] , the perturbations @xmath149 and @xmath22 inherit the structure of the hamiltonian and the coupling operators in ( [ rn ] ) as a quadratic function and a linear function of the system variables , respectively : @xmath150 returning to the general case , we will avoid at this stage technical assumptions on @xmath142 and @xmath143 in ( [ km ] ) , so that the calculations , carried out below for arbitrary perturbations , should be regarded as formal . since the energy operators @xmath76 and @xmath77 completely specify the dynamics of the unitary operator @xmath54 which determines the evolution of the system and output field variables , the response of the latter to the perturbation of @xmath76 and @xmath77 reduces to that of @xmath54 .
therefore , the propagation of the initial perturbations of the energy operators through the subsequent unitary system - field evolution can be described in terms of the operator @xmath151 which satisfies @xmath152 since @xmath153 does not depend on @xmath120 .
the smoothness of dependence of @xmath54 on the parameter @xmath120 is analogous to the corresponding property of solutions of classical sdes ( whose drift and dispersion satisfy suitable regularity conditions @xcite ) and holds at least in the linear case discussed in section [ sec : oqho ] .
[ th : q ] for any time @xmath41 , the operator @xmath154 in ( [ v ] ) , associated with the unitary evolution @xmath54 from ( [ du ] ) , is representable as @xmath155 here , @xmath156 is a self - adjoint operator on @xmath42 which satisfies zero initial condition @xmath157 and is governed by the qsde @xmath158 furthermore , @xmath156 depends linearly on the initial perturbations @xmath142 and @xmath143 of the energy operators in ( [ km ] ) through their unperturbed evolutions in ( [ kmflow ] ) . the differentiation of both sides of the unitarity relation @xmath159 with respect to the parameter @xmath120 and letting @xmath147 leads to the equalities @xmath160 , the second of which implies self - adjointness of the operator @xmath161 thus establishing ( [ vuq ] ) .
now , consider the time evolution of @xmath156 . to this end ,
the differentiation of ( [ du ] ) with respect to @xmath120 yields @xmath162 by left multiplying both sides of ( [ dv ] ) by @xmath163 and recalling ( [ q ] ) , it follows that @xmath164 where use is made of the evolved perturbations ( [ kmflow ] ) . by a similar reasoning
, a combination of ( [ duhl ] ) with ( [ q ] ) leads to @xmath165 where use is also made of ( [ udv ] ) and the quantum ito product rules @xcite in combination with ( [ omega ] ) .
it now follows from ( [ q ] ) and ( [ udv])([dudv ] ) that @xmath166 which establishes ( [ dq ] ) .
the linear dependence of @xmath156 on @xmath142 and @xmath143 follows from the integral representation @xmath167 of the qsde ( [ dq ] ) and the property that the evolved perturbations @xmath149 and @xmath22 in ( [ kmflow ] ) depend linearly on @xmath142 and @xmath143 , respectively . in view of theorem
[ th : q ] , the equation @xmath168 has the structure of isolated quantum dynamics in fictitious time @xmath120 , where @xmath156 plays the role of a hamiltonian associated with the perturbation of the unitary operator @xmath54 . in order to reflect this property , the time - varying operator @xmath169 will be referred to as the _ transverse hamiltonian _ associated with the perturbations @xmath149 and @xmath22 of the energy operators .
its computation is illustrated below . in the absence of perturbation to the system - field coupling ,
when @xmath143 consist of zero operators , the transverse hamiltonian in ( [ qint ] ) reduces to @xmath170 .
moreover , if the system is isolated ( and hence , the unitary process @xmath171 in ( [ du ] ) takes the form @xmath172 ) , then @xmath173 here , use is made of hadamard s lemma @xcite , and @xmath174 is an entire function which plays a part in the solution of systems of nonhomogeneous linear odes with constant coefficients h_2008 .
@xmath175 for the oqho of section [ sec : oqho ] , substitution of the perturbations ( [ kmlin ] ) into ( [ qint ] ) leads to the transverse hamiltonian @xmath176 which depends in a quadratic fashion on the past history of the system variables .
the latter are given by the unperturbed equation ( [ xlin ] ) .
the last term @xmath177 in ( [ qlin ] ) comes from the relation @xmath178 ( following from ( [ omega ] ) ) and the identity @xmath179 which holds for any @xmath180 in view of the ccrs ( [ theta]).@xmath175
since the transverse hamiltonian @xmath156 encodes the propagation of the initial perturbations of the energy operators in ( [ km ] ) through the unitary system - field evolution over the time interval @xmath181 $ ] , it provides a tool for the infinitesimal perturbation analysis of general system operators .
the following theorem is concerned with an extended setting which , in addition to ( [ km ] ) , allows for an initial perturbation in a system operator @xmath74 specified by @xmath182 .
[ th : dashqsde ] for any self - adjoint system operator @xmath74 on the initial space @xmath43 , which smoothly depends on the same scalar parameter @xmath120 as in ( [ km ] ) and is evolved by the flow ( [ flow ] ) , the derivative of its evolved version @xmath81 with respect to @xmath120 is representable as @xmath183\ ] ] at any time @xmath41 , where @xmath156 is the transverse hamiltonian from theorem [ th : q ] . here
, the operator @xmath184 satisfies the qsde @xmath185 + \chi(\sigma))\rd t + \big([q,[\sigma , l^{\rt}]]-i[\sigma , m^{\rt}]\big)\rd w , % \big ( % i ( % [ k -\im ( l^{\rt}\omega m),\sigma ] % + [ q , \cg(\sigma ) ] % ) \\
% \nonumber % & -2\re ( [ \sigma , l^{\rt}]\omega m ) % \big)\rd t\\ % & + \big([q,[\sigma , l^{\rt}]]-i[\sigma , m^{\rt}]\big)\rd w.\ ] ] with zero initial condition @xmath186 , where @xmath83 is the gksl generator from ( [ dsigma ] ) , and @xmath187 is a linear superoperator given by @xmath188 -2\re ( [ \sigma , l^{\rt}]\omega m).\ ] ] by using the leibniz product rule together with ( [ flow ] ) , ( [ v ] ) , ( [ vuq ] ) and the self - adjointness of @xmath156 , it follows that @xmath189,\end{aligned}\ ] ] which establishes ( [ flow ] ) .
the equality @xmath190 follows from @xmath191 .
we will now obtain a qsde for the process @xmath192 .
to this end , by combining the quantum ito lemma @xcite with the bilinearity of the commutator , and using the qsdes ( [ dsigma ] ) and ( [ dq ] ) , it follows that @xmath193 & = [ \rd q , \sigma ] + [ q , \rd \sigma ] + [ \rd q , \rd \sigma]\\ \nonumber & = [ k -\im ( l^{\rt}\omega m),\sigma]\rd t -[\sigma , m^{\rt}]\rd w+ [ q , \cg(\sigma)]\rd t- i[q,[\sigma , l^{\rt}]]\rd w -[m^{\rt}\rd w , i[\sigma , l^{\rt}]\rd w]\\ \nonumber & = \big([k -\im ( l^{\rt}\omega m),\sigma ] + [ q , \cg(\sigma ) ] + 2i \im ( i[\sigma , l^{\rt}]\omega m ) \big)\rd t - \big([\sigma , m^{\rt } ] + i[q,[\sigma , l^{\rt}]]\big)\rd w\\ \label{dqsig } & = \big ( [ k -\im ( l^{\rt}\omega m),\sigma ] + [ q , \cg(\sigma ) ] + 2i \re ( [ \sigma , l^{\rt}]\omega m ) \big)\rd t - \big([\sigma , m^{\rt}]+i[q,[\sigma , l^{\rt}]]\big)\rd w.\end{aligned}\ ] ] here , the quantum ito product rules are combined with the commutativity between adapted processes and future - pointing increments of the quantum wiener process . in the second and third of the equalities ( [ dqsig ] ) ,
use is also made of the relations @xmath194 = [ \alpha,\beta^{\rt}]\rd w$ ] and @xmath195 = 2i\im(\alpha^{\rt } \omega \beta)\rd t$ ] for appropriately dimensioned adapted processes @xmath196 and @xmath197 with self - adjoint operator - valued entries .
these are combined with the identity @xmath198\omega m ) = \re ( [ \sigma , l^{\rt}]\omega m)$ ] in the fourth equality of ( [ dqsig ] )
. the qsde ( [ dphi ] ) can now be obtained my multiplying both sides of ( [ dqsig ] ) by @xmath199 and using ( [ chi ] ) .
the process @xmath200 in ( [ flow ] ) is the response of @xmath81 to the initial perturbation @xmath182 of the system operator under the unperturbed flow ( [ flow ] ) and satisfies the qsde ( [ dsigma ] ) : @xmath201\rd w.\ ] ] the operator @xmath184 in ( [ flow ] ) describes the response of the flow @xmath73 itself to the perturbation of the energy operators and will be referred to as the _ derivative process _ for the system operator @xmath52 .
accordingly , the term @xmath202 $ ] in the drift of the qsde ( [ dphi ] ) is the derivative process for @xmath203 .
the last equality in ( [ flow ] ) is organized as the right - hand side of the heisenberg ode ( in fictitious time @xmath120 ) of an isolated quantum system with the state space @xmath42 and the transverse hamiltonian @xmath156 .
while the unperturbed flow @xmath200 depends linearly on @xmath182 , the derivative process @xmath184 depends linearly on the perturbations @xmath142 and @xmath143 of the energy operators through the transverse hamiltonian @xmath169 and the superoperator @xmath187 in ( [ chi ] ) . in application to the vectors @xmath6 and @xmath95 of the system variables and system - field coupling operators in ( [ xwl ] ) , the qsdes ( [ dphi ] ) , ( [ djsig ] ) and the definition ( [ chi ] ) lead to @xmath204 + [ q , f ] ) + 2\im ( g\omega m ) \big)\rd t\\ \label{x'qsde } & + i([q , g]-[x , m^{\rt}]\big)\rd w,\\ \nonumber \rd l'= & \big ( \cg(m)+ i ( [ k -\im ( l^{\rt}\omega m),l ] + [ q , \cg(l ) ] ) -2\re ( [ l , l^{\rt}]\omega m ) \big)\rd t\\ \label{l'qsde } & + \big([q,[l , l^{\rt}]]-i[l , m^{\rt } ] - i[m , l^{\rt}]\big)\rd w,\end{aligned}\ ] ] where @xmath90 and @xmath91 are the drift vector and the dispersion matrix from ( [ dx ] ) . in ( [ xqsde ] )
, use is also made of the absence @xmath126 of initial perturbations in the system variables , whereby @xmath205 for any @xmath71 .
furthermore , in ( [ lqsde ] ) , we have used the relations @xmath206 in view of ( [ km ] ) , ( [ kmflow ] ) .
since the matrix @xmath94 and the quantum wiener process @xmath63 do not depend on the parameter @xmath120 , the derivative of the output field @xmath8 of the system in ( [ dy ] ) evolves according to the ode @xmath207 where @xmath208 denotes the time derivative , and @xmath209 is governed by the qsde ( [ lqsde ] ) .
in particular , for the oqho of section [ sec : oqho ] with the perturbations ( [ kmlin ] ) of the energy operators , the qsdes ( [ xqsde])([yqsde ] ) lead to the relations ( [ dxdy ] ) , ( [ abc ] ) which were obtained in section [ sec : oqho ] using more elementary techniques .
however , the latter are limited to the class of linear - quadratic perturbations ( [ kmlin ] ) which inherit the structure of the corresponding energy operators , whereas the transverse hamiltonian approach allows the system response to be investigated for general perturbations of the energy operators .
therefore , this approach can be used for the development of optimality conditions in coherent quantum control problems for larger classes of controllers and filters .
following the classical control paradigm for stochastic systems @xcite , suppose the desired operation of the quantum system being considered over an infinite time horizon corresponds to small values of an averaged cost functional @xmath210 provided the limit exists .
the latter assumption guarantees the same cesaro limit @xmath211 .
the expectation @xmath212 is taken over the tensor product @xmath213 of the initial quantum state @xmath214 of the system and the vacuum state @xmath215 in the boson fock space @xmath45 for the external fields .
also , it is assumed that @xmath216 ( which will be referred to as the _ criterion process _ ) is an adapted quantum process given by @xmath217 here , @xmath218 is a function which is extended to the noncommutative system variables and the `` signal '' part @xmath219 of the output field in ( [ dy ] ) so as to make @xmath220 a self - adjoint operator for any @xmath41 ( note that @xmath221 in view of ( [ j ] ) ) .
such extension is straightforward in the case of polynomials and can be carried out through the weyl quantization @xcite for more general functions . in the cqlqg control and filtering problems
@xcite , the function @xmath222 in ( [ z ] ) is a positive semi - definite quadratic form .
now , if the energy operators of the quantum system are subjected to perturbations ( [ km ] ) , then by using the transverse hamiltonian @xmath169 from theorem [ th : q ] in combination with theorem [ th : dashqsde ] , it follows that @xmath223.\ ] ] here , @xmath224 can be nonzero due to the dependence on @xmath95 in ( [ z ] ) despite @xmath225 , and @xmath192 is the derivative process for @xmath216 .
ignoring , at this stage , the technical issues of existence and interchangeability of appropriate limits , ( [ phiz ] ) leads to the following perturbation formula for the cost functional @xmath226 in ( [ cz ] ) : @xmath227 the right - hand side of ( [ cz ] ) is a linear functional of the perturbations @xmath142 and @xmath143 which describes the corresponding ( formal ) gateaux derivative of @xmath226 .
therefore , the quantum system is a stationary point of the performance criterion ( [ cz ] ) with respect to a subspace @xmath228 of perturbations @xmath229 of the energy operators in ( [ km ] ) , if @xmath228 is contained by the null space of the linear functional @xmath230 in ( [ cz ] ) : @xmath231 while @xmath232 in ( [ cz ] ) reduces to averaging over the invariant state of the unperturbed system ( provided the latter is ergodic and certain integrability conditions are satisfied ) , the computation of @xmath233 is less straightforward . due to the product structure of the system - field state @xmath38 ( with the external fields being in the vacuum state ) , the martingale part of the qsde ( [ dphi ] )
does not contribute to the time derivative @xmath234,\ ] ] where @xmath235 is the derivative process for @xmath236 , and the superoperator @xmath187 from ( [ chi ] ) is applied to the criterion process @xmath216 .
the relation ( [ ephidot ] ) is a complicated integro - differential equation ( ide ) .
however , this ide admits an efficient solution , for example , in the case when the system is an oqho , and the function @xmath222 in ( [ z ] ) is a polynomial . in this case , due to the structure of the gksl generator of the oqho
, @xmath236 is also a polynomial of the system variables of the same degree , thus leading to a linear relation ( with constant coefficients ) between the derivative processes @xmath192 and @xmath235 , and algebraic closedness in the ide ( [ ephidot ] ) .
therefore , for oqhos with a hurwitz matrix @xmath1 and a polynomial criterion process @xmath216 , the computation of @xmath230 ( and verification of the stationarity condition ( [ stat ] ) ) reduces to averaging over the unperturbed invariant state ( which is unique and gaussian ) .
[ ex : oqhoquad ] consider the oqho of section [ sec : oqho ] , described by ( [ theta])([abc ] ) with a hurwitz matrix @xmath1 .
suppose the criterion process @xmath216 in ( [ z ] ) is a quadratic form : @xmath237 where the @xmath238 factor is introduced for further convenience , @xmath239 is a given appropriately partitioned weighting matrix , and @xmath240 is an auxiliary matrix : @xmath241 then , in view of @xmath225 , it follows from ( [ zex ] ) and ( [ kmflow ] ) that @xmath242 the second equality in ( [ zex ] ) allows the quantum average of the corresponding derivative process @xmath192 in ( [ phiz ] ) to be represented as @xmath243 , \qquad \xi:= \re(xx^{\rt } ) = xx^{\rt } -i\theta.\ ] ] here , @xmath244 is the expectation of the derivative process @xmath245 $ ] for @xmath246 and hence , satisfies the following ide , similar to ( [ ephidot ] ) : @xmath247 + \be \chi(\xi),\ ] ] where the superoperator @xmath187 from ( [ chi ] ) is applied to @xmath246 entrywise as @xmath248 - 2\re ( [ x_jx_k , l^{\rt}]\omega m ) \\ \label{chixx } & = i[k -\im ( x^{\rt}n^{\rt}\omega m),x_jx_k ] + 4\im ( ( x_k \theta_{j\bullet } + x_j \theta_{k\bullet } ) n^{\rt}\omega m),\end{aligned}\ ] ] with @xmath249 the @xmath250th row of the ccr matrix @xmath115 . by substituting the hamiltonian and coupling operators of the oqho from ( [ rn ] ) into the gksl generator @xmath83 in ( [ dsigma ] ) and ( [ cd ] ) and using the ccrs ( [ theta ] ) together with the state - space matrices ( [ abc ] ) , it follows that @xmath251 substitution of ( [ cgxx ] ) into ( [ eupsdot ] ) reduces the ide to a lyapunov ode : @xmath252 + \be \chi(\xi ) = a\ups + \ups
a^{\rt}+ \be \chi(\xi).\ ] ] since the matrix @xmath1 is hurwitz , the unperturbed oqho is ergodic and its invariant state is gaussian with zero mean @xmath253 and covariance matrix @xmath254 , where @xmath255 is a unique solution of the ale @xmath256 . therefore , under integrability conditions for @xmath257 , to be discussed elsewhere
, the solution of the lyapunov ode ( [ upslode ] ) has a limit @xmath258 which is a unique solution of the ale @xmath259 where @xmath260 can be computed by averaging @xmath257 over the invariant gaussian state of the oqho .
therefore , by assembling ( [ z0ex ] ) and ( [ xiups ] ) into ( [ cz ] ) , it follows that @xmath261 where the limit also reduces to averaging over the invariant gaussian state . the relations ( [ p ] ) , ( [ chixx ] ) , ( [ upsale ] ) and ( [ czex ] ) express the gateaux derivative @xmath230 of the quadratic cost functional ( [ cz ] ) , specified by ( [ zex ] ) , in terms of mixed moments of the system variables @xmath6 and the perturbations @xmath149 and @xmath22 over the invariant gaussian quantum state .
in particular , if @xmath149 and @xmath22 are polynomials of the system variables , such moments can be computed using the isserlis - wick theorem @xcite .
if @xmath149 and @xmath22 are trigonometric polynomials , that is , linear combinations of unitary weyl operators @xmath262 , with @xmath263 , the moments can also be computed using quantum price s theorem @xcite , which will be discussed elsewhere .
we will now outline an application of the transverse hamiltonian technique of sections [ sec : th][sec : func ] to a quantum filtering problem .
let the open quantum system , described in section [ sec : qsde ] , be interpreted as a quantum plant . recall that it has the hamiltonian @xmath33 , system - field coupling @xmath95 , input field @xmath63 , dynamic variables @xmath6 and output field @xmath8 governed by the qsdes ( [ dx ] ) and ( [ dy ] ) .
suppose the plant is cascaded with another open quantum system , further referred to as the filter , which is driven by @xmath8 and a @xmath264-dimensional quantum wiener process @xmath265 on a boson fock space @xmath266 ; see fig .
[ fig : filtering ] .
( 50.00,30.00 ) ( 10,20)(10,10)[cc]filter ( 30,20)(10,10)[cc]plant ( 15,16)(0,1)4 ( 15,16)(1,0)35 ( 52,16)(0,0)[lc]@xmath265 ( 10,25)(-1,0)10 ( 50,25)(-1,0)10 ( 30,25)(-1,0)10 ( -2,25)(0,0)[rc]@xmath267 ( 52,25)(0,0)[lc]@xmath63 ( 25.5,26)(0,0)[cb]@xmath8 -15 mm the filter is endowed with its own initial hilbert space @xmath268 , dynamic variables @xmath269 with a ccr matrix @xmath270 , and an @xmath271-dimensional output field @xmath272 : @xmath273 we denote the filter hamiltonian by @xmath274 and the vectors of operators of coupling of the filter with the plant output @xmath8 and the quantum wiener process @xmath265 by @xmath275 respectively .
the hamiltonian @xmath274 and the coupling operators @xmath276 and @xmath277 are functions of the dynamic variables @xmath278 of the filter and hence , commute with the plant variables and functions thereof , including @xmath33 and @xmath95 .
the plant and filter form a composite open quantum stochastic system , whose vector @xmath279 of dynamic variables satisfies the ccrs @xmath280 = 2i\btheta , \qquad \btheta : = { \begin{bmatrix } \theta & 0\\ 0 & \lambda \end{bmatrix } } , \qquad \cx : = { \begin{bmatrix } x\\ \xi \end{bmatrix}}\ ] ] and is driven by an augmented quantum wiener process @xmath281 with the ito table @xmath282 here , @xmath283 denotes the ito matrix of the quantum wiener process @xmath265 of the filter , which is defined similarly to @xmath284 from ( [ omega ] ) and ( [ j ] ) as @xmath285 where the dimension @xmath264 is also assumed to be even .
the quantum feedback network formalism @xcite allows the hamiltonian @xmath286 of the plant - filter system and the vector @xmath287 of operators of coupling with @xmath281 to be computed as @xmath288 hence , the filter variables are governed by the qsde @xmath289
\rd \cw = ( i[\gamma,\xi ] + \delta(\xi))\rd t - i\big[\xi , { \begin{bmatrix}\phi^{\rt } & \psi^{\rt}\end{bmatrix}}\big]\rd { \begin{bmatrix}y\\ \omega\end{bmatrix}}.\ ] ] here , @xmath290 + \bd(\zeta)$ ] is the gksl generator of the plant - filter system , and @xmath291\bomega \bl - \frac{1}{2 } [ \bl^{\rt}\bomega \bl , \zeta ] $ ] is the corresponding decoherence superoperator , similar to ( [ cd ] ) . in ( [ dxi ] ) , use is also made of the partial gksl decoherence superoperator @xmath292 which acts on the filter variables as @xmath293\omega \phi -[\xi , \psi^{\rt}]\mho \psi - \frac{1}{2 } [ \phi^{\rt}\omega \phi + \psi^{\rt}\mho\psi , \xi]$ ] in view of ( [ cx])([bhbl ] ) .
now , consider a cqf problem which is formulated as the minimization of the mean square discrepancy @xmath294 between @xmath295 linear combinations of the plant variables of interest and filter output variables specified by given matrices @xmath296 and @xmath297 , with @xmath298 , @xmath299 ( note that , similarly to ( [ dy ] ) , @xmath300 is the signal part of @xmath267 in view of ( [ xiomlam ] ) and ( [ cw ] ) ) .
the criterion process @xmath216 in ( [ cqf ] ) is similar to that in ( [ zex ] ) : @xmath301 with only the subvector @xmath6 of @xmath279 from ( [ cx ] ) being present in ( [ zcqf ] ) .
the minimization in ( [ cqf ] ) is carried out over the filter hamiltonian @xmath274 and the vector @xmath302 of the filter - plant coupling operators in ( [ phipsi ] ) .
this problem extends @xcite in that we do not restrict attention to linear filters even if the plant is an oqho and @xmath303 depends linearly on the filter variables @xmath37 .
if @xmath304 and @xmath305 are perturbed in the directions @xmath306 consisting of self - adjoint operators representable as functions of the filter variables , the corresponding perturbations of the plant - filter hamiltonian and coupling operators in ( [ bhbl ] ) are @xmath307 by applying theorem [ th : q ] , and using ( [ cw ] ) , ( [ km0cqf ] ) and ( [ kmcqf ] ) , the corresponding transverse hamiltonian @xmath169 for the plant - filter system satisfies the qsde @xmath308 where use is also made of @xmath309 ( following from ( [ omega ] ) and the commutativity @xmath310 = 0 $ ] ) .
then the gateaux derivative @xmath230 of the cost functional in ( [ cqf ] ) is given by ( [ cz ] ) , where in view of ( [ zcqf ] ) and ( [ kmcqf ] ) ( and similarly to ( [ z0ex ] ) ) , @xmath311 in accordance with ( [ ephidot ] ) , the expectation of the derivative process @xmath312 $ ] satisfies the ide @xmath313 + \be \chi(z),\ ] ] where , in view of ( [ dqcqf ] ) , the superoperator @xmath187 in ( [ chi ] ) is given by @xmath314 -2\re ( [ z,(l+\phi)^{\rt}]\omega m).\ ] ] therefore , in the case when both the plant and the unperturbed filter are oqhos with hurwitz dynamics matrices ( while the perturbations in ( [ km0cqf ] ) are not necessarily linear - quadratic ) , @xmath230 can be computed explicitly along the lines of example [ ex : oqhoquad ] of section [ sec : func ] due to @xmath216 being a quadratic form of the plant - filter variables @xmath279 .
the transverse hamiltonian method , outlined above , has recently been used in @xcite to show that , in the mean square optimal cqf problem for linear plants , linear filters can not be improved locally ( in the sense of first - order optimality conditions ) by varying the energy operators of the filter along the weyl operators .
for a class of open quantum systems governed by markovian hudson - parthasarathy qsdes , we have introduced a transverse hamiltonian and derivative processes associated with perturbations of the energy operators .
this provides a fully quantum tool for infinitesimal perturbation analysis of system operators and averaged cost functionals .
the proposed variational method can be used for the development of optimality conditions in coherent quantum control problems .
we have illustrated these ideas for oqhos with quadratic performance criteria and the mean square optimal cqf problem .
d.dong , and i.r.petersen , quantum control theory and applications : a survey , _ iet contr .
12 , 2010 , pp . 26512671 .
s.c.edwards , and v.p.belavkin , optimal quantum filtering and quantum feedback control , arxiv : quant - ph/0506018v2 , august 1 , 2005 .
g.b.folland , _ harmonic analysis in phase space _ , princeton university press , princeton , 1989 .
c.w.gardiner , and p.zoller , _ quantum noise _ , 3rd ed . ,
springer , berlin , 2004 .
v.gorini , a.kossakowski , and e.c.g.sudarshan , completely positive dynamical semigroups of @xmath119-level systems , _ j. math
17 , no . 5 , 1976 , pp . 821825 .
j.gough , v.p.belavkin , and o.g.smolyanov , hamilton - jacobi - bellman equations for quantum optimal feedback control
b : quant . semiclass . opt .
_ , vol . 7 , 2005 , pp . s237s244
j.gough , and m.r.james , quantum feedback networks : hamiltonian formulation , _ commun .
_ , vol . 287 , 2009 , pp . 11091132 .
a.s.holevo , quantum stochastic calculus , _ j. math .
56 , no . 5 , 1991 , pp .
n.j.higham , _ functions of matrices _ , siam , philadelphia , 2008 .
r.l.hudson , and k.r.parthasarathy , quantum ito s formula and stochastic evolutions , _ commun .
93 , 1984 , pp .
k.jacobs , and p.l.knight , linear quantum trajectories : applications to continuous projection measurements , _ phys .
a _ , vol .
57 , no . 4 , 1998 , pp .
s.janson , _ gaussian hilbert spaces _ , cambridge university press , cambridge , 1997 .
m.r.james , h.i.nurdin , and i.r.petersen , @xmath315 control of linear quantum stochastic systems , _ ieee trans .
53 , no . 8 , 2008 , pp . 17871803 .
i.karatzas , and s.e.shreve , _ brownian motion and stochastic calculus _ , 2nd ed . , springer - verlag , new york , 1991 .
h.kwakernaak , and r.sivan , _ linear optimal control systems _ , wiley , new york , 1972 .
l.d.landau , and e.m.lifshitz , _ statistical physics _ , 3rd ed . ,
elsevier , amsterdam , 1980 .
g.lindblad , on the generators of quantum dynamical semigroups , _ commun . math
48 , 1976 , pp .
e.merzbacher , _ quantum mechanics _ , 3rd ed . ,
wiley , new york , 1998 .
z.miao , and m.r.james , quantum observer for linear quantum stochastic systems , proc .
51st ieee conf .
decision control , maui , hawaii , usa , december 10 - 13 , 2012 , pp
. 16801684 .
h.i.nurdin , m.r.james , and i.r.petersen , coherent quantum lqg control , _ automatica _ , vol . 45 , 2009 , pp . 18371846 .
k.r.parthasarathy , _ an introduction to quantum stochastic calculus _ , birkhuser , basel , 1992 .
k.r.parthasarathy , what is a gaussian state ?
4 , no . 2 , 2010 , pp . 143160 .
k.r.parthasarathy , and k.schmidt , _ positive definite kernels , continuous tensor products , and central limit theorems of probability theory _ , springer - verlag , berlin , 1972 .
i.r.petersen , quantum linear systems theory , proc .
networks syst . , budapest , hungary , july 59 , 2010 , pp .
a.k.sichani , i.g.vladimirov , and i.r.petersen , a gradient descent approach to optimal coherent quantum lqg controller design , _ accepted at american control conference , 2015 _ , preprint : arxiv:1502.00274 , 1 february 2015 .
i.g.vladimirov , and i.r.petersen , a quasi - separation principle and newton - like scheme for coherent quantum lqg control , _ syst .
62 , no . 7 , 2013 , pp . 550559 .
i.g.vladimirov , and i.r.petersen , coherent quantum filtering for physically realizable linear quantum plants , proc .
european control conference , ieee , zurich , switzerland , 17 - 19 july 2013 , pp .
i.g.vladimirov , a quantum mechanical version of price s theorem for gaussian states , proc .
australian control conference , aucc 2014 , ieee , canberra , australia , 17 - 18 november 2014 , pp .
i.g.vladimirov , weyl variations and local sufficiency of linear observers in the mean square optimal coherent quantum filtering problem , _ submitted _ , preprint : arxiv:1506.07653 [ quant - ph ] , 25 june 2015 . | this paper is concerned with variational methods for nonlinear open quantum systems with markovian dynamics governed by hudson - parthasarathy quantum stochastic differential equations .
the latter are driven by quantum wiener processes of the external boson fields and are specified by the system hamiltonian and system - field coupling operators .
we consider the system response to perturbations of these energy operators and introduce a transverse hamiltonian which encodes the propagation of the perturbations through the unitary system - field evolution .
this provides a tool for the infinitesimal perturbation analysis and development of optimality conditions for coherent quantum control problems .
we apply the transverse hamiltonian variational technique to a mean square optimal coherent quantum filtering problem for a measurement - free cascade connection of quantum systems . |
an equilibrium ( no convection ) star might still be a good conductor and contain magnetic fields with the magnetic stress balanced by the fluid pressure and gravity . we want to describe the generic features of such equilibrium magnetic fields .
this problem was clearly formulated many years ago by chandrasekhar and fermi , but it remains unsolved to this day .
recent progress came from numerical work ( braithwaite & nordlund 2005 ) .
the recent numerical work does not point out an important generic feature of the equilibrium
singular current layers .
moreover , we propose that generic equilibrium carries infinitely many singular current layers .
this proposition generic equilibrium carries a volume - filling set of singular current layers is our main thesis .
this is indeed just a proposition .
it is based on very shaky arguments . a numerical proof or refutation
is needed .
the author tried to solve the problem numerically , but succeeded only in 2d , where the singular current lines are isolated rather than area - filling .
still , it appears that modern computers , if handled by professionals , are powerful enough to also solve the 3d problem ( by the relaxation technique , 2 ) . since it is believed that close to equilibrium magnetic fields do exist in neutron stars and in other non - convective stars , a careful numerical study of the singularities should be worthwhile .
if singularities are indeed generic , they might be responsible for some of the observed phenomena say , magnetic field decay and triggering the magnetar flares .
the singular current layer is a local feature , it can be considered without gravity . to simplify the analysis
, we will also assume an incompressible constant density fluid .
then the magnetic fields @xmath0 will be in equilibrium if the magnetic @xmath1 force is balanced by the pressure gardient : @xmath2 we merely want to discuss solutions of eq.([equi ] ) .
a well - known procedure for finding stable magnetic equilibria is the relaxation method . consider a viscous but ideally conducting incompressible fluid
. set an arbitrary initial magnetic field and velocity , and let the system evolve .
the total energy will decrease due to internal friction , and asymptotically the fluid must stop .
the fluid can stop only in a magnetic equilibrium state .
since the fluid is an ideal conductor , the magnetic field topology is not changed by the fluid motion .
conserved topology plus incompressibility guarantee that , generically , the equilibrium state reached by the fluid will be non - trivial , @xmath3 .
the relaxation method proves the existence of magnetic equilibria with arbitrary topology .
the method is constructive it allows one to actually find this equilibrium . the corresponding mathematical problem , in the limit of large viscosity , and after an obvious generalization reads : @xmath4 where @xmath5 is an arbitrary positive operator which commutes with @xmath6 .
the first equation says that the magnetic field @xmath0 is frozen into the fluid moving with velocity @xmath7 .
the second equation says that the fluid incompressibly yields to the magnetic force ; @xmath8 corresponds to internal friction , @xmath9 corresponds to external friction ( @xmath10 ) .
eq.([relax ] ) describes a dissipative evolution .
magnetic energy @xmath11 decreases : @xmath12 the evolution stops only when @xmath13 which is equivalent to eq.([equi ] ) .
we must conclude that for any solenoidal field , there exists an incompressible deformation of this field which is an equilibrium .
the following argument , due to arnold ( 1986 ) , explains why the magnetic equilibrium problem remains unsolved , despite the clear and simple logic of the relaxation method . from eq.([equi ] ) , @xmath14 .
this means that magnetic field lines lie on the surfaces @xmath15=const .
but the magnetic field we started with in the relaxation method generically does not possess magnetic surfaces
generic field lines fill up volumes , not surfaces .
clearly , the isotopological relaxation ( [ relax ] ) can not comb the intertwined volume - filling lines onto the nested tori @xmath15=const . and yet
, the relaxation argument of @xmath16 is still valid . for any solenoidal field there must exist at least one incompressible deformation of the field which is an equilibrium , because asymptotically , the motion described by eq.([relax ] ) stops .
it also appears likely that for suitable operators @xmath5 , the relaxing field remains non - singular at any finite time , meaning that the relaxation method indeed stops at a well - defined asymptotic magnetic field ( the final field may be singular and different for different operators @xmath5 ) .
after the above discussion , one really wants to offer this problem to a computer and see how the computer will resolve this contradiction between topology and equilibrium .
the author was able to obtain trustworthy numerical results only in 2d .
in fact , the 2d case is reach enough to show an important generic feature of magnetic equilibrium
singular current layers .
we discuss the 2d equilibria in this section .
first of all , it appears that in 2d there are no topological obstacles to equilibrium .
indeed , arnold s argument says that magnetic field lines lie on the lines of constant pressure .
but the 2d field lines are always the isolines of the magnetic stream function : @xmath17 gives @xmath18 .
therefore one expects that a smooth equilibrium state will correspond to any initial state @xmath19 . in reality , it has been known for some time that initial states with non - trivial topology relax to equilibrium states with singular current layers . to show how the singular current layers form , consider relaxation with internal friction .
@xmath8 in eq.([relax ] ) gives the `` incompressible diffusion '' equation in 2d : @xmath20 the magnetic energy decreases @xmath21 the evolution stops at equilibrium @xmath22 or , in another form , @xmath23 computer simulations of eq.([relax2 ] ) ( using fast fourier transform ) show that all initial saddles of @xmath19 develop into singular current layers , see figures ( 1 ) and ( 2 ) .
the singularities are of the type @xmath24 , giving the field @xmath25 , and the current @xmath26 singular current layers .
the reason for existence of singular current layers in 2d appears to be clear . the incompressible relaxation process ( [ relax2 ] ) tries to make the field lines as short as possible , while keeping the area enclosed by each field line constant .
the field lines try to deform into circles .
this is impossible for non - trivial initial topology .
then , as illustrated by fig.([fig3 ] ) , one does expect that the saddles ( x - points ) of the initial field split into two y - points with a singular current layer between them .
one can actually transform the cartoon fig.([fig3 ] ) into a rigorous calculation .
take the limit in which only an infinitesimal layer of the field surrounds the separatrix .
assume equal initial energies in the outer and inner parts of the separatrix .
then , the magnetic energy will scale as @xmath27 , where @xmath28 is the inner and @xmath29 is the outer lengths of the separatrix .
minimization of the energy at fixed enclosed areas predicts the final shape of the separatrix two @xmath30 circular arcs and a line segment separating the arcs , for a symmetric separatrix .
the direct numerical simulation confirms this result , fig.([fig4 ] ) .
we have also performed 3d numerical simulations of the incompressible relaxation ( [ relax ] ) with the inner friction , @xmath8 . we see clear evidence for the singular current layers formation but we were unable to see how many singular layers form , how are they distributed in the volume , and how exactly does the system manage to preserve the field line topology , while putting the field lines onto the constant pressure surfaces .
one probably needs a better than our 64@xmath31 resolution , and/or a better handling of the numerical issues .
having failed numerically , we resort to linear perturbation theory .
we start with a translation invariant equilibrium field , put a small perturbation on it , and ask what does the perturbed field relax into .
while this linear problem can indeed be solved analytically , it is unclear what does the solution have to do with the generic equilibrium , which , of course , is fully nonlinear .
we will see , however , that in 2d the linear perturbation theory can be used as a sensible guide to the properties of the generic equilibrium .
we will assume that the same is true in 3d , and this will allow us to propose that the generic magnetic equilibrium carries a volume - filling set of singular current layers .
this kind of reasoning is unsatisfactory .
but this all we can offer .
in 2d , the initial y - translation invariant background field is @xmath32 , corresponding to the magnetic field @xmath33 in the y - direction .
the perturbation is @xmath34 , and we assume that @xmath35 otherwise the perturbation simply re - difines the background field @xmath32
. the final relaxed state will be @xmath36 .
to linear order , the equilibrium condition ( [ equi2 ] ) gives @xmath37 if @xmath38 everywhere , one can write the perturbation as @xmath39
. then @xmath40 with the periodic or decreasing x - boundary conditions , this gives @xmath41 and @xmath42 .
this result is trivial just says that the field lines straighten up during the relaxation .
now consider the case when the background magnetic field @xmath43 has zeros . in the short wavelength limit , for @xmath44 , eq.([pert2 ] ) reads @xmath45 the solution is @xmath46 this is a current singularity .
so far everything was fully rigorous .
what is shaky , is our extrapolation of this result to the non - linear regime .
based on the linear perturbation theory , we `` conclude '' that the generic 2d equilibrium will have isolated singular layers at magnetic field zeros .
this is indeed correct in 2d , if by zeros we mean only the saddle - point zeros . in 3d
, we choose a background field of the form @xmath47 .
the perturbation is @xmath48 .
then the linearized equilibrium condition ( [ equi ] ) is @xmath49 @xmath50 @xmath51 here @xmath52 , and @xmath15 is the pressure perturbation .
the magnetic field perturbation is solenoidal @xmath53 from ( [ pert32 ] ) and ( [ pert34 ] ) we calculate the product @xmath54 in terms of @xmath55 and @xmath15 . from ( [ pert33 ] ) and ( [ pert34 ] ) we calculate the product @xmath56 in terms of @xmath55 and @xmath15 .
plug these into ( [ pert31 ] ) .
the pressure perturbation @xmath15 cancels , and we get @xmath57 this equation is isomorphic to eq.([pert2 ] ) , and we can use the 2d results . if @xmath58 has no zeros , the perturbation is zero . for the relaxation problem
it means that the lines straighten up .
but if @xmath59 does have zeros , each zero will develop into a singular layer .
the difference between 2d and 3d is that in 3d the zeros of @xmath58 depend on wavenumbers .
suppose we work in a @xmath60-periodic cube .
then the wavenumbers @xmath61 , @xmath62 are arbitrary integers , and all points @xmath63 where @xmath64 is a rational number are zeros of @xmath59 for the appropriate choice of @xmath61 and @xmath62 . `` based '' on the above
, we propose that generic magnetic equilibrium contains a volume - filling ( meaning everywhere dense within some domains ) set of singular current layers .
this proposition leaves many questions unanswered : * is this proposition true at all ? + and assuming it is true : * how exactly does this infinite set of singular layers help resolve the topological obstacles to relaxation ?
* where do the layers form ?
, that is on closed field lines . ]
* which current layers , a few strongest ones , or a lot of weaker ones , are responsible for the residual damping , once a small ( probably anomalous in the real world ) resistivity is added to the problem ?
we propose that generic magnetic equilibrium contains a volume - filling set of singular current layers , because topological obstacles to relaxation are everywhere dense .
this is just a proposition .
the proposition leaves many questions unanswered .
the proposition might turn out to be completely false . | we propose that generic magnetic equilibrium of an ideally conducting fluid contains a volume - filling set of singular current layers .
singular current layers should exist inside neutron stars .
residual dissipation in the singular current layers might be the main mechanism for the magnetic field decay .
the slow decay of the field might be the clock responsible for triggering the magnetar flares . |
in the abelian sandpile model , items are randomly assigned to discrete locations on the sandpile until a critical slope , or number of items . at this point ,
the locations that have a slope greater than the critical slope redistribute their items in a cascading fashion onto neighboring locations @xmath4 $ ] .
consider an abelian sandpile described by a closed circle made of n vertices .
randomly distributed among the n vertices are k items , representing an initial distribution . with the goal of distributing the items among the vertices such that no vertex has more than 1 item placed on it , called the terminal set , any vertex a with more than 1 item has 2 items removed from it , with 1 item moving to the vertex right of vertex a and the other moving to the vertex to the left of vertex a. for @xmath5
, the terminal set is reached after a finite number of turns and for @xmath6 , the terminal set is never reached . for @xmath7 ,
the initial distribution of the k items determines whether the terminal set is reached or not .
n terminal set ] * open problem : * suppose @xmath8 .
classify all successful initial distributions of items that lead to the terminal set .
any circular set of vertices can be represented by a cyclic ordered set @xmath9 $ ] . for example
, an @xmath10 circle can be represented by the set : @xmath11 and @xmath12 , where vertex @xmath13 is connected to vertex @xmath14 and vice - versa and @xmath15 is finite @xmath16 $ ] .
since each set is a cyclic order , the following is also true : @xmath17 @xmath18 , which is really just the equivalent of stating that the circle s order is independent of how you rotate it in space , and so the cyclic ordered set representing the circle . for any @xmath8
, the terminal set is reached when an equal distribution of items among all of the vertices is achieved , i.e. where each vertex has 1 item .. ] the first step in determining which initial distributions are successful is to enumerate all possible initial distributions between the set where all items are grouped in one vertex , @xmath19 , and the terminal set @xmath20 , where @xmath21 and @xmath22 .
another way of representing the possible distributions of k in n positions where 0 s and permutations are allowed is the @xmath23 @xmath24 $ ] .
for example , all initial distributions for @xmath25 are : @xmath26 @xmath27 the only initial distributions from this weak composition that reach the terminal set in a finite number of turns are : @xmath28 .
these successful initial distributions yield the terminal set after a finite number of turns greater than or equal to 0 and thus include the terminal set . using this system ,
the original algorithm is : @xmath29 .
* conjecture 1.0 : * all successful initial distributions can be classified as the enumeration of all distributions that are a discrete number of reversed algorithmic turns away from the terminal set .
a reversed algorithmic turn can be described as : @xmath30 .
it is possible to see if each initial distribution reaches the terminal set in a finite number of turns with enough computing power by using the algorithm discussed in the introduction .
however , it is impossible to know with the current method whether or not an arbitrary initial distribution will end in the terminal set .
* proof of conjecture 1.0 : * every successful initial distribution has reached the terminal set in a finite @xmath31 turns , where @xmath32 is the particular successful starting position .
it is true then that the non - terminal set described by @xmath33 turns is a successful initial distribution that is 1 turn away from reaching the terminal set .
this is also the case for the initial distribution that is @xmath34 turns away from the terminal set .
in fact , since all turns are of the same type ( any vertex with more than 1 item moves 1 to the left neighbor and 1 item to the right neighbor ) , by induction we can state that all successful initial distributions between the terminal set and the successful initial distribution @xmath35 can be described as being @xmath36 turns away from the terminal set , where @xmath37 . assuming the successful initial distribution @xmath31 turns away from the terminal set allows for any more reversed algorithmic turns , the initial distribution described by @xmath38 , or the initial distribution that comes 1 turn before @xmath35 is also a valid successful initial distribution
this must also therefore be true for all successful initial distributions described by @xmath39 turns away from the terminal set , where @xmath40 .
in other words , all initial distributions @xmath41 turns from the terminal set are successful initial distributions and therefore all possible successful initial distributions can be enumerated . *
since this is true for any arbitrary successful initial distributions , all successful initial distributions can be classified in this way for the case where * @xmath8 * and no successful initial distribution is outside of this classification by definition . *
* conjecture 2.0 : * every successful initial distribution shares the same characteristic modulus with the terminal set .
* proof of conjecture 2.0 : * for every 2-number item change in the number of items at vertex @xmath42 , there has to be a 1-number item change in @xmath43 and @xmath44 .
therefore the initial distribution described by : @xmath45 , must be equivalent to the initial distribution : @xmath29 , in that they are both successful initial distributions .
it is not enough to know the change in magnitude of each vertex , but also the direction of the change in magnitude of the number of items at each vertex , and so classifying these as successful initial distributions must take both factors into account . taking the sum of the product of the position of the vertex with the number of items on it , we can define the characteristic classification as : @xmath46 , where @xmath47 is the position in the initial distribution , and @xmath48 is the number of items at that vertex .
this can be proven through basic algebra : @xmath49 @xmath50 mod ( n)\ ] ] @xmath51 @xmath52 mod ( n)\ ] ] , where @xmath53 by definition since they are simply magnitude changes in position . therefore : @xmath54 since this applies to any arbitrary @xmath55 and therefore any @xmath56 , and any change in the @xmath35 turns to acquire the terminal set must follow the @xmath29 algorithm in the cyclic ordered set , * this general proof applies to any successful initial distribution in relation to the terminal set and therefore any initial distribution can be checked to see if it ends in the terminal set . *
any initial distribution s classification that does not equal the terminal set s classification is therefore not a successful initial distribution .
[ 1 ] : bak , p. , tang , c. and wiesenfeld , k. ( 1987 ) . ``
self - organized criticality : an explanation of 1/ noise '' .
physical review letters 59 ( 4 ) : 381384 .
bibcode:1987phrvl .. 59 .. 381b .
doi:10.1103/physrevlett.59.381 .
[ 2 ] : huntington , edward v. ( july 1935 ) , `` inter - relations among the four principal types of order '' ( pdf ) , transactions of the american mathematical society 38 ( 1 ) : 19 , doi:10.1090/s0002 - 9947 - 1935 - 1501800 - 1 .
[ 3 ] : courcelle , bruno ( 21 august 2003 ) , `` 2.3 circular order '' , in berwanger , dietmar ; grdel , erich , problems in finite model theory ( pdf ) , p. 12 .
[ 4 ] : eger , steffen ( 2013 ) . `` restricted weighted integer compositions and extended binomial coefficients '' .
journal of integer sequences 16 .
in the appendix , specific examples for each conjecture will be used to demonstrate their proofs . | this paper outlines a closed solution to an open problem in graph theory concerning the classification of the successful initial distributions of k items in n vertices , where @xmath0 , that lead to the terminal set @xmath1 , where @xmath2 and @xmath3 .
first , each successful initial distribution is enumerated using an algorithm .
the closed solution classifies the terminal set in terms of its modulus , and proves that each successful initial distribution can be classified by the same modulus .
graph theory , abelian sandpile , distribution , weak composition , modulus |
metallic ferromagnetism is one of the oldest problems in solid - state physics . despite all progress in solid - state theory over the last century @xcite , we are still far away from a definite theory for transition metals like iron , cobalt and nickel .
it is generally accepted that the magnetic order in metals is a consequence of the interplay between the electrons kinetic energy and their mutual coulomb interactions .
therefore , it does not come as a surprise that a complete theory for this many - particle effect is difficult to develop .
nevertheless , it is commonly believed that at least the ground - state properties of itinerant magnets can be described by ( effective ) one - particle theories such as hartree - fock theory or ( spin-)density - functional theory ( sdft ) .
however , serious doubts are in order , as a closer look on available experimental data reveals .
for example , sdft fails to reproduce even qualitatively the electronic band - structure of nickel , one of the simplest itinerant ferromagnets @xcite ; for a summary , see ref . .
for iron , the agreement is better but can not be considered as satisfactory either . in any case , the same physical mechanism is acting in both systems , and a convincing theory should cover all elements of the iron group equally well .
single - particle theories are not very satisfactory also from a purely theoretical point of view : it is well known that they significantly overestimate the stability of ferromagnetism @xcite .
for example , hartree - fock theory for one - band hubbard models gives a ferromagnetic ground state above some moderate interaction strength .
in contrast , many - body theories show that ferromagnetism in one - band hubbard models occurs , if at all , only for very large coulomb interactions or very peculiar choices of the one - particle density of states @xcite .
this indicates that the stoner mechanism is less prominent in itinerant ferromagnets than implied by single - particle theories . as already argued by van vleck @xcite and
confirmed recently @xcite , an essential ingredient for metallic ferromagnetism is the local ( hunds - rule ) exchange coupling of electrons , a generic feature of electrons in degenerate bands .
in contrast to the one - band case , ferromagnetism occurs for moderate coulomb - interactions in a generic two - band model provided that local hunds - rule terms are included .
therefore , even a qualitative understanding of the 3@xmath0-elements requires a careful consideration of their atomic open - shell structures . in our recent work @xcite
we have developed a variational approach which allows us to examine multi - band hubbard models with an arbitrary number of correlated orbitals per lattice site . for a realistic description of the 3@xmath0-elements a minimal model
includes the 3@xmath0 , 4@xmath1 , and 4@xmath2 bands .
when applied to nickel , our theory provides results which agree much better with experiments than all previous sdft - calculation @xcite .
in particular , our approach reproduces the experimental fermi - surface topology of nickel and removes all other qualitative sdft deficiencies .
since our variational many - body approach provides a reasonable description of the ground state , it may serve as an appropriate starting point for the study of excitations in itinerant ferromagnets . as observed in inelastic neutron scattering experiments @xcite , the spectrum of itinerant ferromagnets displays low - lying spin - wave excitations .
these excitations are generally considered to be the driving force for the transition from the ferromagnetic phase into the paramagnetic phase at the curie temperature .
therefore , their correct description is also very important for our understanding of finite - temperature properties of itinerant ferromagnets .
unfortunately , almost all established theories which are used for the analysis of spin - waves in transition metals are based on effective single - particle theories which do not necessarily provide a satisfactory description of the ground state .
the textbook approach to the problem of itinerant ferromagnetism is a hartree - fock theory for the ground state which is then combined with a random - phase approximation ( rpa ) for the description of spin - excitations @xcite .
for example , in ref . the rpa - method has been applied to iron and nickel starting from a sdft calculation for the ground state .
later , a similar method has been used to analyze data from electron energy - loss spectroscopy for the same materials @xcite .
another common approach starts from energy calculations for static spin deformations of the ferromagnetic ground state within the sdft .
the spin - wave energies are identified as the energy difference between these `` frozen - magnon '' states and the ground state@xcite . despite their conceptual shortcomings both sdft - based approaches
are able to reproduce the experimental data surprisingly well .
very recently @xcite one of us has proposed a spin - wave theory which starts from the variational ground state of multi - band hubbard models as introduced in ref . .
this theory allows us to determine the spin - wave spectrum for systems with an arbitrary number of correlated orbitals per lattice site .
the main advantage of this new approach lies in its foundation on a true many - particle description of the ground state . in this way
, we are able to provide a consistent picture of the ferromagnetic fermi liquid and its spin - wave excitations . in order to obtain a better understanding of our own but likewise of the sdft - based methods it is crucial to apply them for a generic but simple model system .
to our knowledge such an analysis has not been performed yet .
even a comparison between the two well - established sdft methods is still lacking . in this work
we analyze the hartree - fock / rpa theory and compare it with our variational many - body approach .
although the rpa is generally considered as the `` standard '' method for the description of magnetic excitations in itinerant ferromagnets there exists no systematic derivation for multi - band hubbard models . in sect .
[ rand ] , we derive the general rpa equations from a decoupling scheme and a standard diagrammatic viewpoint . in sect .
[ comp ] we provide numerical results for a generic two - band model and show that both methods lead to qualitatively different results . for generic values of the magnetization , the hartree - fock /
rpa treatment incorrectly predicts an instability of the collinear ferromagnetic phase against some spiral ordering . in our variational approach
the spin - waves always have positive energy and their stiffness decreases with increasing correlations .
the case of a one - band model with fully polarized bands allows us to pinpoint the origin of these discrepancies : the hartree - fock theory allows for ferromagnetism with small local moments and large longitudinal spin - fluctuations due to stoner excitations .
these are actually absent in a realistic correlated - electron description where charge fluctuations are strongly suppressed in the ferromagnetic and paramagnetic phases .
conclusions in sect .
[ summ ] close our presentation .
in this section we derive the random phase approximation ( rpa ) for ferromagnetism in multi - band hubbard models . starting from a lda - calculation ,
rpa results for iron and nickel have been reported in ref . .
however , in these calculations the relevant coulomb matrix elements were significantly simplified . in this work we present a more general treatment .
we will address the following general class of multi - band hubbard models @xcite @xmath3 here , @xmath4 creates an electron with combined spin - orbit index @xmath5 at the lattice site @xmath6 of a solid ( @xmath7 ; @xmath8 ; @xmath9 for 3@xmath0 electrons ) . the atomic hamiltonian @xmath10 is assumed to have site - independent interaction parameters @xmath11 . in order to determine the spin - wave properties , we need to consider the imaginary part @xmath12 of the transverse susceptibility@xcite , which is given by the retarded two - particle green function [ 1.3 ] @xmath13 \right| \psi_{0}\right\rangle \;.
\label{1.3b } \end{aligned}\ ] ] here , we introduced the @xmath14-dependent spin - flip operators [ 1.4 ] @xmath15 in the heisenberg picture , where the sum runs over all @xmath16 lattice sites @xmath17 and orbitals @xmath18 .
magnetic excitations are found as poles of the green function @xmath19 , or , equivalently , as peaks in @xmath20 at energies @xmath21 .
the lehmann representation of ( [ 1.3 ] ) , @xmath22 \ ; , \label{1.5}\ ] ] shows that the weights of the poles in @xmath19 at the energies @xmath23 are given by @xmath24 in a ferromagnetic system the state @xmath25 is also a ground state of @xmath26 for @xmath27 because the operator @xmath28 just flips a spin in the spin multiplet of the ground state @xmath29 .
therefore , we can conclude that @xmath30 has one isolated pole for @xmath31 .
this is not surprising since the spin - wave is the gapless goldstone mode in the symmetry - broken ferromagnetic phase . for finite values of @xmath14
, it is an experimental fact that there are also pronounced peaks in @xmath12 at the spin - wave energies @xmath32 the constant @xmath33 is usually denoted as the `` spin - wave stiffness '' .
the one - particle hamilton @xmath34 in ( [ 1.1 ] ) may be written in momentum space as [ 2.4 ] @xmath35 here , we introduced the operators [ 7.2 ] @xmath36 and the respective energy matrix - elements @xmath37 where the elements @xmath38 of the unitary matrix @xmath39 will be identified later .
the transformation ( [ 7.2 ] ) leads to the following expression for the atomic hamiltonian , @xmath40 where [ 7.4 ] @xmath41 we choose the matrix @xmath39 in order to fulfill the corresponding hartree - fock equations @xmath42 \ ; , \label{7.16 } \end{aligned}\ ] ] with [ 7.16h ] [ tmatrices ] @xmath43 our hartree - fock theory is restricted to cases where the translational invariance is conserved but we still allow for a collinear ferromagnetic ground state .
the hartree - fock hamiltonian thus reads @xmath44 in the special case where , ( i ) , all orbitals @xmath18 belong to different representations of the point - symmetry group of the lattice , and , ( ii ) , the hartree - fock ground state @xmath45 is invariant under the respective symmetry operations , eq .
( [ 7.16b ] ) becomes @xmath46 under these conditions , the results of ref .
are recovered ; see below .
we will derive the green function ( [ 1.3 ] ) in the random - phase approximation using the standard equation - of - motion technique@xcite .
first , we rewrite @xmath19 as [ 7.5 ] @xmath47 the equation of motion for @xmath48 can not be decoupled self - consistently .
therefore , we will consider the more general green function @xmath49 which allows us to express ( [ 7.5b ] ) as @xmath50 the equation of motion for @xmath51 has the form [ 7.8 ] @xmath52 \rangle _ { \psi _ { 0 } } \ ; , \label{7.8a } \\
b & = & \langle \langle [ \hat{a}_{\vec{k}+\vec{q};\alpha_{1},\uparrow } ^{+ } \hat{a}_{\vec{k};\alpha_{2},\downarrow } ^ { { } } , \hat{h}_{1}]\,;\ , \hat{c}_{\vec{k}^{\prime } ; b^{\prime } , \downarrow } ^{+ } \hat{c}_{\vec{k}^{\prime } + \vec{q};b^{\prime } , \uparrow } ^ { { } } \rangle \rangle_{e } \ ; , \label{7.8b } \\
c & = & \langle \langle [ \hat{a}_{\vec{k}+\vec{q};\alpha_{1},\uparrow } ^{+ } \hat{a}_{\vec{k};\alpha _ { 2},\downarrow } ^ { { } } , \hat{h}_{\text{at } } ] \,;\ , \hat{c}_{\vec{k}^{\prime } ; b^{\prime } , \downarrow } ^{+ } \hat{c}_{\vec{k}^{\prime } + \vec{q};b^{\prime } , \uparrow } ^ { { } } \rangle \rangle_{e}\;. \label{7.8c } \end{aligned}\ ] ] in rpa we assume that the exact ground state @xmath53 of our hamiltonian ( [ 1.1 ] ) may be replaced by the ( spin - polarized ) hartree - fock ground state @xmath54 of ( [ hfhamilt ] ) . then , the expectation value ( [ 7.8a ] ) becomes @xmath55 the evaluation of ( [ 7.8b ] ) leads to @xmath56 finally , the commutator in ( [ 7.8c ] ) generates two - particle operators , which we decouple according to the rule @xmath57 this approximation appears to be natural because we replaced @xmath58 by the one - particle product state @xmath59 .
note , however , the replacement ( [ 7.11 ] ) does not become exact even if we work with @xmath59 instead of @xmath58 . after the application of the decoupling scheme ( [ 7.11 ] ) , eq .
( [ 7.8c ] ) may be written as [ 7.12 ] @xmath60 \widetilde{g}_{\alpha_{1},\beta^{\prime};b^{\prime}}^{\vec{k},\vec{k}^{\prime},\vec{q}}(e ) \ ; , \label{7.12a } \\ c_2 & = & - \frac{2}{l}\sum_{\beta , \beta ^{\prime } } \sum_{\vec{p};\sigma } \bigl[n_{\vec{p}}^{\beta \sigma } v_{\beta ^{\prime } \alpha_{1}\uparrow } ^{\beta \beta \sigma } ( \vec{k}+\vec{q},\vec{p},\vec{p},\vec{k}+\vec{q}\ , ) -n_{\vec{p}}^{\beta \uparrow } v_{\beta \alpha _ { 1}\uparrow}^{\beta ^{\prime } \beta \uparrow } ( \vec{p},\vec{k}+\vec{q},\vec{p},\vec{k}+\vec{q}\,)\bigr ] \widetilde{g}_{\beta^{\prime},\alpha_{2};b^{\prime } } ^{\vec{k},\vec{k}^{\prime } , \vec{q}}(e ) \ ; , \nonumber \\
\label{7.12b } \\ c_3 & = & \frac{2}{l}\sum_{\beta , \beta ^{\prime } } \sum_{\vec{p } } \left ( n_{\vec{k}}^{\alpha _ { 2}\downarrow } -n_{\vec{k}+\vec{q}}^{\alpha _ { 1}\uparrow } \right ) v_{\beta \alpha _ { 1}\uparrow }
^{\alpha_{2}\beta ^{\prime } \downarrow } ( \vec{p}+\vec{q},\vec{k},\vec{p},\vec{k}+\vec{q}\ , ) \widetilde{g}_{\beta,\beta^{\prime};b^{\prime}}^{\vec{p},\vec{k}^{\prime},\vec{q}}(e ) \;.
\label{7.12d } \end{aligned}\ ] ] when we use the explicit expression ( [ 7.4 ] ) , the sum over @xmath61 can be carried out for @xmath62 and @xmath63 because the green functions do not depend on @xmath61 .
[ 7.14 ] @xmath64 \widetilde{g}_{\alpha_{1},\beta;b^{\prime}}^{\vec{k},\vec{k}^{\prime},\vec{q}}(e ) \label{7.14a } \ ; , \\
c_2&= & 2\sum_{b_{1},b_{2},b_{3},b_{4},\sigma , \sigma ^{\prime } } u_{\sigma \sigma^{\prime } } ^{b_{1},b_{2},b_{3},b_{4 } } \sum_{\beta } \delta _ { \sigma } ^{\uparrow } \widetilde{g}_{\beta,\alpha_{2};b^{\prime}}^{\vec{k},\vec{k}^{\prime},\vec{q}}(e ) \bigl [ t_{b_{1},b_{3}}^{\uparrow } \delta _ { \sigma ^{\prime } } ^{\uparrow } u_{\beta , b_{2}}^{\ast } ( \vec{k}+\vec{q},\uparrow ) u_{\alpha_{1},b_{4}}^{{}}(\vec{k}+\vec{q},\uparrow ) \nonumber \\ & & \hphantom{2\sum_{b_{1},b_{2},b_{3},b_{4},\sigma , \sigma ^{\prime } } u_{\sigma \sigma^{\prime } } ^{b_{1},b_{2},b_{3},b_{4 } } \sum_{\beta } \delta _ { \sigma } ^{\uparrow } \bigl [ } -\left ( \delta_{\sigma^{\prime}}^{\uparrow}t_{b_{2},b_{3}}^{\uparrow } + \delta_{\sigma^{\prime}}^{\downarrow}t_{b_{2},b_{3}}^{\downarrow }
\right ) u_{\beta , b_{1}}^{\ast } ( \vec{k}+\vec{q},\uparrow ) u_{\alpha _ { 1},b_{4}}^{{}}(\vec{k}+\vec{q},\uparrow ) \bigr ] \ ; , \nonumber \\
\label{7.14b}\\ c_3&= & - 2\sum_{b_{1},b_{2},b_{3},b_{4 } } u_{\uparrow\downarrow}^{b_{1},b_{2},b_{3},b_{4 } } \left ( n_{\vec{k}+\vec{q}}^{\alpha _ { 1}\uparrow } -n_{\vec{k}}^{\alpha_{2}\downarrow } \right ) u_{\alpha _ { 2},b_{2}}^{\ast } ( \vec{k},\downarrow ) u_{\alpha _ { 1},b_{4}}^{{}}(\vec{k}+\vec{q},\uparrow ) \nonumber \\ & & \hphantom { - 2\sum_{b_{1},b_{2},b_{3},b_{4 } } } \times \frac{1}{l } \sum_{\vec{p};\beta , \beta ^{\prime } } u_{\beta , b_{1}}^{\ast } ( \vec{p}+\vec{q},\uparrow ) u_{\beta ^{\prime } , b_{3}}^{{}}(\vec{p},\downarrow ) \widetilde{g}_{\beta , \beta^{\prime } ; b^{\prime } } ^{\vec{p},\vec{k}^{\prime } , \vec{q}}(e)\ ; .
\label{7.14c } \end{aligned}\ ] ] with the help of the hartree - fock equations ( [ 7.16 ] ) and relation ( [ 7.7 ] ) we may cast the green function ( [ 7.6 ] ) into the form @xmath65^{-1 } \nonumber \\ & & \bigl\{\delta_{\vec{k},\vec{k}^{\prime } } u_{\alpha _ { 2},b^{\prime } } ^{\ast } ( \vec{k},\downarrow ) u_{\alpha _ { 1},b^{\prime } } ^{{}}(\vec{k}+\vec{q},\uparrow ) \left ( n_{\vec{k}+\vec{q}}^{\alpha _ { 1}\uparrow } -n_{\vec{k}}^{\alpha_{2}\downarrow } \right ) \label{7.17}\\ & & \hphantom{\bigl\ { } -2\!\!\sum_{b_{1},b_{2},b_{3},b_{4 } } g_{\text{ret}}^{b_{1}b_{3},b^{\prime}}(\vec{q},e ) u_{\downarrow \uparrow } ^{b_{1},b_{2},b_{3},b_{4 } } \left ( n_{\vec{k}+\vec{q}}^{\alpha _ { 1}\uparrow } -n_{\vec{k}}^{\alpha _ { 2}\downarrow } \right ) u_{\alpha _ { 2},b_{2}}^{\ast } ( \vec{k},\downarrow ) u_{\alpha _ { 1},b_{4}}^{{}}(\vec{k}+\vec{q},\uparrow ) \bigl\}\;. \nonumber \end{aligned}\ ] ] this equation , together with ( [ 7.7 ] ) , leads to @xmath66 + \sum_{b_{3},b_{4},b_{5},b_{6 } } \left[\gamma_{\text{ret}}(\vec{q},e)_{b_{3}b_{4}}^{b_{1}b_{2}}\right ] \widetilde{u}_{b_{5}b_{6}}^{b_{3}b_{4 } } g_{\text{ret}}^{b_{5}b_{6},b^{\prime } } ( \vec{q},e)\ ; , \label{7.18}\ ] ] where [ 7.19 ] @xmath67 here , we added the infinitesimal increment @xmath68 to ensure the properties of a retarded green function @xcite .
when we consider @xmath69 and @xmath70 in ( [ 7.18 ] ) as matrices with respect to the indices @xmath71 , the solution of eq .
( [ 7.7 ] ) is given by @xmath72_{{\frak { d}}^{\prime } } ^{{\frak { d } } } \gamma_{\text{ret}}(\vec{q},e)_{b^{\prime } b^{\prime } } ^{{\frak { d}}^{\prime } } \;.
\label{7.20}\ ] ] this result , together with eq .
( [ 7.5a ] ) , gives us the green function @xmath19 , whose poles at @xmath73 define the rpa spin - wave dispersion .
we will further analyze it in section [ comp ] .
the result ( [ 7.20 ] ) of the rpa decoupling scheme may also be derived from the standard diagrammatic approach . here
, the matrix ( [ 7.19a ] ) is nothing but the retarded part of the respective _ causal _ transversal susceptibility @xmath74 , @xmath75 where the one - particle lines are evaluated within the hartree - fock approximation ; see below . using this expression and the vertex for spin - flip excitations @xmath76 from ( [ 7.19b ] ) , we can calculate the causal two - particle green function @xmath77 as the usual rpa sum of the `` bare bubbles '' in ( [ 7.21 ] ) , [ 7.22 ] @xmath78 this equation has the same solution as ( [ 7.20 ] ) if we replace @xmath79 and @xmath80 by @xmath74 and @xmath81 , respectively
. the retarded green function @xmath80 can then be re - derived from @xmath82 with the help of the standard relations@xcite [ generalgf ] @xmath83 in order to show the equivalence of both approaches , it remains to calculate @xmath74 in ( [ 7.21 ] ) explicitly . to this end
we first determine the hartree - fock green function @xmath84 .
its diagrammatic evaluation leads to @xmath85 where @xmath86 is the causal green function corresponding to the one - particle operator @xmath87 in ( [ 1.1 ] ) , @xmath88_{bb^{\prime}}^{-1 } = e\delta_{b , b^{\prime}}- \varepsilon_{b , b^{\prime}}(\vec{k } ) \ ; .\ ] ] moreover , we introduced the usual hartree - fock self - energy @xmath89 whose elements are calculated self - consistently , @xmath90 compare ( [ tmatrices ] ) . using a matrix notation in ( [ 7.23 ] ) we find @xmath91^{-1 } = \left [ g_{\sigma } ^{0}(\vec{k},e)\right ] ^{-1 } -\sigma_{\sigma }
^{\text{hf}}= e-\widetilde{e}_{\sigma } ( \vec{k } ) \label{7.25 } \ ; , \ ] ] where the matrix @xmath92 is given by @xmath93 from eqs .
( [ 7.25 ] ) and ( [ 7.26 ] ) we see that @xmath94^{-1 } $ ] is diagonalized by the matrix @xmath39 which solves the hartree - fock equations ( [ 7.16 ] ) .
the eigenvalues are given by @xmath95 .
the diagonalization and inversion of ( [ 7.25 ] ) give the final result @xmath96 again , we added a positive or negative increment @xmath97 depending on whether @xmath95 is larger or smaller than the fermi energy .
this ensures the analytical properties of a causal green function .
we insert ( [ 7.27 ] ) into ( [ 7.21 ] ) , and find @xmath98 from this equation one easily re - derives ( [ 7.19a ] ) with the help of the general relations ( [ generalgf ] ) . in this way , we have shown the complete equivalence of the diagrammatic and the equation - of - motion derivation of the multi - band rpa equations .
in ref . we proposed the following gutzwiller wave - function@xcite for a variational examination of the hamiltonian ( [ 1.1 ] ) : @xmath99 here , @xmath100 is any normalized single - particle product - state and the gutzwiller correlator @xmath101 allows for a variational adjustment of the occupation of local atomic multiplets .
our choice for the correlator @xmath101 ensures that the wave - function ( [ gwf ] ) yields the exact ground state of @xmath26 both in the uncorrelated and the atomic limit .
for all other values of correlation parameters the expectation value of @xmath26 in the wave - function can be determined analytically in the limit of large spatial dimensions , @xmath102 . as in our recent work , this limit must be considered as yet another approximation since we apply our general analytical results to real three - dimensional systems . note that @xmath103-corrections are expected to be small @xcite .
once the optimum variational ground state @xmath104 has been found by minimizing the ground - state energy - functional , the following expression for the variational spin - wave dispersion can be evaluated @xmath105 in ref .
a general analytical expression for @xmath106 has been found in the limit of large spatial dimensions .
the numerical evaluation of this result for real materials like iron or nickel is involved but feasible .
so far , we have applied our method only to a degenerate two - band model .
the results for this model will be compared with the rpa in the next subsection . in ref .
we have discussed the variational spin - wave dispersion for a model with two degenerate @xmath107 orbitals ( @xmath108 ) on a simple - cubic lattice . in this system
the general atomic hamiltonian ( [ 1.2 ] ) becomes @xmath109 & & + j\sum_{\sigma } \hat{c}_{1,\sigma } ^{+}\hat{c}_{2,-\sigma } ^{+}\hat{c}_{1,-\sigma } ^{\vphantom{+}}\hat{c}_{2,\sigma } ^{\vphantom{+}}+j_{\text{c}}\bigl(\hat{c}_{1,\uparrow } ^{+}\hat{c}_{1,\downarrow } ^{+}\hat{c}_{2,\downarrow } ^{\vphantom{+}}\hat{c}_{2,\uparrow } ^{\vphantom{+}}+\hat{c}_{2,\uparrow } ^{+}\hat{c}_{2,\downarrow } ^{+}\hat{c}_{1,\downarrow } ^{\vphantom{+}}\hat{c}_{1,\uparrow } ^{\vphantom{+}}\bigr)\;. \nonumber \end{aligned}\ ] ] in cubic symmetry , the coulomb and exchange - integrals @xmath110 , @xmath111 , @xmath112 and @xmath113 are not independent from each other .
instead we have only two free parameters , because the relations @xmath114 and @xmath115 hold . in ref .
we have discussed the appearance of ferromagnetic order in this model in detail both for the hartree - fock and our gutzwiller theory . as is well known for mean - field theories , the stability of a ferromagnetic solution is drastically overestimated in a hartree - fock treatment . in a hartree - fock description
only the stoner parameter @xmath116 governs the magnetic behavior .
the paramagnet becomes unstable if the stoner - criterion @xmath117 is fulfilled , where @xmath118 is the density of states at the fermi level .
this means that a ferromagnetic transition occurs for any value of the local exchange constant @xmath112 .
this is in striking contrast to our variational approach . in the gutzwiller many - body approach
it becomes evident that a sizeable hunds - rule coupling @xmath112 is crucial for the formation of ferromagnetic order .
thus , for small values of @xmath112 , the hartree - fock theory leads to qualitatively incorrect results . the significant differences for the critical values of the coulomb interaction complicate a comparison between the rpa and the variational spin - wave approach .
for the same set of parameters @xmath112 and @xmath110 the underlying ground states are completely different and , therefore , a comparison of the spin - wave properties would not make sense .
thus , it appears to be more reasonable to consider the results of both methods for the same magnetization per band , @xmath119 the density of electrons per band and spin direction in the paramagnetic phase is given by @xmath120 . in fig .
[ fig1 ] the variational spin - wave dispersion in @xmath121-direction is shown for four different magnetizations .
here we used the same tight - binding parameters as in our analysis of the ground - state properties in .
the average electron density per orbital and spin direction in the paramagnet is @xmath122 .
we keep the ratio @xmath123 fixed and consider different values @xmath124 @xmath125 @xmath126 @xmath127 ( bandwidth @xmath128 ) which correspond to a magnetization per band of @xmath129 @xmath130 @xmath131 @xmath132 .
the last value belongs to the almost fully polarized ferromagnet ( @xmath133 ) . as can be seen from fig .
[ fig1 ] , the spin - wave dispersion drastically depends on the magnetization . the data can be fitted very well to @xmath134 in qualitative agreement with experiments on iron - group metals . in the case of a strong ferromagnet
, we obtain @xmath135 and @xmath136 for @xmath137 and @xmath138 , respectively .
this is the right order of magnitude in nickel where @xmath139 .
the inset of fig .
[ fig1 ] shows that the spin - wave dispersion is almost isotropic which also agrees with experimental observations . in our variational
many - body approach , the spin - wave stiffness _
decreases _ as a function of @xmath110 .
this does not come as a surprise because the effective coupling between the sites decreases when the electron hopping between the sites becomes less effective . in fig .
[ fig2 ] the results for the spin - wave dispersion in the rpa are shown for three different values of @xmath110 .
in contrast to our variational approach , the rpa predicts a _ negative _ slope of the dispersion for non - saturated ferromagnetism , @xmath140 .
note that for @xmath141 the system has just attained the fully polarized state ( compare fig .
7 of ref . ) .
but even in this case , a well - defined rpa spin - wave excitation does not yet exist .
it requires a further increase of @xmath110 until the fully polarized collinear hartree - fock ground state appears to be stable . since the rpa probes the ` local ' stability of a mean - field state it is conceivable that another ferromagnetic hartree - fock state with some spiral order has a lower energy than the collinear state
. however , ferromagnetic phases with spiral order are not generic for transition metals .
we argue that this prediction of exotic ferromagnetic spin structures is an artifact of the hartree - fock approach .
in fact , as we have demonstrated for our two - band model @xcite , hartree - fock theory predicts ferromagnetism in regions of the @xmath142 phase diagram where the paramagnetic phase is actually stable . in the case of a stable collinear ferromagnetic hartree - fock ground state ,
the rpa spin - wave stiffness _ increases _ as a function of @xmath110 , again in contrast to our variational approach . for all generic cases
, @xmath143 holds in the limit of infinite coupling . in our case , @xmath144 , as compared to @xmath145 .
the variational and rpa results differ significantly in the strong - coupling limit , too . nevertheless , there are some intermediate coupling strengths @xmath146 , @xmath147 where the variational and the rpa spin - wave stiffness agree .
such accidental agreements may also occur when rpa results are compared to experimental data . in some special cases where the fully polarized state is a one - particle product state we find @xmath148 for all finite values of @xmath110 , @xmath112 , and @xmath149 for large couplings .
this is also the case in the one - band system ; see below .
the study of the one - band model allows us to elucidate the origin of the qualitatively different results for the spin - wave dispersion . in the case of the one - band model ,
@xmath150 the results of sect .
[ rand ] simplify significantly .
the hartree - fock energies can be written as @xmath151 and the susceptibility ( [ 7.19a ] ) becomes @xmath152 for the spin - wave energies @xmath153 the denominator in @xmath154 vanishes , i.e. , @xmath155 to the leading ( second ) order in @xmath14 we can expand ( [ 8.5 ] ) in powers of @xmath156 because @xmath157 is fulfilled for our low - lying spin - wave excitations , see ( [ 1.6 ] ) .
this expansion leads to @xmath158 \ ; .
\label{8.8 } \end{aligned}\ ] ] this result is to be compared with our variational spin - wave dispersion .
however , such a comparison is somewhat questionable , since the existence of a stable ferromagnetic phase is hard to obtain in our variational approach .
nevertheless , let us assume that a given one - band model with a special density of states may have a fully - polarized ferromagnetic ground state .
then we obtain the following expression for the variational spin - wave dispersion @xmath159 this expression differs from @xmath73 in ( [ 8.8 ] ) by the second term , i.e. , for @xmath160 we find @xmath161 the energy difference @xmath162 is always _ negative _ and vanishes for @xmath163 , i.e. , for @xmath164 .
it is essentially this contribution which eventually leads to negative values of the rpa spin - wave stiffness for finite @xmath110 in the two - band model . from a formal point of view
the origin of the differences between both theories is obvious : for a fully polarized ground state of the one - band model the variational dispersion is identical to the _ exact _ first moment of the spectral function @xmath165 .
however , the rpa predicts the existence of spectral weight around @xmath166 , the so - called `` stoner - excitations '' .
thus , the spin - wave pole of the rpa must have a lower energy than @xmath106 , since , otherwise , the rpa could not give the correct first moment of @xmath12 .
for this reason we need to address the issue of whether or not the rpa - theory correctly describes the _ high - energy physics _ of itinerant ferromagnets .
first , it should be noted that the _ existence _ of the energy scale @xmath167 is an artifact of the hartree - fock theory which survives in the rpa . from ground - state concepts like the magnetic condensation energy
we infer that such an energy scale is an artificial feature of mean - field theories , and irrelevant within a many particle description of itinerant ferromagnets @xcite .
second , the misleading _ relevance _ of this energy scale stems from the overestimation of longitudinal fluctuations in mean - field theories . in order to see the second point
, we consider again the spectral function @xmath12 , ( [ swf ] ) , which provides the energy distribution of the spin excitations . in a spin - wave state
the local moments do not depend on momentum , i.e , @xmath168 .
in particular , this implies for the fully polarized one - band model that there are no doubly occupied lattice sites in @xmath169 .
the temporal development of this state is now crucial : in a strongly correlated many - particle system , longitudinal fluctuations are substantially suppressed because charge fluctuations are energetically too costly at large @xmath110 .
therefore , the length of local spins @xmath170 is basically conserved also as a function of time .
our variational method correctly captures this generic behavior : fast longitudinal fluctuations are small and may be neglected at least for small values of @xmath14 . in contrast , longitudinal fluctuations are always present in mean - field theory since electrons with opposite spin meet each other with non - zero probability .
recall that hartree - fock theory predicts a ferromagnetic ground state for much smaller interaction strengths than a proper many - body treatment .
the presence and importance of considerable longitudinal spin fluctuations within hartree - fock and rpa then leads to the prediction of an instability of the collinear ferromagnetic phase ( @xmath171 ) , as seen in fig .
[ fig2 ] .
these findings also explain the qualitatively different behavior of the spin - wave stiffness as a function of @xmath110 . in the variational approach
we find a stable ferromagnetic ground state with well - defined local moments .
the interaction of these ` spins ' is mediated by hopping processes .
similar to the well - known large-@xmath110 antiferromagnetic coupling @xmath172 in the case of the half - filled hubbard model , we find that the effective spin - coupling @xmath173 in ferromagnets is reduced with increasing @xmath110 because hopping processes are more and more suppressed as they induce charge fluctuations . note
that the coupling @xmath173 remains finite in the limit @xmath163 because we are dealing with a non - integer band - filling where electron transfers are never completely suppressed .
in contrast to this generic behavior , the hartree - fock theory misleadingly predicts a ferromagnetic ground state already for rather small interaction strengths where local moments are only partially generated .
the rpa dynamics of the ` hartree - fock spins ' is considerably determined by longitudinal fluctuations . with increasing interaction strength the local moments
are stabilized even within hartree - fock theory .
consequently , the rpa spin - wave stiffness increases and eventually overshoots the value from our variational approach .
the basic mechanism for a reduction of @xmath173 in our variational description remains completely hidden in the rpa because the reduction of the electron transfer amplitudes due to the electrons mutual coulomb interaction is not contained in a hartree - fock mean - field approach .
in this work we have derived the random - phase approximation ( rpa ) for the analysis of spin - wave properties in general multi - band hubbard models with a collinear ferromagnetic ground state . the equation - of - motion technique and the diagrammatic approach lead to identical expressions .
we have applied our analytical results to model systems with one and two correlated orbitals per lattice site .
this allows us to compare the rpa results with those of a recently introduced variational approach to spin - wave excitations of gutzwiller - correlated many - body ground states .
the numerical analysis of the two - band model reveals significant qualitative differences between both methods . in the rpa ,
the spin - wave stiffness is negative for non - saturated ferromagnetism . with increasing correlations the rpa spin - wave energies
are shifted upwards and the rpa spin - wave stiffness eventually becomes positive in the saturated ferromagnet .
our variational correlated - electron approach shows a completely different behavior .
the spin - wave energies are always positive and the spin - wave stiffness decreases with increasing interactions .
the same discrepancies occur in the one - band case where a fully analytical evaluation of the spin - wave dispersions is possible .
we conclude that the hartree - fock / rpa theory provides an inadequate description for spin - wave properties in correlated electron systems . in our view , the main problem is the mean - field character of the underlying hartree - fock theory . in this approach
the ferromagnetic ground state lacks well - developed local moments , and , therefore , high - energy longitudinal fluctuations of the local spins are large and relevant . as a consequence ,
rpa theory predicts an instability of the collinear ferromagnetic ground state against some spiral order .
our many - body approach reveals quite a different picture . in the region of stable ferromagnetism
we find well - defined local moments .
spin excitations in these systems are transversal fluctuations of the local moments whose length is essentially conserved in space and time .
consequently , the stiffness of the variational spin - wave dispersion is always positive and decreases for increasing interaction strength .
the popular `` frozen - magnon '' approximation for the examination of spin - wave excitations in itinerant ferromagnets can be considered as a mean - field theory in which the conservation of the local spins is introduced by hand . in this way ,
the overestimation of longitudinal fluctuations in the hartree - fock / rpa approach is eliminated by construction .
in fact , it can be shown that the frozen - magnon approximation would lead to the same results as our variational approach if we chose a hartree - fock variational wave - function instead of our gutzwiller wave - function
. a more detailed analysis of the frozen - magnon approach will be the subject of a forthcoming study @xcite . | we derive the random - phase approximation for spin excitations in general multi - band hubbard models , starting from a collinear ferromagnetic hartree - fock ground state .
the results are compared with those of a recently introduced variational many - body approach to spin - waves in itinerant ferromagnets .
as we exemplify for hubbard models with one and two bands , the two approaches lead to qualitatively different results .
the discrepancies can be traced back to the fact that the hartree - fock theory fails to describe properly the local moments which naturally arise in a correlated - electron theory . |
the discovery by hawking @xcite of the result that the area @xmath1 of an arbitrary black hole never decreases , together with the idea of the penrose process @xcite of extraction of energy of rotating black holes , led to the establishment of the thermodynamics of the gravitational field .
the analysis in the literature is normally restricted to the geometrical and dynamical behaviour of the area of the external event horizon of rotating black holes . by means of the penrose process
the initial mass @xmath2 and angular momentum @xmath3 of the kerr black hole vary by @xmath4 and @xmath5 , respectively , such that @xmath6 , where @xmath7 is the angular velocity of the external event horizon of the black hole . in this process ,
the variation of the area @xmath1 of the black hole satisfies @xmath8 . in the final stage of an idealized penrose process , the mass of the black hole becomes the irreducible mass @xmath9 @xcite defined by the relation @xmath10 .
let us denote , as usual , the parameter @xmath11 as the angular momentum per unit mass of the kerr black hole , @xmath12 , and @xmath13 and @xmath14 as the gravitational entropy and temperature , respectively .
the ordinary identification of the gravitational thermodynamic quantities with the parameters @xmath15 of the kerr space - time is the following : @xmath16 and @xmath17 ; @xmath18 is the surface gravity defined by @xmath19 , and @xmath20 is the radius of the external event horizon of the black hole .
the ordinary thermodynamic relation for the kerr black hole reads @xmath21 in this paper , we will obtain an expression for the thermodynamic relation @xmath0 in the context of the teleparallel equivalent of general relativity ( tegr ) .
the expression for the gravitational energy @xmath22 is well defined in the tegr .
we will show in the present analysis that the notion of gravitational pressure @xmath23 over the external event horizon of the black hole is also well defined , and naturally arises from the field equations and from the gravitational energy - momentum tensor defined in the realm of the tegr .
the spatial components of the energy - momentum ( or stress - energy ) tensor yield the standard definition of the gravitational pressure , as we will explain .
the variation @xmath24 is obtained by means of the variation of @xmath25 , when the parameters @xmath2 and @xmath11 of the black hole vary by the amounts @xmath4 and @xmath5 ( @xmath23 turns out to be a density , and therefore @xmath24 is actually given by @xmath26 ) .
we will show that our result for @xmath27 is qualitatively and strikingly similar to eq .
we will restrict our analysis to the external event horizon of the kerr black hole ( defined by @xmath28 ) , but will not consider any property of the area @xmath1 ( or @xmath29 ) of the horizon .
the indications are that the efficiency of the penrose process in the present context is lower than in the ordinary thermodynamic formulation .
we mention that the use of the concept of pressure in the first law of black hole thermodynamics has been made in two recent investigations @xcite , by considering that the cosmological constant plays the role of pressure . in section 2 we present the lagrangian formulation of the tegr and indicate how the definitions of the gravitational energy - momentum @xmath30 , of the gravitational energy - momentum tensor @xmath31 and of the gravitational pressure arise from the field equations of the theory . in section 3 we apply the definition of the gravitational pressure to the external event horizon of the kerr black hole . in section 4
we obtain the expression for @xmath32 that emerges in the present framework , and compare it with eq .
finally , in section 6 we present our conclusions .
notation : space - time indices @xmath33 and so(3,1 ) indices @xmath34 run from 0 to 3 .
time and space indices are indicated according to @xmath35 .
the tetrad field is denoted @xmath36 , and the torsion tensor reads @xmath37 .
the flat , minkowski spacetime metric tensor raises and lowers tetrad indices and is fixed by @xmath38 .
the determinant of the tetrad field is represented by @xmath39 .
in this section , we present a brief summary of the formulation of the teleparallel equivalent of general relativity . in the tegr the gravitational field is represented by the tetrad field @xmath36 only , and the lagrangian density is written in terms of the torsion tensor @xmath37 .
this tensor is related to the antisymmetric part of the weitzenbck connection @xmath40
. however , the dynamics of the gravitational field in the tegr is essentially the same as in the usual metric formulation .
the physics in both formulations is the same .
we first consider the torsion - free , levi - civita connection @xmath41 , @xmath42 the christoffel symbols @xmath43 and the levi - civita connection are identically related by @xmath44 in view of this expression , an identity arises between the levi - civita connection and the contorsion tensor @xmath45 , @xmath46 where @xmath47 , and @xmath48 .
the identity given by eq .
( 3 ) may be used to show that the scalar curvature @xmath49 may be identically written as @xmath50 where @xmath51 is the determinant of the tetrad field . therefore in the framework of the tegr the lagrangian density for the gravitational and matter fields reads @xmath52 where @xmath53 , @xmath54 , @xmath55 and @xmath56 @xmath57 stands for the lagrangian density for the matter fields .
the lagrangian density @xmath58 is invariant under the global so(3,1 ) group .
the absence in the lagrangian density of the divergent term on the right hand side of eq .
( 4 ) prevents the invariance of ( 5 ) under arbitrary local so(3,1 ) transformations .
the field equations derived from ( 5 ) are equivalent to einstein s equations .
they are given by @xmath59 where @xmath60 .
from now on we will make @xmath61 .
the definition of the gravitational energy - momentum may be established in the framework of the lagrangian formulation defined by ( 5 ) , according to the procedure of ref .
@xcite . equation ( 7 ) may be rewritten as @xmath62 where @xmath63 and @xmath64 is defined by @xmath65 in view of the antisymmetry property @xmath66 , it follows that @xmath67=0\ , .
\label{10}\ ] ] the equation above yields the continuity ( or balance ) equation , @xmath68\ , .
\label{11}\ ] ] therefore we identify @xmath64 as the gravitational energy - momentum tensor @xcite , @xmath69 as the total energy - momentum contained in a volume @xmath70 of the three - dimensional space , @xmath71 as the gravitational energy - momentum flux @xcite , and @xmath72 as the energy - momentum flux of matter . in view of ( 8) eq .
( 12 ) may be written as @xmath73 where @xmath74 .
the expression above is the definition for the gravitational energy - momentum presented in ref .
@xcite , obtained in the framework of the vacuum field equations in hamiltonian form .
it is invariant under coordinate transformations of the three - dimensional space and under time reparametrizations .
we note that ( 10 ) is a true energy - momentum conservation equation . by substituting eq .
( 12 ) in the left hand side of eq . ( 11 ) , and assuming that the energy - momentum tensor for matter fields @xmath75 vanishes , which is the case for the kerr space - time , we find @xmath76\ , .
\label{16}\ ] ] considering now eq .
( 8) , we rewrite the right hand side of the equation above in the form @xmath77
now we restrict the lorentz index @xmath11 to be @xmath78 , where @xmath79 , and write eq .
( 17 ) as @xmath80 \ , , \label{18}\ ] ] where @xmath81 equation ( 18 ) is precisely eq .
( 39 ) of ref .
@xcite . the left hand side of eq .
( 18 ) represents the momentum of the field divided by time , and therefore has the dimension of force . since on the right hand side
@xmath82 is an element of area , we see that @xmath83 represents the pressure along the @xmath84 direction , over and element of area oriented along the @xmath85 direction . in cartesian
coordinates the index @xmath86 represents the directions @xmath87 respectively . in the next section
we will consider the kerr solution in terms of the boyer - lindquist coordinates , which are spherical type coordinates . in this case
, we have @xmath88 . in order to obtain the radial pressure over the event horizon we need to consider only the index @xmath89 , which represents the radial direction .
thus in spherical type coordinates we define @xmath90 as @xmath91 and from the expression above we define the radial pressure @xmath23 according to @xmath92 in view of eqs .
( 19 ) and ( 20 ) , we see that @xmath93 is a density . in ref .
@xcite we have written eq . ( 18 ) in vectorial form as @xmath94 where @xmath95 is a spherical open surface of constant radius in the schwarzschild space - time with mass @xmath96 , according to eq .
( 44 ) of ref .
@xcite . considering the surface @xmath95 to be @xmath97 , where @xmath98 is a solid angle ,
we have shown that in the limit @xmath99 the equation above is simplified as @xmath100 finally , we note that all definitions presented in this section follow exclusively from the field equations ( 8) .
in terms of the boyer - lindquist coordinates , the kerr solution is given by @xmath101 with the following definitions : @xmath102 the tetrad field is obtained out of the metric tensor according to the relation @xmath103 .
the two properties that may determine the tetrad field are the following .
first , the tetrad field must be adapted to a field of observers whose trajectories and velocities in space - time are given by @xmath104 and @xmath105 , respectively , where @xmath106 is the proper time .
then we identify @xmath107 . this equation fixes three conditions on the tetrad field : @xmath108 . and second , one may choose the unit vectors @xmath109 , @xmath110 and @xmath111 to be oriented asymptotically along the unit cartesian vectors @xmath112 , @xmath113 , @xmath114
. since the tetrad field is a kind of square root of the metric tensor , it may not be defined in every region of the space - time .
for instance , if we choose the observer to be static in space - time , then the tetrad field is defined only in the region @xmath115 , where @xmath116 represents the external boundary of the ergosphere of the kerr space - time .
this result is justified because inside the ergosphere it is not possible to maintain any observer in static regime . inside the ergosphere
, all observers are necessarily dragged in circular motion by the gravitational field .
the four - velocity of observers that circulate around the black hole inside the ergosphere , under the action of the gravitational field of the kerr space - time , is given by @xmath117 where all functions are defined in eq .
it is possible to show that if we restrict the radial coordinate to @xmath28 , the @xmath118 component of eq .
( 24 ) becomes @xmath119 and @xmath7 is the angular velocity of the external event horizon of the kerr space - time .
the quantity @xmath120 is the dragging velocity of inertial frames .
the tetrad field ( i ) that is adapted to observers whose four - velocities are given by eq .
( 24 ) , i.e. , for which @xmath107 , and consequently is defined in the region @xmath121 , ( ii ) whose @xmath122 components in cartesian coordinates are oriented along the unit vectors @xmath112 , @xmath113 , @xmath114 , and ( iii ) that is asympotically flat , is given by @xmath123 where @xmath124 we emphasize that a relevant feature of the tetrad field above is that it is defined from the spatial infinity up to the external event horizon of the kerr black hole .
it is the unique configuration that satisfies the conditions stated above , since six conditions are imposed on @xmath36 .
therefore we may evaluate the gravitational pressure on the external event horizon .
now we have all necessary quantities to evaluate @xmath93 given by eq .
( 20 ) . for this purpose
we need eqs .
( 6 ) , ( 25 ) and ( 26 ) .
the calculations are rather lengthy , but otherwise straightforward .
we find it relevant to show some intermediate steps .
first , we present the non - vanishing contributions to @xmath125 .
they are given by @xmath126 recalling that @xmath51 is the determinant of @xmath36 .
in the expression of @xmath93 we have neglected all terms that are linear in @xmath127 and @xmath128 , since by integration these terms will not contribute to @xmath23 given by eq .
( 21 ) . in order to evaluate the expression above
we need the following relations : @xmath129\ , .
\label{28}\end{aligned}\ ] ] finally , the torsion tensor components @xmath130 are given by @xmath131 after a large number of calculations and simplifications we evaluate expression ( 20 ) , and arrive at @xmath132\ , , \label{30}\ ] ] where @xmath133 .
let us analyse the case @xmath134 and obtain the radial pressure over the event horizon of the schwarzschild black hole . by making @xmath135
we obtain @xmath136 , when @xmath137 .
applying this expression in eq .
( 21 ) yields @xmath138 .
the latter is the gravitational pressure over the event horizon of the schwarzschild black hole .
the negative sign means that the pressure is exerted towards the centre of the black hole . recall that @xmath93 is a density .
therefore it incorporates a quantity of the type @xmath139 . in the limit @xmath140 , eq .
( 27 ) reduces to @xmath141\,,\ ] ] where @xmath142 .
this expression was obtained in ref .
@xcite , in the context of the schwarzschild black hole .
we note , however , that in eqs .
( 43 ) and ( 44 ) for @xmath93 in ref .
@xcite , the quantity @xmath143 , that multiplies the term @xmath144 on the right hand side of the expression above , is missing . the radial pressure over the external event horizon of the kerr black hole is obtained by integration of the angular variables in eq .
( 30 ) , according to eq .
( 21 ) , and making @xmath28 .
the integration is not difficult , and eventually we obtain @xmath145\ , .
\label{31}\ ] ] it is not difficult to see that the value of the pressure is negative .
this is the expression that we will use in the next section in the analysis of the thermodynamics of the kerr black hole .
in this section , we will consider a kerr black hole of mass @xmath2 and angular momentum @xmath3 in stationary state .
then by means of the penrose process we consider that @xmath2 and @xmath3 vary according to @xmath146 and @xmath147 . the standard thermodynamic relation for black holes is given by eq .
we will obtain here the thermodynamic relation @xmath32 entirely within the framework of the tegr , with no identification between @xmath32 and the variation @xmath29 of the area of the event horizon . since the first law of thermodynamics is given by @xmath148 , we need to calculate separately the variations @xmath149 and @xmath150 in the penrose process .
let us begin with @xmath149 . in ref .
@xcite we have calculated the energy contained within the external event horizon of the kerr black hole , making use of the tetrad field given by eqs .
( 25 ) and ( 26 ) .
the latter is precisely equal to the set of tetrad fields given by expression ( 4.9 ) of ref .
@xcite . the resulting value for the gravitational energy contained within a surface of constant radius @xmath28 is given by eq . ( 5.4 ) of the latter reference .
it reads @xmath151\ , , \label{32}\ ] ] where @xmath152 the parameter @xmath23 used above ( and in ref .
@xcite ) is not to be confused with radial pressure @xmath23 .
the remarkable feature of the expression above is that , for any value of the parameter @xmath153 , it is strikingly close to @xmath154 .
we recall that the energy contained within the event horizon of the schwarzschild black hole of mass @xmath2 is given by @xmath155 . for the kerr black hole one
expects that the energy contained within the external event horizon is given by @xmath154 , which is the amount of gravitational energy that can not be extracted from the black hole in the penrose process .
an analysis of the various gravitational energy expressions for the schwarzschild and kerr black holes has been made in ref .
it was concluded that none of the known expressions yield the value @xmath154 for the kerr black hole , but all of them yield @xmath155 for the schwarzschild black hole .
the almost coincidence between our expression and @xmath154 , for any value of @xmath153 , is displayed by fig . 1 of @xcite . rewriting @xmath22 given by eq .
( 32 ) in terms of @xmath15 only we have @xmath156\ , , \label{33}\ ] ] we will obtain @xmath149 and write it in terms of the differentials @xmath4 and @xmath5 , which are the differentials that arise in eq . ( 1 ) .
thus we will calculate @xmath157 recall that @xmath158 .
the calculation is rather lengthy .
we obtain @xmath159 next we will calculate @xmath150 . since @xmath93 is a density ,
the differential @xmath150 is evaluated as @xmath160dr_+ = p\,dr_+\ , , \label{35}\ ] ] where @xmath13 is the surface of constant radius @xmath28 , and @xmath23 is given by eq .
the differential @xmath161 is given by @xmath162 since we are assuming that @xmath161 , @xmath4 and @xmath5 are infinitesimals , the present analysis is not valid when the squere root @xmath163 approaches zero , i.e. , when @xmath11 is very close to @xmath2 . after a number of calculations and simplifications
we obtain @xmath164 we are now in a position to write the first law of thermodynamics @xmath165 out of eqs .
( 34 ) and ( 37 ) .
we will combine the latter equations and write @xmath32 in the form @xmath166 the functions @xmath167 and @xmath168 are obtained by means of straightforward algebra .
they read @xmath169 \ , , \\
f_2(m , a)&=&{{r_+}\over { 2a}}\biggl [ \biggl({3\over { \sqrt{2mr_+ } } } \biggr ) \nonumber \\ & + & \biggl ( { { 3r_+^2 - 11mr_++4m^2}\over{4ma\sqrt{m^2-a^2 } } } \biggr)\ln\biggl({{\sqrt{2mr_+}+a}\over{\sqrt{2mr_+}-a}}\biggr)\biggr]\ , .
\label{39,40}\end{aligned}\ ] ] expressions ( 38 ) , ( 39 ) and ( 40 ) should be compared with the standard expression for @xmath32 given by eq .
( 1 ) , which may be written as @xmath170 where @xmath171 in order to compare eqs .
( 38 - 40 ) and ( 41 - 43 ) , we will plot in the same graph the functions @xmath167 and @xmath172 , and @xmath168 and @xmath173 . for this purpose ,
we choose three arbitrary values of the parameter @xmath153 , where @xmath174 .
we have chosen @xmath175 , @xmath176 and @xmath177 . as we explained earlier ,
the values of @xmath153 close to @xmath178 are not allowed in the analysis , since in this case the differential @xmath161 is no longer an infinitesimal .
all curves in all graphs are functions of the mass parameter @xmath2 . in all figures , the dotted lines represent @xmath172 and @xmath173 , namely , they represent the standard thermodynamic relation given by eq .
the regular ( continuous ) lines represent our result given by eqs .
( 38 - 40 ) .
the main results are that ( i ) the qualitative behaviour of @xmath167 and @xmath168 is the same in all figures , and ( ii ) @xmath167 and @xmath168 are qualitatively similar to @xmath172 and @xmath173 , respectively , in all figures . at first sight
it is surprising that @xmath167 yields a horizontal line , and that @xmath168 looks like @xmath173 .
we will explain this qualitative behaviour in the next section .
our first conclusion is that the thermodynamic relation given by eq . (
38 ) , obtained entirely out of the definitions in the tegr , is realistic and comparable to eq .
( 1 ) . ] ] ] ] ] ] in the standard treatment of the thermodynamics of black holes one , does not make use of the concept of pressure .
we find it appropriate to analyse the thermodynamic relation for @xmath32 , when the latter is due the variation of the energy only .
thus let us analyse @xmath179 in view of eq .
( 34 ) we find @xmath180\ , , \label{45}\end{aligned}\ ] ] and @xmath181\ ,
. \label{46}\ ] ] ] ] ] ] ] ] we need to compare @xmath182 with @xmath172 , and @xmath183 with @xmath173 .
again we chose the values @xmath184 , @xmath185 and @xmath186 for @xmath187 , and therefore all curves will be functions of the parameter @xmath2 .
once again we observe that @xmath182 is qualitatively similar to @xmath172 , and that @xmath183 is qualitatively similar to @xmath173 .
an immediate conclusion about these figures is the following . by comparing the horizontal lines in fig .
3 ( with pressure , @xmath177 ) and fig . 6 ( with no pressure , @xmath188 )
we observe that the numerical , constant value of the straight line is lower in fig . 3 compared to fig .
admitting that the term @xmath189 overweights the effect of the term @xmath190 in the penrose process ( since the particle may have vanishing angular momentum ) , then we see that the effect of the presence of the gravitational pressure is to increase the efficiency of the penrose process .
the reason is that the smaller the value of @xmath149 , more energy can be extracted from the black hole . however , by inspecting figures 1 , 2 and 3 we see that the efficiency of the penrose process is lower in the thermodynamic process dictated by relation ( 38 ) , compared with the standard relation ( 1 ) or ( 41 ) , since after the variations @xmath191 and @xmath192 , the energy variation @xmath149 is higher in the context of the present analysis , namely , in the context of eq . ( 38 ) .
the higher is @xmath149 , less efficient is the penrose process .
in this section , we will explain the reason for the similarity between @xmath167 , @xmath172 and @xmath182 , and between @xmath168 , @xmath173 and @xmath183 .
the idea is very simple .
we just substitute @xmath193 for @xmath11 in all expressions , and factorize the mass parameter @xmath2 . after this procedure
we see that @xmath167 and @xmath182 depend only on @xmath153 , i.e. , they are independent of @xmath2 , and that @xmath168 , @xmath183 and @xmath173 all depend on @xmath194 times a function that depends only on @xmath153 .
we find @xmath195 \biggr\ } \
, , \label{47}\end{aligned}\ ] ] @xmath196 \biggr\ } \ , , \label{48}\end{aligned}\ ] ] @xmath197 \biggr\ } \ , , \label{49}\end{aligned}\ ] ] @xmath198 \biggr\ } \ , , \label{50}\end{aligned}\ ] ] @xmath199 thus we see that @xmath167 and @xmath182 are independent of @xmath2 , and @xmath168 , @xmath183 and @xmath173 depend on @xmath194 .
these features explain the shape of the curves in the graphs .
in this article , we have investigated the thermodynamics of the kerr black hole in the context of the teleparallel equivalent of general relativity .
we have considered : * ( i ) * a stationary black hole with mass and angular momentum per unit mass @xmath200 ; * ( ii ) * the penrose process by means of which @xmath2 and @xmath3 vary according to @xmath201 , @xmath202 ; * ( iii ) * the expressions for gravitational energy and pressure , @xmath22 and @xmath23 respectively , that arise in teleparallel equivalent of general relativity . in particular , we have considered the expression for @xmath22 given by eq .
the numerical values of the latter , for specific values of @xmath2 and @xmath11 , are strikingly close to @xmath154 , as shown in ref .
