article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
we consider the molecule ( introduced as a set of energy levels ) placed in between two electrodes ( left @xmath5 and right @xmath6 ) and plays a role of channel .
the electrodes are behaved as free electron reservoirs with approximately continuous energy spectra .
the electronic transport properties of molecular junctions are govern by quantum mechanical laws .
one of most important framework for studying theoretical nanoelectronics is a landauer frameworks @xcite .
landauer approach is based on the description of electron transport through elastic scattering model .
the thermalized electrons in reservoirs will be scattered when they come into the channel , but their transport are completely coherent between the electrodes .
one can interpret the conductance of channel as an elastic scatterer , by the quantum mechanical probabilities of transmission @xmath7 , that corresponds to electrons with energy @xmath8 .
because of the importance of transmission function to calculate all transport properties of nanoscale junctions , we will consider this function for our two - level ( toy ) model in following .
let us consider a two - terminal system with @xmath5 and @xmath6 electrodes linked by two - level molecular junction model as fig.[fig2 ] .
for this model , the @xmath9 matrix for an inverse green function defines as @xmath10 where the elements of @xmath9 self energy matrices define as follows @xmath11 in which @xmath12 denote the hybridization matrix elements that connect the two - level , up and down , with the electrodes ; @xmath13 ia a complex valued surface green s function of the uncoupled leads , i.e. , the left and right semi - infinite leads .
the coupling matrix @xmath14 also known as the broadening functions , is related to the self - energies through @xmath15 that from eq . ( 2 ) , one can find @xmath16 so the eq . ( 3 ) can be written as @xmath17 the imaginary part of surface green s functions is related to the contact density of states as @xmath18 so we obtain @xmath19 one can calculate the transmission from the green s function method , using the relation @xmath20 by using of relation ( 7 ) , the above equation converts to @xmath21 since the trace is invariant under cyclic permutations , i.e. , @xmath22 , so we can rewrite the above equation as @xmath23 because @xmath24 is @xmath25 matrix , so the multiplied @xmath26 is a @xmath27 matrix and equation ( 10 ) simplifies as @xmath28 after some calculations and manipulations that are presented in appendices @xmath29 and @xmath30 , the above transmission function can be written as @xmath31 where for a system with energies @xmath32s and level broadening @xmath33s , @xmath34s have a lorentzian definition @xmath35 and interference enters via @xmath36 where constants @xmath37s have following definitions @xmath38 } \\ c_2 & = & 2\gamma_2\frac{\varepsilon_1\varepsilon_2+\gamma_1\gamma_2+\gamma_1 ^ 2-\varepsilon_1 ^ 2}{\gamma_1[(\varepsilon_1-\varepsilon_2))^2+(\gamma_1+\gamma_2)^2]}\end{aligned}\ ] ] and constants @xmath39s will obtain by applying @xmath40 and also @xmath41 exchanges in corresponding @xmath37s . electrode ( black and solid line ) , the right @xmath42 electrode ( red and dashed line ) and the total power dissipated @xmath43 ( blue ans solid line ) as a function of bias voltage . ( a ) upper panel for case-1 where energies levels are chosen fron homo ( @xmath44 and lumo ( @xmath45 ) , ( b ) middel panel for case-2 which both energies are chosen from homos ( @xmath46 , @xmath47 ) and ( c ) bottom panel for case-3 both where energies are chosen from lumos ( @xmath48 , @xmath49 ) .
, title="fig : " ] electrode ( black and solid line ) , the right @xmath42 electrode ( red and dashed line ) and the total power dissipated @xmath43 ( blue ans solid line ) as a function of bias voltage . ( a ) upper panel for case-1 where energies levels are chosen fron homo ( @xmath44 and lumo ( @xmath45 ) , ( b ) middel panel for case-2 which both energies are chosen from homos ( @xmath46 , @xmath47 ) and ( c ) bottom panel for case-3 both where energies are chosen from lumos ( @xmath48 , @xmath49 ) .
, title="fig : " ] electrode ( black and solid line ) , the right @xmath42 electrode ( red and dashed line ) and the total power dissipated @xmath43 ( blue ans solid line ) as a function of bias voltage . ( a ) upper panel for case-1 where energies levels are chosen fron homo ( @xmath44 and lumo ( @xmath45 ) , ( b ) middel panel for case-2 which both energies are chosen from homos ( @xmath46 , @xmath47 ) and ( c ) bottom panel for case-3 both where energies are chosen from lumos ( @xmath48 , @xmath49 ) .
, title="fig : " ] let us suppose that a voltage bias @xmath50 is applied across the system .
it causes that the electrochemical potential of the left and right electrodes shifts such that @xmath51 , where @xmath52 is the electron charge .
we call positive ( negative ) bias for plus ( minus ) sign . on the other words , for positive ( negative ) bias electrons flow from the left ( right ) electrode to the right ( left )
the fact that elastic scattering is not associated with any energy loss in the junction regions implicitly reminds that when an electron of energy @xmath8 tunnels from the left to the right ( in positive bias ) releases its excess energy @xmath53 in the right lead , while the hole , that is created behind the electron , is filled up releasing an energy equal to @xmath54 in the left lead .
more precisely , these released energies dissipate as a heat in two leads . according to the landauer approach
, the electronic contribution to charge current ( @xmath55 ) and energy current ( @xmath56 ) through the junction can be expressed in terms of the transmission function for positive bias @xmath57 , as in the right electrode .
for those electron transmitted via the lumos orbital more power is dissipated than the homos in the right electrode ( see the upper panel ) .
transmitted electrons leave holes in the left electrode , so the hole is filled up by dissipating an energy equal to @xmath54 in the left electrode . for the hole left behind by electron transmitted via the homos orbital more power
is dissipated than the lumos in the left electrode ( see the bottom panel ) . ]
@xmath58 where @xmath59 and @xmath60 is the fermi function of the left ( right ) electrode .
each fermi function depends on the electrode s chemical potential , which in turn is related to the applied bias .
the rate of heat released in the left ( right ) electrode with electrochemical potential @xmath61 is given by @xmath62 using eq .
( 17 ) and ( 18 ) we obtain the rate of heat dissipated in left and right electrodes as @xmath63 thus total power ( heat per unit of time ) dissipated in the system is @xmath64 three limits allow simplifications of the above relation .
first , is the limit that two electrodes to be of the same materials , @xmath65 , and the system is in equilibrium at zero bias and without loss of generality we set @xmath66 .
moreover , we assume that the electrochemical potentials are shifted symmetrically with the bias voltage , i.e. @xmath67 and @xmath68 .
second , is the limit of zero temperature , @xmath69 , where @xmath70 is the heaviside ( step ) function .
the use of plus ( minus ) signs here arises from the positive ( negative ) bias .
third , is the low bias limit , where we can neglect the transmission function s dependence on the bias , @xmath71 .
applying these limits , we arrive @xmath72 if one concentrates on the bias polarity , it is easy to show that @xmath73 the above relation indicates that the power dissipated in one of the electrodes can be obtained from the power dissipated in the other one by simply inverting the bias .
if one checks the behaviour of total heat dissipation with respect to the inversion of the bias voltage , one obtains @xmath74 thus , the total heat dissipation is a symmetric function with respect to the inversion of the bias voltage .
also , for total heat we have @xmath75 this relation implies that the total power dissipated in the model is equal to the sum of the power dissipated for positive and negative bias in a given electrode . by inserting transmission function ( 12 ) in eq.(24 ) and taking the integral , we obtain the explicit expression for the total power dissipation in two - level model as @xmath76 + \frac{ev\gamma_1}{h}c_1\ln\left[\frac{\gamma_1 ^ 2+(ev/2-\varepsilon_1)^2}{\gamma_1 ^ 2+(ev/2+\varepsilon_1)^2}\right ] \\ & + & \mbox{first and second terms with } \left\{c\leftrightarrow d , \varepsilon_1\leftrightarrow \varepsilon_2 , \gamma_1\leftrightarrow\gamma_2\right\}\end{aligned}\ ] ] here , we present our numerical results for the power dissipated of our two - level model between @xmath77-electrodes as a function of applied bias . based on which
levels are dominated during the transmission we consider three cases : one of the levels is chosen from the highest occupied molecular orbital ( homo ) @xmath46 and the other one from the lowest unoccupied molecular orbital ( lumo ) @xmath45 with @xmath78 and @xmath79 , both levels are chosen from homos @xmath46 and @xmath47 below the fermi energy with @xmath80 and @xmath81 and both levels are chosen from lumos @xmath48 and @xmath82 abbove the fermi energy with @xmath83 and @xmath79 ( see fig.[fig3 ] a - b - c respectively ) .
all the energies are measured with respect to the fermi energy of the system , which we set to zero ( @xmath84 ) .
we have borrowed these parameters from table ii in ref@xcite . in the all three cases the power dissipation in left and right electrodes is asymmetric and the condition ( 25 )
is verified .
a detailed elaboration of fig.[fig3]-a shows a cross point between the power dissipated in left and right electrodes @xmath85 which it does not happen in the other two cases .
this cross point means that heat is equally dissipated in both electrodes . from eq.(22 ) and eq.(23 ) it is easy to show that the condition to have equal heat dissipation in both electrodes is that the transmission function fulfills @xmath86 .
this means the transmission does not depend on the bias voltage and it is also energy independent in the transport window .
this is the feature that occurs in any ballistic processes .
another point we should stress out is that before the crossing point , more ( less ) heat is dissipated in the right electrode for positive ( negative ) bias , but after this point two electrodes change their roles and more(less ) heat is dissipated in the left electrode for positive ( negative ) bias ( see fig.[fig3]-a ) .
this behavior can be related to the homo and lumo levels domination during the electron transmission processes . indeed , after turning bias voltage the electrochemical potentials of the left and the right electrodes are shifted and an energy window opens for electrons to cross the junction and it results in a net electron current in the junction .
so for the energy - dependent transmission function which is higher in the lower part of the transport window , the more power is dissipated in the left electrode for positive bias ( case-2 and see fig.[fig3]-b ) . alternatively ,
if the transmission is higher in the upper part of the transport window , more heat is dissipated in the right electrode for positive bias ( case-3 and see fig.[fig3]-c ) .
a cartoon to present the scenario is depicted in fig.[fig3 - 1 ] to further investigation of the portion of given electrode , i.e. , left electrode , in heat dissipation for our toy model , we plotted the @xmath87 as a function of the bias in fig.[fig4 ] .
panel ( a ) of this figure shows that , there is a given positive bias about @xmath88 in which @xmath89 or @xmath90 , i.e. , symmetric heat dissipated in both electrodes .
more precise , for positive bias when @xmath91 ( @xmath92 ) we find @xmath93 ( @xmath94 ) so the portion of left electrode in heat dissipation is less ( more ) than half of the total heat .
this behavior is exactly reversed for negative bias .
briefly , the main conclusion of these results reveals when we plot @xmath95 as a function of @xmath43 for both positive and negative biases in panel ( b ) .
there is a point of intersection that in this point the heat dissipation in left lead is equal for positive and negative bias and after this point for positive ( negative ) bias the left electrode has a major ( minor ) portion in total heat dissipation , although before the intersection point , there is no significant difference between the heat dissipation for positive and negative bias . . ] . ]
we consider a system consists of a @xmath0 molecule attached to one - dimensional gold ( @xmath77 ) electrodes .
the whole system is described within a single electron picture by a tight - binding hamiltonian with nearest - neighbor hopping approximation .
the hamiltonian representing the entire system can be written as @xmath96 where @xmath97 and @xmath98 represent the left / right electrode and the @xmath0 molecule , respectively .
@xmath99 defines the coupling between left / right electrodes and the @xmath0 molecule . in our numerical calculation
we used the well - konw newns - anderson model for gold ( @xmath77 ) electrodes , whose self - energies are given by @xmath100 , where @xmath101 is the surface green s function defined as @xmath102@xcite .
@xmath103 is the nearest - neighboring hopping integral in the left / right electrodes .
the hopping strength in @xmath0 molecule depends on the dimerization of the carbon - carbon bonds ; thus , we consider different hopping integral elements : @xmath104 for the single bonds and @xmath105 for the double bonds . in the numerical calculations we set @xmath106 and @xmath107
. results for one and five contacts have depicted in fig.([fig5 ] ) and fig.([fig6 ] ) , respectively . in fig.([fig5]-a ) we have plotted the logarithmic scale of transmission function versus energy of au-@xmath0-au junctions with one contact . for an electron with energy @xmath8 , that comes from the left connection
, the probability of transmission function reaches its saturated value ( resonance peaks ) for the specific energy values . in order to verify heat dissipation feature in the junction , in fig.([fig5]-d ) we present the power dissipated in left lead as a function of the total power dissipated for negative and positive biases .
our results are in good agreement with the simple two - level model .
it can be seen that an intersection occurs between the power dissipated in left lead with negative and positive bias .
a crossing point in which heat dissipation in leads , finds symmetric features and does not depend on the bias polarity . in fig.[fig6 ] we have considered @xmath108 molecule connected to the leads via its pentagon face .
the result shows that the transmission function does not display any crossing point between the power dissipated in the left and right electrodes .
therefore , in this case , the power dissipated in the electrodes shows the asymmetry feature in the whole range of applied bias .
it shows how heat dissipation in the leads of a two - terminal molecular junction can be specified by its transmission characteristics and from it one can detect which molecular energy levels are more or less dominant in the transmission process .
in summary , we have studied heat dissipation in single - molecule junctions . using the generalized green s function technique and the landauer formalism , we have presented a detailed theoretical and numerical analyses of heat dissipation in a simple two level system .
we have shown how the transmission characteristics can affect on heat dissipation in the electrodes of a two - terminal junction .
notably , we have found some points in which asymmetry characteristics of heat dissipation in the left and the right electrodes failed . in other words , at this point heat dissipation is free of bias polarity . to validate our results on the simple toy model ,
we have chosen c60 molecule sandwiched between au electrodes .
we have simulated a junction based on tight - binding model and landauer approach . in order to have a different transmission spectrum for an equal energy window
we have considered both single and multi contacts cases .
the results show how heat dissipation in the leads of a two - terminal molecular junction can be specified by its transmission characteristics and from it one can detect which molecular energy levels are more or less dominant in the transmission process
we are very grateful to khalil naimi for his valuable stimulation during this work .
we wish you find your health .
suppose a representation with unitary matrix @xmath109 which diagonalizes the green function @xmath110 to @xmath111 as @xmath112 thus equation ( 11 ) can be written as @xmath113 this representation will not necessarily diagonalize @xmath114 indeed , as will show , interesting quantum interference effects often arise from the non - diagonal elements @xmath114 . in this representation ,
the effective hybridization matrices are @xmath115 and @xmath114 is non - diagonal as @xmath116 the expansion of transmission according to effective hybridization matrices elements is @xmath117 the above equation can be expressed with three following terms @xmath118 the two first terms constitute the non - mixing contributions from each energy level @xmath119 interference enters via the third term as @xmath120
according to eq . ( a1 ) it is straightforward that we have @xmath121 the inverse green function elements of eq . ( 1 ) , can be rewritten so that the hermitian and anti - hermitian part of self energy , @xmath122 , appear in inverse green function elements as follows @xmath123 or @xmath124 in which we have used of @xmath125 and @xmath126 is a hermitian part and last term in ( b3 ) is anti - hermitian part of @xmath127 . by using of eqs .
( 2)-(7 ) and ( a3 ) , above equation will be @xmath128 it is clear that both @xmath129 and @xmath130 are hermitian and so each trace is real and we have @xmath131 according to eq.(b4 ) we have @xmath132 and by using of eqs.(b2 ) and ( b9 ) , after the comparison of both sides of above equation , one yields the following conditions @xmath133 if we consider the following definition for @xmath134 quantity @xmath135 where @xmath32s are the energy of the two levels and @xmath33s are the broadening by contacts , then according to eqs .
( b12 ) and ( b13 ) we reach @xmath136 the above conditions indicate that the real part of self - energy causes shifts in the system energy levels , while the imaginary part has the effect of level broadening . in the case of symmetric system - lead coupling
, we take @xmath137 and @xmath138 , and effective hybridization matrices elements should satisfy @xmath139 so according to eq .
( b13 ) we obtain @xmath140 in what follows , we suppose that elements of effective hybridization matrices are real , so above relations can be written as @xmath141 and equations ( a7 ) and ( a8 ) have explicit expression as @xmath142 @xmath143 the above relation for @xmath144 can be expanded according to @xmath145s as @xmath146\ ] ] where constants @xmath37s have a following definitions @xmath147 } \\
c_2 & = & 2\gamma_2\frac{\varepsilon_1\varepsilon_2+\gamma_1\gamma_2+\gamma_1 ^ 2-\varepsilon_1 ^ 2}{\gamma_1[(\varepsilon_1-\varepsilon_2))^2+(\gamma_1+\gamma_2)^2 ] } \end{aligned}\ ] ] and constants @xmath39s will obtain by applying @xmath40 and also @xmath41 exchanges in corresponding @xmath37s .
finally , we can express transmission function @xmath7 as follows @xmath148
10 d. wang , w. shi , j. chen , j. xia and z. shuai , phys .
chem . chem .
phys , * 14 * , 16505 ( 2012 )
. a. v. kovalevsky , a. a. yaremchenko , s. populoh , p. thiel , d. p. fagg , a. weidenkaffb and j. r. fradea , phys .
phys , * 16 * , 26946 ( 2014 ) .
j. li , j. sui , y. pei , c. barreteau , d. berardan , n. dragoe , w. cai , j. he and li - d . zhao energy environ .
, * 5 * , 8543 ( 2012 ) .
a. aviram and m. a. ratner , chem .
* 29 * , 277 ( 1974 ) .
g. yu , j. gao , j. c. hummlen , f. wudl and a. j. heeger , science * 270 * , 1789 ( 1995 ) .
w. u. huynh , j. j. dittmer and a. p. alivisatos , science * 295 * , 2425 ( 2002 ) .
p. reddy , s. y. jang , r. a. segalman and a. majumdar science * 315 * , 1568(2007 ) .
j. c. cuevas and e. scheer .
_ molecular electronics : an introduction to theory and experiment_(world scientific , 2010 ) . m. paulsson and s. datta , phys . rev .
b * 67 * , 241403(r ) ( 2003 ) . c. m. finch , v. m. garcia - suarez , and c. j. lambert , phys . rev .
b * 79 * 033405 ( 2009 ) .
j. p. bergfield , m. a. solis and c. a. stafford , acs nano . * 4 * 5314 ( 2010 ) .
j. r. widawsky , p. darancet , j. b. neaton and l. venkataraman , nano lett .
* 12 * 354 ( 2012 ) .
a. nitzan , m. a. ratner , science * 300 * 1384 ( 2003 ) .
s. guo , g. zhou and n. tao , nano lett . * 13 * 4326 ( 2013 ) .
c. evangeli , k. gillemot , e. leary , m. t. gonzalez , g. r. bollinger , c. j. lambert and nicolas agrait , nano lett .
* 13 * 2141 ( 2013 ) .
s. bilan , l. a. zotti , f. pauly , and j. c. cuevas , phys .
b. * 85 * 205403 ( 2012 ) . s. h. ke , w. yang , s. curtarolo and h. u. baranger , nano lett .
* 9 * 1011 ( 2009 ) .
y. kim , w. jeong , k kim , w. lee and p. reddy , nature nanotechnology .
* 9 * , 881 ( 2014 ) .
g. d. mahan and j. o. sofo , proc .
, usa , vol * 93 * , 7436 ( 1996 ) . a. i. hochbaum , r. chen , r.d .
delgado , w. liang , e. c. garnett , m. najarian , a. majumdar , and p. yang , nature * 451 * , 163 ( 2008 ) .
a. i. boukai , y. bunimovich , j. tahir - kheli , j - k yu , w. a.goddard iii , and j.r .
heath , nature * 451 * , 168 ( 2008 ) .
l. a. zotti , m. brkle , f. pauly , w. lee , k. kim , w. jeong , y. asai , p. reddy and j. c. cuevas , new journal of physics * 16 * , 015004 ( 2014 ) .
w. lee , k. kim , w. jeong , l .
a. zotti , f. pauly f , j. c. cuevas and p. reddy , nature * 489 * , 209 ( 2013 ) .
m. buttiker , y. imry , r. landauer , and s. pinhas , phys .
b * 31 * , 6207 ( 1985 ) .
d. nozaki , h. m. pastawski , and g. cuniberti , new j. phys .
* 12 * , 063004 ( 2010 ) .
a. saffarzadeh , j. appl . phys . * 103 * , 083705 ( 2008 ) . g. granton , c. seiler , a. bagrets , l. venkataraman and f. evers , j. chem . phys .
* 139 * , 234701 ( 2013 ) . | motivated by recent work [ lee et al . nature * 489 * , 209 ( 2013 ) ] , on asymmetry features of heat dissipation in the electrodes of molecular junctions , we put forward an idea as a result of heat dissipation in the electrodes .
based on tight - binding model and a generalized green s function formalism , we describe the conditions under which heat dissipation shows symmetry characteristic and does not depend on the bias polarity .
we also show the power dissipated in the junction can be used to detect which energy levels of molecule junction play more or less role in the transmission process .
we present this idea by studying a simple toy model and au-@xmath0-au junction .
new and powerful ways of energy conversions are a demanding issue which appeals tremendous attentions from wide range of scientists . in this respect ,
thermoelectric devices are highly desirable since they deal with conversion of temperature to voltage differences or vice versa@xcite .
study of charge transport in molecules is very active field , with potential applications in molecular electronics @xcite and energy - conversion devices @xcite .
the efficiency of thermoelectric depends on a combination of material properties quantified by the thermoelectric figure of merit @xmath1 , where @xmath2 is the thermopower or seebeck coefficient , @xmath3 is an electronic conductivity and thermal conductivity @xmath4 contains contributions from electrons as well as phonons .
a good thermoelectric device needs to have a good electrical conductivity @xmath3 , and simultaneously a poor thermal conductivity @xmath4 .
however , in normal bulk material these two parameters are highly correlated via the well - known wiedemann - franz law .
systems that can independently optimize @xmath2 and @xmath3 are scarce . mahan and
sofo @xcite considered the optimization of the figure of merit as a mathematical problem and found that for a system with discrete electronic density of states like delta - function , the figure of merit diverges in the absence of any phonon contribution to the thermal conductivity @xcite .
single - molecule hybrid systems , where a single - molecule is sandwiched between metal electrodes , are one such system where can show a discrete density of states at the frontier molecular orbitals ( mos ) .
understanding and manipulating these hybrid systems could open new pathways for enhancing thermoelectric performance , impossible in bulk semiconductors due to their fundamentally different transport properties . two recent experimental@xcite and theoretical @xcite works reported that heat is not equally dissipated in both electrodes and this asymmetry depends on the bias polarity .
the authors used scanning tunnelling probes with integrated nanothermocouples probed heat dissipation in the electrodes of a molecular junction and showed heat dissipation can be controlled by transmission characteristics .
motivated by these works , we report here a detailed theoretical analysis of the joule heating in current - carrying single - molecule junctions . by combining the tight - binding model and a generalized green s function formalism
, we investigate different conditions of the heat dissipation of molecular junction and its relation with the electron transport . to this end
, we first study transmission and heat dissipation properties of a simple two level systems . using it as a starting point
, we describe the conditions under which heat dissipation shows symmetry characteristic and does not depend on the bias polarity .
this conclusion further is supported by numerical effective single - particle tight - binding model . as test systems
, we consider @xmath0 molecule sandwiched between to metal ( au ) electrodes via single and multiple contacts .
the rest of the paper is organized as follows . in sec .
ii we present a detailed formalism for finding transmission in a toy model and its heat dissipation characteristics . in sec .
iii we present our numerical results for a real au-@xmath0-au junction based on tight - binding model . in sec .
iv we summarize our findings and some technical details are presented in the appendices . |
solar dynamo models predict that the toroidal and poloidal components of the global magnetic field are regenerated from one other .
the poloidal field is transformed into the toroidal field from differential rotation ( the @xmath0-effect ) , and the toroidal field is twisted into the poloidal field ( the @xmath1-effect ) .
these alternate and repeat in a 22 year cycle [ e.g. @xcite , 2008 ] .
this toroidal field component , previously measured by @xcite [ 1994 ] using wso data over the period of 1977 - 1992 , provided evidence for an extended activity cycle of 16 - 18 years . @xcite [ 2005 ] also measured the toroidal field using mount wilson data from 1986 - 2004 and verified the dynamo model for the creation and reversal of the toroidal field .
since we are facing a peculiar solar minimum , we want to investigate whether there is also some peculiar activity in the toroidal fields leading up to the minimum .
three solar cycles of wso magnetic field data are available to us now , as well as a complete cycle observed with soho / mdi [ @xcite , 1995 ] .
we first generated maps of the inclination angle for both the positive and negative field polarities using the wso magnetograms and the mdi synoptic maps .
this is done by tracking the line - of - sight magnetic field , @xmath2 , for a given carrington coordinate as it crosses the disk . , is the angle between @xmath3 and @xmath4 .
we measure @xmath2 as it crosses the disk over various @xmath5 s . ]
we can fit the measurements of @xmath2 for different longitudes to the following equation to calculate the average east - west inclination angle of the field , @xmath1 : @xmath6 where @xmath7 is proportional to the true magnetic field .
a derivation of this fit gives the following formula ( @xcite , 1994 ) : @xmath8 we fit the values of @xmath2 and @xmath5 into the equation for @xmath1 to generate maps of the inclination angles .
following @xcite , individual regions are tracked across the wso magnetograms .
for mdi , measurements of the same features in synoptic maps constructed using data from central meridian , @xmath9 and @xmath10 are compared .
we generate two separate maps for each carrington rotation , one for positive polarity @xmath2 and the other for negative .
the difference of these two maps gives the inclination difference .
this inclination difference map gives us the east - west direction of the magnetic field at each point on the sun for each carrington rotation .
the inclination difference maps for each rotation are averaged over all longitudes to show the inclination differences at each latitude over time , using both wso and mdi data . from these averages
we construct the contour plots shown in figure [ incdiff_wso_all ] , and the line plots ( figure [ incangle_wso_all ] ) for each hemisphere .
the wso data span 34 years , which gives us three solar cycles of coverage .
however , the mdi values are somewhat different than wso , as can be seen in figure [ incdiff_compare ] .
we believe this is due to asymmetric noise characteristics of the mdi instrument . in 2003 , soho began to periodically roll from @xmath11 to @xmath12 , which caused periodic irregularities in our data that need to be better calibrated .
from the inclination difference summary averages , we detected the first emergence of cycle 24 some time during 2003 - 2004 .
this emergence pattern is very much like the pattern observed in the previous two cycles .
the new emerging inclination angles , which are related to the toroidal fields , are consistent with those from the previous cycles .
this suggests that the toroidal field is not a clear indicator of the duration of the following solar minimum or the generation of the poloidal fields at the concurrent minimum .
however , if the toroidal field is generated from the preceding poloidal field , we might expect to see that the next cycle s toroidal field will be weak .
also , the inclination angle that indicates the toroidal field at the sun s surface is only a distant reflection of the toroidal field that is present wherever the true solar dynamo is acting . particularly in the higher resolution
mdi data , the weak field regions that contribute to the analysis may also reflect the spatially averaged characteristics of a surface dynamo process whose contribution to the solar cycle dynamo is uncertain .
further work needs to be done to solve the calibration issues of mdi .
dikpati , m. , de toma , g. & gilman , p.a . ; 2008 , ap.j . , 675 920 .
scherrer , p.h . , et al .
; 1995 , sol .
phys . , 162 , 129 .
shrauner , j. a. , & scherrer , p. h. ; 1994 , sol .
phys . , 153 , 131 .
ulrich , r. k. , & boyden , j. e. ; 2005 , ap.j .
, 620 , 123 . | thirty - four years of wso ( wilcox solar observatory ) and thirteen years of soho / mdi ( michelson doppler imager on the solar and heliospheric observatory ) magnetograms have been studied to measure the east - west inclination angle , indicating the toroidal component of the photospheric magnetic field .
this analysis reveals that the large - scale toroidal component of the global magnetic field is antisymmetric around the equator and reverses direction in regions associated with flux from one solar cycle compared to the next .
the toroidal field revealed the first early signs of cycle 24 at high latitudes , especially in the northern hemisphere , appearing as far back as 2003 in the wso data and 2004 in mdi .
as in previous cycles , the feature moves gradually equatorward .
cycles overlap and the pattern associated with each cycle lasts about 17 years .
even though the polar field at the current solar minimum is significantly lower than the three previous minima , the toroidal field pattern is similar . |
the standard collinear qcd factorization theorems have set the standard for the use of perturbative qcd calculations to probe certain properties of the microscopic structure of matter .
it will be instructive to recall the essential ingredients that lend the standard factorization treatments their predictive power : * unambiguous prescription for calculating pertrubatively well - behaved higher order corrections to the hard scattering .
* correlation functions describing the non - perturbative factors are well - defined and have universality properties so that , once measured , they can be useful for future phenomenological studies . * evolution equations , to allow the correlation functions to be compared at different scales .
the collinear pdfs and fragmentation functions have by now been parametrized by a wide range and variety of experimental data , and have become indispensable tools in general high energy physics . ideally , tmd - factorization , in which the pdfs and fragmentation functions also carry information about the intrinsic transverse momenta of partons , should follow a very similar framework .
however , it is only very recently @xcite that tmd - factorization has reached a level of logical completeness similar to the well - known collinear cases . as long there was ambiguity in the formulation of tmd - factorization
, the procedure for correctly identifying tmds in fits , as well as the correct procedure for implementing evolution has remained unclear . in the absence of a full tmd - factorization treatment ,
there have been several popular but separate approaches for dealing with tmds : * generalized parton model ( gpm ) : a phenomenological approach is to extract tmds from fits to data while assuming a literal parton model interpretation of the tmds .
one typically ignores evolution and therefore the fits correspond to specific scales .
* resummation : begin with a collinear treatment valid for large transverse momentum , and attempt to improve the treatment of lower transverse momentum by resuming logarithms of @xmath1 .
a severe limitation of this approach is that it is bound to fail below some @xmath2 , where many of the most interesting effects of tmd physics are expected to become important .
* models : non - perturbative models of tmds can be used to study specific non - perturbative aspects of hadron structure ( see a. bacchetta s talk for a review of model calculations . ) .
but there are ambiguities in how these tmds are related to the ones used in actual perturbative qcd calculations of cross sections .
in particular , it is unclear what hard scale they should correspond to .
* lattice calculations : lattice calculations of tmds ( see , e.g. , @xcite ) describe the non - perturbative distribution of patrons from first principles , but also require a clear definition for the tmds , and a clear prescription for use in complete cross section calculations .
a useful tmd - factorization treatment should allow the advantages of each of these approaches to be unified within a single , clear formalism for relating tmd studies to observable cross sections .
fortunately , this is now possible following the recent work of ref .
@xcite ( see , esp . ,
chapts . 10 and 13 ) .
( a similar general approach was developed earlier by ji , ma , and yuan @xcite , which built upon the collins - soper - sterman ( css ) @xcite formalism . )
we will show how fits to tmds can be constructed from existing work that follows the above tabulated approaches , but which include evolution and are consistent with a full tmd - factorization .
we apply the tmd - factorization method developed recently by collins in ref .
@xcite . the factorization theorem for the drell - yan ( dy ) process , for example ,
is @xmath3 note the similarity of the first term above to a gpm description ; there is a hard part @xmath4 and a convolution of two tmd pdfs , @xmath5 and @xmath6 .
however , in the full tmd - factorization treatment , they have acquired scale dependence through the renormalization group parameter @xmath7 and @xmath8 and @xmath9 ( which obey @xmath10 ) . moreover , there is no explicit appearance of a soft factor .
the first term in eq . [ dyhadronic ] is appropriate for describing the small @xmath2 region ( @xmath11 ) .
the @xmath12 term corrects the large @xmath2 behavior and can be calculated in terms of normal integrated pdfs . ref .
@xcite derives very specific operator definitions for the tmd pdfs , which include the role of soft gluons and account for all spurious divergences which have hindered efforts to clearly define the tmd pdfs in the past .
the evolution of the individual tmds in transverse coordinate space is governed by the collins - soper ( cs ) equation @xcite , @xmath13 and the renormalization group equations , @xmath14 in eq .
[ csevolution ] , @xmath15 is the kernel for evolution with respect the energy variable @xmath16 , while @xmath17 and @xmath18 are the anomalous dimensions .
these can all be calculated in perturbation theory , though @xmath15 becomes non - perturbative at large @xmath19 and needs to be fit .
it is a universal function , however , both with respect to different processes and with regard to pdfs versus fragmentation functions .
working within a gpm approach , gaussian parametrizations have been fit to various tmds in , for example , ref .
@xcite from low energy sidis measurements . the non - perturbative information corresponding to @xmath15 at large @xmath19 in eq .
( [ csevolution ] ) has been extracted from more traditional applications of the css method , such as in ref .
@xcite . with
the non - perturbative parts constrained , specific tables for the tmds in eq .
( [ dyhadronic ] ) can be generated for any arbitrary scale @xmath20 once the perturbative contributions have been calculated .
we have carried this out explicitly for the unpolarized tmd pdfs and fragmentation functions in ref .
@xcite , and have made the results available at the website : + https://projects.hepforge.org / tmd/. + an example showing the unpolarized up quark pdf for a selection of @xmath20 values is the graph in fig .
[ tmdpdf ] .
note that , once the @xmath12 term is included in eq .
( [ dyhadronic ] ) , the factorization formula exactly valid over the whole range of @xmath0 from @xmath21 to @xmath20 . ,
@xmath22 and @xmath23 gev . ]
a crucial element of the tmd - factorization approach is that the tmds are well - defined and , therefore , can be taken to be genuinely universal .
one consequence of this is that different theories of the non - perturbative properties of hadron structure can be compared with one another , and with experimental results over a range of hard scales .
ultimately , in order to take full advantage of the predictive power of tmd - factorization , requires a concerted effort to improve upon non - perturbative fits , to calculate higher orders to the anomalous dimensions , k , the coefficient functions , and the hard parts , and to provide numerical implementations of evolution in transverse momentum space similar to what has been done in ref .
@xcite , but applied to other tmd pdfs and fragmentation functions , including interesting spin dependent ones .
we have recently been involved in efforts to extend evolution to the sivers function @xcite , and plan to continue this work to generate evolved fits for most of the various spin - dependent tmds .
m. aybat thanks the organizers of the transversity 2011 workshop for the invitation to this interesting and fruitful workshop .
support for this work was provided by the research program of the `` stichting voor fundamenteel onderzoek der materie ( fom ) '' , which is financially supported by the `` nederlandse organisatie voor wetenschappelijk onderzoek ( nwo ) '' .
x. d. ji , j. p. ma and f. yuan , phys .
d * 71 * , 034005 ( 2005 ) [ arxiv : hep - ph/0404183 ] .
x. d. ji , j. p. ma and f. yuan , phys .
b * 597 * , 299 ( 2004 ) [ arxiv : hep - ph/0405085 ] .
b. u. musch , p. hagler , j. w. negele and a. schafer , arxiv:1011.1213 [ hep - lat ] .
collins and d.e .
soper , nucl .
* b193 * , 381 ( 1981 ) ; erratum : ibid * b213 * , 545 ( e ) ( 1983 ) ; j.c .
collins , d.e .
soper , and g. sterman nucl .
* b250 * , 199 ( 1985 ) .
p. schweitzer , t. teckentrup and a. metz , phys .
d * 81 * , 094019 ( 2010 ) [ arxiv:1003.2190 [ hep - ph ] ] . | in this talk , we summarize how qcd evolution can be exploited to improve the treatment of transverse momentum dependent ( tmd ) parton distribution and fragmentation functions .
the methods allow existing non - perturbative fits to be turned into fully evolved tmds that are consistent with a complete tmd - factorization formalism over the full range of @xmath0 .
we argue that evolution is essential to the predictive power of calculations that utilize tmd parton distribution and fragmentation functions , especially tmd observables that are sensitive to transverse spin . |
although thousands of supernovae ( sne ) have been discovered to date , only several dozen have been detected at either or radio wavelengths ( designated xsne or rsne , respectively ) .
many of these xsne and rsne were originally detected simply as variable or radio sources , and only verified as true sne after careful examination of archival optical data or through optical follow - up ( e.g. , * ? ? ?
* ; * ? ? ?
this particular path to discovery has meant that only a handful of these sources have been well - studied during the first several hundred days , a period which is critical for identifying and characterizing the true nature of the sne .
such sne typically turn out to be core - collapse sne of either type ibc , which have been associated with long - duration gamma - ray bursts ( grbs ) , or type ii , whereby the intense and/or radio emission is thought to stem from the interaction between the expanding shock and a dense progenitor wind .
when well - sampled , the additional multi - wavelength constraints from xsne and rsne can provide physical insights into the late evolutionary stages of massive stars that are otherwise impossible to obtain .
for instance , while robust constraints on the overall time - averaged mass - loss in various phases of stellar evolution exist , there has been a long - standing debate over the number and sequence of various evolutionary stages for massive stars , and relatively few observational constraints on the actual evolution of mass - loss within many particular stages ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
because the sn blast wave travels orders of magnitude faster than the stellar wind , studying the interaction between the blast wave and the progenitor stellar wind allows us to probe tens of thousands of years of evolution in a matter of decades .
sne progenitors are considered to provide the bulk of all processed stellar material additionally , and thus the characterization of these sources can likewise help further our understanding of overall galactic chemical enrichment . here
we report on the spectroscopic confirmation of one such serendipitous source , @xcite , and present a multi - wavelength follow - up study using archival data to determine its explosion date and temporal properties .
was originally detected as circinus galaxy ( cg ) x-2 , an ultraluminous x - ray source in the nearby circinus galaxy @xcite , which @xcite found to exhibit many characteristics of a young , type ii sn enshrouded in a dense circumstellar environment .
in particular , it demonstrated a factor of @xmath530 increase in flux between , a @xmath610 kev thermal spectrum with a strong , blended fe emission - line component at 6.85 kev , and spatial coincidence with a strong radio and h@xmath7-emitting point source .
the proximity of this sn affords us a rare opportunity to study in detail a type ii sn which is strongly interacting with its circumstellar medium .
this paper is organized as follows : data and reduction methods are detailed in @xmath8[data ] ; confirmation of as a sn and isolation of its explosion date are provided in @xmath8[confirm ] ; overall temporal and spectroscopic constraints for are investigated in @xmath8[character ] ; and finally conclusions and future prospects are explored in @xmath8[conclude ] . throughout this paper ,
we adopt a distance of @xmath9 mpc to the circinus galaxy ( converted from * ? ? ?
while the circinus galaxy lies close to the galactic plane ( @xmath1038 ) , it is located within a galactic `` window '' with a visual extinction of @xmath11 and a neutral hydrogen column density of @xmath12 @xmath13 ( whereas neighboring regions typically have @xmath14 and @xmath15@xmath16 @xmath13 ; * ? ? ?
* ; * ? ? ?
* ) . due to its @xmath17 inclination
, however , there is significant internal obscuration as well [ ( @xmath18@xmath19 @xmath13 typically ; * ? ? ?
we describe below the extensive observational data used to constrain the properties of .
fortunately , plentiful archival data exist for the circinus galaxy due to the fact that it hosts the second closest compton - thick active galactic nuclei ( agn ) to our own galaxy and exhibits signs of vigorous star formation ( e.g. , * ? ? ?
the resulting multiwavelength dataset provides good constraints on the explosion date and yields useful long - term and radio light curves .
lies 25to the south of the circinus galaxy nucleus .
we adopt the position of ( j2000 ) , determined from the radio observations .
sn1996cr was visited several times in service mode with the vlt fors i spectrograph in two separate campaigns .
the first program was initiated over the course of 2005 - 03 - 06 through 2005 - 03 - 17 but never completed , resulting in a low signal - to - noise spectrum . as such
, we do not provide any further details .
the second program was executed over three nights ( 2006 - 01 - 26 , 2006 - 02 - 02 , and 2006 - 03 - 10 ) , yielding a high - quality spectrum confirming the ambiguous features seen in the spectrum from our first program .
our analyses focus only on the spectrum from the first night , as the observations on the subsequent two nights suffer from significantly worse seeing ( @xmath01520 ) such that the co - addition of these frames failed to improve the signal - to - noise of the final spectrum .
our observations on the first night consisted of four 1200 s exposures taken with the 300v grism and the tek ( 24@xmath20 pixel ) ccd using a slitwidth of 20 .
the grism and ccd combination provided a dispersion of 2.66 pixel@xmath21 and a total useful wavelength coverage of 36007990 .
these images were taken in excellent seeing conditions , in which stellar sources had a full - width half - maximum ( fwhm ) measured at 0607 on the fors i chip over the course of the night .
our target , , had an @xmath22% larger fwhm image size than the point sources in the acquisition image , suggesting that it may be spatially extended or possibly contaminated by a coincident hii region .
we investigated this further by measuring the fwhm orthogonal to the dispersion axis for the strong emission lines [ ] @xmath235007 and h@xmath7 , which had sizes of 074 and 085 , respectively .
the inherent fwhm slowly increased with decreasing wavelength , such that the fwhms quoted above were 5@xmath243% and 35@xmath244% larger than their stellar equivalents , respectively .
thus the [ ] @xmath235007 fwhm is formally unresolved , while the h@xmath7 line is extended .
the strength , extent , and visible asymmetry in the fwhm profile of the extended h@xmath7 are fully consistent with the underlying hii region ( see @xmath8[confirm ] for details ) , although we caution that marginally - resolved light echoes could potentially contaminate and broaden the psf ( e.g. , * ? ? ?
the spectral resolution ( 11 , or 500 km s@xmath21 ) was estimated from the width of the arclines taken with a 10 slit , which matched closely the on - sky image size of .
after bias subtraction and flat - fielding using standard techniques , the co - added spectra were extracted , wavelength calibrated , and flux calibrated ( using the standard ltt4816 ) .
[ fig : sn1996cr - spec ] shows the resulting optical spectrum from 2006 - 01 - 26 .
we conservatively estimate the flux errors on the spectrum at 20% .
-1 cm we searched through the european southern observatory ( eso ) , anglo australian telescope ( aat ) , united kingdom schmidt telescope ( ukst ) , and _ hubble space telescope _
( _ hst _ ) archives .
we only list here data that ultimately were used to constrain the explosion date of .
relevant data are listed in tables [ tab : data_eso][tab : data_ukst ] .
aperture photometry was performed throughout , with psf - fitting employed as a cross - check ; there is only one instance where the psf - fitting magnitude differed from the aperture - measured value at @xmath25 , which we explicitly document below .
the circinus galaxy was observed extensively with susi at the ntt on 1993 april 9 using several narrow - band filters ( # 369 : [ ] , # 430:5100 continuum , # 629:h@xmath7 + [ ] , # 700 : [ ] , # 443:7000 continuum , # 415:[]@xmath237892 ) as well as with the irac2 camera at the eso / mpi 2.2 m telescope on 1994 june 25 using broad - band @xmath26 and @xmath27 filters .
reduced images were kindly provided by e. oliva and a. marconi ( private communication , 2007 ) , for which cutouts of are shown in fig .
[ fig : eso_images ] .
details of the observations and reduction procedures are given in @xcite .
the [ ] , h@xmath7 + [ ] , and [ ] images were the only images provided to us in a flux - calibrated state , and therefore are the only ones for which we measure photometry .
magnitudes for the region in the vicinity of were measured using a 10 radius circular aperture and previously established zero points .
an aperture correction of 0.2 mags was estimated empirically using several bright , isolated point sources in the images .
the resulting aperture - corrected magnitudes are presented in table [ tab : data_eso ] .
the circinus galaxy was observed with both the wfpc2 and nicmos instruments aboard _
hst _ on four separate occasions , as outlined in table [ tab : data_hst ] .
details of the photometric and astrometric reduction of the wfpc2 data are given in @xcite .
briefly , after standard calibration of the _ hst _ images , we used the iraf package daophot to measure aperture - corrected magnitudes using a 02 radius aperture for all sources down to the detection limit .
the _ hst _
images were then aligned to the _
astrometric reference frame to @xmath004 . due to the negligible overlap of the f606w observation with the other filters and complete lack of coverage of itself ,
we do not discuss it further .
for the nicmos data , of which only the nic3 images provided useful imaging , we used the standard pipeline data products .
we performed photometry using a 05 radius aperture and applied band - dependent aperture corrections determined from a set of isolated , unsaturated point sources in the vicinity of .
cutout images of are presented in fig .
[ fig : hst_images ] .
notably , appears to lie at the center of a diffuse patch of h@xmath7 emission in the f656n image and our aperture magnitude here differs from a psf - fitting one at @xmath28 ( 0.11 magnitudes ) .
we thus adopt the psf - fitting magnitude , as well as an additional systematic error of 0.1 magnitudes to reflect the larger uncertainty associated with the deblending .
the circinus galaxy was observed numerous times with the taurus fabry - perot instrument on the aat during 1995 - 02 - 21 to 1995 - 02 - 28 and 1996 - 03 - 15 to 1996 - 03 - 20 .
the data were retrieved through the aat archive .
the observations were comprised primarily of narrow - band imaging and taurus spectral imaging cubes centered on the @xmath234686 , [ ] @xmath295007,4959 , h@xmath30[]@xmath296583,6548 , and [ ] @xmath296731,6716 emission lines .
the original investigators discuss their observational design in @xcite and @xcite , while more complete data reduction procedures are outlined in @xcite . to summarize ,
the taurus instrument was used in the angstrom imaging mode , wherein narrow - band filters were used at different tilt angles to isolate the lines of , [ ] , h@xmath30 [ ] , and [ ] @xcite .
each square pixel subtended 0315 on the sky and the atmospheric seeing at fwhm averaged @xmath01220 . unfortunately , a significant portion of the archived taurus images for the circinus galaxy have only limited accompanying ccd calibration data
; there were often no obvious superbias frames , dark frames , dome - flats , or flux standards , so the photometric quality of the reduced data is limited .
our best - effort reduction proceeded as follows .
the overscan regions were used to subtract the bias from each frame and each image was flat - fielded using a sky flat .
when multiple frames in a given filter were available , the images were combined to reject cosmic rays using the iraf task crrej .
source detection and photometry were performed with sextractor @xcite .
a set of 40 relatively bright , isolated , unsaturated stars , which are detected in all the aat and _ hst _ images , were chosen to register the aat images to the _ hst _ coordinate frame .
the alignment of each aat image to this reference frame is accurate to @xmath00102 ( ) . the zeropoints for the narrow - band images were obtained by bootstrapping the aat magnitudes to the well - determined _ hst _ ones . to this end
, _ hst _ colors were obtained for the 40 stars and matched to standard main sequence stellar templates from the bruzual atlas , modified by a standard galactic dust model with @xmath31 ( the extinction measured toward the circinus galaxy ; e.g. , * ? ? ?
* ) using the iraf package synphot . among the 40 calibration stars , 25 had colors consistent with one of the stellar templates .
these particular stellar templates and the taurus - specific transmission curves were input into synphot to convert between _ hst _ and taurus magnitudes .
we used the iraf task fitparams to compare interactively the _ hst_-derived magnitudes to the sextractor ones and derive reasonable zeropoints for each image .
we typically rejected a few outliers in fitparams to obtain an adequate fit ( 0.30.5 mag error typically on the zeropoint ) .
the , [ ] , and [ ] bands were observed multiple times , so we combined zeropoint estimates to ensure that magnitudes for the detected sources were consistent to within @xmath00.10.2 mags across these images . for bands which sampled comparable spectral windows , we used synphot to convert from one band to another ( for instance , from taurus 5020/30 to _ hst _ f502n or from the vlt fors spectrum to all overlapping bands ) .
the resulting magnitudes are listed in table [ tab : data_aat ] , while fig .
[ fig : aat_images ] shows 12 images taken in 1995 february and 1996 march . lllrll susi / ntt & eso#369 ( [ ] ) & 1993 - 04 - 09 & 900 & 07 & @xmath32 + susi / ntt & eso#430 ( 5108 cont . ) & 1993 - 04 - 09 & 900 & 07 & + susi / ntt & eso#629 ( h@xmath7 + [ ] ) & 1993 - 04 - 09 & 480 & 07 & @xmath33 + susi / ntt & eso#700 ( [ ] ) & 1993 - 04 - 09 & 900 & 07 & @xmath34 + susi / ntt & eso#443 ( 7027 cont . ) & 1993 - 04 - 09 & 900 & 07 & + susi / ntt & eso#415 ( [ ] @xmath237892 ) & 1993 - 04 - 09 & 900 & 07 & + irac2/eso - mpi 2.2 m & j ( 1.25 @xmath20 m ) & 1993 - 06 - 25 & 360 & 09 & @xmath35 + irac2/eso - mpi 2.2 m & h ( 1.65 @xmath20 m ) & 1993 - 06 - 25 & 360 & 09 & @xmath36 + lllrl wfpc2 & f606w & 1996 - 08 - 11 & 200 , 400 & + nic3 & f164n & 1998 - 03 - 16 & 64 & @xmath37 + nic3 & f166n & 1998 - 03 - 16 & 64 & @xmath38 + nic3 & f196n & 1998 - 03 - 16 & 80 & @xmath39 + nic3 & f200n & 1998 - 03 - 16 & 80 & @xmath40 + nic3 & f212n & 1998 - 10 - 16 & 160 & @xmath41 + nic3 & f215n & 1998 - 10 - 16 & 160 & @xmath42 + wfpc2 & f502n & 1999 - 04 - 10 & 900 , 900 & @xmath43 + wfpc2 & f547 m & 1999 - 04 - 10 & 60 & @xmath44 + wfpc2 & f656n & 1999 - 04 - 10 & 800 , 800 & @xmath45 + wfpc2 & f814w & 1999 - 04 - 10 & 40 & @xmath46 + llrll 5020/30 ( @xmath470 ) & 1995 - 02 - 21 & 300 & 12 & @xmath48 + 6583/45 ( @xmath479 ) & 1995 - 02 - 22 & 300 & 12 & @xmath49 + 6583/45 ( @xmath479 ) & 1995 - 02 - 28 & 267 & 11 & @xmath50 + 6618/55 ( @xmath477 ) & 1995 - 02 - 28 & 1442 & 11 & @xmath51 + @xmath52 & 1995 - 02 - 28 & 60 & 12 & @xmath53 + @xmath54 & 1995 - 02 - 28 & 20 & 14 & @xmath55 + 4690/26 ( @xmath475 ) & 1996 - 03 - 16 & 1200 & 35 & @xmath56 + 6730/30 ( @xmath475 ) & 1996 - 03 - 16 & 1200 & 30 & @xmath57 + 4690/26 z - cube ( @xmath475 ) & 1996 - 03 - 16 & & 30 & + 4690/26 ( @xmath470 ) & 1996 - 03 - 17 & 1200 & 31 & @xmath58 + 4690/26 ( @xmath470 ) & 1996 - 03 - 19 & 1000 & 12 & @xmath59 + 5020/30 ( @xmath470 ) & 1996 - 03 - 19 & 120 & 13 & @xmath60 + 6730/30 ( @xmath470 ) & 1996 - 03 - 19 & 1000 & 13 & @xmath61 + the circinus galaxy was observed on numerous occasions with the ukst as detailed in table [ tab : data_ukst ] .
the data were retrieved from the supercosmos archive , although only a subset of the photographic plates have been properly digitized and archived in a usable form .
[ fig : ukst_images ] shows four images taken in 1998 march and april .
the h@xmath7 images from 1998 ( as well as similar degraded ones from 1999 and 2001 ) show a slight enhancement over the 1995 h@xmath7 reference images from the aat and eso . for the four images that were taken after the discovery window of and properly digitized , we performed photometry using a 10 radius circular aperture and previously established zero points .
an aperture correction of 0.2 mags was estimated empirically using several bright , isolated point sources in the images .
the resulting aperture - corrected magnitudes are presented in table [ tab : data_ukst ] .
lllrll or14405 & iiiaf / og590 & 1991 - 07 - 17 & 2400 & ai3 * & + or16161 & iiiaf / og590 & 1994 - 06 - 27 & 2480 & bi3 * & + or16238 & iiiaf / og590 & 1994 - 08 - 12 & 3300 & aid2 & + or17455 & iiiaf / og590 & 1997 - 03 - 11 & 300 & ai * & + or17484 & iiiaf / og590 & 1997 - 03 - 31 & 300 & a * & + ha17930 & 4415/ha659 & 1998 - 03 - 05 & 10800 & a2 & @xmath62 + or17954 & 4415/og590 & 1998 - 03 - 19 & 900 & a & @xmath63 + ha17987 & 4415/ha659 & 1998 - 04 - 05 & 10800 & at2 & @xmath64 + or17988 & 4415/og590 & 1998 - 04 - 05 & 900 & a & @xmath65 + or18316 & 4415/og590 & 1999 - 03 - 11 & 900 & a * & + ha18323 & 4415/ha659 & 1999 - 03 - 12 & 2160 & bu * & + ha18328 & 4415/ha659 & 1999 - 03 - 14 & 10800 & bt * & + ha19197 & 4415/ha659 & 2001 - 05 - 24 & 8400 & au2 * & + the circinus galaxy was observed on 2007 - 03 - 23 with the _ swift _ ultraviolet / optical telescope ( uvot ) for depending on the filter .
the images were reduced using standard pipeline procedures and photometry was performed on the images using a 30 radius circular aperture .
was not detected in any of the six uvot filters , with aperture - corrected upper limits of @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , and @xmath71 .
there are three additional useful upper - limit constraints based on the sne monitoring observations of the circinus galaxy by rev .
robert evans on 1995 - 03 - 31 , 1995 - 08 - 13 , and 1996 - 03 - 03 ( private communication , 2000 ) .
we have adopted upper limits of @xmath72 based on r. evans estimated limiting magnitude of @xmath73 .
lrrrrrrl & 1995 - 02 - 14 & 61.0 & @xmath74 & @xmath75 & @xmath76 & @xmath77 & + _ rosat _ hri & 1995 - 09 - 14 & 4.1 & @xmath78 & & @xmath79 & & + _ rosat _ hri & 1996 - 02 - 18 & 1.1 & @xmath79 & & @xmath80 & & + _ rosat _ hri & 1996 - 09 - 13 & 1.8 & @xmath81 & & @xmath82 & & + _ rosat _ hri & 1997 - 03 - 03 & 26.4 & @xmath83 & & @xmath84 & & + _ rosat _ hri & 1997 - 08 - 17 & 45.9 & @xmath85 & & @xmath86 & & + _ bepposax _ & 1998 - 03 - 13 & 85.2 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & + _ chandra _ acis - s & 2000 - 01 - 16 & 1.0 & @xmath91 & @xmath92 & @xmath93 & @xmath92 & 21% pile - up + _ chandra _ acis - s & 2000 - 03 - 14 & 4.9 & @xmath94 & @xmath95 & @xmath96 & @xmath97 & 2% pile - up + _ chandra _ acis - s & 2000 - 03 - 14 & 23.1 & @xmath98 & @xmath99 & @xmath100 & @xmath99 & 20% pile - up + _ chandra _ hetgs & 2000 - 06 - 15 & 67.1 & @xmath101 & @xmath102 & @xmath103 & @xmath104 & 3% pile - up + _ bepposax _ & 2001 - 01 - 07 & 27.2 & @xmath105 & @xmath106 & @xmath107 & @xmath108 & + _ chandra _ acis - s & 2001 - 05 - 02 & 4.4 & @xmath101 & @xmath109 & @xmath110 & @xmath109 & 21% pile - up + _ xmm - newton _ & 2001 - 08 - 06 & 85.5 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & + _ chandra _ hetgs & 2004 - 06 - 02 & 55.0 & @xmath115 & @xmath116 & @xmath117 & @xmath118 & 3% pile - up + _ chandra _ hetgs & 2004 - 11 - 28 & 59.0 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & 3% pile - up + _ swift _ xrt & 2007 - 03 - 25 & 8.0 & @xmath123 & @xmath124 & @xmath125 & @xmath126 & + details of the various observations are given in table [ tab : data_xray ] .
we describe below our reduction methods for each dataset .
when possible , the fluxes and absorption - corrected luminosities for were calculated via spectral analysis in xspec @xcite using the cash statistic @xcite . unless stated otherwise , errors on spectral parameters are for 68% confidence , assuming one parameter of interest . following the treatment of ( e.g. , * ? ? ?
* ) , we characterized the spectra of with an absorbed variable - abundance non - equilibrium ionization ( nei ) shock model ( _ vpshock _ using nei version 2.0 in xspec ) based on atomdb @xcite .
the best - fitted parameters to the combined _ xmm - newton _ and _ chandra _ hetgs 0th order dataset were as follows : @xmath127 @xmath13 , @xmath128 kev , @xmath129 , @xmath130 , @xmath131 , and @xmath132 .
we note , however , that these specific abundances , and even the estimated temperature to some extent , should be used with caution , as there are known to be significant deviations between adopted models ( e.g. , * ? ? ?
the typical column density , @xmath133 , derived from spectral fits to the data is much larger than the estimated galactic absorption column ( @xmath03@xmath134 @xmath13 ) , implying significant internal absorption either from the disk of the circinus galaxy at large or the immediate vicinity of .
fluxes were measured by varying the normalization of the above model in the or bands .
our adopted model fit all of the spectral data adequately , although we note that there was typically some excess residual emission around prominent emission lines , as well as an apparent soft excess in the 2004 hetgs data .
these issues have a negligible effect , however , on our flux estimates to within errors and will be addressed in more detail in a separate publication on spectral analysis of ( f. bauer et al .
, in preparation ) .
the circinus galaxy was observed on several occasions with _
chandra _ using the backside - illuminated advanced imaging ccd spectrometer ( acis - s ) in the focal plane , both with and without the high - energy transmission grating spectrometer ( hetgs ) .
archival acis - s and hetgs zeroth - order data were retrieved from the _ chandra _ data archive and processed following standard procedures using ciao ( v3.4 ) software .
additionally we removed the 05 pixel randomization , corrected for charge transfer inefficiency ( cti ) , performed standard _
asca _ grade selection , excluded bad pixels and columns , and screened for intervals of excessively high background ( none was found ) .
analysis was performed on reprocessed _ chandra _ data , primarily using ciao , but also with ftools ( v6.3 ) and custom software including acis extract ( v3.128 ; * ? ? ?
spectra were extracted using acis extract with a 95% encircled - energy region derived from the _ chandra _ psf library . as lies at the edge of diffuse emission associated with the agn , a background spectrum was extracted from a local annular region after excluding nearby point sources and the strongest portions of the circumnuclear halo and ionization cone associated with the agn ( the annulus size was grown until approximately 100 counts could be extracted ) .
we generated calibration products ( including an aperture correction ) for the hetgs / acis - s spectra and fit them separately within xspec using our best - fit model to estimate the flux .
the majority of the observations were performed in configurations which mitigate pile - up ; when significant we have noted the estimated pile - up fractions for , and corrected for them using the xspec _ pileup _ model of @xcite .
the circinus galaxy was observed once with _ xmm - newton _ using the epic p - n and mos1/mos2 detectors on 2001 - 08 - 06 for @xmath135 ks .
archival data were retrieved from the _ xmm - newton _ science archive and processed following standard procedures with sas ( v7.0.0 ) .
additionally , we screened for and removed intervals of excessively high background ( @xmath136% for mos1/mos2 and @xmath137% for p - n ) , leaving 85.5 ks , 91.8 ks , and 59.5 ks of useful exposure with the mos1 , mos2 , and p - n instruments , respectively .
spectra were extracted with sas using a circular aperture of radius 110 and local background was extracted in a manner similar to that done for the _ chandra _ observations .
we generated calibration products for the p - n / mos1/mos2 datasets and joint - fit the three spectra within xspec using our best - fit model to estimate the flux .
additionally , we used the _ xmm - newton _ data to place upper limits on the potential flux from in the _ asca _ and _ bepposax _ datasets , both of which lack the spatial resolution to resolve the galaxy into its various components ( e.g. , see fig . 1 of * ? ? ?
to this end , spectra for the entire circinus galaxy were extracted from a 25 radius circular aperture , excluding emission from .
unfortunately , the ultraluminous source is known to vary with a period of 7.5 hr and comprises a non - negligible fraction of the total emission from circinus .
thus we extracted a pessimistic spectrum additionally excluding .
this provides conservative upper limits for the total emission from the circinus galaxy during the _ asca _ and _ bepposax _ observations , under the assumption that the fluxes of the various emission components within the circinus galaxy ( aside from and ) have not changed since 1995 .
based on the spatially resolved observations that we do have , this assumption appears reasonable .
more optimistic upper limits can be estimated by further subtracting the time - averaged flux from .
given ambiguities in determining how the unresolved observations sample the light curve of , we simply provide here the flux of in its `` high '' state ( see * ? ? ?
* ) which is @xmath138 ergs s@xmath21 @xmath13 and @xmath139 ergs s@xmath21 @xmath13 in the 0.52 kev and 28 kev bands , respectively , and note that optimistic upper limits for the flux of would likely be a factor of @xmath023 lower .
the circinus galaxy was observed on several occasions with _
rosat _ using the high - resolution imager ( hri ) .
processed data were retrieved from the high energy astrophysics science archive research center ( heasarc ) , and analysis was performed using ftools and custom software .
we adopted the best - fit model from the combined observations ( see above ) to estimate the _ rosat _ hri upper limits .
upper limits were measured using a 10@xmath140 aperture and a local background extracted to match roughly the amount of diffuse and scattered emission thought to reside within the source aperture .
count rates were converted to fluxes with xspec using the final _ rosat _ calibration products and our adopted best - fit model . adopting similar models does not have a significant impact on our derived flux upper limits .
the circinus galaxy was observed once with the _ swift _ telescope ( xrt ) on 2007 march 23 for 7.4 ks .
processed data were retrieved from _ swift _ archive , and analysis was performed using ftools and custom software .
is visibly separated from the nucleus ( 25separation ) and diffuse circumnuclear emission and appears to be only marginally contaminated .
a spectrum of was extracted from the event list using a 10 pixel ( 235 ) radius circular aperture , masked to exclude emission from the nucleus and diffuse halo to as large an extent as possible .
additionally , a background was extracted in a manner similar to that done for the _ chandra _ observations so as to match roughly the amount of diffuse and scattered emission believed to be present within the source aperture .
low photon statistics , however , limited the fidelity of the background subtraction .
we generated calibration products for the spectrum and fit it within xspec using our best - fit model to estimate the flux .
the circinus galaxy was observed once with _
asca _ using the gas imaging spectrometers ( gis ) and solid state imaging spectrometers ( sis ) on 1995 - 02 - 14 for 61.1 ks .
processed data were retrieved from heasarc , providing of usable data .
analysis was performed using ftools and custom software .
the target appears unresolved with _
asca _ and was treated as a point source for extraction purposes .
spectra for the entire circinus galaxy were extracted from a 25 radius circular aperture , while backgrounds were taken from blank - sky observations at the same position and with identical screening criteria .
contamination from a neighboring source 5@xmath141 away was minimal .
total fluxes were measured via simultaneous spectral fitting of the gis and sis data and upper limits for were estimated by subtracting the flux estimated from the _ xmm - newton _ data as detailed above .
the circinus galaxy was observed on two occasions with _
bepposax _ using the low energy concentrator spectrometer ( lecs ) and medium energy concentrator spectrometer ( mecs ) on 1998 - 03 - 03 for 83.9/138.0 ks and on 2001 - 01 - 07 for 26.9/52.1 ks .
processed data were retrieved from heasarc , providing 83.7/137.2 ks and 26.8/51.7 ks of usable data , respectively .
analysis was performed using ftools and custom software .
the target appears unresolved with _
bepposax _ and was treated as a point source for extraction purposes .
lecs and mecs spectra for the entire circinus galaxy were extracted from a 34@xmath141 radius circular apertures ( limited by the availability of calibration files ) , while backgrounds were taken from blank - sky observations at the same position and with identical screening criteria .
contamination from a neighboring source 5@xmath141 away was minimal .
total fluxes were measured via simultaneous spectral fitting of the lecs and mecs data , and upper limits for were estimated by subtracting the flux estimated from the _ xmm - newton _ data as detailed above .
lrr|rr|rr|rr|rr|rr|l 1995 - 03 - 03 & c204 & 375 & & & 8640 & @xmath50.7 & 4800 & @xmath51.3 & & & & & mosaic field `` a '' + 1995 - 03 - 03 & c204 & 375 & & & 8640 & @xmath50.7 & 4800 & @xmath50.9 & & & & & mosaic field `` b '' + 1995 - 03 - 30 & c363 & 1.5a & & & & & & & & & 1418 & @xmath56.6 & + 1995 - 04 - 16 & c418 & 6c & & & & & & & 2768 & @xmath51.1 & 1418 & @xmath51.2 & + 1995 - 05 - 27 & c204 & 375 & & & 8640 & @xmath50.7 & 4800 & @xmath51.3 & & & & & mosaic field `` a '' + 1995 - 05 - 27 & c204 & 375 & & & 8640 & @xmath50.7 & 4800 & @xmath50.6 & & & & & mosaic field `` b '' + 1995 - 06 - 07 & c204 & 375 & & & 8640 & @xmath50.7 & 4800 & @xmath51.5 & & & & & mosaic field `` a '' + 1995 - 06 - 07 & c204 & 375 & & & 8640 & @xmath51.5 & 4800 & @xmath50.5 & & & & & mosaic field `` b '' + 1995 - 07 - 27 & c204 & 6c & & & & & & & 2368 & @xmath51.3 & 1376 & @xmath53.4 & + 1995 - 08 - 01 & ct09 & 750b & & & & & & & & & 1418 & @xmath517 & + 1996 - 02 - 06 & c466 & 750d & & & & & & & & & 1664 & @xmath51.5 & + 1996 - 08 - 01 & c363 & 375 & & & 8640 & [email protected] & 4800 & @xmath51.2 & & & & & + 1996 - 12 - 15 & c586 & 375 & & & 8512 & [email protected] & 4928 & [email protected] & & & & & mosaic field `` a '' + 1996 - 12 - 15 & c586 & 375 & & & 8512 & [email protected] & 4928 & [email protected] & & & & & mosaic field `` b '' + 1996 - 12 - 05 & c505 & 375 & & & & & & & 2368 & @xmath50.4 & 1384 & @xmath50.7 & + 1997 - 06 - 17 & v100 & 6a & & & 8425 & 14@xmath241 & & & & & & & + 1997 - 06 - 17 & v100 & 6a & & & 8425 & 15@xmath243 & & & & & & & spectral - line mode + 1997 - 12 - 05 & v099 & 6c & & & & & & & 2268 & 22@xmath243 & & & + 1997 - 12 - 05 & v099 & 6c & & & & & & & 2268 & 37@xmath244 & & & spectral - line mode + 1997 - 12 - 31 & c694 & 6c & & & 8500 & 96@xmath2415 & 4800 & 119@xmath2412 & & & & & 8.5 ghz extrapolated + & & & & & & & & & & & & & from 6.0 ghz data + 1999 - 06 - 19 & c788 & 375 & & & & & & & & & 1384 & 89@xmath243 & + 1999 - 06 - 19 & c788 & 375 & & & & & & & & & 1418 & 85@xmath243 & spectral - line mode + 2000 - 06 - 19 & v137 & 6b & 22190 & 76@xmath2421 & & & & & & & & & + 2000 - 06 - 19 & v137 & 6b & 22203 & 72@xmath2420 & & & & & & & & & + 2000 - 06 - 19 & v137 & 6b & 22215 & 69@xmath2421 & & & & & & & & & + 2002 - 07 - 18 & v137 & 1.5 g & 22172 & 63@xmath2410 & & & & & & & & & + 2002 - 07 - 18 & v137 & 1.5 g & 22216 & 71@xmath2410 & & & & & & & & & + 2002 - 07 - 18 & v137 & 1.5 g & 22224 & 70@xmath2410 & & & & & & & & & + 2003 - 09 - 07 & c1224 & ew367 & 23659 & 63@xmath2410 & & & & & & & & & phase cal only , flux forced + 2003 - 11 - 03 & c1049 & h214 & 16960 & 80@xmath2410 & & & & & & & & & + 2003 - 11 - 03 & c1049 & h214 & 19008 & 77@xmath2410 & & & & & & & & & + 2003 - 11 - 04 & c1049 & h214 & 21056 & 79@xmath2410 & & & & & & & & & + 2003 - 11 - 04 & c1049 & h214 & 22796 & 73@xmath2410 & & & & & & & & & + 2003 - 11 - 08 & c1049 & 1.5d & & & 8256 & 163@xmath2410 & 5056 & 227@xmath2411 & & & & & + 2004 - 04 - 02 & c1424 & 6a & & & 8640 & 162@xmath248 & 4800 & 240@xmath2410 & 2368 & 364@xmath2416 & 1384 & 481@xmath2419 & + 2004 - 05 - 20 & cx065 & 1.5b & 22088 & 63@xmath2415 & & & & & & & & & phase cal only , flux forced + 2004 - 06 - 13 & v176 & 750d & 22209 & 57@xmath2413 & & & & & & & & & + 2004 - 06 - 13 & v176 & 750d & 22195 & 59@xmath2411 & & & & & & & & & + 2004 - 08 - 01 & v176b & h168 & 22209 & 66@xmath246 & & & & & & & & & + 2004 - 08 - 01 & v176b & h168 & 22195 & 66@xmath246 & & & & & & & & & + 2005 - 03 - 14 & c1368 & h214 & 22193 & 51@xmath2415 & & & & & & & & & + 2005 - 04 - 20 & c1368 & 750a & 22194 & 68@xmath246 & & & & & & & & & phase cal only + 2005 - 11 - 13 & c1049 & 1.5c & & & 8640 & 159@xmath245 & 4800 & 246@xmath2412 & & & & & + 2006 - 04 - 29 & c1049 & h214 & 18752 & 77@xmath247 & & & & & & & & & + 2006 - 04 - 29 & c1049 & h214 & 21056 & 67@xmath246 & & & & & & & & & + 2006 - 06 - 21 & c1049 & 1.5d & & & 8640 & 154@xmath2411 & 4800 & 261@xmath2410 & & & & & + 2006 - 09 - 14 & c1341 & h75 & & & & & & & & & 1418 & 505@xmath2420 & + 2007 - 06 - 24 & vx013a & & 22316 & 67@xmath246 & & & & & & & & & + radio observations of the circinus galaxy spanning 1995 to 2006 were retrieved from the australia telescope compact array ( atca ) archive and are presented in table [ tab : data_radio ] and fig .
[ fig : sn1996cr - lc ] .
the data were reduced with miriad ( v4.0.5 ) following the procedures outlined in the atnf miriad user manual .
the `` primary '' flux calibrator is 1934@xmath142638 , which is assumed to be constant in time with flux densities of 14.94 , 11.60 , 5.83 , 2.84 , and 1.03 jy at 20 , 13 , 6 , 3 , 1 cm , respectively .
as noted in table [ tab : data_radio ] , there are a few instances in which 1938@xmath142638 is not included as part of the dataset , although the observations are initially `` primed '' using this source ; for such cases , fluxes of the `` secondary '' calibrators were compared to their interpolated historical values from calibration monitoring efforts and were fixed to those values when they strayed by more than @xmath143 from them .
the secondary calibrators , which served as the gain , bandpass , and phase calibrators for the circinus galaxy , varied from observation to observation and include sources 1921@xmath142293 ( bandpass , alternate 1 cm flux ) , 1329@xmath142665 ( bandpass ) , 1549@xmath142790 ( bandpass ) , 1414@xmath14259 ( phase ) , and 1236@xmath142684 ( phase ) .
both source and calibration data were flagged to remove bad time intervals and channels , time intervals strongly affected by interference , and obvious emission and absorption lines .
the calibration tables were applied to the circinus galaxy data , from which deconvolved , primary - beam - corrected images were made . on rare occasions , a phase self - calibration was required . integrated flux densities ( determined from imfit ) or upper limits ( determined from imstat ) were then determined for .
these values are provided in table [ tab : data_radio ] .
there appears to be considerable dispersion in 22.5 ghz flux densities , which were mostly acquired as part of larger australia long baseline array observations .
we lack comparison data at other wavelengths to determine if this is simply additional calibration uncertainty or potentially real flux variations .
however , given that there appears to be dispersion even between observations taken simultaneously , we caution that the dispersion is not intrinsic , but rather due to systematic errors .
we note that the only other strong point source in the field of view was the nucleus of the circinus galaxy , which is likely due to emission from the compton - thick agn or a compact circumnuclear starburst . as a check on the flux of
, we extracted flux measurements from the point - like nucleus in a manner identical to .
while the nucleus could vary intrinsically and may additionally be contaminated at long wavelengths by diffuse emission from the extended disk , it still provides us with a secondary estimate of any systematic error associated with a given observation .
we find that the majority of measurements of the nucleus lie between 1030% of its mean total flux in each band , although a few observations ( mainly at 22.5 ghz where contamination from water maser emission is possible ) vary by a factor of up to four ; we have adopted larger systematic errors for these observations as a result .
thus , aside from the 22.5 ghz data , we are generally confident in our measured fluxes for .
we carried out very long baseline interferometry ( vlbi ) observations of on 2007 - 06 - 24 using the following telescopes of the australian long baseline array : atca , mopra , parkes , hobart , and ceduna .
the observations were phase - referenced to the nearby calibrator .
we observed both senses of circular polarization at a frequency of 22 ghz with a total bandwidth of 65 mhz , and for a total time of @xmath19 h. we report here only our preliminary results and note that a full description is forthcoming ( n. bartel et al .
, in preparation ) .
was detected only on the three baselines involving the parkes , atca and mopra telescopes .
the flux of is consistent with previous atca measurements , while its extent appears resolved by the longest baselines at @xmath28 .
although gaussian and disk models provide adequate fits , we adopt a model consisting of an optically thin spherical shell , as was found to be appropriate for ( see , e.g. , * ? ? ?
* ; * ? ? ?
for a ratio of the outer to the inner shell radius of 1.25 , we find an outer angular radius of @xmath15 mas , which corresponds to @xmath144 cm for an assumed distance of 3.8 mpc .
modeling a smaller outer to inner shell ratio like 1.10 could perhaps narrow the angular radius by @xmath15% , while using a filled - center model ( equivalent to an inner shell radius of 0 ) might increase the angular radius by @xmath120% . among the numerous grb - associated sne that now exist ,
nearly all have been identified as type ibc s , which suggests that they arise from wolf - rayet progenitors .
although sn1996cr is identified as a type iin sn at late times , it is worth searching for any evidence of temporal and spatial gamma - ray emission , as its original type may have been quite different . additionally , there are at least two events in which a good observational case can be made for the association of a grb with a type iin sn : / @xcite and / @xcite .
we find eight batse grb detections over the lifetime of the _ compton _ gamma - ray observatory @xcite which include the circinus galaxy within their error regions .
one of these grbs , , appears to lie within the explosion date constraints of , suggesting a tentative association .
importantly , however , was detected by multiple gamma - ray observatories and has a dramatically smaller error estimate via triangulation from the interplanetary network ( ipn ; * ? ? ?
* ; * ? ? ?
the full ipn 3@xmath145 confidence annulus is only @xmath0033 wide and lies @xmath213 from the sn at its closest tangential point , thus strongly ruling out any identification with .
tables [ tab : data_aat][tab : data_radio ] give a detailed summary of the observations relevant for determining the explosion date of . while was discovered in the x - ray band , the archival data available are too limited and subject to potential geometrical and structural effects ( e.g. , early absorption , cavities ) to narrow down the explosion date to better than an @xmath146 yr window between 19902000 . although is relatively well - sampled at radio frequencies , those observations similarly suffer from the same early - time effects , thus limiting their utility to date the sn .
[ fig : sn1996cr - lc ] demonstrates that the radio emission begins to sharply `` turn on '' in mid-1996 , signaling that must have exploded sometime prior to this .
several narrow - band optical images from the aat fortuitously bracket the sn explosion date between 1995 - 02 - 28 and 1996 - 03 - 16 , allowing us to isolate the sn explosion to within a year .
final confirmation of comes in the form of the optical spectrum , which signals many telltale features of a sn embedded in dense circumstellar material ( csm ) .
we note that s position coincides with a small , powerful hii region .
however , it is clear from a comparison of figs .
[ fig : eso_images ] and [ fig : hst_images ] that much of the h@xmath7 is due to the unresolved sn - csm interaction ( @xmath170% ; see also table [ tab : data_hst ] ) and thus is likely to be associated with photoionized wind material surrounding the progenitor ( e.g. , * ? ? ?
contamination is likely to be negligible for other wavelengths .
the tightest direct temporal constraints arise from comparing the @xmath52-band image taken on 1995 - 02 - 28 with the narrow - band [ sii ] image taken on 1996 - 03 - 16 . as such
, we can not constrain the spectral form of at these early times ( i.e. , whether it is line- or continuum - dominated ) , which make conversions between various bands somewhat uncertain . for simplicity , we adopt here an intrinsically flat continuum absorbed by @xmath147 mags at early times .
this is likely conservative , considering typical early sne are intrinsically quite blue .
the extinction was determined using the three narrow - band discovery - image magnitude constraints and is @xmath00.8 mags more than a standard conversion of our best - fitted @xmath148 @xmath13 from the late - time spectra ; a factor of @xmath12 variance is commonly seen between extinction and @xmath133 in our own galaxy ( e.g. , * ? ? ?
* ) , so this discrepancy is probably not necessarily reason for concern , although it could be indicative of further early - time absorption .
we note that this model should provide adequate conversions to broad - band magnitudes since early - time sne spectra are often dominated by strong , broad lines that blend to form a pseudo - continuum . using synphot in iraf to convert from the narrow aat , [ ] , and [ ] bands
, we find an equivalent discovery magnitude of @xmath149 . at the distance of the circinus galaxy
, this equates to an uncorrected absolute magnitude of @xmath150 , or @xmath151 when corrected for the minimum expected extinction .
this provides a strong lower limit since the sn likely went off prior to the discovery images and the apparent extinction estimated from the discovery magnitudes is substantially higher .
the basic observational picture for core - collapse sne starts at shock breakout ( e.g. , * ? ? ? * hereafter cf94 ) , where the blast wave emerges from the progenitor with a typical velocity of @xmath152 km s@xmath21 and propagates into the csm formed by the pre - sn stellar wind .
the interaction between the ejecta in the forward shock and the csm leads to the formation of a reverse shock , which travels back into the expanding ejecta at least relative to the outward expanding contact discontinuity that separates the shocked ejecta and the shocked csm .
[ fig : sn_diagram ] provides a basic one - dimensional picture of the overall scenario .
the full three - dimensional structure could be much more complicated .
for instance , the contact discontinuity is susceptible to rayleigh - taylor instabilities , which are only resolved in multi - dimensions ( e.g. , * ? ? ? * ; * ? ? ? * ) .
likewise , asymmetries could exist in the csm or sn explosion itself due to the nature of the progenitor or the existence of a binary companion ( e.g. , sn1979c , sn1987a ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
for a typical slow ( @xmath153 km s@xmath21 ) , dense ( @xmath154@xmath155 m@xmath4 yr@xmath21 ) constant progenitor wind , the forward shock / csm interaction produces a very hot shell of shocked csm ( @xmath156 k ) , while the reverse shock / ejecta interaction produces a denser , cooler shell of shocked ejecta ( @xmath157 k ) .
the latter interaction has a much higher emission measure and generates copious and emission which subsequently photoionizes a broad inner ejecta region from which many of the optical lines are believed to originate .
when the csm density is high and/or the power - law density distribution of the ejecta ( @xmath158 ) is steep , the reverse shock will remain radiative and a relatively thin , cool , dense , partially absorbing shell ( cds ) should form between the reverse shock and the contact discontinuity .
uv and radiation associated with the original shock breakout ( detected for the first time from with _
* ) , as well as radiation from the ongoing forward and reverse shocks , may additionally photoionize the outer csm . when the csm is h - rich
, this ionization should give rise to the strong , narrow - line h emission which is thought to epitomize the class ( e.g. , * ? ? ?
* ) . to place within the context of the above scenario
, we examine our temporal and spectral data in detail . to assess the early history of , we compared the @xmath159-band light curves for five , as well as 25 additional sne of varying types , to the optical constraints detailed in @xmath8[data ] and plotted in fig .
[ fig : sn1996cr - lc ] .. ] the three @xmath160 upper limits from r. evans here provide some additional insight into the nature of s light curve . we note that the light curves of sne are relatively uniform and can generally be divided into two distinct parts : ( 1 ) a sharp rise and variable decline rate between @xmath1610.035 per day within the first @xmath1100 days , followed by ( 2 ) a nearly universal decline from @xmath162 with a rate of @xmath1630.01 per day .
the latter is thought to be linked to the decay of @xmath164co @xcite .
the above parameters define our average template light curve , and we use as our pivot point the constraints from 1996 - 03 - 16 to 1996 - 03 - 19 , which lie 0380 days from maximum . as a consistency check
, we note that the slope between our 1996 and 1999 data points falls roughly within the expected range for sne ; we adopt @xmath165 per day to fit our points .
we find that r. evans constraint on 1995 - 03 - 31 excludes earlier explosion maximum dates for all but the faintest observed light curves , while the limit on 1995 - 08 - 13 excludes brighter than average observed light curves up to @xmath150 days prior to this date , and the limit on 1996 - 03 - 03 fails to exclude any additional time .
the dark grey region in the upper panel of fig .
[ fig : sn1996cr - lc ] shows the range of parameter space which is still viable , demonstrating that these upper limits exclude from being an extremely luminous sn over @xmath115% of the explosion window . from this
we argue that ultimately had an uncorrected absolute @xmath159-band magnitude somewhere between @xmath166 and @xmath167 .
once corrected for @xmath24.5 magnitudes of extinction , falls within the typical range for luminous sne . at late times , lies well above an extrapolation of the average decay rate , and is in fact @xmath11.5 magnitudes brighter in 2006 than in 1999 .
this fact demonstrates robustly that must be buoyed either by circumstellar interaction or unresolved light echoes ( e.g. , * ? ? ?
* ; * ? ? ?
we discuss these two possibilities further in @xmath8[opt_spectra_results ] .
several sne ( particularly of the iin sub - type ) have been shown to be strong near - ir emitters , due to reprocessed emission from hot dust ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . whether the hot dust arises from preexisting dust in the circumstellar gas or from newly formed dust in the ejecta is still unclear .
nic3 narrow - band constraints we have demonstrate that any near - ir emission above the earlier eso near - ir imaging is likely to be minimal .
thus if a significant hot - dust component formed , it must have either been short - lived ( e.g. , destroyed by the initial uv flash from the sn or rapidly cooled ) or occurred after the nic3 observations in 1998 .
mid - ir observations will ultimately constrain the dust - content of better , since the bulk of anticipated dust is likely to be at temperatures of @xmath1100 k ( the dust temperature in sn1987a , for instance , is @xmath0180 k ; * ? ? ? * ) . while the pre - sn optical imaging does not provide useful constraints on the progenitor due to the strong extinction toward the circinus galaxy
, the eso near - infrared data does not suffer as much . the @xmath26-band limits in 1994 june are strong enough to exclude stars with @xmath168 , effectively excluding the bright end of the luminous blue variable ( lbv ) distribution . as fig .
[ fig : sn1996cr - spec ] demonstrates , the optical spectrum of exhibits strong narrow emission lines of h and o that typify the sne class ( e.g. , * ? ? ?
* ) superimposed on very broad emission complexes .
the strong narrow lines suggest that circumstellar interaction plays an important role in the overall emission almost from the onset and , because of the strength of this emission , it can often mask the true nature of the photospheric emission .
for instance , @xcite was spectrally a type iin and photometrically a type iil , @xcite and @xcite were spectral `` hybrids '' showing type iin - like h lines superposed on otherwise type ia - like spectra , and @xcite spectrally evolved from a type ic to a type iin on the timescale of a few years .
-0.5 cm in the case of , where we have limited temporal coverage during the early phases , we can not deduce its early photometric or spectral type .
we must rely solely on the late - time sn emission lines in our vlt spectrum to provide diagnostics on ejecta abundances and shock emission processes , as well as potential information on the mass - loss history and evolutionary status of the sn progenitor ( e.g. , cf94 , * ? ? ?
* , * ? ? ?
[ fig : sn1996cr - compare ] shows , aside from the heavy extinction , the spectrum of bears many striking similarities to the classic sne , @xcite , @xcite , and @xcite , and thus can probably be considered to be representative of the class as a whole .
we used the contributed iraf package specfit @xcite to model the various spectral components of .
we began by fitting a power - law continuum , absorbed using the galactic extinction curve of @xcite with @xmath169 fixed at 3.1 .
however , since there is some degeneracy between the power - law slope and degree of extinction , we fixed @xmath170 to a value of 1.8 determined by the narrow - line balmer decrement and then determined a best - fitted power - law slope of 3.4 .
this single powerlaw , however , underestimates the curvature of the continuum between 55007000 by @xmath010% , and thus conversely leads to overestimates in some line fluxes and widths . to fit the continuum
more accurately , we instead adopted a broken power - law model with slopes of 2.65 above and 4.47 below a break at 6400 , respectively , yielding a statistically reliable model of the continuum over the entire spectrum .
note that the narrow balmer lines are thought to arise from the unshocked progenitor wind and thus can be considered as the outer layer of emission .
we therefore consider the extinction derived from the balmer decrement to be a lower limit and caution that some interior components could suffer considerably higher extinction .
we do not , however , see strong evidence of p cygni absorption in any of the emission lines , suggesting that further absorption in the vicinity of is probably minimal .
this may also imply that the bulk of the csm is fully ionized .
we find only marginal evidence for na d absorption ( ) , which can be used to approximate the overall extinction ( e.g. , * ? ? ?
the strength of this absorption appears too small compared to our other estimates , although the spectral resolution , signal - to - noise , and potential for contamination from complex line emission ( see below ) in the vlt spectrum severely compromise any meaningful constraint .
we contend that the lack of broad - band blue and near - uv detections , compared to the strong blue continua typically seen in comparable sne ( e.g. , , , ; * ? ? ?
* ; * ? ? ?
* ) , is fully consistent with our estimated extinction parameters and hence is reasonable .
we proceeded to fit all of the obvious emission lines with narrow gaussian components .
the parameters of each line were originally fit separately , but eventually the central wavelengths and fwhms of the h , he , and heavier element lines were tied together as denoted in table [ tab : line_ids ] to reduce free parameters and improve error estimates ; this was justified by the fact that the free - parameter values overlapped with the nominal fixed ratio values to within errors .
we note that there were , however , several large broad residual wings around the h@xmath7 and o lines .
thus we fit blended narrow and broad gaussians , again linking central wavelengths and fwhms as denoted in table [ tab : line_ids ] .
we describe each element separately below .
lllllll 1 & [ ] @xmath237319 & n & 1.4@xmath24 0.5 & [email protected] & [email protected] & @xmath17129 + 2 & & b & 13.6@xmath24 1.0 & [email protected] & 7324.9 ( @xmath1711 * ) & @xmath17129 + 3 & & c1 & 6.2@xmath24 0.5 & [email protected] & 7420.1 ( @xmath17130 * ) & @xmath17130 + 4 & & c2 & 19.3@xmath24 0.7 & [email protected] & 7251.1 ( @xmath17131 * ) & @xmath17130 + 5 & & c3 & 16.8@xmath24 0.5 & [email protected] & 7190.5 ( @xmath17132 * ) & @xmath17130 + 6 & [ ] @xmath237136 & n & 1.8@xmath24 0.2 & 76.4@xmath24 9.7 & 7136.8 ( f ) & @xmath1718 + 7 & @xmath237065 & n & 3.1@xmath24 0.3 & [email protected] & 7065.7 ( @xmath17121 * ) & @xmath17121 + 8 & [ ] @xmath236731 & n & 3.6@xmath24 0.3 & [email protected] & [email protected] & 741@xmath2424 + 9 & [ ] @xmath236716 & n & 4.0@xmath24 0.1 & 239.0@xmath24 7.8 & 6717.6 ( @xmath1718 * ) & @xmath1718 + 10 & @xmath236678 & n & 1.1@xmath24 0.2 & [email protected] & 6675.7 ( @xmath17121 * ) & @xmath17121 + 11 & [ ] @xmath236583 & n & 19.9@xmath24 0.7 & [email protected] & [email protected] & @xmath1718 + 12 & h@xmath7 & n & 34.5@xmath24 0.8 & [email protected] & [email protected] & 669@xmath2418 + 13 & & b & 9.5@xmath24 1.0 & [email protected] & 6563.0 ( @xmath17112 ) & 4077@xmath24398 + 14 & [ ] @xmath236548 & n & 6.4@xmath24 0.0 & 432.3@xmath24 0.0 & 6547.8 ( @xmath17111 * ) & @xmath1718 + 15 & [ ] @xmath236300 & n & 5.9@xmath24 0.4 & [email protected] & [email protected] & @xmath1718 + 16 & & b & 5.3@xmath24 0.6 & [email protected] & 6296.2 ( @xmath17115 ) & @xmath17129 + 17 & & c1 & 2.1@xmath24 0.3 & [email protected] & 6376.8 ( @xmath17130 * ) & @xmath17130 + 18 & & c2 & 5.8@xmath24 0.4 & [email protected] & 6231.6 ( @xmath17131 * ) & @xmath17130 + 19 & & c3 & 2.9@xmath24 0.3 & [email protected] & 6179.6 ( @xmath17132 * ) & @xmath17130 + 20 & [ ] @xmath236087 & n & 0.5@xmath24 0.2 & [email protected] & 6087.0 ( f ) & @xmath1718 + 21 & @xmath235876 & n & 1.6@xmath24 0.2 & [email protected] & [email protected] & 913@xmath2465 + 22 & [ ] @xmath235755 & n & 1.2@xmath24 0.2 & [email protected] & 5755.0 ( f ) & @xmath1718 + 23 & [ ] @xmath235577 & n & 0.2@xmath24 0.1 & [email protected] & 5579.0 ( @xmath17128 * ) & @xmath1718 + 24 & & b & 0.3@xmath24 0.2 & [email protected] & 5579.0 ( @xmath17128 * ) & @xmath17129 + 25 & & c1 & 0.3@xmath24 0.2 & [email protected] & 5644.5 ( @xmath17130 * ) & @xmath17130 + 26 & & c2 & 0.3@xmath24 0.1 & [email protected] & 5516.0 ( @xmath17131 * ) & @xmath17130 + 27 & & c3 & 0.3@xmath24 0.2 & [email protected] & 5469.9 ( @xmath17132 * ) & @xmath17130 + 28 & [ ] @xmath235007 & n & 5.2@xmath24 0.3 & [email protected] & [email protected] & @xmath1718 + 29 & & b & 3.5@xmath24 0.4 & [email protected] & 5007.9 ( @xmath17128 ) & 2795@xmath24174 + 30 & & c1 & 2.4@xmath24 0.3 & [email protected] & [email protected] & 2049@xmath2419 + 31 & & c2 & 4.5@xmath24 0.3 & [email protected] & [email protected] & @xmath17130 + 32 & & c3 & 1.9@xmath24 0.3 & [email protected] & [email protected] & @xmath17130 + 33 & [ ] @xmath234959 & n & 1.6 ( @xmath17127 * ) & 540.8 ( @xmath17127 * ) & 4960.0 ( @xmath17128 * ) & @xmath17128 + 34 & & b & 1.1 ( @xmath17128 * ) & 362.7 ( @xmath17128 * ) & 4960.0 ( @xmath17128 * ) & @xmath17129 + 35 & & c1 & 0.7 ( @xmath17129 * ) & 225.2 ( @xmath17129 * ) & 5018.9!(@xmath17130 * ) & @xmath17130 + 36 & & c2 & 1.1 ( @xmath17130 * ) & 509.2 ( @xmath17130 * ) & 4905.2 ( @xmath17131 * ) & @xmath17130 + 37 & & c3 & 0.6 ( @xmath17131 * ) & 227.9 ( @xmath17131 * ) & 4866.8 ( @xmath17132 * ) & @xmath17130 + 38 & h@xmath172 & n & 1.9 ( @xmath17111 * ) & 745.0 ( @xmath17111 * ) & 4861.1 ( @xmath17112 * ) & @xmath17112 + 39 & @xmath234685 & n & [email protected] & [email protected] & 4684.5 ( @xmath17121 * ) & @xmath17120 + 40 & h@xmath173 & n & 0.4 ( @xmath17111 * ) & 350.2 ( @xmath17111 * ) & 4340.0 ( @xmath17112 * ) & @xmath17112 + lllllll @xmath174@xmath175 $ ] & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 + @xmath174@xmath175 $ ] & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 + /@xmath174@xmath175 $ ] & @xmath188 & @xmath189 & @xmath190 & @xmath191 & @xmath192 & @xmath193 + @xmath174@xmath175 $ ] & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 + @xmath174@xmath175 $ ] & @xmath200 & @xmath201 & & & & + @xmath174@xmath175 $ ] & @xmath202 & @xmath203 & & & & + @xmath174@xmath175 $ ] & @xmath204 & @xmath205 & & & & + @xmath174@xmath175 $ ] & @xmath206 & @xmath207 & & & & + & @xmath208 & @xmath209 & & & & + & @xmath210 & @xmath211 & & & & + & @xmath212 & @xmath213 & & & & + we should caution that the late - time spectrum we observe for could potentially be contaminated by early - time sn spectral features due to the presence of unresolved light echoes from dust in the immediate circumstellar environment or in the intervening interstellar medium ( e.g. , * ? ? ? * ; * ? ? ?
notably , the spectra of sne near maximum light are the most likely to continue on as light echoes , and thus we would expect such contamination to be isolated primarily to the continuum of since early - time spectra of sne routinely exhibit either blue continua or psuedo - continua comprised of extremely broad emission lines which trace the high velocity ejecta @xcite . for , we find only distinct emission lines of low - to - moderate velocity ( i.e. , @xmath214 km s@xmath21 ) on top of a smooth continuum , with the only notable exception being the region immediately surrounding he @xmath235876 .
the equivalent widths of these emission lines are also comparable to other well - studied sne such as those in fig .
[ fig : sn1996cr - compare ] , suggesting at least that sn1996cr is not unusual .
we additionally included a few early - time light curves of typical sne in our spectral fitting , and found that aside from some degeneracy with our continuum model , such spectra are not generally compatible .
thus we argue that the majority of our line estimates , with the possible exception of h@xmath7 , are unlikely to be strongly contaminated . in the case of h@xmath7
, there are rare exceptions , such as the sn1994w and sn1994y , where early spectra display bright , relatively narrow h lines on top of blue continua from the outset ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . in these cases ,
the narrow lines are interpreted as arising from the sn shock transmitted into a dense , clumpy csm very close in to the sn ( @xmath110@xmath215 cm ) .
our radio and x - ray constraints at early times suggest this is an unlikely scenario for sn1996cr ( see @xmath8[radio_ltcrv ] and @xmath8[xray_ltcrv ] ) .
however , since we lack early optical spectra of , interpretation of the h@xmath7 emission should be viewed with some discretion . -0.2
cm we find exhibits both narrow ( 669@xmath2418 km s@xmath21 ) and broad ( 4077@xmath24398 km s@xmath21 ) components of h@xmath7 , with fluxes of ( [email protected])@xmath216 ergs s@xmath21 @xmath13 and ( [email protected])@xmath217 ergs s@xmath21 @xmath13 , respectively .
this portion of the spectrum is shown in detail in fig .
[ fig : sn1996cr - hacomplex ] .
the narrow line appears to be marginally resolved above our estimated instrumental resolution , although this fact should be regarded with some caution since the line is strongly blended with ; we await confirmation via high - resolution spectroscopic follow - up .
the flux found in our vlt spectrum is consistent to within errors with that measured from the _ hst _ f656n filter seven years earlier , indicating that the overlapping hii region still only contributes at most 30% to the total narrow - line flux above and 10% to the broad - line flux .
accounting for this contamination , we find corresponding extinction - corrected luminosities of ( [email protected])@xmath218 ergs s@xmath21 and ( [email protected])@xmath219 ergs s@xmath21 for the narrow and broad components , respectively . the energy generated in either the narrow or broad h@xmath7 lines
is significantly above that predicted by the decay of @xmath164co@xmath164fe at this late stage ( e.g. , * ? ? ?
* ) , and therefore must instead be related either to mechanical energy associated with the sn ejecta - wind interaction or potential light echoes .
the line width and systemic velocity of the narrow component argue for an origin beyond the blast wave , associated with either the shocked or unperturbed progenitor wind ( e.g. , * ? ? ? * ) .
if the line does arise from a dense , pre - sn , h - rich csm , it must have been photoionized and heated by uv and radiation from the massive progenitor and the initial shock breakout of the sn .
the fact that the h@xmath7 luminosity has remained relatively constant between the epochs of the vlt and _ hst _ observations suggests that this region is now being sustained by ionizing flux from the current ejecta - wind shock zone ( see @xmath8[xray_ltcrv ] ) , although again contaminating echoes from prior epochs is a concern here . for comparison , the h@xmath7 flux of other sne such as , , and decreased by at least an order of magnitude in @xmath14 yrs @xcite .
the h@xmath7 luminosity thus provides an important diagnostic on the total ionizing radiation that complements the x - ray observations , as this line arises as a result of recombination and collisional excitation .
if this h@xmath7 component is associated with csm interaction and is powered by photoionization , then the x - ray luminosity from the forward and reverse shocks should roughly be proportional to the h@xmath7 luminosity following eq . 3.7 in cf94 such that @xmath220 this implies @xmath221 ergs s@xmath21 , however , which is two orders of magnitude larger than the product of observed absorption - corrected luminosity and temperature .
this would argue for the soft x - rays from the reverse shock , which should dominate bolometrically , to perhaps be radiative and hence reprocessed by either unshocked low - density ejecta or high - density material in the cds ( this possibility is discussed further in @xmath8[oxygen_spectra_results ] ) .
alternatively , the narrow line could arise from shocks being driven into dense , slow - moving clumps associated with the progenitor wind .
this would more naturally explain the discrepancy in observed and estimated ionizing flux , in which case the line is instead powered by collisional excitation .
if the narrow h@xmath7 stems from the stellar wind , then the h@xmath7 luminosity can also be related to the mass of the unshocked wind as @xmath222 where @xmath223 is the radius in units of @xmath224 cm , @xmath225 is the wind velocity in units of 10 km s@xmath21 , and @xmath226 is the mass - loss rate in units of @xmath227 m@xmath4 yr@xmath21 @xcite .
we assume an emissivity based on case b recombination at a temperature of 10,000 k , which is in rough agreement with our extinction - corrected h@xmath7/h@xmath172/h@xmath173 ratios .
our h@xmath7 luminosity implies a mass - loss rate of . for a typical red supergiant lifetime of @xmath228 yr
, this implies that at least a few m@xmath4 of hydrogen - rich matter was cast off into the csm .
the high - velocity component of h@xmath7 , on the other hand , must be related directly to the blast wave in order to have achieved its present dispersion .
this component could be associated with the reverse shock or cds behind the reverse shock , perhaps from swept - up wind material or hydrogen that remained on the surface of the progenitor .
the broad - line h@xmath7 velocity is thought to provide a lower limit to the true shock velocity .
however , an observed disperion of @xmath14,000 km s@xmath21 is much lower than the 15,00040,000 km s@xmath21 initially expected for a typical core - collapse sn , and implies that the shock in has already slowed down substantially due to circumstellar interaction .
this observed velocity decrease of at least @xmath2410 implies a circumstellar density increase of @xmath216100 .
this line is almost certainly powered by collisional excitation associated with shocked ejecta .
we additionally see significant narrow - line emission in h@xmath172 and h@xmath173 , as well as marginally in h@xmath229 .
broad emission components could also exist for these lines ; however , for typical line ratios with respect to h@xmath7 they would be marginal at best .
additionally , such emission would be impossible to deblend from the [ ] emission complexes .
we find evidence for one narrow - line and four broad - line velocity components in the profiles of the [ ] @xmath295007,4959 , [ ] @xmath236300 , and [ ] @xmath237319 lines as shown in fig .
[ fig : sn1996cr - complex ] , with velocities of @xmath230 ( `` c1 '' ) , @xmath231 ( `` n '' , `` b '' ) , @xmath232 ( `` c2 '' ) , and @xmath233 km s@xmath21 ( `` c3 '' ) .
we note there is also marginal evidence for similar multi - component emission complexes associated with the [ ] @xmath235577 and [ ] @xmath234363 lines .
this is somewhat in contrast to the late - time spectra of many type ii sne such as , which show evidence for more irregular , `` clumpy '' , newly synthesized o complexes @xcite .
a few type ii sne have shown evidence for three - peaked profiles , which have either been ascribed to inhomogeneous , partially absorbed disk- or torus - shaped csms ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , or from line formation in the cds ( e.g. , * ? ? ?
in such cases , however , the profiles arise not from the o lines , but from h@xmath7 , , or ly@xmath7 ; this may indicate that is o - rich . nonetheless , the first scenario typically yields two relatively symmetric positive and negative velocity peaks indicative of emission from an equatorial ring , while a third central peak could emerge due to a population of dense clouds embedded in a much less dense csm . in the second scenario ,
the cds can only lead to such peaks under a very specific set of conditions such that it is very optically thick to such line emission , it is thin enough that the velocity gradient over the shell is small compared to the thermal velocity , and finally it has a macroscopic velocity larger than its thermal velocity .
the cds could be clumpy , lowering the filling factor and allowing some emission to escape . a final possibility for this velocity structure could arise from the sn blast wave impacting a dense shell produced by a wind - blown bubble .
in this case , several secondary shocks could separate the csm into multiple concentric shells ( e.g. , * ? ? ?
* ) , such that each peak may represent a different velocity shell .
the inner ( red - shifted ) shells would be preferentially absorbed compared to the outer ( blue - shifted ) shells , just as we observe . for
, we see a broad central peak ( `` b '' , fwhm@xmath1712795@xmath24174 km s@xmath21 ) and two roughly symmetric positive and negative peaks ( `` c1 '' and `` c2 '' , fwhm@xmath234 km s@xmath21 ) . the red peak is always fainter than the blue one , implying significant internal extinction .
interestingly , we also see a fourth , faster negative component ( `` c3 '' ) which does not have an obvious positive counterpart ; with our current sensitivity such a symmetric component ( i.e. , `` c4 ? '' ) must be at least a factor of 2.5 fainter ( ) , implying significant additional internal absorption above that of the lower velocity peaks , if it exists .
this `` c3 '' component could be indicative of further asymmetry in the structure of .
we need only look to the nearby , with its triple ring structure , for an example of such potential complexity @xcite . while the properties of the csm of appear to differ dramatically from those of , the spectral features hint at an environment that is equally dynamic and exotic .
another possibility is that the `` c3 '' component represents a light echo from an earlier epoch , where the o lines were shifted by a larger velocity and more heavily absorbed .
we reject this , however since it would require the `` c3 '' oxygen features to peak strongly in the blue part of the spectrum , while we detect them clearly throughout the spectrum .
notably , when we apply our five - component model to the o lines , we find that our constraints on [ ] @xmath236363 are consistent with zero flux if we fit [ ] @xmath236363 and [ ] @xmath236300 separately .
given the degeneracy between the potential broad components of these two lines though , this only provides weak @xmath235 constraints for the various components ( n:@xmath236 , b:@xmath237 , c1@xmath238 , c2@xmath239 , c3@xmath240 ) .
such ratios are all fully consistent with the normal transition probability of 0.3 . for the following ,
we fit only @xmath236300 and use its flux to estimate the total intensity @xmath241 for the various components . as such we may incur some small additional errors associated with flux contamination between different velocity components of the two lines .
we also caution that the complex which we attribute solely to [ ] @xmath237331 is likely in fact to be a blend of [ ] @xmath297319,7331 and [ ] @xmath297291,7325 . using our five - component model , however , we find that additional lines do not significantly improve @xmath242 .
the relative symmetry and constraints on line centers of the individual components imply that other lines do not contribute strongly to the overall emission , although notable small systematic residuals suggest minor blending is likely .
the overlapping broad - line o profiles seen in fig .
[ fig : sn1996cr - complex ] appear to be overestimating the valleys in between distinct component peaks , indicating that either the profiles are not represented well by gaussians or that p cygni - like absorption may be present but masked by the overall complexity of the emission . higher - resolution spectroscopy is needed to place better constraints on these potential emission and absorption contributions .
the fact that the various components show up in all of the o lines strongly argues for a scenario in which they arise from physically associated regions within the sn .
since we see several ionization states though , it is likely that there is strong stratification within these distinct spatial regions ( very likely due to an extended partially - ionized zone ; e.g. , cf94 ) and it is only appropriate to employ nebular - to - auroral line ratios for a given ionic species to estimate physical conditions .
however , our constraints on @xmath243 , @xmath244 , and @xmath245 are poor and do not provide physically interesting temperature and density limits .
table [ tab : ratios ] lists these oxygen ratio constraints .
notably , the relative observed strengths of the o lines are broadly consistent with the model predictions of cf94 ( see figs . 4 and 5 ) ,
whereby the modeled ejecta is photoionized by x - rays from the circumstellar interaction .
finally , there appears to be no sign of the @xmath237774 recombination line either at its systemic value or other velocities seen in the [ ] lines ( although telluric absorption does hinder assessment of the blue components ) .
this suggests that the narrow o resides in a physically distinct region separate from the narrow h@xmath7 , as one typically sees a @xmath246 ratio .
the broad , high - velocity o emission is not mirrored in any other elements , although there is possibly some complexity to the he emission ( see below ) , further arguing that it probably arises from an o - rich shell of processed gas associated with freely expanding ejecta .
sn1996cr shows evidence for four prominent , narrow he lines ( 906@xmath24101 km s@xmath21 ) : he i@xmath237065 , @xmath236678 , @xmath235876 , and @xmath234686 .
the narrow - line width is marginally broader than that of h@xmath7 , hinting that it may be spatially distinct .
the line ratios do not strongly constrain the electron temperature and density of the gas .
the emissivities of @xcite constrain the he gas to be at @xmath247 k and @xmath248 @xmath249 .
we see substantial broad residual emission around @xmath237065 and @xmath235876 which currently defies identification .
unlike the well - formed velocity components associated with the various o lines , there appears to be no spectral consistency between different he lines .
for instance , the emission around @xmath237065 appears box - like and quite strong relative to the systemic narrow line , while the emission around @xmath235876 appears more spread out and much fainter relative to the systemic narrow line .
thus if this broad emission is associated with he , it is likely present in several regions of the sn , and may be far more complex than h or o are .
there may be additional marginal broad features around @xmath234686 , while any potential features around @xmath236678 are masked by the strong , broad h@xmath7 emission .
the narrow he emission likely arises from the same regions of the progenitor wind as the narrow hydrogen emission , whereas the broad unidentified emission , if it comes from helium , could be swept up material or expelled clumps .
either way , its complex profile suggests it is much more irregularly distributed compared to h. we find evidence for strong , narrow n emission in the spectrum of in the form of both the nebular [ ] @xmath296583,6548 ( which appear as a resolvable blend with h@xmath7 ) and the auroral [ ] @xmath235755 lines .
the lines have a best - fitted fwhm of 741@xmath2424 km s@xmath21 .
the line - intensity ratio @xmath250 implies an electron density of @xmath251 @xmath249 for @xmath252 k @xcite .
the high density suggests that clumping of n is likely . as argued for the narrow h emission
, the majority of the emission must arise from the sn - csm interaction given the pre - sn imaging constraints .
the implied velocity of the lines suggest it is likely to have a circumstellar origin , while the strength of the emission hints at potential nitrogen enrichment , as has been notably seen in several other sne such as , , , , and @xcite .
the spectrum shows strong , narrow [ ] @xmath296731,6716 emission lines with a best - fitted fwhm of 741@xmath2424 km s@xmath21 , which again clearly must come from the sn - csm iteraction given the pre - sn imaging constraints .
this feature is relatively uncommon in late - time type ii sne spectra , although it is seen in several comparatively - aged type iin sne spectra such as and ( see fig .
[ fig : sn1996cr - compare ] ) .
the [ ] line - intensity ratio @xmath253 implies an electron density of @xmath254 @xmath249 @xcite .
we additionally see a hint of the [ ] @xmath294076,4069 doublet , although the significance of this feature is marginal and strongly affected by our estimated extinction value .
given the strong density contrasts , this component should reside in a different region from [ ] .
narrow ar emission is present in the spectrum of in the form of both the nebular [ ] @xmath297751,7135 and the auroral [ ] @xmath235192 lines .
our lower limit to the line intensity ratio @xmath255 , however , is too loose to place any useful constraint on the electron temperature or density of this component ( e.g. , * ? ? ?
we find significant high - ionization [ ] @xmath296087,5721 lines , but only a marginal [ ] @xmath235158 line .
additionally , [ ] @xmath236375 may be present , although it is currently lost amid the broad [ ] emission .
the lack of a strong @xmath235158 line constraint unfortunately leads to physically uninteresting electron - density limits ( e.g. , * ? ? ?
* ; * ? ? ?
we also see tentative features at @xmath235840 , @xmath235980 , @xmath237011 , and @xmath237785 which remain unidentified .
an expectation from the interaction between the blast wave and the csm is the production of copious non - thermal synchrotron emission from relativistic electrons and enhanced magnetic fields within the thin shell trailing the forward shock .
the radio emission should be proportional to the injection spectrum of electrons and various loss mechanisms , which are effectively determined by the composition and density of the csm . at early times , however , the radio emission can be absorbed via free - free ( ff ) or synchrotron self - absorption ( ssa ) depending on the magnetic field strength and csm density and structure .
then as the shock wave overtakes more and more of the csm , the radio emission becomes progressively less absorbed , leading to a characteristic frequency - dependent `` turn on '' first at higher frequencies and later at lower ones .
following the prescription of @xcite the optically thin radio luminosity of this shell can be described by @xmath256 where @xmath257 is the radio luminosity at frequency @xmath258 , @xmath259 is the thickness of the synchrotron emitting region at radius @xmath52 , @xmath54 is the strength of the magnetic field , the distribution of accelerated particles is assumed to take a power - law form of @xmath260 , and @xmath261 is the absorption opacity .
both @xmath54 and @xmath262 scale with the thermal pressure @xmath263 , which itself scales with the csm density @xmath264 of the synchrotron emitting region .
the opacity @xmath261 likewise scales with @xmath265 , although in the case of ff absorption it is proportional to the integrated density along the line of sight .
we assume here that all absorbing media are purely thermal , ionized hydrogen with opacities proportional to @xmath266 .
we contend that the ff opacity dominates over the ssa opacity given that ( 1 ) the shock velocity required by ssa , as implied by the radio luminosity and time of peak emission ( see , e.g. , * ? ? ?
* and references therein ) , is many times lower than either the optically inferred late - time shock velocity or the expected early - time shock velocity , and ( 2 ) the absorption does not deviate substantially from @xmath267 where it is possible to constrain it , as expected for ff . by comparison ,
ssa should scale as @xmath268 , which is too steep to fit any of our well - sampled epochs . in this case
, the ff opacity is given by @xmath269 where the emission measure , em , is given as @xmath270 thus the unabsorbed radio luminosity @xmath257 and ff opacity @xmath271 afford us two completely independent opportunities to trace the csm density profile .
-0.1 cm with the above in mind , we now examine the radio data for .
[ fig : sn1996cr - lc ] demonstrates that there are strong upper limits at four observable frequencies from the earliest time of explosion up through the end of 1996 .
these non - detections imply either strong early absorption or the presence of a low - density cavity . over the next year
, we see a dramatic rise in the radio emission , first at higher frequencies and later at lower ones . the inverted rise , however , is atypical and can not solely be the result of dwindling absorption , as is common in other rsne . to make this point clearer , we present the radio spectrum of for several epochs in fig . [
fig : sn1996cr - radiospec ] .
much of the radio data were taken separately , and thus a few of our adopted late - time epochs actually span a several - month window ; such an approach is validated by the fact that strong , rapid radio variability is not expected or observed at late times . from fig .
[ fig : sn1996cr - radiospec ] , we see that the data taken during the 2006 june and 2004 april epochs are only minimally absorbed and thus provide a solid constraint both on the intrinsic synchrotron spectral index ( @xmath272 ) and the ff absorption beyond the synchrotron - emitting region . using our model radio spectra to guide the eye ( described below ) , we see the presence of increasing absorption at the lowest frequencies as we progress back through the 1999 , 1997 , and 1996 epochs .
this is a natural consequence of the radio - emitting region overtaking the csm which absorbs it .
importantly , the observed spectral index between 4.8 ghz and 8.5 ghz does not change dramatically between 1996 and 2006 .
if we assume that the intrinsic spectral index is constant , this implies that the spectrum here is optically thin .
the 8.5 ghz emission is thus only modestly affected by ff absorption ( @xmath273% ) even at early epochs and can be considered a relatively robust tracer of the intrinsic radio luminosity at all times . as such
, the 8.5 ghz flux density appears to jump by a factor of @xmath0150 between 1996 august and 1997 december ( 515 days ) and an additional factor of @xmath01.33 between 1997 december and 2004 april ( 2384 days ) .
given that the radio - emitting region is spread over relatively large scales , the radio emission should vary smoothly in time .
empirically , we find that a broken power law , with indices @xmath274 and @xmath275 between the three highly constrained epochs above , provides an adequate fit to the data from other epochs aside from 1996 august ( @xmath276 ) , although there is clearly some degeneracy with other parameters such as @xmath7 and @xmath271 ( see fig .
[ fig : sn1996cr - radiospec ] ) .
this model should at least provide a qualitative understanding of the light curve during this overall period .
the early upper limits and dramatic increase in the radio luminosity imply a sharp rise in @xmath264 as well , highlighting the possible transition from a fast , sparse stellar wind to a slow , dense one .
such transitions typically lead to the formation of a wind - blown bubble ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and numerous researchers have explored the subsequent interaction between the complex csm associated with a bubble and the sn blast wave ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
quantitative constraints on the density are difficult to obtain , however , since the radio luminosity can not be related to the csm density in the standard manner ( i.e. , self - similar solutions such that @xmath277@xmath278 for @[email protected] ; * ? ? ?
* ) and requires hydrodynamical simulations to account for the interaction properly . our constraints on @xmath271 , on the other hand ,
are more straightforward to calculate . using the levenberg - marquardt least - squares method
we directly constrain @xmath279 to be [email protected] , [email protected] , [email protected] , [email protected] , and [email protected] for the 1996 december , 1997 december , 1999 june , 2004 april , and 2006 june epochs , respectively .
unfortunately , the 1996 august and 1997 june epochs only have 8.5 ghz measurements and do not provide strong constraints , although we are encouraged by the fact that the latter is consistent with our adopted light - curve model .
for the five epochs where we have good spectral constraints , we solve for @xmath280 explicitly in four contiguous shells such that @xmath281 where we have assumed @xmath282 and @xmath28310,000 k. the inner and outer radii @xmath284 and @xmath285 , in pc , are taken from @xmath286 ; and @xmath287 , where @xmath288 will range between [email protected] for most cases .
this would lead to estimated shell radii only @xmath273% larger and should highlight the relative quality of our radial constraints . ]
we note that the outermost radius is constrained to be @xmath289 cm on 2007 - 06 - 24 from our vlbi observation .
the to - be - determined @xmath290-dependence of @xmath291 is encompassed by @xmath292 .
the form of @xmath292 is iteratively determined by ensuring that it is consistent with the slope obtained between adjacent shells of @xmath293 .
we next turn to the evolution of the shock velocity , @xmath294 , which we need in order to determine the shell radii . from our optical spectrum ,
the ejecta ( as traced by the o lines ) and broad h@xmath7 line appear to be expanding at a variety of velocities between @xmath020005800 km s@xmath21 , with the symmetric red and blueshifted velocities of standing out as the most likely values tracing the blast wave .
we thus adopt 3400 km s@xmath21 for the current shock speed . clearly after the shock impacts a high - density region ,
its velocity will be reduced considerably . during its subsequent evolution in such a region ,
however , its velocity should evolve strongly , as can be seen from radio observations of sn1987a @xcite and simulations thereof @xcite .
therefore , without any further information , a first approximation is to assume a constant velocity comparable to that derived from the optical spectrum .
interestingly , the high x - ray temperature , if it is in thermal equilibrium , also yields an equivalent velocity of @xmath03400 km s@xmath21 , although here the velocity is almost certainly related to the reverse rather than forward shock and thus may simply be coincidence .
we therefore adopt an evolution for the blast wave velocity as follows .
the shock began with a high initial speed which it maintained for 12 yrs . by 1996
august it encountered a dense shell of wind - swept material , and the shock velocity dropped rapidly by a factor proportional to the square root of the density gradient it encountered .
this density gradient also initiated a rise in the radio luminosity by a factor of @xmath295 . beyond 1996 august
, we assume the velocity quickly arrived at a constant value of @xmath13400 km s@xmath21 .
these assumptions allow us to determine approximate radii of 0.054 pc , 0.055 pc , 0.059 pc , 0.064 pc , 0.081 pc , and 0.089 pc for the 1996 august , 1996 december , 1997 december , 1999 june , 2004 april , and 2006 june epochs , respectively .
if we take the radius on 1996 august as our shell impact date , then , depending on the 380 day ambiguity in the explosion date , the average initial shock velocity prior to this is constrained to be @xmath137,000139,000 km s@xmath21 ( [email protected]@xmath296 ) .
the lower value is in line with theoretical and observational expectations ( e.g. , * ? ? ?
* ) , while the upper value is probably unrealistic . such a high average shock speed would be difficult to maintain for more than several days in a typical or even underdense csm environment .
if the explosion did come later , then one or more of our assumptions above may be incorrect .
the late - time velocity of could be larger than our adopted value : the largest observed velocity in the optical spectrum is @xmath05,800 km s@xmath21 , for instance , and provides a plausible upper limit to the current shock speed ; adopting this value yields initial velocity constraints a factor of @xmath03 lower .
alternatively , the transition between initial and final velocities could have been gradual , taking many weeks to years to fully drop from one extreme to the other : the proposed csm shell of could have asymmetries akin to those seen in s ring , for instance , which are likely responsible for the onset of hotspots ( e.g. , * ? ? ?
* ) ; a long - lived velocity transition could again lower initial velocity constraints by a factor of a few .
finally , the vlbi radius constraint could be too large , due either to our adopted ring model or our adopted distance to sn1996cr : we note that uncertainties in our adopted vlbi model are estimated to be only @xmath1520% , while our adopted distance could vary by a factor of @xmath2972.7 at most . despite these potential problems ,
though , it is encouraging that our overall velocity model , assuming the lower initial shock velocity at least , is remarkably similar to the actual time evolution of the shock velocity in as deduced from radio observations ( e.g. , * ? ? ?
* ) . in sn1987a ,
the shock expanded at @xmath135,000 km s@xmath21 for its initial three years , before impacting with the hii region and slowing to @xmath14500 km s@xmath21 . using equation [ eq : rho_csm ] , we can now place constraints on @xmath291 in each shell of material overtaken between the five epochs as shown in fig . [
fig : sn1996cr - csmdensity ] .
since we do not know how far out the remainder of the free - free absorbing material extends into the csm , we do not include the last @xmath298 value in this analysis .
@xmath299 observations from the southern h - alpha sky survey atlas @xcite provide constraints on the emission measure toward the circinus galaxy of @xmath010,00025,000 pc @xmath300 , or equivalently @xmath301@xmath302 .
thus for any reasonable temperature of the warm interstellar medium ( ism ) , it appears that much of the remaining free - free absorption from our last measurement probably resides close to the sn in the csm itself .
future low - frequency radio measurements could constrain this potential decline in the absorption . oddly
, we find that the @xmath303 profile derived from the absorption , where @xmath3042.9 , seems at odds with the flat late - time evolution of the radio luminosity itself , which would naively imply @xmath305 . substantially ( e.g. , * ? ? ? * ) .
note , however , than any additional loss mechanism would have to leave the late - time spectrum relatively unperturbed . ]
the strong disconnect between the two may indicate that a large portion of the csm being overtaken is actually neutral , and thus not traced by the radio absorption .
since recombination is a strong function of density , this disconnect may further imply that there is significant clumping within the csm of .
although we are unable to constrain the density profile encountered by the blast wave at early times via @xmath271 , the early implied velocity drop argues for a density increase of @xmath2120 , while the radio luminosity jump supports a similar increase ( even larger if @xmath277@xmath278 holds ) .
thus it appears that we are seeing the transition between two distinctly different density regions separated by a dense shell , possibly associated with a wind - blown bubble ( e.g. , * ? ? ?
if we assume the sn exploded into a sparse stellar wind ( @xmath306 ) initially , then we might expect the shock to adhere to a self - similar solution prior to 1996 august ( i.e. , radio emission evolves as @xmath307@xmath308 ; cf94 , * ? ? ?
thus for explosion dates earlier than 1995 - 06 - 07 ( i.e. , day 98 for @xmath309@xmath1711995 - 02 - 28 ) , our 8.5 ghz upper limit would constrain the csm density in this region to be even lower still if no further ssa or ff absorption is present . a lower density , however , would imply a substantially larger density jump later on , which is difficult to reconcile with our current shock velocity evolution .
this suggests that either the sn went off after this date , the early density did not follow a @xmath310 profile , or there was additional heavy absorption that affected even the 8.5 ghz data . at the assumed radii
, the csm density of the shells equates to @xmath1 0.05 @xmath311 located between ( 1.72.8)@xmath312 cm .
additionally , the sn blast wave took only @xmath3132 yr to reach this high - density region , equating to @xmath1(12)@xmath314 yr in the progenitor time frame . for reasonable wind velocities
, this implies the region was formed a mere @xmath17014,000 yr prior to the explosion . for comparison
, the density for a constant progenitor wind ( i.e. , constant @xmath315 ) can be written as @xmath316 and a few lines of constant @xmath315 are plotted in fig .
[ fig : sn1996cr - csmdensity ] .
notably , the evolution of beyond @xmath110@xmath317 cm appears to be consistent with a constant wind of @xmath318@xmath319 , typical of a wolf - rayet ( wr ) stage .
however , it seems to have transitioned to a much sparser wind of @xmath320 before this ( i.e. , during its last many decades / centuries ) .
there are very few scenarios that can lead to such a csm density distribution , with an initial low density , followed by a large density jump , and then a region of decreasing density .
such a distribution is reminiscent , however , of a wind - blown bubble , wherein the interaction of a star s stellar wind with either a previous stage of the stellar wind or with a constant - density medium such as the ism , results in material being swept up into a thin dense shell between the two regions .
going outwards in radius , the bubble density distribution is described by @xcite : a region of freely expanding wind decreasing as @xmath310 , the wind termination shock , a constant - density region of shocked wind , the thin dense shell of swept - up ambient material and the outer medium , which could be another wind region with density decreasing as @xmath310 .
the circumstellar distribution derived from the radio observations seems to fit this well , with the high - density jump denoting the location of the dense shell , which we may not have seen from free - free absorption if the dense shell was not highly ionized .
interior to the shell is the low density shocked and unshocked wind , for which we only have upper limits .
exterior to the shell is the unshocked ambient medium , presumably also a wind , which we verify from our free - free absorption .
if true , this picture again indicates that the stellar parameters changed significantly just a few years to a few thousand years before the death of the star , and that this region was formed by wind - wind interaction .
although the shell is thin , it is extremely dense , and the shock colliding with this region could be slowed dramatically , and lead to an increase in radio and x - ray emission , as is observed .
sn1996cr can be compared to other observed young rsne , among which ( see * ? ? ?
* ; * ? ? ?
* for reviews ) is again a good starting point because of its rising radio light curve .
was detected at radio wavelengths when it first went off , but this initial outburst was very short - lived @xcite and attributed to shock acceleration of synchrotron - emitting electrons in the stellar wind close to the star at the time of the explosion @xcite .
radio emission was again detected from the supernova after @xmath13 yr in quiescence , at first rising in a dramatic fashion for several hundred days and increasing more or less monotonically thereafter ( see fig . [
fig : sn1996cr - lc ] ) . unlike the early absorption in ,
however , the observed spectral index of has remained relatively constant between @xmath321 to @xmath322 , indicating unabsorbed synchrotron emission at all times during this re - emergence @xcite .
the steady rise in is also much steeper than in and @xmath152 times fainter , indicating a markedly different csm - density distribution , and perhaps sharper density contrasts .
the constant monitoring at radio , optical , and x - ray wavelengths has illuminated the density structure surrounding and directly confirmed multiple evolutionary stages of mass loss @xcite .
unfortunately such consistent monitoring is lacking for , leaving us with only a crude understanding of its density structure to date .
clearly the csm distribution around appears to be much denser and more compact than , in addition to potentially showing a pile - up of material around @xmath323 cm .
we note that more modest structure in the csm density , perhaps from the wind - wind interaction of multiple evolutionary stages or inhomogeneities within a particular stage , has been observed in many well - sampled rsne light curves now , including @xcite , @xcite , @xcite , and @xcite , and demonstrates that perturbations from @xmath306 are probably commonplace . with finer sampling in the future
, we should be able to place stronger constraints on the outlying density structure in .
we turn next to the x - ray light curve of .
the circinus galaxy was observed with a variety of instruments as shown in fig .
[ fig : sn1996cr - lc ] and table [ tab : data_xray ] . from these
, we find that exhibits strong , atypical temporal evolution .
unlike most sne , is not detected at early times despite relatively frequent and sensitive observations . since thermal x - ray emission effectively traces the density of the csm , these early upper limits , especially when compared to the detections later on ,
again indicate the presence of either strong early absorption ( @xmath324 @xmath13 ) or a low - density cavity immediately surrounding the progenitor .
our radio constraints suggest in fact that both are present and likely to affect the x - ray emission : the mass associated with the ff absorption is already marginally enough to obscure an extrapolation of the detected x - ray slope , while the lower density implied by the radio luminosity jump would also result in decreased early x - ray emission . when finally detected on 2000 - 01 - 16 , the and fluxes are best - fitted as @xmath325 @xmath326 and @xmath327 , respectively .
the slopes are consistent with each other to within the errors , suggesting that any apparent emerging soft component is marginal and that the detected portion of the rise is likely to be optically thin .
this strong , continued rise is in stark contrast to the evolution of nearly all other sne detected to date , which are theoretically expected to decline as @xmath325 @xmath328@xmath329 ( e.g. , cf94 ) and observationally follow suit albeit with more scatter ( e.g. , see fig . 2 of * ?
sustained increases in the have only been observed for @xcite , @xcite , and perhaps marginally for ( based upon a single early upper limit ; * ? ? ?
the x - ray luminosity of the forward and reverse shocks at a particular radius @xmath290 is given by @xmath330 where the index @xmath331 refers to values for either the forward or reverse shock and @xmath332 is the temperature - dependent cooling rate @xcite .
the detection of strong he - like fe and si emission lines in the x - ray spectra @xcite suggests the presence of heavy elements in the x - ray emitting material , and therefore that the x - ray emission stems from the shocked ejecta material found behind the reverse shock .
because the progenitor csm distribution likely includes the presence of pre - existing shocks and density discontinuities associated with the wind - blown bubble ( see @xmath8[radio_ltcrv ] ) , the interaction of the sn blast wave with this medium is complex . due to the presence of various discontinuities , the standard model ( e.g *
* ; cf94 ) of the interaction of a sn blast wave with a power - law ambient density medium is not applicable , and we must rely on detailed hydrodynamical models to infer the specific mass - loss properties of the progenitor ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . furthermore the csm may be at least partially photoionized , as is thought to be the case for ( e.g. , * ? ? ? * ) .
as shown in the previous section for a wind - blown bubble , even if the progenitor star had a constant wind mass - loss rate and velocity ( which is not likely ) , the structure of the csm can vary substantially with radius .
thus we can not assume that the csm density directly translates into the wind parameters .
nonetheless , we can make some qualitative arguments based on equation [ eq : lx ] . between days 18004400 , we find that the intrinsic x - ray luminosity scales almost linearly with time
. the radius over this time period should increase by a factor of @xmath01.6 , depending on the exact nature of the ejecta density profile and the csm .
the width of the emitting region can be approximated to be about 0.1 times the radius ( e.g. , cf94 ) , resulting in an overall volume change by a factor of @xmath2975 . from the x - ray spectrum
, we do not see any significant changes in the temperature or absorption , indicating that @xmath333 remains approximately constant .
if the x - ray emission was arising in the shocked ambient medium , the density in equation [ eq : lx ] must decrease by a factor of @xmath02.2 , or equivalently be @xmath334 to produce the observed dependence .
this is significantly different from either a constant density or the @xmath335 dependence expected for a stellar wind with constant parameters .
furthermore , for a density profile decreasing as @xmath336 the velocity would be expected to decrease with time , and therefore the temperature ( which is proportional to @xmath337 ) would drop accordingly .
the fact that this is not seen makes it unlikely that the x - ray emission is arising from the shocked wind ( see fig .
[ fig : sn_diagram ] ) .
we instead contend that the emission is arising from shocked ejecta behind ( in a lagrangian sense ) the reverse shock .
this scenario is strengthened by the presence of heavy element emission lines in the x - ray spectrum ( cf .
fig . 6 of * ? ? ?
* ) , since such elements are far more likely to reside in the processed ejecta than in the progenitor s stellar wind .
if the sn ejecta did collide with a dense shell of csm from a pre - existing wind - blown bubble , then the forward shock would advance quite slowly , and the bulk of the x - ray emission would arise from the reverse - shocked material
. the post - shock temperature just behind the reverse shock could be quite high ( @xmath2@xmath338 k ) if the density were low technically the density could drop to zero for expansion in a constant - density medium . the electron temperature , however , might be an order of magnitude lower if the electrons and ions have not yet reached equilibrium .
the temperature would not change appreciably over this period , consistent with our observational constraints .
the velocity of @xmath03400 km s@xmath21 inferred from the x - ray temperature in this case would refer to the electron temperature in the reverse shock , and thus the actual shock velocity , as traced by the ion temperature , could be considerably higher .
note that a similar linearly increasing light curve profile was found for the hard x - ray emission from sn1987a @xcite .
@xcite has carried out numerical simulations that suggest that this emission arises from the reverse shocked ejecta , as had been postulated by @xcite .
one puzzle which remains , though , is how the strong the h@xmath7 and oxygen emission - line luminosities are powered .
this discrepancy could be reconciled if there were holes in the cds , such that a small fraction of the x - ray emission could escape while the rest went into exciting the strong optical line emission .
we can again gain some further insight by comparing to other sne . as stated above , has the most similarities .
like , was undetected at x - ray wavelengths when it first went off , thereby excluding an explosion directly into a slow , dense wind that is characteristic of a red supergiant .
notably , was detected 130 days after explosion @xcite , although the x - ray emission here was attributed solely to radioactivity and appeared to rise and fall over the following year .
after this brief period , the x - ray emission went undetected for another @xmath0 3 yrs before resurfacing above the noise @xcite . since this time
, the x - ray emission has continued to rise as shown in fig .
[ fig : sn1996cr - lc ] , with its light curve well - modeled by a broken power law @xcite .
the reappearance of the emission at about 3 years was postulated as due to the interaction of the sn blast wave with a region of ionized wind material from the progenitor @xcite .
they suggested that this region is material photoionized from the dense circumstellar shell that surrounds , formed as a result of the interaction of the progenitor blue supergiant wind with a pre - existing red supergiant wind .
although the evolution of the star may be different , a physical scenario broadly like that of , with a region formed by mass - loss from a progenitor star , certainly seems plausible for , although the contrasting light - curve slopes , luminosities , and radial distances clearly highlight the different csm density structure .
the only other sn to exhibit a rise at x - ray wavelengths is , a peculiar typeib .
the light curve of has been ascribed to the interaction of the blast wave with a dense shell of material left over from an lbv - like outburst of the sn progenitor @xmath12 years prior to the explosion @xcite .
the x - ray spectrum for , moreover , is modeled by a hot thermal plasma ( @xmath128 kev ) similar to the early phase of @xcite but substantially hotter than ( @xmath339 kev with an emerging @xmath340 kev component ) , and implies that the conditions within the x - ray - emitting region are somewhat different .
the hotter temperature could indicate a higher shock velocity , or a larger level of equilibration between the electrons and ions .
the x - ray spectrum of is additionally absorbed below @xmath12 kev .
while some fraction of the substantial non - galactic column density could be due to self - absorption , we consider this to be minimal given that other point sources in the circinus galaxy have similar derived column densities @xcite .
the rapid evolution we see here for has been noted in a few other objects , such as @xcite and @xcite . such an evolution has also been invoked to explain the enormous energy output from @xcite , although the lack of obvious csm - interaction measures at x - ray , h@xmath7 , and radio wavelengths here implies that perhaps something more extreme is going on ( e.g. , * ? ? ?
unfortunately , for many of the older , classic x - ray emitting sne such as @xcite , @xcite , and @xcite , early - time behavior of the x - ray emission remains completely unknown and thus the frequency and duration of sudden evolutionary changes in sne progenitors is still an open question .
the x - ray evolution of may allow us to explain the relatively flat late - time x - ray light curves for objects like and , for instance .
we have confirmed as one of the nearest , and subsequently x - ray and radio - brightest , sne to date .
archival optical imaging data have constrained the explosion date to within @xmath01 yr prior to 1996 - 03 - 16 , while a new vlt spectrum has identified it as a type iin sn .
the multi - wavelength constraints detailed above point to a rich progenitor history .
as is the case for most type iin sne , appears to have originated from a massive star that shed its outer layers at some late evolutionary stage , which now act as the target for the outgoing sn shock .
the optical light curve hints at a mild re - brightening at late times , possibly driven by line emission , or by x - ray photons being downscattered and being emitted as optical photons . from the recent optical spectrum , the strong narrow h emission points to a significant accumulation of mass in the csm prior to the sn .
we are additionally able to resolve the complex o emission into a number of distinct high - velocity components with fwhm@xmath120003000 km s@xmath21 .
the origin of such distinct o complexes remains somewhat unclear , perhaps arising from one or more of the following : structural asymmetries , differential absorption , or multiple concentric shocks .
the and radio emission allow us to quantify the csm density , tracing out two apparently distinct regions , which can not be reconciled with a single constant - wind model . from the radio and x - ray upper limits , we infer that the inner region ( @xmath341 cm ) is likely quite sparse .
the outer region [ @xmath342@xmath343 cm ] , as traced by the early radio absorption , radio luminosity jump , and x - ray emission , is not well constrained , and is likely associated with the interaction region between two distinct progenitor winds , which culminates in a thin , dense wind - swept shell .
this second region appears relatively dense ( @xmath344 @xmath249 ) , potentially consistent with a slow - velocity wr - like wind , although the csm density profile traced by the radio absorption ( @xmath345 ) naively seems at odds with the implied csm density profiles from the radio and x - ray luminosities ( @xmath346constant or @xmath336 , respectively ) .
a more detailed understanding of the csm awaits sophisticated hydrodynamical models .
the most plausible explanation is one where the progenitor changed evolutionary states just prior to explosion , likely transitioning from a wr - like wind to an even faster , less dense wind , resulting in portions of the previous csm being swept - up to form a wind - blown cavity surrounded by a dense shell .
such a scenario has been recently proposed for the peculiar @xcite .
the radio and x - ray emission are constant or still rising , demonstrating that the blast wave is still making its way through the dense , swept - up shell at the edge of the cavity .
the collision of the blast wave with the shell would result in a transmitted shock expanding into the shell , and a reflected shock expanding back into the sn ejecta .
this reflected shock should be distinguished initially from the reverse shock of the sn expanding into the ejecta , although it will eventually overtake the reverse shock .
the overall effect could be one of several shocks and rarefaction waves travelling back and forth through the ejecta .
a spectrum would reveal several different velocity structures , perhaps as we observe in the o lines .
the reason why it is seen mainly in o is unclear , but it could indicate that this is an o - rich remnant , similar to casa for example ( see * ? ? ? * and references therein ) . or perhaps it is just that the reverse shock has not penetrated very far into the ejecta , so that only the o - rich layers have been shocked .
we note that if the shock is indeed progressing into shell of constant density , then the resulting density gradients would not lead to the formation of a cds at all ( e.g. , see fig .
14 of * ? ? ?
if so , then the observed o velocity components may instead arise from a disk or ring prehaps similar to sn1987a .
once the blast wave makes its way through the dense shell , it will emerge from the other side , presumably as a weak shock .
this should encounter a relatively pristine wind and start to decline in both the radio and x - ray bands .
further regular monitoring of should reveal this .
once this happens , we will be able to strengthen many of our current constraints and assumptions .
we have compared to the late - time behavior of , which is the only other sn with a documented dramatic increase at and radio wavelengths
. is likely propagating through its csm in a manner similar to expanding into its region and circumstellar ring ( e.g. , * ? ? ?
the and radio luminosities of , however , are several orders of magnitude brighter and comparable to typical type iin sne such as and .
this is presumably because the swept - up csm is more compact , and perhaps , unlike , there is no intervening high - density material encountered which slows down the ejecta until they reach a dense shell , so the density jump is higher and subsequently more dramatic . as we follow over the next few decades , we anticipate that it will transition into an evolution similar to other sne .
it is intriguing to think that might represent a sort of `` bridge '' object between the extremes of and these more luminous type ii sne .
although plagued by small number statistics , the fact that two ( although perhaps more if we count and ) of the @xmath1510 closest sne to have exploded in the past four decades show evidence for wind - blown bubbles implies that the phenomenon is common , and may indicate that current sne searches are somehow biased against them at larger distances .
systematic constraints exist only for type ibc sne @xcite , where late - time radio observations indicate that only @xmath12% of initially undetected targets show late - time detections , while @xmath140% of detections exhibit abrupt light - curve variations .
the large mass associated with the pre - sn csm surrounding could have been the result of explosive mass loss thousands of years before the sn explosion ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , qualitatively similar to @xcite , although the limited and radio data point to a relatively smooth density profile , which implies a more gradual process was at work .
numerous other sne including @xcite , @xcite , @xcite , and @xcite all show signs of dense , enriched csm likely produced by giant mass - loss events just prior to their sne , implying that this form of mass - loss is relatively common .
lbvs have often been invoked to account for such episodes of extreme mass - loss ( e.g. , * ? ? ?
* ; * ? ? ?
indeed a giant outburst was seen in the case of just 2 years prior to core - collapse @xcite .
again it is unclear whether the explosive nature of lbvs can be reconciled with the relatively smooth mass - loss / density gradients observed for .
our analyses clearly demonstrate that is a unique and compelling target . its proximity and flux enable a wealth of observations that would be otherwise impractical for typical sne .
we have reported on initial findings of a vlbi campaign underway to resolve spatially at several radio wavelengths , eventually enabling studies of its morphology and temporal expansion .
this will provide a more coherent picture , allowing us to set strong limits on the expansion velocity from the resolved size and potentially confirm the implied asymmetries we infer from the optical spectrum .
follow - up _ hst _ imaging will provide constraints on any potential light echoes in the vicinity of , possibly allowing refined estimates of its explosion date , distance , and outlying csm structure . likewise , follow - up observations will hopefully elucidate the nature of the strong continued rise in the band and better constrain properties of the reverse shock region . notably , is bright enough to be efficiently monitored using the _ chandra _ hetgs , enabling high - resolution studies of its numerous emission lines and their potential evolution ; an in - depth discussion of the current spectral constraints , including a composite _ chandra _ hetgs spectrum from the data presented here , are forthcoming ( f. bauer et al .
, in preparation )
. additionally , future optical spectroscopic studies should allow investigations into the temporal behavior of complex emission lines found in our discovery spectrum .
infrared imaging and spectroscopy will help address the potential for dust formation ; notably sne are thought to contribute strongly to the overall dust content of galaxies .
looking ahead to the near future , if remains relatively bright at millimeter wavelengths , we anticipate that observations with alma will provide a substantial leap in terms of understanding the structure and composition of and its relation to both and more typical sne .
we thank bjorn emonts and bryan gaensler for help in reducing archival atca observations , roger chevalier , claes frannson , dick mccray , kurt weiler , and christopher stockdale for useful discussions about the nature of the multi - wavelength emission , baerbel koribalski for increasing our awareness of australian online archives and access to proprietary atca observations , ernesto oliva and alessandro marconi for access to reduced ntt data , bruno leibundgut and rob fesen for their spectra of and , philip edwards , steven tingay , and tasso tzioumis for support of the vlbi observations , kevin hurley for directing us to revised ipn results which allowed us to reject an otherwise tentative grb identification , john raymond for helping to improve our emission - line tables , and the anonymous referee for useful comments that improved the content and presentation of the paper .
we gratefully acknowledge the financial support of chandra postdoctoral fellowship award pf4 - 50032 ( feb ) , nsf award ast-0319261 ( vvd ) , nasa stsci grant hst - ar-10649 ( vvd ) , nasa ltsa grant nag5 - 13035 ( wnb ) , the leverhulme trust ( sjs ) and the esf euryi scheme ( sjs ) . , b. m. , staveley - smith , l. , manchester , r. n. , kesteven , m. j. , ball , l. , & tzioumis , a. k. 2007 , in american institute of physics conference series , eds .
s. immler & r. mccray , vol .
937 of _ american institute of physics conference series _ , 86 | we report on new vlt optical spectroscopic and multi - wavelength archival observations of , a previously identified ultraluminous source known as circinus galaxy x-2 .
the spectrum of the optical counterpart confirms as a _
bona fide _
type iin sn , in accordance with its tentative sn classification at wavelengths .
is one of the closest sne ( @xmath03.8 mpc ) in the last several decades and in terms of flux ranks among the brightest radio and sne ever detected .
optical imaging from the anglo - australian telescope archive allows us to isolate the explosion date to between 1995 - 02 - 28 and 1996 - 03 - 16 , while the wealth of optical , , and radio observations that exist for this source provide relatively detailed constraints on its post - explosion expansion and progenitor history , including an preliminary angular size constraint from vlbi .
archival and radio data imply that the progenitor of evacuated a large cavity just prior to exploding via either a sped - up wind or a pre - sn explosion .
the blast wave likely spent @xmath112 yrs in relatively uninhibited expansion before eventually striking the dense circumstellar material which surrounds to become a prodigious and radio emitter .
the and radio emission , which trace the progenitor mass - loss rate , have respectively risen by a factor of @xmath22 and remained roughly constant over the past @xmath3 yr . this behavior is reminiscent of the late rise of , but three orders of magnitude more luminous and much more rapid to onset . may likewise provide us with a younger example of and , both of which exhibit flat x - ray evolution at late times .
the optical spectrum suggests that the progenitor was a massive star that shed some of its outer envelope ( many m@xmath4 ) prior to explosion , while the complex oxygen line emission hints at a possible concentric shell or ring - like structure . taken together
, this implies that a substantial fraction of the closest sne observed in the last several decades have occurred in wind - blown bubbles , and argues for the phenomena being widespread . |
in this paper , we explore the motion of a whip , modeled as an inextensible string .
we prove that the partial differential equation describing this motion is locally well - posed in certain weighted sobolev spaces .
in addition , we are interested in the motion of a chain , modeled as a coupled system of @xmath6 pendula , in the limit as @xmath6 approaches infinity .
we show that the motion of the chain converges to that of the whip .
although the equations of motion are well - known and have been studied by many authors , there are few results known about the general existence and uniqueness problem .
reeken @xcite @xcite proved local existence and uniqueness for the _ infinite _ string in @xmath7 with gravity and initial data sufficiently close ( in @xmath8 ) to the vertical solution , but aside from this , we know of no other existence result . in the current paper
we prove a local well - posedness theorem for arbitrary initial data for the finite string .
one reason this problem is somewhat complicated is that the equation of motion is hyperbolic , nonlinear , nonlocal , degenerate on a spatial boundary , and possibly even elliptic under certain conditions .
if @xmath9 \to \mathbb{r}^d$ ] describes the position @xmath10 of the whip , then one can derive that the equation of motion in the absence of gravity and under the inextensibility constraint @xmath11 is @xmath12 incorporating gravity introduces some complications ; to keep things as simple as possible , we will neglect it .
equation is a standard wave equation ; however , the tension @xmath13 is determined nonlocally , as a consequence of the inextensibility constraint , by the ordinary differential equation @xmath14 with one end fixed and one end free , the boundary conditions are @xmath15 and @xmath16 , along with the compatibility condition @xmath17 .
we use the energy @xmath18 and show that for small time we have local existence and uniqueness in the space for which the energy @xmath19 is bounded .
precisely , for any nonnegative integer @xmath20 , define @xmath21 $ ] to be the space of functions @xmath22\to\mathbb{r}^d$ ] such that @xmath23 is bounded ; then @xmath24 .
we prove the following result : [ theorem1 ] suppose @xmath25\to \mathbb{r}^d$ ] and @xmath26\to \mathbb{r}^d$ ] are restrictions of functions on @xmath27 $ ] satisfying the oddness condition @xmath28 and @xmath29 , and that they have bounded weighted sobolev norms , @xmath30 and @xmath31 .
suppose that in addition we have @xmath32$.}\ ] ] then there is a @xmath33 such that there is a unique solution @xmath34 of the equation in @xmath35 , n_4[0,1])\cap w^{1,\infty}([0,t ] , n_3[0,1])$ ] satisfying @xmath36 and @xmath37 .
we prove this by showing that the corresponding discrete energy @xmath38 for the chain with @xmath6 links is uniformly bounded for small time , independently of @xmath6 .
the solution is then a weak- * limit of the chain solutions in @xmath39 , which converges strongly in @xmath40 and hence in @xmath41 .
one could prove this more directly using a galerkin method , but the present technique allows us to simultaneously discuss convergence of the discrete approximation .
all the higher energies @xmath42 can be bounded in terms of @xmath43 , so that @xmath44 initial conditions yield @xmath44 solutions for short time . as a consequence ,
we derive a simple global existence criterion : if the initial conditions are @xmath44 functions , then a @xmath44 solution exists on @xmath45 $ ] if and only if @xmath43 is uniformly bounded on @xmath45 $ ] .
of course , one expects blowup of the whip equation , at least for some initial data , since the whole purpose of a whip is to construct the initial condition so that the velocity of the free end approaches infinity after a short time .
see mcmillen and goriely @xcite for a discussion of such issues ; although our model neglects some of the phenomena they consider , one expects that the situations are similar in many ways .
for the heuristics of blowup in our situation , see thess et al .
the simplest blowup mechanism appears to be the closing off of a loop along the whip ; as a loop shrinks , there appears a kink in the whip , representing blowup of both the curvature and the angular velocity .
the paper is organized as follows . in section [ basicequationssection ] , we discuss the equations for the whip and derive the corresponding equations for a chain with @xmath6 links , in terms of difference operators , emphasizing the role of odd and even extensions in order to get the fixed endpoint conditions satisfied automatically . in section [ greenfunctionsection ]
we discuss the solution of the tension equation in terms of a green function , showing that the tension is positive except at @xmath46 and deriving a similar result for the chain .
we also derive sharp upper and lower estimates for the green function . in section [ weightedsobolevsection ]
we explain why we need weighted energies , and we derive the analogues of the sobolev and poincar inequalities for weighted norms , which are used throughout the rest of the paper . in section [ tensionboundssection ] ,
we give estimates for the tension @xmath13 in terms of @xmath34 and @xmath47 .
for the @xmath48 norms of @xmath13 we use the bounds on the green function ; for higher derivatives we bound the weighted sobolev norms of @xmath13 in terms of those of @xmath34 and @xmath47 .
these bounds are used in section [ energyestimatesection ] to derive the main energy estimate , to bound the time derivative of one energy in terms of another energy .
section [ existenceuniquenesssection ] contains the proof of theorem [ theorem1 ] .
uniqueness is proved using a low - order estimate for the difference of two solutions .
finally in section [ futureresearchsection ] we discuss related open problems . to make the paper a bit easier to read ,
we have moved all of the longer proofs into an appendix .
victor yudovich found several results on this problem , although he did not publish anything on it to my knowledge .
i learned of this problem from alexander shnirelman , and i would like to thank him for many useful discussions about it .
the study of the inextensible string is one of the oldest applications of calculus , going back to galileo , and yet it is still being studied to this day .
one is especially concerned about kinks in the solution and what the appropriate jump conditions should be ; authors such as oreilly and varadi @xcite , serre @xcite , and reeken @xcite have discussed these issues in detail from differing points of view .
the first problem to be studied was finding the shape of a hanging chain , first solved incorrectly by galileo and then correctly by leibniz and bernoulli , one of the first major applications of the calculus of variations .
the shape of small - magnitude vibrations of a chain hanging straight down ( in a linear approximation ) goes back to the bernoullis and euler @xcite , and is taught in textbooks today as an example of bessel functions ; see johnson @xcite and schagerl - berger @xcite for related problems .
kolodner @xcite , dickey @xcite , luning - perry @xcite , and allen - schmidt @xcite studied the problem of a uniformly rotating inextensible string , one of the few other problems that can be solved more or less exactly .
burchard and thomas @xcite obtained a local well - posedness result for the related problem of inextensible elastica , in which there is a potential energy term reflecting a resistance to bending ; however it is not clear whether the solutions are preserved in the limit as the potential term goes to zero , so this result does not help in the present situation .
many authors have studied the problem of a vertically folded chain falling from rest ; this is a classical problem that appears in several textbooks ( @xcite , @xcite , @xcite , and @xcite ) . in recent years
the problem has been debated in the physics literature , in particular the issue of whether energy is conserved and whether the tip of the chain falls at an acceleration equal to gravity or faster ( @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite ) .
see wong - yasui @xcite or mcmillen @xcite for a good survey of the literature .
mcmillen and goriely ( @xcite and @xcite ) studied a tapered whip theoretically , numerically , and experimentally , showing that the crack comes not from the tip but rather from a loop that straightens itself out .
they use a different model , however , in which the tension depends locally on the configuration .
thess et al .
@xcite studied the blowup problem for the closed inextensible string , especially as a model of the blowup problem for the euler equations for a 3d ideal fluid .
they found evidence of blowup from loops closing off , showing numerically that @xmath49 and @xmath50 , where @xmath51 is the blowup time .
in this section , we present the equations for both whips and chains , assuming no external forces .
our boundary conditions come from the assumption that one end of the whip or chain is held fixed at the origin , while the other end is free .
we describe the whip as a function @xmath52\times [ 0,1 ] \to \mathbb{r}^d$ ] , and describe the chain as a sequence of functions @xmath53\to \mathbb{r}^d$ ] for @xmath54 .
our formulas simplify if we assume the fixed point occurs at @xmath55 , i.e. , @xmath1 for all @xmath56 ; for the chain , we assume @xmath57 for all @xmath56 . we will just present the equations here with a sketch of the derivation ; the reader may refer to @xcite for a detailed derivation and discussion .
schagerl et al .
@xcite and thess et al .
@xcite also present derivations from minimum principles : the basic idea is to find a critical point of the action @xmath58 subject to the constraint @xmath59 .
a variation @xmath60 must satisfy @xmath61 and @xmath62 , and if @xmath34 is a critical point of the action , then @xmath63 for all such @xmath60 .
integrating by parts , we conclude that a critical @xmath34 must satisfy the evolution equation @xmath64 for some function @xmath13 , where @xmath2 . differentiating @xmath65 twice with respect to @xmath56
, we find that @xmath13 is determined by the following boundary - value problem for an ordinary differential equation ( for each fixed @xmath56 ) : @xmath66 the boundary conditions are compatible with the evolution equation as long as @xmath34 can be extended to an odd function through @xmath55 ; in that case @xmath13 can be extended to an even function through @xmath55 , which is where we get the extra boundary condition @xmath67
. see figure [ oddcurve ] .
oddness and evenness give us the correct boundary conditions for all higher derivatives of @xmath34 and @xmath13 at @xmath55 , which is crucial for the a priori estimates .
furthermore there is a discrete analogue of oddness and evenness for the chain which both simplifies the equations and helps greatly in defining the higher discrete energies . , while the fixed end is at @xmath55 .
we imagine the curve extending smoothly through the origin to @xmath68 through an odd reflection such that @xmath69 .
under such an extension , the tension extends to a smooth function satisfying @xmath70 . ]
a sometimes simpler way of dealing with the constraint @xmath71 is to consider @xmath72 as a curve on the unit sphere in @xmath73 .
for simplicity we will assume @xmath74 when doing this , although the technique works in spherical coordinates in any dimension .
we write @xmath75 a straightforward computation verifies that becomes @xmath76 while becomes @xmath77 the fact that @xmath34 is odd through @xmath55 forces us to have @xmath78 even through @xmath55 , so the boundary condition on is @xmath79 .
we could work out all the estimates directly in terms of the system , but the discrete versions of these equations are substantially more complicated than the discrete versions of , even when @xmath74 . if @xmath80 is strictly positive for @xmath81 ,
then equation is a hyperbolic equation with a parabolic degeneracy at @xmath46 ( since we must have @xmath2 ) . as such ,
the only condition necessary to impose at @xmath46 is that @xmath82 remain finite .
we point out that equation can not be an ordinary differential equation on any infinite - dimensional sobolev manifold : the right side is obviously an unbounded operator even in the simplest case .
hence we can not hope to prove existence and uniqueness using the techniques of picard iteration on an infinite - dimensional space , as in ebin - marsden @xcite .
instead we will work directly with the partial differential equation using energy estimates .
we now derive the equations for the finite model , consisting of @xmath83 particles in @xmath84 , each of mass @xmath85 , one of which is held fixed .
the particles are assumed to be joined by rigid links of length @xmath85 , whose mass is negligible .
the position of the @xmath86 particle is @xmath87 for @xmath54 ; we assume the fixed end is the @xmath88 particle , so that @xmath89 for all time . , but our choice makes the tensions @xmath90 proportional to @xmath91 rather than @xmath92 , simplifying some formulas . ] the configuration space is thus homeomorphic to @xmath93 , and is naturally embedded in @xmath94 .
the kinetic energy in @xmath94 is @xmath95 in addition the constraints are given by @xmath96 stationary points of the constrained action satisfy the equations of motion @xmath97 for some lagrange multipliers @xmath98 .
more explicitly , we have @xmath99 for @xmath100 .
the scaling by @xmath101 is chosen so that @xmath102 converges to a function @xmath80 as @xmath103 .
the numbers @xmath13 physically represent the tensions in each link .
we set @xmath104 so the same equation is valid when @xmath105 .
the constraint equations determine the @xmath13 .
differentiating twice with respect to time and using , we get @xmath106 for @xmath107 ( again using @xmath104 ) , while for @xmath108 we get ( using @xmath109 ) @xmath110 we note that if @xmath111 where @xmath112\to\mathbb{r}^d$ ] and @xmath113\to\mathbb{r}$ ] are @xmath44 , then as @xmath114 , the formal limit of is and the formal limit of is .
( note that this discretization of @xmath34 ensures that @xmath115 for all @xmath116 , since @xmath117 for all @xmath118 .
we will refine this in section [ interpolation ] . )
if @xmath119 and @xmath2 , then this choice also gives @xmath120 and @xmath121 , as desired .
hence the chain equations and form a discretization of the whip equations and which conserves energy as well as preserving the geometry .
the analysis of the chain equations becomes much simpler if we can avoid using separate equations for the boundary terms . an easy way to do
this is to extend @xmath122 and @xmath90 beyond @xmath108 by demanding that @xmath34 be odd through @xmath123 and that @xmath13 be even , which is exactly what we had to do for the whip in section [ whipsection ] .
so for @xmath124 we set @xmath125 then it is easy to see that the evolution equation still holds for the fixed point at @xmath123 and that for @xmath108 yields the tension boundary condition .
a further simplification comes from using difference operators .
( see for example @xcite . )
first recall that for a sequence @xmath126 defined on some subset of @xmath127 , the ( forward ) shift operator @xmath128 is given by @xmath129 .
the backward shift is denoted by @xmath130 , so that @xmath131 , and powers of @xmath128 signify composition .
we define the ( forward ) difference operator @xmath132 by @xmath133,\ ] ] so that if @xmath134 denotes the identity operator , then @xmath135 .
it is also sometimes convenient to work with the backward difference operator @xmath136 , defined by @xmath137 $ ] , so that @xmath138 . in this notation for the forward difference and @xmath139 for the backward difference ; we use @xmath132 and @xmath136 instead to avoid confusion with the laplacian on smooth functions , and since our rescaled version is not standard .
we prefer the rescaling since if the sequence @xmath140 converges to a smooth function @xmath141 as @xmath114 , then @xmath142 converges to @xmath143 . ] equations and become @xmath144 where the equations are valid when any subscript @xmath145 is placed on all the terms simultaneously .
we can thus write all the discrete equations without specific reference to subscripts , which simplifies the notation .
the following formulas will be useful when working with difference operators and sums : both follow from the simplest product formula @xmath146 .
@xmath147 we can rewrite in a more useful form , solving for the second difference @xmath148 in terms of everything else , using @xmath149 to simplify the terms .
we obtain @xmath150 and the resemblance to the continuous version is obvious .
at each fixed time , equation is a linear nonhomogeneous ordinary differential equation for the tension @xmath13 .
hence there is a green function @xmath151 depending on @xmath152 , such that @xmath153 similarly , equation can be thought of as a linear nonhomogeneous matrix equation for @xmath13 , for which the solution takes the analogous form @xmath154 for some `` discrete green function '' @xmath155 .
naturally we expect that if @xmath156 and @xmath157 , then @xmath158 as @xmath114 ; this can be proved as a consequence of our general convergence result for @xmath34 .
our goal in this section is to establish properties of these green functions .
in particular we establish that the green function is always nonnegative for a whip in a sufficiently smooth configuration , while the green function is nonnegative for a chain as long as all the angles between links are obtuse .
furthermore we want to establish upper and lower bounds for the ratios @xmath159 and @xmath160 , where @xmath161 , in order to be able to compare the norms weighted by powers of @xmath80 to the norms weighted by powers of @xmath162 .
first we discuss the solution operator of the whip tension . to keep the notation relatively simple , we will suppress the time dependence .
[ sigmasolprop ] for any fixed time @xmath56 , the solution @xmath163 of is given by @xmath164 where @xmath165 is the green function given by @xmath166 the green function is symmetric , i.e. , @xmath167 .
it satisfies @xmath168 whenever @xmath169 and @xmath81 .
in addition if @xmath170 , we have @xmath171 for @xmath172 and @xmath173 for @xmath174 .
the existence of the green function and the symmetry property @xmath167 is a well - known result of the general theory for second - order equations with homogeneous boundary conditions .
see for example courant - hilbert @xcite . to prove the other statements ,
we first show that @xmath175 for any @xmath176 . for any fixed @xmath177 , multiplying by @xmath178 , integrating from @xmath46 to @xmath55 , and using integration by parts with the homogeneous boundary conditions
shows that @xmath179 which forces @xmath180 .
because of the jump condition @xmath181 we can not have @xmath182 identically zero if @xmath170 .
so @xmath183 if @xmath170 .
it is then easy to prove the other statements in the intervals @xmath184 and @xmath185 using the boundary conditions .
now let us do the same for the tension operator for the chain .
the equation , or the more elegant version , makes clear that the vector @xmath186 of tensions comes from inverting a tridiagonal matrix .
since this is one of the easiest matrix types to invert , we get a relatively explicit formula for the solution , which will be useful in constructing estimates on the maximum and minimum tension .
[ solverprop ] the solution of the constraint equations is @xmath187 where the discrete green function @xmath155 is constructed by @xmath188 and @xmath189 satisfies the recursion @xmath190 in we use the convention that the empty product when @xmath191 is @xmath192 . the tensions @xmath90 are positive for every nontrivial choice of @xmath193 if and only if @xmath194 for every @xmath195 .
the system and is of the form @xmath196 , where @xmath197 is a symmetric nonnegative diagonally - dominant tridiagonal matrix and @xmath198 is the vector of angular velocities @xmath199 .
there are several standard algorithms for inverting such a matrix ; the formula is given in the review paper of meurant @xcite .
clearly @xmath200 for all @xmath195 , so that inductively we have @xmath201 for all @xmath195
. hence if all @xmath202 are positive , then all @xmath203 are positive and hence all terms @xmath155 are positive for @xmath204 .
thus if any @xmath205 is nonzero is zero is if the chain is stationary , since @xmath206 always . ]
, then says that all @xmath90 are positive for @xmath145 .
it is easy to see that if @xmath207 for some @xmath195 , then there is some choice of @xmath193 so that some @xmath13 is nonpositive . , @xmath208 , and @xmath209 links .
in the bottom row we plot the corresponding discrete green function @xmath155 as a function of @xmath210 , evaluated at @xmath211 , for each configuration .
notice that when @xmath212 the angles in the chain are acute , which is what allows the tension to become negative in that case .
also notice that as @xmath114 , the discrete green function approaches the green function for the differential equation . ] , @xmath208 , and @xmath209 links .
in the bottom row we plot the corresponding discrete green function @xmath155 as a function of @xmath210 , evaluated at @xmath211 , for each configuration .
notice that when @xmath212 the angles in the chain are acute , which is what allows the tension to become negative in that case .
also notice that as @xmath114 , the discrete green function approaches the green function for the differential equation . ] , @xmath208 , and @xmath209 links .
in the bottom row we plot the corresponding discrete green function @xmath155 as a function of @xmath210 , evaluated at @xmath211 , for each configuration .
notice that when @xmath212 the angles in the chain are acute , which is what allows the tension to become negative in that case .
also notice that as @xmath114 , the discrete green function approaches the green function for the differential equation . ]
+ , @xmath208 , and @xmath209 links . in the bottom row
we plot the corresponding discrete green function @xmath155 as a function of @xmath210 , evaluated at @xmath211 , for each configuration .
notice that when @xmath212 the angles in the chain are acute , which is what allows the tension to become negative in that case .
also notice that as @xmath114 , the discrete green function approaches the green function for the differential equation .
] , @xmath208 , and @xmath209 links . in the bottom row
we plot the corresponding discrete green function @xmath155 as a function of @xmath210 , evaluated at @xmath211 , for each configuration .
notice that when @xmath212 the angles in the chain are acute , which is what allows the tension to become negative in that case .
also notice that as @xmath114 , the discrete green function approaches the green function for the differential equation . ] , @xmath208 , and @xmath209 links .
in the bottom row we plot the corresponding discrete green function @xmath155 as a function of @xmath210 , evaluated at @xmath211 , for each configuration .
notice that when @xmath212 the angles in the chain are acute , which is what allows the tension to become negative in that case .
also notice that as @xmath114 , the discrete green function approaches the green function for the differential equation . ]
proposition [ sigmasolprop ] implies that if @xmath213 , then @xmath214 is a positive function of @xmath162 on @xmath215 $ ] , since @xmath216 .
we now want to know exactly how large or small this positive function can be ; ultimately our interest will be in the quantities @xmath217 } \sigma(s)/s$ ] and @xmath218 } \sigma(s)/s$ ] , which are completely determined by the bounds on the green function .
we are especially interested in the discrete analogues , @xmath219 and @xmath220 .
we end up with the same upper bound in both cases , which is relatively easy to prove , while the lower bound is much more complicated and necessarily weaker in the discrete case .
first we establish the upper bound .
[ greenupperbounds ] if @xmath221 is smooth , then the green function @xmath178 defined by proposition [ sigmasolprop ] satisfies the following bounds .
@xmath222 furthermore ,
suppose @xmath155 , @xmath122 , @xmath223 , and @xmath224 are as defined in proposition [ solverprop ] , and that @xmath225 for all @xmath116 , so that @xmath226 for all @xmath227 .
then if @xmath228 denotes the partial difference @xmath229 , using the convention @xmath230 , then @xmath231 the proof of is easy : by proposition [ sigmasolprop ] , the partial derivative @xmath182 is positive for @xmath232 and nonpositive for @xmath233 , and jumps by @xmath234 at @xmath235 . since @xmath236 whenever @xmath237 , we know @xmath238 is increasing on each interval .
we therefore must have @xmath239 for @xmath232 and @xmath240 for @xmath241 ; either way , @xmath242 . then using the fact that @xmath243
, we have @xmath244 which yields .
the proof of is more complicated , but uses the same basic ideas .
first , from proposition [ solverprop ] we know that @xmath226 for all @xmath116 and @xmath118 since every @xmath245 .
assume first that @xmath246 . then rewriting
, we see that the discrete green function satisfies the equation @xmath247 for @xmath248 , while @xmath249 since @xmath250 with @xmath251 , we have @xmath252 .
thus for @xmath253 we can easily see the second partial difference satisfies @xmath254 since the second partial differences are nonnegative except at the diagonal , the first partial differences are increasing except at the diagonal , i.e. , @xmath255 when @xmath256 we can check that @xmath257 now look at the endpoint terms : at the left endpoint , we know @xmath258 .
at the right endpoint , if @xmath246 then we have @xmath259 , so that @xmath260 . thus combining and , we conclude that if @xmath246 then @xmath261 hence we must have @xmath262 for all @xmath116 , as long as @xmath246 .
if @xmath263 , the situation is slightly different ; in that case we get @xmath264 so that @xmath262 even if @xmath263 .
this completes the proof of .
unfortunately we can not bound @xmath265 from above .
if we denote by @xmath266 the green function when @xmath267 , then we easily compute that @xmath268 , so that @xmath269 is unbounded on @xmath215\times [ 0,1]$ ] .
note that by the sturm comparison theorem , we have that @xmath270 for any green function satisfying .
however it is easy to see that for any @xmath271 , we have @xmath272
. this will be useful in the proof of theorem [ uniqueness ] .
it is easy to check that the discrete green function satisfies the same inequality , @xmath273 , using formula and the fact that @xmath274 and @xmath275 for all @xmath195 .
in fact this bound is valid even if not all @xmath155 are positive .
now we establish the lower bound .
this is the only time in the paper where we get a weaker result for the chain than for the whip ; the reason is that we need to make strong assumptions in order to prevent sharp kinks in the chain , to ensure nonnegative tension .
smoothness of the whip , on the other hand , ensures that the tension in the whip is nonnegative automatically .
[ sigmalowerbound ] suppose @xmath155 , @xmath122 , @xmath223 , and @xmath224 are defined as in proposition [ solverprop ] .
assume the @xmath122 are such that , for some @xmath276 $ ] , we have @xmath277 then for all @xmath204 we have @xmath278 if @xmath165 solves and @xmath221 is a smooth function , then we have @xmath279 the two estimates are proved in slightly different ways , but the main point for both estimates is to show that the minimum is attained at the off - diagonal corner , then estimate this value either using the direct formula ( for the chain ) or through a substitution ( for the whip ) .
the full proof is in appendix [ a1 ] .
the assumption for the discrete case is much stronger than the assumption @xmath280 for the continuous case , but such a pointwise bound is necessary to ensure every @xmath281 in order to get all tensions positive ( by proposition [ solverprop ] ) , even when @xmath105 .
the exponent @xmath282 is important : the exponent @xmath192 would work to prove the estimate , but we can not prove that such an estimate actually holds for all values of @xmath56 ; the exponent @xmath283 is not enough to get a lower bound for the tension .
[ cantweaken ] note that we could easily get a stronger estimate than if we simply assumed an upper bound on @xmath284 , using the sturm - liouville comparison theorem . however , we prefer the weaker assumption that @xmath285 , since it allows for the possibility of the curvature at the free end of the whip approaching infinity ( a possibility not precluded by the equations due to the degeneracy there ) . even if the weighted energy @xmath19 is finite the condition under which we will prove local existence
we will not necessarily have @xmath152 bounded on @xmath215 $ ] ; an example is when @xmath286 for some @xmath287 , using the spherical representation .
see example [ abcnecessary ] for details .
in order to demonstrate existence and uniqueness , we want to apply the usual technique of energy estimates in sobolev spaces . by showing that we have sequences of solutions of the chain equations for which the energy is uniformly bounded
, we can extract a convergent subsequence to establish existence ; the same sort of energy estimates can also be used to establish uniqueness .
several issues arise to complicate this strategy . ordinarily for a wave equation like
, one would try to bound an energy like @xmath288 by computing its derivative and using gronwall s lemma .
it is easy to compute that @xmath289 using the boundary conditions @xmath2 and @xmath290 .
unfortunately we can not bound @xmath291 or even @xmath13 in terms only of @xmath292 .
indeed , it is hard to even make sense of equation unless both @xmath293 and @xmath221 are in @xmath294 , which means we have to consider higher energies .
here a complication arises .
the usual approach would be to consider an energy like @xmath295 its derivative is , using , @xmath296 here the boundary term vanishes since @xmath2 and @xmath297 ( recall we assume @xmath34 extends to an odd function through @xmath55 ) .
furthermore since @xmath298 , we have @xmath299 .
the problem is that if we want to get the right side in terms of @xmath300 alone , we need to use the cauchy - schwarz inequality to get @xmath301 but the right side is _ not _ bounded .
we always have @xmath2 , while we will generally not have @xmath302 .
instead we want an energy for which the integration by parts cancels out this highest - order remainder .
the only such quantity of the form @xmath303 for which this works is @xmath304 with such a choice we get @xmath305 which we can manage once we understand how @xmath13 and @xmath291 behave . the same phenomenon continues for the higher energies as well , which motivates us to define @xmath306 with this definition , we have @xmath307 , \end{split}\ ] ] with the three remaining terms integrating to give @xmath308 due to oddness of @xmath34 through @xmath55 .
our primary goal will be to bound in terms of the energies @xmath309 . more specifically , using the fact that @xmath80 degenerates like @xmath162 near @xmath46 , we want to get bounds in terms of the simpler weighted energies @xmath310 defined by @xmath311 to do this , we will need several estimates .
so our first goal is establishing basic sobolev - type and wirtinger - type inequalities for such weighted norms .
in addition we need to show the energies are equivalent to the tension - dependent energies , which means we have to bound @xmath312 and @xmath313 away from zero .
( the constants in these bounds will also turn out to depend on the energies . )
most of the work for this was done in section [ greenfunctionsection ] .
first let us define the weighted sobolev and supremum seminorms we need .
[ weightednorms ] the weighted sobolev seminorm of a function @xmath22 \to \mathbb{r}^d$ ] is defined by @xmath314 the weighted supremum seminorm of @xmath126 is @xmath315 we want to define a discrete analogue of each of these , for a sequence @xmath316 with values in @xmath84 . for this purpose , it is convenient to set @xmath317 where @xmath318 is the usual gamma function satisfying @xmath319 for @xmath169 and @xmath320 for @xmath116 a natural number .
these are rising factorials , which are more convenient for our purposes than the falling factorials typically used in difference equations ; either is much more convenient in studying difference equations than simply using the powers @xmath321 ; see @xcite .
clearly if @xmath322 is a sequence such that @xmath323 , then we have @xmath324 . recalling the definition of the difference operator @xmath132
, we define our discrete analogues of by @xmath325 and @xmath326 we use the same notation for both norms to emphasize the analogy ; every estimate we prove for the discrete seminorms will have constants independent of @xmath6 , so that we get corresponding estimates for the smooth seminorms . clearly if @xmath126 is smooth and we define @xmath327 for each @xmath6 , then @xmath328 and @xmath329 .
now let us describe the main estimates . for two norms @xmath330 and @xmath331 on functions , we use the notation @xmath332 to mean @xmath333 for some constant @xmath334 independent of @xmath126 . if @xmath126 is instead a sequence , then this notation will imply that @xmath334 is also independent of @xmath6 . for unweighted norms of smooth functions
, we have the wirtinger inequality @xmath335 we also have the sobolev inequality @xmath336 our weighted versions of each are as follows . [ basicinequalitiesthm ]
let @xmath22\to \mathbb{r}^d$ ] be @xmath44 .
then for any @xmath337 the norms and satisfy the weighted inequalities @xmath338 if in addition we have @xmath339 , then these inequalities can be simplified to @xmath340 if @xmath126 is instead a sequence @xmath341 with values in @xmath84 , then the inequalities also hold if the norms are interpreted as and , while the inequalities hold if @xmath342 .
the continuous version of this inequality appears in adams - fournier @xcite .
we prove the discrete version in appendix [ a2 ] , from which the continuous version follows in the limit .
[ nosup ] the example @xmath343 demonstrates that the inequalities and can not be extended to @xmath344 : in that case we have @xmath345 , @xmath346 , and @xmath347 , while @xmath348 and @xmath349 } \lvert f(x)\rvert^2 $ ] are both infinite
. in particular there can not be constants for the discrete versions that are independent of @xmath6 when @xmath344 .
the important thing about is that by remark [ nosup ] , the estimate only works when @xmath337 .
hence in any computation where a supremum norm is required , we will want a positive power of @xmath162 attached to be able to use this result .
this will show up when we need to estimate weighted sobolev norms of products of three functions : we want to pull out a supremum of one and use cauchy - schwarz on the rest , and we will need a little extra weighting in some cases . of course , we could use the usual sobolev inequality to get @xmath350 but requiring _ two _ extra derivatives rather than one is usually not worthwhile ( except once in the proof of theorem [ uniqueness ] ) . frequently our weighting in discrete norms will be slightly off ( for example , we may want to replace @xmath351 with @xmath352 , or we may want to replace @xmath352 with @xmath353 for some @xmath118 ) . in the continuous case
these formulas are trivial , but in the discrete case , bounds such as these come from properties of the gamma function ( in particular the fact that the gamma function is log - convex by the bohr - mollerup theorem ) .
the constants will never be important ; what will matter is that they are independent of @xmath116 and @xmath6 .
the following estimates are easy to prove .
[ weightsprop ] let @xmath354 and let @xmath355 .
let @xmath356 and @xmath357 be positive real numbers .
then the weight function @xmath358 satisfies the following inequalities : @xmath359 we also have @xmath360 for any nonnegative integer @xmath118 .
proposition [ weightsprop ] also gives the following corollary , which is the most useful tool we have for estimating norms of products . to get the higher - difference norms of products
, we will use the product rule for differences together with these formulas .
the proof is trivial .
[ supproductboundcorollary ] suppose @xmath361 and @xmath362 are sequences of real numbers , and let @xmath356 and @xmath357 be nonnegative real numbers .
then @xmath363 and @xmath364 the formulas are also valid if one of the sequences is in @xmath365 and the other in @xmath84 , or if both are in @xmath84 and we use @xmath366 or @xmath367 instead . estimates and are also valid if @xmath126 and @xmath368 are smooth functions with the norms interpreted as and . typically we will extend the seminorms and when used for @xmath34 and @xmath13 ; since we have @xmath104 and @xmath369 , it is more convenient to modify the definitions to @xmath370 this does not affect any of the estimates , but it allows us to incorporate the endpoint information .
this is convenient for example to interpret in terms of @xmath371 in lemma [ acboundslemma ] .
we have already defined the weighted energy and for a whip .
now we want to define corresponding discrete energy for the chain .
the definitions are made much easier if we use the odd extension of @xmath34 to define the differences @xmath372 beyond @xmath373 .
furthermore , by analogy with , we define @xmath374 our time - independent energy will be @xmath375 while the time - dependent energy is @xmath376 recall that we need to use the time - dependent @xmath13-weighted quantities to compute the time - derivative of energy in order to get some cancellation , while only time - independent energies are useful for constructing topologies and relating distinct norms .
clearly if we have sequences @xmath377 and @xmath378 defined for each @xmath354 as in , then @xmath379 = \lim_{n\to\infty } e_m[\eta_n ] \quad \text{and}\quad \widetilde{e}_m[\eta ] = \lim_{n\to\infty } \widetilde{e}_m[\eta_n].\ ] ] so any estimate we obtain on the chain energies @xmath380 and @xmath381 will become an _ a priori _ estimate on the corresponding whip energies @xmath310 and @xmath382 . note that we have @xmath383 in terms of the discrete weighted seminorms .
went from @xmath105 to @xmath384 rather than @xmath385 .
the reason the sums in and contain a few extra terms in the sums is in order to make the derivative estimate of theorem [ discreteenergyestimate ] simpler : with this definition the endpoint terms of the discrete energy derivative always vanish , as they did in . ]
it is also convenient to observe that the lowest - level energy @xmath386 is constant in time . since it will be useful later in lemma [ bboundslemma ] ,
we separate the terms and define @xmath387 [ u0v0constant ] if @xmath34 satisfies with @xmath388 , then both @xmath389 and @xmath390 are constant in time .
the fact that @xmath391 implies that @xmath392 . as a corollary , every energy @xmath380 given by satisfies @xmath393 . for @xmath389 , we just compute @xmath394 using the summation by parts formula along with @xmath395 and the endpoint conditions @xmath104 and @xmath369 , it is easy to show this sum vanishes . for a smooth solution of , we clearly have that the analogous quantities @xmath396 and @xmath397 satisfy @xmath398 and @xmath399 is constant in time .
thus @xmath400 is also constant in time .
our primary use of theorem [ basicinequalitiesthm ] will be the following formulas , which follow easily from .
[ normtoenergy ] for any @xmath401 and @xmath402 , we have @xmath403 we need the extra power of @xmath404 in estimate , since is not valid when @xmath344 .
this is important once at the end of the proof of lemma [ sigmasobolevlemma ] and once at the end of the proof of theorem [ discreteenergyestimate ] .
before bounding the energy itself , we first want bounds for the tension @xmath13 given by either or . to compare the energies @xmath309 and @xmath310 ,
we want upper and lower bounds for @xmath405 . in addition , to compute the time derivative of @xmath309 , we need to know a bound for @xmath406 , by formula . for a smooth solution @xmath407 of ,
we define quantities @xmath197 , @xmath408 , and @xmath334 by the formulas @xmath409 observe that since @xmath2 , we have @xmath410 so that we have @xmath411 generally the bounds will be much more useful to us , although occasionally we will need the actual definition . similarly , we define the discrete analogues of the quantities ; as with the energy , we use upper - case and lower - case for norms of the whip or chain respectively . recall that @xmath161 , while our convention is that @xmath121 , and recall the definition @xmath412 .
we therefore set @xmath413 as above , the fact that @xmath121 means we can write @xmath414 to obtain @xmath415 [ acboundslemma ] if @xmath34 and @xmath13 form a smooth solution of
, then the quantities defined by satisfy the bounds @xmath416 similarly , suppose @xmath417 and @xmath418 form a solution of and with the odd extensions .
suppose also that we have @xmath419 for @xmath420 .
then @xmath421 and @xmath422 given by satisfy the bounds @xmath423 we will just prove the discrete bounds for @xmath421 and @xmath422 in detail ; the bounds for @xmath197 and @xmath334 can be proved using the same techniques , or we can view them as a limiting case of the bounds for @xmath421 and @xmath422 .
the estimate for @xmath421 is easy : by proposition [ solverprop ] , we have @xmath424 by proposition [ greenupperbounds ] , we have @xmath262 , and thus @xmath425 by
. we then have .
the inequality is proved identically .
the bound for @xmath422 is more complicated .
the first step is to differentiate the equation in time to get @xmath426 thus @xmath427 satisfies the same kind of equation as @xmath13 with the same endpoint conditions , so we can use proposition [ solverprop ] to write @xmath428 as above , the fact that @xmath429 implies that @xmath430 , where @xmath431 applying the summation by parts formula to this , using the endpoint conditions @xmath432 , and performing some manipulations with the formula @xmath433 , we get @xmath434.\ ] ] using the bounds @xmath435 and @xmath436 from above , we obtain @xmath437 using lemma [ normtoenergy ] , which gives .
the proof of is almost identical .
[ bboundslemma ] suppose @xmath34 and @xmath13 solve . then @xmath408 defined by satisfies the estimate @xmath438 where @xmath439 is a constant depending on the initial condition .
similarly , suppose @xmath417 and @xmath418 form a solution of and with the odd extensions .
suppose also that @xmath440 .
then the quantity @xmath441 in satisfies the estimate @xmath442 where @xmath389 is defined by ( and is constant due to lemma [ u0v0constant ] ) .
the bound comes directly from : we have by proposition [ sigmasolprop ] that @xmath443 the fact that @xmath444 follows from , since @xmath290 .
the proof of is similar , using and for the discrete green function .
the bound @xmath445 similarly follows from since @xmath446 .
[ abcnecessary ] in terms of the weighted energy , lemmas [ acboundslemma ] and [ bboundslemma ] give upper bounds for @xmath197 and @xmath408 if the energy @xmath447 is finite , while we only get an upper bound for @xmath334 if @xmath19 is finite .
simple examples show that these conditions are necessary : we can have @xmath448 bounded while @xmath197 and @xmath408 are unbounded , and we can have @xmath447 bounded while @xmath334 is unbounded . to obtain the examples for @xmath197 and @xmath408 , we consider ( at time @xmath449 ) the whip position @xmath450 which satisfies @xmath451 , @xmath71 , and @xmath452 .
this corresponds to a whip where the free end sits at @xmath453 despite making infinitely many rotations around it ( as @xmath454 ) .
see figure [ logarithmicspiral ] . infinitely many times .
we have plotted this heuristically in the inset , although the actual curve wraps itself up too tightly for these loops to be visible . ]
the green function can be computed explicitly to obtain @xmath13 from @xmath455 .
if @xmath456 satisfies and @xmath457 , one computes that @xmath448 is finite while @xmath447 and @xmath197 are both infinite . if on the other hand @xmath458 , we easily see that @xmath448 is still finite while @xmath447 and @xmath408 are both infinite .
the example for @xmath334 is a bit more involved .
suppose @xmath34 is given by , where @xmath459 and @xmath460 .
it is easy to verify that @xmath447 is finite at this instant , while @xmath19 is infinite .
we have @xmath461 at this instant by , so that differentiating with respect to time and using gives @xmath462 with boundary conditions @xmath463 and @xmath464 . in this case
we can verify that @xmath334 is infinite .
the fact that we can not bound @xmath334 unless @xmath19 is bounded is one of the main reasons why the energy estimates only close up at @xmath19 . taking a time derivative of as in gives a number of terms of the form @xmath291 which can only be bounded in terms of @xmath334 , and thus in terms of @xmath19 .
we are now ready for an a priori estimate on the tension @xmath13 .
although the norms of @xmath13 and @xmath465 are easier to measure using the supremum , it is convenient to use weighted sobolev norms for the higher derivatives of @xmath13 .
thus we define the squared norm for a whip : @xmath466 the discrete version is defined by the same formula : @xmath467 [ sigmasobolevlemma ] if @xmath13 is a smooth solution of , then the norms can be bounded by the energy via @xmath468 while for @xmath469 we have @xmath470 where @xmath471 depends only on @xmath472 .
similarly if @xmath13 satisfies with the condition @xmath104 , then the norms can be bounded by the energy via @xmath473 while for @xmath469 we have @xmath474 where @xmath471 depends only on @xmath475 .
we will just prove the discrete estimates ; the estimates are proved using the exact same technique .
the full proof is in appendix [ a3 ] ; the basic idea is just to take iterated differences of and estimate using corollary [ supproductboundcorollary ] and lemma [ normtoenergy ] .
[ energyestimatesection ] in order to construct the solution of the partial differential equations , we want to find bounds on all the discrete energies which are independent of the initial conditions and of the number @xmath6 of links . then in section [ convergence ] we will find a subsequence that converges to a solution . as a consequence , we can show that the motion of a chain converges to the motion of a whip as @xmath6 approaches infinity , in the sense that position , velocity , and acceleration all converge .
we now want to estimate the time evolution of the energy @xmath476 . our strategy will be to bound @xmath477 in terms of the energies @xmath380 ; we will then use the fact that @xmath380 and @xmath476 are equivalent ( since @xmath405 is bounded above and below by lemmas [ acboundslemma ] and [ bboundslemma ] ) to get an inequality for @xmath477 in terms of @xmath476 . in proving it we will use lemmas [ acboundslemma][bboundslemma ] and [ sigmasobolevlemma ] in an essential way .
[ discreteenergyestimate ] let @xmath478 , and suppose @xmath417 and @xmath418 form a solution of and with @xmath479 and @xmath480 , and that @xmath34 and @xmath13 extend to sequences satisfying the oddness condition .
then the energies and satisfy the estimates @xmath481 for some @xmath482 independent of the initial data and of @xmath6 .
in addition the higher energies satisfy @xmath483 for every @xmath469 , where @xmath484 depends only on @xmath475 .
analogously , if @xmath34 and @xmath13 form a smooth solution of and , then the energies and satisfy the estimates @xmath485 for some @xmath482 independent of the initial data . in addition
the higher energies satisfy @xmath486 for every @xmath469 , where @xmath484 depends only on @xmath472 .
the proof is in appendix [ a4 ] .
the fact that the energy estimates only close up at @xmath5 is perhaps explained by the following observation , which is easier to understand in terms of the spherical representation .
[ r4 ] let @xmath487 denote the unit ball in @xmath488 .
the pair @xmath489 is a smooth solution of if and only if the functions @xmath490 and @xmath491 defined by @xmath492 are spherically symmetric solutions of the equations @xmath493 with neumann boundary condition @xmath494 for @xmath495 and robin boundary condition @xmath496 for @xmath497 on @xmath498 .
furthermore any smooth solution has @xmath499 everywhere , so that the hyperbolic equation for @xmath495 is nondegenerate . setting @xmath500 and changing variables by @xmath501
, we easily see that
become @xmath502 now the operator @xmath503 is familiar as the spherically symmetric laplacian on @xmath488 , and hence we recognize both terms above as coming from the laplacian on @xmath488 under the assumption that @xmath497 and @xmath78 are both spherically symmetric .
the boundary conditions are easy to check .
the fact that the degeneracy can be removed if we work in a higher - dimensional space , and thus in some sense the equations naturally `` live '' there , is essentially the reason why we need higher than usual sobolev order for the estimates to close .
[ existenceuniquenesssection ] now we can finally prove the local existence theorem for the system of partial differential equations .
the fact that theorem [ discreteenergyestimate ] gives us estimates for @xmath476 in terms of @xmath380 that are independent of @xmath6 allows us to construct the solution as a limit of a subsequence of discrete solutions as @xmath103 , following the technique of ladyzhenskaya @xcite and references therein .
we first need to establish the interior approximation of the space of whips by the space of chains , which allows us to go from estimates on @xmath380 given by to estimates on @xmath310 given by and back .
consider any function @xmath504\to \mathbb{r}^d$ ] such that @xmath505 , with @xmath34 extending to an odd function through @xmath55 , such that the seminorms @xmath506 for @xmath507 are all finite .
for each @xmath354 we want to approximate @xmath34 by a sequence @xmath508 for @xmath145 , extend it for @xmath509 by @xmath510 , have it satisfy @xmath511 , and have uniform bounds on the discrete sobolev seminorms @xmath512 in terms of the smooth seminorms that are independent of @xmath6 .
the complication arises from handling the constraint @xmath71 .
although it is relatively easy to approximate functions by sequences in the norms we need , the typical discrete approximation will not satisfy the condition @xmath388 , which means it does not actually represent a chain .
we deal with this by using the spherical representation @xmath513 as in .
( although this formula works only when @xmath74 , we can use a similar procedure in higher dimensions , using generalized spherical coordinates . ) using @xmath514 , we can easily reconstruct @xmath34 if @xmath78 is known .
we can then approximate the function @xmath78 by a sequence @xmath515 and rebuild @xmath122 using the formula @xmath369 and @xmath516 when @xmath74 , with a similar formula in higher dimensions .
fortunately , the sobolev norms of @xmath34 and @xmath78 are closely related .
[ etatothetasobolev ] if @xmath504\to\mathbb{r}^2 $ ] is related to @xmath517\to \mathbb{r}$ ] by the formula , with @xmath518 and @xmath519 , then boundedness of the squared norm @xmath520 is equivalent to boundedness of the squared norm @xmath521
we easily compute that @xmath522 an integration by parts using @xmath519 shows that and are related by @xmath523 repeated use of the basic weighted sobolev inequalities of theorem [ basicinequalitiesthm ] allows us to express every term on the right side in terms of @xmath197 , so we get an inequality of the form @xmath524 . in the other direction , gives @xmath525 and again using theorem [ basicinequalitiesthm ] gives @xmath526 .
we can derive the same sort of result for difference quotients .
if we write @xmath516 for @xmath145 , then the analogues of are as follows : @xmath527 for sufficiently large @xmath6 , we can proceed as in proposition [ etatothetasobolev ] to show that the discrete squared norms @xmath528 and @xmath529 can each be bounded in terms of the other .
thus for either whips or chains in two dimensions , we can work directly in terms of sobolev norms of @xmath78 .
the most convenient way to map from sobolev spaces of continuous maps to sobolev spaces of discrete sequences is to use orthogonal polynomials .
( a direct approach , using values of the function on a discrete grid , does not work for our purposes since bounds on the differences require more smoothness of the function than we have . )
the only complication is the oddness requirement on @xmath34 ( and the discrete oddness criterion ) . in terms of the spherical variable @xmath78 , oddness of @xmath34 through @xmath55
translates into evenness of @xmath78 , i.e. , there is an extension of @xmath78 to @xmath27 $ ] such that @xmath530 .
similarly the discrete oddness condition translates into the discrete evenness condition @xmath531 .
these conditions are easy to handle if we extend the interval to @xmath27 $ ] ( or extend the sequence to @xmath532 ) and use sobolev seminorms with symmetric weights @xmath533 and @xmath534 these norms are clearly topologically equivalent to the weighted norms we have been using , where the weights are @xmath162 and @xmath210 .
[ orthogonalexpansion ] there are polynomials @xmath535 on @xmath27 $ ] satisfying @xmath536 , and such that @xmath537 in the weighted sobolev seminorms for all @xmath538
. thus we can expand @xmath539 and obtain @xmath540 for all @xmath538 .
there are also , for each @xmath354 , discrete polynomials @xmath541 defined for @xmath542 and @xmath543 , satisfying @xmath544 and such that @xmath545 where @xmath546 is as in . hence if @xmath547 for @xmath145 , then @xmath548 the desired polynomials come from a slight variation on the classical legendre polynomials and the chebyshev polynomials of a discrete variable .
( these are special cases of the jacobi and hahn polynomials respectively , with parameters @xmath549 . )
the desired formulas follow from general properties of continuous and discrete orthogonal polynomials ; see nikiforov et al .
@xcite for a good reference .
to obtain , we set @xmath550 , where @xmath551 is the usual legendre polynomial given by the rodrigues formula and @xmath552 is a constant chosen to make @xmath553 orthonormal when @xmath554 . to obtain , we set @xmath555,\ ] ] where @xmath556\ ] ] is the hahn polynomial given in terms of a discrete rodrigues formula , with @xmath128 denoting the integer shift operator , and again @xmath557 is a constant chosen to give orthonormality when @xmath554 .
checking all the conditions is routine using the formulas in @xcite . for each @xmath354
, we can define the map @xmath558 which takes a continuous angular function @xmath559 to a discrete approximation @xmath515 , and the map @xmath560 which takes a discrete angular sequence @xmath515 to a continuous angular function @xmath559 , by the formulas @xmath561 where the coefficients are obtained using orthonormality by @xmath562 by the formulas and , we can bound the continuous and discrete sobolev norms of any order @xmath118 in terms of each other using this map .
furthermore @xmath560 is an isometry , @xmath563 is the identity , and @xmath564 converges strongly to the identity as @xmath114 in any weighted @xmath565-norm . thus given an initial condition @xmath566 , we can write the discrete initial condition @xmath567 as @xmath568 where @xmath515 is the discretization obtained from . and
conversely , if we solve the discrete chain equations to obtain @xmath569 , we can construct an approximate whip solution by finding , for each @xmath56 , the angles @xmath570 and using to obtain the function @xmath571 , then reconstructing @xmath572 .
we clearly have a similar construction for the velocity @xmath573 in terms of the angular velocity @xmath574 , which works based on the formulas @xmath575 these constructions ensure that we can go back and forth between whips and chains while preserving the sobolev norms as well as the constraint equation .
now suppose that the initial whip conditions @xmath36 and @xmath576 have bounded energy @xmath577 given by , as well as satisfying the constraints @xmath578 and @xmath579 , and have odd extensions through @xmath55 . using the procedure of the preceding section , we know that for each @xmath354 there are discrete initial conditions @xmath567 and @xmath580 such that the discrete energy @xmath38 given by is bounded uniformly , independently of @xmath6 .
these approximate conditions converge strongly in @xmath581 $ ] and @xmath582 $ ] respectively to the actual initial conditions .
[ uniformestimateslemma ] suppose @xmath583 and @xmath198 are initial conditions as in theorem [ theorem1 ] , and suppose discretizations @xmath567 and @xmath580 are defined as in section [ interpolation ] .
let @xmath584 and @xmath585 , for @xmath145 , be the solution of equations and with @xmath586 and @xmath587 .
then there is a @xmath33 such that the discrete energy @xmath588 defined by is bounded uniformly on @xmath45 $ ] and uniformly in @xmath6 .
since the discrete energy @xmath589 is bounded uniformly for all @xmath6 , we conclude by lemma [ normtoenergy ] that @xmath590 is uniformly bounded for all @xmath6 .
in particular we know the hypotheses of lemma [ bboundslemma ] are satisfied for @xmath6 sufficiently large .
fix such an @xmath6 . by lemma [ bboundslemma ]
we have @xmath591 we want an estimate for the evolution of @xmath592 . for any @xmath593 , we have by proposition [ weightsprop ] and lemma [ normtoenergy ] that @xmath594 since this is true for any @xmath116 , we conclude @xmath595 for some constant @xmath596 , independent of @xmath56 and @xmath6 .
this bound also ensures that the hypotheses of lemma [ bboundslemma ] are satisfied for sufficiently large @xmath6 as long as @xmath597 is bounded . by the definitions and
, we clearly have @xmath598 and we conclude by combining and that @xmath599 for some constant @xmath600 which is also independent of @xmath56 and @xmath6 .
let @xmath601 and let @xmath602
. then can be written as @xmath603 and integrating both sides yields @xmath604 now we use theorem [ discreteenergyestimate ] to get @xmath605 for some @xmath482 independent of @xmath56 and @xmath6 . using and , we have @xmath606 so we obtain @xmath607 dividing by @xmath608 and integrating , we obtain @xmath609^{-1/5},\ ] ] where @xmath610 .
another integration gives a bound for @xmath611 on some time interval @xmath45 $ ] , which depends only on @xmath589 , and gives a uniform bound on @xmath612 . combining this with and
, we get a uniform bound on @xmath597 as well on the same time interval .
now having obtained a sequence of chain solutions @xmath613 , bounded uniformly in the discrete weighted sobolev norms uniformly on an interval @xmath45 $ ] , we use the technique of section [ interpolation ] to interpolate . for each @xmath6
we obtain an approximate whip solution @xmath614\times [ 0,1]\to \mathbb{r}^d$ ] for which the energy @xmath43 is bounded on @xmath45 $ ] independently of @xmath6 .
we can then extract a subsequence which converges in the weak- * topology on @xmath35 , n_4[0,1])$ ] . before doing this , we prove one final lemma , a compactness result analogous to the usual rellich theorem .
[ compactnesslemma ] let @xmath615 $ ] denote the space of functions @xmath616\to\mathbb{r}^d$ ] such that the norm @xmath617 is finite .
then @xmath618 $ ] is compact in @xmath615 $ ] for each @xmath619 .
expand @xmath620 , where @xmath621 are the standard legendre polynomials .
then as discussed in section [ interpolation ] , we have @xmath622 hence the embedding @xmath623 is a norm limit of operators with finite - dimensional range , so it is compact . as noted in section [ interpolation ] , for functions on @xmath215 $ ] that are restrictions of odd functions on @xmath27 $ ] , the norm on @xmath615 $ ] given by is equivalent to the norm on @xmath21 $ ] given by , and thus we get compactness of @xmath624 $ ] in @xmath21 $ ] for functions with an odd extension through @xmath55 .
we now establish the existence part of theorem [ theorem1 ] .
[ convergencetheorem ] given initial conditions @xmath583 and @xmath198 as in theorem [ theorem1 ] , there is a @xmath33 such that there is a solution @xmath34 of the system in @xmath35 , n_4[0,1])\cap w^{1,\infty}([0,t ] , n_3[0,1])$ ] . for each fixed @xmath56 and each @xmath478 , construct a continuous approximation of the chain @xmath613 as in section [ interpolation ] , and
call it @xmath625 .
then by lemma [ uniformestimateslemma ] we get a uniform bound on on @xmath43 in some short time interval @xmath45 $ ] ; in other words , the family @xmath626 is bounded in @xmath35 , n_4[0,1])\cap w^{1,\infty}([0,t ] , n_3[0,1])$ ] . by the alaoglu theorem
, there is a subsequence @xmath627 that converges in the weak- * topology to @xmath628 , n_4[0,1])\cap w^{1,\infty}([0,t ] , n_3[0,1])$ ] . by the compactness lemma [ compactnesslemma ] ,
there is a sub - subsequence @xmath629 which converges strongly to @xmath34 in @xmath35 , n_3[0,1])\cap w^{1,\infty}([0,t ] , n_2[0,1])$ ] . for any @xmath630
the convergence is strong in @xmath631 $ ] , and thus by the usual sobolev embedding theorem also in @xmath632 $ ] .
so we can take the limit of the system and pointwise to see that we have a solution of .
the fact that all the estimates close up at the level of @xmath19 , with all other energies satisfying linear differential inequalities , implies that the only way a solution which is initially @xmath44 can fail to be @xmath44 for all time is if @xmath19 becomes infinite in finite time .
this gives a crude blowup criterion .
[ blowupcriterion ] suppose @xmath34 , @xmath13 is a solution of the system @xmath633 where we assume that @xmath583 and @xmath198 satisfy the conditions of theorem [ theorem1 ] .
assume that in some time interval @xmath45 $ ] , the energy @xmath43 is bounded uniformly .
assume further that @xmath634 is bounded for all @xmath469 .
then @xmath42 is also bounded in @xmath45 $ ] for all @xmath469 . by equation
, we have for @xmath635 that @xmath636 furthermore since @xmath637 is bounded , so is @xmath638 by lemma [ bboundslemma ] , and thus @xmath639 for some function @xmath640 .
so by gronwall s inequality , @xmath641 is bounded on @xmath45 $ ] in terms of @xmath642 .
thus finally @xmath643 is also bounded on @xmath45 $ ] .
we now complete the proof of theorem [ theorem1 ] by proving uniqueness .
[ uniqueness ] suppose @xmath583 and @xmath198 are functions on @xmath215 $ ] as in theorem [ theorem1 ] . if @xmath644 and @xmath645 are two solutions of , both in @xmath646 , n_4[0,1])\cap w^{1,\infty}([0,t ] , n_3[0,1]),\ ] ] with the same initial conditions @xmath647 then @xmath648 and @xmath649 for all @xmath650 $ ] and all @xmath651 $ ] .
the proof relies on an energy estimate for the differences at the level of the first energy , @xmath652 = \int_0 ^ 1 \big ( \lvert \partial_t\eta_1-\partial_t\eta_2\rvert^2 + s \lvert \partial_s\eta_1-\partial_s\eta_2\rvert^2\\ + s \lvert \partial_t\partial_s\eta_1-\partial_t\partial_s\eta_2\rvert^2 + s^2 \lvert \partial_s^2\eta_1-\partial_s^2\eta_2\rvert^2 \big ) \ , ds.\end{gathered}\ ] ] we estimate this energy using a gronwall inequality , as in theorem [ discreteenergyestimate ] and lemma [ uniformestimateslemma ] .
the reason this works is that since @xmath653 satisfies a linear pde whose coefficients involve the known quantities @xmath654 , @xmath655 , @xmath656 , and @xmath657 , we can use corollary [ supproductboundcorollary ] in a more effective way to put all the weights on the known terms .
the full proof appears in appendix [ a5 ] .
finally we discuss some refinements of these results .
first , given a solution @xmath34 of with @xmath19 finite , we can check using the differential equation that @xmath658 , @xmath659 , and @xmath660 can all be bounded in terms of @xmath19 .
hence the solution is also in @xmath661,n_2[0,1 ] ) \cap w^{3,\infty}([0,t ] , n_1[0,1 ] ) \cap w^{4,\infty}([0,t ] , n_0[0,1]).$ ] now a well - known general technique ( see e.g. , lemma 11.9 of @xcite ) shows that @xmath34 is continuous as a curve in @xmath581 $ ] , @xmath34 is @xmath48 as a curve in @xmath582 $ ] , etc .
in this paper we considered the whip with one fixed and one free end as boundary conditions .
the other possibilities are to have two free ends , to have two fixed ends , and to have periodicity .
all of the estimates in this paper have analogues in those cases .
when there are two free ends , the tension must satisfy @xmath662 and @xmath663 , so the appropriate weighted norms look like the square root of @xmath664 .
since we have essentially solved the problem with one fixed end by constructing an odd extension in order to turn the problem into a string with two free ends on @xmath27 $ ] , we expect that the same estimates prove existence for an inextensible string with two free ends . when there are two fixed ends , or when the whip is periodic , the problem becomes simpler since we can use ordinary sobolev spaces for the estimates . in this case
we expect the energy estimates to close up at the level of @xmath665 rather than @xmath38 .
the addition of gravity brings some complications .
one is that the boundary conditions change , and oddness through the fixed point is no longer enough to satisfy the conditions automatically .
( this is already an issue even for the wave equation with constant coefficients , if an external force is imposed which does not respect the boundary conditions . )
the other complication is that if the whip is above the fixed point , the tension may become negative : the effect of gravity is to change the boundary condition in to @xmath666 , where @xmath368 is the gravitational acceleration vector , and if @xmath667 then it is possible to have @xmath668 for some @xmath56 and @xmath162 . in that case
the evolution equation becomes elliptic , so the discussion becomes much more complicated .
the blowup criterion corollary [ blowupcriterion ] , that a smooth solution remains smooth up to time @xmath51 iff @xmath669 , can certainly be improved .
once we know a solution exists , we can use alternative methods to get better a priori bounds on it .
thess et al . have speculated that blowup for the periodic loop might be controlled by the @xmath670 norms of @xmath152 and @xmath671 , analogous to the way blowup for the ideal euler equations is controlled by the @xmath670 norm of vorticity .
this is an interesting problem to study , since we have a much greater handle on all aspects of this one - dimensional problem .
we will explore this in a future paper .
in addition , the geometry of the space of inextensible curves is interesting in its own right .
although the geometric objects are not smooth in the sobolev topology ( unlike on the group of volumorphisms ) , the curvature formulas still make sense , and one can compute formally that all sectional curvatures are nonnegative .
we can thus try to study stability of the motion from the geometric point of view ( as in @xcite ) , as well as the geometry of blowup .
see @xcite for details on this . a similar problem in higher dimensions is given by the motion of a flag attached to a pole in @xmath672-space .
here our configuration space would be the space of maps of a rectangle into @xmath7 which are isometric immersions with one side of the rectangle held fixed .
we expect to see a similar nonlocal coupled degenerate system , the only obvious difference being that the ordinary differential equation becomes an elliptic equation in the spatial variables .
the whip - chain equations are interesting partly in and of themselves , but especially as a `` toy model '' of inviscid , incompressible fluids .
there are some structural similarities between the equations and and the euler equation for an ideal fluid , given in lagrangian form by @xmath673 ^ 2\big),\ ] ] with some boundary condition to determine @xmath674 uniquely .
both systems involve a hyperbolic evolution equation for a constrained function , where the right side is given in terms of a function determined by a purely spatial differential equation .
the technique of approximating a continuous system with a discrete system preserving the geometry may be interesting to apply to fluids directly .
for example , in two dimensions we could consider a rectangular grid on a torus , the vertices of which are free to move as long as all quadrilateral areas are preserved .
although such a model may not have global existence ( as edges of a quadrilateral may collapse to give a triangle without changing the area ) , we might still get some useful insight out of it .
suppose @xmath155 , @xmath122 , @xmath223 , and @xmath224 are defined as in proposition [ solverprop ] .
assume the @xmath122 are such that , for some @xmath276 $ ] , we have @xmath675 then for every @xmath204 , we have @xmath676 our strategy for proving will be to first show that the minimum of the ratios @xmath678 and @xmath265 is attained at the off - diagonal corners ; that is , @xmath679 and @xmath680 .
the proofs are nearly identical in both cases , so we will just give the discrete proof .
then we estimate the size of this value ; here the proofs are different , and we can get a sharper estimate for the continuous case . first we show that we can decrease @xmath686 by increasing the larger index . if @xmath687 we know from proposition [ greenupperbounds ] that @xmath688 , so that @xmath689 , and thus @xmath690 . thus @xmath691 next we show that we can decrease @xmath686 by decreasing the smaller index , which is a bit more involved . inspired by , we define for @xmath145 the auxiliary quantity @xmath692 ; then it is easy to compute that @xmath693 , and we conclude using that if @xmath694 then @xmath695 . since @xmath696 , this shows that @xmath697 as long as @xmath698 .
then since @xmath699 for @xmath700 , we have @xmath701 as long as @xmath702 , and hence @xmath703 combining and , and using the fact that @xmath704 , we obtain @xmath705 we finally want to bound @xmath706 from below . using the formula we have that @xmath707 it is easier to estimate sums than products , so we rewrite as @xmath708 recalling that @xmath709 .
first we get an upper estimate for @xmath710 . rearranging and using @xmath711
, we have @xmath712 recalling that @xmath713 for each @xmath116 , we conclude @xmath714 and since @xmath709 , we find @xmath715 next we get a lower estimate for @xmath720 . since @xmath721 , the assumption yields @xmath722 .
since we have assumed @xmath723 , we have @xmath724 for all @xmath725 .
now we want to get a lower bound for @xmath726 ; this is a bit more delicate than an upper bound for the logarithm of @xmath727 .
define @xmath728 in terms of the riemann zeta function .
it is not difficult to verify that when @xmath729 , then @xmath730 .
thus we have @xmath731 for every @xmath116 , from which we conclude @xmath732 now we will just sketch the proof of .
we similarly establish that the infimum of @xmath265 is attained when @xmath46 and @xmath734 , which works the same way as in the discrete case .
so we just need to estimate @xmath735 . letting @xmath736
, we see that @xmath737 satisfies @xmath738 the minimum is then @xmath739 .
set @xmath740}$ ] ; then we can verify by explicit computation that can be rewritten in two ways : @xmath741 & = -s^2 \lambda'(s)^2 + s^2 \lvert \eta''(s)\rvert^2 , \text { and}\\ \frac{d}{ds } \big [ s(1-s)\lambda'(s)\big ] + \lambda'(s ) & = -s(1-s ) \lambda'(s)^2 + s(1-s ) \lvert \eta''(s)\rvert^2.\end{aligned}\ ] ] integrating the first equation from @xmath46 to @xmath55 gives @xmath742 and integrating the second from @xmath46 to @xmath55 gives @xmath743 since @xmath744 and @xmath745}$ ] , estimates and combine to give .
let @xmath22\to \mathbb{r}^d$ ] be @xmath44 .
then for any @xmath337 the norms and satisfy the weighted inequalities @xmath746 if in addition we have @xmath339 , then these inequalities can be simplified to @xmath747 it is clearly sufficient to prove these inequalities when @xmath748 . to derive the discrete versions
, we do the following .
let @xmath356 be any real number .
then for any @xmath749 , we have @xmath750 so that @xmath751 note that if we define @xmath752 in any way at all , the equation is still satisfied at @xmath753 .
furthermore we have the easy - to - verify formulas @xmath754 and @xmath755 , which are valid for @xmath756 . for any real @xmath757 , multiply through by @xmath758 and simplify to get @xmath759 now notice that the last term simplifies to @xmath760 \\
& = p{\nabla_+}\big ( s_k^{(q ) } \lvert f_k\rvert^2 \big ) - pq s_{k+1}^{(q-1 ) } \lvert f_{k+1}\rvert^2,\end{aligned}\ ] ] using @xmath761 .
thus we have @xmath762 now let @xmath195 and @xmath118 be any integers with @xmath763 .
summing all the terms from @xmath764 to @xmath765 and using the telescope formula @xmath766 for any sequence @xmath767 , we obtain @xmath768 after reindexing the first sum on the right side .
this is the basic building block for all the other inequalities in this proof .
now we consider some special cases which will together prove . take any @xmath337 .
* for any integer @xmath195 with @xmath769 , if we set @xmath770 and @xmath263 in , we get @xmath771 and since @xmath772 , we have @xmath773 we use this to obtain ; if @xmath774 we obtain . *
next , if we set @xmath775 , @xmath776 , @xmath777 , and @xmath263 , then we get @xmath778 noting that @xmath779 for any @xmath337 and any @xmath116 , we obtain after solving for @xmath780 that @xmath781 which is used to bound . if @xmath774 we obtain . *
finally we get an upper bound for @xmath782 .
choose @xmath783 and @xmath784 with @xmath777 and @xmath263
. then we have @xmath785 from which we conclude @xmath786 applying the shift operator @xmath128 to , we obtain @xmath791 thus as a first step , we have @xmath792 using the fact that @xmath128 is a bounded operator in any norm .
( the technique we use will make it clear that the norm of @xmath793 is comparable to that of @xmath794 , so there is no reason to study it separately . ) our first goal is to bound @xmath798 in terms of the norms of @xmath13 and the norms of @xmath126 .
note that we have bounds on @xmath13 and @xmath799 in the maximum norm by lemma [ acboundslemma ] , while for higher differences of @xmath13 the bounds are expressed in terms of euclidean - type norms .
so the complication comes from taking this into account . using the general product formula for differences
, we have @xmath800 now the exceptional cases are when @xmath801 or @xmath802 , because there we want to use the maximum norm on @xmath13 directly . in the other cases
, we still need to use the maximum norm , but we will give it some extra weighting so that we can use .
so we have , using @xmath803 and @xmath804 , that @xmath805 using the bounds @xmath806 and @xmath807 from proposition [ weightsprop ] , we obtain @xmath808 to conclude , we need to estimate the other term in .
we will show @xmath810 we will see that the extra half - power in the weighting is only necessary for the norm of @xmath368 ; otherwise we could have worked with @xmath811 , etc . instead .
we have @xmath812 recall from lemma [ acboundslemma ] that @xmath814 , so that @xmath815 .
so when @xmath816 or @xmath817 we get @xmath818 and @xmath819 . from this
we have @xmath820 and @xmath821 .
now we can derive from for @xmath4 the recursive inequality @xmath822 plugging in @xmath5 gives the base case , and induction on @xmath20 gives .
then the energies and satisfy the estimates @xmath823 for some @xmath482 independent of the initial data and of @xmath6 .
in addition the higher energies satisfy @xmath824 for every @xmath469 , where @xmath484 depends only on @xmath475 . as with the proof of lemma [ sigmasobolevlemma ] ,
the estimates for the whip and chain are proved in the same way , so we will just focus on the harder case of the chain ( where nontrivial technical issues such as lemma [ notnaivelemma ] arise ) .
the essential step is the discrete analogue of the computation , together with the integration by parts employed to cancel out the highest - order term .
then we simply estimate the remainder terms using corollary [ supproductboundcorollary ] and lemma [ normtoenergy ] . for ,
if @xmath817 then we have @xmath829 , while if @xmath830 the derivative @xmath831 is a derivative of a product of @xmath832 terms . using @xmath833 and @xmath829 , we clearly have @xmath834 which obviously holds in addition for @xmath816 and @xmath817 .
thus the sum of is bounded by @xmath835 using lemma [ acboundslemma ] .
so our primary concern is .
by analogy with the technique used to derive , we want to pull out the worst terms of and collect them into a single exact difference which will sum to zero .
it is easy to check that if @xmath836 , then @xmath837 the middle term of @xmath838 is precisely the second term of , so we want to show that the difference of the remaining terms is relatively simple . :
we have nothing but @xmath132 in the rest of the formula , so the appearance of one @xmath136 operator ( or , equivalently , of one backward shift @xmath130 ) necessitates rederiving the formulas to get rid of it , rather than using a formula like directly .
the reason we do nt want to see a mix of @xmath136 and @xmath132 operators is because later in the proof we will need to use the fact that @xmath839 , while there is no simple formula for @xmath840 . ] we will do this computation in lemma [ notnaivelemma ] . to check that this all works , we observe that the backward difference @xmath838 sums to zero : we have @xmath841 which can be checked to vanish due to @xmath104 and the oddness conditions .
( recall that this is precisely the reason that the summands in our energies all terminate at @xmath842 . )
hence we have @xmath843 therefore if @xmath846 we have @xmath847 where @xmath848 for @xmath849 . here
we use the fact that by proposition [ weightsprop ] , the shift operators @xmath850 that appear in are bounded for any @xmath118 ; since having or not having the shift operators attached to @xmath13 does nt change the estimates in any way , we may as well ignore them . ( having the shift operators attached to one of the @xmath34 terms _ would _ cause a problem , which is why we need lemma [ notnaivelemma ] . ) obviously when @xmath817 , we must have @xmath851 also in the sum , and in this case every summand in involves @xmath852 , so that @xmath853 if @xmath817
. hence we will assume @xmath830 to estimate @xmath854 .
we will show that for any @xmath849 , @xmath855 for some function @xmath856 of @xmath857 .
the bounds for @xmath858 are all basically the same , while the bound for the @xmath859 term requires another trick .
using corollary [ supproductboundcorollary ] , and lemma [ normtoenergy ] , it is straightforward to verify that @xmath860 and that @xmath861 for @xmath862 . then using lemma [ sigmasobolevlemma ]
, we obtain when @xmath858 .
the last case in is when @xmath859 .
( recall that we are assuming @xmath830 since @xmath863 when @xmath817 . ) since @xmath864 , we have @xmath839 . applying the difference operator @xmath865 to both sides and using the product formula
, we get @xmath866 thus @xmath867 we obviously want to use the cauchy - schwarz inequality on this , and to do so we have to assign the weight @xmath868 to the individual pieces . recall that we have an estimate for @xmath869 from lemma [ sigmasobolevlemma ] , which means we must pull the power @xmath870 out with this term , leaving us with @xmath871 for what remains .
the worst term is the one with @xmath872 , for then we have to estimate @xmath873 .
we want to pull out the supremum norm of @xmath193 , but we need some positive weight on it in order to be able to use .. this is why the sobolev norms of @xmath13 from lemma [ sigmasobolevlemma ] are defined the way they are . ] suppose @xmath583 and @xmath198 are functions on @xmath215 $ ] as in lemma [ uniformestimateslemma ] . if @xmath644 and @xmath645 are two solutions of in @xmath35 , n_4[0,1])\cap w^{1,\infty}([0,t ] , n_3[0,1])$ ] , with the same initial conditions @xmath647 then @xmath648 and @xmath649 for all @xmath650 $ ] and all @xmath651 $ ] .
the quantities @xmath893 and @xmath894 have the same boundary conditions as those for @xmath656 and @xmath657 ; similarly @xmath895 and @xmath896 must both be odd through @xmath55 since @xmath654 and @xmath655 are . furthermore , the fact that @xmath897 implies that @xmath898 .
we now estimate the norms of these quantities ; the primary goal is to estimate the norm of @xmath896 , but we will need the norms of the other terms to do this . for this purpose , we generalize the quantities @xmath197 , @xmath408 , @xmath334 in [ abcdef ] , the quantities @xmath899 from , the quantities @xmath310 from , and the quantities @xmath309 from : we will denote @xmath900 = \sup_{0\le s\le 1 } \frac{\lvert \overline{\sigma}(s)\rvert}{s}$ ] , @xmath901 = \sum_{\ell=0}^m \lvert { \varepsilon_t}\rvert^2_{\ell,\ell } + \lvert \varepsilon\rvert^2_{\ell+1 , \ell+1}$ ] , etc . for the time - dependent energy @xmath309 we use @xmath893 for the weighting .
it is easy to verify that @xmath902 & \le \tfrac{1}{2 } a[\sigma_1 ] + \tfrac{1}{2 } a[\sigma_2 ] , & \qquad b[\overline{\sigma } ] & \le b[\sigma_1 ] + b[\sigma_2]\\ c[\overline{\sigma } ] & \le \tfrac{1}{2 } c[\sigma_1 ] + \tfrac{1}{2 } c[\sigma_2 ] , & \qquad d_m[\overline{\sigma } ] & \le d_m[\sigma_1 ] + d_m[\sigma_2],\\ e_m[\overline{\eta } ] & \le e_m[\eta_1 ] + e_m[\eta_2].&&\end{aligned}\ ] ] we could proceed by imitating the proof of theorem [ discreteenergyestimate ] to get a bound for the energy @xmath903 $ ] ; however it s simpler to use some alternative techniques to get a bound for @xmath904 $ ] .
the reason this works is that we can separate all the estimates into low - derivative norms of @xmath896 by compensating with high - derivative norms of @xmath895 .
first we note that since @xmath905 , we have by the cauchy - schwarz inequality that @xmath906 . in the notation of definition [ weightednorms ] , we conclude that @xmath907 since @xmath905 and @xmath908 , we have using and corollary [ supproductboundcorollary ] that @xmath909 using the fact that @xmath910 $ ] , along with the inequality , we conclude @xmath911 \lvert \varepsilon\rvert_{2,2 } \lvert \overline{\eta}\rvert_{1,2 } + 2\lvert { \varepsilon_t}\rvert_{1,1 } \lvert { \overline{\eta}_t}\rvert_{0,1 } \\ & \lesssim ( 1+a[\overline{\sigma } ] ) \sqrt{e_1[\varepsilon ] } \sqrt{e_2[\overline{\eta } ] } \lesssim ( e_2[\overline{\eta}])^{3/2 } \sqrt{e_1[\varepsilon ] } , \end{split}\ ] ] using lemma [ normtoenergy ] and lemma [ acboundslemma ] .
we compute the energies of @xmath896 using the same technique as in theorem [ discreteenergyestimate ] : we try to bound @xmath912 in terms of @xmath310 for @xmath748 and @xmath913 .
the lowest one is easy : we have by that @xmath914}{dt } & = \frac{d}{dt } \int_0 ^ 1 \big ( \lvert { \varepsilon_t}\rvert^2 + \overline{\sigma } \lvert { \varepsilon_s}\rvert^2\big ) \ , ds \\ & = c[\overline{\sigma } ] e_0[\varepsilon ] \textstyle + 2 \int_0 ^ 1 { \delta_{s}}\langle { \overline{\eta}_s } , { \varepsilon_t}\rangle \ , ds + 2\int_0 ^ 1 \delta \langle { \overline{\eta}_{ss } } , { \varepsilon_t}\rangle \ , ds,\end{aligned}\ ] ] since the boundary term vanishes
. then we can estimate the rest : @xmath915}{dt } & \le c[\overline{\sigma } ] e_0[\varepsilon ] + 2 { \vvvert \overline{\eta}\vvvert}_{0,1 } \lvert \delta\rvert_{0,1 } \lvert { \varepsilon_t}\rvert_{0,0 } + 2 { \vvvert \delta\vvvert}_{-1,0 } \lvert \overline{\eta}\rvert_{1,2 } \lvert { \varepsilon_t}\rvert_{0,0 } \\ & \le c[\overline{\sigma } ] e_0[\varepsilon ] + 2 \lvert \delta\rvert_{0,1 } \sqrt{e_0[\varepsilon ] } \big ( { \vvvert \overline{\eta}\vvvert}_{0,1 } + \lvert \overline{\eta}\rvert_{1,2}\big ) .
\end{split}\ ] ] now from lemma [ normtoenergy ] we have @xmath916}$ ] , while the term @xmath917 is a bit more difficult ( since we have no weighting on the supremum and ca nt use ) .
instead we use the standard sobolev inequality : @xmath918,\ ] ] using lemma [ normtoenergy ] .
this yields @xmath919}{dt } \lesssim e_2[\overline{\eta}]^{3/2 } e_3[\overline{\eta}]^{1/2 } \sqrt{e_0[\varepsilon ] } \sqrt{e_1[\varepsilon]}.\ ] ] we thus can not bound @xmath920 $ ] without also bounding @xmath904 $ ] ; this is not surprising since it s hard to even make sense of equation without @xmath293 and @xmath221 both being in @xmath294 .
our next step is to perform the same estimates for @xmath921 $ ] , at which point the estimates do close up .
we can compute that @xmath922 \lvert { \varepsilon_t}\rvert^2_{1,1 } + 2c[\overline{\sigma}]a[\overline{\sigma } ] \lvert \varepsilon\rvert^2_{2,2 } \\ & \qquad\qquad\qquad + \textstyle 2\int_0 ^ 1 \overline{\sigma } { \overline{\sigma}_{ss}}\langle{\varepsilon_s } , { \varepsilon_{st}}\rangle \ , ds + 2\int_0 ^ 1 \overline{\sigma } { \delta_{ss}}\langle { \overline{\eta}_s } , { \varepsilon_{st}}\rangle \ , ds \\ & \qquad\qquad\qquad \textstyle + 4\int_0 ^ 1 \overline{\sigma } { \delta_{s}}\langle { \overline{\eta}_{ss } } , { \varepsilon_{st}}\rangle \ , ds + 2\int_0 ^ 1 \overline{\sigma } \delta \langle { \overline{\eta}_{sss } } , { \varepsilon_{st}}\rangle \ , ds , \end{split}\ ] ] where again the boundary term vanishes .
we also want to use a trick to simplify the estimates a bit the same trick we used in deriving to reduce the derivatives on @xmath896 to compensate for the high - derivative term @xmath923 .
since @xmath924 , we have @xmath925 , and we use this to write @xmath926 . now becomes @xmath927(1+a[\overline{\sigma } ] ) e_1[\varepsilon ] + 2\int_0 ^ 1 \overline{\sigma } \lvert { \overline{\sigma}_{ss}}\rvert \lvert { \varepsilon_s}\rvert \lvert { \varepsilon_{st}}\rvert \ , ds \\ \textstyle + 4\int_0 ^ 1 \overline{\sigma } \lvert { \delta_{s}}\rvert \lvert { \overline{\eta}_{ss}}\rvert \lvert { \varepsilon_{st}}\rvert \ , ds + 2 \int_0 ^ 1 \overline{\sigma } \lvert \delta\rvert \lvert { \overline{\eta}_{sss}}\rvert \lvert { \varepsilon_{st}}\rvert \ , ds + 2 \int_0 ^ 1 \overline{\sigma } \lvert { \delta_{ss}}\rvert \lvert { \overline{\eta}_{st}}\rvert \lvert { \varepsilon_s}\rvert \ , ds.\end{gathered}\ ] ] using corollary [ supproductboundcorollary ] , we can easily bound all but the last term : @xmath928(1+a[\overline{\sigma } ] ) e_1[\varepsilon ] \\ + 2a[\overline{\sigma } ] \lvert { \varepsilon_t}\rvert_{1,1 } \big ( \sqrt{d_3[\overline{\sigma } ] e_1[\varepsilon ] } + \sqrt{e_3[\overline{\eta } ] } \lvert \delta\rvert_{0,1 } \big ) + \textstyle 2 a[\overline{\sigma } ] \int_0 ^ 1 s \lvert { \delta_{ss}}\rvert \lvert { \overline{\eta}_{st}}\rvert \lvert { \varepsilon_s}\rvert \ , ds.\end{gathered}\ ] ] we can bound @xmath929 \lesssim e_3[\overline{\eta}]^3 $ ] using lemma [ sigmasobolevlemma ] , while bounds @xmath930
. finally we deal with the last term of by plugging in and using the fact that @xmath931 in order to eliminate the seemingly nonlinear dependence on @xmath896 . after a series of computations as above ,
we obtain @xmath932 ^ 3 e_1[\varepsilon].\ ] ] therefore becomes @xmath933 ^ 3 e_1[\varepsilon],\ ] ] and combining this with , we obtain @xmath934}{dt } \lesssim e_3[\overline{\eta}]^3 e_1[\varepsilon].\ ] ] now using the inequality @xmath935\overline{\sigma}(s)$ ] for all @xmath162 , we bound @xmath904 $ ] in terms of @xmath921 $ ] : @xmath936 \le \left(1+b[\overline{\sigma}]\right)^2 \tilde{e}_1[\varepsilon].\ ] ] using @xmath937\le b[\sigma_1]+b[\sigma_2]$ ] and the bound for @xmath938 $ ] and @xmath939 $ ] in terms of @xmath940 $ ] and @xmath941 $ ] respectively , we ultimately find that @xmath942 where @xmath943 is a function depending only on the energies @xmath944 $ ] and @xmath945 $ ] , which are uniformly bounded by assumption . using gronwall s inequality , we conclude that if @xmath946 , then @xmath947 for all time .
in particular we conclude that @xmath948 for all @xmath949 $ ] , so that @xmath950 for all @xmath949 $ ] and @xmath951 $ ] . since @xmath952
, we must have @xmath953 for all @xmath56 and @xmath162 , whence we conclude @xmath954 for all @xmath56 and @xmath162 . the fact that @xmath955 follows . | we study the motion of an inextensible string ( a whip ) fixed at one point in the absence of gravity , satisfying the equations @xmath0 with boundary conditions @xmath1 and @xmath2 .
we prove local existence and uniqueness in the space defined by the weighted sobolev energy @xmath3 when @xmath4 .
in addition we show persistence of smooth solutions as long as the energy for @xmath5 remains bounded .
we do this via the method of lines , approximating with a discrete system of coupled pendula ( a chain ) for which the same estimates hold . |
the nonlinear schrdinger equation is a typical soliton equation.@xcite solitons in optical fibers are described by the nonlinear schrdinger equation and have been intensively studied for the optical communication.@xcite recently , solitons and dark solitons were observed in the bose - einstein condensates ( becs).@xcite dark ( bright ) solitons are found in the becs with repulsive ( attractive ) interaction .
the nonlinear schrdinger equation is called the gross - pitaevskii equation in the research field of becs . in the becs ,
matter waves obey the nonlinear schrdinger equation with external potentials .
the harmonic potentials are used to confine the becs .
interference patterns of laser beams generate external periodic potentials .
solitons in external potentials have been studied by several authors.@xcite especially , solitons in periodic potentials ( optical lattices ) are intensively studied.@xcite however , the scattering of a soliton by a potential wall was not studied in detail , although it is a simple and a fundamental process .
the scattering problem of a quantum particle by the potential wall , which is described by the linear schrdinger equation , is a typical problem of the quantum mechanics . in our previous letter
, we studied the scattering and trapping of solitons by a potential wall and well .
we found that the soliton behaves like a classical mechanical particle under a certain condition , and evaluate the critical energies for the transmission . in this paper
, we study a transition between the classical mechanical behavior and the wavy behavior more in detail .
we will study also the scattering of a dark soliton by the potential wall .
we will further study the scattering of a two - dimensional soliton .
the nonlinear schrdinger equation with a potential term is written as @xmath0 where @xmath1 is the wave function , @xmath2 is the coefficient of the nonlinear term , @xmath3 is an external potential .
first , we consider the case of @xmath4 .
it corresponds to the case of the attractive interaction in the problem of becs .
when @xmath5 , the nonlinear schrdinger equation with @xmath4 has a soliton solution @xmath6 where @xmath7 and @xmath8 . in this section ,
we study the scattering by a simplest potential wall , which is expressed as @xmath9 where @xmath10 is the width of the potential wall and @xmath11 is the height .
we study numerically the scattering of a soliton by the potential wall .
we perform numerical simulations using the split - step fourier method with fourier modes 1024 or 2048 .
the periodic boundary conditions are assumed , and the time step is 0.001 .
figure 1(a ) displays the time evolution of a soliton passing through a potential wall with @xmath12 and @xmath13 .
the system size is @xmath14 and the potential wall locates at @xmath15 .
the initial velocity of the soliton is @xmath16 and the amplitude is @xmath17 . the soliton is split into two pulses by the potential wall .
the transmission coefficient is 0.488 , which is calculated as @xmath18 after the scattering .
figure 1(b ) displays the transmission coefficient @xmath19 as a function of the initial velocity @xmath20 for three values of @xmath21 for @xmath17 and @xmath13 . for @xmath22 , the transmission coefficient changes from 0 to 1 abruptly .
the soliton is completely reflected below the critical value of @xmath20 , and completely transmitted above the critical velocity .
the soliton behaves like a classical mechanical particle .
we showed that the critical kinetic energy depends strongly on @xmath23 as @xmath24 using the lagrangian formalism.@xcite for @xmath25 , the transmission coefficient changes smoothly from 0 to 1 .
this behavior is characteristic of the transmission of waves or the quantum mechanical particle . near @xmath26 , the scattering changes from the particlelike behavior to the wavy behavior for the amplitude @xmath17 .
the transmission coefficient changes in a non - monotonic manner near the critical velocity , although we do not understand the reason yet .
we discussed qualitatively the transition using the conservation law of the energy in the previous paper .
we confirm the conjecture with the direct numerical simulations . for @xmath27 and @xmath16 .
a soliton is splitted into two pulses by the potential wall of @xmath13 located at @xmath28 .
( b ) transmission coefficient @xmath19 as a function of the initial velocity @xmath20 .
( c ) a boundary line between the particlelike behavior and the wavy behavior . under the line
, a soliton behaves like a particle .
the solid line is the theoretical estimate and the rhombi denote the critical value of @xmath21 obtained numerically.,width=491 ] and @xmath29 .
a soliton penetrates into a potential wall , and then is trapped between the double walls .
the walls located at @xmath30 and @xmath31 .
( b ) reflection coefficient @xmath32 ( solid line with rhombi ) , the trapping coefficient @xmath33 ( dashed line with crosses ) and the transmission coefficient @xmath19 ( dotted line with squares ) as a function of the initial velocity @xmath20 .
( c ) time evolution of the peak position of the soliton ( dashed line ) and the theoretical model of eq .
( 6 ) ( solid line).,width=491 ] and @xmath34 .
( b ) generation of a bound state at the collision for @xmath35 and @xmath36 .
( c ) reflectivity @xmath32 ( solid line ) , trapping rate @xmath33 ( dashed line ) and transmissivity @xmath19 ( dotted line ) as a function of @xmath20 for @xmath37 and @xmath38.,width=566 ] we first explain a criterion based on the conservation laws .
we assume that a soliton is located initially at @xmath39 with velocity 0 .
if the height @xmath21 of the potential wall is sufficiently large , the soliton will be split into two pulses .
however , the soliton can not be split , if @xmath21 is not so large owing to the attractive mutual interaction .
equation ( 1 ) has at least two invariants , that is , the total mass @xmath40 and the total energy @xmath41 . if the soliton is split into two counterpropagating solitons : @xmath42 , the amplitude @xmath43 and the momentum @xmath44 satisfy @xmath45 and @xmath46 owing to the conservation laws . if @xmath47 , there is no real solution for @xmath44 , that is , the splitting can not occur and the one - pulse structure is maintained . above the line @xmath48 , the potential energy is used to split a soliton against the attractive interaction .
that is , the above argument suggests that a criterion between keeping one - pulse state and splitting into two pulses is @xmath49 . to check the conjecture of the criterion
, we have performed numerical simulations , changing the value of @xmath21 for several @xmath23 s , and constructed the same type plots as fig .
we have regarded the critical value of @xmath50 for each @xmath23 as the lowest value of @xmath21 , at which the transmission rate @xmath19 has a jump across @xmath51 .
figure 1(c ) displays the critical values @xmath52 for @xmath53 and 2 .
the dashed line is the theoretical estimate @xmath49 .
the transition between the particlelike behavior to the wavy behavior is not so clear - cut , and so the above theoretical curve gives a fairly good criterion .
the splitting of the one - soliton was caused by the perturbation at the potential wall .
it is desirable to determine the reflection and transmission coefficients theoretically using the inverse scattering method or other perturbative methods but we have not succeeded in it yet .
in this section , we consider the scattering of bright solitons by a potential with double walls : @xmath54 there are two potential wall at @xmath55 and @xmath56 .
figure 2(a ) displays a time evolution of a soliton with @xmath38 and the initial velocity @xmath29 .
the system size is 50 , and the parameters of the potential are @xmath22 , @xmath57 , @xmath30 , @xmath31 and @xmath13 .
the soliton passes over the first potential wall , but is reflected by the second wall , and then trapped in an interspace between the two walls .
figure 2(b ) displays the reflectivity @xmath32 , the transmissivity @xmath19 , and the trapping rate @xmath33 , which are defined as @xmath58 , @xmath59 and @xmath60 after the scattering .
the reflectivity and the transmissivity changes from 0 to 1 abruptly , and the soliton behaves like a particle .
the soliton is trapped between the double walls in a narrow parameter region of @xmath61 .
this trapping phenomenon can be described also using the lagrangian formalism .
the nonlinear schrdinger equation is rewritten with the lagrangian form : @xmath62 , where @xmath63 . as is used in the previous paper ,
we assume a solution of the form @xmath64 where @xmath65 and @xmath66 change in time , but @xmath67 is conserved in time . the term @xmath68 is a kind of chirp term .
owing to the inclusion of this term , the width of the soliton changes in time and breathing motion becomes possible .
some energy is given to the breathing motion and the kinetic energy may be decreased .
substitution of eq .
( 5 ) into the lagrangian and its spatial integration yields an effective lagrangian @xmath69 where @xmath70 $ ] for @xmath57 .
the equations of motion for @xmath71 and @xmath72 are @xmath73 this equation is derived from a hamiltonian @xmath74 as @xmath75 .
the classical mechanical energy @xmath76 is not conserved in this equation , and the width @xmath77 changes in time . if the classical mechanical energy @xmath78 decays in time , the soliton can be trapped between the double walls . figure 2(c ) compares the time evolutions of @xmath79 by eq .
( 6 ) with the peak position of the soliton by eq .
( 1 ) for the initial velocity @xmath80 .
the trapping phenomenon may be qualitatively explained by eq .
when @xmath21 and @xmath81 become larger , a soliton is split into two or three solitons .
we show an example of the scattering of a soliton leaving a bound state behind .
figure 3 displays the soliton scattering with amplitude @xmath38 at @xmath82 and @xmath83 .
the system size is 100 .
figure 3(a ) displays the reflection process of the soliton for the initial velocity @xmath34 . for this velocity ,
the reflectivity is almost 1 .
but for @xmath84 , the reflectivity decreases abruptly to 0.675 , and a bound state is created in the potential well .
figure 3(b ) displays the reflection process leaving a bound state behind for @xmath36 .
the bound state has a norm 0.688 and is almost stationary .
the bound state is antisymmetric around the center , therefore , the amplitude @xmath85 is 0 at the center .
the velocity of the reflected pulse is @xmath86 , and it is larger than the initial velocity .
that is , the reflecting pulse is accelerated by leaving a bound state with low energy behind .
figure 3(c ) displays the reflectivity @xmath32 , the transmissivity @xmath19 and the trapping rate @xmath33 as a function of the initial velocity .
there is a jump of the reflectivity at @xmath34 .
the bound state is created abruptly for @xmath87 .
the transmissivity @xmath19 increases continuously from @xmath88 . near @xmath89 , a soliton
is split into three pulses , i.e.,a reflecting pulse , a bound state and the transmitted pulse . for @xmath90 and @xmath91 .
a dark soliton is reflected by the potential wall of @xmath13 located at @xmath92 .
( b ) time evolution of @xmath85 for @xmath93 and @xmath91 .
a dark soliton passes over the potential wall .
( c ) time evolution of the minimum value of @xmath85 for @xmath94 ( solid line ) and @xmath95 ( dashed line ) .
, width=566 ] of the potential wall , above which the dark soliton is reflected .
the solid line is the theoretical estimate using the energy .
the dashed line is the twice of that.,width=226 ] for the double potential walls at @xmath96 and @xmath97 with @xmath98 , @xmath38 and @xmath91 .
a dark soliton goes out from the interspace between the double walls after several reflections .
( b ) time evolution of the minimum point of @xmath85 .
( c ) time evolution of the minimum value of @xmath85.,width=529 ]
if the interaction is repulsive in the becs , the nonlinear schrdinger equation has a form @xmath99 that is , the sign of the nonlinear term becomes positive and the coefficient is assumed to be 1 .
a dark soliton is a special solution to this nonlinear schrdinger equation .
a dark soliton has a form : @xmath100 where @xmath101 , @xmath23 and @xmath102 are parameters which characterizes the dark soliton and the velocity @xmath20 is given by @xmath103 . for @xmath104 ,
the amplitude @xmath85 becomes 0 at the center . for @xmath105 ,
the amplitude does not become 0 , and it is sometimes called a grey soliton .
we consider the case of the potential given by eq .
the energy of the dark soliton is given by @xmath106 , where @xmath79 is the position of the dark soliton .
we investigate numerically the time evolution of this dark soliton .
the initial value of @xmath101 is set to be 0 and the initial velocity is given by @xmath107 .
the width of the potential wall in eq .
( 3 ) is given by @xmath13 , and the system size 50 .
in actual simulations , we have investigated the time evolution of two dark solitons ( another dark soliton is located at @xmath108 ) in a system of double size @xmath14 , since we performs numerical simulations using the periodic boundary conditions .
owing the periodic boundary conditions , the initial wave number needs to be @xmath109 at @xmath110 , where @xmath111 is an integer .
figure 4(a ) displays the reflection process of the dark soliton by the potential wall for @xmath94 , @xmath112 , and @xmath38 .
figure 4(b ) displays the transmission process for @xmath95 , @xmath91 and @xmath38 .
the dark soliton approaches plane waves in the infinity , and the scattering of the plane waves by the potential wall always occurs , before the center of the dark soliton reaches the potential wall . in fig .
4(a ) and ( b ) , the amplitude @xmath85 is slightly depressed owing to the scattering near the position @xmath113 of the potential wall , but the effect of the scattering of the plane waves seems to be small , when the potential height @xmath21 is small .
figure 4(c ) displays the time evolution of the minimum value of @xmath85 for @xmath94 and 0.08 .
if the dark soliton keeps the form ( 8) , the minimum value of @xmath85 represents @xmath114 .
the initial value of @xmath101 is assumed to be 0 . in the case of the reflection
, @xmath101 is decreased to -0.40 , and the velocity of the reflected dark soliton is numerically estimated as -0.20 , which is close to @xmath115 .
the velocity of the reflected dark soliton is larger than the velocity @xmath116 of the initial dark soliton . since the energy and the norm are conserved , @xmath117 and @xmath118 .
the @xmath101-value is expected as -0.404 from the two conservation laws if @xmath102 is fixed to be the initial value , and it is comparable to the numerical value . when the dark soliton passes over the potential wall , the @xmath101-value is decreased to -0.16 near the potential wall , but it recovers to 0 as the dark soliton goes away from the potential wall .
there is a critical value of @xmath21 for the reflection to occur .
figure 5 displays the critical value of @xmath21 as a function of the initial value @xmath119 of @xmath23 above which the reflection is numerically observed for the initial velocity @xmath91 .
the critical value increases as @xmath119 is decreased .
it means that the dark soliton penetrates more easily the potential wall , as the amplitude is decreased .
this effect is similar to the one we have already seen for bright solitons as coherent tunneling .
the critical value may be similarly evaluated using the energy conservation law , if the form of the dark soliton is assumed to be given by eq .
at the critical value of @xmath21 , the dark soliton will stop at @xmath39 .
then , the @xmath101-value is expected to be @xmath120 , since the velocity @xmath121 is 0 .
owing to the energy conservation law , @xmath122 is satisfied .
the critical vale of @xmath21 is evaluated as @xmath123 .
a similar result can be derived from the lagrangian formalism .
we assume a solution of the form @xmath124 substitute it to the lagrangian @xmath125 yields an effective lagrangian @xmath126@xmath127 where @xmath128 and @xmath129 .
we can obtain equations of motion from the effective lagrangian as @xmath130 where @xmath131 $ ] .
similar critical value of @xmath21 can be evaluated using this equation , which is almost the same as the one evaluated using the energy . however , the critical value of @xmath21 estimated by the above theory is close to half of the critical value by numerical simulations as shown in fig .
the above equations of motion are approximated to @xmath132 when @xmath20 and @xmath21 are sufficiently small and then @xmath114 is small , which is the equations of motion for a classical particle of mass 1 . on the other hand , busch and
anglin derived equations of motion of the form @xmath133 when the potential @xmath3 changes very slowly compared to the soliton scale,@xcite using a multiple time scale boundary layer theory . in their theory , the background amplitude @xmath134 and the velocity are assumed to change slowly .
the resultant effective mass of the dark soliton becomes 2 .
considering the factor 1/2 , it is reasonable that the numerical critical value of @xmath21 is twice of the theoretical estimate .
that is , our evaluation based on the energy and the lagrangian fails to estimate the critical value of @xmath21 . probably , it is necessary to involve the effect of the variation of the background component @xmath135 and @xmath102 to the lagrangian formalism , although we have not succeeded yet . for a bright soliton ,
the kinetic energy is somewhat decreased and can be trapped in an interspace between double potential walls , even if the soliton is initially outside of the potential walls .
on the other hand , the dark soliton can get some kinetic energy at the scattering and run away from the interspace between the double walls , because the dark soliton with smaller @xmath101-value is energetically disadvantageous . to show the detrapping , we have used the double potential walls of eq .
the parameters are @xmath22 , @xmath57 , @xmath96 and @xmath97 and @xmath13 .
the initial amplitude @xmath23 is 1 and the initial velocity is @xmath91 , and the initial position is @xmath136 . figure 6(a ) displays the time evolution of the dark soliton .
figure 6(b ) displays the time evolution of the minimum point of @xmath85 .
figure 6(c ) displays the time evolution of the minimum value of @xmath85 .
the maximum of the @xmath114-value increases slowly each time the dark soliton is reflected by the second wall , and finally the dark soliton overcomes the potential wall .
this detrapping phenomenon can be described by eq .
( 9 ) , although it takes more time to overcome the potential walls in eq .
we need to improve the equations of motion to describe the detrapping phenomenon more quantitatively . for the two - dimensional quintic nonlinear schrdinger equation with @xmath137 .
( b ) time evolution of @xmath138 for the initial velocity @xmath139 .
( c ) time evolution of @xmath138 for the initial velocity @xmath140.,width=529 ] and @xmath141 , where @xmath142 is the peak position of the two - dimensional soliton .
( b ) trajectory @xmath143 of the peak position in the @xmath144 plane .
the initial velocity is @xmath145 .
the dashed line denotes a boundary @xmath113 of the stepwise potential.,width=340 ]
the standard two dimensional nonlinear schrdinger equation does not have soliton type solutions .
even if a localized initial condition is assumed , it may extend and changes into radiations or leads to collapse , depending on the initial conditions .
however , two dimensional solitonlike solutions exist in the quintic nonlinear schrdinger equation.@xcite the two - dimensional quintic nonlinear schrdinger equation is not an integrable system , so it may not be suitable to call the localized solution a soliton , but we say it a two - dimensional soliton in this paper by analogy with the one - dimensional system .
we study the scattering and the refraction of the two - dimensional soliton using the quintic nonlinear schrdinger equation : @xmath146 equation ( 10 ) has an axisymmetric localized solution for @xmath147 .
the localized solution has a form @xmath148 , where @xmath149 is the distance from the center of the axisymmetric localized solution .
the amplitude @xmath150 needs to satisfy @xmath151 figure 7(a ) is a numerically obtained solution to the above nonlinear eigenvalue problem for @xmath152 and @xmath137 . the amplitude @xmath153 of the two - dimensional soliton is a free parameter , as is the one - dimensional soliton .
we have checked that the axisymmetric solution is stable by direct numerical simulations of eq .
we can construct a propagating soliton with velocity @xmath154 using the solution of eq .
( 11 ) as @xmath155 .
we have performed numerical simulations of the quintic nonlinear schrdinger equation with a potential wall of @xmath156 for @xmath157 .
that is , the potential wall is parallel to the @xmath158-axis and has a width @xmath13 .
the system size is @xmath159 , and the initial wavenumber is @xmath160 , that is , the initial velocity is perpendicular to the potential wall .
figure 7(b ) and ( c ) display the time evolutions of @xmath138 for @xmath139 and 0.24 at @xmath161 .
we have checked that the form of the two - dimensionally localized structure is maintained at the scattering .
the two - dimensional soliton is reflected for @xmath139 and it has passed through the potential wall for @xmath162 .
the critical value for the transition is @xmath163 in our numerical simulation .
the lagrangian for the quintic nonlinear schrdinger equation is given by @xmath164 .
if the solution is assumed to be @xmath165 , the equation of motion for the center of mass is written using the lagrangian formalism @xmath166 as @xmath167 where @xmath168 .
for the potential wall with height @xmath21 and the width @xmath10 , the critical velocity @xmath169 is estimated from the relation @xmath170 for @xmath13 and @xmath171 , the critical initial velocity is 0.214 for @xmath150 shown in fig.7(a ) , which is consistent with the direct numerical simulation .
the critical energy is sufficiently smaller than the potential height @xmath21 similarly to the one - dimensional case .
it might also be interpreted as the coherent tunneling .
we have performed a numerical simulation for the steplike potential : @xmath5 for @xmath172 and @xmath173 for @xmath174 .
the initial velocity vector is @xmath175 , and the potential height is @xmath171 .
figure 8(a ) displays time evolutions of the peak position @xmath142 of the two - dimensional soliton .
the velocity in the @xmath176-direction changes from 0.3 to 0.1 , and the velocity in the @xmath158-direction does not change .
the velocity @xmath177 in the @xmath176-direction after the scattering satisfies the energy conservation law @xmath178 . as a result
, the trajectory of the two - dimensional soliton is refracted , which is shown in fig .
the angle of the refraction is nearly @xmath179 , and it is larger than the angle of the incidence , although the velocity becomes smaller after the refraction for @xmath174 .
this phenomenon corresponds to newton s particle theory for the refraction .
the soliton behaves like a particle , and therefore newton s theory is revived .
we have studied the scattering of bright solitons and dark solitons by the potential walls .
we have confirmed by direct numerical simulations that the boundary line of the particlelike behavior and the wavy behavior is given by @xmath49 , which was expected in our previous paper . we have found that a bright soliton can be trapped between the double potentials .
when @xmath21 is small , the soliton behaves like a particle and the soliton is completely trapped in a narrow parameter region between the reflection and the transmission .
when @xmath21 is large and @xmath81 is large , there appears a bound state in the interspace between the double walls . at the collision of a soliton to the left wall , a part of the soliton
is changed into the bound state and the rest is reflected .
the velocity of the reflected pulse can be faster than the initial velocity .
we have also studied the scattering of dark solitons . when the dark soliton is reflected ,
the @xmath101-value is decreased from 0 to a certain negative value , that is , the dark soliton is changed into a grey soliton .
we have evaluated the critical value of @xmath21 between the reflection and the transmission .
a simple theoretical estimate is about half of the numerical values .
it may be because the background density and velocity vary in space by the potential wall .
the calculation by busch and anglin is based on the approximation of the slowly varying potential .
for the potential walls whose width are comparable to or smaller than the soliton s scale , their approximation does not seem to be applicable directly , but the factor of 1/2 may be related to the effect .
we need to improve the theoretical estimate by considering the effect .
we have also found that a dark soliton can not be trapped in an interspace between the double walls , and runs naturally away from the trapping region .
we have further studied the scattering and the refraction of a two - dimensional soliton in the quintic nonlinear schrdinger equation .
we have found that the coherent tunneling occurs also for the two dimensional soliton .
we have found a phenomenon corresponding to newton s theory for the refraction .
9 v. i. karpman : _ nonlinear waves in dispersive media _ ( pergamon press , oxford , 1975 ) .
a. hasegawa : _ optical solitons in fibers _
( springer - verlag , berlin , 1990 ) . y. s. kivshar and g. p. agrawal : _ optical solitons : from fibers to photonic crystals _ ( academic press , 2003 ) . j. denschlag , j. e. simsarian , d.l.feder , c. w. clark , l. a. collins , j. cubizolles , l. deng , e. w. hagley , k. helmerson , w. p. reinhart , s. l. rolston , b. i. schneider and w. d. phillips : science * 287 * ( 2000 ) 97 . s. burger , k. bong , s. dettmer , w. etmer , k. sengsock , a. sanpera , g. v. shlyapnikov and m. lewenstein : phys . rev . lett .
* 83 * ( 1999 ) 5198 .
k. e. strecker , g. b. partridge , a.g .
truscott and r. g. hulet : nature * 417 * ( 2002 ) 150 .
l. khaykovich , f. schreck , g. ferrari , t. bourdel , j. cubizolles , l. d. carr , y. castin and c. salomon : science * 296 * ( 2002 ) 1290 .
m. de moura : j. phys . a * 27 * ( 1994 ) 7157 .
h. frauenkron and p. grassberger : phys .
e * 53 * ( 1996 ) 2823 .
a. trombettoni and a. smerzi : phys .
* 86 * ( 2001 ) 2353 .
b. b. baizakov , v. v. konotop and m. salerno : j. phys .
b * 35 * ( 2002 ) 5105 . h. sakaguchi and m. tamura : j. phys .
soc . jpn .
* 73 * ( 2004 ) 503 .
busch and j. r. anglin : phys .
* 84 * ( 2000 ) 2298 . | scattering of solitons and dark solitons by potential walls is studied in the nonlinear schrdinger equation under various conditions .
we investigate the conditions under which solitons are split into two solitons at the potential wall .
we find that a soliton can be trapped in an interspace between two potential walls .
a dark soliton can also be scattered at the potential wall . similarly to a bright soliton , a dark soliton can pass through more easily the potential wall , as the width of the dark soliton is larger .
a dark soliton can run away spontaneously from an interspace between the two potential walls .
we also study the motion of a two - dimensional soliton in a two - dimensional quintic nonlinear schrdinger equation .
we find the coherent tunneling through a potential wall , and the refraction corresponding to newton s refraction theory . |
the @xmath0 reaction constitutes an invaluable tool in the study of many diverse nuclear phenomena .
the experimental setup enables one to determine the 4-momentum of both the virtual photon and the outgoing proton , so that the reaction kinematics are completely determined . as a result
, it was originally believed that the @xmath0 cross section provides direct information about the nucleon momentum distribution in the nuclear medium @xcite . a plane - wave
, independent - particle description of the process suggests that the @xmath0 cross section should be given by a set of discrete peaks each corresponding to knock - out from a given nuclear shell with the momentum distribution obtained by integrating the strength under these peaks .
this description is nearly adequate for states near the fermi surface , yet additional study has revealed that the single - particle strength is strongly fragmented for deeply bound states @xcite .
knock - out from states near the fermi surface does support the notion of unfragmented single - particle strength , but the occupancy of these states is less than one @xcite .
these results show conclusively that the traditional picture of the nucleus , as a collection of particles moving independently in a mean - field potential , is too naive .
consequently , the @xmath0 reaction shows the limitations of an independent - particle picture , yet it provides a fruitful testing ground for various dynamical mechanisms , such as short - range correlations , that go beyond a simple mean - field approach .
the @xmath0 reaction has also been used to study possible medium modifications to the electromagnetic coupling of the nucleon @xcite . that such effects exist
has been suggested by a variety of studies .
the original european - muon - collaboration ( emc ) effect , for example , can be explained by a softening of the quark momentum distribution in the nuclear medium @xcite , which , in turn , modifies the nucleon s electromagnetic coupling .
interest in the @xmath0 reaction has been recently stimulated by suggestions that color transparency may manifest itself as an asymmetry @xcite in the missing - momentum distribution , even in the absence of a modification of the total cross section @xcite .
`` color transparency '' describes the hypothesis that hadrons produced with large laboratory momenta in certain exclusive reactions with nuclear targets interact in an anomalously weak manner with the residual nucleus @xcite .
we shall not concern ourselves directly with color transparency , yet part of our interest in considering conventional relativistic effects in the @xmath0 reaction is to the end of establishing a robust `` baseline '' calculation , against which possible novel effects may be inferred .
the possible existence of color transparency has been discussed in a great many different exclusive processes , yet the focus on the ne18 experiment at slac has made the quasielastic @xmath0 reaction the paradigm @xcite . with the advent of cebaf ,
the focus on this particular reaction is unlikely to diminish .
indeed , several @xmath0 experiments have been approved in the hope of identifying this novel behavior @xcite .
our purpose is to explore the asymmetry in the missing momentum distribution in a conventional @xmath0 calculation . for concreteness ,
let us now quantify what we mean by the `` asymmetry . ''
we consider the @xmath0 reaction in the kinematics where the three - momentum transfer @xmath1 and the transverse missing momentum @xmath2 are fixed ( in all that follows , `` longitudinal '' and `` transverse '' are defined as the components parallel and perpendicular to @xmath1 , respectively ) . with this choice of kinematics , the longitudinal momentum distribution of the bound proton is scanned by varying the energy transfer to the nucleus this results in a varying momentum for the outgoing proton .
the asymmetry is , then , defined as the integrated sum of events with positive longitudinal missing momentum relative to those events having negative missing momentum @xcite . here
we shall consider the reaction specifically in `` parallel kinematics '' , that is , we shall take @xmath3 , so that the knocked - out proton s momentum is parallel to @xmath1 , in all that follows
. studies of the @xmath0 reaction at low momentum transfers ( @xmath4 ) yield missing - momentum ( @xmath5 ) distributions which are asymmetric about the @xmath6 point @xcite .
a calculation in a nonrelativistic plane - wave impulse approximation has no such asymmetry .
the asymmetry at low momentum transfers is understood to arise from the momentum dependence of the distortions that the struck proton suffers in its exit from the nucleus @xcite . in the momentum transfer regime beyond @xmath7 ,
nuclear quasielastic electron scattering is essentially unexplored ; this frontier provides a non - trivial testing ground for phenomenologies constructed in a lower energy regime . at large momenta
, glauber theory has been used to define the `` baseline '' .
the attenuation of the ejected proton through the nucleus is given in this model by the inelastic proton - nucleon cross section .
for proton momenta in excess of 3 gev , the inelastic cross section is approximately constant , so that the asymmetry is essentially zero . the virtue of the asymmetry as a signal of transparency rests wholly on the assumption that such an asymmetry is absent in any conventional ( i.e. , without transparency ) calculation of the @xmath0 reaction at large momentum transfers .
we shall show that this assumption is flawed .
relativistic effects , even in a plane - wave impulse approximation , serve to generate an asymmetry in the @xmath0 reaction . in a previous work
, one of us examined the role of relativity on the spectral function @xcite . in that context , relativistic effects , driven by particle - antiparticle mixing , are responsible for generating features , such as a fragmentation of the single - particle strength , which are associated nonrelativistically with physics beyond the mean - field level .
these relativistic effects exist even if a mean - field ground state is used .
as relativity generates effects similar to what a more complicated nonrelativistic calculation would generate , they can not be regarded as a definitive relativistic signature .
here we consider an effect which is absent in the conventional nonrelativistic formalism at the plane - wave level .
moreover , including distortions as in the glauber theory also results in a small asymmetry at large momentum transfers .
in contrast , relativistic corrections increase with the momentum transfer to the nucleus and can become large .
as these effects are absent in the conventional nonrelativistic theory , they are interesting in their own right
. moreover , their existence limits the utility of the missing - momentum asymmetry as a signal of color transparency @xcite .
we have organized the paper as follows . in sec .
[ secformal ] we review some general properties of the @xmath0 reaction and develop analytic expressions for the nuclear response in a relativistic plane - wave impulse approximation .
these plane - wave expressions identify the lower components of the bound - state spinors as the source of the asymmetry in the missing - momentum distribution .
the only dynamical input required is the relativistic bound state wave function .
consequently , in sec . [ secres ] , we use a @xmath8o ground - state wave function obtained in a relativistic mean - field approximation to the walecka model as an illustrative example .
finally , we offer our conclusions in sec [ secconcl ] .
we begin by writing the inclusive hadronic tensor @xmath9 as @xmath10 the nuclear electromagnetic current @xmath11 is responsible for inducing transitions between the exact nuclear ground state ( @xmath12 ) and an excited state ( @xmath13 ) with excitation energy @xmath14 . in an independent - particle description of the ground state , the inclusive hadronic tensor can be expressed in terms of individual particle - hole transitions ; i.e. , @xmath15 where @xmath16 note that @xmath17 is the energy transfer to the nucleus , and @xmath18 is the excitation energy from the initial single - particle state @xmath19 to the final state @xmath20 . the dynamical information about the process is contained in @xmath21 the transition matrix element between an occupied single - particle state @xmath22 and an unoccupied state @xmath23 .
we assume the impulse approximation valid and employ an electromagnetic current with single - nucleon form factors parameterized from on - shell @xmath24 scattering .
that is , @xmath25 where @xmath26 and @xmath27 are given as in ref .
specifically , for the proton , [ emformf ] @xmath28 ( 1 + 4.97\tau)^{-2 } / ( 1+\tau ) \ ; , \\ f_{2}(q^{2 } ) & = & \kappa_{p } ( 1 + 4.97\tau)^{-2 } / ( 1+\tau ) \;. \end{aligned}\ ] ] note that @xmath29 , and @xmath30 is the proton s anomalous magnetic moment .
the plane - wave limit of the @xmath0 reaction is useful in providing crude guidance into the nature of the reaction .
indeed , the plane - wave impulse approximation ( pwia ) analysis leads one to believe that the @xmath0 reaction can be used to extract the spectral function and , thus , the nucleon momentum distribution @xcite .
this is only approximately correct .
for example , final - state interactions ( fsi ) between the outgoing nucleon and the residual nucleus are known to modify the plane - wave picture .
of particular relevance to our present work is the symmetry in the missing - momentum distribution predicted in the nonrelativistic pwia .
this symmetry is broken by final - state interactions . yet
, the inclusion of distortions in the glauber formalism suggests that the symmetry in the missing - momentum distribution will become approximately restored at large momentum transfers .
this observation supports the claim made in ref .
@xcite that any measured asymmetry in the missing - momentum distribution is due to the formation of a compact state and , thus , constitutes evidence in support of color transparency . in this section
we will show that some notions developed in a nonrelativistic context no longer hold in a relativistic formalism .
in particular , we will prove that the symmetry in the missing - momentum distribution is broken at the plane - wave level due to relativistic effects .
moreover , the asymmetry increases with the momentum transferred to the nucleus and as the momentum of the outgoing nucleon is increased .
consequently , these relativistic `` corrections '' to the asymmetry contaminate any attempt to identify color transparency . in a relativistic plane - wave impulse approximation ( rpwia ) the outgoing proton with momentum @xmath31 and spin projection @xmath32
is described by a plane - wave dirac spinor @xcite : @xmath33 where the on - shell energy of the outgoing proton is given by @xmath34 note that the normalization implied by eq .
( [ freewf ] ) differs from the one in ref .
@xcite in that @xmath35 in the following presentation , we shall use only the dirac ( or vector ) piece of the electromagnetic current , for simplicity .
our final results , presented at the end of this section , include the anomalous piece of the current as well . in a rpwia the dirac piece of the current matrix element for the knock - out of a bound - state nucleon is @xmath36 where @xmath37 is the plane - wave state defined above and where the fourier transform of the relativistic bound - state wave function is given by @xmath38 note that we have also introduced the missing momentum @xmath39 the semi - inclusive hadronic tensor can now be written in terms of a feynman trace , @xmath40 \left [ \overline{\cal u}({\bf p'},s ' ) \gamma^{\nu}{\cal u}_{\alpha}({\bf p } ) \right]^{\star } \nonumber \\ & = & { \rm tr } \left [ \gamma^{\mu } s_{\alpha}({\bf p } ) \gamma^{\nu } \big({{\rlap/{p ' } } + m \over 2e_{{\bf p'}}}\big ) \big({1+\gamma^{5}{\rlap/{s ' } } \over 2}\big ) \right ] \ ; , \label{trace}\end{aligned}\ ] ] involving the spin - dependent projection operator for the outgoing nucleon @xcite , as well as the bound - state `` propagator '' @xmath41 which satisfies the normalization condition @xmath42=1 \;.
\label{normal}\ ] ] to proceed , we need an explicit form for @xmath43 .
to do so , we recognize that in the presence of a spherically symmetric potential the eigenstates of the dirac equation can be classified according to a generalized angular momentum @xmath44 and can be written in a two component representation ; i.e. , @xmath45 \;. \label{dirwfx}\ ] ] the upper and lower components are expressed in terms of spin - spherical harmonics defined by @xmath46 the fourier transform of the above bound - state wave function can now be easily evaluated .
that is , @xmath47 = { 4\pi \over p } ( -i)^{l } \left [ \begin{array}{c } g_{e\kappa}(p ) \\
f_{e\kappa}(p ) ( { \bf \sigma}\cdot{\hat{\bf p } } ) \end{array } \right ] { \cal{y}}_{+\kappa m}(\hat{\bf{p } } ) \ ; , \label{dirwfp}\ ] ] where we have written the fourier transforms of the radial wave functions as @xmath48 [ gfp ] in the above expression we have introduced @xmath49 , the riccati - bessel function , through the relation @xmath50 . by now making use of the following identities @xmath51 [ msum ] we can express the bound - state propagator in closed form .
that is , @xmath52 where we have defined the above mass- , energy- , and momentum - like quantities as @xmath53 \ ; , \\
e_{\alpha } & = & \left({\pi \over p^{2}}\right ) \big[g_{\alpha}^{2}(p ) + f_{\alpha}^{2}(p)\big ] \ ; , \\ { \bf p}_{\alpha } & = & \left({\pi \over p^{2}}\right ) \big[2 g_{\alpha}(p ) f_{\alpha}(p)\hat{\bf p } \big ] \;. \label{epm}\end{aligned}\ ] ] note that they satisfy the `` on - shell relation '' @xmath54 the above representation of the bound - state propagator , given entirely in terms of dirac gamma matrices , enables one to evaluate the semi - inclusive hadronic tensor , @xmath55 \ ; , \label{feyntrace}\end{aligned}\ ] ] with feynman s trace techniques .
this relativistic plane - wave expression for the dirac component of the hadronic tensor is completely general and does not assume any specific kinematical constraint . in the present work , however , we are interested in the missing - momentum asymmetry evaluated in parallel kinematics ( @xmath56 ) . for this case , then , the charge , transverse , and longitudinal components of the hadronic tensor are given , respectively , by @xmath57^{2 } \ ; , \\
w^{11}_{\scriptscriptstyle dd}({\bf q},{\bf p'};\alpha ) & = & \left({e_{\bf p ' } + m \over 2e_{\bf p'}}\right ) \left({4\pi \over p^{2}}\right ) \big[\xi_{\bf p'}g_{\alpha}(p ) - ( \hat{\bf p}\cdot\hat{\bf p } ' ) f_{\alpha}(p)\big]^{2 } \ ; , \\ w^{33}_{\scriptscriptstyle dd}({\bf q},{\bf p'};\alpha ) & = & \left({e_{\bf p ' } + m \over 2e_{\bf p'}}\right ) \left({4\pi \over p^{2}}\right ) \big[\xi_{\bf p'}g_{\alpha}(p ) + ( \hat{\bf p}\cdot\hat{\bf p } ' ) f_{\alpha}(p)\big]^{2 } \ ; , \label{ws}\end{aligned}\ ] ] where we have defined @xmath58 two aspects of these results are particularly noteworthy .
first , we have included the charge as well as the longitudinal components of the hadronic tensor because the electromagnetic current is not conserved at the plane - wave level .
second , the term @xmath59 generates an asymmetry in the missing - momentum distribution .
this has its origin in the relation @xmath60 note , however , that in the absence of lower bound - state components ( i.e. , @xmath61 ) , there is no asymmetry , in agreement with nonrelativistic expectations .
moreover , including distortions as in the glauber theory does not significantly alter this result at large momentum transfers .
the asymmetry at large momenta , then , is a genuinely relativistic effect .
this simple result , namely , the relativistic generation of an asymmetry at the plane - wave level , is the main result of the present work .
one of the advantages of plane - wave analyses is that they suggest that the @xmath0 cross section can be factorized into an ( off - shell ) elementary electron - proton cross section times a spectral function containing all nuclear - structure information . in order to isolate the nuclear structure effects
, we shall introduce a reduced cross section .
this reduced cross section is obtained from the full @xmath0 cross section by dividing out an off - shell electron - proton cross section with some kinematical factors .
unfortunately , the off - shell electron - proton cross section is not uniquely defined @xcite . the choice made can even obscure the dynamical features that one is trying to uncover
indeed , an off - shell electron - proton cross section also generates an asymmetry in the missing - momentum distribution , but this asymmetry is unrelated to nuclear - structure effects . for our present purpose
, we shall adopt the following procedure .
first , we separate the full cross section into longitudinal and transverse response functions .
then , we divide each of these responses by an off - shell electric or magnetic single - nucleon form factor , as appropriate , to yield the reduced cross sections .
this definition has the advantage that the nonrelativistic limit of the reduced cross section does , indeed , yield a symmetric missing - momentum distribution .
this procedure provides , then , a baseline for comparing different theoretical mechanisms of breaking the symmetry , be they relativistic effects , or color transparency
. we shall now introduce longitudinal and transverse rpwia reduced cross sections for the @xmath0 reaction .
the reduced cross sections have been computed with the full electromagnetic current including the anomalous piece and are given by [ rholt ] @xmath62^{2 } \ ; , \label{rhol } \\
\hbox{\raise.8ex\hbox{$\rho$}\kern-2pt}_{t}({\bf q},{\bf p'};\alpha ) & \equiv & { w^{11}({\bf q},{\bf p'};\alpha ) \over w^{11}({\bf q},{\bf p ' } ) } = \left({4\pi \over
p^{2}}\right ) g_{\alpha}^{2}(p ) \left [ 1 \mp \left({f_{1}+(\bar{\omega}-\xi_{\bf p'}\bar{q})f_{2 } \over \xi_{\bf p'}f_{1}+ ( \bar{q}-\xi_{\bf p'}\bar{\omega})f_{2}}\right ) { f_{\alpha}(p ) \over g_{\alpha}(p ) } \right]^{2 } \;. \label{rhot}\end{aligned}\ ] ] the off - shell single - nucleon responses , @xmath63 and @xmath64 , have been chosen as [ slstfree ] @xmath65^{2 } \longrightarrow \left({1 + \tau \over 1 + 2\tau}\right)g_{e}^{2}(q^{2 } ) \ ; , \label{slfree } \\
w^{11}({\bf q},{\bf p ' } ) & = & \left({e_{\bf p ' } + m \over 2e_{\bf p'}}\right ) \big [ \xi_{\bf p'}f_{1}+ ( \bar{q}-\xi_{\bf p'}\bar{\omega})f_{2 } \big]^{2 } \longrightarrow \left({\tau \over 1 + 2\tau}\right)g_{m}^{2}(q^{2 } ) \ ; , \label{stfree}\end{aligned}\ ] ] where the following dimensionless quantities have been introduced @xmath66 in constructing the single - nucleon responses we have demanded that the resulting missing - momentum distribution be symmetric at the plane - wave level in the limit of a vanishing lower bound - state component . note that in this limit @xmath67 the arrow in the previous expressions is used to indicate that the single - nucleon responses in the isolated nucleon limit ( so that @xmath68 ) become proportional to the electric and magnetic form factors of the nucleon .
the off - shell single - nucleon responses have been defined to the end of establishing a baseline calculation for the asymmetry , against which relativistic effects may be assessed .
the longitudinal reduced cross section clearly shows that the symmetry in the missing momentum distribution is broken by relativistic effects .
consequently , the symmetry is restored when all momenta become small relative to the nucleon mass .
the nonrelativistic limit of the transverse reduced cross section is , in contrast , not clearly defined .
the relativistic `` corrections '' are of @xmath69 , so that the transverse response is an intrinsically relativistic observable . in a recent publication ,
bianconi , boffi , and kharzeev suggested @xcite studying the asymmetry in the missing - momentum distribution as a means of uncovering color transparency .
in particular , they suggest comparing the number of events with positive missing momentum ( @xmath70 ) to those events with negative missing momentum ( @xmath71 ) @xcite . in this work
we follow their suggestion and define the following asymmetries . here
we define , however , separate longitudinal and transverse asymmetries : [ asymm ] @xmath72 in all that follows , we will assume that we are studying single - particle transitions from a specific bound state @xmath19 to the continuum , so that the @xmath19 label is suppressed both in the above and in the reduced cross sections .
the asymmetries are obtained from the total number of longitudinal or transverse events having either positive or negative missing momentum : [ pmint ] @xmath73 the arrows indicate two successive limits . in the first the lower bound - state component
is set to zero , while in the second the cut - off @xmath74 goes to infinity . introducing a cut - off
is useful as it enables us to exclude those large missing momentum events that might be sensitive to physics beyond the mean - field picture .
note that when @xmath75 the number of events with positive and negative missing momenta become equal , so that the asymmetry vanishes .
the symmetry is broken , however , by the inclusion of momentum - dependent final - state interactions or by relativistic effects .
the symmetry may also be broken by color transparency @xcite .
color transparency , if it exists , is due to the coherent interference of the distortions suffered by the outgoing proton and other , higher - mass , baryon states produced by the virtual photon
the strong interactions of the individual components conspire to cancel , at least in part .
the initial and final electron 4-momenta fix @xmath76 ( @xmath77 ) so that the longitudinal component of the struck proton momentum depends on whether a proton or a @xmath78 is produced . for @xmath79 , the bound - state momenta @xmath80 required to make a proton ( @xmath81 ) or an excited baryon state ( @xmath78 ) must be of the same sign . satisfying the kinematic constraints for both @xmath81 and @xmath78 production at moderate momentum transfers is unlikely , as bound - nucleon momenta in the tail of the momentum distribution are required . for @xmath82 ,
the required momenta can be of opposite sign , so that the kinematic constraints can be satisfied at moderate @xmath83 with a finite probability .
this is the source of the suggested asymmetry effect @xcite .
bianconi , boffi , and kharzeev @xcite relate the proposed asymmetry in @xmath76 to an asymmetry in the missing - momentum distribution ; they find a positive asymmetry , as we shall find here for the longitudinal asymmetry .
however , their arguments give no insight into how the longitudinal and transverse asymmetries may differ , nor is it helpful in disentangling off - shell form factor effects from the modification of the final - state interactions . yet
, the conventional glauber analysis suggests that the asymmetry should be small at large @xmath83 , supporting the asymmetry as a signal of color transparency , as argued by the authors of ref .
@xcite .
we believe that relativistic effects modify this picture .
indeed , we have shown that relativistic effects generate an asymmetry in the missing - momentum distribution at the plane - wave level .
we now proceed to show relativistic plane - wave results for the missing - momentum asymmetry . in addition
, we estimate how the asymmetry is changed upon the inclusion of relativistic distortions which are essential for restoring electromagnetic gauge invariance . at the plane - wave level , the reduced cross sections , in eq .
( [ rholt ] ) , require only the bound - state wave functions as dynamical input .
we have assumed that the single - nucleon form factors are not modified by the nuclear medium and have simply adopted the parameterization for the on - shell form factors given in eq .
( [ emformf ] ) .
the bound - state wave functions have been obtained in a relativistic mean - field approximation to the walecka model @xcite . in the walecka model nucleons
interact via the exchange of neutral scalar ( @xmath84 ) and vector ( @xmath17 ) mesons @xcite .
the parameters of the model are adjusted , at the mean - field level , to reproduce the binding energy and density of nuclear matter at saturation . with this input ,
the mean - field approximation to the walecka model has been successful in describing the ground - state properties of finite nuclei as well as the linear response of the ground state to a variety of probes @xcite .
this is realized , in part , because the walecka model is characterized by strong scalar and ( timelike ) vector potentials that cancel to generate the observed weak binding energies , but combine to yield a strong spin - orbit splitting for the positive - energy states @xcite . in fig .
[ figzero ] we present the charge response as a function of the missing momentum for the two valence orbitals in @xmath8o for the three - momentum transfers of @xmath85gev . noting eq .
( [ pmint]a ) , the response has been multiplied by the phase - space factors necessary to make the area under the curve directly proportional to the number of integrated events . in addition , we have scaled the response in order to compensate for the rapid falloff of the form factors , since we have chosen to use the same scale here as in fig .
[ figone ] .
these constant factors do not change the asymmetries these are displayed in brackets next to the appropriate momentum - transfer labels .
the asymmetry in the charge response is driven by a combination of nuclear - structure and single - nucleon effects .
as we have discussed , nuclear - structure effects generate an asymmetry in the momentum distribution due to the presence of lower components in the bound - state wave function .
however , even in the absence of relativistic corrections , the charge response , at fixed three - momentum transfer , will display an asymmetry from single - nucleon effects .
the latter asymmetry is driven by the @xmath86 dependence of the form factor .
the change in @xmath83 itself depends on the magnitude and on the direction of the missing momentum . for fixed @xmath87 , the energy transfer to a nucleon moving parallel to @xmath88
must be larger than that to a nucleon moving antiparallel to @xmath1 , in order to yield an outgoing nucleon which is on its mass shell .
hence , the electromagnetic coupling is stronger for the nucleon moving parallel to @xmath1 since its charge form factor is being probed at a lower value of @xmath86 .
this single - nucleon effect is responsible for generating a positive asymmetry in the missing momentum distribution .
its contribution must , then , be eliminated from the response before one can identify the role of the nuclear medium in breaking the symmetry .
eliminating the single - nucleon contribution from the asymmetry is , however , not free from ambiguity . indeed
, certain off - shell prescriptions could accentuate the effect .
we believe that the off - shell prescription adopted here [ eq .
( [ slfree ] ) ] is an appealing choice for the study of the asymmetry .
this choice yields , in particular , a symmetric missing - momentum distribution in the nonrelativistic plane - wave limit .
this is an useful baseline against which interesting behavior may be established . in fig .
[ figone ] we present the longitudinal reduced cross section [ eq .
( [ rhol ] ) ] for the two valence orbitals in @xmath8o for the three - momentum transfers of @xmath85gev . in this case
the area under the curve is directly proportional to the number of events @xmath89 defined in eq .
( [ pmint ] ) .
the reduced cross section clearly shows an asymmetric distribution .
this asymmetry is entirely due to relativistic nuclear - structure effects .
the total asymmetries are displayed in the appropriate brackets .
in particular , we note that removing the single - nucleon contribution to the response yields asymmetries which are slightly smaller than those of fig .
[ figzero ] . in fig .
[ figtwo ] we show the longitudinal asymmetry [ eq .
( [ asymml ] ) ] as a function of the three - momentum transfer @xmath90 and the momentum cut - off @xmath74 . at small three - momentum transfers , the longitudinal asymmetry is negligible because the relativistic corrections are small ( @xmath91 , and @xmath92 ) .
in contrast , at large momentum transfers , the relativistic `` corrections '' are large and so is the asymmetry
. we can understand this by examining the high - energy limit of the longitudinal reduced cross section [ eq .
[ rhol ] ] . at large three - momentum transfers ( @xmath93 ) and small missing momenta ( @xmath94 )
, we obtain the following limits for the other previously defined kinematical variables @xmath95 these relations imply , in particular , that the following ratio of anomalous to dirac form factors becomes @xmath96 \rightarrow { \kappa_{p } \over ( 1+\kappa_{p } ) } \;,\ ] ] so that one finds the following simple expression for the reduced cross section : @xmath97^{2 } \;.\ ] ] this result indicates that the longitudinal asymmetry is enhanced not only by the in - medium wave function ratio of ( @xmath98 ) but also by an additional factor of [ @xmath99 . in fig .
[ figthree ] we show , for completeness , the transverse asymmetry .
the transverse response is an intrinsically relativistic observable , so that the transverse asymmetry is large even at small momentum transfers .
the asymmetry , however , saturates rapidly , as reflected by the high - energy limit of the reduced cross section : @xmath100^{2 } \;.\ ] ] we conclude this section with a brief discussion of distortion effects .
a nonrelativistic analysis of the @xmath0 reaction generates an asymmetry in the missing - momentum distribution when distortion effects are included , at low energies . at high energies , however
, a glauber analysis suggests that the symmetry should be restored .
relativistically , however , the energy dependence of the distorting potential depends strongly on its lorentz character and can be substantially different from its nonrelativistic counterpart @xcite .
distortions play an important role in the conservation of the electromagnetic current .
electromagnetic gauge invariance has been lost as a result of our plane - wave treatment , and distortions play a fundamental role in its restoration .
the included distortions , however , must be consistent with the mean - field potential used to obtain the single - particle spectrum .
hence , in calculating the distorted wave for the outgoing proton , we have employed the same real and energy - independent potential used to generate the mean - field ground state .
unfortunately , this treatment implies that the imaginary part of the nucleon self - energy which is physically important has been suppressed . in spite of this
, we have calculated the nuclear response in a relativistic distorted - wave impulse approximation ( rdwia ) , in order to show what impact the restoration of gauge invariance has on the asymmetry .
as we shall see , the longitudinal asymmetry remains large in the presence of the included distortions .
the incorporation of absorptive and energy - dependent effects in a calculation with electromagnetic gauge invariance remains an important open problem . in fig .
[ figfour ] we show the longitudinal reduced cross section as a function of the missing momentum for the two valence orbitals in @xmath8o at @xmath101 gev .
the dashed line is the relativistic plane - wave result also shown in fig .
[ figone ] .
the symmetric distribution ( dot - dashed line ) results from neglecting the lower component of the bound - state wave function in the rpwia calculation .
this represents our best attempt in generating a nonrelativistic result .
finally , relativistic distorted - wave results are shown by the solid line .
the positive shift in the rdwia distribution relative to the plane - wave result is due to the repulsive character of the real distorting potential .
the self - consistent , mean - field potential employed in the rdwia calculation becomes repulsive for a kinetic energy of @xmath102 mev ( @xmath103 mev ) .
this implies that at a momentum transfer of @xmath101 gev the distorting potential is repulsive over the entire range of missing momenta shown in the figure , thus , producing the observed shift . in table
[ tableone ] we display the sensitivity of the longitudinal asymmetry to distortion effects .
the rdwia asymmetry is consistently larger than the plane - wave results due to the added effect of the repulsive distorting potential .
we have calculated the asymmetry in the missing - momentum distribution from the @xmath0 reaction in a relativistic plane - wave impulse approximation , using a mean - field approximation to the walecka model .
longitudinal and transverse nuclear - response functions have been evaluated in parallel kinematics as a function of the three - momentum transfer to the nucleus . by dividing out off - shell , single - nucleon form factors from the response functions
, we were able to define a nonrelativistic plane - wave limit with a symmetric momentum distribution .
this provides a baseline for comparing different theoretical mechanisms of breaking the symmetry . in a relativistic formalism
the presence of a lower component in the dirac bound - state wave function is sufficient to break the symmetry at the plane - wave level .
we have developed analytic expressions for the plane - wave responses and have illustrated our findings by calculating the missing - momentum asymmetry using @xmath8o as an example .
the ground state for @xmath8o was generated self - consistently in a mean - field approximation to the walecka model .
the mean - field approximation is characterized by the presence of strong ( attractive ) scalar and ( repulsive ) timelike vector potentials .
the scalar potential leads to a reduction of the effective nucleon mass in the medium and is , ultimately , responsible for a substantial enhancement of the ratio of upper to lower components in the bound - state wave function .
this dynamical enhancement of the ratio , together with the increasing importance of the relativistic corrections with momentum transfer , yield longitudinal asymmetries as large as @xmath104 for a three - momentum transfer of @xmath105gev .
these nuclear - structure effects preclude the use of the asymmetry in the missing - momentum distribution as a definitive signature of color transparency .
however , the missing - momentum asymmetry does provide interesting information about the role of relativity in the nuclear medium .
we have also investigated the effects of distortions .
we include distortions to assess the impact of the restoration of gauge invariance on the asymmetry .
the rdwia calculations yield a larger longitudinal asymmetry than in the plane - wave case .
this results from the repulsive character of the real distorting potential .
we have stressed that a distorting potential consistent with the mean field used to generate the nuclear ground state is essential for the conservation of the electromagnetic current .
these mean - field distortions ignore , however , some important physics absorption and energy - dependence associated with the in - medium propagation of the outgoing proton .
we believe that a gauge - invariant treatment of the @xmath0 reaction that successfully describes the distortions of the outgoing proton is an important area for future work .
this research was supported by the u.s .
department of energy contract # de - fg02 - 87er40365 ( s.g . ) , as well as by the florida state university supercomputer computations research institute through the u.s .
department of energy contracts # de - fc05 - 85er250000 and # de - fg05 - 92er40750 ( j.p . ) .
s. frullani and j. mougey , in _ advances in nuclear physics _ , edited by j.w .
negele and e. vogt ( plenum , new york , 1984 ) vol .
quint _ et al .
57 * , 186 ( 1986 ) ; * 58 * , 1088 ( 1987 ) . j.w.a .
den herder , h.p .
blok , e. jans , p.h.m .
keizer , l. lapikas , e.n.m .
quint , g. van der steenhoven , and p.k.a .
de witt huberts , nucl .
a490 * , 507 ( 1988 ) .
g. van der steenhoven _ et al .
* 57 * , 182 ( 1986 ) ; * 58 * , 1727 ( 1987 ) .
close , r.l .
jaffe , r.g .
roberts , and g.g .
ross , phys .
d*31 * , 1004 ( 1985 ) .
jennings and b.z .
kopeliovich , phys . rev
. lett . * 70 * , 3384 ( 1993 ) .
nikolaev , a. szczurek , j. speth , j. wambach , b.g .
zakharov , and v.r .
zoller , kernforschungszentrum - jlich preprint kfa - ikp - th-1993 - 21 .
a. bianconi , s. boffi , and d.e .
kharzeev , university of pavia preprint fnt / t-93 - 19 ; lanl bulletin board nucl - th-9310012 . s.j .
brodsky , in _ proceedings of the thirteenth symposium on multiparticle dynamics _ , edited by w. kittel , w. metzger , and a. stergiou ( world scientific , singapore , 1982 ) .
a. mueller , in _ proceedings of the seventeenth recontre de moriond _ , moriond , 1982 , edited by j. tran thanh van ( editions frontieres , gif - sur - yvette , france , 1982 ) .
l. frankfurt , g.a .
miller , and m. strikman , comm .
nucl . part .
* 21 * , 1 ( 1992 ) and references therein .
cebaf pr 91 - 006 , _ study of nuclear medium effects by recoil polarization up to high momentum transfers _ , spokesperson a. saha .
cebaf pr 91 - 007 , _ measurement of the nuclear dependence and momentum transfer dependence of quasielastic @xmath0 scattering at large momentum transfer _ , spokesperson r. milner .
cebaf pr 91 - 013 , _ the energy dependence of nucleon propagation in nuclei as measured in the @xmath0 reaction _ , spokesperson d.f .
m. bernheim _ et al .
a375 * , 381 ( 1982 ) j. ryckebusch , k. heyde , d. van neck , and m. waroquier , phys .
b216 * , 252 ( 1989 ) .
j. piekarewicz and r.a .
rego , phys .
c*45 * , 1654 ( 1992 ) .
c.j . horowitz and j. piekarewicz ,
c*47 * , 2924 ( 1993 ) .
bjorken and s.d .
drell , _ relativistic quantum fields _ ( mcgraw - hill , new york , 1965 ) .
taber de forest , nucl .
a392 * , 232 ( 1983 ) .
c.j . horowitz and b.d .
serot , nucl .
a368 * , 503 ( 1981 ) .
walecka , ann . of phys .
* 83 * ( 1974 ) 491 .
serot and j.d .
walecka , adv . in nucl
* 16 * , j.w .
negele and e. vogt , eds .
( plenum , n.y .
j. piekarewicz , phys . rev .
c*48 * , 2174 ( 1993 ) .
.longitudinal numbers of events and missing - momentum asymmetries as a function of the three - momentum transfer with @xmath106mev for the @xmath107 state in @xmath8o .
the first(second ) row of numbers are the results of a relativistic calculation that neglects(incorporates ) distortion effects . [
cols="^,^,^,^ " , ] | we calculate the asymmetry in the missing - momentum distribution from the @xmath0 reaction in a relativistic formalism .
longitudinal and transverse response functions are evaluated in parallel kinematics as a function of the three - momentum transfer to the nucleus .
analytic expressions for the responses are obtained in a relativistic plane - wave impulse approximation .
these expressions reveal a large asymmetry in the momentum distribution , at the plane - wave level , induced by the lower components of the bound - state wave functions .
these relativistic effects contaminate any attempt to infer color transparency from a measurement of the asymmetry in the @xmath0 reaction . |
the first confirmed extrasolar planets were found to orbit evolved stars @xcite and the last 5 years has seen a resurgence of interest in this topic due to new discoveries ( e.g. @xcite ) .
the majority of complimentary theoretical analyses has focused on single - planet , single - star evolved systems @xcite or systems with a belt or disc of material @xcite .
these investigations crucially establish physical and analytical frameworks from which to explore more complex systems in greater depth .
just a few studies have considered multi - planet post - main - sequence ( ms ) evolution .
@xcite performed long - term integrations of our solar system assuming a constant solar mass loss rate .
@xcite demonstrated that multiple equal - mass planets on coplanar , circular orbits which are marginally stable on the ms can become destabilised during post - ms evolution , because of the expansion of the hill - stability limit due to mass loss .
@xcite showed that even hill - stable pairs of planets can eventually become unstable , illustrating that instability occurs more readily than previously thought , and that planets do not need to be closely packed to become unstable many gyr later .
@xcite considered the evolution of two planets in cataclysmic variable systems , but did not focus on planetary instability .
although these initial studies have now broached the topic of planet - planet scattering amidst mass loss , the detailed nature of this instability has yet to be explored . here
, we pursue this line of investigation .
we consider a planetary system with two planets in a resonant or nonresonant configuration that is stable on the ms and is perturbed when mass loss takes place .
we will show that the perturbation pushes the system into a chaotic regime through which instability manifests itself after the mass loss ceases .
the consequences of this late time instability may include collisions within the system or escape from the system .
we focus on the later possibility , due to its potential relevance to the purportedly vast free - floating planet population in the milky way @xcite , which is thought to outnumber the galactic population of bound planets .
the existence of such substellar objects , known as free - floating or orphan planets ( e.g. @xcite ) , may help us understand the low - mass end of the initial mass function .
the source of free - floating planets is unknown . @xcite
considered the rate of ejections in planet - planet scattering in ms systems , and found that even in the most optimistic case , the rate is insufficient to generate the observed poputation .
alternatively , @xcite illustrated how a free - floating planet passing a stellar system could trigger dynamical instability and incite planetary escape , leading to another free - floating planet .
@xcite and @xcite considered similar mechanisms for stellar flybys and star - planet flybys in the galactic disc , respectively .
these studies do not reproduce the required escape rate , and potentially suggest that the free - floating planet population is a relic of the initial formation process .
the turbulent early - age birth environment of planets might represent the primary source , as well as the frequent and slow flybys characteristic of young clusters @xcite .
our focus here , however , is on planetary escape in evolved systems , and on bounding the phase space in which this can occur .
the paper is organized as follows . in section 2
we present some analytic and numerical results for the orbital element evolution of planets under stellar mass loss without taking into account the mutual planetary interaction . in section 3
we present results that show the distribution of chaos and order in a system consisting of two massive planets and a star with constant mass ; we discuss the resulting chaotic evolution and the destabilization of the system
. in section 4 we present our results , which represent numerical simulations that combine both planetary interactions and stellar mass loss , and demonstrate the possible system destabilization and planetary escape .
we present our main conclusions in section 5 .
in the classical two body problem , the two bodies describe similar elliptic orbits around their center of mass when the motion is bounded .
the motion is planar and the orbital elements of the two bodies ( or the orbital elements of the relative motion of one body around the other ) are constant . if , however , one of the bodies loses mass isotropically in all directions , then the evolution of the system is quite different .
the orbital elements of the relative motion are no longer constant , and the evolution of the system depends on the mass loss function @xmath2 , where @xmath3 is the sum of the masses of the two bodies . in this sense
, the assumption of @xcite that mass loss by the star and mass gain by the planet are equivalent does not hold .
a particular case , which we shall consider in this work , is a planetary system with a star and a planet with much smaller mass , where the star loses mass isotropically .
the osculating elements of the relative motion are given by the differential equations @xcite : @xmath4 \frac{de}{dt}=-(e+\cos f)\frac{\dot m}{m } \\[0.2 cm ] \frac{d\omega}{dt}=-\frac{\sin f}{e}\frac{\dot m}{m } \\[0.2 cm ] \frac{df}{dt}=\frac{(g m)^{1/2}(1+e\cos f)^2}{[a ( 1-e^2)]^{3/2 } } + \frac{\sin f}{e}\frac{\dot m}{m } , \end{array}\ ] ] where @xmath5 is the semimajor axis of the relative motion , @xmath6 the eccentricity , @xmath7 the argument of pericenter and @xmath8 the true anomaly .
one can readily check that the above equations admit the integral @xmath9 which in fact is the angular momentum integral .
we set @xmath10 in the remainder of the paper . in the following , we assume that the loss of the mass of the star is described by the eddington - jean s law @xmath11 where @xmath12 is , in general , a small positive constant .
the exponent @xmath13 defines the particular law , e.g. for @xmath14 we get a linear variation of mass , @xmath15 , and for @xmath16 the exponential one , @xmath17 .
also , we assume that the star loses mass down to a minimum limit @xmath18 , where the constant @xmath19 defines the final mass loss ratio ( @xmath20 ) .
this minimum mass is obtained at a time @xmath21 , which is found through the solution of ( [ ejlaw ] ) and depends on @xmath22 and the parameters @xmath12 , @xmath19 and @xmath13 .
we introduce the normalized semimajor axis @xmath23 and the normalized mass @xmath24 .
the evolution of the osculating elements in the interval @xmath25 can be given as a function of the normalized mass @xmath26 , which decreases monotonically in time .
then equations ( [ diffoet ] ) are written as @xmath27 \bar{m}\frac{de}{d\bar{m}}=-(e+\cos f ) \\[0.2 cm ] \bar{m}\frac{d\omega}{d\bar{m}}=-\frac{\sin f}{e}\\[0.2 cm ] \bar{m}\frac{df}{d\bar{m}}=-\frac{\bar{m}^{3-n}}{\epsilon } ( 1+e\cos f)^2 + \frac{\sin f}{e } , \end{array } \label{diffoenorm}\ ] ] where @xmath28 from the above equations , and taking into account that @xmath29 and @xmath30 always , we conclude that two orbits with the same initial eccentricity @xmath31 and true anomaly @xmath32 , show an equivalent evolution if the system is characterized by the same value of the parameter @xmath33 .
the initial value of the argument of pericenter @xmath34 affects only the evolution of @xmath35 . in the above analysis
, we considered a star and only one planet .
if there are two , or more , planets , then the evolution of the osculating elements becomes more complicated , because the gravitational interaction between the planets affects in an important way the orbital evolution , as we shall see in the following sections . only in the very special case where
the masses of the planets could be considered as negligible would the evolution of each planet be determined by the system ( [ diffoet ] ) or ( [ diffoenorm ] ) .
it can be proved @xcite that for a slow variation of mass and small eccentricities , the action @xmath36 is an adiabatic invariant , which implies that @xmath37 thus , from the angular momentum integral ( [ integral ] ) , we find that the eccentricity @xmath6 remains constant . from the above
we see that the secular variation of the semimajor axis is given by @xmath38 which implies that the semimajor axis increases monotonically , since the mass decreases . particularly in the case of linear mass loss , @xmath15
, we obtain the adiabatic estimation @xmath39 given by veras et al .
( 2011 ) , while at the minimum mass limit the semimajor axis takes the value @xmath40 the variation of the eccentric anomaly @xmath41 in time , when the star loses mass isotropically , is necessary for estimating the time evolution of the eccentricity and the argument of pericenter , as we will show in the following paragraphs .
the eccentric anomaly obeys the equation ( hadjidemetriou , 1966 ) @xmath42 where the derivatives of the true anomaly and the eccentricity are given by equations ( [ diffoet ] ) . in order to find the secular change of @xmath41
, we ignore in ( [ eadeq ] ) all the periodic terms and then , up to first order terms in eccentricity , we obtain @xmath43 or , by considering the normalization used in equations ( [ diffoenorm ] ) , @xmath44 from the above equations we obtain that @xmath45 always holds true .
the evolution of @xmath41 as a function of @xmath26 depends essentially only on the system s parameters @xmath33 and @xmath13 . for the case of linear mass loss ( @xmath14 )
we obtain @xmath46 where @xmath47 , or @xmath48 hadjidemetriou ( 1966 ) has given a series solution of equations ( [ diffoet ] ) for the eccentricity and the argument of pericenter as a function of the eccentric anomaly @xmath41 , @xmath49 @xmath50 which converge for @xmath51 . assuming @xmath33 as a small parameter we can write , for small eccentricities , an approximate solution for @xmath52 up to @xmath53 as @xmath54 the evolution of the eccentricity in time is given by considering the particular solution @xmath55 of eq .
( [ eadeqest1 ] ) .
apart from the constant term @xmath31 , the first two terms in ( [ solecc ] ) are periodic terms of period @xmath56 ( one planetary revolution ) .
the last two terms indicate also the same periodicity but with a secular variation of the amplitude of the periodic oscillations .
thus , an estimate of the maximum value reached by the eccentricity along the evolution can be approximated by the formula @xmath57 , @xmath58 , @xmath59 and linear mass loss with @xmath60 and @xmath61 .
the dashed line corresponds to the numerical solution of equations ( [ diffoet ] ) .
the solid bold line is the analytic solution ( [ solecc ] ) , where @xmath41 have been substituted by eq . ( [ etlinear ] ) .
the estimate ( [ maxestecc ] ) of the maximum eccentricity is also shown by the thin solid line.,width=302 ] the remaining terms @xmath62 are small for the first planetary revolutions , but they become important as the number of revolutions increases . in fig .
( [ figeccevol ] ) we show an example of the evolution of eccentricity .
we see that the analytic solution ( [ solecc ] ) coincides very well with the real ( numerical ) solution but only for a relatively short time interval .
working in the same way as with eccentricity , we can write an approximate solution for the argument of pericenter @xmath63 where @xmath64 is a secular component of the evolution . note that @xmath65 is zero in the case of a linear mass loss ( @xmath14 ) .
@xmath66{linevolecc.eps } & \includegraphics[width=8cm]{linevolomega.eps } \\ ( a ) & ( b ) \end{array}\ ] ] in the eddington - jean s law given by eq .
( [ ejlaw ] ) . for each case
we give the evolution for @xmath67 , 0.01 and 0.1 , as it is indicated in panel ( d ) .
the initial conditions are @xmath68 , @xmath69.,width=529 ] the analytic estimates found above are sufficiently accurate only for time intervals where the mass loss does not exceed 20 - 30% of the total mass . however ,
the numerical solution show that the eccentricity @xmath6 and the argument of pericenter @xmath7 continue to show oscillations with increasing amplitude for still larger mass loss , until their evolution enters a `` runaway regime '' where the eccentricity will eventually increase monotonically .
the passage of the evolution to the `` runaway regime '' has been studied in detail by veras et al .
( 2011 ) for the case of a linear mass loss . in the following
we present some results obtained by numerical integrations . in fig .
[ figecclinevol ] ( panels a , b and c ) we show some typical examples of the evolution of the eccentricity @xmath6 , where the horizontal axis represents the mass loss ratio @xmath70 , which can be mapped to the time @xmath71 through the particular mass loss law of eq .
( [ ejlaw ] ) . here
we consider the linear law ( @xmath14 ) .
we observe that when the orbits enter the `` runaway regime '' , we soon get @xmath72 and the planet escapes .
such an escape is obtained when the mass loss becomes sufficiently large , equal to a critical value @xmath73 .
this value depends generally on the initial conditions of the system , but when we start with small eccentricity values @xmath31 , @xmath73 depends mainly on the parameter @xmath33 .
we observe that for @xmath74 the orbit enters the runaway regime from the beginning . in the right panels
d , e and f of fig .
[ figecclinevol ] we present the evolution of the argument of pericenter @xmath7 of the same orbits as in the left panels .
again we find that for small values of @xmath33 , @xmath7 librates with an increasing amplitude , as it is indicated by the approximation of eq .
( [ solomega ] ) .
however , in the case of @xmath75 ( panel d ) , @xmath7 increases monotonically . in all cases
@xmath7 completes one revolution at most . by considering @xmath76 , i.e. non - linear mass loss , we find that the evolution of the orbital elements is qualitatively similar to the linear case even though the runaway regime appears in general for larger values of mass loss
. some examples of the evolution of the eccentricity are shown in fig .
[ fignnneccevol ] for the cases @xmath77 , @xmath16 ( exponential decay ) , @xmath78 and @xmath79 . for the case @xmath78 ( panel c ) and for @xmath67 ,
the runaway behaviour appears only when the star mass becomes very small .
an exception is the case @xmath79 ( panel d ) where the eccentricity continues to show small oscillations as @xmath80 and the runaway regime does not exist ( at least for any @xmath81 ) . from the series solution
( [ solecc ] ) we obtain that for @xmath79 the amplitude of the oscillations is constant and the eccentricity is a periodic function of the eccentric anomaly ( see also hadjidemetriou , 1966 ) .
so far we have shown that the planetary eccentricity may take large values either due to its large amplitude oscillations or due to the entrance in the runaway regime .
let us assume that the star mass loss starts at @xmath82 , when @xmath83 , and continues up to a particular mass loss ratio value @xmath19 , which is reached after time @xmath21 .
then , at the end of the star mass loss event , the planetary orbit has reached a new eccentricity value @xmath84 . in the contour maps of fig .
[ figecccontours ] we present the value of @xmath85 in the parameter domain @xmath86 and @xmath87 .
the cases of linear ( left panel ) and exponential ( right panel ) rates of mass loss are given .
the tabulated eccentricity value is the average eccentricity obtained at @xmath21 from 100 orbits with initial eccentricity and true anomaly values that were randomly selected in the intervals @xmath88 and @xmath89 , respectively .
we see that for small values of the parameter @xmath33 ( e.g. @xmath90 ) , a significant increment of the eccentricity is obtained only after a significant mass loss ( e.g. more than 80% in the linear case and more than 90% in the exponential case ) .
as @xmath33 takes larger values , a significant eccentricity increment is observed for smaller amounts of lost mass .
the escape regime ( @xmath72 ) is indicated by the gray shaded region and occupies the region of large values of @xmath19 and @xmath33 , as was expected . by computing the same maps for @xmath78 and @xmath79
, we obtain similar plots .
however , these curves , which correspond to the same eccentricity levels , are obtained for larger @xmath19 values and yield smaller escape regimes .
@xmath91{contourlin.eps } & \includegraphics[width=7cm]{contourexp.eps } \\ ( a ) & ( b ) \end{array}\ ] ] as previously mentioned , when there are at least two planets in the system , the evolution does not depend on the mass loss of the star only ; the gravitational interaction between the planets plays an important role . in order to study this effect , and reveal the extent to which the gravitational interaction influences the evolution , we consider here a planetary system with two planets of negligible mass amidst mass loss . in this case
, no gravitational interaction exists between the planets .
consider a system with two massless planets @xmath92 , @xmath93 , with semimajor axes @xmath94 and angular momenta @xmath95 .
the evolution of each planet depends on the value of the parameter @xmath33 , namely @xmath96 . because @xmath97 is different for each planet
, each planet then follows a different evolution .
if @xmath98 is the outer planet ( @xmath99 ) , then we obtain @xmath100 .
the mean motion resonance is defined by the mean motion ratio @xmath101 , where @xmath102 , which is given by the relation @xmath103 thus two resonat planets with orbits of low eccentricities preserve their mean motion resonance @xmath104 , @xmath105 integers , under the star mass loss and before entering the runaway regime , where the eccentricities take large values .
also , the preservation of the resonance and the libration of the argument of pericenter , mentioned in section [ secunperturbednumoe ] , indicate that the resonant angles @xmath106 and @xmath107 , @xmath93 , should also librate , until a significant stellar mass loss is reached .
@xmath91{linres21evol.eps } & \includegraphics[width=7cm]{linres31evol.eps } \\ ( a ) & ( b ) \end{array}\ ] ] in fig .
[ figresevol ] we present the evolution of the mean motion ratio and of the resonant angles for the 2/1 ( left panel ) and 3/1 ( right panel ) resonance and for @xmath108 , @xmath109 and @xmath110 . for the outer planet
it is @xmath111 and @xmath112 for the 2/1 and 3/1 resonance , respectively . for both resonances
we see that the system evolves in a similar way . for small mass loss parameters @xmath97 , @xmath101 is almost constant before a significant amount of stellar mass loss .
first the outer planet enters the runaway regime and then a rapid increment of @xmath101 is observed .
as we consider larger values of @xmath33 , the system leaves the resonance at a smaller mass loss ratio @xmath70 .
as far as the system remains close to the resonance , the resonant angles show small oscillations around their initial value .
afterwards , the amplitude of the libration increases and the evolution of the resonant angles might end up in circulation . for large values of @xmath33 (
e.g. for @xmath75 ) , we obtain a slow monotonic variation of the angles .
particularly , @xmath113 decreases while @xmath114 increases .
we stop the evolution when the outer planet escapes , i.e. , when @xmath115 , where the mean motion is no longer defined .
here we study the evolution of a two - planet system where the mutual planetary gravitational interaction is included but the star does not lose mass .
thus we establish a dynamical basis with which to compare the mass loss case in the next section .
the study of the dynamics of a system consisting of a star of mass @xmath116 and two planets @xmath92 , @xmath93 of masses @xmath117 can be studied by using the model of the general three body problem ( gtbp ) @xmath118 where @xmath119 indicates the position of the planet and @xmath120 .
we restrict our study to the planar case , @xmath121 where the inertial frame @xmath122 is centered at the barycenter .
also we normalize the units by setting @xmath123 and @xmath124 . in the following
, we should take always the @xmath98 to be initially the inner planet and @xmath125 the outer one in the sense that @xmath126 , where @xmath94 is the semimajor axis of @xmath92 .
although the system appears having four degrees of freedom , we can use the angular momentum integral and a rotating frame and reduce the system to three degrees of freedom ( hadjidemetriou , 1975 ) .
in such a system , regular planetary orbits correspond to quasiperiodic trajectories , which twist on invariant tori in phase space according to the kam theorem
. however the mutual planetary interaction destroys the integrability of the system and chaotic orbits coexist beside the regular ones . ) , chaotic orbits ( @xmath127 , @xmath128 ) and a chaotic orbit ( @xmath129 ) that shows a close encounter after 3.5 ky ( we note that 6.28 time units correspond to about 1 year ) .
the initial conditions are @xmath130 , @xmath131 , @xmath132 , @xmath133 , @xmath134 , @xmath135 and @xmath136 , @xmath137 , @xmath138 and @xmath139 for the orbits @xmath140-@xmath129 , respectively.,width=188 ] many numerical tools for the detection of chaos have been proposed . for example , in planetary dynamics the chaos indicators megno , rli and fli have been used ( godziewski , 2005 ; sndor et al , 2007 ; voyatzis , 2008 , respectively ) . in the present work
we compute the maximal lyapunov characteristic number ( lcn ) , which is the classical measure for the average exponential divergence of nearby orbits .
an example of the evolution of lcn for some different trajectories of the gtbp is shown in fig .
[ figlcnplot ] for @xmath141 time units ( or @xmath142 ky ) .
a regular evolution is described by an lcn evolution that tends to zero , as e.g. the orbit @xmath140 . for chaotic orbits lcn
tends to a positive value , as e.g. in the orbits @xmath127 and @xmath128 .
this value is used to form _ dynamical stability maps _ ( ds maps ) . in these
we define plane grids of initial conditions presented with a color scale representing the lcn value after a particular integration time @xmath143 . for the numerical integration of eq .
( [ gtbpdeqs ] ) we use the bulirsch - stoer integrator . when planetary close encounters occur , the integration may break , in the sense that the integration step becomes very small in order for the method to preserve the requested integration accuracy ( e.g. the orbit @xmath129 ) .
such cases always correspond to strongly chaotic orbits and in the ds maps are presented by the value @xmath144 ( light - coloured regions ) .
b ) the evolution of the energy error along the numerical integration with and without regularization ( solid and dashed curves , respectively ) . , width=264 ] in cases where we need to follow the evolution after a close encounter and without loss of accuracy , we apply a regularization of the collision singularities . in this paper
we apply a levi - civita transformation when the gravitational interaction between any two bodies of the system becomes very strong ( marchal , 1990 ; aarseth , 2003 ) . in fig .
[ figacc ] we present an example of evolution with a sequence of planetary close encounters .
the two planets are placed inside the hill radius and evolve in a satellite configuration . in panel
( a ) we present the planetary distance @xmath145 .
we integrate the system using two methods . in the first we integrate equations ( [ gtbpdeqs ] ) using the bs integrator with accuracy @xmath146 .
when the step becomes very small , instead of stopping the integration , we reduce the requested accuracy . in the second method we use the same integrator as above but we switch on the regularized equations during planetary close encounters . in panel ( b ) of fig [ figacc ] we present the error in energy along the trajectory in both cases .
we can observe that without regularization the error increases after each close encounter , while the integration of the regularized equations preserves the accuracy at a very good level .
we restrict our numerical simulations to planetary systems with the inner planet having mass @xmath147 ( approximately jupiter s mass , @xmath148 ) and semimajor axis @xmath130 au .
the outer planet is either lighter ( @xmath149 ) or heavier ( @xmath150 ) than the inner one and with semimajor axis in the range @xmath151 au . for non - resonant motion ,
the main source of chaos is planetary close encounters , which can happen when the initially elliptic orbits of the two planets intersect . in fig .
[ fignrmaps ] we present ds maps for non resonant motion .
the maps are defined by grids of initial conditions in the plane of eccentricities .
the other orbital elements are fixed and are given in the caption . in panels ( a ) and ( b )
the orbits correspond to an initial mean motion ratio @xmath152 . in case ( a ) , where the outer planet ( @xmath125 ) is more massive than the inner one ( @xmath98 ) , we observe that regular orbits exists only in a well - defined region at small eccentricity values .
some small stable regions appear for @xmath153 up to 0.3 but only for @xmath154 .
a similar picture is obtained also for the case ( b ) , where the inner planet is lighter than the outer one .
now the stable orbits are confined in a region of small eccentricities . above the line @xmath155 ( collision line ) , which is represented by the dashed line in the dynamical maps , the planetary orbits intersect and
strong chaos appears .
we obtain the same dynamical picture in panels ( c ) and
( d ) , where @xmath156 . in this case
the outer planet starts with a larger semimajor axis than in the previous case .
the planetary interactions are weaker , the stable region covers a larger domain and the collision line appears at higher eccentricity values . in fig .
[ figresmaps ] we present ds maps for @xmath157 ( panels a and b ) and @xmath158 ( panels c and d ) . in these cases
we obtain a complex structure in phase space where regions of stable and unstable motion are intermingled ( michtchenko et al , 2008a , b ; hadjidemetriou and voyatzis , 2009 ) .
since resonances may offer phase protection mechanisms , stable long term evolution might appear even for intersecting orbits .
we observe in the ds maps that we can have regular evolution when the orbit of the inner planet is very eccentric .
however when the eccentricity of the outer planet becomes larger than about 0.2 we obtain a wide chaotic sea where trajectories are strongly chaotic .
for the 3/1 resonant case with @xmath159 we observe a zone of regular orbits for @xmath160 .
the existence of such zones and , generally , the distribution of chaotic and regular regimes in phase space is significantly affected by the families of periodic orbits ( symmetric or asymmetric ) that , generally , exist in the resonant regions ( see e.g. voyatzis , 2008 ) .
resonant motion is related with the existence of monoparametric families of periodic orbits in phase space .
such periodic orbits are also called exact resonances ( beaug et al , 2003 ) , which can be stable or unstable and contribute significantly to the topology of phase space and the existence of order and chaos . in particular ,
stable periodic orbits are surrounded by regions of regular motion where the resonant angles @xmath161 , @xmath114 and @xmath162 librate . outside these regions stable motion
can be also obtained , where some of the resonant angles circulate ( michtchenko et al , 2008a , b ; voyatzis , 2008 ) . in fig .
[ figpomaps ] we present maps of dynamical stability which show the distribution of regular motion around 2/1 resonant periodic orbits , which belong to the symmetric family of periodic orbits for @xmath163 and @xmath150 with configuration ( @xmath164 ( family @xmath165 in voyatzis et al , 2009 ) .
the maps are defined on the plane of resonant angles @xmath166 and we see that the regular region , which surrounds the periodic orbit at @xmath167 , shrinks as the eccentricities increase .
resonant families of stable periodic orbits are of particular importance , since they form paths in phase space for a migrating planetary system ( lee and peale , 2002 ; ferraz - mello et al , 2003 , hadjidemetriou and voyatzis , 2010 , 2011b ) .
so , regions around stable periodic orbits are candidates for containing a planetary system after a migrating process and a capture in resonance . in this section
we study the evolution of a planetary system under the gravitational interaction between the two planets , when the system is initially located in a chaotic region .
the mechanism that transfers an initially stable system to a chaotic region will be studied in the next section . and @xmath125 along a strongly chaotic orbit with initial conditions as those in the ds map of figure [ figresmaps]a and for @xmath168 and @xmath169 .
the planet @xmath125 escapes after about 30 ky.,width=340 ] @xmath170{pescevole030.eps } & \includegraphics[width=6cm]{pescevole026.eps}\\ \textnormal{(a ) } & \textnormal{(b ) } \\ \end{array}\ ] ] since the system is of more than two degrees of freedom , invariant tori are not boundaries for the chaotic regimes in phase space and fast or slow diffusion of chaotic motion is possible ( see e.g. lichtenberg and lieberman , 1983 ) .
also , if we take into account that in the gtbp zero velocity curves do not exist to bound the planetary orbits ( marchal and bozis , 1982 ) , then we may conjecture that chaotic motion leads to the escape of a planet after a long term evolution .
this has been established after extensive numerical simulations of the gtbp ( see valtonen and karttunen 2006 , and references therein ) . in the following
we discuss the particular case of a planar two - planet system . in the dynamical maps presented in the previous subsections , we observed wide chaotic regions with strongly chaotic motion .
many numerical integrations of orbits with initial conditions inside these regions showed that , during the evolution , close encounters between planets occur and the less massive planet is scattered to orbits with larger eccentricities .
the continued irregular evolution results in a sequence of close encounters that ejects the planet to large distances and finally to escape .
we classify the evolution as escape when a planet that starts from a distance of order 1 au , moves to a distance larger that 1000 au with eccentricity @xmath72 . in these numerical integrations
we set the accuracy in the bs integrator to 14 digits and we use the regularized equations during close encounters . in fig .
[ figevolrmin ] we present the evolution of the planetary distance for an orbit starting from the strongly chaotic ( light colored ) region appearing in the map of dynamical stability of fig .
[ figresmaps]a at @xmath171 .
the first close encounter appears after about 8 ky .
a sequence of close encounters follows and after 20 ky scattering of the planet @xmath125 to large distances is observed .
finally the planet escapes at about 30 ky . in fig .
[ figevolae]a we present the evolution of the semimajor axis and the eccentricity .
after the first close encounter the eccentricity and the semimajor axis of @xmath125 show a jump to higher values and their variation in time becomes very irregular .
the orbit of the heavier planet @xmath98 shows also the same irregularity , but the variation of its orbital elements is quite smaller .
finally we get @xmath115 and for the remaining planet @xmath172 , @xmath173 .
a second example of chaotic evolution is presented in fig .
[ figevolae]b .
now we start from the point @xmath174 which is closer to the regular ( dark colored ) region of the ds map of fig .