@xcite . with all these quantities we evaluated @xmath203 .
there are two main results in the article .
first , we have evaluated the gravitational pressure over the external event horizon of the black hole .
the radial pressure , with no dependence on the angular variables , is given by eqs .
( 21 ) and ( 31 ) .
the second result is the thermodynamic relation ( 38 ) , with the definitions ( 39 ) and ( 40 ) .
this relation is obtained without any consideration of the area of the external event horizon of the black hole .
it follows entirely from the definitions that arise in the tegr .
although relations ( 38 - 40 ) look like quite different from the standard thermodynamic relation ( 41 - 43 ) , we see from figures 1 - 3 that both thermodynamic relations are qualitatively similar . note , however , that in the present analysis the variations of the gravitational energy and pressure that occur in view of the penrose process are necessarily subject to the conservation equation ( 11 ) . admitting that the higher is @xmath32 , the lower is the efficiency of the penrose process , then it seems that in the context of the tegr the penrose process is less efficient than in the standard process . finally , by analysing figs .
3 - 6 we see that the presence of the gravitational pressure in the thermodynamic relation for @xmath32 does not imply a significant change .
we believe that both the concept of gravitational pressure and our thermodynamic relation @xmath165 given by eq .
( 38 ) are physically consistent and realistic results . | the concept of gravitational pressure is naturally defined in the context of the teleparallel equivalent of general relativity .
together with the definition of gravitational energy , we investigate the thermodynamics of rotating black holes in the teleparallel framework .
we obtain the value of the gravitational pressure over the external event horizon of the kerr black hole , and write an expression for the thermodynamic relation @xmath0 , where the variations refer to the penrose process for the kerr black hole .
we employ only the notions of gravitational energy and pressure that arise in teleparallel gravity , and do not make any consideration of the area or the variation of the area of the event horizon
. however , our results are qualitatively similar to the standard expression of the literature .
( 1 ) instituto de fsica , universidade de braslia , c.p .
04385 , 70.919 - 970 braslia df , brazil ( 2 ) faculdade unb gama , universidade de braslia , 72.405 - 610 , gama df , brazil pacs numbers : 04.20.-q , 04.20.cv , 04.70.dy |
this paper gives a full description of the semiclassical spectral theory of _ quantum toric integrable systems _ in any finite dimension .
the classical limit corresponding to quantum toric integrable systems are the so called symplectic toric manifolds or _
toric systems_. such a system consists of a compact symplectic @xmath0-manifold equipped with @xmath1 commuting hamiltonians @xmath2 with periodic flows .
the paper combines geometric techniques from the theory of toric manifolds , in the complex - algebraic and symplectic settings , with recently developed microlocal analytic methods for toeplitz operators . as a consequence of the spectral theory we develop , we answer the _ isospectrality _ question for quantum toric integrable systems , in any finite dimension : the semiclassical joint spectrum of a quantum toric integrable system , given by a sequence of commuting toeplitz operators acting on quantum hilbert spaces , determines the classical system given by the symplectic manifold and poisson commuting functions , up to symplectic isomorphisms .
this type of symplectic isospectral problem belongs to the realm of classical questions in inverse spectral theory and microlocal analysis , going back to pioneer works of colin de verdire @xcite and guillemin - sternberg @xcite in the 1970s and 1980s
. the question of isospectrality in riemannian geometry may be traced back to weyl @xcite and is most well known thanks to kac s article @xcite , who himself attributes the question to bochner .
kac popularized the sentence : _ can one hear the shape of a drum ? " _ , to refer to this type of isospectral problem .
the spectral theory developed in this paper exemplifies a striking difference with riemannian geometry , where this type of isospectrality rarely holds true , and suggests that _ symplectic _ invariants are much better encoded in spectral theory than _ riemannian _ invariants .
an approach to this problem for general integrable systems is suggested in the last two authors article @xcite .
we refer to section [ sec : remarks ] for further remarks , and references , in these directions . the notion of a quantum integrable system , as a maximal set of commuting quantum observables , dates back to the early quantum mechanics , to the works of bohr , sommerfeld and einstein @xcite .
however , the most basic results in the symplectic theory of classical integrable systems like darboux s theorem or action - angle variables could not be used in schrdinger s quantum setting at that time because they make use of the analysis of differential ( or pseudodifferential ) operators in phase space , known now as microlocal analysis , which was developed only in the 1960s .
the microlocal analysis of action - angle variables starts with the works of duistermaat @xcite and colin de verdire @xcite , followed by the semiclassical theory by charbonnel @xcite , and more recently by v ngc @xcite , zelditch and toth @xcite , charbonnel and popov @xcite , melin - sjstrand @xcite , and many others .
effective models in quantum mechanics often require a compact phase space , and thus can not be treated using pseudodifferential calculus .
for instance the natural classical limit of a quantum spin is a symplectic sphere .
the study of quantum action - angle variables in the case of compact symplectic manifolds treated in this paper was started by charles @xcite , using the theory of toeplitz operators . in the present paper , we present global spectral results for toric integrable systems ; we use in a fundamental way delzant s theorem on symplectic toric manifolds @xcite .
toric integrable systems have played an influential role in symplectic and complex algebraic geometry , representation theory , and spectral theory since t. delzant classified them in terms of combinatorial information ( actually , in terms of a convex polytope , see theorem [ theo : delzant ] ) .
a comparative study of symplectic toric manifolds from the symplectic and complex algebraic view points was given by duistermaat and pelayo @xcite .
a beautiful treatment of the classical theory of toric systems is given in guillemin s book @xcite .
toeplitz operators are a natural generalization of toeplitz matrices ( which correspond to toeplitz operators on the unit disk ) . on @xmath3 ,
toeplitz operators correspond to pseudodifferential operators via the bargmann transform .
of course , such a correspondence can not hold in the case of a compact phase space , but it turns out that toeplitz operators always give rise to a semiclassical algebra of operators with a symbolic calculus and microlocalization properties , which is microlocally isomorphic to the algebra of pseudodifferential operators .
see the book by boutet de monvel and guillemin for an introduction to the spectral theory of toeplitz operators @xcite .
1em in order to state our results , let us introduce the required terminology .
if @xmath4 is a symplectic manifold , a smooth map @xmath5 is called a _ momentum map _ for a hamiltonian @xmath1-torus action if the hamiltonian flows @xmath6 are periodic of period 1 , and pairwise commute : @xmath7 , so that they define an action of @xmath8 .
if this action is effective and @xmath9 is compact .
@xmath0-dimensional and connected , we call @xmath10 a _ symplectic toric manifold_. by the atiyah and guillemin - sternberg theorem , for any torus hamiltonian action on a connected compact manifold , the image of the momentum map is a rational convex polytope @xcite . for a symplectic toric manifold , the momentum polytope @xmath11 has the additional property that for each vertex @xmath12 of @xmath13 , the primitive normal vectors to the facets meeting at @xmath12 form a basis of the integral lattice @xmath14 .
we call such a polytope a _ delzant polytope_. a now standard procedure introduced by b. kostant @xcite and j.m . souriau @xcite to quantize a symplectic compact manifold @xmath15 is to introduce a prequantum bundle @xmath16 , that is a hermitian line bundle with curvature @xmath17 and a complex structure @xmath18 compatible with @xmath19 .
one then defines the quantum space as the space @xmath20 of holomorphic sections of @xmath21 .
the parameter @xmath22 is a positive integer , the semi - classical limit corresponds to the large @xmath22 limit . a description of this procedure , which is called _ geometric quantization _ ,
is given by kostant and pelayo in @xcite from the angle of lie theory and representation theory .
not all symplectic manifold have a complex structure or a prequantum bundle .
however a symplectic toric manifold always admits a compatible complex structure , which is not unique .
furthermore a symplectic toric manifold @xmath9 with momentum map @xmath23 is prequantizable if and only if there exists @xmath24 such that the vertices of the polytope @xmath25 belong to @xmath26 ( see section [ sec : prequantization ] ) .
if it is the case , the prequantum bundle is unique up to isomorphism . in many papers ,
a prequantum bundle is defined as a line bundle with curvature @xmath27 . with this normalization ,
the cohomology class of @xmath19 is integral and the prequantization condition for toric manifolds is that , up to translation , the momentum polytope has integral vertices .
this normalization may look simpler than ours , which includes a @xmath28-factor .
nevertheless , our choice is justified by the weyl law .
indeed , with our normalization , the dimension of the quantum space @xmath29 is @xmath30 associated to such a quantization there is an algebra @xmath31 of operators @xmath32 called _
toeplitz operators_. this algebra plays the same role as the algebra of semiclassical pseudodifferential operators for a cotangent phase space . here
the semiclassical parameter is @xmath33 .
a toeplitz operator has a _ principal symbol _ , which is a smooth function on the phase space @xmath9 . if @xmath34 and @xmath35 are toeplitz operators , then @xmath36 is a toeplitz operator with the same principal symbol as @xmath34 .
if @xmath37 is hermitian ( i.e. self - adjoint ) for @xmath22 sufficiently large , then the principal symbol of @xmath34 is real - valued .
two toeplitz operators @xmath38 and @xmath39 _ commute _ it @xmath37 and @xmath40 commute for every @xmath22 .
we shall also need the following definitions .
if @xmath41 are mutually commuting endomorphism of a finite dimensional vector space , then the _ joint spectrum of _ @xmath41 is the set of @xmath42 such that there exists a non - zero vector @xmath12 for which @xmath43 , for all @xmath44 .
it is denoted by @xmath45 .
the _ hausdorff distance _ between two subsets @xmath46 and @xmath47 of @xmath48 is @xmath49 where for any subset @xmath50 of @xmath48 , the set @xmath51 is @xmath52 .
if @xmath53 and @xmath54 are sequences of subsets of @xmath55 , we say that @xmath56 if @xmath57 for all @xmath58 .
our main result describes in full the joint spectrum of a quantum toric integrable system .
[ theo : spectral ] let @xmath59 be a symplectic toric manifold equipped with a prequantum bundle @xmath60 and a compatible complex structure @xmath18 .
let @xmath61 be commuting toeplitz operators of @xmath62 whose principal symbols are the components of @xmath63 .
then the joint spectrum of @xmath64 satisfies @xmath65 where @xmath66 , @xmath12 is any vertex of @xmath13 and @xmath67 admits a @xmath68-asymptotic expansion of the form @xmath69 where each @xmath70 is smooth .
thus the joint spectrum of a quantum toric system can obtained by taking the @xmath71 lattice points in a polytope @xmath13 ( as in figure [ fig : regular ] ) , and applying a small smooth deformation @xmath72 ( as in figure [ fig : deformed ] ) .
we present next the isospectral theorem for toric systems .
an easy consequence of the previous theorem is that the momentum polytope @xmath13 is the hausdorff limit of the joint spectrum of the quantum system , that is @xmath13 consists of the @xmath73 such that for any neighborhood @xmath74 of @xmath75 , @xmath76 when @xmath22 is sufficiently large .
goes to @xmath77 .
the spectra lie on a plane , so they corresponds to a four - dimensional integrable system with two degrees of freedom . in the hausdorff limit corresponding to @xmath78
, the spectra converges to a polytope ; this is proved for a general quantum toric system in any dimension by theorem [ theo : inverse - spectral ] .
therefore , one can recover the classical system from the semiclassical spectrum ( i.e. the spectrum of the family of toeplitz operators as @xmath79 approaches @xmath77).,height=234 ] recall that two symplectic toric manifolds @xmath80 and @xmath81 are _ isomorphic _ if there exists a symplectomorphism @xmath82 such that @xmath83 by the delzant classification theorem @xcite , a symplectic toric manifold is determined up to isomorphism by its momentum polytope .
furthemore , for any delzant polytope @xmath13 , delzant constructed in @xcite a symplectic toric manifold @xmath84 with momentum polytope @xmath13 .
now we are ready to state our isospectral theorem ( see figure [ fig : deformed ] for an illustration of the semiclassical joint spectrum ) .
[ theo : inverse - spectral ] let @xmath85 be a symplectic toric manifold equipped with a prequantum bundle @xmath60 and a compatible complex structure @xmath18 .
let @xmath61 be commuting toeplitz operators of @xmath62 whose principal symbols are the components of @xmath63 .
then @xmath86 is a delzant polytope and @xmath10 is isomorphic with @xmath87 . in other words , one can recover the classical system from the limit of the joint spectrum .
this type of inverse conjecture is classical and belongs to the realm of questions in inverse spectral theory , going back to similar questions raised ( and in many cases answered ) by pioneer works of colin verdire and guillemin - sternberg in the 1970s and 1980s .
many other contributions followed their works , for instance datchev
ventura @xcite and iantchenko sjstrand
zworski @xcite .
a few global spectral results have also been obtained recently , for instance by v ngc @xcite for one degree of freedom pseudodifferential operators , or in the article by dryden , guillemin , and sena - dias @xcite in which an equivariant spectrum of the laplace operator is considered , and the references therein .
see section [ sec : remarks ] for further references . introducing a metaplectic correction refers to twisting the prequantum bundle ( or its powers ) by a half - form bundle .
the metaplectic correction allows to obtain an easier control of the subprincipal terms in the semiclassical limit . in the following theorem
we improve the previous joint spectral theorem by giving the explicit description of the spectrum up to a @xmath88 .
recall that a half - form bundle of a complex manifold is a square root of its canonical bundle . given a symplectic manifold @xmath89 with a compatible complex structure , a prequantum bundle @xmath60 and a half - form bundle @xmath90 ,
the associated quantum space is @xmath91 .
we can define toeplitz operators in this setting together with their principal symbols .
to state our result , we also need the notion of subprincipal symbol of a toeplitz operator whose definition is recalled in section [ sec : glob - quant - norm ] . two toeplitz operators with the same principal symbol are equal up to @xmath88 if and only if they have the same subprincipal symbols .
as we will see in section [ sec : half - form ] , a symplectic toric manifold with moment polytope @xmath92 has a half - form bundle if and only if there exists a vector @xmath93 such that for any one - codimensional face @xmath94 of @xmath13 , the scalar product of a primitive normal of @xmath94 with @xmath95 is odd .
such a vector , if it exists , is uniquely determined modulo @xmath96 . we denote it by @xmath97 .
[ theo : metapl - corr ] let @xmath59 be a symplectic toric manifold equipped with a prequantum bundle @xmath60 , a compatible complex structure @xmath18 and a half - form bundle @xmath90 .
let @xmath61 be commuting toeplitz operators of @xmath98 whose principal symbols are the components of @xmath63 .
then the joint spectrum of @xmath64 satisfies @xmath99 where @xmath66 , @xmath12 is any vertex of @xmath13 and @xmath67 admits a @xmath68-asymptotic expansion of the form @xmath69 where each @xmath70 is smooth .
furthermore @xmath100 is determined by @xmath101 where @xmath102 , @xmath103 is the subprincipal symbol of @xmath104 and @xmath105 is the hamiltonian flow of @xmath106 . besides the average of the subprincipal symbols , it is interesting to note the shift by @xmath107 so that no eigenvalue lies on the boudary of @xmath108 when @xmath22 is sufficiently large , cf . figure [ fig : deformed_metaplectic ] for the spectrum of a model toric system with metaplectic correction .
a natural question is how to decide whether a given integrable system can be quantized .
a discussion of this problem may be found in garay van straten @xcite and the references therein ( they work with pseudodifferential operators , instead of toeplitz operators ) . concretely , given a prequantizable symplectic manifold endowed with an integrable system @xmath109 , it may not be possible to find a set of _ commuting _ toeplitz operators whose principal symbols are @xmath110 , respectively .
however , in the case toric systems , we will obtain , as a byproduct of the proof of theorem [ theo : spectral ] , the following existence result .
let @xmath111 be a symplectic toric manifold equipped with a prequantum bundle @xmath60 and a compatible complex structure @xmath18 .
then there exists mutually commuting toeplitz operators @xmath61 in @xmath62 whose principal symbols are the components of @xmath63 .
the proofs in the paper combine geometric ideas from the theory of toric manifolds in the complex and symplectic settings with microlocal analytic methods dealing with semi - classical toeplitz operators that were developed by the first author @xcite .
we review the ingredients from the theory of symplectic toric manifolds which we need for this paper , namely the delzant construction .
-sphere @xmath112 is the height function @xmath113 .
the image of @xmath112 under the momentum map @xmath63 is the closed interval @xmath114 $ ] .
note that as predicted by the atiyah - guillemin - sternberg theorem , the interval @xmath115 $ ] is equal to the image under @xmath63 of the set @xmath116 of fixed points of the hamiltonian @xmath117-action on @xmath112 by rotations about the vertical axis.,width=226 ] let @xmath34 be an @xmath1-dimensional torus .
denote by @xmath118 the lie algebra of @xmath34 and by @xmath119 the kernel of the exponential map @xmath120 .
we denote the isomorphism @xmath121 also by @xmath122 .
a _ symplectic toric manifold _
@xmath123 is a symplectic compact connected manifold @xmath89 of dimension @xmath0 with an effective hamiltonian action of @xmath34 with momentum @xmath63 .
when @xmath124 so that @xmath125 , we recover the definition given in the introduction .
two symplectic toric manifolds @xmath126 and @xmath127 are _ isomorphic _ if there exists a symplectomorphism @xmath128 such that @xmath129 .
if it is case , @xmath130 intertwines the torus actions .
we present the construction of delzant @xcite of symplectic toric manifolds as reduced phase spaces .
let @xmath13 be an @xmath1-dimensional convex polytope in the dual lie algebra @xmath132 .
we denote by @xmath133 and @xmath134 the set of all codimension one faces and vertices of @xmath13 , respectively .
every face of @xmath13 is compact . for every @xmath135
, we write @xmath136 the polytope @xmath13 is called a _ delzant polytope _ if it has the following properties , see guillemin @xcite . * for each @xmath137 there exists @xmath138 and @xmath139 such that the hyperplane which contains @xmath94 is equal to the set of all @xmath140 such that @xmath141 , and @xmath13 is contained in the set of all @xmath140 such that @xmath142 .
+ _ note _ : the vector @xmath143 and constant @xmath144 are made unique by requiring that they are not an integral multiple of another such vector and constant , respectively .
* for every vertex @xmath135 , the vectors @xmath143 with @xmath145 form a @xmath146-basis of the integral lattice @xmath147 in @xmath118 .
it follows that @xmath148 also , @xmath149 for every @xmath135 .
let @xmath152 be defined by @xmath153 because , for any vertex @xmath12 , the @xmath143 with @xmath154 form a @xmath146-basis of @xmath147 , we have @xmath155 and @xmath156 .
it follows that @xmath157 induces an epimorphism @xmath158 and we have the corresponding epimorphism @xmath159 .
write @xmath160 , a linear subspace of @xmath161 , and @xmath162 a compact commutative subgroup of the torus @xmath163 .
one can check that @xmath151 is connected , and therefore isomorphic to @xmath164 , where @xmath165 is the integral lattice in @xmath166 of the torus @xmath151 . on the complex vector space @xmath168
, we have the action of the torus @xmath163 , where @xmath169 maps @xmath170 to the element @xmath171 defined by @xmath172 this action is hamiltonian with momentum @xmath173 given by @xmath174 here @xmath175 , with @xmath176 .
furthermore we work with the symplectic form @xmath177 the factor @xmath178 is introduced in order to avoid an integral lattice @xmath179 instead of our @xmath180 . hence @xmath151 acts on @xmath168
hamiltonianly and the corresponding momentum mapping is @xmath181 where @xmath182 denotes the identity viewed as a linear mapping from @xmath183 to @xmath161 , and its transposed @xmath184 is the map which assigns to each linear form on @xmath161 its restriction to @xmath166 .
it follows from guillemin ( * ? ? ?
* theorems 1.6 and 1.4 ) that @xmath77 is a regular value of @xmath186 , hence the zero level set @xmath187 of @xmath188 is a smooth submanifold of @xmath168 , and that the action of @xmath151 on @xmath187 is proper and free .
the @xmath151-orbit space @xmath189 is a smooth @xmath0-dimensional manifold such that the projection @xmath190 exhibits @xmath187 as a principal @xmath151-bundle over @xmath191 .
moreover , there is a unique symplectic form @xmath192 on @xmath191 such that @xmath193 , where @xmath194 is the identity viewed as a smooth mapping from @xmath187 to @xmath168 . on @xmath195 ,
the torus @xmath196 acts effectively and hamiltonianly , with momentum mapping @xmath197 determined by @xmath198 , and @xmath199 . in other words
, @xmath200 is a symplectic toric manifold with momentum map image equal to @xmath13 .
[ theo : delzant ] any abstract symplectic toric manifold @xmath201 with momentum polytope @xmath202 is isomorphic to @xmath203 .
moreover , two symplectic toric manifolds @xmath204 and @xmath205 are isomorphic if and only if @xmath206 .
since the action of @xmath207 preserves the complex structure of @xmath168 , @xmath208 inherits by reduction a complex structure compatible with @xmath209 and invariant by the action of @xmath34 ( cf .
* theorem 3.5 ) ) .
so @xmath208 is a khler manifold .
recall that a _ prequantum bundle _ on a symplectic manifold @xmath210 is a hermitian line bundle with base @xmath9 endowed with a connection of curvature @xmath17 .
automorphism of prequantum bundle _ is an automorphism of hermitian line bundles preserving the connection .
let @xmath211 be a prequantum bundle over @xmath89 .
consider an action of a lie group @xmath212 on @xmath211 by prequantum bundle automorphisms .
this action lifts an action of @xmath212 on @xmath9 .
one proves that the latter action is hamiltonian and has a natural momentum @xmath63 determined by the following condition : the induced action of the lie algebra @xmath213 on @xmath214 is given by the kostant - souriau operators @xmath215 where we denote by @xmath216 the infinitesimal action of @xmath50 on @xmath9 and by @xmath217 the covariant derivative of the prequantum bundle .
if @xmath212 and @xmath9 are connected , the action on @xmath60 is conversely determined by the action on @xmath9 and the momentum @xmath63 .
furthermore if @xmath212 is a torus @xmath34 acting in a hamiltonian way on a connected @xmath9 , then the momentum associated to a lift to @xmath60 , if it exists , is unique up to a translation by a vector of @xmath218 .
let us consider the case of symplectic toric manifolds . in the following result ,
the _ length of an edge _ @xmath219 of the polytope @xmath202 is the smallest positive real @xmath220 such that @xmath221 .
and @xmath222 equipped with scalar multiples of the fubini - study symplectic form.,width=264 ] [ prop : prequantization ] a symplectic toric manifold @xmath223 with momentum polytope @xmath202 admits a prequantum bundle @xmath224 if and only if the edges of @xmath13 belong to @xmath225 . if it is the case , the prequantum bundle is unique up to isomorphism and the action can be lifted to @xmath226 in such a way that each element acts as a prequantum bundle automorphism .
the moments corresponding to the possible lifts are the ones such that the vertices of the associated polytope belong to @xmath227 .
let us check that the condition is necessary .
let @xmath89 be a symplectic toric manifold with momentum @xmath63 .
let @xmath219 be an edge of @xmath228 .
the length of @xmath219 is the smallest positive real @xmath220 such that @xmath221 .
working with the local symplectic charts associated to the vertices of @xmath219 , one checks that @xmath229 is a symplectic 2-sphere embedded in @xmath9 with volume @xmath230 .
because the image by the map @xmath231 of the chern class of a prequantum bundle of @xmath9 is @xmath232 $ ] , if @xmath9 is prequantizable , then @xmath233 is necessarily integral , so @xmath234 . to conclude ,
observe that @xmath235 can be translated to a polytope with integral vertices if and only if its edges have integral lengths .
the uniqueness of the prequantum bundle follows from the fact that any symplectic toric manifold is simply connected .
conversely assume that @xmath236 is a symplectic toric manifold with a momentum such that the lengths of the edges of the momentum polytope @xmath13 are integral multiple of @xmath237 .
modifying this momentum by an additive constant in @xmath238 if necessary , we may assume that the vertices of @xmath239 belong to @xmath240 .
then the delzant construction provides a prequantum bundle @xmath211 and a lift of the action to @xmath60 .
since we need this construction in the sequel , let us give some details .
consider the prequantum bundle of @xmath241 given by @xmath242 with connection @xmath243 here and in the remainder of the proof , we use the same notations as in section [ sec : model - sympl - toric ] .
in particular the symplectic form of @xmath168 is given by ( [ eq : symp_form_c^f ] ) .
since the vertices of @xmath239 belong to @xmath244 , the @xmath245 s defining the faces of @xmath13 are integral multiple of @xmath28 .
so we can lift the action of @xmath246 on @xmath168 to @xmath247 by @xmath248 with a straightforward computation , one checks that this action preserve the prequantum bundle structure and that its associated momentum is @xmath63 defined in ( [ muf ] ) .
recall that the delzant manifold @xmath249 is the quotient of @xmath250 by the action of subtorus @xmath151 of @xmath246
. then quotienting by @xmath151 the restriction of @xmath247 to @xmath187 , we obtain a prequantum bundle @xmath251 over @xmath208 ( cf .
* theorem 3.2 ) ) .
furthermore the group @xmath252 acts on @xmath251 .
this action preserves the prequantum bundle structure , its associated momentum is the application @xmath253 defined in theorem [ theo : delzant ] .
in this section we introduce a quantum model for quantum toric system and compute its spectrum .
consider a delzant polytope @xmath202 with vertices in @xmath254 .
then as explained in section [ sec : prequantization ] , the delzant manifold @xmath255 admits a prequantum bundle @xmath256 unique up to isomorphism .
this line bundle has a unique holomorphic structure compatible with the complex structure of @xmath257 provided by reduction and with the connection .
so for any positive integer @xmath22 , we can define the quantum space @xmath258 which consists of the holomorphic sections of @xmath259 . for any @xmath260 ,
consider the _ rescaled kostant - souriau operators _ @xmath261 this operator is well - defined because the complex structure is invariant by the action of @xmath34 on @xmath191 .
the rescaling has the effect that the @xmath262 s are self - adjoint and that their joint spectrum is the intersection of the polytope @xmath13 and a rescaled lattice .
the precise result is the following .
[ theo : tt ] there is an orthogonal decomposition of the quantum space @xmath263 into a direct sum of lines : @xmath264 such that , for any @xmath265 , @xmath266 the proof uses the delzant construction recalled in section [ sec : model - sympl - toric ] at the quantum level .
consider the same prequantum bundle @xmath267 as in the proof of proposition [ prop : prequantization ] .
the associated quantum space @xmath268 is the space of holomorphic sections @xmath269 of @xmath270 such that @xmath271 where @xmath272 is the liouville measure .
here holomorphic means that the covariant derivative with respect to antiholomorphic vectors vanishes .
such a holomorphic section can be written @xmath273 where @xmath94 is a plain holomorphic function on @xmath274 . here
so @xmath268 can be identified with the usual bargmann space , that is the space of holomorphic functions on a @xmath276 whose square is integrable with respect to a given gaussian weight .
an orthogonal basis of @xmath268 is given by the family @xmath277 consider the kostant - souriau operator associated to momentum @xmath63 given in ( [ muf ] ) @xmath278 if @xmath279 a straightforward computation shows that @xmath280 we deduce that for any @xmath281 and @xmath282 , @xmath283 recall that the delzant space was defined as the symplectic quotient of @xmath241 by the subtorus @xmath151 of @xmath246 .
the corresponding space at the quantum level is @xmath284 we call it the _ reduced quantum space_. we deduce from equation ( [ eq : tx ] ) that a basis of @xmath285 consists of the @xmath286 s such that @xmath287 satisfies @xmath288 for all @xmath289 .
equivalently , @xmath290 runs over @xmath291 .
this set is in bijection with @xmath292 .
[ lemm : bij ] the map @xmath293 from @xmath294 to @xmath295 restricts to a bijection @xmath296 furthermore , for any @xmath297 and @xmath298 we have that @xmath299 we know that @xmath300 is injective with image @xmath301 . using the delzant condition on a vertex ( step 1 in section [ sec : model - sympl - toric ] ) , one sees that @xmath300 restricts to a bijection from @xmath302 to @xmath303 . by the prequantization condition , @xmath304 .
so @xmath293 restricts to a bijection from @xmath305 to @xmath306 .
furthermore , the proof to show that @xmath307 implies that @xmath308 this implies the first part of the lemma .
the second assertion follows from ( [ eq : tx ] ) .
the end of the proof of theorem [ theo : tt ] is an application of the `` quantization commutes with reduction '' theorem of guillemin - sternberg .
lifting the action of @xmath309 to @xmath247 as in the proof of proposition [ prop : prequantization ] , we get an action of @xmath310 on @xmath268 .
the reduced quantum space is the subspace of @xmath268 of @xmath151-invariant vectors , in other words @xmath311 .
now by guillemin - sternberg theorem ( @xcite ) , we have an isomorphism @xmath312 the proof in @xcite given in the compact case extends to our setting by @xcite and @xcite . furthermore
, under the isomorphism @xmath313 , the action of the torus @xmath34 on @xmath314 corresponds to the action of @xmath315 on @xmath285 .
then , passing to the level of lie algebras , we get the following relation between the kostant - souriau operators : @xmath316 this concludes the proof of theorem [ theo : tt ] . from the previous theorem
, we deduce the following quantum normal form . consider a delzant polytope @xmath317 in the lie algebra @xmath48 of @xmath318 .
assume @xmath319 has a prequantum bundle @xmath320 and define the associated quantum spaces @xmath321 .
starting from the canonical basis @xmath322 of @xmath48 , we get @xmath1 operators @xmath323 defined by kostant - souriau formula ( [ eq : txmux ] ) .
[ cor : quantum - model ] for any @xmath22 , @xmath324 are mutually commuting operators with simple joint eigenspaces .
their joint spectrum is @xmath325 where @xmath12 is any vertex of @xmath317
. apply theorem [ theo : tt ] to the polytope @xmath326 , whose vertices are integral .
we briefly review toeplitz operators .
let @xmath4 be a compact connected symplectic manifold with a prequantum line bundle @xmath211 .
assume that @xmath9 is endowed with a complex stucture @xmath18 compatible with @xmath19 , so that @xmath9 is khler and @xmath211 is holomorphic . here
the holomorphic structure of the prequantum bundle is the unique one compatible with the connection .
recall that for a positive integer @xmath22 , @xmath20 is the space of holomorphic sections of @xmath21 . since @xmath9 is compact , @xmath327 is a closed finite dimensional subspace of the hilbert space @xmath328 . here the scalar product is defined by integrating the hermitian pointwise scalar product of sections agains the liouville measure of @xmath9 .
denote by @xmath329 the orthogonal projector of @xmath330 onto @xmath331 .
a _ toeplitz operator _ is any sequence @xmath332 of operators of the form @xmath333 where @xmath334 , viewed as a multiplication operator , is a sequence in @xmath335 with an asymptotic expansion @xmath336 for the @xmath68 topology , and the norm of @xmath337 is @xmath338 .
we denote by @xmath339 the set of toeplitz operators .
the following result corresponds to ( * ? ? ?
* theorem 1.2 ) . [ sc_algebra ]
the set @xmath340 is a semi - classical algebra associated to @xmath89 in the following sense .
the set @xmath341 is closed under the formation of product .
so it is a star algebra , the identity is @xmath342 .
the symbol map @xmath343,$ ] sending @xmath37 into the formal series @xmath344 where the functions @xmath345 are the coefficients of the asymptotic expansion of the multiplicator @xmath346 , is well defined .
it is onto and its kernel is the ideal consisting of @xmath338 toeplitz operators .
more precisely for any integer @xmath220 , @xmath347 furthermore , the induced product @xmath348 on @xmath349 $ ] is a star - product .
we call the formal series @xmath350 the _ contravariant symbol of _ @xmath351 .
the first coefficient @xmath352 is the _ principal symbol _ of @xmath351 .
the _ subprincipal symbol _ of @xmath353 is the function @xmath354 where @xmath13 is the holomorphic laplacian of @xmath9 .
consider two toeplitz operators with principal and subprincipal symbols @xmath355 and
@xmath356 , @xmath357 respectively
. then the principal and subprincipal symbol of their composition is @xmath358 where @xmath359 is the poisson bracket of @xmath89 ( cf .
theorem 1.4 of @xcite ) .
the kostant - souriau operators considered in section [ sec : quantum - model ] are toeplitz operators
. more generally , let @xmath94 be a function of @xmath9 with hamiltonian vector field @xmath50 . applying the tuynman s trick ( @xcite ) ,
one proves that the sequence @xmath360 is a toeplitz operator with principal symbol @xmath94 and subprincipal symbol @xmath361 .
recall that for each delzant polytope @xmath362 , we introduced in section [ sec : model - sympl - toric ] a symplectic toric manifold @xmath363 , a complex structure @xmath364 on @xmath257 compatible with @xmath365 .
assume that @xmath366 has integral vertices for some @xmath367 , so that @xmath319 has a prequantum bundle @xmath320 ( unique up to isomorphism ) .
we defined in section [ sec : quantum - model ] for any positive @xmath22 , commuting operators @xmath368 acting on the hilbert spaces @xmath369 , and described explicitly their spectrum in corollary [ cor : quantum - model ] .
[ theo : normal - form ] let @xmath370 be a symplectic toric manifold equipped with a prequantum bundle @xmath60 and a compatible complex structure @xmath18 . denote by @xmath13 the momentum polytope @xmath371 .
let @xmath61 be commuting toeplitz operators of @xmath62 whose principal symbols are the components of @xmath63 .
then there exists @xmath372 , there exists a sequence @xmath373 of smooth maps @xmath374 , and there exists an operator @xmath375 with @xmath376 invertible for any @xmath22 , such that @xmath377 moreover , @xmath72 admits an asymptotic expansion in the @xmath68 topology of the form @xmath378 if the operators @xmath379 are self - adjoint ( i.e. for any @xmath22 , @xmath380 is a self - adjoint operator ) , then @xmath376 may be chosen such that @xmath381 . for small @xmath22 ,
the dimensions of @xmath327 and @xmath382 might be different .
theorem [ theo : normal - form ] does not give information about small values of @xmath22 .
the proof of theorem [ theo : normal - form ] will require the following technical lemma , which is a global version of a result of eliasson ( * ? ? ?
* corollary page 14 ) . recall that in the case where @xmath383 is a closed half - space , @xmath384 if and only if all the partial derivatives @xmath385 for @xmath386 have a limit at every point in @xmath387 .
( this is equivalent to saying that @xmath72 has a smooth extension in a neighborhood of any point . )
[ lemma : smooth ] let @xmath383 be a vector space or a closed half - space .
let @xmath388 , and let @xmath389 assume that @xmath390 ( here the poisson bracket refers to the symplectic variables @xmath391 ) .
then there exists @xmath392 such that @xmath393 set - theoretically , there is a unique such function @xmath72 .
we need to prove that @xmath72 is smooth .
the taylor expansion of @xmath94 in the @xmath391 variables has to commute with @xmath394 .
this implies that it has the form @xmath395 where @xmath396 .
hence by the taylor formula , for any integer @xmath397 , there is a polynomial @xmath398 in @xmath394 with coefficients in @xmath399 , and a smooth function @xmath400 such that @xmath401 thus we get @xmath402 when @xmath403 , we simply compute the partial derivatives @xmath404 .
they have a limit as @xmath405 as long as @xmath406 .
thus @xmath407 , which proves the lemma .
we divide the proof into several steps .
[ [ step-1 . ] ] * step 1 .
* + + + + + + + + + by theorem [ theo : delzant ] , there exists a symplectomorphism @xmath408 such that @xmath409 .
since @xmath208 is simply connected , the prequantum bundle @xmath410 is unique up to isomorphism .
hence @xmath411 can be lifted to a prequantum bundle isomorphism @xmath412 .
so @xmath411 can be quantized as an operator @xmath413 such that @xmath414 for large @xmath22 and such that for any toeplitz operator @xmath415 with principal symbol @xmath416 , @xmath417 is a toeplitz operator whose principal symbol is @xmath418 .
the operators @xmath376 that we use here have been introduced in ( * ? ?
* chapter 4 ) and similar ones have been considered by @xcite .
they are analogues of fourier integral operators @xcite . replacing @xmath379 by @xmath419
we see that the problem is reduced to the case where @xmath61 are commuting toeplitz operators on @xmath420 , with joint principal symbol equal to @xmath253 .
[ [ step-2 . ] ] * step 2 .
* + + + + + + + + + we now prove the theorem by induction .
assume that , for some @xmath421 , we have @xmath422 where @xmath423 is a vector of @xmath1 toeplitz operators and @xmath424 is polynomial in @xmath22 : @xmath425 and each @xmath70 is a smooth map . for simplicity
we write @xmath426 notice that , for @xmath427 , this is precisely the result of step 1 .
we wish to prove that there exists an invertible toeplitz operator @xmath428 such that @xmath429 the procedure is standard and we only indicate the key points .
it turns out that the case @xmath427 is slightly different from the other cases @xmath430 .
when @xmath427 , we plug in the left - hand side of and multiply on the right by @xmath74 , and obtain @xmath431 + k^{-1}ur_1 = k^{-1}g_1u \mod k^{-2}\mathscr{t}.\ ] ] since both sides of the equation are toeplitz operators of order 1 , the equation is equivalent to the equality of the principal symbols : @xmath432 writing @xmath95 of the form @xmath433 we get the equation @xmath434 for @xmath435 we look for @xmath74 in the form @xmath436 , where @xmath437 is a toeplitz operator . the same calculation as before gives the equation @xmath438 + k^{-(n+1)}(r_{n+1 } - k_{n+1 } ) = 0 \mod k^{-(n+2 ) } \mathscr{t}.\ ] ] ( we use here @xmath439 in order to eliminate higher order terms . ) since @xmath440 the equation is equivalent to the following equation on the principal symbols : @xmath441 [ [ step-3 . ] ] * step 3 . * + + + + + + + + + in order to complete the induction , we need to solve the following cohomological equation , where the unknown functions are @xmath442 and @xmath443 , @xmath444 : @xmath445 the proof follows eliasson s local argument in ( * ? ? ?
* lemma 8) , where he uses a formula due to moser .
here we show that this local argument also works globally . for any smooth function @xmath446 on @xmath208
, we define @xmath447 @xmath448 and @xmath449 are clearly linear operators sending @xmath450 into itself .
notice that , since the flows @xmath451 pairwise commute , the fubini formula ensures that @xmath448 and @xmath452 commute for any @xmath453 .
the following poisson bracket is easy to compute : @xmath454 we shall need the following lemmas .
[ lemma : m ] let @xmath455 be smooth functions on @xmath208 such that for all @xmath456 , @xmath18 , @xmath457 then for all @xmath458 , we have that @xmath459 .
[ lemma : g ] let @xmath460 such that for all @xmath461 we have that @xmath462 .
then there exists @xmath463 such that @xmath464 we have that @xmath465 as desired . since @xmath94 is invariant by the action , set - theoretically , there exists unique function @xmath72 such that @xmath466 .
we want to prove that @xmath72 is smooth . at a regular value of @xmath253
, this follows directly from the action - angle theorem .
let @xmath467 be a critical value of @xmath253 , and let @xmath468 be a small ball around @xmath467 .
let @xmath469 , with @xmath470 be delzant coordinates on @xmath471 .
up to an affine transformation , we can assume @xmath472 by assumption , @xmath94 does not depend on the @xmath473 coordinates , so there is a smooth function @xmath474 such that @xmath475 .
we apply lemma [ lemma : smooth ] to the function @xmath474 with @xmath476 and @xmath477 .
thus there is a smooth function @xmath100 such that @xmath478 we may now apply the same lemma to @xmath479 with @xmath480 and get a smooth function @xmath481 such that @xmath482 we may repeat the argument and finally obtain a smooth function @xmath483 such that @xmath484 this proves that @xmath72 is smooth in @xmath468 . thus @xmath72 is smooth on @xmath485 ( which means that there is a smooth extension of @xmath72 in @xmath3 ) .
we return now to the cohomological equation . by lemmas [ lemma : m ] and
[ lemma : g ] , there exist smooth functions @xmath443 on @xmath3 such that @xmath486 . let @xmath487 notice that for any function @xmath488 , @xmath489 .
hence , since the operators @xmath448 and @xmath452 commute , we get @xmath490 so @xmath442 solves the first equation of .
let @xmath491 we have @xmath492 , and our system becomes @xmath493 with @xmath494 we notice that @xmath495 so by induction @xmath496 solves the complete system .
thus @xmath442 solves .
the construction we have used to solve do not require the functions @xmath497 to be real - valued . in case
they are real - valued , then @xmath442 and @xmath443 will be real - valued as well , and in step 2 we may choose @xmath498 , which is unitary . [ [ step-4 . ] ] * step 4 . * + + + + + + + + + from steps 2 and 3 we obtain , for any positive integer @xmath151 , an invertible operator @xmath499 ( which is unitary in the case of self - adjoint operators @xmath379 ) and a smooth map @xmath500 such that @xmath501 where @xmath423 is a toeplitz operator , and @xmath502 is of the form @xmath503 from step 3 , we have @xmath504 for @xmath505 .
therefore , one can construct by induction a sequence of symbols @xmath506 such that for all @xmath435 , the operator @xmath507 is , modulo @xmath508 , the toeplitz quantization of the symbol @xmath509 by the borel summation procedure , one can find a toeplitz operator @xmath510 whose total symbol has the asymptotic expansion @xmath511 moreover , one can find a smooth map @xmath512 that admits the asymptotic expansion @xmath513 now we let @xmath514 , so that for any @xmath151 , @xmath515 thus from we get , as required : @xmath516 in the case where the operators @xmath379 are self - adjoint , one can change the construction of the sequence @xmath517 in such a way that @xmath507 is , modulo @xmath508 , the _ exponential _ of the toeplitz quantization of the symbol @xmath518
. then we define @xmath519 which is unitary .
in this section we prove theorem [ theo : spectral ] and corollary [ theo : inverse - spectral ] .
recall that the _ joint spectrum of _ of @xmath1 commuting matrix @xmath520 is the set of @xmath42 such that there exists a non - zero vector @xmath12 for which @xmath521 for all @xmath44 .
such an @xmath1-uple @xmath522 will be called a _
joint eigenvalue_. we begin with the following elementary observations .
[ lem : variational1]let @xmath523 be commuting self - adjoint matrices .
let @xmath524 , and let @xmath525 be such that @xmath526 for all @xmath461 .
then there exists @xmath527 such that @xmath528^n.\ ] ] this lemma follows from the usual variational characterization of the largest eigenvalue a matrix , applied to the self - adjoint operator @xmath529 .
it has the following immediate consequences : [ lemm : variational2 ] the following statements hold . 1 .
if @xmath523 are commuting self - adjoint matrices , and @xmath530 is such that @xmath531 for all @xmath532 , then there exists a joint eigenvalue @xmath533 such that @xmath534 for all @xmath461 .
2 . suppose that @xmath535 is another collection of commuting self - adjoint matrices , and assume @xmath536 then the hausdorff distance between @xmath537 and @xmath538 is at most @xmath539 , i.e. @xmath540 the first statement is obtained from lemma [ lem : variational1 ] applied to @xmath541 .
it implies that if @xmath542 is a joint eigenvalue of @xmath543 and @xmath544 for all @xmath461 , then there exists a joint eigenvalue @xmath545 with @xmath546 for all @xmath461 ( and vice - versa ) , which gives the last statement . if @xmath547 are pairwise commuting toeplitz operators , we call _ joint spectrum of @xmath548 _ the sequence of joint spectra of the set of commuting matrices @xmath549 acting on the hilbert space @xmath327 . by theorem [ theo : normal - form ]
applied to @xmath550 , we get an integer @xmath372 , a sequence @xmath373 of smooth maps @xmath551 , and an operator @xmath375 with @xmath376 unitary , such that @xmath552 let @xmath553 .
if we introduce the components of @xmath72 , @xmath554 then by corollary [ cor : quantum - model ] , the joint spectrum of the commuting toeplitz operators @xmath555 is @xmath556 .
equation means that there exists a sequence @xmath557 of real numbers such that @xmath558 then , by virtue of lemma [ lemm : variational2 ] , we have @xmath559 which mean by definition @xmath560 as we wanted to show . recall that we defined the limit of a sequence @xmath561 of subsets of @xmath3 by @xmath562
we denote by @xmath563 the open ball in @xmath3 centered at @xmath467 and of radius @xmath446 . notice that if @xmath564 , for some constant @xmath468 , then @xmath565 .
we use the same notation as in the proof of theorem [ theo : spectral ] : @xmath566 and @xmath567 [ [ step-1.-1 ] ] * step 1 . * + + + + + + + + + the estimate for @xmath568 gives @xmath569 therefore , @xmath570 . [
[ step-2.-1 ] ] * step 2 .
* + + + + + + + + + now let us show that @xmath571 . from theorem [ theo : normal - form ] , we know that @xmath72 admits an asymptotic expansion in the @xmath68 topology of the form @xmath378 therefore , for any compact @xmath572 , there exists a constant @xmath468 such that @xmath573 we may choose @xmath574 large enough so that it contains @xmath13 , and we get the following inclusions : @xmath575 hence @xmath576 from steps 1 and 2 we conclude that @xmath577 , which finishes the proof .
two symplectic toric systems are isomorphic if and only if the limit of their joint spectra coincide .
in particular , if two symplectic toric systems have the same joint spectra , then they are isomorphic .
theorem [ theo : spectral ] and theorem [ theo : inverse - spectral ] can be proven analogously to our proofs in the presence of a half form bundle .
below we explain the corresponding modifications needed .
toeplitz quantization is often considered more natural in presence of a half - form bundle .
let @xmath578 be a compact khler manifold .
a half - form bundle of @xmath579 is a square root of the canonical bundle of @xmath9 .
more precisely we consider a pair @xmath580 consisting of a complex line bundle @xmath581 and an isomorphism @xmath582 here @xmath1 is the complex dimension of @xmath9 .
such a square root does not necessarily exists , and if it exists , the space of half - form bundles up to isomorphism is a principal homogeneous space for the group @xmath583 .
let @xmath16 be a prequantum bundle and @xmath580 be a half - form bundle .
observe that @xmath90 has a metric and a holomorphic structure determined by the condition that @xmath130 is an isomorphism of hermitian holomorphic bundles .
define the quantum space @xmath584 as the vector space of holomorphic sections of @xmath585 .
the space @xmath584 has a natural scalar product obtained by integrating the point wise scalar product of sections of @xmath586 against the liouville measure .
this scalar product is actually defined on the space of @xmath587 sections , and we have an orthogonal projector @xmath329 from the space of @xmath587 sections onto @xmath584 .
the definition of toeplitz operators is the same as before except that we use this new projector .
so a toeplitz operator is any family @xmath588 of endomorphisms of the form @xmath589 where @xmath334 , viewed as a multiplication operator , is a sequence in @xmath335 with an asymptotic expansion @xmath590 for the @xmath68 topology , and the norm of @xmath337 is @xmath338 .
theorem [ sc_algebra ] still holds and we define the principal and subprincipal symbols of a toeplitz operators with the same formula as before .
the rule of composition of these symbols is still given by ( [ eq : comp_symbol ] ) .
we can also define toeplitz operators by using the kostant - souriau formula .
consider a smooth function @xmath94 on @xmath9 whose hamiltonian vector field @xmath50 preserves the complex structure .
then the following operators are well defined @xmath591 here @xmath592 is the lie derivative of half - form , that is the first order differential operator such that for any local section @xmath416 of the half - form bundle one has @xmath593 one also shows that the family @xmath594 is a toeplitz operator with principal symbol @xmath94 and vanishing subprincipal symbol .
consider now the delzant space @xmath191 defined as in section [ sec : model - sympl - toric ] .
since @xmath208 is simply connected , there exists at most one half - form bundle .
recall that we denote by @xmath133 the set of faces of codimension 1 and for each @xmath595 , @xmath596 is the primitive normal vector to the face .
[ prop : existence - half - form ] the delzant space @xmath191 admits an equivariant half - form bundle if and only if there exists @xmath597 such that @xmath598 is odd for any @xmath595 . this criterion may be established using the divisor of the toric variety , cf . @xcite . to each face
@xmath595 corresponds an irreducible divisor @xmath599 of @xmath191 .
it is known that a divisor of the canonical bundle is @xmath600 .
recall also that the divisors @xmath599 s generate the picard group , and that @xmath601 is principal if and only if @xmath602 for some @xmath603 .
so @xmath191 admits a half - form bundle if and only if there exists a divisor @xmath604 such that @xmath605 is principal , that is @xmath606 for some @xmath597 .
assume now that the vertices of @xmath13 belongs to @xmath607 so that @xmath608 admits a prequantum bundle @xmath256 unique up to isomorphism .
assume also that @xmath208 is equipped with a half - form bundle @xmath609 .
for any positive integer @xmath22 , define the quantum space @xmath610 for any @xmath260 , consider the rescaled kostant - souriau operators @xmath611 we can now state the analogue of theorem [ theo : tt ] . introduce @xmath612 such that @xmath613 is odd for any @xmath137 .
[ theo : ttm ] there is an orthogonal decomposition of the quantum space @xmath614 into a direct sum of lines : @xmath615 such that , for any @xmath265 , @xmath266 observe that the points of @xmath616 are all in the interior of @xmath13 . furthermore on a neighborhood of each vertex we recover the usual joint spectrum of @xmath1 harmonic oscillators as follows . for
any vertex @xmath12 denote by @xmath617 the set of one - codimensional faces adjacent to @xmath12 so that @xmath618 is a basis of @xmath619 .
then there exists a neighborhood @xmath74 of @xmath12 such that @xmath620 and @xmath621 let us adapt the proof of theorem [ theo : tt ] .
first we introduce the quantization of @xmath168 with metaplectic correction .
choose an ordering of @xmath133 and define @xmath622 as the wedge product of the @xmath623 s .
@xmath12 is a non vanishing section of the canonical bundle of @xmath168 .
the action of @xmath624 on the canonical bundle sends @xmath625 into @xmath626 .
let @xmath627 be the trivial complex line bundle with base @xmath168 and @xmath628 be the isomorphism from @xmath629 to the canonical bundle of @xmath167 given by @xmath630 .
by proposition [ prop : existence - half - form ] , there exist @xmath631 and @xmath632 such that @xmath633 where @xmath634 is the vector in @xmath295 with all components equal to @xmath635 .
consider the action of @xmath636 on @xmath637 given by @xmath638 then the isomorphism @xmath130 intertwines the action of the subtorus @xmath151 of @xmath636 on @xmath637 with the action of @xmath151 on the canonical bundle .
this condition has the consequence that the quotient of @xmath629 by @xmath151 defines a half - form bundle @xmath639 on the delzant space @xmath191 , cf .
section 8.3 of @xcite .
furthermore one has an isomorphism @xmath640 from the @xmath151-invariant part of the space @xmath641 of holomorphic sections of @xmath642 to the quantum space @xmath643 .
introduce the rescaled kostant - souriau operators @xmath644 then @xmath645 consists of the sections @xmath646 such that @xmath647 for any @xmath648 .
furthermore @xmath649 to conclude the proof let us compute the action of the rescaled kostant - souriau operators .
we have for @xmath279 @xmath650 for any @xmath651 set @xmath652 so that @xmath653 hence the space @xmath654 admits as a basis the family @xmath655 where @xmath656 runs over @xmath657 . to conclude we prove as in lemma [ lemm : bij ] that @xmath658 restricts to a bijection @xmath659 with metaplectic correction.,height=94 ] let us consider now the generalization of theorem [ theo : normal - form ] .
let @xmath370 be a symplectic toric manifold equipped with a prequantum bundle @xmath60 , a compatible complex structure @xmath18 and a half - form bundle @xmath90 . denote by @xmath13 the momentum polytope @xmath371 .
let @xmath61 be commuting self - adjoint toeplitz operators of @xmath584 whose principal symbols are the components of @xmath63 .
denote by @xmath103 the subprincipal symbol of @xmath660 .
there exists @xmath372 , there exists a sequence @xmath373 of smooth maps @xmath374 , and there exists an operator @xmath375 with @xmath376 unitary for any @xmath22 , such that @xmath377 moreover , @xmath72 admits an asymptotic expansion in the @xmath68 topology of the form @xmath661 where the subprincipal term @xmath100 is given by @xmath662 here @xmath663 is the hamiltonian flow of @xmath664 .
the proof is the same as the one of theorem [ theo : normal - form ] .
because of the metaplectic correction , we can choose in the first step the operator @xmath376 quantizing @xmath130 in such a way that for any toeplitz operator @xmath665 , @xmath665 and @xmath666 have the same principal and subprincipal symbols .
this follows from theorem 5.1 in @xcite . from this
we can extract @xmath100 from the cohomological equation @xmath667 .
finally we deduce theorem [ theo : metapl - corr ] in the introduction by following the same method as in section [ sec : inverse ] .
in the present paper we have dealt with isospectrality in the context of integrable systems and symplectic geometry .
the paper settles the _
spectral goal for quantum systems _ for the case of toric systems outlined in the last two authors article @xcite : to prove that large classes of integrable systems are determined by their semiclassical joint spectrum .
this type of inverse question fits in the framework of isospectral questions " : what is the relation between two operators that have the same spectrum ?
the question of isospectrality has been considered by many authors in different contexts , and may be traced back to a more general question of by h. weyl @xcite .
the question is perhaps most famous thanks to kac s article @xcite ( who attributes the question to s. bochner ) , which also popularized the phrase : can one hear the shape of a drum ? " .
corollary [ theo : inverse - spectral ] says that the joint spectrum does indeed determine the system .
this type of conclusion often has a negative answer , at least if one considers it in riemannian geometry . in riemannian geometry
the operator whose spectrum is considered is the laplace operator .
two compact riemannian manifolds are said to be _ isospectral _ if the associated laplace operators have the same spectrum .
bochner and kac s question has a negative answer in this case , even for planar domains with dirichlet boundary conditions ( which is the original version posed in @xcite ) .
there are many works in this direction , see for instance milnor @xcite , sunada @xcite , berard @xcite , and buser @xcite and gordon - webb - wolpert @xcite .
as we have mentioned , in symplectic geometry a few positive results are known .
these results , and the present paper , give evidence that symplectic invariants seem to be more encodable in the spectrum of a quantum integrable system than riemannian invariants in the spectrum of the lapace operator .
inverse type results in the realm of spectral geometry have been obtained by many other authors , see for instance brning - heintze @xcite , colin de verdire @xcite , colin de verdire - guillemin @xcite , croke - sharafutdinov @xcite , guillemin - kazhdan @xcite , mckean - singer @xcite , osgood - phillips - sarnak @xcite , and zelditch @xcite , and the references therein . an interesting general problem ( for instance in the context of toric geometry )
is to what extent information about measures may be recovered from the spectrum , see guillemin - sternberg @xcite for a result in this direction .
in @xcite the last two authors formulated an inverse spectral conjecture for semitoric completely integrable systems ( see @xcite for a classification of semitoric systems in terms of five symplectic invariants ) : the semiclassical joint spectrum of a quantum semitoric system determines the corresponding classical system .
_ semitoric systems _ are four - dimensional integrable systems with two degrees of freedom for which one component of the system generates a @xmath28-periodic flow .
semitoric systems lie somewhere in between toric systems and general integrable systems .
if both components of the semitoric system are @xmath28-periodic , i.e. the system is generated by a hamiltonian @xmath668-torus action , then the system is a _ toric system _ ( strictly speaking after a harmless rescaling of the periods ) .
theorem [ theo : inverse - spectral ] above solves the conjecture in the class of toric systems . in this class
the result is even stronger , since there is no restriction on the dimension , and moreover only the spectrum modulo @xmath669 is needed , whereas in general one expects that an accuracy of order @xmath670 is necessary
. + + + * acknowledgements . *
we thank jochen brning , helmut hofer , peter sarnak , and thomas spencer for fruitful discussions .
the authors are grateful to helmut hofer for his essential support , which made it possible for lc and vns to visit ap at the institute for advanced study during the winter and summer of 2011 , where a part of this paper was written .
additional financial support for the visits was provided by washington university and nsf .
ap was partly supported by an nsf postdoctoral fellowship , a msri membership , an ias membership , nsf grants dms-0965738 and dms-0635607 , an nsf career award , a leibniz fellowship from the mathematisches forschungsinstitut oberwolfach , spanish ministry of science grant mtm 2010 - 21186-c02 - 01 , and by the spanish national research council ( csic ) .
vns was partly supported by the nonaa grant from the french anr and the institut universitaire de france .
y. colin de verdire and v. guillemin : a semi - classical inverse problem i : taylor expansions .
geometric aspects of analysis and mechanics , 81 - 95 , progr . math .
, 292 , birkhuser / springer , new york , 2011 .
b. kostant : on the definition of quantization . with questions by e. onofri , k. gawdzki , r. raczka and a. voros and replies by the author .
gomtrie symplectique et physique mathmatique ( colloq .
cnrs , no .
237 , aix - en - provence , 1974 ) , pp .
187 - 210 .
ditions centre nat .
recherche sci . , paris , 1975 .
s. v ngc : symplectic inverse spectral theory for pseudodifferential operators .
geometric aspects of analysis and mechanics _ progress in mathematics _ birkhuser / springer , new york , 2011 , volume 292 , 353 - 372 .
+ lvaro pelayo + school of mathematics + institute for advanced study + einstein drive + princeton , nj 08540 usa .
+ + and + + washington university , mathematics department + one brookings drive , campus box 1146 + st louis , mo 63130 - 4899 , usa .
+ _ e - mail _ : ` [email protected] ` san v ngoc + institut universitaire de france + + institut de recherches mathmatiques de rennes + universit de rennes 1 , campus de beaulieu + f-35042 rennes cedex , france + _ e - mail : _ ` [email protected] ` | we settle affirmatively the isospectral problem for quantum toric integrable systems : the semiclassical joint spectrum of such a system , given by a sequence of commuting toeplitz operators on a sequence of hilbert spaces , determines the classical integrable system given by the symplectic manifold and poisson commuting functions , up to symplectomorphisms . we also give a full description of the semiclassical spectral theory of quantum toric integrable systems .
this type of problem belongs to the realm of classical questions in spectral theory going back to pioneer works of colin de verdire , guillemin , sternberg and others in the 1970s and 1980s . |
interacting few - body quantum systems have attracted a lot of attention over the past decades . in that respect , quantum dots ( qds ) continue to be of fundamental interest to several fields , e.g. , spintronics , nanoelectronics , and quantum computing .
qds are small solid - state devices , typically containing a few up to several hundred electrons confined in all space directions @xcite .
they can be fabricated and studied using different approaches and materials @xcite , and control over both the charge and the spin degree of freedom of the confined electrons has been reported in experiments . here , we address qds as realized in two - dimensional ( 2d ) semiconductor devices , e.g. , in ultraclean 2d electron gases .
we consider the case of parabolically confined individual qds , which is quite appropriate in most practical cases @xcite .
qds offer the possibility to tune coulomb correlations among electrons via external gates .
spin properties of quantum dots have recently entered the focus of research @xcite , in particular as they are central both to quantum computation @xcite and to spintronics @xcite . here
spin - orbit ( so ) terms have to be taken into account , coupling the spin dynamics to the orbital motion . in general , so coupling is a relativistic effect , and appears to second order in the fine structure constant . in most materials of interest ,
two main mechanisms may be distinguished , namely rashba @xcite and dresselhaus @xcite so couplings .
the rashba so coupling strength @xmath5 is due to the surface inversion asymmetry present in the confinement to a 2d electron gas , and therefore can be tuned by external gates @xcite .
the dresselhaus coupling @xmath6 is generally not tunable but can be important if the host crystal has no bulk inversion symmetry , e.g. , in zincblende semiconductors . for simplicity , we only consider the linear dresselhaus term and neglect various additional spin - orbit contributions , see ref .
@xcite for an extended discussion .
these contributions can in principle be taken into account within our approach .
furthermore , we focus on the case of vanishing magnetic field . in this work , we investigate the behavior of a closed parabolic few - electron qd in the presence of spin - orbit couplings , containing up to @xmath7 interacting electrons . while the numerically exact method employed here allows to study arbitrary interactions in principle , it is probably most useful in the regime of intermediate - to - strong coulomb interactions , where a ` wigner molecule ' @xcite is formed . in terms of the standard electron gas parameter @xmath8 ,
the studied interactions are around @xmath9 to 10 .
then a clear tendency towards wigner crystallization can be observed , but different spatial ` shells' are not yet locked relative to each other as happens for very large @xmath8 @xcite .
in fact , while for @xmath10 , a completely classical situation is encountered @xcite , quantum effects still play a major role for the ` incipient ' wigner molecule of interest here .
in such a case , many standard calculational tools , e.g. , exact diagonalization @xcite , the hartree - fock approximation , the fixed - node @xcite or variational @xcite monte carlo approach , or density functional theory can meet various difficulties ( like artificial symmetry breakings ) or require explicit justification , see ref .
@xcite for a review .
in that situation , finite - temperature path - integral monte carlo ( pimc ) simulations represent an attractive alternative scheme .
the case of no spin - orbit coupling has been studied using pimc in refs .
@xcite . before describing our pimc scheme and the ensuing results ,
let us first discuss previous theoretical approaches to the physics of qds in the presence of so couplings . to study their effect on single - particle energies ,
one may set up perturbation theory for small so couplings . due to the linear dependence on momentum in eq .
( [ modhamso ] ) below , perturbation theory starts at second order and gives a quadratic decrease of single - particle energies with increasing rashba coupling @xmath5 ( or dresselhaus coupling @xmath6 ) @xcite .
single - particle energy level crossings induced by the so coupling have been discussed in refs .
@xcite as a function of an applied magnetic field . here ,
we are mainly concerned with many - body effects due to the coulomb interaction . for two electrons ,
exact diagonalization studies have been carried out for rather strong rashba couplings and weak interactions , as appropriate for insb dots @xcite .
energy spectra were examined , a jump in the magnetization of the dot as a function of magnetic field was found @xcite , and a favoring of exchange over direct interactions as a consequence of so interactions was discussed @xcite .
governale @xcite has employed spin - density functional theory for @xmath11 .
he found that a very strong so coupling leads to new peaks and/or the suppression of hund s rule peaks otherwise present in the addition energy spectrum .
an additional in - plane magnetic field was argued to imply paramagnetic behavior . finally , in ref .
@xcite , several approximate schemes have been employed to study so effects in weakly interacting qds with @xmath12 to 13 electrons . here
we provide results for @xmath1 electrons with strong coulomb interactions and so couplings , where exact diagonalization techniques may not apply anymore .
one should also note that for @xmath13 to 10 , spin effects are very important but necessitate an essentially exact treatment .
for a recent comparison of qmc data to density functional theory in that respect , see ref .
@xcite . in the absence of spin - orbit couplings , pimc simulations for qds
suffer from two well - known problems , namely the fermionic sign problem and the spin contamination problem @xcite .
( the first problem can be relieved to some extent by the multilevel blocking algorithm @xcite . here
we restrict ourselves to a simpler `` brute - force '' approach . )
the spin contamination problem arises for @xmath14 because both the total qd spin @xmath2 and its @xmath15 component @xmath16 are good quantum numbers .
now @xmath16 is in practice fixed during the simulation since there are no spin flip terms in the hamiltonian . at finite temperature ,
one then arrives at the undesirable situation where states with different @xmath2 but the same @xmath16 contribute to the simulation .
this considerably complicates data analysis and represents a well - known problem affecting also other schemes , e.g. , hartree - fock calculations .
in the presence of spin - orbit coupling , however , neither @xmath2 nor @xmath16 are good quantum numbers and the full space of all @xmath17 becomes accessible . with increasing so couplings , we find that the sign problem worsens exponentially , restricting the applicability of our approach to weak so couplings .
since in applications , so effects are usually weak , however , this restriction is not too severe . in order to eliminate the spin contamination problem in the limit of zero so coupling
, we may study a few finite but small values for the so couplings , and then extrapolate @xmath18 .
this allows to reliably compute , for instance , the addition spectrum of the dot , where we find peaks ( corresponding to stability islands of these wigner molecules ) for @xmath3 and @xmath4 electrons .
furthermore , we compute the dependence of the spin state , @xmath19 as a function of particle number @xmath0 , where @xmath20 is the total spin operator .
the structure of the paper is as follows . after introducing the model in section [ sec : model ]
we derive the short - time propagator for interacting fermions in a parabolic qd subject to either rashba or dresselhaus so coupling ( or both of them at the same time ) , and discuss the numerical scheme in some detail .
numerical results are presented in section [ sec : results ] , and we conclude in section [ conc ] . throughout the paper , we put @xmath21 .
we study the @xmath0-electron hamiltonian describing a closed parabolic dot in a 2d electron gas , @xmath22 where @xmath23 is the effective electron mass and @xmath24 is the oscillator frequency . with @xmath25 , the vectors
@xmath26 denote the 2d momenta and space coordinates of all @xmath0 electrons .
the coulomb potential among the electrons contains screening effects of the host material via the dielectric constant , @xmath27 . measuring energies ( lengths ) in units of @xmath24 ( @xmath28 ) ,
a dimensionless coulomb interaction parameter is given by @xmath29 .
for common host materials , @xmath30 is in the range of few up to hundreds of @xmath31 .
the confinement energy @xmath32 is then typically between @xmath33 up to a few @xmath34 , which allows directly to determine the strength of the coulomb interaction @xmath35 .
typical @xmath35 currently realized experimentally are between 0.5 and 5 , the latter value @xcite being already quite close to the regime studied in our paper . for @xmath36 ,
coulomb repulsion starts to dominate over the kinetic energy , and electrons spatially arrange on shells .
this `` wigner molecule '' regime @xcite is studied in this work .
let us then address the so couplings considered here .
typically , two different so couplings are of paramount importance in semiconductor heterostructures , namely the ( linear ) dresselhaus and the rashba so coupling . we allow for both types and consider the spin - orbit hamiltonian ( of the @xmath37th particle ) @xmath38 where the standard pauli matrices @xmath39 act in spin space and @xmath40 note that both types of so coupling can be transformed into each other by a unitary transformation .
hence the spectra for @xmath41 and @xmath42 coincide ( for zero magnetic field ) , see also sec .
[ sec : results ] below .
remarkably , even on the single - particle level , there is no closed solution to the schrdinger equation in a parabolic potential subject to so coupling of any kind .
we mention in passing that for a cylindrical box , the single - particle problem has been solved analytically in ref .
pimc simulations provide a powerful tool to extract numerically exact results for this few - body interacting quantum system .
one starts by discretizing imaginary time ( @xmath43 ) into sufficiently short time intervals @xmath44 , where @xmath45 is the ( integer ) trotter number . for @xmath46 ,
an approximate but accurate short - time propagator can be constructed via the trotter - suzuki decomposition . to that end , we split @xmath47 into the noninteracting part @xmath48 ( including the so couplings ) and the remaining interaction part ( @xmath49 ) . assuming that the spin dynamics is slow on the timescale @xmath50 , some algebra along the lines of ref .
@xcite then gives the short - time single - particle propagator under @xmath48 in the form @xmath51 } \
\left\langle \sigma_2\left| e^{-im^ * ( { \bf r}_2-{\bf r}_1 ) \cdot a\vec\sigma}\right|\sigma_1\right\rangle\end{aligned}\ ] ] where @xmath52 . the electron is described by its position @xmath53 and the @xmath15-component @xmath54 of the spin .
furthermore , @xmath55 denotes the standard 2d oscillator action @xmath56=\frac { \left({\bf r}_1 ^ 2+{\bf r}_2 ^ 2\right)\cosh(\omega_0\tau)-2{\bf r}_1\cdot { \bf r}_2 } { 2l_0 ^ 2\sinh(\omega_0\tau)}.\ ] ] note that in the absence of so couplings , the exact propagator of the harmonic oscillator is reproduced . for small coupling constants
@xmath57 and sufficiently short @xmath58 eq . ( [ single ] ) represents a very accurate short - time approximation to the single - particle propagator .
the spin part in the propagator ( [ single ] ) can be written as @xmath59 where @xmath60 and @xmath61 .
using eq .
( [ single ] ) , the many - body propagator follows in the form of @xmath62 slater determinants .
let us then denote the coordinate and the spin @xmath63 of the @xmath37th electron on time slice @xmath64 ( where @xmath65 ) as @xmath66 and @xmath67 , respectively . with respect to the time direction , we have periodic boundary conditions .
including coulomb interactions , we then obtain the many - particle partition function for given discretization @xmath44 in the form @xmath68 where the @xmath62 matrix @xmath69 has the matrix elements @xmath70 the last term in eq .
( [ partfun ] ) represents the coulomb interaction between all @xmath0 electrons confined to the qd . if there is no so coupling , @xmath71 factorizes into a spin - up and a spin - down part , @xmath16 is a constant of motion , and the spin contamination problem arises .
we mention in passing that the weight in the discretized path integral ( [ partfun ] ) is complex - valued , and one therefore may expect that observables have an imaginary part .
however , all statistical averages for physical observables must have zero imaginary part , and this indeed we find within the standard stochastic error bars . the discretized canonical many - particle partition function ( [ partfun ] ) then allows us to access equilibrium observables of interest . for concreteness , we have chosen a rather low but finite temperature , @xmath72 . furthermore ,
unless stated otherwise , simulations were carried out for @xmath73 and interaction strength @xmath74 , which puts us into the wigner molecule regime .
note that by simply replacing @xmath75 , results for @xmath76 follow .
the main limitation for this type of pimc simulation comes from the fermionic sign problem .
the sign problem generally arises when different paths that contribute to averages carry different signs , or even complex - valued phases , as encountered in the case of non - zero so couplings .
the resulting sign cancellation when sampling fermion paths then manifests itself as a very small signal - to - noise ratio .
for instance , as a consequence of exchange , these phases appear when forming slater determinants . here
, in the presence of so couplings , the sign problem occurs even for a single particle .
unfortunately , as a function of so couplings , we find an exponential decay of the sign , see fig .
[ phivsa ] .
this can be rationalized already on the single - particle level , since the propagator acquires a complex phase factor in the presence of so couplings .
the propagator then resembles a real - time propagator for a single particle , with time corresponding to the so coupling . for this problem ,
the exponential severity of the sign problem is well established . in effect , the parameter regime where reliable simulations are possible is limited to small - to - intermediate so couplings .
we note in passing that the rashba so coupling @xmath77 reported for the ingaas dots of ref .
@xcite are about one order of magnitude larger than our largest value .
however , the so coupling in ingaas is also unusally strong , and our values should apply more directly to gaas dots . simultaneously , the sign problem becomes more severe when increasing @xmath0 or lowering temperature .
for the chosen parameters , the sign problem only allows to study qds containing @xmath1 electrons .
the average sign is @xmath78 in all cases reported below .
we then compute several observables .
first , the energy of the @xmath0-electron dot can be obtained from @xmath79 where the derivative can be explicitly carried out using eq .
( [ partfun ] ) .
knowledge of the @xmath80 determines the addition energy @xmath81 a peak in the addition energy @xmath82 indicates enhanced stability of the @xmath0-electron dot ( `` magic number '' ) @xcite .
note that experimentally observed addition energies are determined by _ free _ energy differences , while we compute the energy .
however , for the low - temperature regime studied here , entropic contributions are smaller by about one order of magnitude , and therefore our values for @xmath82 are of relevance to actual experiments .
another quantity of interest is the total spin @xmath2 , which we extract from the definition @xmath83 where brackets denote the statistical average using eq .
( [ partfun ] ) , and @xmath20 is the total spin operator . finally , spatial ordering can be monitored via the electronic particle density , @xmath84 which is normalized to @xmath85 .
the related charge densities can also be accessed experimentally , e.g. , via capacitance spectroscopy or scanning tunneling microscopy techniques @xcite .
as we work with finite discretization @xmath50 , trotter approximation errors have to be taken into account @xcite . as shown in ref .
@xcite , for small @xmath50 , such errors vanish quadratically for all observables , allowing for simple and efficient extrapolation schemes that completely eliminate this finite-@xmath50 error .
the @xmath86 scaling regime is reached for @xmath87 when computing the energy .
furthermore , particle or spin densities are found to reach the @xmath86 scaling regime already at higher @xmath50 .
results shown here have been carefully extrapolated down to @xmath88 with a linear regression fit , using results from several simulations obtained at @xmath89 .
hence no discretization errors are present , see also ref .
@xcite , and error bars denote just the standard stochastic monte carlo errors .
spin flip moves are an essential ingredient of the algorithm . within the pimc we allow for spin flips as well as for position moves .
single - particle moves were found sufficient to ensure ergodicity .
the average trial step size for position moves was adjusted to give acceptance rates of the order of @xmath90% . on the other hand , typical spin - flip acceptance rates were much lower ( several percent ) and strongly dependent on @xmath91 . the possibility to change @xmath16 arises from the so coupling and can be used to circumvent the spin contamination problem even for the case of no so coupling , namely by extrapolating finite-@xmath5 results ( where spins can be flipped and no spin contamination problem is present ) down to @xmath92 .
such a scheme yields the energy @xmath80 as well as the expectation value @xmath93 for the total spin of the many - body system at finite temperature .
typically , after @xmath94 equilibration passes , @xmath95 mc samples were accumulated for each parameter set .
our code runs at a speed of up to two weeks ( for @xmath7 , @xmath96 ) for a given parameter set per @xmath95 samples on a standard 2 ghz xeon processor .
we have checked our pimc energies for @xmath97 against finite - temperature exact diagonalization results , including both interactions and spin - orbit couplings .
we found excellent agreement , validating our approach .
to verify that for @xmath74 , we indeed have a wigner molecule @xcite , let us start with the radially integrated particle density ( [ chargeden ] ) shown in fig .
[ wigner ] .
the cylindrical symmetry of the qd implies that @xmath98 only depends on the modulus @xmath99 .
the plot indicates that the sixth electron enters the center of the dot , whereas the remaining five electrons arrange on an outer ring in order to minimize the coulomb repulsion .
more electrons are then added to the outermost shell .
finally , for @xmath7 , a second electron enters the center .
this spatial shell filling sequence ( as opposed to orbital shell filling ) is typical of the wigner molecule , which forms the finite - size counterpart of a wigner crystal .
in fact , precisely this spatial filling sequence has been reported from a purely classical analysis ( in particular , disregarding spin effects ) @xcite . at higher temperatures ,
the wigner molecule melts via thermal fluctuations , while for lower @xmath35 , it is eventually destroyed by quantum fluctuations . generally , we find that particle densities are practically independent of the so coupling strengths @xmath5 or @xmath6 , at least for @xmath100 .
although we consider rather small so couplings , the many - body energy @xmath80 can be clearly seen to decrease as a function of @xmath5 ( here , @xmath73 ) .
this trend has been observed in other studies as well @xcite . in fig .
[ ener34 ] , we show this effect for @xmath3 and @xmath101 electrons .
for @xmath102 , the so coupling does not significantly influence addition energies for @xmath103 , see fig .
[ adden ] .
however , there is a slight increase in the addition energy @xmath104 , while the peak for @xmath4 is reduced for @xmath105 ( dotted curve ) .
( in particular , @xmath106 for @xmath107 , while @xmath108 at @xmath105 . )
remarkably , the magic numbers @xmath3 and @xmath4 encountered in the wigner molecule regime are different from the ones for weak interactions ( where a standard fermi liquid phase is present ) .
this is indicated in the inset of fig .
[ adden ] for @xmath109 and @xmath105 .
the magic numbers for @xmath109 can be rationalized in terms of the subsequent filling of energy levels ( orbitals ) of a 2d harmonic oscillator .
this predicts a peak for @xmath97 , where the lowest level is filled , and another peak at @xmath101 reflecting hund s rule behavior , see also ref .
@xcite . for strong coulomb repulsion ,
the peak for @xmath97 is completely absent , while @xmath3 now corresponds to a magic number .
a very distinct peak in the addition energy is observed for @xmath4 . to the best of our knowledge
, this peak is not expected for weak interactions .
the picture of filling up spatial shells in the wigner solid phase ( discussed above ) may suggest that the filled _ spatial _ shell configurations @xmath110 and @xmath111 represent magic numbers .
however , under a classical reasoning , the addition energy should have no pronounced peaks but exhibits a rather smooth and monotonic decay in @xmath82 @xcite .
moreover , the classical prediction for @xmath82 @xcite yields values one order of magnitude smaller than the @xmath82 found here .
this indicates that for @xmath74 , despite the clear onset of spatial ordering , quantum effects are still very important and cause the magic numbers @xmath3 and @xmath4 .
notably , for these @xmath0 , the dot is seen to be partially spin - polarized .
the magic numbers for @xmath3 and @xmath4 can thus be rationalized in terms of a hund - rule type behavior specific to the incipient wigner crystallized regime ( see the discussion below and fig . [ s2b ] ) .
our findings are therefore characteristic for the quantum character of the wigner molecule . a purely classical `` wigner solid '' analysis @xcite is expected to apply only for extremely large @xmath35 . our numerical pimc results for @xmath80 ( and the spin ) as a function of @xmath0 and
the rashba coupling @xmath5 are summarized in table [ valuetab ] .
these data were all obtained for @xmath74 and @xmath112
next we consider the dependence of the energy on the two types of spin couplings .
let us take @xmath101 electrons , again at @xmath74 , and fix @xmath113 .
we then study @xmath114 as a function of @xmath115 , where @xmath116 tunes the relative strength of rashba versus dresselhaus coupling .
the result is depicted in fig .
[ devsg ] , showing a symmetric curve @xmath117 .
the symmetric shape can be explained by the unitary spin rotation which transforms the rashba- into the dresselhaus term and vice versa . with increasing @xmath118 , the relative weights of the dresselhaus and rashba couplings increase , leading to increasing energy gains @xmath119 over the system without so couplings .
finally , we observe that spin plays a major role for @xmath74 . our data for @xmath2 as defined in eq .
( [ s2 ] ) show a drastic influence of temperature .
for such a strongly interacting system , energies for different @xmath2 are typically close by , and a thermal average can then result in different @xmath2 as compared to ground - state results .
for instance , for @xmath97 , we find a mixture of the singlet ground state @xmath120 and the excited triplet state @xmath121 , as pimc gives @xmath122 at @xmath112 . due to the strong coulomb interaction ,
there is a clear trend towards partial spin polarization as a function of @xmath35 , indicated in fig .
[ s2b ] , see also table [ valuetab ] .
moreover , the thermal average for @xmath123 obtained from pimc simulations at @xmath74 increases monotonically with @xmath0 , while the ground - state spin @xcite has rather different values and shows a nonmonotonic dependence on @xmath0 .
it is worth mentioning that the peaks in the addition energy for @xmath3 and @xmath4 correspond to enhanced values of @xmath2 as well .
stability of the @xmath0 electron dot is thus connected to a tendency towards spin polarization , reminiscent of hund s rule behavior . for @xmath124 ,
corresponding to weaker but still sizeable interactions , the dependence of the ground - state spin @xcite is given in the inset .
the shown nonmonotonic dependence on @xmath0 reflects the standard hund s rule physics . comparing the @xmath124 and @xmath74 results
, we observe that interactions tend to further spin - polarize the dot .
the huge thermal effects observed in the expectation value @xmath93 also indicate that unless experiments are carried out at extremely low temperatures , spin blockade phenomena @xcite relying on total spin selection rules will be thermally washed out .
finally , we note that for the so couplings studied here , spin expectation values were not significantly affected .
we have investigated the behavior of up to nine electrons in a quantum dot .
we took into account strong coulomb correlations between the electrons , and also incorporated spin - orbit couplings .
our results were obtained from path - integral quantum monte carlo simulations .
an exponential decrease of the fermionic sign is found with increasing so couplings .
this sign problem appears even for a single electron .
nevertheless , simulations are possible for weak so couplings , where their inclusion can also be used to eliminate the spin contamination problem .
we observe peaks in the addition energy spectrum for @xmath3 and @xmath4 , which are likely to correspond to the stability of partially spin - polarized configurations induced by coulomb interactions .
these peaks are neither expected in the weak - interaction regime nor in the classical ( deep ) wigner solid , where spin effects are negligible .
we hope that this prediction can soon be tested experimentally .
our results were obtained in the regime of weak spin - orbit couplings , since otherwise numerical instabilities associated with the sign problem occur . given this restriction however , pimc offers a powerful tool to analyze the effects of spin - orbit couplings in strongly interacting quantum dots .
we find no dramatic effects , but observable downward shifts in the many - body energy that scale quadratically in the spin - orbit couplings .
spin - orbit couplings also affect addition energies .
reimann and m. manninen , rev .
mod . phys . * 74 * , 1283 ( 2002 ) , and references therein .
y. alhassid , rev .
mod . phys . * 72 * , 895 ( 2000 ) .
kouwenhoven , d.g .
austing , and s. tarucha , rep .
phys * 64 * , 701 ( 2001 ) .
ashoori , nature * 379 * , 413 ( 1996 ) .
s. tarucha , d.g .
austing , y. tokura , w.g .
van der wiel , and l.p .
kouwenhoven , phys .
. lett . * 84 * , 2485 ( 2000 ) .
p. matagne , j.p .
leburton , d.g .
austing , and s. tarucha , phys .
b * 65 * , 085325 ( 2002 ) .
s. lindemann , t. ihn , t. heinzel , w. zwerger , k. ensslin , k. maranowski , and a.c .
gossard , phys .
b * 66 * , 195314 ( 2002 ) . j. nitta , t. akazaki , h. takayanagi , and t. enoki , phys .
lett * 78 * , 1335 ( 1997 ) . th .
schpers , g. engel , j. lange , th .
klocke , m. hollfelder , and h. lth , j. appl .
phys . * 83 * , 4324 ( 1998 ) .
d. grundler , phys .
84 * , 6074 ( 2000 ) .
t. koga , j. nitta , t. akazaki , and h. takayanagi , phys .
lett . * 89 * , 046801 ( 2002 ) .
j. knemann , r.j .
haug , d.k .
maude , v.i .
falko , and b.l .
altshuler , phys .
94 * , 226404 ( 2005 ) .
d. loss and d.p .
divincenzo , phys .
a * 57 * , 120 ( 1998 ) .
i. zutic , j. fabian , and s. das sarma , rev .
mod . phys . * 76 * , 323 ( 2004 ) .
rashba , sov .
state * 2 * , 1109 ( 1960 ) ; y.a .
bychkov and e.i .
rashba , j. phys .
c * 17 * , 6039 ( 1984 ) .
g. dresselhaus , phys . rev . * 100 * , 580 ( 1955 ) .
destefani and s.e .
ulloa , phys .
b * 71 * , 161303(r ) ( 2005 ) .
r. egger , w. husler , c.h .
mak , and h. grabert , phys .
* 82 * , 3320 ( 1999 ) ; _ ibid . _ * 83 * , 462(e ) ( 1999 ) .
mak , r. egger , and h. weber - gottschick , phys .
lett . * 81 * , 4533 ( 1998 ) .
filinov , m. bonitz , and yu.e .
lozovik , phys .
lett . * 86 * , 3851 ( 2001 ) .
b. reusch and r. egger , europhys .
* 64 * , 84 ( 2003 ) .
w. husler , europhys . lett . * 49 * , 231 ( 2000 ) .
j. harting , o. mlken , and p. borrmann , phys .
b * 62 * , 10207 ( 2000 ) .
m. rontani , c. cavazzoni , d. bellucci , and g. goldoni cond - mat/0508111 yu.e .
lozovik , usp .
nauk * 153 * 356 ( 1987 ) [ sov .
usp . * 30 * , 912 ( 1987 ) ] ; yu.e .
lozovik and l.m .
pomirchy , phys .
status solidi b * 161 * , k11 ( 1990 ) .
bolton and u. rssler , superlatt .
* 13 * , 139 ( 1993 ) ; v.m .
bedanov and f.m .
peeters , phys .
b * 49 * , 2667 ( 1994 )
mikhailov , phys .
b * 65 * , 115312 ( 2002 ) .
f. pederiva , c.j .
umrigar , and e. lipparini , phys .
b * 62 * , 8120 ( 2000 ) .
a. harju , s. siljamki , and r.m .
nieminen , phys .
b * 65 * , 075309 ( 2002 ) .
e. tsitsishvili , g.s .
lozano , and a.o .
gogolin , phys . rev .
b * 70 * , 115316 ( 2004 ) .
o. voskoboynikov , c.p .
lee , and o. tretyak , phys .
b * 63 * , 165306 ( 2001 ) .
kuang , c.s .
tang , and w. xu , j. appl .
phys . * 95 * 6368 ( 2004 ) .
t. chakraborty and p. pietilinen , phys .
b * 71 * , 113305 ( 2005 ) . c.f .
destefani , s.e .
ulloa , and g.e .
marques , phys .
b * 69 * , 125302 ( 2004 ) ; phys . rev .
b * 70 * , 205315 ( 2004 ) .
m. governale , phys .
. lett . * 89 * , 206802 ( 2002 ) .
a. emperador , e. lipparini , and f. pederiva , phys .
b * 70 * , 125302 ( 2004 ) .
a. ghosal , c.j .
umrigar , h. jiang , d. ullmo , and h.u .
baranger , phys .
b * 71 * , 241306(r ) ( 2005 ) . j. zinn - justin , _ quantum field theory and critical phenomena _ ,
4th edition , chapter 3 ( clarendon press , oxford , 2002 ) .
h. de raedt and b. de raedt , phys .
a * 28 * , 3575 ( 1983 ) .
fye , phys .
rev b * 33 * , 6271 ( 1986 ) .
d. weinmann , w. husler , and b. kramer , phys .
* 74 * , 984 ( 1995 ) .
.pimc data for @xmath80 and @xmath125 as a function of @xmath5 for @xmath126 and @xmath112 .
the @xmath76 data are taken from extrapolations .
bracketed numbers denote error estimates . [
cols="<,<,<,<",options="header " , ] | we develop path - integral monte carlo simulations for a parabolic two - dimensional ( 2d ) quantum dot containing @xmath0 interacting electrons in the presence of dresselhaus and/or rashba spin - orbit couplings .
our method solves in a natural way the spin contamination problem and allows for numerically exact finite - temperature results at weak spin - orbit coupling . for @xmath1 electrons , we present data for the addition energy , the particle density , and the total spin @xmath2 in the wigner molecule regime of strong coulomb interactions .
we identify magic numbers at @xmath3 and @xmath4 via a peak in the addition energy .
these magic numbers differ both from weak - interaction and classical predictions , and are stable with respect to ( weak ) spin - orbit couplings . |
galaxy clusters are the largest virialized structures in the universe .
these systems are composed of three components behaving differently during collisions : galaxies , hot gas and dark matter . while the hot intergalactic gas presents strong eletromagnetic interactions during clusters enconteurs , both dark matter and galaxies
are predicted to be collisioness , see e.g. @xcite .
hence , clusters are important laboratories for studying the dynamics of the different kinds of matter in the universe .
indeed , galaxy clusters are themselves the result of several mergers of smaller groups of galaxies , see e.g. @xcite , a process still ongoing in many systems like , for instance , the so - called bullet cluster ( 1e 0657 - 56 ) [ @xcite ] , and other clusters like cl 0152 - 1357 and ms 1054 , see @xcite .
also , cluster cl 0024 + 17 has been a target of many studies since its discovery by @xcite .
this is an intermediate - redshift system ( z=0.395 ) with both weak and strong lensing @xcite .
there are several indications that this cluster , an apparently relaxed system , is actually the portrait of a collision of two clusters , along our line of sight @xcite .
the weak lensing analysis presented by @xcite shows a ring - like structure in the projected matter distribution at @xmath4 ( @xmath5 0.4 mpc ) , surrounding a soft , dense core at @xmath6 .
the authors interpret this substructure as the result of a high - speed line - of - sight collision of two massive clusters @xmath51 - 2 gyr ago .
an interesting question in this context refers to the behaviour of the galaxy component with respect to dark matter . in the case of the bullet cluster ,
cl 0152 - 1357 and ms 1054 - 0321 , the distribution of galaxies follows dark matter quite well @xcite .
however , the recent work of @xcite , based on the 2d distribution of galaxies , suggests that the ringlike structure observed in dark matter measurements is not seen in the projected two - dimensional galaxy distribution .
in the present work , i introduce an alternative approach to probe both in radial and angular coordinates the projected galaxy distribution around galaxy clusters , and applied it to the case of cl 0024 + 17 .
the basic aim is to find the scales where galaxies present a significant deviation from an inhomogeneous poisson statistical process and compare that to the dark matter ring radius and the secondary clump present in the field .
the sample of galaxies around cl 0024 + 17 used in this work is taken from the catalog of czoske et al .
this corresponds to 295 galaxies in the range @xmath7 , the most probable cluster members in the @xmath8 field .
the center of the cluster was chosen to be the geometric center of the dark matter ring , see @xcite .
distance @xmath9 for each galaxy was calculated from their equatorial coordinates , in arcsecs .
since the completeness of the sample varies from @xmath1080% at the cluster center to @xmath1150% in the outer regions , the catalog was corrected through 1,000 monte carlo realizations of the sample .
this was done following approximately the completeness variation map of @xcite , adopting a procedure similar to @xcite .
one example of such realizations is presented in figure 1 , where black points are taken from the simulation and green points are the real galaxies from @xcite .
the background in this figure is a density map based on the void distance from each point with respect to a fine grid @xcite .
this is just introduced to illustrate how the simulations fill the empty cells left by the original sample .
, width=359,height=321 ]
galaxy distribution around cl 0024 + 17 is the point pattern we want to probe . in statistical theory , a spatial point pattern ( spp ) @xmath12 is simply a collection of points @xmath13 in some bounded area , and interpreted as a realization of a more generic point process @xmath14 .
analysis and modelling of spp is a challenging topic in modern spatial spatistics .
the mathematical theory was first developed to solve various problems where it is sensible to model the locations of events as random , @xcite .
recently , statistical theory has made many developments with respect to spp analysis .
one of these advances is related to achievements in the study of nonstationary processes . statistical stationarity ( or homogeneity ) means that the distribution of the point process is invariant under translations .
this implies that the intensity @xmath15 , i.e. the expected number of points per unit area of the spp , is constant .
nonstationarity , otherwise , allows for deterministic local variation of the intensity:@xmath16 , where @xmath17 is a given location on the spp .
this is the case we are interested in this work , since galaxy distribution in the sampling window around the cluster is obviously inhomogeneous . in spatial statistics ,
ripley k - function @xcite is a classical tool to analyse spp .
it is defined as the expected number of events within the distance @xmath9 of any given point divided by the intensity @xmath15 .
the k - function , being a measure of the distribution of the inter - point distances , captures the spatial dependence between different regions of a spp .
mathematically , the inhomogeneous k - function is defined as @xmath18.\ ] ] thus , @xmath19 is the expected total weight of all random points within a distance @xmath9 of the point @xmath17 , where the weight of a point @xmath13 is @xmath20 @xcite , and for a poisson process with intensity function @xmath21 , the k - function reduces to @xmath22 .
a standard computational estimator of k is given by @xmath23 where @xmath24 is an edge correction weight , @xmath25 is the observational window , and @xmath26 is an estimate of the intensity function @xmath21 .
although the k - function is the classical tool to probe spp , modern spatial statistics rarely uses k(r ) , but rather its variant , the l - function as introduced by @xcite . in the planar case ,
the l - function estimator is defined as @xmath27 the advantage of using @xmath28 is that it is always proportional to @xmath9 , and in the poisson case @xmath29 , @xcite .
thus , @xmath30 indicates clustering with respect to a poisson process . in this work ,
the l - function was applied to the galaxy distribution around cl 0024 + 17 using the estimation algorithms provided in the library * spatstat * , developed by @xcite , and running under the r statistical package ( www.cran.r-project.org ) .
the edge correction used here is implemented in the tasks * kinhom * and * linhom*. it is briefly described in the next .
edge corrections are used to correct biases in the estimation of k(or l)-function due to the loss of information near to the boundary of @xmath25 . in this work ,
i have applied the following edge correction : @xmath31 where @xmath32 is the distance from @xmath13 to the boundary of the window @xmath25 . among other options available in * spatstat
* , this weighting function is considered the most computationally and statistically efficient when there are large numbers of points , see @xcite .
finally , there remains the question of how to estimate the intensity function @xmath21 .
this is estimated using a ` leave - one - out ' kernel smoother , as described in @xcite .
the estimate @xmath33 is computed by removing @xmath13 from the point pattern , then applying a gaussian kernel smoothing to the remaining points , and finally evaluating the smoothed intensity at @xmath13 ( in this work we set the kernel width equal to @xmath34 ) .
this estimator is also included in * kinhom * and * linhom*.
i have used the routine * linhom * to probe the galaxy distribution around cl 0024 + 17 . for each mc realization of this field
, i found the scales where clustering weakens towards an inhomogeneous poisson process . in figure 1 , regions above the line @xmath29 describes clustered distribution of points . to test the null hypothesis of an inhomogeneous poisson process i also plot in this figure the 95% envelopes computed from @xmath35 independent simulations of this process , with the same intensity as the galaxy field , i.e. , @xmath36 for @xmath37 .
the upper and lower envelopes of these simulated curves are @xmath38 @xmath39 if data come from a poisson process , then @xmath40 and @xmath41, ... ,@xmath42 are statistically equivalent and independent , so the probability that @xmath40 lies outside the envelopes is equal to @xmath43 by symmetry .
this corresponds to the significance level @xmath44 in the test which rejects the null hypothesis .
hence , to compute the 95% envelopes ( @xmath45 ) , we need only 39 mc simulations of the poisson process , see e.g. @xcite . in figure 2 , we see two different regions where the null hypothesis of a poisson process is rejected at the 5% significance level .
the transition scales where the clustered process weakens towards a poisson process are marked with green dashed lines .
a single transition around a cluster indicates its outer limit , the region where galaxy distribution declines in clustering ( or intensity ) and turns indistinguishble from a poisson process .
two ( or more ) transitions indicate a recoverage of the clustering process . any morphological feature could be responsible for that : clumps , rings , tails , etc . in the present analysis , about 88% of the mc simulations present two scales like those we see in figure 2 .
the shapiro - wilk s test fails to detect that the distributions of @xmath46 and @xmath47 transition scales significantly deviate from normal distributions : @xmath48 ( p - value=0.3603 ) and @xmath49 ( p - value=0.7338 ) .
also , bartlett s and kruskal - wallis s tests strongly reject the hypotheses of equal variances and means , respectively , for these two distributions ( p - value @xmath50 in both tests ) .
all these results indicate the presence of two distinct physical scales in the cl 0024 + 17 galaxy field , as we can see in figure 3 .
note that the dark matter ring is located at @xmath4 @xcite , and separations between this radius and the @xmath51 and @xmath52 scales are equal to @xmath53 and @xmath54 , respectively .
this suggests that the dark matter ring is more likely to be associated to the second region .
also in this figure , it is presented the average correlation function for all set of mc simulations ( see the small box ) .
note that the @xmath51 transition scale corresponds just to a narrow diminishment interval in the correlation function , while the dm ring radius and the @xmath52 transition scale are approximately coincidents with the highest peaks of @xmath55 after @xmath56 . not by chance
this scale roughly marks the intersection of the two gaussian distributions .
these results indicate that the system has a core at @xmath57 ( which agrees with @xcite ) and that there is an extended region after this radius with two significant peaks ( @xmath58 above @xmath59 ) .
one of them agrees well with the position of the dm ring radius while the other coincides with the mean of the @xmath52 transition scale .
, width=302,height=302 ] and @xmath52 transition scales .
vertical dashed lines indicate the position of the dm ring radius .
small box : average correlation function for all mc simulations . dm ring , @xmath51 and @xmath52 scales are indicated by green dashed lines.[fig3],width=302,height=302 ]
one should be careful not to overinterpret the previous analysis by concluding that the second scale we see in figure 3 implies that the galaxy distribution around cl 0024 + 17 has a luminous counterpart of the dark matter ring .
indeed , this result may have to do with the way observations were done .
for instance , the completeness map was constructed after a smoothing procedure using a gaussian kernel of @xmath34 @xcite .
this size is similar to that i have found for the first transition scale , but the separation between the two normal distributions is @xmath52.7 times this smoothing scale , so it seems that the width of the kernel is not large enough to invalidate the results . on the other hand , the presence of a secondary peak in the distribution about @xmath60 nw of the cluster center ( see figure 1 and figure 4 of @xcite ) can be the responsible for the second scale detected in data .
this can be seen in figure 4 , where i plot the perspective view of galaxy distribution also smoothed by a gaussian kernel of width @xmath34 .
letters a and b indicate the first and second density peaks in the field .
note that the peaks are very close to each other , although clearly distinct .
less distinguishble , however , can be the effect of a spheroidal clump from a ring in the spatial analysis i have applied to data . one way to measure such morphologic effects
is now presented . .
letters a and b indicate the the first and second density peaks in the field .
[ fig4],width=302,height=377 ] i have built two mock samples made to mimic the cases we want to study . in model
i , we have a central cluster with radius equal to @xmath61 , plus a clump at @xmath3 nw ( radius defined as @xmath62 ) and a poisson background with same intensity of the cl 0024 + 17 field .
both the cluster and the clump have a hernquist profile ( hernquist 1990 ) , but the cluster is set to have @xmath63 times the density of the clump ( after the gaussian smoothing ) .
model ii is the central cluster plus a ring at the second scale , @xmath1,with width equal to @xmath64 .
the density of the cluster is also @xmath63 times the density of the ring . for each case , 1,000 mock fields were generated .
examples are presented in figure 5 , pannels a and c. analysis of these mock fields shows that : ( i ) in both cases two statistically distinct scales are found bartlett and kruskal - wallis tests reject the null hypotheses of equal variances and means at @xmath1099% confidence ( see resultant normal distributions in figure 5 , pannels b and d ) ; ( ii ) however , while in model ii about 94% of the realizations present two scales , this only happens for 61% of the model i realizations ; ( iii ) the @xmath51 gaussian is narrower for model i ( @xmath65 ) while the @xmath52 gaussian is narrower for model ii ( @xmath66 ) ; ( iv ) however , there is no evidence for difference in the results of models 1 and 2 bartlett , anova and tukey tests indicate only a 0.001% probability of difference between the samples .
the morphology test presented here is quite naive in the sense that neither model reproduces the complexity of the real galaxy field .
exactly for this reason , it may indicate that the methodology presented in section 3 is not accurate enough to discriminate different morphologies around the galaxy cluster .
to improve the results , i have applied a wavelet analysis to the galaxy field of cl 0024 + 17 . here
i followed the method of @xcite .
basically , the coordinates of the galaxies are convolved by a radial mexican hat filter on a grid of @xmath67 pixels .
the method leads to a multiscale analysis which starts below the biggest scales , where the whole cluster is detected as a unique structure , and ends with the smallest scale , when approximately only one galaxy lies into the operating area of the wavelet . to obtain the significance level of a given structure
i have applied the replica procedure , where many simulations are made by drawing independently @xmath68 and @xmath69 from the the @xmath70 and @xmath71 distributions of the sample studied to destroy small - scale correlations , see e.g. @xcite .
the analysis is then performed on the `` random fields '' for the same wavelet scale than the true field .
the significance level of a given structure is estimated from the number of images @xmath72 in which maxima greater than the maxima wavelet coefficients are found among the whole set of simulations @xmath73 .
thus , @xmath74 , see @xcite .
in figure 6 we present the isosurfaces of the wavelet coefficients corresponding to the scales @xmath75 , respectively ( in pixels units for a map of @xmath76 pixels ) .
the wavelet scale used is @xmath77 pixels corresponding to 1.0 mpc . in the top
left low resolution image ( @xmath77 pixels ) , we see the prominent central region of the cluster with some density enhament around it , which could suggest an almost symmetric substructure at first sight
. however , with increasing resolution , this feature disappears , while at scales @xmath78 a core + clump structure emerges .
finally , at scale @xmath79 ( bottom right image ) just a central peak remains .
for the two intermediate scales i estimate the significance level of a model core + clump , having found @xmath80 ( @xmath81 ) and @xmath82 ( @xmath83 ) ( for 1,000 mc simulations ) , which suggests that such a model can not be rejected at these scales .
pixels , with @xmath77 pixels corresponding to 1.0 mpc .
the images present decreasing wavelet scales : @xmath75 , respectively .
[ fig5],title="fig:",width=226,height=226 ] pixels , with @xmath77 pixels corresponding to 1.0 mpc .
the images present decreasing wavelet scales : @xmath75 , respectively .
[ fig5],title="fig:",width=226,height=226 ] pixels , with @xmath77 pixels corresponding to 1.0 mpc .
the images present decreasing wavelet scales : @xmath75 , respectively .
[ fig5],title="fig:",width=226,height=226 ] pixels , with @xmath77 pixels corresponding to 1.0 mpc .
the images present decreasing wavelet scales : @xmath75 , respectively .
[ fig5],title="fig:",width=226,height=226 ] , width=321,height=321 ]
spheroidal clumps and rings ( and possibly other shapes ) may produce approximately the same results after spatial analysis .
however , in the case of comparing rings and clumps , there is a remaining property to be tested : the anisotropy of the signal around the cluster .
this second test is now introduced . for the projected distribution
, anisotropy can be probed by the reduced second - order moment measure @xmath84 of a point pattern ( e.g. illian et al . 2008 and references therein ) . in this work
, we estimate @xmath84 with using the library * spatstat * under r statistical package .
the routine * kmeasure * executes the following steps : 1 .
it takes a point pattern , 2 .
and forms the list of all pairs of distinct points in the pattern , 3 . then computes the vectors that join the first point to the second point in each pair , 4 . and treats these vectors as a pattern of ` points ' , 5 .
finally applyies a gaussian kernel smoother to them ( keeping the kernel width @xmath34 , see sections 4 and 5 ) .
the algorithm approximates the point pattern and its window by a binary pixel image , introduces a gaussian smoothing kernel and uses the fast fourier transform to make a density estimate @xmath85 .
the density estimate of @xmath85 is returned in the form of a real - valued pixel image .
the @xmath86 estimator is defined as the the expected number of points lying within a distance @xmath87 of a typical point , and with displacement vector having an orientation in the range @xmath88 $ ] .
this can be computed by summing the entries over the relevant region , i.e , the sector of the disc of radius @xmath89 centred at the origin with angular range @xmath90 $ ] .
hence , we can compute a measure of anisotropy , a , as integrals of the form @xmath91 recall that ripley function ( both in first and second order ) is used to test the hypothesis that a given planar point pattern is a realization of a poisson process . in the case of vectors instead of points ,
the second - order orientation analysis can be done around the center of the field .
a nearly flat anisotropy profile would be consistent with a poisson process @xmath92 .
thus , the idea here is to look for significant anisotropies in integral ( 7 ) for angular steps @xmath93 .
the choice of @xmath94 is arbitrary , and it was set to be @xmath95 , based in the following analysis .
we take 1,000 realizations of a controlled sample corresponding to a point pattern given by a poisson distribution plus a hernquist spheroid ( hernquist 1990 ) located at @xmath96 and making an angle of @xmath97 around the center . to justify our choice of @xmath98 we present in figure 7 the analyses for the case for @xmath99 . for @xmath100
, we still detect the peak , but now there are secondary peaks and a more noisy behaviour for @xmath101 . for @xmath102 ,
the peak is still there , but now it is less significant .
this result suggests that in the limit of too small @xmath98 we have a noisy anisotropy curve ( possibly with false peaks ) , while in the limit of very large @xmath98 the signal can be completely lost .
actually , in the approximate range @xmath103 $ ] the resultant anisotropy profiles are almost indistinguishable , with relative differences @xmath104 . ) .
solid lines for @xmath105 ; dotted lines for @xmath100 ; and dashed lines for @xmath102 .
[ fig],width=302,height=302 ] results of the anisotropy test for the averaged samples are presented in figure 8 .
note that model i ( although very simplified ) reproduces fairly well the anisotropy profile of real data , exhibiting a peak around @xmath106 ( or the supplementary angle @xmath107 ) , a direction corresponding to the axis crossing the nw quadrant . on the contrary , model
ii has a flatter anisotropy profile .
in fact , kolmogorov - smirnov test is consistent with the null hypothesis that data and model i samples come from the same distribution .
this is not true for the ks test comparing data and model ii samples , and model i and model ii samples ( see p - values in figure 8) .
i also have applied a welch two - sample t - test to find if the anisotropy profiles are stochastically higher between each other .
again , the results indicate that both data and model i have profiles significantly higher than that of model ii . also , there is not a significant difference between data and model i profiles ( see p - values in figure 8)
this all means that the galaxy field around cl 0024 + 17 is more consistent with an anisotropic cluster+clump distribution instead of a cluster+ring one .
morphology and anisotropy tests are not totally conclusive , but they suggest that the second scale found in data is more probable to be associated to the clump at @xmath3 nw from the cluster center than to a ring of galaxies , although this feature can not be discarded by spatial analysis alone . ,
by considering the planar galaxy distribution around cl 0024 + 17 as a spp , i have applied some tools of spatial statistics analysis to probe the existence of an extended structured in this cluster .
the catalog was corrected through 1,000 mc realizations of original the sample , following the completeness map of @xcite .
based on the besag l - function curves i have identified two distinct physical scales where the null hypothesis that the data came of a poisson process is rejected at 95% confidence .
values found for the two scales present normal distributions : @xmath108 and @xmath109 .
both bartlett and kruskal - wallis statistical tests indicate these normals are significantly unequal both in variances and means .
this reinforces the idea of two density peaks in the cl 0024 + 17 field .
also , the analysis of the averaged correlation function for all mc realizations suggests that at @xmath56 may be the maximum scale for the core of the system .
this scale is approximately at the intersection of the normal distributions found for the two transition scales .
the dm ring radius and the @xmath52 transition scale are approximately coincidents with the highest peaks of @xmath55 after @xmath56 . however , morphology and anisotropy tests based on simulations of 1,000 mock fields for cluster+clump and cluster+ring models are not conclusive about interpreting the second scale as a ring of galaxies . however , a wavelet analysis of the field points out that a cluster+clump model can not be rejected at scales @xmath81 pixels ( 0.50 mpc ) and @xmath83 pixels ( 0.25 mpc ) .
finally , the anisotropy test also suggests that the second scale found in data is more probable to be associated to the clump at @xmath3 nw from the cluster center than to a ring of galaxies .
summing up this work , the analyses presented in sections 4 , 5 and 6 are not indicative that galaxies follow dark matter around cl 0024 + 17 , but the methodology based on spatial patterns examination seems to have good sensitivity to find physical scales in galaxy systems ( both in radial and angular coordinates ) .
it could turn a useful tool to study galaxy ( and maybe other astrophysical ) systems in general .
spatial analysis ( followed by subsidiary morphology and anisotropy tests ) could be a poweful technique to probe remnant substructures of clusters collisions using galaxy data alone .
analysing a large dataset of galaxy clusters , we could identify candidates to bullet - type clusters by the presence of at least two transition scales . also , the approach can introduce a new estimator for galaxy systems radii based only on the scales detected in data @xcite .
finally , this method can be directly applied to data from n - body simulations of cluster collisions . some sensitivity to initial conditions
could be tested not only looking for bumps in the density projected along the collision axis , but also with the dark matter particles interaction range .
this is an interesting test to be done , since a recent study suggests that the ring - like feature of cl 0024 + 17 ( and similar cases ) would be associated to very improbable initial velocity distributions @xcite .
i thank the referee for very useful suggestions .
i am grateful to a.c . schilling and b. carvalho for helpful statistical discussions .
i also thank the support of cnpq , under grant 201322/2007 - 2 .
finally , i thank the fermilab for the hospitality .
baddeley , a. , analysing spatial point patterns in r , workshop notes ( 2008 ) baddeley , a. , moller , j. and waagepetersen , r. 2000 , statistica neerlandica , 54 , 329 baddeley , a. & turner , r. 2006 , in case studies in spatial point process modeling , lecture notes in statistics 185 , springer - verlag , new york , pp .
23 - 76 besag , j.e .
1977 , j. roy .
b , 39 , 193 broadhurst , t. , huang , x. , frye , b. & ellis , r. 2000 apj , 534 l15 clowe , d. , bradac , m. , gonzalez , a.h . ,
markevitch , a.v . , randall , s.w . , jones , c. & zaritsky 2006 , apj letters , 648 , l109 comerford , j.m . ,
meneghetti , m. , bartelmann , m. & schirmer , m. 2006 , apj , 642 , 39 czoske , o. , kneib , j .-
p . , soucail , g. , bridges , t.j . , mellier , y. & cuillandre , j .- c .
2001 , a & a , 372 391 czoske , o. , moore , b. , kneib , j .-
& sucail , g. 2002 , a & a , 386 , 31 davis , m . ,
efstathiou , g. , frenk , c. , & white , s.d.m .
1985 , apj , 292 , 371 hernquist , l. 1990 , apj , 356 , 359 humason , m.l . & sandage , a. 1957 , in carnegie yearbook 1956 ( washington , dc : carnegie inst .
washington ) , 61 illian , j. , penttinen , a. , stoyan , h. & stoyan , d. 2008 , statistical analysis and modelling of spatial point patterns , ed .
john wiley & sons ltd jee , m.j . , white , r.l . ,
benitez , n. , ford , h.c . ,
blakeslee , j.p . ,
rosati , p. , demarco , r. & illingworth , g.d .
2005a , apj , 618 , 46 jee , m.j . , white , r.l . , ford , h.c . , blakeslee , j.p . ,
illingworth , g.d .
, coe , d.a . & tran , k.v.h .
2005b , apj , 634 , 813 jee , m.j et al .
2007 , apj , 661 , 728 markevitch , m. , gonzalez , a.h . ,
david , l. , vikhlinin , a. , murray , s. , forman , w. , jones , c. & tucker , w. 2002 , apj , 567 , l27 markevitch , m. 2007 , phys . rep .
443 , 1 escalera , e. & mazure , a. 1992 , apj , 388 , 23 ota , n. , pointecouteau , e. , hattori , m. & mitsuda , k. 2004 , apj , 601 120 qin , b. , shan , h .- y . & tilquin , a. 2008 , apj 679 , l81 ribeiro , a.l.b . & lopes , p.a.a .
2009 , in preparation ripley , b.d .
1977 , j. roy .
b , 41 , 368 slezak , e. , bijaoui , a. and mars , g. 1990 , a & a , 227 , 301 tyson , j.a .
, kochanski , g.p . & dellantonio , i.p .
1998 , apj , 498 , l107 zhang , y .- y . , bhringer , h. , meiller , y. , soucail , g. & forman , w. 2005 , a & a , 429 , 85 zuhone , j.a . ,
lamb , d.q . &
ricker , p.m. 2008 , astro - ph/0809.3252 | i present a spatial analysis of the galaxy distribution around the cluster cl 0024 + 17 .
the basic aim is to find the scales where galaxies present a significant deviation from an inhomogeneous poisson statistical process . using the generalization of the ripley , besag , and the pair correlation functions for non - stationary point patterns , i estimate these transition scales for a set of 1,000 monte carlo realizations of the cl 0024 + 17 field , corrected for completeness up to the outskirts .
the results point out the presence of at least two physical scales in this field at @xmath0 and @xmath1 .
the second one is statistically consistent with the dark matter ring radius ( @xmath2 ) previously identified by jee et al .
( 2007 )
. however , morphology and anisotropy tests point out that a clump at @xmath3 nw from the cluster center could be the responsible for the second transition scale .
these results do not indicate the existence of a galaxy counterpart of the dark matter ring , but the methodology developed to study the galaxy field as a spatial point pattern provides a good statistical evaluation of the physical scales around the cluster .
i briefly discuss the usefulness of this approach to probe features in galaxy distribution and n - body dark matter simulation data .
astrophysics , galaxy clusters |
the study of qcd in the high energy limit has a history spanning at least twenty years and is still actively pursued .
within perturbation theory ( pqcd ) an understanding has emerged concerning hadronic processes in the semihard regime , such as near - forward elastic and diffractive scattering , small-@xmath0 deeply inelastic scattering and rapidity gap events .
the evolution of the partonic cascade stemming from the initial state hadrons is known to be described in the leading logarithmic approximation lla(x ) by the bfkl equation either in its original form @xcite or in its improved reformulations @xcite . however , the picture is yet incomplete because of two major problems encountered in this regime .
the first has to do with unitarity @xcite , which is violated by the lla(x ) result .
the second concerns the infrared sensitivity of the lla(x ) result , which is infrared finite but does receive contributions from low momentum regions , where physical observables are sensitive to large non - perturbative corrections .
it is the second problem that we shall consider in this paper . unlike the more conventional high momentum transfer ( @xmath8 ) processes with final state particles in the central rapidity region ,
the semihard kinematic regime is characterised by three ordered scales , @xmath9 .
note that the use of pqcd is justified by the second of the previous inequalities , in the absence of which , regge phenomenology of soft processes is the only available approach so far .
the pqcd factorisation theorem , applicable for high-@xmath8 processes , states that inclusive observables can be expressed as a product of parton distribution functions and a hard scattering cross section ( matrix elements and coefficient function in ope parlance ) . in the semihard regime
this theorem needs amendment . here
, the product is not only a convolution in longitudinal momentum fractions of the partons but also in their transverse momentum .
this is known as @xmath10-factorisation @xcite and in impact parameter space it translates to a convolution in impact parameter of the radiated secondary partons and an off - shell hard amplitude or cross section .
this leads us to the first of the two theoretical inputs of the present study , namely the use of the impact parameter or dipole emission formulation of semihard processes , as developed by mueller @xcite and independently by nikolaev and zakharov @xcite .
the merit of this formalism , apart from the simplicity of the final results , is that it organises the process in terms of sequential soft gluon emissions in the @xmath11-channel , which in the coulomb gauge admit a clear physical interpretation .
they are to be understood as initial state radiation occurring before the hard scattering .
final state interactions are not included in the evolution equation because they only lead to a unitarity rearrangement that does not affect inclusive observables .
to study the infrared sensitivity of a semihard process we are bound to consider a subset of next - to - leading logarithmic corrections , specifically the ones that have to do with the running coupling . in the bfkl approach , the evolution kernel , although infrared finite , is sensitive to the running coupling because of diffusion towards the infrared of the transverse momentum along the exchanged gluon ladder . with specific assumptions about the infrared dynamics ( i.e. long wavelength gluon propagation ) and the scale of the running coupling , the effect on the regge trajectories has been analyzed in refs .
@xcite . in the present study
we take the point of view that in the dipole formalism , just like in timelike parton cascades , the scale of the running coupling is set by the virtuality of the emitted gluon .
the problem of translating this into impact parameter space is resolved through the use of the borel transformation .
we require that the dipole evolution kernel has the same borel singularity structure as the one obtained in momentum space .
this condition is sufficient to determine the scale of the coupling as a function of the impact parameters of the emitted gluon .
therefore , the second theoretical input in our approach is renormalon analysis @xcite . once the dipole kernel with running coupling
is constructed , we study the power corrections that are generated when this kernel is used to evolve the wave function of an initial hadronic state such as a virtual photon in small-@xmath0 d.i.s .
the structure of this paper is as follows . in section 2
we highlight the main concepts and results of the dipole formalism . here
we also review the derivation of the dipole evolution equation with running coupling . in section 3
we discuss the borel singularity structure of the dipole density .
we find a series of infrared renormalon singularities at @xmath12 , @xmath13 as well as a leading singularity at @xmath14 , with @xmath15 the anomalous dimension of the solution .
we show that the borel image of the dipole kernel is analytic in the neighbourhood of @xmath2 and compute its action as a power series expansion in @xmath3 . in the limit @xmath16
the bfkl result is recovered .
iterations of the kernel turn the renormalon poles into branch cuts but do not shift their positions . in section 4
we discuss small-@xmath0 d.i.s . in the dipole formalism using the results derived previously .
the main conclusion is that the leading borel singularity is at @xmath17 , which implies @xmath5 power corrections consistent with the wilson ope expectation .
the dipole evolution equation for fixed coupling can be solved by virtue of its scale invariance and leads to the same spectrum as that of the bfkl equation . for the borel image
this scale invariance is necessarily lost and the resulting evolution equation can only be solved in numerical approximation by successive iterations . in section 5
we present results for the first few iterations of the borel transformed kernel in deeply inelastic scattering , and demonstrate that the singularity conforms with that discussed in section 4 .
we summarise in the final section .
we consider small-@xmath0 d.i.s . as a specific example of a semihard process and
describe it in the rest frame of the nucleon target of mass @xmath18 .
let @xmath19 denote the wave function for @xmath20 to fluctuate into a @xmath21-@xmath22 dipole of transverse size @xmath23 and with the quark carrying longitudinal momentum fraction @xmath24 .
then , the virtual photoabsorption cross section @xmath25 for given polarisation of @xmath20 @xmath26 , can be written as @xcite @xmath27 where @xmath28 denotes the dipole - nucleon total cross section and @xmath29 is the large rapidity parameter to be resummed within pqcd .
the physical interpretation of eq .
( [ impfac ] ) is that the transition @xmath30 , with light cone time scale @xmath31 , occurs much earlier than subsequent interactions with the nucleon target .
the lowest order in @xmath32 dipole cross section @xmath33 is @xmath34 with @xmath35 proportional to the absorptive part of the gluon - gluon - nucleon - nucleon vertex function .
the lla(x ) radiative corrections are generated by emission of soft gluons by the initial @xmath21-@xmath22 dipole with longitudinal momentum fractions strongly ordered as @xmath36 the light cone time scale for the @xmath37-th gluon emission is @xmath38 , therefore it is emitted independently of the previous ones @xmath39 . the @xmath20 wave function @xmath40 for the emission of @xmath41 soft gluons with longitudinal momentum fractions @xmath42 and at transverse positions
@xmath43 can be calculated order by order in pqcd and the inclusive probability distribution for @xmath20 to fluctuate into @xmath21-@xmath22 plus soft radiation , @xmath44 , is obtained by the square of @xmath45 integrated over the phase space of the emitted gluons and summed and averaged over their polarisations @xmath46 and colours @xmath47 .
@xmath48 here and until section 4 we suppress the @xmath20 polarisation indices @xmath49 for brevity .
indeed it is the projection of these polarisation states for the virtual photon that is responsible for the fact that @xmath50 is taken to be a function of @xmath23 only . in the coulomb
gauge the soft radiation can be viewed as a cascade of colour dipoles emanating from the initial @xmath21-@xmath22 dipole .
it is useful to introduce the dipole density @xmath51 @xcite , for emission from the initial @xmath21-@xmath22 dipole of size @xmath23 of a dipole of size @xmath52 and with the smallest longitudinal momentum fraction in the emitted dipole bounded from below by @xmath53 .
the dipole density is independent of the projectile bound state properties .
when convoluted with the @xmath20 wave function squared @xmath54 it generates the number density @xmath55 of dipoles of size @xmath52 and rapidity larger than y as @xmath56 the whole dipole cascade can be constructed from the repeated action of a kernel @xmath57 on the initial density @xmath58 through the dipole evolution equation @xmath59 with driving term @xmath60 the evolution kernel @xmath57 can be calculated from the one soft gluon emission probability @xmath61 .
the result is @xmath62 where hatted vectors will always denote the distance from the second parent emitter , @xmath63 .
the full kernel @xmath57 is @xmath64 the first term of @xmath57 comes from the one loop virtual corrections to the initial @xmath21-@xmath22 dipole of size @xmath23 .
the @xmath65-function denotes the absence of gluon radiated into the final state .
the second term corresponds to real emission and can be read off from eq .
( [ phi1 ] ) taking into account eq .
( [ dipev ] ) . a factor of 2
is included here to account for the fact that after one dipole splits into two , each one of the offsprings can act as parent for subsequent emissions .
the change of integration variables @xmath66 generates the ( triangle ) jacobian @xcite @xmath67 unlike the bfkl kernel , both virtual and real parts of the dipole kernel @xmath57 are separately ir finite in the limit @xmath68 . for fixed @xmath32 and in the limit @xmath69 , for which @xmath70 , @xmath57 is conformally invariant and has the same spectrum as the bfkl kernel , i.e. @xmath71 where @xmath72 is the bfkl spectral function .
indeed , the two approaches lead to the same phenomenological results for inclusive observables .
the equivalence between the dipole and bfkl formalisms at the level of light cone perturbation theory diagrams has been shown explicitly in ref .
@xcite . after the introduction of the dipole density @xmath41 , the dipole cross section @xmath73 of eq .
( [ impfac ] ) takes the form @xmath74 eqs .
( [ impfac ] , [ sigman ] ) provide the factorised form of small-@xmath0 d.i.s . in impact parameter space .
we note that dependence on target mass scale @xmath75 enters through @xmath33 and that the universal part of the photoabsorption cross section @xmath25 is indeed the dipole density @xmath41 , which is independent of projectile or target . the problem of how to include the running coupling in the dipole evolution kernel @xmath76 has been studied in ref .
the form of the kernel with running coupling is fixed by the following two requirements .
the first is that in the emission probability of a soft gluon with transverse momentum @xmath77 the coupling runs as @xmath78 .
the second is that the emission probability in impact parameter space should generate the same singularities in borel space as the borel transformed emission probability in momentum space .
then the form of the one - dipole emission probability @xmath61 is completely determined .
explicitly it is @xmath79 \hat{r}_1 ^ 2 \right .
\nonumber \\ & \quad & \quad \quad \quad \quad \left .
+ \left [ \alpha_s\left(\frac{\lambda^2}{\hat{r}_1 ^ 2}\right ) -\alpha_s \left ( \frac{\lambda^2}{r_1 ^ 2 } \right ) \right ] r_1 ^ 2 \right\ } \ , .
\label{phirun } \end{aligned}\ ] ] where @xmath80 in the limit of fixed @xmath32 the above equation reproduces the expression for the one - dipole emission @xmath81 of eq .
( [ phi1 ] ) .
it turns out that in impact parameter space the scale of the coupling is not a simple function of the distances @xmath82 , @xmath83 but is weighted with three further parameters @xmath84 , @xmath85 and @xmath86 .
the borel image @xmath87 is defined , as usual , by the transformation @xmath88 this transformation can be easily inverted when the one - loop running coupling is used .
then the borel image @xmath89 of @xmath90 is @xmath91 and from eqs .
( [ phirun],[bortransphi ] ) we obtain @xmath92 \ , .
\label{borelphi } \end{aligned}\ ] ] the integrals over @xmath93 , @xmath84 , @xmath85 and @xmath86 can be performed and the result is @xmath94 from eqs .
( [ borelphi ] , [ renormalon ] ) we can construct the dipole evolution kernel in borel space @xmath95 .
it consists of two pieces .
the virtual contribution is the coefficient of the @xmath54 in eq .
( [ renormalon ] ) with the sign inverted as required by unitarity @xcite .
the real contribution is obtained by the coefficient of @xmath54 in eq .
( [ borelphi ] ) . after renaming @xmath96 , changing integration variables as in eq .
( [ chvars ] ) and performing the integration over @xmath97 we obtain the explicit form of the kernel .
it is @xmath98 where @xmath99 and @xmath100 .
finally , the evolution equation for the dipole density in borel space reads @xmath101 with boundary condition @xmath102
to study the borel singularity structure generated by the dipole kernel @xmath103 we consider its action on a test function of the form @xmath104 . defining the function @xmath105 by @xmath106
we obtain the expression of @xmath105 from eq .
( [ runker ] ) , @xmath107 \ , , \label{chiform } \end{aligned}\ ] ] where the first term comes from virtual correction and the second from real emission .
the virtual contribution to @xmath108 contains a series of poles at @xmath109 which are identified with the ir renormalons and correspond to power corrections of @xmath110 , @xmath111 note that these poles are independent of the specific form of the test function .
this set of poles , resulting from the exponentiation of soft radiation , has also been derived in the context of the drell - yan process in ref . @xcite .
in addition , we observe the presence of a series of poles at @xmath112 , @xmath113 generated by the @xmath114 dependence of the real contribution to @xmath105 . for @xmath115
these poles correspond to ir renormalons for @xmath116 .
their ir origin is established by observing that these singularities arise from the @xmath117 integration region of eq .
( [ dipevol ] ) , where the offspring dipole is emitted with size larger than the parent dipole . taken at face value the @xmath118-dependent poles indicate the presence of @xmath119 power corrections . for the moment it suffices to note that in this section we are studying the dipole cascade independently of its embedding in a particular physical process .
the interpretation of the @xmath118-dependent power corrections in terms of ope matrix elements is not an issue here . inspecting the evolution equation ( [ dipevol ] )
we now examine whether the kernel introduces singularity at the lower end of the @xmath120-integration .
the answer is that for fixed @xmath23 the limit @xmath16 of @xmath121 is not singular .
the @xmath2 pole appearing in the virtual contribution to @xmath103 is of uv origin and cancels against a corresponding singularity in the real contribution .
the presence of such a singularity can be seen already in the fixed coupling case .
the integral in eq .
( [ phi1 ] ) is logarithmically divergent in the region @xmath122 .
one way of tracing the cancellation of the uv singularities is to impose a lower cutoff in the impact parameter integration as in ref .
alternatively , in the case of running coupling we can use the borel parameter @xmath3 as a dimensional regulator . defining @xmath123
we note that the uv divergences in the real part of @xmath103 arise from the region @xmath124 .
this is the region where the gluon is emitted at impact parameter almost equal to that of the parent dipole . in the limit
@xmath125 the only uv divergent contribution comes from the first term inside the curly brackets of eq .
( [ runker ] ) .
it is due to the divergence of the hypergeometric function @xmath126 at @xmath16 . to parametrise this divergence we use the linear transformation
@xcite @xmath127 then , in the limit @xmath16 the kernel takes the form @xmath128 the singular factor @xmath129 can be expressed in terms of distributions as @xmath130_+ + ... \label{plusdist}\ ] ] substituting this into eq .
( [ kerzeta ] ) it follows that the uv singular term in the virtual part of @xmath103 cancels against the one in the real part .
since the kernel is regular around @xmath2 we compute the power series expansion in @xmath3 of @xmath108 . from eq .
( [ chiform ] ) we obtain @xmath131 where @xmath132 the @xmath133 singular terms have cancelled as anticipated and the @xmath134 term is the bfkl spectral function @xmath135 , eq .
( [ chifn ] ) . returning to the dipole evolution equation ( [ dipevol ] ) we introduce the usual laplace
transform with respect to the rapidity @xmath136 @xmath137 and the anomalous dimension @xmath118 via the mellin transform @xmath138 then eq .
( [ dipevol ] ) with the boundary condition ( [ bc ] ) through the use of ( [ chidef ] ) becomes an equation for the spectral amplitudes @xmath139 , @xmath140 this equation is universal for all semihard processes .
the scale @xmath141 is only constrained by the requirement @xmath142 and its specific value is otherwise arbitrary . it is identified with the @xmath20 virtuality here because we consider the specific case of small-@xmath0 d.i.s .
the dipole density is reconstructed from the spectral amplitudes as @xmath143 eq .
( [ volterra ] ) is an integral equation of the volterra type .
such equations are known not to have eigenfunctions for bounded kernels .
one way of constructing solutions is by iteration . in our case
, the iterative solution is given by the formal expression @xmath144 with @xmath145 and @xmath146 for @xmath147 .
the index @xmath37 counts the number of iterations of the kernel in eq .
( [ volterra ] ) . having identified the ir renormalon poles of @xmath105 , eq .
( [ chiform ] ) , we can study the borel singularities of the spectral amplitude @xmath139 from eq .
( [ formsol ] ) .
the contribution to @xmath148 from the virtual ( @xmath118-independent ) terms only is @xmath149 because the @xmath3-integrals are nested , for @xmath150 no new singularity is introduced in the solution from these virtual corrections . for @xmath151
some of the @xmath152-function poles will be encountered on the @xmath3-integration paths turning the leading @xmath153 pole into a branch cut .
real contributions introduce singularities on the positive @xmath3-semiaxis for real @xmath118 , and for @xmath154 , the leading one is found from eqs .
( [ init ] , [ formsol ] ) to be at @xmath155 .
we note that for @xmath156 the real contributions to @xmath157 will not generate singularities because none of the poles of the @xmath158-functions in eq .
( [ formsol ] ) are encountered along the nested @xmath3-integration paths . for @xmath159 , as in the case of virtual contributions , poles are found along some of the @xmath3-paths turning the leading @xmath160 pole into a branch cut . from eqs .
( [ chiform ] , [ formsol ] ) it follows that there are also singularities for negative values of @xmath3 .
this means that @xmath155 is the position of the leading singularity for positive @xmath3 _ only _ when @xmath118 is in the range @xmath15 .
if we were to take @xmath118 in the range @xmath161 , then the leading ir singularity would be at @xmath162 , and the singularities for negative @xmath3 would be displaced further to the left on the negative @xmath3-semiaxis .
although these singularities do not affect the estimate of the power corrections from the non - perturbative effects , they reflect the fact that as @xmath118 increases the dipole densities generated from these test functions become increasingly ir divergent .
the above argument can be repeated for the contributions to the solution from cross products of virtual terms for some of the @xmath158-functions and real terms for the rest .
the net result of this analysis is that the region of analyticity in borel space of @xmath139 with @xmath15 contains the interval @xmath163 and the leading singularity is a branch cut at @xmath164 . in the asymptotic limit @xmath165 the anomalous dimension is @xmath166 with @xmath167 the spectral parameter to be integrated over in eq .
( [ inverse ] ) .
such a @xmath168 implies the presence of power corrections of @xmath169 for the dipole density @xmath51 .
these @xmath170 power corrections have also been reported in the context of conventional bfkl approach in ref .
@xcite .
the evolution equation ( [ dipevol ] ) is not conformally invariant because the action of the dipole kernel results in a shift of @xmath118 by @xmath171 in eq .
( [ volterra ] ) .
however , having established above a region of analyticity in borel space that contains @xmath2 we can use the @xmath3-expansion as a measure of deviation from the conformal limit . from eqs .
( [ chiexpand ] , [ formsol ] ) the @xmath3-expansion of the spectral amplitude generated after @xmath37 iterations of the kernel is @xmath172 \right .
\nonumber \\ & \quad & \hspace{2 cm } \left .
-\frac{b^k}{2 } \frac{(k+1)}{(k-1 ) ! } \beta_0 \chi(\gamma)^{k-1 } \frac{{{\mathrm d}}\chi(\gamma)}{{{\mathrm d}}\gamma } + { \mathcal o}(b^{k+1 } ) \right\ } \ , .
\label{bexp } \end{aligned}\ ] ] the full spectral amplitude @xmath173 , defined in eq .
( [ itseries ] ) , can be constructed in the small-@xmath3 region from summing the powers of @xmath3 .
summation of the leading powers of @xmath3 , i.e. retaining only the @xmath174 terms in eq .
( [ bexp ] ) is required in the region @xmath175 this ` resummation ' of the @xmath176 terms is well defined if the @xmath177 terms are small relative to the @xmath174 terms in eq .
( [ bexp ] ) . for the @xmath178 term
this translates to the requirement @xmath179 hence , leading @xmath3-power resummation is valid in the conformal limit of fixed coupling .
in this case the borel transform can be readily inverted in terms of a fixed @xmath32 to yield the well known answer for the bfkl spectral amplitude @xmath180 = \frac{2}{\omega - n_c \alpha_s \chi(\gamma)/\pi } \ , , \label{conflim}\ ] ] and via eq .
( [ inverse ] ) the asymptotic solution for the dipole density is obtained @xcite @xmath181 } { \sqrt{(7/2 ) \alpha_s n_c \zeta(3 ) y } } \ , \exp \left ( -\frac{\pi \ln^2(r/\rho)}{14 \alpha_s n_c \zeta(3 ) y } \right ) \ , , \label{bfklpom}\ ] ] with @xmath182 .
the structure functions in the small-@xmath0 region can be written in the following form @xmath183 using the definition of the dipole - nucleon cross section , eq .
( [ sigman ] ) , the above expression becomes @xmath184 this equation is the factorisation theorem for the structure function in the small-@xmath0 regime . as mentioned in the introduction
, it contains in addition to convolution in the longitudinal momentum fraction a convolution in impact parameter .
@xmath54 is the lowest order transition probability for @xmath185 .
( this is the lowest order impact factor of ref .
it can be calculated reliably in perturbation theory in the limit @xmath186 .
the dipole density @xmath51 contains the universal evolution of the initial @xmath21-@xmath187 dipole into a cascade of dipoles with ordered rapidities .
this is also calculable in perturbation theory as we saw in the previous section .
finally @xmath188 is the cross section for the absorption of a dipole at impact parameter @xmath52 by the nuclear target .
this is a non - perturbative quantity that normalises the structure function and introduces dependence on the mass scale characteristic of the nucleon target .
so far we have derived the evolution equation that determines the dipole density @xmath51 . for completeness
let us briefly review the calculation of @xmath54 defined in eq .
( [ phidef ] ) . in momentum space and to @xmath189 the @xmath20 wave function is @xcite @xmath190 where @xmath191 and @xmath192 are the spinors for the outgoing fermions , @xmath193 is the polarisation vector of @xmath20 and @xmath194 is the @xmath195 green s function @xmath196 with @xmath197 the electric charge of the quark .
performing the numerator algebra and the @xmath198 integration we obtain @xmath199 where @xmath200 after taking the impact parameter transformation defined by @xmath201 and squaring as in eq .
( [ phidef ] ) we find @xmath202 \mu^2 k_1(\mu r)^2 + m_q^2 k_0(\mu r)^2 \right\ } \
, , \nonumber \\
\phi^{(0)}_l(z , r ) & = & \frac{8 n_c e^2 z_q^2}{(2 \pi)^2 } z^2(1-z)^2 q^2 k_0(\mu r)^2 \ , .
\label{phi0 } \end{aligned}\ ] ] these expressions have also been derived in ref .
@xcite from the imaginary part of the @xmath203 forward amplitude . the factorised expression for the structure functions @xmath204 , eq . ( [ struct1 ] ) , allows us to study the effects of soft radiation in the small-@xmath0 region beyond the leading power of @xmath8 . to this end
we shall follow the standard lore of renormalon analysis @xcite .
this means that we shall study the singularity structure of the borel image @xmath205 defined in the usual way , @xmath206 we emphasise that even in the leading power in @xmath8 analysis the structure functions always contain a non - perturbative component , denoted by @xmath207 here .
however , we are looking for the borel singularities that are generated in the subasymptotic @xmath8 regime by the components that are calculable in perturbation theory , @xmath54 and @xmath41 in our case .
this is how perturbation theory is assumed to signal the presence of power corrections . inspecting eq .
( [ struct1 ] ) we see that the source of borel singularities is the dipole density @xmath41 .
@xmath54 has no @xmath32 dependence and therefore it stays unaffected by the borel transformation .
it affects the number density of the produced dipoles @xmath208 by determining the distribution of the initial @xmath21-@xmath187 dipole in impact parameter ( transverse size ) space .
@xmath54 contains infrared regulator set by the scale @xmath209 , eq .
( [ mu ] ) .
this follows from eqs .
( [ phi0 ] ) and the asymptotic behaviour of the bessel-@xmath210 functions for @xmath211 , @xmath212 note that the leading term in this asymptotic expansion is independent of the @xmath20 polarisation .
since @xmath213 is dominated by values of @xmath24 away from the end points @xmath214 and @xmath215 , we can think of the initial @xmath216 fluctuation as having size of @xmath217 .
the object that regulates the emission of dipoles of large sizes at the end of the cascade is the dipole - nucleon cross section @xmath207 .
indeed , dipoles of size @xmath218 , with @xmath219 the nucleon size , will not couple to the target . from this discussion
we can formulate the problem of the borel singularities of the small-@xmath0 structure functions as follows .
the borel image of the structure functions is given by @xmath220 where @xmath221 is a solution to the evolution equation @xmath222 with kernel given in eq .
( [ runker ] ) and boundary condition @xmath223 in the previous section we studied the action of the kernel @xmath224 on the test functions @xmath104 , see eqs .
( [ chidef ] , [ chiform ] ) . then these functions where used as a basis for decomposing the dipole density in terms of the spectral amplitudes , eq .
( [ anomdim ] ) , and we observed that this leads to @xmath118-dependent ir renormalons . if this procedure were applicable for constructing the structure functions @xmath225 it would yield a leading ir renormalon at @xmath226 , which could be to the left of unity for some @xmath227 , corresponding to power corrections of @xmath228 with @xmath229
. this would be in contradiction with the standard ope expansion of the structure functions , which predicts leading power corrections of @xmath230 , coming from the next - to - leading twist operators . in the case at hand such
a contradiction does not arise for the following reason .
we have noted that ir renormalon singularities come from the integration region @xmath231 in eq .
( [ dipevol ] ) , i.e. from emission at large impact parameters .
this means that in eq .
( [ evsigma ] ) we are sampling large values of @xmath232 .
although @xmath233 can not be calculated in pqcd it is controlled by a nucleon wavefunction which is suppressed at large @xmath232 , indicating that the probability to find a sufficiently large primary dipole inside a nucleon is negligible .
we can model @xmath207 ( as we shall do in the next section ) by assuming that it has the same asymptotic impact parameter dependence as that of a virtual photon and mass scale @xmath234 .
such a functional form , see eq .
( [ asyk ] ) , can not be decomposed in the @xmath104 basis .
we emphasise that this does not mean that ir renormalons do not arise from real emission .
it means that the leading ir renormalon is anticipated at @xmath235 leading to @xmath236 corrections consistent with the wilson ope expectation .
we conclude this section with the following remarks .
the conventional ( or wilson ) ope expansion for large @xmath8 can not be used for small-@xmath0 resummation , as we noted in the introduction .
semihard processes involve two large scales and a generalisation of the wilson ope is required .
such an expansion would not only separate soft from hard particles but also high rapidity from low rapidity ones .
this type of formalism has been brought to a considerable degree of maturity by balitsky in ref .
@xcite , although it is still far from the point of being a computational algorithm , like wilson s ope . in the high energy ope ,
evolution equations are non - linear beyond the lla(x ) approximation , whereas eq .
( [ dipevol ] ) is linear .
this is because our evolution equation does not resum the full set of the next - to - leading logarithms of @xmath237 but only a subset that involves the running of the coupling .
hence , our solution is not expected to be unitary but it will parametrise sensitivity with respect to low transverse scales .
in this section we demonstrate by numerical means that the borel transform of a deeply inelastic scattering structure function is regular in the region @xmath238 and has a singularity at @xmath239 . in perturbation theory we can not calculate the cross section @xmath207 of eq .
( [ sigmabc ] ) for electron - proton scattering .
nevertheless we know that it must be controlled by the nucleon wave function , which vanishes rapidly ( assumed exponentially ) for sufficiently large impact parameter . a reasonable model , therefore , is to take the impact factor dependence calculated for the virtual photon , eq .
( [ phi0 ] ) . for the realistic case of d.i.s .
this is indeed a model , whereas if we were considering the scattering of two virtual photons ( onium - onium scattering ) then this would be exact , up to a factor arising from the integration over the longitudinal momentum fraction , @xmath24 , in the wavefunction of the target photon .
we require that the model nucleon wave function be normalisable and this leads us to choose the case of a longitudinal virtual photon since for the transverse photon the behaviour of the wave function for small impact parameter gives rise to an ultraviolet divergence associated with the photon wave function renormalisation .
the scale @xmath209 in eq .
( [ phi0 ] ) is set to a typical hadron scale of @xmath240 gev . to this end
we set the borel transform of @xmath241 in eqs .
( [ sigman ] , [ sigmabc ] ) to @xmath242 we consider this modeling of the dipole - nucleon cross section to be appropriate for studying the ir effects arising from emission and subsequent interaction of large size dipoles .
note that , as stated in the previous section , the large @xmath23 asymptotics of the @xmath20 wave function , which simulates the nucleon target here , is independent of polarisation .
so our choice of @xmath243 bessel function in the previous equation is convenient and plausible .
the form of the input functions prevent an analytic solution to eq .
( [ evsigma ] ) from being found so we perform the integrations over @xmath86 , @xmath232 and the convolution in @xmath120 numerically using standard quadrature methods .
more precisely we expand @xmath244 so as to bring the evolution equation in the form @xmath245 the first convolution of the kernel with the longitudinal input function , @xmath246 exhibits the structure we expect from section 4 , see fig .
( [ conv_1_b ] ) .
we observe that the leftmost singularity in the borel plane appears at @xmath247 and we identify this as the leading ir renormalon .
there is no evidence of any singular behaviour in the region @xmath248 which would have been in contradiction with the ope expectation .
no singularity emerges as @xmath16 either , due to the cancellation between real and virtual parts of the kernel as discussed in section 3 . on the other hand ,
the expected leading singularity at @xmath249 can be seen clearly from the right hand graph of fig .
( [ conv_1_b ] ) . for the first iteration of the borel transformed kernel ,
this singularity occurs as a pole and it is generated entirely by the virtual correction contribution to @xmath250 .
there is a distinct difference between the asymptotic behaviour of @xmath251 and @xmath252 as @xmath253 , fig .
( [ conv_0_1_x ] ) . whereas @xmath251 decays exponentially at large @xmath254 , eq .
( [ asyk ] ) , the asymptotic behaviour of @xmath252 is : @xmath255 examination of the kernel reveals that @xmath250 scales as @xmath256 at large @xmath257 and this is the scaling behaviour which dominates in the convolution to determine the behaviour of @xmath258 . for intermediate values of @xmath254 , the data obtained can be fitted accurately if we include logarithmic correction terms and we find @xmath259 where @xmath260 , @xmath261 and @xmath262 are fit parameters .
this fit is exhibited in fig .
( [ conv_0_1_x ] ) . in section 4 , we demonstrated that the exponential nature of @xmath263 at large @xmath254 was responsible for the borel singularity structure observed in the first convolution .
the power like behaviour of @xmath264 at large @xmath254 does not imply that new singularities will appear in the borel plane to the left of @xmath247 .
it is the action of the convolution in @xmath120 in each subsequent iteration which guarantees this . for large @xmath232 , the convolution integral scales as @xmath265 from the kernel , @xmath266 from the behaviour of @xmath267 and
the integration introduces a factor of @xmath23 ; the result being that the second convolution also exhibits a power dependence of the form given in eq .
( [ asyiter1 ] ) for large @xmath23 .
indeed this large @xmath23 behaviour is sufficient to predict the position of the leading borel plane singularity for subsequent iterations .
since the infrared singularities are determined by the infrared ( large @xmath23 ) behaviour of the function on which the kernel acts , we can consider @xmath268 then , from eq .
( [ chiform ] ) we see that @xmath269 has a leading singularity at @xmath247 , which is where the leading singularity will be found for _ all _ subsequent iterations .
this is demonstrated in fig .
( [ conv_2_b ] ) .
we also note that the convolution in @xmath3 leads to a linear fall off as @xmath270 . for further iterations of the kernel
it is the intermediate @xmath271 range which will be of most significance .
although the leading singularity occurs again at @xmath247 as expected from the discussion above , the shape of the distribution is broader .
this is because for the second and subsequent iterations the leading singularity is converted into a cut with branch point at @xmath272 and receives singular contributions from both the virtual correction and real emission part of the kernel .
the generation of each iteration is a numerically intensive task . after generating a sufficient number of iterations
, we will be in a position to solve eq .
( [ d3 ] ) and construct the structure functions numerically . as explained above , we do not expect the position of the leading singularity to shift .
however , the behaviour of the wave function in the vicinity of the singularity contains important information about the contribution to structure functions from the infrared regions of transverse momentum space . we will report on findings for the structure functions in a future publication
although there exist numerous parametrisations of experimental data for semihard processes that include non - perturbative effects in plausible ways , here we have attempted to study in a systematic fashion the emergence of non - perturbative corrections as signaled by perturbation theory itself . for small-@xmath0 d.i.s .
we have worked within the dipole cascade formalism and introduced the running coupling in a self - consistent way . to study the resulting loss of scale invariance we considered the dipole evolution equation in borel ( @xmath3 ) space .
the variable @xmath3 provides a measure of deviation from the conformal limit @xmath2 , where we have shown that the evolution equation reproduces the well known bfkl result .
we have identified the borel singularities of the dipole density and have studied how these singularities change once the dipole density is convoluted with the dipole - nucleon cross section . in this case the evolution equation generates leading singularity which is a branch cut at @xmath247 .
this was established by numerical calculation for d.i.s . at small @xmath0 .
such a singularity indicates the presence of @xmath5 power corrections to the small-@xmath0 structure functions coming from the emission of dipoles of large transverse size .
these correspond to ` ladder gluons ' of small transverse momentum in the bfkl formalism .
hence our approach is suitable for studying the region of diffusion of transverse momentum towards the infrared .
moreover , it would be interesting to see how the scale dependence of the interactions in the dipole cascade modify the dependence of the structure functions on @xmath237 .
a. sen , phys .
d27 ( 1983 ) 2997 ; + j. c. collins and r. k. ellis , nucl .
b360 ( 1991 ) 3 ; + s. catani , m. ciafaloni and f. hautmann , nucl .
b366 ( 1991 ) 135 ; + s. catani and f. hautmann , nucl .
b427 ( 1994 ) 475 . | starting from the dipole representation of small-@xmath0 evolution we implement the running of the coupling in a self - consistent way .
this results in an evolution equation for the dipole density in borel @xmath1 space .
we show that the borel image of the dipole density is analytic in the neighbourhood of @xmath2 and that it is equal to the bfkl solution at @xmath2 .
we study the borel singularity structure of the dipole cascade emanating from a virtual photon at small @xmath0 and find a branch cut on the positive @xmath3-semiaxis starting at @xmath4 .
this indicates the presence of @xmath5 power corrections to the small-@xmath0 structure functions .
finally we present numerical results in the context of d.i.s .
pacs numbers : 11.55.j , 12.40.m 6.5 in 8.5 in 0.1 in 0.2 in 0.5 in 0.08 in shep 97 - 09 + um - th-97 - 12 + hep - ph/9705466 + 29 may 1997 + k.d .
anderson@xmath6 , d.a .
ross@xmath6 , m.g .
sotiropoulos@xmath7 . |
we deal with the system of conservation laws @xmath0 = 0.\ ] ] the unknown @xmath1 attains values in @xmath2 : @xmath3 and the flux function @xmath4 is defined as @xmath5 + \eta p_1(u ) \phantom{\int } } \\ v^2 \\
\displaystyle{4 \big\ { { v ( v-2)u } - ( v-1 ) w \big\ } + \eta p_3(u ) } \\
\end{matrix}\right).\ ] ] in the previous expression , the parameter @xmath6 attains values in the interval @xmath7 and to simplify the exposition we fix the functions @xmath8 and @xmath9 by setting @xmath10 note , however , that for some of the results discussed in the following the precise expression of the functions @xmath8 and @xmath9 is irrelevant .
system , was introduced by baiti and jenssen in @xcite and it was used to construct an example of a cauchy problem where the initial data have finite , but large , total variation and the @xmath11-norm of the admissible solution blows up in finite time .
more recently , the authors of the present paper used the baiti - jenssen system to exhibit an explicit counter - example which shows that schaeffer s regularity result for scalar conservation laws does not extend to systems , see @xcite .
the counter - example we construct shows , furthermore , that a wave - pattern containing infinitely many shocks can be robust with respect to perturbations of the initial data .
we refer to
[ s : motivation ] in the present paper for a brief overview of these counter - examples .
see also @xcite .
this note aims at establishing new quantitative interaction estimates for the baiti - jenssen systems , .
the estimates we obtain are pivotal to the analysis of the so - called wave front - tracking approximation of the cauchy problem obtained by coupling with an initial datum @xmath12 .
we refer to @xcite for an extended discussion on the wave front - tracking approximation . here
we only mention that the wave front - tracking algorithm is based on the construction of a piecewise constant approximation of the cauchy problem . under suitable conditions on the initial datum @xmath13 and on the flux function @xmath14
, one can show that the wave front - tracking approximation converges to an _ admissible solution _ of the cauchy problem , see in particular the analysis in @xcite . in
@xcite we construct wave front - tracking approximations of the cauchy problems obtained by coupling with suitable initial data .
we then rely on the wave front - tracking approximation to establish qualitative properties of the limit solutions . in the following
we do not consider all the possible interactions one has to handle when constructing the wave front - tracking approximation .
we only discuss those that we encounter in @xcite and that can not be handled by relying on straightforward considerations on the structure of the flux @xmath14 . before going into the technical details
, we make some further remarks .
first , in the present note we fix a very specific system in the wider class considered in @xcite .
the motivation for this choice is twofold : i ) it simplifies the notation and ii ) the analysis in the present note is sufficient for the applications in @xcite .
note , moreover , that in the proof of lemma [ l:22:heuristic ] we use ( although not in an essential way ) the exact expression of the function @xmath14 evaluated at @xmath15 . however , we are confident that our results can be extended to wider classes of systems of the type considered in @xcite .
second , in this note the only occurrences where we explicitly use the precise expression of the functions @xmath8 and @xmath9 is in the results discussed in [ ss : baj ] .
more precisely , the proofs of lemmas [ l:22:heuristic ] and [ l:12:heuristic ] both rely on a perturbation argument : we first show that our system verifies the statement of the lemmas in the case @xmath15 and then we show that the same holds provided @xmath6 is sufficiently small .
the proof of lemma [ l:12:heuristic ] is completely independent of the specific expression of @xmath8 and @xmath9 . in the perturbation argument in the proof of lemma [ l:22:heuristic ]
we use some results from [ ss : baj ] , but we never directly use the specific expression of @xmath8 and @xmath16 .
third , the baiti - jenssen system in not physical , in the sense that it does not admit strictly convex entropies , see @xcite for a proof .
it is natural to wonder whether or not the results established in the present note can be extended to physical systems . very loosely speaking , by combining lemmas [ l:12:heuristic ] and [ l:22:heuristic ] below with the analysis in @xcite we get the following statement . under suitable conditions ,
the only waves generated at the interactions between two shocks are shock waves , or , more precisely , no rarefaction waves are generated at the interaction between two shocks .
there are actually several physical systems that share this property : for instance , one can consider the @xmath17 example discussed by diperna in @xcite and assume that the data have sufficiently small total variation .
we refer to @xcite for the analysis of shock interactions for this system . on the other hand ,
a much more challenging question is whether or not there is any physical system that exhibit the same behaviors as those discussed in @xcite . in other words
, one can wonder whether or not a physical system can i ) exhibit finite time blow up or ii ) violate the regularity prescribed , for scalar conservation laws , by schaeffer s theorem . to the best of the authors knowledge , the answers to the above questions
is presently open .
we now give some technical details about the estimates we establish .
first , we point out that the baiti - jenssen system is strictly hyperbolic in the unit ball , which amounts to say that the jacobian matrix @xmath18 admits three real and distinct eigenvalues @xmath19 for every @xmath20 such that @xmath21 . also , if @xmath22 every characteristic field is genuinely nonlinear .
in other words , let @xmath23 denote the right smooth eigenvectors associated to the eigenvalues @xmath24 .
then @xmath25 for some suitable constant @xmath26 and for every @xmath27 and @xmath21 . in the following , we distinguish three families of shocks : we term a given shock 1- , 2- or 3-shock depending on whether the speed of the shock is close to @xmath28 , @xmath29 or @xmath30 .
we also point out that establishing interaction estimates for system boils down to the following .
consider the so - called riemann problem , namely the cauchy problem obtained by coupling with an initial datum in the form @xmath31 where @xmath32 , @xmath33 are constant states .
the above problem admits , in general , infinitely many distributional solutions : we term _ admissible _ the solution constructed by lax in the pioneering work @xcite , see
[ ss : lax ] for a brief overview . establishing interaction estimates for amounts to establish estimates on the admissible solution of the riemann problem - in the case when @xmath32 and @xmath34 satisfy suitable structural assumptions .
the first case we consider is the case of the interaction of two @xmath35-shocks , see figure [ f:2interazione ] , left part . in other words , we assume that there is a state @xmath36 such that * @xmath32 and @xmath37 are the left and the right states of a lax admissible 2-shock , * @xmath37 and @xmath34 are the left and the right states of a lax admissible 2-shock and * the shock between @xmath32 and @xmath37 has higher speed than the shock between @xmath37 and @xmath34 .
we now give an heuristic formulation of our interaction estimate and we refer to [ s:22 ] for the rigorous statement , which requires some technical notation . here
we only point out that the _ strength _ of a shock is a quantity defined in [ ss : lax ] which is proportional to the modulus of the difference between the left and the right state of the shock .
[ l:22:heuristic ] fix a constant @xmath38 such that @xmath39 and set @xmath40 .
consider the interaction between two 2-shocks and assume that the states @xmath32 and @xmath34 are sufficiently close to @xmath41 .
if the strengths of the interacting 2-shocks are sufficiently small , then the admissible solution of the riemann problem - is obtained by patching together a 1-shock , a 2-shock and a 3-shock .
we remark that the relevant point in the above result is that the solution of the riemann problem that we consider in the statement contains no rarefaction wave .
the second case we consider is the case of the interaction between a @xmath42-shock and a @xmath35-shock , see figure [ f:2interazione ] , right part .
in other words , we assume that there is a state @xmath36 such that * @xmath32 and @xmath37 are the left and the right states of a lax admissible 2-shock , * @xmath37 and @xmath34 are the left and the right states of a lax admissible 1-shock .
the case of the interaction of a @xmath43-shock with a @xmath35-shock is analogous .
we now give an heuristic formulation of our result and we refer to [ s:12 ] for the rigorous statement .
[ l:12:heuristic ] consider the interaction between a @xmath42-shock and a @xmath35-shock and assume both shocks have sufficiently small strength .
then the admissible solution of the riemann problem - is obtained by patching together a 1-shock , a 2-shock and a 3-shock .
also , we establish quantitative bounds from above and from below on the strength of the outgoing shocks , see formulas . note
that the fact that the three outgoing waves are shocks follows from the analysis in @xcite .
also , the bound from above on the strength of the outgoing @xmath43-shocks follows from by now classical interaction estimates , see ( * ? ? ?
* page 133 , ( 7.31 ) ) : the main novelty in lemma [ l:12:heuristic ] is that we have a new bound from below on the strength of the outgoing @xmath43-shock , see the left hand side of formula .
this estimate is important for the analysis in @xcite .
this note is organized as follows . in
[ s : overview ] we go over some previous results . in particular , in [ s : motivation ] we provide some motivation for studying the baiti - jenssen system by describing two counter - examples that use it . in
[ ss : lax ] we recall some results from @xcite and in [ ss : baj ] we apply these results to the baiti - jenssen system . in [ s:22 ]
we discuss the interaction of two @xmath35-shocks and in
[ s:12 ] the interaction of a @xmath42-shock and a @xmath35-shock .
for the reader s convenience , in this section we go over some previous results .
more precisely : * we discuss two counter - examples based on the baiti - jenssen system : the original one in @xcite and a more recent one devised in @xcite .
* we follow the famous work by lax @xcite and we outline the construction of the solution of the riemann problem . * we apply lax s construction to the baiti - jenssen system . this paragraph
is organized as follows : * we discuss the counter - example in @xcite * we discuss the counter - example in @xcite . before dealing with the specific examples , we recall two main features of the baiti - jenssen system : first
, it is strictly hyperbolic , namely holds .
note that strict hyperbolicity is a standard hypothesis for results concerning systems of conservation laws , see @xcite . also , if @xmath22 every characteristic field is genuinely nonlinear , which means that is satisfied for every @xmath27 .
this is a remarkable property because loosely speaking systems where all the characteristic field are genuinely nonlinear are usually better behaved than general systems .
for instance , the celebrated decay estimate by olenik @xcite , which applies to scalar conservation laws with convex fluxes , has been extended to systems of conservation laws where all the characteristic field are genuinely nonlinear , see for instance the works by glimm and lax @xcite , by liu @xcite and , more recently , by bressan and colombo @xcite , bressan and goatin @xcite and bressan and yang @xcite , while for balance laws we refer to christoforou and trivisa @xcite .
consider the general system of conservation laws @xmath44 = 0,\ ] ] where the unknown @xmath45 attains values in @xmath46 , the variables @xmath47 and the flux function @xmath48 is smooth and strictly hyperbolic .
consider furthermore the cauchy problem obtained by coupling with the initial condition @xmath49 under some further technical assumption on the structure of the flux , glimm @xcite established existence of a global in time solution of the cauchy problem provided that @xmath50 , the total variation of the initial datum , is sufficiently small . under the same assumptions , bressan and several collaborators established uniqueness results ,
see @xcite for a detailed exposition .
the requirement that the total variation @xmath50 is small is highly restrictive , but necessary to obtain well - posedness results unless additional assumptions are imposed on the flux function @xmath51 .
indeed , explicit examples have been constructed of systems and data @xmath13 where @xmath50 is finite , but large , and the admissible solution blows up in finite time .
in particular , in @xcite baiti and jenssen constructed an initial datum for system such that the @xmath11-norm of the admissible solution blows up in finite time .
the solution is admissible in the sense that it is piecewise constant and every shock is lax admissible . for further examples of finite time blow up ,
see the references in @xcite and @xcite . in @xcite schaeffer established a regularity result which can be loosely speaking formulated as follows .
consider a scalar conservation law with strictly convex flux , namely equation in the case when @xmath45 attains real values and @xmath52 is uniformly convex , i.e. @xmath53 for some constant @xmath26 .
the work by krukov @xcite establishes existence and uniqueness of the so - called _ entropy admissible solution _ of the cauchy problem posed by coupling and .
it is known that , even if @xmath13 is smooth , the entropy admissible solution can develop shocks , namely discontinuities that propagate in the @xmath54-plane .
schaeffer s theorem states that , for a generic smooth initial datum , the number of shocks of the entropy admissible solution is locally finite .
the word generic " is here to be interpreted in a suitable technical sense , which is related to the baire category theorem , see @xcite for the precise statement . in @xcite
we discuss whether or not schaeffer s theorem extends to systems of conservation laws where every characteristic field is genuinely nonlinear , namely holds .
note that the assumption that every characteristic field is genuinely nonlinear can be loosely speaking regarded as the analogous for systems of the condition ( which applies to scalar equations ) that the flux is strictly convex .
indeed , regularity results for scalar equations with strictly convex fluxes have been extended to systems where every characteristic field is genuinely nonlinear : as we mentioned before , this is the case of olenik s @xcite decay estimate , see for instance @xcite for possible extensions to systems .
also , the @xmath55 regularity result by ambrosio and de lellis @xcite , which applies to scalar conservation laws with strictly convex fluxes , has been extended to systems where every characteristic field is genuinely nonlinear , see @xcite . despite the above considerations , in
@xcite we exhibit an explicit example which rules out the possibility of extending schaeffer s theorem to systems of conservation laws where every characteristic field is genuinely nonlinear .
more precisely , we construct a big " set of initial data such that the corresponding solutions of the cauchy problems for the baiti - jenssen system develop infinitely many shocks on a given compact set of the @xmath56-plane .
the term big " is to be again interpreted in a suitable technical sense , which is related to the baire category theorem , see @xcite for the technical details .
we consider a system of conservation laws and we assume that @xmath57 is strictly hyperbolic and that every characteristic field is genuinely nonlinear , namely holds for @xmath58 .
lemma [ l : eige ] below states that the baiti - jenssen system satisfies these conditions .
the riemann problem is posed by coupling with an initial datum in the form @xmath59 where @xmath60 and @xmath61 are given states in @xmath2 . in @xcite ,
lax constructed a solution of the riemann problem - under the assumptions that the states @xmath60 and @xmath61 are sufficiently close : we now briefly recall the key steps of the analysis in @xcite .
we fix @xmath27 and @xmath62 and we define the _ @xmath63-wave fan curve _ through @xmath64 by setting @xmath65 : = \left\ { \begin{array}{ll } r_i [ s , \bar u ] & s \ge 0 \\ s_i [ s , \bar u ] & s < 0 \\ \end{array } \right.\ ] ] in the previous expression , @xmath66 is the _ @xmath63-rarefaction curve _ through @xmath64 and @xmath67 is the _
@xmath63-hugoniot locus _ through @xmath64 .
the @xmath63- rarefaction curve @xmath66 is the integral curve of the vector field @xmath68 , namely the solution of the cauchy problem @xmath69 = \bar u. \\
\end{array } \right.\ ] ] the @xmath63-th hugoniot locus @xmath67 is the set of states that can be joined to @xmath64 by a shock with speed close to @xmath70 . the @xmath63-hugoniot locus @xmath67 is determined by imposing the rankine - hugoniot conditions .
we term the value @xmath71 _ strength _ of the @xmath63-wave connecting the states @xmath64 ( on the left ) and @xmath72 $ ] ( on the right ) .
note that , owing to , when @xmath73 the @xmath63-wave is a @xmath63-th rarefaction wave , when @xmath74 the @xmath63-wave is an @xmath63-shock satisfying the so - called _ lax admissibility criterion_. the solution of the riemann problem - is computed by imposing @xmath75 \big ] \big]\ ] ] and by using the local invertibility theorem to solve for @xmath76 . from the value of @xmath76
one can reconstruct a solution of the riemann problem - , see @xcite for the precise construction .
this solution is obtained by patching together rarefaction waves and shocks that satisfy the lax admissibility criterion . in the following ,
we refer to this solution as the _ lax solution _ of the riemann problem - .
we collect in this paragraph some features of the baiti - jenssen system . for the proof
, we refer to @xcite .
the first result states that in the unit ball the baiti - jenssen system is strictly hyperbolic whenever @xmath77 . also , when @xmath22 all the characteristic fields are genuinely nonlinear . note that when @xmath78 this last condition is lost because two characteristic fields became linearly degenerate .
see @xcite or @xcite for the explicit computations .
[ l : eige ] assume that @xmath79 and that @xmath20 varies in the unit ball , @xmath80 .
then the baiti - jenssen system with flux is strictly hyperbolic , namely holds true .
if we also have @xmath22 then every characteristic field is genuinely nonlinear , namely is satisfied for @xmath81 we now discuss the structure of the wave fan curves .
we start by giving the explicit expression of the 1- and the 3-wave fan curve . in the statement of the following result , we denote by @xmath82 the components of the state @xmath62 . [
l : wavefan ] consider the flux function , assume that @xmath83 and fix @xmath62 such that @xmath84 .
then the following properties hold true . *
the 1-wave fan curve @xmath85 $ ] is a straight line in the plane @xmath86 , more precisely @xmath87 & = \bar u + \sigma \vec r_1 ( \bar u ) , \\\notag & \text{where } \quad \vec r_1 ( \bar u ) = \left ( \begin{array}{cc } 1\\ 0\\ \bar v\\ \end{array } \right ) .
\end{aligned}\ ] ] note that @xmath88 is the first eigenvector of the jacobian matrix @xmath89 .
also , the states @xmath64 ( on the left ) and @xmath85 $ ] ( on the right ) are connected by a wave which is * * a 1-rarefaction wave when @xmath90 , * * a lax admissible 1-shock when @xmath91 . *
the 3-wave fan curve @xmath92 $ ] is a straight line in the plane @xmath86 , more precisely @xmath93 & = \bar u + \tau \vec r_3 ( \bar u ) , \\\notag & \text{where } \quad \vec r_3 ( \bar u ) = \left ( \begin{array}{cc } 1\\ 0\\ \bar
v- 2 \\ \end{array } \right ) .
\end{aligned}\ ] ] the vector @xmath94 is the third eigenvector of the jacobian matrix @xmath89 . also , the states @xmath64 ( on the left ) and @xmath92 $ ] ( on the right ) are connected by a wave which is * * a 3-rarefaction wave when @xmath95 , * * a lax admissible 3-shock when @xmath96 .
note that , for the 3-wave fan curve , the _ positive _ values of @xmath97 correspond to shocks , the _ negative _ values to rarefaction waves .
this is the contrary with respect to and it is a consequence of the fact that we use the same notation as in @xcite and we choose the orientation of @xmath98 in such a way that when @xmath99 condition is replaced by the opposite inequality @xmath100 we now turn to the structure of the 2-wave fan curve . in the following statement , we use the notation @xmath101 also , we consider _ entropy admissible solutions _ of scalar conservation laws , in the krukov @xcite sense .
[ l : wavefan2 ] assume that @xmath20 is a lax solution of the riemann problem - .
then the second component @xmath102 is an entropy admissible solution of the cauchy problem @xmath103 = 0 \\ v(0 , x ) = \left\ { \begin{array}{ll } v^- & x < 0 \\ v^+ & x > 0
. \\ \end{array } \right .
\end{array } \right.\ ] ] also , we can choose the eigenvector @xmath104 and the parametrization of the 2-wave fan curve @xmath105 $ ] in such a way that the second component of @xmath105 $ ] is exactly @xmath106 .
we first rigorously state lemma [ l:22:heuristic ] [ l:22 ] there is a sufficiently small constant @xmath107 such that the following holds . fix a constant @xmath38 such that @xmath39 and set @xmath108
assume that @xmath109 .
\end{aligned}\ ] ] assume furthermore that @xmath110 \big].\ ] ] then there are @xmath91 and @xmath96 such that @xmath111 \big ] \big]\,.\ ] ] note that by combining with the inequalities @xmath91 , @xmath96 and @xmath112 we get that the three outgoing waves are all shocks .
the proof of lemma [ l:22 ] is organized as follows : * by relying on a perturbation argument , we show that the proof of lemma [ l:22 ] boils down to the proof of the taylor expansion .
* we complete the proof by establishing .
we start with some preliminary considerations .
assume that the states @xmath32 and @xmath34 satisfy .
next , solve the riemann problem between @xmath32 ( on the left ) and @xmath34 ( on the right ) : owing to @xcite , this amounts to determine by relying on the local invertibility theorem the real numbers @xmath113 , @xmath114 and @xmath97 such that @xmath115 \big ] \big].\ ] ] establishing the proof of lemma [ l:22 ] amounts to prove that @xmath116 and that @xmath91 , @xmath117 . to prove that @xmath118 we recall lemma [ l : wavefan2 ] and the fact that the @xmath102 component is constant along the 1-st and the 3-rd wave fan curves @xmath119 and @xmath120 .
we conclude that @xmath121 .
note that @xmath122 and @xmath123 are the second component of @xmath34 and @xmath124 .
we are left to prove that @xmath125 and @xmath117 .
we first introduce some notation : we regard @xmath113 and @xmath97 as functions of @xmath6 , @xmath126 and @xmath127 and @xmath32 and we write @xmath128 and @xmath129 to express this dependence . note that @xmath113 and @xmath97 depend on @xmath6 because the wave fan curve @xmath130 depends on @xmath6 .
owing to the implicit function theorem , the regularity of @xmath128 and @xmath128 is at least the same as the regularity of the functions @xmath119 , @xmath130 and @xmath120 .
also , note that the lax theorem @xcite ( see also @xcite ) states that the wave fan curves @xmath119 , @xmath130 and @xmath120 are @xmath131 .
the reason why we can achieve @xmath132 regularity is because we are actually considering the wave fan curves in regions where they are @xmath132 . to see this ,
we first point out that , owing to and , the wave fan curves @xmath119 , @xmath120 are straight lines and hence they are @xmath132 .
next , we point out that we are only interested in _ negative _ values of @xmath133 .
hence , we can replace the 2-wave fan curve @xmath130 defined as in with the 2-hugoniot locus @xmath134 .
we recall that the 2-hugoniot locus @xmath135 $ ] contains all the states that can be connected to @xmath64 by a shock , namely all the states such that the couple @xmath136 ) $ ] satisfies the rankine - hugoniot conditions .
the 2-hugoniot locus @xmath135 $ ] is @xmath132 and by combining all the previous observations we can conclude that @xmath128 and @xmath129 are both @xmath132 with respect to the variables @xmath137 .
next , we discuss the partial derivatives of @xmath128 and @xmath129 with respect to @xmath138 at the point @xmath139 . by arguing as in the proof of estimate ( 7.32 ) in @xcite we conclude that * for every @xmath32 , for every @xmath99 and every integer @xmath140 we have the following equalities : @xmath141 * for every @xmath32 and for every @xmath99 we also have the following equality concerning the derivatives of second order : @xmath142 this implies that @xmath143 and @xmath144 admit the following taylor expansions @xmath145 in [ ss:22etazero ] we prove that when @xmath15 and @xmath146 the functions @xmath113 and @xmath97 admit the taylor expansions [ e : mclaurin ] @xmath147 next , we use the lipschitz continuous dependence of the derivatives of third order with respect to @xmath6 and @xmath32 and we conclude that @xmath148 provided that @xmath149 and @xmath150 . in the above expression , @xmath151 denotes a universal constant . by plugging the above expressions into and recalling that @xmath152
we can eventually conclude that , if @xmath153 is sufficiently small , then @xmath154 the proof of the lemma is complete .
the proof of the taylor expansion is divided into two parts : * as a preliminary result we determine the structure of the hugoniot locus @xmath155 $ ] * we conclude the proof .
note that in this paragraph we always assume @xmath78 because formula deals with this case .
before giving the technical results , we introduce some notation .
first , we recall that we term @xmath156 the flux function @xmath14 in in the case when @xmath15 . in the following
, we will mostly focus on the behavior of the first and the third component of @xmath20 .
hence , it is convenient to term @xmath157 and @xmath158 the vectors obtained by erasing the second components of @xmath20 and @xmath156 , respectively .
we have the relation @xmath159 where we have also introduced the @xmath160 matrix @xmath161 . finally , we recall that we term @xmath162 $ ] the 2-hugoniot locus passing through @xmath64 , namely the set of states that can be connected to @xmath64 by a ( possibly not admissible ) shock of the second family . also , as usual we denote by @xmath163 , @xmath164 and @xmath165 the first , second and third component of @xmath64 , respectively .
we use the notation @xmath166 . [ ] [ l : twohugoniotlocus ] fix @xmath78 and assume that @xmath167 , then the 2-hugoniot locus through @xmath64 has the following expression : the second component of @xmath168 $ ] is @xmath169 while the first and third components are @xmath170= \widehat{\bar u } + { \boldsymbol{e}}(\bar v , s ) \widehat{\bar u } \end{split}\ ] ] where the @xmath160 matrix @xmath171 is @xmath172 by lemma [ l : wavefan2 ] the second component of @xmath173 $ ] is @xmath169 . to construct @xmath162
$ ] we use the rankine - hugoniot conditions , which are a system of 3 equations .
owing to lemma [ l : wavefan2 ] , the second equation reads @xmath174 and this implies that the speed @xmath175 of the 2-shock is @xmath176 we define the vector @xmath177 by setting @xmath178-\widehat{\bar u}\ ] ] and we point out that to establish lemma [ l : twohugoniotlocus ] we are left to show that @xmath179 the first and the third equations in the rankine - hugoniot conditions can be written as @xmath180-\widehat{j}(\bar v)\widehat{\bar u } , \end{split}\ ] ] where @xmath181 is the same as in .
next , we introduce the @xmath17 matrix @xmath182 and we rewrite as @xmath183 \widehat{\bar u } , \ ] ] which implies provided that @xmath184 \\ \end{split}\ ] ] by recalling that @xmath185 we can compute the explicit expression of the above matrices : @xmath186 the determinant of the matrix @xmath187 is @xmath188 and hence the matrix is invertible when @xmath189
. we can now complete the lemma by computing the explicit expression of @xmath190 , namely @xmath191 we are now ready to establish .
we first recall some notation : we consider the system of conservation laws with flux @xmath156 , see .
we consider the collision between two 2-shocks and we assume that @xmath192 , @xmath37 and @xmath34 are the left , middle and right states before the interaction .
this means that for some @xmath193 , @xmath194 we have @xmath195= d_2 \big [ s_2 , d_2 [ s_1 , u_{\sharp } ] \big ] \\ & = s_2 \big [ s_2 , s_2 [ s_1 , u_{\sharp } ] \big ] . \end{split}\ ] ] in the above expression , @xmath134 represents the @xmath35-hugoniot locus . to establish the last equality we used the fact that @xmath126 and @xmath127 are both negative .
we plug into and we use the equality @xmath196 : we arrive at @xmath197 + { \boldsymbol{e}}(s_1 , s_2 ) \big [ \widehat{u_{\sharp } } + { \boldsymbol{e}}(0 , s_1 ) \widehat{u_{\sharp } } \big ] \\ & = \widehat{u_{\sharp } } + \big [ { \boldsymbol{e}}(0 , s_1 ) + { \boldsymbol{e}}(s_1 , s_2 ) + { \boldsymbol{e}}(s_1 , s_2 ) { \boldsymbol{e}}(0 , s_1 ) \big ] \widehat{u_{\sharp}}. \end{split}\ ] ] next , we focus on the states after the interaction . by arguing as at the beginning of
[ ss:22etanonzero ] , we conclude that it suffices to determine @xmath198 and @xmath199 such that @xmath200 \big]\big].\ ] ] by the explicit expression of @xmath119 and @xmath120 and by applying lemma [ l : twohugoniotlocus ] we infer that the above equality implies @xmath201 + { \boldsymbol{e}}(0 , s_1 + s_2 ) \big [ \widehat{u_{\sharp } } + \sigma\widehat{\vec r_1}(0)\big ] + \tau \widehat{\vec r_3}(s_1 + s_2 ) \\ & = \widehat{u_{\sharp } } + { \boldsymbol{e}}(0 , s_1 + s_2 ) \widehat{u_{\sharp } } + \big [ { \boldsymbol{i}}+ { \boldsymbol{e}}(0 , s_1 + s_2 ) \big ] \sigma \widehat{\vec r_1}(0 ) + \tau \widehat{\vec r_3}(s_1 + s_2 ) \\ & = \widehat{u_{\sharp } } + { \boldsymbol{e}}(0 , s_1 + s_2 ) \widehat{u_{\sharp } } + \mathbf{h } ( s_1 + s_2 ) \left ( \begin{array}{cc } \sigma \\ \tau \\
\end{array } \right ) .
\end{split}\ ] ] in the previous expression we denote by @xmath202 and @xmath203 the vectors obtained from @xmath204 and @xmath205 by erasing the second component .
also , we introduced the matrix @xmath206 : its first column is @xmath207 { \widehat{\vec r_1 } ( 0)}$ ] , the second column is @xmath208 . in the following
, we will prove that @xmath209 is invertible provided that @xmath126 and @xmath127 are both sufficiently close to @xmath210 . by comparing and
we then obtain @xmath211}_{\displaystyle{{\boldsymbol{g}}(s_1 , s_2 ) } } \widehat{u_{\sharp}}.\ ] ] assume that we have established the following asymptotic expansion for @xmath212 : @xmath213 then by plugging both and @xmath214 into we obtain the asymptotic expansion . hence , to conclude the proof of we are left to establish .
first , we point out that , owing to the expression of @xmath190 in the statement of lemma [ l : twohugoniotlocus ] , @xmath215 this implies that when @xmath216 , the matrix @xmath217 vanishes and hence @xmath218 we compute now the asymptotic expansion of @xmath219 by directly computing the sum of the above matrices , we obtain that we can factor the term @xmath220 which multiplies the matrix with coefficients @xmath221 by combining the above computations we obtain the following asymptotic expansion : @xmath222 this establishes and hence concludes the proof of .
we first rigorously state lemma [ l:12:heuristic ] .
[ l:12 ] there is a sufficiently small constant @xmath107 such that if @xmath223 , then the following holds .
assume that the states @xmath32 , @xmath33 satisfy @xmath224 \big]\ ] ] for real numbers @xmath114 , @xmath113 such that @xmath225 furthermore , assume that @xmath226 .
then there are real numbers @xmath227 and @xmath228 such that @xmath229 \big ] \big]\,\ ] ] and @xmath230 note that implies @xmath231 and @xmath232 .
if we combine these inequalities with and @xmath233 we see that the three outgoing waves are all shocks .
establishing the proof of lemma [ l:12 ] amounts to establish . indeed , 1 .
by using lax s construction ( see [ ss : lax ] ) we determine @xmath234 such that @xmath235 \big ] \big].\ ] ] 2 . by combining , and lemma [ l : wavefan2 ] we obtain that @xmath236 . to establish we proceed as follows : * we establish in the case when @xmath78 .
* we conclude the proof by relying on a perturbation argument . more precisely , by using the fact that the flux @xmath14 in smoothly depends on @xmath6 we show that holds provided @xmath6 is sufficiently small .
note that , as we have mentioned in the introduction , the precise expression of the function @xmath8 and @xmath9 plays no role in the proof of lemma [ l:12 ] , what is actually relevant is that estimates hold at @xmath15 with strict inequalities and that @xmath6 is sufficiently small .
we establish in the case @xmath15 .
this part of the proof is actually the same as in @xcite , but for completeness we go over the main steps .
we term @xmath237 , @xmath238 the real numbers satisfying when @xmath78 .
let @xmath122 and @xmath239 denote the second components of @xmath34 and @xmath32 , respectively .
we term @xmath175 the speed of the incoming 2-shock ( which is the same as the speed of the outgoing 2-shock ) , we recall lemma [ l : wavefan2 ] and the fact that the second component varies only across 2-shocks .
we conclude that @xmath240 since by assumption @xmath241 and @xmath242 , then @xmath243 by imposing the rankine - hugoniot conditions on the incoming and outgoing 2-shocks and by arguing as in @xcite , with the choice @xmath244 , we arrive at the following system : @xmath245 if we set @xmath246 and @xmath247 then the above linear system can be recast as @xmath248 .
the explicit expression of the matrix @xmath249 is @xmath250 we solve for @xmath237 and @xmath238 and we obtain @xmath251 by using and the inequality @xmath252 , we obtain @xmath253 and this implies that the estimate holds true in the case when @xmath15 .
we are now ready to complete the proof of lemma [ l:12 ] .
we proceed as follows : * we make some preliminary considerations which reduce the proof of lemma [ l:12 ] to the proof of the fact that a certain map is a strict contraction .
* we conclude the proof by showing that the map is indeed a strict contraction .
we first introduce some notation .
we term @xmath37 the intermediate state _ before _ the interaction , namely @xmath254.\ ] ] also , we term @xmath255 and @xmath256 the intermediate states _ after _ the interaction , namely @xmath257 , \\ & u''_m : = d_2 [ s , u'_m ] = { d_3 \big [ -\tau ' , u_3 \big ] = d_3 \big [ -\tau ' , d_1 [ \sigma , u_m ] \big ] } \\ & { \qquad = d_3 \big [ -\tau ' , d_1[\sigma , d_2 [ s , u_\ell ] ] \big ] } \\ \end{split}\ ] ] next , we use ( * ? ? ?
* eq . ( 5.3)-(5.4 ) ) and we recast the rankine - hugoniot conditions for 2-shocks as a nonlinear system in the form @xmath258 where @xmath259 and @xmath260 are as in and , respectively . also , the vector @xmath261 is defined by setting @xmath262 and the nonlinear term @xmath263 is equal to @xmath264 in the above expression , the functions @xmath8 and @xmath9 are the same as in .
note , however , that the precise expression of @xmath8 and @xmath9 plays no role in the proof , the only relevant point is that @xmath8 and @xmath9 are both regular ( say twice differentiable with lipschitz continuous second derivatives ) .
note furthermore that we can regard @xmath265 as a function of @xmath261 , @xmath32 , @xmath114 and @xmath113 because , owing to and , @xmath37 , @xmath255 and @xmath256 are functions of @xmath261 , @xmath32 , @xmath114 and @xmath113 .
next , we rewrite equation as @xmath266 where the vector @xmath267 is given by and .
we now fix @xmath114 , @xmath113 , @xmath6 and @xmath268 satisfying the assumptions of lemma [ l:12 ] and we define the closed ball @xmath269 in the above expression , @xmath270 is a universal constant that will be determined in the following and @xmath271 and @xmath238 are defined by .
we also define the function @xmath272 by setting @xmath273 assume that @xmath274 is a strict contraction from @xmath275 to @xmath275 .
then the proof of lemma [ l:12 ] is complete : indeed , owing to the fixed point @xmath261 satisfies the rankine - hugoniot conditions .
also , owing to and to we infer that the inequalities are satisfied provided that the parameter @xmath6 is sufficiently small . in this paragraph
we prove that the map @xmath274 defined by is a strict contraction on the closed set @xmath276 defined by .
first , we make some remarks about notation . to simplify the exposition , in the following we denote by @xmath151 a universal constant :
its precise value can vary from occurrence to occurrence .
also , in the following we will determine the constant @xmath277 in and then choose the constant @xmath6 in such a way that @xmath278 .
this choice implies in particular that , when @xmath261 belongs to the set @xmath275 defined as in and the hypotheses of lemma [ l:12 ] are satisfied , then the map @xmath265 attains values on a bounded set and so @xmath265 and all its derivatives are bounded by some constant @xmath151 . finally , note that
, if @xmath279 and the hypotheses of lemma [ l:12 ] are satisfied , then @xmath280 .
we now proceed according to the following steps .
+ step 1 : we point out that to show that the map @xmath274 is a contraction it suffices to show that @xmath281 provided that @xmath279 and the hypotheses of lemma [ l:12 ] hold .
indeed , assume that holds , then @xmath282 and hence @xmath274 attains values in the set @xmath275 defined as in provided that @xmath277 is large enough .
also , @xmath283 provided that the constant @xmath6 is sufficiently small .
this implies that @xmath274 is a contraction and concludes the proof of lemma [ l:12 ] .
+ step 2 : we establish .
first , we point out that , if @xmath279 and the hypotheses of lemma [ l:12 ] are satisfied , then @xmath284 by using an analogous argument , we control the second component of @xmath265 and we arrive at @xmath285 next , we point out that when @xmath286 we have @xmath287 and @xmath288 and by using again the lipschitz continuity of the functions @xmath8 and @xmath9 we conclude that @xmath289 finally , we use the regularity of the function @xmath265 and , by arguing as in the proof of ( * ? ? ?
* lemma 2.5 , p. 28 ) , we combine and to obtain .
this concludes the proof of lemma [ l:12 ] .
the authors wish to thank the anonymous referees for their useful remarks that helped improve the exposition and allowed to shorten and improve the proof of lemma [ l:12 ] .
the second author wish to thank the organizers of the conference
contemporary topics in conservation laws " for the invitation to give a talk and present results related to the topic of the present paper .
both authors are members of the gruppo nazionale per lanalisi matematica , la probabilit e le loro applicazioni ( gnampa ) of the istituto nazionale di alta matematica ( indam ) and are supported by the prin national project `` nonlinear hyperbolic partial differential equations , dispersive and transport equations : theoretical and applicative aspects '' .
this work was also supported by the epsrc science and innovation award to the oxpde ( ep / e035027/1 ) .
l. ambrosio and c. de lellis , a note on admissible solutions of 1d scalar conservation laws and 2d hamilton - jacobi equations , _ j. hyperbolic differ .
_ , * 1 * ( 2004 ) , 813826 , url http://dx.doi.org/10.1142/s0219891604000263 .
p. baiti and h. k. jenssen , blowup in @xmath290 for a class of genuinely nonlinear hyperbolic systems of conservation laws , _ discrete contin .
dynam . systems _ , * 7 * ( 2001 ) , 837853 , url http://dx.doi.org/10.3934/dcds.2001.7.837 .
s. bianchini and l. caravenna , sbv regularity for genuinely nonlinear , strictly hyperbolic systems of conservation laws in one space dimension , _ comm .
_ , * 313 * ( 2012 ) , 133 , url http://dx.doi.org/10.1007/s00220-012-1480-5 .
a. bressan and r. m. colombo , decay of positive waves in nonlinear systems of conservation laws , _ ann .
scuola norm .
( 4 ) _ , * 26 * ( 1998 ) , 133160 , url http://www.numdam.org/item?id=asnsp_1998_4_26_1_133_0 . c. m. dafermos , _ hyperbolic conservation laws in continuum physics _ , vol
. 325 of grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , 3rd edition , springer - verlag , berlin , 2010 , url http://dx.doi.org/10.1007/978-3-642-04048-1 .
j. glimm and p. d. lax , _ decay of solutions of systems of nonlinear hyperbolic conservation laws _ , memoirs of the american mathematical society , no .
101 , american mathematical society , providence , r.i . ,
h. holden and n. h. risebro , _ front tracking for hyperbolic conservation laws _ , vol .
152 of applied mathematical sciences , springer - verlag , new york , 2002 , url http://dx.doi.org/10.1007/978-3-642-56139-9 . | we establish new interaction estimates for a system introduced by baiti and jenssen .
these estimates are pivotal to the analysis of the wave front - tracking approximation . in a companion paper
we use them to construct a counter - example which shows that schaeffer s regularity theorem for scalar conservation laws does not extend to systems .
the counter - example we construct shows , furthermore , that a wave - pattern containing infinitely many shocks can be robust with respect to perturbations of the initial data .
the proof of the interaction estimates is based on the explicit computation of the wave fan curves and on a perturbation argument .
laura caravenna laura v. spinolo |
in 2004 , etheridge and fleischmann @xcite introduced a stochastic spatial model of two interacting populations known as the symbiotic branching model , parametrized by a parameter @xmath1 $ ] governing the correlation between the two driving noises .
the model can be considered in three different spatial setups which we now explain .
first , the continuous - space symbiotic branching model is given by the system of stochastic partial differential equations @xmath2 where @xmath3 denotes the laplace operator and @xmath4 is a fixed constant known as the branching rate .
@xmath5 is a pair of correlated standard gaussian white noises on @xmath6 with correlation @xmath7 $ ] , that is , the unique gaussian process with covariance structure @xmath8 & = & ( t_1\wedge t_2 ) \ell(a_1\cap a_2),\\ \mathbb{e } [ w^2_{t_1}(a_1)w^2_{t_2}(a_2 ) ] & = & ( t_1\wedge t_2 ) \ell(a_1\cap a_2),\\ \mathbb{e } [ w^1_{t_1}(a_1)w^2_{t_2}(a_2 ) ] & = & \varrho ( t_1\wedge t_2 ) \ell(a_1\cap a_2),\end{aligned}\ ] ] where @xmath9 denotes lebesgue measure , @xmath10 and @xmath11 .
note that we work with a white noise @xmath12 in the sense of walsh @xcite .
solutions of this model have been considered rigorously in the framework of the corresponding martingale problem in theorem 4 of @xcite , which states that , under suitable conditions on the initial conditions @xmath13 , a solution exists for all @xmath14 $ ] .
the martingale problem is well posed for all @xmath15 , which implies the strong markov property except in the boundary case @xmath16 .
for a discrete spatial version we consider the system of interacting diffusions on @xmath17 , with values in @xmath18 , defined by the coupled stochastic differential equations @xmath19 where now @xmath20 is a family of standard brownian motions with covariances given by @xmath21_t = \cases { \varrho t , & \quad $ i = j$ and $ n \neq m$,\cr t , & \quad $ i = j$ and $ n = m$ , \cr 0 , & \quad otherwise . } \ ] ] in the discrete case , @xmath3 denotes the discrete laplacian @xmath22 note that in this paper we denote by @xmath23_t$ ] the cross - variation of two martingales @xmath24 .
this is to avoid confusion with @xmath25 which will be defined to be the sum ( resp . , integral ) of the product of @xmath26 and @xmath27 .
finally , the nonspatial symbiotic branching model is defined by the stochastic differential equations @xmath28 again , the noises are correlated with @xmath29_t=\varrho t$ ] . this simple toy - model
( see also @xcite and @xcite ) can be analyzed quite simply and will be used to prove properties of the spatial models . from time to time
we skip the dependence on @xmath30 and @xmath31 if there is no ambiguity .
solutions of @xmath32 and @xmath33 for @xmath34 are called symbiotic branching processes in the recurrent case whereas solutions of @xmath33 for @xmath35 are called symbiotic branching processes in the transient case .
interestingly , symbiotic branching models include well - known spatial models from different branches of probability theory . in the discrete spatial case ( and analogously in continuous - space ) interacting diffusions of the type @xmath36
have been studied extensively in the literature .
some important examples are the following : [ ex1 ] the stepping stone model from mathematical genetics : @xmath37 .
[ ex2 ] the parabolic anderson model ( with brownian potential ) from mathematical physics : @xmath38 .
[ ex3 ] the super random walk from probability theory : @xmath39 . for the super random walk , @xmath40 is the branching rate which in this case is time space independent . in @xcite , a two - type model based on two super random walks with time
space dependent branching was introduced .
the branching rate for one species is proportional to the value of the other species .
more precisely , the authors considered @xmath41 where now @xmath20 is a family of independent standard brownian motions .
solutions are called mutually catalytic branching processes . in the following years
, properties of this model were well studied ( see , e.g. , @xcite and @xcite ) . the corresponding continuous - space version was also treated in @xcite . for correlation @xmath42 ,
solutions of the symbiotic branching model are obviously solutions of the mutually catalytic branching model .
the case @xmath43 with the additional assumption @xmath44 corresponds to the stepping stone model . to see this ,
observe that in the perfectly negatively correlated case @xmath45 which implies that the sum @xmath46 solves a discrete heat equation and with the further assumption @xmath44 stays constant for all time .
hence , for all @xmath47 , @xmath48 , which shows that @xmath49 is a solution of the stepping stone model with initial condition @xmath50 and @xmath51 is a solution with initial condition @xmath31 .
finally , suppose @xmath52 is a solution of the parabolic anderson model , then , for @xmath16 , the pair @xmath53 is a solution of the symbiotic branching model with initial conditions .
the purpose of this and the accompanying paper @xcite is to understand the nature of the symbiotic branching model better .
how does the model depend on the correlation @xmath54 ?
are properties of the extremal cases @xmath55 inherited by some parts of the parameter space ?
since the longtime behavior of the super random walk , stepping stone model , mutually catalytic branching model and parabolic anderson model is very different , one might guess that the parameter space @xmath56 $ ] can be divided into disjoint subsets corresponding to different regimes .
the focus of @xcite is second moment properties . in the discrete
setting , but with a more general setup , growth of second moments is analyzed in detail .
a moment duality is used to reduce the problem to moment generating functions and laplace transforms of local times of discrete - space markov processes . a precise analysis of those is used to derive intermittency and aging results which show that different regimes occur for @xmath57 , @xmath42 and @xmath58 .
in contrast to @xcite , the present paper is not restricted to second moment properties .
the aim is to understand the pathwise behavior of symbiotic branching processes better .
[ rem : migration ] in this paper , we restrict ourselves to the simplest setups which already provide the full variety of results . for the discrete spatial model we thus restrict ourselves to the discrete laplacian instead of allowing more general transitions .
this is not necessary ; see @xcite or @xcite for a construction of solutions and main properties for more general underlying migration mechanisms in the case @xmath42 .
furthermore , we mainly restrict ourselves to homogeneous initial conditions and remark where results hold more generally . here
, for nonnegative real numbers we denote by @xmath59 the constant functions @xmath60 . the paper is organized as follows :
our main results are presented in section [ subsec : smr ] . before proving the results ,
we collect basic properties of the symbiotic branching models and discuss the dualities that we need .
this is carried out in section [ sec : bpd ] .
the final sections are devoted to the proofs . in section [ sec : comnvlaw ] , proofs of the longtime convergence in law are given , and in section [ sec : moments ] we discuss the longtime behavior of moments .
finally , in section [ sec : wavespeed ] we show how to use the results of section [ sec : moments ] to strengthen the main result of @xcite .
before stating the main results , we briefly recall from @xcite that the state space of @xmath61 is given by pairs of tempered functions , that is , pairs of functions contained in @xmath62 where @xmath63 , and we think of @xmath64 as being topologized by the metric given in @xcite , equation ( 13 ) , yielding a polish space .
the state space for @xmath33 is similar .
it was not discussed in @xcite and so we present details in section [ sec : bpd ] .
we begin with a result , generalizing theorem 1.5 of @xcite , on the longtime behavior of the laws of symbiotic branching processes in the recurrent case .
[ prop : convlaw ] suppose @xmath65 is a spatial symbiotic branching process in the recurrent case with @xmath66 , @xmath67 and initial conditions @xmath68 .
let @xmath69 and @xmath70 be two brownian motions with covariance @xmath71_t=\varrho t,\qquad t\ge0,\ ] ] and initial conditions @xmath72 .
further , let @xmath73 be the first exit time of the correlated brownian motions @xmath74 from the upper right quadrant . then , weakly in @xmath75 , @xmath76\rightarrow p^{u , v}[(\bar b^1_{\tau } , \bar b^2_{\tau})\in\cdot]\ ] ] as @xmath77 . here
, @xmath78 denotes the pair of constant functions on @xmath79 , respectively , @xmath17 ( @xmath80 ) taking the values of the stopped brownian motions @xmath81 .
in particular , the proposition shows ultimate extinction of one species in law .
[ all ] for simplicity , proposition [ prop : convlaw ] is formulated for constant initial conditions even though the result holds more generally .
theorem 1.5 of @xcite ( the case @xmath42 ) was extended in @xcite to nondeterministic initial conditions : for fixed @xmath82 let @xmath83 be the set of probability measures @xmath84 on @xmath75 such that @xmath85 and @xmath86\ , d\nu ( a , b ) = 0 \qquad\mbox{for all } x \in{\mathbb{r}}.\ ] ] here , @xmath87 denotes the transition semigroup of brownian motion ( the definition for the discrete case is similar ) .
the proof of @xcite can also be applied to @xmath88 and , thus , proposition [ prop : convlaw ] holds in the same way for initial distributions @xmath89 .
the restriction to @xmath66 arises from our method of proof which exploits a self - duality of the process which gives no information for @xmath90 .
let us briefly discuss the behavior of the limiting distributions in the boundary cases @xmath91 which are well known in the literature and fit neatly into our result .
first , suppose @xmath92 is a solution of the stepping stone model ( see example [ ex1 ] ) and @xmath93 $ ] .
it was proved in @xcite that @xmath94 where @xmath95 ( resp .
, @xmath96 ) denotes the dirac distribution concentrated on the constant function @xmath97 ( resp . , @xmath98 ) .
this can be reformulated in terms of perfectly anti - correlated brownian motions @xmath99 as before : for @xmath43 , the pair @xmath99 takes values only on the straight line connecting @xmath100 and @xmath101 , and stops at the boundaries .
hence , the law of @xmath102 is a mixture of @xmath103 and @xmath104 and the probability of hitting @xmath101 is equal to the probability of a one - dimensional brownian motion started in @xmath105 $ ] hitting @xmath106 before @xmath107 , which is @xmath52 , and hence matches ( [ eq : shigalimit ] ) .
second , let @xmath92 be a solution of the parabolic anderson model with brownian potential ( see example [ ex2 ] ) and constant initial condition @xmath108 . in @xcite it was shown that @xmath109 as discussed above , when viewed as a symbiotic branching process with @xmath16 , this implies @xmath110 from the viewpoint of two perfectly positive - correlated brownian motions , we obtain the same result since they simple move on the diagonal dissecting the upper right quadrant until they eventually get absorbed in the origin , that is , @xmath111 almost surely . to summarize
, we have seen that the weak longtime behavior ( in the recurrent case ) of the classical models connected to symbiotic branching is appropriately described by correlated brownian motions hitting the boundary of the upper right quadrant .
in contrast to extinction in law , the almost - sure behavior is very different . in the recurrent case for the mutually catalytic branching model , cox and klenke
@xcite showed that , almost surely , there is no longtime local extinction of any type , but in fact the locally predominant type changes infinitely often .
it is not hard to see that the same is true for symbiotic branching with @xmath66 .
we do not give a proof since it follows from proposition [ prop : convlaw ] along the same lines as in @xcite .
[ prop : pb ] let @xmath66 , @xmath67 and suppose @xmath65 is a spatial symbiotic branching process in the recurrent case with initial distribution @xmath68 . then
, for all @xmath112 and @xmath113 bounded , @xmath114 = 1,\ ] ] respectively , for @xmath115 bounded , @xmath116 = 1.\ ] ] again , as in remark [ all ] , the result holds for random initial conditions of the class @xmath117 .
note that proposition [ prop : pb ] depends strongly on the spatial structure since in the nonspatial model almost sure convergence holds ( see proposition [ prop : sconv ] ) .
in @xcite the second moments of symbiotic branching processes are analyzed .
this particular case admits a detailed study since a moment duality ( see lemma [ la : mdual ] ) has a particularly simple structure which allows one to reduce the study of the moments to that of moment generating functions and laplace transforms of local times .
here we are interested in the behavior of moments as @xmath118 tends to infinity .
the two available dualities ( self - duality and moment duality ) are combined in two steps .
first , a self - duality argument combined with an equivalence between bounded moments of the exit time distribution and of the exit point distribution for correlated brownian motions stopped on exiting the first quadrant is used to understand the effect of @xmath54 .
it turns out that for any @xmath119 there are critical values , independent of @xmath40 , dividing regimes in which the moments @xmath120 $ ] , @xmath121 $ ] and @xmath122 $ ] are bounded in @xmath118 or grow to infinity .
second , for @xmath123 , a perturbation argument combined with the first step and a moment duality is used to analyze the growth to infinity in more detail .
the following critical curve captures the effect of @xmath54 .
note that the definition is independent of @xmath40 which will become important in the second step .
[ def : cc ] we define the critical curve of symbiotic branching models to be the real - valued function @xmath124 , given by @xmath125 its inverse will be denoted by @xmath126 for @xmath119 .
the critical curve is plotted in figure [ fig : cl ] . here ,
@xmath127 and @xmath128 are marked .
thirty - fifth moments are the key for . ] the improved wavespeed result below and the special case @xmath129 is discussed in @xcite .
we will see in section [ sec : moments ] that this curve is closely connected with the exit distribution of @xmath81 from the upper right quadrant which appeared in proposition [ prop : convlaw ] above .
the first main theorem states that the critical curve separates two regimes ( independently of @xmath130 ) : that of bounded moments and that of unbounded moments .
[ thm : mc ] suppose @xmath65 is a symbiotic branching process @xmath131either nonspatial , continuous space or discrete space in arbitrary dimension@xmath132 with initial conditions @xmath133 .
if @xmath134 , then , for any @xmath67 , the following hold for @xmath119 : in the recurrent case , @xmath135 , { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}[u_t(k)^p ] \mbox { and } { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ u_t(x)^p ] \mbox { are bounded in } t.\ ] ] in the transient case , @xmath136 \mbox { is bounded in } t.\ ] ] due to symmetry the same holds for @xmath137 $ ] , @xmath138 $ ] and @xmath139 $ ] .
note that the theorem provides information about all positive real moments , not just integer moments . in the area below the critical curve in figure [ fig : cl ] ,
the moments remain bounded . on and above the critical curve , in the recurrent case
, the moments grow to infinity .
[ shift ] for @xmath43 the curve could be extended with @xmath140 . in terms of the previous theorem
this makes sense since for @xmath43 , symbiotic branching processes with initial conditions @xmath133 are bounded by @xmath141 .
this is justified by a simple observation : for initial conditions @xmath142 symbiotic branching processes with @xmath43 are solutions of the stepping stone model and , hence , bounded by @xmath106 .
uniqueness in law of solutions implies that solutions @xmath65 with initial conditions @xmath143 are equal in law to solutions @xmath144 times solutions with initial conditions @xmath145 . with this first understanding of the effect of @xmath54 on moments
, we may discuss integer moments for the discrete - space model in more detail .
let us first recall some known results for solutions @xmath92 of the parabolic anderson model ( see example [ ex2 ] ) where only the parameter @xmath40 appears .
using it s lemma , one sees that @xmath146 $ ] solves the ( discrete - space ) partial differential equation @xmath147 with homogeneous initial conditions . here , the potential @xmath148 is given by @xmath149 since @xmath150 is an @xmath151-particle schrdinger operator , many properties are known from the physics literature . in particular
, it is well known that in the recurrent case ( the potential is nonnegative ) exponential growth of solutions holds for any @xmath67 . by contrast , in the transient case the discrete laplacian requires a stronger perturbation before we see exponential growth .
intuitively from the particle picture this should be true since the potential @xmath148 only increases solutions if particles meet , which occurs less frequently in the transient case .
for the transient case ( see , e.g. , @xcite or @xcite for more precise results ) , there is a decreasing sequence @xmath152 such that @xmath153\mbox { is bounded in } t \quad\leftrightarrow\quad\kappa<\kappa(n)\ ] ] and for the lyapunov exponents @xmath154>0 \quad\leftrightarrow\quad\kappa>\kappa(n).\ ] ] these results can be proved with the @xmath151-particle path - integral representation in which solutions are expressed as @xmath155,\ ] ] where @xmath156 are independent simple random walks started in@xmath157 .
coming back to the symbiotic branching model , we ask whether or not the @xmath151th lyapunov exponents @xmath158\ ] ] exist and in which cases @xmath159 is strictly positive . as for the parabolic anderson model ,
there is a system of partial differential equations describing the moments ( see proposition 16 of @xcite for the continuous - space model ) and an @xmath151-particle path - integral representation of the moments .
in addition to the independent motion , the particles now carry a color which randomly changes if particles of the same color stay at the same site ( see lemma [ la : mdual ] ) . with @xmath160 denoting collision times of particles of same color and @xmath161 denoting collision times of particles of different colors , the path - integral representation of moments reads @xmath162={\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho l_t^{\neq } ) } \bigr].\ ] ] this representation is more involved than the path - integral representation for the parabolic anderson model since , in addition to the motion of particles , a second stochastic mechanism is included .
nonetheless , we use it to prove the following theorem which reveals that even in the recurrent case a nontrivial transition occurs .
[ thm : im ] for solutions of @xmath163 , in any dimension , the following hold for @xmath164 : @xmath165 exists for any @xmath7 $ ] , @xmath67 , @xmath166 for any @xmath67 , for any @xmath167 there is a critical @xmath152 such that @xmath168 if @xmath169 . combined with theorem [ thm : mc ] , parts ( ii ) and ( iii ) emphasize the `` criticality '' of the critical curve : for @xmath170 , moments stay bounded , for @xmath171 moments grow subexponentially fast to infinity , and for @xmath167 moments grow exponentially fast if @xmath40 is large enough . as discussed above , for the parabolic anderson model
it is natural that in the transient case perturbing the critical case does not immediately yield exponential growth , whereas perturbing the recurrent case does immediately lead to exponential growth .
it is clear that in the transient case the gap in ( iii ) of theorem [ thm :
im ] is really necessary : for small @xmath40 moments of the parabolic anderson model are bounded . since moments of symbiotic branching are dominated by moments of the parabolic anderson model ( see lemma [ la : mdual ] ) , for small @xmath40 moments are bounded for all @xmath54 .
in the case @xmath172 there seems to be no reason why exponential growth should fail .
unfortunately , in this case there is no moment duality and hence the most useful tool to analyze exponential growth is not available .
[ conj:1 ] in the recurrent case the moment diagram for symbiotic branching ( figure [ fig : cl ] ) describes the moments as follows : pairs @xmath173 below the critical curve correspond precisely to bounded moments , pairs at the critical curve correspond to moments which grow subexponentially fast to infinity and pairs above to the critical curve correspond to exponentially growing moments . a deeper understanding of the lyapunov exponents as functions of @xmath174 remains mainly open ( for an upper bound see proposition [ up ] ) . for second moments
[ @xmath129 ] this is carried out in @xcite .
it is shown that exponential growth holds for @xmath58 and arbitrary @xmath67 in the recurrent case , whereas only for @xmath176 in the transient case . here
@xmath177 denotes the green function of the simple random walk .
the exponential ( and subexponential ) growth rates were analyzed in detail by tauberian theorems .
a direct application of theorem [ thm : im ] is so - called intermittency of solutions .
one says a spatial system with lyapunov exponents @xmath178 is @xmath179-intermittent if @xmath180 intermittent systems concentrate on few peaks with extremely high intensity ( see @xcite ) .
the results above show that as @xmath54 tends to @xmath181 , solutions ( at least for large @xmath40 ) are @xmath179-intermittent for @xmath179 tending to infinity .
this holds since for fixed @xmath54 , the @xmath179th moments are bounded if @xmath182 lies below the critical curve .
increasing @xmath179 ( and @xmath40 if necessary ) there is a first @xmath179 such that the @xmath179th lyapunov exponent is positive .
intermittency for higher exponents suggests that the effect gets weaker .
this is to be expected since for @xmath43 solutions with homogeneous initial conditions are bounded and , hence , solutions do not produce high peaks at all . making this effect more precise ,
in particular combined with the effect of proposition [ prop : convlaw ] , is an interesting task for the future .
let us conclude with a direct application of the moment bounds . here
, we will be concerned with an improved upper bound on the speed of the propagation of the interface of continuous - space symbiotic branching processes which served to some extent as the motivation for this work . to explain this
, we need to introduce the notion of the interface of continuous - space symbiotic branching processes introduced in @xcite .
[ def : ifc ] the interface at time @xmath118 of a solution @xmath65 of the symbiotic branching model @xmath183 with @xmath1 $ ] is defined as @xmath184 where @xmath185 denotes the closure of the set @xmath186 in @xmath79 . in particular , we will be interested in complementary heaviside initial conditions @xmath187 the main question addressed in @xcite is whether for the above initial conditions the so - called compact interface property holds , that is , whether the interface is compact at each time almost surely .
this is answered affirmatively in theorem 6 in @xcite , together with the assertion that the interface propagates with at most linear speed , that is , for each @xmath1 $ ] there exists a constant @xmath188 and a finite random - time @xmath189 so that almost surely for all @xmath190 @xmath191.\ ] ] heuristically , due to the scaling property of the symbiotic branching model ( lemma 8 of @xcite ) one expects that the interface should move with a square - root speed . indeed , with the help of theorem [ thm : mc ]
one can strengthen their result , at least for sufficiently small @xmath54 , to obtain almost square - root speed .
[ cor : wavespeed ] suppose @xmath65 is a solution of @xmath192 with @xmath193 and @xmath67
. then there is a constant @xmath194 and a finite random - time @xmath189 such that almost surely @xmath195\ ] ] for all @xmath196 .
the restriction to @xmath197 is probably not necessary and only caused by the technique of the proof . though @xmath198 is rather close to @xmath181 , the result is interesting .
it shows that sub - linear speed of propagation is not restricted to situations in which solutions are uniformly bounded as they are for @xmath43 . the proof is based on the proof of @xcite for linear speed which carries over the proof of @xcite for the stepping stone model to nonbounded processes .
we are able to strengthen the result by using a better moment bound which is needed to circumvent the lack of uniform boundedness .