[ figresmaps]a . up to 120
ky the evolution of the semimajor axis and eccentricity of both planets seems vary regular .
afterwards chaotic evolution appears and a sequence of close encounters takes place , similar to the previous case .
finally planet @xmath125 escapes and the remaining planet @xmath98 moves on an elliptic orbit with @xmath175 and @xmath173 . by considering initial conditions that correspond to chaotic motion but are closer to the regime of regular motion , we obtain similar planetary destabilization , which , however , occurs after a longer time span . for example , by considering the same initial conditions as above but for @xmath176 and @xmath177 we found that escape appears at 0.9my and 10my , respectively .
however we should remark that the escape time is very sensitive to the numerical integration accuracy .
chaos is a necessary but not sufficient condition for planetary ejection over reasonable time spans ( e.g. ms lifetimes ) .
if we consider the initial conditions in a narrow chaotic zone in the maps of dynamical stability and far from the wide chaotic sea , then chaos may remain bounded throughout long - term evolution .
in other words , consider a system with initial conditions in the chaotic region comprising small eccentricities ( @xmath178 ) of the ds map of fig.[figresmaps]a .
although this system evolves apparently irregularly , up until 100my the system is stable in the sense that no ejection or collision occurs .
the existence of bounded chaos for orbits which are hill stable is also noted by @xcite .
here we present our main results .
we analyze how mass loss could couple with the planets mutual interaction to cause one of them to escape . in the following study
we start with a two - planet system , which is trapped in a stable ( resonant or non - resonant ) configuration .
then we assume that the star begins to lose mass in an isotropic way , but in a manner where the percentage of the mass lost and the rate of mass loss are not large enough to cause the direct ejection of the planet
. however , as we showed in section 2 , even in these cases the orbital elements of the planetary orbits are affected .
so the system migrates in phase space and after the end of the mass loss process it can be found inside a chaotic region . from that point on
, the chaotic evolution of the system suffers from close encounters , allowing for the possible ejection of a planet ( see section [ chaotic ] ) .
we consider a linear rate of mass loss , @xmath15 , which takes place up to time @xmath179 where the mass of star is reduced to @xmath180 ( see section 2 ) . in all the figures showing the evolution of the system ,
the inner planet has the subscript `` 1 '' and is represented by a blue line and the outer planet has the subscript `` 2 '' and is represented by a red line .
@xmath170{zvmi_po1a.eps } & \includegraphics[width=6cm]{zvmi_po1b.eps } \\
\textnormal{(a ) } & \textnormal{(b ) } \end{array}\ ] ] here , the initial total mass of the system is set to one solar mass , namely @xmath181 . as an example we consider the stable 2/1 resonant periodic orbit given in fig .
[ figpomaps]a and we apply a star mass loss @xmath182 and rate @xmath183 . the evolution of the planetary semimajor axes and eccentricities for the first years of the mass loss is shown in fig .
[ figvmipo]a .
the semimajor axes increase for both planets approximately according to the adiabatic estimate given by eq .
( [ semiaxisevol ] ) , but the ratio of mean motions @xmath101 , represented by the horizontal dashed line , is almost constant , as it is suggested by eq .
( [ resevol ] ) . after the end of the mass loss process , which ceases at 120 years
, the eccentricities have increased slightly but the mass of the star has decreased significantly .
such new conditions corresponds now to a chaotic orbit as it becomes evident in fig .
[ figvmipo]b .
we observe that the eccentricities incur well - bounded but weakly chaotic oscillations up to 70 ky .
then the system enters a strongly chaotic region and the eccentricities show large and irregular variations up to about 290 ky . in this interval ,
the variation of the semimajor axis of the planet @xmath125 , which is the heavier one , is significant . after this time interval
the semimajor axis and the eccentricity of the planet @xmath98 increase rapidly and finally the planet is ejected .
it is obvious that if we ignore the gravitational interactions between the planets , the orbits of both planets would be constant ellipses with orbital elements attained at @xmath21 .
a. the initial conditions are those of the periodic orbit except that we set @xmath184 and the mass loss parameters are @xmath185 and @xmath186.,width=302 ] the above is a typical example of destabilisation after stellar mass loss .
the same results are obtained for the periodic orbits in panels ( b ) and ( c ) of fig .
[ figpomaps ] . as we mentioned in section 3 ,
stable periodic orbits are in the centers of islands of stability .
if we consider initial conditions far from the periodic orbit and near a chaotic region , destabilization can happen for smaller values of the parameters @xmath12 and @xmath19 .
an example is shown in fig .
[ figvmipof ] where we use the same initial conditions except that @xmath187= @xmath188 , which means that the orbit is located at the point @xmath189=@xmath190 of the ds map of fig .
[ figpomaps]a .
now the system is destabilized for @xmath185 and @xmath186 .
in section [ secdsm ] we presented ds maps on the plane of eccentricities , which show the distribution of regular and chaotic orbits when the mass of the star is constant . according to the mechanism described above
, the regular orbits can become chaotic if stellar mass loss takes place . in order to make a more extensive study of the possibility of planetary destabilization
, we consider the initial conditions of the regular orbits in the ds maps of figures [ fignrmaps ] and [ figresmaps ] and we follow the evolution of the system by considering star mass loss with @xmath185 and @xmath182 .
for each orbit we compute the lcn and thus we determine which of the orbits become chaotic after the stellar mass loss .
we assume that if lcn is greater than @xmath191 , then the evolution is chaotic .
the results are presented in fig .
[ figvmmaps ] .
the black shaded areas denote the orbits which remain regular after the mass loss process .
the gray shaded area denotes the regions where the initially regular orbits become chaotic . for each map
we indicate the percentage @xmath192 of the orbits that become chaotic . in the resonant case with @xmath193 and the nonresonant
one with @xmath194 , we find that a large fraction of orbits is destabilised .
in contrast , at the @xmath195 resonance and for the nonresonant case @xmath196 , the majority of the orbits remain regular , especially when the outer planet is the less massive one .
we note that in the first cases ( @xmath193 and @xmath194 ) the planets are closer to each other and their gravitational interaction is stronger .
thus , we may claim that planetary orbits are more affected by stellar mass loss as the planetary interactions becomes stronger .
this important claim is not directly assessed in @xcite nor @xcite , as the simulations in those studies are not set up to address this issue .
further , in those studies , the planets are of the same mass and the outer planet is the one that escapes . in our integrations
the planet which escapes is always the lighter one ( see also @xcite ) .
.thermally pulsing asymptotic giant branch properties [ cols="^,^,^,^",options="header " , ] although this study is focused on the dynamical properties of the general three - body problem with mass loss , we can relate the results to real systems .
doing so helps us determine in what cases will significant mass loss affect planet - planet scattering in a non - adiabatic manner . in order to make this relation , we use the sse ( _ single star evolution _ ) code @xcite to create stellar evolutionary tracks . for solar metalicity stars , and a reimers mass loss coefficient of @xmath197
, we trace the time evolution of stars with the following progenitor masses : @xmath198 , @xmath199,@xmath200,@xmath201,@xmath202,@xmath203,@xmath204 and @xmath205 .
we find that for all cases except the @xmath198 case , the vast majority of the mass is lost on the thermally pulsing asymptotic giant branch ( tpagb ) phase .
this phase , which is typically one of the shortest in duration , also typically features the greatest mass loss rates .
this combination is well - suited for observing the type of phenomenon seen in the fiducial simulations .
therefore , we neglect the other phases of evolution and assume that all the most lost occurs during the tpagb . we summarize the output of sse , which includes @xmath19 and @xmath206 ( for the tpagb phase only ) in table 1 .
then we remove the assumption that @xmath181 and compute @xmath12 and record that value as well in the last column .
the table suggests that adopting @xmath207 may reflect real systems with progenitor stellar masses of @xmath208 .
@xmath91{zvmi_m8ee.eps } & \includegraphics[width=7cm]{zvmi_m8esc.eps}\\ \textnormal{(a ) } & \textnormal{(b ) } \\ \end{array}\ ] ] in fig .
[ figevolm8 ] , we show a typical example of the evolution of such a multi - planet system with a star of mass @xmath205 and two planets with @xmath209 and @xmath210 . the planets are initially located in the 2/1 resonance with initial eccentricities @xmath135 and @xmath176 .
the numerical integration shows long term stability if the star does not lose mass .
in particular , the eccentricities oscillate regularly about their initial values .
if we introduce a mass loss process with a rate @xmath211 and @xmath212 then we obtain the evolution given in fig .
[ figevolm8]a . in the time interval ,
@xmath213 the eccentricities show an adiabatic variation besides their fast oscillations .
when the mass loss ceases , the system continues to evolve regularly with the eccentricities oscillating ( with fast and slow components ) about the values 0.13 and 0.19 for the inner and the outer planet , respectively .
the evolution becomes very different if we consider a larger amount of mass loss .
particularly for @xmath182 we get the evolution shown in fig .
[ figevolm8]b .
now , at @xmath21 the planetary system has entered a chaotic region and its evolution for @xmath214 is evidently chaotic .
after 4.1 my close planetary encounters destabilize the system and the mutual gravitational attraction of the planets forces the outer , lighter , planet to be ejected from the system after about 5 my .
evolution in resonance is not a necessary condition for destabilization .
figure [ figevolm4 ] shows the evolution of a non - resonant system ( @xmath215 ) which consist of a 4@xmath216 star and two planets with @xmath217 , @xmath218 , @xmath219 , @xmath220 , @xmath135 and @xmath221 .
note that in this case the inner planet is the smaller one .
now we consider realistic values of @xmath222 and @xmath223 ( see table 1 ) . under the stellar mass loss the semimajor axes increase to @xmath224 and @xmath225 .
the system shows some strong irregularities in the evolution of eccentricities and the orbit of the inner planet becomes very eccentric .
particularly , in the time interval @xmath226 the eccentricity of the outer planet oscillates about its initial value but the orbit of the inner planet shows eccentricity values in the interval @xmath227 . for @xmath228 close planetary approaches force the inner and lighter planet to become ( temporarily ) the outer one in the system .
the evolution becomes very irregular and , finally , the lighter planet is ejected out to a distance larger than 1000 au .
we have also performed integrations ( up to @xmath229 my ) indicating that , without mass loss , this system is stable and regular .
therefore , mass loss is the trigger for the instability in this 4@xmath216 case . in general the numerical simulations
show that the position of the system in phase space determines essentially its robustness to perturbations caused by stellar mass loss .
also , the planetary destabilization seems to be sensitive to the parameters @xmath12 and @xmath19 .
concerning the case of systems of more than two planets and also taking into account the results of @xcite and @xcite , we may claim that these systems may be more sensitive to small perturbations than 2-planet systems .
as soon as more planets are included in a system , the regions of stability in phase space is reduced and instabilities are more likely to occur after a potential decrease of the star s mass . our study might help explain discrepancies in the observed populations of 2-planet systems on the ms with those beyond the ms . to date , no planets orbiting white dwarfs have been confirmed , although white dwarfs are present as distant companions in known planetary systems ( gl 86 : @xcite ; hd 27442 : @xcite ; hd 147513 : @xcite ) .
alternatively , over 30 planets are known to orbit giant stars ( see table 6 of @xcite ) and well over 600 are confirmed to orbit ms stars@xmath230 . among the mechanisms invoked to explain the general lack of known post - ms planets are different modes of formation around giant stars @xcite and direct engulfment into the star itself due to star - planet tides ( e.g. @xcite ) .
we are proposing a third mechanism : scattering in two - planet systems that is induced by mass loss .
the planets in these systems are far away enough from their parent star to remain unaffected by tidal effects . a potential fourth mechanism , albeit one which will be complex to model , is two - planet scattering where the planets are close enough to the star to be affected by both mass loss and tides .
this represents a potentially relevant extension to this work given that at least one planet in the vast majority of currently observed 2-planet systems will likely be affected by tides during post - ms evolution . therefore understanding the fate of multi - planet systems crucially depends on effects such as the one we have modelled here .
we have explored properties of the three - body problem with mass loss .
an astrophysical application for this work is two - planet post - ms systems .
we proposed a mechanism , which combines the effects of stellar mass loss and mutual planetary gravitational interactions , that can destabilize a planetary system .
consequently , the lighter planet escapes and becomes an orphan planet .
in particular , the coupling between mass loss and mutual interactions between the planets is not well - understood and was previously largely unexplored .
we have identified the regions of phase space where the coupling is strong , causing initially stable systems to remain stable during the mass loss but later exhibiting escape .
the stellar mass loss may transfer the system from its initially stable region to a wide chaotic sea in the phase space , which is associated with the final stellar mass .
in such a chaotic region escape of the lighter planet always occurs .
although we find this evolutionary sequence to be robust to small changes in initial conditions , we find that the escape time is sensitive to these changes . however , it seems that chaotic evolution , which is detected by the maximal lcn during an integration interval up to @xmath231 ky , results in most cases to destabilization of the system in less than 100 my .
the general phase space maps presented can be used to characterize a wide variety of real systems . in the numerical simulations of our paper we used a constant rate of mass loss .
however , we have obtained qualitatively similar results for an exponential rate of mass loss : fig .
4 demonstrates excellent agreement between the constant mass loss model and the exponential model . also , we note that mass loss along the asymptotic giant branch closely resembles either constant or exponential mass loss .
we find that the upper - mass end of potential planetary systems that will experience a white dwarf phase ( @xmath232 ) is likely to exhibit the chaotic evolutionary behaviour seen here .
just as importantly , lower - mass planetary systems , such as the currently - observed bound multi - planet exoplanet population , will experience adiabatic planet - planet interactions during post - ms mass loss . claiming that the adiabatic approximation holds in these cases for the three - body problem is useful , and will simplify the analytics for future post - ms three - body studies . ,
@xmath217 and @xmath218 .
the star loses mass linearly with rate @xmath223 and @xmath222 .
the initial conditions are @xmath219 and @xmath220 , @xmath135 , @xmath221 , @xmath233 , @xmath133 , @xmath234 . , width=302 ] hadjidemetriou j.d . , voyatzis g. , 2009 , in contopoulos g. , patsis p.a . , ( eds ) , assp chaos in astronomy , springer .
hadjidemetriou j.d .
, voyatzis g. , 2010 , cmda , 107,3 hadjidemetriou j.d .
, voyatzis g. , 2011a , cmda , 111 , 179 hadjidemetriou j.d . , voyatzis g. , 2011b , ijbc , 21 , 2195 valtonen m. , karttunen h. , 2006 , the three body problem , cambridge university press varvoglis h. , sgardeli v. , tsiganis k. , 2012 , cmda , 113 , 387 veras d. , wyatt m.c .
, mustill a.j .
, bonsor a. , eldridge j.j .
, 2011 , mnras , 417 , 2104 veras d. , mustill a.j . , bonsor a. , wyatt m.c .
, 2012 , submitted to mnras | discoveries of exoplanets orbiting evolved stars motivate critical examinations of the dynamics of @xmath0-body systems with mass loss .
multi - planet evolved systems are particularly complex because of the mutual interactions between the planets . here , we study the underlying dynamical mechanisms which can incite planetary escape in two - planet post - main sequence systems .
stellar mass loss alone is unlikely to be rapid and high enough to eject planets at typically - observed separations .
however , the combination of mass loss and planet - planet interactions can prompt a shift from stable to chaotic regions of phase space .
consequently , when mass loss ceases , the unstable configuration may cause escape . by assuming a constant stellar mass loss rate , we utilize maps of dynamical stability to illustrate the distribution of regular and chaotic trajectories in phase space .
we show that chaos can drive the planets to undergo close encounters , leading to the ejection of one planet .
stellar mass loss can trigger the transition of a planetary system from a stable to chaotic configuration , subsequently causing escape .
we find that mass loss non - adiabatically affects planet - planet interaction for the most massive progenitor stars which avoid the supernova stage .
for these cases , we present specific examples of planetary escape .
[ firstpage ] @xmath1 |
the nearest x - ray emitting cluster of galaxies is the virgo cluster in which the x - ray emission is seen to be sharply peaked towards the cluster center ( schreier , gorenstein & feigelson 1982 ) . in the central few tens of kpc
the temperature of the cluster gas decreases to @xmath5 kev ( canizares et al .
1979 , 1982 ; lea , mushotzky & holt 1982 ; bhringer et al . 1994 ; nulsen & bhringer 1995 ; bhringer et al . 2001 ; belsole et al . 2001 ) and within 10 kpc the cooling time is @xmath6 1 gyr ( stewart et al . 1984 ) , suggesting the presence of a 3 @xmath7 yr@xmath8 cooling flow ( lea et al . 1982 ; canizares et al . 1982 ) .
simple cooling flow models have some problems , however , such as the single temperature phase of the intra - cluster medium in virgo ( bhringer et al .
2001 ) , and the observed absence of x - ray emitting gas with temperature below 1 2 kev in other clusters ( e.g. schmidt , allen & fabian 2001 ; peterson et al .
2001 ; fabian et al . 2001 ; tamura et al .
2001 ) . the giant elliptical galaxy and
powerful radio source m87 ( virgo a ) lies at the core of the virgo cluster . at radio wavelengths
m87 is seen to power spectacular jets and lobes with significant structure extending over many tens of kpc ( owen , eilek & kassim 2000 ) .
since m87 and the core of the virgo cluster are nearby ( 15.9 mpc [ tonry 1991 ] ) , they have a large angular size on the sky ( @xmath9 ) and present an ideal system in which to study the nature of intra - cluster gas and its interaction with the radio source .
the interaction of the radio source with the cluster gas is thought to be important because , at the present epoch , the inferred , model - dependent , kinetic power in the jet is in the range @xmath10 erg s@xmath8 ( bicknell & begelman 1999 ) , which may exceed the total x - ray luminosity of the cooling core ( @xmath11 erg s@xmath8 ; fabricant & gorenstein , 1983 ) . observations with _ einstein _
( feigelson et al 1987 ) revealed an `` arc '' of x - ray emission that is well correlated with the intermediate scale ( nuclear distance @xmath12 ) radio ridges , and subsequent _ rosat _ ( bhringer et al .
1995 ; harris , biretta & junor 1999 ) and _ xmm - newton _ ( belsole et al . 2001 ; bhringer et al .
2001 ) observations find the gas in this arc to be cooler than surrounding cluster gas .
since the x - ray and radio features are quite well aligned , it is assumed there is some causal connection between them
. models which may account for this connection include those of churazov et al .
( 2001 ) , brggen & kaiser ( 2001 ) and reynolds , heinz & begelman ( 2002 , hereafter rhb ) , in which buoyant bubbles of radio plasma interact with the intra - cluster medium .
rhb also find that the transient activity of the radio galaxy sends spherical `` pulses '' into the intra - cluster medium .
we present here our first results on the cluster x - ray emission based on a 40 ks _
observation of m87 and the core of the virgo cluster .
the galactic column density towards m87 is taken to be @xmath13 ( stark et al . 1992 ) .
the core of the virgo cluster was observed by _ chandra _ on 2000 june 29 for 40 ks with m87 at the aim - point of the advanced ccd imaging spectrometer ( acis ) , on the s3 chip . after filtering for times of high background
the total usable exposure time was 36608 s. the acis background was obtained from a compilation of relatively blank fields prepared by maxim markevitch ] , and subtracted from our images and spectra .
the images were corrected for variations in the exposure time across the field caused by chip boundaries and telescope vignetting .
ciao 2.1.3 and caldb 2.7 were used to extract the data and xspec 11.0.1 was used to model the spectra .
at radio wavelengths m87 has a hierarchical morphology , with bright inner lobes at a nuclear distance @xmath14 kpc ( @xmath15 ) , intermediate ridges at a nuclear distance @xmath16 kpc ( @xmath17 ) and an outer diffuse `` halo '' at a nuclear distance @xmath18 kpc ( @xmath19 ) ( we shall refer to these as the `` inner '' lobes , `` intermediate '' ridges , and the outer halo , respectively ) . for reference ,
each acis chip is approximately @xmath20 , so we shall concentrate on the morphology on scales comparable to the inner and intermediate radio structures .
nevertheless , the s3 and s2 chips combined cover almost all of the region corresponding to the outer halo radio emission .
adaptively smoothed 0.5
7 kev x - ray images of the core of the virgo cluster are shown in fig .
[ fig : cxo ] for three different spatial scales .
there is considerable structure on all scales from arcsecond to tens of arcminutes .
concentrating on the central arcminute or so ( bottom panel of fig .
[ fig : cxo ] ) , the brightest structure is the nucleus plus famous jet extending towards the northwest ( wilson & yang 2002 ; marshall et al .
the x - ray surface brightness in the central arcminute is very clumpy and a number of point sources are visible
. there are also major depressions in the x - ray emission ; one to the west of the nucleus just beyond the jet , one immediately to the southeast of the nucleus , and another beyond that ( i.e. 15 @xmath21 from the nucleus ) to the southeast and south . a morphological comparison between the x - ray , 6 cm radio ( hines , eilek & owen 1989 ) and h@xmath4 + [ n ii ] @xmath22 , 6584 ( sparks , ford & kinney 1993 ) emission of the central @xmath23 is shown in fig .
[ fig : cxo_radio_halpha ] .
the radio and x - ray emissions from the jet are well correlated ( wilson & yang 2002 ; marshall et al .
the two inner radio lobes , however , correspond to depressions in the x - ray surface brightness .
such a phenomenon has been seen in other clusters , such as hydra a ( mcnamara et al . 2000 ) and perseus ( fabian et al . 2000 ) , although on much larger spatial scales .
the h@xmath4 + [ n ii ] emission line gas is also excluded from the inner radio lobes with the optical line - emitting gas predominantly on the north side of the radio structure ( sparks 1999 ) .
some knots in the emission line and x - ray gas are associated with each other , especially on the northern and eastern boundaries of the inner eastern radio lobe . as can be seen in fig .
[ fig : cxo ] the decline in x - ray surface brightness with increasing nuclear distance is not smooth . instead , there are quite abrupt `` edges '' in brightness at nuclear distances of @xmath24 45 and @xmath24 3 .
these `` edges '' may also be seen in a plot of the azimuthally averaged surface brightness vs. nuclear distance shown in fig .
[ fig : beta ] .
it is notable that these two roughly circular features are just outside the inner radio lobe and at comparable nuclear distance to the intermediate radio ridges , respectively .
this correspondence suggests the `` edges '' may have resulted from time variability in the active nucleus , perhaps being associated with past epochs of jet production . in the model of rhb , a short episode of jet activity eventually produces buoyant bubbles of radio plasma that rise at @xmath25 times the sound speed , along with a spherical `` pulse '' that expands at approximately the sound speed .
we speculate that the two roughly circular `` edges '' may correspond to the spherical pulses associated with two earlier jet ejecta . while rhb discuss the results of their simulation in the context of powerful fr - ii radio sources , both types of feature the buoyant bubbles and the spherical pulse
are expected to be present in the late - time evolution of a lower power ( fr - i ) radio source , such as m87 .
this model is discussed further in section [ sec : fronts ] .
the most striking feature on large scales is an x - ray `` arc '' running from @xmath24 33 east of the nucleus , across the central region of m87 , to @xmath24 4@xmath26 to the southwest ( figs [ fig : cxo ] and [ fig : color_cxo_radio ] ) . on the east side of the nucleus
, the arc corresponds to a plume of radio emission ( cf .
owen , eilek & kassim 2000 ) , and ends at the center of the `` ear''-shaped eastern lobe of radio emission . on the west side
, the narrow arc passes to the south of the intermediate scale radio ridges 2
3 west of the nucleus . further out , the arc follows diffuse filaments of radio emission that connect the intermediate scale radio ridges with the fainter , outer radio halo .
to get an idea of the spatial variation of the properties of the intra - cluster medium we first constructed an emission - weighted temperature map . to do this , images were made in soft ( 0.5 0.9 kev ) , medium ( 0.9 1.2 kev ) and hard ( 1.2 4 kev ) energy bands .
the background subtracted data were then adaptively binned ( sanders & fabian 2001 ) to ensure comparable fractional errors per bin in each band .
theoretical color ratio tables were computed as a function of temperature and metalicity using the mekal thermal plasma model in xspec , assuming absorption by only the galactic column .
the color ratios of the data were compared with these theoretically computed color ratios to construct a temperature map of the virgo cluster ( fig .
[ fig : temp ] ) .
a few things are immediately apparent from the emission - weighted temperature map : ( i ) gas in the x - ray arc is cooler than surrounding intra - cluster gas at a similar nuclear distance , ( ii ) there is a region of cooler gas immediately surrounding the nucleus extending @xmath27 to the north and @xmath28 to the south , corresponding to the region of higher surface brightness around the nucleus ( see fig . [
fig : cxo ] ) , and ( iii ) there is no evidence of strong shock heating associated with the inner radio lobes .
since a deprojection analysis was not performed , the temperature map shows the emission - weighted mean temperature in each bin .
if the gas distribution is approximately spherically symmetric , with the density rapidly increasing inwards , the properties of the gas in a given bin are dominated by the properties of the gas at the same distance from the nucleus as the projected distance of the bin from the nucleus .
the azimuthally averaged brightness profile of the inner part of the virgo cluster is well described ( i.e. to within @xmath29 at all nuclear distances ; see fig .
[ fig : beta ] ) by an isothermal `` @xmath30-model '' ( sarazin & bahcall 1977 ) in which surface brightness @xmath31 ^ 2)^{-3\beta + 0.5}$ ] with a core radius @xmath32 and @xmath33 , in good agreement with _
hri results ( bhringer 1999 ) .
thus , the x - ray emission from the virgo cluster is strongly peaked towards the center with the possible exception of the inner @xmath34 and the x - ray arcs , so the inferred temperatures are a reasonable approximation to the temperatures of gas at that physical distance from the nucleus . motivated by the temperature map ( fig .
[ fig : temp ] ) , spectra were extracted from a number of @xmath35 square regions , each containing @xmath36 counts , to compare , in detail , the gas properties in the arc and the ambient intra - cluster medium .
an emission - weighted instrument response was made for each region , and the individual spectra were modeled by either ( i ) a single temperature vmekal thermal plasma model or ( ii ) a two temperature vmekal model if the additional component improved the @xmath37/d.o.f . by @xmath38 .
the abundances of o , mg , si , s and fe were allowed to vary but , if a two temperature model was used , both components were constrained to have the same abundances .
the variation of temperature(s ) with nuclear distance for different azimuthal angles is shown in fig .
[ fig : temp_vs_r ] .
the arc to the east and southwest is much better described by a two temperature model , with temperatures of @xmath39 kev and @xmath3 kev .
in contrast , the ambient gas is , in general , better described by a single temperature model with the temperature rising outwards from @xmath40 kev at 6 kpc to @xmath41 kev at @xmath42 kpc .
the region within @xmath43 9 kpc , corresponding approximately to the apparently cooler gas immediately surrounding the nucleus ( see fig . [
fig : temp ] ) , is also better described by a two temperature model , with temperatures of @xmath3 kev and @xmath44 kev .
the properties of the gas in the arc are clearly different from those in the surrounding cluster gas . focusing on the cooler region immediately surrounding the nucleus , we see that the coolest gas in the central arcminute is well correlated with optical h@xmath4 + [ n ii ] @xmath22 , 6584 line emission , which is also concentrated to the north of the nucleus ( sparks et al . 1993 ; our fig .
[ fig : cxo_radio_halpha ] ) . moving farther out
, there is another small filament of h@xmath4 + [ n ii ] emitting gas located near the northwestern corner of the `` ear '' ( gavazzi et al .
this filament corresponds to a region of increased x - ray surface brightness ( fig .
[ fig : cxo ] , upper left ) and cooler ( fig .
[ fig : temp ] ) gas .
the coincidence of the cooler x - ray gas with the h@xmath4 emitting gas suggests these may be regions in which mass is dropping out of the hot intra - cluster medium . to study the pressure and cooling time in a given bin
, the emission measure was converted to an electron density by assuming the path length through the bin is equal to the distance of that bin from the nucleus , an assumption supported by the rapid increase of x - ray surface brightness towards smaller nuclear distances ( see above discussion ) . the electron density ( and gas pressure )
are @xmath45 , where @xmath46 is the path length through the gas so a factor of 2 error in the path length would correspond to only a factor of @xmath47 error in these quantities . to minimize the error in the assumed path length
, we treated the arcs to east and southwest separately and assumed the path length through an arc is 1 kpc for those regions of the arc at a nuclear distance @xmath48 ( @xmath49 kpc ) . the cooling time was computed using the cooling function of sarazin & white ( 1987 ) for solar metallicity gas , correcting for the printing error noted by westbury & henriksen ( 1992 ) .
the rate of cooling ( @xmath50 ) of solar metallicity gas is faster than zero metallicity gas by factors of @xmath40 , 2 and 6 at temperatures of 10 , 2 and 1 kev , respectively . the electron density , total gas pressure and cooling time are shown in fig . [ fig : d_p_ct ] as a function of nuclear distance for a number of different azimuthal angles .
away from the arc , the density and pressure decline rapidly with increasing nuclear distance while the cooling time rises .
the properties of the gas away from the arc are similar to those previously derived for the cluster as a whole , with the run of temperature with nuclear distance being consistent with _ xmm - newton _
( bhringer et al . 2001 ) and the run of density with nuclear distance being consistent with _ rosat _ ( nulsen & bhringer 1995 ) . the gas in the arcs , however , is denser and probably over - pressured with respect to the surrounding cluster gas at a similar nuclear distance ; a deprojection analysis of the ambient cluster emission would provide a more accurate estimate of the pressure of the ambient gas , but would be limited by the lack of symmetry of the x - ray emission .
in addition , the density and pressure decline more slowly with increasing nuclear distance in the arc than in the surrounding cluster gas , although the slope of this decline is sensitive to the assumed path length through the gas in the arc .
a knowledge of the metal abundances of the cooler gas in the arc and the hotter surrounding cluster gas is important for differentiating between various theoretical models of the arc ( see section [ sec : arc ] ) . to study the metal abundances to the east of the nucleus
, counts were extracted from two @xmath35 square regions , one centered on the cooler gas @xmath51 east of the nucleus in the arc ( which we refer to as the `` arc '' gas ) , and another over the nearby cluster gas @xmath51 north of that ( which we refer to as the `` ambient '' gas ) .
there are 8667 and 3323 cts in these two regions , respectively . in order to model the spectra ,
an emission - weighted mean instrument response was used .
this procedure does not provide a perfect description of the instrument response , with residuals corresponding to the iridium m - edge ( 1.8 2.2 kev , as noted by markevitch and vikhlinin [ 2001 ] ) , and at energies corresponding to uncertainties in acis gain maps ( 1.6 and 2.6 kev ) , the effect of which is to increase both the @xmath37 values of our fits and the uncertainty in our determination of the si and s abundances that have strong emission lines in these energy ranges .
the ambient gas to the northeast of the nucleus may be described by a single temperature thermal plasma model ( the vmekal model in xspec ) in which the abundances of o , mg , si , s and fe are allowed to vary , with @xmath37/d.o.f .
= 134/110 .
the parameters of this model are given in table [ tab : spec ] and the data and model are shown in the left panel of fig .
[ fig : spec ] .
the spectrum of the arc to the east of the nucleus is more complicated , and is poorly described by a single temperature thermal plasma model , which has a @xmath37/d.o.f .
. the arc can be described by a two temperature thermal plasma model , with temperatures of 1.77 kev and 1.07 kev .
there is a degeneracy between the relative normalizations and metalicities of the two components , and this can be broken by assuming the warmer component has the same parameters as the ambient gas .
the cooler component is then found to have enhanced mg , si , s and fe abundances relative to the ambient gas and comparable o abundance .
the model parameters are given in table [ tab : spec ] and the model fit shown in the right panel of fig .
[ fig : spec ] with a @xmath37/d.o.f .
= 188/136 . to the west of the nucleus
, counts were extracted from six @xmath52 square regions , three centered on the cooler gas in the arc @xmath51 southwest of the nucleus , and three over the nearby cluster gas a similar distance west of the nucleus .
there are 18065 and 10815 cts in the three regions over the arc and the three regions over the ambient gas , respectively .
these extraction regions are larger than those used to the east of the nucleus because we found that a greater s / n ratio was required to discern variations in the metal abundances .
the spectra of the arc and the nearby cluster gas were then analyzed in the same way as the spectra from the regions east of the nucleus .
the ambient gas to the west of the nucleus can be described by a single temperature thermal plasma model ( see table [ tab : spec ] ) with @xmath37/d.o.f .
= 440/362 .
the spectrum of the arc to the southwest of the nucleus is not well described by a single temperature plasma model ( @xmath37/d.o.f .
= 683/396 )
. the arc can be described by a two temperature thermal plasma model , with temperatures of 2.09 kev and 1.08 kev . if the warmer component has the same parameters as the ambient gas , the cooler component is found to have enhanced mg and si abundances than the ambient gas , comparable o and fe , and comparable or lower s abundance .
the model parameters are given in table [ tab : spec ] with a @xmath37/d.o.f .
= 557/395 . to summarize , the physical properties of the gas in the x - ray arc are significantly different to those of the ambient cluster gas , in particular an additional cooler thermal component is present .
the metal abundances of some of the heavy elements in the cooler gas in the arc , both to the east and west , are consistent with being enhanced relative to those of the ambient gas .
our conclusions are in broad agreement with the _ xmm - newton _ observations of m87 ( belsole et al .
if we assume the `` holes '' in the x - ray gas that correspond to the inner radio lobes ( see fig . [ fig : cxo_radio_halpha ] ) are cavities created by the radio jet then we can make a crude estimate of the minimum power of the jet .
the cavities are approximately spherical with typical diameters @xmath24 1.5 kpc and volumes @xmath53 @xmath54 , and the pressure in the regions around the cavities is @xmath0 2 @xmath55 10@xmath56 dyn @xmath57 .
an estimate of the age of a cavity can be obtained from the free - fall timescale , @xmath58 where @xmath59 is the radius of the cavity , @xmath60 cm , and @xmath61 is the sound speed , @xmath62 cm s@xmath8 . assuming the gas temperature @xmath63 kev gives a jet power @xmath64 erg s@xmath8 .
if the cavities are evacuated faster than @xmath61 , then @xmath65 would be higher , although there appear to be no signs of sharp edges in the x - ray image that might be shocks .
our estimate of @xmath66 is at the lower end of previous estimates ( e.g. bicknell & begelman [ 1999 ] find @xmath67 erg s@xmath8 ) , although there are inevitably uncertainties in our estimates of @xmath68 and other quantities .
the cavity can not be inflated much more slowly than the above estimate or it would have risen too far from the nucleus through buoyancy effects . the bubbles are expected to rise at @xmath69 ( churazov et al .
2001 ) , and can not have traveled more than 2 kpc giving a maximum age of @xmath70 and @xmath71 erg s@xmath8 .
the structure in the radial distribution of x - ray surface brightness ( see figs [ fig : cxo ] [ fig : color_cxo_radio ] ) could be some effect inherent to a cooling flow .
however , we have noted ( section [ sec : morphology ] ) that the two most noticeable sharp declines in azimuthally averaged x - ray surface brightness are just outside the inner radio lobes and at comparable nuclear distance to the intermediate radio ridges .
this similarity between the nuclear distances of the radio lobes and the declines in surface brightness suggests instead that the `` edges '' may be `` fronts '' in the cluster atmosphere created by earlier ejecta from the active nucleus of m87 .
we consider a scenario based on the results of numerical hydrodynamical simulations by rhb of the effects of the jets on the intra - cluster medium .
rhb discuss their simulation results in the context of powerful fr - ii radio galaxies but the end results are expected to be similar in a weaker fr - i radio galaxy .
the end effect of the transient jet activity is that the cluster expands in response to heating of the cluster core .
an approximately spherical pulse propagates outwards at the sound speed ( see fig . 6 of rhb ) , ultimately leaving the core of the cluster at a fractionally lower density and a higher temperature .
the amplitude of the density wave in the rhb simulation is @xmath72 which is dependent upon the initial conditions but is large enough to produce the @xmath73 fluctuations seen in the azimuthally averaged x - ray surface brightness when compared to the best fitting `` @xmath30-model '' shown in fig .
[ fig : beta ] .
the fact that the two fronts are at slightly larger nuclear distances than the inner radio lobes and the intermediate ridges suggests that each of these radio features is in the `` passive '' or buoyant bubble phase that characterizes the late stage of evolution .
such would be consistent with the absence of strong shocks around the cavities inflated by the inner radio lobes .
the @xmath74 ( 12 kpc ) separation of successive fronts corresponds to a sound crossing time of @xmath75 yrs , and indicates there may be relatively brief periods of jet activity every @xmath76 yrs or so .
if this is the case , the time - averaged energy injected into the cluster by the jet may be significantly lower than the present instantaneous power of the jet .
further detailed numerical simulations , tailored to the environment of m87 , are required to determine whether such a model can indeed reproduce the `` fronts '' and what conditions are needed .
in particular , such simulations are required to find out what effect the putative pulse has on the cluster core .
the model of rhb also predicts subtle ( tens of per cent ) changes in the gas temperature associated with the changes in x - ray surface brightness , and while these changes are too small to be detected in our data set , one could look for them in a future observation with a significantly longer integration time .
our observations of the x - ray arcs to the east and southwest and associated with the radio ridges show that they are narrow , with a fwhm of only @xmath77 ( 1 kpc ) .
the length to width ratio of the southwestern arc is @xmath78 .
we have also found ( section [ sec : t_p_ct ] ) that the gas in the arcs is cooler than the surrounding cluster gas , in agreement with previous studies , and appears to contain two components with different temperatures .
the cooler component may have some metal abundances enhanced with respect to the warmer gas in the cluster ( table [ tab : spec ] , section [ sec : abund ] ) .
a number of models have been proposed to explain these arcs .
firstly , it should be noted that although the x - ray emissivity of the gas may be enhanced by compression , shocks or cosmic ray heating processes tend to raise the gas temperature rather than lower it , and can be ruled out ( bhringer et al .
another possibility is that , as the radio ejecta ( which are likely to be in the late `` buoyant plume '' phase [ rhb ] ; cf .
section [ sec : fronts ] ) pass through a multi - phase cooling flow , they disrupt the `` cold '' phase ( consisting of cold , denser clouds ) and mix it with the `` hot '' phase .
this causes a rapid cooling of the hot phase and an enhancement of the density ( bhringer et al .
however , whether the ambient intra - cluster medium is , in fact , multi - phase remains controversial ( we are able to adequately describe most of the intra - cluster gas as a single temperature plasma ) .
a further possibility is that cooler , low entropy gas near the nucleus is mixed into the radio plasma , generating buoyant bubbles of gas ( churazov et al .
2001 ; brggen et al .
2002 ; rhb ) .
these bubbles would rise in the cluster gravitational field and expand until they have a density equal to that of the surrounding medium .
the in - situ disruption of gas in a cold phase , followed by mixing , should leave cooler gas with metal abundances comparable to the surrounding cluster gas . on the other hand ,
if the cool gas was entrained at a small nuclear distance and carried outwards , the buoyant bubbles of cold gas should have metal abundances appropriate to the cluster core .
as discussed above ( section [ sec : abund ] ) , we find that the cooler gas ( see table [ tab : spec ] ) in the arc probably has a higher metalicity for at least some elements than the ambient gas .
these higher abundances favor the buoyant bubble scenario .
the morphology of the x - ray arcs ( i.e. narrow and well collimated ) is consistent with the churazov et al .
( 2001 ) and rhb models in which a narrow column of dense gas is created .
the numerical simulations of both churazov et al .
( 2001 ) and rhb assumed axi - symmetry which may artificially stabilize narrow structures elongated along the symmetry axis , although reasonably narrow , dense structures with an aspect ratio of @xmath79 are also produced in the fully 3d simulations of brggen et al .
the sound - crossing time from the nucleus to a nuclear distance of @xmath80 is a @xmath81 yrs , which is shorter than the cooling time of gas in the arcs , which is @xmath82 yrs and closer to @xmath83 yrs at a nuclear distance of @xmath80 ( see fig . [
fig : d_p_ct ] ) .
this means that there is time for cool gas entrained by a buoyant bubble in the nucleus to be carried to the end of the arc before it has cooled completely . if this were not the case , i.e. if the cooling time of gas in the arcs was shorter than the sound - crossing time , then some in - situ heating mechanism
would be required , or the cool gas must have been transported faster than the sound speed .
the simulation by rhb produces both elongated , buoyant plumes which lift gas from the center of the cluster and approximately spherical pulses traveling at the sound speed that we have identified with the x - ray `` fronts '' . in this sense
, the structure of the intra - cluster medium in the inner part of the virgo cluster conforms qualitatively to the expected effects of repeated injections of energy over a period of 10@xmath84 yrs by radio jets from the nucleus of virgo a.
we have presented the highest resolution x - ray image to date of the core of the virgo cluster .
structure is observed on all spatial scales , from arcsecond ( @xmath85 pc ) to ten arcminutes ( @xmath86 kpc ) .
the inner radio lobes are aligned with depressions in the x - ray surface brightness and there is no evidence of shock heating in the x - ray emission immediately surrounding the inner radio lobes , suggesting that the radio plasma has gently pushed aside the x - ray emitting gas .
these cavities can not have been inflated much slower than the sound speed , however , or they would have risen too far from the nucleus due to buoyancy effects .
we estimate the jet power to be @xmath87 erg s@xmath8 . on larger scales
the most striking feature is the x - ray arc running from the east , across the central regions of m87 , and off to the southwest .
the gas in the arc has at least two temperatures , with one component at the temperature of the ambient intra - cluster medium and a cooler component at @xmath3 kev .
the gas in the arcs is probably over - pressured with respect to , and somewhat more metal rich than , the ambient intra - cluster medium .
abrupt changes in surface brightness or `` fronts '' are seen at nuclear distances slightly larger than the nuclear distances of the inner radio lobes and intermediate radio ridges . within the inner front , at nuclear distances @xmath88 ( @xmath89 kpc ) the gas has at least two temperatures , with the cooler component at @xmath3 kev , similar to the x - ray arc .
this cooler region is concentrated more to the north than the south of the nucleus and is correlated with the h@xmath4 + [ nii ] emission - line distribution .
we suggest that a model based on the hydrodynamical simulations of rhb , scaled to a lower power radio source such as m87 , may explain the observed phenomena .
intermittent jet activity has two effects on the cluster .
firstly , it inflates buoyant bubbles of radio plasma that trail cold gas from the central regions in their wakes as they rise at @xmath90 0.7 times the sound speed , thereby producing the x - ray arcs .
the gas dredged up from the nucleus is expected to have higher metal abundances than the intra - cluster medium at large nuclear distances .
secondly , at late times in the evolution of the radio source , the injection of energy into the cluster core produces a `` pulse '' that expands at the sound speed into the intra - cluster medium as the cluster adjusts to find a new equilibrium .
we identify this pulse with the observed x - ray `` fronts '' .
detailed numerical simulations tailored to the virgo cluster are required to explore this hypothesis .
we thank w. b. sparks for providing the h@xmath4 + [ n ii ] map in electronic form and nrao for supplying the u - v data used to make the 90 cm map .
we also thank chris reynolds for discussions .
this research was supported by nasa grants nag 81027 and nag 81755 .
we are also grateful to d. harris and s. virani of the _ chandra _ science center for their assistance with the observations .
counts & 3323 & 5344 & 10815 & 7250 + @xmath91 ( @xmath92 @xmath57 ) & @xmath93 & @xmath94 & @xmath95 & @xmath96 + @xmath97 ( kev ) & @xmath98 & @xmath99 & @xmath100 & @xmath101 + o & @xmath102 & @xmath103 & @xmath104 & @xmath105 + mg & @xmath106 & @xmath107 & @xmath108 & @xmath109 + si & @xmath110 & @xmath111 & @xmath112 & @xmath113 + s & @xmath114 & @xmath115 & @xmath116 & @xmath117 + fe & @xmath118 & @xmath119 & @xmath120 & @xmath121 + k & @xmath122 & @xmath123 & @xmath124 & @xmath125 + @xmath37/d.o.f . & 134/110 & 188/136 & 440/362 & 557/395 + a color representation of the _ chandra _ image of m87 and the core of the virgo cluster in the 0.5 7 kev energy band .
the images have been background subtracted , exposure map corrected and adaptively smoothed . the upper left panel shows the large scale structure and the upper right and bottom panels show enlargements of the regions outlined by dashed white lines in the upper left and upper right panels , respectively .
the ` + ' sign indicates the nucleus of m87 .
a description of the morphology is given in section [ sec : morphology].,title="fig : " ] a color representation of the _ chandra _ image of m87 and the core of the virgo cluster in the 0.5 7 kev energy band .
the images have been background subtracted , exposure map corrected and adaptively smoothed .
the upper left panel shows the large scale structure and the upper right and bottom panels show enlargements of the regions outlined by dashed white lines in the upper left and upper right panels , respectively .
the ` + ' sign indicates the nucleus of m87 .
a description of the morphology is given in section [ sec : morphology].,title="fig : " ] a color representation of the _ chandra _ image of m87 and the core of the virgo cluster in the 0.5 7 kev energy band .
the images have been background subtracted , exposure map corrected and adaptively smoothed .
the upper left panel shows the large scale structure and the upper right and bottom panels show enlargements of the regions outlined by dashed white lines in the upper left and upper right panels , respectively .
the ` + ' sign indicates the nucleus of m87 .
a description of the morphology is given in section [ sec : morphology ] . ] a color representation of the inner parts of m87 in the 0.5 7 kev energy band with contours of 6 cm radio emission ( green ) and h@xmath4 + [ n ii ] ( blue ) overlayed .
the radio map has been smoothed by a gaussian of fwhm @xmath126 , and the contours are logarithmically spaced .
the h@xmath4 + [ n ii ] image has been smoothed by a gaussian of fwhm @xmath126 , and the contours are linearly spaced .
note the similar nuclear distances of the inner radio lobes and the approximately circular `` front '' in x - rays ( nuclear distance 45 ) . ]
a plot showing the ratio of the azimuthally averaged x - ray surface brightness to the best fitting `` @xmath30-model '' ( see section [ sec : t_p_ct ] ) as a function of nuclear distance .
there are systematic deviations of @xmath127 in surface brightness from the @xmath30-model fit ( shown as the inset ) .
arrows indicate the approximate locations of the sudden declines ( `` edges '' or `` fronts '' ) in surface brightness discussed in the text .
the x - ray emission from the nucleus and jet were excluded from this analysis . ] an x - ray
radio color overlay of the virgo cluster .
the adaptively smoothed _
0.5 7 kev x - ray image is red / yellow , while the 90 cm radio emission ( made from vla archival data ) is blue .
note the similar nuclear distances of the intermediate radio ridges and the approximately circular `` front '' in x - rays ( nuclear distance @xmath17 ) ] temperature map of the core of the virgo cluster .
this map was constructed from adaptively binned _ chandra _ images , using the technique described in section [ sec : t_p_ct ] .
overlayed are contours of 90 cm radio emission ( made from vla archival data ) .
the contours are at 0.01 , 0.05 , 0.1 , 0.15 and 0.2 jy beam@xmath8 . ] plots of temperature vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] , north [ top center ] , west [ bottom left ] and south [ bottom center ] ) and the arc ( towards the east [ top right ] and southwest [ bottom right ] ) . at a given nuclear distance and azimuthal angle the spectrum is described by a single temperature thermal plasma model unless a two temperature thermal plasma model results in an improvement in @xmath37/d.o.f . of @xmath38 , in which case
a two temperature thermal plasma model is used.,title="fig : " ] plots of temperature vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] , north [ top center ] , west [ bottom left ] and south [ bottom center ] ) and the arc ( towards the east [ top right ] and southwest [ bottom right ] ) . at a given nuclear distance and azimuthal angle
the spectrum is described by a single temperature thermal plasma model unless a two temperature thermal plasma model results in an improvement in @xmath37/d.o.f . of @xmath38 ,
in which case a two temperature thermal plasma model is used.,title="fig : " ] plots of temperature vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] , north [ top center ] , west [ bottom left ] and south [ bottom center ] ) and the arc ( towards the east [ top right ] and southwest [ bottom right ] ) . at a given nuclear distance and azimuthal angle the spectrum
is described by a single temperature thermal plasma model unless a two temperature thermal plasma model results in an improvement in @xmath37/d.o.f . of @xmath38 , in which case
a two temperature thermal plasma model is used.,title="fig : " ] plots of temperature vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] , north [ top center ] , west [ bottom left ] and south [ bottom center ] ) and the arc ( towards the east [ top right ] and southwest [ bottom right ] ) . at a given nuclear distance and azimuthal angle the spectrum is described by a single temperature thermal plasma model unless a two temperature thermal plasma model results in an improvement in @xmath37/d.o.f . of @xmath38 , in which case
a two temperature thermal plasma model is used.,title="fig : " ] plots of temperature vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] , north [ top center ] , west [ bottom left ] and south [ bottom center ] ) and the arc ( towards the east [ top right ] and southwest [ bottom right ] ) . at a given nuclear distance and azimuthal angle
the spectrum is described by a single temperature thermal plasma model unless a two temperature thermal plasma model results in an improvement in @xmath37/d.o.f . of @xmath38 , in which case
a two temperature thermal plasma model is used.,title="fig : " ] plots of temperature vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] , north [ top center ] , west [ bottom left ] and south [ bottom center ] ) and the arc ( towards the east [ top right ] and southwest [ bottom right ] ) . at a given nuclear distance and azimuthal angle
the spectrum is described by a single temperature thermal plasma model unless a two temperature thermal plasma model results in an improvement in @xmath37/d.o.f . of @xmath38 , in which case
a two temperature thermal plasma model is used.,title="fig : " ] plots of electron density ( @xmath128 ; dashed line ) , total gas pressure ( @xmath128 kev ; dotted line ) and cooling time ( gyr ; solid line ) vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] and south [ bottom right ] ) and the arc ( towards the east [ top right ] and southwest [ bottom left]).,title="fig : " ] plots of electron density ( @xmath128 ; dashed line ) , total gas pressure ( @xmath128 kev ; dotted line ) and cooling time ( gyr ; solid line ) vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] and south [ bottom right ] ) and the arc ( towards the east [ top right ] and southwest [ bottom left]).,title="fig : " ] plots of electron density ( @xmath128 ; dashed line ) , total gas pressure ( @xmath128 kev ; dotted line ) and cooling time ( gyr ; solid line ) vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] and south [ bottom right ] ) and the arc ( towards the east [ top right ] and southwest [ bottom left]).,title="fig : " ] plots of electron density ( @xmath128 ; dashed line ) , total gas pressure ( @xmath128 kev ; dotted line ) and cooling time ( gyr ; solid line ) vs. nuclear distance for different azimuthal angles corresponding to the ambient cluster medium ( towards the northeast [ top left ] and south [ bottom right ] ) and the arc ( towards the east [ top right ] and southwest [ bottom left]).,title="fig : " ] spectra of the ambient cluster gas to the northeast of the nucleus ( left panel ) and in the arc to the east of the nucleus ( right panel ) .
both spectra were extracted from @xmath35 square regions .
the spectrum of the gas in the arc may be approximated by the sum of the spectrum of the ambient gas and a cooler , more metal rich component , and this sum is plotted in the right hand panel .
see section [ sec : abund ] for more details.,title="fig : " ] spectra of the ambient cluster gas to the northeast of the nucleus ( left panel ) and in the arc to the east of the nucleus ( right panel ) .
both spectra were extracted from @xmath35 square regions .
the spectrum of the gas in the arc may be approximated by the sum of the spectrum of the ambient gas and a cooler , more metal rich component , and this sum is plotted in the right hand panel .
see section [ sec : abund ] for more details.,title="fig : " ] | we report sub - arcsecond x - ray imaging spectroscopy of m87 and the core of the virgo cluster with the chandra x - ray observatory .
the x - ray morphology shows structure on arcsecond ( @xmath0 100 pc ) to ten arcminute ( @xmath0 50 kpc ) scales , the most prominent feature being an `` arc '' running from the east , across the central region of m87 and off to the southwest .
a ridge in the radio map , ending in an `` ear''-shaped structure , follows the arc to the east .
depressions in the x - ray surface brightness correspond to the inner radio lobes and there is no evidence of shock - heated gas surrounding them .
there are also at least two approximately circular ( centered near the nucleus ) `` edges '' in the x - ray brightness distribution , the radii of which are slightly larger than the nuclear distances of the inner radio lobes and intermediate radio ridges , respectively .
we speculate that these discontinuities may be spherical pulses or `` fronts '' driven by the same jet activity as is responsible for the radio structure ; such pulses are found in recent numerical simulations .
all these results provide good evidence that the nuclear activity affects the intra - cluster medium .
we present a temperature map of the intra - cluster medium , and obtain the temperature , pressure and cooling time as a function of nuclear distance for the arcs and the ambient intra - cluster medium .
we show that the gas in the arcs is cooler than , and probably over - pressured with respect to , the ambient intra - cluster medium .
the metal abundances of the cooler gas in the arc are somewhat enhanced relative to the ambient intra - cluster medium , favoring a `` buoyant plume '' origin for the x - ray arc , in which ambient gas near the nucleus is entrained by buoyant radio plasma and carried to larger nuclear distances .
the gas within the inner `` front '' ( nuclear distance @xmath1 kpc @xmath2 ) is at least a two - temperature plasma , with the cool component at @xmath3 kev .
this cool region is concentrated to the north of the nucleus and is strongly correlated with the h@xmath4 + [ nii ] emission - line distribution . # 1to 0pt#1 # 1 # 1 to size # 1 to size # 1 = cmr10 scaled2 = cmbx10 scaled2 = cmti10 scaled2 ps . |
hess measurements of tev fluxes from sgr a@xmath1 @xcite confirm , with high significance , earlier reports by the whipple @xcite and cangaroo @xcite collaborations ( albeit at higher flux levels than hess ) about tev emission from the galactic center .
this result adds important new information about the massive ( @xmath8 ; @xcite ) bh at the center of the galaxy . when combined with the wealth of observational data in the radio to x - ray wavebands @xcite , the tev @xmath0-ray data point to a scenario involving an inner , inefficiently radiating magnetized corona , i.e. the adaf , and a subrelativistic outflow of particles and field from the adaf . as in pulsar plerions like crab nebula ( see @xcite ) , the wind terminates at a shock , at @xmath9 from the gcbh , to create a black - hole powered ` plerion ' where the tev radiation is produced .
the vlbi size of the radio source at the location of sgr a@xmath10is @xmath11as at @xmath12 and @xmath13as at @xmath14 @xcite . using vlba ,
@xcite resolve sgr a@xmath10at 42 ghz to have an angular size of @xmath15 , or @xmath16 for @xmath17 black hole and a gc distance @xmath18 , and find that the size increases with wavelength .
the radio emission of sgr a@xmath10is rather stable , with total luminosity @xmath19 and spectral flux @xmath20 that sharply falls off above @xmath21 @xcite .
the near - infrared ( nir ) flares show factors - of-2 flux variations on timescales of @xmath22min and reach powers of @xmath23 @xcite . the spectra rise in @xmath24 diagrams , and show powers @xmath23 , similar with the ( _ nonsimultaneously _ ) detected x - ray flares @xcite .
the 200s x - ray variability timescale limits the engine size to @xmath25 . the unabsorbed quiescent flux detected by chandra observatory from sgr a@xmath10is @xmath26@xmath27 , and has angular extent @xmath28 @xcite , corresponding to source size @xmath29 .
the tev spectral fluxes in two hess campaigns in june / july and july / august of 2003 are almost exactly the same , extending with photon index @xmath30 - 2.3 from @xmath31gev to 10 tev , with total power @xmath32 .
there is no sign of flux variation within the statistics available in each of these sets @xcite .
given the high statistical significance of the signal in both data sets , these results suggest a quiescent or quasi - stationary tev source coincident within @xmath33 of sgr a@xmath1 .
we propose a model for sgr a@xmath1 where the quasi - stationary radio and flaring x - ray / nir emissions are synchrotron radiation from the adaf .
the mhd wind from the adaf powers a black - hole plerion where the quiescent x - ray and tev emissions are produced by electrons accelerated at the wind shock .
the model explains the existing multi - wavelength data , and makes predictions that can be checked by various x - ray and gamma - ray detectors .
the dimensionless mass - accretion rate ( or eddington ratio ) @xmath34 is defined in units of the eddington luminosity @xmath35 , where @xmath36 is the maximum efficiency of gravitational - to - photon energy conversion by the bh . in the adaf model @xcite , thermal radiation of the accretion plasma at the level @xmath37
is predicted when @xmath38 .
the rate @xmath39 is obtained if one formally equates @xmath40 with the observed @xmath41 . the compact radio emission can not originate from an optically - thick accretion disk which would only be allowed at @xmath42 , or from a much hotter adaf plasma at smaller scales , but is compatible with a synchrotron origin @xcite .
the thermal output of the adaf would be peaked mostly in the hard x - ray / soft gamma - ray domain , where the quiescent luminosity is at least 2 orders of magnitude smaller than @xmath43 .
this results in the accretion rate estimate @xmath44 consistent with adaf model and sgr a@xmath10observations . even for @xmath45 this small
, @xmath46 of accretion power is still either advected into the bh or escapes as an outflow .
adopting gcbh mass @xmath47 , the adaf magnetic field at radii @xmath48 is @xcite @xmath49 where @xmath50 is the ratio of the magnetic to the total pressure in the accreting gas and @xmath51 is the viscosity parameter . using @xmath52 and @xmath53 @xcite , the magnetic field @xmath54 g in the adaf at @xmath55 is found . for these fields ,
the lorentz factor of electrons producing synchrotron radiation peaked at @xmath56 is @xmath57 , and the adaf is still transparent to self - absorption for @xmath58 radiation . a stochastic acceleration model for the radio - mm emission from sgr a@xmath10@xcite
can be quantified using the second - order fermi acceleration theory of @xcite .
the magnetic field is defined through the relation @xmath59 , implying @xmath60 @xmath61 g for a region of size @xmath62 .
equating the acceleration rate of electrons by whistler turbulence with the synchrotron loss rate gives a characteristic electron lorentz factor @xmath63 , assuming a kolmogorov spectrum for the turbulence .
here @xmath64 is the fractional whistler turbulent energy density compared to @xmath65 , and @xmath66 is the thomson depth of the adaf , assumed to be composed of e - p plasma .
values of @xmath67 g and @xmath68 are implied for the adaf when @xmath69 , @xmath70 and @xmath71 .
a compton ( i.e. , leptonic ) origin for non - variable tev flux detected with hess implies that the synchrotron counterpart of that flux is the quiescent component of the x - ray flux .
the magnetic - field energy density in the source should then be @xmath72 , where @xmath73 is the energy density of the target photons at @xmath74 . inside the radio source , at @xmath75 , the radiation energy density saturates at the level @xmath76 .
thus , the origin of tev flux at those small distances would require magnetic fields @xmath77 , far below the adaf model values . in the adios ( `` advection - dominated inflow - outflow solution '' @xcite ) extension of the adaf model , the outflow may carry a significant fraction of the generated kinetic energy of the plasma from the bh vicinity to large distances .
we assume a wind of magnetized accretion plasmas with total power @xmath78 propagating in a two - sided cone with opening angle @xmath79 at speed @xmath80 ; for calculations we use @xmath81 .
this wind terminates at a subrelativistic shock at @xmath82 from the gcbh .
the shock accelerates electrons to energies @xmath83 to produce the observed tev @xmath0-rays by compton upscattering of sub - mm ( @xmath21 ) photons from the adaf .
another important photon target is the emission from the cold ( @xmath84 ) dust ring of sgr a west , with total luminosity @xmath85 in the central arcmin ( few _ pc _ ) region around sgr a@xmath10@xcite .
this radiation has a spectral peak at @xmath86 and estimated energy density @xmath87 .
it becomes the main contributor to the @xmath88 compton flux from the plerion at @xmath89 from sgr a@xmath1 .
synchrotron radiation from these same multi - tev electrons at @xmath90 produces the quiescent x - ray flux from sgr a@xmath1 .
chandra observations @xcite reveal diffuse thermal x - rays with luminosity @xmath91 from the central parsec region of sgr a@xmath1 , produced by @xmath92 plasma with @xmath93 density , where @xmath94 is the volume filling factor . equating the gas pressure @xmath95 in that region with the energy density of the wind gives the shock distance @xmath96 the tev luminosity @xmath97 implies a total electron acceleration power @xmath98 .
assuming that the efficiency of electron acceleration @xmath99 , and letting the rest go to thermal plasma and nonthermal protons ( unlike in pulsar winds where relativistic electrons dominate ) , we arrive at @xmath100 .
note that eq.(3 ) predicts that the maximum power of the outflow sustainable at distances @xmath82 is @xmath101 .
the mean fluid speed @xmath102 in the tev plerion formed downstream of the shock can be derived from @xmath103 , where @xmath104 is the compton cooling time of electrons producing synchrotron x - rays with energy @xmath105 for the estimated magnetic field @xmath106 in the gcbh plerion at scales @xmath107 .
taking @xmath108 for the median distance of the tev plerion , we find @xmath109 at those distances . since relativistic electrons of all energies convect with the fluid at the same speed , the tev plerion model predicts shrinking of the quiescent x - ray source size at higher energies .
note also that the opening angle of the outflow should be @xmath110 rather than @xmath111 , since otherwise @xmath112 would exceed @xmath113 .
this indicates that the energy outflow is in the form of a wind rather than well - collimated jet .
fig . 1 shows numerical calculations of the quiescent radiation components of sgr a@xmath114 .
these include synchrotron and ssc emission from the magnetized adaf within @xmath115 of the gcbh ( _ dashed curves _ ) , x - ray and tev emission from the compact plerion at scales @xmath116 @xmath117 ( _ solid curves _ ) , and the emission from the larger plerion ( _ dot - dashed curves _ ) inflated to _ pc _ scales in the process of convective ( and possibly also diffusive ) propagation of the accelerated electrons on timescales @xmath118
. magnetic field in the latter , @xmath119 , is assumed higher than @xmath120 in the tev plerion .
this is in agreement with the expected increase of the magnetic field downstream of the shock in typical pulsar plerions @xcite , and is explained by deceleration and compression of the plasma .
but it is also possible that @xmath121 , in which case a smaller synchrotron flux from the @xmath122-scale plerion than shown in fig . 1 is predicted .
the inset in fig .
1 shows the nonthermal electron distributions . besides synchrotron and compton losses
, calculations also include coulomb and bremsstrahlung losses in a medium with @xmath123 characteristic for central @xmath122 regions of sgr a@xmath114 .
electrons with a power - law injection index @xmath124 and an exponential cutoff above @xmath125 are injected into 2-sided nebula with total power @xmath126 .
these electrons radiatively cool downstream of the shock and propagate to distances @xmath127 on the timescale @xmath128yr , during which most of the injected energy of the multi - tev electrons ( producing tev and x - ray fluxes observed ) has been already lost .
the radio emission is produced at @xmath129 in @xmath130 field by a relativistic maxwellian electron distribution @xmath131 with @xmath132 resulting from a balance between second - order energy gains and synchrotron losses @xcite .
the stochastic acceleration power of radio electrons is therefore about equal to the observed radio luminosity @xmath43 .
the ssc emission component of these electrons , shown in fig .
1 by the dashed curve , falls in the sub - kev region below the level of the quiescent x - ray flux .
however , if magnetic fields @xmath133 g are assumed , this ssc contribution would rise and also move to 1 kev , and could contribute to radiation at that energies .
it can not , however , explain the entire quiescent flux extending to 10 kev . the powerful x - ray and nir flares observed from sgr a@xmath10are explained in our model by the onset of instabilities of the accretion flow at distances of only few @xmath134 .
they result in strong shocks in the accretion flow and effective acceleration of particles advected with the flow .
the accelerated particles are then easily taken out of this region in the wind / jet outflow .
synchrotron emission of electrons accelerated to @xmath135 explains the x - ray flares with very short variability scales .
furthermore , while propagating through the first few tens of @xmath134 , these electrons have sufficient time , @xmath136 , to cool in the high @xmath137 fields there down to @xmath138 to produce flares in the nir domain .
self - absorbed flares at @xmath139 detected on @xmath140day timescales after the x - ray flares @xcite could be explained by the radiation from these same electrons at later stages / larger distances of the outburst in the ` expanding source ' scenario . in fig .
2 we show the flare fluxes expected in this model . after shock - acceleration close to the bh and injection into collimated wind outflow with speed @xmath141 , electrons propagate through the inner @xmath14210 - 30 region on timescale @xmath143 , after which the radiation losses drop because of wind expansion and decline of @xmath137 .
the solid , dashed , and dot - dashed curves show synchrotron and compton fluxes produced at times 200s , 1200s and 1hr after the onset of the flare .
the electron injection time profile is @xmath144 , with @xmath145 .
the mean magnetic field in adaf outflow could be enhanced at the flaring state , so we take @xmath146 .
the compton fluxes shown ( thin curves ) clearly demonstrate that no detectable tev flares should be expected during powerful x - ray flares .
this is in agreement with the non - detection of tev flux variations during many days of observation with the hess telescopes , whereas the x - ray flares occur with frequency of @xmath147 per day @xcite .
a model consisting of magnetized coronal adaf ( synchrotron ) radio emission within @xmath148 of the gcbh , and of a black - hole plerion powered by a wind from the adaf resolves many puzzling observations of sgr a@xmath1 .
x - ray flares are synchrotron radiation of electrons accelerated through first - order fermi process by shocks within a few @xmath149of the gcbh .
electrons accelerated to @xmath150 at the wind termination shock at @xmath151 cm creates the gcbh plerion .
its synchrotron x - ray emission has been resolved as an @xmath152 source with chandra @xcite .
the multi - tev electrons compton scatter the radio photons from sgr a@xmath10and fir photons from the dust ring of sgr a west to produce the tev emission .
tev emission from the gcbh plerion is nearly stationary because the cooling time of tev electrons is @xmath153 yrs .
a jet - adaf " model for sgr a@xmath10has been proposed by @xcite and @xcite .
our model , though also based on energy outflow from the adaf , differs greatly .
in particular , the quiescent and flaring x - ray components are produced at different sites .
the origin of the x - ray and nir flares are explained as synchrotron emission of electrons accelerated during episodes of instabilities very close to the bh , not as compton radiation in the ssc scenario of the jet - adaf " model .
importantly , the tev flux can not be easily explained if the observed x - rays were due to the ssc mechanism in the vicinity of the gcbh , as suggested by @xcite .
propagation of gev electrons from the plerion on timescales of @xmath154 with speeds @xmath155could significantly contribute to the radio synchrotron flux of sgr a west .
in particular , it could explain the suggested nonthermal origin of radiation in the bar of sgr a west @xcite , formed in the outflow direction .
we also predict that quasi - stationary compton and bremsstrahlung fluxes from the pc - scale plerion , coincident with the central parts of sgr a west , will be significantly detected and possibly resolved with glast at gev energies .
but the expected highly variable compton counterpart of the synchrotron x - ray flares from the gcbh vicinity is too weak to be detectable with glast or hess .
the synchrotron extension of the flare emission in the hard x - ray / soft @xmath0-ray domain explains the @xmath156min episode of a profound increase of flux detected by integral from the direction of sgr a@xmath10@xcite .
the absence of apparent tev flux variations may suggest a proton origin for the tev radiation @xcite .
indeed , proton and ion acceleration through first- and second - processes in the adaf and through first - order acceleration at the wind termination shock could make cosmic rays to produce tev emission through nuclear @xmath157-interactions with @xmath158 dense gas on @xmath122 scales , and could form extended tev emission possibly already detected with hess @xcite .
the hadronic origin of the tev radiation in the adaf itself in the bh vicinity requires , however , an extremely dense gas target or extremely large proton powers @xmath159 .
finally , we note that the unidentified egret source 3eg j1746 - 2851 towards the gc @xcite is significantly displaced @xcite from the direction of the gcbh , and is unlikely to be related to sgr a@xmath1 .
it is probably emission from a young pulsar , though not with the mouse " psr j1747 - 2958 @xcite .
a young pulsar with @xmath0-ray properties like vela but with apparent @xmath0-ray power @xmath160 larger could have been missed in pulsar surveys due to the large dispersion measure towards the gc .
we suggest a deeper , higher radio - frequency and x - ray search at the refined location of 3eg j1746 - 2851 .
mayer - hasselwander , h. a. , et al .
1998 , , 335 , 161 mclaughlin , m. a. , & cordes , j. m. 2003 , in proc .
4th agile science workshop ( astro - ph/0310748 ) melia , f. & falcke , h. 2001 , , 39 , 309 narayan , r. , and yi , i. 1995 , apj 452 , 710 | the hess collaboration recently reported highly significant detection of tev @xmath0-rays coincident with sgr a@xmath1 @xcite . in the context of other galactic center ( gc ) observations ,
this points to the following scenario : in the extreme advection - dominated accretion flow ( adaf ) regime of the gc black hole ( bh ) , synchrotron radio / sub - mm emission of @xmath2 electrons emanates from an inefficiently radiating turbulent magnetized corona within 20 @xmath3 ( schwarzschild radii ) of the gcbh .
these electrons are accelerated through second - order fermi processes by mhd turbulence , as suggested by liu et al .
( 2004 ) .
closer to the innermost stable orbit of the adaf , instabilities and shocks within the flow inject power - law electrons through first - order fermi acceleration to make synchrotron x - ray flares observed with chandra , xmm , and integral . a subrelativistic mhd wind subtending a @xmath4 cone with power @xmath5
is driven by the adaf from the vicinity of the gcbh .
as in pulsar powered plerions , electrons are accelerated at the wind termination shock , at @xmath6 cm from the gcbh , and compton - scatter the adaf and the far infra - red ( fir ) dust radiation to tev energies . the synchrotron radiation of these electrons forms the quiescent x - ray source resolved by chandra .
the radio counterpart of this tev / x - ray plerion , formed when the injected electrons cool on timescales of @xmath7 , could explain the origin of nonthermal radio emission in the _
pc_-scale bar of the radio nebula sgr a west . |
the suppression of oscillation and the revival of oscillation are two opposite but interrelated important emergent phenomena in coupled oscillators . in the suppression of oscillation , oscillators arrive at a common homogeneous steady state ( hss ) or different branches of an inhomogeneous steady state ( ihss ) under some proper parametric and coupling conditions . the former is called the amplitude death ( ad ) state @xcite , whereas the latter is denoted as the oscillation death ( od ) state @xcite . on the other hand ,
the revival of oscillation is a process by which the rhythmic behavior ( or rhythmicity ) of individual nodes in a network of coupled oscillators is restored from an ad / od state without changing the intrinsic parameters associated with the individual nodes @xcite . in the oscillation quenched ( suppressed ) state the dynamic nature of individual coupled oscillators
this has many potential applications : for example , the ad state is important to suppress unwanted oscillations that hinder a certain process , e.g. , in laser systems , chattering in mechanical drilling process , etc @xcite .
similarly , the od state has a significant role in understanding of many biological processes , e.g. , synthetic genetic oscillator @xcite , cardiovascular phenomena @xcite , cellular differentiation @xcite , etc . on the other hand ,
the research in the topic of revival of oscillation is important because many physical , environmental , biological and social processes require _ stable _ and _ robust _ oscillations for their proper functioning : examples include , el nio / southern oscillation in ocean and atmosphere @xcite , brain waves in neuroscience @xcite , electric power generators @xcite , cardiopulmonary sinus rhythm of pacemaker cells @xcite , etc . in these systems
the suppression of oscillation may result in a fatal system breakdown or an irrecoverable physiological malfunction .
thus , it has to be ensured that , if these type of systems are trapped in an oscillation cessation state , that state has to be revoked in order to establish rhythmicity .
a recent burst of publications reveal many aspects of ad and od : in particular , identifying coupling schemes to induce them @xcite , transition from ad to od @xcite , their experimental verifications @xcite , etc .
but only a few techniques have been reported to revoke ad / od and induce rhythmicity in a network of oscillators @xcite . in ref.@xcite several variants of time delay techniques
are discussed in order to revoke death states , whereas in ref.@xcite , network connections are chosen properly in order to revive oscillations .
however , most of these techniques lack the generality to revive oscillations from a death state .
only recently a general technique to revive oscillation from the oscillation suppressed state ( or death state ) has been reported by @xcite .
the authors proposed a simple but effective way to revoke the death state and induce rhythmicity by introducing a simple feedback factor in the diffusive coupling .
they showed that this technique is robust enough to induce rhythmicity in a diffusively coupled network of nonlinear oscillators , such as , the stuart - landau oscillator , brusselators model , chaotic lorenz system , cell membrane model , etc .
they further tested the effectiveness of their proposed technique in conjugate @xcite and dynamic coupling @xcite .
however , in the absence of parameter mismatch or coupling time - delay , simple diffusive coupling can not induce ad in coupled oscillators .
therefore , for identical oscillators under diffusive coupling ( without coupling delay ) no ad is possible and thus the issue of revoking the ad state does not arise .
also , od in diffusively coupled identical oscillators is always accompanied by a limit cycle , thus one needs to choose proper initial conditions to revoke that death state .
regarding the conjugate coupling , it is not always a general technique to induce death : for example , in a first - order intrinsic time - delay system no conjugate coupling is possible .
further , the dynamic coupling has its own pitfalls ( e.g. , its success strongly depends on the intrinsic properties of the oscillators under consideration ) , which has been discussed in detail in @xcite . in this context the mean - field diffusive coupling
is a general way to induce ad / od even in networks of _ identical _ coupled oscillators : it works in any network of oscillators including chaotic first - order intrinsic time - delay systems @xcite .
further , it has been shown in @xcite that , unlike diffusive coupling , the od state induced by the mean - field diffusive coupling is not accompanied by a limit cycle . also , the mean - field diffusive coupling is the most natural coupling scheme that occurs in physics @xcite , biology @xcite , ecology @xcite , etc .
thus , for those systems that always need a robust limit cycle for their proper functioning , the mean - field diffusive coupling is a much stronger `` trap '' to induce death in comparison with the other coupling schemes .
therefore , it is important to study the process of revoking the oscillation suppression state induced by the mean - field diffusive coupling and revive oscillation from the death state .
motivated by the above facts , in this paper we introduce a feedback factor in the _ mean - field diffusive coupling _ and examine its effect in a network of coupled oscillators .
we show that the interplay of the _ feedback factor _ and the _ density of mean - field _ coupling can restore rhythmicity from a death state even in a network of identical coupled oscillators .
thus , unlike ref .
@xcite , here we have two control parameters that enable us to revoke the death state . using rigorous eigenvalue and bifurcation analyses on coupled van der pol and stuart - landau oscillators , separately , we show that the region of the death state shrinks substantially in the parameter space depending upon those two control parameters .
we also extend our results to a network consisting of a large number of mean - field coupled oscillators and show that the revival of rhythmicity works in the spatially extended systems , also .
further , for the first time , we report an experimental observation of the revival of oscillation from death states induced by the mean - field coupling that qualitatively supports our theoretical results .
at first we consider a network of @xmath0 van der pol ( vdp ) oscillators interacting through a modified mean - field diffusive coupling ; the mathematical model of the coupled system is given by [ systemvdp ] @xmath1 here @xmath2 and @xmath3 is the mean - field of the coupled system .
the individual vdp oscillators show a near sinusoidal oscillation for smaller @xmath4 , and relaxation oscillation for larger @xmath4 .
the coupling strength is given by @xmath5 ; @xmath6 is called the mean - field density parameter that determines the _ density of the mean - field _ @xcite ( @xmath7 ) ; it actually provides an additional free parameter that controls the mean - field dynamics : @xmath8 indicates the self - feedback case , whereas @xmath9 represents the maximum mean - field density .
the feedback term @xmath10 controls the rate of diffusion ( @xmath11 ) : @xmath12 represents the maximum feedback and the original mean - field diffusive coupling ; @xmath13 represents the absence of a feedback and thus that of diffusion . any values in between this limit can be treated as the intermediate diffusion rate and thus represent a modified mean - field diffusive coupling . the origin of @xmath10 is well discussed in @xcite where it is speculated that it may arise in the context of cell cycle , neural network and synchronization engineering .
as the limiting case we take two identical vdp oscillators : @xmath14 . from eq .
we can see that there are the following fixed points : the origin @xmath15 and two more coupling dependent fixed points : ( i ) ( @xmath16 , @xmath17 , @xmath18 , @xmath19 ) where @xmath20 and @xmath21 .
( ii ) ( @xmath22 , @xmath23 , @xmath22 , @xmath23 ) where @xmath24 and @xmath25 .
the eigenvalues of the system at the origin are , [ lambdavdp ] @xmath26 from the eigenvalue analysis we derive two pitchfork bifurcation ( pb ) points pb1 and pb2 , which emerge at the following coupling strengths : [ pb ] @xmath27 the ihss , ( @xmath16 , @xmath17 , @xmath18 , @xmath19 ) , emerges at @xmath28 through a symmetry breaking pitchfork bifurcation . the other nontrivial fixed point
( @xmath22 , @xmath23 , @xmath22 , @xmath23 ) comes into existence at @xmath29 , which gives rise to an unique _ nontrivial hss_. further , equating the real part of @xmath30 and @xmath31 to zero , we get two hopf bifurcation points at [ hb12 ] @xmath32 from eqs . and
we see that no hopf bifurcation of trivial fixed point occurs for @xmath33 ; in that case , only pitchfork bifurcations exist .
the eigenvalues of the system at the nontrivial fixed point ( @xmath34 , @xmath35 , @xmath36 , @xmath37 ) , where @xmath38 and @xmath39 are given by : [ ntlambda ] @xmath40 where , @xmath41 , @xmath42 , @xmath43 , @xmath44 .
now , with increasing @xmath6 , @xmath45 moves towards @xmath28 , and at a critical @xmath6 value , say @xmath46 , hb2 collides with pb1 : @xmath47 . for @xmath48 ,
the ihss becomes stable at @xmath49 through a subcritical hopf bifurcation , where @xmath50 this is derived from the eigenvalues of the system at the the nontrivial fixed point ( @xmath16 , @xmath17 , @xmath18 , @xmath19 ) .
@xmath49 actually determines the direct transition from od to a limit cycle , i.e. revival of oscillation .
the second nontrivial fixed point @xmath51 , @xmath23 , @xmath22 , @xmath52 that was created at @xmath53 becomes stable through a subcritical pitchfork bifurcation at @xmath54 : @xmath55 this is derived from the eigenvalues corresponding to @xmath51 , @xmath23 , @xmath22 , @xmath52 .
this nontrivial ad state can also be pushed back to a very large value of @xmath5 by choosing @xmath56 , and thus this ad state can also be revoked effectively .
the above eigenvalue analysis is supported by a numerical bifurcation analysis using xppaut @xcite .
figure [ f:1](a - c ) show the single parameter bifurcation diagram depending on @xmath5 for different @xmath10 for an exemplary value @xmath57 ( throughout the numerical simulation we consider @xmath58 ) .
we observe that the oscillation cessation state ( both ad and od ) moves towards right , i.e. , a stronger coupling strength @xmath5 for a decreasing @xmath10 , and for @xmath59 ( @xmath60 ) [ fig .
[ f:1 ] ( c ) ] the death state moves much further from pb1 in the right direction .
we also verify that at @xmath61 no death state occurs ( not shown in the figure ) .
we further explore the zone of oscillation cessation in a two parameter bifurcation diagram in the @xmath62 space . figure .
[ f:1 ] ( d ) shows the bifurcation curves for @xmath63 .
the hb2 curve determines the transition from oscillation to ad for @xmath64 ; beyond this limit the hbs curve determines the zone of oscillation and thus that of the death region . figure .
[ f:1 ] ( e ) shows that the _ death region shrinks with decreasing @xmath10 _ confirming our theoretical analysis .
finally , we plot the phase diagram in @xmath65 parameter space at @xmath66 [ fig .
[ f:1 ] ( f ) ] : we find that a _ higher _ value of @xmath6 _ or _ a _ lower _ value of @xmath10 support oscillations .
interestingly , even for @xmath63 ( i.e. , a complete mean - field diffusion ) , one can revive rhythmicity by simply increasing the value of @xmath6 ; thus , this coupling scheme offers two control parameters to revive oscillations . finally , we summarize the observations and discuss the following important points : ( i ) hb2 is the inverse hopf bifurcation point where an ad state is revoked and gives rise to a _
stable _ limit cycle .
this point ( or curve in a two parameter space ) determines the revival of oscillation below a critical value @xmath46 , which is determined by @xmath10 . from eq .
we see that by choosing @xmath10 closer to @xmath6 ( @xmath67 ) the death zone shrinks substantially . thus , to revoke a death state one has to choose @xmath56 , and
@xmath61 _ ensures _ that there will be no death state even in the stronger coupling strength ( whatever strong it may be ) .
( ii ) even if complete diffusion is present , i.e. , @xmath63 , one can still achieve the revival of oscillation by choosing @xmath68 .
this is an unique feature of the mean - field diffusive coupling , which is absent in other coupling schemes .
( color online ) network of mean - field coupled van der pol oscillators [ eqs . ] with @xmath57 , @xmath58 , @xmath69 : spatiotemporal plots showing ( a ) and ( b ) ad and its revival , respectively , with decreasing @xmath10 at @xmath66 .
( c ) and d ) show the od and its revival , respectively , with a decreasing @xmath10 at @xmath70 .
the upper rows [ i.e. , ( a ) and ( c ) ] have @xmath63 and the lower rows [ i.e. , ( b ) and ( d ) ] have @xmath71 . the first @xmath72 time is excluded and then @xmath73s for the next @xmath74 are plotted . ] to show that our analysis of two coupled oscillators are valid for a larger network also , we consider the more general case of @xmath69 mean - field coupled identical van der pol oscillators of eqs .
( @xmath58 ) .
[ f:2](a ) shows the spatiotemporal plot of the network in the global ad regime at @xmath63 , @xmath57 and @xmath66 ; here all the nodes arrive at the zero fixed point .
the global ad state is revoked and rhythmicity is restored in the network by decreasing the value of @xmath10 ; as shown in fig .
[ f:2](b ) for an exemplary value @xmath71 .
equivalently , for @xmath70 and @xmath63 ( @xmath57 as before ) we get an od state in the network [ fig .
[ f:2](c ) ] .
it can be seen that the nodes populate the upper and lower branches [ shown with yellow ( light gray ) and brown ( dark gray ) colors , respectively ] of od in a random manner and generate a multi - cluster od state .
oscillation in this network is revived by decreasing @xmath10 ; fig .
[ f:2](d ) shows this for @xmath71 .
note that the values for which ad , od and oscillations are obtained agree with that for the @xmath75 case shown in fig .
[ f:1 ] .
next , we consider @xmath0 stuart - landau oscillators interacting through a modified mean - field diffusive coupling in their real part ; the mathematical model of the coupled system is given by @xmath76 with @xmath2 ; @xmath77 is the mean - field of the coupled system , @xmath78 .
the individual stuart - landau oscillators are of unit amplitude and having eigenfrequency @xmath79 . as the limiting case we take @xmath75 , and rewrite eq . in the cartesian
coordinates : [ system ] @xmath80,\\ \label{y1 } \dot{y}_{i } & = \omega_{i}x_{i}+p_{i}y_{i}.\end{aligned}\ ] ] here @xmath81 @xmath82 , @xmath83 .
we set the oscillators to be identical , i.e. , @xmath84 . from eq .
it is clear that the system has the following fixed points : the trivial fixed point is the origin @xmath15 , and additionally two more coupling dependent nontrivial fixed points : ( i ) ( @xmath16 , @xmath17 , @xmath18 , @xmath19 ) where @xmath85 and @xmath86 .
( ii ) ( @xmath22 , @xmath23 , @xmath22 , @xmath23 ) where @xmath87 and @xmath88 .
( color online ) stuart - landau oscillators , @xmath89 : bifurcation diagram with @xmath5 ( a ) @xmath63 ( @xmath90 ) , ( b ) @xmath91 ( @xmath90 ) .
( c ) two parameter bifurcation diagram in @xmath62 space for different @xmath10 .
( d ) spatiotemporal plots of @xmath69 stuart - landau oscillators showing od at @xmath63 ( upper panel ) and the revival of oscillation at @xmath92 ( lower panel ) : @xmath90 and @xmath93 .
the time scale is same as in fig .
[ f:2 ] . ]
the four eigenvalues of the system at the trivial fixed point @xmath94 are , [ lambda ] @xmath95,\\ \label{lambda3 } { \lambda}_{3,4 } & = 1- \left[\frac{\epsilon \alpha \pm \sqrt{(\epsilon \alpha)^2 - 4{\omega}^2}}{2}\right].\end{aligned}\ ] ] through an eigenvalue analysis and also a close inspection of the nontrivial fixed points reveal that two pitchfork bifurcations ( pb ) occur at : [ pb ] @xmath96 a symmetry breaking pitchfork bifurcation gives birth to the ihss ( @xmath16 , @xmath17 , @xmath18 , @xmath19 ) at @xmath28 .
the second nontrivial fixed point ( @xmath22 , @xmath23 , @xmath22 , @xmath23 ) emerges at pb2 the stabilization of which leads to a _ nontrivial _ ad state that coexists with od .
next , we get the hopf bifurcation point by equating the real part of @xmath31 and @xmath30 to zero , [ epsahb ] @xmath97 from eqs . and
it is clear that for @xmath98 no hopf bifurcations ( of trivial fixed point ) occur , only pitchfork bifurcations govern the dynamics in that case . from eq .
( [ epsahb ] ) it is noticed that for a fixed @xmath10 , @xmath99 is constant , but @xmath45 depends only upon @xmath6 ( but is independent of @xmath100 , where @xmath101 ) . now with increasing @xmath6 , hb2 moves towards pb1 , and at a critical @xmath6 , say @xmath46 , hb2 collides with pb1 : @xmath102 .
the eigenvalues of the system at the nontrivial fixed point ( @xmath34 , @xmath35 , @xmath36 , @xmath37 ) , where @xmath38 and @xmath39 are given by : [ ntlambda ] @xmath103 where , @xmath104 , @xmath105 , @xmath106 , @xmath107 .
we derive the loci of the hbs curve as @xmath108 bifurcation diagrams of fig .
[ f:3](a ) ( @xmath63 ) and [ f:3](b ) ( @xmath109 ) show that with decreasing @xmath10 the death region moves towards a larger coupling strength @xmath5 .
figure [ f:3](c ) shows this in the @xmath62 space ; one can see that a decreasing @xmath10 shrinks the region of death and thus broadens the area of the oscillation state in the parameter space .
finally , we consider a network of @xmath69 identical stuart - landau oscillators with the same coupling scheme [ eq . ] .
we find that the revival of oscillation works here as well .
figure [ f:3](d ) demonstrates this for an exemplary value of @xmath93 and @xmath90 : the upper panel shows the od state in the network for @xmath63 , whereas from the lower panel we see that rhytmicity is restored in the network by reducing the value of @xmath10 to @xmath92 .
( color online ) experimental circuit diagram of the modified mean - field coupled vdp oscillators . a , a1-a4 , and aq
are realized with tl074 op - amps .
all the unlabeled resistors have value @xmath110 k@xmath111 .
c=10 nf , @xmath112 , @xmath113 v. box denoted by
b " are op - amp based buffers ; inverters are realized with the unity gain inverting op - amps .
@xmath114 sign indicates squarer using ad633 . ]
next , we implement the coupled system of van der pol oscillators given by eq . in an electronic circuit ( figure [ ckt ] ) .
we use tl074 ( quad jfet ) op - amps , and ad633 analog multiplier ics .
a @xmath115 v power supply is used ; resistors ( capacitors ) have @xmath116 ( @xmath117 ) tolerance .
the unlabeled resistors have the value @xmath110 k@xmath111 .
the op - amp aq is used to generate the mean - field : @xmath118 , which is subtracted by @xmath119 using op - amps denoted by a. one can see that @xmath120 determines the coupling strength , @xmath121 determines the mean - field density and @xmath122 controls the feedback parameter @xmath10 .
the voltage equation of the circuit can be written as : [ ckteqn ] @xmath123,\\ cr\frac{d{v}_{yi}}{dt}&=\frac{r}{r_a}\left(v_\delta-\frac{v_{xi}^2}{10}\right)\frac{v_{yi}}{10}-v_{xi}.\end{aligned}\ ] ] here @xmath124 .
eqs . is normalized with respect to @xmath125 , and thus now becomes equivalent to eq .
for the following normalized parameters : @xmath126 , @xmath127 , @xmath128 , @xmath129 , @xmath130 , @xmath131 , @xmath132 , @xmath133 , and @xmath134 .
@xmath135 is the saturation voltage of the op - amp . in the experiment
we take @xmath113 v , and @xmath136 nf ; we choose @xmath58 by taking @xmath137 @xmath111 [ using a precision potentiometer ( pot ) ] .
we experimentally observe the revival of oscillation by revoking the oscillation cessation states ( ad and od ) with varying @xmath10 ( i.e. , @xmath138 ) . at
first we consider the case of the ad state : for that we choose @xmath66 and @xmath57 ( by setting @xmath139 k@xmath111 and @xmath140 k@xmath111 , respectively ) and decrease @xmath10 from @xmath63 to a lower value .
figure [ f:5](a ) shows experimental snapshots of the ad state in @xmath141 and @xmath142 at @xmath143 k@xmath111 ( i.e. , @xmath144 ) [ using dso , agilent make , dso - x 2024a , 200 mhz , 2 gs / s ] ; in the same figure we show the revival of oscillation from the ad state at an exemplary value @xmath145 k@xmath111 ( @xmath146 ) .
figure [ f:5](b ) gives the numerical time series at the corresponding @xmath10 values ( fourth - order runge - kutta method with step - size of @xmath147 ) .
the numerical results are in good agreement with the experimental observations : also , the dynamical behaviors at these parameter values are in accordance with fig .
[ f:1](f ) .
next , we choose an od state for @xmath70 and @xmath57 ( i.e. , @xmath148 k@xmath111 and @xmath140 k@xmath111 , respectively ) : experimental snapshots of the od state at @xmath149 k@xmath111 ( @xmath150 ) and the rhythmicity at @xmath151 k@xmath111 ( @xmath152 ) are shown in fig .
[ f:5](c ) .
the corresponding numerical result is given in fig .
[ f:5](d ) .
we see that the experimental and numerical results are in good agreement and in accordance with fig .
[ f:1](e ) .
significantly , despite the presence of inherent parameter fluctuations and noise , which is inevitable in an experiment , it is important that in both the above cases theory and experiment are in good qualitative agreement , which proves the robustness of the coupling scheme in restoring rhythmicity in coupled oscillators .
( color online ) ( a , c ) experimental real time traces of @xmath141 and @xmath142 along with the ( b , d ) numerical time series plots of @xmath153 and @xmath154 . [ ( a ) and ( b ) ] with @xmath139 k@xmath111 ( i.e. @xmath155 ) a decreasing @xmath122 ( @xmath10 ) restores oscillation ( lc ) from ad : ad at @xmath143 k@xmath111 ( @xmath144 ) , lc at @xmath145 k@xmath111 ( @xmath146 ) .
[ ( c ) and ( d ) ] with @xmath148 k@xmath111 ( i.e. @xmath156 ) a decreasing @xmath122 ( @xmath10 ) restores oscillation ( lc ) from od : od at @xmath149 k@xmath111 ( @xmath150 ) , lc at @xmath151 k@xmath111 ( @xmath152 ) .
others parameters are @xmath140 k@xmath111 ( @xmath57 ) and @xmath157 @xmath111 ( @xmath58 ) .
@xmath158 axis : ( a ) 200 mv / div ( c ) 100 mv / div ; @xmath159 axis : 380 @xmath160s / div . ]
we have investigated the effect of a feedback parameter , which controls the diffusion rate , on a network of nonlinear oscillators coupled under mean - field diffusion .
we have shown that , unlike other coupling schemes , here two control parameters exist , namely the density of mean - field diffusion and the feedback parameter .
the interplay of these two parameters can revive rhythmicity from any oscillation cessation state : in fact by controlling the feedback parameter closer to the density of the mean - field one can shrink the region of the oscillation cessation state to a very narrow zone in parameter space .
more interestingly , even in the presence of complete diffusion ( i.e. , in the absence of feedback parameter ) , the density of the mean - field alone can induce rhythmicity from a death state .
thus , it offers a very robust oscillation revival mechanism .
we have extended our study to a network consists of large number of nodes and shown that the oscillation cessation states can be revoked in that case too . finally , we have supported our theoretical results by an experiment with electronic van der pol oscillators and for the first time observed the revival of oscillation from the mean - field - diffusion induced death state . since both the density of the mean - field and the feedback parameter have strong connections with many real biological networks including cell cycle , neural network , synthetic genetic oscillators , etc .
, we believe that the present study will broaden our understanding of those systems and subsequently this study will shed light on the control of oscillation suppression mechanism in several biological and engineering systems .
t. b. acknowledges the financial support from serb , department of science and technology ( dst ) , india [ project grant no .
: sb / ftp / ps-005/2013 ] .
d. g. acknowledges dst , india , for providing support through the inspire fellowship .
j. k. acknowledges government of the russian federation ( agreement no .
14.z50.31.0033 with institute of applied physics ras ) .
31ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo
secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop _ _ , ed .
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) | the revival of oscillation and maintaining rhythmicity in a network of coupled oscillators offer an open challenge to researchers as the cessation of oscillation often leads to a fatal system degradation and an irrecoverable malfunctioning in many physical , biological and physiological systems . recently a general technique of restoration of rhythmicity in diffusively coupled networks of nonlinear oscillators has been proposed in [ zou et al .
nature commun . 6:7709 , 2015 ] , where it is shown that a proper feedback parameter that controls the rate of diffusion can effectively revive oscillation from an oscillation suppressed state .
in this paper we show that the mean - field diffusive coupling , which can suppress oscillation even in a network of identical oscillators , can be modified in order to revoke the cessation of oscillation induced by it . using a rigorous bifurcation analysis
we show that , unlike other diffusive coupling schemes , here one has _
two control parameters _ , namely the _ density of the mean - field _ and the _ feedback parameter _ that can be controlled to revive oscillation from a death state .
we demonstrate that an appropriate choice of density of the mean - field is capable of inducing rhythmicity even in the presence of complete diffusion , which is an unique feature of this mean - field coupling that is not available in other coupling schemes .
finally , we report the _ first _ experimental observation of revival of oscillation from the mean - field induced oscillation suppression state that supports our theoretical results . , |
to elucidate the nature of critical phenomena of two - dimensional disordered ising models has been a subject of a long - standing interest .
harris[1 ] concluded from a heuristic argument that the nature of critical phenomena of a disordered system becomes different from that of the corresponding pure system when the critical exponent of the specific heat , @xmath4 , of the pure system is positive , while it remains the same when @xmath5 .
since the two - dimensional pure ferromagnetic ising model is the marginal case , namely @xmath6 , many authors have investigated the properties of critical phenomena of the two - dimensional disordered ising models[2 - 14 ] .
for the two - dimensional unfrustrated random ising models , many authors concluded that the specific heat diverges double logarithmically at the paramagnetic - ferromagnetic phase boundary[2 - 9 ] , though there are several results which insist that the specific heat remains finite[1,10 - 12 ] .
for the two - dimensional @xmath0 ising model which corresponds to the two - dimensional frustrated random ising model , a few results exist .
inoue[13 ] investigated the system - size dependence of the height of the peak of the specific heat at @xmath7 , and concluded that the logarithmic divergence of the specific heat at the paramagnetic - ferromagnetic phase boundary is most probable in the same way as the pure case .
reis _ et al_[14 ] investigated the system - size dependence of the correlation length at @xmath8 , and concluded that logarithmic corrections do not play a role in contrast with the unfrustrated disordered systems .
they concluded , however , that the specific heat diverges at most logarithmically .
namely , they could not completely exclude the possibility of the double logarithmic divergence of the specific heat .
on the other hand , there exisits the nishimori line , which is defined by the equation , @xmath9 .
( @xmath10 is the boltzmann constant , which we put that @xmath11 from now on . ) on the nishimori line , several rigorous results have been derived[15 ] .
particularly , it has been proved that the specific heat remains finite . from the renormalization group approach
, it is generally believed that , for the two - dimensional @xmath0 ising model , there exist two fixed points on the paramagnetic - ferromagnetid phase boundary above the multicritical point , namely the pure ferromagnetic fixed point and the multicritical fixed point[16 - 17 ] , and there seems to be no random fixed point in contrast with the three - dimensional case[18 ] .
( in this paper , we use the word `` multicritical point '' as the crossing point of the ferromagnetic - nonferromagnetic phase boundary and the nishimori line , though there seems to be no spin glass phase in the two - dimensional case . )
the pure ferromagnetic fixed point is stable , and the multicritical fixed point is unstable .
thus , the critical phenomena on the paramagnetic - ferromagnetic phase boundary above the multicritical point is governed by the pure ferromagnetic fixed point .
therefore , the critical exponent , @xmath4 , should be zero on the whole paramagnetic - ferromagnetic phase boundary above the multicritical point .
the fact mentioned above , however , does not make a help to determine the critical behaviour of the specific heat , since each of a logarithmic divergence , a double logarithmic divergence and a cusp like behaviour belongs to the case , @xmath6 . in this paper
, we investigate the specific heat from a new point of view .
we devide the specific heat of a system into two parts , @xmath12 and @xmath13 .
( for the detail definitions of @xmath12 and @xmath13 , see section 2 . )
the value of @xmath12 is easily found to be non - negative and remain finite at finite temperature . on the other hand ,
the value of @xmath13 becomes negative at @xmath14 when the system is frustrated and the ground state has non - trivial degeneracy .
since we can consider that the negative value of @xmath13 reflects the non - trivial ground state degeneracy , it is an interesting problem to make it clear up to what temperature the property persists .
it is noted that in order that the specific heat may diverge , @xmath13 should become infinite .
thus , we investigate the region in the @xmath1 plane where @xmath13 takes the negative value for the two - dimensional @xmath0 ising model .
our result shows that the region mentioned above is fairly large in the @xmath1 plane .
moreover , it includes the nishimori line .
namely , the region includes a part of the paramagnetic - ferromagnetic phase boundary above the multicritical point . in that region ,
the specific heat can not diverge .
the present result indicates that the specific heat does not diverge at least on a part of the paramagnetic - ferromagnetic phase boundary near and above the multicritical point .
the result is not directly in conflict with previous results[13,14 ] which insist that the specific heat diverges at most logarithmically at the paramagnetic - ferromagnetic phase boundary near the pure ferromagnetic case(@xmath15 ) , since the range of the concentration of the calcultions does not overlap .
it is natural , however , to think that the nature of the phase transition does not change on the paramagnetic - ferromagnetic phase boundary above the multicritical point .
thus , we insist that the specific heat of the two - dimensional @xmath0 ising model does not diverge on the whole paramagnetic - ferromagnetic phase boundary above the multicritical point , though we can not exclude the possibility that there exists a singular point on the paramagnetic - ferromagnetic phase boundary which separates the nature of the specific heat .
we consider the two - dimensional @xmath0 ising model on a square lattice with only nearest neighbor interactions .
the hamiltonian is written as follows : @xmath16 where @xmath17 , and the summation of @xmath18 runs over all the nearest neighbors .
each @xmath19 is determined according to the following probability distribution : @xmath20 in this paper , we put that @xmath21 . the specific heat per bond , @xmath22 , is written as follows ; @xmath23_{p } = \frac { 1}{n_{\rm b}}\frac{\partial } { \partial t } \sum_{(ij)}[<-j_{ij}s_{i}s_{j}>_{t}]_{p},\ ] ] where @xmath24 denotes the thermal average in a given bond configuration , @xmath25 , at temperature , @xmath3 .
@xmath26_{p}$ ] denotes the configurational average at the ferromagnetic bond concentration , @xmath2 , and @xmath27 is the number of bonds .
now , we divide the specific heat of the system , @xmath22 into two parts , @xmath28 and @xmath29 : @xmath30 where @xmath31_{p}\mid_{\{t\}=t},\ ] ] and @xmath32_{p}\mid_{\{t\}=t}.\ ] ] here , we introduce technically , @xmath33 , the local temperature of the bond , @xmath19 .
namely , @xmath28 is considered to be the configurational average of the change of the local energy , @xmath34 , when we infinitesimally increase the correspondeing local temperature , @xmath33 . on the other hand , @xmath29 is considered to be the configuratinal average of the the change of the local energy , @xmath34 , when we infinitesimally increase the temperarture , @xmath35 , which surrounds the local bond , @xmath19 .
it is easily calculated that @xmath36_{p}).\ ] ] namely , @xmath28 has an upper bound ; @xmath37 therefore , at finite temperature , in order that the specific heat may diverge , @xmath29 should become infinite .
now , we consider the case at zero temperature . at zero temperature ,
the specific heat , @xmath38 becomes zero .
when , there is no non - trivial degeneracy in the ground state of the system , each @xmath39_{p}=1 $ ] .
namely , @xmath40 becomes zero .
thus , @xmath41 also becomes zero . on the other hand ,
when the ground state has non - trivial degeneracy , some of @xmath39_{p}$ ] become less than one .
namely , @xmath40 becomes positive infinite .
thus , @xmath41 becomes negative infinite in this case .
the negative value of @xmath29 at finite temperature may be considered to be one of the influence of the non - trivial ground state degeneracy of the frustrated system .
thus , it is an intersting problem to make it clear up to what temperature the property persists .
therefore , in the following sections , we investigate the region in the @xmath1 plane where @xmath29 takes the negative value for the two - dimensional @xmath0 ising model .
in this section , we investigate the region in the @xmath1 plane where @xmath29 takes the negative value by the numerical transfer matrix method .
we have calculated for the lattice size , @xmath42 , at the bond concentration , @xmath43 . in the calculations ,
we take the free boundary condition , and take the configurational average for @xmath44 bond configurations .
figure 1 shows the temperature dependence of @xmath22 , @xmath28 and @xmath29 for @xmath45 at @xmath46 .
we can see that @xmath29 takes the negative value in the low temperature region .
we have estimated the temperature , @xmath47 , where @xmath29 takes the value zero for @xmath42 at @xmath43 .
figure 2 shows the temperature dependence of @xmath29 near the zero point for @xmath48 at @xmath49 . from the figure
, we have estimated the temperture of the zero point , @xmath50 .
the estimated values of @xmath47 for various @xmath51 and @xmath2 are shown in table 1 .
the accuracy of the values of @xmath47 becomes worse as the concentration , @xmath2 , becomes small , since the gradient of @xmath29 near the zero point becomes small .
we can see that , at each concentration , @xmath2 , the size dependence of the values of @xmath47 is very small , and the values are fairly large compared to the value of @xmath52 , the tempearature of the nishimori line at @xmath2 .
@llllllll @xmath51&@xmath53&@xmath54&@xmath55&@xmath46&@xmath56&@xmath49&@xmath57 + 4&3.88(6)&2.85(5)&2.222(16)&1.803(8)&1.4968(28)&1.2553(13)&1.0303(20 ) + 6&3.88(6)&2.845(35)&2.220(12)&1.8045(45)&1.5025(25)&1.2672(12)&1.0517(12 ) + 8&3.88(12)&2.835(65)&2.234(22)&1.8065(85)&1.5020(40)&1.2722(22)&1.0650(20 ) + 10&3.90(10)&2.885(45)&2.230(12)&1.8035(75)&1.5045(30)&1.2744(26)&1.0650(20 ) + 12&3.88(4)&2.83(5)&2.220(14)&1.8035(65)&1.5038(25)&1.2759(19)&1.0823(18 ) + 14&3.88(4)&2.862(42)&2.225(14)&1.8055(35)&1.5050(25)&1.2764(14)&1.0880(20 ) + 16&3.89(7)&2.845(30)&2.225(15)&1.803(7)&1.5047(19)&1.2768(18)&1.0932(12 ) + @xmath52&3.2308&2.3604&1.8205&1.4427&1.1530&0.9102&0.6792 + we have no definite principle to estimate the value of @xmath58 which is the extrapolated value of @xmath47 to @xmath59 .
no clear size dependence , however , can be seen at @xmath60 .
thus , we perform naive extrapolation to @xmath59 in this region . it can be seen that the value of @xmath47 slightly increases as @xmath51 increases at @xmath61 .
we have found that the extrapolation by the @xmath62-law works fairly well at @xmath56 and @xmath49 , where @xmath63 is the total number of the spins of the system .
figure 3 shows a plot of @xmath47 versus @xmath64 at @xmath49 , where we have estimated that @xmath65 .
we have also found that the extrapolation by the @xmath66-law works fairly well at @xmath57 .
the estimated values of @xmath58 for various @xmath2 are shown in table 2 .
strictly speaking , we can not justify the above extrapolaton at each @xmath2 .
we can say , however , that the extrapolated values , @xmath58 , might not change drastically even if we use other extrapolation laws , and there seems to be no possibility that the value of @xmath58 becomes smaller than the temperature of the nishimori line at each @xmath2 .
@lll @xmath2&@xmath58&@xmath52 + 0.65&3.88(6)&3.2308 + 0.7&2.845(65)&2.3604 + 0.75&2.217(19)&1.8205 + 0.8&1.805(9)&1.4427 + 0.85&1.506(3)&1.1530 + 0.9&1.278(3)&0.9102 + 0.95&1.154(3)&0.6792 + figure 4 shows the estimated values of @xmath58 for various @xmath2 in the @xmath1 plane , which are denoted by black squares .
the dashed line denotes the nishimori line , and the open circles denote the paramagnetic - ferromagnetic phase boundary above the multicritical point[19 ] .
we can see that the region where @xmath29 takes the negative value is fairly large .
we can also see that the region mentioned above contains the nishimori line .
there are several numerical calculations about the paramagnetic - ferromagnetic phase boundary of the two - dimensional @xmath0 ising model[13,19 - 22,27 ] .
it is noted that all the calculations are almost consistent with each other and they show that , above the multicritical point , the concentration , @xmath2 , of the phase boundary increases as the temperature increases .
thus , a part of the paramagnetic - ferromagnetic phase boundary is included in the region where @xmath29 takes the negative value . in that region , the specific heat can not diverge .
namely , the specific heat does not diverge at least on a part of the paramgnetic - ferromagnetic phase boundary above the multicritical point .
in order to elucidate the region where @xmath29 takes the negative value near the paramagnetic - ferromagnetic phase boundary , we have performed the monte carlo simulations for more larger lattices , @xmath67 , at @xmath68 . in the calculations , we take the skew boundary condition in one direction and the periodic boundary condition in the other direction . the condition of the simulation is shown in table 3 .
m.c.s . denotes the monte carlo step of the simulation , which is chosen to be @xmath69 , where the relaxation time , @xmath70 , is evaluated by the statistical time - independent method[23 ] , and @xmath71 is the number of the bond configurations .
@llll @xmath2&@xmath51&@xmath72&@xmath71 + 0.88&31&400000&240 + & 61&800000&96 + & 121&800000&120 + 0.89&31&600000&240 + & 61&5000000&120 + & 121&16000000&80 + 0.9&31&800000&320 + & 61&1400000&160 + & 91&30000000&64 + 0.91&31&400000&480 + & 61&10000000&96 + & 121&20000000&32 + 0.92&31&600000&720 + & 61&1400000&120 + & 121&800000&64 + the temperature dependence of @xmath29 near the zero point for @xmath73 at @xmath49 is shown in figure 5 , where we have estimated that @xmath74 .
the estimated values of @xmath47 for various @xmath51 and @xmath2 are shown in table 4 .
@lllll @xmath2&@xmath75&@xmath76&@xmath77&@xmath78 + 0.88&1.3485(55)&1.3535(45)&&1.3555(35 ) + 0.89&1.3075(45)&1.303(5)&&1.3055(55 ) + 0.9&1.286(4)&1.2775(45)&1.2765(45 ) & + 0.91&1.269(4)&1.2665(45)&&1.2685(45 ) + 0.92&1.249(4)&1.2485(45)&&1.2485(35 ) + though @xmath47 takes similar value even if the lattice size , @xmath51 , changes , we have extrapolated the values of @xmath79 to @xmath59 by the @xmath62-law .
we show the estimated values of @xmath58 for various @xmath2 in table 5 with the temperature of the nishimori line , @xmath52 .
@lll @xmath2&@xmath58&@xmath52 + 0.88&1.355(5)&1.0038 + 0.89&1.3045(55)&0.9566 + 0.9&1.275(5)&0.9102 + 0.91&1.267(5)&0.8644 + 0.92&1.2485(55)&0.8188 + the results are also shown in figure 6 .
the black circles and squares denote @xmath58 evaluated by the monte carlo simulations and by the numerical transfer matrix method , respectively .
the dashed line denotes the nishimori line , and the open circles denote the paramagnetic - ferromagnetic phase boundary above the multicritical point[16 ] .
it can be seen that both results by the numerical transfer matrix method and monte carlo simulations are consistent with each other .
the crossing point of the paramagnetic - ferromagnetic phase boundary and the boundary of the region where @xmath29 takes the negative value has been estimated to be 0.8985(15 ) .
therefore , we conclude that the specific heat can not diverge on the paramagnetic - ferromagnetic phase boundary at least for @xmath80 , where @xmath81 is the concentration of the multicritical point
. there are many numerical estimates of @xmath81 : 0.8905(5)[14],0.8872(8)[24],0.886(3)[25 ] , 0.8906[26 ] and 0.8907(2)[27 ] .
recently , there is a conjecture about the exact value of @xmath81 , which insists that @xmath82[28 ] .
we have investigated the property of the specific heat of the two - dimensional @xmath0 ising model by the numerical transfer matrix method and monte carlo simulations from a new point of view .
we have estimated the region where @xmath29 takes the negative value in the @xmath1 plane , which is one of the characteritic of the frustrated system , and reflects the non - trivial degeneracy of the ground state . by the numerical transfer matrix method for @xmath42 , and the monte carlo simulations for @xmath67 , the region
have been found to be fairly large in the @xmath1 plane , which includes the nishimori - line .
moreover , the region includes a part of the paramagnetic - ferromagnetic phase boundary , on which the specific heat can not diverge .
thus , our results indicate that the specific heat can not diverge on the paramagnetic - ferromagnetic phase boundary at least near and above the multicritical point .
on the other hand , there are several literatures that , near the pure ferromagnetic point(@xmath8 ) , the specific heat diverges at most logarithmically[13,14 ] .
both results are not directly in conflict with each other , since the range of the concentration of the calculations does not overlap .
it is natural , however , to think that the nature of the phase transition does not change on the paramagnetic - ferromagnetic phase boundary above the multicriticsl point . in the literatures which insist the divergence of the specific heat , the change of the peak height of the specific heat of various lattice size was investigated[13,14 ] .
it is a very subtle problem whether the specific heat diverges or remains finite , when the divergence is so weak as the logarithmic divergence .
our extrapolation , however , is straightforward and the values of @xmath58 might not change drastically even when there exist other system - size corrections , and there seems to be no possibility that the value of @xmath58 becomes smaller than the temperature of the nishimori line at each @xmath2 .
thus , we insist that the specific heat of the two - dimensional @xmath0 ising model does not diverge on the whole range of the paramagnetic - ferromagnetic phase boundary above the multicritical point , though we can not exclude the possibility that a singular point exists on the paramagnetic - ferromagnetic phase boundary , which separates the nature of the specific heat
. the authors would thank dr .
y. ozeki and dr .
k. hukushima for their useful discussions .
the calculations were made on hitac sr8000 at university of tokyo . and at the institue for solid state physics in university of tokyo .
100 harris a b 1974 _ j. phys .
c _ * 7 * 1671 dotsenko b v s and dotsenko v s 1983 _ advan . phys . _
* 32 * 129 shanker r 1987 _ phys .
* 23 * 2466 .
blackman j a and poulter j 1984 _ j. phys .
c _ * 17 * 107 wang j -s , selke w and dotsenko v s 1990 _ physica a _ * 164 * 221 wiseman s and domany e 1995 _ phys . rev .
b _ * 51 * 3074 stauffer d ,
reis f d a a , de queiroz s l a and dos santos r r 1997 _ int .
c _ * 8 * 1209 selke w , shchur l n and vasilyev o a 1998 _ physica a _ * 259 * 388 ballesteros h g , fernandez l a , mayor v m , sudupe a m , parisi g and ruiz - lorenzo j j 1997 _ j. phys . a _ * 30 * 8379 tamaribuchi t and takano f 1980 _ prog
* 64 * 1212 kim j and patrascioiu a 1994 _ pys
lett . _ * 72 * 2785 zielger k 1991 _ europhys .
_ * 14 * 415 inoue m 1995 _ j. phys .
* 64 * 3699 aarao reis f d a , de queiroz s l a and dos santos r r 1998 _ phys .
b _ * 60 * 6740 nishimori h 1981 _ prog
phys . _ * 66 * 1169 jacobsen j l and picco m 2002 _ phys .
e _ * 65 * 026113 honecker a , picco m and pujol p 2001 _ phys
lett . _ * 87 * 047201 hukushima k 2000 _ j. phys .
* 69 * 631 ito n , ozeki y and kitatani h 1999 _ j. phys .
jpn . _ * 68 * 803 ozeki y and nishimori h 1987 _ j. phys .
. jpn _ * 56 * 3265 kitatani h and oguchi t 1990 _ j. phys .
_ * 59 * 3823 kitatani h and oguchi t 1992 _ j. phys .
. jpn _ * 61 * 1598 kikuchi m and ito n 1993 _ j. phys .
jpn . _ * 62 * 3052 ozeki y and ito n 1998 _ j. phys .
* 31 * 5451 singh r r p and adler j 1996 _ phys .
b _ * 54 * 364 honecker a , picco m and pujol p 2001 _ pjys
lett . _ * 87 * 047201 .
merz f and chalker j t 2002 _ phys .
b _ * 65 * 054425 .
nishimori h and nemoto k 2002 j. phys .
soc . jpn . *
71 * 1198 for various @xmath2 in the @xmath1 plane , which are denoted by black squares .
the dashed line denotes the nishimori line , and the open circles denote the paramagnetic - ferromagnetic phase boundary above the multicritical point.,width=245 ] for various @xmath2 in the @xmath1 plane .
the black circles and squares denote the values of @xmath58 by the monte carlo simulations and by the numerical transfer matrix method , respectively .
the dashed line denotes the nishimori line , and the open circles denote the paramagnetic - ferromagnetic phase boundary above the multicritical point.,width=245 ] | the specific heat of the two - dimensional @xmath0 ising model has been investigated by the numerical transfer matrix method and monte carlo simulations from a new point of view .
the region where a part of the specific heat takes the negative value has been investigated , which is characteristic of frustrated systems and reflects the non - trivial degeneracy of the ground state .
the region mentioned above is found to be fairly large in the @xmath1 plane ( @xmath2 is the concentration of the ferromagnetic bond and @xmath3 is the temperature ) .
moreover , it includes the nishimori line .
namely , it includes a part of the paramagnetic - ferromagnetic phase boundary , on which the specific heat can not diverge .
the present result indicates that the specific heat does not diverge at least on a part of the paramagnetic - ferromagnetic phase boundary above the multicritical point , which is in conflict with previous results . |
the interstellar medium ( ism ) is known to have clumpy density and velocity structures .
scale - free intensity fluctuations in a variety of tracers ( e.g. h i 21 cm emission , dust emission ) is detectable over a wide range of scales .
this scale - free fluctuation is generally interpreted as the signature of a turbulent ism .
the evidence for small scale structures and turbulence in the atomic ism of the galaxy have been found through various observational techniques .
power law scaling for the velocity dispersion of the h i 21 cm emission has been observed to sub - pc scales @xcite . on somewhat larger scales , the intensity fluctuations in h
i 21 cm emission show a scale free power spectrum @xcite .
the h i opacity towards cassiopeia a and cygnus a show fluctuations on scales ranging from @xmath8 au to @xmath9 pc for the absorption in the perseus arm , the local arm , and the outer arm @xcite .
very long baseline interferometry ( vlbi ) observations reveal structures on scales as small as @xmath10 au @xcite .
@xcite , @xcite , @xcite and @xcite have shown h i opacity variations on scales of @xmath1 au from multi - epoch observations of some high velocity pulsars .
@xcite have reported similar opacity variations on @xmath11 au scales towards the microquasar grs @xmath12 .
observations of some other pulsars , however , do not show any opacity variations @xcite , indicating that such tiny scale structures may be rare .
small scale ism structures are not unique to the milky way .
the h i emission from the large magellanic cloud , the small magellanic cloud , several dwarf galaxies in the m81 group and a sample of dwarf and nearby spiral galaxies also show considerable small scale intensity fluctuations @xcite . from an observational point of view , the h i 21 cm opacity fluctuations depend on fluctuations in the density and temperature of the gas .
but , for small density fluctuations , the dependence on temperature is weak @xcite .
earlier , these small fluctuations were believed to be physical structures with densities @xmath13 @xmath14 . however , explaining the existence of these structures in equilibrium with orders of magnitude lower density gas in the surrounding medium , is challenging .
later , it was suggested that the fluctuations at larger spatial scales could , in projection , give rise to the observed small scale transverse fluctuations @xcite . on the other hand ,
there are some direct observational evidence of @xmath15 au size h i structures with a density of @xmath16 @xmath14 @xcite .
recent numerical simulations of the turbulent ism have also indicated the existence of tiny scale structures @xcite .
the short evaporation time - scale ( @xmath17 myr ) of these structures imply that they can survive if either the ambient pressure is much higher than the average ism pressure or they are formed on a similar time - scale . in any case , their physical properties are not yet well understood .
here we present results on tiny scale h i opacity fluctuations towards the quasar 3c 138 derived from vlba , merlin and vla data . combining merlin and vla data with vlba data
has resulted in improved sensitivity and larger spatial dynamic range to detect small optical depth fluctuations and to derive the structure function .
the data analysis method is outlined in section
[ sec : oaa ] .
[ sec : rsl ] contains the results , and the conclusions are presented in section [ sec : con ] .
for this analysis , we have used vlba , merlin and vla data .
results from the merlin and vlba observations were previously presented by @xcite , @xcite and @xcite .
the vlba observation ( project bd0026 ) was carried out on september 10 , 1995 .
the spectral resolution of this data is @xmath18 km s@xmath19 .
the merlin observation was carried out on october 22 and 23 , 1993 , and the spectral resolution in this case is @xmath18 km s@xmath19 .
the vla a - configuration observations were carried out on september 07 and 13 , 1991 ( project as0410 and test ) .
the channel width and spectral resolution in this case are @xmath20 and @xmath21 km s@xmath19 , respectively .
@xcite reported temporal variation of h i opacity towards 3c 138 from the vlba observations taken in 1995 , but also include data from 1999 and 2002 . here
, we have only included the 1995 data , since it is closest to the merlin and vla observation epochs .
the longest baseline was about @xmath22 , @xmath23 and @xmath24 ( angular resolution of about @xmath25 , @xmath26 and @xmath27 mas ) for the vla , merlin and vlba data respectively .
however , even with the combination of the vlba , merlin and vla , sufficient brightness temperature sensitivity was achieved only on angular scales larger than about @xmath28 mas , corresponding to baselines shorter than about @xmath29 .
data analysis was carried out using the nrao astronomical image processing system ( aips ) .
all the data were converted to j2000.0 epoch , and the spectral axis was regrided to a common spectral resolution of @xmath30 km s@xmath19 .
a combined dataset was produced after first reweighting the vlba , merlin and vla data to have similar data weights ( i.e. so they would contribute equally at the imaging stage ) . a continuum and continuum - free dataset
were produced by fitting the line - free channels in the uv - plane .
self - calibration was repeatedly done on the continuum data starting with a vlba image as the initial model , and then including the merlin and the vla baselines at later stages .
the self - calibration solutions obtained for the continuum data were also applied to the spectral line data cube .
the final continuum image , shown in figure ( [ fig : ths1 ] ) , and the spectral line image cube were made using multiscale imaging with a @xmath28 mas restoring beam .
the structures of the continuum and line images agree with the vlba images of @xcite . the spectral image cube and the continuum image were used to obtain the optical depth image cube .
a @xmath31 cutoff was used for blanking the continuum image , and the same blanking window was applied to the spectral data to get rid of possible contribution from high noise or other low level image artifacts .
the optical depth image cube was used to derive the opacity fluctuation structure function .
we note that the structure function amplitude of the line - free channels is more than an order of magnitude lower than the structure function amplitude of the channels with h i absorption signal .
but , for the channels with h i signal , the optical depth noise is significantly higher than that of the line - free channels .
since the rms noise in the continuum image ( @xmath32 ) is significantly lower than the rms noise in line channels ( @xmath33 ) , the optical depth noise can be approximated as @xmath34 , where @xmath35 is the continuum image flux density .
thus , for high optical depth , the noise is scaled up by a factor of @xmath36 .
so , the optical depth noise is highly correlated with the optical depth itself .
numerical analysis shows that , as a result , the noise bias will have a similar power law dependence on all angular scales , and will scale up the structure function amplitude without changing the power law index .
moreover , the h i brightness temperature for the line channels will contribute to the system temperature , and will also increase the image noise . as a result , estimating the noise contribution from the line - free channels will underestimate the errors to the structure function due to optical depth uncertainties .
statistical errors on the derived structure function for individual channels are estimated from the variance of the measurements within the bin , and do not account for the high optical depth measurement uncertainty . if the high opacity pixels are excluded , the effect of this uncertainty and the noise bias to the derived structure function is minimized .
but in that case , the opacity image is modified by a window function , and the structure function can only be derived for about a factor of @xmath9 smaller range of angular scales , and only @xmath37 spectral channels .
hence , first the high opacity pixels are included to derive the initial structure function . with significant noise bias ,
the amplitude of the structure function in this case is only an upper limit .
however , one can use a larger range of angular scale to fit the power law index .
then @xmath38 cutoff is used to minimize the noise bias and to compute the amplitude of the structure function from a limited range of angular scales .
for both cases , statistical errors are estimated using the channel to channel variation of the structure function .
essentially , the opacity values in any channel may be considered as an independent realization drawn from a parent distribution of opacities with associated uncertainty .
so , the channel to channel variation of the structure function will include the opacity uncertainty effect .
the errors are estimated assuming that there is no spectral variation of the structure function .
thus , one can only derive an average structure function from all the spectral channels .
3c 138 is unresolved at the short baselines ( @xmath39 ) of the vla .
thus , by combining the vlba , merlin and vla data , we are able to image 3c 138 maintaining the small scale structures at the vlba resolution , and , at the same time , restoring the total flux density observed by the vla . the intermediate merlin and vlba baselines provide a wide range of uv - coverage , allowing us to probe structures at scales ranging from about ten milliarcsec to about @xmath40 arcsec . adopting a distance to the absorbing gas of 500 pc @xcite
, the range of angular scales translates to @xmath1 au .
the integrated h i 21 cm absorption spectrum is shown in figure ( [ fig : ths2 ] ) .
the shaded region in figure ( [ fig : ths2 ] ) indicates the channels that have adequate signal to noise in optical depth images for further analysis of small scale opacity fluctuations .
we have calculated the structure function for @xmath41 individual spectral channels , each with a velocity width of @xmath18 km s@xmath19 , and used all the measurements together to derive the best estimate of the structure function slope and amplitude . for quantitative study of small scale fluctuations , statistical tools like the correlation function , structure function , and power spectrum
are very useful .
the observed shape of these functions are expected to be related to the physical processes that gives rise to the fluctuations .
the power spectrum of the opacity fluctuations @xmath42 is the fourier transform of the correlation function @xmath43 : @xmath44 where @xmath45 and @xmath46 are the sky coordinate and inverse angular separation respectively . here
, @xmath43 is defined as @xmath47 and the structure function @xmath48 ^ 2 $ ] is given by @xmath49 ^ 2\rangle \,.\ ] ] strictly speaking , @xmath50 and @xmath51 are function of @xmath52 and @xmath53 , respectively , but we have assumed statistical isotropy for the purpose of simplification .
thus , throughout this analysis , @xmath50 and @xmath51 are considered to only be functions of scalar amplitude @xmath54 and @xmath55 , respectively . generally , if the power spectrum is a power law of the form @xmath56 with @xmath57 , the structure function @xmath58 will have a power law index of @xmath59 , and the rms opacity fluctuation @xmath60 will have a power law index of @xmath61 @xcite .
we also note that if the density fluctuations are small ( that is @xmath62 ) , the density fluctuation power spectrum is expected to be very similar to the opacity fluctuation power spectrum @xcite .
the derived structure function of the h i absorption towards 3c 138 for a few spectral channels are shown in figure ( [ fig : ths3 ] ) . the amplitude and the shape of the function changes from channel to channel .
the structure function from one spectral channel with no h i signal is also shown in this figure for a comparison .
the measured structure function for all 36 spectral channels with adequate signal to noise is shown in figure ( [ fig : ths6 ] ) .
the open point symbols are structure function values without any @xmath38 cutoff . clearly , below @xmath63 milliarcsec
, the structures are smoothed by the synthesized beam . above @xmath64 milliarcsec ,
the effects of the spatial window due to the shape of the continuum emission significantly modify and flatten the measured structure function . the mean and rms values of these points in each bin
are shown as filled circles with error bars ( with position shifted slightly to the left for clarity ) .
the best fit power law for the ensemble ( dash - dot lines in the figure ) has a power law index of @xmath65 ( @xmath31 ) , which corresponds to a power spectrum with a power law index @xmath66 .
next , we have used a cutoff based on @xmath38 to minimize the noise bias and obtain a better estimate of the structure function amplitude .
as mentioned earlier , with this cutoff we have a reliable estimate of the structure function only for @xmath67 channels and a limited range of angular scales . here
, we have used a cutoff at @xmath68 with usable range of @xmath69 mas for eight spectral channels .
these estimates ( shown as filled points joined by dotted lines in the same figure ) are also consistent with a power law index of @xmath70 .
since @xmath71 , the signal structure function is at least a factor of @xmath9 higher than the noise bias case .
so , the amplitude of the signal structure function is constrained to be higher than @xmath72 times the observed amplitude .
unfortunately , for any higher cutoff , the flattening due to the spatial window dominates , and essentially makes it impossible to extract the true structure function .
the solid line in the figure is the structure function from @xcite extrapolated from the observed structure function power index of @xmath73 in the range of @xmath74 pc .
clearly the observed structure function amplitude at smaller scales ( @xmath1 au ) is @xmath7 times higher than the extrapolated value . in terms of rms opacity fluctuation ,
the observed values are a factor of @xmath75 times higher than the predicted values . apart from considering measurements from all the channels together for fitting , each channel is considered separately to cross - check if there is any significant channel to channel variation of the structure function .
note that the errors on the structure function derived from single channel reflect only the variance of the measurements , and do not include the effect of large optical depth uncertainties .
the best fit power law amplitude and index for each channel are shown in figure ( [ fig : ths5a ] ) .
error bars for individual channels are @xmath31 error from the power law fit .
the structure function amplitude is a factor of @xmath76 higher for channels with stronger absorption .
this is consistent with the fact that the strong noise bias in these channels are expected to scale up the amplitude .
however , we note that the average value of the power law index @xmath77 ( @xmath31 interval shown in dotted lines ) is consistent with no channel to channel variation of the index .
this best fit value of the power law index is also consistent ( at better than @xmath31 uncertainty level ) with the power law index value of @xmath65 we derived earlier from the ensemble of structure functions from all the channels .
typically , the power spectra from the h i emission and absorption observations have a power law index @xmath78 at scales of @xmath6 au to a few pc @xcite . for 3c 138 , the derived power law index ( @xmath79 ) for the power spectrum is slightly shallower .
this flattening may arise from two reasons .
earlier it has been reported that , in the case of turbulence in a supernova remnant , a transition from 3-d to 2-d turbulence makes the power spectrum shallower @xcite .
so , instead of homogeneous random fluctuations , anisotropic turbulence at smaller scales can result in a shallower spectrum .
these fluctuations may be averaged out for large scale observations with an arcsec beam .
the other plausible reason may be small scale physical processes like viscous damping .
viscous damping for incompressible magnetized turbulence can make the magnetic energy spectrum significantly less steep @xcite . since ,
even with very small ionization fraction , h i is coupled to the galactic magnetic field , any possible correlation of h i density with the magnetic field strength will also flatten the opacity fluctuation power spectrum . in this regard , it will be interesting to check if simulation of compressible magnetohydrodynamic turbulence including damping and anisotropy at small scales can explain these observational results quantitatively .
our results also point out an amplitude mismatch between the observed structure function and that of the structure function extrapolated from large scale . the derived amplitude is more than an order of magnitude higher at tens of au scale for 3c 138 .
the observed higher amplitude at small scales may be due to few possible reasons .
first of all , for our analysis , we have used 3c 138 ( @xmath80 , @xmath81 ) as a background source , whereas @xcite used cas a ( @xmath82 , @xmath83 ) for deriving the structure function at pc scales .
these two lines of sight are widely separated , probing h i in different parts of the galaxy , where the small scale structures may , in principle , be intrinsically different .
alternatively , it may be an effect of the uncertainty in the measurement of structure function towards cas a , and extrapolation over a large angular scale giving rise to this mismatch . in this regard
, we find that extrapolating the shallower structure function for 3c 138 to 0.5 pc scale predicts the large scale structure function amplitude within @xmath84 of the observed value for cas a. in other words , the amplitude of the structure function at au scales is , as expected in this second possibility , completely consistent with the observed rms opacity fluctuations at the pc scale . finally , it is possible that there are physical mechanism(s ) which modify both the amplitude and slope of the structure function at au scales .
for example density fluctuations due to intermittency may in part explain this excess of tiny scale structures .
we note that , @xcite have reported milliparsec scale extreme velocity shears , which are explained as signature of intermittency .
further observations for more lines of sight , and measurements of structure function to bridge the gap between au and pc scales will be very helpful in distinguishing between these different possible scenarios .
we have used high resolution images of h i absorption towards 3c 138 from the combined vla , merlin and vlba data to study the small scale opacity fluctuations .
the fluctuations have a power law structure function with an index of @xmath3 over the scales of @xmath85 au .
this implies a power spectrum @xmath86 , slightly shallower than the earlier reported values of @xmath78 from observations over larger scales .
the amplitude of the structure function at the au scale is found to be significantly higher than the extrapolated amplitude from observations of cas a at pc scale .
the observed rms opacity difference for 3c 138 is about @xmath75 times higher than the prediction from this extrapolation from cas a. however , extrapolating the observed au scale structure function for 3c 138 consistently predicts the observed rms opacity fluctuations for cas a at the pc scale .
physical processes like velocity intermittency , viscous damping or anisotropy of turbulence may be related to this observational indication of a shallower slope and the presence of rich structures at smaller scales .
further measurements of structure function between au and pc scales , and detailed numerical simulations to quantitatively verify these results can improve the understanding of the interstellar turbulence .
we thank the anonymous referee for many useful comments which prompted us into improving this paper substantially .
we also thank john scalo , deputy editor of apj letters , for useful suggestions .
we are grateful to jayaram n. chengalur , avinash a. deshpande , prasun dutta and snezana stanimirovi for helpful discussions and useful comments on an earlier version of this manuscript .
part of this research was carried out at the jet propulsion laboratory , california institute of technology , under a contract with the national aeronautics and space administration .
begum a. , chengalur j. n. , bhardwaj s. , 2006 , mnras , 372 , l33 braun r. , kanekar n. , 2005 , a&a , 436 , l53 brogan c. l. , zauderer b. a. , lazio t. j. , goss w. m. , depree c. g. , faison m. d. , 2005 , aj , 130 , 698 cho j. , lazarian a. , vishniac e. t. , 2002 , apj , 566 , l49 crovisier j. , dickey j. m. , 1983 , a&a , 122 , 282 davis r. j. , diamond p. j. , goss w. m. , 1996 , mnras , 283 , 1105 deshpande a. a. , mcculloch p. m. , radhakrishnan v. , anantharamaiah k. r. , 1992 , mnras , 258 , 19p deshpande a. a. , 2000 , mnras , 317 , 199 deshpande a. a. , dwarakanath k. s. , goss w. m. , 2000 , apj , 543 , 227 dhawan v. , goss w. m. , rodrguez l. f. , 2000 , apj , 540 , 863 diamond p. j. , goss w. m. , romney j. d. , booth r. s. , kalberla p. m. w. , mebold u. , 1989 , apj , 347 , 302 dickey j. m. , mcclure - griffiths n. m. , stanimirovi s. , gaensler b. m. , green a. j. , 2001 , apj , 561 , 264 dieter n. h. , welch w. j. , romney j. d. , 1976 , apj , 206 , l113 dutta p. , begum a. , bharadwaj s. , chengalur j. n. , 2009a , mnras , 397 , l60 dutta p. , begum a. , bharadwaj s. , chengalur j. n. , 2009b , mnras , 398 , 887 elmegreen b. g. , kim s. , staveley - smith l. , 2001 , apj , 548 , 749 faison m. d. , goss w. m. , diamond p. j. , taylor g. b. , 1998 , aj , 116 , 2916 faison m. d. , goss w. m. , 2001 , aj , 121 , 2706 falgarone e. , pety j. , hily - blant p. , 2009 ,
a&a , 507 , 355 frail d. a. , weisberg j. m. , cordes j. m. , mathers c. , 1994 , apj , 436 , 144 green d. a. , 1993 , mnras , 262 , 327 hennebelle p. , audit e. , 2007 , a&a , 465 , 431 johnston s. , koribalski b. , wilson w. , walker m. , 2003 , mnras , 341 , 941 kim s. et al .
, 2007 , apjs , 171 , 419 lee l. c. , jokippii j. r. , 1975 , apj , 196 , 695 minter a. h. , balser d. s. , kartaltepe j. s. , 2005 , apj , 631 , 376 nagashima m. , inutsuka s .- i . , koyama h. , 2006 , apj , 652 , 41 roy n. , peedikakkandy l. , chengalur j. n. , 2008 , mnras , 387 , l18 roy n. , bharadwaj s. , dutta p. , chengalur j. n. , 2009 , mnras , 393 , l26 roy n. , chengalur j. n. , dutta p. , bharadwaj s. , 2010 , mnras , 404 , l45 stanimirovi s. , staveley - smith l. , dickey j. m. , sault r. j. , snowden s. l. , 1999 , mnras , 302 , 417 stanimirovi s. , lazarian a. , 2001 , apj , 551 , l53 stanimirovi s. , weisberg j. m. , hedden a. , devine k. e. , green j. t. , 2003 , apj , 598 , l23 stanimirovi s. , heiles c. , 2005 , apj , 631 , 371 stanimirovi s. , weisberg j. m. , pei z. , tuttle k. , green j. t. , 2010 , apj , 720 , 415 vzquez - semadeni e. , ryu d. , passot t. , gonzlez r. , gazol a. , 2006 , apj , 643 , 245 westpfahl d. j. , coleman p. h. , alexander j. , tongue t. , 1999 , aj , 117 , 868 | the structure function of opacity fluctuations is a useful statistical tool to study tiny scale structures of neutral hydrogen . here
we present high resolution observation of h i absorption towards 3c 138 , and estimate the structure function of opacity fluctuations from the combined vla , merlin and vlba data .
the angular scales probed in this work are @xmath0 milliarcsec ( about @xmath1 au ) .
the structure function in this range is found to be well represented by a power law @xmath2 with index @xmath3 corresponding to a power spectrum @xmath4 .
this is slightly shallower than the earlier reported power law index of @xmath5 at @xmath6 au to few pc scales .
the amplitude of the derived structure function is a factor of @xmath7 times higher than the extrapolated amplitude from observation of cas a at larger scales . on the other hand , extrapolating the au scale structure function for 3c 138 predicts the observed structure function for cas a at the pc scale correctly .
these results clearly establish that the atomic gas has significantly more structures in au scales than expected from earlier pc scale observations .
some plausible reasons are identified and discussed here to explain these results .
the observational evidence of a shallower slope and the presence of rich small scale structures may have implications for the current understanding of the interstellar turbulence . |
for a graph @xmath4 , the chromatic polynomial @xmath5 is an invariant which counts the number of proper colourings of @xmath4 when evaluated at a non - negative integer @xmath6 .
however , the chromatic polynomial has an additional interpretation as the zero - temperature antiferromagnetic potts model of statistical mechanics .
this has motivated research into the zeros of the chromatic polynomial by theoretical physicists as well as mathematicians .
traditionally , the focus from a graph theory perspective has been the positive integer roots , which correspond to the graph not being properly colourable with @xmath6 colours .
a growing body of work has begun to emerge in recent years more concerned with the behaviour of real or complex roots of the chromatic polynomial .
sokal @xcite proved that the set of roots of chromatic polynomials is dense in the complex plane .
in contrast , many other results show that certain regions are free from zeros . for planar graphs , the birkhoff
lewis theorem states that the interval @xmath7 is free from zeros . for more results along these lines , we direct the reader to the work of borgs @xcite , jackson @xcite , sokal @xcite , thomassen @xcite and woodall @xcite
. perhaps one of the most compelling open questions concerning real zeros is to determine tight bounds on the largest real zero of the chromatic polynomial .
one such bound is given in @xcite and depends on the maximum vertex degree .
for recent surveys see @xcite and @xcite . in matroids ,
the corresponding invariant is the characteristic polynomial .
characteristic polynomial _ of a loopless matroid @xmath8 , with ground set @xmath9 and rank function @xmath10 , is defined by @xmath11 where @xmath12 denotes the lattice of flats of @xmath8 and @xmath13 the mbius function of @xmath12 .
when @xmath8 has a loop , @xmath14 is defined to be zero .
observe that for a loopless matroid @xmath8 , @xmath14 is monic of degree @xmath15 and that @xmath8 and its simplification have the same characteristic polynomial .
the projective geometry of rank @xmath10 over @xmath2 is denoted by @xmath16 , and @xmath17 , where @xmath18 , denotes the uniform matroid with rank @xmath10 containing @xmath19 elements . in the uniform matroid , every set of @xmath10 or fewer elements is independent .
the characteristic polynomials of @xmath16 and @xmath17 play important roles in this paper , and these are easily computed . for a prime power @xmath0 , the projective geometry @xmath16 has lattice of flats isomorphic to the lattice of subspaces of the @xmath10-dimensional vector space over @xmath2 .
hence it has characteristic polynomial @xmath20 the largest root of the characteristic polynomial for a projective geometry is therefore @xmath21 .
the characteristic polynomial of the uniform matroid , @xmath17 , is @xmath22 for more background on matroid theory , we suggest that the reader consults @xcite . for the theory of the mbius function and the characteristic polynomial , we recommend @xcite .
perhaps the most compelling open question concerning real zeros in this context is deciding whether there is an upper bound for the real roots of the characteristic polynomial of any matroid belonging to a specified minor - closed class .
welsh conjectured that no cographic matroid has a characteristic polynomial with a root in @xmath23 .
this was recently disproved by haggard et al . in @xcite , and , in @xcite , jacobsen and salas showed that there are cographic matroids whose characteristic polynomials have roots exceeding five .
consequently , determining whether an upper bound exists for the roots of the characteristic polynomials of cographic matroids remains open . in @xcite
, royle conjectured that for any minor - closed class of @xmath2-representable matroids , not including all graphs , there is a bound on the largest real root of the characteristic polynomial .
given the situation with cographic matroids , this is clearly a difficult conjecture to resolve in the affirmative .
in contrast , the situation with graphic matroids has been resolved .
thomassen @xcite noted that by combining a result that he and woodall @xcite had obtained independently with a result of mader @xcite , one obtains the following .
let @xmath24 be a proper minor - closed family of graphs .
then there exists @xmath25 such that the chromatic polynomial of any loopless graph @xmath4 in @xmath24 has no root larger than @xmath26 . for certain minor - closed families of graphs ,
one can find the best possible constant @xmath26 .
one such example is the class of graphs with bounded tree - width , a concept originally introduced by robertson and seymour @xcite .
tree - decomposition _ of a graph @xmath4 comprises a tree @xmath27 and a collection @xmath28 of subsets of @xmath29 satisfying the following properties . 1 . for every edge @xmath30 of @xmath4 ,
there is a vertex @xmath31 of @xmath27 such that @xmath32 .
if @xmath33 and @xmath10 are distinct vertices in @xmath27 , the vertex @xmath34 is in @xmath35 and @xmath0 lies on the path from @xmath33 to @xmath10 in @xmath27 , then @xmath36 .
the width of a tree - decomposition is @xmath37 and the tree - width of a graph is the minimum width of all of its tree - decompositions . as its name suggests ,
graph tree - width measures how closely a graph resembles a tree .
matroid tree - width , which we will define later , measures how closely a matroid resembles a tree . if a graph can be obtained by gluing small graphs together in a tree - like structure , then it has small tree - width . likewise , if a matroid can be obtained by gluing small matroids together along a tree - like pattern , then it has small matroid tree - width .
thomassen @xcite proved the following .
[ graphy ] for positive integer @xmath1 , let @xmath4 be a graph with tree - width at most @xmath1 .
then the chromatic polynomial , @xmath38 , is identically zero or else @xmath39 for all @xmath40 .
thomassen s proof proceeded essentially as follows , using induction on the number of vertices of @xmath4 .
let @xmath4 have tree - width @xmath1 .
take a tree - decomposition of width @xmath1 , with notation as above .
choose @xmath41 and @xmath31 to be neighbouring vertices in @xmath27 .
then @xmath42 is a vertex - cut of @xmath4 .
one may add edges to @xmath4 with both end - vertices in @xmath42 until @xmath42 forms a clique without altering the tree - width .
call this new graph @xmath43 .
the chromatic polynomial of @xmath4 may be written in terms of the chromatic polynomial of graphs with fewer vertices than @xmath4 having tree - width at most @xmath1 and the chromatic polynomial of @xmath43 in such a way that one may apply induction provided the result can be established for @xmath43 .
but since @xmath43 has a clique whose vertices comprise a vertex - cut , the chromatic polynomial of @xmath43 may also be expressed in terms of the chromatic polynomials of graphs with fewer vertices and having tree - width at most @xmath1 . in this paper
, we generalize thomassen s useful technique to matroids .
the @xmath2-representable matroid analogue of a clique is a projective geometry over @xmath2 .
a given simple graph @xmath4 sits inside a clique on @xmath29 in the same way that a simple @xmath2-representable matroid @xmath8 with rank @xmath10 sits inside @xmath16 . in the above technique
, edges are added to an area " of @xmath4 to form a clique restriction , so that the altered graph has a clique vertex - cut .
this can be viewed as adding edges from the clique on @xmath29 to the graph @xmath4 to obtain a clique , across which our graph may be broken . in this paper , we show how to add elements from @xmath16 to a certain area " of @xmath8 in order to get a @xmath2-representable matroid with a certain projective geometry restriction , across which our matroid may be broken . the map that we use to break apart a matroid is a _ tree - decomposition _ , which was established by hlinn and whittle in @xcite .
they developed a matroid analogue of graph tree - width , which we define formally in section [ tree ] . in order to generalize thomassen s technique ,
we first develop some tools for @xmath2-representable matroids of bounded _ matroid tree - width_. we then apply these tools to extend his argument to matroid tree - width , which we shall refer to simply as _ tree - width _ when the context is clear . in this way , we demonstrate the utility of these tools and simultaneously make progress towards royle s conjecture .
an alternative way to prove theorem [ graphy ] is to combine the observation that every graph with tree - width at most @xmath1 has a vertex of degree at most @xmath1 with lemma [ oxco ] below , established by oxley for matroids and rediscovered for the special case of graphs by thomassen @xcite and woodall @xcite .
we show that this proof technique may also be extended to representable matroids .
in fact , this technique extends to a slightly more general class of matroids , namely matroids that exclude long line minors , which are considered in theorem [ nolines ] .
it was shown in @xcite that the tree - width of a matroid is at least equal to the tree - width of each of its minors , thus the class of matroids with tree - width at most @xmath1 is closed under taking minors .
the following result for such a minor - closed class is the main result of this paper .
[ maincourse ] for prime power @xmath0 and positive integer @xmath1 , let @xmath8 be a @xmath2-representable matroid with tree - width at most @xmath1 . then @xmath44 is identically zero or else @xmath45 for all @xmath46 . in the case that @xmath47 , theorem [ maincourse ] follows easily from known results , in particular equation .
however , this case is not especially interesting , because the rank of a matroid is always bounded below by its tree - width .
our result gives a new bound for representable matroids with high rank and low tree - width .
the requirement of representability is essential to the result .
for instance , the characteristic polynomial of the @xmath19-point line , @xmath48 , has a root at @xmath49 .
as @xmath48 has tree - width at most two , the @xmath19-point lines and their minors form a minor - closed class of matroids with bounded tree - width that do not have an upper bound for the roots of their characteristic polynomials . furthermore
, the projective geometry @xmath50 has tree - width @xmath1 and its characteristic polynomial has a root at @xmath3 , hence the bound given is the best possible .
lemma [ twprop ] contains the basic results on tree - width necessary to justify these observations .
given that the line @xmath48 is the simplest counter - example that we know of , it is natural to consider whether @xmath2-representability is necessary , or if excluding a long line minor from a matroid with bounded tree - width is sufficient to bound the roots of the characteristic polynomial , as suggested by geelen and nelson @xcite .
we show that this condition is indeed sufficient in the following theorem .
note that , if @xmath0 is a prime power , then @xmath51 is an excluded minor for matroids representable over @xmath2 .
thus the following theorem applies to a more general class of matroids than theorem [ maincourse ] applies to , although the bound is different .
[ nolines ] for an integer @xmath0 at least two , let @xmath8 be a matroid with tree - width at most @xmath1 and no minor isomorphic to @xmath51 . then @xmath44 is identically zero or else @xmath45 for all @xmath52 . combining theorem [ nolines ] with the observation that the characteristic polynomial of @xmath48 has a root at @xmath49 yields the following dichotomy .
[ dichotomy ] let @xmath53 be a minor - closed class of matroids having tree - width at most @xmath1 .
then either @xmath53 contains all simple matroids of rank two , or there exists @xmath54 such that for any loopless matroid @xmath8 in @xmath53 , @xmath45 for all @xmath55 .
the characteristic polynomial satisfies many identities similar to those satisfied by the chromatic polynomial .
the following is one such identity , which is particularly important for us .
[ condel ] if @xmath56 is an element of a matroid @xmath8 that is neither a loop nor a coloop , then the characteristic polynomial of @xmath8 satisfies @xmath57 from theorem [ condel ] , it is easy to see that a loopless matroid and its simplification have the same characteristic polynomial . the second identity which we will need is a special case of a result of brylawski @xcite .
we first define the generalized parallel connection of two matroids @xmath58 and @xmath59 with ground sets @xmath60 and @xmath61 , respectively , according to @xcite .
let @xmath62 and suppose that @xmath63 . furthermore ,
suppose that @xmath64 is a modular flat of @xmath58 and that each element of @xmath65 is either a loop or parallel to an element of @xmath27 .
let @xmath66 denote the common restriction @xmath67 .
then the _ generalized parallel connection across @xmath66 _ is the matroid @xmath68 whose flats are precisely the subsets @xmath69 of @xmath70 such that @xmath71 is a flat of @xmath58 and @xmath72 is a flat of @xmath59 .
suppose a graph @xmath4 has vertex set @xmath73 and edge set @xmath9 , where @xmath74 , such that @xmath75 and @xmath76 are themselves graphs .
it is a well - known result that , if the graph @xmath77 is isomorphic to @xmath78 , the complete graph on @xmath1 vertices , then the chromatic polynomial @xmath79 is equal to @xmath80
. we now state brylawski s result which generalizes this result to matroids .
[ bry ] let @xmath8 be a generalized parallel connection of the matroids @xmath58 and @xmath59 across the modular flat @xmath66
. then @xmath81
this section is devoted to defining matroid tree - width and developing some techniques for considering matroids of bounded tree - width .
a _ tree - decomposition _ of a matroid @xmath8 is a pair @xmath82 , where @xmath27 is a tree and @xmath83 is an arbitrary mapping . for convenience ,
let @xmath84 and let @xmath85 for all @xmath86 in @xmath87 .
we say that @xmath88 is the _ bag corresponding to @xmath89_. let @xmath90 be the number of components in @xmath91 and let @xmath92 denote the components in @xmath91 .
for @xmath93 , let @xmath94 be the subset of @xmath95 given by @xmath96 .
the vertex @xmath89 is said to _ display _ the subsets @xmath97 of @xmath98 .
note that these subsets are pairwise disjoint .
we say that the _ rank defect _ of @xmath94 , denoted @xmath99 , is equal to @xmath100 .
note that this number is the same as the size of the smallest set @xmath101 such that all of the elements in @xmath102 are in the closure of @xmath103 in the matroid @xmath104 .
clearly @xmath105 is an independent set in @xmath8 .
the rank defect is therefore a measure of the amount of rank contributed to @xmath8 solely by the set @xmath94 .
the _ node width of a vertex @xmath89 _ , written @xmath106 , is equal to @xmath107 .
note that in the degenerate case where @xmath108 , the node width of the single vertex of @xmath27 is equal to @xmath109 .
the _ width of @xmath110 _ is the maximum node width of all vertices in @xmath111 .
the _ matroid tree - width of @xmath8 _ , written @xmath112 , is equal to the minimum width of all tree - decompositions of @xmath8 .
we let @xmath113 be the number of vertices in the smallest tree over all of the tree - decompositions with width equal to the tree - width of @xmath8 . if @xmath82 is a tree - decomposition of @xmath8 with width equal to @xmath112 and if @xmath114 , then we say that @xmath82 is a _ good tree - decomposition _ of @xmath8 . .
each circle is labeled by the vertex of the tree that it represents .
the dots inside each circle represent the matroid elements that are in the bag corresponding to that vertex .
each circle is also labeled with the node width of its vertex . ]
we give a sample tree - decomposition of @xmath115 ( see figure [ tentacles ] ) . due to the symmetry of the matroid elements , it is not necessary to label the elements of the matroid .
we have illustrated the assignment of elements into bags by placing dots within circles .
each dot represents an element in @xmath115 and each circle represents a vertex of the tree in the tree - decomposition .
the vertices of the tree are labeled unambiguously by their names and their node widths .
each dashed region indicates a subtree of the tree @xmath27 and these subtrees comprise the connected components of the tree @xmath116 .
for example , the subtree @xmath117 consists of the vertex set @xmath118 and edge set @xmath119 . as a consequence of each dashed region indicating a connected component of @xmath116 , the matroid elements within the dashed regions are those of the subsets @xmath120 and @xmath121 of @xmath122 displayed by the vertex @xmath123 , where @xmath124 is the single - element bag associated with @xmath123 . to compute the node width for @xmath123 , note that @xmath125 and @xmath126 .
hence @xmath127 .
note that this is not an optimal tree - decomposition of @xmath115 .
for example , a tree - decomposition whose tree is a path where each bag contains exactly one matroid element has width six .
in addition to that used previously in this section , we employ an alternate use of the term `` display '' as follows .
let @xmath128 be an edge of @xmath27 , let @xmath129 and @xmath130 be the two components of @xmath131 containing @xmath132 and @xmath133 respectively , and let @xmath134 and @xmath135 be the sets of matroid elements @xmath136 and @xmath137 .
we say that the edge @xmath56 _ displays _ the sets @xmath134 and @xmath135 .
we now prove a lemma that will lend some structure to good tree - decompositions , which we establish in the following corollary .
[ red ] let @xmath82 be a tree - decomposition of a matroid @xmath8 .
suppose that @xmath27 has an edge @xmath128 that displays the sets @xmath138 , where @xmath139 .
then there exists another tree - decomposition @xmath140 of @xmath8 having width at most the width of @xmath82 , such that @xmath141 .
consider @xmath27 .
let @xmath142 , @xmath143 ,
, @xmath144 be the connected components of @xmath145 , where @xmath146 .
note that @xmath147 .
let @xmath148 be @xmath149 .
we define @xmath150 such that @xmath151 if @xmath152 and @xmath153 if @xmath154 .
clearly , @xmath141 .
take @xmath155 .
if @xmath156 then @xmath41 displays the same subsets of @xmath95 in @xmath140 and @xmath82 , so @xmath157 .
we conclude this proof by showing that @xmath158 . in the original tree - decomposition , @xmath110
, @xmath133 displays the subsets @xmath159 , @xmath160 ,
, @xmath161 , where @xmath162 .
note that @xmath163 . whereas in @xmath140 , @xmath133 displays the subsets @xmath160 , @xmath164 , , @xmath161 . since @xmath165
, we have @xmath166 .
it follows that @xmath158 , as required .
the next result follows immediately from lemma [ red ] . [ hall ]
let @xmath8 be a matroid with tree - width @xmath1 .
if @xmath82 is a good tree - decomposition of @xmath8 , then , for every pair of subsets @xmath134 and @xmath135 of @xmath95 displayed by an edge of @xmath27 , neither @xmath167 nor @xmath168 is equal to @xmath109 . for a good tree - decomposition of a matroid , the preceding result implies that every leaf in the tree corresponds to a set of elements in the matroid that , informally speaking , has some substance .
that is , the set is not in the closure of the rest of the elements in the matroid . in the corollary following the next lemma , we bound the rank of such a set of elements .
the following lemma is a collection of fundamental results for tree - width and tree decompositions .
the first was proved in @xcite .
[ twprop ] let @xmath8 be a matroid
. then 1 .
@xmath169 if @xmath66 is a minor of @xmath8 ; 2 .
@xmath170 , where equality holds if @xmath8 is a projective geometry ; and 3 . in any tree - decomposition with tree @xmath27 ,
the rank of a bag is at most the node width of the corresponding vertex , with equality holding at leaves of @xmath27 .
firstly , ( i ) was proved in @xcite . for ( ii ) ,
consider any tree decomposition @xmath82 of @xmath8 . by definition of node width
, no vertex of @xmath27 can have node width larger than @xmath109 , thus @xmath170 . in the case where @xmath8 is a projective geometry , to demonstrate that @xmath171
, it is sufficient to show that a good tree decomposition for @xmath8 has just one vertex . to that end , suppose that @xmath27 has an edge that displays @xmath172 and @xmath173
. then @xmath174 partitions @xmath95 .
however , for every bipartition of the elements of a projective geometry into sets @xmath134 and @xmath135 , either @xmath175 or @xmath176 , and it follows from corollary [ hall ] that a good tree decomposition for @xmath8 has just one vertex . to prove ( iii ) , using the notation set up in our definition of tree width , first note that since @xmath177 , ... ,
@xmath178 are a collection of pairwise disjoint subsets of @xmath95 , by submodularity of the rank function , @xmath179 by repeatedly applying submodularity , we see that @xmath180 comparing the rank of the bag of matroid elements @xmath181 associated with vertex @xmath89 to the node width of @xmath89 , we have @xmath182 as required . in the case where @xmath89 is a leaf of @xmath27 , equality holds since @xmath183 .
the next result follows immediately from lemma [ twprop ] .
[ octopus ] take @xmath8 with width-@xmath1 tree - decomposition @xmath82 . if @xmath34 is a vertex of @xmath27 , then the set @xmath184 has rank at most @xmath1 . in corollaries [ hall ] and [ octopus ] , we showed that a leaf in the tree of a good tree - decomposition corresponds to a set of elements that has some substance , but not too much substance .
we now find a small cocircuit in the matroid , when it is representable over a finite field .