[ r ] we believe that , at least for @xmath199 , the speed of propagation should be at most @xmath200 , for some suitable constant @xmath201 , that is , for all @xmath202 greater than some @xmath203 , @xmath204.\ ] ] however , it seems unclear how to obtain such a refinement of theroem [ cor : wavespeed ] based on our moment results and the method of @xcite ( resp . , @xcite ) .
as subexponential bounds of higher moments can not be avoided ( see the proof of the fluctuation term estimate lemma [ la : ma ] ) , our results on the behavior of higher moments show that at present , in light of conjecture [ conj:1 ] , one can only hope for stronger results for very small @xmath54 . to overcome this limitation , new methods need to be employed .
the authors think that a possible approach could be based on the scaling property ( lemma 8 of @xcite ) and recent results by klenke and oeler @xcite .
recall that the scaling property states that if @xmath205 is a solution to @xmath206 , then @xmath207 is a solution to @xmath208 ( with suitably transformed initial states @xmath209 ) .
in other words , a diffusive time space rescaling leads to the original model with a suitably increased branching rate @xmath40 .
klenke and oeler @xcite show that , at least for the mutually catalytic model in discrete space , a nontrivial limiting process as @xmath210 exists .
this limit is called `` infinite rate mutually catalytic branching process '' ( see also @xcite for a further discusion ) . in particular , in corollary 1.2 of @xcite they claim that , under suitable assumptions , a nontrivial interface for the limiting process exists , which would in turn predict a square - root speed of propagation in our case .
however , to make this approach rigorous is beyond the scope of the present paper .
note that our results give only limited information about the shape of the interface .
for the case @xmath43 , that is , with locally constant total population size , it is shown in @xcite that there exists a unique stationary interface law , which may therefore be interpreted as a `` stationary wave '' whose position fluctuates at the boundaries , according to @xcite , like a brownian motion , hence explaining the square - root speed ( note that for both results , suitable bounds on fourth mixed moments are required ) . however , for @xmath211 , the population sizes of the interface are expected to fluctuate significantly and it seems unclear how this affects the shape and speed of the interface , in particular the formation of a `` stationary wave . ''
the significance of fourth mixed moments might even lead to a phase - transition in @xmath54 .
this gives rise to many interesting open questions .
in this section we review the setting and properties of the discrete - space model , whereas for continuous - space we refer to @xcite . note that instead of using the state space of tempered functions alternatively we may use a suitable liggett spitzer space .
as the results are only presented for the discrete laplacian this does not play a crucial role . for a discussion of the mutually catalytic branching model in the liggett
spitzer space see @xcite . for functions
@xmath212 we abbreviate @xmath213 . with @xmath214 the space of pairs of tempered sequences
is defined by @xmath215 the space of continuous paths is denoted by @xmath216 similarly , the space of pairs of rapidly decreasing sequences is defined by @xmath217 and the corresponding path space by @xmath218 weak solutions are defined as in @xcite for @xmath42 . in much the same way as for theorems 1.1 and 2.2 of @xcite , we obtain existence and the green - function representation . [
prop : bp1 ] suppose @xmath219 ( resp . , @xmath220 ) , @xmath221 $ ] and @xmath67 .
then there is a weak solution of @xmath222 such that @xmath223 ( resp . , @xmath224 ) and for all @xmath225 ( resp .
, @xmath226 ) @xmath227 where @xmath228 is the semigroup associated to the simple random walk . in particular
, we have @xmath229 the covariation structure of the brownian motions is given by ( [ eq : dcv ] ) .
in fact , ( [ 12 ] ) , ( [ 13 ] ) can be seen as the discrete - space versions of the martingale problem of definition 3 in @xcite .
further , ( [ 14 ] ) , ( [ 15 ] ) are the discrete - space versions of the convolution form given in corollary 20 of @xcite . for the proofs of the longtime behavior of laws and moments , the key step is to transfer to the total mass processes @xmath230 .
to this end , in a similar way to proposition [ prop : bp1 ] , we define @xmath231 and @xmath232 for summable initial conditions we obtain the following crucial martingale characterization . [ prop : tmmart ] if @xmath233 , then each solution of @xmath234 has the following properties : @xmath235 and @xmath236 are nonnegative , continuous , square - integrable martingales with square - functions @xmath237_t= [ \langle v_{\cdot},{\mathbf{1}}\rangle]_t=\kappa\int_0^t\langle u_s , v_s\rangle \,ds\ ] ] and @xmath238_t=\varrho\kappa\int_0^t\langle u_s , v_s\rangle \,ds.\ ] ] we omit the proofs since they are basically standard .
the only step where one needs to be careful is the existence proof .
as usual for such models one first restricts the space to bounded subsets ( boxes ) of @xmath17 , where standard markov process theory applies . enlarging the boxes one obtains a sequence of processes which are shown to converge to a limiting process solving @xmath239 . to prove tightness of the approximating sequence , the moments need to be bounded uniformly in the size of the boxes
more care than for @xmath42 in @xcite is needed .
the uniform moment bound can , for instance , be achieved using a colored particle moment duality for each box similar to the one of lemma [ la : mdual ] .
the symbiotic branching model exhibits an exceptionally rich duality structure , providing powerful tools for the analysis of the longtime properties .
we now recall the two - colors particle moment - duality introduced in section 3.1 of @xcite .
since the dual markov process is presented rigorously in @xcite we only sketch the pathwise behavior . to find a suitable description of the mixed moment @xmath240,\ ] ] @xmath241 particles are located in @xmath17 .
each particle moves as a continuous - time simple random walk independent of all other particles . at time
@xmath107 , @xmath151 particles of color @xmath106 are located at positions @xmath157 and @xmath242 particles of color @xmath141 are located at positions @xmath243 . for each pair of particles , one of the pair changes color when the time the two particles have spent in the same site , while both have same color , first exceeds an ( independent ) exponential time with parameter @xmath40 .
let @xmath244 note that since there are only @xmath241 particles , the infinite product is actually a finite product and hence well defined . the following lemma is taken from section 3 of @xcite .
[ la : mdual ] let @xmath65 be a solution of @xmath245 , @xmath67 and @xmath1 $ ] . then , for any @xmath246 , @xmath47 , @xmath247={\mathbb{e}}\bigl[(u_0,v_0)^{l_t}e^{\kappa(l_t^=+\varrho l_t^{\neq } ) } \bigr],\ ] ] where the dual process behaves as explained above .
note that for homogeneous initial conditions @xmath133 , the first factor in the expectation of the right - hand side equals @xmath106 . in the special case @xmath16 , lemma [ la : mdual ]
was already stated in @xcite , reproved in @xcite and used to analyze the lyapunov exponents of the parabolic anderson model . for @xmath248
, the difficulty of the dual process is based on the two stochastic effects : on the one hand , one has to deal with collision times of random walks which were analyzed in @xcite ; additionally , particles have colors either @xmath106 or @xmath141 which change dynamically .
[ rdual ] similar dualities hold for @xmath61 and @xmath249 . for continuous - space ,
the random walks are replaced by brownian motions and the collision times of the random walks by collision local times of the brownian motions ( see section 4.1 in @xcite ) .
the simplest case is the nonspatial symbiotic branching model where the particles stay at the same site and local times are replaced by real times ( see theorem 3.2 of @xcite or proposition a5 of @xcite ) .
mytnik @xcite introduced a self - duality for the continuous - space mutually catalytic branching model to obtain uniqueness of solutions of the corresponding martingale problem .
this can be extended to symbiotic branching models for @xmath66 as shown in proposition 5 of @xcite .
the discrete - space self - duality for @xmath42 was proved in theorem 2.4 of @xcite .
we first need more spaces of sequences : @xmath250 and @xmath251 in the sequel , the space @xmath252 and its subspaces will be used for @xmath253 .
the duality function for @xmath66 maps @xmath254 to @xmath255 via @xmath256 with this definition the generalized mytnik duality states : [ la : sduality ] for @xmath66 , @xmath67 , @xmath257 and @xmath258 let @xmath65 be a solution of @xmath259 and @xmath260 be a solution of @xmath261 .
then the following holds : @xmath262\\ & & \qquad={\mathbb{e}}^{\tilde{u}_0,\tilde{v}_0 } [ h(u_0+v_0,u_0-v_0,\tilde{u}_t+\tilde v_t,\tilde u_t-\tilde{v}_t ) ] .\end{aligned}\ ] ] analogously , the self - duality relation holds for the nonspatial model with duality function @xmath263 mapping @xmath264 to @xmath255 .
in this section we discuss weak longtime convergence of symbiotic branching models and prove proposition [ prop : convlaw ] .
we proceed in two steps : first , we prove convergence in law to some limit law following the proof of @xcite for @xmath42 .
second , to characterize the limit law for the spatial models , we reduce the problem to the nonspatial model . [ prop : wconv ] let @xmath265 and @xmath65 a solution of either @xmath183 or @xmath266 with initial conditions @xmath68 .
then , as @xmath267 , the law of @xmath65 converges weakly on @xmath75 to some limit @xmath268 .
the proof is only given for the discrete spatial case and the continuous case is completely analogous .
let us first recall the strategy of @xcite for @xmath42 which can also be applied with the generalized self - duality required here .
convergence of @xmath65 in @xmath75 follows from convergence of @xmath269 in @xmath252 . using lemma 2.3(c ) of @xcite
, it suffices to show convergence of @xmath270 $ ] for all @xmath271 .
furthermore , the limit @xmath268 is uniquely determined by @xmath272 $ ] ( see lemma 2.3(b ) of @xcite ) .
hence , it suffices to show convergence of @xmath273={\mathbb{e}}^{{\mathbf{u}},{\mathbf{v } } } \bigl[e^{-\sqrt { 1-\varrho}\langle u_t+v_t,\phi\rangle+i\sqrt{1+\varrho}\langle u_t - v_t,\psi\rangle } \bigr],\hspace*{-32pt}\ ] ] for all @xmath274 .
note that the technical condition of lemma 2.3(c ) of @xcite is fullfilled since due to proposition [ prop : bp1 ] @xmath275=(u+v)\langle { \mathbf{1}},p_t \phi_{-\lambda}\rangle < c<\infty.\ ] ] to ensure convergence of ( [ 456 ] ) we employ the generalized mytnik self - duality of lemma [ la : sduality ] with @xmath276 : @xmath277 \nonumber\\ & & \qquad={\mathbb{e}}^{u_0,v_0 } \bigl[e^{-\sqrt{1-\varrho}\langle u_t+v_t,\tilde { u}_0+\tilde{v}_0\rangle + i\sqrt{1+\varrho}\langle u_t - v_t,\tilde{u}_0-\tilde{v}_0\rangle } \bigr ] \nonumber\\[-8pt]\\[-8pt ] & & \qquad={\mathbb{e}}^{\tilde{u}_0,\tilde{v}_0 } \bigl[e^{-\sqrt{1-\varrho}\langle u_0+v_0 , \tilde{u}_t+\tilde{v}_t\rangle+i\sqrt{1+\varrho}\langle u_0-v_0 , \tilde{u}_t-\tilde{v}_t\rangle } \bigr]\nonumber\\ & & \qquad={\mathbb{e}}^{\tilde{u}_0,\tilde{v}_0 } \bigl[e^{-\sqrt{1-\varrho}(u+v)\langle \mathbf{1 } , \tilde{u}_t+\tilde{v}_t\rangle+i\sqrt{1+\varrho}(u - v)\langle \mathbf{1 } , \tilde{u}_t-\tilde{v}_t\rangle } \bigr].\nonumber\end{aligned}\ ] ] by assumption , @xmath278 have compact support and hence by proposition [ prop : tmmart ] the total - mass processes @xmath279 and @xmath280 are nonnegative martingales . by the martingale convergence theorem @xmath279 and
@xmath280 converge almost surely to finite limits denoted by @xmath281 , @xmath282 .
finally , the dominated convergence theorem implies convergence of the right - hand side of ( [ eq : dualconv ] ) to @xmath283.\ ] ] combining the above , we have proved convergence of @xmath284,\ ] ] which ensures weak convergence of @xmath65 in @xmath75 to some limit which is uniquely determined by ( [ eq : elimit ] ) .
again , as in remark [ all ] , the previous proposition can be proved for nondeterministic initial conditions as in @xcite .
the rest of this section is devoted to identifying the limit @xmath285 in the recurrent case . before completing the proof of theorem [ prop : convlaw ]
we discuss a version of knight s extension of the dubins
schwarz theorem ( see @xcite , 3.4.16 ) for nonorthogonal continuous local martingales .
[ la : eds ] let @xmath286 and @xmath287 be continuous local martingales with @xmath288 almost surely .
assume further that , for @xmath289 , @xmath290_t= [ n_{\cdot},n_{\cdot}]_t \quad\mbox{and}\quad [ m_{\cdot},n_{\cdot}]_t=\varrho [ m_{\cdot},m_{\cdot}]_t\qquad \mbox{a.s.},\ ] ] where @xmath1 $ ] .
if @xmath291_{\infty } = \infty$ ] a.s . , then @xmath292 is a pair of brownian motions with covariances @xmath293_t=\varrho t$ ] , where @xmath294_s > t \}.\ ] ] it follows from the dubins
schwarz theorem that @xmath295 are each brownian motions . further , by the definition of @xmath296 we obtain the claim @xmath297_t= [ m_{\cdot } , n_{\cdot}]_{t(t)}=\varrho [ m_{\cdot},m_{\cdot}]_{t(t)}=\varrho t.\ ] ] if @xmath298_{\infty}<\infty$ ] the situation becomes slightly more delicate but one can use a local version of lemma [ la : eds ] . indeed , define , for @xmath289 , @xmath299 where the time - change @xmath202 is given in ( [ eq : timeshift ] ) and define @xmath70 analogously for @xmath300 ( recall that @xmath301_t= [ n_{\cdot},n_{\cdot}]_t$ ] ) . then the processes @xmath74 are brownian motions stopped at time @xmath302 .
the covariance is again given by @xmath297_{t\wedge t^ * } = \varrho(t\wedge t^*),\qquad t \ge0.\ ] ] for the rest of this section let @xmath295 be standard brownian motions with covariance @xmath303_t=\varrho t\ ] ] started in @xmath304 , denote their expectations by @xmath305 , and let @xmath306 the above discussion can now be used to understand the longtime behavior of symbiotic branching processes .
we start by giving a proof for the nonspatial symbiotic branching model and then modify the proof to capture the corresponding result for the spatial models .
[ prop : sconv ] let @xmath65 be a solution of @xmath307 . then
, as @xmath77 , @xmath65 converges almost surely to some @xmath268 .
furthermore , @xmath308 with @xmath309 from proposition [ prop : convlaw ] .
solutions of the nonspatial symbiotic branching model are nonnegative martingales and hence converge almost surely .
this implies the first part of the claim and it only remains to characterize the limit . obviously , the @xmath310-martingales @xmath311 satisfy the cross - variation structure assumptions of lemma [ la : eds ] and , thus , @xmath312 .
to obtain the result , we need to check that @xmath313 . by definition of @xmath249 ,
the time - change is given by @xmath314_t = \biggl [ \int_0^\cdot\sqrt{\kappa u_s u_s } \,db^1_s , \int_0^\cdot\sqrt{\kappa u_s v_s } \,db^1_s \biggr]_t = \kappa\int_0^t u_s v_s \,ds.\hspace*{-32pt}\ ] ] to see that @xmath315 , first note that @xmath316 for all @xmath47 .
this is true since @xmath317 and solutions of @xmath249 are nonnegative . to argue that @xmath318 increases to @xmath319 ,
more care is needed .
since the martingales converge almost surely , @xmath318 converges to some value @xmath320 .
suppose @xmath321 , then @xmath65 converges to some @xmath322 with @xmath323 .
this yields a contradiction since @xmath324 would increase to infinity .
hence , almost surely , @xmath325 in particular , the proof of proposition [ prop : sconv ] provides an important relation for @xmath326 . as remarked below lemma
[ la : sduality ] , the self - duality also works in the nonspatial model : @xmath327\\ & & \qquad={\mathbb{e}}^{\tilde{u}_0,\tilde{v}_0}\bigl[e^{-\sqrt{1-\varrho } ( u_0+v_0)(\tilde { u}_t+\tilde{v}_t)+i\sqrt{1+\varrho}(u_0-v_0)(\tilde{u}_t-\tilde { v}_t ) } \bigr],\end{aligned}\ ] ] where both @xmath65 and @xmath328 are solutions of @xmath329 with different initial conditions .
as shown in the proof of proposition [ prop : sconv ] , @xmath65 [ resp .
, @xmath328 ] converges almost surely to @xmath331 with initial condition @xmath145 [ resp . , @xmath332 ) ] . using dominated convergence ,
this shows the following duality relation for @xmath326 when started in initial conditions @xmath333 , @xmath334 : @xmath335\nonumber\\[-8pt]\\[-8pt ] & & \qquad = e^{\tilde{u},\tilde{v } } [ h^0(b^1_{\tau}+b^2_{\tau},b^1_{\tau } -b^2_{\tau},u+v , u - v ) ] .\nonumber\end{aligned}\ ] ] proof of proposition [ prop : convlaw ] again , the proof is only presented in the discrete spatial setting since the continuous case is analogous .
we retain the notation of the proof of proposition [ prop : wconv ] where we showed that , as @xmath118 tends to infinity , @xmath336\\ & & \qquad\rightarrow{\mathbb{e}}^{({\phi+\psi})/{2},({\phi-\psi})/{2 } } \bigl[e^{-\sqrt { 1-\varrho}(u+v)\langle\mathbf{1},\tilde{u}_{\infty}+\tilde { v}_{\infty } \rangle+i\sqrt{1+\varrho}(u - v ) \langle\mathbf{1 } , \tilde{u}_{\infty}-\tilde{v}_{\infty}\rangle } \bigr].\end{aligned}\ ] ] let us specify the limit law as for the nonspatial symbiotic branching process .
as seen in proposition [ prop : tmmart ] the total - mass processes @xmath337 and @xmath338 are nonnegative continuous @xmath310-martingales with cross - variations @xmath339_t=\varrho[\bar u_{\cdot } , \bar u_{\cdot}]_t=\varrho[\bar v_{\cdot } , \bar v_{\cdot}]_t , t\ge0 $ ] .
thus , by lemma [ la : eds ] , reasoning as in ( [ eq : invtime ] ) , @xmath340 , where @xmath295 are brownian motions started in @xmath341 , @xmath342 with covariance @xmath343_t=\varrho t$ ] and @xmath344 .
again , we need to show that @xmath345 .
this is much more subtle than in the nonspatial case since the quadratic variation might level off even if both total - mass processes @xmath346 , @xmath347 are strictly positive . in @xcite
it was shown that for @xmath42 , almost surely , this does not happen in the recurrent case [ cf .
the proof of their theorem 1.2(b ) ] .
their proof can be used directly for @xmath66 . hence , almost surely , @xmath348 combining the above discussion with ( [ eq : elimit ] ) , we are able to determine the limit .
first , we derived @xmath349\\ & & \qquad\stackrel{t\rightarrow\infty}{\rightarrow}e^{\langle({\phi + \psi } ) /{2},{\mathbf{1}}\rangle,\langle({\phi-\psi})/{2},{\mathbf{1}}\rangle }
\bigl[e^{-\sqrt{1-\varrho}(u+v ) ( b^1_{\tau}+b^2_{\tau } ) + i\sqrt { 1+\varrho}(u - v ) ( b^1_{\tau}-b^2_{\tau } ) } \bigr].\end{aligned}\ ] ] to use lemma 2.3(c ) of @xcite we manipulate the right - hand side using ( [ dual0 ] ) : @xmath350\\ & & \qquad = e^{\langle({\phi+\psi})/{2},{\mathbf{1}}\rangle,\langle({\phi -\psi } ) /{2},{\mathbf{1}}\rangle } [ h^0 ( { b^1_{\tau}}+{b^2_{\tau}},{b^1_{\tau}}-{b^2_{\tau } } , u+v , u - v ) ] \\ & & \qquad = e^{u , v } [ h^0 ( { b^1_{\tau}}+{b^2_{\tau}},{b^1_{\tau}}-{b^2_{\tau } } , \langle\phi,{\mathbf{1}}\rangle,\langle\psi,{\mathbf{1}}\rangle ) ] \\ & & \qquad = e^{u , v } [ h ( \bar{b}^1_{\tau}+\bar{b}^2_{\tau},\bar{b}^1_{\tau } -\bar{b}^2_{\tau},\phi,\psi ) ] , \end{aligned}\ ] ] where , as in proposition [ prop : convlaw ] , @xmath351 ( resp . ,
@xmath352 ) denotes the constant function taking only the ( random ) value @xmath353 ( resp . ,
@xmath354 ) .
in total we have @xmath355 \stackrel{t\rightarrow\infty}{\rightarrow}e^{u , v } [ h(\bar { b}^1_{\tau } + \bar{b}^2_{\tau},\bar{b}^1_{\tau}-\bar{b}^2_{\tau},\phi,\psi ) ] , \ ] ] which implies weak convergence in @xmath75 of @xmath65 to @xmath356 by lemma 2.3(c ) of @xcite .
in this section we prove theorems [ thm : mc ] and [ thm : im ] . before giving the proofs
we prove an equivalence for moments of correlated brownian motions .
let @xmath66 , @xmath357 and @xmath74 be brownian motions started in @xmath304 with @xmath358 the starting points of brownian motions will be indicated by superscripts in probabilities and expectations .
further , let @xmath359 [ thm : theo2 ] let @xmath360 and @xmath361 . under the above assumptions , the following conditions are equivalent : @xmath362 @xmath363<\infty,\ ] ] @xmath364<\infty.\ ] ] we start with the proof of the equivalence of ( i ) and ( ii ) .
define a cone in the plane with angle @xmath365 by @xmath366 and denote its boundary by @xmath367 .
note that with this definition , the positive real line is always contained in @xmath368 .
further , we define , for @xmath66 , a sector in @xmath369 by @xmath370 and denote by @xmath371 its boundary .
note that this time , the positive imaginary axis is always in @xmath372 and that the angle of the sector at the origin is given by @xmath373 to transform the correlated brownian motions @xmath295 to planar brownian motion we use the simple fact that @xmath374 defines a pair of independent brownian motions started in @xmath375 satisfying @xmath376 . by the definition of @xmath372 , the planar brownian motion @xmath377 started in @xmath378 hits @xmath371 if and only if the correlated brownian motions @xmath295 started in @xmath304 hit @xmath379 .
hence , for @xmath380 as in ( [ eq : tauagain ] ) , we have @xmath381 since planar brownian motion is rotation invariant , @xmath372 may be rotated to agree with the cone @xmath382 , without changing the exit time .
obviously , with the corresponding rotated initial conditions , the law of the first exit time @xmath383 from the cone @xmath382 agrees with the law of @xmath384 .
for planar brownian motion in a cone @xmath382 it is well known ( see @xcite , theorem 2 ) that @xmath385 <
\infty \quad\leftrightarrow\quad p<\frac{\pi}{\theta},\ ] ] independently of @xmath386 .
( [ tt ] ) and ( [ 124 ] ) now imply the equivalence of ( i ) and ( ii ) and independence of @xmath304 .
the proof of the equivalence of ( i ) and ( iii ) is via conformal transformation of the cone @xmath382 to the upper half - plane .
indeed , we are going to calculate the densities of the exit - point distributions @xmath387 we proceed in three steps : after reducing to independent brownian motions in @xmath372 as for the exit time , we rotate @xmath372 to @xmath382 and , finally , stretch the cone to end up with the upper half - plane . recall that the first exit of @xmath388 happens at position @xmath389 if and only if the first exit of @xmath377 takes place at @xmath390 .
hence , ( [ h ] ) transforms to @xmath391\\[-8pt ] & & \qquad= p^{u,({v-\varrho u})/{\sqrt{1-\varrho^2 } } } \biggl(w^1_{\tau ^w}=0,w^2_{\tau^w}\geq\frac{y}{\sqrt{1-\varrho^2 } } \biggr).\nonumber\end{aligned}\ ] ] in a similar fashion one obtains @xmath392\\[-8pt ] & & \qquad= p^{u , ( { v-\varrho u})/{\sqrt{1-\varrho^2 } } } \biggl(w^1_{\tau^w}\leq x , w^2_{\tau^w}=-\frac{\varrho}{\sqrt{1-\varrho ^2}}w^1_{\tau^w } \biggr).\nonumber\end{aligned}\ ] ] we represent the transformed initial conditions @xmath393 in polar coordinates , that is , @xmath394 for the rotation we add the angle @xmath395 to get the new initial condition .
finally , to map the cone @xmath382 conformally to the upper half - plane @xmath396 , we apply the map @xmath397 which maps @xmath382 onto @xmath396 . using conformal invariance of brownian motion ( e.g. , lemma 7.19 of @xcite )
, the problem is reduced to the computation of the exit distribution of planar ( time - changed ) brownian motion from the upper half - plane .
indeed , due to the random time change the ( almost surely finite ) exit time changes but not the distribution of the exit points , which is cauchy ( see theorem 2.37 of @xcite ) .
thus , to obtain the distribution of the exit points explicitly it , only remains to specify the transformed initial condition @xmath398 , which is given by @xmath399 now , let @xmath400 be two independent brownian motions with @xmath401 and @xmath402 then , by ( [ e1 ] ) , ( [ e2 ] ) , @xmath403 explicit manipulations of the cauchy distribution yield @xmath404\\[-8pt ] & & \qquad=\int_{y}^{\infty } \frac{1}{\pi\tilde{z}_2\sqrt{1-\varrho ^2}^{\pi /\theta}}\frac{{\pi}/{\theta}r^{{\pi}/{\theta}-1 } } { 1 + ( ( { ( { r}/{\sqrt{1-\varrho^2 } } ) ^{{\pi } /{\theta } } + \tilde{z}_1})/{\tilde{z}_2 } ) ^2 } \,dr,\nonumber\\ \label{h2 } & & p^{u , v } ( b^1_{\tau^b}\leq x , b^2_{\tau^b}=0 ) \nonumber\\[-8pt]\\[-8pt ] & & \qquad=\int_0^x\frac{1}{\pi\tilde{z}_2\sqrt{1-\varrho^2}^{\pi/\theta } } \frac { { \pi/\theta } r^{{\pi/\theta}-1 } } { 1 + ( ( { ( { r}/{\sqrt{1-\varrho^2 } } ) ^{{\pi /\theta}}- \tilde{z}_1})/{\tilde{z}_2 } ) ^2 } \,dr.\nonumber\end{aligned}\ ] ] finally , noting that @xmath405 if and only if @xmath406 , we deduce from ( [ h1 ] ) and ( [ h2 ] ) that @xmath407<\infty\quad\mbox{if and only if } \quad p<\frac{\pi}{\theta}.\ ] ] the proof relies on a combination of the self - duality based technique of the proof of proposition [ prop : pb ] and the close relation between the moments of the exit - time and exit - point distribution of correlated brownian motions obtained in theorem [ thm : theo2 ] .
proof of theorem [ thm : mc ] we proceed in several steps .
first , the result for the nonspatial model is proved and thereafter the results for the discrete - space and the continuous - space models . finally , we present the argument in the transient case . in the following we use the definition of @xmath295 and @xmath319 from proposition [
prop : convlaw ] . _
suppose @xmath65 is a solution of @xmath408 .
`` @xmath409 '' : we first assume @xmath410 , in which case theorem [ thm : theo2 ] implies@xmath411<\infty$ ] . as argued in the proof of proposition [ prop : sconv ] , @xmath412 is a nonnegative martingale and due to the same arguments satisfies @xmath413_{t}^{p/2 } ] \leq e^{1,1}[\tau^{p/2}]<\infty$ ] for all @xmath47 and @xmath67 .
considering @xmath414 , we apply the burkholder davis gundy inequality to get @xmath415&=&{\mathbb{e}}^{1,1}[(\bar u_t+1)^p]\\ & = & { \mathbb{e}}^{1,1}\bigl[{\mathbf{1}}_{\{\bar u_t\leq1\}}(\bar u_t+1)^p\bigr]+{\mathbb{e}}^{1,1}\bigl[{\mathbf{1}}_{\ { \bar u_t>1\}}(\bar u_t+1)^p\bigr]\\ & \leq & c_p+c_p{\mathbb{e}}^{1,1}[{\bar u_t}^p]\\ & \leq & c_p+c_p{\mathbb{e}}^{1,1}\bigl[\sup_{0\leq s\leq t } \bar u_s^p\bigr]\\ & \leq & c_p+c'_p{\mathbb{e}}^{1,1}[[\bar u_{\cdot}]_t^{p/2}]<\infty\end{aligned}\ ] ] independently of @xmath118 and @xmath40 .
`` @xmath416 . ''
conversely , for @xmath417 , theorem [ thm : theo2 ] implies that @xmath418=\infty$ ] .
using fatou s lemma and almost sure convergence of @xmath412 to @xmath419 , the proof for the nonspatial case is finished with @xmath420\geq{\mathbb{e}}^{1,1}[u_{\infty } ^p]=e^{1,1}[(b_{\tau})^p]=\infty.\ ] ] again , this lower bound is independent of @xmath40 .
_ step _ 2 .
the proof for @xmath421 is started by reducing the moments for homogeneous initial conditions to finite initial conditions . indeed , employing lemma [ la : sduality ] with @xmath422 , where @xmath423 denotes the indicator function of site @xmath424 , gives @xmath425 & = & { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } \bigl[e^{-\sqrt{1-\varrho}\langle u_{t}+v_{t},\phi + \psi \rangle } \bigr]\\ & = & { \mathbb{e}}^{\phi,\psi } \bigl[e^{-\sqrt{1-\varrho}\langle{\mathbf{1}}+{\mathbf{1}},\tilde { u}_{t}+\tilde{v}_{t}\rangle } \bigr]\\ & = & { \mathbb{e}}^{{\mathbf{1}}_{k},{\mathbf{1}}_{k } } \bigl[e^{-\sqrt{1-\varrho}\theta\langle{\mathbf{1 } } , \tilde { u}_{t}+\tilde{v}_{t}\rangle } \bigr],\end{aligned}\ ] ] where we used the argument of remark [ shift ] .
note that , due to our choice of initial conditions , the complex part of the self - duality vanishes . since the above is a laplace transform identity , we have @xmath426 and hence @xmath427 = { \mathbb{e}}^{\mathbf{1}_k,\mathbf{1}_k } [ ( \langle{\mathbf{1 } } , \tilde{u}_t \rangle+ \langle{\mathbf{1 } } , \tilde{v}_t \rangle ) ^p ] .\ ] ] we are now prepared to finish the proof of the theorem for the discrete case . `` @xmath409 . ''
suppose @xmath410 .
let @xmath428 , which due to lemma [ prop : tmmart ] is a square - integrable martingale with quadratic variation @xmath429_t = [ \langle{\mathbf{1 } } , \tilde{u}_{\cdot}\rangle]_t+ [ \langle{\mathbf{1 } } , \tilde{v}_{\cdot}\rangle]_t+2 [ \langle{\mathbf{1 } } , \tilde { u}_{\cdot } \rangle,\langle{\mathbf{1 } } , \tilde{v}_{\cdot } \rangle]_t=(2 + 2\varrho ) [ \langle{\mathbf{1 } } , \tilde{u}_{\cdot } \rangle]_t.\ ] ] to apply the burkholder
gundy inequality , we switch again from @xmath430 to @xmath431 , which is a martingale null at zero
. hence , @xmath432={\mathbb{e}}^{\mathbf{1}_k , \mathbf{1}_k } [ ( \bar{m}_t+m_0)^p ] \leq c_p+c_p{\mathbb{e}}^{\mathbf{1}_k , \mathbf{1}_k } [ \bar{m}_t^p ] .\ ] ] then we get from ( [ 23 ] ) and the burkholder davis gundy inequality @xmath433 & \leq & c_p+c_p{\mathbb{e}}^{\mathbf{1}_k , \mathbf{1}_k } [ \bar{m}_t^p ] \\ &
\leq & c_p+c_p{\mathbb{e}}^{\mathbf{1}_k,\mathbf{1}_k } \bigl [ \sup_{0\leq s\leq t } \bar m_s^p \bigr]\\ & \leq & c_p+c'_p { \mathbb{e}}^{\mathbf{1}_k , \mathbf{1}_k } [ [ \bar m_{\cdot } ] _ t^{p/2 } ] \\ & = & c_p+ c'_p(2 + 2\varrho)^{p/2 } { \mathbb{e}}^{\mathbf{1}_k,\mathbf{1}_k } [ [ \langle { \mathbf{1 } } , \tilde{u}_{\cdot } \rangle]_t^{p/2 } ] \end{aligned}\ ] ] for some constants @xmath434 independent of @xmath118 and @xmath40 . as in the proof of theorem [ prop : convlaw ] , the random time change which makes the pair of total masses a pair of correlated brownian motions is bounded by @xmath319 , that is , @xmath435_t\leq\tau$ ] for all @xmath47 .
this yields by theorem [ thm : theo2 ] @xmath436\leq{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } \bigl[\bigl(u_t(k)+v_t(k)\bigr)^p \bigr]\leq c_p+c'_p(2 + 2\varrho)^{p/2 } e^{1 , 1 } [ \tau^{p/2 } ] < \infty.\ ] ] `` @xmath416 . ''
suppose @xmath417 . as in the proof of theorem [ prop : convlaw ] we use the almost sure convergence of @xmath437 to @xmath438
combining this with fatou s lemma gives @xmath439 & \geq&\liminf_{t\rightarrow\infty}{\mathbb{e}}^{\mathbf{1}_k,\mathbf{1}_k } [ \langle{\mathbf{1 } } , \tilde{u}_t \rangle^p ] \\ & \geq&{\mathbb{e}}^{{\mathbf{1}}_k,{\mathbf{1}}_k } \bigl[\liminf_{t\rightarrow\infty}\langle { \mathbf{1 } } , \tilde{u}_t \rangle^p \bigr ] \\ & = & e^{1,1 } [ ( b^1_{\tau})^p ] .\end{aligned}\ ] ] the right - hand side is infinite due to theorem [ thm : theo2 ] and hence @xmath440 $ ] diverges .
equation ( [ 23 ] ) now shows that @xmath441 $ ] also grows without bound . since symbiotic branching processes
are nonnegative , this is also true for @xmath121 $ ] as can be seen as follows : @xmath442 & \leq&{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}\bigl[(2u_t(k))^p { \mathbf{1}}_{\{u_t(k)\geq v_t(k)\}}\bigr]\\ & & { } + { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}\bigl[(2v_t(k))^p { \mathbf{1}}_{\{u_t(k)<v_t(k)\}}\bigr]\\ & \leq&2^p{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}[u_t(k)^p]+ 2^p{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}[v_t(k)^p]\\ & = & 2^{p+1}{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}[u_t(k)^p],\end{aligned}\ ] ] where we used lemma [ la : mdual ] to see that @xmath121={\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}[v_t(k)^p]$ ] .
_ step _ 3 .
the proof for @xmath443 is slightly more involved since we can not use the indicator @xmath444 to get @xmath445 , where now @xmath446 .
instead we use a standard smoothing procedure . for
fixed @xmath447 let @xmath448 where we skip the dependence on @xmath449 .
the main part is to show that @xmath450\\[-8pt ] & & \qquad\leq\vert u_t(x)-\langle u_t , p_{\varepsilon}\rangle \vert_{l^p}+\vert v_t(x)-\langle v_t , p_{\varepsilon}\rangle\vert_{l^p } \stackrel { \varepsilon\rightarrow0}{\rightarrow } 0,\nonumber\end{aligned}\ ] ] which implies @xmath451 due to symmetry we only consider @xmath452 . to prove ( [ z ] ) we first observe that , due to the green function representation provided in corollary 19 of @xcite , @xmath453 where @xmath454 is a zero - mean martingale measure with quadratic variation @xmath455_t=\kappa\int_0^t\int _ { { \mathbb{r}}}f^2(s , b)u_s(b)v_s(b ) \,ds\,db\ ] ] for test functions @xmath26 such that the integral is well defined ( see lemma 18 of @xcite for details ) . for homogeneous initial conditions
, the first difference vanishes and it suffices to concentrate on the difference of the stochastic integrals . by the burkholder davis
gundy inequality the difference of the integrals can be estimated as @xmath456\\ & & \qquad\leq c\kappa^{p/2}{\mathbb{e}}\biggl [ \biggl ( \int_0^t\int_{{\mathbb{r}}}\bigl(p_{t - s}(x - b)-p_{\varepsilon+t - s}(x - b)\bigr)^2u_s(b)v_s(b ) \,ds\,db \biggr)^{p/2 } \biggr].\end{aligned}\ ] ] now expanding @xmath457 as @xmath458 we get the upper bound ( taking the expectation under the integral is valid since the integrands are nonnegative ) @xmath459 \,ds\,db \biggr],\end{aligned}\ ] ] where we have used that , for @xmath460 , @xmath461 by hlder s inequality . as in @xcite , page 153 , the second term can now be bounded from above by a constant depending only on @xmath179 and @xmath118 .
the first factor can be estimated by @xmath462 due to @xcite , lemma 6.2 .
hence , for fixed @xmath463 and @xmath47 , ( [ z ] ) holds and thus we obtain ( [ z2 ] ) .
the rest of the proof is similar to the discrete case but slightly more technical .
since @xmath464 is rapidly decreasing , we have @xmath465 & = & { \mathbb{e}}^{\theta p_{\varepsilon},\theta p_{\varepsilon } } \bigl [ e^{-2\sqrt { 1-\varrho}\langle{\mathbf{1 } } , \tilde{u}_t+\tilde{v}_t \rangle } \bigr]\\ & = & { \mathbb{e}}^ { p_{\varepsilon } , p_{\varepsilon } } \bigl[e^{-\sqrt{1-\varrho } 2\theta \langle{\mathbf{1 } } , \tilde{u}_t+\tilde{v}_t \rangle } \bigr].\end{aligned}\ ] ] thus , we get @xmath466 and in particular @xmath467 = { \mathbb{e}}^{p_{\varepsilon},p_{\varepsilon } } [ ( \langle{\mathbf{1 } } , \tilde{u}_t \rangle + \langle{\mathbf{1 } } , \tilde{v}_t \rangle)^p ] .\ ] ] we may now finish the proof in a similar way to the discrete case .
`` @xmath409 . ''
due to ( [ z2 ] ) we are done if we can bound @xmath468 $ ] independently of @xmath469 and @xmath47 .
this can be done as before : @xmath470 and @xmath471 are random time - changed correlated brownian motions with initial conditions @xmath472 for all @xmath469 .
using , as before , the auxiliary martingale @xmath473 we obtain ( as in the discrete case ) with the help of the burkholder davis gundy inequality @xmath474 & = & \lim_{\varepsilon\rightarrow0}{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ \langle u_t+v_t , p_{\varepsilon}\rangle^{p } ] \\ & = & \lim_{\varepsilon\rightarrow0 } { \mathbb{e}}^{p_{\varepsilon},p_{\varepsilon } } [ \langle{\mathbf{1}},\tilde { u}_t+\tilde{v}_t \rangle^p ] \\ & \leq & c_p+c_p\lim_{\varepsilon\rightarrow0 } { \mathbb{e}}^ { p_{\varepsilon},p_{\varepsilon } } [ \bar{m}_t^p ]
\\ & \leq & c_p+c'_p\lim_{\varepsilon\rightarrow0 } { \mathbb{e}}^{p_{\varepsilon } , p_{\varepsilon } } [ [ \bar{m}_{\cdot } ] ^{p/2}_{t } ] \\ & \leq & c_p+ c'_p(2 + 2\varrho)^{p/2}e^{1,1}[\tau^{p/2}].\end{aligned}\ ] ] the positive constants @xmath475 are independent of @xmath476 and @xmath118 , whereas @xmath477 and the random time change @xmath478_t$ ] do depend on @xmath476 .
however , the bound @xmath479_{t}\leq \tau$ ] holds for all @xmath469 and @xmath47 since @xmath480 . for @xmath481
the right - hand side is finite by theorem [ thm : theo2 ] and independent of @xmath47 . since @xmath482\leq{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ ( u_t(x)+v_t(x))^p ] $ ]
, the first direction is shown . ``
@xmath416 . ''
first note that by translation invariance of initial condition , spatial motion and white noise @xmath483={\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}\bigl[\bigl(u_t(y)+v_t(y)\bigr)^p\bigr]\ ] ] for fixed time @xmath47 and arbitrary spatial positions @xmath484 implying that @xmath483=\int_{x-1/2}^{x+1/2}{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}\bigl[\bigl(u_t(y)+v_t(y)\bigr)^p\bigr]\ , dy.\ ] ] using fubini s theorem and jensen s inequality we obtain for @xmath119 the lower bound @xmath485&\geq&{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } \biggl [ \biggl(\int _ { x-1/2}^{x+1/2}\bigl(u_t(y)+v_t(y)\bigr ) \,dy
\biggr)^p \biggr]\\ & = & { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}\bigl[\bigl\langle u_t+v_t,1_{(x-1/2,x+1/2)}\bigr\rangle^p\bigr].\end{aligned}\ ] ] we now choose an arbitrary nonnegative ( nontrivial ) smooth function @xmath26 with support contained in @xmath486 that is bounded by @xmath106 and integrates to some @xmath487 , say . a lower bound
is now given by @xmath485&\geq&{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}[\langle u_t+v_t , f\rangle ^p]\\ & = & { \mathbb{e}}^{f ,
f}[\langle\tilde u_t+\tilde v_t,{\mathbf{1}}\rangle^p]\\ & \geq&{\mathbb{e}}^{f , f}[\langle\tilde u_t,{\mathbf{1}}\rangle^p],\end{aligned}\ ] ] where we utilized for the equality the self - duality of proposition 5 of @xcite . finally , as in the discrete case , fatou s lemma and the martingale convergence theorem imply @xmath488\geq e^{c , c}[(b^1_{\tau})^p]=\infty\ ] ] by theorem [ thm : theo2 ] and due to nonnegativity of solutions as well @xmath489=\infty\ ] ] proving the claim .
_ step _ 4 . the first direction of the above proof for @xmath490 also works for the transient case since @xmath491^{p/2}_{\infty } ] \leq e^{1,1}[\tau^{p/2}]$ ] is independent of recurrence / transience .
we now study the `` criticality '' of the critical curve in more detail . as a preliminary result ( mixed ) moments of the nonspatial model are analyzed .
the idea is to combine three different techniques : the martingale argument which led to theorem [ thm : mc ] for @xmath492 $ ] , a perturbation argument based on the moment duality which allows us to deduce exponential increase / decrease of @xmath493 $ ] , and finally moment equations which yield exponential increase / decrease for all mixed moments @xmath494 $ ] . [ 0 ]
the following hold for nonspatial symbiotic branching processes : * for all @xmath67 and @xmath495 : * * @xmath492 $ ] grows to a finite constant if @xmath496 , * * @xmath492 $ ] grows subexponentially fast to infinity if @xmath171 , * * @xmath492 $ ] grows exponentially fast if @xmath167 . * for all @xmath67 , @xmath495 and @xmath497 : * * @xmath494 $ ] decreases exponentially fast if @xmath170 , * * @xmath494 $ ] neither grows exponentially fast nor decreases exponentially fast if @xmath171 , * * @xmath494 $ ] grows exponentially fast if @xmath498 .
_ step _ 1 .
martingale arguments based on the connection of moments of exit times and exit points of correlated brownian motions were carried out in the proof of theorem [ thm : mc ] .
this led to the first part of ( 1 ) . applying hlder s inequality with @xmath499 , @xmath500 , we get the bound @xmath501\leq{\mathbb{e}}^{1,1}[u_t^n]^{(n - m)/n}{\mathbb{e}}^{1,1}[v_t^n]^{m / n}={\mathbb{e}}^{1,1}[u_t^n]\ ] ] by symmetry .
this implies that for @xmath170 all mixed moments stay bounded as well .
_ step _ 2 .
we apply the moment duality for the nonspatial model as explained in remark [ rdual ] . combining the duality with the martingale argument of the first step we can understand the case @xmath170 for mixed moments in a simple way .
note that for mixed moments the dual process starts with @xmath502 particles of one color and @xmath242 particles of the other color at time @xmath107 .
note that for mixed moments @xmath503 , since there is always at least one pair of different color .
now suppose @xmath170 , then for @xmath504 we get @xmath505&=&{\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho l_t^{\neq } ) } \bigr]={\mathbb{e}}\bigl[e^{\kappa(l_t^=+(\varrho+\varepsilon ) l_t^{\neq})}e^{-\kappa
\varepsilon l_t^{\neq}}\bigr]\\ & \leq&{\mathbb{e}}\bigl[e^{\kappa(l_t^=+(\varrho+\varepsilon ) l_t^{\neq } ) } \bigr]e^{-\kappa\varepsilon t}.\end{aligned}\ ] ] since the first factor of the right - hand side is just the moment @xmath494 $ ] for @xmath506 strictly smaller than @xmath507 , this is bounded for all @xmath118 and @xmath40 .
hence , for @xmath170 all mixed moments decrease exponentially fast proving the first part of ( 2 ) .
note that since @xmath508 is a submartingale , the moment @xmath492 $ ] is nondecreasing . for @xmath171
we first consider the pure moments .
again , for the critical case , theorem [ thm : mc ] implies @xmath509<c(\varepsilon ) < \infty\ ] ] for all @xmath469 and @xmath47 . with the crude estimate @xmath510 we get @xmath511\geq{\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho(n)l_t^{\neq } ) } \bigr]e^{-\kappa\varepsilon{n\choose2}t}.\ ] ] since @xmath476 is arbitrary this implies subexponential growth to infinity of @xmath512 $ ] at the critical point .
hence , the second part of ( 1 ) is proven and combined with ( [ df ] ) so is the upper bound of the second part of ( 2 ) .
_ step _ 3 .
a direct application of it s lemma and fubini s theorem yields @xmath513=1+\kappa\pmatrix{n\cr2}\int_0^t{\mathbb{e}}^{1,1}[u_s^{n-1}v_s ] \,ds.\ ] ] since we already know from the martingale arguments that @xmath492 $ ] increases to infinity in the critical case , the mixed moment @xmath514 $ ] can not decrease exponentially fast proving the lower bound of part two of ( 2 ) .
furthermore , with the same arguments as above , for @xmath167 , this leads to @xmath515={\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho(n ) l_t^{\neq } ) } e^{\kappa(\varrho-\varrho(n))l_t^{\neq } } \bigr]\geq{\mathbb{e}}\bigl[e^{\kappa ( l_t^=+\varrho(n ) l_t^{\neq } ) } \bigr]e^{\kappa(\varrho-\varrho(n))t}.\ ] ] since the first factor of the right - hand side equals @xmath514 $ ] at the critical point , it does not decrease exponentially fast .
hence , the product increases exponentially fast .
in particular , due to ( [ df ] ) , this also implies the third part of ( 1 ) .
now it only remains to prove exponential increase for the other mixed moments .
again , using it s lemma and fubini s theorem yields the following moment equations for the mixed moments : @xmath516 & = & 1+\kappa\int_0^t{\mathbb{e}}^{1,1}[u_s^{n-1}v_s ] \,ds+\varrho(n-2)\kappa\int _ 0^t{\mathbb{e}}^{1,1}[u_s^{n-2}v_s^2 ] \,ds\\ & & { } + \pmatrix{{n-2}\cr2 } \kappa\int_0^t{\mathbb{e}}^{1,1}[u_s^{n-3}v_s^3 ] \,ds \ ] ] and similarly for all other mixed moments .
since we already know that@xmath493 $ ] grows exponentially fast in @xmath118 , this implies exponential growth of @xmath517 $ ] .
iterating this argument gives exponential growth of all mixed moments for @xmath498 .
this shows the third part of ( 2 ) and the proof is finished .
now it only remains to prove theorem [ thm : im ] , where some ideas for the nonspatial case are recycled .
proof of theorem [ thm : im ] first , due to lemma [ la : mdual ] , for homogeneous initial conditions , the moments of @xmath518 and @xmath519 are equal for all @xmath47 . for the existence of the lyapunov exponents we use a standard subadditivity argument .
hence , it suffices to show @xmath520 \leq{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ u_{t}(k)^n ] { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ u_{s}(k)^n].\ ] ] using lemma [ la : mdual ] , we reduce the problem to @xmath521 $ ] , where the dual process @xmath522 starts with @xmath151 particles of the same color all placed at site @xmath523 . by the tower property and the strong markov property ,
we obtain @xmath524={\mathbb{e}}^{n_0 } \bigl[e^{\kappa(l_{t}^=+\varrho l_{t}^{\neq})}{\mathbb{e}}^{n_t } \bigl[e^{\kappa(l_{s}^=+\varrho l_{s}^{\neq } ) } \bigr ] \bigr].\ ] ] we are done if we can show that @xmath525\leq{\mathbb{e}}^{n_0 } \bigl[e^{\kappa(l_{s}^=+\varrho l_{s}^{\neq } ) } \bigr]\ ] ] for any given initial configuration @xmath526 of the dual process consisting of @xmath151 particles .
the general initial conditions of the dual process consist of @xmath527 particles of one color and @xmath528 particles of the other color ( @xmath529 ) distributed arbitrarily in space at positions @xmath530 . using the duality relation of lemma [ la : mdual ] , we obtain @xmath531&=&{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1}}}[u_s(k_1 ) \cdot\cdot\cdot u_s(k_{n^1})v_s(k_{n^1 + 1})\cdot\cdot\cdot v_s(k_{n^1+n^2})]\\ & \leq&{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ u_s(k)^{n } ] = { \mathbb{e}}^{n_0 } \bigl[e^{\kappa ( l_{s}^=+\varrho l_{s}^{\neq } ) } \bigr],\end{aligned}\ ] ] where , in the penultimate step , we have used the generalized hlder inequality . having established existence of the lyapunov exponents
, we now turn to the more interesting question of positivity .
the boundedness for @xmath170 in theorem [ thm : mc ] immediately implies that in this case @xmath532 .
now suppose @xmath533 , that is , @xmath534 lies on critical curve .
we use the perturbation argument which we already used for the nonspatial case combined with lemma [ la : mdual ] and theorem [ thm : mc ] to prove that in this case moments only grow subexponentially fast .
this implies that the lyapunov exponents are zero .
again we switch from @xmath535 $ ] to @xmath536 $ ] , where the dual process is started with all particles at the same site and the same color .
since moments below the critical curve are bounded , we can proceed as for the nonspatial model . for any @xmath469
, we get @xmath537 \geq{\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho l_t^{\neq } ) } \bigr]e^{-\kappa \varepsilon{n\choose2}t } \\ & \geq&{\mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ u_t(k)^n ] e^{-\kappa\varepsilon{n\choose 2}t},\end{aligned}\ ] ] where we estimated the collision time of particles of different colors by the collision time of all particles which is bounded from above by @xmath538 .
since @xmath476 on the right - hand side is arbitrary , @xmath539 can not be positive .
finally , we assume @xmath167 .
the idea is to reduce the problem to the nonspatial case which we already discussed in proposition [ 0 ] . actually , we prove more than stated in the theorem since we also show that mixed moments @xmath540 $ ] grow exponentially fast . for @xmath497
the perturbation argument leads to @xmath541={\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho l_t^{\neq } ) } \bigr]={\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho(n ) l_t^{\neq } ) } e^{\kappa(\varrho-\varrho(n))l_t^{\neq } } \bigr].\ ] ] the idea is to obtain a lower bound by conditioning on the event that all particles have not changed their spatial positions before time @xmath118 ( but , of course , have changed their colors ) . under this condition
the particle dual is precisely the particle dual of the nonspatial model .
more precisely , we get the lower bound @xmath542\\ & & \qquad={\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho(n ) l_t^{\neq})}e^{\kappa(\varrho -\varrho(n))l_t^{\neq } } |\mbox{no spatial change of particles before time } t \bigr]\\ & & \qquad\quad { }
\times{\mathbb{p}}[\mbox{no spatial change of particles before time } t ] \\ & & \qquad={\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho(n ) l_t^{\neq})}e^{\kappa(\varrho -\varrho(n))l_t^{\neq } } |\mbox{no spatial change of particles } \bigr]e^{-nt},\end{aligned}\ ] ] where the final equality is valid since the event @xmath543no spatial change of particles before time @xmath544 has probability @xmath545 .
this is true since the event is precisely the event that @xmath151 independent exponential clocks with parameter @xmath106 did not ring before time @xmath118 .
for @xmath546 there is always at least one pair of particles of different colors and , hence , we get the lower bound @xmath547e^{\kappa(\varrho-\varrho(n))t}e^{-nt},\ ] ] which equals @xmath548e^{\kappa(\varrho-\varrho(n))t}e^{-nt}\ ] ] for a nonspatial symbiotic branching process with critical correlation @xmath171 .
choosing @xmath40 such that @xmath549 the result now follows from proposition [ 0 ] .
as mentioned in the course of the proof , we actually proved that for @xmath167 and @xmath550 @xmath551\ ] ] grows exponentially in @xmath118 .
as for the nonspatial model one could ask whether , and if so how fast , mixed moments decrease for @xmath552 . for the second moments it was shown in @xcite that for @xmath553 @xmath554\approx \cases { \dfrac{1}{\sqrt{t } } , & \quad $ d=1$,\cr \dfrac{1}{\log(t ) } , & \quad $ d=2$,\vspace*{2pt}\cr 1 , & \quad $ d\geq3 $ , } \ ] ] where @xmath555 denotes weak asymptotic equivalence as @xmath77 . it would be interesting to see whether or not different rates of decrease appear for moments .
a detailed quantitative study of the lyapunov exponents as functions of @xmath54 and @xmath40 has so far only been carried out for second moments ( see @xcite ) .
in contrast to the parabolic anderson model , where higher lyapunov exponents are well studied ( see @xcite ) , we do not have much insight .
only a first upper bound for the lyapunov exponents in @xmath40 and the distance to the critical curve can be obtained from the perturbation argument of the previous proof . [ up ] if @xmath167 , then @xmath556 . by lemma [ la : mdual ] and theorem [ thm : mc ] for @xmath557 , there are constants @xmath558 such that @xmath559\\ & = & { \mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho l_t^{\neq})}e^{-\kappa ( \varrho-\varrho(n)+\varepsilon)l_t^{\neq } } \bigr]\\ & \geq&{\mathbb{e}}\bigl[e^{\kappa(l_t^=+\varrho l_t^{\neq } ) } \bigr]e^{-\kappa ( \varrho-\varrho(n)+\varepsilon ) { n\choose2}t}.\end{aligned}\ ] ] hence , for all @xmath469 @xmath162\leq c(\varepsilon)e^{\kappa(\varrho -\varrho ( n)+\varepsilon ) { n\choose2}t},\ ] ] yielding the result .
in this section we show how to use the moment bounds of theorem [ thm : mc ] to obtain an improved upper bound on the speed of propagation of the interface as defined in definition [ def : ifc ] .
we will only sketch the crucial parts in the proof of theorem 6 of @xcite that need modification .
note that the method used here is based on mueller s `` dyadic grid technique '' introduced in @xcite .
proof of theorem [ cor : wavespeed ] to prove that the interface will eventually be contained in @xmath560\ ] ] ( for suitable @xmath194 ) , by symmetry , it suffices to show that the right endpoint of the interface @xmath561 up to time @xmath202 can eventually be bounded by @xmath562 . to this end
we define @xmath563 and show that , for suitably chosen @xmath564 , @xmath565 . by the borel
cantelli lemma , this follows from @xmath566 in the following we modify the arguments of @xcite to obtain an upper bound for @xmath567 which is sumamble over @xmath151 .
[ la : m ] for any integer @xmath151 there is a finite constant @xmath568 such that for @xmath569 @xmath570\leq c_n\sqrt{p_t1_{{\mathbb{r}}^-}(x)},\qquad x\in{\mathbb{r } } , t\geq0.\ ] ] first recall from ( 87 ) of @xcite that @xmath571=p_t1_{{\mathbb{r}}^-}(x)$ ] .
we now use hlder s inequality and theorem [ thm : mc ] to reduce the mixed moment to the first moment : @xmath572\\ & & \qquad={\mathbb{e}}^{1_{{\mathbb{r}}^-},1_{{\mathbb{r}}^+ } } [ u_t(x)^{1/2}u_t(x)^{n-1/2}v_t(x)^n ] \\ & & \qquad\leq({\mathbb{e}}^{1_{{\mathbb{r}}^-},1_{{\mathbb{r}}^+ } } [ u_t(x ) ] ) ^{1/2 } ( { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ u_t(x)^{4n-1 } ] ) ^{({2n-1})/({8n-2})}\\ & & \qquad\quad{}\times ( { \mathbb{e}}^{{\mathbf{1}},{\mathbf{1 } } } [ v_t(x)^{4n-1 } ] ) ^{{n}/({4n-1})}.\end{aligned}\ ] ] this follows from the generalized hlder inequality with exponents @xmath573 and @xmath574 .
the first factor yields the heat flow and theorem [ thm : mc ] shows that the latter two factors are bounded by constants for @xmath569 .
we now strengthen the estimate of lemma 23 of @xcite of the stochastic part @xmath575 of the convolution representation of solutions of corollary 20 of @xcite . [
la : ma ] for @xmath197 there is a constant @xmath576 such that for @xmath577 , @xmath578 , the following estimate holds : @xmath579 the proof is along the same lines of @xcite replacing only in ( 116 ) the weaker ( exponentially growing ) moment bound of @xcite by our stronger ( bounded ) moment bound . in the following we sketch the arguments to show where the moments appear . before performing
the `` dyadic grid technique , '' increments of @xmath580 need to be estimated .
first , by definition @xmath581\\ & & \qquad={\mathbb{e}}^{1_{{\mathbb{r}}^-},1_{{\mathbb{r}}^+ } } \biggl [ \biggl|\int_0^t\int_{{\mathbb{r}}}\bigl(p_{t - s}(b - a)-p_{t'-s}(b - a')\bigr)m(ds , db ) \biggr|^{2q } \biggr],\end{aligned}\ ] ] which by burkholder davis
gundy and hlder s inequality gives the upper bound @xmath582 ^ 2 \,db \,ds \biggr|^{q-1}\\ & & \qquad{}\times\int_0^t\int_{{\mathbb{r}}}[p_{t - s}(b - a)-p_{t'-s}(b - a')]^2 { \mathbb{e}}^{1_{{\mathbb{r}}^-},1_{{\mathbb{r}}^+ } } [ ( u_s(b)v_s(b))^q ] \,db \,ds.\end{aligned}\ ] ] using lemma [ la : m ] and classical heat kernel estimates we can derive ( see the calculation on pages 153 , 154 of @xcite ) the upper bound @xmath583\\ & & \qquad\leq c_2\bigl((|t'-t|^{1/2}+|a'-a|)\wedge t^{1/2}\bigr)^{q-1 } \bigl(\sqrt { tp_t1_{{\mathbb{r}}^-}(a)}+\sqrt{t'p_{t'}1_{{\mathbb{r}}^-}(a ' ) } \bigr).\end{aligned}\ ] ] this upper bound corresponds to ( 119 ) of @xcite where they have an additional exponentially growing factor coming from their moment bound .
the dyadic grid technique can now be carried out as in @xcite , choosing @xmath584 , without carrying along their exponential factor .
hence , we may delete the exponential term from their final estimate ( 110 ) .
note that the necessity of @xmath197 comes from our choice @xmath584 and lemma [ la : m ] .
the following lemma corresponds to proposition 24 of @xcite . if @xmath197 then , for some constants @xmath585 , the following estimate holds for @xmath586 and @xmath587 : @xmath588 all we need to do is to argue that proposition 24 of @xcite is valid for @xmath587 instead of @xmath589 .
we perform the same decomposition and note that the estimates of step 2 of @xcite are already given for @xmath587 if @xmath590 is large enough .
the only trouble occurs in their step 3 . up to the estimate ( 154 )
, this step works for @xmath587 but here their ( weaker ) lemma 23 produces an exponential in @xmath202 .
more precisely , they need to justify @xmath591 which is only valid for @xmath592 . as our lemma
[ la : ma ] avoids the exponential on the left - hand side the estimate holds for @xmath593 with suitably chosen @xmath590 and @xmath594 .
the significant distinction of the previous lemma to the result of @xcite is that the inequality is not only valid for @xmath595 but for @xmath593 . at this point
one might hope to obtain a square - root upper bound for the growth of the interface but this fails in the final step in which we validate ( [ abc ] ) : @xmath596 which is finite for @xmath564 large enough .
this work is part of the ph.d .
thesis of the second author who would like to thank the students and faculty from tu berlin for many discussions .
the authors would like to express their gratitude to an anonymous referee for a very careful reading of the manuscript and for pointing out and correcting an error in an earlier version of theorem [ cor : wavespeed ] . | in this paper we introduce a critical curve separating the asymptotic behavior of the moments of the symbiotic branching model , introduced by etheridge and fleischmann [ _ stochastic process .
appl . _
* 114 * ( 2004 ) 127160 ] into two regimes . using arguments based on two different dualities and a classical result of spitzer [ _ trans .
amer .
math .
soc . _
* 87 * ( 1958 ) 187197 ] on the exit - time of a planar brownian motion from a wedge , we prove that the parameter governing the model provides regimes of bounded and exponentially growing moments separated by subexponential growth .
the moments turn out to be closely linked to the limiting distribution as time tends to infinity .
the limiting distribution can be derived by a self - duality argument extending a result of dawson and perkins [ _ ann .
probab . _
* 26 * ( 1998 ) 10881138 ] for the mutually catalytic branching model . as an application
, we show how a bound on the @xmath0th moment improves the result of etheridge and fleischmann [ _ stochastic process .
appl . _
* 114 * ( 2004 ) 127160 ] on the speed of the propagation of the interface of the symbiotic branching model . , and . |
galactic evolution is driven at the end - points of stellar evolution , where copious mass loss returns chemically - enriched and sometimes dusty matter back to the interstellar medium ( ism ) ; the stellar winds of evolved stars and the violent deaths of the most massive stars also inject energy and momentum into the ism , generating turbulence and galactic fountains when superbubbles pop as they reach the `` surface '' of the galactic disc .
the evolved stars are also excellent tracers , not just of the feedback processes , but also of the underlying populations , that were formed from millions to billions of years prior to their appearance
. the evolved phases of evolution generally represent the most luminous , and often the coolest , making evolved stars brilliant beacons at ir wavelengths , where it is also easier to see them deep inside galaxies as dust is more tranpsarent at those longer wavelengths than in the optical and ultraviolet where their main - sequence progenitors shine .
the final stages of stellar evolution of stars with main - sequence masses up to @xmath6 m@xmath1 asymptotic giant branch ( agb ) stars and red supergiants are characterised by strong radial pulsations of the cool atmospheric layers , rendering them identifiable as long - period variables ( lpvs ) in photometric monitoring campaigns spanning months to years ( e.g. , whitelock , feast & catchpole 1991 ; wood 2000 ) .
the local group galaxy triangulum ( hodierna 1654 ) hereafter referred to as m33 ( messier 1771 ) offers us a unique opportunity to study stellar populations , their history and their feedback across an entire spiral galaxy and in particular in its central regions , that in our own milky way are heavily obscured by the intervening dusty disc ( van loon et al .
2003 ; benjamin et al .
our viewing angle with respect to the m33 disc is more favourable ( 56@xmath7 zaritsky , elston & hill 1989 ; deul & van der hulst 1987 ) than that of the larger m31 ( andromeda ) , whilst the distance to m33 is not much different from that to m31 ( @xmath8 mag bonanos et al .
large populations of agb stars have been identified in m33 ( cioni et al .
2008 ) , and red supergiants up to progenitor masses in excess of 20 m@xmath1 ( drout , massey & meynet 2012 ) .
many of them are dusty lpvs ( mcquinn et al . 2007 ; thompson et al .
2009 ) , and these have been found also in the central parts of m33 ( javadi , van loon and mirtorabi 2011a ) .
the main objectives of our project are described in javadi , van loon & mirtorabi ( 2011c ) : to construct the mass function of lpvs and derive from this the star formation history in m33 ; to correlate spatial distributions of the lpvs of different mass with galactic structures ( spheroid , disc and spiral arm components ) ; to measure the rate at which dust is produced and fed into the ism ; to establish correlations between the dust production rate , luminosity , and amplitude of an lpv ; and to compare the _ in situ _ dust replenishment with the amount of pre - existing dust .
paper i in the series presented the photometric catalogue of stars in the inner square kpc ( javadi et al .
2011a ) , and paper ii presented the galactic structure and star formation history in the inner square kpc ( javadi , van loon and mirtorabi 2011b ) .
this is paper iii , describing the mass - loss mechanism and dust production rate in the inner square kpc .
subsequent papers in the series will cover the extension to a nearly square degree area covering much of the m33 optical disc .
to derive the mass - loss rates of the red giant variables we follow a two - step approach .
first , we model the spectral energy distributions ( seds ) of near - ir variables for which mid - ir counterparts have been identified .
we use these results to construct relations between the dust optical depth and bolometric corrections on the one hand , and near - ir colours on the other .
then , we apply those relations to the red giant variables for which no mid - ir counterpart was identified , to derive their mass - loss rates too . images in the j , h and k@xmath9 bands were obtained with the united kingdom infrared telescope ( ukirt ) with the uist instrument , covering an area of approximately @xmath10 corresponding to a square kpc at the distance of m33 .
observations in the k@xmath9 band were made on about a dozen occasions over the period of 20032007 .
the data , point spread function ( psf ) fitting photometry with daophot ( stetson 1987 ) and variability analysis are described in paper i. the full , publicly available catalogue contains 18398 stars ; 812 of these were identified as variables on the basis of the near - ir photometry alone , most of them pulsating red giant stars ( agb stars and red supergiants ) .
images in the 3.6 , 4.5 , 5.8 and 8 @xmath11 m bands were obtained with the _ spitzer _ space telescope with the irac instrument , on six occasions over the period of 20042006 .
we use the photometric catalogue of mcquinn et al.(2007 ) , which is based on psf fitting with daophot to the first five epochs and excludes the less sensitive 5.8 @xmath11 m band .
thompson et al.(2009 ) have published a somewhat deeper catalogue in the 3.6 and 4.5 @xmath11 m bands including the sixth epoch , but they have not made publicly available their photometry obtained at longer wavelengths and we therefore prefer working with the homogeneous catalogue of mcquinn et al .
we cross - matched our near - ir catalogue to their catalogue ( see paper i for details ) and found 523 matches in the 3.6 @xmath11 m band , 471 at 4.5 @xmath11 m and 107 at 8 @xmath11 m ( fig.1 ) .
we noticed ( in paper i ) a discrepancy between the [ 3.6][4.5 ] colours and theoretical isochrones ( from marigo et al .
2008 ) , with the faintest stars being abnormally blue .
this is not apparent in the colour magnitude diagram of m33 as a whole ( see mcquinn et al .
2007 ) and is most likely related to the specific challenges encountered in the crowded central regions of the galaxy ( though strong absorption in the fundamental band of carbon monoxide can depress the 4.5 @xmath11 m brightness , which is a challenge for models to predict accurately ) .
the psf of the mid - ir data has a full width at half maximum of @xmath12 , which is more than twice as large as that of the near - ir data . to gain an appreciation of the severity and effects of blending
, we performed a simple simulation .
first , we picked an individual k@xmath9 band frame and removed all detected stars from it with daophot / allstar ( stetson 1987 ) .
next , we placed each of these stars back into the frame , at their original positions but using a @xmath12 psf to mimic the appearance at 3.6 @xmath11 m .
we assumed that none of these stars have an infrared excess due to circumstellar dust , so their 3.6 @xmath11 m counts can be derived from their k@xmath9 magnitude by applying a typical k@xmath9[3.6 ] colour of a few tenths of a magnitude ( see figure 4 in van loon , marshall & zijlstra 2005 ) .
then , we used daophot / allstar and daomaster to detect and photomeasure the stars on this pseudo-3.6 @xmath11 m frame .
the distributions of input and recovered pseudo-3.6 @xmath11 m magnitudes ( fig.2 ) demonstrate that in the crowded central regions of m33 the 3.6 @xmath11 m photometry becomes severely incomplete for stars fainter than about magnitude 16 , well before the k@xmath9 band data .
hence our simulation also accounts for unresolved emission at 3.6 @xmath11 m from stars too faint to be detected individually .
the recovery rate of stars brighter than magnitude 16 is very high .
stars brighter than magnitude 16 are recovered with reliable photometry , at an accuracy ( much ) better than 0.3 mag and showing no systematic offset ( fig.3 ) .
visual inspection of the few bright sources with large photometric discrepancies revealed that most of these stars were located near the very nucleus of m33 or near the edge of the frame .
stars fainter than magnitude 16 start showing the effects of blending , leading to a reduction in the number of stars that are recovered ( fig . 2 ) and a systematic over - estimation of their brightness ( fig . 3 ) .
a useful number of stars ( 107 ) are detected at 8 @xmath11 m because of excess emission from circumstellar dust , which allows derivation of their mass - loss rates .
the stars form a branch of increasing mid - ir brightness with increasing near-/mid - ir colour ( fig . 4 ) .
the brightest mid - ir objects , @xmath13<10 $ ] mag , with very red colours , ( @xmath14-[8])\approx$ ] 5 mag , deviate from this main branch .
these have not been identified by us as variable stars , and could be non - stellar in origin , for instance background galaxies or compact hii regions within m33 .
we will come back to these and other deviant sources at the end of section 2.3 and in section 3.1 .
rccc i d & @xmath15 & @xmath16 ( l@xmath1 ) & @xmath17 ( m@xmath1 yr@xmath2 ) + + 629 & 0.07 & 4.45 & @xmath18 + 1003 & 0.03 & 4.42 & @xmath19 + 1590 & 0.06 & 4.42 & @xmath20 + ... & ... & ... & ... + + 97 & 0.30 & 3.85 & @xmath21 + 901 & 0.20 & 4.15 & @xmath22 + 2323 & 0.10 & 3.92 & @xmath23 + ... & ... & ... & ... + we modelled the seds of all ukirt variable stars with measurements in at least two near - ir bands ( k@xmath9 and j and/or h ) and two mid - ir bands ( 3.6 , 4.5 and/or 8 @xmath11 m ) , with the publicly available dust radiative transfer code dusty ( based on ivezi & elitzur 1997 ) . because of the scarcity of the photometry we fixed the input temperatures of the star and of the dust at the inner edge of the circumstellar envelope , at 3000 and 900 k , respectively . the density structure is assumed to follow from the analytical approximation for radiatively driven winds ( ivezi & elitzur 1997 )
. based on the estimated birth mass of the star ( see paper ii ) we used amorphous carbon dust ( hanner 1988 ) and a small amount of silicon carbide ( pgouri 1988 ) for @xmath24 , and astronomical silicates ( draine & lee 1984 ) for all other mass ranges , with a gas - to - dust mass ratio of @xmath25 . the optical depth ( @xmath15 )
was varied and the luminosity ( @xmath26 ) was scaled until an acceptable match was obtained , which was decided by visual inspection .
the optical depth and luminosity , and hence mass - loss rate ( @xmath27 ) , were determined for 58 carbon stars and 35 m - type stars ( table 1 ) ; examples are presented in figures 5 ( carbon stars ) and 6 ( m - type stars ) . given the limited number of free parameters even a small number of data points constrain the model fit quite well . in all panels of figs . 5 and 6 we not only show the best fit for the preferred dust species but also that for the alternative dust species .
while it is not always possible to tell , on the basis of the fit , which species of dust is present , the preferred dust species often yields better fits and hardly ever yields worse fits .
our study spans several orders of magnitude in both mass - loss rate and luminosity , and the uncertainties resulting from the assumptions in the fitting procedures are thus relatively unimportant .
a heavily dust - enshrouded pulsating giant star will still be fitted with a model with high optical depth regardless whether it employs carbonaceous or oxygenous dust .
we do note , though , that often the 4.5-@xmath11 m datum is anomalously faint compared to adjacent bands .
this may be due to molecular absorption , which was not included in our sed modelling .
the fundamental ro - vibrational band of co at 4.6 @xmath11 m can be strong especially in the extended atmospheres of pulsating red giant stars ( cf .
nowotny et al.2013 ) .
the 3-@xmath11 m c@xmath28h@xmath28+hcn band is very strong in carbon stars , but it falls largely outside the irac 3.6-@xmath11 m bandpass ; the 3.8-@xmath11 m c@xmath28h@xmath28 band , while strong in metal - poor carbon stars ( van loon et al . 2006 , 2008 ) , is not expected to be nearly as strong in the carbon stars in the central regions of m33 that are of near - solar metallicity . while this approach may seem crude ( by necessity ) , we shall see that the results compare favourably with those obtained for similar populations of stars in the magellanic clouds .
the self - similarity of radiation - driven winds leads to scaling relations ( ivezi & elitzur 1997 ) , with the combination of @xmath29 being approximately constant .
indeed it is , roughly , but with scatter ( @xmath30 dex ; fig .
7 ) due to slight mis - matches between @xmath15 and @xmath31 to the exact shape of the sed .
the amount of scatter is rather modest , though , and suggests that the mass - loss rates for these stars are accurate to well within a factor two ( cf .
srinivasan , sargent & meixner ( 2011 ) ; the luminosities for these stars are obtained with an accuracy of around 1030 per cent ) .
the offset between the carbon stars and m - type stars ( @xmath32 dex @xmath33 ) is expected as this is due primarily to the differences in opacity of the different dust species .
most though not all stars originally classified as carbon stars but fitted with silicates end up in a location slightly off - set from that of the m - type stars .
this suggests that most of these stars are indeed different from the m - type stars and therefore quite possibly genuine carbon stars , but that some may have been misclassified and are in fact m type .
.parameterisations of relations between optical depth ( @xmath15 ) and near - ir colour ( @xmath34 ) of the form @xmath35 for @xmath36 mag ( and @xmath37 for @xmath38 mag ) .
[ cols="<,<,<,<,^ , < ,
< , < " , ] likewise , the bolometric correction to the @xmath39 band ( bc@xmath40 ) shows a dependency on near - ir colour ( fig .
9 ) albeit with considerable scatter and with occasionally large deviations for h k@xmath9 .
this relation depends on the underlying star ( its temperature ) , and the carbon stars fitted with silicates fall in between carbon stars fitted with carbonaceous dust and m - type stars fitted with silicates ; this also drives the difference noted in figure 7 .
luminosities are estimated for the sample at large by applying a parametrisation of the relation between bc@xmath40 and j
k@xmath9 if @xmath41 is available and h
k@xmath9 if @xmath42 ( but not @xmath41 ) is available ( see table 3 ) . the scatter in fig .
9 suggests that an accuracy in the luminosity is achieved of typically 3050 per cent .
there exist pitfalls with this approach , which could lead to sources being erroneously assigned high luminosities and mass - loss rates .
one such case comprises stars that have no j- or h - band counterpart because they are near to the edges of the survey area , or because they are heavily extincted by interstellar dust .
such sources would not normally have been identified as variable in our survey .
we thus decided that we would ignore stars that have no j- or h - band magnitude and are not variable .
this includes the few very bright mid - ir sources mentioned at the end of section 2.2 , that are likely non - stellar in origin .
when they have no j- or h - band magnitude but are variable , inspection of the brightest such sources revealed no suspicion regarding their photometry and so for these stars we do assign a j - band magnitude viz .
equal to a ( conservative ) detection limit of @xmath43 mag . of just twelve such cases , the most extreme example is # 11887 : a large - amplitude variable ( @xmath44 mag ) , with luminosity @xmath45 and mass - loss rate @xmath46 .
the mass - loss rate shows some dependence on luminosity ( fig .
firstly , the highest rates are generally achieved by the most luminous , most massive stars , in agreement with earlier findings in the magellanic clouds ( van loon et al .
1999 ; srinivasan et al .
2009 ) . except for a few more extreme examples that would probably have been missed in our survey , carbon stars in the milky way reach mass - loss rates of a few @xmath47 m@xmath1 yr@xmath2 ( whitelock et al . 2006 ) , whilst those in the large magellanic cloud ( lmc ) reach @xmath48 m@xmath1 yr@xmath2 ( gullieuszik et al .
2012 ) , i.e.comparable to the most extreme carbon stars in our m33 sample . in the solar neighbourhood ,
m - type agb stars reach similar mass - loss rates , a few @xmath47 m@xmath1 yr@xmath2 ( jura & kleinmann 1989 ) , whilst m - type supergiants display rates between @xmath49 m@xmath1 yr@xmath2 with the majority in excess of a few @xmath50 m@xmath1 yr@xmath2 ( jura & kleinmann 1990 ) , again both very similar to what we find in m33 . much of the spread in mass - loss rate at a given luminosity is likely to reflect stellar evolution ( cf
. van loon et al . 1999 , 2005a ) .
the least luminous carbon stars ( @xmath51 ) are probably not yet at the tip of their agb because m - type agb stars are found at the same luminosities but higher mass - loss rates the latter are likely more evolved than the carbon stars ; but these carbon stars are not expected to become m - type again ( marigo et al.2008 ) so they must still evolve to higher luminosities before they end their agb evolution . with @xmath27 of a few @xmath50 m@xmath1 yr@xmath2 they exhibit several times lower mass - loss rates than those near the tip of their agb evolution ( around @xmath52 l@xmath1 ) where they reach @xmath27 in excess of @xmath53 m@xmath1 yr@xmath2 .
this suggests that the mass - loss rate increases as a star climbs the agb . alternatively , the less luminous carbon stars with lower mass - loss rates may be in their inter - thermal pulse luminosity dip and/or the more luminous carbon stars with high mass - loss rates may just be experiencing the aftermath of a thermal pulse ( olofsson et al .
1990 ; vassiliadis & wood 1993 ; mattsson , hfner & herwig 2007 ) . fitting the stars that we originally classified as carbon stars with silicates generally yields higher mass - loss rates because of the lower specific opacity of silicates compared to amorphous carbon grains .
one star , # 12831 ( @xmath54 mag , @xmath55 mag ) has in this way become one of the more extremely mass - losing objects , with @xmath56 nine times higher than if fitted with silicates .
the inferred birth mass would now be 6.7 m@xmath1 .
evolution of red supergiants is mostly in terms of effective temperature ( @xmath57 ) rather than luminosity .
the dusty wind mass - loss rate depends sensitively on @xmath57 ( van loon et al .
2005a ; cf .
bonanos et al.2010 ) ; because red supergiants do not become as cool as massive agb stars the highest rates achieved by red supergiants fall a little behind those achieved by the most extreme agb stars considering that the red supergiants are more luminous ( fig .
yet there is no luminosity gap in the stars with the highest mass - loss rates , confirming our earlier suggestion ( paper ii ) that super - agb stars become dust - enshrouded .
super - agb stars are strictly speaking red supergiants .
however , the most massive agb stars i.e. those that will _ not _ ignite core carbon burning experience hot bottom burning ( hbb ; iben & renzini 1983 ) ; this not only prevents them from turning into carbon stars but also enhances their luminosity up and above the classical core - mass luminosity relation ( boothroyd & sackmann 1992 ) .
these stars may thus reach luminosities that exceed the classical agb limit ( indicated in fig .
10 ) , making it difficult to distinguish between an agb star experiencing hbb and a super - agb star or red supergiant .
they may separate in @xmath57@xmath26 diagrams ( van loon et al .
2005a ) and thus also in @xmath58@xmath26 diagrams ( cf .
wood et al .
1992 ; whitelock et al .
2003 ) and perhaps in @xmath27@xmath26 diagrams too .
it is not very well established what is the maximum luminosity than can be reached under the influence of hbb , but it seems unlikely that it comprises much more than 50 per cent ( 0.2 dex ) . depending on the timing of the ensuing supernova explosion , red supergiants might be found in the dust - enshrouded phase , or a preceding or following phase characterised by a thinner dust envelope .
but if super - agb stars explode they are likely to be dust - enshrouded , explaining explosions such as sn2008s ( botticella et al .
2009 ) .
the sample of stars that were modelled with dusty is not complete , and we therefore include in figure 11 also the ukirt variable stars for which no _ spitzer _ photometry is available as well as non - variable stars .
we note the most extreme datum in this graph , # 13093 , at @xmath59 and @xmath60 . with @xmath61 mag and @xmath62 mag
it is consistent with being a heavily - reddened @xmath63 m@xmath1 supergiant ; we also identified it as a large - amplitude variable ( @xmath64 mag ) .
our k@xmath9-band image shows a point source unaffected by crowding .
no _ spitzer _
photometry is available , neither from mcquinn et al .
( 2007 ) nor thompson et al .
( 2009 ) . inspection of the irac images ( 3.6 , 4.5 , 5.8 and 8 @xmath11 m ) from the _ spitzer _ archive , however , clearly reveals the source , adding credence to it being an extremely dusty evolved star .
it sits next to an extended area of mid - ir emission , though , which may explain why it was not measured .
it is almost twice as bright at 8 @xmath11 m as the highest mass - losing ukirt variable that we modelled with dusty ( # 15552 ) , broadly in line with the factor 2.5 difference in their mass - loss rates . on the other hand , the star with the next - highest mass - loss rate , # 12208 ( @xmath65 , @xmath66 ) is more suspicious . with @xmath67 mag
it is bright , and with @xmath68 mag it is red , but its estimated k@xmath9-band amplitude is moderate ( @xmath69 mag ) .
it is the brightest ukirt variable in this central region of m33 at 8 @xmath11 m , by far ( see fig . 4 ) . while it appears unresolved on the 8-@xmath11 m image it sits on top of more complex emission , within @xmath70 of a catalogued molecular cloud ( bolatto et al .
2008 ) and a supernova remnant ( gordon et al.1999 ) ; this suggests it might be a young stellar object or even an ultra - compact hii region .
a third object worth noting is the candidate @xmath71carin analogue m33 - 8 ( khan , stanek & kochanek 2012 ) .
it is the only such object in their list of nine in m33 that ( just ) falls within our uist / ufti survey .
they combined optical photometry with 2mass near - ir and _ spitzer _ mid - ir data , to derive the spectral energy distribution and luminosity .
they show a _ hubble _ space telescope image in the visual dominated by a single source surrounded by a clustering of fainter stars .
our ufti image reveals the dominant source is in fact a duplet in the k@xmath9 band , which would appear unresolved in the 2mass image .
the uist catalogue lists them as # 18469 , with @xmath72 mag , and # 18441 , with @xmath73 mag .
this compares well with the i - band magnitude ( 0.8 @xmath11 m ) listed in khan et al .
( 2012 ) of @xmath74 mag .
we estimate luminosities of @xmath75 and 3.68 , respectively ; i.e. together they fall well short of accounting for the integrated ir luminosity of @xmath76 derived by khan et al
. the latter must include significant contributions from other sources , especially as it includes the 24-@xmath11 m _ spitzer _ emission which has an angular resolution of only @xmath70 .
we derive mass - loss rates of @xmath77 and @xmath78 for # 18469 and # 18441 , respectively ; substantial , but not extreme .
close pairs of infrared - bright stars are not unique . for instance , # 19002 is an unresolved blend of two roughly equally - bright stars .
we assigned a luminosity of @xmath79 and @xmath80 to the blended object ; certainly at least one of the two stars is very dusty . the mass - loss rate increases with increasing k@xmath9-band amplitude ( fig.12 ) .
the correlation is quite clear in spite of the relative inaccuracy of the amplitudes ( determined from sparsely sampled lightcurves ; see paper i ) .
the correlation is consistent with a rle for stellar pulsation in driving the winds from cool stars ; it also confirms that , whilst the _ relative _ amplitudes ( expressed in magnitudes ) are smaller for more luminous stars ( the m - type stars with @xmath81 m@xmath1 are the most luminous , those with @xmath82 m@xmath1 are generally the least luminous ) , for the same amplitude the mass - loss rates are higher for more luminous stars : mass loss is more directly related to the _ absolute _ amplitude ( expressed in luminosity units ) ( van loon et al .
the k@xmath9 band is relatively insensitive to variations in stellar temperature , circumstellar extinction and dust emission as it is near the peak of the stellar sed and in between the strong attenuation by dust at shorter wavelengths and its emission at longer wavelengths hence also the preference for expressing bolometric corrections in relation to the k or k@xmath9 band .
however , it is important , but difficult in practice , to separate the effects of pulsation and temperature on mass loss , as both the pulsation strengthens and the temperature decreases as the star ascends the rgb , agb or red supergiant branch ( see whitelock , feast & pottasch 1987 ) .
the carbon stars confirm this picture in the sense that their amplitudes are intermediate between the more and less luminous m - type stars , but their mass - loss rates seem a little lower . as expected , the mass - loss rates of presumed carbon stars when fitted with silicates fall in between those of low - mass and higher - mass m - type stars .
this fact neither confirms nor refutes their carbon star nature .
mass - loss rates have often been correlated with infrared colour , to provide a simple recipe for converting infrared colours to mass - loss rates ; however , these methods depend on the star s luminosity as well as the dust properties ( see van loon 2007 for a review ) .
figure 13 exemplifies this once more : the carbon stars become easily reddened by their dust but that does not necessarily imply as high a mass - loss rate as a less reddened m - type star . likewise , among stars with similar dust the less luminous stars are more readily reddened than more luminous ones ( van loon et al .
thus , care must be taken when converting an infrared colour into a mass - loss rate .
the carbon stars luminosities are intermediate between those of the low - mass and higher - mass m - type stars , and indeed their mass - loss rates fall in between those groups mass - loss rates if also the presumed carbon stars are fitted with silicates .
assessments of the integrated mass loss ( or dust production ) from a stellar population are fraught with uncertainties . in small populations ,
stochastic effects are caused by very small numbers of stars which may dominate the budget ( van loon et al .
2005b ; mcdonald et al .
2009 ) , but it is also difficult to accurately account for the many stars that each contribute little .
most red giants undergo mild mass loss and their circumstellar envelopes are not very dusty ; photometric tracers of this mass loss and dust will be affected by photometric uncertainties and scatter in interstellar reddening , as well as uncertainties in the underlying photospheric spectral energy distribution ( cf .
mcdonald , zijlstra & boyer 2012 ) .
interstellar reddening towards the central region of m33 is modest ( see paper i ) ; whilst it is negligible for the dustiest stars it may lead to an over - estimate of the amount of dust produced by stars with little mass loss .
photometric scatter will also cause some stars to appear to have more dust than they really do , but equally cause the opposite effect in others .
as this is of a purely statistical nature we can try to compensate for it by considering the distribution of stars over negative values for the optical depth as one would have derived them from the near - ir colour relation ( fig .
14 , top ) .
a similar distribution of erroneous optical depth values would extent towards positive values , and this distribution can be subtracted to yield the true distribution over optical depth ( fig . 14 , bottom ) .
the fraction of stars that were identified to be variable increases with increasing optical depth , to about unity for the most obscured specimens ( fig .
this is encouraging , suggesting that the dustiest stars produced most of that dust themselves .
we can not be certain , however , that the dust inferred for the non - variables and/or less obscured stars was all produced by themselves . because the mass - loss and dust - loss rate is derived from the optical depth via scaling with the luminosity ,
correcting the mass return is not straightforward .
this correction can be applied statistically in two ways : ( 1 ) by reducing the contribution from each star by a factor @xmath83 where @xmath84 and @xmath85 are the distributions before and after correcting for the symmetric distribution around @xmath37 , where @xmath86 for @xmath87 ; or ( 2 ) by including all stars including those with @xmath87 , and assigning negative mass - loss rates to the latter .
the problem with the second method is that some stars with very negative values and/or very high luminosities could make a large difference to the end result , whereas in the first method the correction is `` smoother '' . on the other hand , the second method accounts , to some extent , for variations of the correction factor with luminosity
should such dependence exist . in what follows
we have adopted the second method .
hence we obtain the binned and cumulative mass - loss rate distributions depicted in figure 15 .
we concentrate on the mass - loss rates of the ukirt variable stars , as these are believed to contribute the majority of the mass loss , but we also show the result including the non - variable stars which are dominated by the low - mass red giants . the latter are numerous and therefore appear to contribute several times as much as the intermediate - mass agb stars and higher - mass red supergiants , but it must be remembered that interstellar dust will be more often the cause of their generally modest reddening and so that result must be regarded as a firm upper limit to the mass return , of @xmath88 m@xmath1 yr@xmath2 ( or @xmath89 m@xmath1 yr@xmath2 if all carbon stars are actually m - type stars )
. the true mass - return rate must be closer to that derived from the ukirt variable stars , @xmath90 m@xmath1 yr@xmath2 ( or @xmath91 m@xmath1 yr@xmath2 if all carbon stars are actually m - type stars ) . accepting that some of the mass loss might have been missed as not all variable stars may have been recognised in our survey we could conceive a total mass - return rate of @xmath40.01 m@xmath1 yr@xmath2 ( up to @xmath92 m@xmath1 yr@xmath2 if a large fraction of the carbon stars are actually m - type stars ) . given the assumptions and uncertainties inherent to the mass - loss determinations we could hardly claim a higher degree of accuracy .
that said , low - mass m - type agb stars appear to make a similar contribution to the mass return in the central regions of m33 as all more massive stars combined . among the latter ,
the massive carbon stars and red supergiants make similar contributions , while the low - mass carbon stars and massive agb stars contribute less . as a result
, the present - day mass - return arises mainly from stars formed in one of three major episodes : @xmath93 myr ago ( @xmath94 ) , @xmath951 gyr ago ( @xmath96 ) and @xmath9710 gyr ago ( @xmath98 ) .
this mainly reflects the star formation history ( paper ii ) .
the contribution from the intermediate - age population , formed @xmath951 gyr ago , is larger if the presumed carbon stars are actually mostly m - type stars , in which case they account for about half of the total mass return .
it is worthy of note that carbon grains make up @xmath99% of the present - day dust - mass return , i.e. the interstellar dust is predominantly oxygen - rich .
this fraction becomes even smaller if the population of carbon stars was overestimated .
the radial profile of the mass return rate ( fig .
16 , in terms of surface density deprojected onto the galaxy plane
see paper ii ) is flat within the inner @xmath100 kpc for the carbon stars .
there appears to be a clear radial gradient for the m - type stars , with the mass return greatest within the central few hundred pc , but large deviations occur around @xmath1010.4 kpc .
a more complete picture is obtained from the full 2-d map of mass return ( fig .
17 , _ not _ deprojected ) , where a smooth level of gas and dust injection underlies dramatic local enhancements due to the concentration of red supergiants ( and super - agb stars ) near to their sites of formation in discrete molecular cloud complexes .
one of these includes # 13093 and # 12208 , the supergiants with the highest mass - loss rates ( see section 3.1 ) . for comparison
we show a _ spitzer _
composite in the left panel of figure 17 note the bright source at ( @xmath102 , @xmath103 ) : this is m33 - 8 ( kahn et al .
2012 ; see section 3.1 ) .
the spatial distribution of mass return does not change markedly if the carbon stars turn out to be mostly m - type stars .
if we believe the model predictions for the duration of the pulsation phase ( @xmath104 , see table 2 in paper ii and marigo et al . 2008 ) then we can calculate the mass lost during the pulsation phase as a fraction of the birth mass if we sample this phase both randomly and sufficiently : @xmath105 obviously , this ratio should be less than unity .
it is therefore somewhat disconcerting to find that this is not the case ( fig .
18 ) , though it does lend support to our belief that the non - variable stars can not contribute much to the total mass - loss budget , or it would seriously aggravate the situation . the integrated mass - loss along the evolution of low- and intermediate - mass stars
is strongly constrained by the initial final mass relation determined from white dwarfs in stellar clusters .
the ratio of our values to those derived from the initial final mass relation from williams , bolte & koester ( 2009 ) varies between @xmath106 and @xmath107 . for massive stars
the situation is more uncertain ; while most of them ( at the lower mass end ) will leave behind a neutron star ( of order a tenth of their birth mass ) it is unclear how much mass is lost as a red supergiant , as a blue supergiant , and during the supernova explosion ( see , e.g. , ekstrm et al .
but what causes the discrepancy ? have the mass - loss rates been over - estimated ? or is the pulsation duration in error ?
if the pulsation duration was over - estimated , then the star formation rate would have been under - estimated . at the time
, we compared our estimate of the recent star formation rate , @xmath1080.007 m@xmath1 yr@xmath2 kpc@xmath3 , with work by wilson , scoville & rice ( 1991 ) who had derived @xmath109 m@xmath1 yr@xmath2 kpc@xmath3 .
we were satisfied that , within the inherent uncertainties , these two estimates were consistent . in light of the above discrepancy between birth mass and lost mass , though , we now revisit these estimates : * it is reasonable to assume that star formation occurs within the gas - rich disc , and so the recent star formation rate ought to be deprojected onto the galaxy s plane .
this depresses the rate by a factor @xmath110 where @xmath111 is the angle of inclination of the normal to the m33 disc with respect to the line - of - sight , for which we take @xmath112 ( zaritsky et al .
the fact that the survey area is square but the position angle is @xmath113 matters , but not a lot .
so our estimate for the recent star formation rate becomes @xmath114 m@xmath1 yr@xmath2 kpc@xmath3 .
* several other groups have estimated star formation rates for the central region of m33 ( as part of larger surveys ) , from h@xmath115 emission or the luminosity in the far - ultraviolet or far - infrared ( hippelein et al.2003 ; engargiola et al . 2003 ; heyer et al .
2004 ; gardan et al .
2007 ; boissier et al . 2007 ; verley et al .
kang et al . (
2012 ) have summarised these works and they determine @xmath116 m@xmath1 yr@xmath2 kpc@xmath3 within the central square kpc of m33 , but possibly as much as 0.030.04 m@xmath1 yr@xmath2 kpc@xmath3 ( heyer et al . 2004 ) .
this is @xmath117 times higher than our estimate .
if we shorten the pulsation duration ( as used in paper ii ) by this factor , then the fractional mass - loss of stars with masses @xmath81 m@xmath1 ( @xmath118 ) becomes @xmath119 .
this reconciles the estimated mass - loss rates and star formation rates , given that some additional mass will be lost by the most massive stars as blue supergiants and supernov .
now that we have regained some confidence in the estimated values of the star formation rate and mass - return rate , we can consider the balance between the ism depletion and replenishment .
kang et al .
( 2012 ) estimated an ism depletion timescale of @xmath120 gyr . but this does not take into account the continuous mass return by evolved stars , which would lengthen the depletion timescale . our estimated mass - return rate ( after deprojection ) of @xmath1210.005 m@xmath1 yr@xmath2 kpc@xmath3 ( up to 0.01 m@xmath1 yr@xmath2 kpc@xmath3
if all carbon stars are actually m - type stars ) is more than our initially estimated recent star formation rate of @xmath122 m@xmath1 yr@xmath2 kpc@xmath3 , but lower than the star formation rate estimated by other groups .
we have just argued that our estimate of the star formation rate needs to be scaled up by a factor @xmath117 , which would mean that the mass - return rate falls short of supplying the fuel for continued star formation by a factor 6 or 7 ( or 3 , if all carbon stars are actually m - type stars ) .
the naive estimate of the depletion timescale of 0.3 gyr would not change by more than 17% as a result of this mass return and by at most 50% if all carbon stars turn out to be m - type stars .
however , additional mass is returned by supernov , hot massive - star winds , luminous blue variable eruptions , et cetera .
most of these contributions come from massive stars , which contribute about one third to the mass returned through dusty stellar winds ( see fig .
15 , top right panel ) . their fractional mass - loss ( after correction for the shorter pulsation duration ) is @xmath0% , i.e. they could at most contribute @xmath123 times as much as estimated here ( accounting also for stellar remnants being left behind ) , which would increase the total mass - return rate ( across all stellar masses ) from @xmath1240.005 to @xmath4 m@xmath1 yr@xmath2 kpc@xmath3 .
note that the contribution from low - mass stars is already fully accounted for by the dusty stellar winds .
thus , the above conclusion does not change , namely that the rate at which the ism is replenished by stellar mass loss falls short by a factor @xmath125 to sustain star formation at the current rate .
this conclusion does not change if all carbon stars turn out to be m - type stars , though the discrepancy is reduced to a factor @xmath126 . for star formation to continue beyond
the next few hundred myr gas must flow into the central regions of m33 , either through a viscous disc or via cooling flows from the circum - galactic medium .
the mass return is not uniform across the central square kpc , so locally the above conclusions might not be accurate .
the mass return by low - mass stars and carbon stars is fairly uniform across the area , with a slight radial gradient .
but there are three areas at @xmath1010.4 kpc from the centre where the mass - return rate is much higher due to the contribution of several massive , very dusty stars
. this could feed and chemically enrich enduring or new star formation in those areas on timescales of a couple of @xmath127 yr . in paper
ii we had found that the star formation rate peaks within the central @xmath128 kpc .
the very centre of m33 , while characterised by high levels of mass return is not where the ism is replenished at the highest rate .
however , if gas from the dense regions at @xmath1010.4 kpc can be accreted into the nucleus of m33 then this could be a mechanism for star formation to occur even within the nuclear star cluster .
indeed , tosaki et al . ( 2011 ) find the molecular gas peaks at @xmath129 kpc from the centre , and this could have been supplied at somewhat greater distances before being channeled towards the nucleus .
they find another peak at @xmath130 kpc which is at the edge of our survey area .
the offsets of the peaks in the dense ism with respect to the peaks in mass return ( and star formation ) suggest that , globally , there is a delay as well as a migration between the epochs and sites of intense mass return , accumulation of dense ism , and star formation .
we thank the staff at ukirt for their excellent support of this programme .
jvl thanks the school of astronomy at ipm , tehran , for their hospitality during his visits .
we are grateful for financial support by the leverhulme trust under grant no .
rf/4/rfg/2007/0297 , and by the royal astronomical society .
finally , we thank the referee for her / his constructive report which prompted us to improve the manuscript .
benjamin r. a. , et al . , 2005 ,
apj , 630 , l149 boissier s. , et al . , 2007 , apjs , 173 , 524 bolatto a. d. , leroy a. k. , rosolowsky e. , walter f. , blitz l. , 2008 , apj , 686 , 948 bonanos a. z. et al . , 2006 ,
apj , 652 , 313 bonanos a. c. , et al . , 2010 , aj , 140 , 416 boothroyd a. i. , sackmann i .-
j . , 1992 , apj , 393 , l21 botticella m. t. et al . , 2009 , mnras , 398 , 1041 cioni m .- r .
l. , et al .
, 2008 , a&a , 487 , 131 deul e. r. , van der hulst j. m. , 1987 , a&as , 67 , 509 draine b. t. , lee h. m. , 1984 , apj , 285 , 89 drout m. , massey p. , meynet g. , 2012 , apj , 750 , 97 ekstrm s. , et al . , 2012 ,
a&a , 537a , 146 engargiola g. , plambeck r. l. , rosolowsky e. , blitz l. , 2003 , apjs , 149 , 343 gardan e. , braine j. , schuster k.f . , brouillet n. , sievers a. , 2007 , a&a , 473 , 91 gordon s. m. , duri n. , kirshner r. p. , goss w. m. , viallefond f. , 1999 , apjs , 120 , 247 gullieuszik m. , et al . , 2012 ,
a&a , 537a , 105 hanner m. s. , 1988 , nasa conf . pub . 3004 , 22 heyer m. h. , corbelli e. , schneider s. e. , young j. s. , 2004 , apj , 602 , 723 hippelein h. , haas m. , tuffs r. j. , lemke d. , stickel m. , klaas u. , volk h.j . , 2003 , a&a , 407 , 137 hodierna g. b. , 1654 , de systemate orbis cometici , deque admirandis coeli characteribus ( about the systematics of the cometary orbit , and about the admirable objects of the sky ) , palermo iben i. jr . ,
renzini a. , 1983 , ara&a , 21 , 271 ivezi , elitzur m. , 1997 , mnras , 287 , 799 javadi a. , van loon j. th .
, mirtorabi m. t. , 2011a , mnras , 411 , 263 ( paper i ) javadi a. , van loon j. th .
, mirtorabi m. t. , 2011b , mnras , 414 , 3394 ( paper ii ) javadi a. , van loon j. th . , mirtorabi m. t. , 2011c , in : why galaxies care about agb stars ii , eds . f. kerschbaum , t. lebzelter & r. f. wing , aspc , 445 , 497 jura m. , kleinmann s. g. , 1989 , apj , 341 , 359 jura m. , kleinmann s. g. , 1990 , apjs , 73 , 769 kang x. , chang r. , yin j. , hou j. , zhang f. , zhang y. , han z. , 2012 , mnras , 426 , 1455 khan r. , stanek k. z. , kochanek c. s. , 2012 , arxiv:1210.6980 marigo p. , girardi l. , bressan a. , groenewegen m. a. t. , silva l. , granato g. l. , 2008 , a&a , 482 , 883 marshall j. r. , van loon j. th . , matsuura m. , wood p. r. , zijlstra a. a. , whitelock p. a. , 2004 , mnras , 355 , 1348 mattsson l. , hfner s. , herwig f. , 2007 , a&a , 470 , 339 mcdonald i. , zijlstra a. a. , boyer m. l. , 2012 , mnras , * 427 , 343 * mcdonald i. , van loon j. th . , decin l. , boyer m. l. , dupree a. k. , evans a. , gehrz r. d. , woodward c. e. , 2009 , mnras , 394 , 831 mcquinn k. b. w. , et al . , 2007 ,
apj , 664 , 850 messier c. , 1771 , mem .
acad . , p448 nowotny w. , aringer b. , hfner s. , eriksson k. , 2013 , a&a , 552a , 20 olofsson h. , carlstrom u. , eriksson k. , gustafsson b. , willson l. a. , 1990 , a&a , 230 , l13 pgouri b. , 1988 , a&a , 194 , 335 srinivasan s. , sargent b. a. , meixner m. , 2011 , a&a , 532a , 54 srinivasan s. , et al . , 2009 , aj , 137 , 4810 stetson p. b. , 1987 , pasp , 99 , 191 thompson t. a. , prieto j. l. , stanek k. z. , kistler m. d. , beacom j. f. , kochanek c. s. , 2009 , apj , 705 , 1364 tosaki t. , kuno n. , ondera s. , et al .
2011 , pasj , 63 , 1171 van loon j. th .
, 2007 , in : why galaxies care about agb stars , eds .
f.kerschbaum , c. charbonnel & r. f. wing , aspc , 378 , 227 van loon j. th . , groenewegen m. a. t. , de koter a. , trams n. r. , waters l. b. f. m. , zijlstra a. a. , whitelock p. a. , loup c. , 1999 , a&a , 351 , 559 van loon j. th . , et al . , 2003 , mnras , 338 , 857 van loon j. th
, marshall j. r. , zijlstra a. a. , 2005b , a&a , 442 , 597 van loon j. th . ,
l. , zijlstra a. a. , loup c. , 2005a , a&a , 438 , 273 van loon j. th .
, marshall j. r. , cohen m. , matsuura m. , wood p. r. , yamamura i. , zijlstra a. a. , 2006 , a&a , 447 , 971 van loon j. th . , cohen m. , oliveira j. m. , matsuura m. , mcdonald i. , sloan g. c. , wood p. r. , zijlstra a. a. , 2008 , a&a , 487 , 1055 vassiliadis e. , wood p. r. , 1993 , apj , 413 , 641 verley s. , corbelli e. , giovanardi c. , hunt l. k. , 2009 , a&a , 493 , 453 whitelock p. a. , feast m. w. , pottasch s. r. , 1987 , in : late stages of stellar evolution , reidel ( dordrecht ) , p.269 whitelock p. , feast m. , catchpole r. , 1991 , mnras , 248 , 276 whitelock p. a. , feast m. w. , van loon j. th .
, zijlstra a. a. , 2003 , mnras , 342 , 86 whitelock p. a. , feast m. w. , marang f. , groenewegen m. a. t. , 2006 , mnras , 369 , 751 williams k. e. , bolte m. , koester d. , 2009 , apj , 693 , 355 wilson c. d. , scoville n. , rice w. , 1991 , aj , 101 , 1293 wood p. r. , 2000 , pasa , 17 , 18 zaritsky d. , elston r. , hill j. m. , 1989 , aj , 97 , 97 | we have conducted a near - infrared monitoring campaign at the uk infrared telescope ( ukirt ) , of the local group spiral galaxy m33 ( triangulum ) .
the main aim was to identify stars in the very final stage of their evolution , and for which the luminosity is more directly related to the birth mass than the more numerous less - evolved giant stars that continue to increase in luminosity . in this third paper of the series ,
we measure the dust production and rates of mass loss by the pulsating asymptotic giant branch ( agb ) stars and red supergiants . to this aim
, we combined our time - averaged near - ir photometry with the multi - epoch mid - ir photometry obtained with the _ spitzer _ space telescope .
the mass - loss rates are seen to increase with increasing strength of pulsation and with increasing bolometric luminosity .
low - mass stars lose most of their mass through stellar winds , but even super - agb stars and red supergiants lose @xmath0% of their mass via a dusty stellar wind .
more than three - quarters of the dust return is oxygenous .
we construct a 2-d map of the mass - return rate , showing a radial decline but also local enhancements due to agglomerations of massive stars .
we estimate a total mass - loss rate of 0.0040.005 m@xmath1 yr@xmath2 kpc@xmath3 , increasing to @xmath4 m@xmath1 yr@xmath2 kpc@xmath3 when accounting for eruptive mass loss ( e.g. , supernov ) ; comparing this to the current star formation rate of @xmath5 m@xmath1 yr@xmath2 kpc@xmath3 we conclude that star formation in the central region of m33 can only be sustained if gas is accreted from further out in the disc or from circum - galactic regions .
[ firstpage ] stars : agb and post - agb stars : carbon stars : mass - loss
supergiants galaxies : individual : m33 galaxies : structure |
correlation functions are important structural descriptors that arise in a variety of disciplines such as solid state physics @xcite , signal processing @xcite , computer vision@xcite , statistical physics@xcite , geostatistics@xcite , and materials science @xcite .
many techniques for structural characterization of materials over a wide range of length scales provide data in the form of correlation functions including , but not limited to , scattering methods @xcite .
other examples are absorption spectroscopy @xcite , energy transfer analysis @xcite , nuclear magnetic resonance @xcite , and also grey - scale image analysis @xcite . moreover , in the case of _ in situ _ studies with a nanometer resolution @xcite , correlation functions are often the only data available experimentally . despite the widespread use of correlation functions ,
the nature of the structural information they contain remains an open area of active research @xcite .
the central question of the present paper can be put as follows : how accurately is a microstructure known when the only data available is a two - point correlation function ?
we shall focus our analysis on two - phase microstructures and the two - point correlation function @xmath0 , which is defined to be the probability that two random points at a distance @xmath1 from one another both belong to the same phase @xcite .
two - point statistics are not sufficient to determine unambiguously a microstructure .
the specification of a given two - point function @xmath0 is equivalent to defining a macrostate of the system , the degeneracy of which can be expressed as a configurational entropy . in particular , all space transformations that preserve distances - translation , rigid rotation and inversion - lead to microstructures with identical two - point statistics . following previous work
, we will call such degeneracies _ trivial _ @xcite .
the focus of the present paper is on _ non - trivially _ degenerate microstructures , which can not be obtained from each other through any of the aforementioned trivial transformations .
examples of non - trivially degenerate microstructures are poisson polyhedra tesselations of three - dimensional space @xcite and debye random media @xcite , which although having distinct microstructures , have identical @xmath0 .
non - trivial degeneracy is not limited to infinite systems .
examples of finite point patterns having the same two - point statistics have been given as early as 1939 , and patterson coined the word `` homometric '' to qualify them @xcite .
very recently , general equations have been derived that can in principle be solved analytically to obtain homometric microstructures @xcite . in the context of crystallography ,
the degeneracy of the structural information contained in correlation functions is referred to as the _ phase problem_. the phase problem , however , is not universally applicable .
a spectacular counterexample is the so - called direct method of crystallography @xcite , for which hauptman and karle received the 1985 nobel prize for chemistry . in the field of computer vision
, it has been shown that finite textures are completely characterized by their orientation - dependent correlation functions @xcite .
many theoretical examples of microstructures with a low degeneracy can be accurately reconstructed from their correlation function alone @xcite .
all these examples have in common that they incorporate orientation information .
the focus of the present work is on radial correlation functions in which orientation information is averaged out .
this simplification is relevant to many experimental contexts , notably small - angle scattering @xcite , where the correlation function is generally rotationally averaged through the measurement of powder scattering patterns , as well as to isotropic disordered systems in general @xcite .
the understanding of the structural information in radial correlation functions has been considerably advanced through the use of reconstruction algorithms , which aim at producing microstructures with a specified correlation function via the minimization of a prescribed energy functional @xcite . in the case of a reconstruction based on two - point correlation functions ,
a natural choice for the energy functional is @xcite @xmath2 ^ 2 \ , \ ] ] where @xmath3 is the target two - point correlation function , @xmath0 is the correlation function of the microstructure , i.e. , the configuration being optimized , and the sum is over all measurable distances .
this definition of the energy is equivalent to a norm-2 error : it is non - negative and it vanishes only for those configurations that satisfy @xmath4 . in this context , the question of the degeneracy associated with a given correlation function is equivalent to determining the number of microstructures having zero energy , i.e. , the _ ground - state degeneracy _ of the energy functional @xcite .
the minimization of eq .
( [ eq : definition_e ] ) is generally done by discretizing the microstructure on a grid with periodic boundary conditions , and by using either a steepest descent @xcite or a simulated annealing @xcite algorithm . in the case of a two - phase microstructure , which can be thought of as an image with black and white pixels , the simulated annealing reconstruction proceeds as follows .
starting from any configuration , with value @xmath5 of the energy functional eq .
( [ eq : definition_e ] ) , a black pixel is chosen randomly and moved to any available white position .
the function @xmath0 is updated and the new energy @xmath6 is calculated .
the move is accepted with probability @xmath7 where @xmath8 is a temperature parameter @xcite .
all energy - decreasing moves are therefore accepted but some energy - increasing moves are accepted as well , depending on the chosen temperature .
the latter moves are necessary to ensure that the entire configuration space be explored in principle , and that the system is not trapped in a local minimum of @xmath9 .
simulated annealing algorithms consist in starting at a high temperature , and progressively decreasing the temperature until convergence is reached ( @xmath10 ) .
this type of approach has been widely used for microstructure reconstruction , in the context of both applications @xcite and theoretical investigations @xcite .
the latter include generalizations to other types of statistical microstructure descriptors besides @xmath0 @xcite , most notably to higher - order correlation functions @xcite as well as to cluster correlation functions @xcite . ) and reconstructed ( @xmath11 ) correlation functions are indistinguishable on the scale of the figure .
the size of the grid is @xmath12 pixels with @xmath13 .
these examples strongly suggest that the two - point function of a single sphere under periodic boundary conditions is only trivially degenerate through translation , but that the two - point degeneracy of a poisson point process has a large non - trivial contribution.,width=264 ] examples of reconstructions of two - phase microstructures under periodic boundary conditions are given in fig .
[ fig : rec_examples ] . in the case of the single disk ,
the reconstructed microstructure is almost identical to the target , except for a translation ( top portion of fig .
[ fig : rec_examples ] ) . in the case of the reconstruction of the hard disks ( middle portion of fig .
[ fig : rec_examples ] ) , the characteristic size of the disks as well as the average distance between them is recovered .
however , an exact reconstruction of the target configuration is not possible ; spurious objects are also formed through the partial merging of neighboring disks . in the case of a realization of a poisson point process
( randomly coloring a pixel black according to a prescribed volume fraction ) , the reconstructed and the target microstructures might look superficially similar because they both appear to be random distributions of black pixels ( bottom portion of fig .
[ fig : rec_examples ] ) .
however , the two microstructures have very little in common if one is interested in the exact configurations of the pixels , although an excellent match is obtained between @xmath0 and @xmath3 .
this illustrates the concept of non - trivial degeneracy @xcite .
in a recent letter @xcite , we presented a general theoretical framework for estimating quantitatively the structural degeneracy corresponding to any specified correlation function .
this was achieved by mapping the problem to the estimation of a ground - state degeneracy through the use of eq .
( [ eq : definition_e ] ) . here
we provide a more comprehensive presentation of the methodology and analyses , including a quantitative characterization of the energy landscape associated with the reconstruction as well as a detailed derivation of the degeneracy metric .
moreover , we show that our results can be expressed in terms of the _ information content _ of the two - point correlation functions .
although the present work focuses on two - dimensional media in euclidean space , our procedure can be applied in any space dimension and generalized to non - euclidean spaces ( e.g. , compact and hyperbolic spaces ) .
the remainder of the paper is organized as follows . in sec .
ii , we discuss the degeneracy of the two - point statistics for a variety of microstructures that are used as benchmarks throughout the rest of the paper .
we consider successively small systems - for which all the configurations can be enumerated - intermediate systems - for which the degeneracy can be determined via a monte carlo method we presented elsewhere @xcite - and large systems for which neither of the aforementioned two methods apply and one needs to use the reconstruction method . in sec .
iii , we devise an analytical method , based on a random walk in configuration space , to characterize the energy landscape associated with reconstruction .
in particular , we determine a characteristic energy profile for the basin of each ground state . in sec .
iv , we show that the ground - state degeneracy of reconstruction problems is related to the roughness of the energy landscape .
we introduce a roughness metric that can be calculated from @xmath3 alone , and we show definitively that it is correlated with the microstructure degeneracy . in sec .
v , the degeneracy is expressed in terms of the _ information content _ of @xmath3 , and a formula is proposed relating the roughness metric to this information content .
the practical usefulness of our results is discussed .
the present paper is restricted to two - phase digitized microstructures , which can be thought of as images with @xmath14 black pixels and @xmath15 white pixels .
however , our analysis can be easily generalized to multiphase microstructures .
we shall first consider the very small microstructures of fig .
[ fig : degenerate ] with @xmath16 .
they will be analyzed in some detail and will serve as a benchmark for analytical methods applicable to larger and more complex microstructures .
@xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 , with @xmath17 the total number of pixels in the grid ( see text ) .
systems c to e have a non - trivial contribution to their degeneracy.,width=283 ] for any finite microstructure it is always possible to refer to the pixels through a linear index , @xmath22 to @xmath17 , independently of the actual dimensionality .
a finite microstructure is therefore completely characterized by a @xmath17-dimensional vector , with components @xmath23 equal to @xmath24 when point @xmath25 is a black pixel , and @xmath26 otherwise .
the two - point correlation function @xmath0 of the black phase is defined as the probability that two random pixels at a distance @xmath1 from one another are both black @xcite .
this can be written formally as @xmath27 where @xmath28 takes the value @xmath24 if the distance between pixels @xmath25 and @xmath29 is @xmath1 , and @xmath26 otherwise .
the quantity @xmath30 is defined as @xmath31 . in eq .
( [ eq : definition_p ] ) the double sum counts the pairs of black pixels separated by a distance @xmath1 , and the pre - factor normalizes that count by the total number of pixel pairs at a distance @xmath1 from one another .
the periodic boundary conditions are incorporated in the definition of the operator @xmath28 .
we assume in the rest of the paper that the discretizing grid is uniform in the sense that @xmath30 is independent of @xmath29 .
the use of a discrete pixel grid is equivalent to a quantizer " problem @xcite , in which every point of the microstructure is quantized to the centroid of its closest pixel .
the distances @xmath1 between pairs of points are therefore approximated by distances that are compatible with the grid .
a square grid is used throughout the present paper . for finite - size systems ,
the quantization naturally introduces some grid - specific artifacts @xcite .
however , the quantization error decreases and becomes zero in the limit of infinitely large microstructures .
the two - point correlation functions of the microstructures of fig .
[ fig : degenerate ] are given in table [ tab1 ] under the form @xmath32 .
the quantity @xmath33 is equal to the number of pairs of points at distance @xmath1 from one another .
note that although configurations @xmath34 and @xmath35 are different , they have identical two - point characteristics .
the same applies to @xmath36 and @xmath37 , as well as to @xmath38 and @xmath39 .
a complete enumeration of all microstructures with @xmath40 shows that there is no other configuration with the same @xmath0 .
lccccccccc @xmath1 & 1 & @xmath41 & 2 & @xmath42 & @xmath43 & 3 & @xmath44 & @xmath45 & @xmath46 + @xmath47 & 0 & 0 & 0 & 0 & 0 & 4 & 0 & 0 & 2 @xmath48 & 0 & 1 & 0 & 4 & 0 & 0 & 0 & 0 & 1 @xmath49 & 2 & 1 & 0 & 2 & 1 & 0 & 0 & 0 & 0 @xmath50 & 0 & 1 & 2 & 0 & 1 & 0 & 2 & 0 & 0 @xmath51 & 1 & 1 & 0 & 2 & 1 & 0 & 1 & 0 & 0 configuration @xmath52 in fig .
[ fig : degenerate ] is uniquely defined by its two - point function , and therefore is only trivially degenerate . on grids with @xmath17 points
the total number of translations is @xmath17 ; the number of rotations is @xmath24 , @xmath53 or @xmath54 , depending on the rotational symmetry of the configuration ; and the number of mirror configurations is @xmath24 or @xmath53 , depending on its chirality . due to the symmetry and chirality of configuration a
, only translation contributes to its degeneracy , which is therefore @xmath55 . in the case of configuration @xmath56
, the two possible orientations contribute an extra factor 2 , i.e. @xmath57 . configurations @xmath34 and @xmath35 are the `` kite & trapezoid '' examples discussed in refs .
@xcite , which are non - trivially degenerate . in this case ,
@xmath58 , where the factor 2 is the non - trivial contribution , and the factor 4 accounts for the possible orientations .
configurations @xmath36 and @xmath37 are also non - trivially degenerate .
configuration @xmath37 is , however , chiral so it has to be counted twice .
this leads to @xmath59 .
finally , non - trivially degenerate configurations @xmath38 and @xmath39 are both chiral .
this leads to @xmath60 .
the complete enumeration of degenerate microstructures is intractable for systems even barely larger than those represented in fig . [
fig : degenerate ] . in the present section ,
we discuss a monte carlo ( mc ) algorithm for estimating @xmath61 , which we introduced previously @xcite .
it can be applied to larger systems .
the approach is based on a general mc algorithm for estimating the density of states ( dos ) developed by wang and landau @xcite and further improved and analyzed by others @xcite .
the algorithm has been applied to a host of problems in condensed matter physics @xcite , in biophysics @xcite , and in logic @xcite .
the dos @xmath62 is defined as the number of states having energy equal to @xmath9 .
the logarithm of @xmath62 is equal to the entropy calculated in the microcanonical ensemble associated with eq .
( [ eq : definition_e ] ) .
the ground - state degeneracy @xmath61 is the value taken by @xmath62 for @xmath63 .
a canonical monte carlo simulation with transition probability given by eq . ( [ eq : metropolis ] ) leads the system to visit any energy with a probability @xmath64 @xcite .
the algorithm of wang and landau is based on the observation that a transition probability of the form @xmath65 would lead the system to visit all energies with equal probability .
the density of states @xmath62 is , however , unknown so the algorithm is iterative .
the starting value is set to @xmath66 for all energies , and the system is let evolve according to eq .
( [ eq : wang_landau ] ) , while updating an histogram @xmath67 .
each time an energy is visited the corresponding bin is updated , @xmath68 , and the estimated density of states is updated according to @xmath69 where @xmath70 is a numerical factor larger than 1 .
the evolution continues according to eq .
( [ eq : wang_landau ] ) with the updated value of @xmath62 .
the evolution is stopped when a flat histogram is obtained . at this point , the histogram @xmath67 is reset to @xmath26 , @xmath70 is reduced to a value closer to 1 , and the evolution starts over again .
the entire procedure is repeated until @xmath70 becomes lower than a prescribed accuracy .
algorithmic details are provided in the supplementary material @xcite .
the accuracy of the mc algorithm was tested by applying it to the microstructures of fig .
[ fig : degenerate ] .
the results are plotted in fig .
[ fig : dos_4points ] in the form of cumulative dos @xmath71 the mc algorithm provides @xmath62 only to within an unknown multiplicative constant , which is determined by imposing @xmath72 to be equal to the total number of configurations @xmath73 .
the latter is equal to the number of different ways in which @xmath14 black pixels can be chosen among a total of @xmath17 possible pixels , i.e. @xmath74 the cumulative dos plotted in fig .
[ fig : dos_4points ] satisfies @xmath75 and @xmath76 . ) , c ( @xmath77 ) , and e ( @xmath78 ) of fig .
[ fig : degenerate].,width=264 ] three independent mc estimations have been calculated for each microstructure in fig .
[ fig : degenerate ] , yielding three independent estimates of @xmath61 .
the results are : @xmath79 for configuration a compared to the exact value @xmath80 ; @xmath81 for configuration b compared to @xmath82 ; @xmath83 for configuration c compared to @xmath84 ; @xmath85 for configuration d compared to @xmath86 ; and @xmath87 for configuration @xmath9 compared to @xmath88
. the exact values are those calculated in sec .
[ sec : small ] with @xmath89 .
the agreement with the mc estimates is excellent .
figure [ fig : dos_13points ] shows mc estimates of the density of states for larger microstructures , with @xmath90 on a @xmath91 grid .
the microstructures are qualitatively similar to those of fig .
[ fig : rec_examples ] , namely a single disk , hard disks , and a poisson point process , all under periodic boundary conditions . in the case of a single disk ,
the mc estimation provides the value @xmath92 , corresponding to the @xmath80 possible translations .
this confirms that the disk is only trivially degenerate .
by contrast , the value found for the poisson point process is @xmath93 , which is orders of magnitude larger than any possible trivial contribution from translation and rotation . in the case of the hard disks , we find @xmath94 . in the case of the disk ( @xmath95 ) , hard disks ( @xmath77 ) , and the poisson point process(@xmath78).,width=264 ] the mc algorithm does not converge for systems larger than about @xmath96 pixels . with larger systems
the criterion for flat histograms is rarely reached , even with as many as @xmath97 mc steps . moreover , when flat histograms are indeed obtained , the estimated value of @xmath61 is much smaller than 1 , which shows that the algorithm explores only a small fraction of the complete configuration space .
these numerical difficulties are consistent with previous observations that flat - histogram algorithms have a convergence time that increases exponentially with system size @xcite .
it is therefore difficult to estimate the @xmath98-degeneracy of systems as large as the one shown in fig .
[ fig : rec_examples ] , except in the particular case where the microstructure is only trivially degenerate .
it has to be stressed that reconstructing exactly a degenerate microstructure is very unlikely .
therefore , whenever a reconstruction leads to a translated , rotated , and inverterted version of the target , this can be considered as very strong evidence that the microstructure is only trivially degenerate . in the remainder of the paper , we shall refer to a microstructure as being _ non - degenerate _
, whenever it has only a trivial degeneracy . in continuous euclidean space under periodic boundary conditions , an example of non - degenerate microstructure
is provided by the single sphere ( composed of a large number of pixels ) .
this results from the observation that @xmath99 is equal to volume fraction of the solid phase and that the negative slope of @xmath0 for @xmath100 is proportional to its surface area @xcite .
a sphere is non - degenerate because it is the microstructure that realizes the lowest possible surface area for a given volume fraction : the two - point correlation function of any microstructure other than a single sphere would have a larger slope at the origin , which would result in a positive energy according to eq .
( [ eq : definition_e ] ) . this observation can be expressed in a way that generalizes to discrete microstructures : for a given number of black pixels @xmath14 , a single sphere is non - degenerate because it is the microstructure that realizes the largest value of @xmath101 , where @xmath102 is a very small distance .
similarly , any configuration with @xmath90 other than the disk of fig .
[ fig : dos_13points ] has a smaller value of @xmath103 , where it is to be recalled that @xmath33 is the number of pairs of points with distance @xmath1 .
the same applies to configuration @xmath52 of fig .
[ fig : degenerate ] , which is not a disk : that particular microstructure is non - degenerate because any other configuration with @xmath40 has a smaller value of @xmath104 .
the origin of the degeneracy of hard - disk systems is touched on in sec .
vi . ) and reconstructed ( @xmath11 ) correlation functions are indistinguishable on the scale of the figure .
the size of the grid is @xmath105 pixels under periodic boundary conditions , and @xmath106.,width=264 ] the analysis of non - degeneracy in terms of extremal values of @xmath33 leads to non - intuitive results . when microstructures are defined on a grid , distances and orientations
are not independent : for instance , a pair of points at a distance @xmath43 from one another is necessarily oriented at 45 with respect to both axes .
a very anisotropic microstructure such as the crystal on the top of fig .
[ fig : rec_crystal ] minimizes @xmath33 for a set of distances corresponding to orientations orthogonal to the stripes .
the figure clearly shows that @xmath0 vanishes for a set of well - defined distances .
it should therefore not be surprising that such a highly anisotropic microstructure is non - degenerate .
the non - degeneracy of the crystal is confirmed by the fact that the reconstructed microstructure in fig .
[ fig : rec_crystal ] is a translated and rotated copy of the target . the vertical discontinuity in the middle of the reconstruction results simply from the target not having the same periodicity as the box . when a large crystal in a periodic box is split into a collection of randomly oriented smaller crystallites ( fig .
[ fig : rec_crystal ] middle and bottom rows ) , its anisotropy is reduced and there are no longer values of @xmath1 at which @xmath33 is extremal .
accordingly the reconstruction becomes less accurate , which means that the microstructure becomes more degenerate .
a more quantitative analysis of this issue is provided in sec . v.
the complete configuration space @xmath107 of two - phase microstructures with @xmath17 pixels is the set of vertices of an @xmath17-dimensional hypercube @xcite .
this results from the properties of the indicator vector , @xmath23 , which can take only values @xmath26 and @xmath24 . moving along a given @xmath17-dimensional direction ( along an edge of the hypercube ) is equivalent to interchanging a white ( black ) with a black ( white ) pixel . in the example of fig .
[ fig : hypercube ] , any movement along the fourth dimension ( joining the outer and inner cubes ) corresponds to changing the upper - left pixel .
of a two - phase microstructure is an @xmath17-dimensional hypercube on which hamming distance can be defined .
any move along a @xmath17-dimensional direction corresponds to changing the color of a particular pixel . in the case of a @xmath108 microstructure
the configuration space is a tesseract , with the fourth dimension represented as the edges joining the outer and inner cubes ( corresponding to the upper - left pixel).,title="fig:",width=170 ] + in the situation relevant to reconstruction , not all the vertices of the hypercube are accessible because the number of black pixels is kept constant , i.e. @xmath109 which means that all realizable microstructures lie on the intersection of the hypercube with a hyperplane .
once a target correlation function @xmath3 is specified , each vertex is assigned an energy through eq .
( [ eq : definition_e ] ) .
what we refer to as the energy landscape is the set of values taken by the energy functional @xmath9 on the vertices of the @xmath17-dimensional hyperplane .
a reconstruction consists in exploring the energy landscape until a vertex is found with @xmath110 .
the dos @xmath62 determined in section [ sec : mc ] is the number of vertices having a given energy @xmath9 .
the problem we address in this section is that of the spatial variability of @xmath9 in configuration space @xmath107 .
this analysis is motivated by the observation , in many fields of physics , that systems with large ground - state degeneracies generally have a rough energy landscape @xcite . if we can characterize the roughness of the energy landscape in terms of @xmath3 we can estimate the ground - state degeneracy @xmath61 . in order to characterize the spatial variability of @xmath9 in configuration space ,
it is necessary to define a distance .
a natural choice is the _ hamming _ distance , which counts the number of edges between any two vertices .
the hamming distance within the hyperplane defined by eq .
( [ eq : hyperplane ] ) takes only even values .
the distance @xmath111 $ ] between two microstructures @xmath52 and @xmath56 is therefore defined as half the hamming distance @xmath112=\frac{1}{2 } \sum_{i=1}^n \left| i_a(i)-i_b(i ) \right| \ , \ ] ] where @xmath113 and @xmath114 are the indicator vectors . in real space , this distance @xmath115 is the smallest number of monte - carlo - like black pixel displacements that are required to pass from @xmath52 to @xmath56 .
the largest possible distance is @xmath116 when the two microstructures have no pixel in common .
the energy landscape can be characterized analytically through a random walk in configuration space .
this is illustrated in fig .
[ fig : random_walk ] . starting from a ground state of a reconstruction problem , with @xmath117 ,
the system moves randomly to any configuration at hamming distance @xmath118 from the current configuration .
this is equivalent to a standard metropolis monte carlo with @xmath119 [ see eq .
( [ eq : metropolis ] ) ] .
when the number of moves @xmath120 increases , the random walk explores the configuration space @xmath107 over increasingly large hamming distances @xmath115 from the starting ground state .
+ the rate at which the average energy @xmath121 visited by the random walk increases with the number @xmath120 of moves characterizes the energy landscape of a given reconstruction problem . in the examples of fig .
[ fig : random_walk ] , the energy curve of the poisson point process is steeper than that of the single disk , which suggests smaller basins .
we now proceed to analytically calculate the values of @xmath121 as a function of the characteristics of the starting ground state .
the only _ a priori _ information about the ground states of a given reconstruction problem is their one - point and two - point characteristics : @xmath122 and @xmath3 .
the other characteristics , in particular the higher - order correlation functions , may differ significantly from one ground state to another .
let us assume for now that the starting ground state of the random walk is perfectly known through its indicator vector @xmath23 .
instead of using @xmath0 , it is convenient to use the equivalent autocovariance @xmath123 defined as @xmath124 where @xmath125 is the fraction of black pixels .
the average energy after @xmath120 random moves can be written in terms of @xmath123 as @xmath126 where we have used the notations @xmath127 , which is associated with the ground state , and @xmath128 for any average value at step @xmath120 . at each step of the random
walk there are @xmath129 possible moves , so that the total number of possible walks of length @xmath120 is @xmath130 : the averages @xmath131 are calculated over all these possible paths .
we now calculate successively @xmath132 and @xmath133 , which are required to calculate and @xmath121 through eq .
( [ eq : e_n_average_general ] ) .
when a black pixel @xmath134 is moved to position @xmath135 previously occupied by a white pixel , the indicator vector becomes @xmath136 where @xmath137 is the kronecker delta function . using the definition of @xmath0 , eq .
( [ eq : definition_p ] ) , the autocovariance is then found to become @xmath138 \big\}\end{aligned}\ ] ] for that particular move .
the average value of @xmath139 is then simply calculated as @xmath140 where the factor @xmath141 accounts for the fact that @xmath134 and @xmath135 are uniformly distributed over the black and white phases , respectively .
combining eqs .
( [ eq : p_prime ] ) and ( [ eq : p_average ] ) , the average autocovariance @xmath123 after a single random move is found to be @xmath142 with @xmath143 . in eq .
( [ eq : evolution_chi ] ) a term of the order of @xmath144 has been neglected , which is justified for large values of @xmath17 .
the complete equation can be found in the supplementary material @xcite .
equation ( [ eq : evolution_chi ] ) is valid only for @xmath145 .
the value for @xmath146 depends only on the fraction of black pixels @xmath147 , @xmath148 , and it is therefore a constant during the random walk . because each random move is independent from the previous one , the analysis leading to eq .
( [ eq : evolution_chi ] ) can be repeated in an iterative way .
taking into account that the starting two - point function is @xmath149 , this leads to the following simple result @xmath150 in the course of the random walk , the average two - point function therefore converges towards that of a poisson point process @xcite with @xmath151 for all @xmath152 .
the convergence is exponential in the number of random moves and it occurs in about @xmath153 moves .
the determination of @xmath154 proceeds along the same lines , but it is more involved ; the details can be found in the supplementary material @xcite . when the expression obtained for @xmath155
is introduced in eq .
( [ eq : e_n_average_general ] ) , the value of the average energy takes eventually the form @xmath156 where @xmath38 , @xmath39 , @xmath157 and @xmath158 are constants that characterize the starting ground state . the constants @xmath159 and @xmath160 depend only on @xmath17 and @xmath14 @xmath161 and @xmath162 has the same meaning as in eq .
( [ eq : evolution_chi ] ) . the constants @xmath38 and @xmath158 depend only on two - point characteristics of the ground states .
they are written as @xmath163 and @xmath164 where the sum is over all the distances that are used in the definition of the energy . as a consequence of eq .
( [ eq : e_n_average ] ) , @xmath165 is the value towards which @xmath121 converges for large values of @xmath120 .
since the random walk is ergodic , any configuration has the same probability of being visited in the long run .
therefore , @xmath165 is the average energy calculated over the entire configuration space , which we refer to hereafter simply as @xmath166 .
the main contribution to @xmath166 is @xmath167 , which is small for disordered microstructures .
this term vanishes in the extreme case of a poisson point process for which the only contribution left is of order @xmath168 , according to eq .
( [ eq : e_infty ] ) .
the shifting of the average energy towards lower values with increasing disorder is clear in figs .
[ fig : dos_13points ] and [ fig : random_walk ] .
the other two constants , @xmath39 and @xmath157 , in the expression of @xmath121 depend on more than the two - point function of the ground state .
they are given by @xmath169 and @xmath170 where the functions @xmath171 and @xmath172 are defined as @xmath173 and @xmath174 we postpone to sec .
[ sec : roughness_metric ] the detailed discussion of the structural meaning of @xmath171 and @xmath175 .
we should only mention here that @xmath171 can be expressed in terms of @xmath0 and that it is therefore common to all ground states ( see the supplementary material @xcite ) .
by contrast , @xmath175 depends on three - point statistics and may differ from one ground state to another . the black lines in fig .
[ fig : random_walk ] have been obtained from eq .
( [ eq : e_n_average ] ) with the constants @xmath38 to @xmath165 evaluated at the starting ground state .
the analysis captures the essential features of the random walk , in particular the steepness of the @xmath121 _ versus _ @xmath120 curves .
an important quantity for the rest of the analysis is the average energy of all configurations at hamming distance @xmath176 from the ground state . setting @xmath177 in eq .
( [ eq : e_n_average ] ) and neglecting terms of the order of @xmath178 , leads to @xmath179 the first two contributions are global characteristics of the configuration space , which depend only on @xmath0 and are therefore common to all ground states .
by contrast , the contribution from @xmath175 may _ a priori _ differ significantly from one ground state to another .
we discuss this point in detail in sec .
[ sec : roughness_metric ] .
the average energies @xmath121 visited after exactly @xmath120 random moves is strictly a property of the random walk , not only of the energy landscape .
the aim of the present section is to use eq .
( [ eq : e_n_average ] ) to calculate the average energy of all microstructures at a given hamming distance @xmath115 from a given ground state . a random walk of length @xmath120 can reach any microstructure at hamming distance @xmath180 from the ground state .
let us call a @xmath115-state a microstructure at distance @xmath115 from the ground state , and @xmath181 the fraction of all the random walks of length @xmath120 that end on a @xmath115-state .
the average energy @xmath182 of all @xmath115-states is related to @xmath183 through @xmath184 if the values of @xmath181 were known , this relation could in principle be inverted to estimate @xmath182 starting from eq .
( [ eq : e_n_average ] ) . for increasing number @xmath120 of steps in the random walk .
the figure is for @xmath185 and @xmath186 , relevant to fig .
[ fig : random_walk].,width=264 ] the distribution @xmath181 can be obtained by noticing that the random walk is a markov process , and by calculating the transition probabilities between various @xmath115-states . in real space ,
the hamming distance is the minimum number of pixel displacements that is necessary to make the state identical to the ground state .
therefore , in a @xmath25-state , the black phase is identical to the ground state but for @xmath25 _ holes _ , and the white phase is identical to the ground state but for @xmath25 _ extra pixels_. starting from an @xmath25-state at step @xmath120 , there are @xmath187 ways to reach a @xmath188-state at step @xmath189 .
these correspond to the number of different ways to take one of the @xmath25 extra pixels and place it into one of the @xmath25 holes .
the transition probability is therefore @xmath190 where the denominator is the total number of possible moves .
there are @xmath191 different ways to reach another @xmath25-state .
these correspond to the @xmath192 different ways of moving the holes within the black phase , plus the @xmath193 different ways of moving the extra pixels within the white phase .
the transition probability is therefore @xmath194 finally , there are @xmath195 different ways of reaching a @xmath196-state , which correspond to the different ways of taking a black pixel and putting it in the white phase .
this leads to @xmath197 the three transition probabilities add up to @xmath24 , which proves that all possibilities have been considered .
the enumerated probabilities define a tridiagonal transition matrix @xmath198 of size @xmath199 , with elements @xmath200 .
the explicit form of @xmath201 is given in the supplementary material @xcite .
writing the values @xmath181 in a vector as @xmath202^t$ ] enables us to write @xmath203 .
the general solution is therefore @xmath204 where @xmath205^t$ ] is the trivial distribution of hamming distances in the ground state . the particular evolution of @xmath181 for @xmath185 and @xmath206 is shown in fig .
[ fig : rw_distance ] .
these values are relevant to fig .
[ fig : random_walk ] . for small values of @xmath120
, the distribution is centered on the value @xmath207 . for large values of @xmath120
, however , @xmath181 converges towards an equilibrium distribution .
it is useful to note that although all states are accessible to the random walk after @xmath208 moves , the energy @xmath121 continues to changes for larger values of @xmath120 ( see fig .
[ fig : random_walk ] ) because the convergence of @xmath181 is asymptotic .
+ using the known values of @xmath181 and of @xmath183 , eq .
( [ eq : e_n_to_d ] ) can be inverted for @xmath209 , yielding the values of @xmath210 .
the procedure is illustrated in fig .
[ fig : profiles_n4 ] for the small - system - size configurations of fig .
[ fig : degenerate ] .
the results are plotted in the normalized form @xmath211 , which we refer to as the energy profiles .
the energy profiles describe quantitatively the average energy landscape surrounding any particular ground state .
they are all initially increasing curves that start from @xmath26 and reach values close to @xmath24 .
the average energy for @xmath118 is equal to @xmath212 , as calculated from eq .
( [ eq : e1 ] ) . for large hamming distances
the energy decreases again because microstructures with large @xmath115 can be thought of as negative imprints of the ground state : for @xmath116 the points that were occupied by black pixels in the ground state are all occupied by white pixels .
figure [ fig : profiles_n13 ] shows the energy profiles of the single disk , the hard disks , and the poisson point process of fig .
[ fig : dos_13points ] . when plotted on logarithmic scales ( insets ) ,
the profiles are seen to satisfy a power law of the type @xmath213 for small values of @xmath115 .
when the resolution is increased - i.e. increasing @xmath17 from @xmath214 to @xmath215 and @xmath216 while keeping @xmath217 constant - the profiles are shifted vertically ( insets ) but the exponent of the power law persists . in the case of the reconstruction of the single disk ,
the exponent @xmath218 is close to @xmath53 ( @xmath219 ) , and for the reconstruction of the poisson point process the exponent is close to @xmath24 ( @xmath220 ) .
the energy profile of the hard disks is not a pure power law : the exponent is that of a single disk for large hamming distances and that of a poisson point process for shorter distances . + the different exponents @xmath218 for the poisson point process and for the single disk hint at a qualitative difference that can be understood with the _ hole _ and _ extra pixel _ interpretation of hamming distance @xmath115 . in the case of the poisson point process
, the energy is proportional to @xmath115 .
this means that any hole added to the ground state contributes additively to the energy , which points to the absence of effective pixel - pixel interaction energy .
by contrast , the quadratic behavior for the single disk points to a collective contribution of the pixels to the overall energy , which can be considered as the signature of a structure .
when comparing the energy profiles in figs .
[ fig : profiles_n4 ] and [ fig : profiles_n13 ] with the densities of states in figs .
[ fig : dos_4points ] and [ fig : dos_13points ] , a striking correlation appears between the sizes of the basins and the ground - state degeneracy @xmath61 .
we observe that the smaller the basin , the more degenerate the reconstruction .
this observation is consistent with the one that large ground - state degeneracies are generally associated with rough energy landscapes @xcite .
however , a major difference between the energy profiles and the ground - state degeneracy is that the latter is a global characteristic of the reconstruction problem but the former are specific to given ground states .
for example , in the case of configuration @xmath221 of fig .
[ fig : degenerate ] , the ground states @xmath34 ( the kite ) and @xmath35 ( the trapezoid ) have slightly different energy profiles ( fig .
[ fig : profiles_n4 ] ) .
the main purpose of the present section is to provide an approximation for the energy profile , common to all ground states , which can be calculated from @xmath3 alone . to understand how the energy profiles depend on the particular ground state , it is necessary to analyze the structural meaning of functions @xmath171 and @xmath175 defined by eqs .
( [ eq : definition_sigma ] ) and ( [ eq : definition_sigma_c ] ) .
we show in the supplementary material @xcite that @xmath171 can be expressed in terms of the two - point function @xmath123 as @xmath222 where @xmath223 is a characteristic of the grid and of the boundary conditions . by definition ,
all ground states have identical two - point statistics .
the contribution of @xmath171 to the energy profile is therefore common to all ground states .
by contrast , it results from eq .
( [ eq : definition_sigma_c ] ) that @xmath175 is a sum of terms of the type @xmath224 which incorporate three - point statistics .
accordingly , the contribution of @xmath175 to the energy profile may differ significantly from one ground state to another . as a consequence of eq .
( [ eq : triangle ] ) , the pixel configurations that contribute to @xmath175 are isosceles triangles with apex @xmath225 and side - length @xmath1 . in the case of configuration @xmath221 of fig .
[ fig : degenerate ] , the kite ( @xmath34 ) has two such triangles , one with @xmath226 and the other with @xmath227 . on the contrary
there is no isosceles triangle in the trapezoid ( @xmath35 ) .
it therefore results from eq .
( [ eq : e1 ] ) that @xmath228 is larger for @xmath34 than for @xmath35 , in agreement with fig .
[ fig : profiles_n4 ] .
the energy profiles of @xmath36 and @xmath37 are identical although these two ground states are different .
this originates in the fact that in both ground states , there is an isosceles triangle of side @xmath229 and another one of side @xmath230 .
the same explanation applies to @xmath38 and @xmath39 . in both ground states
there is a single isosceles triangle with @xmath227 . despite these differences between @xmath171 and @xmath175 ,
the two functions have a strong similarity which can be put in evidence by the following probabilistic interpretation .
consider the set of all pixels at distance @xmath1 from a given pixel @xmath225 , which we refer to as the ring centered on @xmath225 .
the fraction of black pixels in the ring can be written as @xmath231 when @xmath225 is chosen randomly among all pixels ( black and white ) @xmath232 is a random variable having average value @xmath147 .
equation ( [ eq : definition_sigma ] ) shows that @xmath171 is the variance of @xmath232 .
the function @xmath171 can therefore be thought of as a generalized coarseness @xcite . from this probabilistic perspective , the meaning of @xmath175 is equivalent to @xmath171 , only the central pixel @xmath225 is not distributed randomly over the entire space but only over the black pixels . in this case
, the average of @xmath232 is the conditional probability that a pixel of the ring is black , given that the central pixel is black too , i.e. @xmath233 . from eq .
( [ eq : definition_sigma_c ] ) , one sees that @xmath175 is the variance of @xmath232 when the central pixel @xmath225 is randomly distributed over the black phase .
the function @xmath175 can therefore be thought of as a conditional coarseness . for small radii @xmath1 , the values taken by @xmath23 on the ring
are highly correlated with the value in the center , which implies @xmath234 . for large radii ,
the values on the ring and in the center are not correlated at all and therefore @xmath235 .
an example of @xmath171 and @xmath172 , calculated on the realization of a hard - disk system is given in fig .
[ fig : sigma2 ] .
( top ) and @xmath175 ( bottom ) of a hard - disk microstructure .
the inset of the bottom graph shows @xmath175 and its upper bound @xmath236 .
the dashed red line is the approximation @xmath237 obtained through eq .
( [ eq : sigma2c_approximation]).,title="fig:",width=264 ] + the similarity of the probabilistic interpretations of @xmath175 and of @xmath171 , and their strict mathematical equality for large values of @xmath1 , suggest that it should be possible to find an approximation of @xmath175 in terms of two - point functions only .
this would enable us to estimate a single approximate energy profile that would depend only on @xmath3 .
that profile would therefore be common to all ground states of a given reconstruction problem . to find such an approximation
, we observe that the terms between braces in the definitions of @xmath238 and @xmath171 are identical and that they are all positive ( see eqs ( [ eq : definition_sigma ] ) and ( [ eq : definition_sigma_c ] ) ) . however , there are fewer terms in eq .
( [ eq : definition_sigma_c ] ) because @xmath239 can be equal to @xmath26 .
one has therefore @xmath240 which leads to the following upper bound for @xmath175 : @xmath241 which depends only on two - point statistics .
the inset of fig .
[ fig : sigma2 ] compares @xmath175 to @xmath236 in the particular case of a hard - disk microstructure .
the upper bound @xmath236 is a good approximation of @xmath175 only for very small @xmath1 .
we therefore propose the following approximation for @xmath175 @xmath242 which is practically equal to @xmath171 for large @xmath1 , as it should .
[ fig : sigma2 ] shows that @xmath237 is a fair approximation of @xmath175 at all radii .
using @xmath237 in place of @xmath175 enables us to calculate a single energy profile , based on @xmath3 alone .
the red curves in figs .
[ fig : profiles_n4 ] and [ fig : profiles_n13 ] have been calculated in that way : @xmath171 was calculated rigorously from the target @xmath3 through eq .
( [ eq : sigma2_chi ] ) , and @xmath172 was approximated by eq .
( [ eq : sigma2c_approximation ] ) . in the case of the larger microstructure shown in the insets of fig .
[ fig : profiles_n13 ] , the approximate profiles are indistinguishable from the exact profiles on the scale of the figure . and roughness of the energy landscape @xmath243 calculated from @xmath0 alone .
the various microstructures are : disks of different sizes ( @xmath244 ) , realizations of poisson point processes of various densities ( @xmath77 ) , hard - disk microstructures ( @xmath245 ) , the configurations of fig .
[ fig : degenerate ] with @xmath16 ( @xmath95 ) , as well as a configuration with @xmath246 ( @xmath247 ) .
the black line is a guide to the eye ; the insets are sketches of archetypical energy landscapes for large and small values of @xmath243.,title="fig:",width=264 ] + using eq .
( [ eq : sigma2c_approximation ] ) it is possible to estimate a single metric to characterize globally the roughness of the energy landscape .
we propose the ratio @xmath248 where the tilde highlights the fact that @xmath249 is estimated through the approximation @xmath237 .
the quantity @xmath249 is the average energy of all microstructures at distance @xmath250 from the ground state .
because the ground states have zero energy , @xmath249 can be thought of as a laplacian in configuration space @xmath107 .
the ratio @xmath251 is therefore a dimensionless measure of the total curvature of the energy surface in the vicinity of any ground state .
it has to be stressed that @xmath252 is calculated from @xmath0 alone , and that it is therefore not specific to any particular ground state .
figure [ fig : roughness ] shows the quantitative relation between @xmath253 and the normalized ground - state degeneracy @xmath254 for a variety of microstructures defined on a @xmath91 grid .
the microstructures used for the figure are available in the supplementary material @xcite : they comprise both non - degenerate disk - like objects and highly degenerate realizations of poisson point processes .
the ground - state degeneracy of the latter was estimated via the mc algorithm .
the quantity @xmath254 is found to be highly correlated with @xmath255 over more than 14 orders of magnitude .
when passing from small to large values of @xmath243 , the energy landscape changes qualitatively in the way suggested by the insets of fig .
[ fig : roughness ] . for low values of @xmath243 ,
the energy landscape has an overall funnel structure , with low energy barriers , which makes it well suited for optimization problems .
by contrast , for large values of @xmath243 , the landscape is very rough with a large ground - state degeneracy .
it is , however , interesting to note that the rightmost point in fig .
[ fig : roughness ] is obtained for a system with @xmath256 having thus only a trivial degeneracy .
the corresponding energy landscape is extremely rough because any possible energy can be found at a hamming distance as short as @xmath176 from the ground state , but the total number of configurations @xmath73 is also extremely small .
the data referred to as disks in fig .
[ fig : roughness ] is a collection of non - degenerate microstructures with increasing values of @xmath14 .
when increasing @xmath14 , the roughness @xmath243 decreases but the degeneracy remains equal to its trivial translation contribution @xmath55 .
it is noteworthy that the values of @xmath254 of these non - degenerate microstructures span the same curve as the realizations of poisson point processes , for which @xmath61 has a huge non - trivial contribution .
the microstructures considered in fig .
[ fig : roughness ] were limited to @xmath91 grids , which size limit is imposed by the convergence of the mc algorithm . however , the fact that the @xmath254-_versus_-@xmath243 curve does not discriminate trivial from non - trivial degeneracy should not be limited to small microstructures .
this assumption enables us to extend the curve to larger microstructures by using disks of increasing sizes @xmath14 , on grids with increasing sizes @xmath17 , for which the degeneracy is known to be exactly @xmath55 .
this was done in fig .
[ fig : information ] .
the degeneracy is plotted in the form of @xmath257 for reasons that will be explained in next section .
the inset of the figure shows that disks of all sizes and on all grids span a single curve which obeys approximately a power law . in the two - point function and the roughness metric @xmath243 .
the relation was obtained from disks of increasing sizes @xmath14 defined on grids of size @xmath258 ( @xmath244 ) , @xmath259 ( @xmath260 ) , @xmath261 ( @xmath77 ) , @xmath262 ( @xmath95 ) and @xmath105 ( @xmath247 ) .
the inset shows the data on logarithmic scales .
the solid line is eq .
( [ eq : delta_i_s2_fit ] ) obtained by least - square fit.,title="fig:",width=264 ] +
the degeneracy @xmath61 can be analyzed in terms of the information content associated with a given two - point function @xmath0 .
indeed , if a reconstruction problem is non - degenerate , the two - point correlation function is a complete characterization of the microstructure . by contrast , in the case of a large degeneracy the correlation function contains a relatively small amount of microstructural information .
this idea can be made quantitative by borrowing concepts from information theory @xcite . in that context
, a given microstructure is considered to be the outcome of a random process .
more specifically , if nothing is known other than the total number of pixels @xmath17 , then specifying a given microstructure is equivalent to drawing it out of the complete configuration space .
any microstructure is therefore an event having probability @xmath263 .
the self - information ( or so - called _ surprisal _ ) associated with such an event is @xmath264 the use of a base 2 logarithm ensures that the self - information is expressed in units of _ bits_. the self - information can be used as a quantitative measure for the information content of the realization of an event . in this particular case ,
the value is @xmath265 bits , i.e. , 1 bit per pixel , which is quite consistent .
[ cols="<,^,^,^,^,^,^,^,^,^",options="header " , ] if the number of black pixels @xmath14 is known , a microstructure is no longer a draw out of @xmath266 , but rather out of @xmath267 .
this implies that the self - information is reduced by a quantity @xmath268 which can be thought of as the amount of information contained in specifying the value of @xmath14 , compared to knowing merely the system size .
we refer to @xmath269 as the information content of the one - point statistics .
the quantity @xmath269 depends on the particular value of @xmath14 .
it is equal to @xmath17 for @xmath270 and for @xmath271 . in both cases specifying @xmath14
is a complete description of the microstructure , since in those cases the system is either completely black or completely white .
by contrast , @xmath269 is minimum for @xmath272 , because this particular value maximizes @xmath73 .
the same reasoning can be applied to quantify the information content of @xmath0 . once @xmath0 is given , a particular microstructure is drawn out of @xmath61 possibilities , and no longer @xmath73 .
this suggests defining the quantity @xmath273 to measure the information content of @xmath0 , in addition to knowing @xmath14 .
note that the information content of the one - point and two - point statistics can be understood in terms of configurational entropies corresponding to different definitions of the macrostate of the system . in the case of the one - point information ,
the macrostate is specified via the value of @xmath14 , which results in an entropy @xmath274 .
similarly , if @xmath0 is used to define the macrostate , the entropy becomes @xmath275 .
the information content @xmath276 is equal to the entropy reduction that results from incorporating @xmath0 in the definition of the macrostate of the system .
the analysis of sec .
iv suggests that @xmath276 can be accurately calculated from @xmath243 alone .
this is the significance of fig .
[ fig : information ] .
the inset of the figure shows that the dependency is a power - law of the type @xmath277 } \simeq 11.2 \left ( \frac{\tilde e(1)}{\left < e \right > } \right)^{-0.51 } \ , \ ] ] where the numerical coefficients have been obtained by a least - square fit .
it has to be stressed here that the roughness metric @xmath243 is calculated from @xmath0 alone .
therefore , eq . ( [ eq : delta_i_s2_fit ] ) provides a practical means to estimate @xmath276 in any experimental context where the only information about the system is its correlation function @xmath0 . for a reconstruction to be accurate
, the total information available in the form of one - point and two - point statistics , has to be @xmath17 bits .
therefore , the utility of @xmath269 and @xmath276 is best illustrated by normalizing them by @xmath17 and defining the following fractional information contents @xmath278 and @xmath279 , which take values between 0 and 1 .
a reconstruction is accurate whenever the sum @xmath280 is close to one .
table [ tab2 ] analyzes the reconstruction examples of figs .
[ fig : rec_examples ] and [ fig : rec_crystal ] along these directions .
the three microstructures of fig .
[ fig : rec_examples ] all have @xmath206 black pixels on a grid with a total of @xmath281 pixels . in this case
, one estimates through eq .
( [ eq : delta_i_s1 ] ) that the one - point information is @xmath282 .
the two - point information @xmath276 for the various microstructures was calculated from the roughness metric @xmath252 through eq .
( [ eq : delta_i_s2_fit ] ) and the corresponding values of @xmath283 are reported in table ii . in the case of the single disk ,
the total information content of @xmath284 and @xmath98 is close to 1 , which means that a perfect reconstruction is possible .
the fact that the value is slightly larger than 1 results from the limited accuracy of eq .
( [ eq : delta_i_s2_fit ] ) . in the case of the poisson point process , the total information available amounts to only 39% of the information required for the reconstruction :
the reconstruction is therefore impossible .
the case of the hard disks is not so clear - cut : the reconstruction captures many structural characteristics of the target ( fig .
[ fig : rec_examples ] ) although only 77% of the information is available ( table [ tab2 ] ) .
this seems to suggest that a fair reconstruction may be possible with about 20% of missing structural information . the information - theoretic analysis of the polycrystal reconstructions of fig .
[ fig : rec_crystal ] proceeds in the same way . in the case of the single crystal ,
93% of the information is ( see table [ tab2 ] ) , which is consistent with the good quality of the reconstruction . for decreasing crystallite sizes
, the amount of information decreases . in the case of the smallest crystallites ,
the amount of information is only about 50 % and the reconstruction is expectedly inaccurate . the consistency of the information analysis of the single disk reconstruction was expected because fig .
[ fig : information ] and eq .
( [ eq : delta_i_s2_fit ] ) are based on disks of various sizes .
the validity of eq .
( [ eq : delta_i_s2_fit ] ) for microstructures other than disks is established only for very small systems , for which the mc algorithm could be used ( fig .
[ fig : roughness ] ) .
the fact the present analysis enables us to predict the non - degeneracy of the crystal reconstruction strongly supports the generality of eq . ( [ eq : delta_i_s2_fit ] ) .
throughout the paper , we have discussed several cases of trivially and non - trivially degenerate microstructures . we have argued that the geometrical features that contribute to decreasing the non - trivial contribution to @xmath61 are those that lead to extremal values of @xmath0 for given values of @xmath1 .
this is notably the case for a single disk under periodic boundary conditions , which is the microstructure that maximizes @xmath0 for sufficiently small @xmath1 s .
the crystal configuration shown in fig .
[ fig : rec_crystal ] is non - degenerate for similar reasons : because of its anisotropy , that microstructure achieves extremal values of @xmath0 for many different values of @xmath1 .
the opposite situation is that of a poisson point processes , for which @xmath0 takes values close to the average value @xmath285 for all @xmath286 .
this leads to a huge degeneracy ( configurational entropy ) because none of the values of @xmath0 is close to being extremal .
it is interesting to note that the single disk within a periodic box can be thought of as a dilute distribution of disks in all of euclidean space when the infinite number of periodically replicated cells are considered . in the case of impenetrable disks at arbitrary density , it is noteworthy that @xmath0 can be exactly expressed in terms of a one - body or low - density term ( which contains the same shape and surface area information as in the dilute regime ) and a single higher - order two - body term involving pair statistics @xcite .
it is therefore the latter term that is responsible for the degeneracy of such configurations for arbitrary densities .
the trivial contribution to the degeneracy @xmath61 depends on the particular rotational symmetry and chirality of the microstructure , but it is always of the order of total number of pixels in the grid @xmath17 . by contrast , the non - trivial contribution to the degeneracy can be significantly larger .
the monte carlo estimation of @xmath61 shows that even a modestly sized @xmath91 poisson point process can have a degeneracy as large as @xmath287 ( see fig .
[ fig : dos_13points ] ) .
this value is expected to increase exponentially with the size of the system because any possible @xmath98-preserving pixel displacement contributes multiplicatively to @xmath61 . in order to quantitatively
address the question of the degeneracy corresponding to any specified correlation function , we have mapped it to the determination of a ground - state degeneracy .
this mapping led us to two breakthroughs .
first , we now can calculate for the first time the density of states for reconstruction problems via a monte carlo algorithm , and in particular to determine the values of @xmath61 for a few benchmark systems .
second , we built on the general observation throughout physics that large ground - state degeneracies are generally associated with rough energy landscapes , which enabled us to use the roughness of the energy landscape as a proxy for the microstructural degeneracy .
a natural metric for the roughness of the energy landscape is the total curvature of the energy surface , evaluated at the ground - states . using a random walk in configuration space ( see fig .
[ fig : random_walk ] ) , we derived an analytic expression for the total energy - surface curvature in the form of @xmath243 , which can be calculated in terms of @xmath0 alone .
the monte carlo analysis confirms that @xmath243 is indeed highly correlated with the degeneracy of a reconstruction problem , independently of the type or microstructure considered ( fig .
[ fig : roughness ] ) .
it has to be noted that the roughness metric is consistent with the intuitive analysis of degeneracy in terms of extremal values of @xmath0 .
indeed , the main contribution to the denominator @xmath166 is @xmath288 , so that any value of @xmath0 larger or smaller than @xmath285 contributes to decreasing the roughness metric , and hence the degeneracy .
a counterintuitive result of the present study is that the distinction between trivial and non - trivial degeneracy is irrelevant in configuration space @xmath107 .
in particular , the quantitative relationship found between @xmath61 and the roughness metric does not discriminate the two types of degeneracy .
this enabled us to use trivially - degenerate microstructures to generate a single calibration relation for @xmath61 as a function of @xmath243 . that relation applies to a large variety of microstructures of sizes much larger than those analyzable by the monte carlo method ( see fig .
[ fig : information ] ) .
we should point out that although the examples discussed in the present work are all two - dimensional microstructures , the same methodology can be applied in any space dimension .
it is indeed noteworthy that eq .
( [ eq : e1 ] ) and the approximation eq .
( [ eq : sigma2c_approximation ] ) are valid in any space dimension . as a consequence ,
the roughness metric of any higher - dimensional microstructure can be calculated easily from its correlation function @xmath0 alone .
moreover , the observation that we make that the relation between the roughness metric and the ground - state degeneracy does not discriminate trivial from non - trivial degeneracy is also expected to hold in any space dimension .
therefore , higher - dimensional trivially - degenerate microstructures ( e.g. hyperspheres ) can be used to produce a relation equivalent to eq .
( [ eq : delta_i_s2_fit ] ) or fig .
[ fig : information ] in any space dimension . it is also noteworthy that our analytical results do not assume euclidean space : the only restriction is that @xmath289 should be independent of @xmath25 , where @xmath28 is the operator used to define @xmath0 through eq .
( [ eq : definition_p ] ) .
therefore , the mathematical expression of the roughness metric is valid in hyperbolic and spherical spaces as well as in any dimension .
however , the relationship between the degeneracy ( configurational entropy ) and the roughness metric is expected to be space- and dimension - dependent .
the use of information - theoretic concepts allows our methods to be easily applied in practice . as mentioned in the introduction ,
two - point correlation functions are often the only data available experimentally for in situ studies with a nanometer resolution , notably through small - angle scattering measurements @xcite .
the question of the structural ambiguity of small - angle scattering patterns is an old one @xcite , but the recent development of very intense x - ray sources @xcite has ignited a very lively debate about the possibility of reconstructing nanometer - scale objects from scattering patterns @xcite .
our analysis provides a novel and very general approach to address this type of question : an accurate reconstruction is possible whenever the amount of information @xmath280 is close to one .
the examples that we have discussed suggest that a relatively accurate reconstruction is possible with up to 20% missing information , but it is premature to formulate any general rule .
it has to be stressed that , although the present work is based on the reconstruction of microstructures defined on a grid with exact distances , the results apply unchanged to the discrete reconstruction of microstructures starting from experimental ( i.e. , continuous ) correlation functions .
correlation functions of the type of the monocrystal ( fig .
[ fig : rec_crystal ] ) are unrealistic in an experimental context . however , the general relation between @xmath243 and @xmath276 still holds .
the only difference is that experimental correlation functions of disordered systems are generally of the polycrystal type , with very small crystallites . except in some exceptional cases ,
it is therefore expected that experimental correlation functions with no orientation information should be highly degenerate .
the domain of applications of our results is not limited to scattering .
other applications can notably be found in the field of computer vision for texture recognition .
a texture with low degeneracy @xmath61 can in principle be discriminated robustly based on two - point statistics alone , which would make slower three - point characterizations unnecessary @xcite .
besides applications , information - theoretic concepts are also useful conceptually .
it is very natural that a reconstruction be possible whenever the information content of the available data is equal to the amount of information required , i.e. @xmath17 bits where @xmath17 is the total number of pixels . in the cases we considered ,
the information came in the form of one - point statistics @xmath269 and of two - point statistics @xmath276 .
however , the approach could be generalized naturally to higher - order statistics @xcite .
quite generally , a successful reconstruction would require all correlations to be considered up to the @xmath290th order , with @xmath290 satisfying @xmath291 where @xmath292 is the information contained in n - point correlation function @xmath293 in addition to @xmath294 .
there is some evidence supporting the view that @xmath295 does not contain significant information in addition to @xmath98 @xcite . in the present context
this suggests that the series in eq .
( [ eq : information_series ] ) converges slowly .
the approach could be further generalized to other types of statistical descriptors including lineal statistics @xcite , pore - size functions @xcite , and cluster correlation functions @xcite .
we stress that the present work has numerous ramifications in materials sciences and beyond . for instance , an important question concerns the realizability of two - point correlation functions @xcite . it would be interesting to explore whether new necessary conditions for the realizability of @xmath0 can be obtained by expressing that the information content ( in bits ) can not exceed the total number of pixels in the microstructure .
last but not least , other applications can be found in physics . indeed , the hamiltonian of any system with pairwise additive energy can be written as @xmath296 where @xmath297 is the pair interaction potential .
it results from eq .
( [ eq : pairwise_energy ] ) that systems with identical @xmath0 necessarily have the same energy .
therefore the degeneracy @xmath61 calculated from @xmath0 is a lower bound for the physical ground - state degeneracy of any system with pairwise interaction energy .
this includes systems such as frustrated ising models for which the ground - state degeneracy is not trivial @xcite .
another fascinating field of application is that of quasi - crystalline microstructures @xcite , the degeneracy of which could be analyzed with the general results obtained in the present study .
> from a methodological point of view , the general approach we have developed may be valuable in the manifold of fields where complex energy landscapes have to be characterized .
these include protein folding @xcite , complex chemical reactions @xcite , phase equilibria in disordered porous materials @xcite , and glass transitions @xcite .
we hope to investigate some of these aspects in future work .
c.j.g . acknowledges support from the fonds de la recherche scientifique ( f.r.s .- fnrs , belgium ) and from the patrimoine de luniversit de lige ; s.t . and y.j
. were supported by the office of basic energy science , division of materials science and engineering under award de - fg02 - 04-er46108 .
80ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _
_ ( , , ) @noop _ _ , ed .
( , ) @noop _ _ , vol .
( , , ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop _ _ , vol .
( , , ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase { 10.1016/s0006 - 3495(99)77443 - 6 } [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1063/1.1327609 [ * * , ( ) ] @noop * * , ( ) link:\doibase { 10.1016/j.actamat.2007.10.044 } [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase { 10.1063/1.3592709 } [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase { 10.1016/s0042 - 6989(99)00191 - 1 } [ * * , ( ) ] link:\doibase { 10.1103/physreve.63.066701 } [ * * , ( ) ] link:\doibase { 10.1103/physrevlett.89.135501 } [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase { 10.1073/pnas.0905919106 } [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.080601 [ * * , ( ) ] @noop * * , ( ) link:\doibase { 10.1007/s11004 - 008 - 9209-x } [ * * , ( ) ] @noop * * , ( ) link:\doibase { 10.1016/j.advwatres.2010.12.004 } [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase { 10.1103/physreve.75.046701 } [ * * , ( ) ] @noop * * , ( ) link:\doibase { 10.1063/1.1463059 } [ * * , ( ) ] in http://dx.doi.org/10.1007/978-3-642-15396-9_6[__ ] , , vol . ,
( ) pp . @noop * * , ( ) link:\doibase { url will be inserted by publisher } [ ] link:\doibase { 10.1038/29487 } [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.108.093401 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | a two - point correlation function provides a crucial yet an incomplete characterization of a microstructure because distinctly different microstructures may have the same correlation function . in an earlier letter [ gommes , jiao and torquato , phys .
rev .
lett .
108 , 080601 ( 2012 ) ] , we addressed the microstructural degeneracy question : what is the number of microstructures compatible with a specified correlation function ?
we computed this degeneracy , i.e. , configurational entropy , in the framework of reconstruction methods , which enabled us to map the problem to the determination of ground - state degeneracies . here , we provide a more comprehensive presentation of the methodology and analyses , as well as additional results .
since the configuration space of a reconstruction problem is a hypercube on which a hamming distance is defined , we can calculate analytically the energy profile of any reconstruction problem , corresponding to the average energy of all microstructures at a given hamming distance from a ground state .
the steepness of the energy profile is a measure of the roughness of the energy landscape associated with the reconstruction problem , which can be used as a proxy for the ground - state degeneracy .
the relationship between this roughness metric and the ground - state degeneracy is calibrated using a monte carlo algorithm for determining the ground - state degeneracy of a variety of microstructures , including realizations of hard disks and poisson point processes at various densities as well as those with known degeneracies ( e.g. , single disks of various sizes and a particular crystalline microstructure ) .
we show that our results can be expressed in terms of the _ information content _ of the two - point correlation functions . from this perspective
, the _ a priori _ condition for a reconstruction to be accurate is that the information content , expressed in bits , should be comparable to the number of pixels in the unknown microstructure .
we provide a formula to calculate the information content of any two - point correlation function , which makes our results broadly applicable to any field in which correlation functions are employed . |
a star lives a luminous life by burning h into successively heavier elements .
however , as the fe group nuclei near mass number @xmath11 are most tightly bound , no more nuclear binding energy can be released to power the star by burning `` fe . ''
therefore , heavy elements beyond `` fe '' have to be made by processes other than normal stellar burning .
one such process is the rapid neutron capture process , or the @xmath3-process for short .
this process is responsible for approximately half the natural abundance of nuclei with mass numbers @xmath12 .
typical @xmath3-process elements are eu , pt , u , and th .
a crude picture for the @xmath3-process is as follows .
one starts with some seed nuclei and lots of neutrons .
the seed nuclei then rapidly capture these neutrons to make very neutron - rich unstable progenitor nuclei .
after neutron capture stops , the progenitor nuclei successively @xmath13-decay towards stability and become the @xmath3-process nuclei observed in nature .
the @xmath3-process has a lot to do with supernova neutrinos .
neutrinos and antineutrinos of all three flavors are emitted by the neutron star produced in a supernova .
the individual neutrino species has approximately the same luminosity but very different average energy .
as the neutrinos diffuse out of the neutron star , they thermally decouple from the neutron star matter at different radii due to the difference in their ability to exchange energy with such matter . with higher temperatures at smaller radii , @xmath14 , @xmath15 , and their antineutrinos decoupling at the smallest radii
are imprinted with the highest average energy while @xmath1 decoupling at the largest radii are imprinted with the lowest average energy .
the average energy of @xmath2 lies between those of @xmath4 and @xmath1 .
typical average supernova neutrino energies are @xmath16 mev , @xmath17 mev , and @xmath18 mev .
i emphasize that while different supernova calculations give somewhat different numerical values , there is a robust hierarchy of the average supernova neutrino energies : @xmath19 .
this hierarchy is the most crucial aspect of supernova neutrino emission relevant for our discussion .
a few seconds after the supernova explosion , we have a hot neutron star near the center of the supernova .
the neutron star is still cooling by emitting neutrinos . the shock wave which makes the supernova explosion is far away from the neutron star . on its way out to make the explosion , the shock wave has cleared away almost all the material above the neutron star , leaving behind only a thin atmosphere .
close to the neutron star , the temperature is several mev and the atmosphere is essentially dissociated into neutrons and protons . as the neutrinos emitted by the neutron star free - stream through this atmosphere , some of the @xmath1 and @xmath2 are captured by the neutrons and protons and their energy is deposited in the atmosphere .
in other words , the atmosphere is heated by the neutrinos . as a result
, it expands away from the neutron star and eventually develops into a mass outflow a neutrino - driven `` wind '' @xcite .
the capture reactions @xmath20 and @xmath21 not only provide heating to drive the wind , but also interconvert neutrons and protons . with a significant excess of @xmath22 over @xmath23 , neutron production by the @xmath2 dominates neutron destruction by the @xmath1 .
consequently , the wind material is neutron rich .
as this material expands away from the neutron star , its temperature and density decrease and various nuclear reactions take place to change its composition .
when the temperature drops to @xmath24 mev , essentially all the protons are assembled into @xmath25-particles and the material at this temperature just contains neutrons and @xmath25-particles .
as the temperature drops further , @xmath25-particles and neutrons are burned into heavier nuclei ( the @xmath25-process @xcite ) . by the time the coulomb barrier stops all charged - particle reactions at a temperature of @xmath26 mev , nuclei with @xmath27 have been produced .
these nuclei then become the seed nuclei to capture the remaining neutrons during the subsequent @xmath3-process , which occurs at temperatures below @xmath26 mev ( e.g. , @xcite ) .
an absolutely necessary condition for an @xmath3-process to occur in the neutrino - driven wind is that the wind material must be neutron rich .
as @xmath4 have the highest average energy , significant mixing between @xmath4 and @xmath1 would increase the destruction of neutrons by @xmath1 and drive the wind material proton rich @xcite .
( even in the extreme case of @xmath28 , conversion of @xmath4 into @xmath1 can still drive the wind material proton rich as neutron production by the @xmath2 is hindered while neutron destruction by the @xmath1 is aided by the neutron - proton mass difference . ) for @xmath4 with a cosmologically significant ( vacuum ) mass of @xmath29100 ev , matter - enhanced mixing with a lighter @xmath1 would occur below the region where the neutron - richness of the wind material is determined @xcite .
therefore , such mixing is severely constrained if the @xmath3-process indeed occurs in the neutrino - driven wind in a supernova .
figure [ mix ] shows the parameters for mixing between @xmath4 and @xmath1 that are incompatible with supernova @xmath3-process nucleosynthesis @xcite .
note that the @xmath14-@xmath1 mixing parameters @xmath30@xmath31 at @xmath32 several ev@xmath10 reported by the lsnd experiment @xcite lie in the incompatible region .
thus only the lsnd parameters at @xmath33 several ev@xmath10 are compatible with supernova @xmath3-process nucleosynthesis .
the lsnd parameters at low @xmath34 are also consistent with the results from the karmen experiment @xcite .
one may ask what happens if the lsnd parameters at @xmath32 several ev@xmath10 turn out to be true . in this case ,
supernova @xmath3-process nucleosynthesis can still occur if for example , a light sterile neutrino @xmath6 is introduced so that @xmath4 with the highest average energy will be turned into harmless @xmath6 before they can mix with the @xmath1 @xcite .
of course , one can consider many other possibilities of neutrino mixing that are compatible with supernova @xmath3-process nucleosynthesis . in any case ,
supernova @xmath3-process nucleosynthesis provides a potential probe for mixing between @xmath4/@xmath5 and @xmath35/@xmath36 at the level of @xmath8 for @xmath9 ev@xmath10 . to establish this probe on a solid basis requires us to check the supernova laboratory against the standards of terrestrial experiments .
a terrestrial neutrino oscillation experiment consists of a known neutrino source and a detector with unambiguous signals for neutrino oscillations . in the supernova laboratory ,
the hot neutron star is the neutrino source .
the most crucial aspect of this source is the robust hierarchy of average neutrino energies . as the neutrinos move away from the neutron star
, various scenarios of neutrino mixing can occur .
the @xmath1 and @xmath2 emerging from the mixing region are detected by the capture reactions on neutrons and protons , respectively .
the signal that we are looking for is the production of @xmath3-process elements such as eu and u in supernovae .
if this signal is observed , any scenario of neutrino mixing that would cause neutron destruction by the @xmath1 to dominate neutron production by the @xmath2 is forbidden .
note that the use of supernova nucleosynthesis to study forbidden scenarios of neutrino mixing only relies on the _ necessary _ condition for an @xmath3-process to occur in the neutrino - driven wind the wind material must be neutron rich .
provided that we can prove the supernova origin of the @xmath3-process elements , we do not have to understand the exact details of supernova @xmath3-process nucleosynthesis in order to use the supernova laboratory in the `` forbidden '' mode for studying neutrino mixing . on the other hand ,
if we also know the exact characteristics of supernova neutrino emission and of the neutrino - driven wind , we can even determine whether a certain neutrino mixing scenario is required based on the _ sufficient _ conditions for supernova @xmath3-process nucleosynthesis .
however , great improvements in our understanding of supernovae have to be made before the supernova laboratory can be used in the `` required '' mode for studying neutrino mixing . as the supernova origin of the @xmath3-process elements
is the basis for using supernova nucleosynthesis to study neutrino mixing in either the forbidden or the required mode , it will be the focus of the following discussion .
i first present indirect arguments for the supernova origin of the @xmath3-process elements based on recent observations of @xmath3-process elemental abundances in metal - poor stars and consideration of galactic chemical evolution .
two possible direct tests are discussed next .
the astrophysical site for the @xmath3-process has to provide a large neutron abundance .
this can be achieved by having @xmath2 capture on protons dominate @xmath1 capture on neutrons in a supernova .
alternatively , neutron - rich material may be obtained in the merger of a neutron star with another neutron star or a black hole .
in fact , supernovae and neutron star mergers are considered as the two leading candidate sites for the @xmath3-process .
an important distinction between these two sites is the vast difference in the event rate .
massive stars that explode as supernovae are a small fraction of all stars . the progenitor system for neutron star mergers must have two massive stars in a binary .
furthermore , this binary must survive the two supernova explosions that produce the two compact objects for the eventual merger .
a rather high estimate of the neutron star merger rate in the galaxy is @xmath37 ( e.g. , @xcite ) .
this is still @xmath38 times smaller than the galactic supernova rate .
let us assume that supernovae are the major source for the @xmath3-process and consider @xmath3-process enrichment of the interstellar medium ( ism ) .
the ejecta from each supernova is mixed with an average mass @xmath39 of ism ( mostly h ) swept up by the supernova remnant ( e.g. , @xcite ) . for a supernova rate that is proportional to the mass of gas , an average ism in the galaxy
is enriched by supernova ejecta at a frequency of @xmath40 , where the supernova rate per unit mass of gas is estimated using @xmath41 and @xmath42 for the present galaxy .
consequently , an average ism would be enriched with a solar @xmath3-process composition ( denoted by the subscript `` @xmath43 '' ) by @xmath38 supernovae over a period of @xmath44 yr .
this then determines the @xmath3-process abundances resulting from a single supernova , e.g. , ( eu / h)@xmath45 ( eu / h)@xmath46 with eu / h being the abundance ratio of eu to h. in the spectroscopic notation @xmath47 , we have @xmath48 @xcite . the observed eu abundances in many metal - poor stars @xcite are shown in figure [ eu ] .
the `` metallicity '' is defined by [ fe / h ] @xmath49 .
the low values of [ fe / h ] for the stars indicate that they were formed at very early times when the ism had been enriched by only a small number of supernovae .
the lowest eu abundances observed in metal - poor stars are in agreement with the @xmath3-process enrichment resulting from a single supernova discussed above .
the ejecta from each neutron star merger is mixed with approximately the same amount of ism as swept up by a supernova remnant .
however , as the galactic rate of neutron star mergers is @xmath38 times smaller than that of supernovae , an average ism would be enriched by the ejecta from only @xmath29 neutron star merger over a period of @xmath44 yr .
consequently , the eu abundance resulting from a single event would be @xmath50 if neutron star mergers were the major source for the @xmath3-process .
this is in clear disagreement with the data in figure [ eu ] @xcite .
supernovae also provided fe enrichment of the ism at [ fe / h ] @xmath51 @xcite .
figure [ eu ] shows that there is a correlation between the abundances of eu and fe at [ fe / h ] @xmath51 .
this can be explained as the result from mixture of the eu and fe produced by many supernovae if supernovae are the major source for the @xmath3-process . on the other hand ,
an average ism would be enriched in fe by many supernovae between the occurrence of two successive neutron star mergers due to the vast difference in the event rate .
consequently , the correlation between the abundances of eu and fe , especially its early onset at [ fe / h ] @xmath52 , would be very difficult to explain if eu enrichment were provided by neutron star mergers while fe enrichment was provided by supernovae @xcite .
therefore , consideration of galactic chemical evolution and observations of @xmath3-process elemental abundances in metal - poor stars strongly favor supernovae over neutron star mergers as the major source for the @xmath3-process . furthermore , the total amount of @xmath3-process ejecta required from each supernova to explain the observed @xmath3-process elemental abundances in metal - poor stars is consistent with the amount of material ejected in the neutrino - driven wind @xcite . in summary , there is strong evidence for the supernova origin of the @xmath3-process elements .
as described in the introduction , the @xmath3-process initially produces very neutron - rich unstable progenitor nuclei . during the decay towards stability , some progenitor nuclei decay to the excited states of their daughters
the gamma rays from the de - excitation of the daughters constitute the signal for the presence of these @xmath3-process progenitor nuclei .
if such gamma rays are detected from a future supernova , then we will have proven the supernova origin of the @xmath3-process elements .
a supernova becomes transparent to gamma rays after approximately one year of expansion .
therefore , the relevant @xmath3-process progenitor nuclei must have lifetimes of @xmath53 yr .
three most promising nuclei are @xmath54sb , @xmath55ce , and @xmath56os .
the typical gamma - ray flux from a supernova at a distance of 10 kpc is @xmath57 @xmath58 @xmath59 s@xmath60 @xcite . to detect such fluxes requires a future supernova and a new detector with a sensitivity of @xmath61 @xmath58 @xmath59 s@xmath60 .
therefore , while this is the most direct means to prove @xmath3-process production in supernovae , it is also the hardest .
if we can find suitable @xmath3-process progenitor nuclei with lifetimes much longer than one year , we may search for the decay gamma rays from the remnant of a past supernova .
as a longer lifetime means a smaller decay rate , to ensure a substantial gamma - ray flux requires a nearby supernova remnant ( snr ) .
the vela snr is at a distance of @xmath62 pc .
the age of the pulsar in this snr is @xmath63 yr . the relevant @xmath3-process progenitor nucleus for gamma - ray detection is @xmath64sn with a lifetime of @xmath65 yr and several prominent decay gamma rays .
the expected gamma - ray fluxes due to decay of @xmath64sn in the vela snr are @xmath57 @xmath58 @xmath59 s@xmath60 @xcite .
a new snr near vela was discovered recently through its x - ray emission and the gamma rays from decay of @xmath66ti . as @xmath66ti has a lifetime of only @xmath67 yr
, the age of the new snr is @xmath68 yr . in this case , we can search for decay gamma rays from a number of actinides with lifetimes of @xmath38 yr .
the expected fluxes are again @xmath57 @xmath58 @xmath59 s@xmath60 @xcite .
as we can see , to prove @xmath3-process production in supernovae by gamma - ray astronomy requires a new detector with a sensitivity of @xmath61 @xmath58 @xmath59 s@xmath60 . by comparison ,
the sensitivity of the compton gamma - ray observatory just demissioned was @xmath69 @xmath58 @xmath59 s@xmath60 , and that of the integral experiment to be launched within the next few years is @xmath70 @xmath58 @xmath59 s@xmath60 .
perhaps a sensitivity of @xmath61 @xmath58 @xmath59 s@xmath60 can be reached within the next decade .
there is yet another way to prove @xmath3-process production in supernovae if we take advantage of the occurrence of supernovae in binaries .
approximately half of the stars are in binaries .
some binaries initially consist of a massive star and a low - mass star .
after the massive one explodes as a supernova , it is possible for the neutron star or black hole produced in the supernova to remain in orbit around the low - mass star . furthermore , a fraction of the @xmath3-process ejecta from the supernova would be intercepted by the low - mass star .
therefore , @xmath3-process production in supernovae will be proven if we detect @xmath3-process abundance anomalies on the surface of the binary companion to a neutron star or black hole @xcite .
large overabundances of supernova products such as o , mg , si , and s have been observed recently in the binary companion to a black hole @xcite . with the use of the hubble space telescope and the keck observatory , perhaps this kind of observation can be extended successfully to the @xmath3-process elements within the next few years .
i have discussed the role of neutrinos in heavy element production in supernovae , especially the effects of neutrino mixing on supernova @xmath3-process nucleosynthesis .
the neutron - richness of the material in the neutrino - driven wind in a supernova is sensitive to mixing between @xmath4/@xmath5 and @xmath35/@xmath36 at the level of @xmath8 for @xmath9 ev@xmath10 .
a necessary condition for an @xmath3-process to occur in the wind is that the wind material must be neutron rich .
provided that the supernova origin of the @xmath3-process elements can be proven , this necessary condition can be used to eliminate any scenario of neutrino mixing that would cause neutron destruction by the @xmath1 to dominate neutron production by the @xmath2 .
i have presented indirect arguments for the supernova origin of the @xmath3-process elements based on recent observations of @xmath3-process elemental abundances in metal - poor stars and consideration of galactic chemical evolution .
i have also discussed two direct tests for @xmath3-process production in supernovae : detection of gamma rays due to decay of @xmath3-process progenitor nuclei from a future supernova or nearby supernova remnant and observation of @xmath3-process abundance anomalies on the surface of the binary companion to a neutron star or black hole .
hopefully , these tests will prove the supernova origin of the @xmath3-process elements in the near future , thereby establishing supernova @xmath3-process nucleosynthesis as an extremely sensitive probe for neutrino mixing .
this work was supported in part by the department of energy under grant de - fg02 - 87er40328 .
duncan , s.l .
shapiro and i. wasserman , astrophys .
j. 309 ( 1986 ) 141 .
z . qian and s.e .
woosley , astrophys .
j. 471 ( 1996 ) 331 .
woosley and r.d .
hoffman , astrophys .
j. 395 ( 1992 ) 202 .
meyer et al . , astrophys .
j. 399 ( 1992 ) 656 .
k. takahashi , j. witti and h .- th .
janka , astron .
286 ( 1994 ) 857 .
woosley et al . , astrophys .
j. 433 ( 1994 ) 229 .
qian et al . , phys .
71 ( 1993 ) 1965 .
z . qian and g.m .
fuller , phys .
d 51 ( 1995 ) 1479 .
g. mills , these proceedings and references therein .
k. eitel , these proceedings and references therein .
caldwell , g.m .
fuller and y .- z .
qian , phys . rev .
d 61 ( 2000 ) 123005 .
z. arzoumanian , j.m .
cordes and i. wasserman , astrophys .
j. 520 ( 1999 ) 696 .
k. thornton et al .
, astrophys .
j. 500 ( 1998 ) 95 .
z . qian and g.j .
wasserburg , phys . rep .
333334 ( 2000 ) 77 .
wasserburg and y .- z .
qian , astrophys .
j. 529 ( 2000 ) l21 . y .-
qian , astrophys .
j. 534 ( 2000 ) l67 .
r.g . gratton and c. sneden , astron .
astrophys . 287
( 1994 ) 927 .
a. mcwilliam et al . , astron .
j. 109 ( 1995 ) 2757 . c. sneden et al .
, astrophys .
j. 467 ( 1996 ) 819 . c. sneden et al .
, astrophys .
j. 496 ( 1998 ) 235 .
qian , p. vogel and g.j .
wasserburg , astrophys .
j. 506 ( 1998 ) 868 .
qian , p. vogel and g.j .
wasserburg , astrophys .
j. 524 ( 1999 ) 213 .
g. israelian et al . , nature 401 ( 1999 ) 142 . | intense fluxes of neutrinos are emitted by the hot neutron star produced in a supernova .
the average supernova neutrino energies satisfy a robust hierarchy @xmath0 .
the @xmath1 and @xmath2 capture reactions on neutrons and protons , respectively , provide heating to drive a wind from the hot neutron star .
the same reactions also determine the neutron - richness of the wind material .
nucleosynthesis via rapid neutron capture , the @xmath3-process , may occur in the wind material as it expands away from the neutron star .
the neutron - richness of the wind material , and hence , the @xmath3-process nucleosynthesis therein , are sensitive to mixing between @xmath4/@xmath5 and @xmath1/@xmath2 ( or sterile neutrinos @xmath6/@xmath7 ) at the level of @xmath8 for @xmath9 ev@xmath10 . indirect arguments and direct tests for the supernova origin of the @xmath3-process elements are discussed with a goal to establish supernova @xmath3-process nucleosynthesis as an important probe for neutrino mixing . |
neutron stars are laboratories for dense matter physics , because they are composed of the densest form of hadrons and leptons .
however , the detail of neutron star properties , for example , the mass , radius and the particle fractions in the core of neutron star , is not yet theoretically understood .
thus , the observation of the mass and/or radius can provide strong constraints on the equation of state ( eos ) of highly dense matter .
the typical mass of a neutron star is known to be around @xmath5 ( solar mass ) .
however , a few massive neutron stars have recently been observed precisely .
shapiro delay measurements from radio timing observations of the binary millisecond pulsar ( psr j1614 - 2230 ) have indicated the mass of @xmath1 @xcite . however , it is difficult to explain the mass of psr j1614 - 2230 by the eos which has been calculated so far , because the inclusion of hyperons makes the eos very soft and the maximum mass of a neutron star is thus reduced .
therefore , it is very urgent to reconcile this discrepancy .
@xcite , we have extended the quark - meson coupling ( qmc ) model @xcite to include the quark - quark hyperfine interaction due to the exchanges of gluon and pion based on chiral symmetry .
we call this improved version the chiral qmc ( cqmc ) model .
then , we have applied it to a neutron star within relativistic hartree - fock ( rhf ) approximation , and shown that the mass of psr j1614 - 2230 can be explained @xcite . in the present calculation , instead of rhf
, we study the neutron - star properties in relativistic hartree approximation .
the lagrangian density is chosen to be @xmath6\psi_{b } \nonumber \\ & + & \frac{1}{2 } \left ( \partial_{\mu}\sigma\partial^{\mu}\sigma - m_{\sigma}^{2}\sigma^{2 } \right ) + \frac{1}{2 } \left ( \partial_{\mu}\sigma^{\ast}\partial^{\mu}\sigma^{\ast } - m_{\sigma^{\ast}}^{2}\sigma^{\ast2 } \right ) \nonumber \\ & + & \frac{1}{2}m_{\omega}^{2}\omega_{\mu}\omega^{\mu}- \frac{1}{4}w_{\mu\nu}w^{\mu\nu } + \frac{1}{2}m_{\phi}^{2}\phi_{\mu}\phi^{\mu}- \frac{1}{4}p_{\mu\nu}p^{\mu\nu } \nonumber \\ & + & \frac{1}{2}m_{\rho}^{2}\vec{\rho}_{\mu}\cdot\vec{\rho}^{\mu } - \frac{1}{4}\vec{r}_{\mu\nu}\cdot\vec{r}^{\mu\nu } + \sum_{\ell}\bar{\psi}_{\ell}\left[i\gamma_{\mu}\partial^{\mu}-m_{\ell}\right]\psi_{\ell},\end{aligned}\ ] ] where @xmath7 with @xmath8 the baryon ( lepton ) field , @xmath9 the free baryon ( lepton ) mass and @xmath10 the isospin matrix for baryon @xmath11 .
the sum @xmath11 runs over the octet baryons and @xmath0-isobars , while the sum @xmath12 is for the leptons , @xmath13 and @xmath14 .
the baryon - baryon interaction is given through the exchanges of not only the isoscalar ( @xmath15 and @xmath16 ) and isovector ( @xmath17 ) mesons but also the isoscalar , strange ( @xmath18 and @xmath19 ) mesons .
the effective baryon mass , @xmath20 , is calculated by using the simple qmc or cqmc model , and it can be parameterized as a function of @xmath15 and @xmath18 : @xmath21 with @xmath22 , \hspace{0.4 cm } g_{\sigma^{\ast}b}(\sigma^{\ast } ) = g_{\sigma^{\ast}b}b_{b}^{\prime } \left[1-\frac{a_{b}^{\prime}}{2}\left(g_{\sigma^{\ast}\lambda}\sigma^{\ast}\right)\right ] , \label{eq : scalar - coupling - constants}\ ] ] where @xmath23 and @xmath24 are respectively the @xmath15-@xmath25 and @xmath18-@xmath2 coupling constants at zero density . here
, we introduce four parameters , @xmath26 , @xmath27 , @xmath28 and @xmath29 , to describe the mass , and these values are listed in table [ tab:1 ] .
lcccccccc & & + @xmath11 & @xmath26 ( fm ) & @xmath27 & @xmath28 ( fm ) & @xmath29 & @xmath26 ( fm ) & @xmath27 & @xmath28 ( fm ) & @xmath29 + @xmath25 & 0.179 & 1.00 & & & 0.118 & 1.04 & & + @xmath2 & 0.172 & 1.00 & 0.220 & 1.00 & 0.122 & 1.09 & 0.290 & 1.00 + @xmath4 & 0.177 & 1.00 & 0.223 & 1.00 & 0.184 & 1.02 & 0.277 & 1.15 + @xmath3 & 0.166 & 1.00 & 0.215 & 1.00 & 0.181 & 1.15 & 0.292 & 1.04 + @xmath0 & 0.200 & 1.00 & & & 0.197 & 0.89 & & + if we set @xmath30 and @xmath31 , @xmath32 is identical to the @xmath15-@xmath11 coupling constant in quantum hadrodynamics .
this is also true of the coupling of @xmath33 .
to study the role of hyperons ( y ) on the properties of neutron star , it is very important to extend su(6 ) symmetry based on the quark model to the more general su(3 ) flavor symmetry . in general , the interaction lagrangian density in su(3 ) is given by @xcite @xmath34 b ) + ( 1-\alpha ) { \rm tr}(\{{\bar b } , m_8\ } b ) \right ] - g_1 \frac{1}{\sqrt{3 } } { \rm tr}({\bar b}b){\rm tr}(m_1 ) , \label{eq : su(3)-lagrangian}\ ] ] where @xmath35 and @xmath36 are respectively the coupling constants for the meson singlet and octet states , and @xmath37 ( @xmath38 ) denotes the @xmath39 ratio .
for details , see the reference @xcite .
@xcite . ]
the latest extended soft - core ( esc ) model by the nijmegen group suggests that the ratio @xmath40 is 0.1949 for vector mesons , while @xmath41 in su(6 ) symmetry . in the esc model , the mixing angle is chosen to be @xmath42 , which is close to the ideal mixing ( @xmath43 ) .
therefore , in the present study , we use the @xmath44 value in the esc model and assume the ideal mixing to determine the vector - meson couplings to hyperons . for the @xmath15-y coupling constants , as in @xcite , we fix them so as to fit the hyperon potential suggested by the experiments . in fig .
[ fig : tov ] , we present the mass - radius relation for a neutron star , which is calculated using the tolman - oppenheimer - volkoff ( tov ) equation with the eos in the qmc or cqmc model .
the inclusion of the strange mesons ( @xmath18 and @xmath19 ) makes the eos stiff , and the maximum mass of neutron star is thus pushed upwards in both models .
in fact , this enhancement is mainly caused by @xmath19 meson , and the effect of @xmath18 meson is not large .
furthermore , extending su(6 ) to su(3 ) symmetry in the vector - meson couplings ( this case is denoted by beyond su(6 ) " in the figure ) , the maximum mass can reach the massive pulsar mass , @xmath1 @xcite .
the particle fractions in a neutron star are shown in fig .
[ fig : composition ] . in the case of beyond su(6 ) " , the creation densities of @xmath2 and @xmath3 become higher than in the case of su(6 ) .
in contrast , @xmath4 and @xmath0-isobars do not appear in a neutron star . .
the left ( right ) panel is for the case of su(6 ) ( beyond su(6 ) ) in the cqmc model .
, title="fig : " ] .
the left ( right ) panel is for the case of su(6 ) ( beyond su(6 ) ) in the cqmc model .
, title="fig : " ] in the cqmc model , due to the effects of the in - medium baryon structure variation and the quark - quark hyperfine interaction , the potential @xmath0-isobars feel in matter is shallower than in the case of quantum hadrodynamics .
these effects suppress the appearance of @xmath0-isobars .
we have applied the chiral quark - meson coupling model to the eos for a neutron star . using the vector - meson couplings based on flavor
su(3 ) symmetry , we have found that the maximum mass of a neutron star can reach the recently observed mass , @xmath1 , even in relativistic hartree approximation .
furthermore , we have examined how @xmath0-isobars contribute to the eos , and shown that , because of the baryon structure variation in matter and the quark - quark hyperfine interaction , they do not appear in a neutron star .
we finally note that the fock term considerably contributes to the nuclear symmetry energy ( @xmath45 ) @xcite .
99 demorest , p. , pennucci , t. , ransom , s. , roberts , m. , hessels , j. : shapiro delay measurement of a two solar mass neutron star .
nature 467 , 1081 ( 2010 ) nagai , s. , miyatsu , t. , saito , k. , tsushima , k. : quark - meson coupling model with the cloudy bag .
b 666 , 239 ( 2008 ) saito , k. , tsushima , k. , thomas , a.w .
: nucleon and hadron structure changes in the nuclear medium and impact on observables .
prog . part .
58 , 1 ( 2007 ) miyatsu , t. , katayama , t. , saito , k. : effects of fock term , tensor coupling and baryon structure variation on a neutron star .
b 709 , 242 ( 2012 ) katayama , t. , miyatsu t. , saito , k. : eos for massive neutron stars .
arxiv:1207.1554 [ astro-ph.sr ] rijken , t.a . ,
nagels , m.m . , yamamoto , y. : baryon - baryon interactions : nijmegen extended - soft - core models .
185 , 14 ( 2010 ) | we study the properties of neutron star using the chiral quark - meson coupling model , in which the quark - quark hyperfine interaction due to the exchanges of gluon and pion based on chiral symmetry is considered .
we also examine the effects of hyperons and @xmath0-isobars in a neutron star . extending the su(6 ) spin - flavor symmetry to more general su(3 ) flavor symmetry in the vector - meson couplings to baryons
, the maximum mass of neutron star can reach the recently observed , massive pulsar mass , @xmath1 . in this calculation , @xmath2 and @xmath3
are generated in a neutron star , while @xmath4 and @xmath0-isobars do not appear . |
over the last decade , numerical simulations have developed into a cornerstone of astrophysical research . from interactions of vast clusters of galaxies to the formation of proto - planetary discs , simulations allow us to evolve systems through time and view them at every angle .
they are an irreplaceable test - bed of our physical understanding of the universe .
a number of hydrodynamics codes have been developed that are widely used in this field and still more are being developed .
fundamentally , they all do the same job ; they solve the equations of motion to calculate the evolution of matter through time .
whether the matter represents a nebula for the birth of a star or a network of galaxy clusters , the basic technique remains the same .
however , the algorithms used to solve these equations vary from code to code and this results in differences in the resulting data .
understanding the origin of these variations is vital to the understanding of the results themselves ; is an observed anomaly an interesting piece of new physics or a numerical effect ? as observational data takes us deeper into the universe , it becomes more important to pin down the origin of these numerical artifacts .
additionally , it is difficult to compare results from simulations run with different codes . with observations , papers clearly state the properties of the instrument such as the diameter of the mirror and the wavelengths it is most sensitive to . while a brief description of the code is always included in theoretical papers , there exists no obvious conversion to other numerical techniques and therefore the results are more difficult for the reader to interpret .
the problem of code comparison is not new and it is a topic that has recently created a great deal of interest .
the reason for its current importance is a positive one ; improved numerical techniques and increased computer power have resulted in simulations reaching greater resolutions than could have been imagined even a few years ago .
however , this high refinement comes at a price ; as we start to pick out the detail of these complex fluid flows , the physics we need to consider gets dramatically more complicated .
this brings us to the main question code comparison projects are trying to answer ; can we use the same tools for this new regime of problems ?
a number of papers have come out that tackle this .
one of the most famous is the santa barbara comparison project @xcite which compared many of the progenitors of today s codes by running a model of a galaxy cluster forming .
taking a different tack @xcite compared the performance of a dozen different implementations of a single approach ( in this case smoothed particle hydrodynamics , sph ) on standard astrophysical problems that included the sod shock , examining the range of outcomes that were available from a single technique .
they concluded that one of the weaknesses of sph was the weak theoretical grounding which allows several equally viable formulations to be derived .
more recent work includes @xcite , who focus specifically on the formation of fluid instabilities , comparing the formation of kelvin - helmholtz and rayleigh - taylor instabilities in six of the most utilised codes .
additionally , @xcite has completed a direct comparison between two particular codes ( _ enzo _ and _ gadget2 _ , see below ) looking at the formation of galaxies in a cosmological context .
all these projects give detailed insights into the differences between the codes , but are unable to provide a quantitative measure of how well a code performs in a particular aspect .
this is especially true of the cosmological - based tests of @xcite where the problem is not sufficiently well - posed for convergence onto a single answer .
@xcite sets out a simpler problem and compares it to analytical predictions , but the system is still sufficiently complex not to have an exact solution .
additionally , no previous comparison has attempted to quantitatively compare different codes to one another ; asking whether it is possible to obtain identical results and with what conditions . without this crucial piece of information , it is impossible to fully assess a piece of work performed by an unfamiliar code or to judge which code might be the most suited to a given problem type . this has resulted in somewhat general comments being made about the differences between numerical techniques which has led to many myths about a code s ability becoming accepted dogma .
the set of tests we present in this paper are designed to tackle these difficulties with the intention that they might become part of an established test programme all hydrodynamical codes should attempt .
we present four problems that specifically address different aspects of the numerical code all of which have expected ` correct ' answers to compare to .
the first two tests , the sod shock tube and sedov blast , are both strong shock tests with analytical solutions .
the third and forth tests concern the stability of a galaxy cluster and are primarily tests of the code s gravitational solver .
for all four tests we directly compare the codes against the analytic solution and present an estimate of the main sources of any systematic error .
the remainder of this paper is organised as follows : in section 2 we give a short summary of the main features of each of the four codes we have employed . in section 3
we deal with the sod shock and sedov blast tests , setting out the initial and final states and comparing each code against them .
we repeat this exercise for a static and translating king sphere in section 4 .
finally we discuss our results and summarise our conclusions in section 5 .
the two major techniques for modelling gases in astrophysics are smoothed particle hydrodynamics ( sph ) and adaptive mesh refinement ( amr ) . in the first of these ,
the gas is treated as a series of particles whose motion is dictated by lagrangian dynamics . in amr
, the gas is modelled by a series of hierarchical meshes and the flow of material between cells is calculated to determine its evolution .
there are a variety of codes which utilise both techniques and four of the major ones will be used to run the tests presented in the this paper , two of which use sph ( _ hydra _ and _ gadget2 _ ) and two of which use amr ( _ enzo _ and _ flash _ ) . _
enzo _ is a massively parallel , eulerian adaptive mesh refinement code , capable of both hydro and n - body calculations @xcite .
it has two hydro - algorithms which can be selected by the user ; the piecewise parabolic method ( ppm ) and the zeus astrophysical code . the ppm solver @xcite uses godunov s method but with a higher - order spatial interpolation , making it third - order accurate .
it is particularly good at shock capturing and outflows .
the zeus method in _ enzo _ is a three - dimensional implementation of the zeus astrophysical code developed by @xcite .
it is a simple , fast algorithm that allows large problems to be run at high resolution . rather than godunov s method
, zeus uses an artificial viscosity term to model shocks , a technique which inevitably causes some dissipation of the shock front .
we compare both these hydro - schemes in these tests . _
gadget2 _ is a massively parallel , lagrangian , cosmological code that is publicly available from the author s website @xcite .
it is an n - body / sph code that calculates gravitational forces by means of the tree method @xcite and is also able to optionally employ a tree - pm scheme to calculate the long range component of the gravitational interactions . in order to follow the hydrodynamic behaviour of a collisional medium ,
the code uses the entropy - conserving formulation of sph described in @xcite : the main difference of this approach with respect to the standard @xcite formulation of sph resides in the choice of describing the thermodynamic state of a fluid element in terms of its specific entropy rather than its specific thermal energy .
this leads to a tight conservation of both energy and entropy in simulating dissipation - free systems . additionally , _
gadget2 _ employs a slightly modified parametrisation of the artificial viscosity ( by introducing the so called `` signal velocity '' as in @xcite ) .
the user is allowed to set the strength of this artificial viscosity for the specific problem being considered via an input parameter , @xmath0 .
the time stepping scheme adopted by the code is a leap - frog integrator guaranteed to be symplectic if a constant timestep for all particles is employed . in this work
we have exploited the possibility of using fully adaptive individual timesteps for all the particles in the simulation , this being a standard practice .
_ hydra _ is an adaptive particle - particle , particle - mesh code combined with smoothed particle hydrodynamics @xcite .
it has the significant disadvantage that even though massively parallel versions exist @xcite , the publically available version is not a parallel implementation and so this code can not , as released , be used for very large simulations .
akin to _ gadget2 _ , _ hydra _ uses an entropy conserving implementation of sph , but unlike _ gadget2 _ , _ hydra _ does not have fully adaptive individual timesteps .
although the timestep adapts automatically from one step to the next all the particles move in lockstep . _ flash _ is a publicly available massively parallel eulerian amr code developed by the alliances center for astrophysical thermonuclear flashes @xcite .
originally intended for the study of x - ray bursts and supernovae , it has since been adapted for many astrophysical conditions and now includes modules for relativistic hydrodynamics , thermal conduction , radiative cooling , magnetohydrodynamics , thermonuclear burning , self - gravity and particle dynamics via a particle - mesh approach .
_ flash _ uses the oct - tree refinement scheme of the paramesh package , with each mesh block containing the same number of internal zones .
neighbouring blocks may only differ by one level of refinement with each level of refinement changing the resolution by a factor of two .
the hydrodynamics are based on the prometheus code @xcite .
the input states for the riemann solver are obtained using a directionally split ppm solver @xcite and a variable time step leapfrog integrator with second order strang time splitting is adopted @xcite .
this work uses a modified hybrid fftw based multigrid solver to solve poisson s equation and determine the gravitational potential at each timestep .
this results in a vast reduction in time spent calculating the self - gravity of the simulation relative to a conventional multigrid solver .
_ flash s _ refinement and de - refinement criteria can incorporate the adapted @xcite error estimator .
this calculates the modified second derivative of the desired variable , normalised by the average of its gradient over one cell .
one of the greatest differences between simulations performed now versus those undertaken five years ago is the increasing importance of modelling strong shocks accurately .
while it has long been known that the universe is a violent place , with events such as supernovae , galaxy mergers and agn generating blasts which rip through the intergalactic medium , simulations did not have the resolution to see such phenomena in detail , so these sharp discontinuities were largely ignored .
now , as we struggle to understand the effects of feedback in galaxy formation , multiphase media are essential physics @xcite . in order to attack such problems codes must be able to capture shocks with some proficiency .
these two problems , the sod shock test and the sedov blast test , explicitly test the resolution of shock jumps and allow comparison with exact analytical solutions . plane , while the right - hand image shows it oriented along the [ 1,1,0 ] plane .
in actuality our second test is oriented in the [ 1,1,1 ] plane i.e. oblique to all the axes . ]
the shock tube problem @xcite has been used extensively to test the ability of hydrodynamics codes to resolve a sharp shock interface @xcite .
the test set - up is simple , consisting of two fluids of different densities and pressures separated by a membrane that is then removed .
the resulting solution has the advantage of showing all three types of fluid discontinuities ; a shock wave moving from the high density fluid to the low density one , a rarefaction ( sound ) wave moving in the opposite direction and a contact discontinuity which marks the current location of the interface between the two fluids . for this test
the initial conditions are traditionally chosen such that the pressure does not jump across the contact discontinuity .
we extend the traditional one - dimensional shock tube problem to consider two three - dimensional set - ups ; the first of these has the fluid membrane at 90@xmath1 to the x - axis of the box ( [ 1,0,0 ] plane ) , causing the shock to propagate parallel to this axis . in the second test
, the membrane is lined up at 45@xmath1 to each of the @xmath2 and @xmath3 axes ( [ 1,1,1 ] plane ) .
this change in orientation of the shock is designed to highlight any directional dependencies inherent in the code , as illustrated in figure [ fig : shock_setup ] .
all analysis was performed perpendicular to the original shock plane . for our particular
set - up we chose the initial density and pressure jump either side of the membrane to be from @xmath4 to @xmath5 with the fluid initially at rest .
the polytropic index was @xmath6 .
periodic boundary conditions were used and the results were analysed at @xmath7 .
figures [ fig : sod_density ] and [ fig : sod_all ] show the results from all four codes running this test . in the results presented in these figures , both _
enzo _ and _ flash _ used an @xmath8 initial ( minimum refinement ) grid with two levels of higher refinement each of which decreased the cell size by a factor of 2 .
the smallest cell size in this case was therefore @xmath9 of the unit box .
_ enzo _ refined anywhere the gradient of the derived quantities exceeded a critical value whereas _ flash _ placed refinements according to the @xcite error estimator described in section [ sec : flash ] which places subgrids based on the second derivative of the derived quantities . for the sph codes , both _
hydra _ and _ gadget2 _ were run with 1 million particles formed from two glasses containing 1.6 million and 400,000 particles .
the solid black line in all cases is the analytical solution which requires a shock at @xmath10 and a contact discontinuity in the density at @xmath11 at @xmath7 .
what is clear from figures [ fig : sod_density ] and [ fig : sod_all ] is that all of the codes pass the zeroth level test and successfully reproduce the shock jump conditions , although both the sph codes suffer from visible ringing and broadening around any discontinuities .