[ sharktank ] let @xmath8 be a simple @xmath2-representable matroid for some prime power @xmath0 and let @xmath8 have tree - width @xmath1 for some positive integer @xmath1
. then @xmath8 has a cocircuit with at most @xmath3 elements .
let @xmath82 be a good tree - decomposition of @xmath8 .
in the case where @xmath185 , @xmath27 contains a leaf @xmath133 .
let @xmath186 . by lemma [ red ] , @xmath187
is not contained in the flat @xmath188 .
hence this flat is contained in a hyperplane of @xmath8 , whose complement is contained in @xmath187 .
evidently there is a cocircuit @xmath189 contained in @xmath187 .
corollary [ octopus ] implies that @xmath187 has rank at most @xmath1 . as @xmath8 is @xmath2-representable and simple , we know that @xmath187 is a restriction of @xmath50 .
the largest cocircuit in @xmath50 is obtained by deleting a hyperplane , which leaves @xmath3 elements .
hence @xmath190 . in the case
where @xmath191 , we have @xmath192 , thus @xmath8 is a restriction of @xmath50 .
with rank at least @xmath193 , @xmath8 contains a cocircuit , and by the same argument as above , @xmath8 contains a cocircuit @xmath189 with @xmath190 . during the remainder of this paper , for a simple @xmath2-representable matroid @xmath8
, we denote by @xmath194 the projective geometry @xmath195 of which @xmath8 is a spanning restriction .
if @xmath196 , then let @xmath197 denote the restriction of @xmath194 to the elements of @xmath198 .
take @xmath82 , a tree - decomposition of @xmath8 .
for edge @xmath199 in @xmath27 , let @xmath200 and @xmath201 be the subsets of @xmath95 displayed by @xmath199 , where @xmath202 .
let @xmath134 be the subset of elements of @xmath194 obtained by taking the closure @xmath203 , and likewise , let @xmath204 .
we say that the _ neck of @xmath199 with respect to @xmath194 _ , or simply the _ neck of @xmath199 _ when the projective geometry is clear , is the set of elements in @xmath205 .
note that the neck of each edge is a projective geometry over @xmath2 .
we say that the _ external neck of @xmath199 with respect to @xmath194 _ , or simply the _ external neck of @xmath199 _ is the intersection of the neck of @xmath199 with @xmath206 .
[ starfish1 ] let @xmath82 be a tree - decomposition of @xmath8 with width @xmath112 and let @xmath207 be a subset of the external neck of an edge of @xmath27
. then @xmath208 .
let @xmath199 be an edge of @xmath27 whose neck contains @xmath207 .
now @xmath197 has a tree - decomposition @xmath209 obtained from @xmath82 by letting @xmath210 when @xmath211 and by letting @xmath212 when @xmath213 .
thus the decomposition is the same except that we add the elements of @xmath207 to the bag corresponding to @xmath132 .
( we could equally well add them to the bag corresponding to @xmath133 . )
we show that , for each edge of @xmath27 , the corresponding subsets of @xmath95 and @xmath214 displayed by this edge have the same rank defects , and conclude that @xmath8 and @xmath197 have the same tree - width . by the definition of rank defect , if the elements of @xmath207 were added to a set @xmath215 , then @xmath216 .
hence the rank defect of @xmath215 in @xmath8 is equal to the rank defect of @xmath217 in @xmath197 . if the elements of @xmath207 were not added to a set @xmath215 that is displayed by an edge of @xmath27 , then @xmath207 is a subset of the closure of @xmath218 in @xmath197 by construction .
the rank defect again remains unchanged , as @xmath219 .
therefore each vertex in @xmath27 has the same node width in @xmath82 and @xmath220 .
it follows that @xmath221 .
[ starfish2 ] let @xmath8 be a simple @xmath2-representable matroid with tree - decomposition having tree @xmath27 .
let @xmath199 be an edge of @xmath27 and suppose that @xmath222 is the external neck of @xmath199 .
then , @xmath223 by construction , @xmath224 is neither a loop nor a coloop of @xmath225 .
furthermore , @xmath226 is neither a loop nor a coloop of @xmath227 for all @xmath228 . by theorem [ condel ] , @xmath229 . by repeated application of theorem [ condel ] , @xmath230{\vdots } } \\ & = \chi _ { m^{s_1}/s_1}(\lambda ) + \chi _ { m^{\{s_1,s_2\}}/s_2}(\lambda ) + \dots + \chi _ { m^s / s_n}(\lambda ) + \chi _ { m^s}(\lambda).\end{aligned}\ ] ] thus , the lemma holds .
in this section we prove the main theorem in two ways , illustrating different techniques each time .
the first proof requires us to consider separately the case where @xmath191 .
[ bullseye ] let @xmath8 be a loopless , @xmath2-representable matroid of tree - width @xmath1 , for some prime power @xmath0 and some positive integer @xmath1 , with @xmath191 .
suppose that , if @xmath66 is a loopless , @xmath2-representable matroid with tree - width at most @xmath1 and @xmath231 , then @xmath232 for all @xmath233 .
then @xmath234 for all @xmath233 .
we may assume that @xmath8 is simple .
let @xmath82 be a good tree - decomposition of @xmath8 .
the single vertex in @xmath111 must have node width @xmath1 . by lemma [ twprop ] , we have @xmath235 .
let @xmath207 be the set of elements in @xmath206 .
as in the proof of lemma [ starfish2 ] , by repeated application of theorem [ condel ] , @xmath236 . by assumption
, each term of this sum is positive for all @xmath237 with the possible exception of @xmath238 .
as @xmath197 is a projective geometry with rank @xmath239 , it follows that @xmath240 .
thus @xmath241 for all @xmath233 .
the first proof of the main theorem uses basic tools from characteristic polynomials , and exemplifies the tree - decomposition techniques established by hlinn and whittle in @xcite , and further developed in this paper , to generalize thomassen s graph technique .
if @xmath8 has a loop , then its characteristic polynomial is identically zero , so we may assume that @xmath8 is loopless . as @xmath8 and its simplification
have the same characteristic polynomial and the same tree - width , we may assume that @xmath8 is simple .
we proceed by induction on @xmath109 .
suppose that @xmath242 .
then @xmath243 and @xmath244 .
thus @xmath45 if @xmath245 , hence @xmath44 is certainly strictly positive for all @xmath233 .
we now assume @xmath246 .
take @xmath82 , a good tree - decomposition of @xmath8 .
if @xmath27 has a single vertex , then by lemma [ bullseye ] , the result follows .
thus , we may assume that @xmath27 contains a leaf @xmath133 with neighbour @xmath132 .
let @xmath247 be the elements in the external neck of @xmath199 .
by lemma [ starfish1 ] , @xmath208 and by lemma [ starfish2 ] @xmath248 since @xmath8 and , consequently , @xmath197 are simple , each of the matroids appearing in the sum on the right - hand side of is loopless and has rank @xmath249 .
lemma [ twprop ] implies that tree - width is not increased by contracting elements . by induction
the characteristic polynomial of @xmath250 is strictly positive for all @xmath233 , for all @xmath228 .
it remains to consider @xmath238 .
let @xmath251 be the neck of @xmath199 , which is contained in @xmath197 .
clearly @xmath252 for some @xmath253 .
let @xmath187 be the bag corresponding to @xmath133 .
let @xmath254 and @xmath255 .
then @xmath256 . by
* corollary 6.9.6 ) , @xmath251 is a modular flat in @xmath58 . by
* proposition 11.4.15 ) , @xmath197 is the generalized parallel connection of @xmath58 and @xmath59 across @xmath257 . since @xmath8 has tree - width at most @xmath1 , we know that @xmath258 with the last part following from corollary [ octopus ]
. thus , by theorem [ bry ] , @xmath259 using equation , we see that the denominator is strictly positive for all @xmath260 .
hence it is strictly positive for all @xmath233 . by corollary [ hall ] , since @xmath27 has @xmath261 vertices , both @xmath58 and @xmath59 have rank less than @xmath262 . by our inductive hypothesis ,
both @xmath263 and @xmath264 are strictly positive for all @xmath233 .
thus @xmath265 for all @xmath233 , as required .
it is also possible to generalize the second proof of theorem [ graphy ] , outlined in the introduction , to matroids representable over a finite field by using the following result of oxley ( * ? ? ?
* lemma 2.7 ) .
[ oxco ] let @xmath266 be a cocircuit of @xmath8 .
let @xmath267 for all @xmath268 .
then @xmath269 a minor - closed family of matroids @xmath270 has the _ bounded cocircuit property _ if there is a constant @xmath271 such that any simple matroid @xmath8 in @xmath270 has a cocircuit of size at most @xmath272 .
we now apply lemma [ oxco ] to any minor - closed family of matroids with the bounded cocircuit property .
[ le : bdedcoc ] let @xmath53 be a minor - closed family of matroids having the bounded cocircuit property with constant @xmath272 . then for any @xmath8 in @xmath53
, either @xmath273 is identically zero or @xmath274 for all @xmath275 .
let @xmath8 be a matroid in @xmath53 .
we may assume that @xmath8 is simple and that the result is valid if @xmath242 .
we now assume @xmath246 and proceed using induction on @xmath109 . because @xmath53 has the bounded cocircuit property , we know that @xmath8 has a cocircuit @xmath189 with size at most @xmath272 .
let @xmath276 and let @xmath267 for @xmath277 . by lemma [ oxco ] , @xmath44 is equal to the following @xmath278 now @xmath279 and @xmath280 , for all @xmath277 . by induction ,
each of the characteristic polynomials appearing in is either identically zero or strictly positive for @xmath275 .
furthermore @xmath281 is loopless and so @xmath282 for @xmath275 . as @xmath283 ,
we conclude that @xmath284 , and hence @xmath44 , is strictly positive for all @xmath285 .
we now give the alternate proof of theorem [ maincourse ] .
let @xmath53 be the family of @xmath2-representable matroids with tree - width at most @xmath1 .
lemma [ twprop ] implies that @xmath53 is a minor - closed class and lemma [ sharktank ] implies that @xmath53 has the bounded cocircuit property with constant @xmath3 .
the result now follows from lemma [ le : bdedcoc ] .
the argument in the second proof of theorem [ maincourse ] may be extended to any family of matroids with the bounded cocircuit property .
we show the family of matroids with tree - width at most @xmath1 containing no @xmath51 minor is one such family by using the following theorem of kung @xcite .
[ nolineskung ] let @xmath0 be an integer at least two . if @xmath8 is a simple matroid with rank @xmath10 having no @xmath51-minor , then @xmath286 .
let @xmath287 be the largest prime less than or equal to @xmath0 .
when @xmath10 is sufficiently large , geelen and nelson showed that the bound on the number of elements can be obtained by replacing @xmath0 with @xmath287 in the preceding theorem . in @xcite , nelson conjectured that this improvement holds as long as @xmath288 .
if nelson s conjecture holds , then the bound given in theorem [ nolines ] can be improved by replacing @xmath0 with @xmath287 in the case that @xmath289 .
we now prove theorem [ nolines ] , in which we replace the representability condition of theorem [ maincourse ] with the condition that @xmath8 contain no long line minor .
this generalization was suggested by geelen and nelson @xcite .
we claim that if @xmath8 is simple , has tree - width at most @xmath1 and has no @xmath51 minor , then it has a cocircuit of size at most @xmath290 .
the result then follows by noting that the class of matroids with tree - width at most @xmath1 having no @xmath291 minor is a minor - closed class and applying lemma [ le : bdedcoc ] . once again , we may assume that @xmath8 is simple and that the claim is valid if @xmath242 .
we proceed by induction on @xmath109 .
take @xmath82 , a good tree - decomposition of @xmath8 and vertex @xmath292 with degree at most one .
let @xmath293 be the bag corresponding to @xmath34 and let @xmath294 .
lemma [ twprop](iii ) implies that @xmath295 . if @xmath27 consists of a single vertex , then @xmath296 .
furthermore @xmath293 contains a cocircuit since @xmath297 .
suppose then that @xmath27 contains more than one vertex .
then @xmath34 is a leaf vertex .
since @xmath82 is a good tree - decomposition , @xmath293 is not contained in @xmath298 by lemma [ red ] .
thus @xmath293 contains a cocircuit of @xmath8 .
theorem [ nolineskung ] implies that @xmath299 .
thus @xmath8 has a cocircuit @xmath189 of size at most @xmath290 . in corollary
[ dichotomy ] , we completely determine whether there is a bound on the largest real root of the characteristic polynomial of any matroid belonging to a minor - closed family having bounded tree - width .
it would be interesting to find minor - closed classes of matroids that do not have the bounded cocircuit property and determine bounds on the real characteristic roots .
we thank the anonymous referees , jim geelen and peter nelson for several helpful comments .
brylawski , t. , and oxley , j. the tutte polynomial and its applications . _ matroid applications _ , encyclopedia of mathematics and its applications * 40 * , 123225
. cambridge university press , cambridge ( 1992 ) .
royle , g. f. recent results on chromatic and flow roots of graphs and matroids .
_ surveys in combinatorics _ , london mathematical society lecture note series * 365 * , 289327 .
cambridge university press , cambridge ( 2009 ) .
zaslawsky , t. , the mbius function and the characteristic polynomial . _ combinatorial geometries _ , encyclopedia of mathematics and its applications * 29 * , 114138 .
cambridge university press , cambridge ( 1990 ) . | we develop some basic tools to work with representable matroids of bounded tree - width and use them to prove that , for any prime power @xmath0 and constant @xmath1 , the characteristic polynomial of any loopless , @xmath2-representable matroid with tree - width @xmath1 has no real zero greater than @xmath3 .
2016 this manuscript version is made available under the cc - by - nc - nd 4.0 license ` http://creativecommons.org/licenses/by-nc-nd/4.0/ ` ` http://dx.doi.org/10.1016/j.ejc.2016.08.011 ` |
the _ abelian complexity _ of a word @xmath10 is a function @xmath11 that counts the number of pairwise non - abelian - equivalent factors of @xmath10 of length @xmath3 .
the term was introduced by richomme , saari and zamboni @xcite in 2009 , and since then it has been extensively studied @xcite . in one of the first papers on the subject , richomme , saari and zamboni @xcite examined the tribonacci word @xmath12 http://oeis.org/a080843 [ ] , which is the fixed point of the substitution @xmath13 , @xmath14 , @xmath15 , and they showed that @xmath16 for all @xmath3 .
they also characterized those @xmath3 for which @xmath17 , and proposed the following open problem : _ for each @xmath18 , characterize those @xmath3 for which @xmath19_. explicit characterization of @xmath20 of a given infinite word @xmath10 is generally a difficult task , particularly in case of words defined over alphabets consisting of more than two letters .
for example , despite the fact that a recurrent word over a ternary alphabet with constant abelian complexity equal to @xmath21 for all @xmath22 has been already constructed @xcite , there seems to be no other nontrivial example to date of a recurrent @xmath23-ary word with @xmath24 whose abelian complexity function has been precisely determined .
in particular , the problem of precise characterization of the abelian complexity @xmath25 of the tribonacci word @xmath12 , which is a ternary word , has remained open since 2009 .
recently mousavi and shallit @xcite showed that many properties of the tribonacci word , such as the aperiodicity , powers , palindromes etc .
, could be examined purely mechanically with the help of finite automata .
although in principle their method could also be used for the study of the characteristics of the abelian complexity function of the tribonacci word , it turns out to be not computationally feasible . in this paper
we propose a related method that is particularly designed for dealing with abelian properties ( primarily with abelian complexity and the balance properties ) .
our approach is less general than the one of mousavi and shallit , but it is efficient enough to explicitly obtain a finite automaton that computes the function @xmath25 . the automaton in question features a very small set of states , consisting of less than @xmath26 elements .
consequently , it can be easily implemented on any computer ( a powerful machine is not needed ) , and the calculation can be even performed by hand . the automaton takes the tribonacci representation of @xmath3 as its input , which means that the sequence @xmath27 http://oeis.org/a216190 [ ] is @xmath28-automatic ( or `` tribonacci - automatic '' ) in the sense of shallit @xcite .
our approach relies on the technique of abelian co - decomposition @xcite , which was originally developed as a tool for proving that the abelian complexity of a recurrent word attains a certain value infinitely often . as a result
, our construction of the automaton can be generalized to certain other words as well .
the paper is organized as follows . in section [ sect.preliminaries ]
we provide the necessary notation related to abelian complexity , the tribonacci word and finite automata .
section [ sect.codecomposition ] summarizes basic facts about abelian co - decomposition .
section [ sect.z ] contains the core of the paper : we show that there exists a finite number of sets @xmath29 such that each @xmath22 can be associated with a certain @xmath30 via the tribonacci representation of @xmath3 .
since each of those sets is related to a certain value of the abelian complexity , the indices @xmath31 can be regarded as states of a finite automaton that evaluates @xmath25 .
section [ sect.range ] is devoted to the first and easy application of the findings from section [ sect.z ] : we demonstrate that the abelian complexity of the tribonacci word takes values in @xmath32 .
although this fact is already known @xcite , it illustrates the applicability of our approach .
the main result of the paper is presented in section [ sect.eval.ac ] .
we derive a formula for evaluating the abelian complexity of the tribonacci word on the basis of results of section [ sect.z ] .
in particular , we show that the abelian complexity can be calculated by a finite automaton with @xmath33 states .
this result is further improved in section [ sect.simplified ] , where the size of the automaton is reduced from @xmath33 to @xmath34 states .
it is easy to transform this automaton into an automaton that decides , for any @xmath22 , whether @xmath25 attains a given value @xmath35 , which provides a solution of the problem of richomme et al . in section [ sect.mbon ] , we demonstrate that the method allows one to examine the abelian complexity function of other @xmath23-bonacci words . in particular , we present results on the @xmath9-bonacci word ; they show that the abelian complexity of @xmath23-bonacci words gains new properties when @xmath23 exceeds @xmath21 .
the paper is concluded by section [ sect.conclusions ] , in which we discuss other applications and generalizations of the method .
let us consider a set @xmath36 ( _ alphabet _ ) consisting of @xmath23 symbols ( _ letters _ ) @xmath37 .
concatenations of letters from @xmath38 are called _
words_. let @xmath39 denote the free monoid of all finite words over @xmath38 including the empty word @xmath40 .
the _ length _ of a word @xmath41 is the number of its letters , @xmath42 .
the symbol @xmath43 for @xmath44 and @xmath45 stands for the number of occurrences of the letter @xmath46 in the word @xmath47 .
the set of all infinite words over @xmath38 is denoted by @xmath48 .
we say that an infinite word @xmath10 is _ recurrent _ if every factor of @xmath10 occurs infinitely many times in @xmath10 .
a finite word @xmath47 is a _ factor _ of a ( finite or infinite ) word @xmath10 if there exists a finite word @xmath49 and a ( finite or infinite , respectively ) word @xmath50 such that @xmath51 .
if @xmath52 , the factor @xmath47 is called a _ prefix _ of @xmath10 . for any word @xmath45 and @xmath53
we write @xmath54 . similarly , we set @xmath55 .
if a word @xmath56 has the prefix @xmath57 for @xmath53 , then the symbol @xmath58 stands for the word satisfying @xmath59 .
the _ parikh vector _ of @xmath47 is the @xmath23-tuple @xmath60 ; note that @xmath61 . for
any given infinite word @xmath10 , let @xmath62 denote the set of all parikh vectors corresponding to factors of @xmath10 having the length @xmath3 , i.e. , @xmath63 the _ abelian complexity _ of a word @xmath10 is the function @xmath64 defined as @xmath65 where @xmath66 denotes the cardinality .
it is useful to introduce the _ relative parikh vector _
@xcite , which is defined for any factor @xmath47 of @xmath10 of length @xmath3 as @xmath67})\,,\ ] ] where @xmath68}$ ] is the prefix of @xmath10 of length @xmath3 .
since the subtrahend @xmath69})$ ] depends only on the length of @xmath47 and not on @xmath47 itself , the set of relative parikh vectors corresponding to the length @xmath3 , @xmath70 has the same cardinality as @xmath71 .
hence we obtain , with regard to , @xmath72 an infinite word @xmath10 is said to be _
@xmath73-balanced _ for a certain @xmath74 if for every @xmath44 and for every pair of factors @xmath75 , @xmath47 of @xmath10 such that @xmath76 , the inequality @xmath77 holds .
if @xmath10 is a @xmath73-balanced word , the components of relative parikh vectors are bounded by @xmath73 @xcite .
therefore , the set of all relative parikh vectors @xmath78 is finite for any @xmath73-balanced word @xmath10 .
this paper is primarily concerned with the _ tribonacci word _
@xmath12 , which is defined over the alphabet @xmath79 as the fixed point of the substitution @xmath80 i.e. , @xmath81 it is easy to check that @xmath82 for every @xmath83 .
hence , the lengths of factors @xmath84 satisfy the recurrence relation @xmath85 . comparing this relation with the tribonacci recurrence relation @xmath86
, we conclude that @xmath87 for every @xmath88 , where @xmath89 is the sequence of tribonacci numbers http://oeis.org/a000073 [ ] .
any @xmath22 can be written as a sum of tribonacci numbers with binary coefficients , @xmath90 if coefficients @xmath91 are obtained by the greedy algorithm , they form the _ normal @xmath28-representation _ ( also called the _ tribonacci representation _ ) of @xmath3 , which we denote by the symbol @xmath92 : @xmath93 for @xmath94 , we have @xmath95 . table [ t expansions ] shows normal @xmath28-representations of several small integers .
.normal @xmath28-representations of the numbers @xmath96 .
[ cols="^,>,^,>,^,>,^,>,^,>",options="header " , ] let us summarize the results obtained on the @xmath9-bonacci word . the abelian complexity function of the @xmath9-bonacci word has the following properties .
* @xmath97 .
* @xmath98 if and only if @xmath99 is a prefix of @xmath100 . * @xmath101 if and only if @xmath102 .
* for every value @xmath103 there exist infinitely many integers @xmath22 such that @xmath104 .
we are convinced that the existence of gaps in the range of the abelian complexity function , as well as the existence of values that are attained only finitely many times , are common properties of all @xmath23-bonacci words with @xmath105 .
we finish the section by a remark on the minimal value of the abelian complexity function of the @xmath23-bonacci word for a general @xmath23 .
let @xmath10 be an @xmath23-bonacci word . one can easily show that for every @xmath22 and @xmath106 , the factor @xmath107}$ ] is a factor of @xmath10 , thus @xmath108 . at the same time we have @xmath109 . to sum up , @xmath110 for any @xmath23-bonacci word .
results of k. binda s calculations @xcite together with proven formulas and suggest a conjecture on a precise characterization of the numbers for which the abelian complexity function of an @xmath23-bonacci word @xmath10 attains its minimum : @xmath111 the symbol @xmath112 stands here for the @xmath23-bonacci representation of @xmath3 .
we are able to prove the implication @xmath113 in for a general @xmath23 by the abelian co - decomposition method , introduced earlier @xcite .
the implication @xmath114 remains so far open , although we expect that it is probably not difficult to be proven either .
in this paper we focused on the abelian complexity of the tribonacci word ( or , more generally , @xmath23-bonacci words ) , but the method can be easily adapted for application on any simple parry word .
let us consider the fixed point @xmath10 of a substitution .
the calculation naturally takes advantage of the numeration system associated with @xmath115 , i.e. , of the normal @xmath116-representation for @xmath117 .
let us briefly sketch the procedure .
first of all , for the sake of generality , we slightly reformulate the definition of @xmath118 by imposing an additional technical assumption on the decomposition .
namely , we assume that for every @xmath119 , the factor @xmath120 has the prefix @xmath121 , and require that @xmath122 is maximal subject to this condition .
we also need to introduce maps @xmath123 , defined in a way analogous to equations , i.e. , @xmath124 the search for sets @xmath29 starts with calculating @xmath125 by formula for every @xmath126 .
then we take the bunch of sets @xmath127 , which we denote by @xmath128 , and apply the maps @xmath123 on each of the sets , similarly as in the proof of theorem [ subsets ] . in this way
we obtain a new bunch of sets , we apply @xmath123 on them again , and so forth
. however , unlike in the case of @xmath23-bonacci words , one needs to keep track of the admissibility of the normal @xmath116-representations @xmath112 during the calculation .
briefly speaking , if a normal @xmath116-representation , examined at a given moment , can not be validly extended by a certain specific digit @xmath129 , then the map @xmath130 is not applied at that stage .
the procedure ends when the application of @xmath123 generates no new data .
the abelian co - decomposition method also allows us to deal with the other type of parry words , called non - simple parry words , which are fixed points of substitutions @xmath131 where @xmath132 satisfy @xmath133 , @xmath134 for all @xmath44 , and @xmath135 .
although we have not explicitly discussed non - simple parry words in previous sections , the implementation of the procedure would be the same ; we just need to replace the value @xmath23 in the definition of @xmath136 by @xmath137 . to sum up
, the approach is applicable on any parry word ; but one shall keep in mind that in practice it will more likely work well in cases when the image of the abelian complexity function is a set of low cardinality .
nevertheless , it can still give new results for various words for which other methods fail .
those parry words , for which this approach turns out to be inefficient , can be perhaps treated by a newer technique , which replaces pairs @xmath138 by certain conveniently chosen triples @xcite .
that technique is more involved , but it is expected to have smaller memory requirements in most cases and to work faster . a potentially interesting question is whether this approach ( possibly after a certain improvement ) can be used for dealing with a word that depends on a parameter , i.e. , whether one can explore a parametric family of words as a whole .
consider for instance the @xmath23-bonacci word for a general @xmath139 .
we are convinced that the procedure could be implemented with a parameter as well , although the calculation would be of course intricate and lengthy .
the procedure also gives , as a by - product , the optimal balance bound of the examined word .
the optimal bound is equal to the maximal entries of vectors @xmath140 having the form @xmath141 for @xmath142 belonging to the same set @xmath143 .
for example , one can check in this way that the @xmath9-bonacci word is @xmath21-balanced .
consequently , we can regard the method not only as a tool for evaluating abelian complexity , but also as a tool for exploring balance properties of words . in particular , it is possible that this approach can lead to the optimal balance bound for the @xmath23-bonacci word for any @xmath23 .
recall that the optimal bound for the @xmath23-bonacci word is not known yet , despite the fact that an upper bound has been already determined @xcite .
the author thanks j .- p .
allouche for useful comments and suggestions , and the referees for valuable hints and corrections that helped to improve the manuscript . | according to a result of richomme , saari and zamboni , the abelian complexity of the tribonacci word satisfies @xmath0 for each @xmath1 . in this paper
we derive an automaton that evaluates the function @xmath2 explicitly .
the automaton takes the tribonacci representation of @xmath3 as its input ; therefore , @xmath4 is an automatic sequence in a generalized sense .
since our evaluation of @xmath2 uses @xmath5 operations , it is fast even for large values of @xmath3 .
our result also leads to a solution of an open problem proposed by richomme et al . concerning the characterization of those @xmath3 for which @xmath6 with @xmath7 belonging to @xmath8 .
in addition , we apply the same approach on the @xmath9-bonacci word . in this way we find a description of the abelian complexity of the @xmath9-bonacci word , too .
[ theorem]corollary [ theorem]lemma [ theorem]proposition [ theorem]definition [ theorem]example [ theorem]conjecture [ theorem]remark ondej turek + nuclear physics institute + academy of sciences of the czech republic + 250 68 e + czech republic + and + bogolyubov laboratory of theoretical physics + joint institute for nuclear research + 141980 dubna + russia + [email protected] + .2 in p |
according to the hierarchical scenario of structure formation , massive clusters form by merging small groups continuously and accreting field galaxies along the filament @xcite .
optical cluster surveys reveal that many galaxy clusters have evidence for dynamically bound substructures @xcite . a significant fraction ( @xmath640%-50% ) of clusters
show multiple peaks or irregular surface brightness distribution in the x - ray images , indicating that they are still at dynamically active stage , far from equilibrium @xcite . compared with the einstein - de sitter case ,
clusters in early - epoch universe are expected to be more relaxed and less substructured , as supported by many n - body simulation works@xcite .
the fraction of substructured clusters at different redshifts is thus a useful statistical quantity directly relevant to cosmology .
studies on the dynamics of galaxy clusters thus provide a unique tool to put constraint on the models of cluster formation and evolution .
dense environment in galaxy clusters should have produced influence on physical properties and evolutionary path for the member galaxies .
previous studies have found that the observational properties of galaxies correlate strongly with local galaxy environment @xcite .
one of the most well - studied relation in galaxy clusters is the morphology - density relation @xcite .
the core region of a cluster is usually dominated by early - type galaxies ( etgs ) , while the outer region is dominated by late - type galaxies ( ltgs ) .
it is well appreciated that the ltgs gradually lost their gas reservoirs when they were accreted into the core region , and finally evolved into lenticular galaxies ( s0 ) .
this picture of morphology evolution in galaxy clusters has been supported in high - z morphology - density relation studies @xcite .
however , it is still uncertain and controversial how local galaxy environment affects star formation histories of cluster galaxies .
the beijing - arizona - taiwan - connecticut ( batc ) system has spent much time in observing a sample of more than 30 nearby ( @xmath7 ) galaxy clusters at different dynamic statuses , aiming at studying their dynamic substructures , luminosity functions , and the star formation properties of cluster galaxies .
abell 0671 ( a671 ; z=0.0502 ) is one target of the batc galaxy cluster survey .
its abell richness @xmath8 is set to be 0 @xcite , with bautz - mogan type ii - iii @xcite .
the x - ray emission from the cluster center has been detected by the _
observatory and the _ rosat _ all - sky survey ( rass ) .
the x - ray luminosity of a671 detected in the rass 0.1 - 2.4 kev band is @xmath9 @xmath10 , and the x - ray temperature is 3.1 kev @xcite , which confirms that this cluster is a relatively poor system .
figure 1 shows the smoothed contours of the _ einstein _ x - ray image and the radio map at 1.4 ghz from the nrao vla sky survey ( nvss ) , superimposed on the optical image in the batc@xmath11 band .
no radio emission is detected at the center of a671 .
the x - ray surface density contour is quite regular with a single symmetric peak as reported by @xcite .
however , the detail structure of a671 is possibly blurred due to the low resolution ( @xmath6 1 arcmin spatial resolution ) and the large psf ( fwhm @xmath61.5 armin ) of einstein ipc image .
it is easily seen that the x - ray emission peak does not coincide with the central brightest galaxy , ic 2378 , with a positional offset of about 90 kpc .
a671 is included in the cluster sample of the wide - field nearby galaxy - cluster survey ( wings ) , and two substructures in a671 have been found by @xcite recently . for a better understanding of the dynamics of a671 , it is important to construct a large sample of member galaxies , and the faint galaxies ( @xmath12 ) should be taken into account . in this paper , we present a multicolor photometry of the galaxies in a671 region with the beijing - arizona - taiwan - connecticut system .
we try to enlarge the sample of cluster galaxies by applying the photometric redshift technique to the spectral energy distributions ( seds ) of the batc - detected faint galaxies . based on the sdss spectra of bright member galaxies
, we will derive the star formation histories and chemical abundances , and try to find any clues of environmental effects on the physical parameters of cluster galaxies .
image ( @xmath13 kev band ) ( red line ) and the nvss map at 1.4 ghz ( green line ) , superimposed on the batc@xmath11 band image .
the sizes of gaussian smoothing windows are adopted as @xmath14 and @xmath15 for radio and x - ray contours , respectively.,width=321 ] this paper is organized as follows : we present the batc photometric observations and data reduction in section2 . in section3 , we analyze the galaxies with known spectroscopic redshifts in the a671 field . in section4 , we apply photometric redshift technique to select faint member galaxies in a671 . in section5 , dynamic substructures and luminosity function
are investigated based on the enlarged sample of member galaxies . in section6
, we derive the star formation histories and chemical abundances of the etgs in a671 .
finally , we summarize our work in section7 . throughout this paper , we assume the cosmological parameters as @xmath16 , @xmath17 , @xmath18 mpc@xmath19 .
the batc survey is based on the 60/90 @xmath20 @xmath21 schmidt telescope of national astronomical observatories , chinese academy of science ( naoc ) , located at xinglong station .
the batc system contains 15 intermediate - band filters , covering a wavelength from 3000 to 10000 , which are designed to avoid night sky emission lines @xcite .
the transmission curves of batc filters can be found in @xcite . before october 2006
, a ford ccd camera with a format of @xmath22 was mounted on the telescope , and photometric observations in 12 bands , from @xmath23 to @xmath24 , were carried out .
the viewing field was about @xmath25 , with a scale of @xmath26pixel . for pursuing better spatial resolution and higher sensitivity in three blue bands , @xmath27 , a new e2v ccd with @xmath28 pixels
was then equipped .
the field of view becomes larger ( @xmath29 ) with a spatial scale of @xmath30pixel .
the newly equipped ccd camera has a high quantum efficiency of 92.2% . from march 2003 to october 2007
, we totally accumulated 50 hours exposure for a671 with 15 filters ( see the observational information in table 1 ) . with an automatic data - processing software , pipeline i @xcite ,
we carry out the standard procedures of bias subtraction , flat - field correction , and position calibration .
the technique of integral pixel shifting has been used in the image combination during which cosmic rays and bad pixels are removed by comparing multiple images . [ cols="^,^,^,^,^,^,^,^,^ " , ] for detecting and measuring the flux of the sources within a given aperture in the batc images , we convert the @xmath27 combined images to make the pixel size identical with @xmath31 images .
we adopt a radius of 4 pixels as a photometric aperture for all the batc images to the sources detected by the _ sextractor _
codes @xcite . the flux calibration of seds
is performed by using the oke - gunn @xcite standard stars ( hd19445 , hd84937 , bd+26d2606 and bd+17d4708 ) which were observed during photometric nights .
the detailed information about calibration can be found in @xcite .
because we have no calibration images for @xmath32 , @xmath33 , @xmath34 filters , we instead perform the model calibration that has been developed specially for the large - field photometric system by @xcite . as a result
, the seds of 6782 sources have been obtained in our catalog . by cross - identifying the batc sources with the sdss photometric data within a search circle ( defined as a circle of 1.5 arcsec ) ,
all sources are classified into galaxies and stars . as a result ,
985 galaxies brighter than the batc@xmath11 band magnitude limit are found by both surveys , which offers a sample for further analysis .
for studying the dynamics of galaxy cluster a671 , 205 galaxies with known redshifts in our viewing field are extracted from the sdss dr8 galaxy catalog .
figure 2 shows the distribution of spectroscopic redshifts of these galaxies .
the main concentration with a peak at @xmath35 is isolated and less contaminated .
there are 103 galaxies with @xmath36 and they are selected as the member galaxies of a671 , to which we refer as sample i. to characterize the velocity distribution , we convert the spectroscopic redshifts ( ) into the rest - frame velocities ( @xmath37 ) by @xmath38 , where @xmath34 is the light speed , @xmath39 is the cluster redshift with respect to the cosmic background radiation .
we take the ned - given cluster redshift @xmath39 = 0.0502 for a671 .
the velocity distribution can be fitted by a gaussian with a dispersion of @xmath40 , and it is significantly deviated from a standard gaussian function . to qualify the distribution of radial velocities of member galaxies , we use the rostat software @xcite to calculate two resistant and robust estimators , biweight location ( @xmath41 ) and scale ( @xmath42 ) , analogous to the velocity mean and the standard deviation . for these 103 galaxies ,
we achieve @xmath43 , @xmath44 .
the errors are determined from the 68% confidence limits based on 10,000 bootstrp resamplings of the velocity data .
taking a cosmological correction factor of @xmath45 into account , the velocity dispersion of a671 should be @xmath46 .
@xcite studied a sample of 88 nearby clusters .
only 72 galaxies were included in their sample , and a smaller @xmath42 was derived .
they found a671 having @xmath47 and @xmath48 .
our statistics is based on a larger sample , thus is more reliable .
both studies confirm that a671 has a comparatively small velocity dispersion . .
the galaxies in a671 are centered at @xmath35 .
we select the galaxies with @xmath36 as member galaxies .
the smaller panel shows the distribution of rest - frame velocities of member galaxies in detail , and dash line represents a gaussian fit to the histogram .
, width=321 ] because a671 is a nearby cluster , our batc viewing field can not cover whole cluster region .
our photometry focus on a central field of @xmath49 mpc@xmath50 .
the radius of galaxy cluster , , is defined in former studies as the boundary of a cluster , within which the mean inner density is @xmath51 , where @xmath52 is the critical density of the universe @xcite .
we calculate the for a671 following the formula suggested by @xcite .
the is a function of velocity dispersion . by applying the @xmath42 we derived ,
the of a671 is 2.25 mpc , corresponding a slightly larger area than our viewing field . before studying the dynamic structure of a671
, we try to classify the 103 known members into etgs and ltgs .
the early - type members should meet two requirements : ( 1)with no evident emission lines;(2)with no evident galaxy arms .
we firstly extract those galaxies with ew()@xmath53 as early - type candidates .
the ew ( ) values are taken from the mpa / jhu catalog of sdss galaxies .
then we inspect their sdss - given images and removing those with arms . as a result
, the 103 galaxies are classified into 63 etgs and 40 ltgs .
the left panel of figure 3 presents the spatial distribution of 103 known member galaxies within our field of view , with the central position of a671 , r.a.=@xmath54 , dec.=@xmath55 ( for the j2000 equinox ) .
we superpose the contour map of surface density that has been smoothed by a gaussian window with @xmath56 .
as shown in figure 3 , the member galaxies are mainly concentrated in the central region within a radius of 1 mpc .
the irregular contour in the east and north corresponds to the two substructures found by @xcite .
more statistical tests should be performed before we can reach a firm conclusion that there are substructures in a671 . to show the substructures of a671 in both velocity space and projected map , we make use of the @xmath57-test @xcite for the 103 galaxies .
the statistic variable , @xmath58 , is defined to quantify the local deviation on the scale of groups of @xmath59 nearest neighbors .
a larger @xmath58 indicates a greater probability that the local velocity distribution differs from overall velocity distribution .
the probability @xmath60 can be calculated by monte carlo simulations with random shuffling velocities .
when the scale of the nearest neighbors @xmath59 varies from 3 to 9 , the probabilities @xmath60 are nearly zero , which means the substructure appears very obvious at different scales .
the bubble plot at the scale of @xmath61 is given in the right panel of figure 3 . since the bubble size is proportional to @xmath62 ,
the clustering of large bubbles is a good tracer of dynamical substructure . as we can see in the bubble plot , the central region of a671
is dominated by two clumps of bubbles , and a clump of large bubbles in the north - east is also remarkable .
we refer to these three clumps as a , b , and c , which contain 16 , 10 , 8 galaxies , respectively . for confirming whether these clumps trace the real substructures , we present the velocity distributions for subsamples a , b , and c in figure 4 , as well as the velocity distribution of the whole sample .
it is easily seen that the velocity distributions for the three clumps are indeed deviated from that of the whole sample .
the mean velocities of the subsamples a , b , and c are @xmath63 , @xmath64 , and @xmath65 , respectively .
the mean velocities of subsamples b and c are significantly deviated from the mean velocity of the whole sample , 14561 .
though the mean velocity in clump a is similar to that of the whole sample , the galaxies in clump a have a remarkable bimodal velocity distribution , which may imply the existence of two different groups . to quantify the bimodality significance
, we apply the gaussian mixture modeling ( gmm ) method @xcite to the velocity distribution for clump a. the result shows that clump a consists of two components , which peaked at 13829 and 15824 .
the unimodal velocity distribution can be rejected at 99% significance level .
although the velocity distribution of the subsamples are significantly deviated from the whole sample , their masses should also be large enough if they are real substructures .
the masses of the a671 and its clumps can be estimated by applying the virial theorem .
assuming that each subcluster is bound and the galaxy orbits are random , the virial mass ( @xmath66 ) can be derived from the following standard formula @xcite : @xmath67 where @xmath68 is the light - of - sight velocity dispersion , @xmath69 is the cosmological distance of the cluster , @xmath70 is the number of galaxy pairs , and @xmath71 is the angular distance between galaxies @xmath72 and @xmath73 .
the viral masses of a671 is @xmath74 .
clump b is the most remarkable among the 3 subsamples , with viral mass of @xmath75 .
the a clump consists of two groups , with viral mass of @xmath6 @xmath76 for each .
clump c has similar viral mass .
thus we conclude that a671 is not a simple relaxed cluster , but most likely at a dynamically active stage .
the technique of photometric redshift can be used to estimate the redshifts of galaxies by using the sed information covering a wide range of wavelength instead of spectroscopy .
this technique has been extensively applied to the multicolor photometric surveys for detecting the faint and distant galaxies , and for subsequential selection of cluster galaxies @xcite .
based on the standard sed - fitting code called @xmath77 @xcite , for a given object , the photometric redshift , , corresponds to the best fit ( in the @xmath78 -sense ) of its photometric sed with the template sed generated by convolving the galaxy spectra in the template library with the transmission curves of specified filters .
previous work has demonstrated the accuracy of photometric redshift with the batc multicolor photometric data @xcite . in our sed
fitting , only normal galaxies are taken into account in the reference templates .
dust extinction with a reddening law of the milky way @xcite is adopted , and @xmath79 is allowed to be flexible in a range from 0.0 to 0.5 , with a step of 0.05 .
the photometric redshift of a given galaxy is searched from 0.0 to 0.6 , with a step of 0.005 .
we apply this technique to all batc galaxies brighter than @xmath80 .
the procedure of sed - fitting has provided the best - fit photometric redshift and its uncertainty for each galaxy . from the 205 galaxies with known spectroscopic redshifts ( ) , we select 167 galaxies ( including 91 member galaxies ) that are simultaneously detected in at least 12 batc bands to derive their values .
a comparison between and values is shown in the top - left panel of figure 5 .
the error bar of corresponds to 68% confidence level in photometric redshift determination .
the solid line denotes @xmath81 , and the dashed lines show an average redshift deviation of @xmath82 .
it is obvious that our estimates are basically consistent with their values . for 91 member galaxies in our spectroscopic sample ,
the mean value and standard deviation of their values are 0.0505 and 0.0112 , respectively .
there is no systematic offset in the domain with respective to the distribution .
statistically , 85 member galaxies ( about 93 percent ) are found to have their photometric redshifts within @xmath83 deviation , in a range from 0.028 to 0.073 , demonstrating the robustness of our estimate .
this region can be applied as a selection criterion in the following membership determination for faint galaxies .
the top - right panel of figure 5 shows the histogram of photometric redshifts for all galaxies down to @xmath80 .
the black histogram shows distribution for the 167 galaxies with known spectroscopic redshifts mentioned above .
as expected , the peak in the distribution is around z=0.05 . in the bottom - left panel of figure 5
, we show the uncertainties for these 167 galaxies as a function of batc@xmath11 band magnitude .
it is remarkable that the deviation of fainter galaxies tends to be larger . for the faint galaxies with @xmath84 ,
our estimate is still robust , but with larger uncertainty . for the remainder galaxies without values
, we give a plot of their uncertainties versus batc@xmath11 band magnitudes in the bottom - right panel in figure 5 .
the larger uncertainties for faint galaxies are mainly due to larger magnitude errors in photometry . for a reliable membership determination based on the estimate
, we exclude the galaxies fainter than @xmath80 , and take the galaxies with @xmath85 as member candidates . due to the robustness of our photometric redshift technique , it is conservative that our selecting criterion would be able to select 80 - 90% of faint members with least contaminants .
it is well known that there exits a correlation between color and absolute magnitude for the etgs ( c - m relation ) @xcite , in the sense that brighter etgs appear redder , which can be used for verifying the membership selection of the etgs .
the left panel of figure 6 presents the correlation between the color index , @xmath86 , and batc@xmath11 band magnitude for the whole member candidates down to @xmath80 , while the right panel shows the sdss color index , @xmath87 , and @xmath88band magnitude .
the diagrams include the following categories of sources : ( 1 ) spectroscopically confirmed early - type member candidates ( denoted by filled circles ) , ( 2 ) spectroscopically confirmed late - type member candidates ( denoted by filled triangles ) , ( 3 ) newly - selected faint member candidates ( denoted by open triangles ) .
the solid line represents the linear fitting with the 63 spectroscopically confirmed etgs : @xmath89 , and the dashed line represents @xmath90 deviation .
the linear fit of the etgs in the right panel of figure 6 is @xmath91 .
as shown in both figures , the early - type member galaxies follow a very tight color - magnitude relation , faint member candidates also follow the same c - m relation basically , but seem to be more scattered .
this might be caused by some high-@xmath92 galaxies that have been mixed into our faint member candidates .
for excluding these contaminants , we utilize the sdss c - m relation , and remove the etg candidates with color indices @xmath87 0.15 mag redder than the red sequence denoted by the black solid line . finally , we obtain a list of 97 newly - selected member galaxies
. combined with those 103 spectroscopically confirmed members , we form an enlarged sample of 200 galaxies in a671 , to which we refer as sample ii in following investigation .
the left panel of figure 7 shows the projected positions of the galaxies in sample ii , superposed with the contour map of surface density smoothed by a gaussian window with @xmath93 .
the 103 member galaxies with known spectroscopic redshifts are denoted by filled symbols , and 97 photometrically selected galaxies are denoted by open symbols .
red symbols represent the early - type member galaxies and blue symbols represent late - type ones .
our batc multicolor photometry facilitates the finding of large number of faint member galaxies , and makes the underlying substructure along north - east direction more remarkable .
basically , the distribution of faint galaxies traces that of bright ones , and no significant substructures are found with sample ii . the morphology segregation becomes more remarkable in sample ii . both bright and faint etgs
are highly concentrated in the core region , while the ltgs are scattered in the outskirts .
the shape of contour map seems in accord with the x - ray image , which might demonstrate the reliability of our membership selection based on the batc multicolor photometry . for detecting the potential substructures in a671
, we perform the @xmath57-test for sample ii .
the right panel of figure 7 shows the bubble plot that characterizes the degree of difference between the localized velocity distribution , for groups of six nearest neighbors , and the overall velocity distribution .
we performed @xmath94 simulations to estimate probability @xmath60 for different group sizes .
the probability is found to be more than 5% in all cases , which means that no substructure is detected at @xmath95 significance ( see table 2 ) .
this is not consistent with the conclusion that we have achieved based on sample i. we think that the substructure unveiled by the spectroscopic redshifts is surely true . above inconsistency can be well interpreted that the velocities derived from the estimates are not accurate enough to reflect the subtle velocity structure .
the abnormity in velocity distribution of substructures might have been smoothed / swept by the uncertainties of 97 newly selected faint galaxies .
so the @xmath57-test on sample ii might be misleading .
follow - up spectroscopy of these faint member galaxies are needed if one wants to investigate the substructures in a671 in detail .
ccc neighbors & @xmath60 & @xmath60 + size ( @xmath59 ) & sample i & sample ii + 3 & 1.4% & 34.7% + 4 & 0.6% & 35.9% + 5 & 2.8% & 65.3% + 6 & 1.0% & 74.8% + 7 & 0.5% & 71.5% + 8 & 0.7% & 65.7% + 9 & 1.4% & 62.0% + the luminosity function ( lf ) is a key diagnostic for clusters because it is tightly related to dynamical evolution and merging history of galaxy clusters .
the lf has been widely studied in the past decades and it is well described by the schechter function @xcite : @xmath96 where @xmath97 , @xmath98 , and @xmath4 are the normalization parameter , the characteristic luminosity , and the slope parameter at faint end , respectively . in the domain of absolute magnitude , the schechter function can be expressed as : @xmath99dm,\ ] ] where @xmath100 is the characteristic absolute magnitude .
mpc region around a671 .
contour levels are from 1@xmath101 to 5@xmath101 , spaced 1@xmath101 apart .
top right : the region which is used to estimate the local background .
dots denote the galaxies with the sdss r - band magnitude brighter than 21.0 mag .
empty regions represent the removed overdensity area .
bottom left : galaxy number counts extracted within r=1.7 mpc region ( filled circles ) , and those in the local background ( open circles ) for a671 .
bottom right : background - subtracted number counts for a671 .
the solid line represents the best - fit schechter function .
, width=302 ] the most challenging issue in measuring the lfs of galaxy clusters ( gcs ) is that one needs to pick up cluster members out of background galaxies along the line of sight .
ideally , one needs the spectroscopic redshifts for all galaxies to exclude non - members in the cluster field .
unfortunately spectroscopic measurements are rather time - consuming .
our batc photometry enables us to select a set of cluster member candidates by utilizing the photometric redshift technique . as seen in figure 5 ,
the accuracy of is a function of galaxy apparent magnitude .
further corrections are still needed to remove the contribution of contaminant sources and compensate the missing members when investigating the lfs .
unfortunately the exact form of the correction function , particularly at the faint magnitudes , is very difficult to be derived . using the sdss r - band photometric data
, we perform the statistical background subtraction to estimate the contribution of non - members to the number counts of galaxies in the cluster direction , by measuring the projected number counts of field galaxies outside the cluster region .
the background is estimated with a @xmath102 mpc region centered on the cluster centroid , outside the cluster region defined by a radius of 3 mpc , where the contamination from cluster galaxies should be negligible . following the method of @xcite
, we firstly generate a density map of galaxies in the background region by convolving the projected distribution of galaxies with a gaussian kernel of @xmath101=250 kpc in the cluster rest frame ( the typical size of a cluster core ) .
then we mask out all density peaks which are above 3@xmath101 level from the background region .
masked - out regions covered about 3.3% of the whole background area .
finally we calculate the number counts from the remain galaxies to estimate the background number counts in the cluster direction .
recent studies have provided evidences that the lfs of gcs do vary with clustercentric radius @xcite .
a suitable region should be chosen which is large enough to contain most member galaxies and do not include much projected contamination .
we adopt a aperture of r=30centered at the cluster centroid ( about 1.7 mpc at the rest frame of a671 ) .
the results of the background estimation and final luminosity function are showed in figure 8 .
the apparent magnitudes are converted to absolute magnitudes by the relation @xmath103 where @xmath104 is the distance modulus as determined from the redshift assuming a particular cosmology .
@xmath105 is the @xmath106-correction from a galaxy at @xmath92 to @xmath107 .
we estimate the @xmath105 using the software ` kcorrect`(version @xmath108 , @xcite ) .
as showed in the bottom - right panel of figure 8 , a single schechter function can fit the data very well .
the best fit parameters are @xmath97=21.0 , @xmath100=-21.6 , @xmath4=-1.12 , which is in good agreement with @xcite .
no strong `` upturn '' at faint magnitudes is observed for a671 .
as mentioned in the first section , many former studies support that galaxy properties strongly correlate with local environment .
the remarkable morphology - segregation of a671 demonstrates that the morphologies of member galaxies strongly correlate with their local environment . in this section
we would like to investigate the star formation histories ( sfhs ) of the confirmed member galaxies by applying models of stellar population synthesis on their observed spectra .
two models will be taken in this section .
on the one hand , we fit the sdss spectra by the starlight code @xcite to derive the physical parameters based on their sfhs . on the other hand
, we compare the absorption line indices ( the lick / ids indices ) of the member etgs with the model predictions developed by @xcite ( hereafter s07 ) to put constraints on the sfhs and chemical enrichment .
we fit the sdss spectra of member galaxies with the starlight codes , which aims at fitting an observed spectra with a liner combination of theoretical simple stellar populations ( ssps ) .
the model spectrum is given by @xmath109 where @xmath110 is the model spectrum , @xmath111 is the synthesis flux at the normalization wavelength @xmath112 , @xmath113 is the so - called population vector , @xmath114 is the @xmath73th ssp spectrum at @xmath115 , and @xmath116 represents the reddening term .
the @xmath117 is the line - of - sight stellar motions that modelled by a gaussian distribution centered at velocity @xmath118 and with a dispersion of @xmath119 .
@xmath120 is the total number of ssp models . in our work ,
the ssp base is made up of @xmath121 ssps , three metallicities ( @xmath122 , @xmath123 , @xmath124 ) and 15 ages ( from 1 myr to 13 gyr ) , which are taken from evolutionary models in @xcite .
the galactic extinction law of @xcite with @xmath125 is adopted .
all sdss observed spectra are shifted to the rest - frame , and then interpolated into a resolution of 1 @xmath126 before fitting .
wavelength regions of emission lines are masked out .
figure 9 shows the spectral fitting for the brightest cluster galaxy ( bcg ) of a671 .
as demonstrated by this figure , the combination of ssp spectra can fit the observed spectrum very well .
the starlight presents the ssp fraction , intrinsic extinction @xmath79 , velocity dispersion @xmath101 , and stellar mass @xmath127 .
following @xcite , we derive the flux- and mass - weighted average ages , which are defined as @xmath128 where @xmath129 is the flux - weighted population vector ( i.e. , the fraction of flux contributed by certain ssp ) , and @xmath130 is the mass - weighted population vector .
the average metallicities @xmath131z@xmath132@xmath133 and @xmath131z@xmath134@xmath133 can be derived similarly .
as illustrated by @xcite , the individual output vector may be dramatically deviated from the simulated input value . however , the average values of stellar age and metallicity should be more reliable , whatever they are weighted by light or mass .
the flux - weighted age is more sensitive to the young stellar component , so the mass - weighted age is more underlying and intrinsic .
the situation is the same for average metallicity . in this section
, we will take the average ages and metallicities weighted by stellar mass . since stellar population in etgs are dominated by the old components , the etgs have their average ages within a relatively narrow range .
figure 10 presents the derived mass - weighted ages and metallicities as functions of cluster - centric radius r and total stellar mass ( @xmath135 ) assembled in cluster galaxies .
the upper two panels show the mass - weighted ages and metallicities as a function of r. the galaxies in the core region ( r@xmath131400kpc ) of a671 are denoted by filled symbols .
the remarkable morphology - density segregation of a671 is well shown in these two panels .
the ltgs ( denote by blue symbols ) locate in the outskirts ( denoted by open symbols ) , and have younger stellar ages . on the other hand ,
the etgs with older stellar ages are located in the core region . for the etgs of a671
, no correlation is found between metallicities and r. the lower two panels of figure 10 present how the mass - weighted ages and metallicities correlate with stellar mass .
both ages and metallicities are found to be correlated strongly with stellar mass , and such correlations are dependent upon morphology .
in general , the more massive galaxies have older ages and richer metallicities . for the ltgs in a671 , the linear correlations of age and metallicity with stellar mass appear tighter and steeper . even for the etgs with similar stellar mass ,
the etgs in core region tend to have older ages than the outskirt etgs .
however , the etg metallicities seem to not vary with cluster - centric radius .
the age - metallicity degeneracy has haunted stellar population analysis for decades .
nevertheless , the promising approach to break it remains the combined use of multiple absorption - line indices@xcite . in this section , we will measure the absorption lines of etgs in a671 and compare them with state - of - art ssp models in order to infer their ages , metallicities and @xmath4-enhancements .
the bandpasses of the lick indices are defined in table 1 of @xcite .
we measure the lick indices with a modified version of the lick@xmath136ew routine in ez@xmath136ages package developed by @xcite.[multiblock footnote omitted ] the lick@xmath136ew routine reports the errors of each lick index calculated in the way suggested by @xcite .
the sdss spectral resolution ( 69 ) does not match the originally defined resolution of lick indices .
the lick@xmath136ew routine smooths the sdss spectra to the resolution of lick / ids system before measuring the indices .
for the galaxies with high velocity dispersion , the smoothed absorption features in the sdss spectra are at resolution even poorer than the lick resolution , the lick@xmath136ew routine will apply @xmath101-correction for these galaxies .
the output includes the measurements of @xmath101-corrected indices and their errors .
one of the most challenging issues in lick index measurements is emission line contamination .
the balmer absorption features are contaminated by emission from ionized gas , either from star formation , agn activity or interstellar shocks .
balmer line emission was estimated from equivalent of , which are retrieved from the mpa / jhu sdss dr7 catalog . some of previous studies used ew()@xmath137 ew(@xmath1155007 ) for correction @xcite .
we do not use @xmath1155007 for correction because @xcite have found that the relation between and @xmath1155007 for the etgs varies in different mass ranges .
they found that the correlation between and is much tighter .
emission ews for higher order balmer lines are obtained from by assuming standard values from balmer decrement ( in the absence of reddening ) . in this way , ew()@xmath138 ew ( ) , ew()@xmath139 ew ( ) , and ew()@xmath140 ew ( ) .
our goal is to use the s07 model to derive the ssp - equivalent parameters .
firstly we create grid to fit three parameters : ages , [ fe / h ] , and [ mg / fe ] .
we use the solar - scaled isochrones and the salpeter initial mass function suggested by the ez@xmath136ages documents .
the [ o / fe ] is set to be zero , and other @xmath4 elements are tied to mg .
the ages range from 1.2 to 17.7 gyr , and the [ fe / h ] ranges from -1.3 to 0.2 .
other details can be found in s07 , @xcite .
the three parameters are derived following two steps : i ) for each galaxy , we firstly calculate age and [ fe / h ] using and @xmath131fe@xmath133 , at all [ mg / fe ] ; ii ) we compare the pair @xmath131fe@xmath133 ( @xmath131fe@xmath133=0.5(fe5270+fe5335 ) ) and mgb with models at the median age obtained in the first step , and calculate a new [ fe / h ] and [ mg / fe ] .
we then update the ages by interpolating the ages found in the first step at the [ mg / fe ] derived in the second step , and iterate the second step with the new age .
usually two iteration steps are needed before convergence . for those measurements beyond the model grids
, we set the parameters to be the boundaries , i.e , the maximum or minimum of the models . before deriving the stellar population parameters of our sample with the ssp models ,
we compare our measurements to the predictions of each model on the grids of age and metallicity . in figure 11 ,
our measurements of and [ mgfe]@xmath5 are compared with the s07 models assuming [ @xmath4/fe]=0.2 , where the index [ mgfe]@xmath5 is defined as follows : @xmath141^{'}$]=@xmath142 . ]
@xmath5 is a good indicator of metallicity which is almost independent upon @xmath4/fe ration variations @xcite ( hereafter tmb ) . to convert between [ fe / h ] and [ z / h ]
, we adopt the relation given by tmb : [ z / h]@xmath143[fe / h]@xmath1440.94[@xmath4/fe ] , and we assume [ mg / fe]@xmath143[@xmath4/fe ] .
the median age is about 7 gyr but with a large scatter .
most of the core - region galaxies have very small values , and occupy the oldest end of age distribution .
the [ mgfe]@xmath5 is mainly distributed along [ z / h]@xmath145 , and does not have any environmental effects .
fe]=0.2 , shows the @xmath146mgfe@xmath147\arcmin$ ] vs. .
the filled and open circles represent the etgs in the core region and those in outskirts , respectively.,width=302 ] we present the derived six ssp parameters as a function of velocity dispersion ( @xmath101 ) in figure 12 .
galaxies which fall outside model boundary are not included in the figure .
the parameter errors are given by the fitting procedure for each galaxy .
we then compute an average error for each parameter by weighted the output error of individual galaxy with its signal - to - noise ratio .
as presented in figure 11 , majority of the galaxies falling out of the model boundary are located in the core , which makes hard to investigate environmental effect . though only 46 etgs are able to be fitted by the model , the relations between the ssp parameters and velocity dispersion ( @xmath101 ) are still remarkable . in the top - left panel of figure 12 ,
the ssp ages show strong dependence upon velocity dispersion , and the low-@xmath101 galaxies span a wider age range , indicating that the low @xmath101 galaxies have various possibilities of star formation histories compared with the high-@xmath101 ones .
similar results are found in the bright galaxies in coma by @xcite .
it is noteworthy that the ssp ages derived by s07 model are not compatible with the average stellar ages given by the starlight fitting .
firstly , the ssp ages in starlight code range from 1 myr to 13 gyr , while the range of the ssp ages in s07 models is from 1 gyr to 17.7 gyr .
the typical age of etgs from starlight fitting ( see figure 10 ) is older than 7 gyr . in figure 11 , the etg ages derived by the s07 model span a wider range .
additionally , the starlight gives the best - fit ages without errors .
the s07 model determines the ages by comparing several combined - indices with the theoretical model , and the measurement errors can strongly affect the output ages . for the galaxies
whose indices locate near the model boundary , measurement errors of indices will bring in greater uncertainties in the parameter fitting .
the top - middle and top - right panels display the relation of [ fe / h ] and [ mg / fe ] with @xmath101 .
the [ fe / h]@xmath148 correlation is very tight , with a correlation coefficient of @xmath149 , whereas the [ mg / fe]@xmath148 correlation is rather weak .
the former studies have found [ mg / fe ] strongly correlated with @xmath101 @xcite .
we should keep in mind that the tightness and slope of linear correlation strongly depend on the sample size . considering our small sample size and the intrinsic scatter of this correlation ,
a relatively weaker [ mg / fe]@xmath148 relation is still reasonable . in the three bottom panels of figure 12 , we present [ c / fe]@xmath148 , [ n / fe]@xmath148 , and [ ca / fe]@xmath148 relations
. only a weak correlation of [ n / fe]@xmath148 is found , with a correlation coefficient of @xmath150 .
the results of our fitting are similar to those in @xcite , and their results are based on a large sample of about 6000 red sequence galaxies from the sdss .
we have investigated the dynamics of a671 based on the spectroscopically confirmed members , and the result of @xmath57-test strongly suggests that a671 has significant substructures .
many authors have found that a large fraction of galaxy clusters have substructures @xcite , indicating that massive clusters may assemble their masses and grow up by accreting small groups .
the contour map of member galaxies in a671 fits well with the x - ray intensity contour map .
@xcite suggested that there are two substructures in a671 .
location of one substructure is well associated with the potential substructure b , and the other is at the south part of cluster , which is not significant enough to be detected by the @xmath57-test .
their substructure - finding algorithm is based on the projected positions of galaxies , and does not utilize the redshift information .
73% of the clusters in their sample were found to have substructures , and this fraction is higher than most studies .
their magnitude limit of galaxy samples is @xmath151@xmath152 , much fainter than our limiting magnitude @xmath80 , thus the projection effect can not be ignored , and some substructures they found could be untrue . after 97 newly selected galaxies included , the substructure b is enhanced , while another substructure appears less prominent .
the follow - up spectroscopy of these faint galaxies is need for revealing the details of dynamical substructures in a671 .
the remarkable morphology - density relation indicates that the cluster environment indeed have played an important role in evolution of cluster galaxies .
a visual inspection of the c - m diagram shows that a large fraction of bright member galaxies ( @xmath153 ) have evolved to be the `` red sequence '' , and faint member galaxies with @xmath154 are found to have considerable activities of star formation .
when the galaxies are accreted into a cluster , their star formation activities are expected to be suppressed by some important processes , such as tidal stripping and `` harassment '' @xcite , ram pressure stripping of the gas disk @xcite , and removal of gas reservoir surrounding each galaxy @xcite .
we derived the average ages and metallicities for member galaxies by fitting their spectra .
the most remarkable feature of the age distribution is that the etgs in the core region have older ages than those in the outskirts . @xcite
have found that the etgs in dense environment have average ages @xmath155 gyr older than those in field environment .
they derived ages by comparing the lick indices with the prediction of tmb model .
our results confirm their conclusion though the average ages are derived by different methods .
this can be interpreted in two ways . on the one hand
, theoretical work shows that dark matter halos in dense environments were assembled earlier than average @xcite . as a result , the galaxies in the core region of clusters formed earlier than those in the outer regions , and thus have older stellar ages . on the other hand , the older stellar age of the etgs in the core region
could be explained by lack of recent star formation compared with those in the outer region .
galaxy clusters have dense gas with high temperature , and the core - region galaxies lost their gas reservoirs by interacting with the dense intracluster medium ( icm ) ( usually by ram pressure stripping ) . thus the core - region galaxies should be gas - poor , and have less possibilities of recent star - formation . aside from age , the total metallicities of the etgs show subtle dependence on environment , which agrees well with @xcite . a larger sample of cluster etgs is needed to investigate the environmental effects on chemical evolution .
the dependence of age upon stellar mass could be explainable in the downsizing scenario for galaxy formation @xcite . in this scenario ,
star formation lasts longer in less massive galaxies than in more massive galaxies .
thus massive systems will have older ssp ages on average .
evidences of downsizing effect in the local universe have been found by many recent studies on the stellar population of etgs @xcite .
the age-@xmath101 slope in these studies span a range between 0.35 and 0.93 .
the different slope values are mainly due to the different sample property ( e.g. , galaxy type , sample size , etc . ) and methods of index correction ( e.g. , emission infill correction , velocity dispersion correction , etc . ) . @xcite
have tested the roubustness of their observed age-@xmath101 slope against these two factors when they studied the stellar population in coma . in their test , a stricter emission - line cut and different methods of velocity dispersion correction were used . only slight change on the age-@xmath101 slope was found in their tests .
they concluded that their data robustly support the downsizing scenario .
the etgs also have very strong dependence of @xmath4-abundance on @xmath101 .
the @xmath4 elements are mainly from the type ii supernovae , and the iron - peak elements come mainly from the type ia supernovae . the stronger alpha - enhancement in the more massive elliptical galaxies may imply that their star formation timescale is shorter than less massive elliptical galaxies , before the delayed type ia supernovae enrich the star - forming regions with iron - peak elements .
the observed @xmath4-abundance@xmath148 relation fits well with the prediction from the hierarchical models with feedback @xcite .
this paper presents a photometric study of a671 with the beijing - arizona - taiwan - connecticut ( batc ) multicolor system and the sdss data .
the main conclusions can be summarized as follow : \(i ) about 7000 sources are detected in a batc field of @xmath25 centered at a671 , and their seds in 15 intermediate bands are obtained . the 985 galaxies brighter @xmath156 are selected by cross - identifying our batc source catalog with the released catalog of sdss galaxies .
there are 205 galaxies with known spectroscopic redshifts in our viewing field , among which 103 galaxies with @xmath1570.06 are selected as spectroscopically confirmed members of a671 .
the sample of bright member galaxies is composed of 63 etgs and 40 ltgs .
\(ii ) the dynamics of a671 is investigated based on the 103 spectroscopically confirmed members .
the result of @xmath57-test on different scales strongly suggest that a671 have significant substructures .
three potential substructures have been suggested with the method of localized deviation of velocity distribution .
\(iii ) photometric redshift technique is applied to the 985 galaxies for further membership determination .
our photometric redshifts ( ) of the bright members are basically consistent with the spectroscopic redshifts ( ) .
base on the statistics of photometric redshifts , the galaxies with @xmath85 are selected as member candidates .
after further selection by the color - magnitude relation , 97 galaxies down to @xmath80 are picked up as faint members of a671 .
\(iv ) based on the enlarged sample of member galaxies , spatial distribution and velocity structure of a671 are studied .
since the large uncertainty of faint galaxies have smoothed the localized abnormity in velocity distribution , the @xmath57-test of the enlarged sample does not confirm the three substructures mentioned above .
the morphology - segregation becomes very remarkable after the faint members are taken into account .
the luminosity function in the sdss r - band shows a flat slope at faint end , @xmath158 .
\(v ) mass - weighted stellar ages and total metallicities of bright members are derived by fitting their spectra with the spectral synthesis code , starlight .
the etgs in the core region have older ages than those in the outskirts .
the more massive etgs are found to be older than the less massive ones .
no environmental effect is found for the metallicities of the etgs .
strong correlations of mean stellar age and metallicity with stellar mass are confirmed , and such correlations are found to be dependent upon morphology .
the possitive age - mass correlation supports the downsizing scenario .
\(vi ) a set of lick indices of the etgs is measured in order to derived their ssp - equivalent stellar parameters ( such as age , [ fe / h ] , [ mg / fe ] , [ c / fe ] , [ n / fe ] , and [ ca / fe ] ) by utilizing s07 model .
the etgs at cluster center tend to have smaller indices , indicating that central etgs are likely to be older .
the total metallicity indicator [ mgfe]@xmath5 does not show any environmental effects . the relations between the six ssp - parameters and velocity dispersion ( @xmath101 ) are also studied .
the relations between the ssp - parameters and @xmath101 in a671 are in good agreement with previous studies .
this work was funded by the national natural science foundation of china ( nsfc ) ( grant nos . 10803007 , 10873012 , 10873016 and 11173016 ) , the national basic research program of china ( 973 program ) ( grant nos .
2007cb815403 , 2007cb815404 ) , and the chinese universities scientific fund ( cusf ) .
this research has made use of the ned , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
funding for the sloan digital sky survey(sdss ) has been provided by the alfred p.sloan foundation , the participating institutions , the national aeronautics and space administration , the national sience foundation , the us department of energy , the japanese monbukagakusho and the max - planck society .
the sdss web site is http://www.sdss.org .
the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , los alamos national laboratory , the max - planck institute for astronomy(mpia ) , new mexico state university , university of pittsburgh , university of portsmouth , princeton univercity , the united states naval observatory and the university of washington .
abadi m. g. , moore b. , & bower r. g. , 1999 , mnras , 308 , 947 abell g. o. , 1958 , apjs , 3 , 211 aguerri j. a. l. , s@xmath159nchez - janssen r. , & mu@xmath160oz - tu@xmath160@xmath161n , 2007 , a&a , 471 , 17 allen d. a. , 1976 , mnras , 174 , 29 baldry i. k. , balogh r. g. , bower k. , glazebrook k. , nichol r. c. , bamford s. p. , & budavari t. , 2006 , mnras , 373 , 469 balogh m. l. , et al .
, 2002 , mnras , 335 , 10 bautz l. p. , & morgan , w. w. , 1970 , apjl , 162 , l149 beers t. c. , flynn k. , & gebhart k. , 1990 , aj , 100 , 32 beers t. c. , gebhardt k. , forman w. , huchra j. p. , jones c. , 1991 , aj , 102 , 1581 beijersbergen m. , hoekstra h. , van dokkum p. g. , & van de hulst t. , 2002 , mnras , 329 , 385 bertin e. & arnouts s. , 1996 , a&a , 117 , 393 bolzonella m. , miralles j. m. , & pell r. , 2000 , a&a , 363,476 bower r. g. , lucey j. r. , & ellis r. s. , 1992 , mnras , 254 , 589 blanton m. r. et al . , 2003 , apj , 592 , 819 bruzual g. , charlot s. , 2003 , mnras , 344 , 1000 butcher h. & oemler a. jr . , 1978 , apj , 226 , 559 cardelli j. a. , clayton g. c. , mathis j , s. , 1989 , apj , 345 , 245 cardiel n. , gorgas j. , cenarro j. , & gonzalez j. j. , 1998 , a&as , 127 , 597 carlberg r. g. , yee h. k. c. , & ellingson e. , 1997 , apj , 478 , 462 cid fernandes r. , mateus a. , sodr , l. , stasiska g. , gomes j. m. , 2005 , mnras , 358 , 363 cid fernandes r. , asari n. v. , sodr , l. , stasiska g. , mateus a. et al . , 2007 , mnras , 375 , l16 colberg j. m. , white s. d. m. et al . , 2000 ,
mnras , 319 , 209 colless m. , & dunn a. m. , 1996 , apj , 458 , 435 crone m. m. , evrard a. e. , & richstone d. o. , 1996 , apj , 467,489 cowie l. l. , songaila a. , hu e. m. , & cohen j. g. , 1996 , aj , 112 , 839 de lucia g. , springel v. , white s. d. m. , croton d. , & kauffmann g. 2006 , mnras , 366 , 499 de filippis e. , paolillo m. , longo g. , la barbera f. , de carvalho r. r. , gal r. , 2011 , mnras , 414 , 2771 dessler a. , 1980 , apj , 236 , 351 dessler a. , oemler a. jr .
et al . , 1997 , apj , 490 , 577 dressler a. & shectman s. a. , 1988 , aj , 95 , 985 ebeling h. , edge a. c. , b@xmath162hringer h. , allen s. w. , crawford c. s. , fabian a. c. , voges w. , huchra j. p. , 1998
, mnras,301 , 881 fan x. , burstein d. , chen j. s. et al . , 1996 , aj , 112 , 628 fasano g. , poggianti b. m. , couch w. j. , bettoni d. et al . , 2000 , apj , 542 , 673 fernndez - soto a. , lanzetta k. m. , yahil a. , 1999 , apj , 513 , 34 gmez p. l. , nichol r. c. , miller c. j. et al . , 2003 , apj , 584 , 210 gao l. , springel v. & white s. d. m. , 2005 , mnras , 363 , l66 geller m. j. & peebles p. j. e. , 1973 , apj , 184 , 329 gott j. r. i. , 1972 , apj , 173 , 277 goto t. , tamauchi c. , fujita y. , okamura s. et al .
, 2003 , mnras , 346 , 601 graves g. j. , faber s. m. , schiavon r. p. , yan r. , 2007 , apj , 671 , 243 graves g. j. , & schiavon r. p. , 2008 ,
apjs , 177 , 446 gunn j. e. , & stryker l. l. , 1983 , apjs , 52 , 121 hansen s. m. , mckay t. a. , wechsler r. h. , annis j. , sheldon e. s. , & kimball a. , 2005 , apj , 633 , 122 holden b. p. , illingworth g. d. , franx m. , blakeslee j. p. et al .
, 2007 , apj , 670 , 190 ilbert o. , capak p. , salvato m. et al .
, 2009 , apj , 690 , 1236 jones c. & forman w. , 1999 , apj , 511,65 kauffmann g. , white s. d. m. , heckman t. m. et al . , 2004 , mnras , 353 , 713 kong x. , et al . , 2000 ,
aj , 119 , 2745 kong x. , & cheng f.z . , 2001 , mnras , 323 , 1035 kong x. , fang g. , arimoto n. , & wang m. 2009 , apj , 702 , 1458 liu s. f. , yuan q. r. , yang y. b. , ma j. , jiang z. j. , wu j. h. , wu z. y. , chen j. s. & zhou x. , 2011 , aj , 141 , 99 mateus a. , sodr l. , cid fernandes r. et al . , 2006 , mnras , 370 , 721 mohr j. j. , fabricant d. g. & geller m. j. , 1993 , apj , 413 , 492 moore b. , katz n. , lake g. , dressler a. , & oemler a. , 1996 , nature , 379 , 613 muratov a. l. & gnedin o. y. , 2010 , apj , 718 , 1266 nelan j. e. , smith r. j. , hudson m. j. , wegner g. a. , lucey j. r. et al . , 2005 , apj , 632 , 137 oegerle w. r. , & hill j. m. , 1994 , aj , 107 , 857 paollilo m. , andreon s. , longo g. , puddu e. , gal r. r. , scaramella r. , djorgovski s. g. , de carvalho r. , 2001 , a&a , 367 , 59 postman m. , & geller m. j. , 1984 , apj , 281 , 95 postman m. et al . , 2005 , apj , 623 , 721p price j. , phillipps s. , huxor a. , smith r. j. & lucey j. r. , 2011 , mnras , 411 , 2558 ramella m. , biviano a. , pisani a. , varela j. , bettoni d. et al . , 2007 , a&a , 470 , 39 rhee g. f. r. n. , van haarlem m. p. , katgert p. , 1991 ,
a&a , 233,325 schiavon r. p. , 2007 ,
apjs , 171 , 146 schechter p. , 1976 ,
apj , 203 , 297 schuecker p. , bhringer h. , reiprich t. h , & feretti l. , 2001 a&a , 378 , 408 smith g. p. , treu t. , ellis r. s. , moran s. m. , dressler a. , 2005 , apj , 620 , 78s thomas p.
a. et al . , 1998 ,
mnras , 296 , 1061 thomas d. , maraston c. & bender r. , 2003 , mnras , 339 , 897 thomas d. , maraston c. , bender r. , mendes de oliveira c. , 2005 , apj , 621 , 673 trager s. c. , faber s. m. , worthey g. & gonzlez j. j. , 2000 , aj , 119 , 1645 west m. j. , villumsen j. v. , dekel a. , 1991 , apj , 369 , 287 west m. j. , jones c. , forman w. , 1995 , apj , 451 , l5 whitmore b. c. & gilmore d. m. , 1991 , apj , 367 , 64w worthey g. , faber s. m. , gonzalez j. j. & burstein d. , 1994 , apjs , 94 , 687 xia l. f. , zhou x. , ma j.,et al . , 2002 , pasp , 114 , 1349 yang y. b. , zhou x. , yuan q. r. , jiang z. j. , ma j. , wu h. , chen , j .
s. , 2004 , apj , 600 , 141 yuan q. r. , zhou x. , jiang z , j. , 2003 , apjs , 149 , 53 yuan q. r. , zhou x. , chen j. s. , ma j. , wu h. , xue s. j. , zhu j. , 2001 , aj , 122 , 1718 zhang l. , yuan q. r. , yang q. , et al . , 2011 ,
pasj , 63 , 585 zhou x. , chen j. s. , xu w. et al .
, 1999 , pasp , 111 , 909 zhou x. , jiang z. j. , xue s. j. , wu h. , ma j. , chen j. s. , 2001 , chjaa , 1 , 372 zhu g. t. , blanton m. r. & moustakas j. , 2010 , apj , 722 , 491 | in this paper we present a photometric and spectroscopic study of the nearby galaxy cluster abell 0671 ( a671 ) with 15 intermediate - band filters in the beijing - arizona - taiwan - connecticut ( batc ) system and the sloan digital sky survey ( sdss ) data . after a cross - identification between the photometric data obtained from the batc and sdss ,
a list of 985 galaxies down to @xmath0 mag in a view field of 58@xmath1@xmath258@xmath1 is achieved , including 103 spectroscopically confirmed member galaxies .
the photometric redshift technique is applied to the galaxy sample for further membership determination .
after the color - magnitude relation is taken into account , 97 galaxies brighter than @xmath3 mag are selected as new member galaxies .
based on the enlarged sample of cluster galaxies , spatial distribution , dynamics of a671 are investigated .
the substructures of a671 are well shown by the sample of bright members , but it appears less significant based on the enlarged sample , which is mainly due to larger uncertainties in the light - of - sight velocities of the newly - selected faint members .
the sdss r - band luminosity function of a671 is flat at faint magnitudes , with the faint end slope parameter @xmath4=-1.12 .
the sdss spectra allow us to investigate the star formation history of bright cluster galaxies , and the galaxies in the core region are found to be older than those in the outskirts .
no environmental effect is found for metallicities of the early - type galaxies ( etgs ) .
both mean stellar ages and metallicities in bright member galaxies are found to be correlated strongly with their stellar masses assembled , and such correlations are dependent upon morphology .
the possitive correlation between age and stellar mass supports the downsizing scenario . by comparing etg absorption - line indices with the state - of - art stellar population models ,
we derive the relevant parameters of simple stellar population ( such as age , [ fe / h ] , [ mg / fe ] , [ c / fe ] , [ n / fe ] , and [ ca / fe ] ) .
the etgs at cluster center tend to have smaller indices , indicating that central etgs are likely to be older .
the distribution of total metallicity indicator , [ mgfe]@xmath5 , does not show any environmental effects .
the relations between the simple stellar population parameters and velocity dispersion in a671 are in good agreement with previous studies .
[ firstpage ] galaxies : clusters : individual ( a671 ) galaxies : distance and redshifts galaxies : kinematics and dynamics galaxies : evolution methods : data analysis |
the parlor game best known as `` twenty questions '' has a long history and a broad appeal .
it was used to advance the plot of charles dickens _ _ a christmas carol__@xcite , in which it is called `` yes and no , '' and it was used to explain information theory in alfrd rnyi s _ _ a diary on information theory__@xcite , in which it is called `` bar - kochba . '' the two - person game begins with an answerer thinking up an object and then being asked a series of questions about the object by a questioner .
these questions must be answered either `` yes '' or `` no . ''
usually the questioner can ask at most twenty questions , and the winner is determined by whether or not the questioner can sufficiently surmise the object from these questions .
many variants of the game exist both in name and in rules .
a recent popular variant replaces the questioner with an electronic device@xcite .
the answerer can answer the device s questions with one of four answers `` yes , '' `` no , '' `` sometimes , '' and `` unknown . ''
the game also differs from the traditional game in that the device often needs to ask more than twenty questions .
if the device needs to ask more than the customary twenty questions , the answerer can view this as a partial victory , since the device has not answered correctly given the initial twenty .
however , the device eventually gives up after @xmath8 questions if it can not guess the questioner s object .
consider a short example of such a series of questions , with only `` yes , '' `` no , '' and `` sometimes '' as possible answers .
the object to guess is one of the seven newtonian colors@xcite , which we choose to enumerate as follows : 1 .
green ( g ) 2 .
yellow ( y ) 3 .
red ( r ) 4 .
orange ( o ) 5 . indigo ( i ) 6 .
violet ( v ) 7 .
blue ( b ) . a first question we ask might be , `` is the color seen as a warm color ? '' if the answer is `` sometimes , ''
the color is green .
if it is `` yes , '' it is one of colors @xmath9 through @xmath10 .
if so , we then ask , `` is the color considered primary ? '' `` sometimes '' implies yellow , `` yes '' implies red , and `` no '' implies orange .
if the color is not warm , it is one of colors @xmath11 through @xmath12 , and we ask whether the color is considered purple , a different question than the one for colors @xmath9 through @xmath10 .
`` sometimes '' implies indigo , `` yes '' implies violet , and `` no '' implies blue .
thus we can distinguish the seven colors with an average of @xmath13 questions if @xmath14 is the probability that color in question is green .
this series of questions is expressible using code tree notation , e.g. , @xcite , in which a tree is formed with each child split from its parent according to the corresponding output symbol , i.e. , the answer of the corresponding question . a code tree corresponding to the above series of questions is shown in fig . [ codetree ] , where a left branch means `` sometimes , '' a middle branch means `` yes , '' and a right branch means `` no . ''
the number of answers possible is referred to by the constant @xmath15 and the tree is a @xmath15-ary tree . in this case ,
@xmath16 and the code tree is ternary .
the number of outputs , @xmath17 , is the number of colors .
the analogous problem in prefix coding is as follows : a source ( the answerer ) emits input symbols ( objects ) drawn from the alphabet @xmath18 , where @xmath0 is an integer .
symbol @xmath19 has probability @xmath20 , thus defining probability vector @xmath21 .
only possible symbols are considered for coding and these are sorted in decreasing order of probability ; thus @xmath22 and @xmath23 for every @xmath24 such that @xmath25 .
( since sorting is only @xmath26 time and @xmath27 space , this can be assumed without loss of generality . )
each input symbol is encoded into a codeword composed of output symbols of the @xmath15-ary alphabet @xmath28 .
( in the example of colors , @xmath29 represents `` sometimes , '' @xmath30 `` yes , '' and @xmath9 `` no . '' ) the codeword @xmath31 corresponding to input symbol @xmath19 has length @xmath32 , thus defining length vector @xmath33 . in fig .
[ codetree ] , for example , @xmath34 is @xmath35 the codeword corresponding to blue so length @xmath36 . the overall code should be a prefix code , that is , no codeword @xmath31 should begin with the entirety of another codeword @xmath37 . in the game , equivalently
, we should know when to end the questioning , this being the point at which we know the answer . for the variant introduced here , all codewords must have lengths lying in a given interval [ @xmath4,@xmath1 ] . in the example of the device mentioned above , @xmath39 and @xmath40 .
a more practical variant is the problem of designing a data codec which is efficient in terms of both compression ratio and coding speed .
moffat and turpin proposed a variety of efficient implementations of prefix encoding and decoding in @xcite , each involving table lookups rather than code trees .
they noted that the length of the longest codeword should be limited for computational efficiency s sake .
computational efficiency is also improved by restricting the overall range of codeword lengths , reducing the size of the tables and the expected time of searches required for decoding .
thus , one might wish to have a minimum codeword size of , say , @xmath41 bits and a maximum codeword size of @xmath42 bits ( @xmath43 ) .
if expected codeword length for an optimal code found under these restrictions is too long , @xmath4 can be reduced and the algorithm rerun until the proper trade - off between coding speed and compression ratio is found .
a similar problem is one of determining opcodes of a microprocessor designed to use variable - length opcodes , each a certain number of bytes ( @xmath44 ) with a lower limit and an upper limit to size , e.g. , a restriction to opcodes being 16 , 24 , or 32 bits long ( @xmath45 , @xmath46 ) .
this problem clearly falls within the context considered here , as does the problem of assigning video recorder scheduling codes ; these human - readable decimal codes ( @xmath47 ) have lower and upper bounds on their size , such as @xmath48 and @xmath49 , respectively .
other problems of interest have @xmath50 and are thus length limited but have no practical lower bound on length@xcite .
yet other problems have not fixed bounds but a constraint on the difference between minimum and maximum codeword length , a quantity referred to as fringe @xcite . as previously noted , large
fringe has a negative effect of the speed of a decoder . in section [ conclusion ] of this paper
we discuss how to find such codes .
note that a problem of size @xmath0 is trivial for certain values of @xmath4 and @xmath1 .
if @xmath51 , then all codewords can have @xmath4 output symbols , which , by any reasonable objective , forms an optimal code . if @xmath52 , then we can not code all input symbols and the problem , as presented here , has no solution . since only other values are interesting , we can assume that @xmath53 $ ] .
for example , for the modified form of twenty questions , @xmath54 , @xmath39 , and @xmath55 , so we are only interested in problems where @xmath56 $ ] .
since most instances of twenty questions have fewer possible outcomes , this is usually not an interesting problem after all , as instructive as it is .
in fact , the fallibility of the answerer and ambiguity of the questioner mean that a decision tree model is not , strictly speaking , correct .
for example , the aforementioned answers to questions about the seven colors are debatable .
the other applications of length - bounded prefix coding mentioned previously , however , do fall within this model .
if we either do not require a minimum or do not require a maximum , it is easy to find values of @xmath4 or @xmath1 which do not limit the problem .
as mentioned , setting @xmath50 results in a trivial minimum , as does @xmath57 .
similarly , setting @xmath58 or using the hard upper bound @xmath59 results in a trivial maximum value . in the case of trivial maximum values , one can actually minimize expected codeword length in linear time given sorted inputs .
this is possible because , at each stage in the standard huffman coding algorithm , the set of huffman trees is an optimal forest ( set of trees)@xcite .
we describe the linear - time algorithm in section [ refine ] .
if both minimum and maximum values are trivial , huffman coding @xcite yields a prefix code minimizing expected codeword length @xmath60 the conditions necessary and sufficient for the existence of a prefix code with length vector @xmath61 are the integer constraint , @xmath62 , and the kraft ( mcmillan ) inequality @xcite , @xmath63 finding values for @xmath61 is sufficient to find a corresponding code , as a code tree with the optimal length vector can be built from sorted codeword lengths in @xmath27 time and space .
it is not always obvious that we should minimize the expected number of questions @xmath64 ( or , equivalently , the expected number of questions in excess of the first @xmath4 , @xmath65 where @xmath66 is @xmath67 if @xmath67 is positive , @xmath29 otherwise ) .
consider the example of video recorder scheduling codes .
in such an application , one might instead want to minimize mean square distance from @xmath4 , @xmath68 we generalize and investigate how to minimize the value @xmath69 under the above constraints for any penalty function @xmath70 convex and increasing on @xmath71 .
such an additive measurement of cost is called a quasiarithmetic penalty , in this case a convex quasiarithmetic penalty .
one such function @xmath72 is @xmath73 , a quadratic value useful in optimizing a communications delay problem @xcite .
another function , @xmath74 for @xmath75 , can be used to minimize the probability of buffer overflow in a queueing system@xcite . mathematically stating the length - bounded problem , @xmath76 note that we need not assume that probabilities @xmath20 sum to @xmath30 ; they could instead be arbitrary positive weights .
thus , in this paper , given a finite @xmath0-symbol input alphabet with an associated probability vector @xmath77 , a @xmath15-symbol output alphabet with codewords of lengths @xmath78 $ ] allowed , and a constant - time - calculable convex penalty function @xmath72 , we describe an @xmath79-time @xmath27-space algorithm for constructing a @xmath72-optimal code , and sketch an even less complex reduction for the most convex penalty function , @xmath80 , minimization of expected codeword length . in the next section , we present a brief review of the relevant literature . in section [ nodeset ] , we extend to @xmath15-ary codes an alternative to code tree notation first presented in @xcite .
this notation aids in solving the problem in question by reformulating it as an instance of the @xmath15-ary coin collector s problem , presented in section [ cc ] as an extension of the ( binary ) coin collector s problem@xcite .
an extension of the package - merge algorithm solves this problem ; we introduce the reduction and resulting algorithm in section [ algorithm ] .
we make it @xmath27 space in section [ linear ] and refine it in section [ refine ] . the alternative approach for the expected length problem of minimizing ( [ huff ] )
i.e. , @xmath80 is often faster ; this approach is sketched in section [ linpen ] .
algorithmic modifications , applications , possible extensions of this work are discussed in section [ conclusion ] .
reviewing how the problem in question differs from binary huffman coding : 1 .
it can be nonbinary , a case considered by huffman in his original paper@xcite ; 2 .
there is a maximum codeword length , a restriction previously considered , e.g. , @xcite in @xmath81 time @xcite and @xmath82 space , but solved efficiently only for binary coding , e.g. , @xcite in @xmath83 time @xmath27 space and most efficiently in @xcite ; 3 .
there is a minimum codeword length , a novel restriction ; 4 .
the penalty can be nonlinear , a modification previously considered , but only for binary coding , e.g. , @xcite .
there are several methods for finding optimal codes for various constraints and various types of optimality ; we review the three most common families of methods here .
note that other methods fall outside of these families , such as a linear - time method for finding minimum expected length codewords for a uniform distribution with a given fringe@xcite .
( this differs from the limited - fringe problem of section [ conclusion ] , in which the distribution need not be uniform and fringe is upper - bounded , not fixed . ) the first and computationally simplest of these are huffman - like methods , originating with huffman in 1952@xcite and discussed in , e.g. , @xcite .
such algorithms are generally linear time given sorted weights and thus @xmath26 time in general .
these are useful for a variety of problems involving penalties in linear , exponential , or minimax form , but not for other nonlinearities nor for length - limited coding
. more complex variants of this algorithm are used to find optimal alphabetic codes , that is , codes with codewords constrained to be in a given lexicographical order .
these variants are in the hu - tucker family of algorithms@xcite , which , at @xmath26 time and @xmath27 space@xcite , are the most efficient algorithms known for solving such problems ( although some instances can be solved in linear time@xcite ) .
the second type of method , dynamic programming , is also conceptually simple but much more computationally complex .
gilbert and moore proposed a dynamic programming approach in 1959 for finding optimal alphabetic codes , and , unlike the hu - tucker algorithm , this approach is readily extensible to search trees@xcite .
such an approach can also solve the nonalphabetic problem as a special case , e.g. , @xcite , since any probability vector satisfying @xmath84 for every @xmath24 has an optimal alphabetic code that optimizes the nonalphabetic case .
a different dynamic programming approach can be used to find optimal `` 1''-ended codes@xcite and optimal codes with unequal letters costs@xcite .
itai@xcite used dynamic programming to optimize a large variety of coding and search tree problems , including nonbinary length - limited coding , which is done with @xmath85 time and @xmath82 space by a reduction to the alphabetic case .
we reduce complexity significantly in this paper .
the third family is that of package - merge - based algorithms , and this is the type of approach we use for the generalized algorithm considered here
. introduced in 1990 by larmore and hirschberg@xcite , this approach is most often used for binary length - limited linear - penalty huffman coding , although it has been extended for application to binary alphabetic codes @xcite and to binary convex quasiarithmetic penalty functions @xcite .
the algorithms in this approach generally have @xmath86-time @xmath27-space complexity , although space complexity can vary by application and implementation , and the alphabetic variant and some nonquasiarithmetic ( and thus nonlinear ) variants have slightly higher time complexity ( @xmath87 ) .
to use this approach for nonbinary coding with a lower bound on codeword length , we need to alter the approach , generalizing to the problem of interest .
the minimum size constraint on codeword length requires a relatively simple change of solution range .
the nonbinary coding generalization is a bit more involved ; it requires first modifying the package - merge algorithm to allow for an arbitrary numerical base ( binary , ternary , etc . ) , then modifying the coding problem to allow for a provable reduction to the modified package - merge algorithm . at times
`` dummy '' inputs are added in order to assist in finding an optimal nonbinary code . in order to make the algorithm precise , the @xmath79-time @xmath27-space algorithm , unlike some other implementations@xcite
, minimizes height ( that is , maximum codeword length ) among optimal codes ( if multiple optimal codes exist ) .
before presenting an algorithm for optimizing the above problem , we introduce a notation for codes that generalizes one first presented in @xcite and modified in @xcite .
nodeset notation , an alternative to code tree notation , has previously been used for binary alphabets , but not for general @xmath15-ary alphabet coding , thus the need for generalization . _ the key idea : _ each node @xmath88 represents both the share of the penalty ( [ penalty ] ) ( weight ) and the ( scaled ) share of the kraft sum ( [ kraft ] ) ( width ) assumed for the @xmath89th bit of the @xmath19th codeword . by showing that total weight is an increasing function of the penalty and that there is a one - to - one correspondence between an optimal code and a corresponding optimal nodeset , we reduce the problem to an efficiently solvable problem , the coin collector s problem . in order to do this , we first need to make a modification to the problem analogous to one huffman made in his original nonbinary solution .
we must in some cases add a `` dummy '' input or `` dummy '' inputs of infinitesimal probability @xmath90 to the probability vector to assure that the optimal code has the kraft inequality satisfied with equality , an assumption underlying both the huffman algorithm and ours .
the positive probabilities of these dummy inputs mean that codes obtained could be slightly suboptimal , but we later specify an algorithm where @xmath91 , obviating this concern . as with traditional huffman coding@xcite , the number of dummy values needed is @xmath92 , where @xmath93 for all integers @xmath67 ( not just nonnegative integers ) .
such dummy inputs allow us to assume that the optimal tree ( for real plus dummy items ) is an optimal full tree ( i.e. , that @xmath94 , where @xmath95 is as defined in ( [ kraft ] ) ) . for sufficiently small @xmath96
, the code will be identical to that for @xmath97 , and , as in traditional huffman coding , nondummy codewords are identical to the codewords of an optimal code for the original input distribution .
we can thus assume for our algorithm that @xmath94 and @xmath98 . with this we now present nodeset notation :
a node is an ordered pair of integers @xmath99 such that @xmath100 and @xmath101 .
call the set of all possible nodes @xmath102 .
this set can be arranged in an @xmath103 grid , e.g. , fig .
[ nodesetnum ] .
the set of nodes , or nodeset , corresponding to input symbol @xmath19 ( assigned codeword @xmath31 with length @xmath104 ) is the set of the first @xmath105 nodes of column @xmath19 , that is , @xmath106 .
the nodeset corresponding to length vector @xmath61 is @xmath107 ; this corresponds to a set of @xmath0 codewords , a code .
thus , in fig . [ nodesetnum ] , the dashed line surrounds a nodeset corresponding to @xmath108 .
we say a node @xmath88 has width @xmath109 and weight @xmath110 , as shown in the example in fig .
[ nodesetnum ] .
note that if @xmath111 , @xmath112 .
we must emphasize that the above `` nodes '' are unlike nodes in a graph ; similar structures are sometimes instead called tiles@xcite , but we retain the original , more prevalent term `` nodes . ''
given valid nodeset @xmath113 , it is straightforward to find the corresponding length vector and , if it satisfies the kraft inequality , a code .
we find optimal codes by first solving a related problem , the coin collector s problem .
let @xmath114 denote the set of all integer powers of a fixed integer @xmath115 .
the coin collector s problem of size @xmath116 considers `` coins '' indexed by @xmath117 .
each coin has a width , @xmath118 ; one can think of width as coin face value , e.g. , @xmath119 for a quarter dollar ( 25 cents ) .
each coin also has a weight , @xmath120 .
the final problem parameter is total width , denoted @xmath121 .
the problem is then : @xmath122 we thus wish to choose coins with total width @xmath121 such that their total weight is as small as possible .
this problem is an input - restricted variant of the knapsack problem .
however , given sorted inputs , a linear - time solution to ( [ knap ] ) for @xmath43 was proposed in @xcite .
the algorithm in question is called the package - merge algorithm and we extend it here to arbitrary @xmath15 . in our notation
, we use @xmath123 to denote both the index of a coin and the coin itself , and @xmath124 to represent the @xmath116 items along with their weights and widths .
the optimal solution , a function of total width @xmath121 and items @xmath124 , is denoted @xmath125 ( the optimal coin collection for @xmath124 and @xmath121 ) .
note that , due to ties , this need not be unique , but we assume that one of the optimal solutions is chosen ; at the end of section [ linear ] , we discuss how to break ties . because we only consider cases in which a solution exists , @xmath126 for some @xmath127 and @xmath128 . here , assuming @xmath129 , @xmath130 and @xmath131 are the unique pair of a power of @xmath15 and an integer that is not a multiple of @xmath15 , respectively , which , multiplied , form @xmath121 . if @xmath132 , @xmath131 and @xmath130 are not used .
note that @xmath130 need not be an integer .
* algorithm variables * + at any point in the algorithm , given nontrivial @xmath124 and @xmath121 , we use the following definitions : [ cols=">,^ , < " , ] @xmath133 where @xmath134 denotes integer multiples of @xmath15 and @xmath135 denotes that , for all @xmath136 and @xmath137 , @xmath138 .
then the following is a recursive description of the algorithm : @xmath133 * recursive @xmath15-ary package - merge procedure * _ basis . @xmath132 _ : @xmath139 .
@xmath140 and @xmath141 _ : @xmath142 .
_ case 2a .
@xmath143 , @xmath141 , and @xmath144 _ : @xmath145 .
_ case 2b .
@xmath143 , @xmath141 , and @xmath146 _ : create @xmath147 , a new item with weight @xmath148 and width @xmath149 .
this new item is thus a combined item , or package , formed by combining the @xmath15 least weighted items of width @xmath150 .
let @xmath151 ( the optimization of the packaged version ) .
if @xmath152 , then @xmath153 ; otherwise , @xmath154 .
if an optimal solution to the coin collector s problem exists , the above recursive ( package - merge ) algorithm will terminate with an optimal solution .
we show that the package - merge algorithm produces an optimal solution via induction on the number of input items .
the basis is trivially correct , and each inductive case reduces the number of items by at least one . the inductive hypothesis on @xmath155 and @xmath141 is that the algorithm is correct for any problem instance with fewer input items than instance @xmath156 .
if @xmath157 , or if @xmath158 and @xmath159 , then there is no solution to the problem , contrary to our assumption .
thus all feasible cases are covered by those given in the procedure .
case 1 indicates that the solution must contain at least one element ( item or package ) of width @xmath150 .
these must include the minimum weight item in @xmath160 , since otherwise we could substitute one of the items with this `` first '' item and achieve improvement .
case 2 indicates that the solution must contain a number of elements of width @xmath150 that is a multiple of @xmath15
. if this number is @xmath29 , none of the items in @xmath161 are in the solution
. if it is not , then they all are .
thus , if @xmath162 , the number is @xmath29 , and we have case 2a . if not , we may `` package '' the items , considering the replaced package as one item , as in case 2b . thus the inductive hypothesis holds .
[ pm ] presents a simple example of this algorithm at work for @xmath16 , finding minimum total weight items of total width @xmath163 ( or , in ternary , @xmath164 ) . in the figure
, item width represents numeric width and item area represents numeric weight .
initially , as shown in the top row , the minimum weight item has width @xmath165 .
this item is put into the solution set , and the next step repeats the task on the items remaining outside the solution set .
then , the remaining minimum width items are packaged into a merged item of width @xmath166 ( @xmath167 ) , as in the middle row .
finally , the minimum weight item / package with width @xmath168 is added to complete the solution set , which is now of weight @xmath12 .
the remaining packaged item is left out in this case ; when the algorithm is used for coding , several items are usually left out of the optimal set .
given input sorted first by width then weight , the resulting algorithm is @xmath169 time and space .
we now formalize the reduction from the coding problem to the coin collector s problem .
this generalizes the similar reduction shown in @xcite for binary codes with only a limit on maximum length , which is in turn a generalization of @xcite for length - limited binary codes with linear @xmath72 , the traditional penalty function .
we assert that any optimal solution @xmath170 of the coin collector s problem with total width @xmath171 on coins @xmath124 ( identical to the set of all possible nodes @xmath102 ) is a nodeset for an optimal solution of the coding problem .
this yields a suitable method for solving the problem . to show this reduction
, we first define @xmath172 in a natural manner for any @xmath173 : @xmath174 where @xmath175 is the kraft sum ( [ kraft ] ) .
given @xmath176 , all optimal codes have the kraft inequality satisfied with equality ; otherwise , the longest codeword length could be shortened by one , strictly decreasing the penalty without violating the inequality .
thus the optimal solution has @xmath94 and @xmath177 also define : @xmath178 note that @xmath179 since the subtracted term is a constant , if the optimal nodeset corresponds to a valid code , solving the coin collector s problem solves this coding problem . to prove the reduction
, we need to prove that the optimal nodeset indeed corresponds to a valid code .
we begin with the following lemma : [ lllemma ] suppose that @xmath170 is a nodeset of width @xmath180 where @xmath181 and @xmath67 are integers and @xmath182
. then @xmath170 has a subset @xmath183 with width @xmath184 .
let us use induction on the cardinality of the set .
the base case @xmath185 is trivial since then @xmath186 .
assume the lemma holds for all @xmath187 , and suppose @xmath188 .
let @xmath189 and @xmath190 .
we can view item @xmath191 of width @xmath192 as the smallest contributor to the width of @xmath193 and @xmath184 as the portion of the @xmath15-ary expansion of the width of @xmath193 to the right of @xmath194
. then @xmath184 must be an integer multiple of @xmath150 .
if @xmath195 , @xmath196 is a solution .
otherwise let @xmath197 ( so @xmath198 ) and let @xmath199 be the subset obtained from solving the lemma for set @xmath200 of width @xmath201
. then @xmath202 .
we now prove the reduction : [ cceqll ] any @xmath170 that is a solution of the coin collector s problem for @xmath203 has a corresponding length vector @xmath204 such that @xmath205 and @xmath206 .
any optimal length vector nodeset has @xmath207 .
suppose @xmath170 is a solution to the coin collector s problem but is not a valid nodeset of a length vector .
then there exists an @xmath88 with @xmath208 $ ] such that @xmath209 and @xmath210 .
let @xmath211 .
then @xmath212 and , due to convexity , @xmath213 .
using @xmath214 , we know that @xmath121 is an integer multiple of @xmath215 .
thus , using lemma [ lllemma ] with @xmath216 , @xmath217 , and @xmath218 , there exists an @xmath219 such that @xmath220 . since @xmath221 , @xmath222
this is a contradiction to @xmath170 being an optimal solution to the coin collector s problem , and thus any optimal solution of the coin collector s problem corresponds to an optimal length vector . because the coin collector s problem is linear in time and space
same - width inputs are presorted by weight , numerical operations and comparisons are constant time the overall algorithm finds an optimal code in @xmath79 time and space .
space complexity , however , can be decreased .
if @xmath223 , we are guaranteed no particular inequality relation between @xmath104 and @xmath224 since we did not specify a method for breaking ties . thus the length vector returned by the algorithm need not have the property that @xmath225 whenever @xmath226 .
we would like to have an algorithm that has such a monotonicity property .
a monotonic nodeset , @xmath170 , is one with the following properties : @xmath227 in other words , a nodeset is monotonic if and only if it corresponds to a length vector @xmath61 with lengths sorted in increasing order ; this definition is equivalent to that given in @xcite
. examples of monotonic nodesets include the sets of nodes enclosed by dashed lines in fig .
[ nodesetnum ] and fig .
[ abcd ] . in the latter case , @xmath228 , @xmath229 , @xmath230 , and @xmath231 , so @xmath232 . as indicated ,
if @xmath223 for some @xmath19 and @xmath233 , then an optimal nodeset need not be monotonic .
however , if all probabilities are distinct , the optimal nodeset is monotonic .
[ dmlemma ] if @xmath77 has no repeated values , then any optimal solution @xmath234 is monotonic .
the second monotonic property ( [ validlen ] ) was proved for optimal nodesets in theorem [ cceqll ] .
the first property ( [ firstprop ] ) can be shown via a simple exchange argument .
consider optimal @xmath61 with @xmath24 so that @xmath235 , and also consider @xmath236 with lengths for inputs @xmath19 and @xmath233 interchanged , as in @xcite .
then @xmath237 \\
\quad \leq 0 \end{array}\ ] ] where the inequality is to due to the optimality of @xmath61 . since @xmath238 and @xmath72 is monotonically increasing
, @xmath239 for all @xmath24 and an optimal nodeset without repeated @xmath77 must be monotonic .
taking advantage of monotonicity in a package - merge coding implementation to trade off a constant factor of time for drastically reduced space complexity is done in @xcite for length - limited binary codes .
we extend this to the length - bounded problem , first for @xmath77 without repeated values , then for arbitrary @xmath77 .
note that the total width of items that are each less than or equal to width @xmath240 is less than @xmath241 .
thus , when we are processing items and packages of width @xmath240 , fewer than @xmath242 packages are kept in memory . the key idea in reducing space complexity is to keep only four attributes of each package in memory instead of the full contents . in this manner
, we use @xmath27 space while retaining enough information to reconstruct the optimal nodeset in algorithmic postprocessing .
define @xmath243 for each package @xmath244 , we retain only the following attributes : 1 .
@xmath245 2 .
@xmath246 3 .
@xmath247 4 . @xmath248 where @xmath249 and @xmath250 .
we also define @xmath251 . with only these parameters , the `` first run '' of the algorithm takes @xmath27 space .
the output of this run is the package attributes of the optimal nodeset @xmath170 .
thus , at the end of this first run , we know the value for @xmath252 , and we can consider @xmath170 as the disjoint union of four sets , shown in fig .
[ abcd ] : 1 .
@xmath253 = nodes in @xmath254 with indices in @xmath255 $ ] , 2 .
@xmath256 = nodes in @xmath254 with indices in @xmath257 $ ] , 3 .
@xmath258 = nodes in @xmath259 , 4 .
@xmath260 = nodes in @xmath261 . due to the monotonicity of @xmath170 ,
it is clear that @xmath262 \times [ l_{\min}+1 , l_\mid-1]$ ] and @xmath263 \times \{l_\mid\}$ ] .
note then that @xmath264 and @xmath265 .
thus we need merely to recompute which nodes are in @xmath253 and in @xmath260 . because @xmath260 is a subset of @xmath266 , @xmath267 and @xmath268 .
given their respective widths , @xmath253 is a minimal weight subset of @xmath255 \times [ l_{\min}+1,l_{\mid}-1]$ ] and @xmath260 is a minimal weight subset of @xmath269 \times [ l_{\mid}+1,l_{\max}]$ ] .
these are monotonic if the overall nodeset is monotonic .
the nodes at each level of @xmath253 and @xmath260 can thus be found by recursive calls to the algorithm .
this approach uses only @xmath27 space while preserving time complexity ; one run of an algorithm on @xmath270 nodes is replaced with a series of runs , first one on @xmath270 nodes , then two on an average of at most @xmath271 nodes each , then four on an average of at most @xmath272 , and so forth .
an optimization of the same complexity is made in @xcite , where it is proven that this yields @xmath79 time complexity with a linear space requirement .
given the hard bounds for @xmath1 and @xmath4 , this is always @xmath273 .
the assumption of distinct @xmath274 puts an undesirable restriction on our input that we now relax . in doing so
, we make the algorithm deterministic , resolving ties that make certain minimization steps of the algorithm implementation dependent .
this results in what in some sense is the `` best '' optimal code if multiple monotonic optimal codes exist .
recall that @xmath77 is a nonincreasing vector .
thus items of a given width are sorted for use in the package - merge algorithm ; this order is used to break ties .
for example , if we look at the problem in fig .
[ nodesetnum ] @xmath275 , @xmath17 , @xmath16 , @xmath57 , @xmath46 with probability vector @xmath276 , then nodes @xmath277 , @xmath278 , and @xmath279 are the first to be grouped , the tie between @xmath279 and @xmath280 broken by order .
thus , at any step , all identical - width items in one package have adjacent indices .
recall that packages of items will be either in the final nodeset or absent from it as a whole .
this scheme then prevents any of the nonmonotonicity that identical @xmath274 might bring about . in order to assure that the algorithm is fully deterministic , the manner in which
packages and single items are merged must also be taken into account .
we choose to combine nonmerged items before merged items in the case of ties , in a similar manner to the two - queue bottom - merge method of huffman coding@xcite .
thus , in our example , there is a point at which the node @xmath281 is chosen ( to be merged with @xmath282 and @xmath283 ) while the identical - weight package of items @xmath284 , @xmath285 , and @xmath286 is not .
this leads to the optimal length vector @xmath108 , rather than @xmath287 or @xmath288 , which are also optimal .
the corresponding nodeset is enclosed within the dashed line in fig .
[ nodesetnum ] , and the resulting monotonic code tree is the code tree shown in fig .
[ codetree ] .
this approach also enables us to set @xmath96 , the value for dummy variables , equal to @xmath29 without violating monotonicity . as in bottom - merge huffman coding ,
the code with the minimum reverse lexicographical order among optimal codes ( and thus the one with minimum height ) is the one produced ; reverse lexicographical order is the lexicographical order of lengths after their being sorted largest to smallest .
an identical result can be obtained by using the position of the `` largest '' node in a package ( in terms of position number @xmath289 ) in order to choose those with lower values , as in @xcite . however
, our approach , which can be shown to be equivalent via simple induction , eliminates the need for keeping track of the maximum value of @xmath289 for each package .
there are changes we can make to the algorithm that , for certain inputs , result in even better performance .
for example , if @xmath290 , then , rather than minimizing the weight of nodes of a certain total width , it is easier to maximize weight over a complementary total width and find the complementary set of nodes .
similarly , if most input symbols have one of a handful of probability values , one can consider this and simplify calculations . these and other similar optimizations have been done in the past for the special case @xmath80 , @xmath291 , @xmath43@xcite , though we do not address or extend such improvements here .
so far we have assumed that @xmath1 is the best upper bound on codeword length we could obtain .
however , there are many cases in which we can narrow the range of codeword lengths , thus making the algorithm faster . for example , since , as stated previously , we can assume without loss of generality that @xmath292 , we can eliminate the bottom row of nodes from consideration in fig .
[ nodesetnum ] .
consider also when @xmath291 .
an upper bound on @xmath293 can be derived from a theorem and a definition due to larmore : consider penalty functions @xmath72 and @xmath294 .
we say that @xmath294 is flatter than @xmath72 if , for positive integers @xmath295 , @xmath296 .
@xcite .
a consequence of the convex hull theorem of @xcite is that , given @xmath294 flatter than @xmath72 , for any @xmath77 , there exist @xmath72-optimal @xmath297 and @xmath294-optimal @xmath298 such that @xmath297 is greater than @xmath298 in terms of reverse lexicographical order .
this explains why the word `` flatter '' is used .
penalties flatter than the linear penalty i.e. , convex @xmath72 can therefore yield a useful upper bound , reducing complexity .
thus , if @xmath291 , we can use the results of a pre - algorithmic huffman coding of the input symbols to find an upper bound on codeword length in linear time , one that might be better than @xmath1 .
alternatively , we can use the least probable input to find a looser upper bound , as in @xcite .
@xmath133 when @xmath299 , one can still use a modified pre - algorithmic huffman coding to find an upper bound as long as @xmath80 .
this is done via a modification of the huffman algorithm allowing an arbitrary minimum @xmath4 and a trivial maximum ( e.g. , @xmath300 or @xmath301 ) : * procedure for length - lower - bounded ( `` truncated huffman '' ) coding * 1 . add @xmath92 dummy items of probability @xmath29 .
2 . combine the items with the @xmath15 smallest probabilities @xmath302 into one item with the combined probability @xmath303 .
this item has codeword @xmath304 , to be determined later , while these @xmath15 smallest items are assigned concatenations of this yet - to - be - determined codeword and every possible output symbol , that is , @xmath305 .
since these have been assigned in terms of @xmath304 , replace the smallest @xmath15 items with @xmath306 in @xmath77 to form @xmath307 .
3 . repeat previous step , now with the remaining @xmath308 codewords and corresponding probabilities , until only @xmath309 items are left .
4 . assign all possible @xmath4 long codewords to these items , thus defining the overall code based on the fixed - length code assigned to these combined items .
this procedure is huffman coding truncated midway through coding , the resulting trees serving as subtrees of nodes of identical depth . excluding the last step , the algorithm is identical to that shown in @xcite to result in an optimal huffman forest .
the optimality of the algorithm for length - lower - bounded coding is an immediate consequence of the optimality of the forest , as both have the same constraints and the same value to minimize . as with the usual huffman algorithm , this can be made linear time given sorted inputs@xcite and can be made to find a code with the minimum reverse lexicographical order among optimal codes via the bottom - merge variant .
clearly , this algorithm finds the optimal code for the length - bounded problem if the resulting code has no codeword longer than @xmath1 , whether this be because @xmath1 is trivial or because of other specifications of the problem .
if this truncated huffman algorithm fails , then we know that @xmath310 , that is , we can not have that @xmath311 for the length - bounded code .
this is an intuitive result , but one worth stating and proving , as it is used in the next section : [ hclemma ] if a ( truncated ) huffman code ( @xmath312 ) for @xmath4 has a codeword longer than some @xmath313 , then there exists an optimal length - bounded code for bound @xmath314 $ ] with codewords of length @xmath313 .
it suffices to show that , if an optimal code for the bound @xmath315 $ ] has a codeword with length @xmath1 , then an optimal code for the bound @xmath316 $ ] has a codeword with length @xmath317 , since this can be applied inductively from @xmath318 ( assuming @xmath319 is the length of the longest codeword of the truncated huffman code ) to @xmath313 , obtaining the desired result .
the optimal nodeset @xmath170 for the bound @xmath320 $ ] has width @xmath321 .
therefore , in the course of the package - merge algorithm , we at one point have @xmath322 packages of width @xmath215 which will eventually comprise optimal nodeset @xmath170 , these packages having weight no larger than the remaining packages of the same width .
consider the nodeset @xmath200 formed by making each @xmath88 in @xmath170 into @xmath323 .
this nodeset is the solution to the package - merge algorithm for the total width @xmath324 with bounds @xmath325 and @xmath317 .
let @xmath326 denote the number of nodes on level @xmath89 .
then @xmath327 since at most @xmath309 nodes can have length @xmath4 .
the subset of @xmath200 not of depth @xmath325 is thus an optimal solution for bounds @xmath4 and @xmath317 with total width @xmath328 that is , at one point in the algorithm this solution corresponds to the @xmath329 least weighted packages of width @xmath215 . due to the bounds on @xmath330 , this number of packages is less than the number of packages of the same width in the optimal nodeset for bounds @xmath4 and @xmath317 ( with total width @xmath331 ) .
thus an optimal nodeset to the shortened problem can contain the ( shifted - by - one ) original nodeset and must have its maximum length achieved for all input symbols for which the original nodeset achieves maximum length
. thus we can find whether @xmath332 by merely doing pre - algorithmic bottom - merge huffman coding ( which , when @xmath333 , results in reduced computation ) .
this is useful in finding a faster algorithm for large @xmath334 and linear @xmath72 .
a somewhat different reduction , one analogous to the reduction of @xcite , is applicable if @xmath312 .
this more specific algorithm has similar space complexity and strictly better time complexity unless @xmath335 .
however , we only sketch this approach here roughly compared to our previous explanation of the simpler , more general approach .
consider again the code tree representation , that using a @xmath15-ary tree to represent the code .
a codeword is represented by successive splits from the root to a leaf one split for each output symbol so that the length of a codeword is represented by the length of the path to its corresponding leaf .
a vertex that is not a leaf is called an internal vertex ; each internal vertex of the tree in fig .
[ codetree ] is shown as a black circle .
we continue to use dummy variables to ensure that @xmath336 , and thus an optimal tree has @xmath94 ; equivalently , all internal vertices have @xmath15 children .
we also continue to assume without loss of generality that the output tree is monotonic .
an optimal tree given the constraints of the problem will have no internal vertices at level @xmath1 , @xmath322 internal vertices in the @xmath334 previous levels , and @xmath337 internal vertices with no leaves in the levels above this , if any .
the solution to a linear length - bounded problem can be expressed by the number of internal vertices in the unknown levels , that is , by @xmath338 \end{array } \label{alpha}\ ] ] so that we know that @xmath339 if the truncated huffman coding algorithm ( as in the previous section ) fails to find a code with all @xmath340 , then we are assured that there exists an @xmath341 , so that @xmath342 can be assumed to be a sequence of strictly increasing integers .
a strictly increasing sequence can be represented by a path on a different type of graph , a directed acyclic graph with vertices numbered @xmath29 to @xmath322 , e.g. , the graph of vertices in fig .
the @xmath19th edge of the path begins at @xmath343 and ends at @xmath342 , and each @xmath342 represents the number of internal vertices at and below the corresponding level of the tree according to ( [ alpha ] ) .
[ codetree ] shows a code tree with corresponding @xmath344 as a count of internal vertices .
the path length is identical to the height of the corresponding tree , and the path weight is @xmath345 for edge weight function @xmath346 , to be determined .
larmore and przytycka used such a representation for binary codes@xcite ; here we use the generalized representation for @xmath15-ary codes . , @xmath16 , @xmath57 , @xmath347 ( @xmath312 ) ] in order to make this representation correspond to the above problem , we need a way of making weighted path length correspond to coding penalty and a way of assuring a one - to - one correspondence between valid paths and valid monotonic code trees .
first let us define the cumulative probabilities @xmath348 so that there are @xmath349 possible values for @xmath350 , each of which can be accessed in constant time after @xmath27-time preprocessing .
we then use these values to weigh paths such that @xmath351 where we recall that @xmath66 denotes @xmath352 and @xmath353 is necessary for cases in which the numbers of internal vertices are incompatible ; this rules out paths not corresponding to valid trees .
thus path length and penalty are equal , that is , @xmath354 this graph weighting has the concave monge property or quadrangle inequality , @xmath355 for all @xmath356 , since this inequality reduces to the already - assumed @xmath357 ( where @xmath358 for @xmath359 ) .
[ dag ] shows such a graph .
a single - edge path corresponds to @xmath360 while the two - edge path corresponds to @xmath361 . in practice
, only the latter would be under consideration using the algorithm in question , since the pre - algorithmic huffman coding assured that @xmath362 .
thus , if @xmath363 and @xmath364 we wish to find the minimum @xmath181-link path from @xmath29 to @xmath322 on this weighted graph of @xmath365 vertices . given the concave monge property , an @xmath366-time @xmath367-space algorithm for solving this problem is presented in @xcite . thus the problem in question can be solved in @xmath368 time and @xmath369 space @xmath27 space if one counts the pre - algorithmic huffman coding and/or necessary reconstruction of the huffman code or codeword lengths an improvement on the package - merge - based approach except for @xmath370 .
one might wonder whether the time complexity of the aforementioned algorithms is the minimum achievable .
special cases ( e.g. , @xmath371 for @xmath80 , @xmath291 , and @xmath43 ) can be addressed using modifications of the package - merge approach@xcite .
also , @xmath77 often implies ranges of values , obtainable without coding , for @xmath372 and @xmath319 .
this enables one to use values of @xmath4 and @xmath1 that result in a significant improvement , as in @xcite for @xmath373 .
an important problem that can be solved with the techniques in this paper is that of finding an optimal code given an upper bound on fringe , the difference between minimum and maximum codeword length .
one might , for example , wish to find a fringe - limited prefix code in order to have a near - optimal code that can be simply implemented , as in section viii of @xcite . such a problem is mentioned in @xcite , where it is suggested that if there are @xmath374 codes better than the best code having fringe at most @xmath375 , one can find this @xmath376-best code with the @xmath377-time algorithm in @xcite , thus solving the fringe - limited problem .
however , this presumes we know an upper bound for @xmath376 before running this algorithm .
more importantly , if a probability vector is far from uniform , @xmath376 can be very large , since the number of viable code trees is @xmath378@xcite .
thus this is a poor approach in general . instead , we can use the aforementioned algorithms for finding the optimal length - bounded code with codeword lengths restricted to @xmath379 $ ] for each @xmath380 , keeping the best of these codes ; this covers all feasible cases of fringe upper bounded by @xmath375 .
( here we again assume , without loss of generality , that @xmath336 . )
the overall procedure thus has time complexity @xmath381 for the general convex quasiarithmetic case and @xmath382 when applying the algorithm of section [ linpen ] to the most common penalty of expected length ; the latter approach is of lower complexity unless @xmath383 .
both algorithms operate with only @xmath27 space complexity .
the author wishes to thank zhen zhang for first bringing a related problem to his attention and david morgenthaler for constructive discussions on this topic .
m. j. golin and g. rote , `` a dynamic programming algorithm for constructing optimal prefix - free codes for unequal letter costs , '' _ ieee trans .
inf . theory _
it-44 , no . 5 , pp . 17701781 , sept .
1998 . t. c. hu , l. l. larmore , and j. d. morgenthaler , `` optimal integer alphabetic trees in linear time , '' in _ proc .
13th annual european symposium on algorithms_.1em plus 0.5em minus 0.4emspringer - verlag , oct .
2005 , pp . 226237 . t. c. hu and j. d. morgenthaler , `` optimum alphabetic binary trees , '' in _ combinatorics and computer science _
lecture notes in computer science , vol .
1120.1em plus 0.5em minus 0.4emspringer - verlag , aug . 1996 , pp . 234243 .
a. rnyi , _ a diary on information theory_.1em plus 0.5em minus 0.4emnew york , ny : john wiley & sons inc . , 1987 , original publication : _ napl az informcielmletrl _ , gondolat , budapest , hungary , 1976 . | efficient optimal prefix coding has long been accomplished via the huffman algorithm . however , there is still room for improvement and exploration regarding variants of the huffman problem .
length - limited huffman coding , useful for many practical applications , is one such variant , for which codes are restricted to the set of codes in which none of the @xmath0 codewords is longer than a given length , @xmath1 .
binary length - limited coding can be done in @xmath2 time and @xmath3 space via the widely used package - merge algorithm and with even smaller asymptotic complexity using a lesser - known algorithm . in this paper
these algorithms are generalized without increasing complexity in order to introduce a minimum codeword length constraint @xmath4 , to allow for objective functions other than the minimization of expected codeword length , and to be applicable to both binary and nonbinary codes ; nonbinary codes were previously addressed using a slower dynamic programming approach .
these extensions have various applications including fast decompression and a modified version of the game `` twenty questions '' and can be used to solve the problem of finding an optimal code with limited fringe , that is , finding the best code among codes with a maximum difference between the longest and shortest codewords .
the previously proposed method for solving this problem was nonpolynomial time , whereas solving this using the novel linear - space algorithm requires only @xmath5 time , or even less if @xmath6 is not @xmath7 . |
the icecube collaboration recently announced the discovery of extraterrestrial neutrinos @xcite .
the sky distribution of these events is consistent with isotropy @xcite .
such an isotropic distribution could be produced as long as the distance of the source is significantly larger than the size of the galactic plane , thus extragalactic astrophysical objects are usually proposed as their sources , but a galactic halo origin is also possible @xcite .
the proposed extragalactic sources include galaxies with intense star - formation , jets and/or cores of active galactic nuclei ( agns ) , gamma - ray bursts and etc .
the large uncertainty in our current knowledge about the distance of neutrinos is due to that so far no associated astrophysical sources have been identified .
a common way to produce astrophysical high - energy neutrinos is the decay of charged pions created in inelastic hadronuclear ( @xmath1 ) and/or photohadronic ( @xmath2 ) processes of cosmic rays ( crs ) , in which high - energy gamma rays will be also generated from the decay of synchronously created neutral pions .
the emissivity of gamma rays and neutrinos are related through @xmath3 ( for @xmath1 process ) , where @xmath4 represents the emission rate per source .
vhe gamma rays ( @xmath5gev ) , unlike neutrinos which propagate through the universe almost freely , will be significantly absorbed by the extragalactic background light ( ebl ) and cosmic microwave background ( cmb ) during the propagation through intergalactic space if the source distance is larger than the mean free path of these gamma rays . for a tev gamma - ray photon ,
the optical depth would be larger than unity when the source locates at @xmath6 .
therefore , the cumulative flux of vhe gamma rays associated with neutrinos carries the information about the distance of the sources of these neutrinos .
significant progresses have been made in our understanding of the extragalactic gamma - ray background ( egb ) in recent years .
the spectrum of the egb has now been measured with the fermi - lat in the energy range from 0.1 to 820 gev @xcite .
new studies of the blazar source count distribution at gamma - ray energies above 50 gev place an upper limit on the residual non - blazar component of the egb @xcite . in this paper
, we use this upper limit at tev energy to constrain the distance and the evolution of the bulk population of neutrino sources .
the distance information has important implications for the search of correlations between observed neutrino events and nearby gamma - ray sources . if the inferred distances of the majority of neutrino sources are large , the search for nearby correlated sources would be a challenge .
it may also explain the negative result 10 eev uhecr arrival directions .
this can be explained only if the sources are hidden in tev gamma - rays .
] of search for the correlation between neutrinos and ultra - high energy cosmic rays obtained by @xcite , which originate within @xmath7 100 mpc .
in 2 , we first present how we use the tev emission to place constraints . then , in 3 , we give the input conditions and assumptions .
we give our results in 4 .
finally , we give the conclusions and discussions in 5 .
in the astrophysical origin scenarios , neutrinos ( also gamma rays ) are produced in various discrete astrophysical objects , so the total observed neutrino flux are the sum of contributions of each individual source , rather than from truly diffuse emissions .
we assume that the neutrino sources are transparent to tev gamma rays , and only consider the attenuation in the intergalactic space due to ebl and cmb absorption . for simplicity , we assume here a poisson distribution for the closest sources , then the probability density that the @xmath8th - closest source locates at a comoving distance @xmath9 can be expressed as@xmath10
where @xmath11 is the expectation number of the sources within a spherical comoving volume @xmath12 with radius @xmath9 , i.e. @xmath13 . so the expected comoving distance where the @xmath8th - closest source locates is @xmath14 . for distant sources , the effect of fluctuation of distances
are unimportant and the distribution can be treated as a uniform distribution , i.e. @xmath15 . the influence of source number density on the gamma ray background is complicated .
if the spatial number density of the sources is low , given a measured diffuse neutrino flux , the corresponding pionic gamma ray luminosity of each source should be relatively high , so this kind of sources are easier to be resolved by instruments . by contrast , if the spatial number density of the source is high , the luminosity of each source should be smaller . as a result , these sources are more likely to be unresolved and hence the emitted gamma rays contribute to the isotropic diffuse gamma - ray background ( igrb ) .
if more nearby sources are resolved from background , the distant sources are allowed to be brighter without violating the igrb data .
such a requirement can be expressed as , @xmath16 where @xmath17 represents the cumulative flux of unresolved sources , @xmath18 represents the flux of the @xmath8-th closest source and @xmath19 is the point source sensitivity of _ _ fermi-__lat . in the same time
, the resolved sources contribute to egb together with the unresolved ones , so they must satisfy @xmath20 where @xmath21 represents the cumulative flux of all sources , including both resolved and unresolved ones . in our calculation , we use the broadband sensitivity provided by _ fermi_-lat performance in pass 8 . assuming the gamma - ray spectral index @xmath22 , the sensitivity reaches a level of @xmath23ergs @xmath24s@xmath25 . according to this , we can determine whether a point source is resolved or not .
based on the above two requirements ( eqs . 2 and 3 ) , we will study how the maximum neutrino contribution from certain source is constrained by the _ fermi _ data . for simplicity
, we assume that all the sources have the same intrinsic gamma - ray luminosity @xmath26 , which relates to the gamma - ray spectral emissivity @xmath27 by @xmath28 , where @xmath29 and @xmath30 are the maximum and minimum energy of emitted photons respectively . here
the prime denotes quantities measured in the rest frame of the source ( i.e. , @xmath31 ) .
while low - energy gamma rays can propagate to us from the sources freely , vhe gamma rays may be absorbed by the ebl and cmb photons in the intergalactic space .
the produced electron / positron pairs will also interact with ebl and cmb photons and generate secondary gamma rays by inverse - compton scattering .
such a cycle is called cascade , and it will continue until the newly generated photons are not energetic enough to produce electron / positron pairs by interacting with the background photons . as a result ,
the absorbed high - energy gamma rays are reprocessed to a bunch of lower - energy ones .
so the total gamma ray flux after propagation consists of a primary component which is the unabsorbed gamma rays , and a cascade component , i.e. , @xmath32 e^{-\tau(e_{\gamma})}+q_{\gamma , cas}\left ( e_{\gamma}\right ) \right\ } /4\pi\bar{r}^{2},\ ] ] where @xmath33 is the optical depth for a photon of energy @xmath34 . in this paper
, we use the optical depth provided by @xcitefinke / ebl / index.html ] and discuss the effect of other ebl models later .
since the results depend on the density of the neutrino source , we will consider three different cases : 1 ) high - density source case , such as star - forming / starburst galaxies .
the density of starburst galaxies is about @xmath35mpc@xmath36 .
starburst galaxies , due to their high star formation rates , and hence large number of supernova or hypernova remnants therein , are huge reservoirs of cosmic ray ( cr ) protons with energy up to eev @xcite .
these crs produce high - energy neutrinos by colliding with gases in galaxies .
2 ) middle - density case , such as clusters of galaxies .
the density of clusters of galaxies is about @xmath37mpc@xmath36 depending on the mass selection @xcite .
here we choose @xmath38mpc@xmath36 as a reference value .
galaxy clusters have been argued to be able to accelerate crs and considered as possible sources for high - energy neutrinos @xcite .
3 ) low - density case , such as blazars ( e.g. bl lacs and flat - spectrum radio quasars ( fsrqs ) ) .
their density ranges from @xmath39mpc@xmath36 to @xmath40mpc@xmath36 @xcite .
we choose @xmath41mpc@xmath36 as a reference value .
as blazars are powerful gamma - ray sources , there have been extensive discussions about their possibility of being high - energy neutrino sources ( see for a review ) . it should be noted that the three densities we chose are just reference values to study the effect of different source density .
any specific sources should refer to the corresponding results based on their spatial densities .
the latest combined maximum - likelihood analysis of icecube neutrinos gives a best - fit power law spectrum with a spectral index of @xmath42 in the energy range between 25 tev and 2.8 pev , and an all - flavor flux of @xmath43gev@xmath25s@xmath25sr@xmath25@xmath24 at 100 tev @xcite .
interestingly , the icecube collaboration has tested the hypothesis of isotropy by analyzing data in the northern and southern sky respectively .
compared to the all - sky result , the spectrum of the events in the northern sky can be better fitted by a harder power - law ( @xmath44 ) , while the southern one favors a slightly softer spectrum ( @xmath45 ) . however , the result is not conclusive , as the discrepancy could be simply caused by a statistical fluctuation .
alternatively , it could be due to an additional component that is present in only one of the hemispheres ( either an unmodeled background component , or e.g. a component from the inner galaxy ) .
as indicated in some recent studies @xcite , the egb above 50gev is dominated by blazars at a level of @xmath46 .
these are mostly low - luminosity hard - spectrum bl lacs .
if it is correct , this implies a strong suppression of contributions from other sources , which can contribute at most a flux of @xmath47gev @xmath24s@xmath25sr@xmath25 at 50gev .
if the neutrino sources are transparent to gamma rays , the neutrino flux per - flavor is then constrained to be at most @xmath48gev @xmath24s@xmath25sr@xmath25 , which can explain the measured neutrino flux at pev energy , but is insufficient to explain the flux at @xmath49tev . to solve this tension , it has been proposed that tev neutrinos may come from some hidden sources , i.e. they are not transparent to gamma rays @xcite . considering the above uncertainties
, we divide our discussions into two cases . in the first case
, we adopt the non - blazar egb obtained by @xcite to place constraints .
we assume that the gamma - ray background is only relevant to @xmath50 tev neutrinos whose sources are transparent to gamma - rays and @xmath49 tev neutrinos may originate from some hidden sources or from a component of the inner galaxy .
possible hidden sources were suggested .
the gamma - ray spectrum at the source should follow the spectrum of neutrinos , which is assumed to be a flat spectrum with index @xmath51 below 1 pev , as predicted by the fermi acceleration mechanism , and a steeper spectrum with @xmath52 above 1 pev @xcite .
then the injection spectrum of gamma rays can be expressed as @xmath53{ll}q_{\gamma}(e_{\gamma})\propto e_{\gamma}^{-2},e_{\gamma}<2(1+z)\mathrm{pev } & \\ q_{\gamma}(e_{\gamma})\propto e_{\gamma}^{-2.5},e_{\gamma}\geq2(1+z)\mathrm{pev}. & \end{array}\ ] ] in the second case , we relax this requirement by considering the full egb given in @xcite , allowing blazars to contribute to icecube neutrinos .
the neutrino spectrum at the source is assumed to follow a broken power law with the spectral index @xmath22 at @xmath54 tev and @xmath52 at @xmath55 tev @xcite .
then the injection spectrum of gamma rays can be expressed as @xmath56{ll}q_{\gamma}(e_{\gamma})\propto e_{\gamma}^{-2},e_{\gamma}<50(1+z)\mathrm{tev } & \\ q_{\gamma}(e_{\gamma})\propto e_{\gamma}^{-2.5},e_{\gamma}\geq50(1+z)\mathrm{tev}. & \end{array}\ ] ]
when the cumulative tev flux is fixed , as constrained by the igrb and egb data , the total neutrino flux is affected by two factors .
first , as we already mentioned above , the spatial density of the sources is an important factor . if the density is smaller , the luminosity of individual source is larger and hence more nearby sources will be resolved .
gamma rays from these resolved sources will not be counted into igrb while they still contribute to diffuse neutrino flux .
the second factor is the distance of the source or the evolution of the source density with redshift .
tev photons from more distant sources will be more likely to be absorbed during propagation to the earth , while neutrinos will not .
so the neutrino flux will be higher if the fraction of distant sources is higher ( or the density evolution with redshift is stronger ) .
in our calculation , we adopt two forms of redshift evolution for the purpose of illustration : one is the constant density evolution , which means that the comoving density of the sources does not change with redshift ; the other one is the star formation rate ( sfr ) evolution , for which the source density evolves as @xmath57 at @xmath58 , @xmath59 at @xmath60 and @xmath61 at @xmath62 . in this case
, we assume that egb and igrb are only relevant to @xmath50 tev neutrinos and adopt the non - blazar egb obtained by @xcite as an upper limit of the cumulative gamma - ray flux from neutrino sources .
first , we want to study what fraction of the observed neutrino flux is contributed by nearby sources . in fig . 1
, we fix the limit of gamma - ray flux to be @xmath63 at 820gev , which corresponds to @xmath64 of the total egb ( i.e. , the non - blazar egb of the integrated photon flux above 50gev .
we here assume the spectrum of the summed emission of all blazars are identical to that of the observed egb . ] ) , and calculate the maximally allowed neutrino flux at pev as a function of the boundary distance @xmath65 .
the boundary distance @xmath65 means that all the sources contributing to igrb / egb locate within the redshift @xmath65 .
the spectra of gamma rays and neutrinos are assumed to follow eq .
( [ spec_casea ] ) .
the left panel of fig .
1 shows the maximally allowed neutrino flux at 1pev without violating the non - blazar egb .
the figure indicates that , in all three source density cases , only a small fraction of neutrinos come from low redshift sources .
the sources below @xmath66 can account for at most a fraction of @xmath67 of the total neutrino flux .
thus , the majority of neutrinos observed by icecube should come from distance farther than @xmath68 .
we can also see that , to account for the observed neutrino flux , the redshift evolution of the sources should not be slower than that of the cosmic star formation rate ( sfr ) .
the constant evolution with respect to redshift can be ruled out .
the above discussions only consider the constraints by the tev gamma - ray background .
however , in the low source density case , the nearest point sources could have been detected by icecube , given a sensitivity of @xmath69tev s@xmath25@xmath24 for icecube @xcite .
so we should also consider the constraints by icecube observation .
this extra constraint suggests that the source density should not be too low , unless part of gamma ray emissions of the sources does not come from the same hadronic process that produces neutrinos .
considering the latter possibility for the low - density source case and the constraint by icecube non - detection of nearby sources , we re - calculate the maximum neutrino flux , which is shown in the right panel of of fig.1 .
we find that the extra constraint affects the low density case and the requirement of a fast evolution is strengthened .
the requirement of a fast evolution of the source density with redshift can also be seen by comparing the expected cumulative tev gamma - ray emission accompanying the production of neutrinos with the observed gamma - ray background data .
2 shows the cumulative gamma - ray emission for different redshift evolution scenarios . here
we adopt the high source density case for illustration , while the result holds for other source density cases .
we can see that a faster redshift evolution will lead to a steeper gamma - ray spectrum .
since @xcite provide only the fraction of integral contribution to the total egb above 50gev by blazars , the exact fraction of non - blazar egb flux at 820 gev is unclear .
thus , we draw a series of horizontal lines in fig.2 to show the different levels of the non - blazar egb fraction at 820 gev . if the non - blazar egb flux is lower than @xmath70 of the total egb flux at 820gev , the source evolution is required to be faster than that of sfr . in all three density evolution scenarios , however , the flux at 10 - 100 gev are quite similar , which is naturally expected since 10 - 100 gev gamma rays are almost as transparent as neutrinos .
this demonstrates the unique role of tev gamma - ray flux in constraining the evolution of neutrino source density . in this case , we allow blazars to contribute to icecube neutrinos and use the full egb as the upper limit @xcite .
the igrb constraint becomes important for this case and we need to consider this constraint as well . thus , we fix the upper limits of the cumulative gamma - ray flux at 820gev to be @xmath71 for unresolved sources ( igrb ) and @xmath72 for all the sources ( egb ) .
first , we assume that @xmath73 tev neutrinos originate from extragalactic sources that are transparent to tev gamma rays .
3 shows the maximally allowed neutrino flux at @xmath74 tev as a function of the boundary redshift @xmath65 , by assuming a gamma - ray spectrum of eq .
( [ spec_caseb ] ) .
the result suggests a similar preference for high redshift sources . in the high and middle density cases ,
the source density evolution should not be slower than the cosmic sfr evolution .
since the luminosities of these sources are relatively low , we find that icecube non - detection constraint hardly affect the result .
on the other hand , for the low density case , the observed neutrino flux can be achieved even for a constant source density evolution . however , when considering the neutrino non - detection constraint of nearby point sources , only @xmath75 of the egb flux at 820 gev can be associated with neutrinos from the same hadronic process . including this extra constraint by icecube non - detection , we find that the source evolution should be faster than the sfr evolution for the low - density case , as shown in the right panel of fig .
3 . like in the non - blazar case
, we compare the cumulative gamma - ray emission of unresolved sources with the igrb data in fig . 4 for the high source density case .
we find that , in order not to exceed the igrb at 820gev , the source density must evolve faster than the sfr evolution .
however , it has been suggested that the egb above 50 gev is dominated by low - luminosity hard - spectrum bl lacs @xcite , which have a negative evolution with redshift in source density @xcite .
if this is correct , we can exclude the possibility that these bl lacs are the main sources of 10 - 100 tev neutrinos observed by icecube .
this is consistent with the conclusion in a previous study by summing up neutrino flux from individual bl lacs@xcite .
thus , the tension between the fact that the dominant sources producing the gamma - ray background have negative evolution with redshift and our conclusion that the sources of the high energy neutrinos must have a fast , positive evolution with redshift argues for hidden sources for 10 - 100 tev neutrinos .
meanwhile , we should also note that some types of blazars such as fsrqs , have a fast , positive evolution with redshift and they only contribute a small part of egb / igrb flux .
these sources are still possible sources for @xmath50tev icecube neutrinos .
this is consistent with the recent discovery of a pev neutrino which is in temporal and positional coincidence with a high - fluence outburst from the fsrq pks b1424 - 418 at redshift @xmath76 @xcite . for these fsrqs
, the situation is actually quite similar to the low - density sources in the non - blazar case .
in the above calculation , we used the ebl model given by @xcite .
different ebl models might affect the opacity of tev gamma rays and we thus study this effect .
we use the upper and lower bounds on the opacity given by @xcite as the boundary of ebl uncertainties .
5 shows the influence on cumulative gamma - ray emission when varying the ebl opacity within this boundary . for illustration
, we adopt a local source density of @xmath77 with the sfr evolution .
we find that , even for the strongest ebl intensity model , the cumulative tev flux decreases only slightly ( less than 20% ) compared to the case using the ebl model given by @xcite .
thus , we conclude that the uncertainty from ebl models does not change our conclusion that a fast source density evolution is required . in previous sections , we obtained the constraint on the source density evolution assuming that the gamma - ray ( or neutrino ) luminosity is the same for all sources .
if the gamma - ray / neutrino luminosity of the source also varies with redshift , it is then the gamma - ray / neutrino emission rate density , rather than source number density , that is constrained . the emission rate density can be expressed as @xmath78 , where @xmath79 represents the source number density and @xmath80 is the factor accounting for the luminosity evolution of the source .
as long as their product @xmath78 , has a fast evolution with redshift , the requirement is fulfilled .
so the source number density may not need to evolve that fast if the factor @xmath80 evolves fast enough .
the factor @xmath80 could originate from various physical causes .
for example , in the starburst galaxy scenario for icecube neutrinos , cosmic rays may have a higher pion production efficiency in higher - redshift galaxies due to higher gas densities therein , which leads to a larger @xmath80 at higher redshifts @xcite .
besides , we note that , although high - luminosity hard - spectrum bl lacs show only a mild evolution in the redshift range of 1<z<2 , they are severely deficient in low - redshift ( z < 0.5 ) universe ( ajello et al .
these sources can be essentially regarded as a fast evolution case , so they may avoid the excess in the diffusive tev gamma - ray background when accounting for icecube neutrinos . to summarize
, we find that extragalactic tev gamma - ray background is a useful tool to study the distance and density evolution of neutrino sources .
we have considered blazar and non - blazar source models and used different injection spectra correspondingly . in both cases ,
we find that only a small fraction of neutrinos are allowed to come from low - redshift sources in order not to exceed the diffuse tev gamma - ray background limit . to account for the icecube neutrino flux
, the density of neutrino sources must have a fast evolution with redshift .
interestingly , this is consistent with the independent result obtained by the tomographic constraints @xcite .
in addition , even for a fast source density evolution , only @xmath50 tev neutrino flux could be explained . as our result shows that the icecube neutrinos mainly come from distant sources at high redshifts , any models arguing for nearby sources
may be ruled out as long as these sources are transparent to tev gamma rays . also , any search for nearby sources correlated with icecube neutrinos would face a challenge .
instead , we suggest to search for correlations with potential cosmic ray accelerators at high redshifts . we thank paolo padovani for the discussions and the anonymous referee for the invaluable report .
this work is supported by the 973 program under grant 2014cb845800 , the nsfc under grants 11273016 , and the excellent youth foundation of jiangsu province ( bk2012011 ) . in the non - blazar source case .
the red , blue and green lines represent the cases with source density of @xmath35mpc@xmath36 , @xmath38mpc@xmath36 and @xmath41mpc@xmath36 , respectively .
the solid and dashed lines represent , respectively , no evolution and sfr evolution of the source density with redshift .
the grey shaded area represents the observed neutrino flux ( per flavor ) by icecube at 1 pev with errors @xcite . in the right panel , an extra constraint due to the icecube non - detection of nearest point sources
is considered . ,
scaledwidth=100.0% ] at @xmath81 , @xmath82 at @xmath60 , and @xmath83 at @xmath62 . the igrb data from _
fermi_-lat are denoted by grey dots and the horizontal short grey lines shows 10% , 30% , 50% and 70% of egb flux at 820 gev , respectively @xcite .
the icecube data are denoted by black and grey dots @xcite . in the calculation
, we use the case of source density of @xmath77.,scaledwidth=90.0% ] | the nearly isotropic distribution of the tev - pev neutrinos recently detected by icecube suggests that they come from sources at distance beyond our galaxy , but how far they are is largely unknown due to lack of any associations with known sources . in this paper
, we propose that the cumulative tev gamma - ray emission accompanying the production of neutrinos can be used to constrain the distance of these neutrino sources , since the opacity of tev gamma rays due to absorption by the extragalactic background light ( ebl ) depends on the distance that these tev gamma rays have travelled . as the diffuse extragalactic tev background measured by _
fermi _ is much weaker than the expected cumulative flux associated with icecube neutrinos , the majority of icecube neutrinos , if their sources are transparent to tev gamma rays , must come from distances larger than the horizon of tev gamma rays .
we find that above 80% of the icecube neutrinos should come from sources at redshift @xmath0 .
thus , the chance for finding nearby sources correlated with icecube neutrinos would be small .
we also find that , to explain the flux of neutrinos under the tev gamma - ray emission constraint , the redshift evolution of neutrino source density must be at least as fast as the the cosmic star - formation rate . |
agn dominate the extragalactic high - energy @xmath0-ray sky , as revealed by the third _ fermi_-lat gamma - ray source list ( 3fgl , @xcite ) .
the great majority of the sources are blazars , i.e. radio - loud agn with relativistic jets oriented at small angles to the observer s line - of - sight .
their radiation is strongly doppler boosted and shows fast variability .
blazars are among the brightest sources of radiation in the universe , spanning the whole electromagnetic spectrum , from radio to @xmath0-rays .
relativistic jets have been studied for half a century , but their fundamental inner workings are still not well understood , e.g. their acceleration and collimation mechanisms , or the location of the @xmath0-ray emission region .
radio galaxies are believed to be the parent population of blazars , with jets oriented at larger angles @xcite .
because of their misaligned orientation , their radiation is much less doppler boosted with respect to blazars .
therefore , radio galaxies are typically orders of magnitude fainter than their aligned counterpart , across the whole spectrum .
because of the limited sensitivity of @xmath0-ray instruments , this has particularly evident implications at high - energies . indeed , _
fermi_-lat has detected less than 20 radio galaxies so far , while the 3fgl includes more than 1000 associated blazars @xcite . in spite of the small sample size and lower fluxes , which complicate radio galaxy studies at high - energies , they provide us with a view on @xmath0-rays from agn jets which is less biased by relativistic effects , and is complementary to blazar studies @xcite .
the combination of @xmath0-ray data with high - resolution radio observations can be a powerful tool in investigating these questions .
_ fermi_-lat has an energy - dependent angular resolution of the order of @xmath20.1 , therefore almost all agn observed in @xmath0-rays appear as point sources , and it is not possible to directly associate the high - energy emission to a specific morphological component .
vlbi observations , on the other hand , are able to achieve milliarcsecond resolutions . since agn are variable sources , it is possible to identify the @xmath0-ray emission region on vlbi scales by looking for correlated variability in radio morphology or flux , and high - energy emission ( e.g. @xcite ) .
additionally , multi - epoch vlbi observations provide the only direct measure of relativistic jet motion , allowing us to derive relevant jet parameters such as apparent speed and jet orientation angle .
tanami ( tracking active galactic nuclei with austral milliarcsecond interferometry ) is a multi - wavelength monitoring program of agn jets south of @xmath3 declination , including almost @xmath4 sources .
the sample was defined as a hybrid radio flux - limited sample combined with sources detected in @xmath0-rays or reported as likely high - energy emitters @xcite .
the core of the program is vlbi monitoring at 8.4 ghz and 22.3 ghz , since 2007 .
the array is based on the australian long baseline array ( lba ) , supported by associated antennas in south africa , new zealand , antarctica and chile ( a full list of the participating antennas can be found in @xcite ) .
first - epoch 8.4 ghz images at milliarcsecond resolution were presented for an initial sample of 43 sources by @xcite .
the sample was constantly expanded with new @xmath0-ray detections or other interesting new sources , and first - epoch images of the first - ever high - resolution observations , for most sources , will be presented in a forthcoming paper ( mller et al . in prep . ) .
the radio vlbi monitoring is complemented by an excellent multi - wavelength coverage , including nir , optical , uv , x - ray and @xmath0-ray data , providing a quasi - simultaneous broadband view of the sources , as is required for detailed studies of variable sources such as agn .
an overview of the multi - wavelength program and selected tanami results can be found in @xcite .
since the tanami sample was selected based on current or expected @xmath0-ray detections , the majority of the sources are blazars , but several well - known radio galaxies are also monitored .
an example is the closest radio - loud agn , centaurus a , for which tanami data revealed the complex dynamics of the jet s inner parsec , indicating downstream jet acceleration @xcite and a spectral morphology @xcite that can be explained within the spine - sheath scenario @xcite .
the tanami sample includes fifteen additional radio galaxies , including several notable sources for which tanami is able to provide the highest resolution data available : the classic fr ii pictor a , whose jet has been detected and resolved in x - rays @xcite and optical @xcite as well as radio bands @xcite ; the fr i pks 0625@xmath135 , which is the most recent addition to the elusive group of only 5 radio galaxies to be detected in the tev band by cherenkov telescopes @xcite , and shows evidence of superluminal motion in a preliminary kinematic analysis of tanami data @xcite ; the nearby fr i centaurus b , detected in @xmath0-rays by _
fermi_-lat with a notably flat spectral index @xcite ; the peculiar object pks 0521@xmath136 , showing @xmath0-ray variability and originally classified as a bl lac but later shown to have properties consistent with a larger viewing angle @xcite .
the full list of tanami radio galaxies is given in table 1 .
the multi - wavelength properties of these sources are currently under study within the tanami collaboration ( angioni et al . in prep . ) , including images , kinematic analysis and spectral index maps .
here we present preliminary results on one @xmath0-ray bright tanami radio galaxy , pks 0521@xmath136 .
this is a nearby ( @xmath5 ) agn with uncertain classification .
leon et al .
2016 @xcite classify it as a bl lac , and derive limits on the jet viewing angle , speed , and doppler factor using the atacama large millimeter array ( alma ) .
their results suggest a jet viewing angle in the range 16@xmath638 .
dammando et al .
2015 @xcite constrain the same parameters using sed modeling including @xmath0-ray data , obtaining a more aligned jet viewing angle of 6@xmath615 .
these results point to an intermediate jet viewing angle between a blazar and a steep spectrum radio quasar ( ssrq ) or radio galaxy .
previous vlbi observations performed with the vlba and with the southern hemisphere vlbi experiment ( sheve ) at 4.9 ghz and 8.4 ghz provided an upper limit on the apparent speed of jet components @xmath7 @xcite .
this is also consistent with the hypothesis that the jet of pks 0521@xmath136 is not strongly beamed .
the authors mentioned that future more sensitive vlbi data would be able to provide more stringent constraints on the nature of this source . the tanami 8.4 ghz full - resolution image presented in the left panel of fig .
[ maps ] achieves an order of magnitude improvement in sensitivity with respect to previous vlbi data @xcite , allowing to reveal a more extended jet36 was already presented in @xcite . ] .
tanami monitoring provided 9 epochs for this source between 2007 and 2013 , an excellent data set for kinematic analysis .
preliminary results from this analysis are shown in the right panel of fig .
[ maps ] , where the distance of jet components from the core is plotted as a function of time .
the data shown include all tanami epochs and the previous vlbi data from @xcite , providing kinematic information across a @xmath8 years time span .
a linear regression fit is applied to the components , with two of them being cross - identified between the two data sets , yielding an estimate of their apparent speed .
this preliminary analysis confirms the absence of fast jet apparent motions in pks 0521@xmath136 , with the largest possible apparent speed ( for the components that can be reliably identified and fitted ) being @xmath9 for component c7 ( see right panel of fig .
[ maps ] ) .
this , together with estimates of the jet - to - counterjet ratio , constrains the intrinsic jet speed and viewing angle to a narrower region of the parameter space with respect to previous studies , namely @xmath10 and @xmath11 .
this viewing angle estimate agrees with the estimate provided by @xcite , supporting the hypothesis that the jet of pks 0521@xmath136 is oriented at an intermediate angle to the line - of - sight . | radio galaxies are relatively faint at @xmath0-ray energies , where they make up only 12% of all agn detected by _ fermi_-lat .
however , they offer a unique perspective to study the intrinsic properties of agn jets . for this reason ,
the combination of @xmath0-ray and multi - wavelength data with high - resolution vlbi monitoring is a powerful tool to tackle the basic unanswered questions about agn jets .
here we present preliminary results from a sample study of radio galaxies in the southern hemisphere observed by the tanami vlbi monitoring program .
we obtain high - resolution maps at 8.4 and 22.3 ghz , and study the jet kinematics using multi - epoch data .
we present a preliminary kinematic analysis for the peculiar @xmath0-ray agn pks 0521@xmath136 .
* keywords * : vlbi , agn , @xmath0-rays , radio galaxies . |
since the early eighties there has been a substantial revival of interest in quantum cosmology , motivated by the hope of obtaining a complete picture of how the universe could arise and evolve @xcite . by complete we here mean
a theoretical description where , by virtue of the guiding principles of physics and mathematics , both the differential equations of the theory and the associated boundary ( and initial ) conditions are fully specified .
even though modern theoretical cosmology deals with yet other deep issues such as dark matter , dark energy @xcite and cosmic strings @xcite , the effort of formulating the appropriate boundary conditions for the quantum state of the universe @xcite , or at least for its ( one - loop ) semiclassical approximation , plays again a key role , since the universe might have had a semiclassical origin @xcite , and the various orders in @xmath0 in the loop expansion describe the departure from the underlying classical dynamics .
the physical motivations of our research result therefore from the following active areas of research : 0.3 cm ( i ) functional integrals and space - time approach to quantum field theory @xcite .
0.3 cm ( ii ) attempt to derive the whole set of physical laws from invariance principles @xcite .
0.3 cm ( iii ) how to derive the early universe evolution from quantum physics ; how to make sense of a wave function of the universe and of hartle hawking quantum cosmology @xcite . 0.3 cm ( iv ) spectral theory and its physical applications , including functional determinants in one - loop quantum theory and hence the first corrections to classical dynamics @xcite . the boundary conditions that we study are part of a unified scheme for maxwell , yang mills and quantized general relativity at one loop , i.e. @xcite @xmath1_{\mathcal b}=0 , \label{(1)}\ ] ] @xmath2_{\mathcal b}=0 , \label{(2)}\ ] ] @xmath3_{\mathcal b}=0 . \label{(3)}\ ] ] with our notation , @xmath4 is a projector acting on the gauge field @xmath5 , @xmath6 is the gauge - fixing functional , @xmath7 is the ghost field ( or full set of ghost fields ) @xcite .
both equation ( 1 ) and ( 2 ) are preserved under infinitesimal gauge transformations provided that the ghost obeys homogeneous dirichlet conditions as in ( 3 ) . for gravity
, we choose @xmath6 so as to have an operator @xmath8 of laplace type on metric perturbations in the one - loop euclidean theory .
on the euclidean 4-ball , we expand metric perturbations @xmath9 in terms of scalar , transverse vector , transverse - traceless tensor harmonics on @xmath10 . for vector , tensor and ghost modes , boundary conditions reduce to dirichlet or robin @xcite .
for scalar modes , one finds eventually the eigenvalues @xmath11 from the roots @xmath12 of @xcite @xmath13 @xmath14 note that both @xmath12 and @xmath15 solve the same equation .
for example , at small @xmath16 and large @xmath12 , the roots of eq .
( 5 ) with @xmath17 sign in front of @xmath18 read as ( here @xmath19 ) @xmath20 , \label{(6)}\ ] ] where @xmath21 and ( having defined @xmath22 ) @xmath23 @xmath24 @xmath25 as we have found in the second item of @xcite .
from eqs . ( 4 ) and ( 5 ) we obtain the following integral representations of the resulting @xmath26-functions upon exploiting the cauchy theorem and rotation of contour @xcite : @xmath27 where ( here @xmath28 ) @xmath29 @xmath30 @xmath31 being the modified bessel functions of first kind . regularity at the origin is easily proved in the elliptic sectors , corresponding to @xmath32 and @xmath33 .
we now define @xmath36 and consider the uniform asymptotic expansion ( away from @xmath37 , with notation as in @xcite ) @xmath38 the functions @xmath39 being obtained from the olver polynomials for the uniform asymptotic expansion of @xmath31 and @xmath40 @xcite . on splitting @xmath41 with @xmath42 small
, we get an asymptotic expansion of the l.h.s . by writing , _ in the first interval _ on the r.h.s . ,
@xmath43 and then computing @xmath44 the integral @xmath45 is instead found to yield a vanishing contribution in the @xmath46 limit ( second item in @xcite ) . remarkably , by virtue of the spectral identity @xmath47 which holds @xmath48 , we find @xmath49 and @xmath50 where @xmath51 , \label{(20)}\end{aligned}\ ] ] @xmath52 .