_ enzo ( zeus ) _ does not produce the shock jump condition as accurately as _ enzo ( ppm ) _ and _ flash _ , as seen in the plots of energy and entropy in figure [ fig : sod_all ] , where its post - shock values are around 2% lower . in the oblique case ,
the quadratic viscosity term , @xmath12 , in _ enzo ( zeus ) _ was increased from its default value of 2.0 to 10.0 .
the effect of this value is discussed more fully in relation to the sedov blast test in section [ sec : code_spec ] . for the planar case ,
@xmath12 was kept at 2.0 .
pleasingly none of the other codes appear to have any visible directional dependence , performing equally well in both the grid aligned and oblique cases we tried .
all the codes could equally well resolve the location and smooth rise of the rarefaction wave but both the sph codes struggle with the contact discontinuity , with a large overshoot not seen by either of the mesh based codes .
this is partly due to the initial conditions as for the sph codes the sudden appearance of a density jump introduces a local source of entropy . in this paper
we are contrasting the results from the different approaches rather than studying the sod shock problem itself in detail . as it is a standard test case higher resolution results can be found in the individual code s method papers ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
closer inspection of the data reveals differences in each code s capacity to handle strong shocks . in the bottom row of figure
[ fig : sod_density ] , we show a close - up of the density over the shock - front . capturing shocks
accurately is an area that sph codes traditionally struggle with more than their eulerian counterparts due to their inherent nature of smoothing between particles .
indeed , we see in this figure both _ gadget2 _ and _ hydra _ have a smeared out the interface compared to _
enzo _ and _ flash s _ steep drop in density .
small differences between the amr codes are also visible here .
_ enzo ( ppm ) _ spreads the shock front over three cells , whereas _ enzo ( zeus s ) _ use of the artificial viscosity term extends this to five . _ flash s
_ ppm scheme gives very similar results to _ enzo ( ppm ) _ , also spreading the shock front over three cells .
figure [ fig : sod_all ] shows the pressure , internal energy , velocity and computational entropy @xmath13 over the region of interest for the planar [ 1,0,0 ] set - up ( left column ) and the oblique [ 1,1,1 ] set - up ( right column ) .
both _ hydra _ and _ gadget2 _ show signs of post - shock ringing in the velocity plot .
the two lagrangian codes adopt different implementations of artificial viscosity ; with a frequently used choice of the viscosity parameter ( artbulkvisc = 1 ) the ringing features in _ gadget2 _
s profiles appears to be more pronounced than in hydra s ( not shown in the plot ) .
in order for _
gadget2 _ to get closer to hydra s performance , a choice of a significantly higher viscosity parameter has been necessary ( namely @xmath14 ) .
the results produced under such a choice are shown in figure [ fig : sod_all ] .
the sph codes also exhibit a large spike in both internal energy and entropy at the location of the contact discontinuity .
this is driven by the initial conditions , where two independent particle distributions suddenly appear immediately adjacent to one another .
we saw in the previous section that most of the codes model shock development and sound wave propagation with reasonable success and that they show no directional preference to the orientation of the shock interface .
however , differences were apparent between each of the codes , most obviously between the sph and amr techniques ( unsurprising , since their numerical algorithms fundamentally differ ) . in this section
, we quantitatively compare results from each code and attempt to get as close a match between their results as possible .
for simplicity , we confine the amr codes to using static meshes for this comparison .
figure [ fig : sod_comparison ] shows a graphical comparison of the density projection between a _
run of 1 million particles and _ enzo ( ppm ) _ for different grid sizes .
it is clear that a grid size of @xmath15 produces significantly poorer results than the _ hydra _ data whereas a grid size of @xmath8 produces significantly better results particularly in the low density region .
although the situation is confused by the lack of points available with _ enzo _ , a grid size of @xmath15 can produce no more than 20 distinct values across the length of the volume being modelled , it is clear that this is too few to recover this model .
however , even with @xmath16 cells in the box the major features are largely recovered . to make further progress we require a more quantitative way of comparing the results from the codes . to achieve this
, we employ a cubic spline to interpolate the data from all runs to the same 178 @xmath17 points at which we have calculated the analytical values .
the residue between each new curve and the analytical solution was then summed and divided by the number of points .
we considered the residue both across the whole region of interest from [ -0.15 , 0.15 ] and just across the shock - front from [ -0.13 , -0.06 ] .
.residue of from the analytical solution of the sod shock density in the planar [ 1,0,0 ] set - up for different static grid sizes and sph resolutions for _ enzo , hydra _ and _ gadget2_. [ cols="^,^,^ " , ] [ table : times ] while this paper attempts to cover the major features of astrophysical simulations , it does comprise only four tests .
further examples would extend this paper beyond its scope ( and readability ) , but the differences between codes can not be fully cataloged without further testing .
this paper then , is designed as a starting point for a suite of tests to be developed from which codes can be quantitatively compared for the jobs they are intended for . to assist groups wishing to run these tests on their own code and to compare new updates
, we are making these results and initial conditions available on the web at http://www.astro.ufl.edu/codecomparison .
the authors would like to thank alexei kritsuk , volker springel & richard bower for helpful suggestions and advice and lydia heck for computational support .
ejt acknowledges support from a theoretical astrophysics postdoctoral fellowship from dept . of astronomy / clas , university of florida and
both ejt and glb acknowledge support from nsf grants ast-05 - 07161 , ast-05 - 47823 , and ast-06 - 06959 .
dm acknowledges support from the eu magpop marie curie research and training network . many simulations with _ enzo _
were performed at the national center for supercomputing applications and at the university of florida high - performance computing center who also provided excellent computational support .
the _ flash _
software used in this work was in part developed by the doe - supported asc / alliance center for astrophysical thermonuclear flashes at the university of chicago .
hydra _ and _ gadget2 _ simulations were carried out on the nottingham hpc facility .
flash _ simulations were carried out on the virgo consortium computing facility in durham .
couchman h. m. p. , thomas p. a. , pearce f. r. , 1995 , apj , 452 , 797 pearce f. r. , couchman h. m. p. , 1997 ,
newa , 2 , 411 agertz o. , et al . , 2007 , mnras , 380 , 963 barnes j. , hut p. , 1986
, natur , 324 , 446 bryan , g. l. , & norman , m. l. 1997 , arxiv astrophysics e - prints , arxiv : astro - ph/9710187 feng , l .- l . , shu , c .- w . , &
zhang , m. 2004 , apj , 612 , 1 fryxell b. , et al . , 2000 , apjs , 131 , 273 fryxell b. , mller e. , arnett d. , 1989 , nuas.conf , 100 frenk c. s. , et al . , 1999 , apj , 525 , 554 king i. r. , 1966 , aj , 71 , 64 landau , l. d. , & lifshitz , e. m. 1959 , course of theoretical physics , oxford : pergamon press , 1959 , lhner , r. , 1987 , comp .
61 , 323 mitchell , n. l. , mccarthy , i. j. , bower , r. g. , theuns , t. , & crain , r. a. , in prep , mnras monaghan j. j. , 1992 , ara&a , 30 , 543 monaghan j. j. , 2001 , jkas , 34 , 203 oshea b. w. , nagamine k. , springel v. , hernquist l. , norman m. l. , 2005 , apjs , 160 , 1 oshea , b. w. , bryan , g. , bordner , j. , norman , m. l. , abel , t. , harkness , r. , & kritsuk , a. 2004 , arxiv astrophysics e - prints , arxiv : astro - ph/0403044 padmanabhan , t. 2002 , theoretical astrophysics , by t. padmanabhan , pp . 638 .
isbn 0521562422 .
cambridge , uk : cambridge university press , october 2002 . , robertson b. , kravtsov a. , 2007 , arxiv , 710 , arxiv:0710.2102 ryu , d. , ostriker , j. p. , kang , h. , & cen , r. 1993 , apj , 414 , 1 shapiro , p. r. , martel , h. , villumsen , j. v. , & owen , j. m. 1996 , apjs , 103 , 269 sedov , l. i. 1959 , similarity and dimensional methods in mechanics , new york : academic press , 1959 , sod , g. a. 1978 , journal of computational physics , 27 , 1 springel v. , 2005 , mnras , 364 , 1105 springel v. , hernquist l. , 2003 , mnras , 339 , 289 springel v. , hernquist l. , 2002 , mnras , 333 , 649 stone , j. m. , & norman , m. l. 1992 , apjs , 80 , 753 strang , g. , 1968 , siam j. numer .
, 5 , 506 tasker , e. j. , & bryan , g. l. 2008 , apj , 673 , 810 tasker e. j. , bryan g. l. , 2006 , apj , 641 , 878 thacker r. j. , tittley e. r. , pearce f. r. , couchman h. m. p. , thomas p. a. , 2000 , mnras , 319 , 619 wada k. , norman c. a. , 2007 , apj , 660 , 276 wadsley , j. w. , veeravalli , g. , & couchman , h. m. p. 2008 , mnras , 387 , 427 woodward , p. r. , & colella , p. 1984
, journal of computational physics , 54 , 174 | we test four commonly used astrophysical simulation codes ; enzo , flash , gadget and hydra , using a suite of numerical problems with analytic initial and final states .
situations similar to the conditions of these tests , a sod shock , a sedov blast and both a static and translating king sphere occur commonly in astrophysics , where the accurate treatment of shocks , sound waves , supernovae explosions and collapsed haloes is a key condition for obtaining reliable validated simulations .
we demonstrate that comparable results can be obtained for lagrangian and eulerian codes by requiring that approximately one particle exists per grid cell in the region of interest .
we conclude that adaptive eulerian codes , with their ability to place refinements in regions of rapidly changing density , are well suited to problems where physical processes are related to such changes .
lagrangian methods , on the other hand , are well suited to problems where large density contrasts occur and the physics is related to the local density itself rather than the local density gradient . hydrodynamics methods : numerical cosmology : theory |
phase transitions are ubiquitous in nature .
they are a dramatic change in a system s properties triggered by a minuscule shift in its environment .
phase transitions are often associated with spontaneous symmetry breaking , where the transition is between an unordered phase and an ordered , less symmetric phase @xcite . in simple models of phase transitions an order parameter is defined , which is zero in the unordered phase and non - zero in the ordered phase @xcite .
this is the basis for a mean - field approach to phase transitions , where an expansion of the free energy in the order parameter is performed .
many powerful methods have been developed over the years to study phase transitions , especially in the study of universality in so - called second order ( or critical ) transitions , such as the landau - ginzburg theory @xcite and wilson s renormalization group approach @xcite .
for those transitions the order parameter is continuous , thermodynamic quantities obey scaling laws in the vicinity of the critical point , and there is a diverging correlation length . for first order transitions , on the other hand
, there is a jump in the value of the order parameter , there is no diverging correlation length and thus no scaling of the thermodynamic functions near the transition point . during the transition
there can be a mixed phase with a stable interface between the two phases @xcite .
the study of first order transitions is very important for complex systems , especially social or ecological complex systems , because the sudden jump between two phases ( which is discontinuous ) can be quite dramatic @xcite .
while the ginzburg - landau - wilson approach has been tremendously successful in explaining universality in second - order phase transitions , it requires the definition of an order parameter .
different approaches , which do not require an order parameter , can be useful for cases where an order parameter is difficult to identify , or does not exist ( e.g. @xcite ) .
in one such approach we study the probabilistic description of the system while changing the parameters to bring the system across a transition .
the statistical properties of the system in the different phases are very different .
therefore , at the phase transition the shape of the probability distribution function will change drastically .
this is captured by the fisher information matrix ( fim ) through the cramr - rao bound @xcite . a compelling differential - geometric framework to study the changes that the probability distribution undergoes is information geometry ( ig ) @xcite . in ig the family of probability distributions that are parametrized by a set of continuous parameters ( designated here as the vector @xmath0 )
is seen as a differential manifold .
the parameters form a coordinate system on the manifold , and distances are measured by the fisher - rao metric : @xmath1 which is a positive semi - definite , symmetric matrix that changes covariantly under reparametrizations of the probability distribution @xmath2 . here
@xmath3 is a derivative with respect to one of the parameters , indexed by @xmath4 .
the ig of many models in statistical mechanics has been studied , e.g. , in .
of particular interest in these studies is the role of the scalar ( riemannian ) curvature .
it was shown to diverge at critical transition points and on the spinodal curve @xcite , thus effectively preventing geodesics from crossing into the unphysical area of phase space @xcite .
see also @xcite for a general renormalization group analysis of ig near criticality .
a connection between phase transitions in ig and the ginzburg - landau - wilson approach can be made when an order parameter @xmath5 is the derivative of a thermodynamic potential with respect to some thermodynamic variable @xmath6 .
then there exists a collective variable @xmath7 such that @xmath8 @xcite and the fisher information matrix can be shown to obey @xcite : @xmath9 where @xmath10 is the inverse temperature , @xmath11 being boltzmann s constant . at second order phase transitions in the thermodynamic limit
this derivative diverges and therefore a corresponding entry of the fim also diverges . when the system is finite , the fisher information does not diverge but rather attains a maximum . the maximum of the fisher information has been used to accurately find the phase transition point in finite systems @xcite and as a definition of criticality in living systems @xcite .
we have two main goals for the current work : first to test a conjecture set forth by prokopenko _
_ in @xcite that a divergence ( or maximization ) of the entries of the fim can detect phase transitions even in the absence of an order parameter .
second , to measure the fisher information matrix without resorting to the underlying dynamics of the system and without assuming a specific parametric model for equation .
this addresses the problem that often the microscopic dynamics of complex systems are unknown , and an analytic description of the probability density function is missing . to accomplish these two goals we chose to study the specific example of the two dimensional gray - scott ( gs ) reaction diffusion model @xcite .
we chose the gs model for its rich variety of spatial and spatio - temporal patterns @xcite and we consider the transitions between the different patterns as critical transitions . among the different types of patterns one can find self - replicating spots @xcite , spatio - temporal chaos @xcite , and labyrinthine patterns @xcite .
these were first systematically classified by pearson @xcite .
our goal is to use the fisher information matrix to construct a phase map for the gray - scott model , where we expect areas with high values of the fisher information to demarcate the different patterns . as a probabilistic description for the system we chose the blob - size distribution , which we take to be a function of the control parameters of the model @xmath12 and @xmath13 and which we estimate non - parametrically by using image processing on the resulting spatial @xmath14 concentration from our simulations .
the paper is organized as follows : in section [ sec : fisher_criticality ] we discuss the relationship between fisher information and criticality , which forms the motivation for our approach .
we revisit the arguments in @xcite and extend them to our case . in section [ sec : gs ]
we introduce the gray - scott model and discuss some of its properties . in section [ sec : results ] we present the results of the computations and in section [ sec : methods ] we explain the methods we used to compute the fisher information in this settings .
last we discuss our results in section [ sec : conclusions ] .
in this section we summarize the derivation performed in @xcite leading to eq . that relates order parameters and the entries of the fisher information matrix .
this derivation leads to the conjecture @xcite that it is enough to consider the fisher information matrix entries rather than order parameters and is presented here because it is important for our exposition .
the gibbs ensemble can be generically written in the following way : @xmath15\ ] ] with @xmath16 set by normalization and with summation convention over repeated indices , which we use throughout the paper .
for this distribution , the fisher information eq .
is @xcite : @xmath17 where @xmath18 is the gibbs free energy . performing the derivatives we obtain eq . . for many systems ,
the order parameter can be defined by introducing an external field @xmath19 to the free energy that couples to the order parameter @xmath20 @xcite .
the canonical example being the magnetization that couples to the external magnetic field , so that @xmath21 . the external field
being one of the external parameters @xmath6 .
we then have : @xmath22 setting @xmath23 for a particular @xmath24 we obtain : @xmath25 the meaning of eq .
is that if an order parameter @xmath26 is a derivative of the free energy @xmath18 , then there exists a collective variable @xmath27 whose average is proportional to the order parameter @xcite .
it is important to note that many models exist whose order parameter is indeed the average of a collective variable @xcite . combining equations and we obtain eq . .
this links the value of the entries of the fisher information with the derivatives of the order parameters @xmath5 of the system .
since at phase transitions the order parameter or its derivatives becomes non - analytic we can expect the fisher information matrix to have diverging entries at the phase transition point .
for example , at a ferromagnetic transition point , with @xmath28 , the diagonal elements of the fisher information matrix are given by @xcite : @xmath29 since @xmath30 and @xmath31 are proportional to the magnetic susceptibility @xmath32 and heat capacity @xmath33 respectively , we expect both entries to diverge at the point of the magnetic second order phase transition @xcite . more generally , it is easy to show that @xcite @xmath34 this is also called the generalized susceptibility in statistical mechanics @xcite .
this derivation led prokopenko to propose that the maximization of the appropriate fisher information matrix can detect phase transitions , without explicitly defining an order parameter @xcite .
the relation suggests the introduction of an order parameter derived from the fisher information by integrating it from one phase to the next , in the following way : @xmath35 we absorbed the inverse temperature in the definition of @xmath0 , and the integration path starts at @xmath36 which is in one phase and ends at @xmath37 which is in the other phase . in general
this will depend on the integration path and the end - points .
while the derivation above assumes the form for the probabilistic description of the system , the idea of fisher information maximization at phase transition points can be generalized for other probabilistic descriptions based on the cramr - rao bound @xcite that states that the variance an unbiased estimator @xmath38 is bounded from below by the inverse of the fisher information .
we make the following heuristic argument : when a system is said to undergo a phase transition , it means that there is an observable change in some aspect of the system ( often having to do with the symmetries of the system ) .
this means that the statistical properties of the system in the two phases differ significantly .
for example , if we sample the energy per spin of an ising spin system repeatedly in the high - temperature phase , we will obtain a broad distribution of energies .
conversely , at the low - temperature phase the energy distribution is very narrow , since in the low - temperature phase the spins are aligned and the system is in the ground state @xcite .
thus , if we look at the probability density function describing these observables change as a function of the control parameter , it undergoes a drastic change in its functional form .
this , in turn , implies that we can estimate the value of the control parameter at the phase transition point accurately , because of the large change in the behavior of the density function . according to the cramr - rao inequality , the inverse of the value of the fisher information serves as a lower bound on the variance of the estimated parameter .
if this parameter can be estimated accurately then this implies a high value of the fisher information .
we therefore surmise that under very general circumstances the fisher information is maximized at phase transition points .
the gray - scott model is a non - linear reaction - diffusion model of two chemical species @xmath39 and @xmath14 with the reactions @xmath40 @xmath39 is constantly supplied into the system and the inert product @xmath41 removed . we can simulate the reaction using the law of mass action @xcite , where we assume that the rate of each reaction is proportional to the concentration of the reactants at each point .
the resulting non - linear coupled differential equations are : @xmath42 where @xmath43 , @xmath44 are the ( dimensionless ) concentrations of the two chemical species , @xmath45 is the laplacian with respect to @xmath46 , @xmath47 and @xmath48 are diffusion coefficients of @xmath49 and @xmath50 respectively , @xmath12 represents the rate of the feed of @xmath39 and the removal of @xmath39 , @xmath14 and @xmath41 from the system and @xmath13 is the rate of conversion of @xmath14 to @xmath41 . in practice
this is a model of chemical species in a gel reactor where the rate @xmath12 can be relatively easily modified , and @xmath13 is dependent on the temperature of the system . @xmath47 and @xmath48 are more difficult to change and we will consider them constant , with @xmath51 .
we start with the standard stability analysis for the homogeneous system ( i.e.without diffusion ) .
the gray - scott model has a trivial homogeneous steady state solution , referred to as the red state , at @xmath52 = [ 1,0]$ ] which is always linearly stable for positive @xmath12 and @xmath13 @xcite . under the condition that @xmath53 two additional homogeneous steady states solutions appear : @xmath54 = \left[\frac{1}{2}(1-\sqrt{d } ) , \frac{1}{2}\frac{f}{f+k}(1+\sqrt{d})\right]\ ] ] @xmath55 = \left[\frac{1}{2}(1+\sqrt{d } ) , \frac{1}{2}\frac{f}{f+k}(1-\sqrt{d})\right].\ ] ] these are referred to as the blue state and the intermediate state respectively @xcite .
linear stability analysis shows that when these states exist , the intermediate state is always unstable whereas the blue state can be stable .
the two states appear through a saddle - node bifurcation which in the @xmath56 plane is defined by the curve : @xmath57.\ ] ] under certain conditions @xcite the blue state undergoes a hopf bifurcation at @xmath58.\ ] ] for more details see @xcite .
the bifurcation curves are plotted in fig .
[ fig : bifurcation ] together with the area where we performed our simulations . .
] in the vicinity of the bifurcations a variety of inhomogeneous patterns may appear @xcite .
these can be observed by setting the initial state of the system to the red state and adding a finite perturbation that allows the system to reach a different attractor ( for details see sec .
[ sec : methods ] ) .
the first to systematically study these patterns in two dimensions was pearson @xcite .
pearson classified the patterns in @xmath59 types and designated them with greek letters . in this paper
we use the same parameters pearson used for his simulations , except that we will vary the simulation lattice size to include larger patterns , as will be described later .
the different patterns appear close to each other in the @xmath56 space and often mix on the boundaries of regions of different patterns .
one of the patterns that was first discovered by pearson @xcite is that of the self - replicating spots ( e.g. fig .
[ fig : bacteria ] ) .
a single spot of high @xmath14 with a well defined boundary grows until at a certain point it will split into two spots .
the process continues until the whole simulation area is covered with spots @xcite .
depending on the parameters , the self - replicating spots will either reach an asymptotic fixed point or will start to flicker in what is known as spatio - temporal chaos @xcite .
wang and ouyang @xcite derived a probabilistic description of the spot count in the chaotic regime as a function of the rate of spot creation and annihilation which were found to linearly depend on the spot - count .
another group of patterns are the worm - like patterns ( fig .
[ fig : worms ] ) which grow and fill the entire simulation lattice and then are fixed . in the parameter space , between the stable self - replicating spots and the worm - like patterns ,
a mixed pattern appears which contains both spots and stripes .
this pattern ( fig .
[ fig : mixed_phase ] ) is reminiscent of a first order phase transition , where phase co - existence appears in the transition region .
the degree of mixing gradually increases and then decreases again across the transition . + in the examples in fig .
[ fig : patterns ] , ( a ) , ( b ) , ( c ) and ( f ) patterns are at a fixed point of the dynamics .
the patterns ( d ) and ( e ) , on the other hand , are a snapshot of patterns that do not reach a fixed - point .
the patterns that appear in the gray - scott are examples of self - organization because they are macroscopic patterns that appear through the local interactions of the @xmath39 and @xmath14 substances @xcite . in order to construct a phase map using the fisher information matrix , we follow wang @xcite in constructing our probabilistic description of the patterns .
wang treated each spot as an entity and computed the probability for a number of spots to appear at a unit time .
we separate the pattern into similar entities , but regard not only the self - replicating spots but also the stripes and the labyrinthine patterns as entities , which we call `` blobs '' .
we take the lattice of @xmath14 values ( the @xmath39 value lattice is usually very similar to the @xmath14 one ) and treat it as an image . to identify a blob we first binarize the image using otsu s method@xcite .
we use the ` find_blobs ` method from the image processing python library ` simplecv ` to label continuous clusters in the binarized image .
we extract the size of each blob ( in pixels ) and use a non - parametric estimation procedure to obtain a pdf of the sizes of the blobs .
we assume that the blob sizes are characteristic of the different patterns and that by observing their distribution we could , with a high degree of certainty , deduce the type of pattern .
we therefore expect the fisher information to become large when a pdf changes rapidly and therefore we will be able to detect the transitions between the patterns .
the density extraction process is depicted in fig .
[ fig : extraction_process ] .
+ we assume that the extracted pdf is a function of the parameters @xmath12 and @xmath13 and that we can therefore compute derivatives of it with respect to these parameters by using a finite difference scheme .
we use a centered finite difference scheme and the integration is performed by use of ` scipy.integrate.quad ` .
the expression we use for the fisher information is : @xmath60 in this expression @xmath61 represents a small increment in either @xmath12 or @xmath13 , keeping the other fixed ( so , for example , @xmath62 is interpreted as the probability density at @xmath63 ) . following @xcite we set the expression under the integration to zero whenever any of the densities was zero for a given @xmath64 .
an extended discussion about the use of non - parametric density estimation ( such as kernel density estimations ) and using finite difference schemes in the computation of the fisher information is given in @xcite . since
our premise is that the distribution of blob sizes is typical to the different patterns , we use the shannon entropy , defined as : @xmath65(\theta ) \equiv \int p(x;\theta ) \ln \frac{1}{p(x;\theta ) } dx.\ ] ] as a way to validate our method .
the shannon entropy is a measure of uncertainty in the outcome of a random variable which is distributed according to @xmath2 . in our case
it represents our uncertainty about the size of one of the blobs in the pattern . in the stable spots pattern , for example
, the uncertainty is relatively low and so is the shannon entropy . at the stripe pattern it is high , since any given blob can belong to a short or a long stripe .
this gives us a sort of an order parameter which , however , is not necessarily related to a change in the symmetry of the system .
this allows us to compare the predictions from the fisher information to that of the shannon entropy .
we will typically expect a line of high fisher information where the shannon entropy undergoes a large change .
to get an overview of where in parameter space each patterns is located , we first plot the shannon entropy of the blob - size pdf .
the shannon entropy map for a simulation grid of @xmath66 is shown in fig .
[ fig : entropy_patterns ] .
the shannon entropy shows a clear demarcation between different regions in the parameter space near the saddle - node bifurcation curve .
there are two areas with low entropy ( blue regions , one above and one below the saddle - node curve ) .
both blue areas have stable localized spots of roughly the same size , but only the spots in the area containing the ( a ) pattern are created through self replication .
the other spots are created with a different mechanism , see the supplemental material at for movies showing the evolution of these two patterns .
when we move from the area containing ( a ) to the area containing ( d ) there is a well defined boundary where the blue region ends and where the spatio - temporal chaotic regime begins .
the shannon entropy increases slowly as one moves towards lower values of @xmath12 and @xmath13 since the size of the spots becomes less certain and since half formed spots are counted as blobs with smaller and more variable areas .
we refer to this transition as `` transition i '' .
a second interesting transition occurs when moving from ( a ) to ( b ) . in this transition
the spots from pattern ( a ) mix with the stripes of pattern ( b ) to a varying degree , until there are no more spots and only stripes remain .
as we can see the shannon entropy increase is much steeper in this transition , since the blobs turn into stripes of wildly different lengths .
we will refer to this transition as `` transition ii '' .
examples of transitions i and ii can be seen in figs .
[ fig : trans1 ] and [ fig : trans2 ] respectively
. grid .
the location of the different patterns from fig .
[ fig : patterns ] are indicated with @xmath67 .
the color - bar is in logarithmic scale .
gray indicates areas with no inhomogeneous patterns.,scaledwidth=45.0% ] the different components of the fisher information matrix and its trace and eigenvalues are plotted in figure [ fig : fisher ] .
at each point we computed the eigenvalues and plotted separately the larger eigenvalue ( fig .
[ fig : g_ev1 ] ) and the smaller one ( fig .
[ fig : g_ev2 ] ) . + the main features of the fisher information maps we see in fig .
[ fig : fisher ] are that there are many curves with high values of the fisher information ( `` ridges '' ) which separate areas of lower fi .
most ridges follow the general curve of the saddle - node bifurcation line , except for the noticeable ridge separating the self - replicating spots pattern region ( a ) and the spatio - temporal chaotic area ( d ) ( i.e. the ridge representing transition i ) .
almost all versions of the fisher map show this ridge , but with a varying degree of clarity .
transition ii is also clearly represented by a ridge which appears in almost all representations of the fim ( except for the @xmath68 component ) .
this is the rather wide ridge where pattern ( c ) resides .
we constructed the two dimensional phase map by simulating the different patterns that appear in all the different parameter values we were interested in .
often we do not have complete knowledge of the surrounding patterns and only observe one instance of the system at a time . in these cases , it can be useful to look at a one dimensional plot of the fisher information along a given trajectory .
this can happen , for example , when we observe a natural system over time . then we could treat the observation time as the parameter @xmath0 and compute how much we know about the observation time from the distribution @xmath69 . as an example we plot the shannon entropy and the fisher information interval along two trajectories .
one crossing transition i and the other transition ii .
we define the fisher information interval in analogy with special relativity @xcite : @xmath70 it is an invariant measure that represents the squared distance along the path described by @xmath71 . to compute the interval we first defined the start and end of the path @xmath72 and @xmath73 respectively , then computed the unit vector connecting @xmath74 and @xmath75 : @xmath76 . and finally the vector @xmath77 .
@xmath78 and @xmath79 are the lattice spacings in parameter space and @xmath80 is the usual norm in @xmath81 . the first one dimensional transition we analyze is transition i , which we define between the stable , self - replicating , spots and the chaotic spots .
one such trajectory crossing the transition is plotted in fig .
[ fig : trans1 ] .
point @xmath82 ( at parameter values @xmath83 ) is located within the fixed spots area close to the transition and point @xmath84 ( at parameter values @xmath85 ) is located within the chaotic spots area , also not far from the boundary . on the top left of fig .
[ fig : trans1 ] we plot the fisher interval going from point @xmath82 to point @xmath84 . at point
@xmath86 the interval starts increasing and at point @xmath87 it reaches a maximum , after which is slowly decreases again . on the same graph we plot the shannon entropy of the blob size pdf .
the shannon entropy is lower in the stable phase than in the spatio - temporal chaos , as can also be seen in fig .
[ fig : entropy_patterns ] .
this is due to the increased uncertainty in blob sizes in the chaotic regime .
we inspected the patterns on both sides of the transition visually and as a function of time and it does indeed seem to capture the transition in the correct location . to verify this quantitatively
we define an order parameter . since the nature of the transition is dynamic ( i.e. the time dependence of the pattern is different in both phases ) we simulate the system further in time .
we divide the trajectory from @xmath82 to @xmath84 to @xmath88 points at each point run the simulation for @xmath89 warm - up time steps for it to reach the state in which we computed the fisher information .
we then continue the simulation an additional @xmath90 time steps , computing the blob - count every @xmath91 time steps . from these samples
we compute the standard deviation of the blob count .
the rationale is that for the stable spots the variability of blob count should be zero and at the chaotic regime it is non - zero .
this is plotted on the bottom left in fig .
[ fig : trans1 ] , below the fisher interval plot .
we also plotted vertical lines that indicate the position of points @xmath86 and @xmath87 . because of the demanding computation time we performed the validation run on a @xmath92 grid as opposed to the @xmath66 grid from which the fisher interval is computed .
the blob - count indeed remains constant between @xmath82 and @xmath86 .
it then gradually becomes more variable as we go deeper into the chaotic regime .
the increase in blob - count variability coincides with the rise in fisher interval which we interpret as a verification that indeed the transition begins at this point .
the second trajectory we chose crosses what we term transition ii . as the first transition , it starts in the self - replicating spots area ( at a different point @xmath82 which is located at parameter values @xmath93 and @xmath94 ) and crosses to the area of the stripe patterns to point @xmath84 which is located at @xmath95 . along the trajectory
the patterns evolve from completely spots to a mixed spot - stripe region and finally reach a region without any spots . again
we plot the trajectory in fig .
[ fig : trans2 ] , along with the fisher interval and shannon entropy ( top left ) .
the fisher interval clearly increases between points @xmath86 and @xmath87 which this time designate visually selected points of the start and end of the region where the fisher interval is high .
the shannon entropy again increases across the transition and is higher in the stripe phase in comparison with the self - replicating spot phase . below the fisher interval and shannon entropy plots
we draw a part of the patterns that appear at points @xmath82 through @xmath84 .
this helps to provide a visual verification of the position of the transition .
point @xmath86 is the first point with stripes appearing together with the spots and point @xmath87 is one of the last with spots .
it is tempting to compare the shannon entropy and fisher information in terms of how well they capture the transition .
both can be used as indicators , the shannon entropy indicates a transition when its average value changes significantly and the fisher information by rise and subsequent decrease in its value .
we would like to note that , following the discussion earlier , the fisher information acts as a susceptibility measure ( as the magnetic susceptibility would act in an ising model ) and the shannon entropy as an order parameter ( as the magnetization would in an ising model ) .
we can however hypothesize that it is possible to find a transition between patterns in which the shannon entropy is constant but the fisher information peaks .
for example , if the average blob size changes abruptly but the variance in blob sizes remains fixed . in that case
the shannon entropy would not catch the transition but the fisher information would .
the size of the simulation grid , which we varied in our experiments from @xmath96 up to @xmath97 has a large effect on the resulting phase map .
small grid sizes result in a fairly noisy fisher map , whereas larger ones give a smoother map with better defined `` ridges '' which demarcate the different phases .
we suspect this is a result of too few blobs appearing in the smaller grids . because simulating large grid sizes requires considerable computation time ( the @xmath97 took about a month on a cluster running on approximately @xmath98 cores simultaneously )
, one has to find a balance between accuracy and computational time .
we found that a grid size of @xmath66 provided good results . in order to compare the results obtained for each of the grid sizes , we plot the trace of the fisher information matrix obtained from the computation of the different grid sizes
this is presented in fig .
[ fig : grid_sizes ] . at @xmath96
there is quite some noise from large peaks of the fisher information .
the @xmath99 grid already provides a much better resolution of the ridges , which improves even more at the @xmath66 grid .
the @xmath100 grid also shows the highest contrast in fisher information between `` ridge '' and `` valley '' regions .
as we go to the largest grid size we used , @xmath97 , it seems that some of the well - defined ridges become less well defined .
we suspect that this might happen because the system did not have sufficient time to reach equilibrium , even though each simulation was run for @xmath101 time steps .
the fisher information is obtained by performing simulations on the parameter range where inhomogeneous patterns appear in the gray - scott model .
the parameter space was divided into an evenly spaced @xmath102 grid with @xmath103;\ ; k \in [ 0.04 , 0.07]$ ] .
the diffusion coefficients were held constant at @xmath104 such that @xmath105 .
we performed a simulation at each point of parameter space , starting with identical initial conditions ( same seed ) and repeated for the same number of time steps ( depending on the simulation grid size ) .
the simulation was started with an initial condition of the red state @xmath106 with a finite perturbation in the form of a @xmath107 square in the center of the simulation grid in the state @xmath108 and an additional gaussian noise with an amplitude of @xmath109 covering the entire simulation grid .
this initial state was then evolved by integrating numerically eq .
using an euler scheme until the final state was reached .
we repeated the experiment with different simulation grid sizes , ranging between @xmath110 up to @xmath111 .
the simulation times ranged from @xmath89 time steps ( for the smallest grid sizes ) to @xmath101 for the @xmath97 grid .
this was chosen such that the self - replicating stable spots will fill in the entire simulation window .
the simulations were performed on the lisa cluster run by surfsara .
the python code to perform the simulation is based on the code found at . for each simulation we extracted the pdf by following the steps described in section [ sub : probabilistic_description ] .
we used the python package ` simplecv ` for the binarization and blob detection of the images and ` scipy.stats.gaussian_kde ` for the computation of the pdf from the blob sizes .
the computation of the fisher information from the pdf followed the description in @xcite and the code for this computation is available online at .
as mentioned in sec .
[ sec : results ] we also computed the shannon entropy for each pdf we obtained .
this was done by simple integration of the pdf using eq . and the python function ` scipy.integrate.quad ` .
in addition to the gaussian kde , we used the novel density estimation method deft @xcite , and tried two different ways to integrate eq . - once as it is written in eq . and once by first performing the differentiation of the logarithm ( replacing @xmath112 with @xmath113\partial_\mu p$ ] ) before defining the finite difference scheme . as is described in @xcite , the first method yields better results for gaussian kde and the second for deft .
we eventually used the results obtained from the gaussian kde rather than the deft ones because finding the correct parameters for deft for all parameter values was difficult , since both the range of blob sizes and the number of blobs in each simulation in each parameter value varied too much .
in this paper we explore the use of the fisher information matrix to capture transitions between different patterns in the gray - scott model .
the use of fisher information for this purpose is inspired by an analogy with phase transitions in thermodynamic systems and follows the work by prokopenko _
et al . _
the main purpose of the study was to test whether such a description is feasible in a system whose probabilistic description is not derived from the microscopic dynamics and where no statistical model is assumed .
we find that at least in the case of the gs model with the choice of blob - size pdf as a probabilistic description , this seems to be indeed possible . the main difficulty in using this approach to produce the entire phase map is that it is computationally very demanding .
the very smooth phase map only really appears at grid sizes of @xmath66 which requires long computation runs .
a more conceptual difficulty is that once we obtain the maps , their exact interpretation is not trivial .
the value of the fisher information ranges over many orders of magnitude ( between @xmath114 and @xmath115 ) and there is no theoretical value to compare with .
the strength of the method lies in that it manages to capture various types of transitions with the same metric .
we did not have to define specific order parameters for the different patterns .
it also captures our intuitive understanding of a `` phase transition '' as a large change in the statistical properties of a system when the underlying `` external '' parameters are changed . in the one - dimensional cases
, it seems that the fisher information does indeed capture essential features of `` pattern transitions '' in that it gives a clear signal in the form of a peak between two regions of low fisher information .
it is especially interesting to note that the fisher information predicts an exact position for the transition to spatio - temporal chaos , as depicted in fig .
[ fig : trans1 ] , since this transition is dynamic in nature .
further mathematical analysis similar to the one done in @xcite for the two - dimensional gray - scott model might be able to confirm the exact position of the transition line and compare it to the one detected by the fi .
our results suggest that fisher information can indeed serve as a generalized susceptibility measure in the study of complex systems .
because very few assumptions about the underlying dynamics are necessary it is an ideal tool for detecting different regimes even in the absence of an assumed statistical model .
ohs would like to thank lisa jenny krieg from the university of amsterdam for many enlightening discussions and the critical reading of the manuscript .
the research leading to these results has received funding from the european union seventh framework programme ( fp7/2007 - 2013 ) under grant agreement numbers 317534 and 318121 .
rq and pmas acknowledges the financial support of the future and emerging technologies ( fet ) programme within the seventh framework programme ( fp7 ) for research of the european commission , under the fet - proactive grant agreement topdrim , number fp7-ict-318121 .
agh wishes to acknowledge partial funding by the russian scientific foundation , under grant # 14 - 11 - 00826 .
pmas wishes to acknowledge partial funding by the russian scientific foundation , under grant # 14 - 21 - 00137 .
53ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1088/0034 - 4885/50/7/001 [ * * , ( ) ] @noop _ _
( , , ) @noop , ( ) @noop _ _ ( , , ) link:\doibase 10.1016/0370 - 1573(74)90023 - 4 [ * * , ( ) ] link:\doibase 10.1073/pnas.1414708112 [ ( ) ] link:\doibase 10.1063/1.1665530 [ * * , ( ) ] link:\doibase 10.1098/rsta.1922.0009 [ * * , ( ) ] @noop _ _ ( , , ) @noop * * , ( ) http://books.google.com/books?hl=en{&}lr={&}id=euhbluw31hsc{&}pgis=1[__ ] ( , ) p. @noop _ _ ( , ) http://scholar.google.com/scholar?q=related:ksdofjta768j:scholar.google.com/{&}hl=en{&}as{_}sdt=0,5{#}1 http://scholar.google.com/scholar?hl=en{&}btng=search{&}q=intitle:information+thermodynamics+and+differential+geometry{#}0 [ ( ) ] http://scholar.google.com/scholar?q=related:ksdofjta768j:scholar.google.com/{&}hl=en{&}as{_}sdt=0,5{#}0 http://scholar.google.com/scholar?hl=en{&}btng=search{&}q=intitle:information+geometry+of+quantum+statistical+systems{#}0 [ ( ) ] link:\doibase 10.1103/physreva.39.6515 [ * * , ( ) ] link:\doibase 10.1088/0305 - 4470/23/4/016 [ * * , ( ) ] link:\doibase 10.1088/0305 - 4470/23/4/017 [ * * , ( ) ] link:\doibase 10.1103/physreva.20.1608 [ * * , ( ) ] link:\doibase 10.1103/physreva.24.488 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.67.605 [ * * , ( ) ] link:\doibase 10.1103/physreva.41.2200 [ * * , ( ) ] link:\doibase 10.1103/physreve.86.052103 [ * * , ( ) ] http://pre.aps.org/abstract/pre/v51/i2/p1006{_}1 [ * * , ( ) ] link:\doibase 10.1016/s0393 - 0440(02)00190 - 0 [ * * , ( ) ] link:\doibase 10.1088/1751 - 8113/42/2/023001 [ * * , ( ) ] link:\doibase 10.1016/j.physa.2004.01.023 [ * * , ( ) ] link:\doibase 10.1103/physreve.86.051117 [ * * , ( ) ] link:\doibase 10.1103/physreve.92.052101 [ * * , ( ) ] , http://www.ncbi.nlm.nih.gov/pubmed/22181096 [ * * , ( ) ] link:\doibase 10.1162/artl{_}a{_}00041 [ * * , ( ) ] link:\doibase 10.1073/pnas.1319166111 [ * * , ( ) ] \doibase doi : 10.1016/0009 - 2509(84)87017 - 7 [ * * , ( ) ] link:\doibase 10.1126/science.261.5118.189 [ * * , ( ) ] http://journals.aps.org/prl/abstract/10.1103/physrevlett.72.2797 [ * * , ( ) ] http://www.sciencedirect.com/science/article/pii/s0167278900001834 [ * * , ( ) ] http://www.sciencedirect.com/science/article/pii/s0167278902007431 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.214102 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1080/00018730050198152 [ * * , ( ) ] , link:\doibase 10.1016/0378 - 4754(95)00044 - 5 [ * * , ( ) ] link:\doibase 10.1016/s1468 - 1218(03)00020 - 8 [ * * , ( ) ] link:\doibase 10.1016/s0167 - 2789(99)00010-x [ * * , ( ) ] link:\doibase 10.1016/s0167 - 2789(00)00214 - 1 [ * * , ( ) ] link:\doibase 10.1002/cplx [ * * , ( ) ] http://web-ext.u-aizu.ac.jp/course/bmclass/documents/otsu1979.pdf [ * * , ( ) ] http://arxiv.org/abs/1507.00964 [ ( ) ] , @noop _ _ ( , , ) link:\doibase 10.1103/physreve.90.011301 [ * * , ( ) ] | the fisher - rao metric from information geometry is related to phase transition phenomena in classical statistical mechanics .
several studies propose to extend the use of information geometry to study more general phase transitions in complex systems .
however , it is unclear whether the fisher - rao metric does indeed detect these more general transitions , especially in the absence of a statistical model . in this paper
we study the transitions between patterns in the gray - scott reaction - diffusion model using fisher information .
we describe the system by a probability density function that represents the size distribution of blobs in the patterns and compute its fisher information with respect to changing the two rate parameters of the underlying model .
we estimate the distribution non - parametrically so that we do not assume any statistical model . the resulting fisher map can be interpreted as a phase - map of the different patterns .
lines with high fisher information can be considered as boundaries between regions of parameter space where patterns with similar characteristics appear .
these lines of high fisher information can be interpreted as phase transitions between complex patterns . |
the star ab aurigae ( hereafter ab aur ; distance @xmath0 parsecs ; visual magnitude , @xmath1 ; spectral type a0ve ; mass @xmath2 m@xmath3 ; age @xmath4 myr ; also known as hd 31293 , hip 22910 , bd+30 741 ; @xcite ) has been intensively studied and is widely perceived as prototypical of young , intermediate - mass stars , the so - called herbig ae stars .
the star belongs to the taurus - auriga star - forming region , which is thought to be between 1 and 3 myr old , where the dispersion in ages is most likely due to observational or calibration errors ( * ? ? ?
because of its brightness , age and proximity to earth , this star presents an important opportunity for investigating star and perhaps planet formation .
indeed , the star exhibits a fascinating array of phenomena , including a reflection nebula extending thousands of astronomical units ( au ; the earth - sun distance ) away from the star @xcite ; a more compact , annular ( 100 - 750 au ) region mainly composed of gas as detected in @xmath5co , @xmath6co , c@xmath7o , hco@xmath8 , h@xmath9 , and millimeter wavelength continuum imaging ; and an apparent inner region within about 100 au of the star which exhibits gas depletion , where the ratio of the mass in dust to that in gas is large ( based on the surface densities derived by ) .
it seems that the dusty inner disk has an outer radius at about 130 au @xcite , outside of which the gas column density rises .
also , near infrared ( nir ) imaging reveals a spiral structure outside of the inner disk ( from 130 to a few hundred au ; @xcite ) .
ancillary to this study , but of importance in understanding this system , the material also exhibits a central hole of radius between 0.3 and 0.6 au @xcite .
a bright spot has also been detected about 1 - 4 au away from the star @xcite with nir interferometry .
interpretations of these observations are inconsistent with each other in the literature , with some authors describing disk - like structures and others not .
the system is clearly complicated , and it may have properties of both gas - rich and debris disks , based on qualitative consideration of the observations mentioned above . as such , it may be in a stage of evolution that is intermediate between these two states , and it presents an ideal observational opportunity to study disk evolution and the formation of small bodies in such disks .
this means that high resolution observations of the region between 1 and 130 au are particularly important .
however , this region has been difficult to image in light scattered from the dust because of the extreme brightness difference between the star and the dust .
we present the first images of polarized light from this region in the nir with sufficient contrast ( residual starlight after suppression of @xmath10 at radii of 0.3 to 2.2 arcseconds ) and resolution ( 14 au ) to resolve structure in the dust disk .
this level of starlight suppression was achieved using polarimetry in conjunction with high - order adaptive optics ( ao ; for spatial resolution ) and advanced coronagraphy ( for starlight suppression ) under the aegis of the lyot project @xcite .
scattered light from the gas and dust in the vicinity of a star should be linearly polarized with the electric field perpendicular to the plane of incidence as seen from the telescope ( assuming a single scattering process ) .
thus , dust in close proximity to a star will , in general , induce a spatially variable polarization angle and fractional polarization as seen in a polarized light image .
this makes imaging polarimetry of stars a valuable tool for studying star formation .
in fact , numerous imaging polarimetry observations of disks have been made , permitting the observation of extremely faint sources in close proximity to stars , because starlight can be efficiently removed by forming a polarized light image .
indeed , direct light from the star is only minimally polarized by small - angle ( @xmath11 degrees ) forward scattering due to the earth s atmosphere or the interstellar medium .
in contrast , light from the star that has scattered off of material in the near - star environment is imparted with far more polarization due to the large angle of scattering ( near 90@xmath12 for a face - on disk ) .
thus by differencing images obtained in orthogonal polarization states ( _ i.e. _ @xmath13 and @xmath14 , @xmath15 and @xmath16 , or @xmath17 and @xmath18 , using the standard stokes vector notation ) the star light is removed , and an image of polarized light , a `` @xmath19 image , '' can be formed where @xmath20 . the spatial variation of the p image is a sensitive measure of variations in the scattering properties of the near - stellar dust and gas and is insensitive to atmospheric speckle noise .
a further discussion can be found in the appendix of this work .
the series of images acquired ( six sets of two simultaneous images , each with different polarimeter settings ; see appendix ) are also added together to form a separate total intensity or `` @xmath21 '' image .
@xcite describe a double - difference method that minimizes instrumental effects in imaging polarimetry . to derive our @xmath19 images ,
following the @xcite technique , we take differences of simultaneous orthogonal polarization images ( _ e.g. _ @xmath13 and @xmath14 ) .
we then difference these images again after spatially interchanging the polarization states @xcite and then sum them in quadrature to form a `` @xmath19 '' image @xmath54 instrumental and other systematic polarization effects are removed or tend to be spatially constant in this polarized image . to achieve this , a wollaston prism located immediately
after the coronagraph s lyot stop splits the light into two images , viewed simultaneously side - by - side on the kermit s detector , with the two images carrying perpendicular vectors of the polarization .
using liquid crystal variable retarders ( calibrated with a glann - taylor prism ) , the instrument can swap the vectors for each field of view or rotate them . as a result
, we obtain six sets of two images .
the two images are recorded on the same detector simultaneously and with the same exposure time . by swapping the vectors on the two images in a subsequent exposure
the two exposures comprising a `` set''we can remove instrumental effects induced by the different optical paths of the two simultaneous images .
taking the first two exposures , for example , we subtract the left and right images in both and then subtract the resulting final image .
this provides an image in polarized light that we call `` q , '' with `` u '' and `` v '' images derived with the subsequent exposures in the same way . these q , u and v images are from the reference frame of the instrument . in order to derive the q , u , v images in the reference frame of the sky , they have to be reprojected taking into account the mueller matrix of the telescope and instrument as a function of pointing in the sky .
however , this mueller matrix is not only extremely complex , but is also unknown @xcite .
fortunately , the p image is immune to these effects , because it is the quadrature sum of the q , u and v images , but this `` cross - talk '' induced by the telescope requires that we also measure v , which is often excluded in observations similar to these .
our @xmath19 image does not directly reveal the absolute polarization amplitude or direction of the near - star optical scattering , because the intensity image has significant spurious scattered starlight and atmospheric speckle noise ( * ? ? ?
in addition , as mentioned above , the aeos telescope changes both the magnitude and polarization angle of incoming light to the telescope @xcite . fortunately this spurious polarization cross - talk is essentially constant over the telescope field - of - view , meaning that the relative structure in the @xmath19 image is not instrumental in nature .
the observed ab aurigae q , u and v images were all very similar in nature with small relative rotations in the polarization angle and degree of polarization as the telescope changes pointing . however , the quantity @xmath55 was constant during all ten polarization sequences ( 6 dual images for each ) . despite these limitations ,
the spatial variation of the @xmath19 image is a sensitive measure of variations in the scattering properties of the near - stellar dust and gas and is insensitive to atmospheric speckle noise .
over a period of one hour and thirty - eight minutes , commencing at 09:55 on 2006 december 12 utc , we acquired ten polarimetric measurements in h - band ( 1.45 - 1.83@xmath22 m ) with the lyot project coronograph installed on the 3.63 m aeos telescope atop haleakala , maui .
each measurement is a series of six dual - beam polarized images ( 60-s exposures ) with different liquid - crystal retarder settings ( see appendix for details ) , to permit derivation of the p image .
the weather at the observatory was photometric , with winds of less than 8 km / h and relative humidity below 12% .
the local fried parameter , @xmath23 , a measure of the strength of turbulence in the atmosphere above the observatory , spanned the range of 15 to 32 cm , indicating ideal conditions for ao observations at aeos .
aeos has a 941-actuator ao system which suppresses imaging distortions due to the atmosphere @xcite .
the coronagraph @xcite is coupled with a nir camera called `` the kermit '' @xcite , with which we image the 4.7 arcsecond field of view centered on the star at a pixel scale of 13.7 mas pixel@xmath24 .
the diffraction - limited imaging resolution of the system is approximately 94 mas in @xmath25-band .
used here as an experimental `` control '' observation , data on hd 107146 ( a @xmath26 g2v star ) were taken in a manner identical to that of ab aur commencing at 15:14 on 2006 december 12 utc , with a similar range of weather conditions and telescope pointing relative to the horizon .
we observed this star because it has a known debris disk , although the likelihood of our detecting it was extremely low , because the disk is optically very thin and has only been imaged outside of the lyot project field of view @xcite .
furthermore , a lack of infrared excess at 18 @xmath22 m indicates a central hole at least as wide as 31 au ( 1 arcsec ) in radius @xcite .
as expected , hd 107146 s disk was not detected within our field of view ( within 67 au ) .
further discussion of these points can be found in the appendix .
calibration of the system includes standard dark current and flat - fielding gain images ( acquired using incandescent lamps each night ) , bad pixel maps @xcite , and observations of binary stars with well - known orbits to ascertain pixel scale and image rotation fiducials .
the appendix to this work describes the steps in more detail .
1 shows the @xmath21 and @xmath19 images of both hd 107146 and ab aur .
the structure we see in these images is better analyzed after filtering the image with a median pixel selection over a box matched to half the instrument s diffraction limit ( 3.5 pixels , rounded up to 5 ) .
oversampling in the image plane aids speckle removal in postprocessing , but it can dilute real signals in final analyses . the median filtered @xmath19 image of ab aur is shown in fig . 2 .
before analyzing the structure in this image , we must establish quantitative sensitivity limits .
according to a technique we developed previously @xcite , we have derived the difference in magnitude between the occulted star and a 3-@xmath27 point source as a function of position in the @xmath21 images for hd 107146 , to evaluate the sensitivity of the ab aur data without the presence of its disk , or any other source in the field of view .
, a so - called `` dynamic range plot '' @xcite , shows the difference in magnitude ( @xmath28 ) between the occulted star and a point source that would be detected at the 3-@xmath27 level in our @xmath21 image , as a function of radial distance from the star .
the photometric values were calibrated to unocculted images of ab aur taken immediately prior to the occulted images .
an extension of this technique was applied to the polarimetric data to derive the sensitivity limits for various fractional polarizations . for a real source in our data ,
only a lower limit to the fractional polarization is measurable , because of the contamination from the starlight throughout the image ( see appendix ) .
the images of the ab aur system ( figs . 1c , 1d and 2 ) exhibit several important features . in the @xmath21 image
, starlight dominates the entire field of view ( fig .
1c ) and shows some elongation along a position angle ( pa ) of 321 @xmath29 1@xmath12 .
this elongation , which is purely instrumental , and apparent in both the ab aur data and the hd 107146 i -image , is due to a combination of the wind s effect on ao image improvement ( * ? ? ? *
e.g. ) , and some residual image motion compensation errors .
( the wind had an average prevailing direction of 320@xmath12 during these observations . ) in the @xmath19 image ( fig .
1d ) , light is detected throughout the field of view but is mostly constrained to a region 2.1 arcseconds ( 302 au ) in diameter centered near the star . outside of the central region ,
the detection is of low significance , but does exhibit some amplification to the ne , which is where the brightest `` spiral '' feature has been seen in other studies @xcite . in the inner region
significant structure is present .
the photocenter of the inner disk is shifted 88 mas ( 12 au ) along a pa of 336 @xmath29 2@xmath12 .
the outer annulus of material in this inner disk shows a distinctly non - uniform distribution in the azimuthal direction , with a deep depletion of polarization centered 700 mas ( 102 au ) away from the star at a pa of 333 @xmath29 2@xmath12 . in a study conducted at the subaru telescope @xcite , there appears to be a similar depleted region in this outer annulus , although it is right on the edge of their coronagraphic mask .
it is important to note that none of the structure seen in the @xmath19 image correlates with structure in the @xmath21 image , reinforcing the fact that only lower limits to the fractional @xmath19 can be determined .
for example , the @xmath21-image elongation is at least 15 degrees away from the depleted region s azimuthal angle in the outer annulus .
the residual @xmath30 amplitude in hd 107146 was about 1% .
in contrast the @xmath30 structure we observe in ab aur was between 1 and 9% significantly larger than the residual systematic polarization noise we measure in the non - detection around hd 107146 . in the center of the depleted region
, there appears to be a 2.8-@xmath27 amplification of polarization fraction spanning a seven - pixel diameter circular aperture ( @xmath31 , or 94 mas ) .
this appears to be an unresolved point source with @xmath32% .
we investigated all other features in the image and found none to have point source shape or scale ( based on unocculted images ) with significance greater than 2.5-@xmath27 .
no optical ghosts have ever been observed in the nir in our system since it was commissioned in early 2004 , and we have never observed spurious reflections from bright stars outside the field of view .
this putative point source may or may not be real , and despite its marginal detection , we include it in our analysis below , although the principal observations of structure in the disk stand regardless of the reality of the point source . further structure in the p image seems to indicate a drop in polarized light as one proceeds radially inward from the outer region toward the star .
this is particularly pronounced in the nw at a radius of about 570 mas ( 83 au ) .
closer to the star , the polarized light increases again up to the region obscured by the coronagraphic mask
. there may be an annular depletion zone in close proximity to the coronagraphic mask edge , but we state this with less certainty . this annular depletion zone ( which , also , shows no correlation with speckles in the @xmath21 image )
is located at roughly 300 mas ( 43 au ) radius from the star and a pa of about 142@xmath12 .
the azimuthally averaged surface brightness of the circumstellar material provides initial measurements of the disk material .
we use an azimuthal average , because the only effective way to remove the stellar light from the i - image , without restricting study to polarized light , is to use a reference star with no detectable circumstellar material .
ao point spread functions are highly variable and complex , meaning that point - spread function subtraction is not practical ( at best ) .
an azimuthal average permits a first order understanding of the difference between ab aur and a reference star , and the simplest measure of flux from the disk alone . to this end
, we used hd 107146 s azimuthally averaged @xmath21-image subtracted from the same average for ab aur ( fig . 4 ) .
the disk shows surface brightness ranging from about 80 to 3 mjy / arcsec@xmath33 at radii of 50 to 150 au .
this agrees with the values previously derived from 120 au outward in the h - band @xcite .
these values are also useable as a conservative upper limit to hd 107146 s disk flux in the 8 to 66 au region around that star .
interpretation of the morphology present in the p image depends on the notion that that polarization fraction in a given pixel in the image is a function of the column density of dust in the disk .
what exactly this means in terms of the disk s structure is not entirely clear .
a number of other studies suggest that this disk is optically thick in the mid - ir out to radii of about 118 au @xcite , roughly the region we have detected , although data on the spatial scales reported here are lacking .
if the disk is optically thick at the wavelengths we are observing here , then the p image reveals only a map of the inhomogeneities in the surface of the disk , not the mid - plane features .
if the disk is optically thin at these wavelengths and spatial scales , then the p image will map the full structure of the dust in the disk .
this is critical to interpretation , because , as @xcite note , the surface structure in a relatively thick disk will not necessarily reflect the interior density structure .
the full structure of the density is necessary to utilize detailed modeling of disks to provide rigorous theoretical interpretation .
therefore , we need to know whether the disk is optically thick or thin . however , the estimates of optical depth derived thus far are based on extremely simplistic models that assume azimuthal and vertical homogeneity in the disk structure , with a radial dependence .
what is beyond doubt is that these models are inaccurate based on the highly asymmetric morphology of the disk as revealed here .
the conclusion must be that some parts of the disk may be optically thick while others may not be , including some of the regions imaged here .
numerical models , not simple analytical ones are needed .
the primary feature of the observed region is in the annulus at an approximate radius of 102 au with a width equivalent to the image resolution @xmath29 10 au , where there is an apparent hole at a pa of 333@xmath12 .
in addition , there are clumps approximately 60@xmath12 from the gap in the azimuthal direction , and a general reduction of density on the opposite side of the gap .
we note that the very outer edge of this depletion zone is seen in data reporting a spiral structure in the near - ir , right at the edge of the corresponding coronagraphic mask @xcite .
this structure , as noted above , also seems offset from the stellar position by about 12 au .
since we are imaging in polarized light , some of this geometry could be due to the efficiency of scattering due to the inclination angle of the disk , variously reported for this inner region to be between 12@xmath12 and 30@xmath12 , with a major axis near 60 to 80@xmath12 @xcite .
indeed , if these numbers are correct , the annular depletion zone is nearly aligned with the minor axis of the disk and the opposite side would be expected to have reduced polarization fraction because it would be subject to the more weakly polarizing forward scattering by the dust . because of the great uncertainty in the inclination angle and the large range in values depending on the material imaged in a given study
, we did not see the value in adopting a particular inclination angle for the region we have imaged .
thus we have no `` deprojected '' image of the disk , which would also require a model for the scattering material and its mueller matrix .
a veritable zoo of models for structure in circumstellar disks exists , with a dominant competition currently between so - called `` core accretion '' and `` gravitational instability '' theories to explain the formation of low mass objects in disks .
these competing models predict different types of structure in disks . however , as far as could be determined from published material , the structure we see in this disk is most similar to models where a small object is present that dynamically induces mean motion resonance structure .
that structure creates amplifications of density in the azimuthal direction roughly @xmath34 away from the forming object , with a weaker clump opposite the object .
this structure also has clearing just inside the annulus as well as amplified density further in , another set of features that seem to exist in our @xmath19 images .
there are issues with interpreting ab aur s disk in this manner .
these models exhibiting mean motion resonance generally require optically thin disks with no gas present .
we made the point above that the optical depth is uncertain , and other observations show that this inner region is highly depleted of gas .
if this interpretation is correct , a small body would be forming in the gap and should be centered therein .
however , the theoretical modeling effort for disks of this type is not sufficiently mature to provide a complete interpretation .
the observations can most simply be interpreted as showing an amplification of material around the l4 and l5 trojan or lagrange points relative to an object that is situated in the gap .
we discuss the possible detection of that object below .
comparing the ab aur structure with other disks is important .
first we note the large size of this disk and associated other structures ( @xmath35 au ) .
a few low - mass companion formation theories exclude the formation of objects at such large distances as 102 au , due to a presumption of very low dust density ( * ? ? ?
* e.g. and references therein ) .
these models claim that very low mass companions at such distances could only be there through dynamical interactions after forming near the star , and therefore should have highly eccentric orbits .
the core accretion models can form objects at such distances , but generally of extremely low mass @xcite , and interestingly , the gravitational instability models can only form objects in the 50 to 100 au range because the disk is too hot closer to the star @xcite . however , observations of other disks reveal significant structure at scales certainly as large as this , albeit for older systems . for example
, the star fomalhaut has an offset ring of material , with some models claiming that this ring is due to a saturn mass object at an orbital distance of @xmath36 au @xcite with low eccentricity . in some respects
one could speculate that the ab aur structure is an evolutionary precursor to that found around the older fomalhaut , which is very similar in physical properties to ab aur except for its age .
furthermore , the star hd 142527 shows evidence for another offset ( @xmath37 au ) disk of radius 500 au and a hole of about 125 au in radius @xcite . such a structure is much larger than what we report here , and not too dissimilar to a smaller scale structure around epsilon eridani @xcite .
unfortunately few other herbig ae / be stars have been observed at the spatial scales ( @xmath38 100 au ) presented here or with polarimetry , so direct comparison with similarly aged objects is difficult .
there are two possible exceptions : observations of hd 141569 and hd 100546 , which have been observed with , for example hubble stis on larger scales and in optical scattered light ( which traces a variety of phenomena ; @xcite ) . in the case of hd 100546 ,
complex structure is revealed at essentially all probed radii from the star and because it is significantly fainter at larger radii ( @xmath39 au ) from the star than ab aur , it may be more evolved and optically thin .
the observations of this star within 100 au are , unfortunately lacking and so direct comparison with these observations provides little insight . for hd 141569 , which is widely described as a `` transitional '' disk , shows a spiral structure , also at large radii ( @xmath40 300 au ) and a continuous distribution of dust at closer radii with multiple ring - like structures @xcite .
this has direct similarity with the structures we have observed in ab aur , and the star is of similar spectra type and age .
the azimuthal gap structure that we observe , though , does not seem to be observed in any other circumstellar disk material .
one other comparison can be drawn with the star 49 ceti @xcite , which is an a1v star about 10 - 20 myr old .
some authors claim that this object is evolving from the herbig ae stage to the more mature `` vega '' state , and it is clear that there is a transition in dust properties with smaller grains dominating in the 30 - 60 au region @xcite .
this may be a slightly more mature version of the disk we observed here and shows a similar radial hierarchy of structures and dust properties , even though in ab aur the transition in dust properties seems to be at 130 au , rather than at 60 au .
the difference in age suggests that these very similar stars , when compared , provide clues to the evolution of disks around intermediate mass stars .
it may be possible to derive a mass for whatever object is causing the perturbations in the dust density in the disk through detailed modeling of the resonance structure and an estimate of the mass of material in this structure .
this is beyond the scope of this paper , and requires significant theoretical work . here
, we explore the possibility that the point source reported above in the annular depletion zone is a _ bona fide _ companion of ab aur , and provide two possible interpretations of what it is .
this discussion assumes that the detection is real , even though it has low signal - to - noise ratio .
although we have not confirmed that the object detected in the gap is physically associated with ab aur , through the detection of orbital motion or common proper motion , we have determined the probablilty of fortuitous alignment with unassociated background sources .
more details on this calculation are provided in the appendix .
the object s apparent @xmath25 magnitude , @xmath41 , ranges from 12.76 in the brightest possibility where it is just below the detection limit in the i - image ( at a radius of 0.7 arcsec ) , and 17.26 in the faintest case where it is 100% polarized .
these are the sum of the values from the dynamic range curve and @xmath42 for ab aur @xcite .
our model indicates that the probability that a background object happens to lie in this location is less than @xmath43 for the brighter limit and @xmath44 for the fainter one .
the probability is so minute that we assume physical companionship .
there are two ways to interpret this companion .
first , we can derive limits to its mass by assuming that we have detected thermal emission from a cooling body coeval with the star .
( @xmath45 m ) absolute magnitude based on the distance to ab aur is @xmath46 . using a solar luminosity of @xmath47 , the h band luminosity of the object
is between 3.5@xmath48 and 5.6@xmath49 l@xmath3 , the luminosity of the sun . given the 1 to 3 myr age of the system , this companion has a mass of between 5 and 37 m@xmath50 based on theoretical cooling models @xcite .
although these models are not ideal for such young ages and the results are dependent on the initial conditions for the models , there is a lack of alternative theory to constrain these physical parameters
. a second interpretation may be more appropriate in this case , because the object exhibits polarization of greater than 5% .
although planets orbiting stars should exhibit polarization , when viewed in reflected light , as high as 50% ( * ? ? ?
* e.g. ) there is no possibility , due to the brightness , that this source represents light from the star reflected off of a tiny planet at 100 au .
in fact , if it were reflected light , the simple geometry of the reflection would demand that the object have a radius of 0.1 au , about 10 times larger than the primary star , if the albedo were 1 .
the object is in the midst of a rather dusty region , albeit depleted from neighboring regions .
perhaps our images merely detect the infall of dust onto an obscured , accreting companion , rather than the companion itself .
models of planet formation suggest , in some cases , vortices of material surrounding the sites of cores of new planets . such regions would be unresolved at the imaging resolution available to the lyot project , for example . as such , they would appear to be unresolved clumps of dust that could scatter starlight just as the main part of the disk would .
other models , where the disk is denser ( optically thick even ) , predict bright spots and shadows at or slightly offset from the location of planet accretion due to a hole formed by the influx of material onto the forming body ( * ? ? ?
no models exist that permit us to estimate a mass for this object based on the polarimetry presented here .
however , if this is an accretion process , one would certainly expect a high degree of temporal variability as the companion is bombarded at random moments by larger particles or planetessimals as the disk continues to thin out .
ab aur and its environment are indisputably a complex system that provide clues to the physical processes of star and planet formation .
our study provides some insight into a new scale on par with our own solar system .
the combination of observations of this system suggest that ( a ) , because of the 1 - 4 au gap between the star and disk @xcite , the star is mostly formed and has completed its main accretion phase , although there is significant outflow still ; ( b ) planet formation , if it will happen , has already begun , due to the presence of an apparently organized disk with depleted gas , some of which may be adsorbed onto the dust itself ; ( c ) the region imaged here should have a similar temperature to the giant planet region of our solar system , and as such these data could provide constraints on giant planet formation mechanisms ; ( d ) the remaining thin dust and gas at larger radii are either flowing into or out of the disk , as indicated by strong spiral features at larger distances from the star .
in addition , projection effects complicate this interpretation particularly if there really is an organized inner disk ( such as we have imaged ) , with external material distributed more isotropically , or at least not in a series of coplanar disks . regarding the structure we have resolved , models of disks showing mean motion resonances also suggest strong outflows or infalls of material at coeval epochs with significant material trapped in mean motion resonant orbits .
these are the only models that seem to come remotely close to mimicking the structure we observe .
we may be witnessing such a process in action for the first time , or some combination of processes not yet modeled , and the poorly detected point source in our data may be related to the structure in the disk .
higher resolution and more sensitive observations of this fascinating system may be able to constrain planet and star formation models and permit distinguishing between competing theory .
unfortunately such observations can only be expected once new , large sub - mm observatories , such as alma , are in operation .
the critical need here involves spatial resolution superior to 100 mas and dynamic range greater than 12 magnitudes . a major space mission such as the terrestrial planet finder ,
would provide a crystal clear view of what is happening in this fascinating system . in the near term , hst s nicmos instrument can conduct similar polarimetric observations at lower resolution , and possibly worse dynamic range .
the lyot project is based upon work supported by the national science foundation under grant nos .
0334916 , 0215793 , and 0520822 , as well as grant nng05gj86 g from the national aeronautics and space administration under the terrestrial planet finder foundation science program .
the lyot project grateful acknowledges the support of the us air force and nsf in creating the special advanced technologies and instrumentation opportunity that provided access to the aeos telescope .
eighty percent of the funds for that program were provided by the us air force .
this work is based on observations made at the maui space surveillance system , operated by detachment 15 of the u.s .
air force research laboratory directed energy directorate .
the lyot project is also grateful to hilary and ethel lipsitz , the cordelia corporation , the vincent astor fund , judy vale , futdi , and an anonymous donor , who initiated the project .
bro thanks lynne hillenbrand , gilles
chabrier , isabelle baraffe and michal simon for extremely useful comments on early drafts , and misato fukagawa for being open and very generous with her data on ab aur taken with the subaru instrument .
this work has been partially supported by the national science foundation science and technology center for adaptive optics , managed by the university of california at santa cruz under cooperative agreement ast-9876783 .
we used the star hd 76943 for the astrometric calibration , observed at 14:13 on 2006 december 16 utc under similar observing conditions . on this date
, the orbit determination shows that the secondary component is at 0.546 arcseconds at a position angle of 333.54@xmath51 ( n - e)@xcite .
images of that star are shown in fig .
a1 to demonstrate the fidelity of the ao correction , roughly estimated via the strehl ratio , @xmath52 , where @xmath27 is the root mean square of the residual wave front error .
@xmath53 exceeded 85% in these images , as determined through comparison with laboratory - based images with the coronagraph @xcite .
this performance is similar in acquisition images of ab aur taken immediately before the primary data acquisition .
to demonstrate how completely the starlight ( and speckle noise ) is removed with our imaging polarimetry process , we examine , in more detail , the observations of the star hd 107146 which is used in the main article as a `` control '' non - detection .
it is bright in the coronagraphic i image ( fig .
1a ) , with a total of 1.33@xmath56 photons detected in a 360 second sequence of exposures .
however , the star is almost entirely absent in the p image ( fig .
1b ) with 1.60@xmath57 photons , most of which , due to diffraction , are behind the coronagraphic occulting mask , a region of no consequence to scientific results .
the polarimetry method reduces the residual starlight by a factor of 843 in this case and brings the vast majority of the field of view to the instrumental throughput limits
the @xmath19 image shows a typical readnoise - limited background .
no speckle structure remains in the region usable for scientific investigations .
combined with the coronagraph , the central star s light has been suppressed by a factor of about 10@xmath58 over the entire field of view , and possibly by much more since our images were not limited by the starlight but rather instrument sensitivity .
it is important to note that none of the speckles in the @xmath21 image are present in the p image .
this is true for the ab aur data as well .
to determine our sensitivity in the @xmath19 image , artificial point sources with a given intensity and fractional polarization were placed radially into the constituent images prior to composing the @xmath19 image , with appropriate adjustments in the pairs of simultaneous images so that the relevant fractional polarization was correctly represented by the fake sources .
these images were double differenced as described above ( polarimetry ) .
the total intensity of these point sources in the original images was adjusted until a 3-@xmath27 detection was obtained in the processed @xmath59 , @xmath60 , and @xmath61 data .
( @xmath19 images have non - gaussian noise properties so noise must be derived from the constituent images . ) the total intensity ( not just the flux in net - polarized light ) for a given fractional polarization is plotted in the dynamic range plot , although for any real source detected , only a lower limit to the fractional polarization is measurable , because of the contamination from the starlight throughout the image .
we constrained the probability that a point source would be co - aligned with features in ab aur s disk using our detection limits and a model for the density of objects on the sky . deriving a star count - magnitude relationship from the two - micron all - sky survey point source catalog @xcite and using a fit to galaxy count - magnitude relations
@xcite , we find that sources in the region of sky around ab aur have a sky surface density of @xmath62 where @xmath63 and @xmath64 are the number of stars and galaxies per square degree , respectively , and @xmath41 is the apparent h - band magnitude . because we have detected this source in polarized light , we must modify these equations to reflect the fact that the majority of celestial sources have immeasurable ( @xmath38 1% ) polarization with our technique @xcite . based on results of a survey of polarized starlight @xcite
, we very conservatively assume that 50% of all stars and galaxies are polarized to the point where they would be detected in our p - image .
this adds a constant of -0.30 to the equations above . to apply this , we make two estimates of the object s @xmath41 , which are 12.76 in the brightest possibility where it is just below the detection limit in the @xmath21-image ( at a radius of 0.7 arcsec ) , and 17.26 in the faintest case where it is 100% polarized .
these are the sum of the values from the dynamic range curve and @xmath42 for ab aur @xcite . from equations ( 1 ) and ( 2 ) and the area of the depleted region in the outer disk annulus , a circle of radius @xmath65 arcsec , the probability that a background object happens to lie in this location is less than @xmath43 for the brighter limit and @xmath44 for the fainter one .
the probability is so minute that we assume physical companionship . , h .- w . ,
mccarthy , p. j. , marzke , r. o. , wilson , j. , carlberg , r. g. , firth , a. e. , persson , s. e. , sabbey , c. n. , lewis , j. r. , mcmahon , r. g. , lahav , o. , ellis , r. s. , martini , p. , abraham , r. g. , oemler , a. , murphy , d. c. , somerville , r. s. , beckett , m. g. , & mackay , c. d. 2002 , _ astrophys .
j. _ , 570 , 54 , m. , krist , j. e. , ardila , d. r. , golimowski , d. a. , hartig , g. f. , ford , h. c. , illingworth , g. d. , bartko , f. , bentez , n. , blakeslee , j. p. , bouwens , r. j. , broadhurst , t. j. , brown , r. a. , burrows , c. j. , cheng , e. s. , cross , n. j. g. , feldman , p. d. , franx , m. , gronwall , c. , infante , l. , kimble , r. a. , lesser , m. p. , martel , a. r. , menanteau , f. , meurer , g. r. , miley , g. k. , postman , m. , rosati , p. , sirianni , m. , sparks , w. b. , tran , h. d. , tsvetanov , z. i. , white , r. l. , & zheng , w. 2003 , , 126 , 385 , m. , hayashi , m. , tamura , m. , itoh , y. , hayashi , s. s. , oasa , y. , takeuchi , t. , morino , j .- i . , murakawa , k. , oya , s. , yamashita , t. , suto , h. , mayama , s. , naoi , t. , ishii , m. , pyo , t .- s . , nishikawa , t. , takato , n. , usuda , t. , ando , h. , iye , m. , miyama , s. m. , & kaifu , n. 2004 , _ astrophys . j. lett .
_ , 605 , l53 , m. , itoh , y. , oasa , y. , kudo , t. , fujiwara , h. , tamura , m. , hayashi , m. , & hayashi , s. 2007 , in in the spirit of bernard lyot : the direct detection of planets and circumstellar disks in the 21st century , ed .
p. kalas , c. a. , polomski , e. f. , henning , t. , stecklum , b. , woodgate , b. e. , telesco , c. m. , pia , r. k. , gull , t. r. , boggess , a. , bowers , c. w. , bruhweiler , f. c. , clampin , m. , danks , a. c. , green , r. f. , heap , s. r. , hutchings , j. b. , jenkins , e. b. , joseph , c. , kaiser , m. e. , kimble , r. a. , kraemer , s. , lindler , d. , linsky , j. l. , maran , s. p. , moos , h. w. , plait , p. , roesler , f. , timothy , j. g. , & weistrop , d. 2001 , , 122 , 3396 , c. a. , woodgate , b. e. , bowers , c. w. , gull , t. r. , sitko , m. l. , carpenter , w. j. , lynch , d. k. , russell , r. w. , perry , r. b. , williger , g. m. , roberge , a. , bouret , j .- c . , & sahu , m. 2005 , , 630 , 958 , h. 2007 , in proceedings of the conference in the spirit of bernard lyot : the direct detection of planets and circumstellar disks in the 21st century .
june 04 - 08 , 2007 .
university of california , berkeley , ca , usa . edited by paul kalas . , ed .
p. kalas , r. , monnier , j. d. , berger , j .-
p . , traub , w. a. , schloerb , f. p. , pedretti , e. , benisty , m. , carleton , n. p. , haguenauer , p. , kern , p. , labeye , p. , lacasse , m. g. , malbet , f. , perraut , k. , pearlman , m. , & thureau , n. 2006 , _ astrophys .
_ , 645 , l77 , b. r. , digby , a. p. , newburgh , l. , brenner , d. , shara , m. , mey , j. , mandeville , c. , makidon , r. b. , sivaramakrishnan , a. , soummer , r. , graham , j. r. , kalas , p. , perrin , m. d. , roberts , jr . , l. c. , kuhn , j. r. , whitman , k. , & lloyd , j. p. 2004 , in proc .
spie , vol .
5490 , advancements in adaptive optics , ed .
d. bonaccini calia , b. l. ellerbroek , & r. ragazzoni , 433442 , jr . ,
l. c. , perrin , m. d. , marchis , f. , sivaramakrishnan , a. , makidon , r. b. , christou , j. c. , macintosh , b. a. , poyneer , l. a. , van dam , m. a. , & troy , m. 2004 , in proc .
spie , vol .
5490 , advancements in adaptive optics , ed .
d. bonaccini calia , b. l. ellerbroek , & r. ragazzoni , 504515 , m. f. , cutri , r. m. , stiening , r. , weinberg , m. d. , schneider , s. , carpenter , j. m. , beichman , c. , capps , r. , chester , t. , elias , j. , huchra , j. , liebert , j. , lonsdale , c. , monet , d. g. , price , s. , seitzer , p. , jarrett , t. , kirkpatrick , j. d. , gizis , j. e. , howard , e. , evans , t. , fowler , j. , fullmer , l. , hurt , r. , light , r. , kopan , e. l. , marsh , k. a. , mccallon , h. l. , tam , r. , van dyk , s. , & wheelock , s. 2006 , _ astron . j. _
, 131 , 1163 , r. , oppenheimer , b. r. , hinkley , s. , sivaramakrishnan , a. , makidon , r. b. , digby , a. , brenner , d. , kuhn , j. , perrin , m. d. , roberts , jr . , l. c. , & kratter , k. 2006 , in eas publications series , 199212 | the young star ab aurigae is surrounded by a complex combination of gas - rich and dust dominated structures .
the inner disk which has not been studied previously at sufficient resolution and imaging dynamic range seems to contain very little gas inside a radius of least 130 astronomical units ( au ) from the star . using adaptive - optics coronagraphy and polarimetry
we have imaged the dust in an annulus between 43 and 302 au from the star , a region never seen before .
an azimuthal gap in an annulus of dust at a radius of 102 au , along with a clearing at closer radii inside this annulus , suggests the formation of at least one small body at an orbital distance of about 100 au .
this structure seems consistent with crude models of mean motion resonances , or accumulation of material at two of the lagrange points relative to the putative object and the star .
we also report a low significance detection of a point source in this outer annulus of dust .
this source may be an overdensity in the disk due to dust accreting onto an unseen companion .
an alternate interpretation suggests that the object s mass is between 5 and 37 times the mass of jupiter .
the results have implications for circumstellar disk dynamics and planet formation . |
transverse lattice gauge theory @xcite is a promising approach for ab - initio calculations of the hadronic light - cone wavefunctions .
all successful calculations so far have been performed on coarse transverse lattices with effective ` colour - dielectric ' degrees of freedom , keeping the longitudinal coordinates @xmath2 continuous .
although this has required some parameters in the effective hamiltonian to be fit phenomenologically , it has the great computational advantages that hadrons are dominated by states containing few partons and there is a clean separation between the longitudinal and transverse dynamics .
in fact , in this case , it is not much harder to analyse the dependence of wavefunctions on the longitudinal coordinates than in @xmath3-dimensional gauge theories . while the full wavefunction of a boundstate typically requires even in 1 + 1 dimensions a numerical solution , the behaviour when parton lightcone momenta @xmath4 vanish can be largely understood analytically .
the prototype example is @xmath3-dimensional large-@xmath1 qcd @xcite , where the behaviour of the meson wavefunction for vanishing quark @xmath4 was found exactly .
a similar analysis can be performed to some extent even in @xmath5-dimensional large-@xmath1 qcd @xcite .
the key strategy in all cases is to demand cancellation of singularities in the lightcone boundstate schrodinger equation when parton lightcone momenta vanish .
typically , wavefunctions do not simply vanish in this high free - energy limit , as they would at large transverse momentum . rather , there occur ` ladder ' relations between states with different numbers of partons that couple the analytic structure of the infinite tower of lightcone wavefunctions . in this paper
, we analyse in a similar vein the corresponding lightcone boundstate schrodinger equation for coarse transverse lattice gauge theory at large-@xmath1 . by demanding finiteness of the boundstate energy
, we determine the form of meson wavefunctions at small lightcone momentum .
these depend on two exponents , @xmath6 and @xmath7 , whose values are independent of all details of the wavefunctions , which encode the leading non - analytic behaviour . in general , naive intuition based on
the free lightcone kinetic energy is wrong . in particular , wavefunctions become more singular as more momenta vanish .
quark distributions diverge as @xmath8 and all fock sectors contribute as @xmath9 . in the next section ,
we briefly review the colour - dielectric formulation of transverse lattice gauge theory and set up the boundstate schrodinger equation to lowest non - trivial order of the colour - dielectric expansion that is used in the following two sections .
sections [ onelink ] and [ analytic ] determine systematically the behaviour of lightcone wavefunctions of mesons at small parton lightcone momentum . at first ,
in section [ onelink ] , we use the approximation where up to one gauge link - variable is allowed in a state .
then , in section [ analytic ] , we relax this approximation and generalise many of the results , giving a simple ansatz that incorporates all the boundary conditions . applications of the results to the behaviour of quark generalised parton distributions are given in section [ appl ] and conclusions are discussed in section [ last ] .
we introduce continuum light - cone co - ordinates @xmath10 and discretize the transverse coordinates @xmath11 on a square lattice of spacing @xmath12 .
lorentz indices @xmath13 are split into light - cone indices @xmath14 and transverse indices @xmath15 .
quark fields @xmath16 in the fundamental representation and longitudinal gauge potentials @xmath17 in the adjoint representation of the gauge group @xmath18 are associated to the sites of the transverse lattice .
flux link fields @xmath19 , which we choose to be in a complex matrix representation of @xmath18 , are associated with the directed link from @xmath20 to @xmath21 , where @xmath22 is a unit vector in direction @xmath23 . in the transverse continuum limit @xmath24
, these would be related to transverse gauge potentials @xmath25 .
subsequent analysis is done to leading order of the @xmath26 expansion , which allows one to drop the site argument @xmath20 of the fields @xcite . for finite transverse lattice spacing @xmath12
, the lagrangian may contain any operator invariant under transverse lattice gauge symmetries , poincar symmetries manifestly preserved by the lattice cut - off , and renormalisable by dimensional counting with respect to the continuum co - ordinates @xmath27 .
one may control the subsequent proliferation of operators either by approaching the transverse continuum limit , in which case only renormalisable operators with respect to co - ordinates @xmath20 survive , or by ` strong - coupling ' methods on a coarse lattice , such as the colour - dielectric expansion @xcite . all calculations to date have used the colour - dielectric form @xcite and it is this approach that we analyse here .
this involves fixing to the light - cone gauge @xmath28 , classically eliminating non - dynamical fields @xmath29 , @xmath30 , then expanding the resulting lightcone hamiltonian in powers of the remaining dynamical fields @xmath31 , @xmath32 .
truncation of such an expansion is a valid approximation provided wavefunctions of interest ( typically those of the lightest eigenstates ) are dominated by few - body fock states .
this is achieved by the choice of sufficiently large masses for the dynamical fields .
when the remaining couplings in the truncated hamiltonian are tuned to optimise symmetries broken by the lattice cutoff , the choice of large masses corresponds to choice of a coarse lattice spacing @xmath12 .
the lagrangian density @xmath33 we consider , which was previously investigated in ref .
@xcite , contains all allowed terms up to order @xmath34 and @xmath35 .
( there are additional terms involving multiple traces of products of @xmath36 fields that contribute only to the glueball sector at large-@xmath1 @xcite . )
@xmath37[{\rm h.c.}]\right\ } \nonumber \\ & & - \mu_{b}^2 { { \rm \ , tr \!}}\left\{m_r m_r^{{\dagger}}\right\ } - v[m ] + { \rm i } \overline{\psi } { \gamma}^{\alpha } ( \partial_{\alpha } + { \rm i } a_{\alpha } ) \psi - \mu_f \overline{\psi}\psi \nonumber\\ & & + { \rm i } \kappa_a \left ( \overline{\psi } { \gamma}^{r } m_r \psi - \overline{\psi } { \gamma}^{r } m_{r}^{{\dagger } } \psi \right ) + \kappa_s \left ( \overline{\psi } m_r \psi+\overline{\psi } m_{r}^{{\dagger } } \psi\right ) \ , \label{ferlag}\end{aligned}\ ] ] where @xmath38 is the continuum field strength in the @xmath39 plane and the link - field potential is @xmath40 & = & - \frac{\beta}{a^2 n_c } \sum_{r \neq s } { { \rm \ , tr \!}}\left\ { m_{r } m_{s } m_{r}^{{\dagger } } m_{s}^{{\dagger } } \right\ } + \frac{\lambda_1}{a^{2 } n_c } \sum_r { { \rm \ , tr \!}}\left\ { m_r m_r^{{\dagger } } m_r m_r^{{\dagger } } \right\ } \nonumber \\ & & + \frac{\lambda_2}{a^{2 } n_c}\sum_r { { \rm \ , tr \!}}\left\ { m_r m_r m_r^{{\dagger } } m_r^{{\dagger } } \right\ } + \frac{\lambda_4}{a^{2 } n_c } \sum_{\sigma=\pm 2 , \sigma^\prime = \pm 1 } { { \rm \ , tr \!}}\left\ { m_\sigma^{{\dagger } } m_\sigma m_{\sigma^\prime}^{{\dagger } } m_{\sigma^\prime } \right\ } \ ; . \label{pot1}\end{aligned}\ ] ] we have defined @xmath41 and hold @xmath42 , @xmath43 , and @xmath44 finite as @xmath45 . in the chiral representation , @xmath46 decomposes into two - dimensional complex spinors @xmath47 the helicity subscript @xmath48 denotes the sign of the eigenvalue of @xmath49 . using pauli matrices ,
we may then write various bilinears @xmath50 we also introduce the matrices @xmath51 which act on the 2-spinors : @xmath52 in light - cone gauge @xmath53 , @xmath54 and @xmath55 are non - dynamical ( independent of light - cone time @xmath56 ) and their constraint equations of motion are used to eliminate them at the classical level .
the lightcone hamiltonian may then be obtained from the action ( [ ferlag ] ) in the standard way in terms of the remaining dynamical fields @xmath57 , @xmath58 , and @xmath59 . if we define @xmath60 the lightcone hamiltonian one finds is @xmath61 \ . \label{ham } \end{aligned}\ ] ] under certain reasonable assumptions @xcite , the hamiltonian ( [ ham ] ) is a truncation of the most general hamiltonian to order @xmath34 and @xmath62 .
it also contains some , but not all , allowed operators at order @xmath63 and @xmath64 ; most importantly , it contains those responsible for confinement @xcite . introducing the mode expansions @xmath65 we have @xmath66 & = & \delta_{ik}\ , \delta_{jl}\ , \delta_{\lambda \rho}\ , \,\delta(k^+-\tilde{k}^+ ) \ ; , \\
\left[a_{\lambda , ij}(k^+),a_{\rho , kl}(\tilde{k}^+)\right ] & = & 0 \ ; , \\
\left\{b_{h , i}(k^+ ) , b_{h',j}^{*}(\tilde{k}^+)\right\ } & = & \delta_{ij}\ , \delta_{hh'}\ , \,\delta(k^+-\tilde{k}^+ ) \ ; , \\ \left\{b_{h ,
i}(k^+ ) , b_{h',j}(\tilde{k}^+)\right\ } & = & 0 \ ; , \end{aligned}\ ] ] where @xmath67 are colour indices , @xmath68 and @xmath69 denote the orientations of link variables in the @xmath70 plane , and @xmath71 .
the sequence of orientations @xmath72 of link variables together with the @xmath73 momentum fractions @xmath74 and quark helicities @xmath75 are sufficient to encode the structure of fock states contributing to the boundstate at zero total transverse momentum @xmath76 .
only gauge singlet combinations under residual gauge transformations in @xmath77 gauge can contribute to states of finite energy @xcite , so a boundstate containing quarks at large @xmath1 decomposes in terms of fock states as @xmath78 the state is normalised by @xmath79 if @xmath80 the object of this paper is to determine the behaviour of the wavefunctions @xmath81 when one or more @xmath82 are small .
the fock space matrix elements of the invariant mass operator @xmath83 resulting from ( [ ham ] ) were given in ref .
@xcite . for ease of reference and to correct some typing errors in interaction _
( ii ) _ , we reproduce these in a slightly more efficient notation as fig .
[ vert ] and table [ table2 ] in appendix [ appendix ] . projecting @xmath84 , where @xmath85 is the invariant mass eigenvalue , onto individual fock states , one obtains integral equations for the wavefunctions .
to begin analysing the small @xmath0 behaviour of the boundstate equations , we make the approximation of restricting the sum in eq .
( [ fock ] ) to @xmath86 , i.e. at most one flux link in a state .
we correspondingly restrict to interactions @xmath87 ( see appendix [ appendix ] ) .
the @xmath88 or @xmath89 meson state can then be written as @xmath90 to obtain the @xmath91 spin projection component of the @xmath92 meson , one replaces @xmath93 in the above .
the wavefunctions of the @xmath94 components of the @xmath92 meson are more complicated : @xmath95 this is the state that is odd under reflections @xmath96 ; the even state is obtained by removing the @xmath97 .
the bound state equations for these four states are almost identical .
the @xmath98 bound state equation is @xmath99 \psi(x ) \nonumber \\ & & - \frac{g^2n_c}{2 \pi } \int_0 ^ 1 \frac{dy}{(x - y)^2 } \left(\psi(y)-\psi(x)\right ) \nonumber \\ & & - \frac{\mu_f \kappa_s \sqrt{n_c}}{\sqrt{\pi } } \left [ \int_0^x dy\ , \frac{(2x - y)\,\psi_s(x - y , y)}{\sqrt{y}\ , x ( x - y ) } + \int_0^{1-x } dy\ , \frac{(2 - 2x - y)\,\psi_s(x , y ) } { \sqrt{y } ( 1-x ) ( 1-x - y ) } \right ] \nonumber \\ & & + \frac{\mu_f \kappa_a \sqrt{n_c}}{\sqrt{\pi } } \left [ \int_0^x dy\ , \frac{\sqrt{y}\,\psi_a(x - y , y)}{x ( x - y ) } \right .
\nonumber \\ & & \left .
+ \int_0^{1-x } dy\ , \frac{\sqrt{y}}{(1-x ) ( 1-x - y ) } \left\{\begin{array}{cl } \psi_a(x , y ) & \mbox{for pion } \\
-\psi_a(x , y ) & \mbox{for rho , $ { \cal j}_3=0 $ } \\
-\psi_a^\prime(x , y ) & \mbox{for rho , $ { \cal j}_3=\pm1 $ } \end{array}\right .
\right ] \ ; . \label{two}\end{aligned}\ ] ] for each of the wavefunctions , @xmath100 , @xmath101 , @xmath102 , the @xmath103 bound state equation is @xmath104 \;. \label{three}\end{aligned}\ ] ] for @xmath105 , we include @xmath106 \ ; .\end{aligned}\ ] ] for @xmath107 in the pion and the rho @xmath91 state , we include @xmath108 \ ; .\end{aligned}\ ] ] for @xmath107 in the rho @xmath109 states , we include @xmath110 for @xmath111 in the rho @xmath109 states , we include @xmath112 next , we analyze the singularities as one or more of the momenta go to zero
. we will consider each of the following limits : @xmath113 where @xmath114 , and @xmath115 .
see figure [ psixy ] . the remaining limit of interest , @xmath116 with @xmath117 constant , is equivalent to the limit ( [ xylimit ] ) by charge conjugation . for @xmath105
, we obtain : @xmath119\nonumber\\ & & + o(x^a)\ ; .\end{aligned}\ ] ] for @xmath107 , we obtain : @xmath120\nonumber\\ & & + o(x^a)\ ; .\end{aligned}\ ] ] in order to cancel the divergence of second term , we must set @xmath121 .
we then find that the integral term , which is also @xmath122 , goes to zero .
simplifying , we get : @xmath123 for this limit , the resulting equation is the same for @xmath100 and @xmath101 because the differing terms are non - leading .
@xmath125\nonumber\\ & & + o(y^{-\frac{1}{2}})\mbox{. } \end{aligned}\ ] ] we find the constraint @xmath126 , so that the integrals converge , in which case the @xmath127 terms are non - leading .
we also have the constraint @xmath128 , so that the expectation value of @xmath129 is finite .
thus , we obtain : @xmath130 \ ; .\end{aligned}\ ] ] simplifying , we derive : @xmath131\ , g^\prime\!(x)\,y^{\beta-1}\mbox{.}\label{twogluon}\ ] ] note that @xmath132 is not fixed by this analysis .
( [ twogluon ] ) is the same expression that one obtains for the two - link bound state equation without quarks @xcite . in this limit
, we obtain , for both @xmath135 and @xmath136 : @xmath137 \nonumber\\ & & -\frac{g^2 n_c}{4\pi}\ , p^{b-1}\int^1_0\frac{\mathrm{d}q}{(q - w)^2 } \left[\frac{(w+q-2 q w)\,f(q)}{\sqrt{(1-w)(1-q ) } } \left(\frac{w}{q}\right)^\alpha -2w\,f(w)\right ] \nonumber\\ & & + \frac{\mu_f c_2 \sqrt{n_c } } { \sqrt{\pi } w}\,p^{\alpha-3/2 } \left\{\begin{array}{cl } -\kappa_s \frac{(w+1)}{\sqrt{1-w } } & \mbox{for $ f_s(w)$ } \\ \kappa_a \sqrt{1-w } & \mbox{for $ f_a(w)$ } \end{array}\right . \nonumber\\ & & + o\left(p^{-1/2}\right ) \ ; . \label{fw}\end{aligned}\ ] ] in order for the last term in the equation to be cancelled in the limit @xmath138 , we must set @xmath139 .
thus , @xmath140 taking the @xmath118 limit of eq .
( [ two ] ) and using ( [ alpha ] ) , we obtain @xmath142\nonumber\\ & & + x^{\alpha-1 } \int^1_0 \frac{\mathrm{d}z}{1-z}\left[\frac { \kappa_a^2 n_c c_2}{\pi}+\frac{\mu_f \kappa_a \sqrt{n_c } } { \sqrt{\pi}}(\sqrt{z}\,f_{\!a}\!(1-z))\right]\nonumber\\ & & + o(x^0 ) \ ; . \label{twoeqn}\end{aligned}\ ] ] for the expectation value of the energy to be finite , we must have @xmath143 . [ cols="^,^ " , ] to determine @xmath6 , one must find @xmath144 and @xmath145 that self - consistently solve equations ( [ fw ] ) and ( [ twoeqn ] ) .
we obtained solutions numerically by discretizing @xmath146 on the interval [ 0,1 ] . using numerical calculations of @xmath144 and @xmath145
, one can study the dependence of @xmath6 on the couplings in the hamiltonian ; see figures [ figa][figb ] .
note that @xmath6 is very nearly constant with respect to link field mass @xmath147 ( fig .
[ figa ] ) . also , note that @xmath6 is outside of its allowed range for some areas of the coupling space , corresponding generally to large values of @xmath44 and @xmath43 ( fig .
[ figb ] ) .
thus , the endpoint behavior of the wavefunction can restrict the allowed values of the couplings .
@xmath6 + @xmath148 @xmath149 @xmath6 + @xmath150 @xmath151
when relaxing the truncation on @xmath152 and including the other interactions of fig . [ vert ] , most of the complication occurs in the analytic auxiliary functions , such as @xmath153 , @xmath154 , @xmath155 in the previous section . if we wish to determine only the form of the leading non - analytic behaviour of wavefunctions at small momenta , the analysis is somewhat simpler .
we note that this is independent of spin , represented by quark helicity and link field transverse orientation .
consider the effect of including @xmath156 wavefunctions @xmath157 and the other interactions of fig .
[ vert ] . in the following
we ignore the spin dependence , finding only the momentum dependent part of these contributions .
accordingly , we label wavefunction components @xmath158 by the number of partons @xmath152 and their lightcone momenta only . to the @xmath98 boundstate equation ( [ two ] ) we must add interactions @xmath159 : @xmath160 from @xmath161 , and @xmath162 from @xmath163 .
if we assume that these contribute at the leading @xmath164 order as @xmath8 in the analsyis of section [ twoxzero ] , then @xmath165 in the @xmath103 equation ( [ three ] ) , we must include interaction @xmath166 , that connects @xmath103 to @xmath103 and was previously omitted : @xmath167 this is subleading in the @xmath8 analysis of section [ threexzero ] and the @xmath168 analysis of section [ threeyzero ] ( recall that @xmath169 ) .
it contributes at the leading order @xmath170 in the analysis of section [ threexyzero ] , but does not affect the relation @xmath139 . to the @xmath103 equation ( [ three ] ) we must also add terms involving @xmath157 via interaction @xmath171 : @xmath172 if we assume that these contribute at the leading @xmath173 order of the analysis in section [ threexzero ] , then @xmath174 in the @xmath168 analysis of section [ threeyzero ] , ( [ ii ] ) does not contribute at leading order since , by normalisability , @xmath175 and @xmath176 can not diverge faster than @xmath177 when @xmath168 . in the analysis of section [ threexyzero ] , property ( [ 3zero ] ) implies that ( [ ii ] ) contributes at the leading @xmath178 order as @xmath179 .
finally , there will be a new @xmath156 boundstate equation for @xmath180 .
the momentum dependent parts of these contributions are given in table [ table1 ] .
@xmath181 repeating the analysis of sections [ threexzero ] and [ threeyzero ] for @xmath156 , we again find from comparing interactions @xmath182 and @xmath171 that @xmath183 since all other interactions are subleading , while from interactions @xmath184 @xmath185 we confirm directly , from the @xmath156 amplitudes of table i , the correctness of eq .
( [ asump ] ) and eq .
( [ 3zero ] ) , thus verifying the assumptions that went into deriving them . in a similar way we also deduce as @xmath186 : @xmath187 and @xmath188 we can summarize all these results in the following forms for @xmath158 , which have the correct endpoint behaviour when any number of momenta vanish : @xmath189^{\beta-\alpha+1/2 } } \label{ansatz } \\ \psi_{4}(x , y , z ) & = & c_4 \frac{(yz)^{\beta } [ ( 1-y)(1-z)]^{\alpha } [ ( x+y+z)(1-x)]^{\alpha+\beta-1/2 } } { [ ( x+ y)(1-x - y)]^{\beta + 1/2}(y+z)^{2\beta } } \nonumber \end{aligned}\ ] ] the exponents @xmath6 and @xmath7 are independent of @xmath152 , quark helicity , and the transverse shape , while the auxiliary functions @xmath190 are dependent on all of these factors . if the @xmath190 are made from a complete set of functions of @xmath82 , non - singular in the limit when any number of momenta vanish , then any boundstate wavefunction should be expressible in the form ( [ ansatz ] ) .
we have also examined the effect of higher orders in @xmath191 .
the forms ( [ ansatz ] ) remain invariant , although the exponents @xmath6 and @xmath7 are renormlised in general because of parton self - energy and vacuum polarization .
wavefunctions @xmath158 analogous to ( [ ansatz ] ) become increasing complicated at higher @xmath152 and we have not found a simple general expression ; no new exponents beyond @xmath6 and @xmath7 occur however . of special interest
are the limits @xmath192 which follow from comparing interactions @xmath182 and @xmath171 .
these govern the behaviour of hadronic structure functions at large and small bjorken @xmath0 , as discussed in section [ appl ] .
the forms ( [ ansatz ] ) demonstrate that lightcone wavefunctions become more singular as more lightcone momenta vanish , even though this is a ` high - energy ' limit .
intuition from the free lightcone kinetic energy ( sum of all terms @xmath182 and @xmath193 in table [ table2 ] of appendix [ appendix ] ) would have suggested the complete opposite . rather than simply vanishing at the edges of phase space , as indicated by the free theory ,
the interactions have forced relations between wavefunctions with different numbers of partons to cancel divergences in the boundstate equations .
this was noted for continuum qcd for a single vanishing @xmath82 in ref .
@xcite , but for the transverse lattice we are able to generalise it to any number of vanishing momenta .
our analysis rules out a number of naive models for the lightcone wavefunctions , including products of individual parton wavefunctions and functions of the free kinetic energy .
based on the previous analysis , we can make definite statements about the behaviour of quark parton distributions in mesons at the edges of phase space .
some other recent studies of these distributions in various models have been made in refs .
we use the generalised quark distributions @xcite , which in lightcone gauge are defined as @xmath194 @xmath195 bars generally indicate an average over in and out states . in eq .
( [ gpd ] ) , we define the boundstates at non - zero transverse momentum @xmath196 by boosting the fock states with the poincar generators @xmath197 : @xmath198 b_{h}^{\dagger}(x_{1}p^+ ) a^{\dagger}_{\lambda_2}(x_{2}p^+ ) \cdots a^{\dagger}_{\lambda_{n-1}}(x_{n-1}p^+ ) d_{h'}^*(x_n p^+ ) |0\rangle \nonumber \\ & = & { \rm exp}\left [ { \rm i } { \bf p}. { \bf c } \right ] b_{h}^{\dagger}(x_{1}p^+ ) a^{\dagger}_{\lambda_2}(x_{2}p^+ ) \cdots a^{\dagger}_{\lambda_{n-1}}(x_{n-1}p^+ ) d_{h'}^*(x_n p^+ ) |0\rangle \ , \label{boost}\end{aligned}\ ] ] @xmath199 where @xmath200 is a transverse unit vector in the direction of the @xmath201 link .
the boundaries of interest are typically at large and small @xmath202 and at the transition point @xmath203 .
we find from eqs .
( [ gpd ] ) and ( [ uno ] ) that each @xmath158 contributes @xmath205 to @xmath206 as @xmath204 .
this is interesting because the conventional wisdom , based upon the free kinetic energy , is that @xmath207 should dominate in this region all but one parton has small momentum , so the free kinetic energy is minimized by the minimum number of partons .
although the exponent _ is _ correctly given by the lowest fock wavefunction , which governs the distribution amplitude , on the transverse lattice all higher fock states contribute to the structure function with the same power .
we find from eqs .
( [ gpd ] ) and ( [ nought ] ) that @xmath158 contributes @xmath208 , @xmath209 , to the traditional quark distribution function @xmath210 as @xmath211 .
thus , the sum over all fock sectors @xmath152 rises faster than any power of @xmath212 , i.e. at least like a power of @xmath0 .
this is consistent with the regge behaviour of structure functions in the small-@xmath0 region .
unfortunately , we are not able to explore the region @xmath214 with the strict large-@xmath1 wavefunctions . that region is governed by their @xmath26 corrections , since it depends upon quark pair production .
however , we find from eq .