\label{(21)}\ ] ] the spectral cancellation ( 17 ) achieves three goals : ( i ) vanishing of @xmath53 coefficient in eq .
( 19 ) ; ( ii ) vanishing of @xmath54 in eq .
( 19 ) ; ( iii ) regularity at the origin of @xmath34 .
to cross - check our analysis , we evaluate @xmath55 and hence obtain @xmath56 for all @xmath57 .
only @xmath58 contributes to @xmath59 , and we find @xmath60 \kappa_{2l+1}^{(3 ) } \nonumber \\ & = & { 206 \over 45}+2={296\over 45 } , \label{(22)}\end{aligned}\ ] ] in agreement with eq .
( 19 ) , where @xmath61 are the four coefficients on the right - hand side of @xmath62 within this framework , the spectral cancellation reads as @xmath63 which is a particular case of @xmath64 interestingly , the full @xmath65 value for pure gravity ( i.e. including the contribution of tensor , vector , scalar and ghost modes ) is then found to be positive : @xmath66 @xcite , which suggests a quantum avoidance of the cosmological singularity driven by full diffeomorphism invariance of the boundary - value problem for one - loop quantum theory @xcite .
several open problems should be brought to the attention of the reader , and are as follows . 0.3 cm ( i )
we have encountered a boundary - value problem where the generalized @xmath26-function remains well defined , even though the mellin transform relating @xmath26-function to heat kernel does not exist ( see further comments below ) , since strong ellipticity is violated @xcite ( see also @xcite ) .
are the spectral cancellations ( 17 ) and ( 25 ) a peculiar property of the euclidean 4-ball , or can they be extended to more general riemannian manifolds with non - empty boundary ? 0.3 cm ( ii ) what is the deeper underlying reason for finding @xmath67 ?
is it possible to foresee a geometrical or topological or group - theoretical origin of this result ?
0.3 cm ( iii ) is it correct to say that our positive @xmath65 value for pure gravity engenders a quantum avoidance of the cosmological singularity at one - loop level ?
@xcite does the result remain true in higher - loop calculations or on using other regularization techniques for the one - loop correction ? 0.3 cm ( iv ) the whole scheme might be relevant for ads / cft in light of a profound link between ads / cft and the hartle hawking wave function of the universe @xcite
. 0.3 cm ( v ) what happens if one considers instead non - local boundary data , e.g. those giving rise to surface states for the laplacian ?
@xcite 0.3 cm as far as item ( i ) is concerned , we should add what follows . the integral representation ( 11 ) of the generalized @xmath26-function
is legitimate because , for any fixed @xmath16 , there is a countable infinity of roots @xmath68 and @xmath69 of eqs .
( 4 ) and ( 5 ) , and they grow approximately linearly with the integer @xmath57 counting such roots .
the functions @xmath70 and @xmath71 admit therefore a canonical - product representation @xcite which ensures that the integral representation ( 11 ) reproduces the standard definition of generalized @xmath26-function @xcite . furthermore ,
even though the mellin transform relating @xmath26-function to integrated heat kernel can not be exploited when strong ellipticity is not fulfilled , it remains possible to define a generalized @xmath26-function .
for this purpose , a weaker assumption provides a sufficient condition , i.e. the existence of a sector in the complex plane free of eigenvalues of the leading symbol of the differential operator under consideration @xcite . to make sure
we have not overlooked some properties of the spectrum , we have been looking for negative eigenvalues or zero - modes , but finding none . indeed , negative eigenvalues @xmath72 would imply purely imaginary roots @xmath73 of eq .
( 5 ) , but such roots do not exist , as one can check both numerically and analytically ; zero - modes would be non - trivial eigenfunctions belonging to zero - eigenvalues , but all modes ( tensor , vector , scalar and ghost modes ) are combinations of regular bessel functions @xcite ( since we require regularity at the origin of the left - hand side of eqs .
( 1)(3 ) ) for which this is impossible .
as far as we can see , we still find sources of singularities at the origin in the generalized @xmath26-function resulting from lack of strong ellipticity , but the particular symmetries of the euclidean 4-ball background reduce them to the four terms in eq .
( 24 ) , which add up to zero despite two of them are non - vanishing .
we have proposed to interpret the result @xmath66 for pure gravity as an indication that full diffeomorphism invariance of the boundary - value problem engenders a quantum avoidance of the cosmological singularity . indeed , on the one hand , the work by schleich @xcite had found that , on restricting the functional integral to transverse - traceless perturbations , the one - loop semiclassical approximation to the wave function of the universe diverges at small volumes , at least for the boundary geometry of a three - sphere .
the divergence of the wave functional does not imply , by itself , that the probability density of the wave functional diverges at small volumes , since the probability density @xmath74 $ ] on the space of wave functionals @xmath75 $ ] is given by @xmath74=m[h]|\psi|^{2}[h]$ ] , where @xmath76 $ ] is the measure on this space , the scaling of which is not known in general . on the other hand , in our manifestly covariant evaluation of the one - loop functional integral for the wave function of the universe , it seems incorrect to assume that the measure @xmath76 $ ] scales in such a way as to cancel exactly the contribution of the squared modulus of @xmath77 , which is proportional to the three - sphere radius raised to the power @xmath78 .
thus , we find that our one - loop wave function of the universe vanishes at small volume
. the normalizability condition of the wave function in the limit of small three - geometry , which is weaker than requiring it should vanish in this limit , was instead formulated and studied in @xcite .
the years to come will hopefully tell us whether our calculations may be viewed as a first step towards finding under which conditions a quantum theory of gravity is singularity free in cosmology @xcite . for this purpose
, it might also be interesting to study diffeomorphism - invariant boundary conditions for @xmath79 theories of gravity , recently studied at one - loop level on manifolds without boundary @xcite .
we are grateful to gerald dunne , dmitri fursaev , mariel santangelo and antonello scardicchio for stimulating questions , and to emilio elizalde and sergei odintsov for having organized such a beautiful conference .
the work of g esposito has been partially supported by prin _ sintesi_. k kirsten is grateful to the baylor university research committee , to the max - planck - institute for mathematics in the sciences ( leipzig , germany ) and to the infn for financial support .
the work of a yu kamenshchik was partially supported by the russian foundation for basic research under the grant no 02 - 02 - 16817 and by the scientific school grant no 2338.2003.2 | a vanishing one - loop wave function of the universe in the limit of small three - geometry is found , on imposing diffeomorphism - invariant boundary conditions on the euclidean 4-ball in the de donder gauge .
this result suggests a quantum avoidance of the cosmological singularity driven by full diffeomorphism invariance of the boundary - value problem for one - loop quantum theory .
all of this is made possible by a peculiar spectral cancellation on the euclidean 4-ball , here derived and discussed . |
the gravitational physics has been successfully understood in terms of general relativity . however , since the non - gravitational physics has been accurately explained by the principles of quantum mechanics , it seems necessary that general relativity is merged with quantum mechanics @xcite . for the quantum theory of gravity @xcite ,
there have been attempts such as string theory @xcite . in this paper , we present a particle model of our spacetime , and explain the origin of gravity ( i.e. , general relativity ) , as follows : in the ambient or bulk spacetime @xmath10 ( @xmath1 ) , there exist new particles called space quanta . "
a multitude of space quanta form a nearly three - dimensional object , which is called the quasi-3d object . " within this quasi-3d object , the average distance @xmath11 between nearest - neighbor space quanta satisfies @xmath12 , where @xmath13 is the planck mass @xmath14 . at low energies
@xmath15 , we can use a continuum approximation @xcite that the quasi-3d object is replaced with a 3d continuum called the space 3-brane . " like a bosonic string , this space 3-brane sweeps out its 4d world volume " @xmath16 in the ambient spacetime @xmath10 .
this world volume @xmath17 is described by an embedding @xmath3 , which produces the induced metric @xmath4 on @xmath17 .
this emergent spacetime ( @xmath9 ) from the many space quanta is proposed as the particle model of our spacetime .
the dynamics of the embedding @xmath18 is provided by an effective theory @xmath19 = s_{\rm emb}^{\rm ( 3 br)}[f^a ] + \cdots$ ] , where the latter ellipsis @xmath20 denotes the action for the matter sector ( e.g. , the standard model ) .
to study our spacetime ( @xmath9 ) , we construct the aim - at - target ( aat ) method , " which introduces an action for a 4d metric @xmath7 , namely , @xmath21 = s_{\rm met}^{\rm ( ovlp)}[g_{\mu \nu } ] + \cdots$ ] , where the latter ellipsis @xmath20 denotes the action for the matter sector .
this new metric @xmath22 is used for finding the embedding @xmath23 through the equality @xmath24 , as follows : for an easy understanding of the aat method , we consider a simple situation that the universe contains only the space - quantum sector ( i.e. , the space 3-brane)the absence of the matter sector . in this situation
, we only have to study the two simpler actions @xmath25 $ ] and @xmath26 $ ] , which are the actions without the matter sector .
for example , when the ambient spacetime @xmath27 is the minkowski spacetime @xmath28 of the flat metric @xmath29 , the induced metric @xmath4 is represented as [ intro - ind - met ] _ = _
f_sol^a _ f_sol^b _
ab^bulk , where @xmath30 is a solution for the @xmath18 equation of motion @xmath31 . for the above metric action @xmath32 $ ] , when a solution @xmath33 of @xmath34 satisfies the equality [ intro - met = ind ] g_^sol = _ , eqs .
( [ intro - ind - met ] ) and ( [ intro - met = ind ] ) require that the solution @xmath30 of @xmath35 should also be a solution of the partial differential equation ( pde ) for @xmath18 [ intro - rifle - pde ] _
f^a _ f^b _
ab^bulk = g_^sol .
in other words , when the new metric @xmath33 satisfies @xmath36 , the solution @xmath30 of the equation of motion @xmath37 can be found by solving the pde @xmath38 . note that the embedding @xmath18 is similar to the locally inertial coordinates @xmath39 of general relativity , because the pde @xmath40 is similar in form to @xmath41 , where @xmath42 is the vierbein @xcite . to sum up
, the aat method using the metric action @xmath43 $ ] consists of two main steps : ( i ) finding a solution @xmath33 of @xmath44 = 0 $ ] , and next ( ii ) finding a solution @xmath45 of @xmath40 . in case of @xmath46 ( @xmath47 )
, we can find a solution @xmath48 satisfying @xmath49 , where @xmath30 is what we really want to know .
then , as far as the equality @xmath50 remains true , the combination " of [ combi ] ( a ) the metric action s_met [ g _ ] and ( b ) the pde _ f^a _ f^b _ ab^bulk = g_^sol can be used instead of the 3-brane action @xmath51 $ ] .
this is the essential feature of the aat method . at low enough energies , the metric action @xmath52 $ ] in eq .
( [ combi ] ) can be well approximated by the einstein - hilbert action @xmath53 of general relativity .
then , this einstein - hilbert action @xmath53 can be a good low - energy description for the 3-brane action @xmath25 $ ] in the absence of the matter sector .
when the matter sector is present , the whole action of general relativity can be a good low - energy description for the above universe action " @xmath54 $ ] . in this manner ,
the aat method can produce general relativity at low enough energies this explains the origin of gravity ( i.e. , general relativity ) .
since , in the aat method , general relativity can be subsidiary to the universe action @xmath19 $ ] ( see around eq .
( [ combi ] ) ) , we must not forget that , at the most fundamental level , our universe should be studied in terms of _ physical laws within the ambient spacetime _
@xmath27 which govern both the space - quantum and matter sectors of our universe
. these physical laws within the ambient spacetime @xmath10 can be represented as quantum field theories defined on @xmath10this may shed some light on the quantum theory of gravity @xcite . meanwhile , like usual many - particle systems ( e.g. , superconductors ) , our universe as a system in the ambient spacetime @xmath10 consists of an enormous number of particles such as space quanta .
thus , useful ideas for the study of our universe in @xmath10 can be found by surveying _ physics in our spacetime _
@xmath17 , for example , condensed matter physics @xcite .
the rest of this paper is organized , as follows : in sec .
[ sec : sp - qu - hypothesis ] , the wave - particle duality of quantum mechanics is applied to the gravitational field . since the particle nature of the gravitational field implies the existence of the new particle ( i.e. , space quantum ) , the spacetime manifold @xmath55 of general relativity is assumed to consist of ( very many ) space quanta this is called the space - quantum hypothesis . in sec .
[ sec : cont - approx ] , to maintain the wave nature of a single space quantum ( even without any other space quanta ) , we assume that there exists the ambient spacetime @xmath27 , which surrounds the spacetime @xmath56 of general relativity . to explain the 3d space part of the gr spacetime @xmath56 , we assume that space quanta in @xmath10 form the quasi-3d object , whose continuum limit is the space 3-brane . in sec .
[ sec : 3br - eff - action ] , we deal with the kinematics of the space 3-brane , whose world volume @xmath17 is described by an embedding @xmath23 .
the induced metric @xmath4 on the world volume @xmath17 can be approximated by the gr metric @xmath57 . for simplicity
, we consider the effective theory @xmath25 $ ] _ only _ for the space 3-brane ( the action for matter will be studied in sec .
[ sec : eft - univ ] ) . in sec .
[ sec : aat - method ] , we present the aim - at - target ( aat ) method for studying the effective theory @xmath51 $ ] of the space 3-brane . this aat method using a metric action @xmath58 $ ]
contains the coupled equations @xmath59 and @xmath60 . as far as @xmath24 holds good
, the metric action @xmath58 $ ] can replace the 3-brane action @xmath61 $ ] . in sec .
[ sec : form - met ] , in terms of symmetries , we study the forms of the metric action @xmath58 $ ] used in the aat method .
the @xmath62-invariant action @xmath63 $ ] can explain the einstein - hilbert action @xmath64 $ ] , which is an essential part of general relativity . in sec .
[ sec : eft - univ ] , since the universe contains the matter sector , we consider the more general action @xmath54 = s_{\rm emb}^{\rm ( 3 br)}[f^a ] + \cdots$ ] for the inclusion of matter . by applying the aat method similarly , we can obtain general relativity at low enough energies .
quantum mechanics explains many phenomena of nature very well .
thus , we can try to combine gravity with it ( i.e. , a quantum theory of gravitation ) .
because quantum mechanics has the wave - particle duality as its signature property , we further think about the basic concept * particle * : since the wave - particle duality has been successfully applied to ordinary sensible objects like light and matter , considering these ordinary objects ( rather than graviton ) is helpful in understanding the concept of particle .
an ordinary material object ( e.g. , a bearing ball ) has a substance " characterized by ( i ) stuff material ( e.g. , metal ) and ( ii ) shape in space ( e.g. , ball or sphere ) , which may correspond to matter " ( _ hyle _ in greek ) and form " ( _ eidos _ or _ morphe _ ) of the ancient greek philosophy , respectively .
therefore , an object is called a particle " if the shape of its substance is point - like in space while the object exists . the particle nature of material objects like electron is evident . because the important quantum phenomena like the photoelectric effect and the davisson - germer experiment have been observed by laboratory frames ( e.g. , @xmath65 of fig
. 1(a ) ) under the influence of gravity , the wave - particle duality of quantum mechanics must be observed by the rest frame @xmath65 .
in addition , through the general covariance , the wave - particle duality is also observed by the freely - falling frame @xmath66 of fig .
1(a ) . in the weak - field situation of general relativity ( gr ) , there exists a
nearly lorentz ( nl ) coordinate system " @xmath67 relative to which the metric @xmath57 of a slightly curved gr spacetime @xmath68 has the components at every point @xmath69 of the spacetime @xmath68 [ weakfield ] * g * _ = _ + * h * _ with | * h * _ | 1 at every p _
gr^weak , where @xmath70 in the _ mostly plus _ convention is called the flat background metric , " and @xmath71 a small perturbation " @xcite .
the correspondence between two different situations distinguished by the existence of a gravitational source @xmath72 ( e.g. , the earth ) , namely , ( a ) the _ source - present _ , and ( b ) the _ source - absent _ situations .
there are two kinds of two - frame equivalences , " ( i ) the non - inertial equivalence " between @xmath65 and @xmath73 , and ( ii ) the inertial equivalence " between @xmath66 and @xmath74.,width=340 ] for the above nl coordinates @xmath67 , the einstein tensor @xmath75 has the series expansion in powers of the perturbation @xmath71 [ series - g ] _ * g * _ _ ( _ + * h * _ ) = _ * g*__^ ( 1 ) + o(*h*^2 ) with _ * g*__^ ( 1 ) ( _ _ _ ^+ ) / 2 .
then , the _ vacuum _
einstein s equation @xmath76 has its first - order approximation [ lin - e - eq ] _ * g*__^ ( 1 ) = 0 . under a background lorentz transformation " with @xmath77 , the perturbation @xmath71 in eq .
( [ weakfield ] ) transforms like @xmath78 as if it were a lorentz tensor defined on the flat minkowski spacetime @xmath79 .
this leads to the flat - spacetime fiction " that the tensor @xmath71 belongs to a _ theory in the flat spacetime _
@xmath79 @xcite .
this fiction is supported by the fierz - pauli ( f - p ) theory , where gravity is described by a symmetric tensor on the flat spacetime @xmath79 @xcite .
because the f - p theory shares @xmath80 with general relativity , the curved spacetime @xmath81 of general relativity can be interpreted as the combination of ( i ) the flat spacetime @xmath82 , and ( ii ) the field @xmath83 propagating in this @xmath79 .
this interpretation about @xmath68 is expressed as [ cl - equiv ] s_gr^weak ^4 * h * _ . since the linearized vacuum einstein s equation @xmath84 in eq .
( [ lin - e - eq ] ) has plane - wave solutions , its solution @xmath83 in eq .
( [ cl - equiv ] ) can be the superposition of plane - wave solutions [ pl - wave ] * h*_(x_nl ) = _ d^ 3 k , where @xmath85 is a polarization tensor for wave vector @xmath86 and helicity @xmath87 @xcite . as in the field quantization of the f - p theory , the amplitudes @xmath88 and @xmath89 in eq .
( [ pl - wave ] ) are replaced with the annihilation and creation operators @xmath90 and @xmath91 for a particle " called the graviton the wave - particle duality is applied to the wave " @xmath71 .
the graviton for the field operator @xmath92 is a massless spin-2 particle moving in the flat spacetime @xmath79 .
after the field quantization , the classical field " @xmath93 in the expression @xmath94 corresponds to gravitons , " whose number is denoted by @xmath95 ( @xmath96 )
. then , the classical relation @xmath94 in eq .
( [ cl - equiv ] ) is replaced with its semi - classical counterpart [ semi - equiv ] s_gr^weak ^4 gravitons , which means that the curved spacetime @xmath97 of general relativity is the combination of ( i ) the flat spacetime @xmath79 and ( ii ) the @xmath95 gravitons moving in this @xmath79 . in other words ,
the curved gr spacetime @xmath97 is formed by adding the gravitons ( i.e. , _ particles _ ) to the flat spacetime @xmath79 .
the gravitons " in eq .
( [ semi - equiv ] ) can be regarded as the building blocks of the difference between @xmath68 and @xmath79 .
for example , the gr spacetime @xmath98 depends on the number @xmath95 and the locations of the gravitons .
of course , the flat spacetime @xmath79 in eq .
( [ semi - equiv ] ) may be such a bizarre entity that it does not contain any particles unlike the curved spacetime @xmath99 . however , this @xmath79 shares the same name spacetime manifold " with the @xmath97 , which surely contains particles ( i.e. , the @xmath95 gravitons in eq .
( [ semi - equiv ] ) ) .
thus , the analogical reasoning based on its sharing the same name favors the opposite opinion that the @xmath79 contains particles like the @xmath99 .
moreover , since the quantum theory of gravitons is possible for non - flat background spacetimes " ( e.g. , an expanding universe ) @xcite , the flat spacetime @xmath79 can not be the only " background spacetime for the definition of gravitons .
therefore , we assume that the background spacetime @xmath79 is composed of _ particles _ , whose number is denoted by @xmath100 ( @xmath96 ) .
this implies , through @xmath101 " in eq .
( [ semi - equiv ] ) , that the full gr spacetime " @xmath99 is also composed of particles ( e.g. , the @xmath102 particles@xmath103the @xmath95 gravitons ) .
this conclusion that @xmath68 consists of particles is based on the particular form of the metric @xmath104 in eq .
( [ weakfield ] ) , which is unchanged only for special types of coordinate transformations among all the transformations of general relativity @xcite . despite this
, the conclusion about @xmath68 can be true for all the other coordinate transformations , because our theory can produce general relativity as a prediction ( see sec .
[ sec : eft - univ ] ) . to explain that the flat and curved spacetimes @xmath79 and @xmath68 of general relativity are composed of particles , we make a hypothesis that [ grav - hyp ] every spacetime manifold s_gr of general relativity is composed of _ particles _ , which has the _ meaning _ that [ grav - hyp2 ] & & every point p of the gr spacetime s_gr has a threedimensional ( 3d ) spacelike & & neighborhood n_space^ 3d(p ) which is a continuum approximation " to a discrete & & system composed of particles . since the concepts like substance and shape are basically defined at a constant time , the 3d spacelike neighborhood @xmath105 " in eq .
( [ grav - hyp2 ] ) can represent ( partially ) the _ substance of the spacetime _ @xmath106 .
for example , the substance of a robertson - walker spacetime @xmath107 is wholly represented by the 3d spacelike hypersurface of a constant cosmic time @xmath108 , which approximately describes a discrete system composed of particles according to eq .
( [ grav - hyp2 ] ) .
next , we consider a question : is graviton a _
fundamental _ building block for the substance of the gr spacetime @xmath55 ? " according to eq .
( [ grav - hyp2 ] ) , the substance of the spacetime @xmath56 is a discrete system like solid materials ( cf . sec .
[ sec : cont - approx ] ) .
thus , for analysis , we can use an analogy that the substance of the spacetime @xmath56 corresponds to a crystal composed of many lattice atoms . this atomic crystal can experience a large - scale deformation of its lattice . in the quantum - mechanical framework , the lattice deformation of longer wavelengths than the lattice spacing(s )
can be analyzed by introducing a quantized normal mode called the
phonon " @xcite .
this bosonic quasi - particle , phonon , differs much in moving range from the lattice atom , which is confined to a small region around its equilibrium position .
if the graviton corresponds to the lattice atom in the above analogy , then ( i ) the graviton ( e.g. , a plane - wave solution moving at the speed of light ) should be confined to a small region like the lattice atom , and ( ii ) there exist collective vibrational motions of many
lattice gravitons , " i.e. , the lower - energy excitations corresponding to the phonon in the analogy .
since these two conclusions do not seem plausible , the graviton does not correspond to the lattice atom but to the phonon in the analogy . therefore , we formulate the space - quantum hypothesis " that [ sq - hyp ] & & every point p of the gr spacetime s_gr has a 3d spacelike neighborhood n_space^ 3d(p ) & & which is a continuum approximation to a discrete system * _ syst_*_sq composed of & & particles called * space quanta * , which is the final meaning of the hypothesis in eq .
( [ grav - hyp ] ) . like the phonon ,
the graviton is an _
emergent phenomenon _ arising through interactions among space quanta ( see secs .
[ sec : form - met ] and [ sec : eft - univ ] ) , implying each of these space quanta is different and more fundamental than the graviton .
if the substance of every space quantum has a point - like shape , the space quantum is a particle .
however , the point - like shape of the space quantum ( and the other kinds of quanta ) may be only an approximation based on the smallness of its substance compared with the observational precision .
then , the space quantum may be a spatially @xmath109-dimensional object such as a string ( @xmath110 ) , or a composite system made up of two or more objects which interact weakly and/or strongly .
however , in this paper , the space quantum is regarded as a particle of point - like shape ( i.e. , @xmath111 ) , if the assumption of @xmath112 produces general relativity as a low - energy effective theory ( see secs . [
sec : form - met ] and [ sec : eft - univ ] ) .
since the space - quantum hypothesis implies the space part of the gr spacetime @xmath56 is essentially a discrete object , the hypothesis is different from the basic axiom of general relativity that spacetime is a differentiable manifold ( i.e. , a continuous object ) .
this difference may not be sufficiently studied when the particle nature of gravity receives much less attention than its wave nature .
however , the discrete system of many space quanta ( e.g. , @xmath113 ) can be approximated by a 3d continuous object , when the precision of length measurement is sufficiently larger than the average distance @xmath11 between nearest - neighbor space quanta ( see sec . [
sec : cont - approx ] ) . this philosophy has been successfully used in the continuum mechanics @xcite .
when space quanta forming the discrete system @xmath113 in eq .
( [ sq - hyp ] ) change their positions , the system @xmath113 undergoes a deformation .
this implies , due to the space - quantum hypothesis , that the spacelike subset @xmath114 also deforms .
this deformation of the subset @xmath115 is similarly found in general relativity ( e.g. , the schwarzschild metric ) .
in addition , the deformation of the system @xmath113 can affect the motions of other objects ( e.g. , lights and matters ) _ within _ the system @xmath113 .
this is similar to the deflection of light in general relativity .
these two similarities to general relativity suggest the relationship between the space - quantum hypothesis and general relativity ( see secs . [
sec : form - met ] and [ sec : eft - univ ] ) .
when @xmath116 space quanta form a gr spacetime @xmath117 , the motion of a _ single _ space quantum @xmath118 within @xmath119 can be described by its background spacetime @xmath120 ( @xmath121 ) , which is formed by the _
@xmath122 space quanta
. however , if we consider the limiting case that there are no space quanta except the single quantum @xmath118 ( i.e. , @xmath123 ) , then the wave kinematics using the background spacetime @xmath120 ( @xmath124 ) is not possible any longer , implying the space quantum @xmath118 loses its wave nature . in other words ,
the wave - particle duality ( and thus quantum mechanics ) can not be applied to the single particle @xmath125 in this limiting case .
if we want to maintain the quantum mechanics ( e.g. , the wave nature ) of the particle @xmath118 , a simple solution to the wave - nature problem for @xmath118 is to introduce another spacetime @xmath10 of dimension @xmath126 ( @xmath127 ) within which the space quantum @xmath118 moves like a particle moving within a gr spacetime @xmath56 . in other words ,
the motion of the single space quantum @xmath118 is defined by the * ambient ( i.e. , surrounding ) spacetime * @xmath0 , without considering any other space quanta . in general , any number of space quanta can occupy the ambient spacetime @xmath10 .
since the spacetime @xmath56 of general relativity has the metric @xmath57 of the lorentzian signature ( @xmath128 ) , the ambient spacetime @xmath10 can have its own @xmath126-dimensional lorentzian metric @xmath129 ( @xmath130 ) .
for simplicity , the ambient spacetime @xmath10 with the bulk metric @xmath129 is assumed to be the @xmath131-dimensional minkowski spacetime @xmath132 , where the flat bulk metric @xmath29 is the diagonal matrix in the _ mostly plus _ convention [ bulk - metric ] _ ab^bulk = diag(-1 , + 1 , , + 1 ) everywhere for the inertial bulk - coordinates " @xmath133 ( @xmath134 ) .
these bulk - coordinates @xmath133 are used by an inertial bulk observer " @xmath135 who lives in the ambient spacetime @xmath28 . because it is natural that any particle performs a time evolution in every minkowski spacetime , all space quanta occupying
the ambient spacetime @xmath28 must execute time evolutions , producing their own world lines @xmath136 in the spacetime @xmath28 . here
, the physics of space quantum is studied in the ambient spacetime @xmath28 . to explain the observation that the space part of our universe is three - dimensional
, we assume that space quanta in the spacetime @xmath28 form a nearly 3d object , which is called the * quasi-3d object * of the many space quanta . if the average distance @xmath11 between nearest - neighbor space quanta is sufficiently smaller than the precision @xmath137 of the length measurement , we can apply the continuum approximation to the quasi-3d object in the spacetime @xmath28 , as in the continuum mechanics @xcite . the validity condition " of the continuum approximation
@xcite is [ valid ] d _ sq^ 3 v_sq ( l_obs)^3 , where @xmath138 is the volume of a 3d spacelike region @xmath139 ( @xmath140 ) which contains space quanta .
since there are @xmath141 space quanta inside the region @xmath139 , the validity condition in eq .
( [ valid ] ) implies the region @xmath139 contains many space quanta ( i.e. , @xmath142 ) .
we assume that the quasi-3d object satisfying the validity condition behaves like a 3d continuously - distributed system , which is called the * space 3-brane * ( i.e. , another name of space ) .
in other words , the space 3-brane in the spacetime @xmath28 is the continuum approximation of the quasi-3d object having many space quanta , as in the continuum mechanics for solids and fluids .
mathematically , the space 3-brane composed of many space quanta is represented by a 3d spacelike submanifold of the ambient spacetime @xmath28 .
since the continuum approximation is applied to both of solids and fluids , we need to discuss whether the quasi-3d object ( or its space 3-brane ) is like a solid or a fluid : because space quanta in the fluid phase move faster , the brownian motion can be a crucial criterion distinguishing between the two phases of the quasi-3d object . in the brownian motion ,
the root - mean - square displacement @xmath143 of a big " particle ( e.g. , proton ) colliding with quick space quanta can be proportional to the square root of the elapsed time @xmath144 , namely , r_rms _ e^1/2 .
this long - term behavior implies that the quasi-3d object ( i.e. , space ) behaves like a medium which exerts random forces on the above big particle .
however , the brownian motion caused by the fluid phase of space quanta is rejected by ( i ) newton s first law imposing @xmath145 on every free particle , and ( ii ) the rectilinear propagation of light in vacuum .
for example , if lights from ( more ) distant stars were ( more ) deflected by the above brownian motion , we would observe the ( larger ) disk - like images of the stars .
in fact , the images of stars are point - like .
therefore , the quasi-3d object of many space quanta is like a solid in our observation region .
this _ solid - like _
quasi-3d object ( a ) can have a crystal lattice ( e.g. , simple cubic ) or a non - crystalline structure like an amorphous glass , and ( b ) can deform elastically or plastically in response to stimuli like ordinary solid materials @xcite . the deformations or strains of the quasi-3d object can be ( approximately ) determined by einstein s equation @xmath146 ( see secs .
[ sec : aat - method ] , [ sec : form - met ] and [ sec : eft - univ ] ) .
since a space quantum within the solid - like quasi-3d object is not an isolated particle in the ambient spacetime @xmath147 , the mass @xmath148 of the space quantum may differ considerably from its effective mass @xmath149 which is affected by interactions like ( i ) the effective mass of electron in a solid and ( ii ) the constituent quark mass in a hadron .
because space quanta in the quasi-3d object have inter - particle spacings of @xmath150 , the physical quantities of the space quanta ( e.g. , energy density ) can vary significantly over spatial distances @xmath151 .
thus , since the quasi-3d object resembles a discontinuously - distributed system at a high observational precision @xmath152 , the above continuum approximation breaks down at the high precision @xmath153 .
this requires the lower bound @xmath154 ( @xmath155 ) in order for the continuum approximation to be acceptable .
thus , the continuum approximation considers only the larger - scale ( i.e. , lower - energy ) behaviors of the space - quantum system , ignoring its smaller - scale ( i.e. , higher - energy ) physics .
then , the critical precision @xmath156 for the continuum approximation plays a similar role to the uv cutoff " @xmath157 of an effective field theory @xcite , whose example is the wilsonian effective action obtained by integrating out higher - energy modes than a uv cutoff . therefore , the continuum approximation of the quasi-3d object can be regarded as a low - energy effective theory of the space - quantum system , which has its own uv cutoff @xmath158 ( @xmath159 ) satisfying [ cut - cont ] _ cont = _ cont ( 1/d _ sq ) with _ cont o(10 ^ -1 ) .
the space 3-brane corresponds to the continuum limit " @xmath160 of the quasi-3d object which consists of many space quanta .
then , like a bosonic string , the space 3-brane sweeps out a 4d manifold @xmath17 in the ambient spacetime @xmath28 .
the * world volume * @xmath16 of the space 3-brane is a continuum approximation to the discrete set of the world lines @xmath136 of all space quanta forming the space 3-brane . as in the case of the 2d world sheet @xmath161 of a bosonic string
, we can assume that the world volume @xmath17 of the space 3-brane is a _
4d submanifold _ of the ambient spacetime @xmath28 ( see refs . for mathematical treatments ) : since this submanifold @xmath17 is a subset of @xmath28 ( i.e. , @xmath162 ) , there exists the inclusion map @xmath163 of the world volume @xmath17 , which is defined as a function [ inc - map0 ] & & * _ i _ * : wv_sq ( ^d_amb ) ^d_amb , satisfying + [ inc - map ] & & * _ i_*(p ) = p ^d_amb for every point p _ sq . since @xmath17 is a submanifold of @xmath147 , the inclusion map @xmath163 is an immersion , i.e. , [ immers - inc ] its derivative at p , * _ i_*_p^ : t_p wv_sq t_p ^d_amb , is injective for every p _ sq , where @xmath164 ( @xmath165 ) denotes the tangent vector space of @xmath166 at @xmath69 .
in addition , the inclusion map @xmath163 is of constant rank 4 everywhere on @xmath2 , i.e. , [ rank - inc ] rank ( * _ i_*(p ) ) rank ( * _ i_*_p^ ) = 4 for every p _ sq , where @xmath167 .
then , @xmath168 in eq .
( [ rank - inc ] ) guarantees that the image @xmath169 of the brane tangent space " @xmath170 under the map @xmath171 in eq .
( [ immers - inc ] ) [ pf - tang - sp ] * _ i_*_p^ ( t_p wv_sq ) \ { * _ i_*_p^ ( v ) for v t_p wv_sq } t_p ^d_amb is a 4d subspace of the bulk tangent space " @xmath172 .
we are studying the submanifold @xmath17 within its ambient manifold @xmath28 , which has a coordinate chart @xmath133 at every point @xmath173 : since the submanifold @xmath17 is also a manifold , the set @xmath17 as a 4d manifold has its own coordinate chart @xmath174 ( @xmath175 ) at every point @xmath176 .
therefore , we need to consider _ two _ kinds of charts at every point @xmath69 of @xmath17 : ( i ) a brane - chart " @xmath174 of @xmath17 , and ( ii ) a bulk - chart " @xmath133 of @xmath28 . then , a coordinate transformation @xmath177 between two brane - charts @xmath174 and @xmath178 of @xmath2 is called a
brane - to - brane ( b@xmath179@xmath180 ) transformation .
" moreover , a coordinate transformation @xmath181 between two bulk - charts @xmath133 and @xmath182 of @xmath28 is called a bulk - to - bulk ( b@xmath179@xmath183 ) transformation . "
none of these coordinate transformations @xmath184 and @xmath181 change the point @xmath69 of @xmath17 at all this passive - viewpoint property is shared by any coordinate transformation between two charts in differential geometry .
relative to the brane - chart @xmath174 of @xmath17 , and the bulk - chart @xmath133 of @xmath28 , the equality @xmath185 in eq .
( [ inc - map ] ) has its coordinate representation [ point - inc - map ] y^a(p ) = ( y^a * _ i _ * ( x^)^-1 ) ( x^(p ) ) , where @xmath186 and @xmath187 .
then , for the @xmath174-and-@xmath133 coordinate representation of @xmath163 ( see eq .
( [ point - inc - map ] ) ) [ rep - inc - map ] f^a y^a * _ i _ * ( x^)^-1 , eq . ( [ point - inc - map ] ) defines a new kind of transformation @xmath188 , called the brane - to - bulk ( b@xmath179b ) transformation , " [ b - b - tr ] y^a = f^a(x^ ) f^a x^ at the point p of wv_sq . through the representation @xmath189 in eq .
( [ rep - inc - map ] ) , @xmath190 in eq .
( [ rank - inc ] ) has its @xmath174-and-@xmath133 coordinate representation [ rep - rank ] rank ( * _ i_*_p^ ) = rank(_f^a ) | _ at x^(p ) , where the @xmath191 matrix @xmath192 is the jacobian matrix of the above b@xmath179b transformation @xmath193 . by using the metric bulk - tensor @xmath194 of the ambient manifold @xmath28 , the _ pullback map
_ @xmath195 of the inclusion map @xmath163 in eqs .
( [ inc - map0 ] ) and ( [ inc - map ] ) induces a symmetric tensor @xmath4 on the submanifold @xmath17 in the brane - coordinates " @xmath174 [ ind - metric ] & & _ ( * * _ i_**^ * ^bulk ) _ = ( f^ * _ ab^bulk ) _
= _ f^a _ f^b _ ab^bulk satisfying + [ ind - metric2 ] & & _ v^w^ = ^bulk ( * _ i _ * _ * v , * _ i _ * _ * w ) = _ ab^bulk ( f _ * v)^a ( f _ * w)^b for v , w t_p wv_sq , where the two maps @xmath196 and @xmath197 from @xmath198 are the coordinate representations of the pullback and the pushforward maps @xmath195 and @xmath199 ( e.g. , @xmath171 in eq .
( [ pf - tang - sp ] ) ) in the brane- and bulk - charts @xmath174 and @xmath133 ( cf
. refs . ) .
then , @xmath200 is a tensor defined on the tangent space @xmath201 at a point @xmath176 , whereas @xmath202 is a tensor defined on @xmath172 at the same point @xmath203 as an element of @xmath147 .
then , besides the constraint in eq .
( [ rank - inc ] ) ( equivalently , @xmath204 ) , we assume another constraint on @xmath23 that the symmetric tensor @xmath205 in eq .
( [ ind - metric ] ) is non - degenerate everywhere on the world volume @xmath17 , i.e. , [ nondeg ] det ( _ ) 0 , which means that the pullback @xmath4 becomes a metric tensor " on @xmath17 ( called the induced metric " ) .
note that @xmath206 is a sufficient condition for @xmath207 in eq .
( [ rank - inc ] ) .
relative to the bulk metric @xmath194 , the 4d subspace @xmath169 of the bulk tangent space @xmath208 in eq .
( [ pf - tang - sp ] ) contains both * _ timelike _ bulk - vectors @xmath209 ( i.e. , @xmath210 ) due to the time evolution in the ambient spacetime @xmath28 , and * _ spacelike _ bulk - vectors @xmath211 ( i.e. , @xmath212 ) due to the three - brane nature of the space 3-brane
. thus , the restriction @xmath213 of the bulk - tensor @xmath194 to the subspace @xmath214 has the ( 3 + 1)-dimensional lorentzian signature @xmath215 .
this signature @xmath215 is shared by the induced metric @xmath216 , because @xmath4 as the pullback of @xmath29 satisfies , for example , @xmath217 with @xmath218 ( see eq .
( [ ind - metric2 ] ) ) . therefore , through eq .
( [ ind - metric2 ] ) , the induced metric @xmath4 ( i.e. , the pullback @xmath219 of @xmath29 by @xmath18 ) becomes a lorentzian metric having the 4d lorentzian signature @xmath215 .
then , the 4d lorentzian manifold ( @xmath9 ) is interpreted as a ( 3 + 1)-dimensional spacetime .
this spacetime manifold ( @xmath9 ) is an emergent entity , " because ( @xmath9 ) arises through interactions among many space quanta which occupy the ambient spacetime @xmath28 . to sum up , the 4d manifold ( @xmath220 ) is the ( 3 + 1)-dimensional emergent spacetime which occupies the ambient spacetime @xmath28 . since the spacetime of our universe is ( 3 + 1)-dimensional like ( @xmath6 ) , we assume that the emergent spacetime ( @xmath9 ) occupying @xmath28 forms the spacetime of our universe(@xmath9 ) is the model of our spacetime .
note that the emergent spacetime ( @xmath220 ) is the _ exact or true _
spacetime of our universe .
thus , when we say that a spacetime and a metric of the universe are observed ( or measured ) , the observed spacetime and the observed metric should be identical to @xmath17 and @xmath4 within the measurement precisions . then , since general relativity has accurately explained the spacetime of our universe , we can think that the spacetime @xmath55 and the metric @xmath57 of general relativity are at least the good approximations of the world volume @xmath17 and the induced metric @xmath216 ( see around eqs .
( [ gr - met = g-2 ] ) and ( [ gr - st = st-2 ] ) ) , i.e. , [ s - gr = wv ] s_gr & & wv_sq , + [ gr - met = ind ] * g * _ & & _ (= _ f^a _ f^b _ ab^bulk ) . because einstein developed general relativity without considering the space quanta and the ambient spacetime ,
general relativity is a _ phenomenological _ theory of spacetime like the meson theory which yukawa developed without considering quarks and gluons . until now
, we have established the kinematics for the space 3-brane : the world volume @xmath17 of the space 3-brane in the ambient spacetime @xmath28 is a 4d submanifold of @xmath28 , which is described by the brane - to - bulk transformation @xmath221 satisfying [ lemb - emb ] & ( i ) & an embedding ( i.e. , an immersion and an injection ) , + [ lemb - sign ] & ( ii ) & the 4d lorentzian signature of the induced metric _ = _ f^a _ f^b _ ab^bulk . in eq .
( [ lemb - emb ] ) , the immersion condition is replaced with @xmath206 contained in eq .
( [ lemb - sign ] ) ( see below eq .
( [ nondeg ] ) ) , and the injection condition may be omitted in the case of eccentric behaviors of the space 3-brane ( e.g. , self - intersections ) .
note that the function @xmath23 describing the world volume @xmath2 is neither an arbitrary function of @xmath174 nor a general embedding , but it is a special kind of embedding called the 4d - lorentzian ( 4dl ) embedding , " which is defined as a function satisfying the conditions in eqs .
( [ lemb - emb ] ) and ( [ lemb - sign ] ) . based on the above kinematics for the space 3-brane
, we have to consider its dynamics : an effective theory for the space 3-brane can be defined by the 3-brane action " [ 3br - eff - action ] s_emb^(3 br)[f^a ] = _
wv_sq d^ 4 x _ emb ( f^a , _
f^a , ) , where the lagrangian density @xmath222 can contain the uv cutoff @xmath158 in eq .
( [ cut - cont ] ) . in eq .
( [ 3br - eff - action ] ) , the symbol @xmath223 in @xmath222 does not denote the operator nature of @xmath222 unlike the same symbol used in , e.g. , @xmath224 of sec .
[ sec : sp - qu - hypothesis ] .
note @xmath225 without the symbol @xmath223 is called the lagrangian " ( see below eq .
( [ jac - det ] ) ) .
because the full theory of the effective theory @xmath226 $ ] is not known , we use the bottom - up approach to building an effective theory , i.e. , writing out the most general set of lagrangians consistent with the _ symmetries _ of the theory @xcite .
then , the crucial step is to find the symmetries satisfied by the effective action @xmath51 $ ] for the embedding @xmath227 .
to find the symmetries of the 3-brane action @xmath51 $ ] , we will use a generalization based on the special case of a 0-brane ( i.e. , point particle ) in the 4d minkowski spacetime @xmath79 , as follows : similarly to the space 3-brane in the ambient spacetime @xmath28 , the 0-brane in @xmath79 produces a 1d world line @xmath228 in @xmath79 , which is described by a 1d - lorentzian embedding @xmath229 of the world - line parameter @xmath230 ( @xmath231 ) .
it is well known that the effective action @xmath232 $ ] for the 0-brane has two kinds of symmetries under ( i ) the 4d poincar@xmath233 group @xmath234 with @xmath235 , and ( ii ) the 1d diffeomorphism group @xmath236 with @xmath237 , where @xmath238 for @xmath239 .
note that the @xmath234 symmetry is required by the special principle of relativity ( i.e. , special covariance ) in @xmath79 .
therefore , by using the generalization from the 0-brane in @xmath240 " to a
@xmath241-brane in @xmath242 , " the effective action @xmath51 $ ] for the space 3-brane in @xmath28 ( i.e. , @xmath243 and @xmath244 ) has two corresponding symmetries : the first symmetry is the invariance of the 3-brane action @xmath25 $ ] under the bulk poincar@xmath233 group @xmath245 with [ bulk - poin ] f^a(x^(p ) ) = _ b^a f^b(x^(p ) ) + c^ a at a point
p of wv_sq & & + [ bulk - poin2 ] for f^a = y^a * _ i _ * ( x^)^-1 and f^a = y^a * _ i _ * ( x^)^-1 , & & where @xmath246 and @xmath247 denote each element of the bulk lorentz group @xmath248 , and each translation in @xmath28 , respectively . due to eq .
( [ bulk - poin2 ] ) , the transformation @xmath249 in eq .
( [ bulk - poin ] ) uses only the b@xmath179@xmath183 transformation @xmath250 while keeping the brane - chart @xmath174 fixed .
in other words , the above @xmath245 is exactly the same as the set of all coordinate transformations @xmath250 of the ambient manifold @xmath28 .
the second one is the invariance of the 3-brane action @xmath61 $ ] under the 4d _ local_-reparametrization group @xmath62 ( i.e. , the symmetry group of general relativity ) with [ diff4 ] x^(p ) = _ 4d^(x^(p ) ) and f^a(x^(p ) ) = f^a(x^(p ) ) at the same point p & & + [ diff4 - 2 ] for f^a = y^a * _ i _ * ( x^)^-1 and f^a = y^a * _ i _ * ( x^)^-1 , & & where @xmath251 corresponds to every general coordinate transformation of general relativity . due to eq .
( [ diff4 - 2 ] ) , the transformation @xmath252 in eq . ( [ diff4 ] ) uses only the b@xmath179@xmath253 transformation @xmath184 while keeping the bulk - chart @xmath133 fixed .
in other words , the above @xmath62 is exactly the same as the set of all coordinate transformations @xmath177 of the submanifold @xmath17 .
the insertion of eq .
( [ diff4 - 2 ] ) into eq .
( [ diff4 ] ) produces @xmath254 , which means the invariance of @xmath255 under @xmath256 .
this transformation law @xmath257 under @xmath258 implies that each of @xmath259 is a scalar field under @xmath62 .
note that the nambu - goto action for a bosonic string is the @xmath260 case in the above generalization , having the similar kinds of symmetries .
first , we deal with the @xmath62 invariance of the 3-brane action @xmath25 $ ] more closely : since the world volume @xmath17 exists in the ambient spacetime @xmath28 irrespective of the b@xmath179@xmath180 transformation @xmath261 in eq .
( [ diff4 ] ) , the pair @xmath262 and @xmath227 should be simultaneously the solutions for the equation of motion .
thus , if the unprimed map @xmath227 is an extremum point of the unprimed action @xmath51 $ ] ( i.e. , @xmath227 obeys hamilton s principle @xmath263 = 0 $ ] ) , then the primed map @xmath262 is an extremum point of the primed action @xmath264 $ ] in the primed system ( @xmath265 ) [ primed-3br - eff - action ] s_emb^(3 br ) [ f^a ] = _
wv_sq d^ 4 x^ _ emb^ ( f^a ,
_ ^ f^a , ) , and vice versa .
the situation that both of @xmath262 and @xmath227 are the solutions can be realized by the sameness in the values of the two actions ( called the value invariance of the action " ) [ value - inv ] s_emb^(3 br ) [ f^a ] = s_emb^(3 br)[f^a ] , which leads to [ transf - lag ] _
emb^ ( f^a , _ ^ f^a , ) = det(x^/ x^ ) _
emb(f^a , _
f^a , ) . due to @xmath257 in eq .
( [ diff4 ] ) , the primed metric @xmath266 follows the usual transformation law @xmath267 for a ( @xmath268)-type tensor , which together with @xmath269 in eq .
( [ nondeg ] ) implies [ jac - det ] det(x^/ x^ ) = / .
then , the lagrangian @xmath225 defined as @xmath270 is a scalar under @xmath271 due to @xmath272 unlike the scalar density @xmath222 of weight @xmath273 .
note that @xmath274 is a function of @xmath192 due to the definition @xmath275 in eq .
( [ ind - metric ] ) .
in addition , when the form invariance of the lagrangian density " [ form - inv ] _ emb^ ( f^a , _ ^ f^a , ) = _ emb ( f^a , _ ^ f^a , ) ( thus @xmath276 ) is fulfilled , the euler - lagrange equation for the primed map @xmath262 has the same form as that for the unprimed map @xmath227 .
similarly , the @xmath245 invariance of @xmath61 $ ] consists of two parts , ( a ) the value invariance of the action , and ( b ) the form invariance of the lagrangian density . due to these value and form invariances
, the invariance under a translation @xmath277 from @xmath245 means that the lagrangian density @xmath222 ( thus @xmath225 ) does not contain any derivative - free terms of @xmath18 ( e.g. , @xmath278 ) , implying @xmath279 . from now on , the effective action @xmath25 $ ] in eq .
( [ 3br - eff - action ] ) is expressed as the form using the @xmath62-invariant volume element @xmath280 [ final - eff - s ] s_emb^(3 br)[f^a ] = _
wv_sq d^ 4 x l_emb ( _ f^a , ) , where the lagrangian @xmath225 can contain the uv cutoff @xmath158 in eq . ( [ cut - cont ] ) .
the lagrangian @xmath225 of the 3-brane action @xmath61 $ ] can have the form of [ f - lag ] l_emb ( _ f^a , ) = - t_3 br + l_emb^(der)(_f^a , ) , where @xmath281 is the energy density " or three - brane tension " of the space 3-brane , which corresponds to the nambu - goto action for a three - brane .
the derivative lagrangian " @xmath282 in eq .
( [ f - lag ] ) does not have any constant term .
this derivative lagrangian @xmath283 can be originated from ( i ) internal interactions between space quanta ( e.g. , elastic forces ) , and/or ( ii ) external interactions of space quanta with other kind(s ) of particles .
this lagrangian @xmath284 can contain the einstein - hilbert term @xmath285 , where @xmath286 is the ricci scalar built from the induced metric @xmath216 ( @xmath287 : constant ) .
the embedding scalars " @xmath18 of the 3-brane action @xmath61 $ ] are _ different _ in two ways from ordinary scalars ( e.g. , pions @xmath288 ) in general relativity , as follows : first , unlike the ordinary scalars , the embedding scalars @xmath227 appear in the metric tensor @xmath4 of the world volume @xmath17 through the definition @xmath275 . as a result
, the embedding scalars @xmath18 also appear in the quantities depending on @xmath4 , for example , ( i ) @xmath289 in the action @xmath51 $ ] , ( ii ) the christoffel symbols @xmath290 for the covariant derivative @xmath291 , and ( iii ) the inverse metric @xmath292 for contraction .
second , unlike the ordinary scalars , the solution @xmath293 of the equation @xmath294 = 0 $ ] is not an arbitrary function , but a 4dl embedding .
this 4dl embedding @xmath295 makes the induced metric @xmath4 a 4d metric of the signature @xmath296 .
then , the non - zero value of the composite field @xmath297 may be interpreted as the condensation " for the covariant four - vectors @xmath192 .
now , we want to find the 4d - lorentzian embedding @xmath227 which makes the world volume @xmath17 _ globally flat _ , that is , the induced metric [ emb - flat - sol ] _ ( x^(p ) ) = _ at every point p of wv_sq . due to the definition of @xmath4 , the equality @xmath298 in eq .
( [ emb - flat - sol ] ) can be represented as the partial differential equation ( pde ) for the 4dl embedding @xmath18 [ f - pde ] _
f^a _ f^b _
ab^bulk = _ .
this pde for the embedding @xmath18 can be solved , when its derivatives @xmath192 satisfy [ df - conden ] _ f^a = _
b^^a everywhere on wv_sq , where @xmath299 , @xmath300 ( i.e. , the bulk time ) , and three different @xmath301 .
a simple example is @xmath302 , where @xmath303 comes from the bulk kronecker delta . for @xmath304 in eq .
( [ df - conden ] ) , the induced metric is expressed as [ gl - flat - ind - met ] _ = _
b^b^^bulk , which corresponds to the 4d minkowski spacetime @xmath79 .
because @xmath305 everywhere on @xmath2 , the integration of eq .
( [ df - conden ] ) leads to a linear function of @xmath174 [ lin - emb ] f_lin^a(x ) = _ b^^a x^ + d^a , where @xmath306 are independent of @xmath174 .
then , our remaining task is to check whether this linear embedding @xmath307 is a solution of the euler - lagrange ( e - l ) equation , as follows : hamilton s principle using the lagrangian @xmath308 in eq .
( [ f - lag])[hamil - emb ] [ f^a ] = 0 produces the equation of motion for the space 3-brane ( i.e. , the e - l equation ) [ f - eq ] \ { _ } + = 0 with _ = _
f^a _ f^b _
ab^bulk , where the ellipsis @xmath20 denotes the contribution of the derivative lagrangian @xmath284 . if the energy density @xmath281 of the space 3-brane is independent of @xmath174 , then each term of eq .
( [ f - eq ] ) can contain only the second or higher derivatives of @xmath18 ( e.g. , @xmath309 ) . as a result , each term of eq .
( [ f - eq ] ) vanishes for any linear function of @xmath174 , for example , the linear embedding @xmath310 in eq .
( [ lin - emb ] ) .
therefore , for the @xmath174-independent energy density @xmath311 , the linear embedding @xmath307 is the solution of the e - l equation in eq .
( [ f - eq ] ) , implying the world volume @xmath17 is the 4d minkowski spacetime @xmath79 due to the flat metric @xmath312 induced by @xmath307 . since the linear embedding @xmath307 satisfies the e - l equation _ irrespective of the value _ of the @xmath174-independent @xmath281 , the flat metric @xmath298 of the world volume @xmath2 exists for any value of the uniform energy density @xmath281 .
this interesting feature distinguishes the 3-brane action @xmath25 $ ] from general relativity .
in sec . [ sec : 3br - eff - action ] , since the space 3-brane occupies the ambient spacetime @xmath28 , the effective theory @xmath25 $ ] of the space 3-brane was built for the ambient spacetime @xmath147 through the field @xmath313 .
the world volume ( @xmath9 ) of the space 3-brane is described by the solution @xmath30 for the equation of motion @xmath35 . since this world volume ( @xmath5 ) is the exact or true spacetime of our universe , it is important to know the solution @xmath30 describing our spacetime ( @xmath9 ) . in this section ,
we want to show a methodology for studying the world volume ( @xmath9 ) by using a _
metric action _
@xmath314 $ ] ( see table [ out - aat ] ) .
the key point of this methodology is that the solution @xmath30 of @xmath35 can be found by solving the different equation @xmath315 when the new metric @xmath7 satisfies @xmath316 ( @xmath317 ) .
an example of the methodology is the flat - metric case @xmath318 , whose treatment is shown below eq .
( [ emb - flat - sol ] ) .
the details of our methodology are shown , as follows : to study the world volume ( @xmath9 ) of the space 3-brane , we consider the solution set @xmath319 for the equation of motion @xmath320 = 0 $ ] in eq .
( [ hamil - emb ] ) [ target - set ] _ target \ { f_sol^a : 4dl embedding | ( s_emb^(3 br ) / f^a)[f_sol^a ] = 0 } f_space , where @xmath321 is the function space , and the element @xmath30 is called the 4d - lorentzian ( 4dl ) solution . " for the given * original action * @xmath61 $ ] , the solution set @xmath322 is a _
fixed _ set in the function space @xmath323 .
note that @xmath324 for a uniform energy density @xmath281 ( see around eq .
( [ hamil - emb ] ) ) . by the way , since our methodology to study the solution set @xmath325 is similar to a bullet fired at a fixed target in space " ( see the discussions around eq .
( [ methodology ] ) ) , the solution set @xmath326 is called the * target * ( or target set ) in the * space * @xmath327 . for an easy understanding
, we show the tenors and the vehicles in the bullet - target ( b - t ) metaphor [ metaphor ] _ bullet , _ target , f_space bullet , target , space , where the bullet " @xmath328 will be defined in eq .
( [ bullet - set ] ) .
moreover , there are discussions about the rifle " ( above eq .
( [ rifle - pde ] ) ) and the aim - of - the - rifle " ( above eq .
( [ methodology ] ) ) . through the definition @xmath329 in eq .
( [ ind - metric ] ) , the target @xmath325 in eq .
( [ target - set ] ) produces the set @xmath330 of the induced metrics @xmath4 with the signature @xmath331 [ set - ind ] _ ind \ { _ | _ = _ f_sol^a _ f_sol^b _ ab^bulk for every f_sol^a _ target } .
note that @xmath332 for a uniform energy density @xmath281 ( see around eq .
( [ gl - flat - ind - met ] ) ) .
conversely , this induced - metric set " @xmath333 can produce the target set @xmath326 , because ( i ) the former set @xmath334 produces the solution set @xmath335 of the partial differential equation ( pde ) for the embedding @xmath18 [ sol - pde ] _
pde^(sol ) \ { f^a | _ f^a _ f^b _ ab^bulk = _ for every
_ _ ind } , and ( ii ) this pde solution set " @xmath335 contains the target set @xmath336 , i.e. , [ ind - cont - emb ] _
pde^(sol ) _ target . the result @xmath337 in eq .
( [ ind - cont - emb ] ) suggests a hint about how to know the target set @xmath338 ( @xmath339 ) , implying the importance of studying the induced - metric set @xmath334 . to study this set @xmath334 , since its element
@xmath4 is a 4d lorentzian metric , we consider the theory @xmath314 $ ] of a 4d lorentzian metric @xmath7 ( rather than @xmath4 ) [ met - action ] s_met [ g _ ] = _
s_met^4d d^ 4 x l_met ( _ met ; g _ ) , where ( @xmath340 ) is a 4d spacetime manifold , @xmath341 is the lagrangian containing the derivatives of @xmath22 , and @xmath342 is the uv cutoff of the metric action @xmath314 $ ] ( see table [ out - aat ] for the role of @xmath314 $ ] ) . since the metric action @xmath314 $ ] in eq .
( [ met - action ] ) neglects the _ microscopic _ behaviors of individual space quanta ( i.e. , the underlying discreteness of the space 3-brane ) like the original action @xmath51 $ ] , we can assume the uv cutoff @xmath342 of the metric action @xmath314 $ ] satisfies [ cut - met < cont ] _ met o(_cont ) , where @xmath158 is the uv cutoff of the original action @xmath25 $ ] ( see eqs .
( [ cut - cont ] ) and ( [ 3br - eff - action ] ) ) .
the spacetime manifold ( @xmath343 ) for the metric action @xmath43 $ ] can be a good approximation of the exact or true spacetime manifold ( @xmath9 ) , which is an emergent object arising from many space quanta in @xmath28 ( see eqs .
( [ met - sol = ind ] ) , ( [ st - met = wv ] ) and ( [ equal - st ] ) ) .
then , we can approach the induced - metric set @xmath344 by using the solution set @xmath345 of the equation @xmath346 = 0 $ ] [ sol - met - set ] _
met^(sol ) \ { g_^sol | ( s_met / g _ ) [ g_^sol ] = 0 } , which is called the * cartridge * ( see above eq .
( [ rifle - pde ] ) ) . the solution @xmath33 in eq .
( [ sol - met - set ] ) is called the solution metric . " in order to study the target @xmath347 in the space @xmath323 , we define the * bullet * @xmath328 ( corresponding to the pde solution set @xmath335 in eq .
( [ sol - pde ] ) ) [ bullet - set ] _ bullet \ { f^a | _ f^a _ f^b _ ab^bulk = g_^sol for every g_^sol _ met^(sol ) } f_space by using the cartridge @xmath348 .
since the bullet set @xmath349 can overlap the target set @xmath350 in the function space @xmath351 like the bullet " of the b - t metaphor a bullet fired at a fixed target in space " ( see eq .
( [ metaphor ] ) ) , the set @xmath352 is called the bullet . in eq .
( [ bullet - set ] ) , the pde @xmath353 defines a transformation @xmath354 like the rifle of the above b - t metaphor , which transforms its cartridge into the metallic bullet .
thus , the solution - metric set " @xmath348 and the pde @xmath40 are called the cartridge ( see below eq .
( [ sol - met - set ] ) ) , and the * rifle * firing the bullet @xmath352 , respectively .
this rifle pde , " and the pde for @xmath355 in eq .
( [ sol - pde ] ) have the same form [ rifle - pde ] _ f^a _ f^b _ ab^bulk = q _ ( q
_ = g_^sol , _ ) , which is invariant under the bulk poincar@xmath233 group @xmath245 with @xmath356 ( see sec . [
sec : 3br - eff - action ] ) .
thus , if the
metric intersection ( mi ) " @xmath357 contains an element [ g = gamma ] g_^mi = g_^sol = _ ^ ( = _ f_sol ^a _ f_sol ^b _ ab^bulk with f_sol ^a _ target = \ { f_sol^a } ) , then the bullet @xmath352 shares the 4dl solution @xmath358 with the target @xmath325 . since the converse of this proposition is true , we have the equivalence that [ mi - bt ] _
mi = _ met^(sol ) _ ind _ if and only if _ _ b _ bullet _ target , where @xmath359 is called the bullet - target ( b - t ) overlap . " in eq .
( [ mi - bt ] ) , the b - t overlap @xmath360 describes the manner in which the target @xmath336 of the original action @xmath25 $ ] is overlapped by the bullet @xmath361 of the metric action @xmath314 $ ] .
since the original action @xmath25 $ ] is given to us ( i.e. , not changed arbitrarily by us ) , the target @xmath319 in the b - t overlap @xmath362 is treated as a _ fixed set _ in the function space @xmath327 .
then , the b - t overlap @xmath359 represents the _
maximum knowledge about the fixed target @xmath338 _ which we can obtain by using the bullet @xmath328 of the chosen action @xmath363 $ ] .
for example , the maximum b - t overlap @xmath364 ( i.e. , @xmath365 ) means that we can know the whole of the target @xmath338 by using the maximum bullet @xmath366 .
however , the minimum overlap @xmath367 ( i.e. , @xmath368 ) means that we can not know the target @xmath338 through the minimum bullet @xmath369 .
fortunately , unlike the target @xmath338 , the bullet @xmath328 changes _ depending on _ which metric action @xmath314 $ ] we choose ( see eqs .
( [ sol - met - set ] ) and ( [ bullet - set ] ) ) .
thus , this metric action @xmath314 $ ] corresponds to the aim of the rifle at the target " of the above b - t metaphor a bullet fired at a fixed target in space . "
then , the metric action @xmath63 $ ] is called the * aim - of - the - rifle*. through the dependence of @xmath328 on @xmath63 $ ] , the b - t overlap @xmath370 in eq .
( [ mi - bt ] ) depends on the metric action @xmath363 $ ] . in other words ,
the metric action @xmath363 $ ] produces the bullet @xmath328 , and this bullet @xmath328 produces the b - t overlap @xmath371 , i.e. , [ methodology ] s_met[g _ ] _ bullet _ b (= _ bullet _ target ) . since this sequence in eq .
( [ methodology ] ) implies that the metric action @xmath314 $ ] determines the b - t overlap @xmath359 , various metric actions @xmath314 $ ] are classified by their b - t overlap @xmath359 into two types , ( i ) the overlapping " metric action @xmath32 $ ] satisfying @xmath372 , and ( ii ) the non - overlapping " metric action @xmath373 $ ] satisfying @xmath374 .
because a non - overlapping action @xmath375 $ ] has its empty b - t overlap @xmath376 ( @xmath377 ) , the solution set @xmath326 of the original action @xmath61 $ ] can not be known by using the non - overlapping bullet @xmath378 ( see above ) .
thus , this _
action @xmath373 $ ] should be modified .
then , through the dependence of @xmath328 on @xmath63 $ ] , we can find a suitable overlapping action @xmath379 $ ] by making @xmath380 , namely , by changing ( i ) the form of the metric action @xmath314 $ ] and ( ii ) the values of its parameters . as a result
, we can ( partially ) know the target @xmath325 by using the bullet @xmath381 of the overlapping action @xmath379 $ ] .
this methodology for knowing the target @xmath338 by trying the aim - of - the - rifle @xmath314 $ ] is called the aim - at - target ( aat ) method . "
cll + action & name & role + + @xmath226 $ ] & original action & the true theory " of the space 3-brane within @xmath28 + + @xmath379 $ ] & overlapping action & @xmath382 + [ out - aat ] @xmath383 ~ ( { \rm see ~ the ~ text } ) .
\\ \end{array } \right]$ ] for a better understanding of this aat method , we summarize it for a non - empty b - t overlap @xmath384 ( see table [ out - aat ] ) , as follows : first , the aat method uses the two actions @xmath25 $ ] and @xmath314 $ ] .
second , since the original action @xmath25 $ ] shows the truth ( e.g. , the equation of motion in @xmath28 ) about the space 3-brane occupying @xmath28 , the 3-brane action @xmath25 $ ] is the true theory " of the space 3-brane in @xmath28 .
third , when the b - t overlap @xmath362 is not empty ( i.e. , @xmath372 ) , the metric action @xmath314 $ ] can be desirable .
depending on the mode of existence " ( mathematical / physical ) , the overlapping metric action @xmath379 $ ] has two different implications : * \(a ) an overlapping action @xmath32 $ ] has only the mathematical existence ( me ) " unlike the original action @xmath25 $ ] : this overlapping action is called the me action " @xmath385 $ ] . due to its mathematical existence , this me action
@xmath386 $ ] can not affect the occurrence of any element @xmath30 of the target @xmath387 through the b - t overlap @xmath388 .
in other words , irrespective of the b - t overlap @xmath389 , every element @xmath30 of the target @xmath319 still can occur in the ambient spacetime @xmath28 as a motion of the space 3-brane .
therefore , the me action @xmath390 $ ] making the overlap @xmath391 is a mere tool " for knowing the target @xmath319 of the true theory
@xmath25 $ ] .
* \(b ) an overlapping action @xmath32 $ ] has the physical existence ( pe ) " like the original action @xmath25 $ ] : this overlapping action is called the pe action " @xmath392 $ ] . due to its physical existence ,
this pe action @xmath393 $ ] allows ( forbids ) the occurrence of an element @xmath394 of the target @xmath338 , when this 4dl solution @xmath30 does ( not ) belong to the b - t bullet @xmath395 . in other words ,
only the element @xmath396 of the b - t overlap @xmath397 ( @xmath398 ) can occur in the ambient spacetime @xmath28 as a motion of the space 3-brane . since this decrease in set of possible motions " from @xmath338 to @xmath399 ( @xmath398 ) is similarly found for constitutive equations ( e.g. , ohm s law ) , the pe action @xmath392 $ ] making the overlap @xmath395 produces a constitutive equation " specific to the space 3-brane in @xmath28 .
our aat method using the metric action @xmath26 $ ] seems similarly found in general relativity : in general relativity , the locally inertial coordinates ( lic ) @xmath39 are studied by solving the pde @xmath400 with @xmath401 = 0 $ ] ( this situation corresponds to eqs .
( [ sol - met - set ] ) and ( [ bullet - set ] ) ) .
moreover , these lic @xmath39 are similar to the embedding @xmath18 of the action @xmath25 $ ] , because ( i ) @xmath39 appear in the gr metric " @xmath402 like @xmath18 in @xmath275 , and ( ii ) @xmath39 form an immersion of @xmath174 like @xmath18 due to @xmath403 , where @xmath404 is the vierbein . then , due to these _ similarities _ between @xmath39 and @xmath18 , the analogical reasoning can support that the embedding @xmath18 has the metric action @xmath379 $ ] like the lic @xmath39 having @xmath405 $ ] . for the use of these actions
@xmath379 $ ] and @xmath405 $ ] , see between eqs .
( [ sol - subst ] ) and ( [ equiv - var ] ) . mathematically , the aat method using the overlapping action @xmath379 $ ] consists of two main steps : ( i ) finding a solution @xmath33 of @xmath406 = 0 $ ] as in eq .
( [ sol - met - set ] ) , and next ( ii ) finding a solution @xmath407 of the rifle pde @xmath353 as in eq .
( [ bullet - set ] ) .
since this two - step aat method " is a method for solving the two coupled equations @xmath408 = 0 $ ] and @xmath409 , we can try a different method , i.e. , the insertion of the latter equation @xmath60 into the former [ sol - subst ] .
[ _ f^a _ f^b _ ab^bulk ] = 0 , where the symbol @xmath410 denotes the replacement @xmath411 .
this new equation @xmath412 is called the replaced - equation " ( cf .
( [ oq - eq - repl ] ) and ( [ u - eqs - repl ] ) ) . because solving eq .
( [ sol - subst ] ) is the same as solving the rifle pde @xmath413 of the two - step aat method , the solution set of eq .
( [ sol - subst ] ) [ sol - ovlp ] _
ovlp^(sol ) \ { f^a | s_met^(ovlp ) / g _ @xmath381 of the overlapping action @xmath379 $ ] , namely , [ sol - ovlp = bullet ] _
ovlp^(sol ) = _ bullet^(ovlp ) . in eq . ( [ sol - subst ] )
, the replacement @xmath410 was applied after the functional derivative @xmath414 . here
, we apply the replacement @xmath410 before the derivative @xmath415 .
this produces a new _ functional _ of the embedding @xmath18 [ new - action ] _
met^(ovlp)[f^a ] s_met^(ovlp)[g _ ] |_repl = s_met^(ovlp)[_f^a _ f^b _ ab^bulk ] , implying the lagrangian @xmath416 of this new functional @xmath417 $ ] satisfies ( cf .
( [ final - eff - s ] ) ) [ new - lag ] _
met^(ovlp)(_f^a , ) l_met^(ovlp ) ( g _ , ) |_repl .
of course , it is possible that the original action @xmath61 $ ] in eq .
( [ final - eff - s ] ) takes the form of the new functional @xmath418 = s_{\rm met}^{\rm ( ovlp)}[\partial_\mu f^a \partial_\nu f^b \eta_{ab}^{\rm bulk}]$ ] ( see eq .
( [ s - emb = eq ] ) ) . after the replacement in eq .
( [ new - action ] ) [ g = ind ] g _
= _ f^a _ f^b _
ab^bulk , the functional derivative @xmath415 in eq .
( [ sol - subst ] ) is replaced with @xmath419 .
thus , @xmath420 in eq .
( [ sol - subst ] ) is expressed as [ var - new - action ] [ f^a ] = 0 , which is different from the _ usual _ variational equation [ usual - var ] [ f^a ] = 0 . since @xmath421 in eq .
( [ var - new - action ] ) is the same as @xmath422 in eq .
( [ sol - subst ] ) , the solution set @xmath423 of the former equation [ sol - ovlp2 ] _
ovlp^(sol ) \ { f^a | _
met^(ovlp ) / ( _ f^a _ f^b _ ab^bulk ) = 0 } satisfies [ sol - ovlp2=1 ] _ ovlp^(sol ) = _
ovlp^(sol ) = _ bullet^(ovlp ) ( see eq .
( [ sol - ovlp = bullet ] ) ) . due to the equality @xmath424 in eq .
( [ sol - ovlp2=1 ] ) , the bullet @xmath381 is also produced by the odd - looking equation @xmath425 in eqs .
( [ var - new - action ] ) and ( [ sol - ovlp2 ] ) .
therefore , this equation @xmath426 can replace the equation of motion @xmath427 _ within the overlapping b - t overlap @xmath428 _ ( @xmath429)the equivalence " between these two equations within @xmath430 .
this conclusion becomes more evident , when we compare the bullet @xmath431 with the target @xmath432 in eq .
( [ target - set ] ) . in the above conclusion
, we should be careful in interpreting the odd - looking equation @xmath425 : this odd - looking equation must not be interpreted as the _ equation of motion _ for the space 3-brane , because ( i ) the space 3-brane already has its own equation of motion @xmath433 , and ( ii ) the solution set @xmath434 ( @xmath435 ) in eq .
( [ sol - ovlp2 ] ) may not be equal to the target @xmath338 ( see eq .
( [ eq < emb ] ) ) .
then , the odd - looking equation @xmath436 may be interpreted , at best , as the constitutive equation specific to the space 3-brane ( see above ) . despite this , if we try to know the target @xmath338 by using the new functional @xmath417 $ ] , the odd - looking equation @xmath436 should be used rather than the usual variational equation @xmath437 .
however , in general relativity , it is well known that @xmath438 _ if and only if _
@xmath439 , where @xmath42 is the vierbein satisfying @xmath440 .
( for the mathematical proof , see ref .
@xcite . ) from the similarities of @xmath18 to @xmath39 ( see above ) , we easily confirm the equivalence that [ equiv - var ] [ f^a ] = 0 _ if and only if _ [ f^a ] = 0 .
for the special case of [ s - emb = eq ] s_emb^(3 br)[f^a ] = _ met^(ovlp)[f^a ] = s_met^(ovlp)[_f^a _ f^b _ ab^bulk ] ( see eq .
( [ new - action ] ) ) , the usual variational equation @xmath441 in eq .
( [ usual - var ] ) becomes the equation of motion for the space 3-brane . according to [ var - rel ] _
met^(ovlp ) / f^a & = & + & & - 2 _ \ { _ f^m _ am^bulk } , the odd - looking equation @xmath426 is not a necessary but _ sufficient _ condition for the equation of motion @xmath442 .
this means [ eq < emb ] _ ovlp^(sol ) sp _
target^sp ( thus _ b ^(ovlp ) sp _
target^sp ) , where the superscript sp " denotes the special case @xmath61 = \widetilde{s}_{\rm met}^{\rm ( ovlp)}[f^a]$ ] , and @xmath443 = 0 \ , \}$ ] is the target set for this special case .
the inequality @xmath444 in eq .
( [ eq < emb ] ) supports the above statement that @xmath445 must not be interpreted as the equation of motion for the space 3-brane .
due to @xmath446 , the special case @xmath25 = \widetilde{s}_{\rm met}^{\rm ( ovlp)}[f^a]$ ] in eq .
( [ s - emb = eq ] ) does not have the defect " of [ defect ] _
ovlp^(sol ) - _ target , which means that the bullet set @xmath423 in eq .
( [ sol - ovlp2 ] ) contains elements outside the target set @xmath336 .
this suggests defining the contained metric action " @xmath447 $ ] as an overlapping action @xmath448 $ ] satisfying @xmath449 ( i.e. , without the defect of @xmath450 ) .
the overlapping action @xmath379 $ ] of the special case in eq .
( [ s - emb = eq ] ) is an example of the contained action . as implied in eq .
( [ defect ] ) , for evaluating an me action @xmath390 $ ] , we may use the bullet - target ( b - t ) difference " [ bt - diff ] _ bt _ bullet - _ target together with the b - t overlap @xmath362 in eq . ( [ mi - bt ] ) .
for example , a large " b - t overlap @xmath359 and a small " b - t difference @xmath451 can result in a good " me action @xmath386 $ ] .
now , in terms of symmetries , we study the _ forms _ of the overlapping metric action @xmath379 $ ] ( see table [ sym - aat ] ) .
the b - t overlap @xmath452 represents the maximum knowledge which we can obtain about the target @xmath336 of the original action @xmath25 $ ] by using the chosen metric action @xmath32 $ ] .
ccccc + action & & @xmath245 & & @xmath62 + + @xmath25 $ ] & & o & & o + + @xmath453 $ ] & & o & & o ( x@xmath454 ) + [ sym - aat ] + (o: preserving , x : breaking ) + .symmetries in the aim - at - target ( aat ) method [ cols= " < " , ] first , we consider the symmetries of the target @xmath326 of the original action @xmath61 $ ] , as follows : since the original action @xmath25 $ ] is invariant under @xmath455 and @xmath62 , the definition of the invariance of this action @xmath25 $ ] implies that the solution set @xmath325 of the action @xmath25 $ ] is also invariant under @xmath245 and @xmath62 ( see sec .
[ sec : 3br - eff - action ] ) . by definition ,
the induced - metric set @xmath334 in eq .
( [ set - ind ] ) is invariant under @xmath245 and @xmath62 .
note @xmath456 under @xmath245 , and @xmath457 under @xmath62 .
next , we consider the symmetries of the bullet @xmath458 of the overlapping action @xmath379 $ ] , as follows : since @xmath459 in eq .
( [ g = ind ] ) is already invariant under @xmath460 , the action @xmath32 $ ] has the @xmath245 invariance .
this implies its solution - metric set @xmath461 also has the @xmath245 invariance .
then , since both @xmath33 and @xmath462 are invariant under @xmath460 , @xmath463 is equivalent to @xmath464 , which means the bullet @xmath465 has the @xmath245 invariance like the target @xmath326 .
therefore , since the intersection of two @xmath466-invariant sets is @xmath466-invariant ( @xmath466 : a group ) , the b - t overlap @xmath452 is invariant under @xmath245 . before studying the @xmath62 symmetry properties , we need to re - consider the ( i ) mathematical and ( ii ) physical existences of the overlapping action @xmath379 $ ] ( see the summary of the aat method in sec .
[ sec : aat - method ] ) : first , we study an me action @xmath385 $ ] , which is a mere tool for knowing the target @xmath319 . since this mere tool @xmath385 $ ] can not forbid any element @xmath30 of the target set @xmath325
, we can use a @xmath271-breaking or a @xmath62-preserving action @xmath386 $ ] as long as this me action @xmath386 $ ] provides a considerable information about the target @xmath338 .
for example , we can use a @xmath62-breaking me action @xmath390 $ ] as a mere tool for @xmath338 .
since this action @xmath467 $ ] is not invariant under @xmath62 unlike the original action @xmath25 $ ] , the solution set @xmath468 of the metric action @xmath390 $ ] has an element @xmath469 satisfying [ me - sym - br ] g_^sol _ met^(sol ) me but g_^sol _ met^(sol ) me for an element _
4d^ of diff(4 ) , where @xmath470 with @xmath471 .
this means the solution - metric set @xmath472 breaks the @xmath62 invariance . despite this , if the @xmath62-breaking set @xmath473 contains a
@xmath62 gauge slice " @xmath474 of the @xmath62-preserving set @xmath344 , we can still know the target @xmath326 by , for example , ( i ) finding a solution @xmath30 ( @xmath475 ) of the rifle pde @xmath476 , and ( ii ) applying @xmath62 to this solution @xmath394 , which forms its
@xmath62 gauge orbit " @xmath477
. this aspect is similarly found in a gauge theory , where the gauge invariance is broken by adding a gauge - fixing term .
thus , the breaking of the @xmath62 invariance by the me action @xmath390 $ ] may not be a serious problem for knowing the target @xmath338 ( see the symbol x in table [ sym - aat ] ) .
of course , we can use a @xmath62-preserving me action @xmath467 $ ] as another mere tool for the target @xmath338 .
next , we study a pe action @xmath392 $ ] , which has the physical existence unlike the me action @xmath386 $ ] .
then , since the pe action @xmath478 $ ] produces the constitutive equation , only the element @xmath396 of the b - t overlap @xmath479 ( @xmath480 ) can be a motion of the space 3-brane in @xmath28 , as said in sec .
[ sec : aat - method ] .
the pe action @xmath392 $ ] can determine the @xmath62 symmetry property of the b - t overlap @xmath481 through its bullet @xmath482 : for example , we consider a @xmath271-breaking pe action @xmath393 $ ] , whose solution - metric set @xmath483 has an element @xmath484 satisfying [ pe - sym - br ] g_^sol _ met^(sol ) pe but g_^sol _ met^(sol ) pe for an element _
4d^ of diff(4 ) , where @xmath485 with @xmath486 .
this means the breaking of the @xmath271 invariance by the solution - metric set @xmath487 .
suppose that a 4dl solution @xmath488 ( @xmath489 ) of the original action @xmath51 $ ] satisfies [ f*-in - b ] _
ab^bulk = g_^sol ( i.e. , f_sol ^a _
bullet^(pe ) ) , which means the induced metric @xmath490 ( cf . eq .
( [ ind - metric ] ) ) has the equality [ gam*=g * ] _ ^ = g_^sol .
then , the transformed 4dl solution @xmath491 with @xmath492 satisfies [ f*-b2 ]
_ ^ f_sol ^a _ ^ f_sol ^b _ ab^bulk = g_^sol , where @xmath493 ( @xmath475 ) is an element of the @xmath62 gauge orbit @xmath494 .
due to @xmath495 in eq .
( [ pe - sym - br ] ) , the transformed 4dl solution @xmath493 ( @xmath489 ) does not belong to the bullet @xmath496 ( i.e. , @xmath497 ) unlike the original 4dl solution @xmath488 in eq .
( [ f*-in - b ] ) .
thus , like the bullet @xmath482 , the b - t overlap @xmath498 describing the space 3-brane breaks the @xmath62 invariance , because [ br - ovlp ] _ b ^(pe ) f_sol ^a but _ b ^(pe ) f_sol ^a . to sum up
, the @xmath62-breaking pe action @xmath478 $ ] may imply the @xmath271-breaking b - t overlap @xmath499 .
however , the @xmath62-breaking b - t overlap @xmath500 can cause a physical problem of being contrary to the observed general relativity : since only the element @xmath396 of the b - t overlap @xmath501 can occur in the ambient spacetime @xmath28 as a motion of the space 3-brane ( see sec .
[ sec : aat - method ] ) , the latter result @xmath502 in eq .
( [ br - ovlp ] ) forbids @xmath503 to occur in @xmath28 _ unlike _ the former @xmath504 , which allows @xmath505 to occur in @xmath28 .
thus , due to the approximation @xmath506 in eq .
( [ gr - met = ind ] ) , @xmath507 ( @xmath508 ) can not occur in @xmath28 unlike @xmath509 ( @xmath510 ) .
this means that the primed gr metric @xmath507 can not be a solution of general relativity unlike the unprimed one @xmath511 . as a result ,
general relativity should be a @xmath62-breaking theory , which is falsified by observations .
therefore , it is natural to use only a @xmath62-preserving pe action @xmath392 $ ] which produces the @xmath62-preserving b - t overlap @xmath512 .
in addition , since it is not compulsory that the me action @xmath390 $ ] breaks the @xmath271 invariance , we can choose to use a @xmath271-preserving me action @xmath385 $ ] . to sum up ,
we use only a @xmath62-invariant case of the overlapping action @xmath379 $ ] , irrespective of whether it is an me or pe action . for a @xmath62-invariant overlapping action
@xmath448 $ ] , due to the definition @xmath513 , every element @xmath396 of the b - t overlap @xmath514 satisfies [ ove = bul = sol ] f_ove^a = f_bul^a = f_sol^a with f_bul^a _ bullet^(ovlp ) and f_sol^a _ target , which results in [ df2=g - sol = ind ] _
f_ove^a _ f_ove^b
_ ab^bulk = g_^sol = _ . from eq .
( [ df2=g - sol = ind ] ) , we obtain the equality for the solution metric [ met - sol = ind ] g_^sol = _ for every element f_ove^a of _ b ^(ovlp ) ( cf .
( [ gr - met = ind ] ) ) .
this equality @xmath36 within the b - t overlap @xmath430 means that , below the metric cutoff " @xmath342 , the metric @xmath7 can describe the
emergent field " @xmath216 ( @xmath515 ) , which is derived from the locations ( @xmath516 ) of space quanta occupying @xmath28 .
moreover , due to the equality @xmath50 , the spacetime @xmath517 having this metric @xmath33 is exactly the same as the world volume @xmath17 of the space 3-brane , i.e. , [ st - met = wv ] s_met^ 4d = wv_sq for every element f_ove^a of _ b ^(ovlp ) ( cf .
( [ s - gr = wv ] ) ) .
this equality @xmath518 within the b - t overlap @xmath519 means that , below the cutoff @xmath342 , the spacetime @xmath517 with the metric @xmath520 can describe the
emergent spacetime " @xmath2 , which is formed by the world lines @xmath521 ( @xmath140 ) of many space quanta in @xmath28 . in sum , by eqs .
( [ met - sol = ind ] ) and ( [ st - met = wv ] ) , we have the equality for the two spacetime manifolds [ equal - st ] ( s_met^ 4d , g_^sol ) = ( wv_sq , _ ) within _ b ^(ovlp ) ( below _ met ) .
exact values for the spacetime measurements are provided by the exact or true spacetime @xmath522 , which is the 4d emergent spacetime occupying the ambient spacetime @xmath28 .
now , for the @xmath62-preserving overlapping action @xmath448 $ ] , we consider the form of its lagrangian @xmath523 more closely : as in usual effective theories , this lagrangian @xmath524 having its own uv cutoff @xmath342 ( cf .
( [ met - action ] ) ) can be expressed as [ met - series ] l_met^(ovlp)(_met ; g _ )
= c_k , where the coefficient @xmath525 has no mass dimension , and the local operator @xmath526 of mass dimension @xmath527 consists of the metric @xmath7 and its derivatives . to make @xmath528 invariant under @xmath62
, we assume every operator @xmath526 is invariant under @xmath62 .
since the @xmath62 invariance of the overlapping action @xmath26 $ ] is shared by general relativity , we easily expect this metric action @xmath379 $ ] to contain the einstein - hilbert action ( see eq .
( [ gen - met - lag ] ) ) . generally speaking ,
since the metric lagrangian @xmath529 having the derivatives of @xmath7 can contain at least one dimensionful parameter ( say , @xmath530 ) to maintain its mass dimension @xmath531 = 4 $ ] , the lagrangian @xmath524 becomes the function of the parameter @xmath530 , which has a laurent series for @xmath530 .
thus , this laurent series with @xmath532 can lead to the series like eq .
( [ met - series ] ) , even when the effective - theory nature of @xmath533 is not considered .
if we ( i ) observe at an energy @xmath534 ( @xmath535 ) , and ( ii ) neglect all the operators with @xmath536 , then the error @xmath537 has a size of @xmath538 , implying [ ind - d - negl ] d _ negl 4 + .
this leads to the approximate predictive power that a computation with the error @xmath537 requires only a finite number of operators @xmath526 up to the maximally allowed mass dimension @xmath539 ( @xmath540 ) . when the operator @xmath526 in eq .
( [ met - series ] ) contains @xmath541 derivatives @xmath542 and @xmath543 metrics @xmath7 , the @xmath62 invariance requires the operator @xmath526 to possess @xmath544 inverse metrics @xmath545 for contraction . the mass dimension of @xmath526 satisfies [ ind - d - k ] d_k = [ o_k ] = [ ( ^-1)^ n_+ n_g ^ n
_ ^n_g ] = [ ( dx dx)^- n _ ] = n _ , where the four symbols have correspondences @xmath546 , @xmath547 , @xmath548 and @xmath549
. since the number @xmath544 of inverse metrics @xmath545 should be an integer ( @xmath550 ) , [ even - n ] n _ = 2 ( integer ) , implying @xmath527 is an _ even _ integer due to @xmath551 in eq .
( [ ind - d - k ] ) .
thus , the overlapping lagrangian " @xmath552 in eq . ( [ met - series ] ) has the derivative expansion [ ind - deriv - exp ] l_met^(ovlp)(_met ; g _ ) = _ d_k : even c _ d_k _ met^4 o ( ) ^d_k , where @xmath527 are non - negative even integers . the overlapping lagrangian @xmath553 in eq .
( [ ind - deriv - exp ] ) can have the form of [ gen - met - lag ] l_met^(ovlp ) = c _ 0 _ met^4 + c _ 2 _ met^2 r + c _
4^(1 ) r^ 2 + c _ 4^(2 ) r _
r^ + c _ 4^(3 ) g^ _ r _ r + , where all the coefficients ( e.g. , @xmath554 , @xmath555 ) are dimensionless , and both of the covariant derivative @xmath291 and the curvature quantities ( e.g. , @xmath286 ) are built from the metric @xmath7 ( see ref .
according to observations , our universe contains various particles ( e.g. , leptons ) which are different in kind from space quanta . to distinguish those particles from the space quanta ,
we coin a new term * occupant quantum * ( oq ) denoting any particle which ( i ) differs from space quanta , and ( ii ) occupies the space 3-brane without departing from it ( i.e. , the confinement of the occupant quantum to the space 3-brane ) . to sum up
, our universe can be regarded as a _ composite system _ which consists of space quanta and occupant quanta , moving within the ambient spacetime @xmath28 .
since space quantum is more fundamental than graviton , there can be a scenario that every particle of the standard model ( sm ) is a bound state of occupant quanta
. however , there can be another scenario that each sm particle is identified with a single occupant quantum .
besides these , there can be various other scenarios . despite this ,
from now on , we will consider only the _ low - energy _ spectrum ( e.g. , the sm particles ) of occupant quanta which can be observed at low enough energies : since each of these observable occupant quanta is confined to the world volume @xmath2 of the space 3-brane , it is described by a function @xmath556 whose domain is the world volume @xmath16 . for a brane - chart @xmath174 of @xmath17 ,
the brane - field " @xmath556 on @xmath17 is represented as the function @xmath557 of the four coordinates @xmath174 .
the _ value _ @xmath558 at a point @xmath559 is either ( i ) a brane - tensor of a type " ( e.g. , a scalar ) of @xmath17 , or ( ii ) a brane - spinor " ( e.g. , a weyl spinor ) of the brane lorentz group " @xmath560 at the point @xmath69 .
the vierbein @xmath561 satisfying @xmath562 can be used in the action for brane - spinors .
suppose that the bosons and fermions of the standard model are described by their corresponding brane - fields @xmath563 . like the induced metric @xmath564 , all the sm brane - fields @xmath565 are invariant ( i.e. , bulk - scalars " ) under every b@xmath179@xmath183 transformation @xmath566 between the bulk - charts @xmath133 and @xmath182 of @xmath28 .
the action @xmath567 for the observable occupant quanta @xmath568 can be expressed as [ br - oq - act ] s_oq^(3 br)[_oq , f^a ] _
wv_sq d^ 4 x l_oq^(3 br ) ( _ oq , _
f^a , ) , where @xmath569 .
this action @xmath570 $ ] is assumed to be invariant under @xmath571 and @xmath62 like the 3-brane action @xmath25 $ ] .
of course , the action @xmath572 in eq .
( [ br - oq - act ] ) may depend on a bulk - field " @xmath573 of the bulk spacetime @xmath28 , whose field point @xmath133 should satisfy @xmath221 .
for example , when @xmath573 is a bulk - tensor ( e.g. , a @xmath126-dimensional vector ) , it can appear in the action @xmath567 through its pullback @xmath574 at sufficiently low energies .
finally , the * original " universe action * @xmath575 $ ] at low energies is written as [ br - univ - act ] s_univ^(3 br)[f^a , _ oq ] = s_emb^(3 br)[f^a ] + s_oq^(3 br)[_oq , f^a ] , where the integral for the 3-brane action @xmath51 $ ] shares the same set @xmath17 with that for @xmath576 $ ] in eq .
( [ br - oq - act ] ) . for the original action
@xmath577 $ ] in eq .
( [ br - univ - act ] ) , its * universe target * @xmath578 is defined as [ u - target ] _ target^(univ ) \ { f_sol^a : 4dl embedding where ( @xmath579 ) is a solution of the coupled euler - lagrange ( e - l ) equations [ br - eqs ] ( s_univ^(3 br ) / f^a)[f^a , _ oq ] = 0 and ( s_univ^(3 br ) / _ oq)[f^a , _ oq ] = 0 .
although the element @xmath30 of the set @xmath580 satisfies the different equation ( i.e. , @xmath581 ) from @xmath582 for the target @xmath338 in eq .
( [ target - set ] ) , the solution @xmath30 in eq .
( [ u - target ] ) is still called a 4d - lorentzian ( 4dl ) solution . " since the universe target @xmath583 in eq .
( [ u - target ] ) is defined similarly to the target @xmath338 , we can similarly apply the aat method in order to know the universe target @xmath584 , as follows : as in sec .
[ sec : aat - method ] , the knowledge about the universe target @xmath585 is related to the * universe induced - metric set * [ u - set - ind ] _
ind^(univ ) \ { _ | _ = _ f_sol^a _ f_sol^b _ ab^bulk for every f_sol^a _ target^(univ ) } . to study this universe induced - metric set @xmath586 as in sec .
[ sec : aat - method ] , we impose three requirements on the * overlapping " universe action * @xmath587 : * for the study of @xmath588 , the overlapping action @xmath589 is a functional of the 4d lorentzian metric @xmath7 on the 4d manifold @xmath590 . * the spacetime @xmath590 for the action @xmath591 satisfies [ equal - u - st ] s_univ^ 4d = wv_sq ( cf . eqs .
( [ st - met = wv ] ) and ( [ equal - st - u ] ) ) . * the solution @xmath592 ( called the u - metric " ) of the equation @xmath593 ( see eq .
( [ u - cartridge ] ) ) satisfies [ equal - u - met ] g_^sol u = _ ( cf .
( [ met - sol = ind ] ) and ( [ equal - st - u ] ) ) .
thus , since this induced metric @xmath4 depends on the observable occupant quanta through the 4dl solution @xmath394 due to eqs .
( [ u - target ] ) and ( [ u - set - ind ] ) , it is natural to assume that the overlapping universe action @xmath594 depends on these occupant quanta .
therefore , we consider the overlapping universe action of the form [ univ - act ] s_univ^(ovlp)[g _ , _ oq ] s_met^(ovlp)[g _ ] + s_oq^(ovlp)[_oq , g _ ] , where the occupant - quantum ( oq ) action " [ oq - act ] s_oq^(ovlp)[_oq , g _ ] _ s_univ^4d d^ 4 x l_oq^(ovlp)(_oq , g _ , ) , and the metric action [ univ - met - act ] s_met^(ovlp)[g _ ] _ s_univ^4d d^ 4 x l_met^(ovlp)(_met ; g _ ) . because this metric action @xmath379 $ ] defined for @xmath595 will be chosen to be invariant under @xmath596 and @xmath62 ( see below ) , its lagrangian @xmath553 in eq .
( [ univ - met - act ] ) has the same form as the lagrangian in eqs .
( [ ind - deriv - exp ] ) and ( [ gen - met - lag])we use the same notations .
like @xmath7 describing @xmath4 through @xmath597 , each occupant - quantum ( oq ) field " @xmath598 in the overlapping action @xmath599 $ ] describes its counterpart @xmath556 through [ equal - oq ] _ oq^sol = _
oq^sol ( see eqs .
( [ u - eqs ] ) and ( [ u - cartridge ] ) ) , where @xmath600 is a part of the solution ( @xmath601 ) of the coupled e - l equations [ u - eqs ] ( s_univ^(ovlp ) / g_)[g _ , _ oq ] = 0 and ( s_univ^(ovlp ) / _ oq)[g _ , _ oq ] = 0 .
when @xmath602 in eq .
( [ u - eqs ] ) is compared with @xmath603 in eq .
( [ br - eqs ] ) , we can find a simple method for achieving the above equality @xmath604 under the assumption @xmath605 in eq .
( [ equal - u - met ] ) , as follows : the original oq action @xmath576 $ ] in eq .
( [ br - oq - act ] ) can satisfy , at least at low enough energies , [ oq - act - repl ] s_oq^(3 br)[_oq , f^a ] = _ oq^(ovlp)[_oq , f^a ] s_oq^(ovlp)[_oq , _ f^a _
ab^bulk ] , which is obtained by the replacements ( i ) @xmath606 and ( ii ) @xmath607 in the overlapping oq action @xmath608 $ ] in eq .
( [ oq - act ] ) . due to
( [ oq - act - repl ] ) , the solution @xmath609 of @xmath610 = 0 $ ] from eq .
( [ br - eqs ] ) is also a solution of the replaced - equation from eq .
( [ u - eqs ] ) [ oq - eq - repl ] s_oq^(ovlp ) / _ oq |_repl ( s_oq^(ovlp ) / _ oq)[_oq , _ f_sol^a _ f_sol^b _ ab^bulk ] = 0 , which contains @xmath30 unlike other replaced - equations in eqs .
( [ sol - subst ] ) and ( [ u - eqs - repl ] ) due to the assumption @xmath605 . in this manner , the equality @xmath611 in eq .
( [ equal - oq ] ) is achieved .
for this equality @xmath612 , the oq field @xmath598 shares the same @xmath613 and @xmath62 symmetry properties with its corresponding brane - field @xmath556 .
for example , the oq field @xmath598 is a @xmath62-tensor or @xmath560-spinor of the spacetime @xmath595 like its counterpart @xmath556 . of course
, the equality @xmath614 may have a limited validity like @xmath36 in eq .
( [ met - sol = ind ] ) , which is valid only for the b - t overlap @xmath430 ( @xmath480 ) .
due to eq .
( [ oq - act - repl ] ) , the original universe action @xmath615 $ ] in eq .
( [ br - univ - act ] ) can satisfy [ new - br - univ - act ] s_univ^(3 br)[f^a , _ oq ] = s_emb^(3 br)[f^a ] + _ oq^(ovlp)[_oq , f^a ] at low enough energies .
principle @xmath616 gives the equation of motion for the space 3-brane [ univ - f - eq ] _ ( t_3 br ^ _
ab^bulk ) + = _ ( t_oq^ _ f^b _ ab^bulk ) , where [ br - oq - em ] t_oq - .
since the oq action @xmath617 $ ] is added to the 3-brane action @xmath51 $ ] , the equation of motion in eq .
( [ univ - f - eq ] ) is changed from eq .
( [ f - eq ] ) .
as in sec .
[ sec : aat - method ] , for the overlapping universe action @xmath599 $ ] in eq .
( [ univ - act ] ) , its * universe cartridge * @xmath618 is defined as [ u - cartridge ] _
univ^(sol ) \ { g_^sol
u | ( s_univ^(ovlp ) / g_)[g_^sol u , _
oq^sol ] = 0 } , where ( @xmath619 ) is the solution of the coupled e - l equations in eq .
( [ u - eqs ] ) . then , the * universe bullet * @xmath620 of the overlapping action @xmath621 $ ] is defined as [ u - bullet ] _
bullet^(univ ) \ { f^a | _ f^a _ f^b _ ab^bulk = g_^sol u for every g_^sol u _
univ^(sol ) } f_space . like the universe target @xmath622 in eq .
( [ u - target ] ) , the universe bullet @xmath623 depends on the occupant quanta through the u - metric @xmath624 in eq .
( [ u - bullet ] ) , because this solution metric @xmath592 depends on the occupant quanta @xmath598 through , e.g. , the @xmath598-dependent equation @xmath625 = 0 $ ] in eq .
( [ u - eqs ] ) . because the action @xmath626 $ ] in eq . ( [ univ - act
] ) is an _ overlapping _ one , the universe b - t overlap " @xmath627 is not the empty set , i.e. , [ u - ove ] _
b t^(univ ) , where @xmath628 is assumed to contain a low - energy motion ( e.g. , @xmath629 ) which the space 3-brane can perform in the ambient spacetime @xmath28 . as in sec .
[ sec : aat - method ] , the universe b - t overlap @xmath630 is the maximum knowledge which we can obtain about the universe target @xmath578 by using the overlapping universe action @xmath631 $ ] .
until now , we have presented the two - step aat method " for the overlapping universe action @xmath631 $ ] ( cf .
[ sec : aat - method ] ) : * step 1 : finding a solution @xmath592 of the coupled equations @xmath632 = 0 $ ] and @xmath633 = 0 $ ] in eq .
( [ u - eqs ] ) , and next * step 2 : finding a solution @xmath30 of the new rifle pde @xmath634 . instead of this two - step aat method , as in sec .
[ sec : aat - method ] , we try another method of eliminating the metric @xmath7 from those e - l equations @xmath635 and @xmath636 by inserting the pde @xmath315 into them .
namely , we solve the coupled replaced - equations [ u - eqs - repl ] s_univ^(ovlp ) / g _ |_repl = 0 and s_univ^(ovlp ) / _ oq |_repl = 0 ( cf .
( [ sol - subst ] ) ) , where @xmath637 $ ] for @xmath638 .
the former replaced - equation @xmath639 is expressed as @xmath640 , where @xmath641 \stackrel{\rm def}{= } s_{\rm univ}^{\rm ( ovlp)}[\partial_\mu f^a \partial_\nu f^b \eta_{ab}^{\rm bulk } , \psi_{\rm oq}]$ ] ( cf .
( [ new - action ] ) and ( [ var - new - action ] ) ) .
since solving the coupled replaced - equations in eq .
( [ u - eqs - repl ] ) is the same as solving the new rifle pde @xmath642 of the two - step aat method , the solution set for eq .
( [ u - eqs - repl ] ) [ u - sol - ovlp ] _
univ^(sol ) \ { f^a | s_univ^(ovlp ) / g _ |_repl = 0 and s_univ^(ovlp ) / _ oq |_repl = 0 } is equal to the universe bullet @xmath620 in eq .
( [ u - bullet ] ) , namely , [ u - sol = bullet ] _ univ^(sol ) = _ bullet^(univ ) ( cf .
( [ sol - ovlp = bullet ] ) ) .
this equality @xmath643 means that the universe bullet @xmath644 is also produced by the coupled replaced - equations @xmath645 .
therefore , these replaced - equations @xmath646 can be used instead of the e - l equations @xmath647 in eq .
( [ br - eqs ] ) _ within the universe b - t overlap @xmath648 _ ( see below eq .
( [ sol - ovlp2=1 ] ) ) . in other words , within this b - t overlap @xmath648 , the replaced - equations from @xmath599 $ ] are equivalent " to the e - l equations from @xmath649 $ ] .
when a single " overlapping action @xmath650 $ ] is discovered as a result of investigation , we can assume that the replaced - equations from this discovered action @xmath626 $ ] are applied , at least , to many and various motions which the space 3-brane can perform in the ambient spacetime @xmath28 at low enough energies .
namely , the aat method using the single discovered action @xmath650 $ ] is valid for those many and various low - energy motions of the space 3-brane .
( for a further study , see our next paper @xcite . ) of course , the discovered action @xmath599 $ ] can change , depending on observation energies .
suppose a low - energy motion of the space 3-brane is described by a 4dl solution @xmath295 .
then , each momentum @xmath651 in the fourier transform of @xmath295 satisfies @xmath652 for all @xmath653 . in this fourier - transform context
, the low - energy motion @xmath293 is expressed as [ le - motion ] || _ cont .
for this low - energy motion @xmath30 of @xmath654 , the u - metric @xmath592 ( @xmath515 by eq .
( [ equal - u - met ] ) ) is also expressed as @xmath655 . as in sec .
[ sec : form - met ] , we choose the invariant case " that the overlapping action @xmath631 $ ] in eq .
( [ univ - act ] ) is invariant under @xmath596 and @xmath62 like the original one @xmath615 $ ] , irrespective of whether @xmath599 $ ] is an me or pe action .
thus , the metric lagrangian @xmath528 in @xmath621 $ ] shares the same form with that in eqs .
( [ ind - deriv - exp ] ) and ( [ gen - met - lag ] ) .
then , the @xmath62-invariant universe action @xmath599 $ ] can contain ( i ) the einstein - hilbert action and ( ii ) the action for matter ( i.e. , occupant quanta ) , both of which are the essential parts of general relativity . within the region @xmath655 in eq .
( [ le - motion ] ) , we have the equality for the two spacetime manifolds ( see eqs .
( [ equal - u - st ] ) and ( [ equal - u - met ] ) ) [ equal - st - u ] ( s_univ^ 4d , g_^sol u ) = ( wv_sq , _ ) .
note the 4d emergent spacetime @xmath656 is determined by the 4dl solution @xmath30 for the e - l equations @xmath657 in eq .
( [ u - target ] ) .
this emergent manifold @xmath658 occupying @xmath28 is the exact or true spacetime which provides exact values for our spacetime measurements ( see below eq .
( [ equal - st ] ) ) . as said below eq .
( [ univ - met - act ] ) , since the metric lagrangian @xmath523 in eq .
( [ univ - met - act ] ) shares the same form with that in eqs .
( [ ind - deriv - exp ] ) and ( [ gen - met - lag ] ) , we use the same notations . due to the power - law behaviors
@xmath659 in eqs .
( [ ind - deriv - exp ] ) and ( [ gen - met - lag ] ) , the most dominant term in @xmath660 for @xmath629 is the @xmath661 lagrangian @xmath662 , which contributes to a cosmological constant .
moreover , the next dominant term is the @xmath663 lagrangian @xmath664 , which contains only the two - derivative terms of the metric @xmath7 .
as assumed before , the aat method using the overlapping universe action @xmath599 $ ] is applied to various low - energy motions ( i.e. , @xmath654 ) of the space 3-brane . within the region @xmath655 , we can find a low - energy region @xmath629 ( @xmath665 ) ) in which the overlapping action @xmath621 $ ] has the approximation [ 2d - approx - u ] s_univ^(ovlp)[g _ , _ oq ] s_univ^ ( 2)[g _ , _
oq^low ] , where [ 2d - act - u ] s_univ^ ( 2)[g _ , _
oq^low ] & & s_met^ ( 2)[g _ ] + s_oq^(low)[_oq^low , g _ ] , + [ 2-der - act ] s_met^ ( 2 ) [ g _ ] & & _ s_univ^4d ( 2 ) d^ 4 x ( c _ 0 _ met^4 + c _ 2 _ met^2 r ) . in eq .
( [ 2d - act - u ] ) , the new oq action @xmath666 $ ] is the low - energy approximation of its full theory @xmath667 $ ] in eq .
( [ oq - act ] ) .
namely , @xmath668 $ ] contains only the low - dimension interactions of @xmath669 $ ] which are not negligible in the low - energy region @xmath670 .
of course , some heavy oq fields ( say , @xmath671 ) appearing in @xmath608 $ ] may be decoupled from its low - energy approximation @xmath672 $ ] . in eq .
( [ 2-der - act ] ) , the integral for @xmath673 $ ] undergoes the replacement @xmath674 , which is also undergone by the integral for @xmath675 $ ] . due to the approximate equality @xmath676 in eq .
( [ 2d - approx - u ] ) , the solution ( @xmath677 ) of the exact " equations @xmath678 and @xmath679 in eq .
( [ u - eqs ] ) has the approximate equalities [ approx - g - oq - u ] g_^sol u g_^sol ( 2 ) and _
oq^sol _ oq^sol ( 2 ) , where ( @xmath680 ) is the solution of the approximate " equations @xmath681 and @xmath682 .
in addition , @xmath683 in eq .
( [ approx - g - oq - u ] ) implies the approximate equality for the spacetime [ approx - st - u ] s_univ^ 4d s_univ^ 4d ( 2 ) . in the low - energy region @xmath629 , the approximate equation
@xmath684 is the same as einstein s equation with the three parameters @xmath554 , @xmath555 and @xmath342 [ eins - eq ] g _ - g _
= t_^(low ) , where @xmath685 corresponds to @xmath686 of the ordinary einstein s equation , and [ oq - em ] t_^(low ) - . for @xmath687 , the tensor @xmath688 in eq . (
[ br - oq - em ] ) satisfies @xmath689 at the low energies @xmath629 .
suppose that the @xmath690 " term in eq .
( [ eins - eq ] ) is negligible as in the observed @xmath691cdm model @xcite .
then , a spherical massive object can produce the schwarzschild metric @xmath692 , which leads to the gravitational potential @xmath693 in the newtonian limit .
since this potential @xmath694 ( @xmath695 ) depends on @xmath696 strongly , the value of @xmath696 can be easily determined by the comparison with observed data . in eq .
( [ 2-der - act ] ) , when the coefficient @xmath697 in the integrand satisfies [ cutoff ] c _ 2 _ met^2 = 1 / 16 g_n ( i.e. , _ met = m_p/ ) , the approximate metric action @xmath698 $ ] may not be distinguished from the einstein - hilbert action @xmath699 $ ] with a dark energy ( de ) density @xmath700 [ eh+de ] s_eh^(de ) [ * g * _ ] _
s_gr d^ 4 x ( _ * r * _ / 16 g_n - _ de ) , where the ricci scalar @xmath701 of general relativity ( gr ) is built from the gr metric @xmath57 . to be more concrete ,
we consider in what situation general relativity is valid : since using general relativity of the @xmath702 terms means neglecting all the higher - order @xmath703 terms of the lagrangian @xmath523 in eq .
( [ univ - met - act ] ) , it is important to estimate the size of @xmath704 . as a measure of @xmath705 , kretschmann scalar @xmath706
is used due to @xmath707 . in this context
, we deal with an extremely strong gravity related to a schwarzschild black hole of mass @xmath708 , whose kretschmann scalar is @xmath709 @xcite . outside the schwarzschild radius @xmath710 ( i.e. , @xmath711 ) , the scalar satisfies @xmath712 , which leads to @xmath713 due to @xmath714 and @xmath715 .
then , for @xmath716 ( @xmath717 ) , the result @xmath718 implies that general relativity is valid outside the event horizon at @xmath719 .
meanwhile , inside this event horizon , there is a radius @xmath720 satisfying @xmath721 , which produces @xmath722 ( @xmath723 ) . for @xmath724 , @xmath725
( i.e. , @xmath726 ) implies that general relativity is not valid far inside the event horizon .
similarly , for @xmath727 with @xmath728 , @xmath729 implies that the continuum approximation of the quasi-3d object breaks down the above black hole may not have the singularity at its center @xmath730 . to sum up , in the low - energy region @xmath670 , general relativity can be a good approximation of the overlapping universe action @xmath631 $ ] , which is an essential part of the aat method for studying the original universe action @xmath731 $ ] ( i.e. , the principle governing the motions of the space 3-brane ) .
note that our spacetime @xmath17 ( @xmath732 ) can have its own gravity ( i.e. , general relativity ) although the ambient spacetime @xmath10 does not have any bulk gravity " ( i.e. , @xmath733 ) .
in the case of [ met-2=eh ] s_met^ ( 2)[*g * _ ] = s_eh^(de)[*g * _ ] , the solution metric @xmath734 and the spacetime @xmath56 of general relativity have the equalities [ gr - met = g-2 ] * g*_^sol & = & g_^sol ( 2 ) ( g_^sol u = _ due to eqs .
( [ equal - st - u ] ) and ( [ approx - g - oq - u ] ) ) , + [ gr - st = st-2 ] s_gr & = & s_univ^ 4d ( 2 ) ( s_univ^ 4d = wv_sq due to eqs .
( [ equal - st - u ] ) and ( [ approx - st - u ] ) ) . according to eqs .
( [ gr - met = g-2 ] ) and ( [ gr - st = st-2 ] ) , the spacetime ( @xmath735 ) of general relativity can be at least a good approximation of the exact or true spacetime ( @xmath9 ) , which is formed by many space quanta occupying the ambient spacetime @xmath28 .
this supports the space - quantum hypothesis in eq .
( [ sq - hyp ] ) . if the exact equality @xmath736 really happens instead of the approximate one in eq .
( [ 2d - approx - u ] ) , the approximate equality " signs @xmath737 in eqs .
( [ approx - g - oq - u ] ) , ( [ approx - st - u ] ) , ( [ gr - met = g-2 ] ) and ( [ gr - st = st-2 ] ) are replaced with the equality signs @xmath738 .
finally , until now , we have considered only the special situation that the ambient spacetime @xmath10 is the flat manifold @xmath739 , in which the inertial bulk observer @xmath135 uses the inertial bulk - coordinates @xmath133 ( see sec .
[ sec : cont - approx ] ) .
however , the ambient spacetime @xmath10 can be a _
curved _ manifold having a general bulk metric @xmath740 , implying the replacements [ rep - s ] ^d_amb & & s^d_amb , + [ rep - g ] _
ab^bulk & & g_ab^bulk ,
+ [ rep - diff ] iso(1 , d_amb - 1 ) & & diff(d_amb ) . for these replacements ,
our previous studies can be extended similarly .
the topology of the ambient spacetime @xmath10 may be , for example , @xmath741 or @xmath742 , where @xmath743 is a ( @xmath744)-dimensional spacelike torus . for @xmath745 , when the topology of @xmath10 is @xmath746 , the topology of the space 3-brane can be @xmath747 , implying the world volume @xmath17 of this space 3-brane may be spatially flat .
then , due to @xmath748 in eq .
( [ gr - st = st-2 ] ) , the corresponding spacetime @xmath56 of general relativity may be spatially flat , which can agree with the observed @xmath691cdm model @xcite . for @xmath749
, the same conclusions can be reached even for the topology @xmath750 of @xmath0 , when the size of this torus @xmath743 is much smaller than the distance @xmath11 between space quanta at low energies @xmath751 , the bulk spacetime @xmath10 can be observed as if its topology were @xmath746 .
we would like to thank prof .
k. y. lee , prof .
i. park , dr . j .- h .
park , dr . h .- s .
lee , and prof .
e. j. chun for reading this paper and expressing their opinions .
y. c. fung , _ a first course in continuum mechanics : for physical and biological scientists and engineers _ , 3rd ed .
( prentice hall , 1994 ) ; p.
haupt , _ continuum mechanics and theory of materials _
( springer - verlag , 2000 ) ; and i. g. currie , _ fundamental mechanics of fluids _ , 2nd ed .
( mcgraw - hill , new york , 1993 ) .
planck collaboration , _ planck _ 2015 results .
cosmological parameters , " arxiv:1502.01589 [ astro-ph.co ] ; k. a. olive _ et al . _ ( particle data group ) , chin . phys .
c , * 38 * , 090001 ( 2014 ) ; and c. l. bennett _ et al .
_ , nine - year wilkinson microwave anisotropy probe ( _ wmap _ ) observations : final maps and results , " arxiv:1212.5225 [ astro-ph.co ] . | we build a model of our spacetime by assuming new particles called space quanta . " in the ambient or bulk spacetime @xmath0 ( @xmath1 ) , a multitude of space quanta form a nearly three - dimensional object , whose continuum approximation is called the space 3-brane . the world volume @xmath2 of this space 3-brane is described by an embedding @xmath3 , which produces the induced metric @xmath4 on the world volume @xmath2 .
this emergent spacetime ( @xmath5 ) from the many space quanta is proposed as the particle model of our spacetime . to study our spacetime ( @xmath6 ) , we construct what we call the aim - at - target ( aat ) method , which introduces an action for a 4d metric @xmath7 .
this metric action from the aat method can lead to general relativity at low enough energies .
the spacetime ( @xmath8 ) of general relativity is , at least , a good approximation to the exact or true spacetime ( @xmath9 ) of our universe . 0 |
for last couple of decades , magnetic shape memory alloys ( msmas ) have been fascinating the researchers because these materials combine different degrees of freedom like structural , magnetic , elastic , caloric etc .
, in a single phase .
coupling of micro - structure with magneto - crystalline anisotropy enables increase of magnetic field - induced strain ( mfis ) , which makes this class of smart materials suitable for micro - mechanical automotive sensors and actuators applications .
other interesting multifunctional properties exhibited by this class of materials are conventional and inverse magneto - caloric effect @xcite , barocaloric @xcite , elastocaloric @xcite , giant magnetoregistance @xcite , exchange bias @xcite , kinetic arrest @xcite , spin - glass @xcite and strain - glass @xcite . in the heart of these applications
is the magnetic - field driven martensitic transformation .
the magnitude and sign of the magnetization difference ( @xmath1 ) between the high temperature austenite phase and the low temperature martensite phase is a key quantity responsible for the functional properties mentioned above .
for example , zeeman energy plays a crucial role in realizing the martensitic transformation in ferromagnetic shape memory alloys .
an externally applied magnetic field performs two different tasks inside msma .
firstly , if the direction of the applied field is different from the easy magnetic axis of the material , the field tries to rotate the magnetization direction along it against the force associated with magnetic anisotropy .
secondly , the field generates driving force across the twin boundaries between martensitic variants , the associated energy is known as zeeman energy . since the energy required to move martensitic domains is lower than the magnetocrystralline anisotropy energy in msma , the pressure created by zeeman energy increases the volume fraction of favorably oriented martensitic variants and lowers the energy of the product phase .
it has been found that a large @xmath1 between the two phases in the presence of moderate external magnetic field @xmath2 , i.e. , the zeeman term @xmath3 , drives the motions of the martensitic domains facilitating the martensitic transformation @xcite . on the other hand ,
the sign of @xmath1 is related to the conventional or inverse magnetocaloric effect @xcite . in case of the former ,
a magnetic field will cause a decrease in entropy when applied isothermally and an increase in temperature when applied adiabatically . in case of the later , an increase in entropy will effect a decrease of temperature leading to `` magnetic cooling '' , a phenomenon which can be exploited for green technology in refrigeration . for the inverse magnetocaloric effect
, the magnetization in the martensitic phase has to be lower than that in the austenite phase .
recent research on msmas thus , have focused on exploring materials where @xmath1 can be suitably tuned in order to obtain large zeeman energy with a small field and/or a substantial inverse magneto - caloric effect .
a decade ago , ni@xmath0mnga was discovered as the first promising msma @xcite .
however , ni@xmath0mnga in the stoichiometric composition of 2:1:1 shows only conventional magneto - caloric effect .
moreover , @xmath1 for the system is rather small .
other materials in the ni@xmath0mn_x _ ( _ x_= al , in , sn ) series do not even posses shape memory properties at their stoichiometric compositions .
a lot of research has taken place since then to obtain shape memory and/or inverse magneto - caloric effect in ni@xmath0mnx systems by tuning the electron to atom ( @xmath4 ) ratio by either changing concentration ratio of ni , mn and x @xcite or by doping with another element @xcite .
these approaches led to the investigations into the potentials of mn@xmath0ni_x _ ( _ x_= al , ga , in , sn ) systems .
the exploration in mn@xmath0ni_x _ materials was initiated in 1987 by helmholdt _
et al . _ with their investigation in the structural and the magnetic properties of mn@xmath0nisn form neutron diffraction data @xcite .
their study concluded that unlike the ni@xmath0mn_x _ materials which form the usual heusler structure , mn@xmath0nisn crystallizes in inverse heusler structure . the later structure forces one of the tetrahedral mn atoms to interchange its position with octahedral ni atom making the two mn atoms nearest neighbors .
the next significant step was investigations into the shape memory related properties of mn@xmath0niga , conducted by liu _
the material exhibits an excellent two - way shape memory effect where the martensitic transformation temperature ( t@xmath5 270 k ) is close to room temperature and the curie temperature ( t@xmath6 588 k ) is quite high @xcite .
the last two parameters of mn@xmath0niga are better than those of ni@xmath0mnga .
this study hinted at the better functional properties of this material compared to the prototype and thus , generated immense interests into investigations of the mn@xmath0ni_x _ materials with the possibilities of discovering new msmas suitable for practical applications .
it is noteworthy to mention that mn@xmath0nisn also has very high curie temperature ( t@xmath6 530 k ) @xcite .
x - ray diffraction experiments confirmed that mn@xmath0niga also forms an inverse heusler structure in the austenite phase .
experimental reports on the structure and shape memory property of mn@xmath0nial and mn@xmath0niin are not yet available .
however theoretical calculations on mn@xmath0ni_x _ confirm that all the materials exhibit shape memory effect and prefer inverse heusler structure @xcite .
the direct consequence for crystallizing in this structure is lower saturation magnetization due to opposite and unequal moments of nearest neighbor mn atoms , as compared to prototype ni@xmath0mnga , where all the constituent atoms couple ferromagnetically . for mn@xmath0nial and mn@xmath0niga , total
magnetization reported is @xmath71 @xmath8 and for mn@xmath0niin and mn@xmath0nisn those are @xmath70.5
@xmath8 @xcite , while ni@xmath0mnga posses a high moments of @xmath74 @xmath8 .
since the mfis is directly proportional to the total magnetization , lowering of the later would results in a smaller mfis in mn@xmath0niga ( @xmath74@xmath9 ) , whereas largest mfis are observed in nearly stoichiometric composition of nimnga msma is about 10@xmath9 @xcite . on the contrary ,
helmholdt _ et al .
_ observed a higher value of total magnetization , nearly 2.5 @xmath8 , in cubic austenite phase of mn@xmath0nisn @xcite .
they showed that within the inverse heusler structure , there is significant anti - site disorder among the ni and mn atoms at the tetrahedral positions .
recent neutron diffraction studies @xcite and first principles density functional theory ( dft ) calculations on mn@xmath0niga @xcite and mn@xmath0nisn @xcite also confirmed the presence of anti - site disorder in these materials .
dft calculations conclusively showed that such anti - site disorder is responsible for a significantly large magnetic moment in the high temperature phase .
the calculations also showed that mn@xmath0niga would exhibit inverse magneto - caloric effect @xcite .
motivated by these discoveries , in this communication , we perform dft calculations to investigate the magnetic properties of the four mn@xmath0ni_x _ alloys which are potential msmas , in crystal structures relevant to the austenite and martensite phases , with and without anti - site disorder .
we specifically looked at the trends in @xmath1 across the series .
we analyze the results by computing the electronic structures and the inter - atomic exchange interactions for each of the cases .
the outcome of this systematic study indicates that the mn@xmath0ni_x _ alloys may have better functionalities than the prototype msmas , which is driven by the anti - site disorder inherent to these systems , an important feat relevant in todays technology .
the paper is organized as follows : in section ii , we discuss the computational details .
section iii contains the results and subsequent discussions on the magnetic properties , the electronic structure and the inter - atomic exchange interactions .
summary and importance of this work are described in the conclusions section .
the electronic structures of the materials concerned have been calculated using full - potential based spin - polarized relativistic korringa - kohn - rostoker ( spr - kkr ) green s function method @xcite .
the local spin density approximation ( lsda ) as parameterized by vosko - wilk - nusair ( vwn ) was used as the exchange - correlation part of the potential to solve the kohn - sham equation @xcite . the angular momentum cut - off to the plane wave
was taken to be @xmath10=3 .
the brillouin zone integrations have been carried out on a uniform 24@xmath1124@xmath1124 _ k_-mesh .
the green s function was calculated for 30 complex energy points distributed exponentially on a semicircular contour .
the energy convergence criterion was set to 10@xmath12 ry for the self - consistent cycles .
the coherent potential approximation ( cpa ) was used to incorporate the effects of disorder @xcite .
the magnetic pair exchange interactions have been calculated with multiple scattering green function formalism as implemented in spr - kkr code @xcite .
spin fluctuation theories for metals map the complicated itinerant electron systems onto an effective heisenberg hamiltonian having the classical spins as @xmath13 the indices @xmath14 and @xmath15 represent different sublattices , _ i _ and _ j _ denote the atomic positions and @xmath16 is the unit vector along the direction of magnetic moments at site _
i _ belonging to sublattice @xmath14 .
the size of the magnetic moments are included in the exchange interaction parameter @xmath17 .
the exchange parameters are computed from energy difference due to the small orientation of a pair of spins resulting a perturbation in spin - density which within the formulation of lichtenstein _ et al . _
@xcite based on magnetic force theorem @xcite takes the following form @xmath18 where @xmath19= @xmath20-@xmath21 , @xmath22 is the spin index , @xmath23 is the single site scattering matrix and @xmath24 is the scattering path operator related to the off - diagonal elements of the green s function . @xmath25
is the trace over the orbital indices of the scattering matrix .
positive ( negative ) values for j@xmath26 indicate ferromagnetic ( antiferromagnetic ) coupling between atoms @xmath27 and @xmath28 .
.sublattice occupancies corresponding to the configurations used .
the details are described in the text . [ cols="<,^,^,^,^ " , ] the central quantity related to multi - functionalities in the present context , @xmath1
is defined as @xmath29 , the total moment in the martensite phase with respect to that in the austenite phase for a given occupancy of the sublattices . calculated values of @xmath1
are also presented in table iv .
the results show the following trends : \(i ) the magnetic moments in the martensitic phase is lower than that in the austenite phase for both `` ordered '' and `` disordered '' configurations and for all three alloys under consideration , making @xmath30 .
\(ii ) for the `` disordered '' configurations , that is , both * dc * and * dt * , the magnetic moments are substantially higher than their `` ordered '' counterparts , i.e. , * oc * and * ot * respectively .
the increment is significantly greater in austenite phases than that in the martensitic phases .
the total moment increases by @xmath31 and @xmath32 for mn@xmath0niga , mn@xmath0niin and mn@xmath0nisn respectively as configuration changes from * oc * to * dc*. in the martensitic phases , the increase in moments as configuration changes from * ot * to * dt * are @xmath33 and @xmath34 for mn@xmath0niga , mn@xmath0niin and mn@xmath0nisn respectively .
this results in a huge gain in @xmath1 in the `` anti - site disordered '' configurations over that in the `` ordered '' configurations .
the behavior of the magnetic moments across crystal structures , configurations and materials , are driven by the crystallographically inequivalent mn atoms .
the results suggest that the ferrimagnetic ground states are achieved due to the anti - parallel alignments of the mni and mnii atoms . in
the `` ordered '' configurations , the magnitude of the moments compensate each other substantially resulting in relatively low moments .
drastic changes are observed in cases of anti - site disorder .
the substantially high moments in * dc * and * dt * configurations are driven by considerable quenching of mn moments at the sites where they form alloys with ni .
the moments of the other mn atom remain nearly unaltered .
the ni moments , on the other hand , increases significantly , bringing in an increase in the net moments .
the changes are more dramatic in the austenites where the mni moment changes between @xmath35 as compared to changes between @xmath36 in the martensites .
the significant changes in the mni moments driven by the anti - site disorder coupled with a substantially larger changes in the total moments in the austenites lead to larger @xmath1 when anti - site disorder affects the sublattices with tetrahedral symmetry in the inverse heusler structure .
this is a significant result as it shows that all three mn@xmath0nix alloys may exhibit inverse magneto - caloric effects and a large zeeman energy can be achieved with relatively small magnetic field making them attractive from the point of view of shape memory applications . it may be noted that the prototype msma ni@xmath0mnga has @xmath37 and the magnitude of @xmath1 is only @xmath38 @xcite .
thus , mn@xmath0nix alloys turn out to be msmas with better functional parameters even with the stoichiometric composition of 2:1:1 . in the sections
iiid and iiie , we explore the origin of these magnetic properties by discussing results on electronic structures and magnetic exchange interactions
. figs . 2 - 4 show a comprehensive comparative presentation of the densities of states in various configurations for the three systems considered . in each figure
, panels ( a)-(d ) show comparisons of total and partial densities of states ( of magnetic components ) for * oc * and * dc * configurations .
panels ( e)-(h ) of each figure show comparisons of same quantities for * dc * and * dt * configurations .
the comparisons between * oc * and * ot * configurations for these systems had already been done @xcite .
niga . in panels ( a)-(d )
black curves stand for * dc * configuration and red curves stand for * oc * configuration . in panels
( e)-(h ) , black curves denote the * dc * configuration and red curves denote the * dt * configuration .
( a)-(d ) and ( e)-(h ) panels display total , mni , mnii and ni densities of states , respectively.,width=321 ] the densities of states of all three alloys in the * oc * configuration show certain common features : the mni densities of states have characteristic unfilled majority bands as can be inferred by presence of substantial states in the unoccupied part , while mnii have same characteristics associated with their minority bands .
this explains the reasons for getting substantial compensation of mn moments leading to rather small total moments .
the features in the minority bands near the fermi level emerge due to hybridizations between mni and ni 3@xmath39 states . for mn@xmath0niga and mn@xmath0niin ,
such hybridizations give rise to a small peak around -0.25 ev in the minority bands while for mn@xmath0nisn , a prominent peak around -0.5 ev emerges .
the prominent peaks in the occupied part of the minority bands such as the ones between -1 ev and -2 ev also arise from the hybridizations of same states .
the majority bands for all three systems too have common characteristics ; the features in the occupied parts arise due to hybridizations between ni and mnii states .
iin panels ( a)-(d ) black curves stand for * dc * configuration and red curves stand for * oc * configuration . in panels
( e)-(h ) , black curves stand for * dc * configuration and red curves stand for * dt * configuration .
( a)-(d ) and ( e)-(h ) panels display total , mni , mnii and ni densities of states , respectively.,width=321 ] significant modifications to the densities of states in the * oc * configurations occur due to anti - site disorder between mni and ni sites .
expectedly , major changes in the electronic structure come from the mni and the ni densities of states .
the total densities of states in the majority spin channel become rather featureless and flat in the * dc * configurations . the densities of states in the minority bands , although still retain some of the structures , but become smooth in general . for example , the peaks near the fermi level in the minority bands of mn@xmath0niga and mn@xmath0niin and the peaks at the fermi level in the minority bands of mn@xmath0nisn are all broadened in the * dc * configuration .
mni densities of states are affected most followed by the ni ones .
the minority spin mni peaks ( around -1.5 ev for mn@xmath0niga and mn@xmath0niin , and around -2 ev for mn@xmath0nisn ) broaden considerably and move closer to fermi level ( around -1.4 ev for mn@xmath0niga and mn@xmath0niin , and around -1.3 ev for mn@xmath0nisn ) .
the peaks near or at the fermi level in the minority channels are now destroyed producing a continuously increasing densities of states near the fermi levels .
the peaks in the unoccupied parts of the majority bands of mni atoms are severely modified : their intensities decrease considerably and they flatten out into broad plateaus extending into the occupied parts .
thus , the occupied parts in the majority bands start to get filled with states .
consequently , the magnetic moments of mni atoms decrease in magnitudes as compared to the ones in the * oc * configurations .
densities of states in both spin channels associated with ni too get modified with sharp peaks in * oc * configurations getting broadened in general .
the states in the minority channels start to shift towards the fermi level , while there is hardly any shift in the states in the majority channels ( comparison of panels ( d ) and ( h ) illustrate these clearly ) .
these features explain the slight increases in the ni moment in the * dc * configurations in comparison to the * oc * configurations of all three alloys .
the anti - site disorder between mni and ni sites do not affect mnii densities of states substantially.the major features of the mnii densities of states in * oc * configuration do not change in * dc * configurations , the peaks only broaden explaining why the mnii moments remain almost intact in spite of anti - site disorder . to summarize , the drastic re - distribution of electronic states in both spin channels of primarily mni atoms , brought about by the anti - site disorder between the sites of same point group symmetry , quenches the mni moments substantially reducing the exchange splitting and thereby increasing the total magnetic moment significantly as compared to the * oc * configurations.the degree of this increment depends on the degree of changes in the mni densities of states .
the maximum changes occur in mn@xmath0nisn , where , the peak at -2 ev in the mni minority band moves substantially towards the fermi level , relocating itself at -1.25 ev as one goes from * oc * to * dc * , the other peaks too make such a move , affecting the electron distribution considerably , explaining why the changes in the moments are the largest in the series .
nisn . in panels ( a)-(d )
black curves stand for * dc * configuration and red curves stand for * oc * configuration . in panels ( e)-(h )
, black curves stand for * dc * configuration and red curves stand for * dt * configuration .
( a)-(d ) and ( e)-(h ) panels display total , mni , mnii and ni densities of states , respectively.,width=321 ] the reason behind obtaining a large @xmath1 when anti - site disorder is present , as compared to the `` ordered '' configurations is that the magnetic moments do not change as substantially as the systems undergo martensitic transformations .
the percentage changes quoted in section iiic exemplify this . in order to understand the reasons behind this
, one needs to inspect the panels ( e)-(h ) in fig .
2 - 4 where comparisons between densities of states in the * dc * and * dt * configurations are done .
the results suggest that for all three materials , the total densities of states at the fermi level in the * dt * configurations are less than that in the * dc * configurations .
this provides a clue to the stability of the martensitic phases .
in fact , the densities of states are suppressed prominently in the minority bands of between -0.3 ev and the fermi level in the * dt * configurations .
in contrast , the minority densities of states in * dt * configurations for all three materials get elevated as compared to the * dc * configurations between -0.3 ev and -1 ev . these undoubtedly point to the fact that electron states are transferred to the lower energies as the systems undergo martensitic transformations , explaining the stabilizations of the martensitic phases . from the atom projected densities of states , it is clear that the hybridizations of ni - mni 3@xmath39 minority electrons are responsible for such re - distributions of electronic states when the system undergoes a tetragonal deformation . due to the tetragonal distortion , the ni - mni hybridizations strengthen due to the significant reductions of the ni - mni bond distances .
table v shows the bond distances between various mn atoms in the austenite and in the martensitic phases .
the results clearly demonstrate that only the ni - mni inter - atomic distances reduce by almost @xmath40 in all the cases . on the other hand , the changes in the majority band electronic structures due to tetragonal distortions are overwhelmingly due to the mni states . a comparison between * dc * and * dt * configurations show that states in the mni majority bands are pushed into the unoccupied part as the systems undergo martensitic transformations .
this coupled with the shifting of minority states towards lower energies produce a larger exchange splitting of mni resulting in an increase of the net mni moment as compared to the austenitic phases .
thus the total moment in the martensitic phases do not change as much as they do in the austenite phases upon changes in the sublattice occupancies due to anti - site disorder in the sites with tetrahedral symmetries , resulting in a larger @xmath1 in the `` disordered '' configurations . ) as a function of inter - atomic distance d for mn@xmath0niga .
panel ( a ) , ( b ) , ( c ) and ( d ) represent ordered cubic ( * oc * ) , ordered tetragonal ( * ot * ) , disordered cubic ( *
dc * ) and disordered tetragonal ( * dt * ) configurations , respectively.,width=321 ] in this section , we show comparative results on inter - atomic exchange interactions across structures , configurations and materials in order to understand the trends in the magnetic properties of mn@xmath0ni_x _ series . in fig .
5 - 7 we present results on inter - sublattice and intra - sublattice exchange interactions .
the results suggest that in all three materials , the magnetic properties are governed by competitions between two interactions : ni - mnii and mni - mnii .
the mni - mnii interactions are anti - ferromagnetic while the ni - mnii interactions are ferromagnetic . comparisons of * oc * and * ot * configurations show that upon tetragonal distortions , the anti - ferromagnetic mni - mnii interactions strengthen while the ferromagnetic ni - mnii interactions weaken , resulting in a net loss of mnii moments and subsequent low moments in the * ot * configurations .
drastic modifications in the exchange interactions are observed due to anti - site disorder .
a comparison between * oc * and * dc * configurations show that for all three alloys , the ni - mnii interactions strengthen that is become more ferromagnetic while the mni - mnii interactions weaken substantially .
the weakening of the most prominent anti - ferromagnetic interaction and strengthening of the most prominent ferromagnetic interaction in presence of anti - site disorder , strengthen the ferromagnetic interactions in the systems resulting in the enhancement of the overall magnetic moments when anti - site disorder is present .
however , comparative assessments of * dc * and * dt * configurations show that upon tetragonal distortions , the ferromagnetic ni - mnii interactions weaken and the antiferomagnetic mni - mnii interactions strengthen.the weakening of the former being substantial , the anti - ferromagnetic interactions in the * dt * configurations are more significant than those in the * dc * configurations , resulting in a lower total moment in the former configurations in comparison to the later ones . ) as a function of inter - atomic distance d for mn@xmath0niin .
panel ( a ) , ( b ) , ( c ) and ( d ) represent ordered cubic ( * oc * ) , ordered tetragonal ( * ot * ) , disordered cubic ( * dc * ) and disordered tetragonal ( * dt * ) configurations , respectively.,width=321 ] the intra - sublattice exchange interactions are much weaker than the inter - sublattice ones and thus do not contribute enough to understand the trends in the magnetic properties . however , unlike the inter - sublattice ones , no common trend across the materials is observed in some of these interactions .
the ni - ni interactions are the weakest and are weakly ferromagnetic for all materials and for all configurations .
the mni - mni interactions are oscillatory in the * oc * configurations and become slightly anti - ferromagnetic in the * dc * configurations . in the * dt * configurations , they again become oscillatory with first neighbor being strongly ferromagnetic and second and third neighbors being slightly anti - ferromagnetic ; the only exception being mn@xmath0nisn where the second and third neighbor anti - ferromagnetic interactions are stronger than the first neighbor ferromagnetic one .
the ni - mni interactions become weak and predominantly anti - ferromagnetic as one goes from * oc * to * dc * configurations where they are predominantly ferromagnetic . under tetragonal distortions in the * dc * configurations ,
the first neighbor interactions change to become slightly ferromagnetic offering no other significant changes .
the mnii - mnii interactions vary qualitatively quite a bit across materials . in the * oc * configurations , they are oscillatory in case of mn@xmath0nisn with dominant interactions being ferromagnetic . for the other two materials , the interactions are primarily ferromagnetic and weaker in comparison to mn@xmath0nisn .
the anti - site disorder keeps the interactions largely intact except that they are weaker in mn@xmath0nisn .
the tetragonal distortions modify these interactions significantly by making them more oscillatory . however , the strengths of the ferromagnetic and anti - ferromagnetic components in the oscillatory exchange interactions are nearly equal and thus compensate .
to summarize , the magnetic exchange interactions in these materials are thus influenced by the inter - sublattice interactions and changes in their relative strengths upon changes in structures and configurations help make connections to the trends in the magnetic properties . ) as a function of inter - atomic distance d for mn@xmath0nisn .
panel ( a ) , ( b ) , ( c ) and ( d ) represent ordered cubic ( * oc * ) , ordered tetragonal ( * ot * ) , disordered cubic ( * dc * ) and disordered tetragonal ( * dt * ) configurations , respectively.,width=321 ]
by means of _ ab initio _ calculations , we have investigated compositional and structural stabilities along with the magnetic properties of inverse heusler mn@xmath0ni_x _ alloys in connection to the improvement of shape memory and magneto - caloric effects .
we find that the three out of four alloys considered here are potentially better shape memory materials as a large value of @xmath1 can be achieved even in these compositions whereas in prototype ni@xmath0mn_x _ systems , one has to vary the ratio of ni , mn and _ x _ judiciously to achieve the same .
we also find that these materials would show inverse magneto - caloric effect , which is technologically desirable for green environment .
our analysis reveal that these properties emerge due to the inverse heusler structure where the two mn atoms are inequivalent and an anti - site disorder exists between one of the mn and the ni sublattices .
it is also found that the electronic structures associated with the mn atom , which makes an alloy with ni , are primarily responsible for the dramatic changes in the magnetic properties , and consequently for improved functionalities .
this work , thus , show that the mn@xmath0ni_x _ alloys in the inverse heusler structure can be considered as potential functional materials and that more experimental verifications of their functional properties are required .
financial assistance from the swedish research links ( vr - sida ) is acknowledged .
the swedish national computing facilities , computation facilities from c - dac , pune , india and from department of physics , iit guwahati funded under the fist programme of dst , india are also acknowledged .
99 f. x. hu , b. g. shen , and j. r. sun , appl .
. lett . * 76 * , 3460 ( 2000 ) .
l. pareti , m. solzi , f. albertini , and a. paoluzi , eur .
j. b * 32 * , 303 ( 2003 ) .
t. krenke , e. duman , m. acet , e. f. wassermann , x. moya , l. ma@xmath41osa , and a. planes , nature mater . * 4 * , 450 ( 2005 ) .
a. planes , l. ma@xmath41osa and m. acet , j. phys .
condens . matter .
* 21 * , 233201 ( 2009 ) .
l. ma@xmath41osa , d. gonzlez - alonso , a. planes , e. bonnot , m. barrio , j. tamarit , s. aksoy and m. acet , nat mater * 9 * 478 ( 2010 ) .
p. o. castillo - villa , d. e. soto - parra , j. a. matutes - aquino , r. a ochoa - gamboa , a. planes , l. ma@xmath41osa , d. gonzlez - alonso , m. stipcich , r. romero , d. rios - jara , and h. flores - zu@xmath41iga , phys .
b * 83 * , 174109 ( 2011 ) . c. biswas , r. rawat , and s. r. barmana , appl . phys .
lett . * 86 * , 202508 ( 2005 ) . v. k. sharma , m. k. chattopadhyat , k. h. b. shaeb , a. chouhan and s. b. roy , appl .
89 * , 222509 ( 2006 ) .
z. li , c. jing , j. chen , s. yuan , s. cao , and j. zhang , appl .
. lett . * 91 * , 112505 ( 2007 ) . v. k. sharma , m. k. chattopadhyay , and s. b. roy , phys .
b * 76 * , 140401r ( 2007 ) .
d. y. cong , s. roth , j. liu , q. luo , m. potschke , c. hurrich , and l. schultz , appl .
* 96 * , 112504 ( 2010 ) .
y. wang , c. huang , h. wu , j. gao , s. yang , d. wang , x. ding , x. song , and x. ren , appl . phys
. lett . * 102 * , 141909 ( 2013 ) .
l. ma , w. h. wang , c. m. zhen , d. l. hou , x. d. tang , e. k. liu and g. h. wu , phys .
b * 84 * , 224404 ( 2011 ) .
h. luo , f. meng , z. feng , y. li , w. zhu , g. wu , x. zhu , c. jiang and h. xu , j. appl .
phys . * 107 * , 013905 ( 2010 ) .
k. ullakko , j. k. huang , c. kanter , r. c. ohandley and v. v. kokorin , appl .
. lett . * 69 * , 1966 ( 1996 ) .
a. sozinov , a. a. likhachev , n. lanska and k. ullakko , appl .
. lett . * 80 * , 1746 ( 2002 ) .
p. entel , a. banneberg , m. siewert , h. c. harper , m. e. gruner , d. comtesse , h. elmers and m. kallmayer , metall . mater .
a * 43 * , 2891 ( 2012 ) .
p. entel , m. siewert , m. e. gruner , a. chakrabarti , s. r. barman , v. v. sokolovskiy and v. d. buchelnikov , j. alloys compd .
, * 577 * , s107 ( 2013 ) . r. b. helmholdt and k. h. j. buschow , j. less - common met .
* 128 * , 167 ( 1987 ) .
g. d. liu , j. l. chen , z. h. liu , x. f. dai , g. h. wu , b. zhang , and x. x. zhang , appl .
phys . lett . *
87 * , 262504 ( 2005 ) .
n. lakshmi , k. pandey and n. venugopalan , _ bull . mater .
_ * 25 * , 309 ( 2002 ) .
s. paul , and s. ghosh , j. appl .
phys . * 110 * , 063523 ( 2011 ) .
p. j. brown , t. kanomata , k. neumann , k. -u .
neumann , b. ouladdiaf , a. sheikh and k. r. a. ziebeck , j. phys . condens . matter .
* 22 * , 506001 ( 2010 ) .
s. singh , r. rawat , s. e. muthu , s.w .
dsouza , e. suard , a. senyshyn , s. banik , p. rajput , s. bhardwaj , a. m. awasthi , rajeev ranjan , s. arumugam , d. l. schlagel , phys .
lett . * 109 * , 246601 ( 2012 ) . s. w. dsouza , a.chakrabarti and s. r. barman http://arxiv.org/abs/1310.6130v1 ( 2013 ) .
s. paul , s. ghosh and b. sanyal , j. phys .
condens . matter .
* 25 * , 236005 ( 2013 ) .
h. ebert , d. kdderitzsch and j. minar , rep .
phys . * 74 * , 096501 ( 2011 ) see http://olymp.cup.uni-muenchen.de/ak/ebert/sprkkr .
spr - kkr is a high quality density functional theory based electronic srtucture code , co - ordinated by h. ebert .
s. h. vosko , l. wilk and m. nusair , can .
* 58 * , 1200 ( 1980 ) .
p. soven , phys . rev . * 156 * , 809 ( 1967 ) .
i. liechtenstein , m. i. katsnelson , v. a. antropov and v. p. gubanov , j. magn .
magn . mater . *
67 * , 65 ( 1987 ) .
m. methfessel and j. k@xmath42ubler , j. phys .
* 12 * , 141 ( 1982 ) .
g. grimvall , phys .
* 13 * , 59 ( 1976 ) p. sderlind and b. johansson , thermochim .
acta * 218 * , 145 ( 1993 ) g. kresse and j. furthmuller , phys . rev . * b54
* , 11169 ( 1996 ) g. kresse and d. joubart , phys . rev .
* b59 * , 1758 ( 1999 ) a. zunger , s .- h .
wei , l. ferreira , and j. e. bernard , phys .
* 65 * , 353 ( 1990 ) a. van de walle , p. tiwary , m. de jong , d.l .
olmsted , m. asta , a. dick , d. shin , y. wang , l .- q .
chen and z .- k .
liu , calphad , * 42 * , 13 ( 2013 ) v. v. godlevsky and k. m. rabe , phys . rev . * b63 * , 134407 ( 2001 ) . | recent first - principles calculations have predicted mn@xmath0ni_x _ ( _ x_=al , ga , in , sn ) alloys to be magnetic shape memory alloys .
moreover , experiments on mn@xmath0niga and mn@xmath0nisn suggest that the alloys deviate from the perfect inverse heusler arrangement and that there is chemical disorder at the sublattices with tetrahedral symmetry . in this work , we investigate the effects of such chemical disorder on phase stabilities and magnetic properties using first - principles electronic structure methods .
we find that except mn@xmath0nial , all other alloys show signatures of martensitic transformations in presence of anti - site disorder at the sublattices with tetrahedral symmetry .
this improves the possibilities of realizing martensitic transformations at relatively low fields and the possibilities of obtaining significantly large inverse magneto - caloric effects , in comparison to perfect inverse heusler arrangement of atoms .
we analyze the origin of such improvements in functional properties by investigating electronic structures and magnetic exchange interactions . |
a large variety of systems in physics , chemistry , biology , population genetics , or traffic flows can be modeled by reaction - diffusion processes .
these processes are stochastic out - of - equilibrium models describing particles of one or several species which randomly diffuse on a lattice and can undergo some reactions at given rates when they meet @xcite .
these models are of fundamental importance both from a phenomenological and from a theoretical point of view .
they generically feature pattern formation and non - equilibrium phase transitions , such as the ubiquituous transition to an absorbing state which traps the system @xcite .
reaction - diffusion processes hence provide simple models to study non - equilibrium scaling and phase transitions . whereas the one - species processes are relatively well understood @xcite , the multi - particle processes are far less explored .
in this work , we consider the diffusive epidemic process ( dep ) @xcite , which is the two - species ( @xmath2 and @xmath3 ) reaction - diffusion system with reactions : @xmath4 the species @xmath2 and @xmath3 diffuse independently with diffusion constants @xmath5 and @xmath6 .
they can be interpreted as healthy and sick individuals respectively , undergoing infection on contact at rate @xmath7 and spontaneous recovery at rate @xmath8 .
a salient feature of this process is that the total number of particles is conserved .
this model was theoretically studied using field theory and renormalization group ( rg ) in @xcite .
an action for dep was derived using the doi - peliti formalism @xcite and analyzed perturbatively . for equal diffusion rates
, the authors found that at a critical population density , a continuous absorbing phase transition occurs with upper critical dimension @xmath0 , which does not belong to the ubiquitous directed percolation ( dp ) universality class but to a new class ( kss ) introduced in @xcite to model the effect of pollution on a population .
the kss universality class is endowed with the critical exponents @xmath9 , @xmath10 and @xmath11 ( that is @xmath12 ) to first - order in @xmath13 . for @xmath14 ,
the authors of @xcite predicted a continuous transition of yet a new ( woh ) universality class ( distinct from dp and kss ) with exponents @xmath9 , @xmath10 and @xmath15 ( @xmath16 ) to all orders in @xmath17 . for @xmath18 , a fluctuation - induced first order transition was conjectured and was seemingly confirmed in @xcite by analytical arguments and numerical simulations in @xmath19 .
.critical exponents of dep from monte carlo simulations and field theoretical analyses .
the values in grey ( red online ) are deduced assuming @xmath20 .
the parameter @xmath21 is defined as @xmath22 . [
cols="^,^,^,^,^,^,<",options="header " , ] let us make a point clear about issue iii ) in the introduction , regarding the value of the exponent @xmath23 .
one of the outcomes of the previous analysis is that for dp - c , the identities @xmath20 and @xmath24 imply @xmath25 exactly for @xmath26 .
hence the dp - c action can not account for a non - trivial value of @xmath23 , as seems to be observed in this case in lattice simulations @xcite .
the symmetry ( [ symiii ] ) provides a possible scenario for the transition from dep to dp - c .
let us examine an infinitesimal version of ( [ symiii ] ) : @xmath27 .
the associated ward identity implies that the corresponding symmetry is transposed to the effective action .
accordingly , the effective action is invariant under @xmath28 .
we expect @xmath29 to have a positive scaling dimension so that in terms of adimensioned fields , one gets @xmath30 .
consequently , when the renormalization group scale @xmath7 flows to zero , the constant shift term becomes dominant and the ( time independent version of the ) symmetry ( [ symi ] ) is recovered .
in fact , we present below results from the integration of the nprg flow which confirm to some extent this scenario . however , let us note that this property is not sufficient to ensure that the dep transition belongs to the universality class of dp - c . indeed , the exponents of dp - c follow from the conjunction of both the space - time gauged shift ( [ symi ] ) and the duality relation ( [ symii ] )
however , none of these symmetries is preserved for dep , since a renormalization of the @xmath31 propagator is allowed .
consequently @xmath32 can still depart from @xmath33 even though the shift symmetry is restored in the infrared at the dep fixed point .
of course , these symmetries can always appear as accidental symmetries of the fixed point .
indeed , the field theoretical results @xcite suggest that this is the case for spacial dimensions close to the upper critical dimension , where the only relevant terms are the cubic ones , and the cubic action satisfies these gauged symmetries .
however , it is unclear whether the gauged identities remain attractive at the relevant infrared fixed point in arbitrary dimensions .
this connects with the original question raised in @xcite : the dp - c fixed point has _ a priori _ more symmetries than the dep one . whether the dp - c fixed point continues to control the dep transition in low dimensions ( as it does near the upper critical dimension ) , is a dynamical question to be investigated .
our task in the remainder of the article is to study both dp - c and dep through their respective symmetries with the help of nprg techniques .
the nprg procedure for equilibrium systems @xcite can be straightforwardly generalized to out - of - equilibrium phenomena @xcite .
the starting point is to modify the generating functional ( [ partfunc ] ) by adding a scale dependent quadratic regulator .
generically , @xmath34
\equiv \int \mathcal{d}[\phi ] \ , e^{-s -\frac{1}{2}r_k^{\alpha\beta}\phi_{\alpha}\phi_{\beta } + j^{\alpha}\phi_{\alpha}}\ , , \ ] ] where , as before , greek indices stand for coordinate and internal indices .
the role of the regulator @xmath35 is to smoothly replace momenta @xmath36 in the propagator with @xmath7 ( denoting @xmath37 ) . here , both the @xmath38 and @xmath39 propagators are regulated ; more precisely @xmath40 with @xmath41 and @xmath42 , where @xmath43 and @xmath44 are running coefficients defined in eq .
( [ ansatzdpc ] ) . in this work ,
we choose to use the cutoff function @xmath45 proposed by litim @xcite @xmath46 which allows for analytical integration on momenta .
the regulated cumulant generating functional and its ( modified ) legendre transform are then defined as @xmath47 where @xmath48 , generalizing previous definitions in presence of the infrared regulator .
the evolution of @xmath49 with the rg scale @xmath7 is given by the exact flow equation @xcite : @xmath50 which smoothly interpolates between the mean - field effective action @xmath51 and the exact effective action @xmath52 , when @xmath7 is lowered from a microscopic uv cutoff @xmath53 to @xmath54 .
at this point , in order to integrate ( [ wetterich ] ) , some approximations have to be made based on the properties of @xmath35 and on identities coming from the symmetries of @xmath49 . in order to propose appropriate anstze for @xmath49
, its general form must be constrained by the symmetries of the action .
this is achieved as previously through the use of ward identities , generalized to account for the presence of the infrared regulator @xmath35 ( see appendix [ ward ] ) .
in particular , the ward identity ( [ wardidsansreg ] ) remains true for @xmath49 , and yields that @xmath49 endows the same general form as ( [ gammagen ] ) @xmath55\ , , \ ] ] with the explicit kinetic terms unrenormalized and @xmath56 invariant under ( [ symi ] ) .
this implies that the nprg flow maintains the property @xmath57 and @xmath58 .
moreover , if @xmath26 , then the flow also preserves @xmath59 , _ i.e. _ @xmath24 .
furthermore , using the duality transformation ( [ symii ] ) , one obtains the constraint ( see appendix [ ward ] ) @xmath60 where @xmath61 is the effective potential , which corresponds to @xmath49 evaluated for homogeneous static fields ( divided by the space - time volume ) . within an adequate parametrization of @xmath61
, one can deduce from ( [ ward0 ] ) that @xmath20 exactly , and at the same time identify the relevant direction of the flow , see appendix [ ward ] .
hence , within the nprg framework , the modified ward identities also fix the exact values for the critical exponents of dp - c , gathered in table [ exact ] . using the above contraints
, we propose an ansatz for @xmath49 . in this work
, we restrict to the local potential approximation , or lpa , modified minimally to account for the renormalization of the fields @xcite , a modification which is referred to as lpa. it means that the only momentum dependence of @xmath49 is contained in the bare propagator ( up to the scaling factor for the field ) , while the potential part @xmath61 is renormalized : @xmath62 furthermore , while some of the results presented below are valid for an arbitrary potential , the integration of the flow and the search for a fixed point are performed by taylor expanding @xmath61 ( around a minimum for @xmath63 , see appendix [ ward ] ) to a finite order ( from cubic up to sextic ) and truncating the corresponding flow equations at the same order .
the derivation of the flow equations is explained in appendix [ nprg ] .
the lpa ( or lpa ) approximations are simple to implement , and straightforward to relate to known perturbative renormalization results .
it gives a fairly accurate view of the physics in most problems .
however , it also has limitations , as emphasized in the following . for the cubic order ( which corresponds to a renormalization of the coupling constants of the vertices of the dp - c bare action )
, we find a critical fixed point for any dimensions below @xmath64 and for any value of @xmath21 .
it corresponds to the fixed point precedently described perturbatively in @xcite . for @xmath65
, we also find that a fixed point exists , but the flow can not reach it since this would require for the vertex @xmath66 to change sign , which is prevented in our scheme by the expansion around the minimum .
this is manifest in eq .
( [ eqmin ] ) , the corresponding coupling @xmath67 can not vanish . from the quartic to sextic order , we find a fixed point with one unstable direction for all values of @xmath21 , but only for @xmath68 .
let us first discuss the case @xmath69 .
the values of @xmath70 at the corresponding fixed point are displayed as a function of the dimension in fig .
[ result_mu0 ] .
a striking difference is manifest between the lowest and higher orders . for @xmath71
, we obtain the value @xmath72 at quartic order . for higher orders ,
the values of @xmath70 at different dimensions are similar to those of the quartic order , but the fixed point solution is lost at a dimension slightly higher than @xmath71 . if extrapolated to @xmath73 , these values seem to be in accordance with the simulations in @xmath73 , but the fixed point solution disappears in dimension @xmath74 .
unfortunately , it is computationally too costly to implement the septic and higher orders . the way to progress would be to deal with richer anstze , for example using the full effective potential without truncation . at this point , as discussed below , it is not clear whether the disappearance of the fixed point is an artifact due to the finite - order truncation of the lpa , or if it is related to the intrinsic nature of the dp - c fixed point and its relation to dep . ( image ) at ( 0,0 ) as a function of the dimension @xmath75 for @xmath69 calculated from the lpa truncations compared to the one - loop result ( dashed line ) and to the simulation @xcite ( isolated point at @xmath73 ) .
thick saltires mark the disappearance of the fixed point .
( colors online),title="fig:",scaledwidth=50.0% ] ; in the @xmath76 case , this problem seems even sharper ( see fig . [ result_mu ] ) , since the domain of existence of the fixed point with one unstable direction describing the transition around @xmath77 is not under control as the truncation order increases .
finally , when @xmath65 , the same fixed point as for @xmath76 is present , but it can not be reached from the bare action . besides this fixed point , from the order four in the field expansion , we find a new fixed point in the physical region of parameters .
in contrast with the perturbative calculations but in agreement with the numerical simulations , this finding supports the existence of a second order phase transition for @xmath65 . however , we are not able to convincingly assess the convergence of this result with the order of the truncation in the field expansion , since the domain of existence of this fixed point also varies significantly as the order is increased .
let us note that for @xmath26 , as @xmath24 , the only non - trivial exponent for the dp - c class is @xmath78 ( see table [ exact ] ) .
this exponent is related to @xmath79 , the critical initial slip exponent .
hence , it would be interesting to compare the value of @xmath78 obtained for @xmath80 in @xmath71 to simulations in order to test the relevance of the corresponding fixed point .
unfortunately no such determination exists in the literature for @xmath81 . in both cases ,
the disappearance of the dp - c fixed point gives rise to the same question as in the @xmath69 case .
either it is an artifact to be imputed to the finite - order truncation , or it signals that the transition can not be described by the dp - c class in low dimensions .
( image ) at ( 0,0 ) as a function of the dimension for @xmath82 ( resp .
@xmath83 ) calculated from the lpa truncations compared to the one - loop result ( dashed line ) .
thick saltires mark the disappearance of the fixed point .
( colors online),title="fig:",scaledwidth=50.0% ] ; ( image ) at ( 0,-5 ) as a function of the dimension for @xmath82 ( resp .
@xmath83 ) calculated from the lpa truncations compared to the one - loop result ( dashed line ) .
thick saltires mark the disappearance of the fixed point .
( colors online),title="fig:",scaledwidth=50.0% ] ; although the lpa for dp - c shows promising results , the confirmation of the existence of a fixed point with one unstable direction driving the absorbing transition in @xmath73 does not seem to be reachable within the kind of approximation implemented in this paper . moreover , given that the dp - c fixed point seems to be characterized by a larger symmetry group than the dep one , it is unclear if one should expect the infrared dep fixed point to be the same as the dp - c one for dimensions far from @xmath84 . on top of this ,
as already mentioned , some dp - c predictions for critical exponents ( such as @xmath85 for @xmath26 ) seem to be inconsistent with the results of lattice simulations . independently of this last point
, it is unsatisfactory from a nprg point of view not to pinpoint how the dp - c symmetries can emerge along the flow from a microscopic action lacking these symmetries . as an attempt to clarify this issue
, we studied the flow of a potential @xmath86 symmetric under the transformation ( [ symiii ] ) related to the dep action , taylor expanded both in @xmath29 and @xmath87 around a minimum value .
we find that , in dimensions @xmath88 where the dp - c fixed point exists , the effective action preserves the rescaling symmetry ( [ symiii ] ) along the flow , which progressively behaves as the shift symmetry ( [ symi ] ) , eventually recovered at long distance , for any value of @xmath21 , confirming the scenario proposed in sec .
[ scenario ] .
however , this scenario only holds for values of @xmath75 where the dp - c fixed point exists and has the appropriate number of relevant directions in order to control the critical behaviour ( which depend on @xmath21 and on the order of the truncation ) . below this value
, the flow leads to another fixed point which is unphysical or diverges .
this new fixed point is presumably spurious but its existence may be the reason why the dp - c fixed point ceases to exist below a certain dimension .
we conjecture that if a more elaborate approximation describing a broader theory space than dp - c was used , the same mechanism would occur in small dimensions but the flow would lead to a genuine fixed point , which is absent in our truncation .
let us explain why the lpa does not suffice to resolve this issue .
first , let us note that the analysis of the diagrams involved in the perturbative renormalization of the theory shows that the conjunction of the @xmath89 term as a multiplicative factor of the potential and the it@xmath90 s prescription @xcite ( which forbids tadpoles in the perturbative formalism and also within the lpa approximation ) implies that in practice the @xmath91 propagator can not be renormalized and hence no fixed point other than the dp - c one can be reached starting from the dep action .
the same conclusion holds within the lpa approximation , regardless whether the effective potential is expanded or not . in order to study the possible differences between dep and dp - c , a more elaborate approximation , allowing in particular for a renormalization of the @xmath91 propagator
has to be implemented . in this respect
, the grassberger formulation seems to be a better starting point because , as already said , its variation under the time - gauged shift ( [ symi ] ) is linear .
futhermore , in the case @xmath69 , the kpz - like tilt symmetry of ( [ tilt ] ) could be exploited .
this may constitute the subject of a future work .
we analyzed the action of dp - c and dep , and provided a detailed analysis of their respective symmetries , unveiling in particular new symmetries that were not identified before .
we then studied the dp - c action in the framework of non - perturbative renormalization , using a modified local potential approximation lpa. we recovered the results of previous field theoretical studies at one - loop in the @xmath17-expansion @xcite , that is the existence of a dp - c fixed point for @xmath69 and @xmath76 .
however , it turns out that this result can not be simply extended to @xmath73 and @xmath92 .
indeed , we observed ( within the lpa ) that this fixed point no longer exists below a certain dimension @xmath74 . at this stage ,
two possible explanations can be contemplated .
the disappearance of the dp - c fixed point can be an artifact of the lpa approximation implemented in this work , which means that finding the dp - c fixed point is low dimensions requires a higher - order approximation .
the alternative interpretation is that the dp - c fixed point ceases to control the absorbing transition in low dimensions , and thus that the transition in the dep model is driven by another fixed point , which has less symmetries than the dp - c one .
investigating this scenario also requires a more elaborate approximation , able to distinguish between the dep and dp - c renormalizations .
this interpretation means that the transitions observed in lattice simulations in @xmath73 or 2 do not correspond to the dp - c universality class , but to a genuine dep class , where the exponent values are not fixed by the exact dp - c identities .
this scenario could explain the mismatch between the exact predictions coming from the symmetries of the dp - c fixed point , and results obtained in lattice simulations .
in conclusion , the study of dep and of its possible coarse - grained counterpart , dp - c , is far from being a settled subject .
we hope to be able to answer the questions raised in this article in future studies .
within the nprg framework , in the presence of the infrared regulator @xmath35 , ward identities can be derived by considering an infinitesimal change of variable in @xmath44 @xcite : @xmath93 if the jacobian of the change of variables is one , and using @xmath94 , one obtains @xmath95 where @xmath96 and @xmath97 .
it means that the variation of @xmath49 is equal to the mean of the variation of @xmath98 ( up to a possible regulator term ) .
the right - hand side can be made explicit if @xmath99 is linear in the field and the regulator term is invariant .
this is a standard result for a global symmetry of the action .
another interesting case is when a global shift ( @xmath100 ) symmetry of the action is gauged . under the condition that the momentum dependence of @xmath98 is only contained in its quadratic part
, one can derive a local constraint on @xmath49 .
this is the case for the symmetry ( [ symi ] ) of dp - c .
the variation of @xmath49 under its gauged version reads @xmath101 this identity means that these two terms are not renormalized , and the remaining part of @xmath49 must be invariant under the transformation ( [ symi ] ) .
that is , it can only depend on @xmath102 , hence the parametrization ( [ ansatzdpc ] ) . the exact value for the exponent @xmath32 is then fixed by the duality ( [ symii ] ) .
indeed , performing the infinitesimal change of variable @xmath103 in @xmath44 , one obtains the ward identity @xmath104 evaluating this identity for a uniform and static field configuration then yields @xmath105 since @xmath61 is analytic , let us consider the expansion of @xmath61 around a minimum @xmath106 for @xmath63 @xmath107 such that @xmath108 ( notice that the property @xmath109 can be shown to be conserved by the flow , which is why the minimum is defined after derivation with respect to @xmath110 and @xmath111 ) .
the identity ( [ wardu ] ) then yields @xmath112 by evaluating this equation at @xmath113 , using @xmath114 and @xmath115 as well as the analyticity of the potential , one obtains @xmath116 ( [ ward2 ] ) means that @xmath106 is the only relevant coupling , so that the flow can be projected on the critical surface by considering all couplings except @xmath106 . as usual , the fixed point is fully attractive within the critical surface .
now , if one considers the action detuned from criticality by the infinitesimal quantity @xmath117 , one has @xmath118 . combining this consideration with ( [ ward1 ] ) ,
one gets @xmath119
within the derivative expansion , one focuses on the small momentum and frequency regime of vertex functions . in this regime and in the lpa approximation , which corresponds to the ansatz ( [ ansatzdpc ] ) ,
the scale - dependence of the effective action is restricted to the potential @xmath61 and to the scaling factors @xmath44 and @xmath43 , _ i.e. _ to low order derivatives of propagators at zero momenta and frequencies , evaluated at a fixed configuration of constant fields . in the general case , for two species , one has to calculate a @xmath120 matrix of 2-point correlation functions for @xmath110 , @xmath29 and their response fields . however , for dp - c , the only propagator which can undergo renormalization in the lpa approximation is @xmath89 . as a consequence ,
only two scale - dependent numbers are needed ( on top of potential couplings ) : @xmath44 , which encode the renormalization of the fields , and @xmath43 , which gives the scaling between time and space , as encoded in ( [ ansatzdpc ] ) .
as the @xmath91 propagator is not renormalized , @xmath43 does not have a scaling dimension but it can still have a sub - dominant dependence on @xmath7 . another approximation is implemented for the potential @xmath61 . instead of treating its full field dependence , it is expanded in a polynomial of the fields around the minimum @xmath106 of @xmath63 defined in eq .
( [ ukexp ] ) , and truncated at a given order .
let us first define the @xmath89 inverse propagator : @xmath121 where @xmath122 denotes the static and uniform configuration of the fields @xmath123 and `` ft '' the fourier - transformation , performed after differentiation and evaluation at the minimum configuration .
furthermore translational invariance is used so that the left - hand side only depends on one momentum and frequency .
the scaling factors can then be defined as @xmath124 and the anomalous dimensions are obtained as @xmath125 the respective values of each anomalous dimension is provided by the symmetries : @xmath24 for @xmath126 and @xmath127 for @xmath128 .
the nprg flow equation for the inverse propagator reads : where @xmath130 is the 4@xmath1314 matrix of second derivatives of the potential with respect to the fields @xmath132 and @xmath111 and @xmath133 is the matrix of connected correlation functions , obtained by inverting the 4@xmath1314 matrix @xmath134 . the @xmath17 subscript account for the it@xmath90 s choice of discetization and indicates that the time at which the response field is evaluated is shifted by @xmath17 @xmath135 in fourier space , this leads for example to @xmath136 see @xcite for the justification of this procedure . from eq .
( [ dkgam2 ] ) , one then deduces the flow equations for @xmath44 and @xmath43 following the definitions ( [ deflambdaz ] ) . note that additional tadpole diagrams ( proportional to fourth derivatives of the potential ) , which vanish within the lpa in it@xmath90 s convention , are not included in eq .
( [ dkgam2 ] ) .
the flow equation for the potential reads @xmath137 with it@xmath90 s shift convention ( [ ito ] ) , the integration on @xmath138 amounts to a sum on the residues of @xmath139 in the upper - half plane , operation which is denoted by @xmath140 in the following . using the ansatz ( [ ansatzdpc ] ) for dp - c one obtains by differentiating @xmath61 and evaluating the result at the minimum configuration defined in ( [ gamma2 ] ) ,
one is left with only two possible poles : @xmath145 and @xmath146 .
more precisely , the integrand is now a rational function of @xmath147 and their respective complex conjugate .
the non - zero contributions to the integral arise from one or more @xmath148 factors in the denominator . with the notation @xmath149 one obtains for example for @xmath150 @xmath151 + q^2 \partial_k ( \lambda_k z r ) \big [ \frac{2 q^2 u_{111 } \left(2 \mu u_{012 } \gamma_{\lambda } ^*+\left(\lambda^2 - 1\right ) q^2 ( r+1)^2 u_{111 } z\right)}{\gamma ( \lambda -1 ) z^3 \left(\gamma_{\lambda}^*\right)^3 \gamma^*}\nonumber\\ & + \frac{4 u_{012 } \left((\lambda -1 ) u_{021 } \gamma ^*- \mu q^2
u_{111}\right)}{(\lambda -1 ) \gamma_{\lambda } z^3 \left(\gamma_{\lambda } ^*\right)^2 \gamma^ * } - 2\frac{z\gamma_{\lambda } ^ * \left(u_{121 } \gamma ^*+\mu q^2 u_{112}\right)+ q^2 u_{111}^2 \left((r+1 ) z-\mu \right)}{z^3 \left(\gamma _
{ \lambda } ^*\right)^3 \gamma ^ * } \big]\bigg\}\nonumber\\ = \int_{{\bf q}}\bigg\{&q^2 \partial_k ( \lambda z r ) \frac { \lambda ^2 ( r+1 ) u_{111}^2 z+u_{012 } \left((\lambda + 1)^2 ( r+1 ) u_{021}+ ( 2 \lambda + 1 ) \mu u_{111}\right ) } { \lambda ^2 ( \lambda + 1)^2 q^4 ( r+1)^3 z^3 } + q^2 \partial_k r \frac{u_{111 } \left ( \lambda ( r+1 ) u_{111 } z+\mu u_{012}\right)}{\lambda ( \lambda + 1)^2 q^4 ( r+1)^3 z^2 } \bigg\}\ , , \end{aligned}\ ] ] where the dependence in @xmath7 has been omitted for notational simplicity . at this point ,
the procedure is the same as for nprg applied to equilibrium systems . with our choice of regulator , the momentum integral can be performed analytically and , following the previous example , one is left with @xmath152}{(d+2 ) ( d+4)\lambda ^2 ( \lambda + 1)^2 z^2}+\nonumber\\ \frac { u_{012 } u_{021 } \big[\lambda \left(2 + d -{\bar{\eta}}-\eta\right)+k \partial_k\lambda\big]}{d ( d+2 ) \lambda ^2 z^2}+ \frac { u_{111}^2 \big[\lambda \left(2 + d -{\bar{\eta}}-\eta \right)+d+k
\partial_k\lambda+2\big]}{d ( d+2 ) ( \lambda + 1)^2 z}\big\}\ , , \end{aligned}\ ] ] the flow equation of the rescaled minimum is deduced from the implicit equation @xmath155 in terms of the rescaled variables . for @xmath69 ,
one obtains @xmath156 in the case @xmath80 , one obtains @xmath157 to get the full set of coupled ordinary differential equation for @xmath158 , @xmath159 and the coefficients @xmath160 up to the truncation order , the above procedure is systematically implemented using mathematica .
we verified that @xmath158 decouples from the other couplings , which implies that it is an instable direction of the flow with eigenvalue @xmath161 and that , once its flow is excluded , the flow of the other couplings is fully attractive ( if a critical fixed point exists ) . in consequence ,
the integration of the flow can be simply done numerically , without the need to perform any fine - tuning .
45ifxundefined [ 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 10.1080/00018730050198152 [ * * , ( ) ] , link:\doibase 10.1103/revmodphys.76.663 [ * * , ( ) ] @noop _ _
( , ) @noop * * , ( ) @noop _ _ ( , , ) link:\doibase 10.1016/s0378 - 4371(97)00603 - 1 [ * * , ( ) ] http://stacks.iop.org/0305-4470/9/i=9/a=009 [ * * , ( ) ] \doibase 10.1051/jphys:019850046090146900 [ * * , ( ) ] link:\doibase 10.1103/physreva.39.2214 [ * * , ( ) ] \doibase 10.1023/a:1018697024371 [ * * , ( ) ] link:\doibase 10.1103/physreve.64.058102 [ * * , ( ) ] \doibase doi:10.1016/s0378 - 4371(01)00050 - 4 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/19/6/065143 [ * * , ( ) ] link:\doibase 10.1088/742 - 5468/2010/04/p04027 [ * * , ( ) ] link:\doibase 10.1103/physreve.64.058101 [ * * , ( ) ] link:\doibase 10.1016/j.physa.2007.08.038 [ * * , ( ) ] link:\doibase 10.1088/1751 - 8113/41/40/405002 [ * * , ( ) ] link:\doibase 10.1103/physreve.61.6330 [ * * , ( ) ] link:\doibase 10.1016/0370 - 2693(93)90726-x [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(94)90446 - 4 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 1573(00)00137-x [ * * , ( ) ] , \doibase http://dx.doi.org/10.1016/s0370-1573(01)00098-9 [ * * , ( ) ] , http://www.springer.com/physics/book/978-3-642-05093-0[__ ] , lecture notes in physics ( , , ) `` , '' in link:\doibase 10.1007/978 - 3 - 642 - 27320 - 9_2 [ _ _ ] , ( , , ) pp .
link:\doibase 10.1103/physrevlett.92.195703 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.255703 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.100601 [ * * , ( ) ] link:\doibase 10.1103/physreve.89.010102 [ * * , ( ) ] , link:\doibase 10.1103/physreve.87.052132 [ * * , ( ) ] , link:\doibase 10.1007/jhep03(2016)201 [ * * , ( ) ] , link:\doibase 10.1016/0378 - 4371(91)90063-i [ * * , ( ) ] link:\doibase 10.1103/physreve.62.r5875 [ * * , ( ) ] link:\doibase 10.1103/physreve.64.056104 [ * * , ( ) ] link:\doibase 10.1103/physreve.82.042103 [ * * , ( ) ] link:\doibase 10.1103/physreve.93.042117 [ * * , ( ) ] https://arxiv.org/abs/1607.01635 [ ( ) ] , link:\doibase 10.1007/bf01313803 [ * * , ( ) ] link:\doibase 10.1103/physreve.74.030101 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.56.889 [ * * , ( ) ] link:\doibase 10.1103/physreve.84.061128 [ * * , ( ) ] link:\doibase 10.1088/1751 - 8113/44/49/495001 [ * * , ( ) ] , link:\doibase 10.1142/s0217751x01004748 [ * * , ( ) ] link:\doibase 10.1103/physreve.91.053004 [ * * , ( ) ] | we consider the diffusive epidemic process ( dep ) , a two - species reaction - diffusion process originally proposed to model disease spread within a population .
this model exhibits a phase transition from an active epidemic to an absorbing state without sick individuals .
field - theoretic analyses suggest that this transition belongs to the universality class of directed percolation with a conserved quantity ( dp - c ) .
however , some exact predictions derived from the symmetries of dp - c seem to be in contradiction with lattice simulations .
here we revisit the field theory of both dp - c and dep .
we discuss in detail the symmetries present in the various formulations of both models , some of which had not been identified previously .
we then investigate the dp - c model using the derivative expansion of the non - perturbative renormalization group formalism .
we recover previous results for dp - c near its upper critical dimension @xmath0 , but show how the corresponding fixed point seems to no longer exist below @xmath1 .
consequences for the dep universality class are considered . |
in the last two decades , the nonperturbative renormalization group ( nprg ) approach has been successfully applied to many areas of physics,@xcite from high - energy physics to statistical and condensed - matter physics .
it has proven to be a powerful tool to study not only the low - energy long - distance properties in the vicinity of second - order phase transitions but also non - universal quantities .
in particular , the nprg approach has been implemented in lattice models and used to compute the transition temperature and the magnetization in classical spin models ( ising , xy and heisenberg models).@xcite this implementation of the nprg is referred to as the lattice nprg .
the strategy of the nprg is to build a family of models indexed by a momentum scale @xmath2 , such that fluctuations are smoothly taken into account as @xmath2 is lowered from a microscopic scale @xmath3 down to 0 . in practice
this is achieved by adding to the action @xmath4 of the system an infrared regulator term @xmath5 which vanishes for @xmath6 . for a scalar field theory ,
the regulator term is a mass - like term @xmath7=\half \sum_\q \varphi_{-\q } r_k(\q ) \varphi_\q$ ] , where the cutoff function @xmath8 is chosen such that @xmath9 for @xmath10 and @xmath11 for @xmath12 , which effectively suppresses the low - energy modes @xmath10 .
one can then define a scale - dependent partition function @xmath13 $ ] and a scale - dependent effective action @xmath14 $ ] defined as a slightly modified legendre transform ( see sec .
[ subsec_ea ] for the precise definition ) of @xmath15 $ ] . here
@xmath16 is an external source which couples linearly to the @xmath17 field and @xmath18/\delta j(\r)$ ] . in the standard implementation of the nprg , at the microscopic scale @xmath19 ,
all fluctuations are frozen by the @xmath20 term so that @xmath21=s[\phi]$ ] as in landau s ( mean - field ) theory of phase transitions .
the effective action of the original model is obtained for @xmath6 ( @xmath22 ) and can be determined by ( approximately ) solving the rg equation satisfied by @xmath23.@xcite the lattice nprg differs from the standard implementation in the initial condition.@xcite the cutoff function @xmath8 is chosen such that at the microscopic scale @xmath19 the action @xmath24 corresponds to the local limit of decoupled sites .
local fluctuations are therefore included from the very beginning of the rg procedure .
the intersite coupling is then gradually restored as @xmath2 decreases from @xmath3 down to 0 . in the low - energy limit @xmath25 , @xmath8 acts as an infrared regulator suppressing fluctuations with momenta @xmath10 .
the lattice nprg is then equivalent to the standard nprg and yields identical results for the critical properties .
the hallmark of the lattice nprg is thus to take into account both local and critical fluctuations in a nontrivial way . in this paper
, we present a nprg study of the bose - hubbard model@xcite at zero temperature and in dimension @xmath26 or @xmath27 .
this model has been intensively studied in the last years following the experimental observation of the superfluid mott - insulator transition of an ultracold bosonic gas in an optical lattice.@xcite phase diagram and thermodynamic quantities are known from the numerically exact lattice quantum monte carlo ( qmc ) simulations.@xcite on the other hand few studies have addressed the critical behavior at the superfluid mott - insulator transition,@xcite and most of our understanding goes back to the seminal work of fisher _ et al._@xcite the standard nprg scheme does not capture the superfluid mott - insulator transition in the bose - hubbard model .
the reason is that near the transition the mean - field solution is too far away from the actual state of the system to provide a reliable initial condition for the nprg procedure .
the two - pole structure of the local ( on - site ) single - particle propagator is crucial for the very existence of the transition ( see , e.g. , ref . ) .
it is however impossible to reproduce this structure from a rg approach starting from the mean - field ( bogoliubov ) theory within standard approximations of the rg equation satisfied by the effective action @xmath23 .
this prevents a straightforward generalization of recent nprg studies@xcite of interacting bosons to the bose - hubbard model .
by contrast the lattice nprg , which takes into account local fluctuations , is able to describe the superfluid mott - insulator transition.@xcite since the starting action @xmath24 is purely local , this approach is to some extent reminiscent of various @xmath28 expansions of the bose - hubbard model.@xcite moreover , the lattice nprg is not restricted to the computation of thermodynamic quantities and allows us to study the critical behavior at the superfluid mott - insulator transition and compare with the predictions of fisher _ et al._@xcite based on scaling arguments .
in addition to the phase diagram and the critical behavior at the superfluid mott - insulator transition , the nprg approach can also address the superfluid phase .
deep in the superfluid phase , localization effects are negligible and we expect the bogoliubov theory to provide a good description of the system . however , even in this weak correlation limit , it is known that the bogoliubov approximation breaks down below a characteristic ( ginzburg ) momentum scale @xmath0 . in perturbation theory about the bogoliubov approximation ,
the ginzburg scale manifests itself by the appearance of infrared divergences below three dimensions ( @xmath29 ) .
although these divergences cancel out in local gauge invariant quantities ( condensate density , sound mode velocity , etc.),@xcite they do have a physical origin : they result from the coupling between longitudinal and transverse ( phase ) fluctuations and reflect the divergence of the longitudinal susceptibility@xcite a general phenomenon in systems with a continuous broken symmetry.@xcite the normal and anomalous self - energies , @xmath30 and @xmath31 , are non - analytic functions of @xmath32 and @xmath33 when @xmath34 and @xmath29 ( @xmath35 denotes the velocity of the sound mode ) , while @xmath36 vanishes,@xcite in marked contrast with the bogoliubov approximation where the linear spectrum and the superfluidity rely on a finite value of the anomalous self - energy .
a weakly - correlated superfluid is defined by the condition @xmath37 where the healing scale @xmath1 is the inverse of the healing length @xmath38.@xcite in this case , the bogoliubov theory applies to a large part of the spectrum where the dispersion is linear ( @xmath39 ) and breaks down only at very low momenta @xmath40 .
the goldstone regime @xmath40 , dominated by phase fluctuations , is conveniently described by popov s hydrodynamic theory ( free of infrared divergences ) based on a density - phase representation of the boson field @xmath41.@xcite the nprg approach yields a unified description of superfluidity which includes both bogoliubov theory ( valid for @xmath42 ) and popov s hydrodynamic approach ( valid for @xmath39).@xcite the bose - hubbard model gives us the opportunity to understand the fate of the weakly - correlated superfluid phase as we increase the strength of the interactions and move closer to the mott insulating phase in the phase diagram .
the paper is organized as follows . in sec .
[ sec_nprg ] , we derive the lattice nprg formalism for the bose - hubbard model .
we introduce the scale - dependent effective action @xmath23 and compute its initial value @xmath43 by solving the single - site bose - hubbard model .
we show that @xmath43 reproduces the result of the strong - coupling random - phase approximation ( rpa).@xcite we also discuss the approximations used to solve the flow equation satisfied by @xmath23 .
the phase diagram obtained from the nprg equations is in very good quantitative agreement with the qmc results ( sec .
[ sec_phase_dia ] ) .
furthermore , the critical behavior derived from the nprg analysis is in complete agreement with the predictions of fisher _ et al . _
based on scaling arguments ( sec .
[ sec_crit]).@xcite we find multicritical points in the universality class of the @xmath44-dimensional xy model where the transition takes place at constant density .
the xy critical behavior is observed in the mott gap , the condensate density , the compressibility and the superfluid stiffness when a multicritical point is approached at constant chemical potential by varying the ratio @xmath28 between the hopping amplitude and the local repulsion between particles . at a generic transition point
, we observe mean - field behavior , with logarithmic corrections in dimension @xmath26 ( corresponding to the upper critical dimension ) .
the superfluid phase is discussed in sec .
[ sec_sf ] . in the dilute limit , the renormalization - group flow
is controlled by the bogoliubov fixed point down to a characteristic ( ginzburg ) momentum scale @xmath0 which is much smaller than the inverse healing length @xmath1 .
the goldstone regime @xmath45 , dominated by phase fluctuations , is characterized by a ( relativistic ) lorentz invariance of the effective action @xmath23.@xcite in the vicinity of the multicritical points , when the density is commensurate , we identify a sharp crossover from a weakly- to a strongly - correlated superfluid phase where the condensate density and the superfluid stiffness are strongly suppressed and both @xmath0 and @xmath1 are of the order of the inverse lattice spacing .
the main results are summarized in sec .
[ sec_conclu ] .
the bose - hubbard model on a @xmath46-dimensional hypercubic lattice is defined by the ( euclidean ) action @xmath47 \nonumber \\ & - t \sum_{{\langle \r,\r ' \rangle } } \left(\psi_\r^ * \psi_{\r'}+\mbox{c.c.}\right ) \biggr\rbrace , \label{action}\end{aligned}\ ] ] where @xmath48 is a complex field and @xmath49 $ ] an imaginary time with @xmath50 the inverse temperature .
@xmath51 denotes the @xmath52 sites of the lattice .
@xmath53 is the on - site repulsion , @xmath54 the hopping amplitude between nearest - neighbor sites @xmath55 and @xmath56 the chemical potential . in the following
we will sometimes write the boson field @xmath57 in terms of two real fields @xmath58 and @xmath59 .
we set @xmath60 and take the lattice spacing as the unit length throughout the paper . for @xmath19 , @xmath61 and @xmath6 with cutoff function ( [ cutoff ] ) .
the ( green ) dashed line shows the bare dispersion @xmath62 . @xmath63 and @xmath54
is taken as the energy unit.,title="fig:",width=100 ] for @xmath19 , @xmath61 and @xmath6 with cutoff function ( [ cutoff ] ) .
the ( green ) dashed line shows the bare dispersion @xmath62 . @xmath63 and @xmath54
is taken as the energy unit.,title="fig:",width=94 ] for @xmath19 , @xmath61 and @xmath6 with cutoff function ( [ cutoff ] ) .
the ( green ) dashed line shows the bare dispersion @xmath62 . @xmath63 and @xmath54
is taken as the energy unit.,title="fig:",width=94 ] following the general strategy of the nprg , we consider a family of models with action @xmath64 indexed by a momentum scale @xmath2 varying from a microscopic scale @xmath3 down to 0 .
the regulator term is defined by @xmath65 where @xmath66 is the fourier transform of @xmath67 and the sum over @xmath32 runs over the first brillouin zone @xmath68-\pi,\pi]^d$ ] of the reciprocal lattice . the cutoff function @xmath8 modifies the bare dispersion @xmath69 of the bosons .
@xmath70 is chosen such that the effective ( bare ) dispersion @xmath71 vanishes.@xcite the action @xmath72 then corresponds to the local limit of decoupled sites ( vanishing hopping amplitude ) . by choosing @xmath73 , rather than @xmath74 as in ref .
, we ensure that @xmath70 does not modify the chemical potential but only the kinetic energy . in practice , we choose the cutoff function @xmath75 with @xmath76 , @xmath77 , @xmath78 and @xmath79 the step function ( see fig . [ fig_disp ] ) .
the @xmath2-dependent constant @xmath80 is defined below ( @xmath81 ) . since @xmath82 ,
the action @xmath83 coincides with the action ( [ action ] ) of the original model . for small @xmath2 ,
the function @xmath8 gives a mass @xmath84 to the low - energy modes @xmath10 and acts as an infrared regulator as in the standard nprg scheme.@xcite the scale - dependent effective action @xmath85 = { } & - \ln z_k[j^*,j ] + \inttau \sum_\r ( j^*_\r\phi_\r+\mbox{c.c . } ) \nonumber \\ & - \delta s_k[\phi^*,\phi ] \end{aligned}\ ] ] is defined as a ( slightly modified ) legendre transform which includes the explicit subtraction of @xmath86 $ ] .
here @xmath87 $ ] is the partition function obtained from the action @xmath88 , @xmath89 a complex external source which couples linearly to the bosonic field @xmath67 , and @xmath90}{\delta j^*_\r(\tau ) } , \quad \phi^*_\r(\tau)=\frac{\delta\ln z_k[j^*,j]}{\delta j_\r(\tau)}\ ] ] the superfluid order parameter .
the variation of the effective action with @xmath2 is governed by wetterich s
equation,@xcite @xmath91 = \half { { \rm tr}}\biggl\lbrace \partial_k r_k\left(\gamma^{(2)}_k[\phi^*,\phi ] + r_k\right)^{-1 } \biggr\rbrace , \label{wetteq}\ ] ] where @xmath92 is the second - order functional derivative of @xmath23 . in fourier space ,
the trace in ( [ wetteq ] ) involves a sum over momenta and frequencies as well as the two components of the complex field @xmath93 .
we are primarily interested in two quantities .
the first one is the effective potential defined by @xmath94 \biggl|_{\phi\;{{\rm const}}}\ ] ] where @xmath93 is a constant ( uniform and time - independent ) field .
the u(1 ) symmetry of the action implies that @xmath95 is a function of @xmath96 .
its minimum determines the condensate density @xmath97 and the thermodynamic potential ( per site ) @xmath98 in the equilibrium state .
the second quantity of interest is the two - point vertex @xmath99}{\delta\phi_{i\r}(\tau ) \delta\phi_{j\r'}(\tau ' ) } \biggl|_{\phi\;{{\rm const}}}\ ] ] which determines the one - particle propagator @xmath100 . here
the indices @xmath101 refer to the real and imaginary parts of @xmath93 [ see eq .
( [ psi12 ] ) ] .
because of the u(1 ) symmetry of the action ( [ action ] ) , the two - point vertex in a constant field takes the form@xcite @xmath102 in fourier space , where @xmath103 , @xmath33 is a matsubara frequency and @xmath104 the antisymmetric tensor . for @xmath105
, we can relate @xmath92 to the derivative of the effective potential , @xmath106 so that @xmath107 parity and time - reversal invariance imply@xcite @xmath108 the one - particle propagator @xmath100 can be written in a form analogous to ( [ gam2 ] ) or in terms of its longitudinal and transverse components , @xmath109 where @xmath110 with @xmath111 .
note that the single - particle propagator entering the flow equation ( [ wetteq ] ) is defined by @xmath112 , which is the propagator associated with the true legendre transform , rather than @xmath113 .
since the action @xmath114 corresponds to the local limit , the initial value of the effective action reads @xmath115 = \gamloc[\phi^*,\phi ] + \inttau \sum_\q \phi^*(\q ) t_\q \phi(\q ) , \label{gamlambda}\ ] ] where @xmath116 = -\ln \zloc[j^*,j ] + \inttau \sum_\r ( j^*_\r\phi_\r+\mbox{c.c . } ) \label{gamloc}\ ] ] is the legendre transform of the thermodynamic potential @xmath117 $ ] in the local limit . in eq .
( [ gamloc ] ) , @xmath16 is related to @xmath93 by the relation @xmath118/\delta j^*_r(\tau)$ ] and @xmath119 is the partition function obtained from @xmath120 .
it is not possible to compute the functional @xmath121 $ ] for arbitrary time - dependent fields .
one can however easily obtain the effective potential @xmath122 and the two - point vertex @xmath123 in a time - independent field @xmath93 .
these quantities are sufficient to specify the initial conditions of the flow within the approximations that we consider in sec .
[ subsec_rgeq ] . vs
external source @xmath16 ( here assumed real ) in the local limit for various values of the chemical potential @xmath56 .
( bottom ) effective potential @xmath124.,width=245 ] vs external source @xmath16 ( here assumed real ) in the local limit for various values of the chemical potential @xmath56 .
( bottom ) effective potential @xmath124.,width=245 ] to obtain @xmath125 and @xmath123 in a time - independent field , it is sufficient to consider a single site with time - independent complex external source @xmath16 .
the corresponding hamiltonian reads @xmath126 where @xmath127 ( @xmath128 ) is a creation ( annihilation ) operator and @xmath129 .
in the basis @xmath130 [ @xmath131 with @xmath132 integer ] , the hamiltonian is represented by a tridiagonal matrix , @xmath133 \nonumber \\ & - \delta_{m+1,m ' } j^ * \sqrt{m+1 } - \delta_{m-1,m ' } j \sqrt{m } , \end{aligned}\ ] ] which can be numerically diagonalized in the truncated hilbert space @xmath134 .
the low - energy eigenstates are independent of @xmath135 if the latter is large enough . if we denote by @xmath136 the source - dependent eigenstates and eigenvalues with @xmath137 the ground state
we obtain the superfluid order parameter @xmath138 and the effective potential @xmath139 ( @xmath96 ) in the zero - temperature limit @xmath50 .
figure [ fig_vloc ] shows the superfluid order parameter @xmath93 as a function of the external source @xmath16 , and the local effective potential @xmath124 obtained by numerically inverting ( [ phij ] ) .
the special case where the ground state in the local limit is degenerate for @xmath140 ( @xmath141 integer ) is discussed in appendix [ app_local ] .
vs @xmath142 for various values of @xmath143 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right ) . in the large frequency limit , @xmath148 [ eqs .
( [ dinf])].,title="fig:",width=151 ] vs @xmath142 for various values of @xmath143 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right ) . in the large frequency limit , @xmath148 [ eqs .
( [ dinf])].,title="fig:",width=151 ] vs @xmath142 for various values of @xmath143 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right ) . in the large frequency limit , @xmath148 [ eqs .
( [ dinf])].,title="fig : " ] vs @xmath142 for various values of @xmath143 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right ) . in the large frequency limit , @xmath148 [ eqs .
( [ dinf])].,title="fig : " ] vs @xmath142 for various values of @xmath143 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right ) . in the large frequency limit , @xmath148 [ eqs .
( [ dinf])].,title="fig : " ] vs @xmath142 for various values of @xmath143 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right ) . in the large frequency limit , @xmath148 [ eqs .
( [ dinf])].,title="fig : " ] to determine the two - point vertex @xmath149 , we start from the ( source - dependent ) normal and anomalous local green functions @xmath150 where @xmath151 and @xmath152 is a time - ordering operator .
the fourier transforms @xmath153 and @xmath154 are easily expressed in terms of the eigenstates @xmath155 of the hamiltonian , @xmath156 , \\
\gan(i\w ) & = - \sum_{\alpha\neq 0 } { \langle\alpha| } \hat b { |0\rangle}{\langle0| } \hat b { |\alpha\rangle } \frac{2(e_\alpha - e_0)}{\w^2+(e_\alpha - e_0)^2 } . \end{split}\ ] ] from the relation @xmath157 , we obtain @xmath158 , \\
\gamma_{{\rm loc},b}(i\w;n ) & = \frac{\gan(i\w)}{nd } , \\ \gamma_{{\rm loc},c}(i\w;n ) & = \frac{i}{2d } [ \gn(i\w)-\gn(-i\w ) ] , \end{split}\ ] ] where @xmath159 .
@xmath160 is expressed in terms of the condensate density @xmath143 ( rather than the external source @xmath16 ) by inverting ( [ phij ] ) .
the large frequency limit of the two - point vertex is given by@xcite @xmath161 where @xmath162 and @xmath163 are assumed real ( which corresponds to a real external source @xmath16 ) , @xmath164 and @xmath165 the asymptotic forms ( [ dinf ] ) are reached for @xmath166 .
@xmath167 and @xmath168 are shown in figs .
[ fig_deltaa ] and [ fig_deltac ] . to obtain the large - field limit of the effective potential @xmath122
, we must compute the local partition function @xmath169 $ ] for @xmath170 . using a loop expansion about the saddle - point approximation ( see appendix [ app_largefield ] ) , we find @xmath171 where @xmath172 .
the initial effective action @xmath43 [ eq . ( [ gamlambda ] ) ] treats the local fluctuations exactly but includes the intersite hopping term at the mean - field level , thus reproducing the strong - coupling rpa.@xcite the effective potential reads @xmath173 while the two - point vertex takes the rpa - like form @xmath174 expanding @xmath175 about @xmath176 , we find @xmath177 n + \calo(n^2 ) , \label{vexpand}\ ] ] where @xmath178 is determined by the local green function @xmath179 for vanishing source ( @xmath180 ) . here
@xmath181 is the number of bosons per site in the local limit : @xmath182 if @xmath183 and @xmath184 if @xmath185 .
the ground state is a mott insulator as long as @xmath186 .
thus the transition to the superfluid state is determined by the criterion @xmath187 , i.e. @xmath188 which reproduces the mean - field ( or strong - coupling rpa ) phase diagram.@xcite equation ( [ mfcrit ] ) can also be obtained from the condition @xmath189 , which signals the appearance of a pole at zero momentum and frequency in the one - particle propagator @xmath190 . in the strong - coupling rpa ,
the condensate density @xmath191 in the superfluid phase is determined by @xmath192 the hopping amplitude @xmath54 acts as a source term for the local potential @xmath122 . for @xmath193 ,
i.e. deep in the superfluid phase , the source term is large and we are effectively in the large field limit discussed in sec . [ subsubsec_largefield ] . from eqs . , and the fact that @xmath194 when @xmath193 , we then obtain @xmath195 $ ] ( with @xmath196 ) , which is nothing but the result of the bogoliubov approximation.@xcite the strong - coupling rpa reduces to the bogoliubov theory in the limit @xmath193.@xcite table [ table_init ] compares the initial conditions given by the bogoliubov theory and the strong - coupling rpa .
it should be noted that the true legendre transform is @xmath197 for @xmath19 since the action @xmath198 is local .
the lattice nprg is an expansion about the local limit .
the scale - dependent effective action @xmath23 is however the right quantity to consider to analyze the physical properties of the system with action @xmath4 ( without the regulator term ) . in @xmath23 ,
the regulator term @xmath5 is compensated , in a mean - field manner , by subtracting @xmath86 $ ] from the true legendre transform .
it follows that the physical quantities at scale @xmath2 , such as the condensate density @xmath97 , are obtained from @xmath23 rather than from the true legendre transform .
.initial conditions given by the bogoliubov approximation and the strong - coupling rpa . in the weak - coupling limit the two approximations become equivalent . [
cols="^,^,^",options="header " , ] it is instructive to compare the infrared behavior in the superfluid phase or the mott insulating phase with the critical behavior at the superfluid mott - insulator transition ( table [ table_ir ] ) .
both in the superfluid phase and at the multicritical points , the infrared behavior is characterized by a ( relativistic ) lorentz invariance . at fixed commensurate density @xmath199 .
@xmath1 : healing scale , @xmath0 : ginzburg scale , @xmath200 : josephson scale.,width=207 ] at fixed commensurate density @xmath199 .
@xmath1 : healing scale , @xmath0 : ginzburg scale , @xmath200 : josephson scale.,width=226 ] figure [ fig_density ] shows @xmath0 , @xmath1 , @xmath191 , @xmath201 , @xmath202 and @xmath35 vs @xmath28 at fixed density @xmath203 .
we see a sharp crossover between a weakly - correlated ( @xmath37 , @xmath204 and @xmath205 ) and a strongly - correlated ( @xmath206 , @xmath207 and @xmath208 ) superfluid phase as @xmath28 is decreased .
close to the multicritical point , there is a critical regime where the flow of @xmath209 , @xmath210 , @xmath211 and @xmath212 shows plateaus characteristic of critical behavior ( sec .
[ sec_crit ] ) .
the critical behavior ends at the josephson momentum scale @xmath200 ( sec .
[ subsec_multicrit ] ) , and for @xmath213 we recover the goldstone regime of the superfluid phase .
the behavior in the superfluid phase at fixed commensurate density @xmath199 ( @xmath199 integer ) is summarized in fig .
[ fig_carac ] . except for lattice effects ( which force @xmath1 , @xmath0 and @xmath200 to be at most of order @xmath3 )
we recover the behavior of the @xmath44-dimensional o(@xmath52 ) model.@xcite , superfluid stiffness @xmath202 and characteristic scales @xmath1 and @xmath0 vs density @xmath199 for @xmath214 ( @xmath26 ) .
black dots show fits @xmath215 and @xmath216 [ eq . ( [ dilute4a],[khlambh])].,width=207 ]
figure [ fig_density1 ] shows @xmath1 and @xmath0 at fixed @xmath28 for @xmath199 varying between 0 and 1 . in the small density limit @xmath217
, our numerical results for @xmath1 and @xmath0 agree with eqs .
( [ dilute2a],[khlambh ] ) . for @xmath218 , @xmath0 and @xmath1
become of the same order and @xmath219 as expected for a strongly - correlated superfluid phase .
the behavior near the mott insulating phase ( @xmath220 ) is similar to the low - density limit and reflects the fact that the transitions from the superfluid phase to the vacuum or the mott insulating phase @xmath221 belong to the same universality class . in particular , near the transition to the mott insulating phase @xmath203 , the system is effectively in the weakly correlated limit ( @xmath222 ) in agreement with the fact that the quantum critical point is gaussian for @xmath223 .
the analogy between the limits @xmath224 and @xmath225 leads to interesting consequences which will be discussed elsewhere.@xcite the knowledge of the infrared limit of the one - particle green s function enables us to obtain the spectral function@xcite @xmath226 \end{aligned}\ ] ] in the low - energy limit . from eqs .
( [ green1 ] ) we deduce @xmath227 \nonumber \\ & + \frac{{{\rm sgn}}(\w)}{4\pi n_0c } \frac{\theta(\w - c|\q|)}{\sqrt{\w^2-c^2\q^2 } } \label{aspec } \end{aligned}\ ] ] for @xmath34 .
in addition to the delta peak due to the goldstone mode , the spectral function exhibits a continuum of excitations which is a direct consequence of the singularity of the longitudinal propagator @xmath228 .
while the sound mode extends up to @xmath229 , the continuum is observed only at momenta and energies @xmath230 . in the weak - coupling limit , where the lattice does not play an important role , these results are in complete agreement with popov s hydrodynamic theory.@xcite the latter gives @xmath231 so that the ratio of spectral weights carried by the continuum and the sound mode is extremely small in the weak - coupling limit . of the sound mode [ eq . ( [ weight ] ) ] vs @xmath28 ( @xmath203 and @xmath26).,width=226 ] it should be noted that there is no qualitative difference between the weakly- and strongly - correlated superfluid phases regarding the low - energy single - particle spectrum [ eq .
( [ aspec ] ) ] . in the strong - coupling regime
however , the continuum of excitations due to the singular longitudinal propagator is expected to extend up to momenta of order @xmath232 ( i.e. over most part of the brillouin zone ) and carry a significant fraction of spectral weight . this expectation is confirmed by the suppression of spectral weight of the sound mode as the ratio @xmath28 is decreased at fixed density .
figure [ fig_spectral ] shows the spectral weight @xmath233 of the sound mode for @xmath203 , normalized by its value in the weakly - correlated limit @xmath234 , @xmath235 @xmath236 remains close to one in the weakly - correlated superfluid phase but is strongly suppressed in the strongly - correlated regime .
it vanishes at the transition to the mott insulating phase ( @xmath237 ) with a critical exponent @xmath238 , @xmath239 for @xmath240 .
being equivalent to the strong - coupling rpa , the initial effective action @xmath43 predicts the existence of a gapped mode in addition to the sound mode.@xcite whether this gapped mode is a true characteristic of the spectrum ( which would then show up in the propagator @xmath241 ) is an interesting question which however requires a more refined nprg analysis @xcite beyond the derivative expansion [ eq .
( [ gamde])].@xcite and @xmath203.,width=207 ] in the dilute limit @xmath217 ( sec .
[ subsec_dilute ] ) , the initial conditions of the rg flow are given by @xmath242 and reproduce the bogoliubov approximation .
the flow for @xmath243 is logarithmic and therefore very slow .
this explains why the ginzburg scale is exponentially small in the dilute limit [ eq . ( [ dilute4a ] ) ] and irrelevant for most purposes . in the infrared limit , @xmath244 and @xmath245
are given by ( [ green1 ] ) , while the longitudinal propagator @xmath246 diverges logarithmically .
a typical rg flow in the strong - coupling limit , where @xmath247 , is shown in fig .
[ fig_sf3d ] for @xmath203 .
thermodynamic quantities very rapidly converge to their @xmath6 values . on the other hand ,
the flow of @xmath248 and @xmath249 is logarithmic .
we have presented a detailed nprg study of the bose - hubbard model .
although we have only considered the zero - temperature limit , it is straightforward to extend the analysis to finite temperatures .
the lattice nprg seems to be the only available technique which treats fluctuations at all length scales on equal footing : * the lattice nprg takes into account on - site correlations which are responsible for the very existence of the superfluid mott - insulator transition .
it is exact in the local limit ( vanishing hopping amplitude ) , the latter corresponding to the initial condition of the nprg at the microscopic scale @xmath76 . *
the lattice nprg also takes into account critical fluctuations at the superfluid mott - insulator transition . in this respect
it is very similar to the standard implementation of the nprg in continuum models , the cutoff function @xmath8 playing the role of an infrared regulator .
* as already known from previous studies in continuum models,@xcite the nprg is a method of choice to study the superfluid phase .
it is free of infrared divergences , satisfies the hugenholtz - pines theorem , and is able to describe both the ( perturbative ) bogoliubov regime @xmath250 and the ( nonperturbative ) goldstone regime @xmath45 .
the latter is characterized by a vanishing anomalous self - energy and a diverging longitudinal propagator . in the strong - coupling limit where @xmath251
, there is no bogoliubov regime and the whole rg flow becomes nonperturbative .
our results agree with known results on the bose - hubbard model .
in particular we reproduce the phase diagram obtained from qmc calculations with a typical accuracy of 1 - 3 % and at a very modest numerical cost.@xcite moreover , we recover the two universality classes of the superfluid mott - insulator transition .
the lattice nprg enables a detailed study of the critical behavior near multicritical or generic transition points , which confirms the original predictions of fisher _ et al._@xcite based on scaling arguments .
we would like to thank b. delamotte for useful discussions , b. capogrosso - sansone for providing us with the qmc data shown in figs .
[ fig_dia3d ] and [ fig_dia2d ] , and p. anders for the dmft data shown in fig .
[ fig_dia3d ] .
and occupation number @xmath199 vs @xmath141 in the local limit ( with vanishing external source).,width=188 ] in this appendix , we discuss the solution of the local hamiltonian ( [ hloc ] ) when the ground state is degenerate for vanishing external sources ( @xmath141 integer ) .
this degeneracy has important consequences for the effective potential @xmath122 . when @xmath180 , the hamiltonian is diagonal in the basis @xmath130 ( sec .
[ subsec_init ] ) .
the ground state is the vacuum state @xmath252 for @xmath253 , and @xmath254 for @xmath255 .
the ground state energy @xmath256 and occupation number @xmath199 are shown in fig .
[ fig_local ] as a function of @xmath141 .
there is a quantum phase transition whenever @xmath141 is integer due to a level crossing .
when @xmath257 ( @xmath132 integer ) , the states @xmath254 and @xmath258 are degenerate . for an infinitesimal external source and @xmath257 ( @xmath132 integer ) , it is sufficient to consider the degenerate states @xmath254 and @xmath258 to determine the ground state of the hamiltonian ( [ hloc ] ) . in this subspace ,
@xmath259 with @xmath260 .
diagonalizing ( [ app6 ] ) , we find the two states @xmath261 with eigenvalues @xmath262 where @xmath263 denotes the phase of the complex source @xmath264 .
the occupation number in the ground state @xmath265 is @xmath266 while the superfluid order parameter @xmath267 is finite .
we conclude that the u(1 ) symmetry is spontaneously broken whenever @xmath141 is integer , @xmath268 although @xmath269 for @xmath270 .
( top ) and effective potential @xmath122 ( bottom ) in the local limit .
the source @xmath16 is taken real.,width=226 ] ( top ) and effective potential @xmath122 ( bottom ) in the local limit .
the source @xmath16 is taken real.,width=226 ] figure [ fig_vloc_degenerate ] shows the superfluid order parameter @xmath271 obtained from the numerical diagonalization of the full hamiltonian ( [ hloc ] ) . for @xmath272 , we find @xmath273 in agreement with eq .
( [ app7 ] ) .
the effective potential @xmath122 takes the usual form in a system with a spontaneous broken symmetry , with a flat part ( @xmath274 ) for @xmath275 .
by contrast , for @xmath276 , the superfluid order parameter @xmath277 vanishes and @xmath278 for all values of @xmath143 .
of the local effective potential for various values of @xmath56 .
the dotted lines show the large - field limit ( [ app21]).,width=226 ] the large - field limit of the effective action @xmath279 $ ] is obtained by considering the partition function @xmath280 = \int\cald[\psi^*,\psi ] e^{-\sloc[\psi^*,\psi]+\inttau ( j^*\psi+j\psi)}\ ] ] for @xmath170 . in this limit
, we expect the field to weakly fluctuate about its saddle - point value @xmath281 defined by @xmath282 let us compute the effective action by including gaussian fluctuations about the saddle - point solution @xmath281 ( one - loop order ) .
the calculation is standard and gives @xmath283 = \sloc[\phi^*,\phi ] + \half { { \rm tr}}\ln \calg^{-1}_c[\phi^*,\phi ] , \label{app20}\ ] ] where @xmath284 = -\delta(\tau-\tau ' ) \\ \times
\left ( \begin{array}{cc } \partial_{\tau'}-\mu+2u|\phi(\tau)|^2 & u \phi(\tau)^2 \\ u \phi^*(\tau)^2 & -\partial_{\tau'}-\mu+2u|\phi(\tau)|^2 \end{array } \right ) \end{gathered}\ ] ] is the inverse classical ( local ) propagator . by performing the trace in ( [ app20 ] ) for a time - independent field @xmath93
, we easily obtain the effective potential @xmath285^{1/2}+\mu - 2un \bigr\rbrace \nonumber \\ = { } & - \bar \mu n + \frac{u}{2 } n^2 + \calo(n^0 ) , \label{app21}\end{aligned}\ ] ] where @xmath286 to one - loop order , the effective potential is given by the microscopic action @xmath287 with a shift of the chemical potential .
it is straightforward to verify that higher - order contributions ( e.g. those coming from two - loop diagrams ) are at most of order @xmath288 in the large - field limit .
equation ( [ app21 ] ) is in very good agreement with the numerical calculation of @xmath122 ( fig .
[ fig_pot_der ] ) .
vs @xmath142 for various values of @xmath143 .
the dotted lines show the derivative expansion @xmath289 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right).,title="fig:",width=151 ] vs @xmath142 for various values of @xmath143 .
the dotted lines show the derivative expansion @xmath289 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right).,title="fig:",width=151 ] vs @xmath142 for various values of @xmath143 .
the dotted lines show the derivative expansion @xmath289 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right).,title="fig : " ] vs @xmath142 for various values of @xmath143 .
the dotted lines show the derivative expansion @xmath289 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right).,title="fig : " ] vs @xmath142 for various values of @xmath143 .
the dotted lines show the derivative expansion @xmath289 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right).,title="fig : " ] vs @xmath142 for various values of @xmath143 .
the dotted lines show the derivative expansion @xmath289 .
@xmath144 , @xmath145 , @xmath146 and @xmath147 ( from top left to bottom right).,title="fig : " ] figures [ fig_deltaa_de ] and [ fig_deltac_de ] show the local vertices @xmath290 and @xmath291 ( also shown in figs . [ fig_deltaa ] and [ fig_deltac ] ) together with their derivative expansions @xmath292 the derivative expansion is remarkably accurate whenever the chemical potential is negative or the condensate density large . in both limits , @xmath293 and @xmath294 .
since a negative @xmath56 or a large @xmath143 corresponds to a system deep in the superfluid phase , we conclude that the derivative expansion is fully justified in this limit . more generally , we see that the derivative expansion is always valid in the limit @xmath295 . as argued in sec
. [ subsec_rgeq ] , except deep in the mott phase ( where the strong - coupling rpa is a good approximation to the @xmath6 results ) , @xmath53 is a very large energy scale in the strong - coupling limit , and the knowledge of the vertices at energies @xmath295 is sufficient to solve the flow equations .
we therefore expect the derivative expansion to be justified also in the strong - coupling limit .
figure [ fig_vazcloc ] shows @xmath296 and @xmath297 for various values of the chemical potential @xmath56 . and @xmath298 vs @xmath143 for various values of the chemical potential @xmath56.,width=226 ] and @xmath298 vs @xmath143 for various values of the chemical potential @xmath56.,width=226 ]
the flow equations in the bmw scheme can be found in appendix c.1 of ref . .
when @xmath299 and @xmath300 are approximated by their values at the minimum @xmath97 of the effective potential , the equations simplify into @xmath301 , \\
\dl z_{c , k}(n ) = { } &
-2 n v''_k(n)^2 \dw\left [ 3j_{\text{ll},\text{lt}}(q , n)-3 j_{\text{lt},\text{ll}}(q , n ) + j_{\text{lt},\text{tt}}(q , n)- j_{\text{tt},\text{lt}}(q , n)\right]_{q=0 } \nonumber\\ & -4 n^2 v^{(3)}_k(n ) v''_k(n ) \dw\left [ j_{\text{ll},\text{lt}}(q , n)- j_{\text{lt},\text{ll}}(q , n)\right]_{q=0 } -4 n^2 v^{(3)}_k(n ) z_{c , k}'(n ) j_{\text{ll},\text{ll}}(q , n ) \nonumber\\ & -\half z_{c , k}'(n ) \left[i_{\text{tt}}(n ) + i_{\text{ll}}(n)\right]-i_{\text{ll}}(n ) n z_{c , k}''(n ) -2 n v''_k(n ) z_{c , k}'(n ) j_{\text{ll},\text{tt}}(0,n ) \nonumber \\ & -6 n v''_k(n ) z_{c , k}'(n ) j_{\text{ll},\text{ll}}(0,n ) \\
\dl v_{a , k } = { } & -n_{0,k}\lamb_k^2 \dw^2\left[\jbarlltt(q ) + 2 \jbarltlt(q ) + \jbarttll(q)\right]_{q=0 } + 2 n_{0,k } \jbarllll(0 ) z_{c , k}'(n_{0,k})^2 \nonumber \\ & + 4 n_{0,k } \lamb_k z_{c , k}'(n_{0,k } ) \dw \left[\jbarlllt(q ) -\jbarltll(q)\right]_{q=0 } , \\
\eta_{a , k } = { } & 2\lamb_k^2 n_{0,k } z_{a , k }
\eps_k k^2 \llbrace \frac{\delta_{d,2}}{2\pi } + \int_\q \theta(\eps_k-\eps_\q ) \left[\frac{\partial^2_{q_x } \eps_\q}{\eps_\q } - \frac{(\partial_{q_x}\eps_\q)^2}{\eps_\q^2 } \right ] \rrbrace \int_\w \left ( \frac{1}{d_-^2 } + \frac{1}{d_+^2 } \right ) , \label{etaa } \end{aligned}\ ] ] where @xmath302 [ z_{a , k}(4dt-\eps_k)+ v_{a , k } \w^2 + 2n_{0,k } \lamb_k]+z_{c , k}(n_{0,k})^2\w^2 , \end{split}\ ] ] @xmath303 , @xmath304 , @xmath305 , and @xmath306 .
when @xmath307 is approximated by @xmath308 and @xmath95 truncated to quadratic order [ eq .
( [ trunc ] ) ] , we obtain @xmath309 , \\ \dl n_{0,k } & = \frac{3}{2 } \ibarll + \half \ibartt \quad \mbox{if } \quad n_{0,k}>0 , \\ \dl \delta_k & = -2 \lamb_k \ibarll \quad \mbox{if } \quad n_{0,k}=0 , \\ \dl \lambda_k & = - \lambda_k^2 \bigl[9\jbarllll(0 ) - 6\jbarltlt(0 ) + \jbartttt(0 ) \bigr ] , \\ \dl z_{c , k } & = 2 \lamb_k^2 n_{0,k } \frac{\partial}{\partial \w } \bigl [ \jbarttlt(q ) - \jbarlttt(q ) -3 \jbarlllt(q ) + 3\jbarltll(q \bigr]_{q=0 } \\ \dl v_{a , k } & = - 2 \lamb_k^2 n_{0,k } \frac{\partial}{\partial \w^2 } \bigl [ \jbarlltt(q ) + \jbarttll(q ) + 2\jbarltlt(q ) \bigr]_{q=0 } , \end{split } \label{floweq}\ ] ] with @xmath211 given by ( [ etaa ] ) .
we have introduced the coefficients @xmath310 g_{k,\beta}(q+q';n ) , \end{split } \label{ijdef}\ ] ] where @xmath311 . to alleviate the notations ,
we have omitted the subscript @xmath2 in @xmath312 and @xmath313 .
the notation @xmath314 , @xmath315 and @xmath316 means that these quantities are evaluated for @xmath317 .
the green functions @xmath318 in ( [ ijdef ] ) are defined as @xmath319 , with @xmath92 approximated by its derivative expansion ( [ gamde],[gamde1 ] ) . with @xmath299 and @xmath300
approximated by their values at the minimum @xmath97 of the effective potential , this gives @xmath320 and @xmath321 where @xmath322 ^ 2 + 2 n \gamma_{b , k}(q;n ) \nonumber \\ & \times [ \gamma_{a , k}(q;n ) + r_k(\q ) ] + \gamma_{c , k}(q;n)^2 .
\end{aligned}\ ] ] the lattice cutoff function ( [ cutoff ] ) differs from cutoff functions used in the continuum , in particular due a symmetric treatment of the low- and high - energy parts of the spectrum ( fig .
[ fig_disp ] ) .
we can rewrite @xmath8 in the form @xmath323 where @xmath324 and @xmath325 this gives @xmath326 \nonumber \\ & = z_{a , k}\eps_k { { \rm sgn}}(t_\q ) \theta(1-y_\q ) [ \eta_{a , k}(1-y_\q)-2 ] . \label{app1}\end{aligned}\ ] ] @xmath8 enters the flow equations always in the combination @xmath327 where @xmath328 for @xmath329 , and @xmath330 for @xmath331 .
equations ( [ app1],[app2],[app3 ] ) lead to a significant simplification of the coefficients @xmath332 and @xmath333 . a typical contribution to @xmath312 or
@xmath313 reads @xmath334 where @xmath335 is a product of propagators @xmath336 .
since @xmath337 restricts the momentum integral to the domain @xmath338 where @xmath339 is independent of @xmath32 , @xmath340 introducing the lattice density of states @xmath341 we obtain @xmath342 \label{app5}\end{gathered}\ ] ] and @xmath343 .\end{gathered}\ ] ] since the hypercubic lattice density of states is symmetric , @xmath344 , the last equation gives @xmath345 which enables us to rewrite as @xmath346 , \end{gathered}\ ] ] with the momentum integral given by . in the limit @xmath25 ( or @xmath347 ) , the function @xmath348 involves propagators with a large gap and
is therefore negligible with respect to @xmath349 ; the flow is then governed only by the low - energy modes . moreover , for @xmath350 , the lattice does not matter and we can approximate @xmath351 , which leads to @xmath352 and @xmath353 where @xmath354 .
we then obtain @xmath355 which is the usual form for models in the continuum limit with the theta cutoff function.@xcite
let us first recall how the @xmath356-wave scattering length @xmath357 is computed from the low - energy behavior of the @xmath358 matrix in the continuum . for a contact interaction @xmath53 , the retarded @xmath358 matrix is defined by @xmath359 where @xmath360 ( @xmath361 denotes the principal part ) and @xmath362 is the dispersion of the free bosons . in three dimensions ,
@xmath363 while in two dimensions @xmath364 where @xmath365 is euler s constant .
equations ( [ app12 ] ) and ( [ app13 ] ) define the @xmath356-wave scattering length @xmath357 in three and two dimensions , respectively .
a scattering length can be defined similarly in the bose - hubbard model . at low energy ( @xmath366 )
the lattice does not matter and one can approximate the boson dispersion @xmath367 by @xmath368 .
the bosons then behaves as free particles with an effective mass @xmath369 .
thus , equations ( [ app10]-[app13 ] ) allow us to define a scattering length provided that we replace @xmath132 by @xmath370 . in the low - energy limit @xmath371 , @xmath372 where @xmath373 is the density of states of the cubic lattice .
the last integral in ( [ app14 ] ) can be computed numerically while @xmath374 for small @xmath33 .
this gives @xmath375 and @xmath376 for @xmath377 with @xmath378 and @xmath379 . to compute @xmath380 in two dimensions ,
we use @xmath381 since the first integral in the rhs of eq .
( [ app15 ] ) is convergent , we can set @xmath382 , which gives @xmath383 where @xmath384 is the catalan constant . as for the last integral in eq .
( [ app15 ] ) we obtain @xmath385 for @xmath386 . since @xmath387
we finally obtain @xmath388\ ] ] for @xmath377 with @xmath389 the low - energy limit @xmath391 of the coupling constant @xmath390 in vacuum ( @xmath392 ) can be expressed in terms of the scattering length @xmath357 .
@xmath390 can be obtained from the rg equation ( [ floweq ] ) ( see sec . [ subsec_rgeqvac ] ) or more simply from @xmath393 where @xmath394 is given by eqs .
( [ app2],[app3 ] ) with @xmath395 . for @xmath25
, we can ignore the effect of the cutoff function on the high - energy part of the spectrum ( @xmath331 ) , @xmath396 in two dimensions , we rewrite eq .
( [ app16 ] ) as @xmath401 where @xmath402 is obtained from @xmath8 by replacing @xmath403 with @xmath368 . since the last integral in ( [ app17 ] )
is convergent for @xmath404 , we can set @xmath6 and use the result ( [ app14a ] ) . the first integral in ( [ app17 ]
) can be expressed as @xmath405 we deduce @xmath406 , \end{aligned}\ ] ] where @xmath357 is the two - dimensional scattering length ( [ app14b ] ) . in the vacuum , @xmath408 , @xmath409 and @xmath410 ,
the two - point vertex is defined by @xmath411 the rg equation satisfied by @xmath390 takes the simple form @xmath412 \nonumber \\ & -10[\eps_\q+r_k(\q)]^3 \bigr\rbrace , \label{lambeq}\end{aligned}\ ] ] where @xmath413 ^ 2+\w^2 $ ] .
we deduce _ k = _ and @xmath414 @xmath415 can be computed from the local action @xmath416 in the vacuum ( @xmath417 ) , @xmath418 where @xmath419 is the local ( normal ) propagator in vacuum . from eqs .
( [ app18 ] ) and ( [ app19 ] ) , we recover eq .
( [ app16 ] ) .
we now consider the rg equations at finite density but in the dilute limit ( @xmath421 ) for @xmath420 . to leading order in @xmath422
, @xmath390 satisfies the equation ( [ lambeq ] ) . to obtain the equation satisfied by @xmath97
, we must expand the propagator to first order in @xmath422 , @xmath423 ^ 2 , \\
g_{\rm tt}(q ) & = - \frac{\eps_\q+r_k(\q)}{d } - \frac{2\lamb_k n_{0,k}}{d^2 } \w^2 , \end{split}\ ] ] where @xmath424 is defined in sec .
[ subsec_rgeqvac ] .
we will show below that the flow of @xmath80 , @xmath425 and @xmath249 leads to higher - order corrections .
we can therefore set @xmath80 , @xmath425 and @xmath249 to their vacuum values .
this gives @xmath426 \nonumber \\ & -3 [ \eps_\q+r_k(\q)]^3 \bigr\rbrace .
\label{n0eq}\end{aligned}\ ] ] eq .
( [ lambn0 ] ) follows from ( [ lambeq ] ) and ( [ n0eq ] ) .
let us now show that the flow of @xmath80 , @xmath425 and @xmath249 give subleading contributions to eq .
( [ n0eq ] ) . to leading order in @xmath422 ,
the rg equation of @xmath425 reads @xmath427 ^ 3 } \nonumber \\ & = -\lamb_k^2 n_{0,k } \int_\q \theta(-t_\q ) \dl r_k(\q ) \left [ \frac{1}{\eps_k^3 } - \frac{1}{(4dt-\eps_k)^3 } \right ] .\end{aligned}\ ] ] to estimate the order of magnitude of @xmath428 , we can ignore the term @xmath429 ( which is smaller than @xmath430 ) and use the approximate density of states ( [ dosapp ] ) .
this gives @xmath431 where we have used @xmath432 to leading order .
integrating this equation between @xmath19 and @xmath25 , we obtain z_c , k -1 ~ ( ignoring the dependence of @xmath390 on @xmath2 ) . in the dilute limit and for @xmath433 ,
the rhs is always small .
for example , in three dimensions , one finds @xmath434 .
similarly , we obtain @xmath435 ^ 4 } \nonumber \\ & \sim - 6\lamb_k \frac{v_d}{dt^2 } k_h^2 k^{d-6 } \label{vaeq}\end{aligned}\ ] ] and therefore v_a , k ~ [ vaeq1 ] for @xmath25 . since the frequency integrals extend up to @xmath436 while @xmath437 , the @xmath438 term in the propagators can be neglected if @xmath439 .
equation ( [ vaeq1 ] ) shows that this is indeed the case .
for example , in three dimensions @xmath440 .
85ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty \doibase doi:10.1016/s0370 - 1573(01)00098 - 9 [ * * , ( ) ] @noop link:\doibase 10.1103/physreve.82.041128 [ * * , ( ) ] link:\doibase 10.1103/physrevb.40.546 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.81.3108 [ * * , ( ) ] \doibase doi:10.1038/415039a [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.130403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.080404 [ * * , ( ) ] link:\doibase 10.1103/physrevb.75.134302 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.015602 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/14/7/003 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/80/50007 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.190401 [ * * , ( ) ] link:\doibase 10.1103/physreva.80.043627 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.064504 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.053603 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.120601 [ * * , ( ) ] link:\doibase 10.1103/physreva.82.063632 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.172501 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/26/7/012 [ * * , ( ) ] link:\doibase 10.1103/physrevb.53.2691 [ * * , ( ) ] link:\doibase 10.1103/physreva.79.053631 [ * * , ( ) ] link:\doibase 10.1103/physreva.72.013614 [ * * , ( ) ] link:\doibase 10.1103/physreva.79.013614 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.100503 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.224515 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/18/41/019 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.024301 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.134507 [ * * , ( ) ] @noop @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrev.116.489 [ * * , ( ) ] link:\doibase 10.1016/0003 - 4916(64)90200 - 3 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreva.8.1111 [ * * , ( ) ] link:\doibase 10.1142/s0217751x99001615 [ * * , ( ) ] link:\doibase 10.1103/physrevb.59.14054 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.027203 [ * * , ( ) ] link:\doibase 10.1103/physreve.83.031120 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.69.024513 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/12/4/043010 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) link:\doibase 10.1209/0295 - 5075/22/4/004 [ * * , ( ) ] link:\doibase 10.1103/physreva.63.053601 [ * * , ( ) ] link:\doibase 10.1103/physreva.71.033629 [ * * , ( ) ] link:\doibase 10.1103/physreva.73.033617 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.235120 [ * * , ( ) ] \doibase doi:10.1016/0370 - 2693(93)90726-x [ * * , ( ) ] @noop @noop @noop @noop @noop _ _ , second edition , ( , , ) link:\doibase 10.1016/j.physletb.2005.10.086 [ * * , ( ) ] link:\doibase 10.1103/physreve.80.030103 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.024431 [ * * , ( ) ] @noop @noop \doibase doi : 10.1140/epjb / e2007 - 00296-x [ * * , ( ) ] @noop @noop link:\doibase 10.1103/physrevlett.105.096402 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevb.63.214503 [ * * , ( ) ] \doibase doi:10.1016/0031 - 9163(66)90088 - 6 [ * * , ( ) ] from fig .
[ fig_rhokappa ] , we deduce @xmath443 and @xmath444 , which gives @xmath445 .
@noop \doibase 10.1051/jp4:2004116001 [ * * , ( ) ] @noop link:\doibase 10.1103/physreva.3.1067 [ * * , ( ) ] link:\doibase 10.1103/physrevb.37.4936 [ * * , ( ) ] schick s results@xcite have also been obtained from an nprg approach ( in a continuum model ) in ref . .
@noop @noop link:\doibase 10.1103/physrevb.78.035127 [ * * , ( ) ] link:\doibase 10.1103/physrevb.75.085106 [ * * , ( ) ] @noop @noop @noop _ _
( , , ) link:\doibase 10.1103/physreva.77.023626 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0370-2693(00)00748-6 [ * * , ( ) ] | we present a nonperturbative renormalization - group approach to the bose - hubbard model . by taking as initial condition of the renormalization - group flow the ( local ) limit of decoupled sites , we take into account both local and long - distance fluctuations in a nontrivial way .
this approach yields a phase diagram in very good quantitative agreement with quantum monte carlo simulations , and reproduces the two universality classes of the superfluid mott - insulator transition .
the critical behavior near the multicritical points , where the transition takes place at constant density , agrees with the original predictions of fisher _ et al . _
[ phys .
rev .
b * 40 * , 546 ( 1989 ) ] based on simple scaling arguments . at a generic transition point , the critical behavior is mean - field like with logarithmic corrections in two dimensions . in the weakly - correlated superfluid phase ( far away from the mott insulating phase ) , the renormalization - group flow is controlled by the bogoliubov fixed point down to a characteristic ( ginzburg ) momentum scale @xmath0 which is much smaller than the inverse healing length @xmath1 . in the vicinity of the multicritical points , when the density is commensurate , we identify a sharp crossover from a weakly- to a strongly - correlated superfluid phase where the condensate density and the superfluid stiffness are strongly suppressed and both @xmath0 and @xmath1 are of the order of the inverse lattice spacing . *l * * u * * v * * h * *l * *o * *p * * s * |
in the era of big data analytics , cloud computing and internet of things , the growing demand of massive data processing challenges existing network resource allocation approaches .
huge volumes of data acquired online with distributed sensors in the presence of operational uncertainties caused by , e.g. , renewable energy sources , call for scalable and adaptive network control schemes
. scalability of a desirable approach refers to low complexity and amenability to distributed implementation , while adaptivity implies capability of online adjustment to dynamic environments .
allocation of network resources can be traced back to the seminal work of @xcite . since then , popular allocation algorithms operating in the dual domain are first - order methods based on dual gradient ascent , either deterministic @xcite or stochastic @xcite . due to the simplicity in computation and implementation , these approaches have attracted a great deal of recent interest and have been successfully applied to cloud , transportation and power grid networks ; see , e.g. , @xcite .
however , their major limitation is slow convergence , which results in high network delay . here
, the delay can be viewed as workload queuing time in a cloud network , traffic congestion in a transportation network , or energy level of batteries in a power network . to address this issue
, recent attempts aim at accelerating first- and second - order optimization algorithms @xcite .
specifically , momentum - based accelerations over first - order methods were investigated using nesterov @xcite , or , heavy - ball iterations @xcite .
though these approaches work well in static settings , their performance degrades with online scheduling , as evidenced by the increase in accumulated steady - state error @xcite . on the other hand ,
second - order methods such as decentralized quasi - newton approach and its dynamic variant developed in @xcite and @xcite , incur high overhead to compute and communicate the decentralized hessian approximations . capturing prices of resources ,
lagrange multipliers play a central role in stochastic resource allocation algorithms @xcite . given abundant historical data in an online optimization setting , a natural question arises : _ is it possible to learn the optimal prices from past data , so as to improve the quality of online resource allocation strategies ? _
the rationale here is that past data contains statistics of network states , and learning from them can aid coping with the stochasticity of future resource allocation .
a recent work attempting to address this question is @xcite , which considers resource allocation with a _ finite _ number of possible network states and allocation actions .
the learning procedure , however , involves constructing a histogram to estimate the underlying distribution of the network states and explicitly solves an empirical dual problem . while constructing a histogram is feasible for a probability distribution with finite support , quantization errors and prohibitively high complexity are inevitable for a continuous distribution with infinite support . in this context
, the present paper aims to design a novel online resource allocation algorithm that leverages online learning from historical data for stochastic optimization of the ensuing allocation stage .
the resultant approach , which we term `` learn - and - adapt '' stochastic dual gradient ( la - sdg ) method , only doubles computational complexity of the classic stochastic dual gradient ( sdg ) method . with this minimal cost
, la - sdg mitigates steady - state oscillation , which is common in stochastic first - order acceleration methods @xcite , while avoiding computation of the hessian approximations present in the second - order methods @xcite .
specifically , la - sdg only requires one more past sample to compute an extra stochastic dual gradient , in contrast to constructing costly histograms and solving the resultant large - scale problem @xcite .
the main contributions of this paper are summarized next .
1 . targeting a low - complexity online solution , la - sdg only takes an additional dual gradient step relative to the classic sdg iteration .
this additional dual gradient step plays the role of adapting the future resource allocation strategy through learning from historical data .
meanwhile , la - sdg is linked with the stochastic heavy - ball method , nicely inheriting its fast convergence in the initial stage , while reducing its steady - state oscillation .
2 . we analytically establish that the novel la - sdg approach , parameterized by a positive constant @xmath0 , yields a very attractive cost - delay tradeoff @xmath1 $ ] , which is better than the standard tradeoff @xmath2 $ ] of the classic sdg method @xcite .
numerical tests further corroborate the performance gain of la - sdg over existing resource allocation schemes .
_ outline_. the rest of the paper is organized as follows .
the stochastic network resource allocation problem is formulated in section ii , along with the standard sdg approach introduced in section iii .
the la - sdg method is the subject of section iv .
performance analysis of la - sdg is carried out in section v. numerical tests are provided in section vi , followed by concluding remarks in section vii . _
notation_. @xmath3 denotes the expectation operator , @xmath4 stands for probability ; @xmath5 stands for vector and matrix transposition , and @xmath6 denotes the @xmath7-norm of a vector @xmath8 .
inequalities for vectors , e.g. , @xmath9 , are defined entry - wise .
the positive projection operator is defined as @xmath10^+:=\max\{a,0\}$ ] , also entry - wise .
in this section , we start with a generic network model and its resource allocation task in section [ subsec.pf ] , and then introduce a specific example of resource allocation in cloud networks in section [ subsec.exp ] .
the proposed approach is applicable to more general network resource allocation tasks such as geographical load balancing in cloud networks @xcite , traffic control in transportation networks @xcite , and energy management in power networks @xcite .
consider discrete time @xmath11 , and a network represented as a directed graph @xmath12 with nodes @xmath13 and edges @xmath14 .
collect the workloads across edges @xmath15 in a resource allocation vector @xmath16 .
the @xmath17 node - incidence matrix is formed with the @xmath18-th entry @xmath19 we assume that each row of @xmath20 has at least one @xmath21 entry , and each column of @xmath20 has at most one @xmath21 entry , meaning that each node has at least one outgoing link , and each link has at most one source node . with @xmath22 collecting the random workload arrival rates of all nodes per slot @xmath23 , the aggregate ( endogenous plus exogenous ) workloads of all nodes are @xmath24 . if the @xmath25-th entry of @xmath24 is positive , there is service residual queued at node @xmath25 ; otherwise , node @xmath25 over - serves the current arrival . with a workload queue per node ,
the queue length vector @xmath26^{\top}\in \mathbb{r}_+^{i}$ ] obeys the recursion @xmath27^{+}\!,~\forall t.\ ] ] defining @xmath28 as the aggregate network cost parameterized by the random vector @xmath29 , the local cost per node @xmath25 is @xmath30 , and @xmath31 .
the model here is quite general .
the duration of time slots can vary from ( micro-)seconds in cloud networks , minutes in road networks , to even hours in power networks ; the nodes can present the distributed front - end mapping nodes and back - end data centers in cloud networks , intersections in traffic networks , or , buses and substations in power networks ; the links can model wireless / wireline channels , traffic lanes , and power transmission lines ; while the resource vector @xmath32 can include the size of data workloads , the number of vehicles , or the amount of energy .
vector @xmath33 can represent user requests in data queues , cars waiting at road intersections , and amounts of energy in batteries . concatenating the random parameters into a random state vector @xmath34^{\top}$ ] , the resource allocation task is to determine the allocation @xmath32 in response to the observed ( realization ) @xmath35 `` on the fly , '' so as to minimize the _ long - term average _ network cost subject to queue stability at each node , and operation feasibility at each link .
concretely , we have where @xmath37 is the optimal objective of problem given all the future information ; @xmath3 is taken over @xmath35 as well as possible randomness of optimization variables @xmath32 and @xmath33 ; constraints ensure the queue stability ; and constrains the instantaneous allocation variables to be within a time - invariant box constraint set @xmath38 , which is specified by , e.g. , link capacities , or , server / generator capacities .
the queue dynamics in couple the optimization variables over an infinite time horizon , which implies that the decision variable at current slot will have effect on all the future decisions .
therefore , finding an optimal solution to calls for dynamic programming @xcite , which is known to suffer from the `` curse of dimensionality '' and intractability in an online setting . in section [ subsec.reform ] , we will circumvent this obstacle by relaxing - to limiting average constraints , and employing dual decomposition techniques .
the geographic load balancing task in cloud networks @xcite takes the form of with @xmath39 being the set of mapping nodes ( e.g. , dns servers ) , @xmath40 data centers , and @xmath41 links connecting mapping nodes with data centers . per time
@xmath23 , each mapping node @xmath42 collects the amount of user data requests @xmath43 , and forwards the amount @xmath44 to data center @xmath45 constrained by the bandwidth availability . each data center @xmath45 schedules workload processing @xmath46 according to its resource availability . to match the definition of node - incidence matrix in , we assume a virtual outgoing node from each data center @xmath45 , so that @xmath46 is the amount of resources on this outgoing link .
the bandwidth limit of link @xmath47 is @xmath48 , and the resource capability of data center @xmath45 is @xmath49 . in this case
, we have @xmath50^{\top}$ ] , @xmath51^{\top}$ ] , and @xmath52^{\top}$ ] . at each mapping node and data center ,
undistributed or unprocessed workloads are buffered in queues obeying ; see fig .
[ fig : system ] for a system diagram .
, mapping node @xmath42 has an exogenous workload @xmath43 plus that stored in the queue @xmath53 , and schedules workload @xmath44 to data center @xmath45 .
data center @xmath45 serves an amount of workload @xmath46 out of the assigned @xmath44 as well as that stored in the queue @xmath54 .
the thickness of each edge is proportional to its capacity.,scaledwidth=49.0% ] [ fig : system ] performance is characterized by the aggregate cost of power consumed at the data centers plus the bandwidth costs at the mapping nodes , namely @xmath55 the power cost @xmath56 , parameterized by the random vector @xmath57 , captures the local marginal price and the renewable generation at data center @xmath45 during time period @xmath23 .
the bandwidth cost @xmath58 , parameterized by the random vector @xmath59 , characterizes the heterogeneous cost of data transmission due to spatio - temporal differences . with the goal of minimizing the time - average of
, geographical load balancing fits the formulation in .
in this section , the dynamic problem is reformulated to a tractable form , and classical stochastic dual gradient ( sdg ) approach is revisited , along with a brief discussion of its online performance .
recall in section [ subsec.pf ] that the main challenge of solving resides in time - coupling constraints and unknown distribution of the underlying random processes .
regarding the first hurdle , combining with , it can be shown that in the long term , workload arrival and departure rates must satisfy the following necessary condition ( * ? ? ? * theorem 2.8 ) @xmath60\leq \mathbf{0}\ ] ] given that the initial queue length is finite , i.e. , @xmath61 .
in other words , on average all buffered delay - tolerant workloads should be served .
using , a relaxed version of is @xmath62 ~~ \text{s.t.}~ \eqref{eq.probl}~\text{and}~\eqref{queue - relax}\end{aligned}\ ] ] where @xmath63 is the optimal objective for the relaxed problem .
compared to , problem eliminates the time coupling across variables @xmath64 by replacing and with .
since is a relaxed version of with the optimal objective @xmath65 , if one solves instead of , it will be prudent to derive an optimality bound on @xmath37 , provided that the sequence of solutions @xmath66 obtained by solving is feasible for the relaxed constraints and . regarding the relaxed problem , using arguments similar to those in ( * ? ? ?
* theorem 4.5 ) , it can be shown that if the random state @xmath35 is independent and identically distributed ( i.i.d . ) over time @xmath23 , there exists a _
stationary _ control policy @xmath67 , which is a pure ( possibly randomized ) function of the realization of random state @xmath35 ( or the _ observed _ state @xmath35 ) ; i.e. , it satisfies , as well as guarantees that @xmath68 = \tilde{\psi}^{*}$ ] and @xmath69\leq \mathbf{0}$ ] .
as the optimal policy @xmath67 is time invariant , it implies that the _ dynamic _ problem is equivalent to the following time - invariant _ ensemble _ program where @xmath71 , @xmath72 , and @xmath73 ; set @xmath74 is the sample space of @xmath35 , and the constraint holds almost surely .
observe that the index @xmath23 in can be dropped , since the expectation is taken over the distribution of random variable @xmath35 , which is time - invariant .
leveraging the equivalent form , the remaining task boils down to finding the optimal policy that achieves the minimal objective in and obeys the constraints and .
, which can be time - independent ( * ? ? ?
* theorem 4.5 ) . ]
note that the optimization in is with respect to a stationary policy @xmath75 , which is an infinite dimensional problem in the primal domain .
however , there is a finite number of expected constraints [ cf . ] .
thus , the dual problem contains a finite number of variables , hinting to the effect that solving is tractable in the dual domain @xcite . with @xmath76 denoting the lagrange multipliers associated with , the lagrangian of is @xmath77\ ] ] where the instantaneous lagrangian is @xmath78 and constraint remains implicit .
notice that the instantaneous objective @xmath79 and the instantaneous constraint @xmath24 are both parameterized by the observed state @xmath34^{\top}$ ] at time @xmath23 ; thus , the instantaneous lagrangian can be written as @xmath80 , so that @xmath81 $ ] . correspondingly , the lagrange dual function is defined as the minimum of the lagrangian over the all feasible primal variables @xcite , given by [ eq.dual-func ] @xmath82.\end{aligned}\ ] ] due to the linearity of the expectation operator and the separability of the instantaneous lagrangian defined by the realization @xmath35 , we can interchange the minimization and the expectation , and re - write the dual function in the following form @xmath83\!=\!\mathbb{e}\!\left[\min_{\mathbf{x}_t \in { \cal x}}{\cal l}_t(\mathbf{x}_t,\bm{\lambda})\right]\!.\!\!\end{aligned}\ ] ] likewise , for the instantaneous dual function @xmath84 , the dual problem of is @xmath85.\end{aligned}\ ] ] in accordance with the ensemble primal problem , we will henceforth refer to as the _ ensemble _ dual problem .
if the optimal lagrange multiplier @xmath86 associated with were known , then optimizing and consequently would be equivalent to minimizing the lagrangian @xmath87 or infinitely many instantaneous @xmath88 , over the set @xmath38 ( * ? ? ?
3.3.4 ) and @xcite .
we restate this assertion as follows .
[ prop.closedform ] consider the optimization problem in .
given a realization @xmath35 , and the optimal lagrange multiplier @xmath86 associated with the constraints , the optimal instantaneous resource allocation decision is @xmath89 where @xmath90 accounts for possibly multiple minimizers of @xmath91 .
when the realizations @xmath92 are obtained sequentially , one can generate a sequence of optimal solutions @xmath93 correspondingly for the dynamic problem . to obtain the optimal allocation in however , @xmath86 must be known .
this fact motivates our novel `` learn - and - adapt '' stochastic dual gradient ( la - sdg ) method in section [ sec.la-sdg ] . to this end
, we will first outline the celebrated stochastic dual gradient iteration ( a.k.a .
lyapunov optimization ) . to solve
, a standard gradient iteration involves sequentially taking expectations over the distribution of @xmath35 to compute the gradient .
this is challenging because the distribution of @xmath35 is usually unknown in practice .
even if the joint probability distribution functions were available , finding the expectations is not scalable as the dimensionality of @xmath35 grows .
a common remedy to this challenge is stochastic approximation @xcite , which corresponds to the following sdg iteration @xmath94^{+},\;\forall t\end{aligned}\ ] ] where @xmath0 is a positive ( and typically pre - selected constant ) stepsize .
the stochastic gradient @xmath95 is an unbiased estimate of the true gradient ; that is , @xmath96=\nabla{\cal d}(\bm{\lambda}_t)$ ] .
hence , the primal @xmath32 can be found by solving the following instantaneous sub - problems , one per @xmath23 @xmath97 the iterate @xmath98 in depends only on the probability distribution of @xmath35 through the stochastic gradient @xmath99 .
consequently , the process @xmath100 is markov with invariant transition probability when @xmath35 is stationary .
an interesting observation is that since @xmath101 , the dual iteration can be written as [ cf . ]
@xmath102^{+},\ ; \forall t\end{aligned}\ ] ] which coincides with for @xmath103 ; see also @xcite for the virtual queue interpretation of this parallelism .
thanks to its low complexity and robustness to non - stationary scenarios , sdg is widely used in various areas , including adaptive signal processing @xcite , stochastic network optimization @xcite , and energy management in power grids @xcite . for network management in particular
, this iteration entails a cost - delay tradeoff as summarized next ; see e.g. , @xcite .
[ prop.sdg ] if @xmath104 is the optimal cost in under any feasible control policy with the state distribution available , and if the sdg recursions in - achieve an @xmath105-optimal solution in the sense that @xmath106 \leq { \psi}^*+{\cal o}(\mu)\ ] ] where @xmath107 denotes the decisions obtained from , then it necessarily incurs a steady - state queue length @xmath108 , namely @xmath109={\cal o}\left(\frac{1}{\mu}\right).\ ] ] proposition [ prop.sdg ] asserts that sdg will asymptotically yield an @xmath110-optimal solution , and have steady - state queue length @xmath111 inversely proportional to @xmath0 . intuitively , sdg iteration with a constant stepsize @xmath0 will converge to a neighborhood of the optimum @xmath86 @xcite . under mild conditions ,
the optimal multiplier is bounded , i.e. , @xmath112 , so that the steady - state queue length @xmath113 naturally scales with @xmath114 ; see . as a consequence however , to achieve the near optimality ( sufficiently small @xmath0 ) , sdg suffers large average queue lengths , thus undesired average delay by little s law @xcite . to overcome this limitation
, we will develop next an _ online _ approach , which can improve sdg s cost - delay tradeoff , while still preserving its affordable complexity and adaptability .
our main approach is derived in this section , by nicely leveraging both learning and optimization tools .
its decentralized implementation is also developed . the intuition behind our learn - and - adapt stochastic dual gradient ( la - sdg ) approach is to incrementally learn network state statistics from observed data while adapting resource allocation driven by the learning process .
a key element of la - sdg could be termed as `` foresighted '' learning because instead of myopically learning the exact optimal argument from empirical data , la - sdg maintains the capability to hedge against the risk of `` future non - stationarities . ''
* initialize : * dual iterate @xmath115 , empirical dual iterate @xmath116 , queue length @xmath117 , control variable @xmath118 , and proper stepsizes @xmath0 and @xmath119 .
* resource allocation ( 1st gradient ) : * construct the effective dual variable via , observe the current state @xmath35 , and obtain resource allocation @xmath120 by minimizing online lagrangian . update the instantaneous queue length @xmath121 via @xmath122^{+},\;\forall t.\ ] ] * sample recourse ( 2nd gradient ) : * obtain variable @xmath123 by solving online lagrangian minimization with sample @xmath35 via . update the empirical dual variable @xmath124 via .
the proposed la - sdg is summarized in algorithm [ algo1 ] .
it involves the queue length @xmath33 and an empirical dual variable @xmath125 , along with a bias - control variable @xmath126 to ensure that la - sdg will attain near optimality in the steady state [ cf .
theorems [ the.queue-stable ] and [ gap - onlinela - sdg ] ] . at each time slot @xmath23 , la - sdg obtains two stochastic gradients using the current @xmath35 : one for online resource allocation , and another one for sample learning / recourse . for the first gradient ( lines 3 - 5 ) , contrary to sdg that relies on the stochastic multiplier estimate @xmath127 [ cf . ]
, la - sdg minimizes the instantaneous lagrangian @xmath128 which depends on what we term _ effective _ multiplier @xmath129 , given by @xmath130 variable @xmath129 also captures the effective price , which is a linear combination of the empirical @xmath125 and the queue length @xmath33 , where the control variable @xmath0 tunes the weights of these two factors , and @xmath126 controls the bias of @xmath129 in the steady state @xcite . as a single pass of sdg `` wastes '' valuable online samples , la - sdg resolves this limitation in a learning step by evaluating a second gradient ( lines 6 - 8 )
; that is , la - sdg simply finds the stochastic gradient of at the previous empirical dual variable @xmath125 , and implements a gradient ascent update as @xmath131^{+},\;\forall t\ ] ] where @xmath132 is a proper diminishing stepsize , and the `` virtual '' allocation @xmath123 can be found by solving @xmath133 notice that the multiplicative constant @xmath0 in allows for adaptation even in the steady state ( i.e. , @xmath134 ) , but the diminishing @xmath132 is for online learning , as we shall discuss below . for a better illustration of the effective price , we call @xmath125 the statistically learnt price to obtain the exact optimal argument of the expected problem
. we also call @xmath135 ( which is exactly @xmath127 as shown in ) the online adaptation term since it can track the instantaneous change of system statistics
. intuitively , a large @xmath0 will allow the effective policy to quickly respond to instantaneous variations so that the policy gains improved control of queue lengths , while a small @xmath0 puts more weight on statistical learning from historical samples so that the allocation strategy will incur less variance in the steady state .
in this sense , the proposed la - sdg can attains both statistical efficiency and adaptability . distinctly different from sdg with constant stepsize that involves a pure adaptation step ,
la - sdg takes an additional learning step [ cf .
] , which adopts gradient ascent with diminishing stepsize to find the `` best empirical '' dual variable from all observed network states .
this pair of complementary gradient steps endows la - sdg with its attractive properties . in its transient stage
, the extra gradient evaluations and empirical dual variables accelerate the convergence speed of sdg ; while in the steady stage , the empirical dual variable approaches the optimal multiplier , which significantly reduces the steady - state queue lengths . readers familiar with algorithms on statistical learning and stochastic network optimization can recognize their similarities and differences with la - sdg .
( p1 ) sdg in @xcite involves only the first part of la - sdg ( @xmath136st gradient ) , where the allocation policy purely relies on stochastic estimates of lagrange multipliers or instantaneous queue lengths , i.e. , @xmath137 .
in contrast , la - sdg further leverages statistical learning from streaming data .
( p2 ) several schemes have been developed recently for statistical learning at scale to find @xmath125 , namely , sag in @xcite and saga in @xcite .
however , directly applying @xmath138 to allocate resources causes infeasibility . for a finite time @xmath23
, @xmath125 is @xmath139-optimal is @xmath139-optimal if @xmath140 , and likewise for @xmath139-feasibility . ] for , and the primal variable @xmath123 in turn is @xmath139-feasible with respect to that is necessary for . since @xmath33 essentially accumulates online constraint violations of , it will grow linearly with @xmath23 and eventually become unbounded .
the heavy - ball iteration belongs to the family of momentum - based first - order methods , and has well - documented acceleration merits in the deterministic setting @xcite .
motivated by its convergence speed in solving deterministic problems , stochastic heavy - ball methods have been also pursued recently @xcite .
the stochastic version of the heavy - ball iteration is @xcite @xmath141 where @xmath142 is an appropriate constant stepsize , @xmath143 denotes the momentum factor , and the stochastic gradient @xmath99 can be found by solving using heavy - ball iterate @xmath127 .
this iteration exhibits attractive convergence rate during the initial stage , but its performance degrades in the steady state .
recently , the performance of momentum iterations ( heavy - ball or nesterov ) with constant stepsize @xmath0 and momentum factor @xmath144 , has been proved equivalent to sdg with constant @xmath145 per iteration @xcite . since sdg with a large stepsize converges
fast at the price of considerable loss in optimality , the momentum methods naturally inherit these attributes .
to see the influence of the momentum term , consider expanding the iteration as @xmath146\nonumber\\ & = \bm{\lambda}_t+\underbrace{\mu \textstyle\sum_{\tau=1}^t\beta^{t-\tau}\nabla{\cal d}_{\tau}(\bm{\lambda}_{\tau})}_{\rm accumulated~gradient}+\underbrace{\beta^t(\bm{\lambda}_1\!-\!\bm{\lambda}_0)}_{\rm initial~state}.\end{aligned}\ ] ] the stochastic heavy - ball method will accelerate convergence in the initial stage thanks to the accumulated gradients , and it will gradually forget the initial state . as @xmath23 increases however , the algorithm also incurs a worst - case oscillation @xmath147 , which degrades performance in terms of objective values when compared to sdg with stepsize @xmath0 .
this is in agreement with the theoretical analysis in ( * ? ? ?
* theorem 11 ) .
different from standard momentum methods , la - sdg nicely inherits the fast convergence in the initial stage , while reducing the oscillation of stochastic momentum methods in the steady state . to see this ,
consider two consecutive iterations [ eq.dual-effect1 ] @xmath148 and subtract them , to arrive at @xmath149 here the equalities in follows from @xmath150 in @xmath33 recursion , and with a sufficiently large @xmath126 , the projection in rarely ( with sufficiently low probability ) takes effect since the steady - state @xmath33 will hover around @xmath151 ; see the details of theorem [ the.queue-stable ] and the proof thereof .
comparing the la - sdg iteration with the stochastic heavy - ball iteration , both of them correct the iterates using the stochastic gradient @xmath152 or @xmath153 .
however , la - sdg incorporates the variation of a learning sequence ( also known as a reference sequence ) @xmath154 into the recursion of the main iterate @xmath129 , other than heavy - ball s momentum term @xmath155 .
since the variation of learning iterate @xmath125 eventually diminishes as @xmath23 increases , keeping the learning sequence enables la - sdg to enjoy accelerated convergence in the initial ( transient ) stage compared to sdg , while avoiding large oscillation in the steady state compared to the stochastic heavy - ball method .
we formally remark this obervation next .
la - sdg offers a fresh approach to designing stochastic optimization algorithms in a dynamic environment .
while directly applying the momentum - based iteration to a stochastic setting may lead to unsatisfactory steady - state performance , it is promising to carefully design a reference sequence that exactly converges to the optimal argument .
therefore , algorithms with improved convergence ( e.g. , the second - order method in @xcite ) can also be incorporated as a reference sequence to further enhance the performance of la - sdg .
this section introduces a fully distributed implementation of la - sdg by exploiting the problem structure of network resource allocation .
for notational brevity , collect the variables representing outgoing links from node @xmath25 in @xmath156 with @xmath157 denoting the index set of outgoing neighbors of node @xmath25 .
let also @xmath158 $ ] denote the random state at node @xmath25 .
it will be shown that the learning and allocation decision per time slot @xmath23 is processed locally per node @xmath25 based on its local state @xmath159 . to this end , rewrite the lagrangian minimization for a general dual variable @xmath76 at time @xmath23 as [ cf . and ]
@xmath160 where @xmath161 is the @xmath25-th entry of vector @xmath162 , and @xmath163 denotes the @xmath25-th row of the node - incidence matrix @xmath20 .
clearly , @xmath163 selects entries of @xmath32 associated with the in- and out - links of node @xmath25 .
therefore , the subproblem at node @xmath25 is @xmath164 where @xmath165 is the feasible set of primal variable @xmath166 . in the case of , the feasible set @xmath38 can be written as a cartesian product of sets @xmath167 , so that the projection of @xmath32 to @xmath38 is equivalent to separate projections of @xmath166 onto @xmath165 .
note that @xmath168 will be available at node @xmath25 by exchanging information with the neighbors per time @xmath23 .
hence , given the effective multipliers @xmath169 ( @xmath42-th entry of @xmath129 ) from its outgoing neighbors in @xmath170 , node @xmath25 is able to form an allocation decision @xmath171 by solving the convex programs with @xmath172 ; see also .
needless to mention , @xmath173 can be locally updated via , that is @xmath174^{+}\ ] ] where @xmath175 are the local measurements of arrival ( departure ) workloads from ( to ) its neighbors . likewise , the tentative primal variable @xmath176 can be obtained at each node locally by solving using the current sample @xmath177 again with @xmath178 . by sending @xmath176 to its outgoing neighbors ,
node @xmath25 can update the empirical multiplier @xmath179 via @xmath180^{+}\ ] ] which , together with the local queue length @xmath181 , also implies that the next @xmath182 can be obtained locally . compared with the classic sdg recursion - , the distributed implementation of la - sdg incurs only a factor of two increase in computational complexity .
next , we will further analytically establish that it can improve the delay of sdg by an order of magnitude with the same order of optimality gap .
this section presents performance analysis of la - sdg , which will rely on the following four assumptions .
[ assp.iid ] the state @xmath35 is bounded and i.i.d . over time @xmath23 .
[ assp.primal ] @xmath79 in is @xmath183-strongly convex , and has @xmath184-lipschitz continuous gradient . also , @xmath79 is non - decreasing w.r.t .
all entries of @xmath32 over a convex set @xmath38 .
[ assp.slater ] there exists a stationary policy @xmath75 satisfying @xmath185 for all @xmath35 , and @xmath186\leq -\bm{\zeta}$ ] , where @xmath187 is a slack vector constant .
[ assp.dualgrad ] for any time @xmath23 , the magnitude of the constraint is bounded , that is , @xmath188 .
assumption [ assp.iid ] is typical in stochastic network resource allocation @xcite , and can be relaxed to an ergodic and stationary setting following @xcite .
assumption [ assp.primal ] ensures uniqueness of the optimal solution .
non - decreasing costs with increased resources are easily guaranteed with exponential and quadratic functions .
in addition , assumption 2 ensures that the dual function has favorable properties , which are important to ensure stability .
assumption [ assp.slater ] is slater s condition , which guarantees the existence of a bounded optimal lagrange multiplier @xcite , and is also necessary for queue stability @xcite .
assumption [ assp.dualgrad ] guarantees boundedness of the gradient of the instantaneous dual function , which is common in performance analysis of stochastic gradient - type algorithms @xcite .
building upon the desirable properties of the primal problem , we next show that the corresponding dual function satisfies both smoothness and quadratic growth properties @xcite , which will be critical to the subsequent analysis .
[ lemma.qg ] under assumption 2 , the dual function @xmath189 in is @xmath190-smooth , where @xmath191 , and @xmath192 denotes the spectral radius of @xmath193 .
in addition , if the dual variable @xmath162 lies in a compact set , there always exists a small constant @xmath194 such that @xmath195 satisfies the following quadratic growth property @xmath196 where @xmath86 is the optimal multiplier for the dual problem
. see appendix a. we start with the convergence of the empirical dual variables @xmath125 .
note that the update of @xmath125 is a classical learning process from historical data , and it is not affected by future resource allocation strategies .
therefore , the theoretical result on sdg with diminishing stepsize is directly applicable ( * ? ? ?
[ error - emp ] let @xmath125 denote the empirical dual variable in algorithm [ algo1 ] , and @xmath86 the optimal argument for the dual problem .
if the stepsize is chosen as @xmath197 , with a constant @xmath198 , a sufficient large constant @xmath199 , and @xmath200 as in assumption [ assp.dualgrad ] , then it holds that @xmath201\leq \max\{\alpha,\alpha^{-1}\}\,\frac{dm}{\sqrt{t}}\ ] ] where the expectation is over all the random states @xmath35 up to @xmath23 .
lemma [ error - emp ] asserts that using a diminishing stepsize , the dual function value converges sub - linearly to the optimal value in expectation . in principle , @xmath202 is the radius of the feasible set for the dual variable @xmath162 ( * ? ? ?
2.2 ) . however , as the optimal multiplier @xmath86 is bounded according to assumption [ assp.slater ] , one can always estimate a large enough @xmath202 , and the estimation error will only affect the constant of the sub - optimality bound through the scalar @xmath203 .
the sub - optimality bound in lemma [ error - emp ] holds in expectation , which averages over all possible sample paths @xmath204 . as a complement to lemma [ error - emp ] ,
the almost sure convergence of the empirical dual variables is established next to characterize the performance of each individual sample path .
[ emp - dual ] for the sequence of empirical multipliers @xmath154 in algorithm [ algo1 ] , if the stepsizes are chosen as @xmath205 , with constants @xmath206 defined in lemma [ error - emp ] , it holds that @xmath207 where @xmath86 is the optimal dual variable for the expected dual function minimization .
the proof follows the steps in ( * ? ? ?
* proposition 8.2.13 ) , which is omitted here .
building upon the asymptotic convergence of empirical dual variables for statistical learning , it becomes possible to analyze the online performance of la - sdg .
clearly , the online resource allocation @xmath32 is a function of the effective dual variable @xmath129 and the instantaneous network state @xmath35 [ cf . ] .
therefore , the next step is to show that the effective dual variable @xmath129 also converges to the optimal argument of the expected problem , which would establish that the online resource allocation @xmath32 is asymptotically optimal .
however , directly analyzing the trajectory of @xmath129 is nontrivial , because the queue length @xmath208 is coupled with the reference sequence @xmath154 in @xmath129 . to address this issue ,
rewrite the recursion of @xmath129 as @xmath209 where the update of @xmath129 depends on the variations of @xmath125 and @xmath33 .
we will first study the asymptotic behavior of queue lengths @xmath33 , and then derive the analysis of @xmath129 using the convergence of @xmath125 in , and the recursion .
define the time - varying target @xmath210 , which is the optimality residual of statistical learning @xmath211 plus the bias - control variable @xmath126 . per theorem
[ emp - dual ] , it readily follows that @xmath212 . by showing that @xmath33 is attracted towards the time - varying target @xmath213
, we will further derive the stability of queue lengths .
[ lem.drift ] with @xmath33 and @xmath0 denoting queue length and stepsize , there exists a constant @xmath214 , and a finite time @xmath215 , such that for all @xmath216 , if @xmath217 , it holds in la - sdg that @xmath218\leq \left\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\right\|-\sqrt{\mu},\ ; { \rm w.p.1}.\ ] ] [ pf.queue-drift ] see appendix b. lemma [ lem.drift ] reveals that when @xmath33 is large and deviates from the time - varying target @xmath213 , it will be bounced back towards the target in the next time slot . upon establishing this drift behavior of queues , we are on track to establish queue stability .
[ the.queue-stable ] with @xmath219 defined in , there exists a constant @xmath220 such that the queue length under la - sdg converges to a neighborhood of @xmath151 as @xmath221 in addition , if we choose @xmath222 , the long - term average expected queue length satisfies @xmath223={\cal o}\left(\frac{\log^2(\mu)}{\sqrt{\mu}}\right),~{\rm w.p.1}.\ ] ] [ pf.queue-stable ] see appendix c. theorem [ the.queue-stable ] in asserts that the sequence of queue iterates converges ( in the infimum sense ) to a neighborhood of @xmath224 , where the radius of neighborhood region scales as @xmath225 .
in addition to the sample path result , demonstrates that with a specific choice of @xmath126 , the queue length averaged over all sample paths will be @xmath226 .
together with theorem [ emp - dual ] , it suffices to have the effective dual variable converge to a neighborhood of the optimal multiplier @xmath86 ; that is , @xmath227 . notice that the sdg iterate @xmath127 in will also converge to a neighborhood of @xmath86 .
therefore , intuitively la - sdg will behave similar to sdg in the steady state , and its asymptotic performance follows from that of sdg .
however , the difference is that through a careful choice of @xmath126 , for a sufficiently small @xmath0 , la - sdg can improve the queue length @xmath228 under sdg by an order of magnitude .
in addition to feasibility , we formally establish in the next theorem that la - sdg is asymptotically near - optimal .
[ gap - onlinela - sdg ] let @xmath104 be the optimal objective value of under any feasible policy with distribution information about the state fully available .
if the control variable is chosen as @xmath222 , then with a sufficiently small @xmath0 , the proposed la - sdg algorithm yields a near - optimal solution for in the sense that @xmath229 \leq { \psi}^*+{\cal o}(\mu),\;{\rm w.p.1}\ ] ] where @xmath120 denotes the real - time operations obtained from the lagrangian minimization .
see appendix d. combining theorems [ the.queue-stable ] and [ gap - onlinela - sdg ] , we are ready to state that by setting @xmath230 , la - sdg is asymptotically @xmath110-optimal with an average queue length @xmath231 .
this result implies that la - sdg is able to achieve a near - optimal cost - delay tradeoff @xmath232 $ ] ; see @xcite . comparing with the standard tradeoff
@xmath2 $ ] under sdg , the learn - and - adapt design of la - sdg markedly improves the online performance in terms of delay .
note that a better tradeoff @xmath233 $ ] has been derived in @xcite under the so - termed local polyhedral assumption .
observe though , that the considered setting in @xcite is different from the one here . while the network state set @xmath234 and the action set @xmath235 in @xcite
are discrete and countable , la - sdg allows continuous @xmath234 and @xmath235 with possibly infinite elements , and still be amenable to efficient and scalable online operations . )
.,scaledwidth=31.0% ] ) .,scaledwidth=31.0% ] ) .,scaledwidth=31.0% ] ) .,scaledwidth=31.0% ]
) .,scaledwidth=32.0% ] this section presents numerical tests to confirm the analytical claims and demonstrate the merits of the proposed approach .
we consider the geographical load balancing network of section [ subsec.exp ] with @xmath236 data centers , and @xmath237 mapping nodes .
performance is tested in terms of the time - averaged instantaneous network cost in , namely @xmath238 where the energy price @xmath239 is uniformly distributed over @xmath240 $ ] ; samples of the renewable supply @xmath241 are generated uniformly over @xmath242 $ ] ; and the per - unit bandwidth cost is set to @xmath243 , with bandwidth limits @xmath244 generated from a uniform distribution within @xmath245 $ ] .
the capacities at data centers @xmath246 are uniformly generated from @xmath245 $ ] .
the delay - tolerant workloads @xmath247 arrive at each mapping node @xmath42 according to a uniform distribution over @xmath242 $ ] . clearly , the cost and the state @xmath35 here satisfy assumptions 1 and 2 . finally , the stepsize is @xmath248 , the trade - off variable is @xmath249 , and the bias correction vector is chosen as @xmath250 by default , but manually tuned in figs .
[ fig.obj-all]-[fig.delay-all ] .
we introduce two benchmarks : sdg in ( see e.g. , @xcite ) , and the projected stochastic heavy - ball in and @xmath251 by default ( see e.g. , @xcite ) .
slots).,scaledwidth=31.0% ] slots).,scaledwidth=31.0% ] performance is compared in terms of the time - averaged cost , and the instantaneous queue length in figs .
[ fig.obj ] and [ fig.queue ] . for the network cost ,
the three algorithms converge to almost the same value .
la - sdg and heavy - ball exhibit faster convergence than sdg as their running - average costs quickly arrive at the optimal operating phase by leveraging the learning process or the momentum acceleration . under this parameter configuration ( @xmath252 )
, la - sdg exhibits a much lower delay as its aggregated queue length is only 10% of that for heavy - ball and 4% of that for sdg .
clearly , our learn - and - adapt procedure improves the delay performance .
the same metrics are also compared by setting @xmath253 ; see figs .
[ fig.obj2 ] and [ fig.queue2 ] . with a larger momentum factor @xmath254 ,
the heavy - ball method exhibits a pronounced optimality loss relative to sdg in fig .
[ fig.obj2 ] , while la - sdg still converges to a time - average cost similar to that of sdg . however , by using a larger @xmath144 , the heavy - ball method incurs a much lower queue length relative to that in fig .
[ fig.queue ] . comparing to the la - sdg s queue length however , it is still much larger in the steady state .
recall that the instantaneous resource allocation can be viewed as a function of the dual variable ; see proposition [ prop.closedform ] . hence , the performance differences in figs .
[ fig.obj]-[fig.queue ] and [ fig.obj2]-[fig.queue2 ] can be also anticipated by the different behavior of dual variables . in fig .
[ fig.dual ] , the evolution of stochastic dual variables are plotted under two considered settings ; that is the dual iterate in for sdg , the momentum iteration in for the heavy - ball method , and the effective multiplier in for la - sdg .
as illustrated in , the performance of momentum iterations is similar to sdg with larger stepsize @xmath145 .
this is corroborated by fig .
[ fig.dual ] , where the stochastic momentum iterate with @xmath251 behaves similar to the dual iterates of sdg and la - sdg , but its oscillation becomes prohibitively high with a larger factor @xmath254 , which nicely explains the higher cost in fig .
[ fig.obj2 ] .
since the cost - delay performance is sensitive to the choice of parameters @xmath0 and @xmath144 , extensive experiments are further conducted among three algorithms using different values of @xmath0 and @xmath144 in figs .
[ fig.obj-all ] and [ fig.delay-all ] .
the steady - state performance is evaluated by running algorithms for sufficiently long time , up to @xmath255 slots .
the steady - state costs of all three algorithms increase as @xmath0 becomes larger , and the costs of la - sdg and the heavy - ball with small momentum factor @xmath256 are close to that of sdg , while the costs of the heavy - ball with larger momentum factors @xmath257 and @xmath254 are much larger than that of sdg .
considering steady - state queue lengths ( network delay ) , la - sdg exhibits an order of magnitude lower amount than those of sdg and the heavy - ball with small @xmath144 , under all choices of @xmath0 .
note that the heavy - ball with a sufficiently large factor @xmath254 also has a very low queue length , but it incurs a much higher cost than la - sdg in fig .
[ fig.obj-all ] due to a higher steady - state oscillation observed in fig .
[ fig.dual ] .
fast convergent resource allocation and low service delay are highly desirable attributes of stochastic network management approaches . leveraging recent advances in online learning and momentum - based optimization , a novel online approach termed la - sdg was developed in this paper .
la - sdg learns the network state statistics through an additional sample recourse procedure .
the associated novel iteration can be nicely interpreted as a modified heavy - ball recursion with an extra correction step to mitigate steady - state oscillations .
it was analytically established that la - sdg achieves a near - optimal cost - delay tradeoff @xmath1 $ ] , which is better than @xmath258 $ ] of sdg , at the cost of only one extra gradient evaluation per new datum .
our future research agenda includes novel approaches to further hedge against non - stationarity , and improved learning schemes to uncover other valuable statistical patterns from historical data .
the authors would like to thank profs .
xin wang , longbo huang and jia liu for helpful discussions . [ prop.primal-dual ] under assumptions 1 - 3 , for the constrained optimization with the optimal policy @xmath67 and its optimal lagrange multiplier @xmath86 , it holds that @xmath259=\mathbf{0}$ ] with @xmath260 , and accordingly that @xmath261 .
@xmath262+\mathbf{a}^{\top}\bm{\lambda}^*\right)^{\top}(\mathbb{e}[\mathbf{x}_t-\mathbf{x}_t^*])&\geq 0,\;\forall \mathbf{x}_t\in { \cal x}\label{eq.kkt1}\\ ( \bm{\lambda}^*)^{\top}\mathbb{e}[\mathbf{a}\mathbf{x}_t^*+\mathbf{c}_t]&=0\label{eq.kkt2}\\ \mathbb{e}[\mathbf{a}\mathbf{x}_t^*+\mathbf{c}_t]\leq \mathbf{0};\;\bm{\lambda}^*&\geq \mathbf{0}\label{eq.kkt3 } \end{aligned}\ ] ] to establish the claim , let us first assume that there exists entry @xmath45 that the inequality constraint is not active ; i.e. , @xmath263=-\zeta$ ] with the constant @xmath264 , and @xmath265 denoting the @xmath45-th row of @xmath20 .
as each row of @xmath20 has at least one entry equal to @xmath21 , we collect all indices of entries at @xmath45-th row with value @xmath21 in set @xmath266 so that @xmath267 .
since @xmath268 is feasible , we have @xmath269 , and thus @xmath270=\mathbb{e}\left[\sum_{e\in { \cal e}}\mathbf{a}_{(k , e)}(x_t^e)^*+c_t^k\right]=-\zeta\end{aligned}\ ] ] which implies that @xmath271\!=\!\zeta+\mathbb{e}[c_t^k\!+\!\textstyle\sum_{e\in { \cal e}\backslash{\cal e}_{k}^{-1}}\mathbf{a}_{(k , e)}(x_t^e)^*]>0.\!\!\!\ ] ] according , it further follows that @xmath272 since @xmath273=-(\lambda^k)^*\cdot\zeta=0 $ ] .
now we are on track to show that it contradicts with .
since @xmath274>0 $ ] , there exists at least an index @xmath42 such that @xmath275>0,\,j \in { \cal e}_{k}^{-1}$ ] . choose @xmath276 $ ] with @xmath277=\mathbb{e}[(x_t^{\tilde{j}})^*],\forall { \tilde{j}}\neq j$ ] and @xmath278=0 $ ] , to have @xmath279=[0,\ldots,-\mathbb{e}[(x_t^j)^*],\ldots,0]^{\top}$ ] .
hence , we arrive at ( with @xmath280 denoting @xmath42-th entry of gradient ) @xmath281+\mathbf{a}^{\top}\bm{\lambda}^*\right)^{\top}\mathbb{e}[(\mathbf{x}_t-\mathbf{x}_t^*)]\nonumber\\
= & -\mathbb{e}\big[\nabla_j\psi_t(\mathbf{x}_t^*)(x_t^j)^*-\sum_{i\in{\cal i } } ( \lambda^i)^*\mathbf{a}_{(i , j)}(x_t^j)^*\big]\nonumber\\ \stackrel{(a)}{=}&\underbrace{-\mathbb{e}[\nabla_j\psi_t(\mathbf{x}_t^*)(x_t^j)^*]}_{<0}-\!\!\sum_{i\in{\cal i}\backslash k } \underbrace{(\lambda^i)^*\mathbf{a}_{(i , j)}\mathbb{e}[(x_t^j)^*]}_{\geq 0}<0\end{aligned}\ ] ] where ( a ) uses @xmath272 ; the first bracket follows from assumption [ assp.primal ] since @xmath280 is monotonically increasing and @xmath282 , thus for @xmath275>0 $ ] it follows @xmath283>0 $ ] ; and the second bracket follows that @xmath284 and each column of @xmath20 has at most one @xmath21 and @xmath285 .
the proof is then complete since contradicts .
* proof of lipschitz continuity : * under assumption [ assp.primal ] , the primal objective @xmath79 is @xmath183-strongly convex , and the smooth constant of the dual function @xmath189 , or equivalently , the lipschitz constant of gradient @xmath286 directly follows from ( * ? ? ?
* lemma ii.2 ) , which equals to @xmath191 , with @xmath192 denoting the maximum eigenvalue of @xmath193 .
we omit the derivations of this result , and refer readers to that in @xcite .
[ lemma.errorbd ] ( * ? ? ?
* lemma 2.3 ) consider the dual function in and the feasible set @xmath38 in with only linear constraints .
for any @xmath162 satisfying @xmath287 and @xmath288 , we have@xmath289 where the scalar @xmath290 depends on the matrix @xmath20 as well the constants @xmath183 , @xmath184 and @xmath190 introduced in assumption [ assp.primal ] .
lemma [ lemma.errorbd ] states a _
local _ error bound for the dual function @xmath195 .
the error bound is `` local '' since it holds only for @xmath162 close enough to the optimum @xmath86 , i.e. , @xmath288 . following the arguments in @xcite however , if the dual iterate @xmath162 is artificially confined to a compact set @xmath291 such that @xmath292 with @xmath202 denoting the radius of @xmath291 ,
, one can safely find a large set @xmath291 with radius @xmath202 to project dual iterates during optimization .
] then for the case @xmath293 , the ratio @xmath294 , which implies the existence of @xmath290 satisfying for any @xmath295 .
lemma [ lemma.errorbd ] is important for establishing linear convergence rate without strong convexity @xcite .
remarkably , we will show next that this error bound is also critical to characterize the steady - state behavior of our la - sdg scheme .
[ lemma.pl ] under assumption [ assp.primal ] , the local error - bound in implies the following pl condition , namely @xmath296 where @xmath190 is the lipschitz constant of the dual gradient and @xmath290 is as in . using the @xmath190-smoothness of the dual function @xmath195 , we have for any @xmath162 and @xmath297 that @xmath298 choosing @xmath299 , and using proposition [ prop.primal-dual ] such that @xmath261 , we have @xmath300 where inequality ( a ) uses the local error - bound in .
* proof of quadratic growth : * the proof follows the main steps of that in @xcite . building upon lemma [ lemma.pl ] , we next prove lemma [ lemma.qg ] . define a function of the dual variable @xmath162 as @xmath301 . with the pl condition in , and @xmath302 denoting the set of optimal multipliers for , we have for any @xmath303 that @xmath304 which implies that @xmath305 . which describes the continuous trajectory of @xmath307 starting from @xmath308 along the direction of @xmath309 .
by using @xmath305 , it follows that @xmath310 is bounded below ; thus , the differential equation guarantees that we sufficiently reduce the value of function @xmath311 , and @xmath312 will eventually reach @xmath302 . in other words ,
there exists a time @xmath313 such that @xmath314 .
formally , for @xmath315 , we have @xmath316 since @xmath317 , we have @xmath318 , which implies that there exists a finite time @xmath313 such that @xmath319 . on the other hand ,
the path length of trajectory @xmath307 will be longer than the projection distance between @xmath320 and the closest point in @xmath302 denoted as @xmath86 , that is , @xmath321 and thus we have from that @xmath322 where ( b ) follows from . choosing @xmath313 such that @xmath323 , we have @xmath324 squaring both sides of , the proof is complete , since @xmath194 is defined as @xmath325 and @xmath320 can be any point outside the set of optimal multipliers .
since @xmath125 converges to @xmath326 according to theorem [ emp - dual ] , there exists a finite time @xmath327 such that for @xmath328 , we have @xmath329 . in such case , it follows that @xmath330 , since @xmath331 .
therefore , we have @xmath332^+-[\tilde{\bm{\theta}}_t/\mu]^+\|^2\\ \stackrel{(a)}{\leq}&\|\mathbf{q}_t+\mathbf{a}\mathbf{x}_t+\mathbf{c}_t-\tilde{\bm{\theta}}_t/\mu\|^2\nonumber\\ \stackrel{(b)}{\leq } & \|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|^2 + 2(\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu)^{\top}(\mathbf{a}\mathbf{x}_t+\mathbf{c}_t)+m^2\nonumber \end{aligned}\ ] ] where ( a ) comes from the non - expansive property of the projection operator , and ( b ) is due to the upper bound @xmath200 in assumption [ assp.dualgrad ] .
the rhs of can be upper bounded by @xmath333 where ( c ) uses the definitions @xmath334 , and @xmath335 . since @xmath24 is the stochastic subgradient of the concave function @xmath195 at @xmath336 [ cf .
] , we have @xmath337\leq { \cal d}(\bm{\gamma}_t)-{\cal d}(\bm{\lambda}^*).\ ] ] taking expectations on - over the random state @xmath35 conditioned on @xmath33 and using , we arrive at @xmath338\!\leq\!\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|^2\!+\frac{2}{\mu}\left({\cal d}(\bm{\gamma}_t)-{\cal d}(\bm{\lambda}^*)\right)+m^2\ ] ] where we use the fact that @xmath339 $ ] in . using the quadratic growth property of @xmath195 in of lemma [ lemma.qg ]
, the recursion further leads to @xmath340 \leq\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|^2\!-\frac{2\epsilon}{\mu}\|\bm{\gamma}_t-\bm{\lambda}^*\|^2+m^2\nonumber\\ \stackrel{(d)}{\leq } & \|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|^2 - 2\mu\epsilon\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|^2+m^2\end{aligned}\ ] ] where equality ( d ) uses the definitions @xmath334 and @xmath335 , implying that @xmath341 . now considering ( cf . )
@xmath342 and plugging it back into yields @xmath343 & \leq \|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|^2 - 2\sqrt{\mu}\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|+\mu\nonumber\\ & = \big(\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|-\sqrt{\mu}\big)^2.\end{aligned}\ ] ] by the convexity of @xmath344 , we further arrive at @xmath345 ^ 2&\leq \mathbb{e}\big[\|\mathbf{q}_{t+1}- \tilde{\bm{\theta}}_t/\mu\|^2\big]\nonumber\\ & \leq\left(\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\|-\sqrt{\mu}\right)^2\end{aligned}\ ] ] which directly implies the argument in the lemma . by checking vieta s formulas for second - order equations , there exists @xmath346 such that for @xmath217 , inequality holds , and thus the lemma follows readily .
* proof of in theorem [ the.queue-stable ] : * theorem [ emp - dual ] asserts that @xmath125 eventually converges to the optimum @xmath347 . hence , there always exists a finite time @xmath348 and an arbitrarily small @xmath349 such that for @xmath350 , it holds that @xmath351 . using the definition @xmath210
, it then follows by the triangle inequality that @xmath352 which also holds for @xmath121 . using and the conditional drift in lemma [ lem.drift ] , for @xmath353 and @xmath354
, it holds that
@xmath355\leq \mathbb{e}\left[\left\|\mathbf{q}_{t+1}-\tilde{\bm{\theta}}_t/\mu\right\|\big|\mathbf{q}_t\right]+\rho\nonumber\\ \leq & \left\|\mathbf{q}_t-\tilde{\bm{\theta}}_t/\mu\right\|-\sqrt{\mu } + \rho\leq \left\|\mathbf{q}_t-\bm{\theta}/\mu\right\|-\sqrt{\mu } + 2\rho.\end{aligned}\ ] ] choosing @xmath349 such that @xmath356 , then for @xmath353 and @xmath357 , we have @xmath358\leq \left\|\mathbf{q}_t-\bm{\theta}/\mu\right\|-\sqrt{\tilde{\mu}}.\ ] ] clearly , @xmath359 tracks the distance between @xmath33 and @xmath151 until the distance becomes smaller than @xmath363 for the first time ; and @xmath361 stops until @xmath364 for the first time as well . with the definitions of @xmath359 and @xmath361
, one can easily show that the recursion implies @xmath365\leq a_t - b_t\ ] ] where @xmath366 is the so - termed sigma algebra measuring the history of two processes .
as @xmath359 and @xmath361 are both nonnegative , allows us to apply the super - martingale convergence theorem ( * ? ? ?
* theorem e7.4 ) , which almost surely establishes that : ( i ) the sequence @xmath359 converges to a limit ; and ( ii ) the summation @xmath367 .
note that ( ii ) implies that @xmath368 , @xmath369 . since @xmath370
, it follows that the indicator function of @xmath361 eventually becomes null and thus @xmath371 which establishes that @xmath33 will eventually visit and then hover around a neighborhood of the reference point @xmath151 . defining the lyapunov drift as @xmath374 and taking expectations on over @xmath35 conditioned on @xmath33 , we have @xmath375\leq \mathbb{e}\big[(\bm{\gamma}_t-\bm{\lambda}^*)^{\!\top}\!(\mathbf{a}\mathbf{x}_t\!+\!\mathbf{c}_t)\big]\nonumber\\ & \qquad\qquad\qquad\qquad~+\mathbb{e}\big[(\bm{\theta}\!-\!\hat{\bm{\lambda}}_t)^{\!\top}\!(\mathbf{a}\mathbf{x}_t+\mathbf{c}_t)\big]\!+\!{\mu m^2}/{2}\nonumber\\ & \stackrel{(b)}{\leq } { \cal d}(\bm{\gamma}_t)\!-\!{\cal d}(\bm{\lambda}^*)\!+\mathbb{e}\big[(\bm{\theta}\!-\!\hat{\bm{\lambda}}_t)^{\!\top}\!(\mathbf{a}\mathbf{x}_t\!+\!\mathbf{c}_t)\big]\!+\!{\mu m^2}/{2 } \end{aligned}\ ] ] where
( b ) follows from .
summing both sides over @xmath376 , taking expectations over all possible @xmath33 , and dividing both sides by @xmath313 , we arrive at @xmath377-\mathbb{e}\left[\|\mathbf{q}_{1}-\bm{\lambda}^*/\mu\|^2\right]\right)\leq\\ & \frac{1}{t}\!\sum_{t=1}^t\mathbb{e}[{\cal d}(\bm{\gamma}_t)]\!-\!{\cal d}(\bm{\lambda}^*)\!+\!\frac{1}{t}\!\sum_{t=1}^t\mathbb{e}\big[(\bm{\theta}\!-\!\hat{\bm{\lambda}}_t)^{\!\top}\!(\mathbf{a}\mathbf{x}_t\!+\!\mathbf{c}_t)\big]\!\!+\!\frac{\mu m^2}{2}.\nonumber\end{aligned}\ ] ] first , it is easy to show that @xmath378-\mathbb{e}\left[\left\|\mathbf{q}_{1}-{\bm{\lambda}^*}/{\mu}\right\|^2\right]\right)\nonumber\\ \stackrel{(c)}{\geq } & - \lim_{t\rightarrow\infty } \frac{\mu}{2t}\mathbb{e}\left[\left\|\mathbf{q}_{1}-{\bm{\lambda}^*}/{\mu}\right\|^2\right]\stackrel{(d)}{=}0\end{aligned}\ ] ] where ( c ) holds since @xmath379 , and ( d ) follows from the boundedness of @xmath380 .
we next argue that the following equality holds @xmath381={\cal o}(\mu).\ ] ] rearranging terms in leads to @xmath382\\ = & \lim_{t\rightarrow \infty } \frac{1}{t } \sum_{t=1}^{t}\mathbb{e}\left[\left(\bm{\theta}-\bm{\lambda}^*+(\bm{\lambda}^*\!-\!\bm{\theta}\!-\!\hat{\bm{\lambda}}_t\!+\!\bm{\theta})\right)^{\top}(\mathbf{a}\mathbf{x}_t+\mathbf{c}_t)\right].\nonumber \ ] ] since @xmath125 converges to @xmath383 according to theorem [ emp - dual ] , there always exists a finite time @xmath348 such that for @xmath350 , which implies that @xmath384 .
hence , together with the cauchy - schwarz inequality , we have @xmath385 where ( d ) follows since @xmath386 and constant @xmath200 is as in assumption [ assp.dualgrad ] .
plugging into , it follows that @xmath387\\ \leq&\lim_{t\rightarrow \infty } ( { 1}/{t})\ ; \textstyle \sum_{t=1}^{t } \mathbb{e}\left[(\bm{\theta}-\bm{\lambda}^*)^{\top}(\mathbf{a}\mathbf{x}_t+\mathbf{c}_t)\right]+{\cal o}(\rho)\nonumber\\ \stackrel{(e)}{\leq}&\|\bm{\lambda}^*-\bm{\theta}\|\cdot \left\|\lim_{t\rightarrow \infty } ( { 1}/{t})\ ; \textstyle \sum_{t=1}^{t } \mathbb{e}\left[-\mathbf{a}\mathbf{x}_t-\mathbf{c}_t\right]\right\|+{\cal o}(\rho)\nonumber\end{aligned}\ ] ] where ( e ) simply follows from the cauchy - schwarz inequality . building upon
, one can follow the arguments in ( * ? ? ? * theorem 4 ) to show that there exist constants @xmath388 , and @xmath389 , for any @xmath390 , to obtain a large deviation bound as @xmath391 where @xmath220 as in . intuitively speaking , upper bounds the probability that the steady - state @xmath33 deviates from @xmath151 , and implies that the probability that @xmath392 is exponentially decreasing in @xmath393 . using the large deviation bound in
, it follows that @xmath394\\ & \stackrel{(g)}{\leq}\!\lim_{t\rightarrow \infty } \frac{1}{t } \sum_{t=1}^{t}\mathbf{1}\!\cdot\!{m}\,\mathbb{p}\left(\mathbf{q}_t < m\right)\stackrel{(h)}{\leq}\ !
md_1 e^{-d_2(\bm{\theta}/\mu-\tilde{b}-m)}\nonumber\end{aligned}\ ] ] where ( f ) holds because taking expectation in over all @xmath390 implies that the expected queue length is finite [ cf .
] , which implies the necessary condition in ; ( g ) follows from ( * ? ?
* lemma 4 ) which establishes that negative accumulated service residual @xmath395 $ ] may happen only when @xmath396 and the maximum value is bounded by @xmath397 in assumption [ assp.dualgrad ] ; and ( h ) uses the bound in by choosing @xmath398 .
setting @xmath399 in , there exists a sufficiently small @xmath0 such that @xmath400 .
together with and @xmath388 , the latter implies that @xmath401\right\|\leq \|\mathbf{1}\cdot md_1\mu^2\|={\cal o}(\mu).\ ] ] plugging into , setting @xmath402 in , and using @xmath403 , we arrive at . letting @xmath404 in
, it follows from and that @xmath405\!-\!{\cal d}(\bm{\lambda}^*)\!+{\cal o}(\mu)+\!\frac{\mu m^2}{2}\nonumber\\ & \stackrel{(h)}{\leq } { \cal d}\left(\lim_{t\rightarrow \infty}\frac{1}{t}\!\sum_{t=1}^t\mathbb{e}[\bm{\gamma}_t]\right)\!-\!{\cal d}(\bm{\lambda}^*)\!+{\cal o}(\mu)+\!\frac{\mu m^2}{2}.\end{aligned}\ ] ] where inequality ( h ) uses the concavity of the dual function @xmath195 .
defining @xmath406 $ ] , and using @xmath407 in lemma [ lemma.qg ] , implies that @xmath408 where ( i ) follows since constants @xmath200 and @xmath194 are independent of @xmath0 . from
, we can further conclude that @xmath409 . recalling the definition @xmath335
, we have that @xmath410\!-\!\frac{\bm{\lambda}^*}{\mu}\!=\!\lim_{t\rightarrow \infty}\!\frac{1}{t}\!\sum_{t=1}^t\mathbb{e}\!\left[\mathbf{q}_t\!+\!\frac{\hat{\bm{\lambda}}_t}{\mu}\right]\!-\!\frac{\bm{\lambda}^*}{\mu}\!-\!\frac{\bm{\theta}}{\mu}\nonumber\\ \stackrel{(j)}{=}&\lim_{t\rightarrow \infty}\frac{1}{t}\!\sum_{t=1}^t\mathbb{e}\left[\mathbf{q}_t\right]\!-\!\frac{\bm{\theta}}{\mu}\stackrel{(k)}{\leq } \frac{1}{\mu}\|\bm{\lambda}^*-\bm{\varphi}\|={\cal o}\left(\frac{1}{\sqrt{\mu}}\right)\end{aligned}\ ] ] where ( j ) follows from the convergence of @xmath125 in theorem [ emp - dual ] , and inequality ( k ) uses the definition of @xmath411 and @xmath412 .
recalling that @xmath399 in completes the proof . defining the lyapunov drift as @xmath413 , and squaring the queue update
, we obtain @xmath414 where ( a ) follows from the definition of @xmath200 in assumption [ assp.dualgrad ] .
multiplying by @xmath415 and adding @xmath79 , yields @xmath416 where ( b ) uses the definition of @xmath129 , and ( c ) is the definition of the instantaneous lagrangian . taking expectations on the both sides of over @xmath35 conditioned on @xmath33
, it holds that @xmath417+\mathbb{e}\left[\psi_t(\mathbf{x}_t)\big|\mathbf{q}_t\right]\nonumber\\ \stackrel{(d)}{=}&{\cal d}(\bm{\gamma}_t)+\mathbb{e}\left[(\bm{\theta}-\hat{\bm{\lambda}}_t)^{\top}(\mathbf{a}\mathbf{x}_t+\mathbf{c}_t)\big|\mathbf{q}_t\right]+{\mu m^2}/{2}\nonumber\\ \stackrel{(e)}{\leq}&{\psi}^{*}+\mathbb{e}\left[(\bm{\theta}-\hat{\bm{\lambda}}_t)^{\top}(\mathbf{a}\mathbf{x}_t+\mathbf{c}_t)\big|\mathbf{q}_t\right]+{\mu m^2}/{2}\end{aligned}\ ] ] where ( d ) follows from the definition of the dual function , while ( e ) uses the weak duality that @xmath418 , and the fact that @xmath419 ( cf . the discussion after ) .
taking expectations on both sides of over all possible @xmath33 , summing over @xmath376 , dividing by @xmath313 , and letting @xmath404 , we arrive at @xmath420\nonumber\\ \!\!\ !
\stackrel{(f)}{\leq}&{\psi}^{*}\!\!+\!\!\lim_{t\rightarrow \infty } \!\frac{1}{t } { \sum_{t=1}^{t}\mathbb{e}\!\left[(\bm{\theta}\!-\!\hat{\bm{\lambda}}_t)\!^{\top}\!(\mathbf{a}\mathbf{x}_t\!+\!\mathbf{c}_t)\right]}\!\!+\!\frac{\mu m^2}{2}\!+\!\!\lim_{t\rightarrow \infty}\!\!\!\frac{\mu\|\mathbf{q}_{1}\|^2}{2t}\nonumber\\ \!\!\!\stackrel{(g)}{\leq } & { \psi}^{*}\!+\!\lim_{t\rightarrow \infty } \frac{1}{t } { \sum_{t=1}^{t}\mathbb{e}\left[(\bm{\theta}-\hat{\bm{\lambda}}_t)^{\top}(\mathbf{a}\mathbf{x}_t+\mathbf{c}_t)\right]}\!+\!\frac{\mu m^2}{2}\!\end{aligned}\ ] ] where ( f ) comes from @xmath421\geq 0 $ ] , and ( g ) follows because @xmath422 is bounded .
one can follow the derivations in - to show , which is the second term in the rhs of .
therefore , we have from that @xmath423-{\psi}^ { * } \leq \!{\cal o}(\mu)+\frac{\mu m^2}{2}={\cal o}(\mu)\end{aligned}\ ] ] which completes the proof .
l. tassiulas and a. ephremides , `` stability properties of constrained queueing systems and scheduling policies for maximum throughput in multihop radio networks , '' _ ieee trans .
_ , vol .
37 , no . 12 , pp .
19361948 , dec . 1992 .
t. chen , x. wang , and g. b. giannakis , `` cooling - aware energy and workload management in data centers via stochastic optimization , '' _
ieee j. sel .
topics signal process . _ ,
10 , no . 2 ,
402415 , mar .
2016 .
j. gregoire , x. qian , e. frazzoli , a. de la fortelle , and t. wongpiromsarn , `` capacity - aware backpressure traffic signal control , '' _ ieee trans .
control of network systems _ , vol . 2 , no . 2 ,
pp . 164173 , june 2015 .
a. beck , a. nedic , a. ozdaglar , and m. teboulle , `` an @xmath424 gradient method for network resource allocation problems , '' _ ieee trans .
control of network systems _ , vol . 1 , no . 1 ,
6473 , mar .
j. liu , a. eryilmaz , n. b. shroff , and e. s. bentley , `` heavy - ball : a new approach to tame delay and convergence in wireless network optimization , '' in _ proc .
ieee infocom _
, san francisco , ca , apr .
2016 .
a. g. marques , l. m. lopez - ramos , g. b. giannakis , j. ramos , and a. j. caamao , `` optimal cross - layer resource allocation in cellular networks using channel - and queue - state information , '' _ ieee trans . veh .
_ , vol .
61 , no . 6 , pp . 27892807 , jul
. 2012 .
n. l. roux , m. schmidt , and f. r. bach , `` a stochastic gradient method with an exponential convergence rate for finite training sets , '' in _ proc .
advances in neural information processing systems _ , lake tahoe , nv , dec . 2012 , pp . 26632671
. a. defazio , f. bach , and s. lacoste - julien , `` saga : a fast incremental gradient method with support for non - strongly convex composite objectives , '' in _ advances in neural info . process .
_ , montral , canada , dec .
2014 , pp . 16461654 . | network resource allocation shows revived popularity in the era of data deluge and information explosion .
existing stochastic optimization approaches fall short in attaining a desirable cost - delay tradeoff . recognizing the central role of lagrange multipliers in network resource allocation , a novel learn - and - adapt stochastic dual gradient ( la - sdg ) method is developed in this paper to learn the empirical optimal lagrange multiplier from historical data , and adapt to the upcoming resource allocation strategy .
remarkably , it only requires one more sample ( gradient ) evaluation than the celebrated stochastic dual gradient ( sdg ) method .
la - sdg can be interpreted as a foresighted learning approach with _ an eye on the future _ , or , a modified heavy - ball approach from an optimization viewpoint .
it is established - both theoretically and empirically - that la - sdg markedly improves the cost - delay tradeoff over state - of - the - art allocation schemes .
first - order method , stochastic approximation , statistical learning , network resource allocation . |
the use of atmospheric erenkov telescopes to make observations at large zenith angles was suggested by sommers and elbert @xcite as a potentially efficient method of measuring gamma rays at the highest energy .
the underlying principle is that at large zenith angles , the footprint of an air shower as seen in optical erenkov light ( which is a penetrating component of the cascade ) becomes large .
this means that , although the threshold energy of a telescope will increase as the zenith angle increases , the accompanying increase in collecting area results in a useful gamma ray detection rate .
sommers and elbert discussed this large zenith angle technique in the context of first - generation ( non - imaging ) gamma ray telescopes .
markarian 501 ( mrk 501 ) is a relatively close bl lac object ( at a red shift @xmath1 of 0.034 ) which shows strong , variable , non - thermal emission from radio to x - ray wavelengths . since its discovery as a vhe gamma ray source @xcite
, mrk 501 has been extensively monitored in the vhe waveband . in 1997
, the object went into a prolonged active state at all wavebands .
the vhe activity was detected by a number of northern hemisphere facilities employing imaging techniques ( e.g. @xcite ) .
this state of high activity lasted from 1997 march october and the source exhibited a complex time history with flaring activity on a time scale as short as a few hours @xcite .
multiwavelength studies of mrk 501 during this time have been reported @xcite and the energy spectrum at tev energies has been established @xcite .
the upper bound to the tev energy spectrum of an agn is of considerable interest as it , together with the distance to the source , leads to a value for the density of intergalactic infra red photons ( see e.g. @xcite ) .
the highest energy photons observed from this source to date have been detected by the hegra group and are in excess of 25 tev @xcite .
we report here observations of mrk 501 made from the southern hemisphere at zenith angles in excess of @xmath2 using the university of durham mark 6 imaging telescope .
our measurements of cascades which develop through three atmospheres demonstrate that imaging techniques may be applied at these large zenith angles with the prospect of improving the significance of the gamma ray signal at high energies .
the university of durham mark 6 telescope is located at narrabri nsw australia and has been described in detail @xcite .
it uses the established imaging technique to separate gamma rays from background cosmic rays and incorporates a robust , noise - free trigger involving signals from three parabolic flux collectors of diameter 7 m mounted on a single alt - azimuth platform .
the imaging camera comprises 109 pixels ( 91 with @xmath3 diameter and 18 with @xmath4 diameter ) and is mounted at the focus of the central flux detector .
it combines with two low resolution cameras ( 19 pixels each of diameter @xmath4 ) which are mounted at the foci of the left and right flux collectors to provide a trigger demanding a simultaneous temporal ( 10 ns gate ) and spatial ( @xmath4 ) coincidence of the erenkov light detected in the three cameras .
this trigger system allows the telescope to detect small light flashes , and hence low energy gamma rays , free of the triggers due to light produced by individual local muons .
this is particularly important for observations at very large zenith angles when the background from local muons mimicking the signatures of gamma rays can become important .
mrk 501 was observed in 1997 july and august under clear and moonless skies during an interval of extreme activity when detection rates for gamma rays @xmath5 gev ( using conventional imaging telescopes ) exceeded 10 photons per min ( see e.g. bradbury @xcite ) .
data were recorded at zenith angles of @xmath0 using the standard mark 6 telescope procedure which involves recording data in 15 min segments .
off source observations were taken by alternately observing regions of the sky which differ by @xmath6 min in right ascension from the position of mrk 501 so ensuring that the on and off segments possess identical azimuth and zenith profiles .
this observing pattern , which involves reversal of the order of the on and off source measurements , eliminates any first order changes in telescope performance due to residual secular changes in atmospheric clarity , temperature etc .
an observing log is shown in table [ table : observing_log ] .
the background count rate of the telescope at @xmath2 to the zenith was about 70 cpm , some 10% of that at the directions close to the zenith where the telescope would normally be operated .
a preliminary estimate of the threshold for gamma ray detection with the mark 6 telescope at @xmath2 to the zenith is @xmath7 tev which may be compared with the value of @xmath8 gev for zenith angles @xmath9 .
we note that sommers and elbert @xcite estimate a factor of @xmath10 in threshold energy difference between zenith angles of @xmath11 and @xmath12 compared with the value of @xmath13 from our estimate .
@lc date & no . of scans
+ & on source + 1997 july 4 & 3 + 1997 july 5 & 5 + 1997 july 7 & 4 + 1997 july 9 & 3 + 1997 july 23 & 1 + 1997 july 28 & 5 + 1997 july 30 & 5 + 1997 july 31 & 3 + 1997 august 1 & 4 + 1997 august 5 & 4 + data were accepted for analysis only if the sky was clear and stable according to the telescope background count rate and a boresighted fir radiometer which provides a measure of sky clarity @xcite .
the gross counting rates in each pair of on - off segments used were consistent at the @xmath14 level .
a total of 9.25 hrs of data for on - source observations and an equal amount off source meet these criteria .
the reduction and analysis of accepted data follows a well established routine @xcite .
the gains and pedestals of all 147 pmts and digitizer electronics are calibrated within each 15 min segment using embedded laser and false coincidence events .
pmt noise is equalised for the on and off source segments using the software padding technique @xcite prior to identifying the accurate location of the source in the camera s field of view using the axial ccd camera .
events confined to within @xmath15 of the centre of the camera and which have in excess of 600 digital counts ( to ensure reliable image reconstruction ) are considered suitable for further analysis .
finally , the spatial moments of each image relative to the source position are evaluated and those events are rejected which are unlikely to have been initiated by gamma rays .
in addition to the background rejection based upon the image analysis , a measure of the fluctuations between the centroids of the samples recorded by the left and right collectors of the mark 6 telescope ( @xmath16 ) provide an additional discriminant @xcite .
gamma rays are identified on the basis of image shape and left / right fluctuation and then plotting the number of events as a function of the pointing parameter _
alpha_. gamma ray events from a point source appear as an excess at small values of _ alpha_. the selection criteria appropriate to observations at large zenith angles ( @xmath17 ) are summarised in table [ select_table ] .
small changes have been allowed from the standard set of parameters identified for the analysis of data taken with the mark 6 telescope at zenith angles @xmath18 to accommodate e.g. the narrower images of cascades propagated through three atmospheres .
@lccccc parameter&ranges&ranges&ranges&ranges + ( d.c.)&@xmath19&@xmath20&@xmath21&@xmath22 + _ distance_&@xmath23&@xmath23&@xmath23&@xmath23 + _ eccentricity_&@xmath24&@xmath24&@xmath24&@xmath24 + _ width_&@xmath25&@xmath26&@xmath27&@xmath27 + _ concentration_&@xmath28&@xmath28&@xmath28&@xmath29 + @xmath16&@xmath30&@xmath25&@xmath31&@xmath32 + the number of events in the on and off samples after the application of the selections described above are summarised in table [ result_table ] .
the _ alpha _ plot of the differences of the on and off distributions is shown in figure [ fig : alphaplot ] .
the excess of events at small values of _ alpha _ suggest the detection of a source with an excess of events with _ alpha _ @xmath33 at a significance of @xmath34 .
the width of the _ alpha _ distribution , with excess events with _ alpha _ of up to @xmath35 , may be a consequence of the effects of magnetic field of @xmath36 g , acting on the cascade over unusually large linear distances @xcite .
@lrrrr & on & off & difference & significance + number of events & 40869 & 41232 & @xmath37 & @xmath38 + + number of size and & 21570 & 21286 & 284 & @xmath39 + distance selected events & & & & + + number of shape & 1935 & 1750 & 185 & @xmath40 + selected events & & & & + + number of shape and & 647 & 475 & 172 & @xmath41 + selected events & & & & + the on and off source data have been re - analysed using a matrix of assumed source positions to provide a `` false source '' analysis . in figure
[ fig : mapleplot ] we show that the excess of gamma ray events originates from the source which is located in the right ascension , declination plot at a position consistent with mrk 501 .
( the camera centre does not coincide with the source position . )
observations of mrk 501 during outburst made at zenith angles @xmath42 using a southern hemisphere imaging telescope have demonstrated the efficacy of the large zenith angle act technique suggested by sommers and elbert @xcite .
the selection criteria used to reject background in observations with our imaging telescope at large zenith angles are similar to those employed for observations at small zenith angles .
the main difference is in the maximum value of the width of images accepted , which are smaller than those in observations near the zenith due to the increased distance to cascade maximum . in the absence of a network of gamma ray observatories giving full sky coverage
, this detection of mrk 501 has demonstrated that observations at large zenith angles may allow more continuous monitoring of strong tev gamma ray sources .
pending completion of a simulation study of the detailed response of the mark 6 telescope at zenith angles of @xmath43 it is not possible to ascribe a reliable value to the highest energy gamma rays detected in this short exposure observation .
thus it has not yet been possible to extend the spectrum beyond the limiting energy available from more extensive observations using atmospheric erenkov telescope techniques at smaller zenith angles from northern hemisphere observatories .
we are grateful to the uk particle physics and astronomy research council for support of the project and the university of sydney for the lease of the narrabri site .
the mark 6 telescope was designed and constructed with the assistance of the staff of the physics department , university of durham .
the efforts of mrs s e hilton and mr k tindale are acknowledged with gratitude . | tev gamma rays from the blazar markarian 501 have been detected with the university of durham mark 6 atmospheric erenkov telescope using the imaging technique at large zenith angles .
observations were made at zenith angles in the range @xmath0 during 1997 july and august when markarian 501 was undergoing a prolonged and strong flare . |
the spreading of a liquid droplet on a surface proceeds as the droplet balances the interfacial tensions between the solid , liquid , and vapor phases . on a homogenous substrate , completely wetting droplets ( i.e. where the equilibrium contact angle @xmath4 ) often form a monolayer film or precursor foot on the surface that spreads ahead of the main droplet .
this has been observed in experiments @xcite as well as simulation @xcite .
this foot grows diffusively with an effective diffusion constant @xmath5 that depends on the square root of the initial radius @xmath6 of the droplet @xcite . while much has been revealed about the kinetics of droplets spreading on homogenous surfaces , spreading on chemically patterned surfaces
is much less understood .
in particular the influence of the pattern feature size on the spreading rate of both the precursor foot and main droplet is unknown . with the advent of modern lithography techniques , it is now possible to chemically pattern a substrate with lyophobic and lyophilic surface domains on the nano to micron scale @xcite .
these chemically structured surfaces can be useful in controlling the wetting of liquid droplets .
depending on the size of the pattern and the contact angle in the two domains , wetting transitions can occur in which the liquid droplet can change its shape or morphology @xcite .
the equilibrium shapes of liquid droplets on a variety of chemically patterned surfaces has been studied both experimentally and theoretically @xcite .
introduction of a lyophobic grid can be used to confine liquid droplets and has been shown to improve the quality of ink jet printing @xcite . in most of these studies ,
the liquid droplet is confined by the pattern of lyophilic and lyophobic domains and does not spread .
an alternate situation occurs if one places a long liquid line transverse to a chemically striped surface of alternating lyophilic and lyophobic domains .
this is the situation we study here for the case of liquid polymer droplets of varying chain length . for small @xmath2 ,
the main droplet partially wets both regions of the surface equivalently and does not spread even in the wetting region . for sufficiently large @xmath2 the main droplet spreads on the wetting region of the substrate by pulling material from non - wetting regions . in most cases ,
a precursor film spreads on the wetting portion of the substrate at a rate strongly dependent upon interaction strength even when the main droplet does not spread .
here we present extensive molecular dynamics ( md ) simulations of the spreading of short chain polymer droplets spreading on a chemically patterned surface .
the polymers are modeled by a coarse grained bead spring model in which each polymer contains @xmath7 beads per chain with @xmath8 and @xmath9 . as the entanglement length @xmath10 for this model @xcite ,
the chains are unentangled .
the polymer is represented by spherical beads of mass @xmath11 attached by springs , which interact with a truncated lennard - jones ( lj ) potential , @xmath12 & r\leq r_{c}\\ 0 & r > r_{c}\end{array}\right .
\label{eq : ljcut}\ ] ] where @xmath13 and @xmath14 are the lj units of energy and length and the cutoff @xmath15 .
the monomer - monomer interaction @xmath13 is used as the reference and all monomers have the same diameter @xmath14 . for bonded monomers
, we apply an additional potential where each bond is described by the fene potential @xcite with @xmath16 and @xmath17 .
the substrate is modeled as a flat surface since it was found previously @xcite that with the proper choice of thermostat , simulations using a flat surface exhibit the same behavior as a realistic atomic substrate with greater computational efficiency .
the interactions between the surface and the monomers in the droplet at a distance @xmath18 from the surface are modeled using an integrated lj potential , @xmath19 & z\leq z_{c}\\ 0 & z > z_{c}\end{array}\right .
\label{eq : ljwall}\ ] ] where @xmath20 is the monomer - surface interaction strength and @xmath21 extending the range of this surface interaction to infinity shifts the wetting transition to a lower energy ; however , the qualitative spreading behavior is identical to the @xmath22 case @xcite . for @xmath23 ,
the critical wetting strength ( i.e. @xmath24 ) is @xmath25 for @xmath8 and @xmath26 for @xmath27 @xcite .
the system is periodic in the @xmath28 direction with length @xmath29 and open in the other two directions .
the chemical pattern consists of infinite strips of wetting and non - wetting regions of equal width @xmath30 as shown in figs . 1 - 2 and fig . 4 .
in the present simulations the monomer / wall interaction @xmath31 depends on the position of each monomer in the @xmath28 direction . for the non - wetting region ,
@xmath31 was chosen so that @xmath32 , which for @xmath8 is @xmath33 and for @xmath27 is @xmath34 @xcite . in the wetting region ,
two cases were studied for @xmath8 : @xmath35 and @xmath36 , corresponding to weak and strong wetting , respectively . for @xmath27 , @xmath37 , which is in the weak wetting regime .
all of the droplets presented here are modeled as hemicylinders as described previously @xcite with initial droplet radii @xmath38 .
the effect of varying @xmath39 for homogeneous droplets has been studied previously @xcite . to provide a realistic representation of the transfer of energy in the polymer droplet
, a langevin thermostat is applied only near the surface , @xmath40 where @xmath41 is the mass of monomer @xmath42 , @xmath43 is the position of monomer @xmath42 , @xmath44 is the friction coupling term for the langevin thermostat , @xmath45 is the force acting on monomer @xmath42 due to the potentials defined above , and @xmath46 is a gaussian white noise term .
we use a langevin coupling term with a damping rate that decreases exponentially away from the substrate @xcite , @xmath47 . here
@xmath48 and @xmath18 is the distance from the substrate .
the effect of varying @xmath49 for homogeneous surfaces has been discussed in ref .
@xcite . by including only a coupling of monomers near the surface to the thermostat
, we avoid the unphysical effect of screening the hydrodynamic interactions in the droplet .
we also damp monomers near the surface stronger than those in the bulk , which is a more physically reasonable description of dissipation mechanisms due to the surface interaction . in the simulations @xmath50
varied from @xmath51 to @xmath52 , depending on the periodicity of the pattern @xmath2 . for @xmath53 , @xmath54 while for @xmath55 , @xmath56 .
results for a homogenous wetting substrate , labeled @xmath57 are for @xmath54 .
the total size of the droplets varied from @xmath58 to @xmath59 monomers depending on the value of @xmath50 .
the vapor pressure is low so that spreading does not occur via vaporization and condensation .
however for small @xmath2 , particularly for @xmath8 , a few chains separate from the precursor foot and randomly diffuse on to the non - wetting region of the substrate and sometimes evaporate from the substrate .
these chains are subsequently removed from the simulation once they move away from the surface a distance several times @xmath39 .
the equations of motion are integrated using a velocity - verlet algorithm with a time step of @xmath60 where @xmath61 .
the simulations are performed at a temperature @xmath62 using the lammps code @xcite , modified to include the chemically patterned surface .
the instantaneous contact radius for the main droplet @xmath63 and the radius of the precursor foot @xmath64 are determined every @xmath65 according to the procedure described previously @xcite .
briefly , the system is divided into layers of thickness @xmath66 parallel to the interface .
data presented for foot kinetics are obtained from the layer closest to the surface ( layer @xmath67 ) and data for bulk kinetics are obtained from layer @xmath68 , which is far enough above the substrate to provide an accurate description of bulk droplet behavior .
time sequences for @xmath8 and @xmath27 are shown in fig . 1 for @xmath69 .
for the main droplet to spread ( bulk spreading ) , the gain in interface energy must exceed the cost of increasing the surface area of the droplet .
this balance depends on @xmath2 , @xmath7 , and @xmath31 . for @xmath8 ,
bulk spreading occurs on the lyophilic strip as material is transfered from the non - wetting to the wetting region . as such , the cross - section of material in the non - wetting region shrinks while maintaining a hemi - cylinder shape with contact angle near @xmath70 . at later time in the simulation
, material is completely removed from non - wetting regions . given periodicity in the @xmath28 direction
, this demonstrates a case where an intial line of liquid ( in @xmath28 ) transforms into a series of parallel liquid strips ( in @xmath71 ) with separation distance in @xmath28 equal to @xmath30 . for @xmath27 , evidence of @xmath7 dependence
is observed in that bulk spreading does not occur on the wetting region . a precursor film or foot
is observed to spread ahead of the main droplet on the lyophilic strip in both systems .
thus , even when the main droplet does not spread , a precursor foot may spread on the wetting region of the substrate . for all cases where
a foot was observed to spread , the contact radius follows @xmath72 where @xmath5 is the effective diffusion coefficient . for @xmath8 and @xmath69 ( fig . 1 ) , the effective diffusion constant @xmath5 of the precursor foot is essentially equal to that for the pure wetting case @xmath57 .
spreading of the precursor foot for @xmath27 causes depletion of material on the wetting region .
furthermore , bulk material is unable to move quickly enough from the non - wetting region to the wetting region , resulting in the formation of a saddle shape at the center of the droplet . at late times , the precursor film spreads by polymer moving perpendicularly from the non - wetting to the wetting region , then continuing along the wetting region . in this case
@xmath5 is @xmath73 of that for @xmath57 .
although we do not access such time scales in our simulations , we expect that material will completely wick away from the non - wetting region as the precursor film continues to spread .
figure 2 presents snapshots of @xmath8 droplets at late time for varying @xmath2 .
the snapshots are shown at different times chosen so that the size of the precursor foot , when the foot spreads ahead of the droplet , is approximately the same size in each snapshot . in the weak wetting case ( fig . 2 , left panels ) , the main droplet increases in size only slightly before saturating for @xmath74 as shown in fig . 3 .
for @xmath75 , the bulk droplet saturates at a radius @xmath76 in both the wetting and non - wetting regimes . as @xmath2 increases , the radius of the droplet in the non - wetting region decreases while the radius of the droplet in the wetting regime increases slightly . for @xmath77 , the main droplet spreads .
similar to the behavior seen in fig . 1 for @xmath69 , the @xmath78 system in fig .