( [ ansatz ] ) that @xmath207 contributes zero , while @xmath215 and @xmath157 contribute finite constants to @xmath206 as @xmath216 from above , provided @xmath217 .
this is presumably true for higher @xmath152 also and , provided these constants fall sufficiently rapidly with @xmath152 , suggests that @xmath218 is finite .
by demanding finiteness of the boundstate energy , we have determined the form ( [ ansatz ] ) of meson wavefunctions at small lightcone momentum in a coarse transverse lattice gauge theory at large @xmath1 .
the forms depend on two exponents @xmath6 and @xmath7 , independent of the other details of the wavefunction such as spin and number of partons , which encode the leading non - analytic behaviour .
as more momenta vanish , wavefunctions become more singular , contrary to the naive intuition from free lightcone kinetic energy , which diverges in these limits .
the analytic forms lead to predictions for the behaviour near their endpoints of quark distributions .
in particular , we see regge behaviour at small @xmath0 and contributions from all fock states at large @xmath0 . the use of coarse lattice wavefunctions in transverse position space simplified the analysis , allowing us to separate the lightcone momentum problem from the transverse structure .
the large-@xmath1 limit also aided us by restricting the diagrammatic rules to planar ones and suppressing quark pair production .
we would need to reformulate the problem at finite @xmath1 in order to study the small lightcone momentum behaviour in baryons and in the sea quark part of wavefunctions .
there is also the question of higher orders of the colour - dieletric expansion ; that is , higher powers of dynamical quark and link fields in the lightcone hamiltonian .
we have investigated some of these possible terms , but found none that altered the form ( [ ansatz ] ) , although the values of @xmath6 and @xmath7 are affected .
finally , it would be interesting to investigate similar questions in continuum lightcone qcd .
@xmath219 m. v. polyakov and c. weiss , phys . rev .
d * 60 * , 114017 ( 1999 ) [ arxiv : hep - ph/9902451 ] ; + i. v. anikin _
et al _ , nucl .
a * 678 * , 175 ( 2000 ) [ arxiv : hep - ph/9905332 ] ; + a. mukherjee , i. v. musatov , h - c .
pauli , and a. v. radyushkin , phys .
d * 67 * , 073014 ( 2003 ) [ arxiv : hep - ph/0205315 ] ; + w. broniowski and e. ruiz arriola , phys .
b * 574 * , 57 ( 2003 ) [ arxiv : hep - ph/0307198 ] ; in _ lightcone physics : hadrons and beyond _ , p182 ( durham ippp , 2003 ) [ arxiv : hep - ph/0310048 ] ; + l. theussl , s. noguera and v. vento , eur .
j. a * 20 * , 483 ( 2004 ) [ arxiv : nucl - th/0211036 ] ; + f. bissey _
et al . _ ,
b * 587 * , 189 ( 2004 ) [ arxiv : hep - ph/0310184 ] ; + m. praszalowicz and a. rostworowski , acta phys.polon .
b * 34 * , 2699 ( 2003 ) [ arxiv : hep - ph/0302269 ] . | we study the behaviour of lightcone wavefunctions , in coarse transverse lattice gauge theory , when one or more parton lightcone momenta are small .
this probes the limit of hadron structure at large and small bjorken @xmath0 .
finite - energy boundary conditions on boundstates allow one to derive the analytic form of wavefunctions in this region .
this leads to simple , universal predictions for the behaviour of quark generalised parton distributions near their endpoints .
for the first few meson wavefunctions at large @xmath1 , we give the simplest ansatz that incorporates all the boundary conditions .
+ .2 in |
as a successful geometric theory of gravitation , einstein s theory of general relativity ( gr ) has been confirmed in all observations devoted to its testing to date @xcite , in particular in famous experiments @xcite .
however , the pursuit of testing gravity at much higher precision has ever continued in the past decades , including measurements of the earth - moon separation as a function of time through lunar laser ranging @xcite . on the other hand , formulating and
quantitatively interpreting the test of gravity is another question and an interesting proposal in this respect has been formulated in the frameworks of the parameterized post - newtonian ( ppn ) framework @xcite .
different from the original physical indication @xcite , scale independent post - newtonian parameter denoted by @xmath4 , with @xmath7 representing gr , may serve as a test of the theory on large distances .
this paper is focused on the quantitative constraints of the gr as a theory of gravity , using the recently - released large sample of galaxy scale strong gravitational lensing systems discovered and observed in slacs , bells , lsd , and sl2s surveys @xcite .
up to now , most of the progress in strong gravitational lensing has been made in investigating cosmological parameters @xcite , the distribution of matter in massive galaxies acting as lenses @xcite , and the photometric properties of background sources at cosmological distances @xcite . all the above mentioned results
have been obtained under the assumption that gr is valid . using strong lensing systems @xcite reported the value for the present - day matter density @xmath8 ranging from 0.2 to 0.3 at 99% confidence level .
this initial result , confirmed in later strong lensing studies ( e.g. @xcite ) is consistent with most of the current data including precision measurements of type ia supernovae @xcite and the anisotropies in the cosmic microwave background radiation @xcite .
currently , the concordance @xmath1cdm model is in agreement with most of the available cosmological observations , in which the cosmological constant contributing more than 70% to the total energy of the universe is playing the role of an exotic component called dark energy responsible for accelerated expansion of the universe .
however , there appeared noticeable tensions between different cosmological probes .
for example , regarding the @xmath9 there is a tension between the cmb results from planck @xcite and the most recent type ia supernovae data @xcite .
similarly , the @xmath10 parameter derived from the cmb results from planck @xcite turned out to be in tension with the recent tomographic cosmic shear results both from the canada france hawaii telescope lensing survey ( cfhtlens ) @xcite and the kilo degree survey ( kids ) @xcite .
these tensions partly motivate the test of gr performed in the present paper . with reasonable prior assumptions and independent measurements concerning background cosmology and internal structure of lensing galaxies
, one can use strong lening systems as another tool to constrain the ppn parameters describing the deviations from the gr .
this idea was first adopted on 15 slacs lenses by @xcite , who found the post - newtonian parameter to be @xmath11 based on priors on galaxy structure from local observations .
more recently , @xcite re - examined the expanded slacs sample @xcite and obtained a constraint on the ppn parameter @xmath12 .
having available reasonable catalogs of strong lenses : containing more than 100 lenses , with spectroscopic as well as astrometric data obtained with well defined selection criteria @xcite , the purpose of this work is to use a mass - selected sample of 80 early - type lenses compiled from slacs , bells , lsd , and sl2s to provide independent constraints on the post - newtonian parameter @xmath4 . throughout this paper
we assume a flat @xmath1cdm cosmology with parameters based on the recent _ planck _ observations @xcite .
our goal will be to constrain deviations from general relativity at the level of @xmath4 post - newtonian parameter .
the ppn form of the schwarzschild metric can be written as @xmath13 general relativity corresponds to @xmath2 . from the theory of gravitational lensing @xcite , for a specific strong lensing system with the intervening galaxy acting as a lens
, multiple images can form with angular separations close to the so - called einstein radius @xmath14 : @xmath15 where @xmath16 is the mass enclosed within a cylinder of radius equal to the einstein radius , @xmath17 is the distance to the source , @xmath18 is the distance to the lens , and @xmath19 is the distance between the lens and the source .
all the above mentioned distances are angular - diameter distances . rearranging terms with @xmath20 ( @xmath21 is the cylindrical radius perpendicular to the line of sight the @xmath22-axis )
, we obtain a useful formula : @xmath23 which indicates that only the matter within the einstein ring is important according to the gauss s law .
on the other hand , spectroscopic measurements of central velocity dispersions @xmath24 in elliptical galaxies , can provide a dynamical estimate of this mass , based on power - law density profiles for the total mass density , @xmath25 , and luminosity density , @xmath26 @xcite : @xmath27 here @xmath28 is the spherical radial coordinate from the lens center : @xmath29 . in order to characterize anisotropic distribution of three - dimensional velocity dispersion pattern , one introduces @xcite an anisotropy parameter @xmath30 @xmath31 where @xmath32 and @xmath33 are , respectively , the tangential and radial components of the velocity dispersion . in the current analysis
we will consider anisotropic distribution @xmath34 and assume , as it almost always is assumed , that @xmath30 is independent of @xmath28 .
following the well - known spherical jeans equation @xcite , the radial velocity dispersion of the luminous matter @xmath35 in the early - type lens galaxies can be expressed as @xmath36 where @xmath30 is a constant velocity anisotropy parameter . combining the mass density profiles in eq .
( [ eq : rhopl ] ) , we obtain the relation between the mass enclosed within a spherical radius @xmath28 and @xmath16 as @xmath37 where by @xmath38 we denoted the ratio of respective euler s gamma functions . simplifying the formulae with the notation :
@xmath39 taken after @xcite , we obtain a convenient form for the radial velocity dispersion by scaling the dynamical mass to the einstein radius : @xmath40 \frac{2}{\sqrt{\pi}\left(\xi- 2 \beta \right ) \lambda(\alpha ) } \left(\frac{r}{r_e}\right)^{2 - \alpha}\ ] ] in all strong lensing measurements we use , the _ observed _ velocity dispersion is reported , which is a projected , luminosity weighted average of the radially - dependent velocity dispersion profile of the lensing galaxy .
its theoretical value can be calculated from the eq .
( [ eq : binney ] ) with the assumption that the relationship between stellar number density and stellar luminosity density is spatially constant .
this assumption is unlikely to be violated appreciably within the effective radius of the early - type lens galaxies under consideration . for the purpose of our analysis , the angular diameter distances @xmath48 between reshifts @xmath49 and @xmath50
were calculated using the best - fit matter density parameter @xmath8 given by _
collaboration assuming flat frw metric @xcite .
moreover , we allow the luminosity density profile to be different from the total - mass density profile , i.e. , @xmath51 , and the stellar velocity anisotropy exits , i.e. , @xmath34 . based on a well - studied sample of nearby elliptical galaxies from @xcite ,
the anisotropy @xmath30 is characterized by a gaussian distribution , @xmath52 , which is also extensively used in the previous works @xcite .
more recently , @xcite measured the stellar velocity anisotropy parameter @xmath30 and its correlations with redshifts and stellar velocity dispersion , based on the illustris simulated early - type galaxies with spherically symmetric density distributions .
it is worth noting from their results that @xmath30 markedly depends on stellar velocity dispersion and its mean value varies from 0.10 to 0.30 for intermediate - mass galaxies ( @xmath53 ) , which is consistent with the values used in our analysis .
following our previous analysis @xcite concerning power - law mass and luminosity density profiles of elliptical galaxies , we used a mass - selected sample of strong lensing systems , taken from a comprehensive compilation of strong lensing systems observed by four surveys : slacs , bells , lsd and sl2s .
the sample has been defined by restricting the velocity dispersions of lensing galaxies * to * the intermediate range : @xmath53 .
lenses of this sub - sample are located at redshifts ranging from @xmath54 to @xmath55 .
original data about these strong lenses were derived by @xcite , and more comprehensive data concerning these systems can be found in table 1 of @xcite . fig .
[ fig1 ] shows the scatter plot for this sample in the plane spanned by the redshift of the lens and its velocity dispersion . ]
because @xmath46 , and @xmath47 could not be independently measured for each lensing galaxy , we firstly treated them as free parameters and inferred @xmath46 , @xmath47 , @xmath4 , simultaneously . performing fits on the strong lensing data - set ,
the 68% confidence level uncertainties on the three model parameters are @xmath56 fig .
[ fig2 ] shows these constraints in the parameter space of @xmath46 , @xmath47 , and @xmath4 .
it is obvious that fits on @xmath46 and @xmath47 are well consistent with the analysis results of @xcite , which are characterized by gaussian distributions : @xmath57 more importantly , the degeneracy between the two parameters , @xmath4 and @xmath47 , is apparently indicated by the results presented in fig .
[ fig2 ] , i.e. , a steeper luminosity - density profile for the lensing early - type galaxies will lead to a larger value for the parameterized post - newtonian parameter .
this tendency could also be seen from the sensitivity analysis shown below . now the parameters characterizing the total mass - profile shape , velocity anisotropy , and light - profile shape of lenses are set at their best measured values . performing fits on @xmath4
, we find the resulting posterior probability density shown in fig .
the result @xmath0 ( 1@xmath24 confidence ) is consistent with @xmath7 and also with previous results of @xcite obtained with strong lensing systems .
the scatter of galaxy structure parameters is an important source of systematic errors on the final result . taking the best - fitted values of the structure parameters as our fiducial model
, we investigated how the ppn constraint is altered by introducing the uncertainties on @xmath46 , @xmath30 , and @xmath47 as listed in eq .
( [ eq : alphabeta_pars ] ) .
therefore , firstly , we perform a sensitivity analysis , varying the parameter of interest while fixing the other parameters at their best - fit values . in general
, one can see from table [ priorresult ] and fig .
[ fig7 ] that constraint on @xmath4 is quite sensitive to small systematic shifts in the adopted lens - galaxy parameters . by comparing the contribution of each of these systematic errors to the systematic error on @xmath4 , we find that the largest sources of systematic error are the mass density slope @xmath46 , followed by the anisotropy parameter of velocity dispersion @xmath30 and and the luminosity density slope @xmath47 . secondly , by considering the intrinsic scatter of @xmath46 , @xmath30 , and @xmath47 into consideration
, we found @xmath4 varying from 0.845 to 1.240 at 1@xmath24 confidence level .
it means that systematic errors might exceed @xmath3 of the final result .
the large covariances of @xmath4 with @xmath46 and @xmath47 seen in fig .
[ fig2 ] motivate the future use of auxiliary data to improve constraints on @xmath46 , @xmath30 and @xmath47 .
for example , @xmath46 can be inferred for individual lenses from high resolution imaging of arcs @xcite , while constraints on @xmath30 and @xmath47 can be improved with integral field unit ( ifu ) data @xcite , without the assumption of general relativity ( gr ) .
parameter , the total - mass and luminosity density parameters obtained from the sample of strong lensing systems .
blue crosses denote the best - fit values .
[ fig2 ] ] parameter obtained with rigid priors on the nuisance parameters ( @xmath46 , @xmath30 , @xmath47 ) .
[ fig3 ] ] another issue which should be discussed is by how much is @xmath4 affected by the uncertainty of cosmological parameters of the @xmath1cdm model used in our study . for this purpose , we also considered wmap9 result of @xmath58 in order to make comparison with
_ planck _ observations .
not surprisingly , the result was that differences were negligible .
it could have been expected because cosmology intervenes here through the distance ratio @xmath59 , which is very weakly dependent on the value of @xmath8 and in flat cosmology does not depend on @xmath9 at all .
the next generation wide and deep sky surveys with improved depth , area and resolution may , in the near future , increase the current galactic - scale lens sample sizes by orders of magnitude @xcite .
such a significant increase of the number of strong lensing systems will considerably improve the constraints on the ppn parameter .
now we will illustrate what kind of result one could get using the future data from the forthcoming large synoptic survey telescope ( lsst ) survey , which may detect 120000 lenses for the most optimistic scenario @xcite . in order to make a good comparison with the results derived with current strong lensing systems ( fig .
2 ) , we firstly turn to the simulated lsst population containing @xmath60 lensing galaxies with intermediate velocity dispersions ( @xmath61 ) . performing fits on this simulated strong lensing data - set
, we obtain the constraints in the parameter space of @xmath46 , @xmath47 , and @xmath4 shown in fig .
it is apparent that from the simulated lsst strong lensing data , we may expect the total - mass density parameter @xmath46 to be estimated with @xmath62 precision .
however , the degeneracy between the ppn @xmath4 parameter and the luminosity density parameter @xmath47 still needs to be investigated with future high - quality integral field unit ( ifu ) data @xcite . in the next section
, we will apply a cosmological - independent method to study the degeneracy @xcite between cosmic curvature and parameterized post - newtonian parameter @xmath4 .
.[priorresult ] sensitivity of constraints on @xmath4 with respect to the galaxy structure parameters .
[ cols="^,^",options="header " , ] parameter , the total - mass and luminosity density parameters obtained from the simulated lsst strong lensing data .
in a homogeneous and isotropic universe , the dimensionless distance @xmath63 can be written as @xmath64,\ ] ] where @xmath65 is the expansion rate , and @xmath66 is the spatial curvature density parameter ; @xmath67 for @xmath68 , @xmath69 for @xmath70 , and @xmath71 for @xmath72 , respectively . for a strong lensing system with the following notation : @xmath73 , @xmath74 , @xmath75 , and @xmath76 , a simple sum rule could be easily obtained as @xmath77 [ the case of eq .
( [ dls2 ] ) is given in , e.g. , @xcite , p336 .
] this fundamental formula provides an model - independent probe to test both the spatial curvature , in combination with weak lensing and baryon acoustic oscillations ( bao ) measurements @xcite and the flrw metric , in combination with strong lensing systems and sne ia observations @xcite . for the purpose of our analysis
, we determined the dimensionless distances @xmath78 and @xmath79 of all `` observed '' strong lensing systems ( taken from the lsst simulation by @xcite ) by fitting a polynomial to the union2.1 sn ia data covering the redshift range @xmath80 @xcite .
therefore we bypassed the need to assume any specific cosmological model . by using eq .
( [ dls2 ] ) we were able to calculate the distance ratio @xmath81 depending only on the curvature density parameter @xmath82 .
the reported statistical and systematic uncertainties of the distance modulus for individual sne ia are considered in the fitting procedure . in the union2.1 sn ia compilation ,
light - curve fitting parameters which are used for distance estimation are constrained in a global fit .
however , compared to the uncertainties in the modeling of the strong lensing systems , the model - dependence of the sne ia analysis is likely subdominant @xcite .
then we assessed the distance ratios @xmath81 from the strong lensing data ( einstein radius and velocity dispersion ) according to the eq .
[ eq : plsig ] . for this purpose
we used the simulated observations of forthcoming photometric lsst survey @xcite . using the simulation programs available on the github.com/tcollett/lenspop , we obtained 53000 strong lensing systems meeting the redshift criterion @xmath83 in compliance with snia data used in parallel .
the simulated catalog is derived on the base of realistic population models of elliptical galaxies acting as lenses , with the mass distribution approximated by the singular isothermal ellipsoids .
following the assumptions underlying the simulation , we fixed @xmath84 and @xmath85 in our analysis .
we took the fractional uncertainty of the einstein radius at the level of 1% and the observed velocity dispersion at the level of 10% .
secondary lensing contribution from the matter along the line - of - sight was neglected in our analysis .
uncertainties in the einstein radii @xcite .
however , according to the recent analysis by @xcite , the lines - of - sight for monitorable strong lenses ( especially for quadruply imaged quasars ) might be biased at the level of 1% . some attempts to account for the line - of - sight secondary lensing for quasars
can also be found in @xcite .
[ fig4 ] displays the fitting results in the @xmath86 plane , thus illustrating the dependence between the cosmic curvature and the ppn @xmath4 parameter .
it is apparent that a flat universe together with the validity of gr ( @xmath70 , @xmath2 ) is strongly supported .
more importantly , it is interesting to note that there exists a significant degeneracy between the spatial curvature of the universe and the ppn parameter , which captures how much space curvature is provided by unit rest mass of the objects along or near the path of the particles .
similar degeneracy between @xmath4 and the other cosmological parameters ( the matter density fraction , @xmath8 and the equation of state of dark energy , @xmath87 ) can also be seen from fig .
[ fig5 ] . ] ] one can easily check that reduction of the error of @xmath82 would lead to more stringent fits of @xmath4 , which encourages us to consider the possibility of testing ppn at much higher accuracy with future surveys of strong lensing systems .
we now set a prior on the cosmic curvature with @xmath88 , according to the latest cmb data and baryon acoustic oscillation data @xcite , and get a constraint on the ppn parameter : @xmath6 .
when we changed the fractional uncertainty of the einstein radius to the level of 1% and the observed velocity dispersion to the level of 5% , the resulting constraint on the ppn parameter became : @xmath89 .
the posterior probability density for @xmath4 is shown in fig .
one can see from this plot that much more severe constraints would be achieved , and one can expect @xmath4 to be estimated with @xmath90 precision . from _
planck_. [ fig6 ] ]
based on a mass - selected galaxy - scale strong gravitational lenes from the slacs , bells , lsd and sl2s surveys and a well - motivated fiducial set of lens - galaxy parameters , we tested the weak - field metric on kiloparsec scales and found a constraint on the post - newtonian parameter @xmath0 under the assumption of a flat universe from _ planck _ observations .
therefore it is in agreement with the general relativity value of @xmath2 with @xmath91 accuracy . considering systematic uncertainties in total mass - profile shape , velocity anisotropy , and light - profile shape ,
we estimate systematic errors to be @xmath3 . furthermore , we illustrated what kind of result we could get using the future data from the forthcoming large synoptic survey telescope ( lsst ) survey @xcite .
we applied a cosmological model independent method to study the degeneracy @xcite between cosmic curvature and the parameterized post - newtonian parameter @xmath4 .
it is apparent that spatially flat universe with the conservation of gr ( @xmath70 , @xmath2 ) is strongly supported .
moreover , the reduced uncertainty of @xmath82 leads to more stringent fits of @xmath4 .
this opens up the possibility of testing ppn with much higher accuracy using strong lensing systems discovered in the future surveys . by setting a prior on the cosmic curvature with @xmath88 , assumed according to the latest cmb plus baryon acoustic oscillation data @xcite , the accuracy of @xmath4 determination reached @xmath90 precision .
therefore , we conclude that samples of strong lensing systems with measured stellar velocity dispersions , much larger than currently available , may serve as an important probe to test the validity of the gr , provided that mass - dynamical structure of lensing galaxies is better characterized and constrained in the future surveys . this work was supported by the ministry of science and technology national basic science program ( project 973 ) under grants nos .
2012cb821804 and 2014cb845806 , the strategic priority research program
the emergence of cosmological structure " of the chinese academy of sciences ( no .
xdb09000000 ) , the national natural science foundation of china under grants nos . 11503001 , 11373014 and 11073005 , the fundamental research funds for the central universities and scientific research foundation of beijing normal university , china postdoctoral science foundation under grant no .
2015t80052 , and the opening project of key laboratory of computational astrophysics , national astronomical observatories , chinese academy of sciences .
part of the research was conducted within the scope of the hecols international associated laboratory , supported in part by the polish ncn grant dec-2013/08/m
/ st9/00664 - m.b . gratefully acknowledges this support
. m.b . obtained approval of foreign talent introducing project in china and gained special fund support of foreign knowledge introducing project .
ade , p.a .
r. , et al .
2014 , a&a , 571 , a16 amanullah , r. , et al .
2010 , apj , 716 , 712 ashby , n. 2002 , phys . today , 55 , 41 auger , m. w. , et al .
2009 , apj , 105 , 1099 barnab , m. , et al .
2013 , mnras , 436 , 253 bernstein , g. 2006 , apj , 637 , 598 bertotti , b. , iess , l. , & tortora , p. 2003 , nature , 425 , 374 biesiada , m. 2006 , prd , 73 , 023006 biesiada , m. , pirkowska , a. , & malec , b. 2010 , mnras , 406 , 1055 binney , j. 1980 , mnras , 190 , 873 bolton , a. s. , rappaport , s. , & burles , s. 2006 , prd , 74 , 061501 bolton , a. s. , et al .
2008 , apj , 682 , 964 bonvin , v. , et al .
2016 , arxiv:1607.01790 brownstein , et al .
2012 , apj , 744 , 41 cardone , v. f. , piedipalumbo , e. , scudellaro , p. 2016
mnras , 455 , 831 - 837 .
cao , s. , & zhu , z .- h .
2012 , a&a , 538 , a43 cao , s. , covone , g. , & zhu , z .- h .
2012 , apj , 755 , 31 cao , s. , pan , y. , biesiada , m. , godlowski , w. , & zhu , z .- h .
2012 , jcap , 03 , 016 cao , s. , zhu , z .- h . , & zhao , r. 2012 , prd , 84 , 023005 cao , s. , & zhu , z .- h .
2014 , prd , 90 , 083006 cao , s. , et al .
, 2015a , apj , 806 , 185 cao , s. , et al .
, 2015b , aj , 149 , 3 cao , s. , et al . ,
2016 , mnras , 461 , 2192 chae , k .- h . 2003 , mnras , 346 , 746 chae , k .- h . , chen , g. , ratra , b. , & lee , d .- w .
2004 , apj , 607 , l71 collett , t. e. , et al .
2013 , mnras , 432 , 679 collett , t. e. & auger , m. w. 2014 , mnras , 443 , 969 collett , t. e. 2015 , arxiv:1507.02657 collett , t. e. & cunnington , s. d. 2016 , mnras , 462 , 3255 dyson , f. w. , eddington , a. s. , & davidson , c. 1920 , phil .
r. soc . , 220 , 291 gerhard , o. , kronawitter , a. , saglia , r. p. , & bender , r. 2001 , aj , 121 , 1936 golse , g. , kneib , j .-
p . , & soucail , g. 2002 , a&a , 387 , 788 greene , z. s. , et al .
2013 , apj , 768 , 39 grillo , c. , lombardi , m. , & bertin , g. 2008 , a&a , 477 , 397 heymans , c. , et al .
2012 , mnras , 427 , 146 hilbert , s. , et al .
2009 , a&a , 499 , 31 hildebrandt , h. , et al .
2016 , preprint [ arxiv:1603.07722 ] jin , k .- j . , zhang , y .- z .
, & zhu , z .- h . 2000 ,
pla , 264,335 keeton , c. r. 2001 , apj , 561 , 46 koopmans , l.v.e , & treu , t. 2002 , apj , 583 , 606 koopmans , l.v.e .
2006 , in eas publications series , ed .
g. a. mamon , f. combes , c. deffayet , & b. fort , vol .
20 , 161 kuhlen , m. , keeton , c. r. , & madau , p. 2004
, apj , 601 , 104 maccrann , n. , zuntz , j. , bridle , s. , jain , b. , becker , m. r. 2015 , mnras , 451 , 2877 mao , s. d. , & schneider , p. 1998
, mnras , 295 , 587 marshall , p. , blandford , r. , & sako , m. 2005 , nar , 49 , 387 mitchell , j. l. , keeton , c. r. , frieman , j. a. , & sheth , r. k. 2005 , apj , 622 , 81 ofek , e. o. , rix , h .- w . , & maoz , d. 2003 , mnras , 343 , 639 oguri , m. , et al .
2008 , aj , 135 , 512 peebles , p. j. e. 1993 , principles of physical cosmology ( princeton university press , princeton , nj , 1993 pound , r. v. , & rebka , g. a. 1960 , prl , 4 , 337 rsnen , s. , bolejko , k. , & finoguenov , a. 2015 , prl , 115 , 101301 riess , a. g. , et al .
2016 , apj , 826 , 56 rusu , c. e. , et al .
2016 , arxiv:1607.01047 schneider , p. , ehlers , j. , & falco , e. e. 1992 , gravitational lenses ( springer - verlag , new york ) schwab , j. , bolton , a. s. , & rappaport , s. a. 2010 , apj , 708 , 750 shapiro , i. i. 1964 , phys .
lett . , 13 , 789 sonnenfeld , a. , gavazzi , r. , suyu , s.h . ,
treu , t. , marshall , p.j . 2013a , apj , 777 , 97 [ arxiv:1307.4764 ] sonnenfeld , a. , treu , t. , gavazzi , r. , suyu , s.h . ,
marshall , p.j .
, auger , m.w . ,
nipoti , c. , 2013b , apj , 777 , 98 suyu , s. h. , et al . 2007 aas / aapt joint meeting , american astronomical society meeting 209 , id.21.02 ; bulletin of the american astronomical society , vol .
38 , p.927 taylor , j. h. , fowler , l. a. , & mcculloch , p. m. 1979 , nature , 277 , 437 thorne , k. s. , & will , c. m. 1971 , apj , 163 , 595 treu , t. , & koopmans , l.v.e .
2002 , apj , 575 , 87 treu , t. , & koopmans , l.v.e .
2004 , apj , 611 , 739 treu , t. , et al . 2006a , apj , 640 , 662 vegetti , s. , koopmans , l. v. e. , bolton , a. , treu , t. , & gavazzi , r. 2010 , mnras , 408 , 1969 williams , j. g. , turyshev , s. g. , & boggs , d. h. 2004 , prl , 93 , 261101 wong , k. c. , suyu , s. h. , & matsushita , s. 2015 , apj , 811 , 115 xu , d. d. , et al .
2016 , arxiv:1610.07605v1 zhu , z .- h .
, & wu , x .-
1997 , a&a , 324 , 483 zhu , z .- h .
2000 , mpla , 15 , 1023 zhu , z .- h .
2000 , ijmpd , 9 , 591 zhu , z .- h . , & sereno , m. 2008 , a&a , 487 , 831 zhu , z .- h . , et al .
2008 , a&a , 483 , 15 | based on a mass - selected sample of galaxy - scale strong gravitational lenses from the slacs , bells , lsd and sl2s surveys and using a well - motivated fiducial set of lens - galaxy parameters we tested the weak - field metric on kiloparsec scales and found a constraint on the post - newtonian parameter @xmath0 under the assumption of a flat @xmath1cdm universe with parameters taken from _
planck _ observations .
general relativity ( gr ) predicts exactly @xmath2 .
uncertainties concerning the total mass density profile , anisotropy of the velocity dispersion and the shape of the light - profile combine to systematic uncertainties of @xmath3 . by applying a cosmological model independent method to the simulated future lsst data
, we found a significant degeneracy between the ppn @xmath4 parameter and spatial curvature of the universe . setting a prior on the cosmic curvature parameter @xmath5
, we obtained the following constraint on the ppn parameter : @xmath6 .
we conclude that strong - lensing systems with measured stellar velocity dispersions may serve as another important probe to investigate validity of the gr , if the mass - dynamical structure of the lensing galaxies is accurately constrained in the future lens surveys . |
back in 1968 , ashcroft predicted that according to the bcs theory@xcite , dense hydrogen would not only be metallic@xcite , but more importantly also a high - temperature superconductor@xcite .
since recently it has been shown that dissociation is a necessary condition for the metallization of hydrogen@xcite , the necessary pressure to cause bandgap closure has remained impractical high , so that metallic hydrogen has only been realized at finite - temperature@xcite . yet , an appealing way to circumvent the high pressures required to metallize
hydrogen is to precompress it in hydrogen - rich systems@xcite , since , in general , the electronic density of states is high and the electron - phonon interactions are strong@xcite .
in fact , pressurized hydrogen - rich materials have demonstrated to be rather promising candidates for high - t@xmath3 superconductivity @xcite . in particular , drozdov and eremets
reported that at pressures around 200 gpa , dense hydrogen sulphide becomes metallic and superconducting with a critical temperature ( @xmath4 ) of 203 k@xcite , which is well above the highest t@xmath3 of 133 k and 164 k that were achieved in cuprates@xcite at ambient@xcite and high pressures@xcite respectively .
recent experimental results indicate that the superconducting state of h@xmath2s adopts a body - centered cubic ( bcc ) structure@xcite .
however , nearly all of the recent calculations on hydrogen - rich systems are based on the single - particle mean - field theories such as density - functional theory ( dft)@xcite . even though formally exact , the exact exchange and correlation ( xc ) functional is unknown from the outset and needs to be approximated , which affects both the relative stabilities of the different crystal structures .
indeed , for dense hydrogen - rich materials a significant dependence on the particular of xc functional was established @xcite .
therefore , in this paper we revisit the stability of the individual products originating from the various proposed decompositions of h@xmath0s for pressures above 150 gpa by means of highly accurate diffusion monte carlo ( dmc ) simulations . using the dmc method@xcite ,
the electronic many - body schrdinger equation is solved stochastically , which have yielded very accurate total energies for atoms@xcite , molecules@xcite , crystals@xcite including hydrogen - rich materials at very high pressure@xcite .
at first , all of the examined structures were determined by relaxing the internal parameters of each phase within dft at fixed external pressure .
the dft calculations were all conducted within the pseudopotential and plane - wave approach using the castep code@xcite . specifically , ultrasoft pseudopotentials were employed together with an energy cutoff of 1000 ev@xcite .
the exact xc functional was substituted by the generalized gradient approximation of perdew - burke - ernzerhof ( pbe).@xcite the geometry and cell optimizations were conducted using a dense @xmath5 @xmath6-point mesh to sample the brillouin zone , while the nuclear forces and components of the stress tensor were converged to 0.01 ev/@xmath7 and 0.01 gpa , respectively .
the casino code was used to perform fixed - node dmc simulations with a trial wave function of the slater - jastrow ( sj ) form@xcite , @xmath8 \det[\psi_{n}({\bf r}_i^{\uparrow } ) ] \det[\psi_{n}({\bf r}_j^{\downarrow } ) ] , \label{eq1}\ ] ] where @xmath9 is a @xmath10-dimensional vector containing the positions of all @xmath11 electrons , @xmath12 the position of the @xmath13th spin - up electron , @xmath14 the position of the @xmath15th spin - down electron , @xmath16 $ ] a jastrow factor , while @xmath17 $ ] and @xmath18 $ ] are slater determinants made of spin - up and spin - down one - electron wave functions . these orbitals were obtained from pbe - dft calculations performed with the castep plane - wave code@xcite , in conjunction with trail - needs hartree - fock pseudopotentials@xcite . for the purpose to approach the complete basis set limit@xcite , a large energy cut - off of 4000 ev have been chosen .
the resulting plane - wave orbitals were subsequently transformed into a localized `` blip''polynomial basis@xcite . the jastrow factor within eq .
[ eq1 ] is a positive , symmetric , explicit function of interparticle distances .
the employed jastrow factor includes polynomial one - body electron - nucleus ( 1b ) , two - body electron - electron ( 2b ) and three - body electron - electron - nucleus ( 3b ) terms , as well as plane - wave expansions of the electron - electron separation known as @xmath19 terms@xcite .
these @xmath19 terms build long - ranged correlations into the jastrow factor and thus significantly improve the wave function and variational energy .
we also investigated the effect of the inhomogenous backflow ( bf ) coordinate transformation on the vmc and dmc total energies@xcite .
our bf transformation includes electron - electron and electron - proton terms and is given by @xmath20 where @xmath21 is the transformed coordinate of electron @xmath13 , which depends on the full configuration of the system @xmath22 .
the vector functions @xmath23 and @xmath24 are the electron - electron and electron - proton backflow displacements of electron @xmath13 .
they are parameterized as @xmath25 and @xmath26 where @xmath27 and @xmath28 are polynomial functions of electron - electron and electron - proton distances that contains variational parameters .
all adjustable parameters in the jastrow factor and backflow terms were optimized by minimizing the variance as well as the variational energy at the vmc level@xcite .
if not explictly stated otherwise , all of our calculations were conducted using the sj trail wave function including 1b , 2b , 3b and p terms augmented by the bf coordinate transformation . beside the usage of twist - averaged boundary conditions ( tabc ) to correct finite - size errors@xcite , we extrapolated the energy per atom to the thermodynamic limit by fitting our twist - averaged dmc results for different system sizes to @xmath29 , where @xmath30 and @xmath31 are fitting parameters and @xmath32 is the eventual energy per atom in the infinite - system limit . depending on simulation cell size , we used 8 , 12 and 16 randomly chosen twists@xcite .
the enthalpy was evaluated by differentiating the polynomial fit of our the finite - size - corrected dmc energies as a function of volume .
.[hs2c2 m ] total energies of the c2/m phase of hs@xmath0 at the dmc level of theory for four different pressures .
the energies are given in ev / atom and are calculated for @xmath33 , @xmath34 and @xmath35 particles in the unit cell , respectively .
the extrapolated dmc energy at the infinite system size limit is denoted by e(@xmath36 ) . [ cols="^,^,^,^,^",options="header " , ] the energy gain of the various trial wave functions with respect to the sj(1 + 2b ) approach using the vmd and dmc methods for the two different systems
are shown in fig .
it is apparent that the addition of the three - body term and more so the @xmath19 term substantially reduces the vmc energy , but tht the dmc is only marginally affected since the initial nodal structure is identical .
the usage of the bf transformation , however , improves the nodal surface and hence entails an energy lowering at the vmc and the dmc levels .
interestingly , the gain in energy for hs@xmath0 is much more pronounced than for h@xmath2s , which demonstrates the importance of an accurate trail wave function for the former . for the im-3 m phase of h@xmath2s the energy gain is -1.634 and -0.56 ev at the vmc and dmc levels of theory , while for the c2/m hs@xmath0 structure , the energy gain can be as high as -4.898 and -1.643 ev . ( colour online ) the reduction of variational energy for the hs@xmath0 and h@xmath2s systems using different trial wave functions with respect to the sj(1b+2b ) approach using the vmc and dmc methods.,scaledwidth=50.0% ] the increased accuracy of the bf wave function , however , comes at a rather high computational cost , which is due to the necessity to evaluate the orbitals and their first two derivatives and moreover also the collective bf coordinates , because every element of the slater matrix must be updated every time a single electron is moved .
even though it is partially compensated by the fact that the less complex bsj(1b+2b ) trail wave function is equally accurate than the much more complex sj(1b+2b+3b+@xmath19 ) calculations , we have elected to use the former instead of the latter in the following dmc calculations of the enthalpy - pressure phase diagram calculations . in order to compute the enthalpy - pressure phase diagrams for the different structures
, we fitted our extrapolated dmc total energies as a function of volume @xmath37 against a model equations of state @xmath38 .
we found that a quadratic polynomial is a sufficiently accurate representation of our actual dmc energies . using this model ,
it is straightforward to calculate the pressure @xmath39 as a function of @xmath37 and thus also the enthalpy per atom @xmath40 .
in previous dft calculations including zpe correction it was predicted that at 200 gpa 5h@xmath0s decomposes into 3h@xmath2s and hs@xmath0 , where hs@xmath0 adopts c2/c symmetry , but undergoes a phase transition to the more stable c2/m structure at 250 gpa@xcite .
however , the present dmc enthalpies indicate that the c2/c hs@xmath0 structure is more stable than the c2/m phase up to 440 gpa , as it is shown in fig .
[ hphs2 ] .
in fact , the enthalpy difference between the c2/c and c2/m phases of hs@xmath0 is much larger than the zpe correction as estimated by dft@xcite .
( colour online ) the dmc enthalpy as a function of pressure of the c2/c hs@xmath0 structure relative to the c2/c phase . the estimated uncertainties in the dmc enthalpies due to statistical and systematic errors
are represented by the widths of the corresponding lines.,scaledwidth=40.0% ] ( colour online ) the dmc enthalpy of the c2/m and i4@xmath1/amd phases of hs as a function of pressure .
the corresponding error bars are smaller than the size of the data points.,scaledwidth=40.0% ] moreover , dft - based crystal structure searches predict that the energetically most favorable phase of hs at 200 gpa is i4@xmath1/amd and at 300 gpa c2/m@xcite .
the present dmc enthalpy - pressure curves of hs are shown in fig .
[ hphs ] . at variance to dft
, they suggest that the c2/m phase of hs is more stable than the i4@xmath1/amd hs structure over the whole pressure range from 150 to 400 gpa .
in fact , the difference in enthalpy between the c2/m and i4@xmath1/amd phases is more than 600 mev / atom and is even larger with increasing pressure .
( colour online ) the dmc enthalpy of the im-3 m and r3 m phases of h@xmath2s as a function of pressure .
the corresponding error bars are smaller than the size of the data points.,scaledwidth=40.0% ] previous crystal structure prediction calculations using dft indicate that the r3 m h@xmath2s structure is stable at 130 gpa and that the im-3 m phase of h@xmath2s is the most likely candidate at pressures larger than 200 gpa . yet
, our dmc enthalpy - pressure calculations again show that the im-3 m h@xmath2s structure is the more favorable candidate up to at least 320 gpa , which is in agreement with very recent experimental results.@xcite this is to say that altogether , our dmc results leads to a revision of the enthalpy - pressure phase diagram between 150 and 320 gpa .
specifically , we find that the c2/c hs@xmath0 , c2/m hs and im-3 m h@xmath2s structures are energetically most favorable within the considered pressure range .
however , it is well - known that the zpe corrections plays an important role in the numerical determination of the phase diagram hydrogen - rich systems@xcite .
nevertheless , dft calculations of others have shown that the energy difference between h@xmath0s and s+h@xmath0 due to zpe is about 6 mev / f.u . at 160 gpa@xcite , which is much smaller than our dmc enthalpy differences .
we thus predict that the effect of zpe on our dmc results is negligible .
in summary , using highly accurate dmc calculations , we find that in the pressure range between 150 and 320 gpa , the c2/c hs@xmath0 , c2/m hs and the superconducting im-3 m h@xmath2s structures are the enthalpically most favorable products of the decomposition of h@xmath0s .
nevertheless , we conclude by noting that instead of the often proposed decomposition of h@xmath0s into h@xmath2s , hs@xmath0 , hs , or rather elemental @xmath41 , the dissociation into h@xmath2s@xmath42 and hs@xmath43 would be chemically much more sensible and still compatible with the recently performed xrd measurements@xcite .
this work made use of computing facilities provided by archer , the uk national super computing service , and by the high performance computing service of imperial college london , as well as of the oculus system of the paderborn center for parallel computing ( pc@xmath44 ) .
in particular , we acknowledge that the results of this research have been achieved using the prace-3ip project ( fp7 ri-312763 ) resource archer based in uk .
support from the thomas young centre under grant number tyc-101 and ep / n50869x/1 is also thankfully acknowledged .
j. m. mcmahon , m. a. morales , c. pierleoni and d. m. ceperley , rev .
84 * , 1607 ( 2012 ) .
s. j. clark , m. d. segall , c. j. pickard , p. j. hasnip , m. j. probert , k. refson and m. c. payne , z. kristallogr . * 220 * , 567 ( 2005 ) . | we use the diffusion quantum monte carlo to revisit the enthalpy - pressure phase diagram of the various products from the different proposed decompositions of h@xmath0s at pressures above 150 gpa .
our results entails a revision of the ground - state enthalpy - pressure phase diagram . specifically , we find that the c2/c hs@xmath0 structure is persistent up to 440 gpa before undergoing a phase transition into the c2/m phase .
contrary to density functional theory , our calculations suggest that the c2/m phase of hs is more stable than the i4@xmath1/amd hs structure over the whole pressure range from 150 to 400 gpa .
moreover , we predict that the im-3 m phase is the most likely candidate for h@xmath2s , which is consistent with recent experimental x - ray diffraction measurements . |
optimal control theory has become one of the most dominant and indispensable techniques for analyzing dynamical systems in which optimal decisions are sought at each moment . surely , the principal part in the establishment of the theory as an important and rich area of applied mathematics arises in the strong utilization of the subject area in a great breadth of applications and research areas such as engineering , computer science , astronautics , biological sciences , chemistry , agriculture , business , management , energy , path planning problems , and a host of many other areas ; cf .
@xcite . the most popular analytical methods for solving optimal control problems such as the calculus of variations , pontryagin s principle , and bellman s principle , can generally solve only fairly simple problems .
however , such methods are largely deficient to handle the increasing complexity of optimal control problems since the advent of digital computers , which led to a revolution in the development of numerical dynamic optimization methods over the past few decades . among the popular numerical methods for solving optimal control problems , the so - called `` direct orthogonal collocation methods '' and `` direct pseudospectral methods '' have become two of the most universal and well established numerical dynamic optimization methods due to many merits they offer over other competitive methods in the literature ; cf .
@xcite . both classes of numerical dynamic optimization methods convert the continuous optimal control problem into a finite dimensional constrained optimization problem based on the elegant spectral and pseudospectral methods , which are known to furnish exponential / spectral convergence rates faster than any polynomial convergence rate when the problem exhibits sufficiently smooth solutions ; cf .
@xcite .
direct hp - pseudospectral methods were specifically designed to handle optimal control problems with discontinuous or nonsmooth states and controls ; cf .
such methods generally recover the prominent exponential convergence rates of pseudospectral methods by dividing the solution domain into an increasing number of mesh intervals and increasing the degree of the polynomial interpolant within each mesh interval .
in particular , a local @xmath0-refinement is a suitable technique on regions where the solution is smooth , while a local @xmath1-refinement is preferable on elements where the solution is discontinuous / nonsmooth . to avoid high computational costs , adaptive strategies strive to control the locations of mesh intervals , minimize their number , and answer the question of whether increasing the number of collocation points within each mesh interval is necessary or not to achieve a certain accuracy threshold . in the most general formulation of an hp finite element method ,
the solution over each element is approximated by an arbitrary degree polynomial .
the spectral element method uses instead a high - degree piecewise polynomial defined by an appropriate set of interpolation nodes or expansion modes . to achieve the highest interpolation accuracy ,
the interior interpolation nodes are distributed at positions corresponding to the zeros of certain families of orthogonal polynomials ; cf . @xcite .
while direct hp - pseudospectral methods were thoroughly investigated in the past few years , comparable literature for direct adaptive spectral element methods for solving special classes of optimal control problems is rather very few , and to the best of our knowledge , it seems that such methods do not exist for solving more general nonlinear optimal control problems .
we acknowledge though the existence of some posteriori error analyses of hp finite element approximations of special forms of convex optimal control problems ; cf .
posteriori error estimates for the spectral element approximation of a linear quadratic optimal control problem in one dimension was recently presented by @xcite .
however , all three papers lacked any adaptive strategies to efficiently implement their numerical schemes .
perhaps , the earliest and sole adaptive spectral element method for solving a special class of optimal control problems described by a quadratic cost functional and linear advection - diffusion state equation was put forward by @xcite . in their presented work , an approximate saddle point of the lagrangian functional
is sought by iterating on the karush - kuhn - tucker optimality conditions to seek their satisfaction numerically using a galerkin spectral element method for the space discretization .
the adaptive algorithm relies on a posteriori error estimate of the cost functional , from which the parameters of the spectral element discretization are selected .
the main purpose of this paper is to derive high - order numerical solutions of nonlinear optimal control problems exhibiting smooth / nonsmooth solutions using a novel direct adaptive gegenbauer integral spectral element ( gise ) method . in particular
, the proposed method converts the nonlinear optimal control problem into an integral multiple - phase optimal control problem .
the multiple - phases are then connected using state continuity linkage conditions with easily incorporated control continuity linkage conditions when the control functions are assumed continuous .
the numerical discretization is carried out using truncated shifted gegenbauer series expansions and a novel numerical quadrature defined on each mesh interval henceforth called the @xmath2th elemental shifted optimal barycentric gegenbauer quadrature ( kesobgq) based on the stable barycentric representation of lagrange interpolating polynomials .
such a quadrature can produce excellent approximations while significantly reducing the number of operational costs required for the evaluation of the involved integrals .
the proposed method is further invigorated by a novel adaptive strategy that uses a multicriterion for locating the mesh intervals where the state and control functions are smooth / nonsmooth based on information derived from the residual of the discrete dynamical system equations , and the magnitude of the last coefficients in the state and control truncated series .
in fact , the idea of using the spectral coefficients of the state trajectories as a measure to verify the convergence of the computed solution was previously presented by @xcite . nonetheless , in this article
, we shall exploit the spectral coefficients instead to check the smoothness of the approximate solutions on the interval of interest .
the proposed method generally produces a small / medium - scale nonlinear programming problem that could be easily solved using the current powerful numerical optimization methods .
the current paper casts further the light on the judicious choice of the shifted gegenbauer - gauss collocation points set to be utilized on each mesh interval during the discretization process of optimal control problems based on numerical simulations .
the remaining part of the paper is organized as follows : in section [ sec : ps ] , we describe the optimal control problem statement under study . in section [ sec :
tgsem ] , we present our novel gise method . a novel adaptive strategy is presented in section [ subsec : as1 ] . section [ subsec : err ] is devoted for the error analysis and convergence properties of the kesobgq . in section [ sec : ne ] , two test examples of nonlinear optimal control problems are included to demonstrate the efficiency and the accuracy of the proposed gise method followed by some concluding remarks illustrating the advantages of the proposed gise method in section [ conc ] .
consider the nonlinear time - varying dynamical system @xmath3 where @xmath4 and @xmath5 are the state and control vector functions , respectively ; @xmath6 is the vector of first - order time derivatives of the states ; @xmath7 is the initial time , @xmath8 is the terminal time .
the problem is to find the optimal control @xmath9 and the corresponding state trajectory @xmath10 satisfying eq . while minimizing the cost functional @xmath11 subject to the mixed state and control path constraints @xmath12 and the boundary conditions @xmath13 where @xmath14 is the terminal cost function , @xmath15 is the lagrangian function , @xmath16 is a nonlinear vector field , @xmath17 is a mixed inequality constraint vector on the state and control functions ; @xmath18 are constant specified vectors ; @xmath19 is a boundary constraint vector .
here it is assumed that @xmath20 , and each system function @xmath21 are nonlinear continuously differentiable functions with respect to @xmath22 .
it is also assumed that the nonlinear optimal control problem has a unique solution with possibly discontinuous / nonsmooth optimal control .
we shall refer to the above optimal control problem in bolza form by problem 1 .
using the affine transformation @xmath23 we could easily rewrite problem 1 as follows : [ pr:2 ] @xmath24,\\ & { \bm{c}_{\min } } \le \tilde { \bm{c}}(\tilde { \bm{x}}(\tau ) , \tilde { \bm{u}}(\tau ) , \tau ) \le { \bm{c}_{\max } } , \quad \tau \in [ - 1,1],\\ & \psi \left ( { \tilde { \bm{x } } ( - 1),{t_0},\tilde { \bm{x}}(1),{t_f } } \right ) = \bm{0},\end{aligned}\ ] ] where @xmath25 .
we refer to the optimal control problem described by eqs . by problem 2 .
one of the primary advantages of spectral element methods is the ability to resolve complex geometries and problems exhibiting discontinuous / nonsmooth solutions with high - order accuracies through the decomposition of the solution interval into small mesh intervals or elements `` @xmath1-refinement , '' and approximating the restricted solution function on each mesh interval with high - order truncated spectral expansion series `` @xmath0-refinement . ''
considering the solution interval @xmath26 $ ] , we can partition it into @xmath27 mesh intervals @xmath28 using @xmath29 mesh points @xmath30 distributed along the interval @xmath26 $ ] : @xmath31 = \bigcup\limits_{k = 1}^k { { \mkern 1mu } { \bm{\omega } _ k } } , \quad { \bm{\omega } _ k } = [ { \tau _ { k - 1}},{\tau _ k}],\quad - 1 = { \tau _ 0 } <
{ \tau _ 1 } < \ldots < { \tau _ k } = 1.\ ] ] we denote the state and control vector functions in the @xmath2th element by @xmath32 and @xmath33 , respectively . based on this initial setting , we can put problem 2 into its multiple - interval form as follows : [ pr:3 ] @xmath34 to take advantage of the well - conditioning of numerical integration operators , we further rewrite eq . in its integral formulation so that @xmath35 to impose the states continuity conditions , the following conditions must be fulfilled at the interface of any two consecutive mesh intervals : @xmath36 or equivalently , @xmath37 if the control vector function is assumed to be continuous , then we further add either of the following two sets of constraints : @xmath38 we refer to the optimal control problem , , , , provided with any coupled sets of conditions or and conditions or by problem 3 .
let @xmath39 be the @xmath40th - degree shifted gegenbauer polynomial defined on the mesh interval @xmath41 henceforth referred to by the @xmath40th - degree @xmath2th elemental shifted gegenbauer polynomial , where @xmath42 is the classical @xmath40th - degree gegenbauer polynomial associated with the real parameter @xmath43 ; cf .
moreover , let @xmath44 denote the set of the zeroesth elemental shifted gegenbauer - gauss ( kesgg ) nodes in @xmath41 .
] of the @xmath45th - degree @xmath2th elemental shifted gegenbauer polynomial , @xmath46 , for some @xmath47 , and set @xmath48 .
the @xmath2th elemental shifted gegenbauer polynomials @xmath49 form a complete @xmath50-orthogonal system with respect to the weight function , @xmath51 and their orthogonality relation is defined by the following weighted inner product : @xmath52 where @xmath53 is the kronecker delta function , @xmath54 is the normalization factor , and @xmath55 is as defined by ( * ? ? ?
* eq . ( 2.6 ) ) . for @xmath56 and @xmath57 , we recover the shifted chebyshev polynomials of the first kind and the shifted legendre polynomials , respectively , on each mesh interval @xmath58 .
let @xmath59 , @xmath60\forall l \in \mathbb{z}_0^ + , \ ] ] and denote the identity matrix of order @xmath61 by @xmath62 .
moreover , define @xmath63 $ ] and @xmath64 $ ] as the spectral coefficient vectors obtained through collocating the state and control vectors at the augmented kesgg nodes @xmath65 , respectively , where @xmath66^t},{\bm{b}}_s^{(k ) } = { \left [ { b_{s,0}^{(k)},b_{s,1}^{(k ) } , \ldots , b_{s,{l_{u , k}}}^{(k ) } } \right]^t}\ ; \forall k \in \mathbb{k } , r = 1 , \ldots , n_x ; s = 1 , \ldots , n_u$ ] , and `` vec '' denotes the vectorization of a matrix .
as promised by the sturm - liouville theorem , we can represent any square integrable function as an infinite series in the shifted gegenbauer polynomials ; therefore , we can approximate the state and control vector functions as follows : [ eq : statesandcontrolsapprox1 ] @xmath67 where `` @xmath68 '' denotes the kronecker product .
let @xmath69 denote the first - order optimal barycentric gegenbauer integration matrix ( obgim ) ; cf . @xcite . with a simple mathematical manipulation
, we can easily show that the first - order @xmath2th elemental shifted optimal barycentric gegenbauer integration matrix ( kesobgim ) , @xmath70 , on @xmath71 is related to the obgim by the following useful relation : @xmath72 we shall refer to the set @xmath73 by the set of adjoint kesgg points on @xmath74 such that @xmath75 are the associated optimal gegenbauer parameters ; cf .
denote the @xmath76th row of the kesobgim , @xmath77 $ ] , by @xmath78 .
the sought discrete cost function @xmath79 can be written as @xmath80 where @xmath81 is the all ones vector , @xmath82 is the all alternating ones vector for all @xmath83 , and @xmath84^t.\end{aligned}\ ] ] to account for the state continuity conditions , say eqs .
, the discrete integral dynamical system equations on the elemental domains ( or simply elements ) can be approximated by [ eq : discdynsys1 ] @xmath85 @xmath86 where @xmath87^t.\end{aligned}\ ] ] furthermore , the discrete path and boundary constraints are given by @xmath88 @xmath89 the discrete control continuity constraints are imposed as follows : @xmath90 hence , the optimal control problem has been reduced to a nonlinear programming problem in which we seek the minimization of the objective function @xmath91 defined by eq .
subject to the generally nonlinear constraints , , , and the linear constraints .
the present gise method adopts both collocation and interpolation techniques to obtain the sought approximations .
in particular , the spectral coefficient vectors are determined through collocation at the augmented kesgg nodes on each element @xmath71 , while the kesobgims are constructed through interpolation at the kesgg nodes . in this section
, we present a multicriterion for locating the elements where the state and control functions are smooth / nonsmooth based on : (
i ) the maximum residual of the discrete dynamical system equations ; i.e. , checking whether the state and control variables at the midpoints of each segment joining two consecutive discretization points on the same element meet the restrictions of the dynamical system equations , ( ii ) the magnitude of the last coefficients in the state and control truncated series . to illustrate the proposed adaptive technique ,
let us begin by defining the elemental midpoints vector @xmath92^t } : \check\tau _ { { n_k},i}^{(k),\alpha } = \frac{1}{2}\left ( { \hat \tau _ { { n_k},i}^{(k),\alpha } + \hat \tau _ { { n_k},i + 1}^{(k),\alpha } } \right),\ ; i = 0 , \ldots , { n_k } ; k \in \mathbb{k}.$ ] let @xmath93 , and denote the kesobgim constructed using the integration nodes vector @xmath94 by @xmath95 ; cf . @xcite
. on each element @xmath71 , define the absolute residual matrix @xmath96:$ ] @xmath97 where `` @xmath98 $ ] '' is the vertical matrix concatenation along columns , and @xmath99^t.\end{aligned}\ ] ] now , let @xmath100 .
if @xmath101 falls below @xmath102 , a user - specified threshold for the size of the residual error , then the obtained state and control spectral coefficients could be optimal . as a further measure of the quality of the determined coefficients , we check the magnitude of the last coefficients in the state and control truncated series .
in fact , for sufficiently smooth functions , the @xmath76th coefficient of the spectral expansion must decay faster than any inverse power of @xmath103 ; cf . @xcite .
this fundamental property of spectral methods can be exploited and combined with the residual error measurement to form a strong tool used to decide whether to accept the obtained approximations , or not .
we shall refer to the condition @xmath104 by condition @xmath105 and refer to the inequalities @xmath106 by condition @xmath107 , where @xmath108 is a user - specified threshold for the size of the last spectral coefficients . if both conditions @xmath105 and @xmath107 are satisfied , the obtained approximations are considered acceptable . if not , then we need to decide whether to divide the current mesh interval @xmath71 , or increase the number of collocation points and spectral coefficients .
to this end , we slightly follow the approach presented by @xcite . in particular
, we determine the column vector , @xmath109^t$ ] from the largest element of @xmath110 , and calculate the arithmetic mean , @xmath111 , of the elements of @xmath112 as follows : @xmath113 .
finally , we find the residual vector @xmath114 via calculating @xmath115 . now , let @xmath116 be a user - specified threshold for the size of the elements @xmath117 of the vector @xmath114 , and define a discrete local maximum ( peak ) @xmath118 of @xmath114 by the data sample that is larger than its two neighboring samples ; i.e. , the value @xmath119 .
let @xmath120 be the row vector of the local maxima of @xmath114 .
we have the following three cases : @xmath121 : : if @xmath122 , then set @xmath123 $ ] .
@xmath124 : : if @xmath125 , then set @xmath126 $ ] .
@xmath127 : : if @xmath128 , then set @xmath129 $ ] .
if the error is nonuniform , we break the domain @xmath71 at the elemental midpoints @xmath130 .
otherwise , the error is considered uniform , so we increase the number of collocation points and spectral coefficients by some constant values as long as the degree of the gegenbauer polynomial interpolant remains below a maximum allowable degree . in particular , we choose some positive integer numbers @xmath131 , and update the values of @xmath132 , and @xmath133 . in the former case , the interval partitioning is only allowed for a maximum number of divisions @xmath134 . moreover , to prevent the division of a relatively small domain , we introduce the `` edge spacing '' parameter @xmath135 , so that further domain partitioning is forbidden if @xmath136 , where @xmath137 denotes the length of the interval @xmath74 . in this case , only increasing the number of collocation points and spectral coefficients is allowed . on the other hand ,
if @xmath138 , for some @xmath139 , and the computational algorithm locates some edge points that are sufficiently close from the endpoints of @xmath71 in the sense that the distance between each point of them and an endpoint of @xmath71 is less than the prescribed @xmath135 , then these located edge points are to be discarded , and we break the interval at the remaining located edge points . if no other edge points exist , then we divide the domain using a similar partitioning technique to that adopted by the popular golden section search method . in particular , we break the interval at @xmath140 , where @xmath141 represents the golden ratio .
this section is devoted for analyzing the truncation error of the kesobgq constructed through interpolation at the adjoint kesgg nodes , since it constitutes a crucial numerical tool in the discretization procedure .
let @xmath142 , for any real - valued function @xmath143 defined on @xmath144 .
let also @xmath145 denote the space of all polynomials of degree at most @xmath61 , for some @xmath146 .
the following theorem highlights the truncation error of the kesobgq , after successfully locating the discontinuities or the `` points of nonsmoothness . ''
[ sec : erranalysgp1 ] let @xmath147 , and consider any arbitrary integration nodes @xmath148 .
furthermore , let @xmath149 , be a real - valued function approximated by the truncated @xmath2th elemental shifted gegenbauer polynomials expansion series such that the @xmath2th elemental shifted gegenbauer spectral coefficients are computed at each node @xmath150 by interpolating the function @xmath143 at the adjoint kesgg nodes @xmath151 . then there exist some real numbers @xmath152 , such that @xmath153 where @xmath154 is the truncation error of the kesobgq , and @xmath155 is the leading coefficient of the @xmath40th - degree @xmath2th elemental shifted gegenbauer polynomial
. for each node @xmath150 , set the error term of the @xmath2th elemental shifted gegenbauer interpolation as @xmath156 where @xmath157 is the @xmath2th elemental @xmath158th - degree shifted gegenbauer interpolant of the function @xmath143 on @xmath71 , and @xmath159 are the @xmath2th elemental shifted gegenbauer spectral coefficients .
now construct the auxiliary function @xmath160 since @xmath149 , and @xmath161 it follows that @xmath162 . for @xmath163 ,
we have @xmath164 since @xmath165 , are zeroes of @xmath166 .
moreover , @xmath167 thus @xmath168 and @xmath169 is zero at the @xmath170 distinct nodes @xmath171 . by the generalized rolle s theorem
, there exists a number @xmath172 in @xmath173 such that @xmath174 .
therefore , @xmath175 where @xmath176 is the leading coefficient of the gegenbauer polynomial , @xmath177 .
since @xmath178 is identically zero , then we have @xmath179 the proof of the theorem is established by rewriting the @xmath2th elemental @xmath158th - degree shifted gegenbauer interpolant , @xmath180 , in its equivalent lagrange form ( nodal approximation ) , and integrating both sides of eq . on the interval
@xmath181 $ ] .
the following theorem marks the error bounds of the quadrature truncation error given by the above theorem on each element @xmath41 .
[ thm:2 ] given the assumptions of theorem [ sec : erranalysgp1 ] such that @xmath182 , where the constant @xmath183 is dependent on @xmath2 but independent of @xmath158 .
then there exist some constants @xmath184 and @xmath185 , dependent on @xmath186 and independent of @xmath158 such that the quadrature truncation error , @xmath187 , on each element @xmath71 is bounded by @xmath188 where @xmath189 ; the constants @xmath190 and @xmath191 are dependent on @xmath186 , but independent of @xmath158 .
the proof can be established easily using ( * ? ? ?
* lemmas 4.1 and 4.2 ) .
it is noteworthy to mention that the accuracy achieved by a spectral differentiation / integration matrix used by traditional pseudospectral methods in the literature is usually constrained by the number of collocation points .
therefore , increasing the number of collocation points on each domain @xmath71 , requires a similar grow in the size of the spectral differentiation / integration matrix , which could result in a significant grow in the total computational cost of the method . on the other hand , a notable merit of the present method as shown by theorem [ thm:2 ]
occurs in taking advantage of the free rectangular form of the kesobgim .
in particular , regardless of the small / large number of collocation points used to determine the approximate states and controls , the present method endowed with the kesobgim can achieve almost full machine precision approximations to the integrals involved in the optimal control problem using relatively moderate values of the parameters @xmath192 and @xmath193 on each element @xmath194 ; thus achieving excellent approximations while maintaining a low operational cost . we shall demonstrate this virtue further in the next section .
in this section , we report the results of the present gise method on two nonlinear optimal control problems well studied in the literature .
the nonlinear programming problems were solved using snopt software @xcite with the major and minor feasibility tolerances , and major optimality tolerance all set at @xmath195 .
the numerical experiments were conducted on a personal laptop equipped with an intel(r ) core(tm ) i7 - 2670qm cpu with 2.20ghz speed running on a windows 10 64-bit operating system and provided with matlab r2014b ( 8.4.0.150421 ) software .
[ [ example-1 ] ] * example 1 * + + + + + + + + + + + consider the following nonlinear optimal control problem : [ ex1:1 ] @xmath196,\\ & u(t ) \in [ 0,1],\;x(0 ) = 1,\;x(1 ) = 0.\end{aligned}\ ] ] this problem was numerically solved in a series of papers ; cf .
the work presented in the latest article of this series , @xcite , adopted a linearization of the nonlinear dynamical system via a linear combination property of intervals followed by a `` _ _ random _ _ '' interval partitioning ( three switching points were chosen randomly ) and an integral reformulation of the multidomain dynamical system .
the transformed problem was then collocated at the legendre - gauss - lobatto points and the involved integrals were approximated using the legendre - gauss - lobatto quadrature rule .
moreover , the control and state functions were approximated by piecewise constants and piecewise polynomials , respectively .
we applied the present gise method for solving the problem numerically using the parameter settings @xmath197 , and @xmath198 .
all state and control coefficients were initially set to one .
figure [ fig : ex1statecontrol ] shows a sketch of the obtained approximate optimal state and control profiles on @xmath199 $ ] with a reported approximate optimal cost function value of @xmath200 .
in contrast with the work of @xcite , the present gise method endowed with the proposed adaptive strategy finds no points of discontinuities / nonsmoothness , as the state and control functions appear to be sufficiently smooth on the interval @xmath199 $ ] . therefore , the adaptivity of the method enables a fast implementation using a single collocation grid without any domain partitioning ; thus @xmath201 .
figure [ fig : ex1fun1 ] shows the plot of the approximate optimal cost functional @xmath202 , for several values of @xmath203 and @xmath204 . as observed from the figure , the reported approximate optimal cost function values approach @xmath205 for increasing values of collocation points and spectral coefficients in a close agreement with the results obtained by @xcite .
figure [ fig : ex1coeff1 ] manifests further the corresponding exponential ( spectral ) decay of the last optimal coefficients in the state and control truncated series , @xmath206 and @xmath207 .
in fact , figure [ fig : ex1coeff1 ] shows an interesting behavior of the gise method .
in particular , the figure shows that the last optimal coefficients in the state and control shifted gegenbauer truncated series generally decay faster for negative values of the gegenbauer parameter @xmath204 than for positive values , and this deterioration phenomenon seems to happen monotonically as the value of @xmath204 approaches @xmath208 .
this numerical simulation is in close consensus with the work of @xcite on the numerical solution of the second - order one - dimensional hyperbolic telegraph equation using a shifted gegenbauer pseudospectral method . in particular , the latter showed theoretically that the coefficients of the bivariate shifted gegenbauer expansions decay faster for negative @xmath204-values than for non - negative @xmath204-values , but the asymptotic truncation error as the number of collocation points grows largely is minimized in the chebyshev norm exactly at @xmath56 ; i.e. , when applying the shifted chebyshev basis polynomials . figure [ fig : ex1fun1 ] indicates that collocations at negative values of @xmath204 close to @xmath208 is not to be endorsed for increasing values of collocation points and expansion terms . in particular , while the values of @xmath202 seem to be matching for almost all of the @xmath204-values used in the numerical simulation , a peak in the surface of the approximate optimal cost function is clearly observed at @xmath209 , for @xmath210 , indicating a poor approximation in this case . on the other hand , @xcite pointed out that the gegenbauer quadrature ` may become sensitive to round - off errors for positive and large values of the parameter @xmath204 due to the narrowing effect of the gegenbauer weight function , ' which drives the quadrature to become more extrapolatory .
in particular , @xcite identified the range @xmath211 , as a preferable choice to construct the gegenbauer quadrature , for some relatively small positive number @xmath212 and @xmath213 $ ] .
we shall refer to the interval @xmath214 $ ] by _ `` the gegenbauer collocation interval of choice , '' _ and denote it by @xmath215 .
figure [ ex1fun1_expandedperfect ] shows a sketch of the approximate optimal cost functional @xmath216 for @xmath217 , and @xmath218 , where we can clearly see the rise of hills in the surface profile for increasing values of @xmath219 demonstrating poor approximations for such @xmath204-values .
this formation of hills for increasing values of @xmath219 is salient as well in the surface profiles of the corresponding magnitudes of the last coefficients in the state and control truncated series ; cf .
figure [ ex1coeff1expandedperfect ] .
in general , we largely endorse the following rule of thumb . [ [ rule - of - thumb ] ] * rule of thumb * + + + + + + + + + + + + + + + _ it is generally advantageous to collocate problem 3 for values of @xmath220 for small / medium numbers of collocation points and gegenbauer expansion terms ; however , collocations at the shifted chebyshev - gauss points should be put into effect for large numbers of collocation points and gegenbauer expansion terms if the approximations are sought in the infinity norm ( chebyshev norm ) . _
+ we shall further examine experimentally this rule of thumb in the next example , where the exact control function is given in closed form .
$ ] using @xmath221 , and @xmath222 .
the plot was generated using @xmath223 linearly spaced nodes from @xmath224 to @xmath225 .
] for @xmath217 , and @xmath226 . ] , and @xmath226 . ] for @xmath217 , and @xmath218 . ] , and @xmath218 . ]
[ [ example-2 ] ] * example 2 * + + + + + + + + + + + consider the popular breakwell problem @xcite : [ ex2:2 ] @xmath227,\\ & { { \dot x}_2}(t ) = u(t),\quad t \in [ 0,1],\\ & { x_1}(0 ) = { x_1}(1 ) = 0,\\ & { x_2}(0 ) = - { x_2}(1 ) = 1,\\ & { x_1}(t ) \le 0.1.\end{aligned}\ ] ] the exact control function to this problem is given by @xmath228,\\ 0,\quad t \in [ 0.3,0.7],\\ - 200\,t/9 + 140/9,\quad t \in [ 0.7,1 ] , \end{array } \right.\ ] ] and the optimal cost function value is @xmath229 .
this problem was numerically solved by @xcite using a direct legendre pseudospectral , and the collocation was performed at the legendre - gauss - lobatto quadrature nodes ; however , the observed convergence rate was very slow due to the lack of smoothness of the optimal solution .
later , @xcite numerically solved the problem using a direct composite chebyshev finite difference method based on a hybrid of block - pulse functions and chebyshev polynomials , and the implementation was carried out using chebyshev - gauss - lobatto points .
we implemented the present gse for solving the problem numerically using the parameter settings @xmath230 , and @xmath198 .
all state and control coefficients were initially set to zero .
the computational algorithm breaks the transformed interval @xmath26 $ ] into the three domains @xmath231 , \bm{\omega}_2 = [ -0.3906,0.3906]$ ] , and @xmath232 $ ] ; thus @xmath233 .
these three mesh intervals correspond to the three domains @xmath234 , \bm{\omega}_{2,\text{orig } } = [ 0.3047,0.6953]$ ] , and @xmath235 $ ] of the original optimal control problem .
figure [ ex2_all ] shows the plots of the approximate state functions , @xmath236 and @xmath237 , exact and approximate control functions , @xmath238 and @xmath239 , respectively , on the interval @xmath199 $ ] . the corresponding approximate optimal cost function value , @xmath240 .
figure [ ex2_u_err ] shows the absolute error of the control function values , @xmath241 , on the interval @xmath199 $ ] in log - lin scale , where a rapid error decay is clearly seen .
in contrast , figure [ ex2_u_err_single ] shows the slow convergence of the absolute error on the interval @xmath199 $ ] in log - lin scale using the same parameter settings , but with a single collocation grid .
we expect to attain faster convergence rates of the present gse by optimizing the parameters @xmath242 , and @xmath108 relative to the initial inputs @xmath132 , and @xmath243 , so that the determined edge points are sufficiently close from the discontinuities or points of nonsmoothness . however , we prefer to leave this interesting topic as a future research .
it is also intriguing to study the numerical effect of different values of @xmath204 on the numerical scheme .
to this end , and for fairness of comparisons , we implemented the gise method in the absence of the proposed adaptive strategy , and broke the transformed domain @xmath26 $ ] at the exact edge points @xmath244 and @xmath245 , which correspond to the original edge points @xmath246 and @xmath247 in @xmath199 $ ] .
figure [ ex2costn0p4to1 ] shows the plot of the approximate optimal cost functional @xmath216 for @xmath248 , and @xmath249 such input values lead to a small - scale nonlinear programming problem of dimension @xmath250 .
obviously , the cost functional profile decreases as @xmath204 progresses away from @xmath208 , then remains nearly steady at about @xmath251 till @xmath204 reaches the value @xmath225 .
figure [ ex2coeff0p4to1 ] shows the corresponding magnitudes of the last coefficients in the state and control truncated series on the mesh intervals @xmath252 .
on the other hand , figure [ ex2cost1to20 ] shows a sketch of the approximate optimal cost functional profile for @xmath253 , where we see a significant rise in the functional value reported at @xmath254 .
this increase in the functional value is captured very clearly in the profile of the corresponding magnitudes of the last coefficients in the state and control truncated series on the mesh intervals @xmath252 ; cf .
figure [ ex2coeff1to20 ] .
in particular , we can see some few jumps and increases in the last coefficients magnitudes exactly at the reported @xmath204-values .
therefore , for the given parameter settings , we can consider the set @xmath255 as a feasible gegenbauer collocation interval of choice . to investigate further the convergence of the gise method for the same two sets of possible @xmath204-values
, we implemented the method using the relatively medium values @xmath256 , which lead to a medium - scale nonlinear programming problem of dimension @xmath257 .
figure [ ex2costn0p4to122 ] shows the plot of the approximate optimal cost functional @xmath216 for @xmath249 , where a reduction in its value is generally observed for increasing values of @xmath204 till it arrives at a nearly uniform state at about @xmath258 for @xmath259 .
moreover , the corresponding magnitudes of the last coefficients in the state and control truncated series remain bounded below @xmath260 and @xmath261 , respectively , on the mesh intervals @xmath252 ; cf . figure [ ex2coeff0p4to122 ] . on the other hand ,
figure [ ex2costn0p4to12244 ] shows a sketch of the approximate optimal cost functional profile for @xmath253 , where we see some wild escalations in the functional value at @xmath262 and @xmath263 accompanied by another significant rise for @xmath264 .
this serious degradation in the accuracy of the functional value is also observed very clearly in the profile of the corresponding magnitudes of the last coefficients in the state and control truncated series on the mesh intervals @xmath265 ; in particular , the last coefficients magnitudes generally soar rapidly as @xmath204 increases ; cf .
figure [ ex2coeff0p4to12244 ] .
hence , for the given parameter settings , the set @xmath266 is a feasible gegenbauer collocation interval of choice .
figure [ fig : all ] shows further the absolute errors of the control function values , @xmath241 , on the interval @xmath199 $ ] in log - lin scale using @xmath267 .
high - order control approximations are achieved in all cases , except for @xmath268 , and @xmath269 , where a significant recession in accuracy is reported near the time boundary points @xmath270 and @xmath271 in the former case , while a tangible error growth in the vicinity of the right endpoint @xmath271 is reported in the latter two cases .
a typical hp - pseudospectral method would normally apply a square differentiation / integration matrix of size @xmath272 to compute the derivatives / integrals involved in the optimal control problem ; thus requires @xmath273 flops to evaluate the derivatives of a real - valued differentiable function at a set of @xmath274 collocation points for each @xmath139 , or the definite integrals of an integrable function using the same sets of collocation points as the upper limits of the integrations .
in contrast , the kesobgim requires @xmath275 flops to evaluate the needed integrals . to prevent an enormous amount of calculations
, we can set @xmath192 at a relatively medium value , say , @xmath276 usually sufficient to achieve nearly full machine precision approximations to the integrals of well - behaved functions for large values of @xmath277 .
for instance , using @xmath278 and @xmath279 , we can evidently count a substantial difference of @xmath280 flops between the developed kesobgim and a standard operational matrix of differentiation / integration for each derivative / integral per mesh interval . to visualize the big picture , notice that the discretization of the present optimal control problem requires the evaluation of a single integral for the cost functional , and @xmath281 integrals involved in eqs .
, for a total of @xmath282 integral evaluations .
working out the mathematics , it is not hard to realize a remarkable gap of @xmath283 flops in favor of the present gise method endowed with the kesobgim !
the current study casts the light on the judicious choice of the kesgg collocation points set , @xmath284 , to be utilized on each mesh interval @xmath74 during the discretization process of optimal control problems .
in particular , the current work supports collocations performed at @xmath285 , for small / medium numbers of collocation points and gegenbauer expansion terms .
nonetheless , it would be extremely beneficial to determine theoretically the optimal collocation sets @xmath286 for each domain @xmath74 * a question which yet remains open*. and @xmath237 ( left and middle ) , exact and approximate control functions , @xmath238 and @xmath239 ( right ) , respectively , on the interval @xmath199 $ ] using the initial values @xmath287 , and @xmath288 .
the plots were generated using @xmath269 linearly spaced nodes in each domain @xmath252 . ] for @xmath248 , and @xmath249 . ] , \bm{\omega}_{2,\text{orig } } = [ 0.3,0.7]$ ] , and @xmath289 $ ] using @xmath248 , and @xmath249 . ] for @xmath248 , and @xmath253 . ] , \bm{\omega}_{2,\text{orig } } = [ 0.3,0.7]$ ] , and @xmath289 $ ] using @xmath248 , and @xmath253 . ] for @xmath256 , and @xmath249 . ] , \bm{\omega}_{2,\text{orig } } = [ 0.3,0.7]$ ] , and @xmath289 $ ] using @xmath256 , and @xmath249 . ] for @xmath256 , and @xmath253 . ] , \bm{\omega}_{2,\text{orig } } = [ 0.3,0.7]$ ] , and @xmath289 $ ] using @xmath256 , and @xmath253 . ] , on the interval @xmath199 $ ] in log - lin scale using @xmath256 , and @xmath267 .
all plots were generated using @xmath269 linearly spaced nodes in each domain @xmath252.,width=680,height=566 ]
the current gise method was tested on only two numerical test problems in an attempt to reduce the size of the manuscript .
however , further test problems may be necessary to verify further the power of the proposed method .
moreover , a further theoretical study may be conducted to analyze the convergence of the gise method .
motivated by the spectral accuracy offered by spectral element methods , we have proposed a fast , economic , and high - order algorithm for the solution of nonlinear optimal control problems exhibiting smooth / nonsmooth solutions .
the coalition of information derived from the residual of the discrete dynamical system equations and the magnitude of the last coefficients in the state and control truncated series forms a powerful multicriterion adaptive strategy to boost the accuracy of the state and control approximations .
another major source for the strength of the proposed method lies in the free rectangular form of the kesobgim , which allows for excellent approximations to integrals with accuracy approaching machine precision .
remarkably , this significant result is achieved regardless of the number of collocation points used in the discretization process .
the numerical experiments support collocations of nonlinear optimal control problems performed at @xmath290 , for small / medium numbers of collocation points and gegenbauer expansion terms .
the proposed method can be easily extended to different problems and applications .
37 natexlab#1#1[2]#2 , , , ( ) .
, , , ( ) . , , , , , ( ) . , , , , , ( ) . . , , , ( ) . , , ,
, ( ) . , , , , , , , , ( ) . , , , ( ) . , , , , in : , pp . . , , ph.d
thesis , school of mathematical sciences , faculty of science , monash university , .
, , , ( ) . , , , , , ( ) . , , , , , , , , ( ) . , , , ( ) .
, , , , , springer series in computational physics , , .
, , , , , , , , ( ) . , , , , ( ) .
, , , , , ( ) . , , , , ( ) .
, , , ( ) . , , , .
, , , ( ) . , , , , ( ) .
, , , ( ) . , , , , ( ) .
, , , ( ) . , , . .
, , , , , , department of mathematics , university of california , san diego , , .
, , , , ( ) . , , in : , , , pp . .
, , , ( ) . , , , , , in : , volume , .
, , , ( ) . , , , ( ) . | in this work , we propose an adaptive spectral element algorithm for solving nonlinear optimal control problems .
the method employs orthogonal collocation at the shifted gegenbauer - gauss points combined with very accurate and stable numerical quadratures to fully discretize the multiple - phase integral form of the optimal control problem .
the method brackets discontinuities and `` points of nonsmoothness '' through a novel local adaptive algorithm , which achieves a desired accuracy on the discrete dynamical system equations by adjusting both the mesh size and the degree of the approximating polynomials .
a rigorous error analysis of the developed numerical quadratures is presented .
finally , the efficiency of the proposed method is demonstrated on two test examples from the open literature . |
our understanding of the first few billion years of cosmic time has increased significantly in recent years , thanks to the _ hubble space telescope s _ ( ) wide - field camera 3/infrared channel ( wfc3/ir , * ? ? ?
* ) as well as the _
spitzer space telescope s _ infrared array camera ( irac , * ? ? ?
* ) . until recently
, the hubble ultra deep field @xcite has provided our deepest view of the universe , revealing a considerable number of galaxy candidates at @xmath12 , including one candidate at @xmath13 @xcite .
the cosmic epoch of @xmath13 is important to study as it marks the dawn of galaxy formation and the beginning of reionization of the intergalactic medium .
however , galaxies at that redshift are extremely faint , making it difficult to discover and study the abundant population of galaxies below @xmath14 , the knee of the luminosity function .
fortunately , the magnification boost afforded by gravitational lensing combined with s exquisite imaging capabilities in the near - infrared ( nir ) , provides an avenue for both discovering and characterizing the intrinsic properties of galaxies around @xmath15 , when the universe was less than half a billion years old .
the cluster lensing and supernova survey with hubble ( clash , * ? ? ?
* ) carried out imaging of 25 galaxy clusters in 16 broad bands between @xmath16 @xmath17 m to a depth of ab magnitude @xmath18 with a total of 20 orbits per cluster .
the clash program has led to many interesting discoveries of magnified , intrinsically faint galaxies .
several hundred dropout galaxies have been uncovered in the range @xmath19 , with a few notable examples at higher redshifts of @xmath20 ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , helping to motivate dedicated deeper lensing surveys .
the hubble frontier fields ( hff ) is a new initiative now being carried out to observe the distant universe to an unprecedented depth , combining the power of deep imaging and gravitational lensing . in s cycles 21 and 22 ,
560 orbits of director s discretionary time have been allocated to observe four clusters .
the observations are carried out with four wfc3/ir filters ( f160w , f140w , f125w , f105w ) and three acs filters ( advanced camera for surveys , * ? ? ?
* f814w , f606w , f435w ) .
it is anticipated that 280 orbits will be allocated in cycle 23 to observe two additional clusters .
in addition , deep _
spitzer _ and _ chandra _ observations are planned for the six hff fields .
these coordinated observations will enable us to probe the star formation rate density at @xmath21 , study the faint end of the galaxy population at @xmath22 , and map the dark matter in these clusters in unprecedented detail via many multiple images of background sources ( hubble deep fields initiative 2012 science working group report ) .
we report the discovery of 24 candidate lyman - break galaxies ( lbgs ) at @xmath23 in the field of , based on the hff observations and archival data .
the faintest sources detected are around ab magnitude 29 .
these objects , listed in tables [ tbl - z9]-[tbl - z7 ] , have `` secure '' photometric redshifts greater than 7 and a negligible probability ( @xmath24 ) of being at lower redshift .
we adopt a concordance cosmology with @xmath25 , @xmath26 and @xmath27 , and the ab magnitude system throughout .
abell 2744 ( @xmath1 ) is the first hff target in s cycle 21 .
it is one of the most actively merging galaxy clusters known @xcite , displaying a large critical curve of roughly @xmath28 .
the six hff clusters have been selected to maximize the lensing boost , which means that systems with highly complex mass distributions ( _ e.g. , _ clusters in the process of merging ) have been selected @xcite .
the hff observations of ( go / dd 13495 , pi : lotz ) were carried out between 2013 oct . 25 and 2014 jul .
1 . additional wfc3/ir images obtained in 2013 aug . and 2014 jun .- jul .
( go 13386 , pi : rodney ) and acs images of 2009 ( go 11689 , pi : dupke ) are retrieved from the mikulski archive for space telescopes ( mast ) and used .
table 1 lists the exposure times and limiting magnitudes for all the imaging used in our analysis .
we process the data using aplus @xcite , an automated pipeline modified from the apsis package @xcite with an enhanced capability of processing wfc3 data and aligning them with the acs data .
we retrieve the calibrated images from the instrument pipelines , namely the _ flc _
images for acs ( corrected for the detector charge transfer efficiency ) and _ flt _ images for wfc3/ir .
recently , we have updated aplus so that images of individual exposures are aligned using _ drizzlepac
_ @xcite , achieving an astrometric precision of @xmath29 or better .
figure [ fig - fov ] displays a composite color image of the cluster field . using aplus ,
we align , resample , and combine all the available imaging in each filter to a common pixel scale of @xmath30 , which is half of wfc3/ir s pixel scale and slightly larger than that of acs .
we then create detection images from the inverse - variance weighted sum of the wfc3/ir and acs images , respectively , and run sextractor @xcite in dual - image mode .
the @xmath23 candidates are first selected from the nir catalog derived from the wfc3/ir detection image , with a threshold of 1.5 times the signal - to - noise value over a minimum of 4 pixels .
we choose colors measured from isophotal magnitudes to select our high - redshift candidates ( see [ sec : selection ] ) , as they balance the need between depth and photometric precision @xcite .
the @xmath31 limiting magnitude in the wfc3/ir bands is @xmath32 in a @xmath33 diameter aperture ( see table 1 ) , and @xmath34 for the observed - frame optical acs bands . as part of the hff campaign ,
spitzer_/irac images were obtained in 2013 sep . and 2014 jan.@xmath35feb . in channels 1 and 2 at wavelengths
@xmath36 and @xmath37 , respectively , using director s discretionary time ( program 90257 , pi : soifer ) . the effective exposure time in each channel
, including that of the archival data ( program 84 ; pi : rieke ) obtained in 2004 , is @xmath38 ksec .
the irac corrected basic calibrated data ( cbcd ) images are processed with mopex @xcite and sampled to a final pixel scale of @xmath39 . in order to perform background matching of the individual cbcd frames we run all steps in the overlap module , where we use the sextractor background estimation with a window size of 25 pixels . to create the mosaic images we run all three outlier modules ( _ i.e. _ dual outlier , mosaic outlier and box outlier ) and we use the default interpolation , with the fine resolution parameter = 0 .
the estimated @xmath40 limiting magnitude is @xmath41 for irac channel 1 ( irac1 , 3.6 ) and @xmath42 for channel 2 ( irac2 , 4.5 ; see table 1 ) .
we search for lbgs using their distinct color around @xmath43 . for example , at @xmath44 , the lyman break is at @xmath45 , between the f814w and f125w bands .
our selection criteria , in units of magnitude , are as follows : @xmath46@xmath47@xmath48these color cuts are similar to those utilized in previous work such as @xcite . for @xmath49 ,
the break is at @xmath50 , between the f105w and f140w bands : @xmath51@xmath52@xmath53 and for @xmath13 , the break is between the f125w and f160w bands : @xmath54 we require that a candidate must not be detected above @xmath40 in a summed image blueward of the selection bands defined above . for objects at @xmath55 ,
this requires a non - detection in a summed image of the f606w and f435w bands , while for candidates at @xmath56 this requires a non - detection in the stacked optical detection image .
in addition to the color selection criteria described above , we also exclude candidates lying within one arcsecond of the detector edges , in order to mitigate potentially spurious photometry . we also exclude candidates lying near stellar diffraction spikes , which are difficult to remove because hff wfc3/ir exposures were obtained at the same position angle .
finally , we also identify and remove candidates with a color decrement of f160w - irac @xmath57 , as they are most likely extremely red objects at lower redshift ( @xmath58 ) .
our photometry is measured within an isophotal aperture , but aperture - corrected to a total flux using the @xmath59 difference in the f160w band . in a few cases where source blending is significant ,
we visually inspect the images and choose an aperture that is larger than the source s full - width at half - maximum ( fwhm ) , but not so large as to be affected by nearby sources , and use the corresponding aperture magnitude in place of @xmath60 .
we also verify that our aperture colors ( and therefore our list of high - redshift candidates ) are not affected by image artifacts if we use the publicly released mosaics based on the _ mosaicdrizzle _ pipeline @xcite .
the irac images of our candidates suffer from crowding due to the instrument s large point spread function ( psf , fwhm @xmath61 ) , such that simple aperture photometry might result in inaccurate fluxes due to contamination from nearby sources . to address this issue
, we use a deblending technique whereby contaminating neighbors are subtracted using galfit @xcite by performing a fit to the objects of interest and all their close neighbors simultaneously in a @xmath62 fitting window around the source of interest @xcite .
the irac psf is determined from the same image by stacking 10 bright , isolated point sources .
positions and radial profiles of neighboring sources in this region are derived from the higher resolution f160w - band mosaic , while the initial input magnitudes are obtained by running sextractor on the irac images . during the fitting process
, all input parameters are allowed to vary , except for the positions of the objects of interest .
this process is similar to that of @xcite and @xcite .
four of our candidates are so heavily blended by nearby bright sources with complex radial profiles that galfit fails to satisfactorily deblend them .
this yields a total of 16 sources for which photometry or upper limits from galfit are possible ( see table [ tbl - irac ] ) .
we carry out extensive tests to check the reliability of our irac photometry for each source .
first , we place a simulated point source of magnitude 25 near the candidate and run galfit with different fitting windows and background levels until the expected magnitude of each simulated source is recovered ( with a magnitude difference @xmath63 mag compared to the input value , consistent with the photometric errors ) .
we then proceed to fit the flux of each candidate , using the fitting window and background level on the image that recovered the brightness of the simulated source .
we repeat these tests at three different positions for the simulated source to verify our measurement of the source magnitude . to account for the uncertainties in estimating the background , we choose the average magnitude of three measurements as the source magnitude , which are reported in table [ tbl - irac ] . in figures [ fig - z9]-[fig - z7 ] we show cutout images of all the candidates . [ fig - z8 ]
as part of the hff initiative , seven independently derived gravitational lensing models of the abell 2744 field were developed and publicly released through the mast archive . for our analysis we adopt the zitrin `` nfw '' model as our fiducial lensing model , and utilize the other six models to help quantify the systematic uncertainties in our magnification estimates .
the zitrin `` nfw '' model assumes a pseudo - isothermal elliptical mass distribution for each cluster galaxy , scaled by the galaxy luminosity , and an elliptical nfw @xcite halo for the dark matter . in the case of ,
two such elliptical nfw halos are used , centered on the two central , brightest cluster galaxies to represent the global dark - matter component .
these are combined with the galaxies component to generate , via a long monte - carlo - markov - chain minimization , the best - fitting model for the total projected mass ( see @xcite , and references therein ) . to estimate the systematic uncertainty in the magnification of each of our high - redshift candidates , we exclude the highest and lowest magnification factors and then calculate the difference in the second - highest and second - lowest magnifications from the different models .
this procedure is designed to mitigate potential extremes in the model predictions , and better reflect the true systematic uncertainties .
the magnification factors listed in tables [ tbl - z9]-[tbl - z7 ] are the best - fit values at the tabulated redshift based on the zitrin `` nfw '' model , while the uncertainties are the quadrature sum of the systematic and statistical uncertainties . figure [ fig - fov ] shows the composite color image of the abell 2744 field , overlaid with the critical curves and identification numbers for all our candidates .
we calculate photometric redshifts using the bayesian photometric redshift code bpz ( bayesian photometric redshifts ; * ? ? ?
* ; * ? ? ?
* ) , adopting the same template library used by the clash collaboration @xcite .
the template set consists of five elliptical galaxy templates , two spiral galaxy templates , and four starburst galaxy templates with moderately strong emission lines .
the templates were originally based on the pgase stellar population synthesis models @xcite , but have been recalibrated using spectroscopic redshifts of galaxies with deep , multiband photometry from the fireworks survey @xcite .
we assume ignorant ( _ i.e. _ , flat ) priors on both galaxy type and redshift in the range @xmath64 .
using bpz , we identify 24 candidates that satisfy our color selection criteria ( see [ sec : selection ] ) and whose photometric redshifts place them at @xmath65 . in tables
[ tbl - z9]-[tbl - z7 ] we list the coordinates , photometric redshifts , _ hst _ photometry , and magnifications of these 27 candidates . in order to infer the physical properties of our high - redshift candidates ( see [ sec : properties ] ) , and as an additional check on the bpz - based photometric redshifts , we use the bayesian spectral energy distribution ( sed ) modeling code isedfit @xcite . using a monte carlo technique ,
we generate @xmath66 model seds with a broad range of star formation histories , ages , stellar metallicities , dust content , and nebular emission - line strength .
we use the flexible stellar population synthesis models ( fsps , v 2.4 ; * ? ? ?
* ; * ? ? ?
* ) based on the miles stellar library @xcite and assume the @xcite initial mass function from @xmath67 .
we adopt _ delayed _ star - formation histories , @xmath68 , where sfr is the star formation rate , @xmath69 is the time since the onset of star formation ( `` age '' ) , and @xmath70 is the characteristic time for star formation .
the advantage of this parameterization is that it allows for both linearly rising ( @xmath71 ) and exponentially declining ( @xmath72 ) star formation histories , which may be important for modeling the seds of galaxies at the highest redshifts ( e.g. , * ? ? ?
* ) . for our photometric redshift calculations we adopt uniform priors on age @xmath73 $ ] gyr , star formation timescale @xmath74 $ ] gyr , stellar metallicity @xmath75 $ ] , and rest - frame @xmath76-band attenuation @xmath77 $ ] mag , assuming the time - dependent attenuation curve of @xcite .
each model also includes nebular emission lines whose luminosity is tied self - consistently to the number of hydrogen - ionizing photons .
we find that isedfit and bpz yield statistically consistent photometric redshifts for the majority of the candidates ; the mean difference is @xmath78 ( isedfit minus bpz ) , which is well within our quoted photometric redshift uncertainties . in a few cases
isedfit prefers a lower - redshift solution , @xmath79 ; however , in every case these lower - redshift solutions require a highly unlikely combination of physical properties , namely low stellar masses , low star - formation rates , and large amounts of dust attenuation .
secondary peaks in the redshift probability distribution from isedfit , on the other hand , place these candidates at @xmath65 , in agreement with bpz s primary redshift probability peaks .
to help corroborate the high - redshift nature of our 24 candidates , we search for potential counter images near the locations predicted by the gravitational lensing model . among the candidates that are inside or near the @xmath80 critical curves , jd1 @xcite and zd7 are the only two cases where multiple images are found . for the others ,
the predicted counter images are either behind bright foreground galaxies , or too faint to be confirmed .
two of the counter images predicted for source zd10 are near that of zd11 , which have been discussed by a14 ( system 5 ) .
the triple system zd7 is shown in figure [ fig - arc ] and table [ tbl - ms ] .
image a is an arclet made of two components that are separated only by @xmath81 kpc in the source plane , assuming a magnification of six .
this is similar to a case in abell 1689 where a pair of lbgs at @xmath82 may be merging @xcite .
image b is behind a bright foreground galaxy , which we subtract before carrying out photometry .
we make use of the pure geometric scaling induced by strong lensing to estimate a purely geometric distance for this triply imaged case .
the zitrin `` nfw '' lensing model described in
[ sec : models ] is based on eleven sets of multiply lensed galaxies between @xmath83 , including a spectroscopic redshift of system 6 at @xmath84 ( j. richard , in preparation ) .
this spectroscopic redshift provides a normalization of the model so that the deflection field induced in the lens plane , @xmath85 , can be scaled to any redshift via the lensing source distance ratio @xmath86 to provide the observed deflection field @xmath87 .
hence only a simple scaling of the relative lensing distance ratios is required to relate deflections at any given redshift to the lensing distance of the normalization used to calibrate the lens model , which in our case is @xmath88 .
we find that this factor is @xmath89 for the triple system which minimizes the location of the observed images relative to that generated by the model , and this corresponds to a best bpz estimate of @xmath2 .
a possible faint fourth image d is noted between images a and b , close to another bright galaxy ( figure 5 ) .
our paper serves as an independent verification of dropout objects in other reports , including the recent work of @xcite ( a14 hereafter ) , @xcite and @xcite .
most of the 15 candidates in a14 are at @xmath90 , and therefore have only limited overlap with ours .
three of their candidates have been independently identified by us , but with somewhat different photometric redshifts .
object zd9 in table [ tbl - z7 ] , with @xmath91 , corresponds to a14 s object 561 ( @xmath92 ) .
object zd2 ( @xmath93 ) is a14 s object 2070 ( @xmath94 ) , object y1 ( @xmath95 ) in @xcite and object 2493 - 2561 in @xcite
. object zd11 ( @xmath91 ) is a14 s object 5.2 ( @xmath96 ) and system 17 ( @xmath97 ) in @xcite .
the close pair of zd3/zd6 ( @xmath98 ) are marked by a14 as object 2070 ( @xmath94 ) and object 2555 - 2516 in @xcite .
objects yd4 , yd7 , yd8 , yd10 , zd2 , zd3 , zd5 are objects 2493 - 2561 , 2481 - 2561 , 2306 - 3090 , 2355 - 2566 , 2508 - 2497 , 2555 - 2516 , 2136 - 2432 in @xcite , respectively . the irac2 flux of object zd2 ( y1 in @xcite ) is more than three times the irac1 value , which suggests a strong balmer break .
a similar case at @xmath99 was reported in abell 383 @xcite in which both the irac1 and irac2 magnitudes are higher than the photometry by 1.6 magnitude . at redshift @xmath100 ,
the balmer break is in the irac1 band , and the model prediction is higher than the measured irac1 flux .
in addition to reporting on the discovery of our high - redshift candidates , we can also begin to characterize their physical properties .
we defer a more detailed analysis of the full sample to a forthcoming paper . here
, we focus on the ten objects at @xmath65 with the highest - confidence photometric redshifts which have well - measured photometry in at least one irac channel .
photometry redward of the balmer break is particularly important for placing meaningful constraints on the stellar mass and age of the stellar populations in these distant objects . to infer the physical properties of these galaxies we use isedfit , but adopt a more restricted set of priors than the ones used to estimate photometric redshifts ( see [ sec : photoz ] ) .
specifically , we adopt uniform priors on galaxy age @xmath101 $ ] myr , @xmath102 $ ] myr , stellar metallicity @xmath103 $ ] , and we assume no dust attenuation ( _ e.g. _ , * ? ? ?
recall that the age of the universe at @xmath104 in our adopted concordance cosmology is @xmath105 myr .
we find that our @xmath65 candidates have demagnified stellar masses of around @xmath106 , and sfrs of approximately @xmath107 yr@xmath10 . these results imply an average doubling time of around @xmath108 myr , which is comparable to the age of the universe at @xmath109 . of the stellar mass formed
is returned to the interstellar medium via supernovae and stellar winds . ]
the ages of the galaxies in our sample are less well constrained given the uncertainties in our irac photometry ; nevertheless , we find a median sfr - weighted age for the sample of @xmath110 myr ( @xmath111 confidence interval ) , corresponding to a typical formation redshift of @xmath112 .
figure [ fig - sed ] presents the seds of four representative galaxies in our sample , sorted by decreasing redshift , as well as the maximum likelihood fits derived using isedfit .
an apparent concentration of candidates northeast of the cluster center is shown in figure [ fig - fov ] .
nine objects at @xmath113 are found within a region of @xmath7 .
since the average magnification in that area is not high ( @xmath114 ) , this apparent overdensity is likely intrinsic rather than being due to lensing .
in addition , figure 7 shows a small region `` quintet '' where five candidates are located within approximately @xmath115 of one another : objects yd1 , yd4 , yd6 , yd7 and zd1 . these objects have similar estimated photometric redshifts of between 7.9 and 8.6 , and their projected separations in the source plane are within @xmath116 kpc .
given the uncertainties in our photometric redshifts , it is therefore possible that these sources are physically associated . to estimate the uncertainties in irac photometry for these sources whose separations are close to the irac psf size , we place 500 sets of simulated sources of different brightness and compared the input and fitted fluxes . in more than 68% of the cases , the fitted results are within 0.5 mag of the inputs for three brighter sources in the f160w band , while for the faintest source the output flux decreases with the f160w - irac color . due to the source confusion
, our simulations suggest that the de - blending with galfit works better for the brighter objects , while it is biased to fainter objects in the sense that their fluxes are likely to be overestimated .
this is because that galfit often crashes under a very faint flux level .
we therefore conclude that we are able to deblend these sources with reasonable accuracy , except for the faint source zd1 with large uncertainties .
overdensities in the high - redshift domain have been previously reported .
for example , @xcite identified four candidates with @xmath100 within @xmath117 in the field borg58 .
@xcite found seven lbgs at @xmath55 in the wfc3/ir field ( @xmath118 ) of abell 1703 .
however , our finding of nine lbgs at @xmath119 within @xmath7 ( @xmath120 kpc in the source plane ) is unique , suggesting that the cosmic variance in source density @xcite is more significant than anticipated .
it appears that cosmic variance increases with redshift @xcite , as evidenced by different results from surveys .
this trend might explain both the large number of clustered @xmath100 sources and the deficiency of @xmath21 sources and illustrate the need for observations in more fields .
we find 24 lbg candidates at @xmath121 in the hff imaging of , reaching an intrinsic magnitude of @xmath122 .
one source at @xmath2 is lensed into three images .
significant clustering is observed on the intrinsic scale of @xmath123 kpc .
thanks to gravitational lensing , we are able to carry out spitzer / irac photometry for 16 of the sources . sed fitting to the brightest candidates suggests stellar masses of @xmath124 , star - formation rates of @xmath125 per year , and a typical formation redshift of @xmath126 . the redshift distribution of our sample is not a smoothly declining function towards higher redshift . in particular , our redshift distribution does not extend smoothly beyond @xmath127 , clustering notwithstanding , considering the effect of cosmic variance , the number density in our sample is consistent with that derived from other studies , _ e.g. _ , @xcite .
given the level of clustering that we see in , it will be important to average over more hff , and to perform a luminosity function analysis so that the redshift dependence can be better related to galaxy mass .
the work presented in this paper is based on observations made with the nasa / esa _ hubble space telescope _ , and has been supported by award ar-13079 from the space telescope science institute ( stsci ) , which is operated by the association of universities for research in astronomy , inc . under nasa contract nas 5 - 26555 .
it is also based on data obtained with the _
spitzer space telescope _ , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
this work utilizes gravitational lensing models produced by p.i.s bradac , ebeling , merten & zitrin , sharon , and williams , funded as part of the frontier fields program conducted by stsci .
these models were calibrated using arcs identified in archival imaging by @xcite , spectroscopic redshifts of arcs obtained using the vlt / fors2 spectrograph ( j. richard et al .
2014 , in prep . ) , and vlt and subaru / suprimecam imaging of the abell 2744 field @xcite .
xs acknowledges support from fp7-space-2012-astrodeep-312725 , and nsfc grants 11103017 and 11233002 .
support for az is provided by nasa through hubble fellowship grant hst - hf-51334.01-a awarded by stsci .
feb acknowledges support from basal - cata pfb-06/2007 , conicyt - chile grants ( fondecyt 1141218 , alma - conicyt 31100004 , gemini - conicyt 32120003 , anillo act1101 ) , and the millennium institute of astrophysics ( project ic120009 ) .
jmd acknowledges support from the spanish consolider project cad2010 - 00064 and aya2012 - 39475-c02 - 01 .
we thank m. meneghitti for helpful comments .
atek , h. , richard , j. , kneib , j .-
2014 , , 786 , 60 ( a14 ) beckwith , s. v. w. , stiavelli , m. , koekemoer , a. m. et al .
2006 , , 132 , 1729 bentez , n , 2000 , , 536 , 571 bertin , e. , & arnouts , s. 1996 , , 117 , 393 blakeslee , j. p. , anderson , k. r. , meurer , g. r. , bentez , n. , & magee , d. 2003 , in astronomical data analysis software and systems xii , asp conference series , vol .
295 , eds .
h. e. payne , r. i. jedrzejewski , & r. n. hook , ( san francisco : asp ) 257 bouwens , r. j. , bradley , l. d. , zitrin , a. et al . 2012@xmath128 , arxiv 1211.2230 bouwens , r. j. , illingworth , g. d. , labbe , i. et al .
2011 , nature , 469 , 504 bouwens , r. j. , illingworth , g. d. , oesch , p. a. et al .