2d ( left ) shows that the cross - section of material in the non - wetting region shrinks as the bulk droplet in the wetting region continually spreads . for @xmath78 ,
the radius of the droplet in the wetting regime continues to grow at late time as shown in fig .
3 . between @xmath79 and @xmath80
the gain in surface energy exceeds the cost of increasing the surface area and material is able to be drawn continually from the non - wetting region to the wetting region .
figure 2 ( right ) presents @xmath8 results for the strong wetting case .
compared to the weak wetting case where spreading of a precursor foot was nearly absent for @xmath75 , fig .
2a ( right ) clearly shows precursor spreading .
a separate simulation of the strong wetting system with @xmath81 showed a precursor foot spreads even for this very small feature size .
it is also apparent , especially for smaller @xmath2 , that the stronger interaction in the wetting region causes more spreading in the non - wetting region .
the @xmath78 case for strong wetting is not shown in fig . 2 ; however , similar to the weak wetting case , the minimum width on which the bulk droplet spreads is between @xmath82 and @xmath80 .
this indicates that the bulk wetting transition is related more to @xmath2 than to the strength of interaction in the wetting region @xmath83 . for reference ,
@xmath57 is shown for the strong wetting case in fig .
2d ( right ) .
@xmath27 results for different @xmath2 are shown in fig .
4 . at @xmath84 , which is about 4 times the average end - to - end distance of the chain in the melt ,
the droplet spreads at early time to a radius @xmath85 before reaching its equilibrium size .
this final size is comparable to what is observed for a homogenous surface with a mean surface interaction strength , @xmath86 . for @xmath79 ( fig .
4b ) and @xmath52 ( fig . 1 )
, the bulk radii in the wetting region saturate at lower values , @xmath87 .
these data demonstrate that , for sufficiently small feature size , a mean interaction description is fairly accurate but becomes less so for increasing feature size .
for @xmath88 a precursor foot spreads ahead of the droplet while for @xmath89 no foot is observed .
the width @xmath2 above which the main droplet spreads on the lyophilic strip is larger than we can access with present computational resources . for comparison purposes , fig .
4c presents data for @xmath57 and it is obvious that the main droplet spreads for this limit of pattern size
. the contact radius of the precursor foot increases as @xmath72 for all cases studied . the effective diffusion constant @xmath5 normalized by its value for the homogeneous wetting case is shown in fig .
a result observed in all cases is that @xmath5 increases monotonically with @xmath2 . of note
is an apparent cross - over in observed behavior with decreasing @xmath2 for the @xmath8 strong and weak wetting systems .
data imply that , at large pattern width , foot transport for the stronger wetting system is more affected by the presence of a lyophobic region than it is for the weaker wetting system .
we hypothesize that this results from the very aggressive rate of transport for the foot on the stronger wetting substrate .
the system must supply material to the foot from the bulk of the droplet .
the rate of foot advancement is large enough in the strong wetting case that it becomes limited by transport of material from the non - wetting region to the wetting region , and on to the foot .
this is only true for large pattern width @xmath90 where the foot is comprised of a more significant amount of polymer material .
note that foot transport along the lyophilic strip is observed for @xmath91 in the strong wetting system whereas it is not for the weaker interaction . in our previous work
@xcite studying wetting on homogeneous substrates , we found that @xmath5 increased with the size of the initial droplet as @xmath92 .
while we have not varied the size of the initial droplet in this study , we expect a similar dependence of @xmath5 on the initial size of the droplet .
chemically patterning a surface offers a unique opportunity to control the spreading of liquid nanodroplets . here we have presented simulation results which demonstrate the interplay between the pattern width @xmath2 , the length of the polymer chain @xmath7 , and the interaction strength of the lyophilic strips in determining spreading behavior .
we find that there exists a minimum @xmath2 , which is a function of @xmath7 , below which spreading of the main droplet does not occur .
for @xmath8 , a bulk spreading transition with increasing @xmath2 was observed , while for @xmath27 , the transition occurs for @xmath2 larger than is presently accessible . in a number of cases ,
a precursor foot is observed to spread even when the main droplet does not spread .
the observed effective diffusion constant of this precursor foot decreases monotonically with @xmath2 .
future work will reveal the dependence of observed kinetics on the initial size of the droplet and specify the relation between @xmath7 , @xmath2 , and wetting transition .
sandia is a multiprogram laboratory operated by sandia corporation , a lockheed martin company , for the united states department of energy s national nuclear security administration under contract de - ac04 - 94al85000 . | controlling the spatial distribution of liquid droplets on surfaces via surface energy patterning can be used to control material delivery to specified regions via selective liquid / solid wetting . while studies of the equilibrium shape of liquid droplets on heterogenous substrates exist , much less
is known about the corresponding wetting kinetics .
we make significant progress towards elucidating details of this topic by studying , via large - scale atomistic simulations , liquid nanodroplets spreading on chemically patterned surfaces .
a model is presented for lines of polymer liquid ( droplets ) on substrates consisting of alternating strips of wetting ( equilibrium contact angle @xmath0 ) and non - wetting ( @xmath1 ) material .
droplet spreading is compared for different wavelength @xmath2 of the pattern and strength of surface interaction on the wetting strips . for small @xmath2 , droplets partially spread on both the wetting and non - wetting regions of the substrate to attain a finite contact angle less than @xmath3 . in this case , the extent of spreading depends on the interaction strength in the wetting regions .
a transition is observed such that , for large @xmath2 , the droplet spreads only on the wetting region of the substrate by pulling material from non - wetting regions . in most cases ,
a precursor film spreads on the wetting portion of the substrate at a rate strongly dependent upon interaction strength . |
quantum decoherence presents a fundamental limitation to the realization of practical quantum computing and of other technological devices which actively exploit quantum phenomena . in 2002 , a ground - breaking study established the usefulness of so - called optimal working points ( owps):@xcite these are parameter regimes where the system becomes - to first order - insensitive to fluctuations of external classical fields .
we consider here the effect of owps in a system where decoherence of a central spin system arises from interactions with a fluctuating bath of surrounding spins - a scenario that is of considerable significance in the field of quantum information .
@xcite a promising approach for silicon - based quantum - information processing ( qip ) involves combined electron and nuclear spins of donor atoms in si , which are amenable to high - fidelity manipulation by means of electron paramagnetic resonance ( epr ) and nuclear magnetic resonance ( nmr ) , respectively .
most studies have considered phosphorus ( @xmath2 ) donors in si .
@xcite more recently , several different groups have investigated another group v donor , @xmath3 .
these studies not only showed that bismuth donors have similar properties to si : p , such as long electron spin coherence times @xmath4 of the order of several ms at low temperatures ( @xmath1 k),@xcite but that they also offer new possibilities for qip .
for example , strong optical hyperpolarization was demonstrated,@xcite allowing for efficient initialization of the nuclear spin .
the si : bi spin system has an electron spin @xmath5 and a large nuclear spin @xmath6 as well as an atypically strong hyperfine coupling constant , @xmath7 ghz .
the strong state - mixing occurring for magnetic fields
@xmath8 t where the hyperfine interaction competes with the electronic zeeman energy allows transitions which are forbidden at high magnetic fields,@xcite observed recently in ref . .
in refs . , a set of minima and maxima were found in the @xmath9 parameter space of dipole - allowed transitions at frequencies @xmath10 .
these @xmath11 points were identified as owps : line narrowing and reduced sensitivity to temporal and spatial noise in @xmath12 over a broad region of fields ( closely related to @xmath13 extrema ) were found .
however , to date , their effectiveness for reducing decoherence in the real environment of a spin bath remains untested . in natural si , 4.67% of lattice sites are occupied by the nuclear spin - half @xmath0 isotope , rather than the spinless @xmath14 .
flip - flopping of the @xmath0 spins provides the dominant mechanism of decoherence for both si : p and si : bi systems at low temperatures .
the decay of the donor hahn spin echo for these systems is typically fitted to @xmath15 $ ] , where @xmath16 ms characterizes the nuclear spin diffusion , with @xmath17.@xcite other relaxation processes , such as those arising from donor - donor interactions , are represented by @xmath4 .
since @xmath18 , nuclear spin diffusion remains the main channel of decoherence at low temperatures.@xcite in this work , we investigate the nature of the bi-@xmath0 interaction by means of pulsed electron - nuclear double resonance ( endor).@xcite to obtain an endor spectrum , an epr spin echo is detected as a function of a radio frequency ( rf ) excitation .
when the rf radiation is resonant with an nmr transition , changes are seen in the epr signal if the populations of the relevant energy levels change .
previous endor studies of si : bi used rf frequencies of at least several hundreds of mhz,@xcite and thus could not probe the weak couplings to a spin - bath .
in contrast , here , rf frequencies of a few mhz were used . with this approach
, we have measured the bismuth spin bath superhyperfine ( shf ) couplings and determined their anisotropy .
we also present the results of cluster correlation expansion ( cce ) simulations .
@xcite this model has been used with considerable success to model central spin decoherence in si : p.@xcite in ref . , weak state - mixing in si : bi was investigated by simply allowing for the variation of an effective gyromagnetic ratio . here
we adapt the cce simulations to include , for the first time , the strong state - mixing seen near the owps . a key finding is the demonstration of near - complete suppression of nuclear spin diffusion , even in natural si : this occurs in extremely narrow regions , where @xmath19 is in effect divergent,@xcite in contrast to the broader effect expected from the form of @xmath20.@xcite a successful means of controlling decoherence is to employ isotopically enriched
samples,@xcite which can exhibit long @xmath4 times up to the order of seconds.@xcite thus , the owps represent a potentially complementary technique , effective for both natural si and partially enriched samples .
in addition , our work suggests that the owps may also be effective in suppressing residual effects such as donor - donor interactions which are responsible for shortening @xmath4 times from seconds to the ms timescale.@xcite
we investigate a spin system with total hamiltonian @xmath21 the first term denotes the isotropic hyperfine interaction between the bismuth electronic and nuclear spins . for si : bi , the usual high - field reduction to ising form @xmath22 can not be made at the fields of interest here .
the second term represents the zeeman interaction with the external field , including the donor spins and summed over bath spins , @xmath23 here , @xmath24 is the electronic zeeman frequency ( @xmath25 jt@xmath26 ) while @xmath27 and @xmath28 are the ratios of the nuclear to electronic zeeman frequencies for the donor and @xmath0 spins , respectively .
the spin - bath interaction term @xmath29 represents the shf couplings between the donor and bath spins , in general of tensor form ( for anisotropic couplings ) .
finally , dipolar coupling between each pair of @xmath0 spins is represented by the bath term @xmath30 where @xmath31 denotes the relative position vector of bath spins at lattice sites @xmath32 and @xmath33 .
writing @xmath34 for a pair of spins , the components of the dipolar tensor @xmath35 are given by @xmath36 with @xmath37 na@xmath38 and @xmath39 .
the experimental endor studies reported here served to investigate and characterize the isotropic / anisotropic character of the spin - bath interaction term , namely a set of distinct @xmath40 values - shf couplings - in eq .
( [ eq : interactionh ] ) , corresponding to occupancy of inequivalent lattice sites by @xmath0 impurities .
pulsed endor experiments were performed using the davies endor pulse sequence.@xcite we applied the pulse sequence @xmath41 , where the microwave ( mw ) frequency is chosen to excite one epr transition and the rf is stochastically varied between @xmath42 mhz or @xmath43 mhz to excite all nuclear spin transitions in this region .
we used @xmath44 ns long @xmath45-pulses and a @xmath46 ns long @xmath47-pulse . for optimal signal - to - noise ratio and resolution , we used a @xmath48-pulse of @xmath49 .
pulse delays were set to @xmath50 , @xmath51 , and @xmath52 and a shot repetition time of 1.3 ms was employed to give a good signal - to - noise ratio .
all experiments were carried out at 15 k on an e580 pulsed epr spectrometer ( bruker biospin ) equipped with pulsed endor accessory ( e560d - p ) , a dielectric ring endor resonator ( en4118x - md4 ) , a liquid helium flow cryostat ( oxford cf935 ) , and an rf amplifier ( eni a-500w ) .
we used a donor concentration of @xmath53 @xmath54 and the magnetic field was directed perpendicular to the [ 111 ] plane .
while not offering the higher frequency resolution attainable with continuous - wave endor,@xcite the pulsed endor measurements permit us to adequately constrain and to demonstrate the reliability of the numerical simulations .
in particular , we established that isotropic couplings to the spin bath dominate the decoherence dynamics . while not the focus of this study , a further motivation is to investigate the feasibility of an alternative possibility for qip : to simultaneously manipulate the @xmath0 atoms as spin - half qubits , along with the donors.@xcite measured endor spectra at @xmath55 ghz are presented in fig . [
fig : endor ] , together with a list of shf couplings . for the magnetic field range @xmath56 t in fig .
[ fig : endor ] , there is significant mixing of the high - field si : bi energy eigenstates @xmath57 .
the mixed eigenstates , @xmath58 , correspond to doublets ( at most ) of constant @xmath59 : @xmath60 where @xmath61 and @xmath33 is an integer , @xmath62 . such mixing leads to a complex epr spectrum for bismuth with @xmath13 extrema .
the minima correspond to transitions between states corresponding to adjacent avoided level crossings , of which there are four .
the disparity between the electronic zeeman and hyperfine energy scales and shf energy scales means that the tensor coupling in eq .
( [ eq : interactionh ] ) reduces to simpler form @xmath63 written for coupling to a single @xmath0 at site @xmath64 , where @xmath65 $ ] and @xmath66 with @xmath67 and @xmath68 the isotropic and anisotropic parts of the molecular - frame shf tensor , respectively , and @xmath69 the angle between the external field and the line connecting the bismuth site and site @xmath64 .
nonsecular terms involving @xmath70 and @xmath71 can be neglected.@xcite diagonalization of the resulting 2-dimensional hamiltonian , and setting @xmath72 for a purely isotropic coupling leads to a simple expression for the endor resonance frequency for donor level @xmath73 : @xmath74 the above expression is in perfect agreement with full numerical diagonalization . the couplings in fig .
[ fig : endor ] were extracted from the measured spectra by fitting to the data gaussians of equal width and using eq .
( [ eq : isoendor ] ) .
the same expression and a single set of couplings gave excellent agreement with data at 10 different fields .
in particular , the observed pattern of half a dozen highest frequency @xmath0 resonances moving to a minimum at @xmath75 t , then increasing again , is directly attributable to mixing of the states of the bismuth donor : i.e. , here @xmath76 has a minimum .
ten out of the twelve couplings extracted from data were found to be purely isotropic .
the highest - field spectrum was measured for a range of crystal orientations and only three weak intensity lines showed orientation - dependent frequencies and hence anisotropy .
two are indicated by @xmath77 and @xmath78 in fig .
[ fig : endor ] : the corresponding two couplings with nonzero anisotropy were found to have ( @xmath79 , @xmath80 ) mhz and ( @xmath81 , @xmath82 ) mhz by fitting the more general form of eq .
( [ eq : isoendor ] ) with nonzero @xmath83 .
a previous eseem ( electron spin echo envelope modulation ) study identified a single anisotropic coupling,@xcite attributed to @xmath84-shell ( nearest neighbor ) @xmath0 .
the third line we identify is fitted by coupling constants consistent with the anisotropic coupling in ref . .
for most crystal orientations , this line is masked by much higher intensity lines arising from isotropic couplings .
nuclei at inequivalent lattice sites .
the isotropic superhyperfine couplings were extracted from the spectrum at the highest magnetic field . as the field is varied , the smooth lines follow the resonance positions according to eq .
( [ eq : isoendor ] ) .
solid and dotted lines distinguish between the two peaks observed for each coupling , each corresponding to one of the two donor levels involved in the epr transition .
only the peaks labeled @xmath77 and @xmath78 , in addition to a third pair not resolved here , were found to show anisotropy from performing endor as a function of crystal orientation.,width=324 ] at fields where @xmath76 becomes small ( this occurs close to the @xmath13 minima as shown in refs . ) , eq .
( [ eq : isoendor ] ) tends to the @xmath0 zeeman frequency @xmath85 .
it is straightforward to extend eq .
( [ eq : isoendor ] ) to the anisotropic case and show that the latter statement also holds for anisotropic couplings .
in effect , at these points , the donor might be said to approximately decouple from the bath .
for example , for the epr transition @xmath86 ( labeling the eigenstates @xmath87 in increasing order of energy ) , @xmath88 at @xmath89 mt for level @xmath90 and @xmath91 mt for @xmath92 .
we note that there is however no @xmath12-field value where both the upper and lower levels have @xmath88 : as we see below , this is not actually essential for complete suppression of spin diffusion .
the actual owp is at @xmath93 mt , where @xmath94 .
this is extremely close to where @xmath13 , which occurs when @xmath95
in order to model the full dynamics , we assume that the temperature and donor concentrations are low enough so that phonon - induced decoherence and decoherence due to interactions between donors are negligible . the hahn spin echo decay of a central donor electron coupled to the bismuth nucleus in a bath of @xmath96si was calculated using the cce.@xcite denoting the spin echo intensity by @xmath97 , let @xmath98 be @xmath97 computed including only spins in some set or `` cluster '' of bath spins @xmath99 .
the quantity @xmath100 is defined as @xmath101 where the product is over all clusters . applying this definition to @xmath98 and factoring out @xmath100 , an explicit form for the @xmath100 is obtained in terms of the @xmath102 and the @xmath103 in subsets @xmath104 of @xmath99 , @xmath105 the problem of calculating @xmath97 is reduced into independent components each for a distinct cluster of bath spins . the exact solution to @xmath97
is obtained if the @xmath100 from all clusters are combined using eq .
( [ eq : cce ] ) and the approximation to @xmath106 up to a maximum cluster size of @xmath107 is defined as @xmath108 which involves calculating reduced problems for all clusters each containing at most @xmath107 spins .
the @xmath109 ( 2-cluster ) calculation is a good approximation to @xmath97 when considering only dipolar interactions in the bath affecting the spin echo , as these are at most a few khz and hence perturbative compared to the shf interactions in the mhz range involving the donor electron .
the cce is exact , but not always convergent .
we calculated the 2-cluster ( @xmath110 ) approximation to the cce and obtained convergence for up to @xmath111 .
spin bath decoherence for the @xmath86 epr transition .
( a ) simulated endor and nuclear spin diffusion coherence times @xmath19 as a function of magnetic field @xmath12 , showing collapse of the superhyperfine couplings and a sharp increase in @xmath19 as the field approaches the @xmath93 mt optimal working point ( owp ) .
the dashed line is a fit .
( b ) simulated endor at the @xmath93 mt owp ( upper panel ) and experimental spectrum at 9.755 ghz ( lower panel ) .
( c ) calculated donor hahn spin echo decays from which coherence times in fig .
[ fig : suppression](a ) were extracted .
, width=324 ] the hahn echo sequence evolves the combined system - bath state to time @xmath112 : @xmath113 where @xmath114 is the pauli-@xmath115 gate acting on the donor and @xmath116 denotes the bath identity .
we assume that the time taken for a @xmath117-pulse is small compared to @xmath118 .
the initial state was written as a product of the initial donor and bath states , the former chosen as an equal superposition of states @xmath90 and @xmath92 .
the donor subsystem is recovered after tracing over the bath and the modulus of the normalized off - diagonal element of the donor reduced density matrix is proportional to the intensity of the echo at time @xmath112 .
the reduced problem of the 2-cluster bath was solved for each of the four initial 2-product bath states and the average intensity obtained .
2-clusters were formed by pairing @xmath0 spins separated by up to the 3rd nearest neighbor distance in a diamond cubic lattice of side length 160 .
convergence was obtained as the lattice size and the separation between the two bath nuclei were extended .
it was assumed that @xmath12 was large enough to conserve the total @xmath0 zeeman energy .
thus , the dipolar interaction between the two bath spins took the form of a combination of ising ( @xmath119 ) and flip - flop ( @xmath120 ) terms .
the kohn - luttinger electronic wavefunction was used to calculate the isotropic fermi contact shf strength with an ionization energy of 0.069 ev for the bismuth donor .
calculated couplings were of the same order as those obtained from data .
the data suggest that isotropic couplings predominate ; hence anisotropic couplings were neglected and the simulations were largely insensitive to orientation .
finally , we obtained the average @xmath109 over 100 spatial configurations of @xmath0 occupying 4.67% of lattice sites .
the resulting decay curves were fitted to @xmath15 $ ] , obtaining @xmath121 and values of @xmath17 .
the results of our cce simulations are presented in fig . [
fig : suppression ] .
figure [ fig : suppression](a ) shows the behavior around the @xmath93 mt owp , associated with the @xmath86 epr line .
the calculated coherence time @xmath19 ( orange dashed line ) is superposed on a color map showing the shf spectrum : the latter shows endor spectra simulated as a function of @xmath12 , using eq .
( [ eq : isoendor ] ) and centered about the @xmath0 nuclear zeeman frequency .
strikingly , as @xmath12 approaches the owp , the `` comb '' of shf lines narrows to little more than the width of a single line .
this suggests a drastic reduction in the value of the shf couplings , indicating that the bismuth has become largely decoupled from the @xmath0 spin bath .
the collapse in the shf couplings is illustrated further in fig .
[ fig : suppression](b ) .
the lower panel shows the measured spectrum at 9.755 ghz . using our experimentally determined shf couplings , the corresponding spectrum at the owp
is shown in the upper panel of fig .
[ fig : suppression](b ) , demonstrating clearly the narrowing of the spectrum [ corresponding to the same parameters as fig .
[ fig : suppression](a ) but at the precise field value of the owp ] .
the behavior of @xmath19 is also quite striking and unexpected : the coherence time predicted by cce simulations increases asymptotically at the owp by several orders of magnitude .
away from the owp , the results agree well with experimentally measured values of approximately 0.7 ms .
@xcite in ref .
, in a regime of weak state - mixing , simulations using an effective gyromagnetic ratio indicated that @xmath19 was slightly reduced ( by about 5% ) in a regime corresponding to lower @xmath20 .
the present study , on the other hand ( which in contrast to ref . employed a full treatment of the quantum eigenstate mixing ) , shows rather an effect very sharply peaked about the owp : nuclear spin diffusion is predicted to be largely suppressed , but over an extremely narrow magnetic field range .
figure [ fig : suppression](c ) shows a sample of cce spin echo decays from which @xmath19 times were extracted , and also serves to further illustrate the sharp increase in @xmath19 .
similar suppression is present for other owps in si : bi .
there are @xmath13 minima for the @xmath122 , @xmath123 , @xmath124 , @xmath86 , and @xmath125 transitions in the frequency range @xmath126 ghz and two maxima for @xmath127 and @xmath128 close to @xmath129 ghz .
the decoupling from the spin bath is also expected to lead to suppression of decoherence arising from the interaction with a bath of donors .
in conclusion , we present measurements of the shf couplings between a bismuth donor and a bath of @xmath0 impurities which suggest that isotropic couplings dominate .
we further demonstrate the suppression of couplings at owps .
finally , the spin echo decay of the donor is calculated as a many - body problem and sharp divergence of the spin diffusion time is found at an owp .
our study motivates experimental epr studies in the range @xmath130 ghz corresponding to the regions of suppressed decoherence .
we acknowledge bernard pajot for the si : bi crystal used here and thank the epsrc for financial support through the compasss grant . m.b.a.k . is supported by a wellcome trust studentship .
g.w.m is supported by the royal commission for the exhibition of 1851 .
sandia national laboratories is a multiprogram laboratory operated by sandia corporation , a wholly owned subsidiary of lockheed martin corporation , for the u.s .
department of energy s national nuclear security administration under contract no .
de - ac04 - 94al85000 .
d. vion , a. aassime , a. cottet , p. joyez , h. pothier , c. urbina , d. esteve , and m. h. devoret , science * 296 * , 886 ( 2002 ) .
r. de sousa and s. das sarma , phys .
b * 67 * , 033301 ( 2003 ) .
r. de sousa and s. das sarma , phys .
b * 68 * , 115322 ( 2003 ) .
e. abe , k. m. itoh , j. isoya , and s. yamasaki , phys .
b * 70 * , 033204 ( 2004 ) .
s. takahashi , r. hanson , j. van tol , m. s. sherwin , and d. d. awschalom , phys .
* 101 * , 047601 ( 2008 ) .
e. abe , a. m. tyryshkin , s. tojo , j. j. l. morton , w. m. witzel , a. fujimoto , j. w. ager , e. e. haller , j. isoya , s. a. lyon , m. l. w. thewalt , and k. m. itoh , phys .
b * 82 * , 121201(r ) ( 2010 ) .
w. m. witzel , m. s. carroll , a. morello , l. cywinski , and s. das sarma , phys .
lett . * 105 * , 187602 ( 2010 ) .
n. bar - gill , l. m. pham , c. belthangady , d. le sage , p. cappellaro , j. r. maze , m. d. lukin , a. yacoby , and r. walsworth , nat .
commun . * 3 * , 858 ( 2012 ) .
n. zhao , s. w. ho , and r. b. liu , phys .
b * 85 * , 115303 ( 2012 ) . g. de lange , t. van der sar , m. blok , z .- h . wang , v. dobrovitski , and r. hanson , sci
. rep . * 2 * , 382 ( 2012 ) .
b. e. kane , nature ( london ) * 393 * , 133 ( 1998 ) .
s. r. schofield , n. j. curson , m. y. simmons , f. j. rue , t. hallam , l. oberbeck , and r. g. clark , phys .
* 91 * , 136104 ( 2003 ) .
a. m. stoneham , a. j. fisher , and p. t. greenland , j. phys . : condens .
matter * 15 * , l447 ( 2003 ) .
a. m. tyryshkin , s. a. lyon , a. v. astashkin , and a. m. raitsimring , phys .
b * 68 * , 193207 ( 2003 ) .
k .- m . c. fu , t. d. ladd , c. santori , and y. yamamoto , phys .
b * 69 * , 125306 ( 2004 ) . g. w. morley , d. r. mccamey , h. a. seipel , l .- c .
brunel , j. van tol , and c. boehme , phys .
* 101 * , 207602 ( 2008 ) .
a. morello , j. j. pla , f. a. zwanenburg , k. w. chan , k. y. tan , h. huebl , m. mttnen , c. d. nugroho , c. yang , j. a. van donkelaar , a. d. c. alves , d. n. jamieson , c. c. escott , l. c. l. hollenberg , r. g. clark , and a. s. dzurak , nature ( london ) * 467 * , 687 ( 2010 ) .
t. greenland , s. a. lynch , a. f. g. van der meer , b. n. murdin , c. r. pidgeon , b. redlich , n. q. vinh , and g. aeppli , nature ( london ) * 465 * , 1057 ( 2010 ) .
s. simmons , r. m. brown , h. riemann , n. v. abrosimov , p. becker , h .- j .
pohl , m. l. w. thewalt , k. m. itoh , and j. j. l. morton , nature ( london ) * 470 * , 69 ( 2011 ) .
m. steger , t. sekiguchi , a. yang , k. saeedi , m. e. hayden , m. l. w. thewalt , k. m. itoh , h. riemann , n. v. abrosimov , p. becker , and h .- j .
pohl , j. appl .
phys . * 109 * , 102411 ( 2011 ) .
l. dreher , f. hoehne , m. stutzmann , and m. s. brandt , phys .
lett . * 108 * , 027602 ( 2012 ) .
m. fuechsle , j. a. miwa , s. mahapatra , h. ryu , s. lee , o. warschkow , l. c. l. hollenberg , g. klimeck , and m. y. simmons , nat . nanotechnol .
* 7 * , 242 ( 2012 ) .
g. w. morley , m. warner , a. m. stoneham , p. t. greenland , j. van tol , c. w. m. kay , and g. aeppli , nature mater .
* 9 * , 725 - 729 ( 2010 ) .
r. e. george , w. witzel , h. riemann , n. v. abrosimov , n. ntzel , m. l. w. thewalt , and j. j. l. morton , phys .
* 105 * , 067601 ( 2010 ) .
t. sekiguchi , m. steger , k. saeedi , m. l. w. thewalt , h. riemann , n. v. abrosimov , and n. ntzel , phys .
lett . * 104 * , 137402 ( 2010 ) .
m. h. mohammady , g. w. morley , and t. s. monteiro , phys .
lett . * 105 * , 067602 ( 2010 ) .
m. h. mohammady , g. w. morley , a. nazir , and t. s. monteiro , phys .
b * 85 * , 094404 ( 2012 ) . g. w. morley , p. lueders , m. h. mohammady , s. j. balian , g. aeppli , c. w. m. kay , w. m. witzel , g. jeschke , and t. s. monteiro , arxiv:1109.4269 .
w. m. witzel and s. das sarma , phys .
b * 74 * , 035322 ( 2006 ) .
a. m. tyryshkin , j. j. l. morton , s. c. benjamin , a. ardavan , g. a. d. briggs , j. w. ager , and s. a. lyon , j. phys . : condens .
matter * 18 * , s783 ( 2006 ) .
a. m. tyryshkin , s. tojo , j. j. l. morton , h. riemann , n. v. abrosimov , p. becker , h .- j .
pohl , t. schenkel , m. l. w. thewalt , k. m. itoh , and s. a. lyon , nature mater . * 11 * , 143 ( 2012 ) .
a. schweiger and g. jeschke , _ principles of pulse paramagnetic resonance _
( oxford university press , oxford , 2001 ) .
w. yang and r. b. liu , phys .
b * 78 * , 085315 ( 2008 ) ; * 78 * , 129901(e ) ( 2008 ) ; * 79 * , 115320 ( 2009 ) . in the sense that decay of the spin echo is no longer discernible in our converged cluster calculations . m. steger , k. saeedi , m. l. w. thewalt , j. j. l. morton , h. riemann , n. v. abrosimov , p. becker , and h .- j .
pohl , science * 336 * , 1280 ( 2012 ) . c. d. weis , c. c. lo , v. lang , a. m. tyryshkin , r. e. george , k. m. yu , j. bokor , s. a. lyon , j. j. l. morton , and t. schenkel , arxiv:1202.1560 . g. wolfowicz , s. simmons , a. m. tyryshkin , r. e. george , h. riemann , n. v. abrosimov , p. becker , h .- j .
pohl , s. a. lyon , m. l. w. thewalt , and j. j. l. morton , arxiv:1207.3776 .
although mims endor offers higher selectivity to small shf couplings , it would suffer from blind spots .
g. feher , phys . rev . * 114 * , 1219 ( 1959 ) . e. b. hale and r. l. mieher , phys . rev . *
184 * , 739 ( 1969 ) .
w. akhtar , v. filidou , t. sekiguchi , e. kawakami , t. itahashi , l. vlasenko , j. j. l. morton , and k. m. itoh , phys .
lett . * 108 * , 097601 ( 2012 ) .
m. belli , m. fanciulli , and n. v. abrosimov , phys .
b * 83 * , 235204 ( 2011 ) . | we present pulsed electron - nuclear double resonance ( endor ) experiments which enable us to characterize the coupling between bismuth donor spin qubits in si and the surrounding spin bath of @xmath0 impurities which provides the dominant decoherence mechanism ( nuclear spin diffusion ) at low temperatures ( @xmath1 k ) .
decoupling from the spin bath is predicted and cluster correlation expansion simulations show near - complete suppression of spin diffusion , at optimal working points .
the suppression takes the form of sharply peaked divergences of the spin diffusion coherence time , in contrast with previously identified broader regions of insensitivity to classical fluctuations .
endor data suggest that anisotropic contributions are comparatively weak , so the form of the divergences is largely independent of crystal orientation . |
the giant magnetoresistance effect in magnetic metal multilayers@xcite occurs because of the coupling of external magnetic fields to band electron spins through their collective spin polarization .
the utility of this effect for information storage and field sensing devices has increased interest in exploring related spin - dependent transport properties in both ferromagnetic and paramagnetic semiconductors . at the same time , progress in the homo and hetero epitaxy of magnetically doped semiconductors is creating new possibilities for engineered material geometries in which new spin - dependent transport effects are likely to occur .
@xcite large band electron spin polarizations can occur@xcite in diluted magnetic semiconductors ( dmss),@xcite for example in ii - vi compounds with mn substituted on the group ii site . among the striking phenomena already demonstrated in mn
doped ii - vi quantum structures are magnetically tunable quantum well barriers and interwell couplings , @xcite spin - dependent dynamics of polarized excitons in spin superlattices , @xcite optically - probed spin coherence , @xcite and the injection of highly polarized spin currents into gaas / algaas light emitting diodes .
@xcite in addition , @xmath0-doping of wide - gap ii - vi magnetic semiconductor quantum wells ( zn@xmath1cd@xmath2mn@xmath3se ) has been achieved,@xcite yielding two - dimensional electron gas ( 2deg ) systems that are exchange coupled to magnetic ions . @xcite the exchange interaction , @xmath4 , between @xmath5-electrons in the conduction band and mn@xmath6 @xmath7 local moments , results in band spin splittings larger than @xmath8 , the landau level splitting . the spin - splittings can reach values as high as @xmath9 mev.@xcite in fact , complete spin polarization can be achieved in quantum wells at relatively low magnetic fields ( @xmath10 t ) .
@xcite although the study of electronic transport properties in dms heterostructure systems is still in its initial stages , interesting predictions have already been made .
@xcite the present study is motivated in part by the recent growth of a modulation - doped znse/(zn , cd , mn)se multiple quantum well ( mqw ) system.@xcite in non - magnetic mqw systems , growth direction transport phenomenology is enriched by an interplay between charge accumulation and resonant inter - well tunneling effects that results in the formation of electric field domains . in this paper
we report on a theory of the influence of exchange coupling with mn ion spins on electric field domain formation and on the sensitivity of this influence to spin - relaxation rates within the quantum wells .
the formation of electric field domains is the hallmark of dc - biased transport in weakly - coupled semiconductor superlattices .
spontaneously generated inhomogeneities in the spatial distribution of voltage drops , were proposed as an explanation for overall conductance ( @xmath11 ) oscillations , @xcite discovered in pioneering growth direction mqw transport studies .
this early hypothesis was later confirmed by direct photoluminiscence measurements .
@xcite subsequently , highly doped gaas / algaas superlattices that present sawtooth - like current voltage ( @xmath12@xmath13 ) characteristics in the negative differential conductance ( ndc ) region were studied in detail . along branches of the sawtooth two approximately constant electric field regions
develop in the sample , separated by a layer of accumulated electrons . in the following ,
we follow common usage in referring to these layers with higher 2d electron density as _ monopoles_.
this non - equilibrium configuration enables resonant tunneling between ground and excited subbands in the high field region , minimizing the total resistance of the superlattice .
increases in external voltage in this regime lead to sharp decreases in current , followed by discrete jumps of the monopole region from a well to its upstream neighbor , extending the high field domain over an additional period and increasing the current .
past work has studied the dependence of domain formation and evolution on magnetic fields applied along the growth direction and on far - infrared radiation wavelength . in the former case , the formation of landau levels and scattering between them introduces a new voltage scale for domain formation .
@xcite in the latter case , photonic sidebands sustain the formation of the electric field domains .
@xcite the inclusion of the electronic spin in the study of perpendicular transport in mqws can also be expected to alter electric field domain formation physics , and exchange coupling to mn spins should make it possible to tune these effects with relatively weak external magnetic fields .
this is the possibility that we explore at greater length in this paper . in all these cases
the electron - electron interaction , although small in the ground state in comparison with typical energies of the system , can not be neglected since it is the poisson equation relating charge accumulation to field variations that is at the heart of field - domain formation .
for example it permits the experimentally observed@xcite multistability of distinct stationary physical states at a fixed bias voltage . nonetheless , the mean - field hartree approximation is sufficient to capture this physics in typical samples .
the non - linearity of the current versus voltage relationship between neighboring quantum wells , coupled with the non - locality of electron - electron interaction effects leads to transport equations that can be solved only numerically , and also to results that are sometimes difficult to interpret . in this paper
we deal with the formation of electric field domains in ii - vi mqw systems with one or more ( ii , mn)vi quantum wells . the main ingredients of our self - consistent theoretical model are : ( i ) a theory for the tunneling current between two spin - polarized 2degs ; ( ii ) a continuity equation that accounts for relaxation of non - equilibrium spin populations ; ( iii ) a relationship between the up and down chemical potentials and their densities ; ( iv ) the application of simple hartree mean field theory to account for the coulomb interaction ; and ( v ) a mean - field theory for the interaction between 2deg electrons and mn spins whose average polarization is very sensitive to external magnetic fields at low temperature .
we shall demonstrate that new features appear in the @xmath12@xmath13 curve that depend on temperature and mn spin concentration , and explain why spin bottlenecks turn out to have a strong influence in the instability regions .
the paper is organized as follows . in section [ sec - model ] our theoretical model is thoroughly explained .
[ sec - tech ] is devoted to a discussion of technical details important for the numerical integration of the rate equations that describe the time dependent charge , spin , and current distributions . in sec .
[ sec - res ] we give numerical results and discuss their interpretation .
finally , sec .
[ sec - con ] contains our conclusions .
following sucessful@xcite early work on bulk systems by kossut,@xcite bastard,@xcite and gaj,@xcite we account for the presence of mn ions in dms quantum wells , by combining a phenomenological exchange model with a virtual crystal approximation and mean - field - theory . the lattice parameters and the band hamiltonian parameters of a ii - vi heterostructure
are assummed to change smoothly as mn@xmath6 spins are introduced in the system and a @xmath7 quantum spin is assumed to be added to the low energy degrees of freedom for each mn spin .
the band electron system and the local moments are coupled by a ferromagnetic exchange interaction that favors parallel alignment of the local moment and band electron spins .
the total hamiltonian of the system is : @xmath14 the first four terms in the right - hand side of the equation describe a conventional superlattice system with many weakly coupled quantum wells : 1 .
@xmath15 is the hamiltonian for independent electrons in @xmath16 _ isolated _ quantum wells .
its energy spectrum is purely single - particle - like , and the quasiparticle spectrum is that of an isolated quantum well 2deg @xmath17 , where @xmath18 is the wave vector parallel to the mqw heterointerfaces , @xmath19 , @xmath20 is the effective mass , and @xmath21 is the quantum well subband index .
we will take @xmath18 to be a continuous index , disregarding landau - level formation in the weak magnetic fields we will consider .
2 . @xmath22 contains the tunneling amplitudes that couple quasiparticles in different quantum wells . in weakly - coupled superlattices
it is a good approximation to treat this term by leading order perturbation theory as we discuss in sec .
[ sec - tun ] .
@xmath23 contains the scattering terms within a quantum well that allow a non - equilibrium quasiparticle to relax its excess energy , but does not contain terms that permit the quasiparticle system to bring its spin - subsystems into equilibrium .
( these terms are absorbed in @xmath24 . ) because it is difficult to describe these scattering processes accurately , or even to know what they are in particular systems , we will use a phenomenological relaxation time approximation . the time scale associated with these processes
is typically rather short ( @xmath25 ps ) .
@xcite 4 .
@xmath26 is the electron - electron interaction in the conduction band for which we will use a hartree mean - field approximation .
( see sec .
[ sec - elec ] ) .
the remaining terms in eq .
( [ eq - ham ] ) describe _ spin_-related physics : 1 .
@xmath27 is the exchange interaction between @xmath5 conduction band electrons and mn local moments , an interaction that turns out to be ferromagnetic in ii - vi mqws . when the mean - field and virtual crystal approximations are employed , the effect of this coupling is to make the subband energies spin - dependent in those quantum wells that contain mn ions : @xmath28 .
2 . @xmath29 represents the antiferromagnetic super exchange interaction between mn spins on neighboring lattice sites that has been found to be important in modelling bulk dms systems.@xcite since our intention here is to address the qualitative physics of field - domains in dms mqw systems , we neglect @xmath29 .
we do expect , however , that these interactions will be important for detailed modelling of specific experimental systems .
3 . @xmath24 contains the microscopic processes that allow equilibrium to be established between spin subsystems within a quantum well .
the fact that spin relaxation can be quite slow in the conduction band@xcite is one of the motivations for this work .
relaxation times in excess of @xmath30 ns have been established experimentally @xcite in ii - vi semiconductor qws without mn . in ii - vi dms qws these times are reduced to tens of picoseconds ( but still larger than @xmath31).@xcite we discuss the role of these terms at greater length in [ sec - spin ] .
the standard theory of tunneling relates the electric current between weakly - coupled subsystems to tunneling matrix elements and subsystem spectral functions .
@xcite in our case we will apply this theory to describe the current flowing between one quantum well and its neighbor . since elastic and inelastic scattering times in the quantum wells are shorter than any other time scale of the problem , we can follow the standard lines of tunneling theory and assume that the electrons in each well are in quasi equilibrium between succesive tunneling events and that their temperature is that of the lattice .
we ignore interwell spin - flip processes , so that currents are carried between wells by the two spin subsystems _ in parallel_. accordingly , the current per spin from the @xmath32th well to the ( @xmath33)st well is given by the following general expression : @xmath34 \ , , \label{eq - curr1}\end{aligned}\ ] ] where @xmath35 is the conduction electron spin index , @xmath36 is the 2d density of states per spin , @xmath37 is the transmission coefficient between particular wavevector states in the two quantum wells , @xmath38 is the voltage drop across the @xmath32th barrier and @xmath39 $ ] is the fermi factor .
@xmath40 denotes the chemical potential in well @xmath32 measured from the bottom of well @xmath32 .
a commonly used lorentzian - shape function is chosen to represent the influence of disorder on quasiparticles in the @xmath21th subband within the @xmath32th quantum well : @xmath41 this form for the spectral function results from neglecting the real part of the disorder self energy , which introduces an unimportant rigid shift of the quasiparticle energies , and the energy dependence of its imaginary part .
@xmath42 is treated as a phenomenological parameter whose value may vary substantially from sample to sample and is to be taken from experiment .
( @xmath43 mev ) . in weakly - coupled superlattices
the broadening due to scattering is much larger than the miniband width so that tunneling between quantum wells is _ sequential _ rather than band like . in this regime
an electron undergoes many scattering events in one well before tunneling to the next well .
because of the epitaxial nature of the samples in question , we assume that the tunneling process conserves parallel momentum , _ i.e. _ , effects such as interface roughness are not taken into account .
this approximation is made for the sake of definiteness and does not influence the qualitative conclusions we will reach . in addition , the typical electronic densities in particular quantum wells are assumed to be @xmath44@xmath45 so that only the lowest subband is appreciably populated in the quasi - equilibrium state of a quantum well .
therefore , when current flows dominantly by a transition from the ground state of well @xmath32 to the first excited state in well @xmath33 , rapid relaxation to the lowest subband ( via , e.g. , emission of a lo phonon ) is assumed . this simplifications lead us to an analytical expresion for eq .
( [ eq - curr1 ] ) at @xmath46 k : @xmath47 the sum in eq .
( [ eq - curr2 ] ) is extended over all subands in the @xmath48st well .
the transmission coefficients from the ground subband to the @xmath21th subband of the neighbouring well , @xmath49 , are calculated by means of the transfer hamiltonian method.@xcite since this approach involves only orbital degrees of freedom , @xmath49 does not depend explicitly on @xmath50 .
we are therefore implicitly restricting ourselves to the case when the spin splittings are not so large as to mix different subbands .
the function @xmath51 expresses the width of the energy `` window '' available for tunneling . for tunneling between the lowest subbands : @xmath52 where @xmath53 is the first subband energy in the @xmath32th - well and so on .
the expressions for the energy window for tunneling to higher subbands are similar and differ only through the absence of pauli blocking effects in the target layer .
the exchange interaction between the conduction band electrons and the mn@xmath6 ions produces a giant spin splitting of the 2deg even in the presence of a small magnetic field .
the bottom of the band for each spin rigidly shifts accordingly : @xmath54 where @xmath55 is the @xmath21th miniband center , @xmath56 for @xmath57(@xmath58 ) , @xmath4 is the exchange integral , @xmath59 is the density of mn@xmath60 ions with spin equals to @xmath7 , @xmath61 is the brillouin function , and @xmath62 is an effective temperature which can , in general , include a correction due to antiferromagnetic interactions between neighboring mn ions.@xcite we have assumed here that the magnetic field ( @xmath63 ) orientation is such that the quasiparticle energy is lowered for _ up _ spins . note that our virtual crystal approximation for the mn ions implies that the mean - field experienced by band electrons of spin @xmath50 is spatially uniform . in this paper
we take the field experienced by the local moments in eq .
( [ eq - delta ] ) to be the external magnetic field .
in fact , the mean - field approximation we employ can be extended@xcite to include the contribution of spin - polarized band electrons to the total effective magnetic field experienced by the local moments .
when this is done , ferromagnetism results . that extention of the models and approximations employed here does in fact appear to account for the main features of the carrier induced ferromagnetism that occurs@xcite in doped diluted ferromagnetic semiconductors .
the transition temperatures in these systems can be substantial@xcite for p - type systems and for higher carrier densities .
in the case of relatively low - density n - type systems , the ferromagnetic transition temperatures will be low and ferromagnetism does not necessarily occur , when direct interactions between the mn local moments are included in the theory .
experimentally , most n - type dms 2degs show no evidence for ferromagnetism . in this paper
we will assume that the contribution of the spin - polarized conduction band system to the effective field experienced by the mn local moments is negligible .
if ferromagnetism did occur , the non - linear field - domain transport physics we discuss in this paper would be further enriched . in modelling spin - relaxation within a quantum
well we start by nelecting the transport currents @xmath64 .
@xcite for instantaneous spin - splittings smaller than the fermi energy we use the following phenomenological spin - relaxation rate equation within each quantum well : @xmath65 where @xmath66 is the spin opposite to @xmath50 and @xmath67 is the spin - scattering time .
( we neglect the dependence of @xmath67 on @xmath68 , which may be important in certain samples@xcite ) .
this equation implicitly assumes the linear non - interacting 2deg relationship between density and chemical potential , so that correlation effects are not taken into account:@xcite @xmath69 in the absence of a driving bias voltage , the solution of eq .
( [ eq - rate1 ] ) at @xmath70 is @xmath71 and @xmath72 . here
@xmath73 is the total density of carriers in the @xmath32th well .
we can see from this asymptote that eq .
( [ eq - rate1 ] ) is valid only for spin splittings smaller than the chemical potential . for @xmath68 greater than the chemical potential , eq . ( [ eq - rate1 ] )
must be modified using eq .
( [ eq - ni ] ) : [ eq - rate2 ] @xmath74 for large enough @xmath68 , eq . ( [ eq - rate2 ] ) leads to an equilibrium state with full spin polarization .
( see fig .
[ fig1 ] ) . adding transport currents to these considerations leads to the following discrete continuity equations for the spin population in each quantum well : @xmath75 for the case @xmath76 . otherwise , eq . ( [ eq - rate2 ] )
must replace the second term on the right - hand side of eq .
( [ eq - rate3 ] ) . in large area heterostructures , the coulomb interaction is usually included in a hartree mean - field approximation .
the poisson equation relates the electrostatic potential drop across mqw barriers , @xmath77 , to the charge distribution among the quantum wells : @xmath78 where @xmath79 is the superlattice period , @xmath80 is the sample average permittivity , and @xmath81 denotes the doping density within the wells .
( experimentally , doping is usually accomplished by placing a zncl@xmath82 layer in the barrier layers ; this difference in electrostatics compared to our model has no important consequences . )
@xcite by inspection of eqs .
( [ eq - rate3 ] ) and ( [ eq - poi ] ) it is obvious that a set of boundary conditions must be provided for @xmath83 and @xmath84 . within our model
these layers play the role of source and drain , respectively . a simple way to represent source and
drain @xcite is to fix the density in both layers at high values .
we take @xmath85 where @xmath86 is an adjustable parameter .
more sophisticated models @xcite have proven that a proper description of the contacts can have a strong effect on the selection of the the transport equation solution when multistability occurs , especially when dynamical solutions are allowed .
@xcite however , we choose not to delve into these effects in detail here since our main interest is on spin effects . for the sake of definiteness ,
the contacts are taken to be unpolarized throughout our calculations ; including spin - polarized injection in our theory would be straightforward and indeed this may be a very interesting avenue to explore in future experimental and theoretical studies . with this representation of the source and drain , fixing the overall bias voltage , @xmath87 closes the set of equations .
our model contains @xmath88 unknown functions of time ( @xmath89 chemical potentials , @xmath90 , @xmath89 electronic densities , @xmath91 , and @xmath92 voltage drops , @xmath77 ) .
these unknown functions are determined by the @xmath89 constitutive equalities ( eq . ( [ eq - ni ] ) ) , @xmath89 rate equations ( eq . ( [ eq - rate3 ] ) ) , @xmath16 poisson relations ( eq . ( [ eq - poi ] ) ) , and the total bias condition ( eq . ( [ eq - bias ] ) ) .
thus , giving a physically sensible initial profile , the system of algebraic - differential equations can be integrated to yield a definite solution .
standard numerical methods are employed in solving eq .
( [ eq - rate3 ] ) , being careful to use the appropriate spin - relaxation equation ( either eq . ( [ eq - rate1 ] ) or eq . ( [ eq - rate2 ] ) ) at each time step .
the total current density @xmath93 traversing the sample at time @xmath94 is determined by the following procedure .
differentiate eq .
( [ eq - poi ] ) with respect to time and substitute the result into the sum of eq .
( [ eq - rate3 ] ) over spin indices in order to eliminate densities and chemical potentials .
this leads to the following current which has the same value when evaluated for any well index , @xmath32 : @xmath95 the first term of the right - hand member of the previous equation is the displacement current whereas the second term is the tunneling current , @xmath96 . for static steady state solutions discussed in this paper
only the latter term is finite .
we focus on znse / zn@xmath1cd@xmath2mn@xmath3se dms mqws .
a value of @xmath97 has been chosen to be consistent with barriers ( @xmath98 mev ) sufficiently high to capture more than one subband in the quantum wells .
for definiteness , we focus on the case where mn has been incorporated only in the _ central _ well of the superlattice . in experimental samples the value of @xmath99 can be varied over a wide range .
we expect that the field - tuned field - domain effects we discuss will be strongest at moderate mn densities , large enough to give rise to sizable spin - splittings but not so large as to increase the spin scattering rate excessively .
the experimental samples reported on in ref . possess znse barriers too thick ( @xmath100 nm ) for perpedicular transport . under those conditions , the coefficients @xmath101 are so small that electron tunneling would likely occur via impurity channels in the barriers or through different symmetry points in the host semiconductor band structure , violating the assumptions of our theory . for these model calculations
we choose a smaller barrier width : @xmath102 nm . the remaining phenomenological parameters have been fixed on the basis of available experimental data : @xmath103 nm , @xmath104 , @xmath105 @xmath45 , @xmath106 ps , and @xmath107 .
@xmath108 is the well width and @xmath109 specifies the source and drain densities ( see eq .
( [ eq - bc ] ) ) . in sec .
[ sec - lin ] we study self - consistent steady - state solutions for low voltages .
this will help us understand the more complicated behaviors that result from instabilities in the ndc voltage regime .
( see sec .
[ sec - ndc ] below . ) the behavior of the spin - dependent current density in eq .
( [ eq - curr2 ] ) depends on many variables . in figs .
[ fig2 ] and [ fig3 ] we have plotted @xmath64 , for the case where @xmath110 and @xmath111 have been set to their equilibrium values ( @xmath110=@xmath111=@xmath112 ) .
thus the spin - dependent chemical potential , measured from the bottom of the well , will depend on the value of @xmath68 ( see fig . [ fig1 ] ) .
[ fig2](a ) plots the up - spin current from a _ _ n__on - magnetic well to a _ _
m__agnetic one , @xmath113 , _
i.e. _ , net electron flow from a magnetic well to an non - magnetic one . for @xmath114 mev and low voltages
the behavior is ohmic , as expected . a larger bias results in the alignment of the ground states of both wells within @xmath42 , giving rise to a first maximum in the current .
after this bias is exceeded , eq .
( [ eq - curr2 ] ) implies the appearance of a ndc region due to subband mismatch and of a second peak at higher biases when the first subband is aligned with the second subband of the next well . increasing @xmath68 decreases the value of the first peak since the up - spin density in the magnetic well increases and fewer states are available for tunneling .
the peak corresponding to tunneling from first@xmath115second subband is not affected in magnitude but its position is shifted to lower bias voltages because the bottom of the up - spin subband goes down . for a given value of the spin splitting ( @xmath116 mev ) , the magnetic well
is fully polarized so that the first peak magnitude can no longer vary its value , a displacement due to subband lowering is observed instead . in fig .
[ fig2](b ) the up - spin current from a magnetic well to a non - magnetic one is shown . here
the first peak becomes larger as @xmath110 grows since more electrons take part in the tunneling .
this increase ceases once the fully polarized situation is achieved .
in addition larger voltages are needed to align the energy levels as @xmath68 increases and the peak moves to higher voltages .
field - domain physics is at heart controlled by the interplay of electrostatics and the layer - to - layer non - linear @xmath12@xmath13 relationships .
these examples illustrate that the layer - to - layer non - linear @xmath12@xmath13 s can be altered by @xmath68 ( and hence an external magnetic field ) when one of the layers contains mn local moments .
it follows that the field - domain structure must be @xmath68 and field dependent as we show below .
the corresponding down - spin current is plotted in fig .
the considerations explained above for the up spin case may be invoked to understand this figure . for @xmath117 mev
( see fig .
[ fig3](b ) ) the down - spin current flowing from the magnetic to the non - magnetic well ( @xmath118 ) vanishes for any value of the applied bias since in the fully polarized regime , no down - spin carriers are present in the magnetic well .
a numerical calculation for @xmath16=3 gives rise to the mqw current voltage characteristics shown in fig .
@xmath68 has been set to 3 mev . at low voltages
the behavior is linear , up to the peak which marks alignment to within @xmath119 of the first subbands of all wells .
then , after entering the ndc region , an increase of the current occurs because an electric field domain has formed .
the underlying mechanism can be understood in the following terms .
as the voltage is increased , the system prefers to mantain regions of low and high electric fields where intra - subband and inter - subband resonant tunneling can occur , rather than maintaining a constant field .
the high field regime forms in the downstream side of the sample . in this way
, the superlattice minimizes the total resistance and current increases . in order to build the inhomogenous field , according to poisson equation ( [ eq - poi ] ) an excess electron density , or a monopole , must accumulate in the well at the boundary between low and high field regions .
as the voltage is further increased , configurations with particular monopole locations sucessively become unstable and the monopole moves by one quantum well to increase the width of the high field region .
monopole motion gives rise to an abrupt decrease of the current , which is followed by a gradual increase as the voltage increases further .
these successive monopole position jumps lead ultimately to the sawtooth - like @xmath12@xmath13 curve shape of fig .
the inset in this figure shows the corresponding @xmath114 mev @xmath12@xmath13 curve for comparison .
the most obvious change is the appearance of a new branch that develops close to the first@xmath115second subband transition ( see below ) .
the somewhat complicated behavior of the bias voltage dependence of the steady state spin polarization is depicted in fig .
[ fig5 ] for @xmath120 and @xmath121 mev .
notice that in fig .
[ fig5](a ) a non - zero steady state spin polarization is induced in the non - magnetic wells by the spin - dependent transport currents .
we have assumed here that the spin - relaxation time constant @xmath67 is the same in all quantum wells ( 10 ps ) . in practice
these times are likely to be considerably larger in the non - magnetic wells , and the induced spin - polarization in these wells will have an even larger importance . for low voltages , the inset of fig .
[ fig5](b ) shows a reduction of the magnetic well spin polarization .
this effect can be understood by realizing that the minority spin current from a non - magnetic well to a magnetic one , @xmath122 , grows when the system is driven out of equilibrium by a small bias voltage .
this occurs also in @xmath113 but at a smaller rate because of the spin subband displacement in the magnetic well ( compare figs .
[ fig2](a ) and [ fig3](a ) ) . therefore , at low voltages the magnetic well polarization must decrease .
correspondingly , the upstream - well polarization must increase ( see the inset of fig .
[ fig5](a ) ) since down spins leak out of it .
the enlargement of the downstream - well spin polarization has the same origin .
( see figs .
[ fig2](b ) and [ fig3](b ) ) . as explained above , once the bias voltage reaches the ndc regime , electric field domains form in the sample . as we now discuss , their formation strongly influences spin - polarizations in both magnetic and non - magnetic quantum wells , with discontinuities associated with every break in the @xmath12@xmath13 curve . the magnetic well polarization ,
@xmath123 , varies particulary strongly , especially when the monopole moves through the magnetic well , and becomes stronger as the spin - splitting is increased .
this can be seen in fig .
[ fig6 ] , which describes the general behavior of @xmath123 with voltage and spin splitting .
if @xmath68 is further increased ( @xmath68=6 mev ) , new branches appear in the @xmath12@xmath13 curve ( see fig .
[ fig7 ] ) .
the extra branch that appeared close to the @xmath124 transition at smaller @xmath68 is now fully developed .
in addition , a new branch forms _ before _
the @xmath125 transition . to better understand the subtle interplays that control these features ,
we have studied the self - consistent steady - state well - dependent spin - polarizations at the particular voltages marked in fig .
the series of mqw electrostatic profiles illustrated in fig .
[ fig8 ] are dominated by classical field - domain physics not qualitatively influenced by the spin - dependent nature of the transport .
[ fig8](a ) ( @xmath126 ) is the highest voltage at which intrasubband ( @xmath125 ) resonant tunneling can be maintained .
the electric field drops almost linearly along the system . in fig .
[ fig8](b ) ( @xmath127 ) the formation of a high electric field domain in the last barrier is clearly observed , favouring a resonant condition between the third well and the collector first - excited subband .
@xcite the second branch in the ndc region ( @xmath128 ) involves the generation of a larger high field domain ( see fig . [ fig8](c ) ) .
the domain wall is now located in the magnetic well . a jump to the first well ( fig . [
fig8](d))(@xmath129 ) is accompanied by further expansion of the high field domain . in this situation
all tunneling within the wells takes place between the ground and the excited states , followed by a rapid relaxation to the first subband .
[ fig9 ] illustrates spin - populations near @xmath12@xmath13 features where the spin - dependent element introduced by the magnetic quantum well plays a qualitative role .
[ fig9](a ) describes the position of energy levels for the voltage @xmath130 marked in fig .
the subband energies in the non - magnetic wells ( @xmath131 and @xmath132 ) are quasidegenerate .
notice that resonant tunneling occurs between @xmath133 and @xmath134 .
further increase of the voltage , however , results in a decrease of the current since now @xmath135 is then off - resonance ( see fig .
[ fig9](b ) ) .
the current is then increased again since @xmath136 starts to match @xmath137 .
this explains why @xmath123 does not show the behavior observed for smaller @xmath68 ( @xmath68=3 mev ) in the linear regime .
( in contrast to the inset of fig .
[ fig5](b ) , the inset of fig .
[ fig7 ] presents a flat polarization at low voltages . )
@xmath123 starts to increase only when @xmath13 is such that @xmath133 reaches within @xmath42 of @xmath137 .
the splitting of the branch at higher bias voltage where all transport occurs via intersubband ( @xmath124 ) resonant tunneling has a similar explanation .
the fig .
[ fig9](c ) is stabilized by alignment of @xmath133 with @xmath138 levels .
subband mismatch at higher bias gives rise to a sharp reduction of the current , which increases later , as the voltage is increased and resonant tunneling between @xmath136 and @xmath139 levels is achieved .
incidentally , the dips in @xmath123 illustrated in fig . [ fig5 ] can be explained in this way .
when the latter alignment occurs , a large flow of down - spin carriers streams towards the magnetic well , causing a sharp decrease of the polarization .
the remaining features in fig .
[ fig5 ] can be understood in similar terms .
so far all results were calculated by sweeping voltages up .
steady state solutions in the ndc bias voltage region are in general multistable .
we can obtain different solutions at a given bias voltage by evolving solutions following different histories , @xcite for example by decreasing voltages from a high initial bias .
for a given voltage different values of the current with different density and spin polarization profiles may be achieved . to amplify effects ,
in studying this possibility , we have set @xmath106 ns in a @xmath140 superlattice .
the change in @xmath123 with voltage is now so greatly increased ( see fig .
[ fig10 ] ) that even _ reversed _ polarization can be observed at somewhat larger values of @xmath67 in the @xmath124 resonant tunneling regime ( not shown here ) .
this is a direct consequence of the emergence of dominant spin bottlenecks.@xcite the alterations in polarization are more visible for voltages greater than the one corresponding to the @xmath124 resonance , but we choose not to show them here since these involve transitions to higher excited energy levels , close to the top of the barrier , where our model breaks down .
however , for sufficently high barriers it would be natural to obtain such behavior .
[ fig10 ] shows three different values of spin polarization in the magnetic well which can be obtained at a particular bias voltage , depending on the sample history , up sweep from zero voltage , down sweep from a high voltage , and up sweep to an intermediate voltage followed by down sweep .
we emphasize that this kind of hysteretic phenomena between magnetic states is driven here by _
electric _ fields .
we have introduced and studied a simple model for growth direction non - linear transport in multiple quantum well systems containing magnetically doped layers .
our analysis is based on a tunneling hamiltonian expression for the current between spin - polarized quantum wells and on a phenomenological expression for spin - relaxation within quantum wells .
numerical studies of this model show that it predicts rich behavior due to its non - linearity and due to the additional degrees of freedom introduced by spin - dependent transport .
nonlinearity manifests itself in the formation of electric field domains when the differential conductivity between neighboring layers is negative .
we find that the spin polarization of electrons in magnetic wells can change substantially when the system jumps between different branches of the @xmath12@xmath13 curve .
when current flows finite electron spin - polarization extends from the magnetic quantum wells to non - magnetic quantum wells .
when large equilibrium spin splitting , the @xmath12@xmath13 curve is strongly affected by the appearance of extra branches , due to tunneling into and out of spin polarized well subbands and these effects become more and more prominent when the characteristic time for spin - relaxation is longer .
the effects addressed in this paper could be investigated experimentally by studying transport properties and by studying the polarization dependence of inter - band optical absorption and photoluminescence in mqw systems containing magnetically doped layers .
the predicted sensitivity of transport properties to external magnetic fields , suggests that these systems could potentially be useful for magnetic field sensors , most likely in geometries with a relatively small number of quantum wells .
the sensitivity to external fields will be strongest at low temperatures where the the mn ions are easily polarized to produce the maximum equilibrium spin - splitting field . to illustrate the sort of effects that we expect to occur , we have considered only relatively simple geometries with a single magnetic layer .
other effects will occur in larger mqw systems with particular geometries .
in general there is considerable lattitude for designing the mn density distribution in the mqw system to realize desired magnetoresistance effects that could be described with the type of model we have introduced here .
analogs of the magnetotransport effects we discuss will also occur in ferromagnetic multiple quantum well systems , similar to the delta - doped layered ( ga , mn)as systems studied by kawakami _ et al . _
@xcite these systems are ferromagnetic and the carriers are holes rather than electrons , leading to strain - sensitive spin - orbit - coupling induced magnetic anisotropy@xcite and coercitivities .
these properties suggest a rich interplay between the hysteretic magnetoresistance effects common in thin film itinerant electron magnets@xcite and the hysteretic effects discussed here , which have their roots in electric field domain structures.@xcite had ferromagnetism been taken into account in our model calculations , by solving for @xmath68 self - consistenly in eq .
( [ eq - delta ] ) , not only the electric field domain structure but also the magnetic state configuration would have been sensitive to the bias voltage history . exploring these possibilities
appears to be a promising avenue for future experimental and theoretical work .
one of us ( d.s . ) thanks the hospitality of indiana university and the university of texas at austin where most of this work was completed .
this work was supported by the spanish dges grant no . pb96 - 00875 , by the european union tmr contract fmrx - ct98 - 0180 , and by the indiana 21st century fund , the welch foundation and darpa / onr award no .
n00014 - 00 - 1 - 0951 .
the authors acknowledge valueable assistance from tomas jungwirth .
baibich , j.m .
broto , a. fert , f. nguyen van dau , f. petroff , p. eitenne , g. creuzet , a. friederich , and j.c .
chazelas , phys .
61 * , 2472 ( 1988 ) ; g. binash , p. grnberg , f. saurenbach , and w. zinn , phys .
b * 39 * , 4828 ( 1989 ) .
p. kacman , semicond .
. technol .
* 16 * , r25 ( 2001 ) ; j.k .
furdyna and j. kossut , _ diluted magnetic semiconductors _ , vol .
25 of _ semiconductor and semimetals _ ( academic press , new york , 1988 ) ; t. dietl , _ diluted magnetic semiconductors _ , vol .
3b of _ handbook of semiconductors _ ( north - holland , new york , 1994 ) .
j. knig , h.h .
lin , and a.h .
macdonald , e - print : http://arxiv.org/abs/cond-mat/0010417 , to be published in _ interacting electrons in nanostructures _ , eds .
r. haug and h. schoeller , ( springer series , 2001 ) .
macdonald , in _ statistical and dynamical aspects of mesoscopic systems ( proceedings of the xvi sitges conference ) _ , eds .
d. reguero , g. platero , l.l .
bonilla , and j.m .
rub , ( springer - verlag , berlin , 2000 ) .
actually this is not a realistic case since in an experimental sample the contacts are 3d regions .
thus the boundary condition of eq .
( [ eq - bc ] ) fails to account for the experimental evidence : the number of branches in the @xmath12@xmath13 characteristic is _ not _ equal to the number of wells .
proper description of the contact regions @xcite results in a good agreement with the experiment .
however , eq . ( [ eq - bc ] ) is qualitatively a good starting point .
this effect differs qualitatively from the electric - field control of the microscopic properties of ferromagnetic semiconductors recently demonstrated by h. ohno , d. chiba , f. matsukura , t. omiya , e. abe , t. dietl , y. ohno , and k. ohtani , nature * 408 * , 944 ( 2000 ) .
kawakami , e. johnston - halperin , l.f .
chen , m. hanson , n. gubels , j.s .
speck , a.c .
gossard , and d.d .
awschalom , appl .
77 * , 2379 ( 2000 ) ; j. fernandez - rossier and l.j .
sham , preprint [ cond - mat/0106548 ] ( 2001 ) . | we develop a theory of non - linear growth direction transport in magnetically doped ii - vi compound semiconductor multiple - quantum - well systems .
we find that the formation of electric field domains can be controlled by manipulating the space dependence of the band electron spin - polarization , using its exchange coupling to local moments .
we emphasize the importance of band electron spin relaxation in limiting the strength of these effects . 2 |
k. g. , a. h. and k. a. p. were supported by the pan - cnrs program for french - polish collaboration .
moreover , k. g. thanks for support from mnisw ( warsaw , poland ) , `` iuventus plus 2015 - 2016 '' , program no ip2014 013073 . | we introduce and study an extension of the heat equation relevant to relativistic energy formula involving square root of differential operators .
we furnish exact solutions of corresponding cauchy ( initial ) problem using the operator formalism invoking one - sided lvy stable distributions .
we note a natural appearance of bessel polynomials which allow one the obtention of closed form solutions for a number of initial conditions .
the resulting relativistic diffusion is slower than the non - relativistic one , although it still can be termed a normal one .
its detailed statistical characterization is presented in terms of exact evaluation of arbitrary moments and is compared with the non - relativistic case .
the ordinary diffusion problem is customarily treated in terms of parabolic differential equation , linking first order derivative in time and second order derivative in space of a distribution function , which in the case of the heat equation , represents a physical quantity like the behavior of temperature .
the extension of the equation to include relativistic effects has been the topic of intense debate in the past , see @xcite , and particularly @xcite for extensive bibliography .
the authors of @xcite have developed and interesting point of view displaying how the brownian motion underlying ordinary diffusion is extended to relativistic realm . according to @xcite the problem
is ruled by the telegrapher equation , thus finding agreement with the original cattaneo s treatment of the problem @xcite .
the interest of their analysis stems from the fact that they emphasize the intimate connection between the stochastic process described by the telegrapher equation and the dirac equation .
in this note we develop an alternative point of view which takes advantage of the concepts of anomalous diffusion @xcite governed by the evolution equations involving square root pseudo - differential operators .
we postulate and study in this note the following one - dimensional differential equation for the distribution function @xmath0 which we believe would be appropriate in the relativistic ( r ) situation : @xmath1^{1/2}\ } \psi(\xi , \tau),\ ] ] with @xmath2 and @xmath3 respectively space and time variables , and where @xmath4 is the rest mass and @xmath5 has the dimension of action . the conventional diffusion equation @xcite @xmath6 can be obtained from the non - relativistic ( nr ) schrdinger equation by means of the wick rotation @xmath7 .
we observe that eq .
can be analogously produced by performing the wick rotation on the r schrdinger ( i.e. the salpeter @xcite ) equation with substracted rest energy @xmath8 .
the nr limit of eq .
, i.e. the @xmath9 expansion in the spirit of 33 of @xcite , results in eq . with the identification @xmath10
. subsequently we introduce in eq .
the dimensionless variables @xmath11 and @xmath12 and we end up with @xmath13 which is formally of infinite order in @xmath14 . for the purpose of this work we shall refer to eq . as a relativistic heat equation . in quantum mechanics the difficulty to treat the square root in eqs . and
( so - called pseudo - differential operators @xcite ) is usually resolved by introducing the wave functions with several components , like in pauli and in dirac equations @xcite .
however the interest for equations involving pseudo - differential operators arose even before the inception of the dirac equation .
it can be traced back to h. weyl @xcite . the fact that eqs . and correctly reproduce the nr limit is reminiscent of the r extension of the analogy between the schrdinger and the heat equations , carried out in 1984 in @xcite , by identifying the underlying poisson processes .
their real version obeys the telegrapher s equation , which , when analytically continued , produce the dirac equation . to see how this circumstance mirrors in our context , we square the operator of eq .
and we get : @xmath15 where the ensemble of solutions of @xmath16 contains the solutions of eq . .
note that eq .
is similar but not identical to the telegrapher s equation @xcite , also referred to as the cattaneo equation @xcite .
we can shed another light on eqs . and by modifying the definition of constants in eq . :
introduce @xmath17 and @xmath18 with dimensions of time and length respectively
. then @xmath19f(x , t ) \approx ( \delta^{2 } \partial_{x}^{2}/2)f(x , t)$ ] , leading to @xmath20 . if @xmath21 then it is a kind of free path in the characteristic time @xmath17 .
the expansion in @xmath18 means that @xmath18 is smaller than the typical diffusion lenght .
( with the above notation eq .
( 9 ) in @xcite becomes @xmath22h(x , t)$ ] which however can not be treated by our method of eqs . - . )
therefore our eq . provides an alternative to the procedure of @xcite . here
, we keep the scalar nature of @xmath23 in eq . and set out to solve it exactly . since eq .
is linear in @xmath24 its full solution is obtained by the action of the evolution operator @xmath25 on the initial condition ( i.c . )
@xmath26 : @xmath27 f(x).\ ] ] the objective of the present work is to elucidate the properties of @xmath25 and to derive a number of exact solutions of eq . .
to this end we make contact with the exponential generating function of bessel polynomials @xmath28 @xcite : @xmath29.\ ] ] then we define the lvy - smirnov function @xcite @xmath30/(2\sqrt{\pi } \xi^{3/2})$ ] , @xmath31 , the canonical case of one - sided lvy stable distributions @xcite , whose laplace transform is known @xcite : @xmath32 for @xmath33 and @xmath34 .
combining eqs .
and we obtain the relation between @xmath28 and @xmath35 : @xmath36 using now eq . for @xmath37 the following form of eq .
emerges @xmath38 f(x).\ ] ] for many i. c. @xmath39 the series of eq . can be summed explicitly , as the bessel polynomials ( in carlitz form ) are known : @xmath40 . if @xmath39 is defined for @xmath41 then the solution of the nr heat equation , eq . for @xmath42 , is given by @xmath43 @xmath44 , which is the gauss - weierstrass ( gw ) transform @xcite of @xmath39 .
in contrast , the r case , for _ the same _ initial condition @xmath39 , according to eqs . and , is given by @xmath45 i.e. it is expressed through an integral operator involving the @xmath46 function , thus linking the r and nr solutions .
we shall compare now the nr and r equations for i.c .
@xmath47 , i.e. the normalized gaussian .
( the superscripts in @xmath48 enumerate the i.c.s considered in this work . )
as it is well known the nr time - evolved gaussian is again a gaussian , as seen from the glaisher formula after @xmath49 : @xmath50 @xcite . by the same token , the corresponding r exact solution of eq .
is given by the following integral : @xmath51 which is a kind of integral convolution with @xmath46 .
( color online ) comparison of r and nr distributions with i.c .
@xmath52 for @xmath53 : @xmath54 has less evolved than @xmath55 .
this fact holds for all @xmath56 . ] in fig .
[ fig1 ] we compare @xmath52 , @xmath57 and @xmath58 from eq . .
we observe that the nr evolution is more rapid than the r one : for all times the r curves are less flattened that the corresponding nr ones .
similar results hold for i.c . which are not gaussians , s. fig .
[ fig5 ] below .
additional informations about the so obtained distributions are contained in the averages of type @xmath59 , @xmath60 .
for nr case the calculations of @xmath61 are standard and for a general i.c .
@xmath39 they yield @xmath62 , implying that both integrals @xmath63 , @xmath64 should be converging if the description in terms of @xmath65 is adopted . in these cases
@xmath66 , as can be seen from eq . by elementary integration .
it means that the r diffusion is normal .
below we shall also display explicit solutions in which @xmath67 is diverging for all @xmath56 . according to eq .
the r averages read : @xmath68 for @xmath69 and @xmath52 , eq . furnishes @xmath70 , indicating that both nr and r diffusions are of normal type .
more refined characteristics of the r diffusion can be deducted from higher - order averages for @xmath71 , like kurtosis @xmath72^{2 } - 3 $ ] , see @xcite .
it can be evaluated using @xmath73 which leads , for general i. c. @xmath39 , to @xmath74 .
this last result implies that for arbitrary , but the same i. c. @xmath39 , @xmath75 .
we proceed by constructing exact solutions of eq . for a number of i. c. other than the gaussian .
the first instance is @xmath76 , @xmath77 , @xmath78 . as an intermediate step we first treat the unnormalized i.c .
@xmath79 , @xmath78 , @xmath60 where @xmath80 are the conventional hermite polynomials @xcite .
we use now eq . , and observing that @xmath81 = h_{p + 2n}(x ) \exp(-x^{2})$ ] ( s. eq .
( 1.20.1.11 ) of @xcite ) and employing eq .
we obtain the r evolution of @xmath82 : @xmath83\!\!\operatorname{d\!}y \nonumber\\ & = \operatorname{e}^{t}\int_{0}^{\infty}\!\ ! g_{1/2}(y)\!\ !
\left[\frac{\exp(-y t^{2}-\frac{x^{2}}{1 + 4t^{2}y})}{(1 + 4t^{2 } y)^{(p+1)/2 } } h_{p}\left(\!\!\ulamek{x}{\sqrt{1 + 4t^{2 } y}}\!\right)\!\right]\!\!\operatorname{d\!}y,\end{aligned}\ ] ] where in deriving eq .
we used eq .
( 5.12.1.4 ) of @xcite .
eq .
permits one to deduce the exact normalized solution @xmath84 with i.c .
@xmath85 by expressing @xmath86 as a finite sum of @xmath80 s , compare eq .
( 18.18.20 ) of @xcite , yielding @xmath87
the time evolution of eq . is exemplified for @xmath88 and for several values of @xmath56 on fig .
[ fig2 ] .
( color online ) plot of @xmath89 given by eq . for @xmath90 ( black line , i ) , @xmath91 ( red line , ii ) , @xmath53 ( blue line , iii ) , and @xmath92 ( green line , iv ) . ]
we observe a curious behavior at @xmath93 : the time evolution lifts up the values of solutions at @xmath93 : @xmath94 for all @xmath44 . the choice of i.c .
@xmath95^{-1}$ ] , @xmath41 , ( cauchy function ) also allows for exact solution .
however here @xmath96 is diverging .
going back to eq .
we first observe that @xmath97 = ( 2n ) ! ( 1+x^{2})^{-(n+1 ) } u_{2n}(-x/\sqrt{1+x^{2}})$ ] , where @xmath98 , @xmath99 , are the chebyshev polynomials @xcite , and the solution can be recast in the form @xmath100 \operatorname{d\!}y.\end{aligned}\ ] ] the inner summation in eq .
( in obvious notation ) , i.e. @xmath101 , can be repre- + sented in integral form using @xmath102^{-1}\int_{0}^{\infty } \operatorname{e}^{-u } \\\times u^{n } h_{2n}(\tilde{x}\sqrt{u } ) \operatorname{d\!}u$ ] @xcite and the previously quoted eq .
( 5.12.1.4 ) of @xcite .
the result is , for @xmath103 : @xmath104}{\sqrt{1 + \tilde{t}u } } \operatorname{d\!}u,\ ] ] furnishing , with the auxiliary function @xmath105 , ( @xmath106 , @xmath41 ) , the final form : @xmath107 the time evolution of eq .
is illustrated for several values of @xmath56 on fig .
[ fig3 ] .
( color online ) plot of @xmath108 given by eq . for @xmath90 ( black line , i ) , @xmath109 ( red line , ii ) , @xmath110 ( blue line , iii ) , and @xmath111 ( green line , iv ) . ] the treatment of the next set of solutions evolving from the ( non - normalizable ) i.c .
@xmath112 , @xmath41 , is in a sense similar to the preceding case . the key tool to evaluate eq .
here is eq .
( 1.1.2.42 ) of @xcite which reads : @xmath113 = ( 2n ) ! ( 1+x^{2})^{-(n+1/2)}p_{2n}(\ulamek{x}{\sqrt{1+x^{2}}})$ ] , with @xmath114
the legendre polynomials @xcite , @xmath115 .
accordingly , the solution is given in integral form as @xmath116\operatorname{d\!}y,\end{aligned}\ ] ] whose inner summation @xmath117 can be performed with the use of eq .
( 4 ) from 108 of @xcite and eq . ( 5.12.1.4 ) of @xcite , with the result : @xmath118 @xmath103 . defining now
@xmath119 we end up with @xmath120 several examples of @xmath121 are displayed on the fig .
[ fig4 ] .
( color online ) plot of @xmath122 given by eq . for @xmath90 ( black line , i ) , @xmath109 ( red line , ii ) , @xmath110 ( blue line , iii ) , and @xmath111 ( green line , iv ) . ] after having worked out the r evolution from the usual set of i.c .
@xmath48 , @xmath123 we shall apply our formalism to a less common i.c .
, namely @xmath124 , @xmath125 , with @xmath126 diverging .
( this is fully justified as @xmath127 for all @xmath128 . )
the application of eq . immediately yields : @xmath129g_{1/2}(\xi ) \operatorname{d\!}\xi\right\}\operatorname{d\!}u \\ & \qquad = \frac{t\operatorname{e}^{t}}{2\pi^{3/2 } } \int_{0}^{\infty } \frac{k_{1}(\sqrt{t^{2 } + ( x-\xi)^{2}})}{\xi^{3/2}\sqrt{t^{2 } + ( x-\xi)^{2 } } } \operatorname{e}^{-\frac{1}{4\xi } } \operatorname{d\!}\xi , \label{30/07 - 21}\end{aligned}\ ] ] where in eq .
we switched the order of integrations over @xmath130 and @xmath2 , and then employed sommerfeld s representation of modified bessel functions @xmath131 , s. formula ( 8.432.6 ) of @xcite .
observe that eq .
is fully consistent with the path - integral evaluation of the corresponding quantity in @xcite . in fig .
[ fig5 ] we display @xmath132 and @xmath133 for several values of @xmath56 .
observe that @xmath134 , initially confined to @xmath78 , spreads out to negative @xmath135 for _ any _ arbitrarily small @xmath56 . here again , the r evolution is `` slower '' as the nr one , as seen by comparing the r curves with the @xmath133 solutions for the _ same _ values of @xmath56 , obtained via the gw transform of eq . with @xmath134 .
+ ( color online ) plot of @xmath136 given by eq .
( solid curves ) and @xmath133 given by eq .
( dashed curves ) with @xmath137 as the i.c .
( black solid line , i ) for @xmath138 ( red line , ii and ii ) , @xmath139 ( blue line , iii and iii ) , and @xmath140 ( green line , iv and iv ) . ]
the problem of a consistent extension of the conventional diffusion to the relativistic situation has been extensively discussed for a long time .
for a detailed chronology of these efforts consult @xcite .
the main reason was to properly describe the heat waves with finite speeds @xcite .
amongst other subjects to which the relativistic - type diffusion equations would be relevant we mention the landau theory of the second sound in helium ii @xcite and in more general phonon gases @xcite , as well as the thermoelasticity @xcite .
the initial hope that the cattaneo - type equations would be appropriate to treat this range of problems was dissipated once it turned out that this approach is plagued with inherent inconsistencies @xcite .
the current prevailing opinion is that the problem is still open @xcite and therefore the study of other options is evidently legitimate .
the salient feature of our exact analysis based on eq . is the observation that the r diffusion propagates more slowly that the nr one while still retaining its normal " character
. it will be intriguing to verify if this feature persists in more realistic settings .
the natural extension of our formalism would involve considerations in higher dimensions and the explicit inclusion of the potential energy , which constitutes a true challenge . |
the study of the low - lying excited states of the @xmath0 isotones offers the best ground to describe the gradual filling of the active proton orbitals lying above the @xmath7 closure , since these states are free of any neutron contribution up to an excitation energy of about 4 mev .
particularly , the description of the high - spin states mainly involve the effective interactions of the protons when occupying the valence orbits , particularly the high-@xmath6 ones , @xmath9 , @xmath10 , and @xmath11 , which can thus be tested .
many years ago , a lot of low - spin states of several @xmath0 isotones were used to define the two - body part of shell - model hamiltonian @xcite , which was adjusted some times later @xcite using new data on some high - spin states of the light
@xmath0 isotones produced by spontaneous fission of actinides @xcite .
this set of empirical interactions was then tested on the new high - spin states of @xmath4cs , identified in a deep - inelastic reaction up to i@xmath12=(31/2@xmath13 ) @xcite . by combining appropriate fusion - fission reactions to a powerful @xmath1-detection array
, it is possible to populate and to study the high - spin states of many @xmath0 isotones . in this paper
, we present our results obtained on the high - spin states of five isotones , @xmath14xe , @xmath15cs , @xmath16ba , @xmath17la , and @xmath18ce , which have been identified in two euroball experiments .
the level schemes of four of them have been well extended to higher - spin values .
the high - spin structures of @xmath2la have been identified for the first time , displaying an isomeric state with @xmath5 ns located at 1800-kev excitation energy .
the states of these five @xmath0 isotones are then compared to the results of shell - model calculations using the set of empirical interactions already used to describe the lightest isotones .
in addition , groups of states lying above 4 mev excitation energy in @xmath3xe , @xmath4cs , and @xmath8ba are assigned to the excitation of the neutron core .
the @xmath0 isotones of interest were obtained as fission fragments in two experiments .
first , the @xmath19c + @xmath20u reaction was studied at 90 mev incident energy , with a beam provided by the legnaro xtu tandem accelerator . secondly , the @xmath21o + @xmath22pb reaction was studied with a 85 mev incident energy beam provided by the vivitron accelerator of ires ( strasbourg ) .
the gamma - rays were detected with the euroball array @xcite .
the spectrometer contained 15 cluster germanium detectors placed in the backward hemisphere with respect to the beam , 26 clover germanium detectors located around 90@xmath23 and 30 tapered single - crystal germanium detectors located at forward angles .
each cluster detector consists of seven closely packed large volume ge crystals @xcite and each clover detector consists of four smaller ge crystals @xcite . in order
to get rid of the doppler effect , both experiments have been performed with thick targets in order to stop the recoiling nuclei ( 47 mg/@xmath24 for @xmath20u and 100 mg/@xmath24 for @xmath22pb targets , respectively ) . the data of the ( c+u ) experiment were recorded in an event - by - event mode with the requirement that a minimum of five unsuppressed ge detectors fired in prompt coincidence .
a set of 1.9@xmath25 three- and higher - fold events was available for a subsequent analysis .
for the ( o+pb ) experiment , a lower trigger condition ( three unsuppressed ge ) allowed us to register 4@xmath26 events with a @xmath1-fold greater or equal to three . the offline analysis consisted of both multi - gated spectra and three - dimensional cubes built and analyzed with the radware package @xcite .
more than one hundred nuclei are produced at high spin in such experiments , and this gives several thousands of @xmath1 transitions which have to be sorted out .
single - gated spectra are useless in most of the cases .
the selection of one particular nucleus needs at least two energy conditions , implying that at least two transitions have to be known .
the identification of transitions depopulating high - spin levels which are completely unknown is based on the fact that prompt @xmath1 rays emitted by complementary fragments are detected in coincidence @xcite .
moreover , the isotopes of interest being produced from two different fissionning compound nuclei , the complementary fragments are different in the two reactions .
this gives a fully unambiguous assignment of transitions seen in both experiments .
as reported in previous papers @xcite , another experiment was performed using the saphir heavy - ion detector @xcite , here composed of 32 photovoltaic cells , in order to identify new isomeric states in the fission fragments .
placed in the target chamber of euroball , saphir was used to detect the escaping fission - fragments of the @xmath19c ( 90 mev ) + @xmath20u reaction from a thin 0.14 mg/@xmath24 uranium target .
thus , all recorded @xmath1 rays emitted at rest by a fragment implanted in saphir correspond to the decay of an isomeric state , for which the half - life can be measured in a range of several tens to several hundreds of nanoseconds .
it is well known that the @xmath1 rays emitted by fusion - fission fragments do not show any anisotropy in their angular distributions with respect to the incident beam .
however , angular correlations of two successive transitions are meaningful . in order to determine the spin values of excited states ,
the coincidence rates of two successive @xmath1 rays are analyzed as a function of @xmath27 , the average relative angle between the two fired detectors .
the euroball spectrometer had @xmath28=28441 combinations of 2 crystals , out of which only @xmath29 2000 involved different values of relative angle within 2@xmath23 .
therefore , in order to keep reasonable numbers of counts , all these angles have been gathered around three average relative angles : 22@xmath23 , 46@xmath23 , and 75@xmath23 .
the coincidence rate is increasing between 0@xmath23 and 90@xmath23 for the dipole - quadrupole cascades , whereas it decreases for the quadrupole - quadrupole or dipole - dipole ones .
the theoretical values of several coincidence rates for the euroball geometry have been already given in previous papers @xcite .
the method has been checked by correctly reproducing the expected angular correlations of @xmath1-transitions having well known multipole orders and belonging to various fission fragments .
the yrast states of @xmath4cs were obtained up to @xmath32 mev from the analysis of the @xmath33 cm spontaneous - fission study @xcite .
more recently a few transitions were added to this level scheme , using a @xmath34cf source @xcite .
nevertheless the spin - parity values were only based on results of shell - model calculations . in our work ,
@xmath1 rays emitted by @xmath4cs are seen in the two fusion - fission reactions , with a slightly higher production in the ( c+u ) reaction compared to the ( o+pb ) one , especially for the high - spin states .
the high - spin level scheme of @xmath4cs has been extended up to @xmath35 mev excitation energy and spin ( @xmath36 ) , thanks to the observation of about 25 new transitions ( see fig . [ schema_cs137 ] ) .
a 681-kev transition , already identified in the two previous experiments , is also oberved in our work , being located just above the 2783-kev level ( structure b ) .
the coincidence relationships of this @xmath1 line involve the four yrast transitions , but we could not observe any transition located above the 3464-kev level , although the intensity of the 681-kev transition would have allowed us to see a higher - spin part of the cascade .
this level is likely populated by several high - energy transitions which have escaped from our detection apparatus .
we do not confirm two of the six new transitions above the 1893-kev level , which were proposed by the authors of ref .
@xcite in addition to the previous level scheme @xcite . the 2223.6- and 1609.4-kev @xmath1 rays
are not observed in coincidence with the first yrast transitions .
moreover the decay of the 3495-kev level toward the 3302-kev one by means of a 193-kev transition is not observed .
the most noticeable information obtained in this work is the observation of two structures composed of low - energy transitions ( therefore presumably m1 ) located above the 4350- and 5452-kev levels ( structures c and d in fig . [ schema_cs137 ] ) .
the decays of their low - lying states are very fragmented and involve high - energy transitions .
an example of double - gated spectrum showing the first transitions of structure d and its links to structure e is given in fig .
[ spectre_cs137 ] .
rays in coincidence with two transitions of @xmath4cs showing the first transitions of structure d in the low - energy part ( a ) and the links between structures d and e in the high - energy part ( b ) .
the first gate is set on the 712-kev transition and the second gate is set on one of the four yrast transitions ( 1184 , 487 , 222 , and 890 kev ) .
transitions emitted by @xmath37tc , the complementary fragments of @xmath4cs , are labeled by tc .
the peaks marked with a star are contaminants .
, width=321 ] angular correlations of successive @xmath1 rays have been extracted for the most intense transitions of @xmath4cs .
the experimental results are given in table [ correl_cs ] .
the three first results indicate that the transitions at 1184 kev , 487 kev , and 890 kev have the same multipole order , while the last one indicates that the multipole order of the 222-kev transition is different from that of the three others .
an @xmath38 character is ruled out for the 222-kev transition , because the 1893-kev level would then be an isomeric state , with a half - life of a few tens to a few hundreds of ns , at variance with the fact that the 1184 - 487 - 222 kev cascade is not observed in the saphir experiment .
thus the 222-kev transition is assigned to be a dipole transition , while the 1184- , 487- , and 890-kev transitions have an @xmath38 character .
these arguments result in the spin values of 11/2@xmath39 , 15/2@xmath39 , 17/2@xmath39 , and 21/2@xmath39 for the levels at 1184 , 1671 , 1893 , and 2783 kev , belonging to structure b. .coincidence rates between the low - lying @xmath1 rays of @xmath4cs as a function of their relative angle of detection , normalized by the ones obtained around 75@xmath23 . [ cols="^,^,^,^",options="header " , ] all the experimental states observed below 4116 kev have a theoretical counterpart ( see the stars and the crosses in fig .
[ 139la_exp_sm ] ) . for the 15/2@xmath39 states ,
only one experimental level is proposed at 2032 kev , which is close to the prediction of the 15/2@xmath40 state .
the fact that the 15/2@xmath41 state was not populated in our experiment could be explained by its configuration,@xmath42 , which is different from the one of the states lying above it : the 17/2@xmath40 , 19/2@xmath41 , and 19/2@xmath40 states have the same configuration , @xmath43 .
it is important to note that the experimental 21/2@xmath40 , 23/2@xmath39 and 25/2@xmath39 states are located at lower energies than the calculated ones , the deviations being @xmath29 100 kev ( table [ e_exp_sm_139ba ] ) .
this could be related to interactions between the states due to proton excitations among the @xmath9 and @xmath10 orbits ( belonging to the valence space used in the sm calculations ) and the states of structure c ( see fig .
[ schema_la139 ] )
. finally , the states of structure c would be likely due to the breaking of the neutron core , as the corresponding states of @xmath4cs .
nevertheless since the high - spin states having a broken proton pair in the @xmath11 orbit are predicted at lower energy than in @xmath4cs ( about 600 kev less ) , we expect large mixings of these two excitation modes in @xmath2la .
this set of empirical effective interactions within the proton valence space including all the orbits of the 50 - 82 major shell describes well the yrast states of the five isotones studied in the present work , both the even-@xmath30 and the odd-@xmath30 ones .
the deviations between experimental and calculated energies are mostly below 50 kev , that is well better than the values obtained for sm calculations using effective interactions derived from nucleon - nucleon potential , such as done for one prediction of the yrast states of @xmath4cs ( see table ii of ref .
@xcite ) . the latter approach is broader since it can be used in all the valence spaces , thus it would be interesting to compare the numerical values of the two sets in order to find whether some particular tbme are at the origin of the better description and to understand why they would be unsatisfactorily calculated from nucleon - nucleon potential .
the neutron excitation of the doubly - magic @xmath44sn nucleus gives rise to its first positive - parity excited states : the multiplet from the @xmath45 configuration was measured in the 4 - 5 mev energy range , with spin values extending from 2@xmath39 to 9@xmath39 @xcite . in @xmath46te , states with higher spin values are obtained as this first neutron - core excitation can be coupled to the breaking of the proton pair @xcite [ see the red circles in fig .
[ coeurbrise](a ) ] .
gap , i.e. having the main configuration , @xmath47 .
( a ) @xmath48 for @xmath44sn ( black circles ) , @xmath49 for @xmath46te ( red circles ) , @xmath50 for @xmath3xe ( blue triangles ) , and @xmath51 for @xmath8ba ( green triangles ) .
( b ) @xmath52 for @xmath53i ( red circles ) and @xmath54 for @xmath4cs ( blue triangles ) ( data from ref . @xcite and this work ) .
, width=302 ] the states of @xmath3xe , drawn with crosses in fig . [ evenz_exp_sm](a ) , extend from 4379 kev [ i@xmath12 = ( 8@xmath39 ) ] to 7946 kev [ i@xmath12 = ( 17@xmath39 ) ] .
they fit well with those due to the neutron - core breaking of @xmath46te [ compare the blue triangles and the red circles in fig .
[ coeurbrise](a ) ] . finally , the states of @xmath8ba extending from 5740 kev [ i@xmath12 = ( 11@xmath39 ) ] to 6656 kev [ i@xmath12 = ( 14@xmath39 ) ] are in good agreement with those of the lighter isotones [ see the green triangles in fig . [ coeurbrise](a ) ] .
the closeness of the states of @xmath46te , @xmath3xe , and @xmath8ba involving the neutron - core excitation is due to the stability of the gap in energy between the @xmath55 and the @xmath56 orbits , as shown in fig .
38 of ref .
@xcite . for the breaking of the neutron core in the odd - z isotones
, we compare in fig .
[ coeurbrise](b ) the states of @xmath53i lying between 4241 kev and 5576 kev @xcite and those of structure c of @xmath4cs , both of them being not described by sm calculations within the proton space ( see ref .
@xcite and sec .
[ sm_odd ] ) .
they show the same behavior and are in good agreement with the corresponding states of the even-@xmath30 isotones .
even though the configurations of the highest - spin states of @xmath4cs and @xmath8ba observed in the present work can be @xmath57 and @xmath58 , respectively ( see sects .
[ sm_odd ] and [ sm_even ] ) , we can not exclude other configurations involving the breaking of the neutron core , such as @xmath59 and @xmath60 , respectively .
five @xmath0 isotones have been produced as fission fragments in two fusion reactions , @xmath19c + @xmath20u at 90 mev and @xmath21o + @xmath22pb at 85 mev .
their high - spin level schemes have been built by analyzing the triple @xmath1-ray coincidence data obtained with the euroball array . in order to identify the new transitions of @xmath2la ,
we have used their coincidences with the lines emitted by its complementary fragments . moreover ,
the use of the fission - fragment detector , saphir , has allowed us to identify a new isomeric state in @xmath2la , at 1800-kev excitation energy .
spin and parity values of several excited states of @xmath3xe and @xmath4cs could be assigned from the results of @xmath61 angular correlations .
all the states observed in these @xmath0 isotones have been compared to results of shell model calculations performed in the 50 - 82 proton valence space and using empirical effective interactions which had been fitted on previous experimental data .
most of the yrast states of the five @xmath0 isotones are very well described by this approach .
in addition , the excitation of the @xmath0 core gives rise to several structures which have been clearly identified in @xmath3xe , @xmath4cs , and @xmath8ba .
the euroball project was a collaboration among france , the united kingdom , germany , italy , denmark and sweden .
the first experiment has been performed under u.e .
contract ( erb fhgect 980 110 ) at legnaro .
the second experiment has been supported in part by the eu under contract hpri - ct-1999 - 00078 ( euroviv ) .
we thank many colleagues for their active participation in the experiments , drs .
a. bogachev , a. buta , j.l .
durell , th .
ethvignot , f. khalfalla , i. piqueras , a.a . roach , a.g .
smith and b.j .
we thank the crews of the vivitron .
we are very indebted to m .- a .
saettle for preparing the pb target , p. bednarczyk , j. devin , j .-
gallone , p. mdina , and d. vintache for their help during the experiment .
b. h. wildenthal , in _ understanding the variety of nuclear excitations : proceedings of the 3rd international spring seminar on nuclear physics , ischia , 1990 _ , edited by a. covello ( world scientific , singapore , 1991 ) p. 35 .
j. blomqvist , acta phys . polon . *
b30 * , 697 ( 1999 ) .
c.t . zhang _
et al . _ ,
lett . * 77 * , 3743 ( 1996 ) .
et al . _ ,
c * 59 * , 3066 ( 1999 ) .
r. broda _ et al .
c * 59 * , 3071 ( 1999 ) .
j. simpson , z. phys .
a * 358 * , 139 ( 1997 ) and f. a. beck prog . part
a * 28 * , 443 ( 1992 ) .
j. eberth _
et al . _ ,
methods * a 369 * , 135 ( 1996 ) .
g. duchne _
et al . _ ,
methods * a 432 * , 90 ( 1999 ) .
d. radford , nucl .
methods * a 361 * , 297 and 306 ( 1995 ) .
et al . _ ,
nucl . phys . *
a530 * , 111 ( 1991 ) .
et al . _ , acta phys .
polonica b * 27 * , 179 ( 1996 ) .
r. lucas _ et al .
j. a * 15 * , 315 ( 2002 ) .
j. a * 24 * , 39 ( 2005 ) .
_ , _ proceedings of the 2nd international workshop on nuclear fission and fission product spectroscopy , seyssins , france _ ,
april 1998 , aip conf . proc . * 447 * , 143 ( 1998 ) . a. astier _ et al .
j. a * 30 * , 541 ( 2006 ) .
et al . _ ,
c * 84 * , 054305 ( 2011 ) .
ensdf database , http://www.nndc.bnl.gov / ensdf/. j.k hwang _ et al .
c * 58 * , 3252 ( 1998 ) .
h. prade , w. enghardt , h.u .
jger , l. kubler , h .- j .
keller , and f. stary , nucl.phys .
* a370 * , 47 ( 1981 ) .
_ , nucl.phys .
* a472 * , 381 ( 1987 ) .
et al . _ , acta phys .
polonica b * 36 * , 1313 ( 2005 ) .
w. enghardt , h.u .
jger , l. kubler , h .- j .
keller , h. prade and f. stary , nucl .
a449 * , 417 ( 1986 ) .
i. talmi , _ simple models of complex nuclei _ ( harwood academic publishers , chur , switzerland , 1993 ) . a. astier and m
porquet , phys .
c * 83 * , 034302 ( 2011 ) .
j. a * 20 * , 245 ( 2004 ) . | five @xmath0 isotones have been produced in two fusion - fission reactions and their @xmath1 rays studied with the euroball array .
the high - spin states of @xmath2la have been identified for the first time , while the high - spin yrast and near - to - yrast structures of the four others have been greatly extended . from angular correlation analysis , spin values
have been assigned to some states of @xmath3xe and @xmath4cs .
several cascades involving @xmath1 rays of @xmath2la have been found to be delayed , they deexcite an isomeric state with @xmath5 ns located at 1800-kev excitation energy
. the excited states of these five @xmath0 isotones are expected to be due to various proton excitations involving the three high-@xmath6 subshells located above the @xmath7 shell closure .
this is confirmed by the results of shell - model calculations performed in this work .
in addition , high - spin states corresponding to the excitation of the neutron core have been unambiguously identified in @xmath3xe , @xmath4cs , and @xmath8ba . |
it is well known that severe plastic deformation allows obtaining of materials combining high strength with plasticity .
one of such methods is twist extrusion ( te ) that is extruding of a prismatic billet though the die with twist channel .
channel geometry enables the deformed billet to preserve the size and the form identical to the initial state .
the maximum accumulated deformation per one pass is @xmath0 ( @xmath1 is the descent angle of the twist line ) @xcite .
earlier , we analyzed the effect of te on the redistribution of alloying elements in low - carbon construction steel and formulated general conceptions of morphological structure changes @xcite . to obtain more detailed description of structure changes in low - carbon steels under te , the method of automatic analysis of images of electron backscattering diffraction ( ebsd )
was used here @xcite .
as investigated material was low - carbon steel @xmath2 of the following composition , mass in percent : @xmath3 .
the samples were obtained by warm forging ( @xmath4 ) and succeeding milling down to cross - section of @xmath5 .
then the samples were annealed at @xmath6for @xmath7 hour and cooled in air .
the extrusion was carried out with hydraulic press performing three passes at @xmath8 and backpressure of @xmath9 . before the first pass ,
the sample was heated up to @xmath10 ; before the second and the third pass , the heating was done up to @xmath4 ; the temperature of the equipment was @xmath11 in all the cases .
the total accumulated deformation was @xmath12 .
electron scanning microscope jsm-6490lv ( jeol , japan ) with special holder for bulk samples of @xmath13@xmath13@xmath14 mm in size was used for ebsd analysis of structure changes .
the principle of the analysis is well known ; electron beam scans the selected surface of the sample and kikuchi images consisting of kikuchi bands are plotted for every point .
every band is associated with a definite group of crystal planes . using software of @xmath15 , we established the position of each kikuchi band , compared it with the theoretical data and calculated three - dimensional crystallographic orientation of the lattice .
the obtained three - dimensional information was used for reconstruction of the microstructure . to accumulate a representative sample
, we had to analyze at least @xmath16 grains on the specimen .
thus , the following procedure was applied : magnification of @xmath17 was used , scanning area was @xmath18@xmath19 , scanning step was @xmath20 ( at least @xmath13 measuring points per a grain ) .
the degree of indicating was at least @xmath21 percent , the required result was not achieved at a lower value .
the method allowed both qualitative analysis by mapping and qualitative analysis using statistical data . besides
, texture analysis was done . for more detailed structure reconstruction ,
the threshold of noise misorientation was established at @xmath22 . at lower threshold
, sub - grain boundaries would be taken into account and the analysis of ebsd - maps would be complicated . nevertheless ,
when analyzing distributions of grains and subgrains and drawing pole figures , the whole data set was accounted including misorientations below the threshold of @xmath22 .
for the visualization of grain morphology , contrast maps were created where grains were depicted in gray tones and grain boundaries were represented by black lines ( fig .
[ f1 ] , a , b ) .
contrast maps are similar to the structure images obtained with the optical microscope .
the maps of small - angle and high - angle boundaries were drawn for detailed analysis of the grain boundaries ( fig .
[ f1 ] , c , d ) .
boundaries with the misorientation angle less than @xmath14 are colored in green and those above the given threshold are colored in black .
we have obtained also the maps characterizing the part of strained ( deformed ) and unstrained ( recrystallized ) grains ( fig .
[ f1 ] , e , f ) in steel after te : unstrained recrystalized grains are marked by blue color , polygonized grains are marked by yellow color and deformed grains are red ( these maps have been built only for bcc - iron . beside bcc - iron ,
steel contains some fcc - iron and cementite . according to ebsd data , the initial composition was as follows : @xmath23 percent of bcc iron , @xmath24 percent of fcc iron and @xmath25 percent of @xmath26;the composition in the deformed state was : @xmath27 percent of bcc iron , @xmath28 percent of fcc iron and @xmath29 percent of @xmath26 . ) .
the basic idea of these maps is that if there are no misorientations in the neighbor pairs of the analyzed points within a grain , the grain is identified as an unstrained recrystallized one . in the case
when misorientations between two neighbor subgrains exceed @xmath28 , the grain is polygnized and if they are above this value , the grain is a deformed one .
homogeneity of deformation was analyzed with the use of schmids maps .
the slip system @xmath30 was selected as the most informative one in bcc iron .
schmid s factor is evaluated as @xcite : @xmath31 , where @xmath32 is the angle between the slip direction and the deformation axis , @xmath33 is the angle between the normal to the slip plane and the deformation axis .
the maximum value of schmid s factor is achieved at @xmath34 .
a macroscopic shift occurs when the shear strain in the given slip system reaches the maximum value @xmath35 , that is called the critical shear strain .
shear strain @xmath36 is related to schmid s factor as @xmath37 .
thus , schmid s maps shows , what grain will be deformed earlier under uniaxial loading ( these grains are colored in light , fig .
[ f1 ] , j , h ) .
correlation of esbd - maps of different types is explicitly seen .
all of them demonstrate grain refinement from @xmath38 to @xmath39 after te and substantial part of high - angle boundaries .
this fact is confirmed by data of optical and electron scanning microscopy .
after warm deformation by te , recrystallized ferrite grains are detected in the material ( fig .
[ f1 ] e , f ) .
they are located mostly near high - angle boundaries but never within large grains .
similar location of recrystallized volume near grain boundaries was observed in the deformed nickel alloy @xcite ; a suggestion was made that recrystallization nuclei are formed in the most distorted areas of the lattice , i.e. at high - angle grain boundaries .
nevertheless , small recrystallized grains were found within a large polygonized grain . in this case
, we suppose that they were formed from a few adjoining subgrains by defect flow toward the polygonized grain boundary inside the grain .
the result was the reduction of defect density in the subgrain and increased misorientation of the boundary up to the high - angle one ( fig.[f1 ] f ) .
this fact means that two processes ( dynamical polygonization and dynamical recrystallization ) progress simultaneously .
these processes can occur jointly or by turns .
thus , the analysis of ebsd maps gives evidences that the structure of the tested steel consists of fragmented , polygonized and recrystallized grains . on the basis of statistical analysis of the data of ebsd maps , graphs of the frequency distribution of the grain size and
misorientation angles were constructed .
it follows from fig .
[ f2 ] that the metal structure in the initial frequency distribution of grains by size , width=3 ] state was characterized by bimodal grain distribution with the maximums at @xmath40 and @xmath41 , i.e. by explicit difference in the grain size . after te ,
the distribution became more homogeneous with the maximum at @xmath41 confirmed by ebsd - maps ( see .
[ f1 ] , a , b ) . according to the distribution of
grain boundary misorientation angles ( fig .
[ f3 ] ) , the specific part of high - angle boundaries in the initial frequency distribution of misorientation angles of grain boundaries , width=3 ] material was about @xmath42 percent .
this parameter was retained at high level ( @xmath43 percent ) in the deformed material , too .
the formation of small - angle boundaries could be explained by specific features of the material .
for instance , in copper under te , formation of small - angle boundaries prevails at the initial deformation stages , but their amount is reduced at developed deformation @xcite .
we can suggest that high - angle grain boundaries will dominate in steel at accumulation of higher deformation degree .
[ f4 ] illustrates parts of different grains depending on the state of the material .
the data were obtained by numerical component distribution of steel microstructure : @xmath7 recrystallized grains , @xmath28 polygonized grains ; @xmath44 deformed grains , width=3 ] processing of ebsd maps ( fig.[f1 ] e , f ) .
so , we can make a conclusion that the content of recrystallized grains in the initial state of the sample was @xmath45 percent , and this amount was reduced to @xmath46 percent after the deformation . as the deformation of the sample in the course of processing is substantial , all grains are deformed with succeeding passing the stages of defect density accumulation , fragmentation , dynamical polygonization and dynamical re - crystallization
. thus , occurrence of @xmath46 percent of recrystallized areas can be explained by active relaxation processes .
the effect of dynamical recrystallization of steel at severe plastic deformation was observed in a number of experiments : for example , at ecap @xcite , hpt@xcite . as distinct from traditional x - ray methods used for the tests of large areas of a specimen , ebsd method allows detailed analysis of texture evolution in the selected zones
[ f5 ] and [ f6 ] demonstrate pole figures registered in the areas of about @xmath47 in size .
the analysis of inverse pole figures gives information not only about the presence of the texture in the sample but also about the mechanisms involved in the course of the processing . in the case of one of another prevailing pole density , we can make conclusions about the dominating relaxation mechanism ( intragrain or intergrain sliding ) , and texture evolution can be related to the change of deformation mechanisms or development of dynamical recrystallization @xcite .
if the maximum pole density is located near @xmath48and @xmath49 orientation , the mechanisms of twinning progressed actively within the sample @xcite .
if the intensity of the outlet of normal lines is reduced , ( the texture is smearing ) and the distribution becomes more homogeneous as in the case considered above ( fig .
[ f5 ] ) , we can say about activated non - crystallographic mechanisms of plastic deformation .
texture analysis can not be accomplished without the analysis of direct pole figures characterizing the density of the outlet of normal lines of definite crystallographic planes ( here these are @xmath50 planes ) .
the direct pole figures confirm the fact that axial features of the texture are reduced in the tested steel after te , being weakly expressed in general .
two factors can be associated with this phenomenon , these are peculiarities of the deformation scheme ( transversal flow of metal ) and relaxation processes .
accidentally arranged sharp texture maximums present in pole figures can indicate to the appearance of recrystallized grains @xcite .
because of possible ambiguity , this effect requires additional investigations .
the use of ebsd method has revealed effect of twist extrusion of the structure of low - carbon construction steel . besides grain refinement and texture smearing , such treatment results in more homogeneous structure with a substantial part of high - angle boundaries .
this fact is explained by peculiarities of the deformation scheme at twist extrusion and evolution of relaxation processes .
the authors are very grateful to prof .
dobatkin for the material granted for the experiments , to dr .
efros for productive discussion and to providing engineer b.. stebletsov for the aid in construction of the sample holder . | the structure of low - carbon steel after twist extrusion is tested with using electron backscattered diffraction .
it has been shown that warm twist extrusion results in grain refinement with conservation of a substantial part of high - angle boundaries , smearing of the texture , more homogeneous distribution of grains and development of the dynamic recrystallization processes . |
j. f. has greatly benefited from conversations with cheng - yi sun .
this work was supported by nsfc , 973 program ( 2010cb922904 ) .
w.l.y . acknowledges the support of nsfc ( grant nos .
11075126 and 11031005 ) .
y.z.z . acknowledges the support of the australian research council through dp110103434 .
u. h. danielsson , phys .
d 66 ( 2002 ) 023511 ; j. high energy phys . 07 ( 2002 ) 040 ; phys . rev .
d 71 ( 2005 ) 023516 ; n. kaloper , m. kleban , a. lawrence and s. shenker , phys .
d 66 ( 2002 ) 123510 ; k. goldstein and d. a. lowe , phys .
d 69 ( 2004 ) 023507 ; r. easther , b. r. greene , w. h. kinney , and g. shiu , phys . rev .
d 66 ( 2002 ) 023518 ; j. martin and r. brandenberger , phys .
d 68 ( 2003 ) 063513 ; k. schalm , g. shiu , and j. p. van der schaar , j. high energy phys .
04 ( 2004 ) 076 . | we investigate the quantum teleportation between a conformal detector alice and an inertial detector bob in de sitter space in two schemes , ( i ) one uses free scalar modes and ( ii ) one utilizes cavity to store qubit .
we show that the fidelity of the teleportation is degraded for bob in both cases .
while the fidelity - loss is due to the gibbons - hawking effect associated with his cosmological horizon in the scheme ( i ) , the entanglement decreases in the scheme ( ii ) because the ability to entangle the cavities is reduced by the spacetime curvature . with a cutoff at planck - scale , comparing with the standard bunch - davies choice , we also show that the possible planckian physics cause extra modifications to the fidelity of the teleportation protocol in both schemes .
+ one of the main challenges in modern physics is to find a complete theory of quantum gravity which merges quantum mechanics and general relativity into a unified framework .
evidences from some candidate theories ( e.g. string theory ) have shown @xcite that typical quantum gravitational phenomena should be a unitary process .
this however conflicts with the semiclassical analysis that predict the information loss during the process . to resolve this paradox ,
it has become increasingly clear @xcite that non - locality , the basic feature of quantum information theory , should be employed .
this indeed allows one to understand some quantum gravitational effects in a quantum - information framework .
recently , a new fast growing field called relativistic quantum information ( rqi ) ( see ref.@xcite for a review ) has shed new light on this issue .
the insight from rqi is the novel observer - dependent character of quantum correlations like entanglement . for a bipartite entangled system in flat space , this means @xcite an accelerated observer would experience decrement of quantum entanglement he shares initially with an inertial partner due to the celebrated unruh effect @xcite .
such kind of environmental decoherence has later been generalized to curved background @xcite . for a static observer nearby the black hole ,
a degradation of quantum correlations provoked by hawking radiation from event horizon would be detected .
the entanglement produced in the formation of a black hole has also been studied and provides a quantum information resource between the field modes falling into the black hole and those radiated to infinity . by imposing proper final - state boundary conditions at the singularity @xcite
, this non - locality could transmit information outside the event horizon via a teleportation - like process and restore the unitarity of black hole evaporation process . on the other hand ,
it has been emphasized @xcite that even the standard teleportation protocol @xcite is highly non - trivial in the rqi framework . in
flat space , as a result of entanglement degradation , the fidelity of a teleportation process would also suffer a reduction for an observer with uniform acceleration .
moreover , quantum teleportation process in a black hole background was also investigated @xcite and it was shown that the fidelity is considerably reduced for the fixed observer near the horizon . an analogues experiment using sonic black hole is proposed @xcite to test this phenomenon in a suitable laboratory setting . in this letter
, we investigate a quantum teleportation process in de sitter space , which idealizes the inflation epoch of early universe and plays a fundamental role in quantum gravity theory ( see ref .
@xcite for a review ) .
we propose a protocol to teleport an unknown qubit @xmath0 from a conformal observer alice to her inertial partner bob while they initially share a bell state .
more specifically , we investigate two schemes of our protocol , ( i ) one using free scalar modes from which a qubit can be truncated @xcite , and ( ii ) another scheme utilizing moving cavities to store the field modes related with respective observers . in the scheme ( i ) , unlike the standard teleportation protocol , we will show that the decoherence , provoked by the gibbons - hawking effect associated with bob s cosmological horizon @xcite , would reduce the fidelity of the teleportation process .
while de sitter space provide the best scenario to the so - called trans - planck problem @xcite , we also discuss its influence on our teleportation scheme from the existence of some fundamental scales ( like planckian or even stringy ) . with a cutoff on physical momentum of field mode at planck - scale , comparing with the standard bunch - davies choice , we calculate the extra modifications to the fidelity of the teleportation process from the possible high - energy new physics . in the scheme ( ii )
, since the field modes are localized in cavities by the reflecting mirrors , the entanglement between the cavities could be protected from the degradation once it has been prepared @xcite . however , as we will show that , even using of localized cavity states could save the protocol from the gibbons - hawking radiation , the ability to entangle the different cavity mode is still decreased by the curvature of spacetime , that results a similar qualitatively dependence as in ( i ) between the fidelity and the hubble parameter @xmath1 . to proceed , we first recall the thermal feature of quantum field theory in de sitter space .
consider the mode expansion of a free scalar field in de sitter space @xmath2 \label{modeboson}\ ] ] the vacuum state , which respects the spacetime isometries , is defined by @xmath3 . to specify the mode functions @xmath4 ,
the coordinate systems affiliated to different observers should be employed prior to solve the field equation .
a conformal observer in de sitter space adopts the planar coordinates which reduce the spacetime metric to @xmath5 where @xmath6 is conformal time , and the coordinates cover the upper right triangle of the carter - penrose diagram ( both regions i and ii ) , as depicted in fig.[telep - ds ] .
( or the entanglement decrease during entangling cavities ) , the fidelity of teleporting a qubit @xmath7 would be suppressed . ]
since the space undergoes an accelerated expansion , it follows that the wavelength of field mode could become arbitrarily small if one goes backwards in @xmath8 long enough , where any distinction between de sitter space and minkowski space could be safely ignored .
therefore , an essentially unique bunch - davies vacuum @xmath9 could be defined , by requesting it approaching the conformal vacuum of minkowski space in the limit @xmath10 . however , the existence of the fundamental scales , where the quantum gravitational effects become unignorable , prevents us from following a mode back unlimited ( or equivalently , to the arbitrary short distance ) @xcite . imposing a reasonable cutoff on physical momentum as @xmath11 ,
the latest time with quantum gravity dominant is @xmath12 , where @xmath13 refers to the planck energy scale .
this results a modified vacuum state of the conformal observer as @xmath14 , which is in general different from the bunch - davies choice and could be formally realized as a squeezed state @xmath15 where the superscript @xmath16 indicates the bunch - davies choice and throughout . without a complete theory of quantum gravity ,
this new vacuum of conformal observer can provide a typical signature of planck - scale physics .
for instance , it was shown ( see @xcite and the references therein ) that the inflation power spectrum @xmath17 with respect to the new vacuum would be modified as @xmath18 , which is expected to be observed in the wmap or planck satellite experiments .
more ambitious view is that above argument indeed provides a one - parameter family of vacua with the @xmath13 predicted by various quantum gravity theories , e.g. an energy scale interpolated between planckian and stringy scales .
equivalently , this leads the so - called @xmath19-vacua which have been known for a long time @xcite . introducing the new mode basis related to the bunch - davies one by the mottola - allen ( ma ) transformation @xmath20\ ] ] where @xmath19 is an arbitrary complex number with re@xmath21 , @xmath22 .
the one - parameter family of vacua is defined as @xmath23 , where @xmath24 \label{maboson}\ ] ] are the corresponding annihilation operators .
these @xmath19-vacua preserve all @xmath25 de sitter isometries , and clearly include the bunch - davies vacuum as one element since @xmath26 if re@xmath27 .
moreover , the condition ( [ bc2 ] ) can now be explicitly resolved as |0^_k=|0^_k[bcboson ] in a realistic model , the value of @xmath19 could be strictly constrained .
first , for a theory consistent with @xmath28-invariance , @xmath19 should be real . therefore
, we henceforth adopt @xmath29re@xmath19 for simplicity . on the other hand , rather than the bunch - davies choice ,
if a non - trivial @xmath19-state ( @xmath30 ) is chosen as an alternative initial state of inflation , the modified power spectrum of inflationary perturbations requires that @xcite @xmath31 .
it was shown @xcite that the vacuum state defined by the conformal observer would be nonempty in the view of a static observer .
while the bunch - davies vacuum appears thermal with the gibbons - hawking temperature @xmath32 , it is clear that these @xmath19-vacua would exhibit non - thermal feature encoding the quantum gravitational corrections . in terms of the static coordinates ,
de sitter metric becomes ds^2=(1-r^2h^2)dt^2-(1-r^2h^2)^-1dr^2-r^2d^2 [ static ] where @xmath33 is the cosmic time .
the coordinates only cover the region i in fig .
[ telep - ds ] , half of the planar coordinates dose .
the hypersurface on @xmath34 is a cosmological horizon for an observer situated at @xmath35 . since the existence of the killing vector @xmath36
, a static vacuum @xmath37 could be defined unambiguously . to analyze the thermality of this vacuum , we employ the particular useful painlev coordinates @xcite , which reduce the metric ( [ confmetric ] ) into ds^2=(1-h^2r^2)dt^2 + 2hrdrdt-dr^2-r^2d^2 covering both regions i and ii , and @xmath38 , admitting the same killing vector as in ( [ static ] ) throughout the whole space but change its character from timelike to spacelike upon passing through the horizon .
the most attractive feature that we utilized is that the coordinates admit the same bunch - davies vacuum , while the out - state can be split into a tensor product @xmath39 , with @xmath40 , after tracing over the degrees in region ii unaccessible for the static observer in region i. denoting by @xmath41 and @xmath42 the particle annihilation operators in regions i and ii , we can relate them with those in the conformal framework by the bogoliubov transformations a^_k = r
a_k^i - r a_-k^ii [ bogoboson ] with the squeezing operator @xmath43 $ ] , the bunch - davies vacuum for the conformal observer can be realized as a squeezed state in the view of inertial observer @xmath44 obtained from the gibbons - hawking effect .
this means that particles are created in pairs on either side of event horizon , and only the one in region i could be detected as de sitter radiation by an inertial observer . for the general @xmath19-vacua
, deviations from thermality should be included . from ( [ bcboson ] ) and ( [ bogoboson ] )
, it follows that @xmath27 , these corrections can be neglected and a pure thermal de sitter radiation associated with the bunch - davies choice is recovered .
the one - particle excitation in @xmath19-vacua can also be obtained as we now investigate our teleportation protocol in de sitter space . as stated in the introduction , the scheme ( i ) of the protocol
utilizes the free field modes with infinite levels , from which a a qubit can be truncated @xcite .
we work with so - called dual rail encoding which is used in linear - optical quantum computing schemes and the experimental scheme of quantum teleportation protocol @xcite .
the bipartite system contains two conformal observers alice and bob , sharing a bell state when they coincide .
suppose that each one supports two orthogonal modes , with the same frequency , labelled @xmath45 , @xmath46 with @xmath47 , the total state of alice and bob encodes a two - qubit bell state @xmath48 , @xmath49 , with similar expression for bob .
here the superscript @xmath19 indicates a general choice of vacuum state with cutoff at certain fundamental scales in de sitter space . after their coincidence , we suppose bob becomes inertial while alice maintains her comoving motion with respect to conformal time @xmath8 . as we analyzed before , for the inertial observer bob , the vacuum state chosen by alice becomes thermal and can be written as the two - mode squeezed states ( [ bosealpha ] ) and ( [ onealpha ] ) .
to establish a teleportation protocol between alice and bob , which means to teleport an unknown state from conformal detector alice to now inertial detector bob , alice possesses an additional qubit @xmath50 in dual - rail basis .
the input state to system is then @xmath51 , which can be expanded in the bell basis associated with cavities @xmath52 and @xmath53 .
if alice makes a joint projective measurement on her two logical qubits with the result @xmath54 , the full state could be projected into |=(|i|j)_ac|_ij_b where bob s state is @xmath55 , @xmath56 , @xmath57 , @xmath58 . depending on the results of measurement @xmath59 sent from alice by classical channel
, bob can recover the unknown state by applying a proper unitary transformation on his qubit and complete the protocol in his local frame .
the main point is , such a teleportation process in de sitter space should be indeed influenced by the local gibbons - hawking radiation detected by bob .
moreover , additional modification from the existence of minimal fundamental scale should also be taken into account if one starts from general conformal vacuum like ( [ bosealpha ] ) rather than the bunch - davies choice .
the way to probe the deviations of our scheme from the standard teleportation process in flat space is to calculate the fidelity of the teleported state of bob .
we propose that alice should not yet cross bob s event horizon at @xmath60 when she sends her measurement result out , otherwise bob would never receive it via a classical channel .
since bob has no access to field modes beyond his cosmological horizon , his state would be projected into a mixed one by tracing over the states in region ii . from ( [ bosealpha ] ) ,
( [ onealpha ] ) and ( [ cstate ] ) we have & & _ ij^i=_k , l=0^ _ iik , l|_ij_b_ij|k , l_ii + & & = ( 1-^2r^2)^3_n=0^_m=0^n[bobstate ] which can also be written in block diagonal form as @xmath61 , with the coefficients p_0&=&0 , p_1=(1-^2r^2)^3 , + p_n&=&(^2r^2)^n-1(1-^2r^2)^3 it should be emphasized that , by applying a proper unitary operation , bob can only turn his state ( [ bobstate ] ) into a region i analogue of the unknown state from alice , like @xmath62 , expanded in the restricted rail basis @xmath63 .
this deviation from the standard teleportation protocol can be measured by the fidelity defined as f^i_i(|_i_i|^i)=_i|^i|_i=(1-^2r^2)^3 [ fidelity ] which indicates the probability that the bob s state @xmath64 will pass a test to identify it as the desired teleported state @xmath7 .
recall the relation @xmath44 and ( [ bosoncorrec ] ) , above result can be depicted as in fig.[fidelity - ds ] . .
] our first observation is that , the fidelity of the teleportation process in de sitter space is closely related to the hubble parameter @xmath1 ( or the curvature radius @xmath65 of spacetime ) . as we demonstrated before , this phenomenon roots from the information loss via de sitter radiation detected by inertial observer bob , similar to the unruh effect for accelerated frame in flat space or hawking radiation from black hole .
for the bunch - davies state which has @xmath66 with pure thermal spectrum , we could reach an conclusion compatible to the remarkable results of @xcite where the hubble parameter @xmath1 is replaced by the acceleration of the unruh detector or the surface gravity of black hole .
moreover , we also observe that , if one consider the possible cutoff at fundamental scales , the resulting quantum gravitational effect could be encoded in the pattern of fidelity evolution for our teleportation scheme . comparing with the standard bunch - davies choice , the fidelity is suppressed for all @xmath19-vacua choice with @xmath30 . in a realistic model @xcite with @xmath67gev and @xmath68gev
, the typical value of @xmath19 can be estimated to be @xmath69 . by proper tuning of the remaining parameters ,
this choice could result in a significantly modification in the degradation pattern of fidelity for bob .
we now turn to the scheme ( ii ) which utilizes cavity to store the qubit for each observer .
the reflecting mirrors impose the boundary conditions on the field , which result a localized mode restricted inside the cavity .
while the exact solutions of field equation in curved background are highly involved , for simplicity , we demonstrate our model in a 2-dimensional de sitter space .
we assume alice possesses cavity @xmath52 described in ( [ confmetric ] ) with two mirrors position at @xmath70 and the length @xmath71 .
the solution for the scalar field equation @xmath72 which satisfied the dirichlet boundaries @xmath73 is ^_a , k(,z)=n_ah^(2)_(k_n)realizing the bunch - davies vacuum , where @xmath74 and @xmath75 is the hankel function with the order @xmath76 . the further simplification could be made for ignoring the tiny mass , then a general mode related with @xmath19-vacua is^_a , k(,z)=(e^-ik_n+e^e^ik_n ) [ bdsolution ] similarly , the localized mode in bob s cavity @xmath77 situated at @xmath78 could be solved in static coordinates by the hypergeometric function @xcite . within the same limitation
, we have _ b , k(t , r)=e^-ik_nt [ staticsolution ] where @xmath79 and the static vacuum has been admitted . by the transformations between different charts ( [ confmetric ] ) and ( [ static ] ) = -,z = r [ atomcoordtransf ] the cavities length in different observers view could be related by @xmath80 .
since we assume perfectly reflecting mirrors , the field modes inside a cavity are protected and the detector should not be kicked @xcite . as the resource for quantum information tasks ,
the entanglement between the cavities now should be maintained once it was prepared . by this meaning
, the teleportation protocol could be saved from hawking radiation .
however , since entangling cavities is a very non - trivial process in curved background , the ability to prepare the entangled state between alice and bob would result a curvature - dependence in protocol .
we generalize the robust scheme in flat space @xcite to prepare the entangling cavities in the curved de sitter background , which employees a two - level excited atom , passing through cavities @xmath52 and @xmath77 in their ground states .
after subsequently measurement on the atom , a photon has been emitted into one of the cavities if the atom is found in its ground state .
the interaction between the atom and cavity modes is described by the hamiltonian _
i=()[_a()_a(x())+_b()_b(x ( ) ) ] where @xmath81 is atom s proper time and the monopole operator is @xmath82 .
the switching functions @xmath83 and @xmath84 model the effective interaction time for the atom passes through the length of cavity . by mode functions ( [ bdsolution ] ) , ( [ staticsolution ] ) ,
the final entangled state between alice and bob s cavities can be given & = & _ k(c_a^_k+c_b^i_k)|0_a^|0^s_b + & = & _ k(c_a|1_a^|0^s_b+c_b|0_a^|1^s_b ) [ entangling ] which depends on the spacetime curvature in terms of @xmath1 , and clearly is not a maximally entangled state as ( [ maxentg ] ) .
once the entangled state is prepared , same local measurements could be made as in scheme ( i ) . however , while the localized cavity states have been protected from the gibbons - hawking radiation , the degradation of entanglement during entangling cavity results an imperfect teleported state for bob , similar as ( [ cstate ] ) but with the coefficients @xmath85 , @xmath86 , @xmath87 , @xmath88 .
therefore , the fidelity loss in this scheme depends on our ability in preparing the entangled state .
we note that the planckian modification has been included in the possibility amplitude @xmath89 . to calculate it ,
we specify the atom s trajectory , @xmath90 in conformal coordinates , and can be expressed in the static coordinates as @xmath91 , -\frac{z}{h\tau}\big ) $ ] , where @xmath92 is the spatial location of the atom . for alice s cavity ,
we choose a gaussian - like switching function @xmath93 , where @xmath94 denotes the time that atom passing through the cavity center , and @xmath95 depends on both the cavity geometry and the atom s transverse velocity . from ( [ bdsolution ] ) , ( [ entangling ] )
, we have c_a&=&-iw n_a ( ) + & & + where @xmath96 . as @xmath27 ,
the bunch - davies choice is recovered , a counterpart of the remarkable results in flat space @xcite . in summary
, we have demonstrated that the fidelity loss of a teleportation protocol in de sitter space is unavoidable , either due to the gibbons - hawking radiation in scheme ( i ) , or the imperfection in entangling different cavities in scheme ( ii ) . with a cutoff at planck scale ,
the possible planckian physics cause some extra modification on the fidelity in both schemes of our teleportation protocol . since the important role of planck - scale physics in early universe
, we hope that those discussions on the rqi process in cosmological background @xcite may shed some light on our understanding of quantum fluctuation decoherence during / after inflation .
even it takes a certain amount of foolhardiness to do that directly , we can at least simulate these planckian modifications by analogue gravity experiments , like using ion trap .
for instance , in detector picture @xcite , to simulate planckian modification , the detector s response function represents the probability for an ion excited by interaction with the field . the ion analogue of the wightman function @xmath97 should be evaluated in some motional - state@xcite , since the @xmath19-vacua can be interpreted as squeezed states over the bunch - davies vacuum state . |
over the last decades observational and numerical evidence has substantiated the picture of a filamentary large - scale structure in the universe . in principle the large - scale tidal field is expected to induce large - scale correlations between the orientations of halos and galaxies that are embedded within these filaments ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ?
* ; * ? ? ?
on the other hand , the subsequent accretion onto larger systems , such as groups and clusters of galaxies , may alter the orientations of these ( sub-)structures in response to the local tidal field @xcite .
cosmological n - body simulations provide a valuable tool to differentiate the various contributions to the halo / galaxy alignments within overdense regions .
observationally , various types of alignment between galaxies and their environment have been detected on a wide range in scales , from super - cluster systems down to the distribution of the satellite galaxies in our milky way ( mw ) .
on cluster scales various different types of alignment are discussed in the literature : alignment between neighboring clusters , between brightest cluster galaxies ( bcgs ) and their parent clusters , between the orientation of satellite galaxies and the orientation of the cluster @xcite , and between the orientation of satellite galaxies and the orientation of the bcg @xcite . according to these studies the typical scales over which clusters reveal signs for alignment range up to @xmath6 ,
which can be most naturally explained by the presence of filaments . with large galaxy redshift surveys , such as the two - degree field galaxy redshift survey ( 2dfgrs ,
* ) and the sloan digital sky survey ( sdss , * ? ? ?
* ) , it has recently also become possible to investigate alignments on group scales using large and homogeneous samples .
this has resulted in robust detections of various alignments : @xcite , @xcite and @xcite all found that satellite galaxies are preferentially distributed along the major axes of their host galaxies , while @xcite and @xcite noticed that satellite galaxies tend to be oriented towards the galaxy at the center of the halo . in contradiction to the studies above , @xcite found that satellites around isolated late type galaxies preferentially lie along the minor axis of the disc .
subsequent studies , however , were unable to confirm this so - called ` holmberg effect'@xcite .
recently @xcite reported a holmberg effect at large projected distances around blue host galaxies , while on smaller scales the satellites were found to be aligned with the major axis of their host galaxy and @xcite claim that a careful selection of isolated late - type galaxies reveals the the tendency for the satellites to align with the minor axis of the galactic disc .
investigating the companions of m31 @xcite find little evidence for a holmberg effect . yet , the milky way ( mw ) seems to exhibit a holmberg effect even on small scales , in that the 11 innermost mw satellites show a pronounced planar distribution oriented close to perpendicular to the mw disc .
numerical simulations have been employed to test alignment on a similar range in scales , from super - clusters down to galaxy - satellite systems .
all studies focusing on cluster size halos report a correlation in the orientations for distances of at least @xmath7 ; some studies observe a positive alignment signal up to @xmath8 .
these findings are interpreted as the signature of the filamentary network which interconnects the clusters .
the preferential accretion along these filaments causes the clusters to point towards each other .
also , for galaxy and group - sized halos a tendency to point toward neighboring halos is detected . according to @xcite
the alignments for such intermediate mass objects are caused by tidal fields rather than accretion along the filaments .
consequently , the mechanisms responsible for the alignment of the orientations depend on halo mass .
further evidence for a mass dependence of alignment effects comes from the examination of the halos angular momenta . @xcite and @xcite
find that the spins of galaxy size halos tend to be parallel to the filaments whereas the spins of group - sized halos tend to be perpendicular .
this behavior may originate in the relative sizes of halos with respect to the surrounding filaments .
on subhalo scales basically three different types of alignments have been discussed : the alignment of the overall subhalo distribution with the orientation of the host halo ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the alignment of the orientations of subhalos among each other ( e.g. , * ? ? ?
* ) and , very recently , the orientation of the satellites with respect to the center of the host @xcite .
again , accretion along the filaments and the impact of tidal fields have been invoked as explanations for the former and the latter , respectively .
thus , on all scales tidal fields and accretion along filaments are considered to be the main contributers to the observed alignment signals .
here we attempt to isolate the different contributions .
in particular we focus on the continuous transition from subhalo to halo scales meaning we examine the alignment of ( sub)structure on distance scales between 0.3 and 6 times the virial of groups sized halos .
faltenbacher et al .
( 2007 , hereafter paper i ) applied the halo - based group finder of @xcite to the sdss data release four ( dr4 ; * ? ? ?
* ) and carried out a study of the mutual alignments between central galaxies ( bcg ) and their satellites in group - sized halos . using the same data set consisting of over @xmath9 galaxies three different types of alignment have been investigated : ( 1 ) the ` halo ' alignment between the orientations of the bcg and associated satellite distribution ; ( 2 ) the ` radial ' alignment between the direction given by the bcg - satellite connection line and the satellite orientation ; ( 3 ) the ` direct ' alignment between the orientations of the bcg and the satellites .
the study presented in this paper focuses on the same types of alignment and is aimed to compare the observational results with theoretical expectations derived from n - body simulations .
there are a variety of dynamical processes which can contribute to the alignments of satellites associated with groups , the most important are : ( 1 ) a possible pre - adjustment of satellites in the filaments , which for distances of a few times the virial radius commonly point radially towards the group ; ( 2 ) the preferential accretion along those filaments ; ( 3 ) the change of the satellite orbits in the triaxial group potential well ; ( 4 ) the continuous re - adjustment of satellite orientations as they orbit within the group .
basically , the first two points can be attributed to the large scale environment of the groups whereas the latter two are more closely associated with the impact of the group potential on small scales .
the purpose of the present analysis is to separate the different contributions to the observed alignment signals , therefore we analyse the mutual orientations of satellites within 6 times the virial radius of the groups . since the tidal forces are closely related to the dynamics of the satellites additional insight into the generation of alignment can be gained by considering the satellite velocities .
therefore , we also investigate the direction of the satellite velocities with respect to their orientations , which constitutes an indirect way to infer the impact of the dynamics onto the orientation of the satellites .
a more direct way to work out the interplay between the dynamics and the orientations would be to trace the orbits of individual satellites , however such an approach goes beyond the scope of the present study .
the paper is organized as follows . in [ sec : sim ] we introduce the simulation and describe the halo finding procedure .
[ sec : size ] deals with some technical aspects , namely the determination of the size and orientation of the substructures . in [ sec:3d ]
we present the signals of the three dimensional spatial and velocity alignments and in [ sec:2d ]
we repeat the analysis based on projected data .
finally , we conclude with a summary in [ sec : diss ] .
for the present analysis we employ an n - body simulation of structure formation in a flat @xmath10cdm universe with a matter density @xmath11 , a hubble parameter @xmath12 , and a harrison - zeldovich initial power spectrum with normalization @xmath13 .
the density field is sampled by @xmath14 particles within a @xmath8 cube resulting in a mass resolution of @xmath15 .
the softening length was set to @xmath16 , beyond which the gravitational force between two particles is exactly newtonian .
the density filed is evolved with 5000 time steps from an initial redshift of @xmath17 using a pppm method .
an extensive description of the simulation can be found in @xcite where it is quoted as lcdma realization . as detailed in the following two paragraphs
the host halos and its satellites are found in two subsequent steps with two different techniques , first the main halos are located thereafter the associated satellite halos are detected . in order to identify the host halos we first run a fof algorithm @xcite on the simulation output at @xmath18 .
we set the fof linking length to 0.1 times the mean particle separation , which selects regions with an average overdensity of @xmath19 .
note that , this linking length is a factor of two smaller than the commonly used value of 0.2 , consequently only the central part of the host halo ( and occasionally large substructures ) are selected .
subsequently , the virial radius , @xmath0 , is defined as the radius of the sphere centered on the most bound fof particle which includes a mean density of 101 times the critical density , and we simply define the virial mass of each halo as the mass within @xmath0 . if the virial regions of two halos overlap , the lower mass halo is discarded . in what follows
we only focus on the 515 halos with a virial mass in the range from @xmath20 to @xmath21 ( corresponding to halos with more than 16,000 particles ) .
since this is the typical mass scale of galaxy groups , we will refer to these halos as ` groups ' . in a second step we search for
self - bound ( sub)structures using the skid halo finder @xcite applied to the particle distribution within group centric distances of @xmath1 . as discussed in @xcite
skid adequately identifies the smallest resolvable substructures when using a linking length @xmath22 equal to twice the softening length , i.e. four times the spline softening length .
we therefore adopt @xmath23 . throughout
we will distinguish between `` group central substructures '' ( gcss ) , which are located at the center of our groups , and satellites which are all the other ( sub)structures , no matter whether they lie within or beyond @xmath0 . according to this definition every group hosts one , and only one , gcs at its center while it may have numerous satellites outside the volume occupied by the gcs .
satellites are allocated to all groups from which they are separated less than @xmath1 .
hence , a satellite may be assigned to more than one gcs .
before describing the computation of the orientation we determine the typical sizes of the gcss and the satellites .
knowledge about the physical sizes of the ( sub)structures provides a crucial link for the comparison to observational data .
the physical interpretation of the size of the gcs is not straightforward .
for one thing , it depends on the skid linking length used .
however , for our purposes it is sufficient to notice that the gcs represents the dense inner region of the group which , largely due to numerical reasons , is free of substructure .
consequently , any radial dependence of satellite properties can only be probed down to the size of the gcs . in order to express the sizes of the gcs and the satellites we use the rms of the distances between the bound particles , @xmath24 .
the advantage of this size measure is that it provides a direct estimate of the ( momentary ) size without having to make any assumption regarding the actual density distribution . in the case of an isolated nfw halo @xmath25 , with only a very weak dependence on the concentration parameter .
figure [ fig : covi ] displays the @xmath26 distribution of the gcss in units of the group s virial radius , @xmath0 . the distribution peaks at @xmath27 and has a mean of @xmath28 . the aim of the present analysis is twofold : ( 1 ) to assess the impact of the group tidal field on the satellite orientations , and ( 2 ) to compare the alignment signals in our n - body simulation to observations of galaxy alignments . the impact of the group tidal field is stronger at larger satellite - centric radii
on the other hand , since galaxies reside at the centers of their dark matter halos , the central parts of the satellites are more of interest when comparing the alignment signals with those observed for galaxies . to meet both requirements we therefore measure the orientation of the satellite mass distribution within two radii . in analogy to the measurement of gcs sizes ,
we determine these radii with reference to the spatial dispersion @xmath26 . more precisely , we choose the particles within 1.0 and @xmath29 as the basic sets for the subsequent determination of the satellite orientation ( see section [ sec : ori ] below ) . figure [ fig : rkpc ] displays the distributions of the corresponding physical sizes .
the @xmath29 sample probes the matter distribution of the satellites within @xmath30 , which is comparable to the sizes of elliptical galaxies .
the mean , physical radii of the @xmath31 sample is @xmath32 .
if not quoted otherwise we will display the results for the @xmath29 sample , since this may most closely resemble the properties of observable galaxy distributions ( outside of the very central part of the host halo ) .
there are a few different ways found in the literature ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ) to model halos as ellipsoids .
they all differ in details , but most methods model halos using the eigenvectors from some form of the inertia tensor .
the eigenvectors correspond to the direction of the major axes , and the eigenvalues to the lengths of the semi - major axes @xmath33 .
following @xcite we determine the main axes by iteratively computing the eigenvectors of the _ distance weighted _ inertia tensor .
@xmath34 where @xmath35 denotes the @xmath36th component of the position vector of the @xmath37th particle with respect to the center of mass and @xmath38 is the elliptical distance in the eigenvector coordinate system from the center to the @xmath37th particle
. the square roots of the eigenvalues of the inertia tensor determine the axial ratios of the halo ( @xmath39 ) .
the iteration is initialized by computing the eigenvalues of the inertia tensor for the spherically truncated halo . in the following iterations
the length of the intermediate axis is kept unchanged and all bound particles within the ellipsoidal window determined by the eigenvalues of the foregoing iteration are used for the computation of the new inertia tensor .
the iteration is completed when the eigenvectors have converged .
the direction of the resulting major axis is identified as the orientation .
the advantage of keeping the intermediate axis fixed is that the number of particles within the varying ellipsoidal windows remains almost constant . instead ,
if the longest ( shortest ) axis is kept constant the number of particles within the ellipsoidal windows can decrease ( increase ) substantially during the iteration .
note that we apply this truncation to all ( sub)structures , both satellites and gcss , and that the orientation of each sub(structure ) is measured within this truncation radius .
throughout we only consider those sub(structures ) that comprise at least 200 bound particles within the volume of the final ellipsoid ( corresponding to a lower limit in mass of @xmath40 ) .
for the satellites this implies that a smaller truncation radius results in a smaller sample .
for example , there are 772 @xmath29 satellites within the virial radii of our groups whereas the @xmath31 sample comprises 1431 satellites .
since all 515 gcss contain more than 200 particles within @xmath29 their sample size is independent of the truncation radius used . figure [ fig : shape ] displays the distribution of the shape parameter @xmath41 .
the shading corresponds to different truncation radii as listed .
there is a weak indication that satellites become more spherical with increasing truncation radii .
a similar behavior was found for isolated halos ( e.g. , * ? ? ?
* ; * ? ? ?
as discussed by @xcite the exact determination of individual shapes may need as many as 7000 particles , so that the resolution of the present simulation is not suited for the analysis of ( sub)structure shapes .
however , for the determination of the orientations , which is the focus of this paper , a particle limit of 200 can be considered conservative ( cf .
* ; * ? ? ?
a study examining the shapes of substructure in a single high - resolution milky way - sized halo can be found in @xcite .
for both classes of objects , gcss and satellites , the orientations are determined according to the approach described above . a third orientation - like quantity
is given by the direction of the line connecting a gcs - satellite pair . throughout
we refer to the orientation of the gcs , the satellite and the connecting line as @xmath42 , @xmath43 and @xmath44 , respectively .
these quantities are unit vectors , such that the scalar product of two vectors yields the cosine of the angle between them .
we will focus on three different types of alignment , ( 1 ) the _ halo alignment _ between the orientations of the gcss and the connecting lines , ( 2 ) the _ radial alignment _ between the orientations of the satellites and the connecting lines and ( 3 ) the _ direct alignment _ between the orientation of the gcs and that of its satellites .
in addition , we also consider various alignments based on the proper velocity , @xmath45 , of the satellite with respect to its gcs .
in particular , we discuss ( 4 ) the _ radial velocity alignment _ between @xmath45 and @xmath44 , ( 5 ) the _ halo velocity alignment _ between @xmath46 and @xmath45 , and finally ( vi ) the _ auto velocity alignment _ between the orientations , @xmath43 , and velocities , @xmath45 , of the satellites . here
@xmath45 is the unit vector indicating the direction of the _ proper _ velocity of the satellite ( including the hubble flow ) relative to the host .
since all the other quantities also represent unit vectors the scalar products yield the cosines of the enclosed angles . in order to measure the alignment between the gcs and the satellite distribution we use @xmath46 and @xmath44 ( the orientation of the gcs and the position of the satellite with respect to its gcs ) .
figure [ fig : a11 ] displays the radial dependence of @xmath47 within @xmath1 , where @xmath48 denotes the mean value within a bin of @xmath49 .
the error bars indicate the @xmath50 bootstrap confidence intervals based on 1000 bootstrap samples for each distance bin . over the entire range of distances probed , the mean values of the cosines deviate significantly from a isotropic distribution .
the strength of the alignment , i.e. the deviation from @xmath51 , increases with group centric distance and reaches a maximum at @xmath52 . the subsequent decline
, however , is very weak and even at @xmath1 the alignment is still very pronounced ( @xmath53 ) , with no clear indication of a downward trend .
the fact that there is strong alignment over such a long range suggests that the halo intrinsic alignment is closely connected to the filamentary structure in which the groups are embedded in .
since here we focus on the transition between group and environment dominated areas we do not aim to map out the entire range of the radial alignment .
the weakening of the signal at small scales may be attributed to the fact that the information about the filamentary origin is washed away once the satellites start to orbit within the groups ( i.e. , once non - linear effects kick in ) . yet
, the orientation of the group itself is closely correlated with the surrounding filamentary network , so that a residual alignment is maintained by the overall distribution of satellites orbiting in the potential well of the group ( cf .
* ; * ? ? ?
* ; * ? ? ?
additionally , if one assumes that filaments are approximately cylindrical in shape and the gcs is aligned with the orientation of the cylinder , then the mean angles between the orientation of the gcs and the satellites position become larger at smaller group - centric radii .
in fact , at distances smaller than the radius of the cylinder the distribution will converge to isotropic .
finally , some contribution to the decrease of the alignment strength on small scales may come from the fact that satellites on nearly radial orbits are filtered out during their epicenter passage .
they get severely stripped and consequently the number of particles that remains bound can easily fall below the detection criterion ( minimum of 200 particles ) , thus weakening the alignment signal . the radial alignment , @xmath54 , probes the orientations of individual satellites , @xmath43 , relative to the direction pointing towards their gcs , @xmath44 .
figure [ fig : ax1 ] displays @xmath54 for distances up to @xmath1 .
the line styles represent different truncation radii of the satellites . over
the entire range of group - centric distances probed , the data reveal a significant anisotropic distribution .
the signal is most pronounced on small scales , where it also shows a strong dependence on the truncation radii . the @xmath31 sample , which includes the behavior of the outer mass shells of the satellites , clearly exhibits a stronger deviation from isotropy .
within @xmath55 there is a pronounced decline of the radial alignment signal , while it remains remarkably constant at larger radii . for distance in the range between @xmath56
we detect a weak but significant signal , @xmath57 , inconsistent with isotropy at 95% confidence level in good agreement with @xcite . in a recent study , @xcite detected no radial alignment for distances @xmath58 .
however , their analysis is based on a resimulation of a single galaxy - sized host halo . since
this halo is rather isolated , in that it has not experienced any major merger after redshift @xmath59 , it is likely that its ambient filaments have already been drained . at large distances satellites
preferentially reside in filaments ( as discussed in the context of figure [ fig : a11 ] ) which point radially towards the groups .
consequently , the signal on scales @xmath60 indicates an alignment between the satellite orientations and the filaments in which they are embedded .
such an alignment may be caused by accretion of matter along those filaments or by the local tidal fields generated by the mass distribution within the filaments .
the group tidal field is not likely to be responsible for the observed large scale alignment signal due to its rapid decline with distance .
on small scales , however , the tidal field can substantially alter the orientations of the satellites . as shown by @xcite
the time scale on which a prolate satellite can adjust its orientation to the tidal field of a group is much shorter than the hubble time , but longer than its intrinsic dynamical time .
therefore , the adjustment of the satellite orientations parallel to the gradients of the group potential offers a convincing explanation for a radial alignment signal on small scales .
this perception is further supported by the dependence of the alignment strength on the truncation radii of the satellites . for the largest radii , which are strongest affected by tidal forces , the alignment signal is strongest .
the strong signals for halo and radial alignment may lead to the expectation of a comparably pronounced signal for the direct alignment between the orientation of the gcs , @xmath46 , and the orientations of its satellites , @xmath43 .
however , as can be seen in figure [ fig : ac1 ] , the signal is weak .
there is only a weak trend for positive alignment up to @xmath4 .
the significance found at distances between @xmath61 and @xmath62 seems to be somewhat higher ( @xmath63 confidence ) .
based on an analytical model @xcite predict a certain degree of parallel alignment between host and satellite orientations due to the evolution of the satellites within the tidal shear field of host .
the signal for the direct alignment may be a relic of this effect . to summarize , we find positive alignment signals for all three types of alignment tested here .
however , they differ in strength and radial extent .
the halo alignment is the strongest and reaches far beyond the virial radii of the groups ( @xmath64 ) .
the radial alignment is most pronounced at small scales , where it reveals a strong dependence on the radial extent of the satellite over which its orientation has been measured .
although the radial alignment is weak beyond @xmath65 , the signal stays remarkably constant out to @xmath66 .
finally , the least prominent signal comes from the direct alignment .
this ranking of the alignment strengths is in good agreement with the observational results reported in paper i. if tidal forces give rise to the radial alignment on small scales , as displayed in fig .
[ fig : ax1 ] , the satellite orientations should be related to their actual velocities and the local gradients of the host potential .
for instance a satellite moving radially towards the gcs will show an enhanced radial alignment since the gradient of the potential and the actual velocity are pointing in the same direction inducing an orientation in radial direction . on the other side
the orientations of satellites moving perpendicular to the gravitational field ( i.e. tangentially with respect to the gcs ) will lie in between their velocities and the gradients of the potential well . to gain some more insight into the dynamical origin of the alignments ,
we include the directions of satellite velocities into the alignment study .
we will consider three different kinds of alignments : the radial velocity alignment , @xmath67 , the halo velocity alignment @xmath68 and the auto velocity alignment @xmath69 . to facilitate the interpretation of the velocity alignments
, we split the satellites according to whether their net motion is inward ( @xmath70 ) or outward ( @xmath71 ) with respect to their group .
figure [ fig : io ] shows the fraction of inward moving satellites , @xmath72 , as a function of their group centric distances .
note that @xmath72 reaches a maximum around @xmath73 , beyond which the hubble flow gradually starts to become more and more important .
in fact , at sufficiently large radii , where the hubble flow dominates , one expects that @xmath74 , and all satellites reveal an outward motion . for satellites that are in virial equilibrium within the group potential ( i.e. , at @xmath75 ) , one expects roughly equal numbers of inward and outward moving systems ( i.e. , @xmath76 ) .
however , on these small scales one has an additional contribution from the infall region around the group , causing @xmath77 . in addition , a substantial fraction of satellites get stripped below the detection limit ( 200 particles ) at their peri - centric passage , so that they no longer contribute to the signal on their outward motion ( cf .
* ) . at @xmath0 ,
the outgoing satellite fraction is about 40% , which is ( within the errors ) consistent with the value @xmath78 determined by @xcite .
if one assumes an average ratio of 6:1 between apo- and peri - center distances for typical satellite orbits @xcite the majority of these satellites must have passed the central parts of the group before ( cf .
the upper panel of figure [ fig : vel ] displays the radial velocity alignment , @xmath79 , as a function of @xmath49 .
@xmath80 indicates that the distribution of angles between @xmath44 and @xmath45 is not isotropic , instead , on average they preferentially point in radial directions .
this behavior is in agreement with earlier studies of the velocity anisotropy of subhalos which is usually expressed by the anisotropy parameter @xmath81 ( e.g. , * ? ? ?
* ) , where @xmath82 and @xmath83 denote the velocity dispersions of the satellites in the tangential and radial direction , respectively . note , @xmath79 is closely related to @xmath84 . if one assumes a relaxed ( steady - state ) halo the above mentioned tendency towards radial motions translates into a higher radial velocity dispersion compared to the tangential one @xmath85 ( where @xmath86 and @xmath82 are the one and two dimensional tangential velocity dispersions , respectively and tangential isotropy is assumed ) .
thus @xmath80 on small scales ( @xmath87 ) suggest that @xmath88 , in good qualitative agreement with numerical simulations which have shown that @xmath89 for subhalos within the virial radius of their hosts . @xcite . in accordance with the spherical collapse model the signal extends out to @xmath90 , which roughly reflects the distance of turnaround . at @xmath91
the distribution is close to isotropic suggesting that at these distances the impact of the group potential is negligible and the satellite motions are dominated by local potential variations arising from the filaments and dark matter halos within these filaments .
note that the presence of this filamentary structure in the vicinity of groups is clearly evident from figure [ fig : a11 ] . finally , the increase of the radial velocity alignment on large scales , @xmath92 , is simply due to the hubble flow ( i.e. , @xmath93 at @xmath94 ) .
the middle panel of figure [ fig : vel ] shows @xmath79 for the inward moving satellites only .
the radial trend within @xmath4 is somewhat enhanced compared to the upper panel . at larger radii , the inward moving satellites have a velocity structure that is consistent with isotropy .
the lower panel of figure [ fig : vel ] reveals a marked difference in the behavior of @xmath79 for the outward moving satellites .
it indicates a slightly radial trend for satellites within @xmath95 which is much lower than seen in the upper two panels . within @xmath96
it drops below 0.5 , indicating a preference for tangential velocities . together with the information derived from figure [ fig : io ]
this suggests that a substantial fraction of outward moving satellites located at @xmath96 currently are close to their apo - center passage after having crossed the more central regions of the group .
finally , on large scales the outward moving satellites clearly reveal the hubble flow .
figure [ fig : c1v ] displays the radial dependence of @xmath97 which measures the cosines of the angels between the satellite velocities and the orientation of the gcs . on large scales the radial outward motion caused by the hubble flow exceeds the internal velocities of the satellites within the filaments . since the gcs is strongly aligned with these filaments over the entire radial range shown ( cf .
figure [ fig : a11 ] ) , one has that @xmath98 on scales where the hubble flow becomes important ( @xmath92 ) .
the strong alignment signal on small scales indicates that the satellites tend to move parallel to the orientation of the gcs . according to @xcite and @xcite the principal axes of the velocity anisotropy tensor
are strongly correlated with the principal axes of the satellite distribution .
therefore , the alignment found for @xmath47 ( figure [ fig : a11 ] ) actually implies an analogous signal for @xmath97 .
however , in contrast to the halo alignment , @xmath47 , the velocity halo alignment , @xmath97 , only extends out to @xmath99 . beyond this radius
a substantial fraction of the satellites shows relatively large angles between their velocities and the orientation of the gcs which is consistent with the picture of tangential motions associated with the apo - center passage of the satellites , as discussed in the context of figure [ fig : vel ] .
finally we consider the auto velocity alignment , @xmath100 , which reflects the distribution of the cosines between the satellite velocities and their orientations , @xmath101 .
fig [ fig : a1v ] displays the variation of @xmath100 with @xmath49 .
the signal for @xmath100 shows a maximum at @xmath102 . at larger distances
it decreases quickly .
beyond @xmath5 it is roughly in agreement with an isotropic distribution .
a possible reason for the slight central dip is , that satellites on their peri - center passages move perpendicular to the gradients of the group potential .
figure [ fig : ax1 ] , however , revealed a preferential radial orientation of these satellites .
thus , during the peri - center passages the angles between satellite orientations and velocities can become large .
the degree of the radial alignment depends on the ratio between the internal dynamical time of the satellite , with which it can adjust its orientation , and the duration of the peri - center passage .
if the peri - center passage occurs too quickly the time may be too short for a ` perfect ' radial alignment ( cf .