2010 , , 708 , l69 2012@xmath129 , , 737 , 90 2014 , arxiv 1403.4295 bradley , l. d. , bouwens , r. j. , ford , h. c. et al .
2008 , , 678 , 647 bradley , l. d. , bouwens , r. j. , zitrin , a. et al . 2012 , , 747 , 3 bradley , l. d. , zitrin , a. , coe , d. et al .
2014 , , in press , arxiv 1308.1692 chabrier , g. 2003 , , 115 , 763 charlot , s. & fall , s. m. 2000 , , 539 , 718 coe , d. , bentez , n. , snchez , s. f. , jee , m. , bouwens , r. & ford , h. 2006 , , 132 , 926 coe , d. , zitrin , a. , carrasco , m. et al .
2013 , , 762 , 32 coe , d. , bradley , l. & zitrin , a. , 2014 , arxiv 1405.0011 conray , c. & gunn , j. e. 2010 , , 712 , 833 conray , c. , gunn , j. e. & white , m. 2009 , , 699 , 486 cypriano , e. s. , sodr , l. jr . , kneib , j .-
& campusano , l. e. 2008 , , 613 , 95 ellis , r. s. , mclure , r. j. , dunlop , j. s. et al .
2013 , , 763 , l7 fazio , g. g. , hora , j. l. , allen , l. e. et al . , 2004 , ,
154 , 10 ferguson , h. c. & mcgaugh , s. s. 1995 , , 440 , 470 fioc , m. , & rocca - volmerange , b. 1997 , , 326 , 950 ford , h. c. & the acs science team 1998 , in space telescopes and instruments v , spie 3356 , eds .
y. bely & j. b. breckinridge , 234 hack , w. j. , dencheva , n. , fruchter , a. s. , 2013 in astronomical data analysis software and systems xxii , asp conf . ser .
475 , ( san francisco : asp ) ed .
d. freidel , 49 illingworth , g. d. , magee , d. , oesch , p. a. et al .
2013 , , 209 , 6 jouvel , s. , hst , o. , lehav , o. , et al .
2014 , , 562 , a86 kimble , r. a. , mackenty , j. w. , oconnell , r. w. , & townsend , j. a. 2008 , in space telescopes and instrumentation 2008 : optical , infrared , and millimeter , spie 7010 , eds .
j. m. oschmann , jr .
, m. w. m. de graauw , & h. a. macewen , 70101e koekemoer , a. m. , ellis , r. s. , mclure , r. j. et al .
2013 , , 209 , 3 labb , i. , bouwens , r. , illingworth , g. d. , & franx , m. 2006 , , 649 , l67 labb , i , gonzlez , v. , bouwens , r. j. et al .
2010 , , 716 , 103 lam , d. , broadhurst , t. , diego , j. m. et al .
2014 , arxiv 1406.6586 laporte , n. , streblyanska , a. , clement , b. et al .
2014 , , 562 , 8l makovoz , d. & khan , i. 2005 , in astronomical data analysis software and systems vi , asp conf .
132 , eds .
p. l. shopbell , m. c. britton , & r. ebert ( san francisco : asp ) , 81 merten , j. , coe , d. , dupke , r. et al .
2011 , , 417 , 333 moustakas , j. , coil , a. , aird , j. et al .
2013 , , 767 , 50 moster , b. p. , somerville , r. s. , newman , j. a. & hans - walter , r. 2011 , , 731 , 113 navarro , j. f. , frenk , c. s. & white , s. d. m. 1996 , , 462 , 563 oesch , p. a. , bouwens , r. j. , illingworth , g. d. et al .
2010 , , 709 , l16 2013 , , 773 , 75 okabe , n. , okura , y. & futamase , t. 2010 , , 713 , 291 okabe , n. , takada , m. , umetsu , k. et al .
2010 , , 62 , 811 okabe , n. & umetsu , k. 2008 , , 60 , 345 overzier , r. a. , shu , x. , zheng , w. et al .
2009 , 704 , 548 papovich , c. , finkelstain , s. l. , ferguson , h. c. , lotz , j. m. , & giavalisco , m. 2011 , , 412 , 1123 peng , c. y. , ho , l. c. , impey , c. d. & rix , h .- w .
2010 , , 139 , 2097 postman , m. , coe , d. , bentez , n. et al .
2012 , , 199 , 25 redlich , m. , bartemann , m. , waizmann , j .- c . &
fedeli , c. 2012 , , 547 , 66 richard , j. , kneib , j .-
p . , ebeling , h. et al .
2011 , , 414 , l31 sanchez - blzquez , p. , peletier , r. f. , jimnez - vicente , j. et al .
2006 , , 371 , 703 torri , e. , meneghetti , m. , bartelmann , m. et al .
2004 , , 349 , 476 trenti , m. , bradley , l. d. , stiavelli , m. et al .
2012 , , 746 , 55 trenti , m. & stiavelli , m. 2008 , , 676 , 767 wuyts , s. , labb , i. , schreiber , n. et al .
2008 , , 682 , 985 zheng , w. , bradley , l. d. , saraff , a. et al .
2012 , in seventh conference on astronomical data analysis , http://ada7.cosmostat.org/proceedings.php , 17 zheng , w. , postman , m. , zitrin , a. et al .
2012 , nature , 489 , 406 zitrin , a. , broadhurst , t. , umetsu , k. et al .
2009 , , 396 , 1895 zitrin , a. , menanteau , f. , hughs , j. p. et al .
2013@xmath129 , , 770 , l15 zitrin , a. , meneghetti , m. , umetsu , k. et al .
2013@xmath128 , , 762 , 30 zitrin , a. , zheng , w. , broadhurst , t. et al .
2014 , arxiv 1407.3769 ccccc & f160w&2013 aug.@xmath352014 jul .
& 71664 & 28.7 + & f140w&2013 oct.,nov .
& 28140 & 28.5 + & f125w&2013 oct.@xmath352014 jul . & 36620 & 28.6 + & f105w&2013 aug.@xmath352014 jul .
& 71209 & 29.0 + & f814w&2014 may@xmath35jul . ,
& 117518 & 29.4 + & f606w&2014 may , 2009 oct .
& 36882 & 29.2 + & f435w&2014 jun.@xmath35jul . , 2009 oct .
& 61909 & 29.3 + & irac1 & 2013 sep . , 2014 jan.,feb .
& 339291 & 25.5 + & & 2004 jun.,nov . &
& + & irac2 & 2013 sep . , 2014 jan.,feb . &
339291 & 25.3 + & & 2004 jun.,nov . & & + lccccccccc jd2 & @xmath130 & 3.597037 & @xmath131 & @xmath132 & @xmath133 & @xmath134 & @xmath135 & @xmath136 & @xmath137 + jd3 & @xmath138 & 3.603344 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 & @xmath143 & @xmath144 lccccccccc yd1 & @xmath145 & 3.603856 & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & @xmath152 + yd2 & @xmath153 & 3.572515 & @xmath154 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath136 & @xmath159 + yd3 & @xmath160 & 3.603858 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath143 & @xmath136 & @xmath165 + yd4 & @xmath166 & 3.603864 & @xmath167 & @xmath168 & @xmath169 & @xmath170 & @xmath171 & @xmath136 & @xmath152 + yd5 & @xmath172 & 3.579479 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 + yd6 & @xmath180 & 3.604005 & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath151 & @xmath152 + yd7 & @xmath186 & 3.603397 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & @xmath191 & @xmath136 & @xmath152 + yd8 & @xmath192 & 3.596096 & @xmath193 & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath151 & @xmath198 + yd9 & @xmath199 & 3.572902 & @xmath200 & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath136 & @xmath205 + yd10 & @xmath153 & 3.598108 & @xmath206 & @xmath207 & @xmath208 & @xmath209 & @xmath210 & @xmath151 & @xmath211 + yd11 & @xmath212 & 3.600947 & @xmath213 & @xmath214 & @xmath215 & @xmath216 & @xmath143 & @xmath136 & @xmath217 + lccccccccc zd1 & @xmath218 & 3.603582 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath143 & @xmath151 & @xmath152 + zd2 & @xmath223 & 3.604520 & @xmath224 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 + zd3 & @xmath231 & 3.606477 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath151 & @xmath237 + zd4 & @xmath238 & 3.605263 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath151 & @xmath230 + zd5 & @xmath244 & 3.588985 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath144 + zd6 & @xmath251 & 3.606575 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & @xmath151 & @xmath257 + zd7 & @xmath258 & 3.592285 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 + zd8 & @xmath266 & 3.579668 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 & @xmath136 & @xmath272 + zd9 & @xmath273 & 3.603208 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath178 & @xmath279 + zd10 & @xmath280 & 3.581282 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath136 & @xmath286 + zd11 & @xmath273 & 3.585321 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & @xmath291 & @xmath136 & @xmath292 + lcc jd2 & @xmath293 & @xmath294 + yd1 & @xmath293 & @xmath295 + yd2 & @xmath293 & @xmath294 + yd4 & @xmath296 & @xmath297 + yd6 & @xmath298 & @xmath299 + yd7 & @xmath300 & @xmath295 + yd8 & @xmath301 & @xmath294 + yd9 & @xmath293 & @xmath294 + yd10 & @xmath293 & @xmath294 + yd11 & @xmath293 & @xmath294 + zd1 & @xmath302 & @xmath303 + zd2 & @xmath304 & @xmath305 + zd3 & @xmath300 & @xmath306 + zd4 & @xmath293 & @xmath294 + zd6 & @xmath307 & @xmath308 + zd9 & @xmath309 & @xmath310 + lcccccccc zd7a1 & 3.592410 & -30.409897 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath316 + zd7a2 & 3.592160 & -30.409925 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 + zd7b & 3.588430 & -30.410340 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 + zd7c & 3.600940 & -30.400824 & @xmath329 & @xmath330 & @xmath331 & @xmath332 & @xmath327 & @xmath333 + | we report the discovery of 24 lyman - break candidates at @xmath0 , in the hubble frontier fields ( hff ) imaging data of abell 2744 ( @xmath1 ) , plus _ spiter_/irac data and archival acs data .
the sample includes a triple image system with a photometric redshift of @xmath2 .
this high redshift is geometrically confirmed by our lens model corresponding to deflection angles that are 12% larger than the lower - redshift systems used to calibrate the lens model at @xmath3 .
the majority of our high - redshift candidates are not expected to be multiply lensed given their locations in the image plane and the brightness of foreground galaxies , but are magnified by factors of @xmath4 , so that we are seeing further down the luminosity function than comparable deep field imaging .
it is apparent that the redshift distribution of these sources does not smoothly extend over the full redshift range accessible at @xmath5 , but appears to break above @xmath6 .
nine candidates are clustered within a small region of @xmath7 across , representing a potentially unprecedented concentration .
given the poor statistics , however , we must await similar constraints from the additional hff clusters to properly examine this trend .
the physical properties of our candidates are examined using the range of lens models developed for the hff program by various groups including our own , for a better estimate of underlying systematics .
our spectral - energy - distribution fits for the brightest objects suggest stellar masses of @xmath8 , star - formation rates of @xmath9 yr@xmath10 , and a typical formation redshift of @xmath11 . |
a two - user genie - aided cognitive interference channel ( cic ) is a two - user interference channel ( ic ) in which one of the transmitters ( termed the secondary transmitter , here @xmath0 ) knows the other transmitter s message ( termed the primary transmitter , here @xmath1 ) noncausally ( i.e. , by a genie ) .
devroye , mitran , and tarokh ( dmt ) in their premier paper , titled `` achievable rates in cognitive radio channels , '' derived an achievable rate region for the discrete memoryless cic ( [ [ dmt : cic ] , th . 1 ] ) .
we observe that the coding scheme proposed in @xcite is correct but unfortunately the derived achievable rate region is incorrect because of occurring some mistakes in decoding and analysis of error probability .
we first intuitively show that some rate - terms in the dmt rate region seem to be incorrect ( in fact , they are incomplete ) .
then , we correct the dmt achievable rate region and thereby show that the corrected achievable rate region includes the dmt rate region given in @xcite .
the two - user discrete memoryless cic ( dm - cic ) , denoted by @xmath2 consists of four finite alphabets @xmath3 and a collection of conditional probability mass functions @xmath4 on @xmath5 .
the channel is memoryless in the sense that @xmath6 . in this channel transmitter @xmath7 ,
wants to send a message @xmath8 , uniformly distributed on @xmath9 , to its respective receiver .
the primary transmitter @xmath1 generates the codeword @xmath10 as @xmath11 , and the secondary transmitter @xmath0 , being non - causally aware of the primary message , generates the codeword @xmath12 as @xmath13 . the decoding function @xmath14 is given by @xmath15 . a pair @xmath16 of non - negative real values is called an achievable rate for the dm - cic if for any given @xmath17 and for any sufficiently large @xmath18 , there exists a sequence of encoding functions @xmath19 , and a sequence of decoding functions @xmath20 , such that l p_e^(n)=p_r\ { g_1(y_1^n ) m_1 g_2(y_2^n ) m_2 | ( m_1,m_2 ) } where @xmath21 is the average probability of error .
the closure of the set of all achievable rate pairs @xmath16 is called the capacity region . in @xcite ,
devroye _ et al . _
, by using rate splitting , divided each message @xmath8 , @xmath22 , into two independent sub - messages : * common sub - message @xmath23 at rate @xmath24 ( to be sent from @xmath25 ) , * private sub - message @xmath26 at rate @xmath27 ( to be sent from @xmath28 ) , such that @xmath29 . in this
paper , auxiliary random variables ( rvs ) @xmath30 and @xmath31 represent the sub - messages @xmath23 and @xmath26 , respectively . moreover , rv @xmath32 is time sharing rv which is independent of all other rvs .
we now present the dmt achievable rate region for the two - user genie - aided dm - cic .
_ theorem 1 [ [ dmt : cic ] , th .
1 ] : _ let @xmath33 be the set of all joint distributions @xmath34 that factor as l p(q , u_1c , u_1p , u_2c , u_2p , x_1,x_2)= + p(q)p(u_1c|q)p(u_1p|q)p(x_1|q , u_1c , u_1p)p(u_2c|q , u_1c , u_1p)p(u_2p|q , u_1c , u_1p)p(x_2|q , u_2c , u_2p ) .
[ pdf_dmt ] for any @xmath35 , let @xmath36 be the set of all quadruples @xmath37 of non - negative real numbers
such that there exist non - negative real @xmath38 satisfying @xmath39 then * @xmath40 is an achievable rate region for the genie - aided dm - cic in terms of @xmath37 , * @xmath41 is the implicit description of the dmt achievable rate region where @xmath42 is the set of all pairs @xmath16 of non - negative real numbers such that @xmath43 and @xmath44 for some @xmath45 .
we first saw that the rate - terms ( 2.7)(2.9 ) and ( 2.13)(2.16 ) of @xmath36 intuitively seem to be incomplete because of not utilizing some dependencies among rvs .
for example , in ( 2.7 ) we have the main term @xmath46 . by considering the coding exploited in @xcite and @xmath33 and also since in the main term @xmath46 ,
rv @xmath47 is known ( or given ) and rvs @xmath48 are unknown , we expect that the dependency between known rv @xmath47 and unknown rvs @xmath49 as well as the dependency between unknown rvs @xmath50 and @xmath51 help communication and boost the rates . as we observe in ( 2.7 ) , the term @xmath52 is added to the main term but unfortunately , the term @xmath53 is not .
similarly , in ( 2.8 ) the dependency between @xmath47 and @xmath51 , in ( 2.9 ) the dependency between @xmath54 and @xmath51 , in ( 2.13 ) the dependency between @xmath55 and @xmath51 , in ( 2.14 ) the dependency between @xmath55 and @xmath47 , in ( 2.15 ) the dependency between @xmath47 and @xmath51 , and in ( 2.16 ) the dependencies among @xmath47 , @xmath55 and @xmath51 can help communication and boost the rates , while they are overlooked in @xmath36 . in this section
, we present the corrected version of the dmt rate region that utilizes the aforementioned dependencies among auxiliary rvs in boosting the rates . _ theorem 2 [ corrected th . 1 ] : _ for any @xmath35 , let @xmath56 be the set of all quadruples @xmath37 of non - negative real numbers
such that there exist non - negative real @xmath38 satisfying @xmath57 then * @xmath58 is an achievable rate region for the genie - aided dm - cic in terms of @xmath37 , * @xmath59 is the implicit description of the corrected dmt achievable rate region where @xmath60 is the set of all pairs @xmath16 of non - negative real numbers such that @xmath43 and @xmath44 for some @xmath61 .
the generation of the codewords @xmath64 and @xmath65 can be performed independently of @xmath66 and @xmath67 by using binning scheme . in other words ,
the codebook is generated according to the distribution * generate @xmath68 independent and identically distributed ( i.i.d . ) _
n_-sequences @xmath69 @xmath70 , each according to @xmath71 ; * generate @xmath72 i.i.d .
_ n_-sequences @xmath73 @xmath74 , each according to @xmath75 ; * generate @xmath76 i.i.d .
_ n_-sequences @xmath77 @xmath78 and @xmath79 , each according to @xmath80 ; ( i.e. , @xmath81 bins and @xmath82 sequences in each bin ) * generate @xmath83 i.i.d .
_ n_-sequences @xmath84 @xmath85 and @xmath86 each according to @xmath87 ; ( i.e. , @xmath88 bins and @xmath89 sequences in each bin ) the cognitive transmitter @xmath0 , being non - causally aware of @xmath95 and @xmath96 , to send @xmath92 and @xmath93 , first looks for indices @xmath99 and @xmath100 in bins @xmath92 and @xmath93 , respectively , such that l e^enc2_1=\{(q^n , u^n_1p(1),u^n_1c(1),u^n_2c(1,l_2c))a^(n)_(q , u_1p , u_1c , u_2c ) for all l_2c\{1 , , 2^nr^_2c } } + e^enc2_2=\{(q^n , u^n_1p(1),u^n_1c(1),u^n_2p(1,l_2p))a^(n)_(q , u_1p , u_1c , u_2p ) for all l_2p\{1 , , 2^nr^_2p } } [ enc.error ] l e^dec1_1=\{(q^n , u^n_1p(m_1p),u^n_1c(1),u^n_2c(1,l^*_2c),y^n_1)a^(n)_1 for m_1p 1 } + e^dec1_2=\{(q^n , u^n_1p(1),u^n_1c(m_1c),u^n_2c(1,l^*_2c),y^n_1)a^(n)_1 for m_1c 1 } + e^dec1_3=\{(q^n , u^n_1p(1),u^n_1c(1),u^n_2c(m_2c , l_2c),y^n_1)a^(n)_1 for m_2c 1 and l_2c l^*_2c } + e^dec1_4=\{(q^n , u^n_1p(m_1p),u^n_1c(m_1c),u^n_2c(1,l^*_2c),y^n_1)a^(n)_1 for m_1p 1 and m_1c 1 } + e^dec1_5=\{(q^n , u^n_1p(m_1p),u^n_1c(1),u^n_2c(m_2c , l_2c),y^n_1)a^(n)_1 for m_1p 1 , m_2c 1 and l_2c l^*_2c } + e^dec1_6=\{(q^n , u^n_1p(1),u^n_1c(m_1c),u^n_2c(m_2c , l_2c),y^n_1)a^(n)_1 for m_1c 1 , m_2c 1 and l_2c l^*_2c } + e^dec1_7=\{(q^n , u^n_1p(m_1p),u^n_1c(m_1c),u^n_2c(m_2c , l_2c),y^n_1)a^(n)_1 for m_1p 1 , m_1c 1 , m_2c 1 + and l_2c l^*_2c } [ dec1.errors ] where , for simplicity , @xmath116 is denoted by @xmath117 .
note that the probability of decoding error events will be evaluated by considering : ( i ) the encoding distribution ( i.e. , @xmath33 ) and the actual transmitted sequences , and ( ii ) the codebook generation distribution , the correctly decoded sequences and how to generate the sequences . as we mentioned earlier , for decoder 1 only the rate - terms ( 2.7)(2.9 ) are wrong , therefore we only evaluate the probabilities of @xmath118 , @xmath119 and @xmath120 .
the probability of the event @xmath121 can be bounded as @xmath122hence , @xmath123 goes to zero as @xmath113 if ( 3.7 ) is satisfied .
similarly , @xmath124 and @xmath125 can be bounded as @xmath126 @xmath127hence , @xmath128 and @xmath129 as @xmath113 if ( 3.8 ) and ( 3.9 ) are satisfied , respectively .
l e^dec2_1=\{(q^n , u^n_2p(m_2p , l_2p),u^n_2c(1,l^*_2c),u^n_1c(1),y^n_2)a^(n)_2 for m_2p 1 and l_2p l^*_2p } + e^dec2_2=\{(q^n , u^n_2p(1,l^*_2p),u^n_2c(m_2c , l_2c),u^n_1c(1),y^n_2)a^(n)_2 for m_2c 1 and l_2c l^*_2c } + e^dec2_3=\{(q^n , u^n_2p(1,l^*_2p),u^n_2c(1,l^*_2c),u^n_1c(m_1c),y^n_2)a^(n)_2 for m_1c 1 } + e^dec2_4=\{(q^n , u^n_2p(m_2p , l_2p),u^n_2c(m_2c , l_2c),u^n_1c(1),y^n_2)a^(n)_2 for m_2p 1 , l_2p l^*_2p , m_2c 1 + and l_2c l^*_2c } + e^dec2_5=\{(q^n , u^n_2p(m_2p , l_2p),u^n_2c(1,l^*_2c),u^n_1c(m_1c),y^n_2)a^(n)_2 for m_2p 1 , l_2p l^*_2p , m_1c 1 } + e^dec2_6=\{(q^n , u^n_2p(1,l^*_2p),u^n_2c(m_2c , l_2c),u^n_1c(m_1c),y^n_2)a^(n)_2 for m_2c 1 , l_2c l^*_2c , m_1c 1 } + e^dec2_7=\{(q^n , u^n_2p(m_2p , l_2p),u^n_2c(m_2c , l_2c),u^n_1c(m_1c),y^n_2)a^(n)_2 for m_2p 1 , l_2p l^*_2p , m_2c 1 , + l_2c l^*_2c and m_1c 1 } [ dec2.errors ] where , for simplicity , @xmath130 is denoted by @xmath131 .
as we mentioned earlier , for decoder 2 only the rate - terms ( 2.13)(2.16 ) are wrong , therefore we only evaluate the probabilities of @xmath132 , @xmath133 , @xmath134 and @xmath135 . the probability of the event @xmath136 can be bounded as where , is obtained by considering this fact @xmath55 and @xmath51 are dependent in general and the encoding distribution says that they are independent only when @xmath138 are given , i.e. , @xmath139 form a markov chain .
hence , @xmath140 goes to zero as @xmath113 if ( 3.13 ) is satisfied .
similarly , @xmath141 , @xmath142 and @xmath125 can be bounded as @xmath143 @xmath144@xmath145hence , @xmath146 , @xmath147 and @xmath148 as @xmath113 if ( 3.14 ) , ( 3.15 ) and ( 3.9 ) are satisfied , respectively .
this completes the proof of theorem 2 . 6 [ dmt : cic ] n. devroye , p. mitran , v. tarokh , `` achievable rates in cognitive radio channels , '' _ ieee trans .
inform . theory _
5 , pp . 18131827 , may 2006 .
[ gelfand : pinsker ] s. gelfand and m. pinsker , `` coding for channels with random parameters , '' _ probl . contr . and
theory , _ vol .
1 , pp . 1931 , 1980 . | in a premier paper on the information - theoretic analysis of a two - user cognitive interference channel ( cic ) @xcite , devroye _ et al . _ presented an achievable rate region for the two - user discrete memoryless cic . the coding scheme proposed by devroye _
et al . _ is correct but unfortunately some rate - terms in the derived achievable rate region are incorrect ( in fact incomplete ) because of occurring some mistakes in decoding and analysis of error probability .
we correct and complete the wrong rate - terms and thereby show that the corrected achievable rate region includes the rate region presented in @xcite .
achievable rate region , cognitive interference channel , gelfand-pinsker coding . |
the distortion of sound waves in materials with spatially varying index of refraction can be controlled using concepts from rays and high frequency propagation .
for instance , all ray paths can be made to behave according to a prescribed pattern , e.g. convergence at a point on the other side .
transformation acoustics ( ta ) goes further in making the material replicate an equivalent volume of `` virtual '' acoustic fluid which faithfully mimics the wave equation itself rather than some asymptotic approximation .
such ta - based gradient index ( grin ) lenses fall under the umbrella of acoustic metamaterials , a field which has seen tremendous innovation in recent years . however , to build a ta - grin lens in the laboratory often demands compromise between the frequency range of operation , transmission loss and lensing effectiveness , particularly in water , the acoustic medium of interest here .
a successful ta - grin lens simultaneously displaying high transmission and accurate wave steering can be achieved in water using a sonic crystal ( sc ) array of elastic scatterers .
quasi - periodic scs are capable of filtering , guiding and/or steering an incident wave based on a gradient of effective properties @xcite . unlike phononic crystals
, scs can not support shear waves in the bulk , hence , energy loss to mode conversion is minimized .
the localized effective acoustic properties of a sc element are an average of the fluid and contained elastic scatterer .
these depend on the shape , filling fraction , the effective bulk modulus and the effective density of the scatterer . in order to display the inhomogeneity required for a grin lens
, the properties of the elements have to differ in a quasi - continuous manner .
this has been successfully achieved in air @xcite and in water @xcite by fixing the lattice constant and varying the filling fraction of solid cylinder scatterers in the fluid unit cell . for air - based scs the cylinders can be modeled as rigid @xcite . for water - based scs
the elasticity of the scatterer is not only non - negligible , but essential in the modeling of such structures .
typical engineering materials , such as metals , are much denser than water leading to impedance mismatch and undesired scattering .
a solution is to use a hollow air - filled elastic shell which has an effective density and bulk modulus much closer to that of water as compared to the solid material .
the effective acoustic properties ( speed , impedance ) depend on the material of the shell and , in particular , on its thickness @xcite .
the sensitivity of the effective compressibility to shell thickness is a consequence of hoop stress in thin shells , which combined with the dependence of the effective density , results in the fact that thin shells have effective sound speed that is independent of thickness @xcite .
the effective impedance , on the other hand , is a linear function of thickness in the same thin shell approximation .
these two basic facts together indicate that by choosing the material and the thickness , it is possible to achieve a wide range of effective properties , as illustrated in the chart in figure [ f1 ] ( motivated by earlier work in @xcite ) .
this is the central idea in the present work . and
bulk modulus @xmath0 of hollow cylindrical shells for ten commonly available materials normalized relative to water , from eq . .
each curve shows the properties as a function of the relative thickness to radius ratio @xmath1 , from small to large as indicated by the arrow .
circles indicate the values for @xmath2 .
diagonal dashed lines indicate where the effective acoustic speed and impedance coincide with those of water .
( color online),width=326 ] the purpose of the present paper is to demonstrate the potential for ta - based grin lens design in water using the wide variety of shells available .
the transformation acoustics example considered in detail here is the cylindrical - to - plane wave lens discussed by layman et al .
@xcite it works by steering waves from a monopole source at the center away from the corners to the faces of the lens .
the sc of ref .
@xcite is based on constructive multiple scattering from finite embedded elastic materials in a fluid matrix , something previously investigated by torrent and sanchez - dehesa @xcite .
the grin lens device considered here expands the possibilities in ref .
@xcite by increasing the range of achievable properties over those presented by martin et al .
@xcite .
the cylindrical - to - plane wave lens is designed to increase radiation in specific directions .
enhanced directionality has also been experimentally observed @xcite for an acoustic source placed inside a two - dimensional square lattice phononic crystal operating at the band - edge frequency @xcite .
highly directional acoustic wave radiation is also possible in 2d pcs at pass band frequencies far away from the band edge states , as shown in simulations of a square lattice of steel cylinders in water @xcite .
the use of the band structure of a periodic square array to produce directional water wave radiation was proposed by @xcite , and subsequently demonstrated in experimental measurements on a 6@xmath36 array of surface - breaking cylinders @xcite with a monopolar source at the array center .
directional radiation has been demonstrated in air using a non - periodic array of cylinders to produce scattering enhancement in the forward direction @xcite .
martin et al .
@xcite produced acoustic grin focusing by changing the lattice constant in a pc with elastic shell elements .
parallel zigzag rigid screens have also been proposed as potential focusing and directional beaming devices @xcite . while the spatial filtering device described in this paper uses a fluid matrix , morval et al .
@xcite show directional enhancement of a monochromatic acoustic source into a surrounding water medium using a square array of cylinders in a solid matrix ; the 2-dimensional quadropolar collimation effect is based on square - shaped equifrequency contours of the phononic crystal @xcite . although the solid matrix has obvious practical advantage , the narrow frequency device of @xcite yields decreased amplitude in the preferential directions as compared with the free field radiation .
the ta - based device described here does not have these limitations , and shows for the first time as far as we are aware , broadband positive gain in a neutrally buoyant square grin lens , with obvious implications for low loss underwater application .
the outline of the paper is as follows .
transformation acoustics and the mapping for the cylinder - to - square lens are described in section [ sec2 ] .
acoustical properties of cylindrical shells are discussed in section [ sec3 ] and the proposed design using available cylindrical tubes is presented .
the experimental setup is described and acoustical measurements are discussed in section [ sec4 ] , with concluding remarks in section [ sec5 ] .
the transformation of a circular region to a square one can be achieved using a conformal change of coordinates .
conformal mapping is a special case of the general theory of transformation acoustics ( ta ) .
usually , in ta one can expect the material properties associated with a spatial transformation to display anisotropy .
this could be in the density @xcite or the bulk modulus @xcite , or in both simultaneously @xcite , but usually something has to become anisotropic .
conformal maps are unique in ta in that they do not require anisotropy . in this case both the inertial @xcite and the pentamodal @xcite forms of ta converge , and there is no ambiguity or degrees of freedom , a feature that distinguishes ta from its electromagnetic counterpart . at the same time , there is some confusion in the application of ta for conformal mappings , e.g. @xcite , so we briefly review the correct procedure @xcite .
we are concerned with a background fluid ( water ) of density @xmath4 and bulk modulus @xmath5 in which the acoustic pressure @xmath6 satisfies @xmath7 where @xmath8 is the speed of sound and time harmonic dependence @xmath9 is understood . under a conformal transformation @xmath10
the laplacian @xmath11 in the original variables becomes @xmath12 .
if we define the pressure as @xmath13 then @xmath14 satisfies the helmholtz equation in the mapped coordinates with transformed acoustic speed @xmath15 where @xmath16 .
this means that the transformed parameters are indeed isotropic , but it does not provide unique expressions for the individual parameters @xmath17 and @xmath18 , only the combination @xmath19 .
the necessary second relation comes from the requirement that the pressure in the transformed fluid arises from a particle displacement field @xmath20 which satisfies the momentum equation @xmath21 and the pressure constitutive relation @xmath22 .
eliminating @xmath23 gives the transformed helmholtz equation for @xmath24 if and only if @xmath18 is constant , which can be assumed equal to the original density . in summary ,
the transformed parameters are _ 1 = , k_1 = | z_1 ( z)|^2 k. the lens is based the transformation of a circle of diameter @xmath25 into a square of side @xmath25 , with the precise form of the circle - to - square mapping given in the appendix . in particular
, we note from eqs . and that the mapped value of the bulk modulus associated with the original point @xmath26 in the circle is k_1 = . along the principal directions
@xmath27 the bulk modulus decreases from the center of the square to a global minimum at the center of the sides . along the diagonals
@xmath28 it increases from its value at the center as it becomes unbounded at the four corners of the square .
the overall trend is illustrated in figure [ f2 ] .
[ h ! ] for the cylindrical - to - square mapping .
( color online),title="fig : " ]
consider a cylindrical shell of thickness @xmath29 and outer radius @xmath30 made of uniform solid with density @xmath31 , shear modulus @xmath32 , and poisson s ratio @xmath33 .
the interior is air filled , which in the context of water as the ambient medium in the exterior means that we can safely ignore the inertia and stiffness of the interior . the shell s effective density is the average value taken over the circular region of radius @xmath30 .
the effective bulk modulus is the value for which the radial compression of a uniform circular region of fluid under external pressure is the same as that of the shell under the same pressure , which follows from plane strain elasticity @xcite . in summary ,
_ & = ( 2h / a-(h / a)^2)_s , + k_&= _ s / ( 2(1-_s ) _ s/ _ - 1 ) .
the unit cell of the square array , shown in figure [ f3 ] , consists of a solid cylindrical shell surrounded by a complementary region of water .
the equivalent density and bulk modulus , @xmath34 , @xmath35 , of the unit cell depend on the properties of the surrounding fluid as well as the effective shell properties , according to [ -1 ] @xmath36 here @xmath37 is the shell volume fraction in the unit cell , where @xmath38 is the cylinder spacing as well as the side length of the unit cell .
since the required density from ta is @xmath39 , it follows that the shell effective density is also constant , @xmath40 .
the effective bulk modulus of the shell necessary to achieve the equivalent value from ta is = ( k^-1 + ( k_^-1 - k^-1 ) f^-1 ) ^-1 . as a function of the effective bulk modulus of the tuned shell @xmath0 for several filling fractions .
( color online),width=316 ] the equivalent bulk modulus of the unit cell is significantly affected by the surrounding fluid . with the exception of @xmath41
, all in - plane modes produce no volume change and hence do not change the effective bulk modulus of the unit cell .
no significant volume altering modes were observed in the frequency range considered .
shells of radius @xmath42 cm with a relatively tight packing of @xmath43 yields a filling fraction of @xmath44 . in this case , in order to have the effective quasi - static bulk modulus of the unit cell @xmath45 , the effective bulk modulus of the shell - springs - mass system must be @xmath46 , see figure [ f4 ] .
the proposed array contains 7 by 7 unit cells of size @xmath43 with @xmath42 cm giving a lens side length of @xmath47 cm .
the central square element is left empty , requiring 48 cylinders .
this was considered the minimal number necessary to provide both a reliable and an accurate gradient index effect .
the spacing was chosen to reduce the overall dimension of the lens as much as possible , without making the filling fraction unduly large .
inter - cylinder spacing in the fabricated lens was controlled from the two ends using preformed holders , see figure [ f8 ] below .
figure [ f5 ] shows the discretized values for the equivalent stiffness of each unit cell @xmath35 as determined from figure [ f2 ] by spatial averaging .
the effective properties of the shells are obtained from equation with @xmath44 , using the required equivalent stiffness of each unit cell @xmath35 in figure [ f2 ] . as noted above
, this means that effective properties of the shells must be more extreme than those implied by the mapping alone .
the effective density of each shell is tuned to water . in the 7x7 array .
the central element is absent ( i.e. water ) in the constructed design.,width=307 ] the three primary design criteria were : 1 ) that the shells are readily available , 2 ) the effective density of each shell approximately matches water and 3 ) the effect is apparent in the designated frequency range of interest : near 20 to 25 khz .
the shells must be sub - wavelength in dimension . furthermore all shells are required to have nearly the same outer diameter ; therefore
, the common outer diameter of 0.5 inches is selected as practical . fixing this outer dimension
leaves two parameters : the shell material and its relative thickness @xmath48 .
the range of effective properties as a function of both the shell material and thickness are succinctly summarized in figure [ f1 ] .
several features are apparent from the chart in figure [ f1 ] .
first , it is clear that the ten materials considered provide a comprehensive range of effective properties . for each material ,
the effective properties are approximately linear functions of the shell thickness for thin shells @xmath49 , with some curvature at larger values of @xmath1 .
the present design requires shells with effective density equal to that of water , which restricts values of @xmath1 to those near the vertical dotted line .
table [ table1 ] summarizes the properties of available shells which have nearly the same density as water , but varying effective bulk moduli . [
h ! ] .readily available shells ( i.e. tubes and pipes ) with different effective bulk moduli that have effective density close to that of water .
all properties are normalized to water . [
cols="<,^,^,^,^,^,^,^,^",options="header " , ] for the final design we considered only commercially available tubes made from a variety of materials with standard values of radius and thickness . as figure [ f1 ] illustrates , this provides a surprisingly wide range of possible properties , with the added advantage of allowing us to fabricate the lens with minimal effort and cost .
based on the available candidates from table [ table1 ] we selected nine different shells as shown in figure [ f6 ] for the fabricated lens . .
the actual thicknesses of the individual shells are indicated .
( color online),width=259 ] + the total pressure field of the cylindrical - to - square lens made of elastic shells was obtained by numerical computation using comsol .
figure [ f7 ] shows a simulation for a monopole source of frequency 22 khz in the center of the lens .
also shown for comparison are the pressure fields for the lens with the unit cells replaced by the effective acoustic medium and the free field radiation of the monopole source .
the simulations indicate that the cylindrical - to - plane wave lens made from the distribution of nine distinct empty shells performs very well as compared the optimal case of each unit cell having the prescribed effective acoustic properties directly from the conformal mapping .
it is also evident that the transmission is enhanced in the four principal directions .
the use of empty shells in water acoustics opens up the likelihood of exciting flexural resonances .
this is not normally a concern when dealing with isolated shells because the flexural waves are subsonic in speed and hence do not radiate .
the present design places the shells in close proximity , leading to the possibility of coherent flexural wave interaction , which can lead to strong scattering .
this effect can have positive or negative consequences , depending on one s immediate goal . in the present situation the shells are of varying thickness and comprised of different materials , with the result that the flexural resonances are spread over many frequencies , which decreases the possibility for coherent interaction .
in particular , we note that no such coherent effects were observed in the experiments ( see next section )
. related and surprising constructive interference effects resulting from coherent interaction of flexural waves in closely packed arrays of shells in water are described elsewhere @xcite .
the device pictured in figure [ f8 ] was fabricated to validate the cylindrical - to - plane wave lens design .
the device tested has the cylinder positions , radii , and material properties provided in figure [ f6 ] and table [ table1 ] .
although the model presented section [ sec3 ] is strictly two - dimensional ( 2d ) , which would suggest experimental validation using a 2d water waveguide , the test facilities available to the authors required a three dimensional ( 3d ) test configuration .
details of the configuration and rationale for their selection are provided here .
the as - tested lens is constructed from cylindrical rods 1 m in height and sealed at either end with urethane end - caps to prevent water intrusion .
the cylinders are clamped between 2 cm thick acrylonitrile - butadiene - styrene ( abs ) plates using tensioned monofilament .
the plates were machined to precisely locate the top and bottom of the cylinders in the positions dictated by the design .
note that an added benefit of the urethane end - caps is that they provide some level of vibration isolation between the end plates and the cylinders
. one key challenge to accurately measure the performance predicted in section [ sec3 ] was to minimize the effects of the finite height of the lens and thus observe its 2d response .
the associated practical difficulty encountered was in the selection and placement of the appropriate acoustic source .
validation of the lens design implies the need for an axis - symmetric source pressure along the vertical axis in the 3d lens , but no such source was available to the authors nor could one be easily constructed .
acoustic reciprocity , described below , was invoked to resolve this difficulty .
reciprocity is a fundamental principal of quiescent acoustic media , first fully described for acoustics by rayleigh @xcite ; it states that the interchange of source and receiver will lead to the same measured acoustic field if the environment is un - perturbed . specifically , if one excites acoustic waves at some point , _ a _ , then `` the resulting velocity potential at a second point , _ b _ , is the same both in magnitude and in phase , as it would have been at _ a _ had _ b _ been the source of sound@xcite . ''
applying this principal to the problem at hand , it is possible to replace the axis - symmetric source at the center of the cylindrical - to - plane wave lens with a point receiver and then measure the acoustic field at the center of the lens due to a plane wave incident from a specified radial angle . by varying the angle incidence of the plane wave , it is thus possible to construct the far - field radiation pattern expected from an axis - symmetric source placed at the center of the lens .
the only remaining problem is the generation of plane waves at a specified angle of incidence .
this is achieved using a spherical wave source located sufficiently far away from the lens such that the phase of the pressure field impinging on the lens aperture has variations less than 1@xmath50 across the entire frequency band of interest . for the lens geometry and frequencies considered ,
this can be achieved by at least 10 m separation between the spherical source and the lens . .
( color online ) ] the experiment was conducted at the lake travis test station ( ltts ) of the applied research laboratories ( arl ) at the university of texas at austin .
an a48 hydrophone and associated pre - amplification electronics , which was fabricated and calibrated by the underwater sound reference division ( usrd ) of the naval undersea warfare center ( nuwc ) , was located at the center of the lens .
this hydrophone has less than 1 db of variation across the entire frequency range of interest for this experiment , which was 1540 khz .
the acoustic source was an omni - directional itc-1032 fabricated by channel technologies group .
the source and lens with internal hydrophone were submerged to a depth of 5.5 m with a separation distance of 10 m. the lens was attached to a column capable of angular rotation through 360@xmath50 .
the source is then driven with 2 ms tone bursts from 1540 khz at 2.5 khz intervals and the time - series voltage output from the hydrophone was collected from 0 - 360@xmath50 at approximately 0.5@xmath50 intervals using a sampling frequency of 512 khz .
this process was then repeated for the hydrophone without the lens as a reference and referred to as the bare hydrophone case .
representative results from the series of experiments conducted on the cylindrical - to - plane wave lens are summarized in figures [ f9 ] and [ f10 ] .
the results were obtained by performing post - processing of the time - series data output from the hydrophone , described next .
for each angle and frequency , the steady - state portion of the tone - burst is identified through inspection of the time - domain voltage signal and a time gate is set so that only the steady state portion is considered .
the magnitude of the signal at each frequency and angle combination is then found by averaging the magnitude of the complex envelope of the received voltage signal during its steady - state response .
this process is carried out for the both measurement configurations ( hydrophone in the lens and bare hydrophone ) .
the frequency- and angle - dependent gain is then calculated as @xmath51 $ ] .
representative polar plots for @xmath52 15 , 22.5 , 25 , and 40 khz from experimental data and 2d finite element models are shown in figure [ f9 ] .
agreement between model and measurement for the both gain and angular dependence ( beam pattern ) match very well , with the location of the main lobes observed at 4@xmath50 , 91@xmath50 , 176@xmath50 , and 268@xmath50 on average across all frequencies inspected ( with the exception of the 20 khz case as described below ) .
unexpected variations in beam pattern between predicted and measured performance are likely owing to imperfections in the constructed device .
one very important observation of this data is the broadband performance of this metamaterial lens .
the broadband nature of the response is clearly demonstrated by the results provided in figure [ f10 ] , which shows the measured half - power beam width ( -3 db points ) and on - axis gain averaged at across all four main lobes .
the data clearly show that the as - tested lens provides broadband on - axis gain and beam - widths ranging from approximately 15@xmath50 -
30@xmath50 for frequencies from 22.5 - 40 khz , respectively . finally , it is important to note that the red shaded region in figures [ f10 ] indicates a regime of flexural tube resonances @xcite where the lens behavior was significantly degraded .
this experiment provides clear validation of the broadband impedance matched lensing effect provided by hollow cylinder metamaterial elements . .
the shaded region ( 17.5-22.5 khz ) denotes a flexural tube resonance regime predicted in @xcite .
broadband gain and narrow beamwidth is apparent over the entire range of frequencies inspected .
( color online ) ]
the results of this paper have shown the practical potential of using cylindrical elastic shells as elements in acoustic metamaterial devices .
the demonstration test device considered is a cylindrical - to - plane wave structure for which the required element properties are determined from transformation acoustics .
the size and material composition of the elements in the square array are chosen based on availability of shells , minimizing fabrication difficulties .
the device has the added advantage that is neutrally buoyant by virtue of the transformation acoustics design .
simulations indicated the operating frequency response of the final design would display a surprisingly broadband effect , which is verified in the experimental findings .
the underwater measurements show effective conversion of the monopolar source to quadropolar radiation over an octave band ( 20 to 40 khz ) with positive gain in the desired directions , all despite the minimal number of elements used .
these features have been demonstrated for the first time in a water - based acoustic lens device .
future research will consider other device designs using cylindrical shell passive amm elements .
the schwartz - christoffel conformal transformation of the unit disk to a square has been used previously for lens design via transformation optics @xcite and transformation acoustics @xcite . here
we provide a simpler form of the transformation than that given in @xcite .
our objective is a transformation from the plane of the unit circle , defined by the complex variable @xmath53 , to the plane containing the mapped square , defined by the complex variable @xmath54 ( for `` square '' ) .
we first map the interior of the unit circle to the upper half plane of the variable @xmath55 through a bilinear transformation as @xmath56 with @xmath57 .
the mapping that takes the upper half of the @xmath55-plane to the @xmath54-plane containing the square is a special case of the more general mapping known for mapping to polygons .
thus , consider @xmath58 where @xmath59 taking @xmath60 , @xmath61 , @xmath62 , we find @xmath63 where @xmath64 is the incomplete elliptic integral of the first kind .
the parameters @xmath65 and @xmath66 are found by setting @xmath67 , @xmath68 , and using @xmath69 , @xmath70 , where @xmath71 is the complete elliptic integral of the first kind and @xmath72 is the gamma function .
hence , in terms of the original @xmath53-plane containing the unit circle @xmath73[1 - 14 ] equation and its inverse map the boundary points in the n , s , e , w , ne , nw , se and sw directions in the circle and square plane to one another .
the density and bulk modulus are functions of the derivative of the mapping function .
the derivative of is found from @xmath74 and @xmath75 , which gives @xmath76 .
hence , for @xmath77 , @xmath78 and @xmath79 , latexmath:[\[\label{1 - 15a } this work was supported by onr through muri grant no .
n00014 - 13 - 1 - 0631 and uli grant no .
n00014 - 13 - 1 - 0417 .
many thanks to dr .
maria medeiros of onr ( code 333 ) and dr .
stephen oregan of nswccd ( code 7220 ) . m. ke , z. liu , p. pang , c. qiu , d. zhao , s. peng , j. shi , and w. wen .
experimental demonstration of directional acoustic radiation based on two - dimensional phononic crystal band edge states .
, 90(8):083509 , 2007 .
t. p. martin , c. j. naify , e. a. skerritt , c. n. layman , m. nicholas , d. c. calvo , g. j. orris , d. torrent , and j. sanchez - dehesa .
transparent gradient - index lens for underwater sound based on phase advance .
, 4(3 ) , sep 2015 .
b. morvan , a. tinel , j. o. vasseur , r. sainidou , p. rembert , a .- c .
hladky - hennion , n. swinteck , and p. a. deymier . ultra - directional source of longitudinal acoustic waves based on a two - dimensional solid / solid phononic crystal .
, 116(21):214901 , dec 2014 .
c. a. rohde , t. p. martin , m. d. guild , c. n. layman , c. j. naify , m. nicholas , a. l. thangawng , d. c. calvo , and g. j. orris .
experimental demonstration of underwater acoustic scattering cancellation .
, 5:13175 , aug 2015 .
v. romero - garcia , c. lagarrigue , j .-
groby , o. richoux , and v. tournat .
tunable acoustic waveguides in periodic arrays made of rigid square - rod scatterers : theory and experimental realization .
, 46:305108 , 2013 .
j. snchez - prez , d. caballero , r. mrtinez - sala , c. rubio , j. snchez - dehesa , f. meseguer , j. llinares , and f. glvez .
sound attenuation by a two - dimensional array of rigid cylinders .
, 80(24):53255328 , jun 1998 .
a. s. titovich and a. n. norris .
acoustic scattering from an infinitely long cylindrical shell with an internal mass attached by multiple axisymmetrically distributed stiffeners .
, 228:134153 , march 2015 .
j. o. vasseur , b. morvan , a. tinel , n. swinteck , a .- c .
hladky - hennion , and p. a. deymier .
experimental evidence of zero - angle refraction and acoustic wave - phase control in a two - dimensional solid / solid phononic crystal . , 86(13 ) , oct 2012 . | the use of cylindrical elastic shells as elements in acoustic metamaterial devices is demonstrated through simulations and underwater measurements of a cylindrical - to - plane wave lens .
transformation acoustics ( ta ) of a circular region to a square dictates that the effective density in the lens remain constant and equal to that of water . piecewise approximation to the desired effective compressibility is achieved using a square array with elements based on the elastic shell metamaterial concept developed in @xcite .
the size of the elements are chosen based on availability of shells , minimizing fabrication difficulties .
the tested device is neutrally buoyant comprising 48 elements of nine different types of commercial shells made from aluminum , brass , copper , and polymers .
simulations indicate a broadband range in which the device acts as a cylindrical to plane wave lens .
the experimental findings confirm the broadband quadropolar response from approximately 20 to 40 khz , with positive gain of the radiation pattern in the four plane wave directions . |
for the development of the long - term economic strategy and for realization of appropriate investment policy one needs the theoretical model of economic growth .
all developed countries pay a considerable attention to researching of such theoretical models , as well as to developing of corresponding instrumental means that are important for calculation of the concrete prognoses and programs [ 1 ] .the practical experience , accumulated by many countries , demonstrates that the most effective tool for the development of strategic directions of an economic policy on a long - run period are the special economic - mathematical models of small dimension ( macromodels ) .
such models are elaborated on the base of the theory of economic growth [ 2 ] . in traditional macromodels the principal attention
is payed to the estimation of future dynamics of the investments that determine trajectories of the economic growth .
however , economic growth depends not only on scales of investment resources , enclosed in economics : these trajectories are also determined by a number of the so - called quality factors [ 3 ] .
moreover , the final economic growth results depend basically on these factors . the orientation of economy to the extensive growth by escalating of wide area investments only can not ensure an achievement of the useful final results without paying attention to the quality factors .
the history of development of many countries , including the former ussr , shows the possibility of economical development by the principle of production for the sake of production .
economical system absorbs huge investment resources and augments volumetric parameters .
nevertheless this type of investment policy is not able to raise essentially standards of living .
the limited prospects of the economic development on the base of only extensive factors are demonstrated in the solow model [ 4 ] .
the estimation of influence of technical progress is made on the basis of the solow model with technical progress [ 2 ] .
this model accounts for the contribution of technical progress in the simplest way .
it is based on the rather relative concept of autonomous progress advance .
the intensity of the autonomous progress effect on production growth is completely determined by the time factor .
a quantitative estimations of such effect are received on the basis of the production functions method .
the parameters for this method are calculated by means of econometric processing of dynamic rows that define change of production volumes .
a row of successive numbering of time periods ( years , quarters or other periods ) is used as a dynamic row , that corresponds to the change of technical progress . in essence another approach to define the quality factors contribution to the growth of production volumes has been realized in growth model with effectiveness .
this approach does not use insecure parameters of rather relative econometric models .
it is based on a direct estimation of the change of an economic system material capabilities that are indispensable for the realization of social purposes .
the growth model with effectiveness is most adequate for simulation of economic growth in the present economical situation for the republic of belarus .
the main purpose of this work is to formulate the optimal planning problem for the growth model with effectiveness and solution of this problem : such solution is especially important for researches of economic growth problems , because it answers the main question that appears when choosing the investment policy .
the essence of this question consists in choosing between maintenance of current demand ( consumption ) and maintenance of the future demand ( capital investment ) .
the solution of the optimal planning problem allows to point such investment policy , at which the economic system works at best . the formulation of the depends on the purposes that the economic system has . in this work one
takes the most realistic version of such purpose - maximization of the welfare integral which is consumption per man during the modeled period of time .
the article is organized as follows : section ii deals with the concept of production effectiveness . growth model with effectiveness is described in section iii . in section
iv one considers optimal planning problem formulation .
maximum principle is applied in section v. solution of the optmal planning problem is analyzed in section vi .
policy options are considered in section vii .
main achievements of this work are briefly concluded in section viii .
the growth model with effectiveness uses modern approach to predict the economic growth trajectory and define the quality factors contribution in increasing of production volumes .
this approach does not rest on accident - sensitive parameters of rather conditional econometric models , but it is grounded on the direct estimation of those changes of material capabilities of an economical system that are necessary for realization of the chosen purposes . on the basis of such estimation it is expedient to draw a conclusion about the advance , reached in economics : one admits the existence of `` advance '' only in the case , when the capabilities for implementation of the social and economic purposes are extended . at the same time , from the standpoint of the target approach , opening capabilities scales shows the degree of the obtained advances . within the framework of such an approach there is a more successful name for the identification of the advance , reached in economics .
this name is `` increase of production effectiveness '' ( or , accordingly , its reduction , if the capabilities for the implementation of the social and economic purposes are reduced ) .
this term is used in this work to reflect all quality factors influence on economic growth .
thus , the primary task of this chapter is to design a criteria index that should characterize change of production effectiveness on the macrolevel .
such an index should be used as one of the most important variables of the dynamic macromodel , which reflects the relationship of this macromodel with other macroeconomic indexes that describe the intensity of production and use of resources , as far as accumulation and non - productive consumption . within the framework of the dynamical macromodel of economical growth
, the submodel of the production effectiveness has been developed from the standpoint of the target approach [ 5 ] .
the economical system under consideration is supposed to be closed in the sense that the total amount of the needs is fulfiled at the expense of the production only .
the system does not move its production into other countries in the debt or gratuitously ( that does not eliminate barter with an environment on the equivalent basis for the coordination of commodity pattern of production with consumption pattern ) . to define an effectiveness criterion of production it is necessary to reveal the certain economical form of the operational outcome of the economical system and resources , used for its achievement . for this purpose
one needs to define not only spatial - organizational , but also temporal boundaries of this system in order to reveal and to agree final output indexes with the initial input factors involved in the process of production .
if we abstract from the natural and external economical factors and consider the effectiveness of an economical system , without taking into account its temporal boundaries , then the labour force is the only used resource , and the amount of created material benefits and services intended only for the non - productive consumption is the outcome of the system operation . in this case
it would be possible to define the effectiveness index on the basis of the simple comparison of the total final consumption and used labour force .
it is always necessary to associate an estimation of effectiveness with the particular time frame , therefore time also should play a role of the factor that limits the frameworks of the economical system .
the inputs and outputs of the system reflect not the connection with the environment .
the evolution of the system depends on the past and defines the future evolution of the system .
the relations of the resources reproduction costs to their volumes are the relevant characteristics of the reproduction process .
for example the consumption per man @xmath0 reflects the level of workers material benefits .
any change of this index substantially determines the dynamics of all indexes of the population welfare .
the relation of gross investment @xmath1 to a volume of accumulated productive capital @xmath2 @xmath3 predetermines the rate of capital growth in a decisive measure .
the estimations @xmath0 and @xmath4 are connected with the indexes of labour productivity @xmath5 and capital per man @xmath6 as follows @xmath7
this ratio is a consequence of the balance identity @xmath8 , where @xmath9 - output , @xmath10 - consumption , which originates from the assumption about closure of the considered economical system ( see fig.1 ) . to derive the equation ( 2 ) from the balance identity
, one has to divide it by the labour force volume @xmath11 and take into consideration that @xmath12 relative resource estimations @xmath0 and @xmath4 can be accepted as objective functions that describe the implementation degree of two main purposes of the reproductive process .
these purposes form one general purpose of an economical system . here
we denote it as `` integral purpose '' .
it implies the maximization of satisfaction of current and future needs of society .
the index of effectiveness should serve as the objective function that permits to estimate quantitatively the integral purpose implementation . for the concrete definition of the integral purpose and the integral index of effectiveness
it is necessary to distribute the priorities between two introduced primary purposes .
this distribution leads to their coordination and resolving of the contradiction between the primary purposes .
such priorities are reflected quantitatively at distribution of the gross internal product to parts intended for reproduction of two kinds of manufacturing resources .
the purposes of the society define the proportions of such distribution and concretize the effectiveness index formula that should serve as the integral object function of the reproduction process .
below all variables are counted for the given year . here
we assume that during t - th year the usage of the manpower quantity @xmath13 and capital of the volume @xmath14 makes the gross internal product of the volume @xmath15 .
the corresponding values of the relative indexes of labour productivity and capital per man are peer @xmath16 and @xmath17 , respectively .
the indexes @xmath16 and @xmath17 are the most important quantitative characteristics of the process of reproduction as generalized technology on the macrolevel .
though their values do nt determine base - line values of the object functions @xmath18 and @xmath19 uniquely , but they limit the area of their possible values . in accordance with the general limiting condition
the above indicated main balance identity reads @xmath20 the particular values of estimations @xmath18 and @xmath19 can be derived from the area , restricted by the equality ( 4 ) .
this area depends on the distribution of the made gross internal product , which arrests the norm of accumulation @xmath21 .
the last is defined as the fraction of the gross accumulation in the total amount of the gross internal product @xmath22 at the given norm of accumulation and the certain labour productivity level as well as capital per man , the values of object functions are determined uniquely according to the formulas @xmath23 their values are the starting point for the analysis of the effectiveness dynamics .
here we consider the values of variables for the next year @xmath24 .
the capital per man increases to the value of @xmath25 .
the labour productivity for the given capital per man reaches the value @xmath26 .
the dynamics of the reproduction process effectiveness characterizes the change of the implementation capabilities of two main purposes that have to be reflected in the change of area of acceptable values of the object functions @xmath4 and @xmath0 . in order to estimate the degree of the indicated change quantitatively one needs to compare labour productivity in current year with the value @xmath27 , which describes that minimal production volume per man
this allows to keep the values of the object functions @xmath4 and @xmath0 at the level of the basic year .
@xmath28 the last corresponds to the equal effectiveness of production in current and basic years .
it means that the values of both object functions can be saved at a basic level and both can not be increased at once .
note that actual values of estimations @xmath29 and @xmath0 in one year @xmath24 can differ from @xmath30 and @xmath31 , but one of them will be more basic , and another will be less . assume the labour productivity level in one year @xmath32 surpasses the bound @xmath33 therefore in this year there appears the possibility for simultaneous increase of values of two object functions in contrast to the year @xmath34 .
as mentioned before these object functions describe the level of implementation of the main purposes of the reproduction process .
it gives the ground to suppose that in one year @xmath32 the value of the integral object function is augmented and , therefore , the level of efficiency is raised .
moreover , the value of a difference @xmath35 allows to judge that there appeared capability to increase the values of object functions @xmath4 . therefore this difference can serve as the characteristic of the effectiveness increase degree .
the actual increment of labour productivity @xmath36 can be decomposed in two parts @xmath37 the value @xmath38 represents the increment of the labour productivity , at which the level of effectiveness remains invariable .
it can also be interpreted as the increment of theproductivity reached at the expense of the extensive increase of the capital per man ( at the basic level of the effectiveness ) .
another part of the increment of labour productivity @xmath39 quantitatively characterizes the additional capabilities of implementation of the main purposes of production appeared in one year @xmath40 .
we start from the reason that the growth of the effectiveness is the only source of increase of such capabilities .
therefore it is possible to consider the value @xmath41 as the increment of the productivity reached at the expense of the increase of efficiency .
the ratio of the increment @xmath41 to the base - line value of the labour productivity @xmath42 represents the rate of the productivity increment at the expense of the effectiveness increase .
this ratio can be identified with the rate of the increment of the effectiveness of production . in equation ( 10 )
we replace @xmath26 with @xmath43 and @xmath44 with @xmath45 and simplify it using the balance identity ( 2 ) .
the rate of the increment of effectiveness depends on the pure increments of the labour productivity @xmath46 and capital per man @xmath47 @xmath48 the equation ( 11 ) can be rewritten using the absolute increases of the corresponding indexes to the rates of their change . for this purpose one
substitutes the intensity of reproduction of the capital from the formula ( 6 ) into equation ( 11 ) : @xmath49 here we introduce the notations for rates of increment of the corresponding indexes @xmath50 and @xmath51 .
then the estimation of dynamics of a production effectiveness takes the form @xmath52 thus , the conducted analysis results in a conclusion that the reference point for the society that seeks to achieve the social and economic purposes , ca nt serve the increase of labour productivity , which traditionally was considered to be the main criteria index in our economics .
the obtained two equivalent dynamical formulas of the index of a production effectiveness ( 11 ) and ( 12 ) demonstrate that for the estimation of outcomes of the economical development it is necessary to correct the growth of labour productivity allowing for the change of the capital per man .
the bigger part of the made gross internal product is routed to the reproduction of the productive capital , the more is the deviation of the dynamics of the criteria index of effectiveness from dynamics of labour productivity .
the last can serve as the main reference point of control only in that case , when the capital per man remains invariable .
in this work we consider the usage of the effectiveness index as the preferred reference point of control .
the following six main macroeconomic indexes have been selected as variables of the growth model with effectiveness @xmath6 - capital per man , @xmath5 - output per man , @xmath0 - consumption per man , @xmath4 - saving per capital , @xmath53 - sum of amortization rate and population rate , @xmath54 - the rate of economics effectiveness . for simplification of model and its analysis
the following preconditions are adopted : the number of workers in economics changes with constant rate of increment @xmath55 , ( so that the dynamics of employment can be recorded with the help of the function @xmath56 ) .
the amortization rate @xmath57 is also invariable . besides the differential form of notation
this form is more convenient for solution and analysis and widely used in the literature on economic - mathematical modeling [ 5 ] . for conversion to the differential form one supposes , that the values of the considered economical indexes are continuously differentiable functions of time .
the increments of indexes per unit of time , selected as a step for the analysis , are substituted by derivative from functions that describe their dynamics .
the main equation of the model is based on the definition of the effectiveness index .
the formula ( 11 ) , derived earlier for the estimation of dynamics of effectiveness and introduced in the incremental form , can be converted to the following equivalent equation in the differential form @xmath58 the differential equation ( 13 ) is included in the developed macroeconomic model and plays the role of production function , which links indexes of capital per man and labour productivity .
besides , the model includes the main balance identity ( 2 ) @xmath59 the ratio ( 6 ) can be rewritten in differential form as follows @xmath60 the differential equation for @xmath6 comes from the equation of capital dinamics @xmath61 to pass from the increment of capital to the index of the increment of capital per man , it is necessary to differentiate the formula for capital per man @xmath62 here we introduce the population rate @xmath63 .
the last assumption allows to rewrite equation ( 17 ) in the following form : @xmath64 now in ( 18 ) instead of @xmath65 we put equivalent expression from ( 3 ) , and for the sum of two constants @xmath66 we enter new identification @xmath67 .
then it is possible to enter into the model the equation , that reflects the correlation between capital per man and saving per capital .
@xmath68 equations ( 13 ) - ( 15 ) and ( 19 ) form the set of four equations , which describes the correlations between six abovelisted main macroeconomic indexes .
it is nessesary to emphasize that this model is represented by six variables and four equations .
therefore , the obtained set of equations ( 13)-(15),(19 ) is incomplete . in order to close the set of equations
one introduces control variables that describe investment policy .
it is possible to do by assuming @xmath69 and @xmath70 is a control variable
. then one can formulate optimal planning problem .
first , it is necessary to construct a target functional .
the task of the central planning establishment is to select a feasible trajectory @xmath70 that is the optimum for achievement of some economic target .
the economic target of central planning organ should be based on the standards of living , estimated by the consumption level . in particular
, it is presumed that the central planning organ has an utility function , which determines utility @xmath71 at any moment of time as a function of consumption per man @xmath72 $ ] .
we assume that the utility function is doubly differentiable and that the marginal utility is positive and non - increasing function .
therefore , the utility function is concave and monotonically increasing function .
the utility function determines utility in a certain year .
however , the problem of the central planning organ is to select the whole trajectory of consumption per man . for this purpose
it is necessary to compare indexes of utility that correspond to different years .
suppose , that utilities in different years do not depend on each other .
the utility in any year does not depend on either consumption or utility in any other year directly .
utilities for different years are supposed to be additive .
this assumption is based on the fact , that the proximate consumption is more important , than distant .
we suppose that the norm of discounting @xmath73 is constant and positive .
the bigger norm of discounting testifies the greater preference of the utilities closest in time . suspecting the exponential behavior of the discounting
, we receive the value of utility in year @xmath34 equal to @xmath74 $ ] . during the indicated time period from @xmath75
to @xmath76 the welfare @xmath77 , corresponding to the pathway of consumption per man @xmath78 , is determined by integrating of all instantaneous utilities over the whole interval . the time planning horizon @xmath76 can be finite or infinite . in case
this time is finite , it is necessary to set the minimally acceptable value of capital per man in final year to ensure the possibility of consumption outside the given horizon of time .
here we try to avoid difficulties , connected with the definition of the minimal value of capital per man in a final moment of time .
consider , that @xmath76 is indefinite , so the control trajectory is selected for all times in the future .
however , in this case the welfare integral can miss .
the convergence of an integral is guaranteed , if the following conditions are satisfied : the initial value of capital per man is less than the maximal accessible level @xmath79 and the norm of discounting is positive . in this case @xmath80 and @xmath81 so the integral of welfare is bounded above .
we choose @xmath82 and @xmath83 as state variables . we express consumption per man through control and state variables from ( 14 ) .
then we receive optimal planning problem for model with effectiveness @xmath84
in this section we derive the equation for the saving per capital trajectory . when solving optimal planning problems with the help of the maximum principle , for each coordinate of state variables vector @xmath85 a costate variable is used .
the hamiltonian function is @xmath86 , where @xmath87 stands for the integrand of a target functional , @xmath88 for the vector of motion equations right parts , @xmath89 for control .
we find functions @xmath90 , @xmath91 , @xmath92 that satisfy the following conditions @xmath93 .
% ( 25 ) \end{aligned}\ ] ] where @xmath94 , @xmath95 are costate variables . finally , by using of eqs .
( 22 ) - ( 24 ) we can write the problem to be solved in the following form @xmath96 the system of five equations with five unknown functions is then obtained .
four equations of this system are differential , one is algebraic . solution of this system of equations is equivalent to the solution of the problem ( 21 ) .
the set of equations ( 26 ) - ( 30 ) has the solution for any kind of the utility function .
however in this work we perform the analysis of the elementary case @xmath97 . in this case
( 29 ) and ( 30 ) for costate variables @xmath94 and @xmath95 do not depend on state ones @xmath6 and @xmath5 . in order to find an optimal savings trajectory @xmath98 we have to solve only three equations ( 26 ) , ( 29 ) , and ( 30 ) .
after that one can find state variables @xmath6 and @xmath5 . as a first step we obtain from eq .
( 26 ) an explicit formula for @xmath98 written as a function of costate variables @xmath94 and @xmath95 : @xmath99 where @xmath100 is an exponential function .
note , that the solution of eq.(30 ) for @xmath95 can be written in the analytical form @xmath101 where @xmath102 .
having differentiated the left and right parts of eq .
( 31 ) , one arrives at @xmath103 by rewriting eq .
( 31 ) in the following form @xmath104 and substituting it in the right part of eq .
( 33 ) , we find @xmath105 then we replace @xmath106 and @xmath107 in the right part of eq .
( 29 ) and ( 30 ) respectively with the intermediate result @xmath108 } } { { y_p } } + \frac{1}{2 } \frac { { ( - \chi - y_p \mu ) } } { { y_p } } ( \lambda - 2\omega ) .
( 36)\ ] ] finally , we substitute the ratio of costate variables @xmath109 found from eq .
( 34 ) in the equation ( 36 ) and obtain the following differential equation for saving per capital @xmath110 found equation together with eq .
( 32 ) is equivalent to the optimal planning problem ( 21 ) .
solution of this equation gives an optimal savings trajectory @xmath98 .
the differential linear equation ( 38 ) is riccati equation which can be solved by standard methods .. for the parameters values @xmath111 , @xmath112 ( corresponding to decreasing significance of consumption in e times for 5 years ) , @xmath113 , and the initial value @xmath114 , the solution of the riccati equation ( 38 ) has been found numerically by the runge - kutta method .
the obtained solution is shown in figure 2 .
the equation ( 38 ) enables analytical consideration in the asymptotic limit when @xmath115 , i.e. when @xmath116 and @xmath117 . in this case
we can neglect those terms in eq .
( 38 ) proportional to @xmath118 with a result @xmath119 where @xmath120 .
note , that this equation is known as equation of interacting masses .
having integrated it we obtain an asymptotic optimal trajectory @xmath121 where the constant @xmath122 is defined by the values @xmath123 and the value of @xmath124 taken as `` initial '' value for the asymptotic regime of evolution @xmath125 as it follows from obtained solution an asymptotic trajectory will be slightly increased if @xmath126 , being restricted by the value of @xmath127 . an asymptotic behavior of state variable @xmath6 is obtained from eq . ( 27 ) by taking into account eq .
( 40 ) @xmath128 where @xmath129 is a rate - fixing constant .
there are two main applications of this analysis : recommendations to the government for investment policy development and prediction of a long - run period macroeconomic system development .
to define recommended investment volume in the current year , we need to obtain the stock of capital statistical data for this year and to find the optimal saving per capital value using the optimal planning problem solution
. then we can evaluate the investment volume as a product of @xmath2 and @xmath4 .
the next year we change the optimal planning problem parameters to improve the accuracy of calculation .
the procedure of recommended investment volume evaluation remains invariable .
thus this method allows the optimization of the investment policy during any period of time .
variables from economic growth models play important role in models of other important fields of economic life too .
a financial programming method combines these models .
the optimal planning problem for the model with effectiveness can serve as a part of financial programming systems [ 7 ] .
the solution of the optimal planning problem may be useful for evaluating financial programming system parameters .
to predict long - run period macroeconomic system development we put optimal saving per capital trajectory into the growth model with effectiveness and then calculate all macroeconomic variables from this model .
table 1 shows predictions based on the republic of belarus economic parameters .
optimal saving per capital trajectory for this model is the result of the optimal planning problem solution from this work . table 1 shows that this variant of macroeconomic strategy allows the extension of productivity potential by increasing the capital per man by 72.4 % and the output per man by 55.5 % .
the total growth of consumption per man is 51.6 % .
the main result of this article is the optimal investment trajectory for belarus for the period of 2000 - 2020 .
the number of additional results may be of use for the investment policy development in other countries and other periods .
the nonlinear system of equations allows to analyze dependencies between trajectories of main macroeconomic variables .
the asymptotic solution shows long - run perspectives of macroeconomic system .
equation from this work may be used in financial programming systems .
a possible improvment of the present model is connected with the splitting of capital in two parts - private industrial capital and public overhead capital . besides
, a more realistic utility function can be chosen for the model .
the results of the inprovements will be published elsewhere .
useful advises of professor komkov are gratefully acknowledged . | the optimal planning trajectrory is analysed on the basis of the growth model with effectiveness .
the saving per capital value has to be rather high initially with smooth decrement in the future years . |
at the end of the last century , there have been performed remarkable experiments @xcite , that demonstrated the possibility to reduce the group velocity of light in a medium by many orders of magnitude .
the idea of `` slow light '' has rapidly become extremely popular , gained a high prestige , and soon turned into a separate direction of quantum optics . in the first experiments on `` slow light '' , the anomalously narrow dispersion feature of the atomic medium was created using the effect of electromagnetically induced transparency @xcite . later on , a different scheme of strong reduction of the group velocity of light has been proposed based on the so - called effect of coherent population oscillations ( cpo ) @xcite .
the classical cpo effect is observed with the use of two monochromatic light beams , namely , the probe and the pump .
the pump beam creates the spectral hole detected by the probe @xcite . in the observations of the cpo - based `` slow light '' ,
the pump beam is not explicitly used , but , nevertheless , the time delay of the light pulse is ascribed to the cpo - based hole - burning effect . in the papers @xcite , it has been shown that all manifestations of the cpo - based `` slow light '' perfectly match the simplest model of saturable absorber ( see , e.g. , @xcite ) , which does not imply spectral hole - burning with appropriate modification of the group velocity of light .
still , the studies on the cpo - based `` slow light '' proceed , with the new interpretation of the pulse delay in a saturable absorber being regarded as self - evident . in our opinion , the possibility of such a revision of interpretation of a well - known effect of classical nonlinear optics is related to insufficiently clear understanding of physical nature of the hole - burning effect under conditions of the coherent population oscillations .
the goal of this paper is to demonstrate , in the most straightforward way , the fact that all manifestations of `` slow light '' in a saturable absorber , as well as the cpo effect are controlled exclusively by frequency dependence of its _ intensity - related _ response .
it is important to emphasize that in terms of the intensity spectra , which allow one to describe this problem in a comprehensive way , neither the susceptibility of the system , nor the optical perturbation exhibit any spectral features at optical frequencies .
moreover , these effects can not be considered , by any means , as specifically optical .
as an example , we present a simple electric circuit with a nonlinear resistor that realizes a full analogue of the saturable absorber in the field of the light wave including the effect of hole- burning in the absorption spectrum ( converted into the effect of hole - burning in the frequency dependence of of the resistor s resistance ) and the effect of `` slow light '' ( converted into the effect of `` slow current '' ) .
the proposed analogy makes it possible , in our opinion , to better understand physical content of the effects under consideration . in this paper
, we will talk only about `` slow '' light , keeping in mind that all the results are equally valid for the so - called `` fast '' light ( the light with a superluminal or even `` negative '' group velocity ) upon sign inversion of the intensity - related nonlinearity of the system ( i.e. , upon replacing the bleachable absorber by the inverse one ) .
by the saturable absorber is usually meant a layer of an optical medium whose transmission @xmath0 varies with intensity @xmath1 of the light , following , as a rule with some delay , its variations .
as follows from the name of the saturable absorber , its absorption exhibits saturation with growing light intensity , though here this fact will be of no importance . in the most general form , the relation between intensities of the incident ( @xmath1 ) and transmitted ( @xmath2 ) light can be expressed by the formula @xmath3 and the dynamics of establishment of the equilibrium transmission is given by the conventional relationship @xmath4 where @xmath5 is the relaxation time of the absorber . to describe the effects under study
, it suffices to restrict oneself to the case of small variations of the light intensity @xmath6 in the vicinity of @xmath7 and to consider a `` linear '' saturable absorber , with its stationary transmissivity @xmath8 linearly varying with @xmath6 : @xmath9 in the framework of these simple assumptions , one can easily find the amplitude and phase characteristics of the intensity - related response of the saturable absorber ( fig .
1 ) . these curves represent frequency dependences of the amplitude @xmath10 and phase @xmath11 of the light intensity modulation at the exit of the saturable absorber under condition of harmonic modulation of the light intensity at the entrance .
as seen from the figure , the spectrum of such an `` intensity susceptibility '' displays a single feature lying in the range of low frequencies comparable with the inverse relaxation time @xmath5 ( for more detail see @xcite ) .
we want to attract special attention to the fact that this approach uses only the notion of the light intensity ( the field amplitude squared ) , whose spectrum has no spectral peculiarities at optical frequencies . for this reason , this solution does not contain explicitly either optical characteristics of the light beam ( spectral composition , degree of monochromaticity ) , or spectral properties of the absorber at optical frequencies ( structure of the absorption spectrum , the nature of the band broadening , etc . ) . at the same time
, this model describes comprehensively the cpo effect , as well as all manifestations of `` slow light '' .
indeed , the phase and amplitude characteristics of transmission for the intensity- modulated light passing through a saturable absorber studied in the papers on the hole - burning in a homogeneously broadened band and on `` slow light '' @xcite , precisely agree with the above model and do not require to invoke the effect of hole - burning in the optical spectrum of the absorber .
the same can be said about the light pulse delay in the saturable absorber
the effect that is regarded as the main evidence for the group velocity reduction and for the `` slow light '' effect . since in the range of low modulation frequencies
@xmath12 , the phase delay @xmath13 linearly varies with frequency , the time delay @xmath14 ceases to be frequency - dependent . as a result , the light pulse , whose intensity spectrum lies within the interval @xmath15 , displays practically pure temporal shift with no reshaping . in the conventional cpo effect
, the weak modulation of the light passing through the saturable absorber is created using two monochromatic waves with close frequencies , one of them strong ( pump wave ) and the other weak ( probe wave ) .
this intensity - modulated light in this case , evidently , exhibits all the amplitude and phase changes mentioned above . but now , one detects variations in the optical spectrum ( rather than in the intensity spectrum ) of the transmitted light .
this possibility is provided by the fine spectral structure of the light .
now , it becomes important that for the sufficiently low modulation frequencies ( for the probe light frequency being close enough to that of the pump ) , the transmissivity of the medium is modulated at the difference frequency , and the light passing through such a modulator acquires additional spectral components , with the frequency of one of them being exactly coincident with that of the probe wave . as a result , the probe beam exhibits , in the vicinity of the pump wave frequency , an effective dip of the absorption spectrum ( the appropriate spectral feature is also displayed by the phase behavior of the probe wave ) .
it is this dip that is considered to be responsible for all manifestations of `` slow light '' in a saturable absorber .
in fact , we may conclude that physical content of the effects of cpo and `` slow light '' is confined to transformation of the intensity spectrum of the light transmitted through the saturable absorber in accordance with the transfer function shown in fig
it should be noted that observation of the cpo effect , generally , does not necessarily imply real population oscillations , because the mechanism of the nonlinear dependence ( 1 ) may be not related to the light - induced changes in population of the initial state of the optical transition .
such a dependence may result from , say , a temperature shift of the optical spectrum that changes the overlap of the spectrum of the light beam with the absorption spectrum of the medium with increasing pump power , provided that the medium is heated due to nonradiative relaxation of the excitation . in other words , the saturable absorber , under conditions of the modulated pump , is capable of demonstrating coherent oscillations of losses , rather than populations .
moreover , this effect is , by no means , specifically optical and can be observed on nonlinear systems of other physical nature . to get a clearer idea about the so - called effect of `` hole - burning '' in the optical spectrum of a saturable absorber and about attendant effects of `` slow light '' ,
let us turn to a simple nonlinear system that has nothing to do with optics but realizes an isomorphic model of the saturable absorber .
consider the electric circuit shown in fig .
2 . the key element of the circuit ( which actually simulates the saturable absorber ) is the resistor @xmath16 , whose resistance varies with temperature which , in turn , follows ( with some temporal delay ) the power released on the resistor . by the input signal
, we will mean the voltage @xmath17 applied to the circuit and by the output , the voltage @xmath18 falling on the resistor @xmath19 ( proportional to the current in the circuit ) : @xmath20 here , the transfer coefficient @xmath0 is the counterpart of transmissivity of the saturable absorber ( see eq .
the dynamics of this coefficient is controlled by the equation @xmath4 here , @xmath5 is the relaxation time and @xmath8 is the equilibrium value of the transfer coefficient .
we will assume that for small variations of the input voltage @xmath21 in the vicinity of some value @xmath22 , the equilibrium value of the coefficient @xmath8 varies linearly with @xmath21 : @xmath23 one can easily see that the relation between the input and output voltage , in this case , is completely equivalent to that between intensities of the input and output light in the saturable absorber ( see eqs .
( 13 ) ) , and the quadripole described above realizes analog of the saturable absorber in the field of the light wave .
indeed , the amplitude and phase transformations of the input voltage modulation , in this circuit , are controlled by the same transfer functions ( fig .
1 ) , whereas a sufficiently long pulse ( in the scale of inverse relaxation times ) of the input voltage modulation will exhibit a time delay , thus demonstrating the effect of `` slow current '' ( if the resistance falls with increasing temperature , then the pulse of the current is delayed with respect to the voltage ) . note that the fullness of the analogy is not hindered by the absence of the high - frequency carrier , which excludes any possible role of the `` hole - burning '' effect .
as one can easily see , however , no essential changes occur if the dc modulated voltage is replaced by a ac modulated voltage by filling the envelope with a high - frequency carrier . in this case
, all specific features of the low - frequency response of the quadripole to the modulation of the effective voltage will remain the same ( including the effect of `` slow current '' ) , but simultaneously there will appear the possibility to analyze transformation of the spectra of the high - frequency carrier by `` slow '' nonlinearity of the system . using this circuit , one can easily make the experiment similar to that described in classical papers on observation of the hole - burning effect under conditions of the coherent population oscillations . for this purpose , one should apply to the input of the circuit two high - frequency signals with close frequencies ( the `` pump '' voltage and the `` probe '' voltage ) , and detect at the exit only the weak `` probe '' voltage . in this kind of experiment
, one will observe a narrow peak of current in the range of the `` probe '' frequencies close to that of the `` pump '' .
this peak can be interpreted as the effect of `` hole - burning '' in the frequency dependence of resistance of the nonlinear resistor @xmath16 .
figure 2 ( a d ) schematically shows characteristic types of transformation of a modulated signal by the above electric circuit .
in this paper , we have made an attempt to vividly demonstrate the fact that the effect of coherent population oscillations , as well as all manifestations of `` slow light '' result exceptionally from the `` resonance '' of the intensity - related susceptibility of the absorber at zero frequency .
this spectral feature of the intensity susceptibility is really capable of modifying spectral composition of the light passing through the absorber ( to the extent of modification of its intensity spectrum ) , which is revealed in the effect of `` hole - burning '' in the homogeneously broadened absorption spectrum .
however , the question about possibility to treat this hole as a usual spectral feature of a linear absorption spectrum and , in particular , to apply to it the kramers - kronig relations needs special consideration .
specifically , as we already pointed out in @xcite , the possibility to detect the `` hole '' using standard absorption spectroscopy may depend on the internal phase structure of the light beam . in any case , we are not aware of experiments in which the effect of the pulse delay was observed in the presence of the pump beam . as for the `` slow light '' effect observed with a single light pulse , which
, according to @xcite , combines the functions of the pump and probe beams , we see no grounds , in this case , to invoke the effect of coherent population oscillations for explanation of the pulse delay . we believe that the physical nature of the effects considered above is convincingly demonstrated by the proposed analogy between the saturable absorber and the nonlinear resistor in the electric circuit .
this circuit simulates the effects of `` slow '' and `` fast '' light ( depending on the sign of the temperature nonlinearity of the resistor ) , as well as the above effect of `` hole - burning '' .
this model simultaneously shows that the light pulse delay is not connected with effects of propagation and that it is inappropriate to invoke the notion of group velocity to interpret the effect .
a. kasapi , maneesh jain , g. y. yin , and s. e. harris , phys .
lett . , 74 , 2447 ( 1995 ) .
l. v. hau , s. e. harris , z. dutton , and c. h. behroozi , nature , 397 , 594 ( 1999 ) .
s. e. harris , phys . today
, 50 , 36 ( 1997 ) .
bigelow m.s . ,
lepeshkin n.n .
, boyd r.w .
2003 . v. 90 .
p. 113903 .
bigelow m.s . ,
lepeshkin n.n .
, boyd r.w .
// science .
schwarz s.e .
phys . lett .
v.s.zapasskii , g.g.kozlov , //
optics and spectroscopy , v.100 , no.3 , pp 419 - 424,(2006 ) .
e.b.aleksandrov , v.s.zapasskii , physics
usppekhi 49 ( 10 ) 1067 1075 ( 2006 ) .
selden a.c .
j. appl . phys .
selden a c , physics/0512149 baldit e et al .
95 143601 ( 2005 ) ; ccsd-00004377 ( 2005 )
zhang yun - dong et al .
chinese phys .
21 87 ( 2004 ) | it is shown that the effect of hole - burning under conditions of coherent population oscillations as well as the light pulse delay in a saturable absorber ( a modification of the `` slow light '' effect ) can be interpreted , in a comprehensive way , in terms of _ intensity spectrum _ of the light and _ intensity - related susceptibility _ of the non - linear medium .
the physical content of these effects is illustrated by a simple electric circuit with a non - linear resistor which realizes a full analog of the saturable absorber . in this case
the effect of hole - burning in the absorption spectrum of the medium is converted in to the effect of hole - burning in the frequency dependence of resistance of the non - linear resistor and the effect of `` slow light '' in to the effect of `` slow current '' .
institute of physics , st .- petersburg state university , st .- petersburg , 198504 russia .
-mail : [email protected] |
nonthermal radiation , usually having a power - law emission spectrum , have been observed from astrophysical systems containing relativistic jets , e.g. , active galactic nuclei ( agns ) , gamma - ray bursts ( grbs ) , galactic microquasar systems , and crab - like supernova remnants ( snrs ) . in most of these systems ,
the emission is thought to be generated by accelerated electrons through the synchrotron or inverse compton mechanisms .
radiation is observed from these systems from the radio through the gamma - ray region .
the most widely known mechanism for the acceleration of particles in astrophysical environments characterized by a power - law spectrum is fermi acceleration .
this mechanism for particle acceleration relies on the shock jump conditions at relativistic shocks ( e.g. , gallant 2002 ) .
most astrophysical shocks are collisionless since dissipation is dominated by wave - particle interactions rather than particle - particle collisions .
diffusive shock acceleration ( dsa ) relies on repeated scattering of charged particles by magnetic irregularities ( alfvn waves ) to confine the particles near the shocks . however , particle acceleration near relativistic shocks is not due to dsa because the propagation of accelerated particles ahead of the shock can not be described as spatial diffusion .
anisotropies in the angular distribution of the accelerated particles are large , and the diffusion approximation for spatial transport does not apply ( achterberg et al .
particle - in - cell ( pic ) simulations may shed light on the physical mechanism of particle acceleration that involves the complicated dynamics of particles in relativistic shocks .
recent pic simulations using counter - streaming relativistic jets show that acceleration is provided in situ in the downstream jet , rather than by the scattering of particles back and forth across the shock as in fermi acceleration ( frederiksen et al .
two independent simulation studies have confirmed that the counter - streaming jets excite the weibel instability ( weibel 1959 ) , which generates magnetic fields ( medvedev & loeb 1999 ; brainerd 2000 ; pruet et al .
2001 ; gruzinov 2001 ) , and consequently accelerates electrons ( silva et al .
2003 ; frederiksen et al .
2003 ) . since these simulations are performed with counter - streaming jets , shock dynamics involving the jet head , where fermi acceleration may take place , has not been investigated .
in this paper we present new simulation results of particle acceleration and magnetic field generation in relativistic jets using 3-d relativistic electromagnetic particle - in - cell ( remp ) simulations with and without initial ambient magnetic fields . in our simulations , an electron - ion relativistic jet with lorentz factor
, @xmath0 ( corresponds to 5 mev ) is injected into an electron - ion plasma in order to study the dynamics of a relativistic collisionless shock .
we illustrate the features of the collisionless shock generated at the head of the relativistic jet injected into magnetized and unmagnetized ambient plasma .
the weibel instability is excited in the downstream region and accelerates electrons and ions . in section 2 the simulation model and initial conditions are described .
the simulation results are presented in the section 3 , and in section 4 we summarize and discuss the results .
the code used in this study is a modified version of the tristan code , which is a relativistic electromagnetic particle ( remp ) code ( buneman 1993 ) .
descriptions of pic codes are presented in dawson ( 1983 ) , birdsall & langdon ( 1995 ) , and hickory & eastwood ( 1988 ) .
this code has been used for many applications including astrophysical plasmas ( zhao et al .
1994 ; nishikawa et al .
1997 ) .
the simulations were performed using @xmath1 grids with a total of 55 to 85 million particles ( 27 particles@xmath2cell@xmath2species for the ambient plasma ) . both periodic and radiating boundary conditions
are used ( buneman 1993 )
. the ambient electron and ion plasma has mass ratio @xmath3 .
the electron thermal velocity @xmath4 is @xmath5 , where @xmath6 is the speed of light .
the electron skin depth , @xmath7 , is @xmath8 , where @xmath9 is the electron plasma frequency ( @xmath10 is the grid size ) . in this study
three different cases are simulated to investigate the fundamental characteristics of the weibel instability in electron - ion plasmas .
one simulation is performed injecting a thin jet into a magnetized ambient plasma .
the other two simulations consider flat jets ( infinite width ) injected into a magnetized ambient plasma and then into an unmagnetized ambient plasma in order to compare this simulation with previous simulations ( silva et al .
2003 ; frederiksen et al .
two kinds of jets have been simulated : a `` thin '' jet with radius @xmath11 and a flat " ( thick ) jet that fills the computational domain in the transverse directions ( infinite width ) . in the simulations ,
jets are injected at @xmath12 .
first , we describe the simulation results for the thin jet with initial ambient magnetic fields .
the number density of the thin jet is @xmath13 , where @xmath14 is the density of ambient ( background ) electrons .
the average jet velocity @xmath15 , and the lorentz factor is 5 .
the time step @xmath16 , the ratio @xmath17 , the alfvn speed @xmath18 , and the alfvn mach number @xmath19 .
the gyroradii of ambient electrons and ions are @xmath20 , and @xmath21 , respectively . as shown in fig .
1 the jet electrons are twisted due to a two - stream - type instability .
since the size of jet is of the order of the electron skin depth , it is not clear if the weibel instability is excited .
as will be shown in the flat jet simulation , the density perturbation is similar to that induced by the weibel instability .
since the size of the jet is extremely small compared to realistic jets , further study of thin jets will be performed including simulations with larger jet sizes .
the jet density of the flat jet is also nearly @xmath13 .
other parameters are the same as in the case of the thin jet . in this case
, the jet makes contact with the ambient plasma at a 2d interface spanning the computational domain .
therefore , only the dynamics of the jet head and the propagation of a shock in the downstream region is studied .
this simulation system is similar to simulations performed previously using effectively infinite counter - streaming jets .
the important difference between this simulation and previous simulations is that the evolution of the weibel instability is examined in a more realistic spatial way including motion of the jet head .
furthermore , the density of the jet relative to the ambient plasma can be changed easily .
the weibel instability is excited and the electron density is perturbed as shown in fig .
the electrons are deflected by the perturbed ( small ) transverse magnetic fields ( @xmath22 ) via the lorentz force : @xmath23 , generating filamented current perturbations ( @xmath24 ) , which enhance the transverse magnetic fields ( weibel 1959 ; medvedev and loeb 1999 ) .
the complicated current structures due to the weibel instability are shown in fig .
the sizes of these structures are nearly the electron skin depth ( @xmath8 ) .
this is in good agreement with @xmath25 ( medvedev & loeb 1999 ) . here
, @xmath26 is a thermal lorentz factor , @xmath27 is the electron plasma frequency .
the shapes are elongated along the direction of the jet ( the @xmath28-direction , horizontal in fig .
the acceleration of electrons has been reported in previous work ( silva et al . 2003 ; frederiksen et al .
2003 ) . as shown in fig .
3 , the kinetic energy ( parallel velocity @xmath29 ) of the jet electrons is transferred to the perpendicular velocity via the electric and magnetic fields generated by the weibel instability . the strongest deceleration of electron flow ( fig .
3b ) and strongest transverse acceleration ( fig .
3a ) is between @xmath30 and takes place around the maximum amplitude of perturbations due to the weibel instability at @xmath31 ( see figs . 2 and 4 ) .
on the other hand , jet ions are slightly accelerated ( not shown ) .
however , due to the larger mass of the ions , the energy gained by ion acceleration is similar to that gained by the electrons .
furthermore , at the jet front some jet electrons are accelerated and some are decelerated .
this acceleration and deceleration is shown by the slanting in the parallel velocity distribution at the jet head ( fig .
3b ) . this may indicate that fermi acceleration is taking place at the jet front as described in previous work ( e.g. , achterberg et al .
2001 ; gallant 2002 ; ellison & double 2002 ) .
the weibel instability is excited just behind the jet head as shown in fig .
the three curves are measured at three different locations ( @xmath32 , and 48 ) about the electron skin depth apart . as a result
the phase of instability is different , but the amplitudes are similar .
the growth rate of the weibel instability is calculated to be , @xmath33 ( @xmath34 ) ( medvedev & loeb 1999 ) .
this is in good agreement with the simulation results with the jet head located at @xmath35 .
figure 3 suggests that the shock " has a thickness from about @xmath36 behind the shock front .
possibly , the turbulence " assumed for the diffusive shock acceleration ( dsa ) corresponds to this shock region .
the transverse acceleration of jet electrons takes place in the shock , as shown in fig . 3a .
subtle structures are recognized at the jet head ( @xmath35 ) . as shown in fig .
4a , the width of the jet head is nearly the electron skin depth ( @xmath8 ) .
additionally , figure 4b shows that the @xmath28-component of current density becomes slightly negative .
this results from the fact that some jet electrons are ahead of the ions .
this negative current at @xmath37 is clearly shown in fig . 5b .
based on figs . 2 and 4
the size of perturbations along the jet around @xmath38 is nearly twice the electron skin depth .
this result is consistent with the previous simulations by silva et al .
( 2003 ) .
the weibel instability is excited just behind the jet head located at @xmath35 ( see fig .
5c ( @xmath39 ) ) , and the accelerated jet electrons are creating the negative current at @xmath37 ( fig .
5b ) just in front of the jet head .
this total negative current is indicated by three curves dipped uniformly at @xmath37 in fig .
the weibel instability creates elongated shell - type structures which are also shown in counter - streaming jet simulations ( silva et al .
2003 ; frederiksen et al .
the size of these structures transverse to the jet propagation is nearly the electron skin depth ( @xmath8 ) as shown in figs . 5 and 6 .
it should be noted that the size of perturbations at @xmath38 is larger than those at different locations , since the smaller scale perturbations merge to larger sizes in the nonlinear stage at the maximum amplitudes ( silva et al .
2003 ) . in the previous simulations of silva
et al . ( 2003 ) and frederiksen et al .
( 2003 ) ambient magnetic fields are not included initially . in order to investigate the generation of magnetic fields ,
a simulation has been performed without ambient magnetic fields but otherwise with the same parameters .
the structure of perturbations to the electron density and @xmath28-component of the current density are similar to those with initial ambient magnetic fields ( @xmath40 ) , as shown in fig .
the unmagnetized jet generates magnetic fields due to the current structures produced in the weibel instability , as shown in figs .
7b and 8b .
the peak values of the perturbations due to the weibel instability for an unmagnetized ambient plasma is larger that those for a magnetized ambient plasma .
however , the amplitude of these perturbations is similar ; therefore the effects of initial ambient magnetic fields are not apparent in these cases .
further investigation is required for a systematic analysis of the effects of magnetic fields .
the generation of magnetic fields both with and without an initial magnetic field suggests that synchrotron emission ( or jitter radiation ) is relevant in grb afterglows and crab - like pulsar winds ( medvedev 2000 ) .
figure 9 shows the structures of perturbations along the @xmath28-direction in electron density , @xmath24 , @xmath41 , and @xmath42 , without an initial magnetic field .
these are very similar to those with an initial magnetic field , shown in fig .
the three curves are measured at three different locations ( @xmath43 , and 48 ) each separated about one electron skin depth .
the phase of the instability is different along each line , but the amplitudes are similar .
furthermore , the different seed perturbations ( thermal noise by the initial loading of particles ) and the unmagnetized plasma make the amplitudes and shape of perturbations at the nonlinear phase different from those with the magnetized plasma ( as shown in fig .
4 ) . from the previous simulations without an initial ambient magnetic field ( silva et al . 2003 ; frederiksen et al .
2003 ) , highly nonuniform , small - scale magnetic fields are generated due to the weibel instability .
we have performed the first self - consistent , three - dimensional relativistic particle simulations of electron - ion relativistic jets propagating through magnetized and unmagnetized electron - ion ambient plasmas .
the weibel instability is excited in the downstream region behind the jet head , where electron density perturbations and filamented currents are generated .
the nonuniform electric field and magnetic field structures slightly decelerate the jet electrons and ions , while accelerating ( heating ) the jet electrons and ions in the transverse direction , in addition to accelerating the ambient material .
the origin of the weibel instability relies on the fact that the electrons are deflected by the perturbed ( small ) transverse magnetic fields ( @xmath22 ) , and subsequently enhance the filamented current ( weibel 1959 ; medvedev and loeb 1999 ; brainerd 2000 ; gruzinov 2001 ) .
the deflection of particle orbits due to the lorentz force increases as the magnetic field perturbation grows in amplitude .
the amplified magnetic field is random in the direction perpendicular to the particle motion , since it is generated from a random seed field . the perturbed electron density and filamented currents have a complicated three - dimensional structure .
the transverse size of these structures is nearly equal to the electron skin depth but is larger if there are no ambient magnetic fields .
however , the size along the direction of jets is larger than the transverse scales . at the termination of our simulation , the thickness of the unstable region along the jet direction ranges from @xmath44 to @xmath45 .
the perturbation size in the transverse direction become largest around @xmath46 , where nonlinear effects lead to the merging of the smaller scale filaments that first appear behind the jet front .
this result is similar to previous counter - streaming simulation results ( silva et al .
2003 ) in which smaller scale filaments first appear and then merge into larger scale filaments at a later time .
now we see the temporal development appear as a spatial development .
the simulation results show that the initial jet kinetic energy goes to the magnetic fields and transverse acceleration of the jet particles through the weibel instability .
the properties of the synchrotron or
jitter " emission from relativistic shocks are determined by the magnetic field strength , @xmath47 and the electron energy distribution behind the shock .
the following dimensionless parameters are used to estimate these values ; @xmath48 and @xmath49 ( medvedev & loeb 1999 ) . here
@xmath50 , @xmath51 are the magnetic and electron energy densities , and @xmath52 is the total thermal energy density behind the shock , where @xmath53 is the ion mass , @xmath54 is the ion number density , and @xmath26 is the mean thermal lorenz factor of ions . based on the available diagnostics the following values are estimated ; @xmath55 and @xmath56 ( typical values for grb afterglows ) .
these estimates are made at the maximum amplitude ( @xmath57 ) .
the instability is aperiodic , i.e. , @xmath58 = 0 ( convective ) .
thus , it can be saturated only by nonlinear effects and not by kinetic effects , such as collisionless damping or resonance broadening .
hence the magnetic field can be amplified to very high values .
the produced magnetic field is randomly oriented in the direction perpendicular to the flow .
the lorentz forces randomize particle motion over the pitch angle , and hence introduce an effective scattering into the otherwise collisionless system .
1. the smaller scale filaments appear immediately behind the jet front .
the nonlinear effect is recognized by the fact that smaller filaments merge into larger filaments behind the jet front , as shown in figs . 5 and 6 .
2. the instability is self - saturating ( spatially ) behind the jet front and continues until all of the free energy due to the particle distribution function anisotropy is transferred to the magnetic field energy .
the largest amplitude of perturbations is located near the region with the maximum transverse acceleration ( fig .
3a ) with deceleration in parallel velocity ( @xmath59 ) ( fig .
3b ) , as shown in figs . 2 , 5d , and 6d .
3. the produced magnetic field is randomly oriented in the shock " plane ( three - dimensional structure , as shown in figs .
2b and 5 ) .
the random lorentz forces result in randomization and deceleration of the jet particles , an acceleration of the ambient particles , and transverse acceleration ( heating ) of jet and ambient particles inside the `` shock '' region .
previous microphysical analyses of the energy conversion in relativistic pair outflows interacting with the surrounding interstellar medium consisting of cold protons and electrons have been performed ( e.g. , brainerd 2000 ; schlickeiser et al .
these analyses have demonstrated that the beam excites both electrostatic and low - frequency magnetohydrodynamic alfvn - type waves via a two - stream instability in the background plasma .
they have also provided the time evolution of the distribution functions of beam particles and the generated plasma wave turbulence power spectra .
while the jet front showed evidence of fermi acceleration , the main acceleration of electrons may take place in the downstream region ( e.g. , brainerd 2000 ; schlickeiser et al .
2002 ; ostrowski & bednarz 2002 ) .
our present simulation study has provided the framework of the fundamental dynamics of a relativistic shock generated within a relativistic jet . while some fermi acceleration may occur at the jet front
, the majority of electron acceleration takes place behind the jet front and can not be characterized as fermi acceleration . since the shock dynamics is complex and subtle , further comprehensive study is required for better understanding of the acceleration of electrons and the associated emission as compared with current theory ( e.g. , rossi & rees 2002 ) .
this further study will provide more insight into basic relativistic collisionless shock characteristics .
the fundamental characteristics of such shocks are essential for a proper understanding of the prompt gamma - ray and afterglow emission in gamma - ray bursts , and also to an understanding of the particle reacceleration processes and emission from the shocked regions in relativistic agn jets .
is a nrc senior research fellow at nasa marshall space flight center .
this research ( k.n . ) is partially supported by nsf atm 9730230 , atm-9870072 , atm-0100997 , and int-9981508 .
the simulations have been performed on origin 2000 and ibm p690 ( copper ) at ncsa which is supported by nsf .
achterberg , a. , gallant , y. a. , kirk , j. g. , & guthmann , a. x. 2001 , mnras , 328 , 393 birdsall , c. k. , & langdon , a. b. , 1995 , plasma physics via computer simulation , mcgrow - hill , second edition brainerd , j. j. 2000 , , 538 , 628 buneman , o. , 1993 , tristan , in computer space plasma physics : simulation techniques and software , edited by h. matsumoto matsumoto & y. omura , p. 67
, terra scientific publishing company , tokyo dawson , j. m. 1983 , rev .
, 55 , 403 ellison , d. c. & double , g. p. 2002
, astroparticle phys .
18 , 213 frederiksen , j. t. , hededal , c. b. , haugblle , & nordlund , .
2003 , proc . from 1st nbsi on beams and jets in gamma ray bursts , held at nbifafg / nordita , copenhagen , denmark , august , 2002 , astro - ph/0303360 gallant , y. a. , 2002 , particle acceleration at relativistic shocks , in relativistic flows in astrophysics , eds .
w. guthmann , m. georganopoulos , a. marcowith , & k. manolokou , lecture notes in physics , springer verlag .
astro - ph/0201243 gruzinov , a. 2001 , apj , 563 , l15 hickory & eastwood , j. w. , 1988 , computer simulation using particles , mcgraw - hill , second edition medvedev , m. v. 2000 , apj , 540 , 704 medvedev , m. v. & loeb , a. 1999 , apj , 526 , 697 nishikawa , k .-
i . , zhao , j. , sakai , j. i. , & neubert , t. 1997 , adv .
19 , ( 1)117 ostrowski , m. , & bednarz , j. 2002 , a&a , 394 , 1141 pruet , j. , abazajian , k. , fuller , g. m. 2001 , phys .
d. 64 , 063002 - 1 rossi , e. & rees , m. j. mnras , 2002 ( astro - ph/02044406 ) .
schlickeiser , r. , vainio , r. , bttcher , m. , lerche , i. , pohl , m. , & schuster , c. 2002 , a&a , 393 , 69 silva , l. o. , fonseca , r. a. , tonge , j,.w . ,
dawson , j. m. , mori , w.b . , & medvedev , m. v. , 2003 , submitted to apj weibel , e. s. 1959 , phys .
lett . , 2 , 83 zhao , j.,sakai , j. i. , nishikawa , k .-
i . , & neubert , t. 1994 , phys .
plasmas , 1 , 4114 | shock acceleration is an ubiquitous phenomenon in astrophysical plasmas .
plasma waves and their associated instabilities ( e.g. , the buneman instability , two - streaming instability , and the weibel instability ) created in the shocks are responsible for particle ( electron , positron , and ion ) acceleration . using a 3-d relativistic electromagnetic particle ( remp ) code ,
we have investigated particle acceleration associated with a relativistic jet front propagating through an ambient plasma with and without initial magnetic fields .
we find only small differences in the results between no ambient and weak ambient magnetic fields .
simulations show that the weibel instability created in the collisionless shock front accelerates particles perpendicular and parallel to the jet propagation direction .
while some fermi acceleration may occur at the jet front , the majority of electron acceleration takes place behind the jet front and can not be characterized as fermi acceleration .
the simulation results show that this instability is responsible for generating and amplifying highly nonuniform , small - scale magnetic fields , which contribute to the electron s transverse deflection behind the jet head .
the `` jitter '' radiation ( medvedev 2000 ) from deflected electrons has different properties than synchrotron radiation which is calculated in a uniform magnetic field .
this jitter radiation may be important to understanding the complex time evolution and/or spectral structure in gamma - ray bursts , relativistic jets , and supernova remnants . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.