* ) . on large scales ( @xmath96 )
a similar mechanism may take place .
above we have argued that within this distance range a substantial fraction of satellites are close to their apo - center passage . during this phase
the velocities are again perpendicular to the gradient of the potential but , as indicated by figure [ fig : ax1 ] , the satellites are oriented radially .
the comparison between the signal for @xmath54 and @xmath100 suggests that , in a statistical sense , the ( spatial ) radial alignment is maintained during the entire orbit of the satellite within the potential well of the groups , which in turn causes a suppression of @xmath100 , at its apo- and peri - center .
to facilitate a comparison with observations , in particular with the results presented in paper i , we repeat the foregoing analysis using projected data , i.e. we project the particle distribution along one of the coordinate axes and compute the second moment of mass for the projected particle distribution . accordingly , for the distances between gcs and satellites we use the projected values ( all satellites within a sphere of @xmath1 about the gcs are projected ) , which we label as @xmath103 ( the physical distances are labelled as @xmath104 ) .
since the projections along the three cartesian coordinate axes are independent we include all three projections of each host - satellite in our 2d sample . to reduce the contamination by satellites associated with massive ambient groups we exclude those host - satellite systems where another skid group more massive than the gcs ( which is most likely the center of an ambient host - satellite system )
is found within a sphere of @xmath1 .
after rejection of ` contaminated ' groups we obtain 1034 and 543 satellites for the @xmath31 and @xmath29 samples with 3d distances to the gcs @xmath105 ( for all groups irrespective of their environment we found 1431 and 772 , see [ sec : ori ] . ) furthermore , since ( due to technical reasons ) we project satellites located within a _ sphere _ of @xmath1 the projected volume at large projected distances shrinks substantially .
therefore , we analyze the 2d data only for projected distances @xmath106 which roughly resembles the projection of all satellites within a cylinder with a radius @xmath107 and length of @xmath108 along the ` line of sight ' .
thus , in an approximate manner , uncertainties in the determination of group membership based on redshift measurements are accounted for .
the resolution of the simulation does not permit to probe alignment below @xmath109 .
other authors ( using semi - analytical techniques , e.g. , * ? ? ?
* ) have bypassed this problem by introducing so - called _ orphan galaxies _ , i.e. galaxies which are associated with the once most bound particle of a satellite halo which subsequently has become undetectable due to the stripping by tidal forces .
here we do not adopt this technique since it does not provide us with information about the orientation of a satellite .
both approaches , considering only satellite halos with a minimum number of particles and the introduction of orphan galaxies , have certain disadvantages .
the former does not account for galaxies which are hosted by strongly stripped subhalos whereas the latter ignores the dynamical differences of galaxies and ( once most bound ) particles .
the application of a fixed lower particle limit excludes satellites from the analysis which still constitute distinct objects .
in particular satellites which are strongly tidally stripped may fall below the selection criterion even if the galaxy , which is assumed to sit at the center , may still be observable .
thus , we caution that our satellite sample may be somewhat biased toward more recently accreted satellites compared to a hypothetical galaxy population .
this effect appears whenever a fixed lower particle limit is imposed . in analogy to paper
i we define the angles @xmath110 , @xmath111 and @xmath112 to address the projected halo , radial and direct alignments ( same definitions as in
[ sec:3d ] but for the 2d data , see fig . [
fig : sketch ] ) and the projected orientations are referred to as position angles ( pas ) .
it is not straightforward to derive galaxy properties , such as luminosity and color , from the dark matter distribution . in particular ,
if the satellite halo hosts a late type galaxy , it is not obvious how to accurately determine the orientation of the disk ( but see e.g. , @xcite and @xcite for attempts ) . on the other side ,
if one focuses on early type galaxies the orientation of the central dark matter distribution is very likely correlated with the orientation of the stellar component ( see the evidence from gravitational lensing , e.g. , @xcite ) .
the lower particle limit for the satellites results in a lower mass of @xmath113 within @xmath114 .
assuming a dynamical mass - to - light ratio of a few @xcite within this radius yields a stellar component which roughly resembles @xmath115 galaxies .
therefore , our findings in the current paper may be best compared with results based on bright early - type satellite galaxies . however , as we have pointed out in paper i , our observational results were only marginally dependent on the luminosity / mass of satellite galaxies .
therefore , a comparison with observations based on somewhat fainter satellites is viable as well .
figure [ fig : d2a11 ] shows the results obtained for the angle @xmath110 between the orientation of the gcs and the line connecting the gcs with the satellite .
the short horizontal line on the left indicates the result for the innermost bin if only the satellites within @xmath95 are projected .
the sample shows @xmath116 for the entire distance range .
the error bars give the @xmath50 bootstrap confidence intervals for the mean angles within each bin .
the alignment strength within @xmath0 is @xmath117 , in good agreement with the findings of @xcite , @xcite . in paper
i we found a mean value @xmath118 within @xmath119 which is very close to the values we obtain for the innermost bin , in particular if only the satellites within @xmath95 ( short horizontal lines on the left ) are projected .
as also shown by @xcite the alignment signal extends beyond the virial radius .
the strongest amplitude is found outside the virial radius at @xmath120 .
currently there are no available observations covering the same distance range .
the analysis in paper i , for instance , is based on galaxies within the virial radius whereas we use all galaxies with projected distances @xmath106 .
according to our findings a search for alignment of satellite distribution in group environments for distances larger than @xmath0 may be a promising proposition .
figure [ fig : d2ax1 ] displays the mean angle @xmath111 between the pa of the satellite and the line connecting the satellite with its gcs . for all group
centric distances there is a clear and significant signal for the major axes of the satellites to point towards the gcs ( i.e. , @xmath121 ) .
the projection of only those satellites within @xmath95 increases the central signal by about @xmath122 ( differences between the innermost data points and the short horizontal lines ) .
the mean angle for the @xmath29 sample within the innermost bin is @xmath123 and according to paper i the mean value for the red sdss satellites within @xmath119 is very close to this value .
however , the observations suggest a significant alignment for red galaxies only out to @xmath102 whereas the n - body data indicate that radial alignment extends beyond @xmath107
. the discrepancy may be caused by the observational confinement to galaxies within the virial radius .
figure [ fig : d2ac1 ] displays the results for the direct alignment , based on the angle @xmath112 between the orientations of gcss and satellites . the alignment signal is significant at a @xmath124 confidence level for distances @xmath125 . in paper
i we obtained @xmath126 for red satellite with in @xmath119 which indicates a somewhat weaker alignment than we find here .
since the 3d analysis shows no increase of @xmath47 at small scales ( figure [ fig : a11 ] ) the central enhancement displayed here has to be interpreted as a result of projection effects .
in summary for all three types of alignments we find good agreement between numerical data presented here and the observational results from paper i. in particular the relative strength among the different alignments is well reproduced in the numerical analysis .
due to limited resolution the range below @xmath95 is only sparsely sampled thus no detailed information about the radial dependence of the alignment signal on small scales can be derived .
however , the signal for @xmath110 increases with distance which is only marginally implied by the sdss results presented in paper i. also for @xmath111 , the dependence on the distance disagrees between simulations and observations .
it is currently unclear whether this is due to shortcomings from the numerical or observational side .
based on a sample of 515 groups with masses ranging from @xmath127 to @xmath21 we have investigated the halo alignment , @xmath47 , the radial alignment , @xmath54 and the direct alignment @xmath128 , between the central region of each group ( the gcs ) and its satellite halos out to a distance of @xmath1 .
here @xmath46 , @xmath43 and @xmath44 denote the unit vectors associated with the orientation of the gcs , the satellites and the line connecting both of them .
additionally , we have employed the directions of the satellite velocities @xmath45 to probe the alignments @xmath79 , @xmath97 and @xmath100 , referred to as radial , halo and auto velocity alignments , respectively .
our main results are : * halo , radial and direct alignment differ in strength .
the halo alignment is strongest followed by the radial alignment . by far the weakest and least significant signal comes from the direct alignment .
this sequence is found in the 3d analysis as well as for the projected data and agrees well with our recent analysis of galaxy alignments in the sdss ( cf . , paper i ) . *
the signal for the halo alignment , @xmath47 , reaches far beyond the virial radii of the groups ( @xmath129 ) which we interpret as evidence for large scale filamentary structure . *
the signal for the radial alignment , @xmath54 , is largest on small scales .
after a rapid decline with distance it flattens , such that a relatively small @xmath130 , but significant deviation from isotropy is detected out to @xmath2 .
whereas the small scale signal more likely owes to the group s tidal field , the weak but significant signal on large scales suggests that satellites tend to be oriented along the filaments in which they reside . *
the 3d signal for the direct alignment , @xmath128 , shows a weak trend for parallel orientations on scales @xmath3 .
the projected data indicate an increasing signal for distances @xmath125 which is likely caused by projection effects . *
all kinetic alignment signals are highly significant at small scales .
the signal for @xmath79 is basically constant within @xmath131 , beyond which it rapidly drops . in the subset of outward moving satellites
we find a tendency for tangential motions which can be attributed to the satellites which have been accreted earlier and are currently passing their peri- or apo - centers .
the signal for @xmath97 is maximal at the center , drops rapidly with distance and disappears at @xmath95 .
finally , @xmath100 shows a slight dip at the center , reaches a maximum at @xmath102 , and becomes consistent with isotropy at @xmath5 .
all these features support the interpretations advocated for the spatial alignments .
the simulation analyzed here clearly demonstrates that tidal forces cause a variety of alignments among neighboring , non - linear structures . on large scales ,
the tidal forces are responsible for creating a filamentary network , which gives rise to a halo alignment out to at least @xmath1 .
the same tidal forces also cause an alignment between filaments and ( sub)structures within the filaments ( cf .
* ; * ? ? ?
* ) which in turn results in a large scale radial alignment with the virialized structures at the nodes of the cosmic web . within these virialized structures ,
tidal forces are responsible for a radial alignment of its substructures , similar to the tidal locking mechanism that affects the earth - moon system .
this is further supported by the fact that the auto velocity alignment @xmath100 reveals a dip on small scale , indicating that at peri - centric passage satellites tend to be oriented perpendicular to the direction of their motion ( cf .
* ) . this behavior also explains , why the direct spatial alignment , @xmath128 , is so weak . a possible direct alignment originating from the co - evolution of group and satellites , as proposed by @xcite , is quickly erased as the satellites orbit in the potential well of the group . for future work it will be instructive to trace the orbits of individual satellites and consider more closely how their shapes and orientations evolve with time . the infall regions around virialized dark matter halos cause a radial velocity alignment out to @xmath73 , and an enhancement of inward moving ( sub)structures . at around the same scale , the ( sub)structures with a net outward movement have a tendency to move tangentially .
this most likely reflects the apo - centric passage of substructures that have previously fallen through the virialized halo . within a virialized region , the orientation of orbits
is naturally aligned with that of its gcs . since ( sub)structures reveal at most a weak velocity bias with respect to dark matter particles ( e.g. , * ? ? ?
* ) , this causes a strong halo velocity alignment on scales @xmath132 .
the halo velocity alignment is also strong on large scales ( @xmath133 ) , which reflects the hubble flow combined with the filamentary , non - isotropic distribution of ( sub)structures on these scales . a one - to - one comparison between the n - body results discussed here and the observations presented in paper i is not straightforward .
although we have employed the same mass range for the groups in both studies the resolution of the current simulation only allows to resolve satellites which are expected to host @xmath134 galaxies .
these are bright compared to our sdss sample for which a lower magnitude limit of @xmath135 has been adopted . nevertheless , the qualitative agreement between the relative strengths of the different types of spatial alignment is promising .
supplementary to the observational results of paper i we find a strong halo alignment and a somewhat weaker radial alignment out to at least @xmath1 which we will investigate further .
finally , the weak but significant detection of radial alignment out to @xmath1 may contaminate the cosmic shear measurements on these scales .
this correlation has to be considered , either by simply removing or down - weighting pairs of galaxies within this distance range
. this may be particularly important for applications of weak gravitational lensing for the purposes of precision cosmology .
this work has is supported by nsfc ( 10533030 , 0742961001 , 0742951001 , 973 program no .
2007cb815402 ) and the knowledge innovation program of the chinese academy of sciences , grant no .
kjcx2-yw - t05 .
the cas research fellowship for international young researchers ( af ) , the local support of the chinese academy of sciences ( hjm and sm ) and the alexander von humboldt foundation ( sm ) is gratefully acknowledged .
hjm would like to acknowledge the support of nsf atp-0607535 , nasa aisr-126270 , and nsf iis-0611948 . ,
m. , bacon , r. , bureau , m. , damen , m. c. , davies , r. l. , de zeeuw , p. t. , emsellem , e. , falcn - barroso , j. , krajnovi , d. , kuntschner , h. , mcdermid , r. m. , peletier , r. f. , sarzi , m. , van den bosch , r. c. e. , & van de ven , g. 2006 , , 366 , 1126 , m. , dalton , g. , maddox , s. , sutherland , w. , norberg , p. , cole , s. , bland - hawthorn , j. , bridges , t. , cannon , r. , collins , c. , couch , w. , cross , n. , deeley , k. , de propris , r. , driver , s. p. , efstathiou , g. , ellis , r. s. , frenk , c. s. , glazebrook , k. , jackson , c. , lahav , o. , lewis , i. , lumsden , s. , madgwick , d. , peacock , j. a. , peterson , b. a. , price , i. , seaborne , m. , & taylor , k. 2001 , , 328 , 1039 | based on a cosmological n - body simulation we analyze spatial and kinematic alignments of satellite halos within six times the virial radius of group size host halos ( @xmath0 ) .
we measure three different types of spatial alignment : halo alignment between the orientation of the group central substructure ( gcs ) and the distribution of its satellites , radial alignment between the orientation of a satellite and the direction towards its gcs , and direct alignment between the orientation of the gcs and that of its satellites .
in analogy we use the directions of satellite velocities and probe three further types of alignment : the radial velocity alignment between the satellite velocity and connecting line between satellite and gcs , the halo velocity alignment between the orientation of the gcs and satellite velocities and the auto velocity alignment between the satellites orientations and their velocities .
we find that satellites are preferentially located along the major axis of the gcs within at least @xmath1 ( the range probed here ) .
furthermore , satellites preferentially point towards the gcs .
the most pronounced signal is detected on small scales but a detectable signal extends out to @xmath2 .
the direct alignment signal is weaker , however a systematic trend is visible at distances @xmath3 .
all velocity alignments are highly significant on small scales .
the halo velocity alignment is constant within @xmath4 and declines rapidly beyond .
the halo and the auto velocity alignments are maximal at small scales and disappear beyond 1 and @xmath5 respectively .
our results suggest that the halo alignment reflects the filamentary large scale structure which extends far beyond the virial radii of the groups .
in contrast , the main contribution to the radial alignment arises from the adjustment of the satellite orientations in the group tidal field .
the projected data reveal good agreement with recent results derived from large galaxy surveys . |
accurate abundances in hii regions can be derived via the classic t@xmath6 method , t@xmath6 being the electron temperature of the hii region . in this method
, the reliability of the equations for determining the o@xmath7/h@xmath8 and o@xmath9/h@xmath8 ionic oxygen abundances depends to a great part on the reliability of the atomic data . @xcite
have compared the ionic oxygen abundances derived with a recent set of equations using the latest atomic data ( @xcite ) , with those determined from an earlier set of equations ( @xcite ) and have found close agreement .
thus , the equations for the determination of the ionic oxygen abundances in hii regions appear to be robust , in the sense that , for a given set of measured electron temperatures t@xmath4 and t@xmath3 , relations determined by different authors result in close oxygen abundances in hii regions . the electron temperatures @xmath0 within the [ oiii ] zone and @xmath1 within the [ oii ] zone are determined from diagnostic line ratios when they are available .
the intensity ratio of the [ oiii]@xmath10 nebular line to the [ oiii]@xmath11 auroral line is used to determine t@xmath4 . as for t@xmath3 , the intensity ratio of the [ oii]@xmath12 nebular line to the [ oii]@xmath13 auroral line is used .
based on a sample of hii regions where all the necessary oxygen lines are detected , @xcite found that the t@xmath3 temperatures show a large scatter and that they are nearly uncorrelated with t@xmath4 temperatures .
they noted that the measurements of the faint [ oii]@xmath13 auroral lines may contain large random errors .
furthermore , @xcite have pointed out that the [ oii]@xmath12/[oii]@xmath13 ratio may be affected by recombination .
@xcite have derived t@xmath3 and t@xmath4 from the above ratios for a sample of hii regions .
they concluded that the two temperatures generally follow the t@xmath3
t@xmath4 relation obtained from photoionization models ( e.g. stasinska 1982 , 1990 ) , but that the scatter of the data points is very large .
the large scatter was attributed to large flux errors of the weak [ oii]@xmath13 emission lines .
the intensity ratio of the [ nii]@xmath14 nebular line to the [ nii]@xmath15 auroral line is also used to determine t@xmath3 .
it seems to give more reliable values .
however , the contribution of recombination to the excitation of the [ nii]@xmath15 line may affect temperatures derived from the [ nii ] nebular to auroral line ratios @xcite . the ( [ nii]@xmath14)/[nii]@xmath15 ratio is measured in only a few hii regions with a detected ( [ oiii]@xmath10)/[oiii]@xmath11 line ratio . when only a single temperature measurement is available , a t@xmath3 t@xmath4 relation based on grids of hii region models is usually used .
several versions of such a @xmath1
@xmath0 relation have been proposed ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but the agreement between them is not very good .
the available measurements do not provide an undisputable evidences in favour of any out of suggested relations , i.e. the choice of the relation is in fact arbitrary .
thus , the @xmath1
@xmath0 relation seems to be the weakest link of the classic t@xmath6 method .
we will address this problem here .
the main goal of this paper is to derive a t@xmath3
t@xmath4 relation which is model - independent .
this allows one to relax the arbitrarity in the choice of the @xmath1
@xmath0 relation .
the basic idea is the following .
usually , the set of equations used for the determination of the ionic o@xmath7/h@xmath8 and o@xmath9/h@xmath8 abundances in hii regions is applied to the whole nebula .
if we apply the equation for o@xmath7/h@xmath8 only to the o@xmath7 zone , then this would yield , not the ionic o@xmath7/h@xmath8 abundance , but the total o / h oxygen abundance instead .
there is no reason to suspect that the oxygen abundances in the o@xmath7 and o@xmath9 zones differ .
we therefore require that the equation for o@xmath7/h@xmath8 applied to the o@xmath7 zone and the one for o@xmath9/h@xmath8 applied to the o@xmath9 zone result in exactly the same value of the oxygen abundance .
this condition allows us to derive a relation between t@xmath3 and t@xmath4 .
we can apply the equation for o@xmath7/h@xmath8 to the o@xmath7 zone alone , only if we know the contribution of the o@xmath7 zone to the measured h@xmath16 flux of the hii region .
we will be using the ff relation @xcite for this purpose . the observational sample to be used
is described in section 2 . the strategy for determining the t@xmath3
t@xmath4 relation is discussed in section 3 .
a model - independent t@xmath3
t@xmath4 relation is derived in section 4 .
we discuss the results in section 5 , and summarize our conclusions in section 6 .
we will be using the following notations throughout the paper : r@xmath3 = i@xmath17 \lambda 3727 + \lambda 3729}}$]/i@xmath18 , r@xmath4 = i@xmath19 \lambda 4959 + \lambda 5007}}$]/i@xmath18 , r = i@xmath19 \lambda 4363}}$]/i@xmath18 , r@xmath20 = r@xmath3 + r@xmath4 . with these definitions ,
the excitation parameter p can be expressed as : p = r@xmath4/(r@xmath3+r@xmath4 ) .
a large sample of high - precision measurements of hii regions is at the base of the present investigation .
we include first the sample of @xcite who have carried out an extensive search of the literature to compile a list of more than 700 individual spectra of hii regions in irregular and spiral galaxies , with the requirement that they all possess a detected [ oiii ] @xmath11 emission line .
while we have tried to include as many sources as possible , we do not claim our search to be exhaustive .
since the majority of extragalactic hii regions are in the low - density regime @xcite , only such hii regions will be considered here . of the objects with the electron density - sensitive ratio r@xmath21 = [ sii ] @xmath22/[sii ] @xmath23 available , those with r@xmath21 @xmath24 1.3 were excluded .
we have also included the sample of @xcite who have extracted from the data release 3 of the sloan digital sky survey ( sdss ) around 4500 spectra of hii regions with an [ oiii ] @xmath11 emission line detected at a level better than 1@xmath25 , and carefully measured the line intensities in each spectrum .
the sdss hii regions are also in a low - density regime .
yuri izotov and natalia guseva have kindly provided us with the total list of their measurements , as only part of these have been published @xcite . in total ,
our sample consists of around 5200 hii region spectra . following @xcite
, we can use the the ff relation @xcite to select out the hii regions with the highest quality measurements .
we can define a `` discrepancy index '' , equal to the difference between the logarithm of the observed flux r@xmath26 in the [ oiii]@xmath54363 line and the logarithm of the flux r@xmath27 of that line derived from the strong [ oii]@xmath53727 , [ oiii]@xmath5@xmath54959,5007 lines using the ff relation : @xmath28 it is well known that the relation between the oxygen abundance and the strong oxygen line intensities is double - valued , with two distincts parts , traditionally known as the upper high - metallicity branch and the lower low - metallicity branch of the r@xmath20 o / h diagram . following @xcite , we adopt the value of 12+log(o / h ) = 8.25 as the boundary between the upper branch and the transition zone .
the exact boundary is difficult to establish , but we have chosen that value because we will be using later the ff relation which is applicable at metallicities above 12+log(o / h ) @xmath29 8.25 . we have extracted from our total sample a subsample of high - metallicity hii regions with an oxygen abundance 12+log(o / h ) @xmath2 8.25 .
the oxygen abundances have been calculated with the equations of @xcite for the t@xmath6 method , and with the t@xmath3
t@xmath4 relation of @xcite .
[ figure : gist ] shows the cumulative number of individual hii region measurements with the absolute value of the discrepancy index d@xmath30 less than a certain value .
solid and open circles show data for hii regions with respectively positive and negative values of the discrepancy index .
it is seen that the agreement between the cumulative numbers of measurements with positive and negative values of the discrepancy index d@xmath30 is good for values of d@xmath30 less than 0.05 , but gets increasingly worse at higher values .
this is the result of a selection effect .
indeed , a large positive value of d@xmath30 means that the observed r@xmath26 is significantly overestimated .
this leads to a substantial overestimate of the electron temperature and an underestimate of the oxygen abundance , bringing it lower than our metallicity cut - off of 12+log(o / h ) = 8.25 .
this effect thus systematically depletes the number of high - metallicity hii regions with a large positive d@xmath30 . to free ourselves from this systematic effect , we select only measurements with the absolute value of d@xmath30 less than 0.05 .
this sample contains 372 data points and will be referred to as the standard sample .
the hii regions in it have line intensity errors that are random .
for abundance determination in an hii region , a two - zone model for its temperature structure is usually adopted .
@xcite have recently published a set of equations for the determination of the oxygen abundance in hii regions in the context of such a two - zone model . according to those authors , the electron temperature @xmath0 within the [ oiii ] zone , in units of 10@xmath32k ,
is given by the following equation @xmath33 the quantity c@xmath34 is defined by : @xmath35 where @xmath36 and @xmath37 as for the ionic oxygen abundances , they are given by the following equations : @xmath38 \lambda 4959 + \lambda 5007}/i_{h_{\beta } } ) + \nonumber \\ 6.200 + \frac{1.251}{t_3 } - 0.55 \log t_3 - 0.014\,t_3 , \label{equation : o3}\end{aligned}\ ] ] and @xmath39 \lambda 3727 + \lambda 3729}/i _ { h_{\beta } } ) + \nonumber \\ 5.961 + \frac{1.676}{t_2 } - 0.40 \log t_2 -0.034\,t_2 + \log ( 1 + 1.35x_2 ) , \label{equation : o2}\end{aligned}\ ] ] where @xmath40 here @xmath41 is the electron density in @xmath42 .
the total oxygen abundances are then derived from the following equation : @xmath43 the electron temperature @xmath1 of the [ oii ] zone is usually determined from an equation which relates @xmath1 to @xmath0 , derived by fitting hii region models .
several versions of this @xmath1
@xmath0 relation have been proposed .
a widely used relation is the one by @xcite ( see also @xcite ) based on the hii region models of @xcite : @xmath44 another relation has been proposed by @xcite , also based on hii region models of @xcite ) @xmath45 @xcite , fitting also the hii region models of @xcite , have proposed the following expression @xmath46 based on hii region model calculations by @xcite , @xcite have suggested the following relation @xmath47 @xcite have found that the campbell et al .
relation is reasonable for @xmath48 @xmath2 1.0 .
however at lower temperatures , the models are more consistent with an isothermal nebula .
they consequently adopted the formulation , @xmath49 using the system of equations eqs.([equation : t3])@xmath50([equation : otot ] ) and one of the t@xmath3 t@xmath4 relations chosen among eqs.([equation : ttcam])@xmath50([equation : ttoey ] ) ) , the oxygen abundance ( o / h)@xmath51 in an hii region can then be determined .
examination of eq.([equation : o3 ] ) shows that the fluxes used to calculate r@xmath4 and in the h@xmath16 line originate from different volumes of the nebula .
the h@xmath16 flux comes from the whole hii region while the r@xmath4 fluxes come from the volume where the oxygen is in the o@xmath7 stage .
if eq.([equation : o3 ] ) is applied to the o@xmath7 zone only , then it gives the total @xmath52 instead of the ionic @xmath53 oxygen abundance as in the traditional approach . in that case , eq.([equation : o3 ] ) takes the form @xmath54 \lambda 4959 + \lambda 5007}/ ( w \times i_{h_{\beta } } ) ) \nonumber \\ + 6.200 + \frac{1.251}{t_3 } - 0.55 \log t_3 - 0.014\,t_3 , \label{equation : o3w1}\end{aligned}\ ] ] where _ w _ is the fraction of the h@xmath16 flux in the o@xmath7 zone . with the adopted notations , eq.([equation : o3w1 ] ) can be rewritten as @xmath55 similarly , eq.([equation : o2 ] ) can be rewritten as @xmath56 \lambda 3727 + \lambda 3729}/ [ ( 1-w ) \times i _ { h_{\beta } } ] \nonumber \\ + 5.961 + \frac{1.676}{t_2 } - 0.40 \log t_2 -0.034\,t_2 \nonumber
\\ + \log ( 1 + 1.35x_2 ) .
\label{equation : o2w1}\end{aligned}\ ] ] or @xmath57 + 5.961 + \frac{1.676}{t_2 } \nonumber \\ - 0.40 \log t_2 -0.034\,t_2 + \log ( 1 + 1.35x_2 ) .
\label{equation : o2w}\end{aligned}\ ] ] there is no reason to expect the oxygen abundance in the o@xmath7 zone to differ from that in the o@xmath9 zone within the same hii region .
in other words , the oxygen abundance derived from eq.([equation : o3w ] ) must be equal to the one derived from eq.([equation : o2w ] ) . equating the right - hand sides of these two equations result in @xmath58 + 5.961 + \frac{1.676}{t_2 } - 0.40 \log t_2 \nonumber \\ - 0.034\,t_2 + \log ( 1 + 1.35x_2 ) .
\label{equation : bas}\end{aligned}\ ] ] eq.([equation : bas ] ) can be solved for t@xmath59 if the values of t@xmath4 , n@xmath6 and _ w _ are known .
the majority of extragalactic hii regions are in the low - density regime @xcite .
therefore we adopt an electron density n@xmath6 = 100 @xmath42 for all the hii regions in our sample .
thus , eq.([equation : bas ] ) allows to derive the electron temperature @xmath1 within the [ oii ] zone in a hii region if the electron temperature @xmath0 within the [ oiii ] zone and the fraction _ w _ of h@xmath16 flux in the o@xmath7 zone are known .
the fraction _ w _ of h@xmath16 flux in the o@xmath7 zone can be estimated in the following way .
we have investigated recently the relationship between the observed auroral and nebular oxygen line fluxes in spectra of hii regions @xcite .
we have found a relation ( called the ff relation ) that is metallicity - dependent at low metallicities , but independent of metallicity above 12+logo / h @xmath29 8.25 , i.e. there is one - to - one correspondence between the auroral and nebular oxygen line fluxes in spectra of high - metallicity hii regions . using a compilation of recent high - precision measurements of oxygen lines fluxes in high - metallicity hii regions ,
the following ff relation was derived @xcite @xmath60 fig .
[ figure : f - f ] shows the relation corresponding to eq.([equation : ffp ] ) in the r@xmath4 r diagram for various values of the excitation parameter p. the solid line , called hereafter as the basic line , corresponds to p = 1.0 , while the dotted lines correspond , from right to left , to p = 0.8 , 0.6 , 0.4 , 0.2 .
the dashed lines show the loci of hii regions for several values of the electron temperature t@xmath4 . from top to bottom , t@xmath4 = 1.2 , 1.1 , 1.0 , 0.9 , 0.8 , 0.7 in units of 10@xmath32 k. the circles show individual measurements of hii regions from the present sample .
let us consider a sequence of hii regions with the same value of t@xmath4 .
the fraction of radiation of the hii region in the o@xmath7 zone increases along this sequence from low to high values of p , and reaches its maximum value , _ w _ = 1 , at p = 1 , i.e. at the intersection point of a dashed line with a given electron temperature with the basic solid line .
then , the fraction of radiation in the o@xmath7 zone can be estimated as @xmath61 such an estimate is not airtight since not only _ w _ can change along a sequence of hii regions with a fixed value of t@xmath4 , but other characteristics of hii regions , such as , for example , their oxygen abundance , can do so as well .
therefore , the value of t@xmath62 derived from eq.([equation : bas ] ) and eq.([equation : w ] ) may carry some systematic error _ e_@xmath63 .
that error should be small for high - excitation hii regions ( they are located close to the basic line ) but increase with decreasing p. in the following , we discuss how to estimate that error and correct t@xmath62 for it .
the top panel in fig .
[ figure : t2-t3 ] shows the value of t@xmath62 as derived from eq.([equation : bas ] ) and eq.([equation : w ] ) for the objects in our standard subsample of high - metallicity hii regions . as noted above
, t@xmath62 has a systematic error _
e_@xmath63 that depends on the value of the excitation parameter p. to illustrate the point , we show in the bottom panel of fig . [
figure : t2-t3 ] high quality data ( those with @xmath64d@xmath65 0.01 ) for hii regions with 0.9 @xmath2 p @xmath2 0.8 ( filled circles ) and for those with 0.5 @xmath2 p @xmath2 0.4 ( open circles )
. inspection of the bottom panel in fig .
[ figure : t2-t3 ] shows that indeed the t@xmath62 t@xmath4 relations are slightly different for hii regions with different values of p : at a given value of t@xmath4 , t@xmath62 = t@xmath3 + e@xmath66 is slightly smaller , on average , for low than for high excitation hii regions . the true t@xmath3
t@xmath4 relation can be found by extrapolation of the t@xmath62 t@xmath4 relation to p = 1 .
the bottom panel in fig .
[ figure : t2-t3 ] shows that the systematic error _
e_@xmath63 is small .
therefore , we will approximate it by a linear expression , i.e. the true t@xmath3 and the derived t@xmath62 values are related by @xmath67 examination of fig .
[ figure : t2-t3 ] shows also that the @xmath1
@xmath0 relation can be parametrized as @xmath68 the coefficients _ a _ and _ b _ in eq.([equation : tt1 ] ) and the coefficient _ c _ in eq.([equation : e1 ] ) can be found simultaneosly by fitting the data in the top panel of fig .
[ figure : t2-t3 ] ) by an expression of the type @xmath69 carrying out a standard least - squares fit to the data , we obtain the following values of the coefficients ; _ a _ = 0.716 @xmath70 0.012 , _
b _ = 0.264 @xmath70 0.013 , _ c _ = 0.042 @xmath70 0.007 .
then , the true t@xmath3
t@xmath4 relation is @xmath71 the value of _ w _ given by eq.([equation : w ] ) is the fraction of radiation in the o@xmath7 zone if the condition t@xmath3=t@xmath4 holds .
this is not however the case . since the h@xmath16 emission coefficient , e@xmath72 , is dependent on the electron temperature @xcite @xmath73 then the contribution _
w_@xmath74 of the [ oiii ] zone to the total flux of the nebula in the h@xmath16 line is given by the following expression @xmath75 we have recomputed the values of t@xmath62 using _
w_@xmath74 instead of _
w_. we found the differences between the t@xmath62 values derived with _
w _ and _ w_@xmath74 to be less than 0.01 .
therefore , we will neglect this small correction , and will use the value of _ w _ given by eq.([equation : w ] ) .
we now check whether the linear forms of the analytical expressions adopted for the t@xmath3 t@xmath4 and t@xmath76 t@xmath3 relations are justified .
the top panel of fig .
[ figure : dt2 ] shows the difference @xmath77t@xmath76 between the value of t@xmath76 derived from eq.([equation : bas ] ) and eq.([equation : w ] ) and the one derived from eq.([equation : e ] ) as a function of the excitation parameter p. the points are individual measurements of hii regions from the standard subsample , and the line is the linear best fit to those data obtained through the least - squares method .
the middle panel in fig .
[ figure : dt2 ] shows the same difference @xmath77t@xmath76 as a function of the electron temperature t@xmath4 .
examination of the top and middle panels of fig .
[ figure : dt2 ] shows that @xmath77t@xmath76 does not coorelate either with p or t@xmath4 , justifying our adoption of linear forms for the t@xmath3 t@xmath4 and t@xmath76 t@xmath3 relations . the bottom panel in fig .
[ figure : dt2 ] shows @xmath77t@xmath76 as a function of the discrepancy index d@xmath30 .
there is an anticorrelation of the two quantities . since the discrepancy index appears to be an indicator of the error in the auroral line r measurements , this suggests that the t@xmath3
t@xmath4 correlation is rather tight , and that any scatter in this correlation is caused mainly by uncertainties in the measurements . the low uncertainty in the values of the coefficients in eq.([equation : tt ] )
may be a consequence of the large number of points ( 372 ) used .
the following consideration can tell us something about the true accuracy of the derived t@xmath3
t@xmath4 relation .
we have extracted from our standard sample a subsample of 86 high - quality measurements of hii regions with @xmath64d@xmath78 @xmath24 0.01 . the derived t@xmath76
t@xmath4 diagram for this subsample is shown in the top panel of fig .
[ figure : t2-t3-df ] .
a fit to those data gives @xmath79 the corresponding t@xmath3
t@xmath4 relation is then @xmath80 this relation is shown by the solid line in the top panel of fig .
[ figure : t2-t3-df ] . for comparison , the bottom panel of fig .
[ figure : t2-t3-df ] shows the t@xmath76
t@xmath4 diagram and the t@xmath3
t@xmath4 relation for the whole standard sample .
[ figure : dt2dff ] shows the differences @xmath77t@xmath76 between the values of t@xmath76 derived from eq.([equation : bas ] ) and eq.([equation : w ] ) and the one derived from eq.([equation : tte01 ] ) as a function of p ( top panel ) , t@xmath4 ( middle panel ) and d@xmath30 ( bottom panel ) .
the points are individual hii regions with 0.01 @xmath2 d@xmath30 @xmath2 0.01 , the lines are linear least - square fits to those data .
comparison of fig .
[ figure : dt2 ] and fig .
[ figure : dt2dff ] shows clearly that the scatter in t@xmath3 at a fixed value of t@xmath4 decreases in the subsample of hii regions with more precise measurements .
this confirms our conclusion that the t@xmath3
t@xmath4 correlation is rather tight , and that the scatter in it is caused mainly by measurement uncertainties .
examination of eq.([equation : tt ] ) and eq.([equation : tt01 ] ) shows that the t@xmath3 t@xmath4 relations derived from the two different samples of hii regions are very similar and agree within the formal uncertainties .
thus the derived t@xmath3
t@xmath4 relation is rather robust . in the following
, we will adopt as the t@xmath3
t@xmath4 relation @xmath81 we note that it has been generally accepted that there is a one - to - one correspondence between t@xmath3 and t@xmath4 , i.e. that the t@xmath3 t@xmath4 relation does not depend on an additional parameter . if there is a dependence of the t@xmath3
t@xmath4 relation on the excitation parameter ( from general considerations , this can not be excluded ) , that dependence can not be revealed by our approach and can influence the derived relation . in summary
, there is a tight correlation between the electron temperature @xmath0 within the [ oiii ] zone and the electron temperature @xmath1 within the [ oii ] zone in high - metallicity hii regions .
this correlation can be well approximated by a linear expression and its form is rather robust .
we now compare the t@xmath3 t@xmath4 relation obtained here with those obtained by other authors .
[ figure : compar ] shows our t@xmath3
t@xmath4 relation together with those from @xcite . since our t@xmath3
t@xmath4 relation is derived for cool high - metallicity hii regions then the high - temperature low - metallicity part of relation is not considered here . fig .
[ figure : compar ] shows that our t@xmath3
t@xmath4 relation is most similar to the one by @xcite .
we consider next how the obtained t@xmath3 t@xmath4 relation may affect the derived oxygen abundances in hii regions .
since the t@xmath3 t@xmath4 relation of @xcite has found wide acceptance and use in abundance determinations in hii regions , we will compare oxygen abundances derived with the @xcite relation and ours .
for this purpose , we use the data for hii regions in the spiral galaxy ngc 4254 obtained by @xcite .
since the auroral [ oiii]@xmath11 line was not detected in those hii regions , o / h)@xmath30 abundances based on the line fluxes r@xmath27 have been determined from the ff relation , eq.([equation : ffp ] ) , for every hii region , following @xcite .
the ( o / h)@xmath82 abundances in the hii regions of ngc 4254 determined using the t@xmath3
t@xmath4 relation of @xcite are shown as a function of galactocentric distance in fig .
[ figure : ngc4254 ] by open circles .
the dashed line is the linear least - squares best fit to those data : @xmath83 we have adopted a distance of 16.14 mpc for ngc 4254 @xcite .
the filled circles are ( o / h)@xmath82 abundances determined with our t@xmath3 t@xmath4 relation .
the solid line is the linear least - squares best fit to those data : @xmath84 fig .
[ figure : ngc4254 ] shows that the abundances derived with our t@xmath3
t@xmath4 relation are slightly higher ( up to @xmath29 0.1 dex ) than those derived with the relation by @xcite .
comparison of eq.([equation : ngccam ] ) and eq.([equation : ngcpil ] ) leads to the same conclusion .
we suggest a new way to establish the relation between the electron temperature @xmath0 within the [ oiii ] zone and the electron temperature @xmath1 within the [ oii ] zone in high - metallicity ( 12+log(o / h ) @xmath2 8.25 ) hii regions . the basic idea is the following .
if we apply the equation used to calculate the ionic abundance o@xmath7/h@xmath8 not to the entire hii region but only to the o@xmath7 zone , then this would yield , not the ionic o@xmath7/h@xmath8 abundance but the total o / h oxygen abundance instead .
we require that the equation for o@xmath7/h@xmath8 applied to the o@xmath7 zone and the one for o@xmath9/h@xmath8 applied to the o@xmath9 zone result in exactly the same value of the oxygen abundance .
this condition allows us to derive a relation between t@xmath3 and t@xmath4 .
we have applied this method to a sample of 372 hii regions selected to have high - quality measurements by using the ff relation @xcite .
we find that the correlation between t@xmath3 and t@xmath4 is tight and can be approximated by a linear expression . the so derived @xmath1
@xmath0 relation is independent of photoionization models of hii regions .
the derived relation can be used to determine t@xmath3 and accurate abundances in high - metallicity hii regions with a measured t@xmath4 temperature .
it can be also used in conjunction with the ff relation of @xcite for the determination of the t@xmath4 and t@xmath3 temperatures and oxygen abundances in high - metallicity hii regions where the [ oiii]@xmath54363 auroral line is not detected . our t@xmath3
t@xmath4 relation is close to the widely used relation of @xcite .
however , the abundances derived with our t@xmath3
t@xmath4 relation are slightly higher ( up to @xmath29 0.1 dex ) than those derived with the @xcite relation .
we thank yuri izotov and natalia guseva for providing us with their measurements of the line intensities in spectra of hii regions extracted from the data release 3 of the sloan digital sky survey ( sdss ) , the majority of which are not published .
we are grateful to bernard pagel for a numerous constructive comments over all the period of the performance of this study .
we thank the anonymous referee for helpful comments .
this research was made possible in part by award no .
up1 - 2551-kv-03 of the us civilian research & development foundation for the independent states of the former soviet union ( crdf ) .
t.x.t . has been partially supported by nsf grant ast-02 - 05785 .
thanks the hospitality of the institut dastrophysique in paris and of the service dastrophysique at saclay during his sabbatical leave .
he is grateful for a sesquicentennial fellowship from the university of virginia .
all the authors acknowledge the work of the sdss team .
funding for the sdss has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the korean scientist group , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
aller l.h .
, 1984 , physics of thermal gaseous nebulae , d. reidel publishing company , dordrecht .
bresolin f. , schaerer d. , gonzlez delgado r.m .
, stasiska g. , 2005 , a&a , 441 , 981 campbell a. , terlevich r. , melnick j. , 1986 , mnras , 223 , 811 deharveng l. , pea m. , caplan j. , costero r. , 2000 , mnras , 311 , 329 garnett d.r . , 1992 ,
aj , 103 , 1330 henry r.b.c . , pagel b.e.j
. , chincarini g.l . , 1994 , mnras , 266 , 421 izotov y.i . , stasiska g. , guseva n.g . , thuan t.x . , 2004 ,
a&a , 415 , 87 izotov y.i .
, stasiska g. , meynet g. , guseva n.g .
, thuan t.x . , 2006 ,
a&a , 448 , 955 izotov y.i . ,
thuan t.x .
, lipovetsky v.a . , 1997 ,
apjs , 108 , 1 kennicutt r.c .
, bresolin f. , garnett d.r . , 2003 ,
apj , 591 , 801 mccall m.l . ,
rybski p.m. , shields g.a . , 1985 ,
apjs , 57 , 1 oey m.s . , shields j.c .
, 2000 , apj , 539 , 687 pagel b.e.j . , simonson e.a . ,
terlevich r.j . , edmunds m.g . , 1992 , mnras , 255 , 325 pilyugin l.s . , 2005 ,
a&a , 436 , 1l pilyugin l.s .
, thuan t.x . , 2005 ,
apj , 631 , 231 pilyugin l.s . ,
thuan t.x .
, vlchez j.m . , 2006 , mnras , 367 , 1139 pilyugin l.s . , vlchez j.m . , contini t. , 2004 , a&a , 425 , 849 rubin r.h . , 1986 ,
apj , 309 , 334 shields g.a . ,
skillman e.d . ,
kennicutt r.c . , 1991 , apj , 371 , 82 stasiska g. , 1982 , a&as , 48 , 299 stasiska g. , 1990 , a&as , 83 , 501 stasiska g. , schaerer d. , 1997 , a&a , 322 , 615 tsamis y.g . ,
barlow m.j . , liu x .- w . , danziger i.j . ,
storey p.j . , 2003 , mnras , 338 , 687 zaritsky d. , kennicutt r.c . , huchra j.p . , 1994 , apj , 420 , 87 | we suggest a new way to establish the relation between the electron temperature @xmath0 within the [ oiii ] zone and the electron temperature @xmath1 within the [ oii ] zone in high - metallicity ( 12+log(o / h ) @xmath2 8.25 ) hii regions .
the t@xmath3
t@xmath4 diagram is constructed by applying our method to a sample of 372 hii regions .
we find that the correlation between t@xmath3 and t@xmath4 is tight and can be approximated by a linear expression . the new @xmath1
@xmath0 relation can be used to determine t@xmath3 and accurate abundances in high - metallicity hii regions with a measured t@xmath4 .
it can also be used in conjunction with the ff relation for the determination of t@xmath4 and t@xmath3 and oxygen abundances in high - metallicity hii regions where the [ oiii]@xmath54363 auroral line is not detected .
the derived @xmath1
@xmath0 relation is independent of photoionization models of hii regions .
galaxies : abundances ism : abundances hii regions |
in spite of the exciting progresses carried out recently , the central engine that powers gamma - ray bursts ( grbs ) remains unknown .
following the detection of an absorption system in the optical spectrum of the optical transient ( ot ) related to the gamma - ray burst grb 970508 ( metzger et al .
1997 ) , it is believed now that most grbs , if not all , arise at cosmological distances
. however , the mechanism responsible for the emission is not yet known , and it could be still possible the existence of recurrent transient optical emission .
if this is the case , archival searches are extremely important , because they will allow to search for the possible recurrent optical emission .
the detection of optical transient emission for grb 970228 ( van paradijs et al .
1997 , guarnieri et al . 1997 ) and grb 970508 ( bond 1997 , djorgovski et al .
1997 , castro - tirado et al . 1998 ) confirms this fact .
however , for some other bursts no optical emission was detected although prompt and deep follow up observations were carried out ( castro - tirado et al .
1997 , groot et al .
1998 ) .
several ots have been already found by means of archival searches . however , the confirmation of optical recurrent emission associated to grbs is still unclear . among them , ot 1901 , ot 1928 and ot 1944 ( schaefer 1981 , schaefer et al .
1984 ) are located inside small grb error boxes .
for instance ot 1928 is a prominent object 5.8 mag over the plate limit , within a 8 arcmin@xmath1 error box .
but the reality of these objects has been under debate ( ytkow 1990 , hudec et al .
other firm candidates are the three objects found by hudec et al .
( 1990 ) , at the same position of the sky .
they are well above the plate limit and the structure is compatible with a star - like object in all three cases .
however , they are outside the grb 790325b error box . moskalenko s object ( ot 1959 , moskalenko et al .
1989 ) was found inside the grb 791101 error box as a 13 mag stellar image ( or 6.6 mag if a 1-s flash is assumed ) on two different plates ( b and v filters ) taken simultaneously .
a double exposure can not be ruled out because both plates were taken with cameras on the same mount .
moreover , the low quality and the steepness of the image profile cast doubts about the reality of the object ( greiner and moskalenko 1994 ) .
the ot reported by greiner et al .
( 1991 ) , represents one of the best candidates detected so far .
it was found on three plates taken simultaneously in 1966 in the grb 781006b error box .
the magnitudes of the objects are similar to moskalenko s one , at @xmath2 mag ( or 6.3 mag if a 1-s flash is assumed ) .
although some preliminary explanations were given ( asteroids , satellite glints , head - on meteors , aircraft air lights or a dwarf nova ) , spectroscopic observations provided evidence that the ot was due to a large amplitude flaring dme star , probably unrelated to the grb ( greiner and motch 1995 ) .
the only ot found near - simultaneous to a grb event , is the one reported by borovicka et al .
they found a star - like object at the edge of the grb 790929 error box on a plate taken only 7.1 h after the high energy event .
also , two additional objects were found at the same position on another plates .
the objects are consistent with the star hde 249119 , but are relatively faint ( at the utmost 1 mag above the plate limit ) , so a plate fault or a grain cumulation can not be completely excluded .
further optical studies detected additional optical eruptions of hde 249119 , reporting ten optical brightenings with amplitudes greater than 0.5 mag ( stpn and hudec 1996 ) . besides the above - mentioned ot candidates , hudec et al .
( 1994b ) reported the presence of a very bright ot ( 5 mag above the plate limit ) inside grb 910219 watch/_granat _ error box ( sazonov et al .
1998 ) coincident with the position of a confirmed quasar with redshift z=1.78 ( vrba et al .
however the position of the object is not consistent with the small error box derived by the third interplanetary network ( hereafter ipn , hurley 1997 ) .
more recently hudec et al .
( 1997 ) reported two ots that may be connected to grb 910522 and grb 920406 . in the former case
the ot seems to be related to a dme flare star whereas a quiet counterpart with colours consistent with a quasar / agn was found in the second one .
other ot candidates have been found ( scholz 1984 , greiner and flohrer 1985 , flohrer et al .
1986 , greiner et al .
1987 , hudec et al .
1987 , hudec et al . 1988 ) but none of them can be conclusively proven as a real ot . therefore , despite strong efforts carried out during the last two decades there is no a convincing prove that ots found in plate archives and grbs are related to the same physical phenomenon .
comprehensive reviews on ots found so far can be seen in hudec et al .
( 1993a , 1993b , 1994c , 1995 ) .
the most recent archival search has been performed by gorosabel and castro - tirado ( 1998a ) .
they have examined 8004 plates at the harvard college observatory plate collection searching for optical transient emission from the gamma - ray burst grb 970228 .
this has been the first archival search carried out so far for a gamma - ray burst with known transient optical emission .
the total exposure time amounted to @xmath3 yr .
no convincing optical activity has been found above @xmath4 mag at the expected position of the grb 970228 optical counterpart .
table 1 displays a summary of the most exhaustive archival searches performed to date .
.the most exhaustive grb archival searches carried out to date .
[ cols="<,^,^,^ " , ] ubvri images of the field containing the ot are shown in fig .
the content of the @xmath5 astrometrical error circle is shown in fig .
4 . magnitudes and positions of all objects are displayed in table 6 .
photometrical errors in the magnitudes are overestimated due to the presence of several bad ccd columns .
object # 1 seems to be the most interesting object within the 1@xmath6 error circle , with u @xmath7 , b = @xmath8 , v = @xmath9 , r = @xmath10 , i = @xmath11 .
the colours are consistent with that of a k7/m0 dwarf star or a k5/m1 giant , although they could be also consistent with a reddened galaxy ( see fig 5 ) .
the object on plate am 24589 appears to be @xmath12 mag over the plate limit and is absent on plate am 24592 ( obtained 3 hours and 5 minutes after ) .
therefore , the object should have declined at a rate of @xmath13 mag / hr , which is similar to same decline rates in flares stars . could object # 1 be classified as a flare star ? . to estimate
the probability of having a k7/m0 dwarf star or a k5/m1 giant coincident with the 1@xmath6 astrometrical error is difficult because of various statistical biases in star catalogues .
taking into account the number of stars per square degree of magnitude b=22 at galactic latitude @xmath14 and the fraction of k7/m0 spectral - type main sequence and k5/m1 giant stars ( allen 1973 ) , we expect @xmath15 k7/m0 main sequence stars and @xmath16 k5/m1 giants within a @xmath17 error circle ( the size of the 1@xmath6 astrometrical error radius ) .
therefore , in the absence of stronger constraints we can not prove that this star has caused the object detected on the plate am 24589 .
considering that in the solar neighbourhood the fraction of m0-m1 type stars that can be considered as flare stars is 5% ( shakhovskaya 1994 , mirzoyan 1994 ) , the probability that a m0-m1 flare star is located by chance inside the 1@xmath6 astrometrical error would be @xmath18 .
therefore , if object # 1 was a flare star , it is likely to be the responsible of the object found on the plate am 24589 . in this case
, the flare amplitude would be @xmath19 mag which is slightly higher than the extreme flares of the cz cnc stars ( greiner and motch 1994 , toth et al .
it should be noted that m flare stars have been found at ots position inside two grb error boxes ( greiner and motch 1995 , hudec et al .
in any case an optical spectrum would be desirable in order to clarify its nature .
ots have been found at the edge or slightly outside grb error boxes , like grb 910219 ( hudec et al .
1994b ) , grb 790325b ( hudec et al .
1990 ) or grb 791006b ( greiner et al .
in fact several authors suggest a strong relationship between quasars and ots ( vrba et al .
1994 , hudec et al .
1997 ) , but statistical studies do not show a convincing correlation between quasars and grbs ( gorosabel et al .
1995 , schartel et al .
1997 , gorosabel and castro - tirado 1998b , gorosabel and castro tirado 1998c , hurley et al .
1998 ) .
the object is inside the watch error circle ( sazonov et al . 1998 ) , but @xmath20 outside the improved ipn error box for grb 910122 ( hurley et al .
1997 ) and not included in any of the three annuli determined by the _ granat _ , _ pvo _ and _ ulysses _ spacecraft .
even if we assume an extended halo radius equal to 150 kpc and a high - velocity neutron star ( at 1000 km / s ) , its apparent motion on the sky would be only @xmath0 0.14 in 100 yr .
therefore , theories based on high - speed neutron stars in the halo will not allow to place the optical source that far away the ipn error box ( li and dermer 1992 ) .
the maximum allowable distance to detect such proper motion of @xmath20 in 46 yr ( from 1945 to 1991 ) would be 35 pc , but such a nearby neutron star would have been easily detected in previous all - sky x - ray surveys .
therefore we can set a lower limit of 0.55 yr for any recurrent optical transient activity ( above 14.1 mag ) related to grb 910122 .
this can be compared with the 2.1 yr in the case of grb 910219 ( hudec et al .
further results on the optical content of the grb 910122 ipn error box will be published elsewhere ( gorosabel , castro - tirado and lund 1998 ) .
we have examined @xmath0 3995 plates at hco for grb 910122 covering @xmath0 0.55 yr .
no convincing evidence of optical transient emission was found within the grb 910122 error box .
additional plates at roe did not show any suspicious object either .
however , a possible ot was found at hco on the plate am 24589 , located @xmath20 outside the improved error box provided by the ipn .
multicolour optical observations were carried out , finding 3 sources consistent , within the @xmath21 object astrometrical error box .
one of them shows colours typical of a m dwarf star or a reddened galaxy . in any case
it seems to be unrelated to the grb because of its location outside the ipn error box .
the authors warmly thank the harvard college observatory plate curator m. l. hazen and a. doane for their help at the hco plate stacks .
we are indebted to the referee r. hudec for valuable comments and advices .
also , we would like to thank prof .
ofer bar - yosef , at the stone age laboratory , department of anthropology ( peabody museum - harvard university ) to have given us the opportunity of using the reflected - transmitted light microscope .
the authors also wish to thank n. lund for the facilities given in order to perform the optical observations at la silla , s. tritton for the support provided at the royal observatory of edinburgh , and g.v .
williams for the information provided by the iau minor planet center .
we are grateful to k. hurley for kindly providing the new coordinates of the grb 910122 error box , and to l. metcalfe , j. greiner and b. montesinos for fruitful conversations .
this work has been partially supported by spanish cicyt grant esp95 - 0389-c02 - 02 .
guziy , s. , shlyapnikov , a. , gorosabel , j. et al .
1997 , in : proceedings of the 2nd integral workshop `` the transparent universe '' , eds : c. winkler , t.j .- l .
courvoisier & ph .
durouchoux , esa sp-382 , 485 .
hudec , r. , luginbuhl , c.b . ,
vrba f.j . ,
hartmann , d.h .
& hurley , k. 1997 , in : proceedings of the 2nd integral workshop `` the transparent universe '' , eds : c. winkler , t.j .- l .
courvoisier & ph .
durouchoux , esa sp-382 , 481 .
moskalenko , e.i . ,
karnashov , a.n . ,
kramer , e.n .
1992 , in : `` gamma ray bursts observations , analyses and theories'',ho , c .
, epstein r. i. & fenimore , e. e. , eds . , cambridge university press , cambridge , p. 127 . | we present here the results of a study carried out at the harvard college observatory plate collection .
we examined 3995 plates covering the error box of the gamma - ray burst grb 910122 , over a time span of 90 yr ( from 1889 to 1979 ) .
the total exposure time is @xmath0 0.55 yr . no convincing evidence of optical transient emission was found within the grb 910122 ipn error box .
however , a possible ot was found _ outside _ the grb 910122 error box .
optical ground based observations have revealed a [email protected] object consistent with the position of the new object .
the colours of the object are typical of a k7/m0 star or a reddened galaxy , which could have caused the ot , but the fact that the object is far away from the grb error box makes both events unrelated . |
in the last decades , the investigation about the transition from quantum to classical dynamics has progressed enormously .
this was induced in part by the development of the experimental techniques , especially in quantum optics @xcite , and in part by the possibility of appearance of technology in quantum information processing @xcite , @xcite .
such research aims to understand how the typically quantum effects disappear in the dynamics of macroscopic systems . according to a popular theoretical model , one believes that the emergence of the classical world from quantum mechanics is a consequence of the unavoidable coupling between the macroscopic system and its environment .
for example , in accordance with this proposal @xcite , environmental coupling is responsible by the rapid evolution of coherent quantum superpositions of macroscopically distinguishable states ( schrdinger cat states ) into statiscal mixtures , a phenomenon known as decoherence . despite these theoretical advances ,
some aspects of the quantum to classical transition remain subtle and controversial , especially in the case of classically nonlinear or chaotic systems . for some authors ,
the departure of the quantum mean values of observables from the corresponding classical ones ( correspondence breakdown ) in chaotic systems occurs in a very small time scale ( see , for example , @xcite ) and they sustain the idea that the coupling with purely diffusive environment can reduce these discrepancies and increase the break time by decoherence . on the other hand , for other authors ( see @xcite ) decoherence is not necessary to explain the classical behavior of macroscopic systems ( including the chaotic ones ) since the observed discrepancies between quantum and classical mean values of observables are negligible for any current realistic measurement . in order to shed some light on this debate
, we turn our attention to the one - dimensional nonlinear or anharmonic oscillator ( aho ) coupled to a purely diffusive reservoir .
three are the reasons for the choice of this model .
firstly , in the limit of vanishing environmental coupling , its quantum dynamical evolution exhibits several effects without analogous in its classical counterpart , such as revivals and appearance of coherent superpositions of states @xcite .
secondly , by virtue of its relative simplicity , it is possible to obtain the exact solutions of the equations of motion for quantum @xcite and classical versions of the model even in the presence of the reservoir .
last , but not least , the recent technical advances in trapping and controlling cold atoms suggest the bose - einstein condensates ( bec ) as potential candidates to implement experimental tests of this model .
in fact , in the single - mode approximation , a bec trapped in a optical lattice is suitably described by the quantum nonlinear oscillator @xcite .
hence , dissipative aho became a largely studied model in the literature . in ref .
@xcite , daniel and milburn obtained the exact evolution of the _ q _ function associated to an initial coherent state of an aho subject to attenuation or amplification .
the authors showed that the effects yielded by nonlinearity , such as revivals and squeezing , gradually vanish if a non - unitary mechanism is taken into account .
these results were extended by pe@xmath0inov and luk@xmath1 @xcite to an arbitrary initial state . in ref .
@xcite , kheruntsyan obtained the steady state of wigner function of a single driven damped cavity mode in the presence of a kerr medium .
the author studied a model of reservoir that included two - photon absortion , besides the usual one - photon absorption .
closely related to the present contribution , chaturvedi and srinivasan @xcite found an exact solution of a class of master equations governing the dynamics of a chain of coupled dissipative aho .
the authors used thermofield dynamics notation in order to map a master equation into a schrdinger equation with a non - hermitean hamiltonian .
the form of the general solution of this equation was expressed as an ordered product of exponentials of operators acting on an arbitrary initial condition . in this contribution
, we obtain the exact time evolutions of the density operator ( section 2 ) and of the classical distribution function ( section 3 ) in the quantum and classical versions of the nonlinear oscillator coupled to a purely diffusive environment .
both results are obtained by the application of the lie algebraic techniques , in particular , the baker - campbell - hausdorff ( bch ) formulas used to expand a lie exponential in an ordered product of exponentials @xcite . in the quantum case , the result allows us to find the corresponding wigner function in terms of an expansion in associated laguerre polynomials .
it is important to mention that the wigner function should smoothly approach the corresponding classical distribution in the appropriate limit @xcite .
thus , we take the classical limit of the partial differential equation ( pde ) that governs the time evolution of the wigner function and we obtain a classical fokker - planck ( fp ) equation .
the algebraic structure of the master equation in the quantum case is preserved by the corresponding fp equation in the classical case , so that we can extend the methods applied to obtain the solution of the first in the finding of the solution of the last . in ref .
@xcite , oliveira and co - workers investigated the diffusive aho and showed that the break times ( _ i.e. _ , the characteristic times of departure of the quantum and classical dynamics ) depend strongly on observable and initial condition . a more fair " comparison between the two dynamics shall be given by the evaluation of the distance between the corresponding distributions in the phase space @xcite .
hence , the exact solutions of the equations of motion for the wigner function and classical distribution function will allow to define analytically the break time for the diffusive aho and its dependence in terms of the nonlinearity strength and diffusion constant .
we conclude this work by comparing the quantum and classical evolutions of the wigner and the classical distribution functions for an initial coherent state in the aho with and without diffusion ( section 4 ) .
the results are preliminar and deserve a more careful analysis , but they suggest that , as expected , in quantum case , inclusion of diffusion reduces the phase space interferences and therefore prevent the appearance of quantum coherent superpositions .
the area of the regions where the wigner function should be negative is reduced too . as time goes by
, the wigner distribution gradually takes the form of an annular volume around the origin of the phase space . for later times , this volume becomes more `` fat '' and `` flat '' . in the classical case ,
the fine - structured whorl yielded by the distribution in the diffusionless regime is destroyed .
as it happens in the quantum case , diffusion turns the classical distribution more `` fat '' and `` flat '' around the origin .
these results suggest that the wigner function of the quantum diffusive aho converges gradually to the distribution function of the corresponding classical version of the model in non - unitary evolution .
let us consider the aho coupled to a thermal bath of oscillators in equilibrium at temperature @xmath2 .
assuming that the nonlinearity strength is small and the coupling to the reservoir degrees of freedom is weak , we obtain the following master equation in the interaction picture : @xmath3 + k\left(\bar{n } + 1\right)\left[2\a \rot{t}\adag - \adag\a \rot{t } - \rot{t}\adag\a\right ] \notag \\ & & + k\bar{n}\left[2\adag \rot{t}\a - \a\adag \rot{t } - \rot{t}\a\adag\right ] \notag \ , , \end{aligned}\ ] ] where @xmath4 and @xmath5 are the damping and nonlinearity constants , respectively , and @xmath6 is the average number of thermal photons in the mode @xmath7 of reservoir ( @xmath7 is the natural frequency of the oscillator ) .
the density operator @xmath8 represents the state of the system at time @xmath9 ; @xmath10 and @xmath11 are the annihilation and creation operators , respectively .
we are interested in the so called diffusive limit of the above equation .
this limit is obtained by taking the damping constant going to zero , @xmath12 , and the number of thermal photons going to infinite , @xmath13 , keeping the product @xmath14 finite .
thus , the master equation ( [ master1 ] ) becomes @xmath15 \notag \\ & & + 2\kappa\left[\a \rot{t}\adag + \adag \rot{t}\a - \adag\a \rot{t } - \rot{t}\adag\a- \rot{t}\right ] \notag \\ & = & -ig\left ( \mathcal{m}^{2}-\mathcal{p}^{2}\right ) \rot{t } + 2\kappa \left [ \mathcal{j}+\mathcal{r}-\left ( \mathcal{m}+\mathcal{p}% + 1\right ) \right ] \rot{t } \notag \ , .\end{aligned}\ ] ] using the notation given in ref .
@xcite , the super - operators ( sup - op ) that appear in the last line of above equation are defined by [ supop ] @xmath16 here , @xmath17 , @xmath18 , @xmath19 , @xmath20 represent the left and right actions of the creation and annihilation operators on a generic operator @xmath21 : @xmath22 at this point , it is convenient to introduce the following nomenclature .
we call @xmath23 the set of operators that act on the oscillator space state @xmath24
. the elements of @xmath23 can be assigned to vectors of an extended hilbert space constructed by direct product between the original space state and its dual @xmath25 , _ viz .
_ @xmath26 .
this extended hilbert space is frequently called hilbert - schmidt space or liouville space @xcite . the formal solution of eq .
( [ master2 ] ) is given by @xmath27 where @xmath28 represents the initial state of the aho .
the evolution of a generic initial state @xmath29 can be evaluated by expanding the exponential @xmath30 in an ordered product of exponentials .
usually , this task is achieved by the systematical application of the lie algebraic methods , in particular , using the baker - campbell - hausdorff ( bch ) formulas @xcite . for the dissipative aho
, this expansion was obtained in ref .
@xcite , and we reproduce in detail the procedure here . to carry out this expansion , we begin by evaluating the commutation relations between the sup - op defined in ( [ supop ] ) .
they are listed in table [ comutsupop ] .
.commutation relations between the sup - op defined in eq .
( [ supop ] ) .
the @xmath31 entry in the table is the result of the commutation between the sup - op of the @xmath32-th row and the sup - op of the @xmath33-th column . [
cols="^,^,^,^,^",options="header " , ] in the master equation ( [ master2 ] ) , the unitary nonlinear term introduces the sup - op @xmath34 .
the inclusion of this sup - op to the set considered in table [ comutsupop ] yields an infinite lie algebra when we evaluate its commutation relations with another sup - op . the trick used to find the solution of eq .
( [ master2 ] ) consists formally in considering the sup - op @xmath35 as a c - number , since it commutates with the rest and the functions obtained in the expansion of the correpondent lie exponential are , in fact , functions of this operator .
we can determine the quantum state in time @xmath36 , @xmath37 , by expanding the initial state @xmath38 in terms of the eigenfunctions of @xmath35 .
in the same way , the nonlinear hamiltonian term in fokker - planck equation ( [ fpquartic2 ] ) introduces products of differential operators and they also yield an infinite lie algebra
. however , these products are in the form @xmath39 , @xmath40 . since @xmath41 commutates with the rest of the elements defined in eq .
( [ opdif ] ) , we can employ an analogous trick to that employed in the solution of ( [ fpquartic2 ] ) , i.e. , we can formally consider @xmath41 a c - number and evaluate the functions in the correspondent lie series as functions of this operator .
however , the solution @xmath42 determined in this way will be useful " if the action of these differential operators on elements in @xmath43 is known . instead and analogously to the procedure adopted in the quantum version of the problem
, we prefer to find the eigenfunctions of @xmath41 and to express the initial state in terms of them .
if we know how each operator in eq .
( [ opdif ] ) acts on these eigenfunctions , the classical distribution @xmath42 can be written as an expansion in terms of them with time dependent coeficients .
we can easily determine the eigenfunctions of @xmath44 , since they are related with the eigenfunctions of the sup - op @xmath45 by the weyl - wigner transform .
remembering , the eigenfunctions of this sup - op are @xmath46 , with eigenvalues @xmath47 . by eq .
( [ corresp ] ) , the weyl - wigner transform of the sup - op @xmath45 yields the differential operator @xmath41 , i.e. , @xmath48 = y_0 \mathcal{w}\left(\hat{o}\right)$ ] , where @xmath49 is a generic operator acting on @xmath24 .
making @xmath50 , we have @xmath51=y_0 \mathcal{w}\left(\left|m\right\rangle \left\langle n\right|\right)=\left(m - n\right)\mathcal{w}\left(\left|m \right\rangle \left\langle n\right|\right).\ ] ] therefore , the eigenfunctions of @xmath44 are the weyl - wigner transform of @xmath52 , namely @xmath53 , with eigenvalues @xmath54 . since the rest of the differential operators defined in eq .
( [ corresp ] ) are related with the sup - op defined in eq .
( [ supop ] ) , and the action of the last on elements of set @xmath46 is known , we directly obtain the corresponding action of the first on the functions @xmath53 .
in fact , we have @xmath55 in the problems that we are interested , we need to compare the time evolutions of the wigner function and of the classical distribution associated to a given initial state is a probability function that the marginal distributions coincide with the corresponding ones produced by the wigner function @xmath56 . ]
@xmath57 . for our purposes
, it is interesting to express such function as an expansion in eigenstates of @xmath41 . using the properties of the weyl - wigner
transform @xmath58 we have @xmath59 hence , @xmath60 consider the initial state @xmath61 .
if the corresponding wigner function @xmath62 , given by eq . ( [ expansion ] ) , represents a valid classical distribution function , we can make @xmath63 .
the solution of eq .
( [ fpquartic2 ] ) for this initial condition is @xmath64\exp\left[2y_z \ln g_z\left(t\right ) \right]\exp\left[g\left(t\right)y_-\right ] \label{sol } , \end{aligned}\ ] ] where [ parameters ] @xmath65 note that @xmath66 and @xmath67 are functions of operator @xmath41 .
the action of @xmath68 $ ] on the initial condition @xmath69 yields @xmath70 w_0 = \sum_j \sum_{r , s } \rho_{r+j , s+j}\frac{v_{r - s}^j\left(t\right)}{j ! } \sqrt{\frac{\left(r+j\right)!\left(s+j\right)!}{r!s ! } } \pi_{r , s } .\ ] ] here , @xmath71 is obtained from the function @xmath66 substituting the operator @xmath41 by @xmath72 in eq .
( [ g ] ) . since @xmath53 is eigenfunction of @xmath73 with eigenvalue @xmath74 , the action of @xmath75 $ ] on this function produces @xmath76 \pi_{m , n } = g_z^{m + n + 1}\left(t\right)\pi_{m , n}.\ ] ] hence , @xmath77 \exp\left[g\left(t\right)y_-\right]w_0 & = & \sum_j \sum_{r , s } \rho_{r+j , s+j } u_{r - s}^{r+s+1}\left(t\right)\\ & & \times \frac{v_{r - s}^j\left(t\right)}{j ! } \sqrt{\frac{\left(r+j\right)!\left(s+j\right)!}{r!s ! } } \pi_{r , s } , \end{aligned}\ ] ] where @xmath78 is obtained from @xmath67 substituting the operator @xmath41 by @xmath79 in eq .
( [ gz ] ) .
finally , the action of the exponencial @xmath80 $ ] on the above result gives @xmath81 for the initial condition ( [ expansion ] ) , the fokker - planck equation ( [ fpquartic2 ] ) for the diffusive aho has the following solution @xmath82 compare this solution with the corresponding one obtained for the density operator , eq .
( [ solution1 ] ) , or for the wigner function , eq . ( [ wigner2 ] ) , in the quantum mechanical version of the problem .
_ mutatis mutandis _ , the form of the solutions is identical . substituting the functions @xmath83 and @xmath84 in eq .
( [ wigner2 ] ) by @xmath85 and @xmath86 , we find the result given in eq .
( [ solution2 ] ) .
in order to gain some insight into the differences between quantum and classical dynamics of the nonlinear oscillator , let us compare the time evolution of a common initial condition .
consider a coherent state @xmath87 .
the associated wigner function in the complex phase space coincides with the corresponding classical distribution , _
i.e. _ , a gaussian with variance equal to the unity centered in @xmath88 .
an example is shown in fig .
the other figures represent the time evolution of this initial state in the quantum and classical models and for the regimes with and without diffusion .
it is important to mention that the time evolutions were obtained with the solutions ( [ wigner2 ] ) and ( [ solution2 ] ) .
the classical hamiltonian evolution is such that any point of the phase space moves around the origin with angular frequency proportional to @xmath89 , where @xmath90 is the coordinate of this point @xcite
. therefore , points over the the initial distribution will rotate with an angular velocity that depends on their distance to the origin . as consequence , the distribution will continuously spiral around the origin , as shown in fig .
the distribution yields a fine - structure in phase space @xcite , which is gradually destroyed if diffusion is included ( see fig .
[ fig3 ] ) . and @xmath91 .
in the quantum case , this distribution corresponds to an initial coherent state @xmath92 with @xmath93 , whereas in the classical case it corresponds to a gaussian distribution centered in @xmath94 with unitary variance . ] corresponding to the hamiltonian evolution ( @xmath95 ) of the initial condition shown in fig .
[ fig1 ] . for this case , @xmath96 .
note the fine - structure yielded by the continuous spiraling of the distribution around the origin of the phase space . ]
( top ) and @xmath97 ( bottom ) corresponding to the diffusive evolution ( @xmath98 ) of the initial condition shown in fig .
[ fig1 ] . for this case , @xmath96 .
note that the fine - structure that should be yielded in the diffusionless regime ( see fig .
[ fig2 ] ) is gradually destroyed.,title="fig : " ] ( top ) and @xmath97 ( bottom ) corresponding to the diffusive evolution ( @xmath98 ) of the initial condition shown in fig .
[ fig1 ] . for this case , @xmath96 .
note that the fine - structure that should be yielded in the diffusionless regime ( see fig .
[ fig2 ] ) is gradually destroyed.,title="fig : " ] in the quantum version of the model , the unitary evolution of the wigner function exhibits a very different behavior that the classical one . for times equal to @xmath99 , where @xmath100 is an integer
, the nonlinearity leads to the quantum superpositions of states ( @xmath100 is odd ) or revivals and anti - revivals ( @xmath100 is even ) .
these effects were already reported by yurke and stoler @xcite and examples of them are given in figs .
[ fig4 ] and [ fig5 ] . corresponding to the unitary evolution ( @xmath101 ) of the initial condition shown in fig .
[ fig1 ] . for this case , @xmath96 .
note that the appearance of phase space interference due to the coherent superposition of states ( schrdinger cat state ) . at this time , the state exhibits some squeezing as well @xcite .
compare with the corresponding classical one shown in fig .
[ fig2 ] . ]
corresponding to the unitary evolution ( @xmath101 ) of the initial condition shown in fig .
[ fig1 ] . for this case , @xmath96 . at this instant
, the oscillator exhibits an anti - revival since it is found in a coherent state with amplitude @xmath102 , where @xmath88 is the amplitude of the initial coherent state . ]
when the diffusion is included , the quantum effects discussed above are gradually suppressed . the interference in phase space
is reduced , and the regions where the wigner function is negative diminish ( see fig .
[ fig6 ] ) . for later times , the wigner function and the classical distribution take a form of an annular volume around the origin of the phase space .
the annular region grows with time but its maximum value diminishes in order to mantain constant the integral of @xmath103 or @xmath104 over the phase space .
this suggests that the wigner function of the quantum diffusive aho converges gradually to the corresponding classical distribution function .
non - unitary effects due to quantum dynamics of the open aho were investigated by milburn and holmes @xcite , and by daniel and milburn @xcite , considering the coupling to a null and non - null temperature reservoir , respectively . in both works ,
the authors evaluates the time evolution of the husimi function ( @xmath105 function ) , another phase space representation of the density operator .
their results show a similar behavior for the @xmath105 function , in qualitative agreement with the one reported here .
( top ) and @xmath97 ( bottom ) corresponding to the diffusive evolution ( @xmath98 ) of the initial condition shown in fig .
[ fig1 ] . for this case , @xmath96 . comparing with figs .
[ fig4 ] and [ fig5 ] one notes that the phase space interferences and the revivals are gradually suppressed.,title="fig : " ] ( top ) and @xmath97 ( bottom ) corresponding to the diffusive evolution ( @xmath98 ) of the initial condition shown in fig .
[ fig1 ] . for this case , @xmath96 . comparing with figs .
[ fig4 ] and [ fig5 ] one notes that the phase space interferences and the revivals are gradually suppressed.,title="fig : " ] these results serve to illustrate the procedure and to show that the results are physically consistent .
however , a detailed study of the quantum to classical transition in the aho must take into account the role played by the parameters of interest , namely the nonlinearity strengh , the diffusion constant , and the amplitude of the initial coherent state ( a measure of classicallity of the initial state ) .
this work is in progress .
the author thanks to a. c. oliveira and m. c. nemes for fruitful discussions on this topic . | we present the general solutions for the classical and quantum dynamics of the anharmonic oscillator coupled to a purely diffusive environment . in both cases ,
these solutions are obtained by the application of the baker - campbell - hausdorff ( bch ) formulas to expand the evolution operator in an ordered product of exponentials .
moreover , we obtain an expression for the wigner function in the quantum version of the problem .
we observe that the role played by diffusion is to reduce or to attenuate the the characteristic quantum effects yielded by the nonlinearity , as the appearance of coherent superpositions of quantum states ( schrdinger cat states ) and revivals .
pacs : 03.65.yz : decoherence ; open systems ; quantum statistical methods , 02.20.sv : lie algebras of lie groups |
the ever increasing demand for more secure , ecosystem friendly and broadband underwater communication and exploration necessitates comprehensive studies on underwater wireless optical communication ( uwoc ) . compared to its traditional counterpart , namely acoustic communication ,
uwoc provides better security , lower time delay and higher bandwidth @xcite , and is more compatible with underwater ecosystem .
despite all these , propagation of light under water is affected by three main degrading effects , namely absorption , scattering and turbulence , which limit the viable communication ranges of uwoc systems to typically less than @xmath4 . in recent years , many valuable researches have been carried out in the context of uwoc , from channel modeling to system design . in @xcite petzold
s experimental data @xcite are used to simulate the uwoc channel impulse response through monte carlo numerical method . however ,
these prior works only focused on the absorption and scattering effects of the channel , many relevant studies have been performed on the turbulence characterization .
for example , in @xcite an accurate power spectrum has been derived for fluctuations of turbulent seawater refractive index .
rytov method has been used in @xcite to evaluate the scintillation index of optical plane and spherical waves propagating in underwater turbulent medium . and in @xcite , the on - axis scintillation index of a focused gaussian beam is formulated for weak oceanic turbulence and log - normal distribution is considered for intensity fluctuations to evaluate the average bit error rate ( ber ) in such systems . on the other hand , a cellular topology based on optical code division multiple access
( ocdma ) technique has been recently proposed in @xcite for underwater wireless optical networks , while the potential challenges and applications of such a network is discussed in @xcite . besides , beneficial applications of multi - hop transmission ( serial relaying ) on the up - and downlink performance of underwater users in an ocdma network and also in point - to - point uwoc links have been investigated in @xcite and @xcite , respectively , with respect to all the degrading effects of uwoc channels .
furthermore , the authors in @xcite have focused on the turbulence - induced fading effects of uwoc channels and proposed multiple - input multiple - output ( mimo ) transmission to mitigate turbulence effects through spatial diversity gain of mimo technique .
although , the statistical distribution of fading for free - space optical ( fso ) channels is thoroughly investigated and characterized @xcite , its probability density function ( pdf ) for uwoc channels is not yet well determined .
prior works on the context of uwoc mainly focused on weak oceanic turbulence and adopted the same statistical distribution as fso channels , i.e. , log - normal distribution for this regime of underwater turbulence @xcite .
however , significant differences between the nature of atmospheric and oceanic optical channels may result in a different pdf for underwater fading statistics .
the research in this paper is inspired by the need to more specifically study the intensity fluctuations through uwoc channels . in this paper
, we experimentally investigate the effect of various phenomena on the fluctuations of the received optical signal within a laboratory water tank . since the severity of optical turbulence rapidly increases with the link range , the short link range available with a laboratory water tank may typically result negligible fluctuations on the received optical signal .
therefore , at the first step , we artificially add air bubbles in our water tank to enlarge the channel s temporal fluctuations . then , for each channel condition , we take sufficient samples from the received optical signal , normalize them to the received power mean , and plot the normalized samples histogram to obtain the intensity fluctuations distribution and the channel coherence time ; the time period in which the channel fading coefficient remains approximately constant .
it is worth mentioning that since many of underwater vehicles , divers and submarines generate air bubbles , study on the intensity fluctuations due to the presence of air bubbles has significant importance in precise uwoc channel modeling and system design .
the rest of the paper is organized as follows . in section
ii the uwoc channel model is presented .
the used evaluation metrics , and general statistical distributions for optical turbulence are briefly described in sections iii and iv , respectively .
section v provides the system model and experimental set up , section vi presents experimental results for various scenarios and section vii concludes the paper .
propagation of light under water is affected by three major degrading effects , i.e. , absorption , scattering and turbulence .
in fact , photons of a propagating light wave may encounter with water molecules and particles . during this collision , energy of each photon
may be lost thermally , which is specified as absorption process and is characterized by absorption coefficient @xmath5 , where @xmath6 is the wavelength of the propagating light wave .
furthermore , in the aforementioned collision process direction of each photon may be changed that is defined as scattering process and is determined by scattering coefficient @xmath7 . besides , total energy loss of non - scattered light is described by extinction coefficient @xmath8 @xcite .
the channel fading - free impulse response ( ffir ) , @xmath9 , can be obtained through monte carlo simulations similar to @xcite .
although , this ffir includes both absorption and scattering effects , comprehensive characterization of the uwoc channel impulse response requires fading consideration as well . to do so
, the channel ffir can be multiplied by a fading coefficient @xcite .
however , log - normal distribution is mainly used for fading coefficients pdf @xcite , inspired by the behaviour of optical turbulence in atmosphere , an accurate characterization of fading coefficients statistical distribution necessitates more specific studies , that is carried out in this paper . in order to specify the fading strength ,
it is common in the literature to define the scintillation index of a propagating light wave as @xcite ; @xmath10 -{{\operatorname{\mathbb{e}}}^2\left[i\left(r , d_0,\lambda \right)\right]}}{{{\operatorname{\mathbb{e}}}^2\left[i\left(r , d_0,\lambda \right)\right]}},\end{aligned}\ ] ] in which @xmath11 is the instantaneous intensity at a point with position vector @xmath12 , where @xmath13 is the propagation distance and @xmath14 $ ] denotes the expected value of the random variable ( rv ) @xmath15 .
in order to investigate the uwoc channel s fading statistic , we first , experimentally measure the channel coherence time for different channel conditions . to do so
, we measure the time in which the temporal covariance coefficient of irradiance , defined as eq . , remains approximately constant @xcite .
@xmath16 where @xmath17 with @xmath18 .
moreover , @xmath19 , and @xmath20 is defined as the covariance of irradiance for two points @xmath21 and @xmath22 at different time instants @xmath23 and @xmath24 . as the next step of our work , we specify the channel fading statistical distribution .
we use the following two well - known metrics to evaluate the accordance of different statistical distributions with the experimental data .
this metric determines how well the considered distribution predicts the experimental data , and is defined as ; @xmath25 where @xmath26 is the number of measured samples , @xmath27 is the measured intensity through the experiment and @xmath28 is the predicted intensity by the considered distribution .
this metric , which is also known as @xmath29 measure , is used to test the distribution s fitness and is expressed as ; @xmath30 in which @xmath31 is the considered distribution s sum of square errors , i.e. , @xmath32 , where @xmath33 is the number of bins of the acquired data histogram , and @xmath34 and @xmath35 are , respectively , the measured and the predicted probability values for a given received intensity level .
and @xmath36 is the sum of the squares of distances between the measured points and the mean of them , i.e. , @xmath37 , where @xmath38 . as the value of the @xmath29 measure , for a given distribution , approaches its maximum ( i.e. , @xmath39 )
, the distribution is considered to better fit the measured data .
for the sake of brevity , we only take into account the following statistical distributions which are widely accepted for various regimes of atmospheric turbulence @xcite .
moreover , we propose a combined exponential and log - normal distribution to better characterize the fading statistics for @xmath3 . _
1 ) log - normal distribution : _ such a distribution is commonly used in literature for weak atmospheric turbulence regime , characterized by @xmath40 .
let @xmath41 be the channel fading coefficient ( @xmath42 ) with log - normal pdf as @xcite ; @xmath43 therefore , the fading log - amplitude @xmath44 has a gaussian distribution with mean @xmath45 and variance @xmath46 . to ensure that fading neither amplifies nor attenuates the average power ,
we normalize fading coefficients such that @xmath47=1 $ ] , which implies that @xmath48 @xcite .
it can be shown that for log - normal fading the log - amplitude variance relates to the scintillation index as @xmath49 @xcite .
_ 2 ) k distribution : _ this distribution is mainly used for strong atmospheric turbulence in which @xmath50 . the pdf of k distribution is given by @xcite ; @xmath51 where @xmath52 is the gamma function , @xmath53 is a positive parameter which relates to the scintillation index as @xmath54 @xcite .
and @xmath55 is the @xmath56th order modified bessel function of the second kind .
_ 3 ) gamma - gamma distribution : _ such a statistical model , which factorizes the irradiance as the product of two independent random processes each with a gamma pdf , has been introduced in literature to describe both small - scale and large - scale atmospheric fluctuations .
the pdf of gamma - gamma distribution is expressed as @xcite ; @xmath57 in which @xmath53 and @xmath58 are parameters related to effective atmospheric conditions .
the scintillation index for gamma - gamma distribution is given by @xmath59 @xcite .
_ 4 ) proposed combined exponential and log - normal distribution : _ when the received optical signal has a large dynamic range or equivalently the received signal lies either in large or small levels , general single - lobe distributions can not specify the fading statistical distribution .
in such circumstances , a two - lobe statistical distribution is required to interpret the channel intensity fluctuations .
therefore , we propose the combination of an exponential and a log - normal distributions with pdf of ; @xmath60 where , @xmath61 determines the proportion between the exponential and the log - normal distributions , @xmath62 is the exponential distribution mean , and @xmath63 and @xmath64 are the constants of the log - normal distribution .
fading coefficients normalization implies that @xmath65=k\gamma+(1-k)\exp\left(\mu+{\sigma^2}/{2}\right)=1 $ ] .
moreover , the scintillation index and the @xmath66th ( @xmath66 is a positive integer ) moment for the above distribution can be obtained as @xmath67 and @xmath68=k\gamma^nn!+(1-k)\exp\left(n\mu+n^2{\sigma^2}/{2}\right)$ ] , respectively .
our setup is implemented in a @xmath69 black colored water tank . in the transmitter side , a @xmath70 nm green laser diode with the maximum output power of @xmath4 mw
is driven by a mos transistor to ensure a constant optical irradiance . in the receiver side , an ultraviolate - visible photodetector ( pd )
is used ; the pd is covered by a convex outer lens in order to collect the maximum amount of irradiance .
additionally , both laser and pd are sealed by placing in transparent boxes .
pd s generated current is amplified through ad840 operational amplifier .
the magnified signal , which is proportional to the received optical power relates to the received optical power @xmath71 as @xmath72 , where @xmath73 is the pd s quantum efficiency , @xmath74 is the planck s constant and @xmath75 is the light frequency .
moreover , the amplified signal is proportional to the pd s current and hence to the received optical signal .
therefore , normalized samples of the amplified photo - detected signal are good representatives of the received optical power normalized samples . ] , is sampled and monitored by an hp infiniium oscilloscope . for each test
, we have collected @xmath76 samples with the sampling rate of @xmath77 . along with our setup
, an alcohol thermometer is attached to the lower side of the tank to ensure that the water temperature remains constant around @xmath78 degrees celsius . finally , to produce air bubbles , a tunable air blower with the maximum blowing capacity of @xmath79 is employed .
[ fig : set_up ] shows our experimental setup .
[ cols= " < , < , < , < , < , < , < , < , < , < , < , < " , ] in this section , we present our experimental results for the channel coherence time and the fading statistical distribution , under various underwater channel conditions . for each scenario
we evaluate the validity of different distributions using the defined metrics in section iii .
the considered channel conditions include free - space link as well as fresh and salty underwater channels with and without bubbles . for each scenario , we adjusted the transmitter laser power to ensure that a considerable power reaches the receiver . figs . [ first_fig](a ) and [ first_fig](b ) illustrate the received signal through the free - space and fresh water links , respectively . as it can be seen , for fso and uwoc channels with low link ranges ( like our water tank ) the received signal is approximately constant over a large period of time , i.e. , fading has a negligible effect on the performance of low - range fso and uwoc systems . as a consequence ,
the channel impulse response for such scenarios can be thoroughly described by a deterministic ffir . in order to investigate salinity effects on the underwater optical channels , we added @xmath80 salt to our water tank .
we observed significant loss on the received signal in comparison to fresh water link .
therefore , we increased the transmitter power to observe a notable signal at the receiver , and hence to better investigate the fading fluctuations due to water salinity .
[ first_fig](c ) shows the received signal through the salty water .
as it is obvious , water salinity does not impose severe fluctuations on the propagating signal .
we also used a water pump to flow water within the tank .
our extra experiments indicated that as the velocity of moving medium increases , the received signal experiences faster and more severe fluctuations .
this is mainly due to the fact that as the water flow increases , particles within the water move more rapidly and this causes more randomness on the transmitting signal path . according to figs .
[ first_fig](a)-(c ) , due to the low link range available by the laboratory water tank , the received signal fluctuations are typically negligible . in other words ,
optical turbulence manifests its effect at longer link ranges , i.e. , for uwoc and fso channels with typically @xmath81 and @xmath82 link ranges , respectively @xcite .
therefore , an air compressor is used to produce bubbles and hence to induce a severe fading on the channel
. figs . [ first_fig](d ) and [ first_fig](e ) illustrates the received signal through the bubbly fresh water , while fig .
[ first_fig](f ) shows the same result for the bubbly and salty underwater link .
as it is obvious from these figures , the presence of bubbles within the channel causes the received signal to severely fluctuate .
this is mainly due to the random arrangement of bubbles encountering the propagating signal across the channel , that causes propagating photons to randomly scatter in different directions and leave their direct path .
table i summarizes the results of @xmath29 test and rmse measure for different statistical distributions and for various channel conditions and @xmath53 is a positive quantity , the evaluation of k distribution is only possible when @xmath2 .
moreover , for very small values of @xmath0 the the gamma - gamma distribution parameters , i.e. , @xmath53 and @xmath58 have large values hampering the analytical evaluation of gamma - gamma distribution for such regimes of scintillation index . ] .
as it can be seen , for very small values of scintillation index , i.e. , @xmath83 log - normal distribution has an excellent goodness of fit and a negligible rmse .
however , both log - normal and gamma - gamma distributions acceptably predict the irradinace fluctuations when @xmath84 , as @xmath0 increases and approaches @xmath1 both of these distributions lose their accuracy
. meanwhile , both k and gamma - gamma distributions can aptly fit the experimental data for strong turbulent channels , characterized by @xmath2 , while log - normal distribution fails to describe the intensity fluctuations for such a regime of scintillation index .
[ hists ] better illustrates the fitness of different statistical distributions with the acquired data histograms for various channel scenarios .
based on the summarized results of table i , neither of the discussed general statistical distributions are capable of predicting the intensity fluctuations when @xmath3 . as the received intensity shapes in figs .
[ first_fig](d)-(f ) and the histograms of the acquired data in figs .
[ hists](d)-(f ) show the presence of air bubbles causes the received intensity to mainly lie either in large or small values .
hence , typical single - lobe distributions can not appropriately fit the experimental data and generally a two - lobe statistical distribution is required .
therefore , we proposed a combined exponential and log - normal distribution as discussed in section iv .
as table i and fig .
[ hists ] illustrate the proposed distribution excellently fits the acquired experimental data for all the regions of scintillation index with an acceptable rmse .
[ tcc ] illustrates the temporal covariance coefficient of irradiance for different scenarios .
as it can be seen , even in strongest fading condition of bubbly channels the aforementioned coefficient is larger than @xmath85 seconds confirming the flat fading , i.e. , the same fading coefficient over thousands up to millions of consecutive bits .
moreover , this figure shows that the channel temporal variation slightly increases with the increase on the scintillation index value , i.e , the higher the value of @xmath0 , the less the value of @xmath86 .
in this paper , we experimentally investigated the statistical distribution of intensity fluctuations for underwater wireless optical channels under different channel conditions .
our experimental results showed that salt attenuates the received signal while air bubbles mainly introduce severe intensity fluctuations .
moreover , we observed that log - normal distribution precisely fits the acquired data pdf for scintillation index values less than @xmath1 , while gamma - gamma and k distributions aptly predict the intensity fluctuations for @xmath2 .
meanwhile , neither of these distributions are capable of predicting the received irradiance for @xmath3 .
therefore , we proposed a combination of an exponential and a log - normal distributions to perfectly describe the acquired data pdf for such regimes of scintillation index .
furthermore , our study on the channel coherence time have shown larger values than @xmath85 seconds for the channel temporal covariance coefficient implying flat fading uwoc channels .
c. gabriel , m .- a .
khalighi , s. bourennane , p. lon , and v. rigaud , `` monte - carlo - based channel characterization for underwater optical communication systems , '' _ journal of optical communications and networking _ , vol . 5 , no . 1 ,
pp . 112 , 2013 .
f. akhoundi , j. a. salehi , and a. tashakori , `` cellular underwater wireless optical cdma network : performance analysis and implementation concepts , '' _ ieee trans .
_ , vol .
63 , no . 3 , pp .
882891 , 2015 .
f. akhoundi , m. v. jamali , n. banihassan , h. beyranvand , a. minoofar , and j. a. salehi , `` cellular underwater wireless optical cdma network : potentials and challenges , '' _ arxiv preprint arxiv:1602.00377 _ , 2016 . m. v. jamali , f. akhoundi , and j. a. salehi , `` performance characterization of relay - assisted wireless optical cdma networks in turbulent underwater channel , '' _ accepted for publication in ieee trans .
wireless commun . _ , 2016 .
m. v. jamali and j. a. salehi , `` on the ber of multiple - input multiple - output underwater wireless optical communication systems , '' in _ 4th international workshop on optical wireless communications ( iwow)_.1em plus 0.5em minus 0.4emieee , 2015 , pp .
2630 .
m. al - habash , l. c. andrews , and r. l. phillips , `` mathematical model for the irradiance probability density function of a laser beam propagating through turbulent media , '' _ optical engineering _ , vol .
40 , no . 8 , pp . 15541562 , 2001 .
m. a. kashani , m. uysal , and m. kavehrad , `` a novel statistical channel model for turbulence - induced fading in free - space optical systems , '' _ j. lightw
. technol .
_ , vol .
33 , no . 11 , pp .
23032312 , 2015 . | in this paper , we experimentally investigate the statistical distribution of intensity fluctuations for underwater wireless optical channels under different channel conditions , namely fresh and salty underwater channels with and without air bubbles .
to do so , we first measure the received optical signal with a large number of samples . based on the normalized acquired data the channel coherence time and the fluctuations probability density function ( pdf ) are obtained for different channel scenarios .
our experimental results show that salt attenuates the received signal while air bubbles mainly introduce severe intensity fluctuations . moreover
, we observe that log - normal distribution precisely fits the acquired data pdf for scintillation index ( @xmath0 ) values less than @xmath1 , while gamma - gamma and k distributions aptly predict the intensity fluctuations for @xmath2 .
since neither of these distributions are capable of predicting the received irradiance for @xmath3 , we propose a combination of an exponential and a log - normal distributions to perfectly describe the acquired data pdf for such regimes of scintillation index .
underwater wireless optical communications , intensity fluctuations , fading statistical distribution , channel coherence time . |
social network analysis is the study of the structure of relationships linking individuals ( or other social units , such as organizations ) and of interdependencies in behavior or attitudes related to configurations of social relations .
the observational units in a social network are the relationships between individuals and their attributes . whereas studies in medicine typically involve individuals whose observations
can be thought of as statistically independent , observations made on social networks may be simultaneously dependent on all other observations due to the social ties and pathways linking them .
accordingly , different statistical techniques are needed to analyze social network data .
the focus of this chapter is _ sociocentric data _ , the case when relational data is available for all pairs of individuals , allowing a fully - fledged review of available methods .
two major questions in social network analysis are : 1 ) do behavioral and other mutable traits spread from person - to - person through a process of induction ( also known as _ social influence _ , _ peer effects _ , or _ social contagion _ ) ; 2 ) what exogeneous factors ( e.g. , shared actor traits ) or endogeneous factors ( e.g. , internal configurations of actors such as triads ) are important to the overall structure of relationships among a group of individuals .
the first problem has affinity to medical studies in that individuals are the observational units . in medicine ,
the health of an individual is paramount and so individual outcomes have historically been used to judge the effectiveness of an intervention .
a study of social influence in medicine may involve the same outcome but the treatment or intervention is the same variable evaluated on the peers of the focal individual ( referred to as _ alters _ ) .
an important characteristic of studies of social influence is that individuals may partly or fully share treatments and one individual s treatment may depend on the outcome of another .
for example , an intervention that encourages person a to exercise in order to lose weight might also influence the weight of a s friends ( b and c ) because they exercise more when around a. hence , a s weight intervention may also affect the weight of b and c. a consequence is that the total effect of a s treatment must also consider it s effect on b and c , the benefit to individuals to whom b and c are connected , and so on . such _ interference _ between observations violates the stable - unit treatment value assumption ( sutva ) that one individuals treatment not affect anothers outcome @xcite , which presents challenges for identification of causal effects .
interference is likely to result in an incongruity between a regression parameter and the causal effect that would be estimated in the absence of interference .
the second problem is important in sociology as social networks are thought to reveal the structure of a group , organization , or society as a whole @xcite .
for example , there has always been great interest in determining whether the triad is an important social unit @xcite .
if the existence of network ties a - b and a - c makes the presence of network tie b - c more likely then the network exhibits _ transitivity _ , commonly described as `` a friend of a friend is a friend '' .
thus , just as an individual may influence or be influenced by multiple others , the relationship status of one dyad ( pair of individuals ) may affect the relationship status of another dyad , even if no individuals are common to multiple dyads .
accounting for between dyad dependence is a core component of many social network analyses and has entailed much methodological research .
network science is a parallel field to social network analysis in that there is very little overlap between researchers in the respective fields despite the similiarity of the problems . whereas solutions to problems in social networks have tended to be data - oriented in that models and statistical tests are based on the data , those in network science have tended to be phenomenon - oriented with analogies to problems in the physical sciences often providing the backbone for solutions .
methods for social network analysis often have causal hypotheses ( e.g. , does one individual have an effect on another , does the presence of a common friend make friendship formation more likely ) motivating them and involving micro - level modeling .
in contrast , methods in network science seek models generated from some theoretical basis that reproduce the network at a global or system level and in so - doing reveal features of the data generating process ( e.g. , is the network scale - free , does the degree - distribution follow a power - law ) .
one of the goals of this chapter is to address the lack of interaction between the social network and network science fields by providing the first joint review of both . by enlarging the range of methods at the disposal of researchers , advances at the frontier of networks and health
will hopefully accelerate .
the computer age has enabled widespread implementation of methods for social network and network science analysis , particularly statistical models . at the same time , a diverse range of applications of social network analysis have appeared , including in medicine . because many medical and health - related phenomena involve interdependent actors ( e.g. patients , nurses , physicians , and hospitals ) , there is enormous potential for social network analysis to advance health services research @xcite .
the layout of the remainder of the chapter is as follows .
this introductory section concludes with a brief historical account of social networks and network science is given .
the major types of networks and methods for representing networks are then discussed ( section 2 ) . in section 3 formal notation
is introduced and descriptive measures for networks are reviewed . social influence and social selection
are studied in sections 4 and 5 , respectively .
our focus switches to methods akin with network science in section 6 , where descriptive methods are discussed .
the review of network science methods continues with community detection methods in section 7 and generative models in section 8 .
the chapter concludes in section 9 . in the 1930 s ,
a field of study involving human interactions and relationships emerged simultaneously from sociology , psychology , and anthropology .
moreno is credited for inventing the sociogram @xcite , a visual display of social structure .
the appeal of the sociogram led to moreno being considered a founder of sociometry , a precursor to the field known as _
social networks_. a number of mathematical analyses of network - valued random variables in the form of sociograms followed @xcite .
other important contributions were to structural balance @xcite ; the diffusion of medical innovations @xcite ; structural equivalence @xcite ; and social influence @xcite
. refer to for a detailed historical account .
early network studies involved small networks with defined boundaries such as students in a classroom , or a few large entities such as countries engaging in international trade . because the typical number of individuals in such studies was small ( e.g. , @xmath0 ) , relationships could be determined for all possible pairs of individuals yielding complete _ sociocentric _ datasets .
furthermore , the often enclosed nature of the system ( e.g. , a classroom or commune ) reduced the risk of confounding by external factors ( e.g. , unobserved actors ) . sociological theory developed over time as sociologists provided intuitive reasoning to support various hypotheses involving social networks and society @xcite . in the specific area of individual health , at least five principal mediating pathways through which social relationships and thus social networks may influence outcomes have been posited @xcite .
prominent among these is social support , which has emotional , instrumental , appraisal ( assistance in decision making ) , and informational aspects @xcite . beyond social support , networks may also offer access to tangible resources such as financial assistance or transportation .
they can also convey social influence by defining norms about such health - related behaviors as smoking or diet , or via social controls promoting ( for example ) adherence to medication regimes @xcite .
networks are also channels through which certain communicable diseases , notably sexually transmitted ones , spread @xcite and certain network structures have been hypothesized to reduce exposure to stressors @xcite . a field known as mathematical sociology complemented social theory by attempting to derive results using mathematical rather than intuitive arguments .
in particular , statistical and probability methods are used to test for the presence of various structural features in the network .
other key areas of mathematics that have been used in network analysis include graph theory and algebraic models .
develop tests of dependence within dyads ( pairs of actors ) while ( ) develop tests of triadic dependence .
in general , results were descriptive or based on simple models making strong assumptions about the network . with the advent of powerful computers ,
mathematical contributions have taken on more importance as so much more can be implemented that in the past .
for example , computer simulation has recently been used to test and develop theoretical results @xcite . in the mid - late 1990s ,
network science emerged as a discipline . whereas social networks were the domain of social scientists and a growing number of statisticians ,
network scientists typically have backgrounds in physics , computer science or applied mathematics .
the use of physical concepts to generate solutions to problems is common as evinced by the large domains of research focusing on the adaptation of ( e.g. ) a particular physical equation to network data .
for example , several procedures for partitioning a network into disjoint groups of individuals ( `` communities '' ) rely on the _ modularity _ equation , which was developed in the context of spin - theory to model the interaction of electrons .
while much of the initial work focused on the properties of the solution at different values of the parameters there recently been has increased attention to using these methods to provide valuable insight on important practical problems .
social networks are comprised of units and the relationships between them .
the units are often individuals ( also referred to as _ actors _ ) but can include larger ( e.g. , countries , companies ) and smaller ( e.g. , organisms , genes ) entities . in sociocentric studies ,
data is assembled on the ties linking all units or actors within some bounded social collective @xcite .
for example , the collection of data on the network of all children in a classroom or on all pairs of physician collaborations within a medical practice constitutes a sociocentric study .
relationships can be shared or directional , and quantified by binary ( tie exists or not ) , scale ( or valued ) , or multivariate variables . by measuring all relationships , sociocentric data
constitutes the highest level of information collection and facilitates an extensive range of analyses including accounting for the effects of multiple actors on actor outcomes or the structure of the network itself to be studied @xcite
. a weaker form of relational data is collected in egocentric studies where individuals ( `` egos '' ) are sampled at random and information is collected on at least a sample of the individuals with direct ties to the egos ( `` alters '' ) . because standard statistical methods such as regression analysis can generally be used to analyze egocentric data , herein egocentric data are not featured .
relational data is often binary ( e.g. , friend or non - friend ) .
one reason is that other types of relational data ( e.g. , nominal , ordinal , interval - valued ) are often transformed to binary due to the convenience of displaying binary networks .
another is the greater range of models available for modeling binary data .
many studies involve two distinct types of units , such as patients and physicians , or physicians and hospitals , authors and journal articles or books , etc . in these two - mode networks
, the elementary relationships of interest usually refer to affiliations of units in one set with those in the other ; e.g. , of patients with the physician(s ) responsible for their care , or of physicians with the hospital(s ) at which they are admitted to practice .
two - mode networks are also known as affiliation or _
_ networks .
they can be viewed as a special - case of general sociocentric network data in that the relationship of interest is between heterogeneous types of actors .
the advent of high - powered computers has enabled the analysis of large networks , which has benefitted fields such as health services research that regularly encounter large data sets .
a challenge facing analyses of large networks is that it may be infeasible for all actors to be exposed to each other actor and thus for a relationship to have formed .
therefore , statistical analyses for large networks essentially use relational data representing the joint event of individuals meeting and then forming a tie , not the network of ties that would be observed if all pairs of individuals actually met .
accordingly , analyses of large networks may underestimate effect sizes unless information on the likelihood of two individuals meeting is incorporated .
let the status of the relationship from @xmath1 to @xmath2 be denoted by @xmath3 , element @xmath4 of the adjacency matrix @xmath5 . in a directed network
@xmath3 may differ from @xmath6 while in a non - directed network @xmath7 , implying @xmath8 .
a network constructed from friendship nominations is likely to be directed while a network of coworkers is non - directed . in the case of immutable relationships ( e.g. , siblings ) ,
@xmath9 will only change as actors are added or removed ( e.g. , through birth or death ) , as relationship status is otherwise invariant . in the following ,
assume the network is binary unless otherwise stated .
matrices and graphs are two common ways of representing the status of a networks at a fixed time . in a matrix representation ,
rows and columns correspond to units or actors ; the matrix is square for one - mode and rectangular for two - mode networks
. elements of the matrix contain the value of the relationship linking the corresponding units or actors , so that element @xmath4 represents the relationship from actor @xmath1 to actor @xmath2 . with binary ties ( 1 = tie present , 0 = tie absent ) ,
the matrix representation is known as an adjacency matrix .
irrespective of how the network is valued , the diagonal elements of the matrix representing the network equal 0 as self - ties are not permitted .
several network properties can be computed through matrix operations . .
note : self - ties are not relevant in studies involving relationships.,width=432 ] in graphical form , units or actors are vertices and non - null relationships are lines .
non - directed relationships are known as `` edges '' and directed ones as `` arcs '' ; arrows at the end(s ) of arcs denote their directionality .
value - weighted graphs can be constructed by displaying non - null tie values along arcs or edges , or by letting thinner and thicker lines represent line values .
such graphical imagery is a hallmark of social network analysis @xcite . two - mode ( or bipartite ) networks may be represented in set - theoretic form as hypergraphs consisting of a set of actors of one type , together with a collection of subsets of the actors defined on the basis of a common actor of the second type @xcite . this representation highlights the multi - party relationships that may exist among those actors of one type that are linked to a given actor of the other type ; e.g. , the set of all physicians affiliated with a particular clinic or service . in matrix form ,
element @xmath4 of an affiliation matrix @xmath9 indicates that actor @xmath1 of the first type is linked to actor @xmath2 of the second type .
affiliation networks may usefully be represented as bipartite graphs in which nodes are partitioned into two disjoint subsets and all lines link nodes in different sets .
an induced one - mode network @xmath9 may be obtained by multiplying an affiliation matrix @xmath10 by its transpose , @xmath11 ; entry @xmath4 of the outer - product @xmath12 gives the number of affiliations shared by a pair of actors of one type ( see figure [ fig : biparproj ] , which emulates a figure in ) .
dually , the inner - product @xmath13 yields a one - mode network of shared affiliations among actors of the second type @xcite .
the diagonals of the outer and inner matrix products give the degree of the actors ( i.e. , the number of ties to actors of the other mode ) . )
bipartite adjacency matrix @xmath10 .
a one - mode projection of the doctor - patient network is obtained by multiplying the bipartite adjacency matrix @xmath10 by its transpose , @xmath14 , to yield a @xmath15 symmetric one - mode adjacency matrix @xmath9 , whose elements indicate the number of patients the two physicians have in common .
the diagonal elements of @xmath9 correspond to the number of patients the given physician `` shares with themselves '' ( i.e. , the number of patients they care for).,width=432 ] in health services applications , an investigator is often interested in a one - mode network that is not directly observed but rather is induced from a two - mode network .
such one - mode projection networks are motivated theoretically by a claim that shared actors from the other mode act as surrogates for ties between the actors .
for example , physicians with many patients in common might have heightened opportunities for contact through consultations or sharing of information about those patients and thus the number of shared patients is a surrogate for the actual extent of interaction between pairs of physicians .
examples of provider ( physician , hospital , health service area ) networks obtained as one - mode projections of bipartite networks in health services research are given in ( ) .
an often overlooked feature of bipartite network analysis is the mechanism by which network data is obtained .
networks obtained from one - mode projections have different statistical properties from directly - observed one - mode networks .
consider a patient - physician bipartite network and suppose a threshold is applied to the physician one - mode projection such that true social ties are assumed to exist or not according to whether one or more patients are shared .
then a patient that visits three physicians is seen to induce ties between all three physicians .
the same complete set of ties between the three physicians is also induced by three patients that each visit different pairs of the three physicians .
however , the projection does not preserve the distinction ( see section [ sec : bipar ] for further comment ) .
the number of units or actors ( @xmath16 ) is known as the order of the network .
a common network statistic is network density ( @xmath17 ) , defined as the number of ties across the network ( @xmath18 ) divided by the number of possible ties ; for directed networks @xmath19 and for non - directed networks @xmath20 .
thus , density equals the mean value of the binary ( 1 , 0 ) ties across the network . the same definition can be used for general relational data , in which case the resulting measure is sometimes referred to as _
strength_. while results in this chapter are generally presented for binary networks , corresponding measures for weighted networks often exist ( ) .
the tendency for relationships to form between people having similar attributes is known as homophily .
homophily involves subgroup - specific network density statistics . with high homophily according to some attribute , networks tend toward segregation by that attribute - the extreme case occurs when the network consists of separate components ( i.e. , no ties between actors in different components ) defined by levels of the attribute . in the other direction , one obtains a bipartite network where all ties are between different types of actors ( extreme heterophily ) .
the out- and in - degree for an actor @xmath1 are the number of ties from , @xmath21 ( column sum ) , and to , @xmath22 ( row sum ) , actor @xmath1 .
these are also referred to as _ expansiveness _ and _ popularity _ , respectively .
for example , a positive correlation between out- and in - degree suggests that popular individuals are expansive .
the number of ties ( or value of the ties ) in a network is given by @xmath23 , where @xmath24 denotes the mean degree ( or strength ) of an individual , implying the density of the network is given by @xmath25 .
this result is not specific to in- or out - degree due to the fact that the total number of inward ties must equal the total number of outward ties , implying mean in - degree equals mean out - degree .
the variance of the degree distribution measures the extent to which tie - density ( or connectedness ) varies across the network @xcite .
often actors having higher degree have prominent roles in the network @xcite .
a special type of homophily is the phenomenon where individuals form ties with individuals of similar degree , commonly referred to as _
assortative mixing_. in directed networks , assortative mixing can be defined with respect to both out - degree and in - degree @xcite .
the opposite scenario to a network with the same degree for all actors is a @xmath26-star a network configuration with @xmath26 relationships are incident to the focal actor ( figure [ fig : config ] ) in which there are no ties between the other actors .
the length of a path between two actors through the network is defined as the number of ties traversed to get from one actor to the other .
the elements of the adjacency matrix multiplied by itself @xmath27 times , denoted @xmath28 , equal the number of paths of length @xmath26 between any two actors with the number of @xmath26-cycles ( including multiple or repeated loops ) on the diagonal .
the shortest path between two actors is referred to as the _ geodesic distance_. certain subnetworks have particular theoretical prominence .
the first step - up from the trivial single actor subnetwork , also known as an isolated node , is the network comprising two actors ( a `` dyad '' ) .
the presence and magnitude of a tendency toward symmetry or reciprocity in a directed network can be measured by comparing the number of mutual dyads ( ties in both directions ) to the number expected under a null model that does not accommodate reciprocity .
if the number of mutual dyads is higher than expected , there is a tendency towards reciprocation .
a _ triad _ is formed by a group of three actors .
figure [ fig : config ] shows a `` transitive triad '' , so - named as it exhibits the phenomenon that a `` friend of a friend is a friend . '' non - parametric tests for the presence of transitivity or other forms of triadic dependence are based on the distribution of the number of closed and non - closed triads conditional on the number of null ( no ties intact ) , directed ( one tie intact ) , and mutual dyads ( both ties intact ) collectively known as the _ dyad census _ ; the degree distribution ; and other lower - order effects ( e.g. , homophily of relevant individual characteristics ) in the observed network . such tests are described in .
centrality is the most common metric of an actor s prominence in the network and many distinct measures exist .
they are often taken as indicators of an actor s network - based `` structural power . ''
such measures are often used as explanatory variables in individual - level regression models .
different centrality measures are characterized by the aspects of an actor s position in the network that they reflect .
for example , degree - based centrality the degree of an actor in an undirected network and in- or out - degree in a directed network reflects an actor s level of network connectivity or involvement in the network .
betweenness centrality computes the frequency with which an actor is found in an intermediary position along the geodesic paths linking pairs of other actors .
actors with high betweenness centrality have high capacity to broker or control relationships among other actors . a third major centrality measure , closeness centrality ,
is inversely - proportional to the sum of geodesic distances from a given actor to all others .
the rationale underlying closeness measures is that actors linked to others via short geodesics have comparatively little need for intermediary units , and hence have relative independence in managing their relationships .
closeness measures are defined only for networks in which all actors are mutually related to one another by paths of finite geodesic distance ; i.e. , single component networks .
finally , eigenvalue centrality is sensitive to the presence or strength of connections , as well as those of the actors to which an actor is linked @xcite .
it assumes that connections to central actors indicate greater prominence than do ( similar - strength ) connections to peripheral actors .
the key component of the measure is the largest eigenvalue of an adjacency or other matrix representation of the network @xcite .
network - level centrality indices @xcite are network - level statistics that resemble the degree variance whose values grow larger to the extent that a single actor is involved in all relationships ( as in the `` star '' network shown in figure [ fig : config ] ) .
the assignment of actors to groups is an important and growing field within social networks .
the rationale for grouping actors is that it may reveal salient social distinctions that are not directly observed .
the general statistical principle adhered to is that individuals within a group are more alike than individuals in different groups .
groups are typically formed on the basis of network ties alone , the rationale being that the similarity of individuals positions in the network is in - part revealed by the pattern of ties involving them .
thus , actors in densely connected parts of the network are likely to be grouped together .
a related concept to a group is a clique , a maximal subset of actors having density 1.0 ( i.e. , ties exist between all pairs of individuals in a binary network ) .
the larger the clique the stronger the evidence that the collective individuals are in the same group .
grouping algorithms based on maximizing the ratio of within - group to between group ties are unlikely to split large cliques as doing so creates a lot of between group ties .
however , a clique need not be its own group . components of a network are defined by the non - existence of any paths between the actors in them . often a network is comprised of one large component and several small components containing few individuals .
a more practical way of grouping individuals than by cliques is through @xmath26-connected components @xcite , a maximal subset of actors mutually linked to one another by at least @xmath26 node - independent paths ( i.e. , paths that involve disjoint sets of intermediary actors who also lie within the subgraph ) .
such a criterion is related to @xmath26-coreness , a measure of the extent to which subgraphs with all internal degrees @xmath29 occur @xcite in a network .
there are several other ways for grouping the actors in a network .
model - based methods include mixed - membership stochastic block models @xcite and latent - class models in which the group is treated as a categorical individual - level latent variable @xcite while non - parametric methods used in network science include _ modularity _ and its variants .
these methods are discussed in section [ sec : physcd ] , where the grouping of actors is referred to as _ community detection_. in practice two - mode networks are rarely directly analyzed .
if one of the modes instigates ties or is of primary interest , the network involving just those actors is often analyzed as a single - mode network .
for example , in a physician - patient referral network , the physicians often instigate ties through patient referrals while patients are chiefly responsible for who they see first . the projection from a two - mode network to a one - mode network links nodes in one mode ( e.g. , physicians )
if they share a node of the other mode ( e.g. , patients ) . a weighted network can be formed with the number of shared actors of the other mode ( or function thereof ) as weights . in describing networks obtained from a projection of a two - mode network ,
the usual practice is to use unipartite descriptive measures .
however , several layers of information are lost , including the number of actors in the other mode underlying a tie and the degree distribution of the actors in the other mode , from treating a one - mode projection as an actual network . even if the two - mode network is completely random , ties in a one - mode projection that arise from a single ( e.g. ) patient with ties to ( e.g. ) three physicians are not separate events .
more generally , a patient who visits @xmath26-physicians generates a @xmath26-clique among those physicians and tells us nothing about whether physician sharing of one patient is correlated with physician sharing of another patient the question of primary interest in the study of the diffusion of treatment practices .
thus , @xmath26-cliques for @xmath30 may be excluded from measures of transitivity in two - mode networks .
descriptive measures for two - mode networks may be computed that parallel those for one - mode networks @xcite .
centrality measures based on the bipartite network representation are covered in .
review visualization , subgroup detection , and measurement of centrality for two - mode network data .
more descriptive measures for two - mode networks have recently been proposed .
for example , a two - mode measure of transitivity defined as the ratio of the total number of six cycles ( closed paths of six ties through six nodes ) in the two - mode network divided by the total number of open five - paths through six nodes @xcite . in the context of the patient - physician network , _ physician transitivity _ exists if physicians a and b sharing a patient and physicians b and c sharing a patient makes it more likely for physicians a and c to share a patient .
it is only if the two pairs of physicians have different patients in common that the physician triad may be transitive and only if the third pair share a different patient from the first two that the event can be attributed to transitivity .
the involvement of distinct patients makes the physician - physician ties distinct events and thus informative about clustering of physicians ( and patients ) . in general , the matrix equation @xmath11 in which a bipartite network adjacency matrix @xmath31 is multiplied by its transpose yields a weighted one - mode network ( the elements contain the number of shared actors of the other mode ) .
to avoid losing information about the number of actors leading to a tie between primary nodes , weights can be retained or monotonically transformed in the projected network .
weighted analogies of descriptive measures of binary networks can be evaluated on the weighted one - mode projection .
for example , the calculation of degree is emulated by summing the weights of the edges involving an individual , yielding their _ strength_. degree and strength together distinguish between actors with many weak ties and those with a few strong ties .
analogous measures of centrality can also be computed for the weighted one - mode projection @xcite . however , whether ties between @xmath26 physicians arise through them all treating the same patient , from each pair of physicians sharing a unique patient , or some in - between scenario can not be determined post - transformation ; thus , the projection transformation expends information .
a further strategy is to set weights for the bipartite network prior to forming the projection .
for example , in co - authorship networks , the tie connecting an author to a publication might receive a weight of @xmath32 where @xmath33 is the number of authors on paper @xmath2 @xcite .
( only papers with at least two authors are used to form such networks . )
the rationale is that the greater the number of authors the lower the expected interaction between any pair ( a similar logic underlies the example weight matrix described in section [ sec : influ ] ) .
the sum of the weights across all publications common to two authors is then the basis of their relationship in the author network .
if the events defining the bipartite network occur at different times ( e.g. , medical claims data often contain time - stamps for each patient - physician encounter ) a directed one - mode network may be formed
. the value of the a - b and b - a ties in the physician - physician network could be the number of patients who visited a before b and b before a , respectively . in the resulting directed network
each physician has a flow to and from each other physician .
subsequent transformation of the flows to binary values yields dyads with states null , directed , and mutual as in a directed unipartite binary network .
because medical claims and surveys are frequent sources of information about one entity s experience ( e.g. , a patient ) with another entity ( e.g. , a health plan or physician ) , bipartite network analysis is an area that promises to have enormous applicability to health services research .
hence , new methods for bipartite network analysis are needed .
we now consider the use of statistical models in social network analysis .
particular emphasis is placed on methods for estimating social influence or peer effects and models for analyzing the network itself , including accounting for social selection through the estimation of effects of homophily .
reported claims about peer effects of health outcomes such as bmi , smoking , depression , alcohol use , and happiness have recently tantalized the social sciences . in large part , the discussion and associated controversies have arisen from the statistical methods used to estimate peer effects @xcite .
let @xmath34 and @xmath35 denote a scalar outcome and a vector of variables , respectively , for individual @xmath36 at time @xmath37 ( @xmath35 includes 1 as its first element to accommodate an intercept ) .
in this section , the relationship status of individuals @xmath1 and @xmath2 from the perspective of individual @xmath1 ( denoted @xmath3 ) , is assumed to be time - invariant . for ease of notation no distinction is made between random variables and realizations of them .
the vector @xmath38 and the matrices @xmath39 and @xmath9 are the network - wide quantities whose @xmath1th element , @xmath1th row , and @xmath4th element contain the outcome for individual @xmath1 , the vector of covariates for individual @xmath1 , and the relationship between individuals @xmath1 and @xmath2 as perceived by individual @xmath1 , respectively .
the representation of an example adjacency matrix , denoted @xmath9 , is depicted in figure [ fig : netrep1 ] .
regression models for estimating peer effects are primarily concerned with how the distribution of a dependent variable ( e.g. a behavior , attitude or opinion ) measured on a focal actor is related to one or more explanatory variables . when behaviors , attitudes or opinions are formed in part as the result of interpersonal influence
, outcomes for different individuals may be statistically dependent .
the outcome for one actor will be related to those for the other actors who influence her or him , leading to a complex correlation structure . in social influence analyses
the weight matrix , @xmath40 $ ] in figure [ fig : netrep2 ] , apportions the total influence acting on an individual evenly across the individuals with whom they have a netwok tie .
typically 1 .
@xmath41 : non - negative weights
@xmath42 : no self - influence .
@xmath43 : weights give relative influences ( because its row - sums equal 1 , @xmath44 is said to be row - stochastic ) .
let @xmath45 denote the influence - weighted average of the outcome @xmath46 across the network after excluding ( i.e. , subtracting ) individual @xmath1 from the set of individuals to be averaged over .
similarly , let @xmath47 denote the vector containing the corresponding influence weighted covariates , often referred to as _
contextual variables_. the most common choice for @xmath44 is the row - stochastic version of @xmath9 . for illustration ,
suppose that @xmath9 is binary ( the elements are 1 and 0 ) . then the off - diagonal elements on the @xmath1th row of @xmath44 equal @xmath48 if @xmath49 and @xmath50 otherwise ( figure [ fig : netrep2 ] ) .
this framework assumes that an individual s alters are equally influential . in general , influence might only transmit through outgoing ties ( e.g. , those individuals viewed as friends by the focal actor - a scenario consistent with figure [ fig : netrep2 ] ) , or might only transmit through received ties ( e.g. , individuals who view the focal actor as a friend ) , or might act in equal or different magnitude in both directions .
( right ) .
a directed edge from @xmath1 to @xmath2 means that node ( or individual ) @xmath1 has a relationship to node @xmath2 while element @xmath4 of @xmath44 quantifies the extent that individual @xmath1 is influenced by individual @xmath2 .
although the mathematical form of influence depicted here assumes that influence only acts in the direction of the edge , influence may in general act in the absence of a tie ( e.g. , people who consider me as a friend might influence me even if i do not consider them a friend).,width=432 ] network - related interdependence among the outcomes may be incorporated in two distinct ways .
first , an outcome for one actor may depend directly on the lagged outcomes or lagged covariates of the alters to whom she or he is linked . for example , consider the model : @xmath51 where @xmath52 is a scalar parameter quantifying the peer effect ; @xmath53 is a @xmath54-dimensional vector of parameters of peer effects acting through the @xmath54 covariates in @xmath55 , @xmath56 is a vector of other regression parameters for the within - individual predictors , and @xmath57 is the independent error assumed to have mean 0 and variance @xmath58 .
the notation used in equation [ eq : dynam1 ] is adopted through this section ; hence , @xmath59 and @xmath60 denote peer effects and within - individual effects , respectively .
equation ( [ eq : dynam1 ] ) is known as the `` linear - in - means model '' @xcite due to the conduit for peer influence being the trait averaged over the alters of each focal actor .
the model has a symmetric appearance in that it contains corresponding peer effects for each of the within individual predictors .
a common alternative model assumes @xmath61 ; in other words , that peer effects only act through the same variable in the alters as the outcome .
another set of variants arises in the case when there are multiple types of alters with heterogeneous peer effects .
such a situation may be represented in a model by defining distinct influence matrices for each type of peer .
let @xmath62 denote the weight matrix formed from the adjacency matrix for the network comprising only alters of type @xmath63 and let @xmath64 for @xmath65 , where @xmath66 is the number of distinct types of alters .
then an extension of the linear - in - means model to accommodate heterogeneous peer effects is : @xmath67 in the special case where @xmath68 and @xmath69 , ( [ eq : dynam2 ] ) reduces to ( [ eq : dynam1 ] ) .
an alternative to ( [ eq : dynam2 ] ) is to fit separate models for each type of peer , which would yield estimates of the overall ( or marginal ) peer effect for each type of peer as opposed to the independent effect of each type of peer above and beyond that of the other types . , which is an intermediary between @xmath70 and @xmath34 .
( because the point made here does nut depend on @xmath71 and @xmath72 they are not depicted . )
if @xmath70 ( or @xmath73 ) is conditioned on , the path @xmath74 is unblocked and therefore confounds @xmath75 , whose effect is the peer effect of interest .
although the dag looks like a digraph of a network , a dag is a different construction.,width=192 ] failing to account for all alters may lead to biased results if the alters are interconnected .
figure [ fig : netconfound ] presents a simple directed acyclic graph ( dag ) , which is a device for determining whether or not an effect is identifiable , involving three individuals @xmath1 , @xmath2 and @xmath26 .
the nodes represent the variables of interest ( a trait measured on each individual such as their bmi ) and the arrows represent causal effects ( the origin of the arrow is the cause and the tip is the effect ) .
consider the peer effect of individual @xmath2 at @xmath76 on individual @xmath1 at @xmath77 .
a causal effect is identifiable if it is the only unblocked path between two variables .
because individual @xmath26 is a cause of both individual @xmath2 and individual @xmath1 , the peer effect of @xmath2 on @xmath1 will be confounded by individual @xmath26 unless unless the analysis conditions on @xmath70 .
the scenario depicted in figure [ fig : netconfound ] does not present any major difficulties as long as effects involving individual @xmath26 are accounted .
however , if individual @xmath26 is not known about or is ignored , then the analysis may be exposed to unmeasured confounding .
this point has particular relevance to social network analyses as networks are often defined by specifying boundaries or rules for including individuals as opposed to being finite , closed systems @xcite . in situations where such boundaries break true ties ,
influential peers may be excluded , potentially leading to biased results . from a practical standpoint , it may be infeasible to use a model with only lagged predictors such as ( [ eq : dynam1 ] ) .
for instance , the time points might be so far apart that statistical power is severely compromised .
therefore , it is tempting to use a model with contemporaneous predictors such as : @xmath78 where adjusting for @xmath79 seeks to isolate the peer effect acting since @xmath76 .
however , because @xmath80 is correlated with the outcome variables of other observations , ols will be inconsistent .
therefore , methods to account for the endogeneity arising from the correlation between @xmath80 and @xmath81 for @xmath82 in network science parlance the state of @xmath80 is said to be an internal product or consequence of the system as opposed to an external ( exogenous ) force . in ,
the most widely cited of the christakis - fowler peer effect papers , the endogeneity problem is resolved using a novel theoretical argument .
they purported that it is reasonable to assume in a friendship network that the influence acting on the focal actor ( the ego ) is greatest for mutual friendships , followed by ego - nominated friendships , followed by alter - nominated friendships , and finally is close to 0 in dyads with no friendships .
furthermore , they reasoned that because unmeasured common causes should affect each dyad equally . because the estimated peer effects declined from large and positive for mutual friendships to close to 0 for alter and null friendships , consistent with their theory ,
it was suggested that this constituted strong evidence of a peer effect . despite the compelling argument , revealed that unobserved factors affecting tie - formation ( homophily ) may confound the relationship and thus lead to biased effects .
the estimation of peer effects is a topic of ongoing vigorous debate in the academic and the popular press .
alternative approaches to the theory - based approach of christakis and fowler are now described .
a parametric model - based solution to endogenous feedback is to specify a joint distribution for @xmath83 .
then the reduced form of the model satisfies @xmath84 for @xmath38 to yield @xmath85 .
the resulting model emulates a spatial autocorrelation model @xcite .
one way of facilitating estimation is by specifying a probability distribution for @xmath86 .
however , relying on the correctness of the assumed distribution for identification may make the estimation procedure sensitive to an erroneous assumed distribution .
a semi - parametric solution is to find an instrumental variable ( iv ) , @xmath87 ; a variable that is related to @xmath80 but conditional on @xmath80 and @xmath88 does not cause @xmath34 . if @xmath89 is excluded from ( [ eq : dynamcf ] ) , its elements can potentially be used as ivs @xcite .
however , iv methods can be problematic if the instrument is weak or if the assumption that the iv does not directly impact @xmath34 ( the exclusion restriction ) is violated , an untestable assumption .
thus , in fitting a model with contemporaneous peer effects , one faces a choice between assuming a multivariate distribution holds , relying on the non - existence of unmeasured confounding variables , or relying on the validity of an iv .
none of these assumptions can be evaluated unconditionally on the observed data .
while joint modeling and iv methods provide theoretical solutions to the estimation of contemporaneous peer effects , the notion of causality is philosophically challenged when the cause is not known to occur prior to effect . therefore , longitudinal data provide an important basis for the identification of causal effects , in particular in negating concerns of reverse causality . if the observation times are far apart the use of lagged alter predictors may , however , substantially reduced the power of an analysis .
if the dyads consist of mutually exclusive or isolated pairs of actors there are no inter - dyad ties and influence only acts within dyads .
an example of such a situation occurs when individuals can have exactly one relationship and the relationship is reciprocated , as is the case with spousal dyads .
the network influence models of section [ sec : influ ] reduce to _ dyadic influence models _ in which the predictors are based on individual alters . for example , the dyadic influence model analogous to ( [ eq : dynamcf ] ) is obtained by replacing the subscript @xmath90 with @xmath2 .
that is , @xmath91 the model in ( [ eq : dyad1 ] ) may be estimated using generalized estimating equations ( gee ) , avoiding specifying a distribution for @xmath57 . however ,
if any relationships are bidirectional , standard software packages will yield inconsistent estimates of the peer effects as they do not account for the statistical dependence introduced by individuals who play the dual role of ego and alter at time @xmath77 @xcite .
there has recently been a lot of interest and discussion concerning causal peer effects .
issues that have been discussed include the use of ordinary least squares ( ols ) for the estimation of contemporaneous peer effects @xcite and the identification of peer effects independent of homophily @xcite .
the discussion has helped elevate social network methodology to the forefront of many disciplines .
for example , show that ols still provides a valid test of the null hypothesis that the peer effect is zero when the true peer effect is zero .
therefore , ols can be used to test for peer effects despite the fact that ols estimates are inconsistent under the alternative hypothesis .
use tie directionality to account for unmeasured confounding variables under the assumption that their effect on relationship status is the same for all types of relationships .
the rationale is that the estimated peer effect in dyads where the relationship is not expected to be conducive to peer influence ( `` control relationships '' ) provides a baseline against which to identify the peer effect for other types of relationships .
however , this test fails to offer complete protection against unmeasured homophily @xcite , reflecting the vulnerability of observational data to unmeasured sources of bias .
however , sensitivity analyses that evaluate the effect - size needed to overturn the results may be conducted to help support a conclusion by illustrating that the confounding effect must be implausibly large to reverse the finding @xcite .
instrumental variable ( iv ) methods have also been used to estimate peer effects .
a common source of instruments is alters attributes other than the one for which the peer effect is estimated @xcite .
potential ivs must predict the attribute of interest in the alter but must not be a cause of the same attribute in other individuals .
attributes that are invisible such as an individuals genes appear to be ideal candidate genes . for instance , an individual with two risk alleles of an obesity gene is at more risk of increased bmi but conditional on that individual s bmi their obesity genes should not affect the bmi of other individuals .
however , if the obesity genes are revealed through another behavior ( a phenomenon known as _ pleiotropy _ ) that is associated with bmi then , unless such factors are conditioned on , genes will not be valid ivs .
sociocentric network studies assemble data on the ties representing the relationship linking a set of individuals , such as all physicians within a medical practice .
models for such data posit that global network properties are the result of phenomena involving subgroups of ( most commonly ) four or fewer actors @xcite .
examples of such regularities are actor - level tendencies to produce or attract ties ( homophily and heterophily ) , dyadic tendencies toward reciprocity , and triadic tendencies toward closure or transitivity .
a relational model , in essence , specifies a set of micro - level rules governing the local structure of a network . in this section , models for cross - sectional relational data
are consider first followed by longitudinal counterparts of them .
the simplest models for sociocentric data assume dyadic independence . under the random model ,
all ties have equal probability of occurring and the status of one has no impact on the status of another @xcite .
more general dyadic models were developed in and later were extended in . because independence is still assumed between dyads , the information from the data about the model parameters accumulates in the form of a product of the probability densities for the status of the dyadin observation over each dyad : @xmath92 where @xmath93 and @xmath94 are vectors of actor - specific parameters representing the actors expansiveness ( propensity to send ties ) and popularity ( propensity to receive ties ) , respectively , and @xmath95 is a vector of covariates relevant to @xmath3 ( this may include covariates specific to either actor and combined traits of both actors ) .
it is important to realize that covariates can be directional ; thus , @xmath95 need not equal @xmath96 .
although the model may include other parameters , @xmath97 and @xmath98 play an important role in network analysis due to their relationship to the degree distribution of the network and so are explicitly denoted .
when relationship status is binary , the distribution of @xmath99 is a four - component multinomial distribution .
the probabilities are typically represented in the form of a generalized logistic regression model ( an extension of the logistic regression model to @xmath100 categories ) having the form @xmath101 where @xmath102 and @xmath103 , @xmath104 and @xmath105 are functions of @xmath106 and @xmath107 .
the term @xmath103 includes factors associated with the likelihood that @xmath108 but not necessarily the likelihood that @xmath109 . in an non - directed network
the predictors can be directional and so it is likely that @xmath110 .
however , the only covariates included in @xmath105 must be non - directional as they affect the likelihood of @xmath111 ; the sign of @xmath105 indicates whether a mutual tie is more ( if @xmath112 ) or less ( if @xmath113 ) likely to occur than predicted by the density terms and so is a measure of _ reciprocity _ or _
mutuality_. null mutuality is implied by @xmath114 . in dyadic models , the terms @xmath103 , @xmath104 and @xmath105 account for the local network about actors @xmath1 and @xmath2 through the inclusion of @xmath106 .
furthermore , other effects can be homogeneous across actors or actor - specific .
for example , the p@xmath115 model @xcite assumes @xmath116 and @xmath117 , implying the covariate - free joint probability density function of the network given by @xmath118 where @xmath119 , @xmath120 , @xmath121 , and @xmath122 .
thus , the p@xmath115 model depends on @xmath123 network statistics and associated parameters .
if the p@xmath115 model holds within ( ego , alter)-shared values of categorical attributes , a stochastic block model is obtained by allowing block - specific modifications to the density and reciprocity of ties @xcite .
an extension would allow reciprocity to also vary between blocks .
because the stochastic blockmodel extension of the p@xmath115 model is saturated at the actor - level due to the expansiveness and popularity fixed effects , no assumption is made about differences in the degree - distributions of the actors in different blocks .
stochastic block models are the basis of mixed - membership and other recent statistical approaches for node - partitioning social network data @xcite .
individuals in the same block of a stochastic block model are often referred to as being structurally equivalent .
a criticism of dyadic independence models is that they fail to account for interdependencies between dyads .
the @xmath124 or exponential random graph model ( ergm ) generalizes dyadic independence models @xcite .
an ergm has the form @xmath125 where @xmath9 denotes a possible state of the network , @xmath126 denotes a network statistic evaluated over @xmath9 ( e.g. , the number of ties , the number of reciprocated ties ) , @xmath127 and @xmath128 is the set of all @xmath129 possible realizations of a directed network . in general , the scale factor @xmath130 that sums over each distinct network does not factor into a product of analogous terms . as a result , it is computationally infeasible to exactly evaluate the likelihood function of dyadic dependent ergms for even moderately - sized @xmath16 ( e.g. , @xmath131 is problematic @xcite ) .
the key feature of the p@xmath115 model that allows the probability of the network to decompose into the product of dyadic - state probabilities is that it only depends on network statistics @xmath126 that sum individual ties or pairs of ties from the same dyad .
if dyads are independent unless they share an actor , the network is a _
markov random graph _
markov random graphs may include terms for density , reciprocity , transitivity and other triadic structures , and @xmath26-stars ( equivalent to the degree distribution ) these terms contain sums of the products of no more than three ties .
such terms may be multiplied with actor attribute variables to define interaction effects .
( an interaction is the effect of the product of two or more variables ; e.g. , if males and females have different tendencies to reciprocate ties then gender is said to interact with reciprocity . )
networks that extend markov random graphs by allowing four - cycles but no fifth- or higher - order terms are _ partially conditionally dependent_. in such networks , a sufficient condition for dependence of @xmath3 and @xmath132 is that @xmath133 or @xmath134 .
thus , two edges may be dependent despite not having any actors in common .
partial conditional dependence is the basis of the new parameterizations of network statistics developed by @xcite that have led to better fitting ergms ( see below ) . under ergms ,
the conditional likelihood of each tie given the other ties in the network has the logistic form : @xmath135^{-1 } , \label{eq : condlike}\ ] ] where @xmath136 is @xmath9 with @xmath3 excluded , @xmath137 is the vector of changes in network statistics that occur if @xmath3 is 1 rather than 0 .
thus , the parameters of an ergm are interpreted as the change in the log of the odds that the tie is present to not being present conditional on the status of the rest of the network @xcite .
a large positive parameter suggests that more configurations of the type represented in the network statistic appear in the observed network more often than expected by chance , all else equal @xcite . due to the factorization of the likelihood function in ( [ eq : likefn ] ) , likelihood - based estimators of dyadic independence models
have desirable statistical properties such as consistency and statistical efficiency .
however , if the model for the network includes predictors based on three or more actors , no such factorization occurs and markov chain monte carlo ( mcmc ) is required to optimize the likelihood function for ( [ eq : ergm ] ) , which for each observation involves making computations on @xmath138 ( @xmath139 if directed and @xmath140 if non - directed ) distinct networks .
ergms have been demonstrated to be estimable on networks with @xmath141 @xcite , but computational feasibility depends on the terms in the model and the amount of memory available .
the ergm ( `` exponential random graph model '' ) package that is part of the statnet suite in r , developed by the statnet project , estimates ergms .
other estimation difficulties include failure of the optimization algorithm to converge and the fitted model producing nonsensical `` degenerate '' predicted networks . _
degeneracy _ arises because for certain specifications of @xmath126 the network statistics are highly collinear or there is unaccounted effect heterogeneity across the network . as a result , under the fitted model the local neighborhood of networks around the observed network may have probability close to 0 and those networks with positive probability ( often the empty and complete graphs )
may be radically different from each other and thus the observed network .
although the average network over repeated draws has similar network statistics to the observed network , the individual networks generated under the fitted model do not bear any resemblence to the observed network . because an actor of degree @xmath142 contributes @xmath26-stars for @xmath143 ,
@xmath26-star configurations are nested within one another and thus are highly correlated .
therefore , when multiple @xmath26-stars are predictors , extensive collinearity results .
however , the estimated coefficients of successive @xmath26-star configurations ( e.g. , 2-star , 3-star , 4-star ) tend to decrease in magnitude and have alternating signs , an observation often seen when multiple highly colinear variables are included in a regression model .
this observation led to the development of the _ alternating @xmath26-star _
@xcite , given by @xmath144 where @xmath145 denotes the number of @xmath26-stars , being used in place of multiple individual @xmath26-star terms in ( [ eq : ergm ] ) .
a positive estimate of the coefficient of as(@xmath146 ) suggests that the degree distribution is skewed towards higher degree nodes while a negative coefficient implies large degrees are unlikely .
the value of @xmath146 can be specified or estimated from the data @xcite .
network statistics for triadic configurations
the triangle ( a non - directed closed triad ) in non - directed networks and transitive triads , three - cycles , closed three - out stars , closed three - in stars in directed networks are the most prone to degeneracy .
one reason is that heterogeneity in the prevalence of triads across the network , leads to heterogeneity in the density of ties across the network @xcite .
a model that assumes homogeneous triadic effects across the network is unable to describe networks with regions of high and low density ; the generated networks are either dominated by excessive low density regions or by excessive high density regions .
this observation suggests a hierarchical modeling strategy where the first step is to use a community detection algorithm ( see section [ sec : physcd ] ) to partition the network into blocks of nodes .
then fit an ergm ( or other model ) to the sub - network corresponding to each community , allowing the network statistics to have different effects within each community . the just - described modeling strategy combines methods of network science and social network analysis .
a similar approach has been used to overcome severe computational difficulties that often occur when one or multiple triadic ( triangle - type ) terms are included in the model .
a @xmath26-triangle is a set of @xmath26 triangles resting on a common base .
for example , if individuals @xmath1 , @xmath2 , and @xmath26 are one closed triad and individuals @xmath1 , @xmath2 , and @xmath147 are another then the four individuals form a 2-triangle with the edge
@xmath148 common to both . let @xmath149 denote the number of @xmath26-triangles in the network .
thus , @xmath150 denotes the total number of closed triads , @xmath151 the total number of 2-triangles , and so on .
the _ alternating @xmath26-triangle statistic _
@xmath152 was developed to perform for triadic structures what as(@xmath146 ) performs for @xmath26-stars
@xcite . the presence of @xmath146 makes at(@xmath146 ) nonlinear in the triangle count , giving lower probability to highly clustered structures . by making the number of actors who share @xmath26 partners the core term , at(@xmath146 )
can be re - written as a geometrically weighted edgewise shared partner ( gwesp ) statistic @xcite .
the as(@xmath146 ) and at(@xmath146 ) statistics do not differentiate between outward and inward ties .
recently , directed forms of these statistics have been introduced @xcite .
the directed versions of the @xmath26-star are threefold , corresponding to two - paths , shared destination node ( activity ) , shared originator node ( popularity ) .
the directed versions of the @xmath26-triangle represent transitivity , activity closure , popularity closure , and cyclic - closure .
an alternative approach to modeling a one - mode projection ( by construction a non - directed network ) from a two - mode network is to directly model the two - mode network .
an advantage of direct modeling is that all the information in the data is used .
ergms or any other model applied to bipartite data need to account for the fact that ties can only form in dyads including one actor from each mode . in a dyadic independence model
this is recognized simply by excluding all same mode dyads from the dataset . in general , the denominator @xmath130 in ( [ eq : ergm ] ) only sums over networks in which there are no within mode ties . if the number of actors in the two modes are @xmath16 and @xmath153 , there are @xmath154 distinct non - directed networks .
the density and degree distributions may be represented in a bipartite ergm as in a unipartite ergm .
however , with two modes it may be that two types of each network statistic and other predictor is needed .
representations of homophily in two - mode networks are defined across modes . likewise , because there are no within mode ties , statistics that account for closure must also depend only on inter - mode ties .
the smallest closed structure in a bipartite graph is a four - cycle ( closed four - path ) .
an example of a four - cycle is the path a1c2a in figure [ fig : biparproj ] ; it includes four distinct actors and four edges are traversed to return to the initial actor .
a simple measure of closure contrasts the number of closed four - cycles out of all three paths containing four unique actors with the overall density of ties . a simple model for testing whether clustering ( closure ) is present in a bipartite network includes density , both sets of @xmath26-stars , three - path , and four - cycle statistics as predictors .
a significant positive effect of the four - cycle statistic suggests that two actors of degree two in one mode that have one of the actors in the other mode in common are more likely to also have the second actor in common , relative to two randomly selected actors of degree two from the same mode .
for example , in a physician - patient network , clustering implies having one patient in common increases the likelihood of having another patient in common .
physicians a and c both have patients 1 and 2 in common , hence they provide evidence for bipartite closure .
however , physicians e and f have patient 3 in common ; despite being eligible to exhibit bipartite closure they do not , hence they provide evidence against bipartite closure .
analogies of ergms and solutions to problematic issues exist for bipartite networks .
for example , to avoid problems of high colinearity between the @xmath26-star terms , alternating @xmath26-star statistics can be used in place of them .
let @xmath155 denote the number of ties from one mode to the other , @xmath156 and @xmath157 denote the alternating @xmath26-star statistics for each mode , @xmath158 denotes the number of three - paths , and @xmath159 denote the number of closed four - cycles for a network @xmath10 .
the resulting bipartite ergm for @xmath10 has the form : @xmath160 where @xmath130 sums over the @xmath161 possible bipartite graphs .
the statistic @xmath162 is the proportion of times that two patients each visit the same two physicians out of all the occurrences where two patients both have one visit to one physician and one patient visits the other physician .
the coefficient @xmath163 is the effect associated with this lowest - order form of closure in a two - mode sense ( but should not be thought of as reciprocity because the network is non - directed ) .
the development of relational models has primarily focused on cross - sectional data .
however , extensions of ergms to longitudinal scenarios have been developed
most often involving a markov assumption to describe dependence across time .
the first longitudinal ergms treated tie - formation and tie - dissolution as equitable events in the evolution of the network @xcite .
a more general formulation treats tie - formation ( attractiveness in the context of network science ) and tie - duration ( the complement of tie - duration referred to as fitness in network science ) as separable processes , thereby allowing the same network statistic to impact tie - formation and tie - dissolution differently @xcite .
like ergms for cross - sectional data , longitudinal ergms are defined by statistics that count the number of occurrences of substructures in the network . however , in addition to the current state of the network , such statistics may also depend on previous states . under markovian dependence , network statistics only depend on the current and the most recent state ; for example , the number of ties that remain intact from the preceding observation .
the recently released tergm ( `` temporal exponential random graph model '' ) package in the statnet suite in r estimates ergms for discrete temporal ( i.e. , longitudinal ) sociocentric data @xcite .
an alternative approach for modeling network evolution is the actor - oriented model @xcite .
this centers on an objective function that actors seek to maximize and which may be sensitive to multiple network properties , including reciprocity , closure , homophily , or contact with high - degree actors .
the model assumes that actors control their outgoing ties and change them in order to increase their satisfaction with the network in one or more respects as quantified by the objective function .
it resembles a rationale choice model in which each agent attempts to maximize their own utility function .
estimated parameters indicate whether changes in a given property raise or lower actor satisfaction .
an important distinction of actor - oriented models from ergms is that the relevant network statistics in the actor - oriented model are specific to individuals rather than being aggregations across the network .
however , like ergms , estimation is computationally intensive .
the siena package in stocnet @xcite uses a stochastic approximation algorithm but struggles with networks of appreciable size ( e.g. , thousands of individuals ) .
because they only resemble ergms in the limiting steady - state case , actor - oriented models may also suffer from degeneracy but the problem is less profound @xcite .
a virtue of the actor - oriented modeling framework in siena is that an actor s relationships can be modeled jointly with the social - influence effects of an actor s peers on their own traits . if the model is correctly specified , it has the potential to account for unmeasured confounding factors that affect both the evolution of relationship status and the values of individuals attributes , yielding unbiased estimates of the effects of observed variables affecting social influence and the evolution of the network .
such a model was developed by steglich and colleagues @xcite but to date work in this area is limited . in ergms
a huge increase in computational complexity occurs between the dyadic independent and dyadic dependent models .
a second concern about ergms is that in general they are not consistent under sampling in the sense that statistical inferences drawn from the network for the sample do not generalize to the full network @xcite .
the few ergms to exhibit such consistency include the dyadic independent p@xmath115 and stochastic block models .
an alternative modeling strategy provides a more graduated transition between independence and dependence scenarios by using random effects to model dyadic dependence and also ensures consistency between the results of analyzing the sample and the population of interest .
random effects are used to account for dyadic independence in the p@xmath164 model @xcite introduced below . the p@xmath164 model is much like the p@xmath115 model except that the expansiveness @xmath165 and popularity @xmath166 parameters are random as opposed to fixed effects .
typically , @xmath167 is assumed to be bivariate normal with covariance matrix @xmath168 .
therefore , the p@xmath164 model is given by @xmath169 @xmath170 and @xmath171 .
thus , @xmath172 and @xmath173 includes a subset of covariates that are symmetric ( @xmath174 ) in reflection of the fact that reciprocity is a symmetric phenomenon .
conditional on @xmath167 the model implies that the relationship status of one dyad does not depend on that of another .
a positive off - diagonal element of @xmath168 implies that expansive individuals also tend to be popular .
the p@xmath164 model can be extended to account for more general forms of dyadic dependence than the latent propensity of an individual to send or receive ties .
let each individual have a vector of latent variables , denoted @xmath175 in the case of individual @xmath1 , that together with the same for individual @xmath2 affects the value of the relationship between @xmath1 and @xmath2 .
the dependence of tie - status on @xmath176 is generally represented using a simple mathematical function .
the major types of models are latent class models @xcite , latent distance models @xcite , and latent eigenmodels @xcite .
these models are characterized by the form of the latent variable @xmath177 which is included as an additional predictor in @xmath103 . in ( [ eq : latvble ] ) the form and interpretation of @xmath176 changes from denoting a scalar @xmath178 categorical latent variable in the latent class model ( first row ) , to a position in a continuously - valued multi - dimensional space in the latent distance and latent eigenmodels ( second and third rows , respectively ) .
the term @xmath178 can be added to either the @xmath103 or @xmath105 components of the p@xmath164 model to allow higher - order dependence to moderate the effect of density and reciprocity , respectively . in the latent class specification
the array of values of @xmath179 form a symmetric @xmath180 matrix @xmath181 .
a basic specification is @xmath182 if @xmath183 ( nodes in same partition ) and @xmath184 if @xmath185 ( nodes in different partitions ) ( ) .
latent class models extend stochastic - block models to allow latent clusters as well as observed clustering variables .
this family of models is suited to network data exhibiting structural equivalence ; that is , under the model individuals are hypothesized to belong to latent groups such that members of the same group have similar patterns of relationships . in the latent distance specification the most common values for @xmath186 are 1 and 2 , corresponding to absolute and cartesian distance , respectively .
the distance metric accounts for latent homophily the effect of unobserved individual characteristics that induce ties between individuals . in this model
, @xmath176 can be interpreted as the position of individual @xmath1 in a social space @xcite .
this model accounts for triadic dependence ( e.g. , transitivity ) by requiring that latent distances between individuals obey the triangle inequality .
latent distance models are available in the latentnet package in r @xcite .
the latent eigenmodel is the most general specification and accounts for both structural equivalence and latent homophily .
furthermore , the parameter space of the latent eigenmodel model of dimension @xmath187 generalizes that of the latent class model of the same dimension and weakly generalizes the latent distance model of dimension @xmath188 .
conversely , the latent distance model of dimension @xmath187 does not generalize the one - dimensional latent eigenmodel model @xcite . the closeness of the latent factors @xmath189 and @xmath190 quantifies the structural equivalence of actors @xmath1 and @xmath2 positions in the network ; a tie is more likely if @xmath189 and @xmath190 have a similar direction and magnitude , allowing for more clustering than under ( [ eq : p2mod ] ) . on the other hand ,
latent homophily is accounted for by the diagonal elements of @xmath191 , which can be positive or negative ( allowing for heterophily as well as homophily ) .
the model constrains the extent to which the quadratic - forms @xmath192 , @xmath193 , and @xmath194 constructed from the latent vectors vary from one another .
the greater the magnitude of @xmath195 the greater the extent to which ties are expected to cluster and form cliques .
the latent eigenmodel model is appropriate if a network exhibits clustering due to both structural equivalence and unmeasured homophily . in and models
are specified at the tie - level with reciprocity ( in directed networks ) represented as the within - dyad correlation between two tie - specific latent variables . modeling reciprocity as a latent process differs from the p@xmath164 model , in which reciprocity
is represented as a direct effect @xcite .
therefore , an alternative family of latent variable models for networks is obtained by augmenting the density term in the p@xmath164 model with ( [ eq : latvble ] ) .
an advantage of specifying a joint model at the dyad level is that the resulting ( extended - p@xmath164 ) model involves @xmath196 fewer latent variables , possibly alleviating computational issues such as non - identifiability of parameters or multiple local optima . the challenges of estimating models involving latent variables resemble those of factor analysis or other dimension - reduction methods .
first , an appropriate value of @xmath187 may not be able to be specified from existing knowledge of the network and estimating @xmath187 from the data is not straightforward .
second , computational challenges in estimating the latent variables can make the method difficult to apply to large networks .
however , such issues are more easily overcome than degeneracy in ergms .
degeneracy is avoided in these models as the model for a dyad determines the distribution of the network . in other words ,
the factorization of the likelihood into a product of like terms ensures asymptotically that networks sampled under the model are almost surely in the neighborhood of the observed network , increasingly so as @xmath16 increases .
another contrast with ergms is that the model describes a population as opposed to the single observed network .
thus , in latent variable models the data generating process is modeled whereas ergms are specific to the observed network and so have more in common with finite population inference .
another advantage of conditional independence models over ergms is that the same types of models can be applied to valued relational data .
analogous to generalized linear models , the link function and any parametric distributions assumptions that define a conditional independence network model can be tailored to the type of relationship variable ( scale , count , ratio , categorical , multivariate ) . however
, a recent adaptation of ergms has been proposed for modeling count - valued sociocentric data @xcite .
offsetting the above advantageous features of conditional independence models is that terms such as @xmath178 are limited from the hypothesis testing and interpretational standpoint in that they do not distinguish particular forms of social equivalence or latent homophily .
for example , the effect of transitivity is not distinguished from that of cyclicity or higher - order clustering , such as tetradic closure .
therefore , the choice of model in practice might depend on the importance of testing specific hypotheses about higher - order effects to obtaining a model whose generative basis allows it to make predictions beyond the data set on which the model was estimated .
longitudinal counterparts of conditional independence models are obtained by introducing terms that account for longitudinal dependence ( e.g. , past states of the dyad ) .
a simple markov transition model was developed in with tie - formation and tie - dissolution treated as unrelated processes .
conditional on the past state of the dyad and the sender and receiver random effects , the value of each tie is assumed to be statistically independent of that of any other tie .
a more general formulation extends the p@xmath164 model , allowing dependence between ties within a dyad ( reciprocity ) , heterogeneous effects in the formation and dissolution of ties , and the inclusion of higher - order effects ( e.g. , third - order interactions to account for transitivity ) as lagged predictors @xcite .
the approach in is notable for attempting to capture the best of both worlds : it allows localized ( actor or dyadic ) versions of the higher - order predictors available in ergms to be included as predictors , but avoids degeneracy by using their lagged values as opposed to their current values as predictors .
therefore , conditional on the observed and latent predictors , dyads are cross - sectionally independent but longitudinally dependent on prior states of other dyads ( in addition to their own past states ) in the network .
an extension that builds on is to incorporate the latent class , distance or eigenfactor terms in ( [ eq : latvble ] ) in the model .
such a model was entertained in but has not yet been developed .
we now switch attention to methods that have been derived and used in the field of network science . in general ,
network science approaches avoid assumptions about distributions in models .
for example , to test whether a network exhibits a certain property , the commonly - employed approach is to use a permutation test to develop a null distribution for a statistic that embodies the property in question and then evaluate how extreme the observed value of the statistic is with respect to the null distribution .
this technique is the cornerstone of the procedure used to evaluate the degree of separation to which social clustering can be detected in .
network science focuses not only on social networks , but also covers information networks , transportation networks , biological networks , and many others .
most of the networks studied within network science are non - directed as ties are typically thought of as connections as opposed to measures for which the distinction between instigator and receiver is relevant .
thus , the networks in this section are assumed to be non - directed unless stated otherwise .
network science has taken a somewhat different approach to modeling networks than the social sciences or statistics .
essentially all models developed within network science are _ generative models _ , sometimes also known as forward models , in contrast to probabilistic models such as ergms .
these models start from a set of simple hypothesized mechanisms , often functioning at the level of individual nodes and ties , and attempt to describe what types of network structures emerge from a repeated application of the proposed mechanisms .
many of the models describe growing networks , where one starts from a small connected seed network consisting of a few connected nodes , and then grows the network by subsequent addition of nodes , usually one at a time .
the _ attachment rules _ specify how exactly an incoming node attaches itself to the existing network .
generative models are commonly exploratory in nature .
if they reproduce the type of structure observed in an empirical network , it is plausible that the proposed mechanisms may underlie network formation in the real world .
the main insight to be gained from a generative model is a potential explanation for why a network possesses the type of structure it does .
many of the models are simple in nature , which occasionally leads to analytical tractability , but the main reason for simplicity is the potential to expose clearly the main mechanism(s ) driving the phenomenon of interest .
it is not uncommon for generative models to possess only two or three parameters , yet occasionally simple generative mechanisms can explain some of the key features surprisingly well .
once a model can explain the main features , it can be fine - tuned by adding more specific or nuanced mechanisms .
a few examples of generative models are now described .
cumulative advantage refers to phenomena where success seems to breed success , such as in the case of accumulation of further wealth to already wealthy individuals . in networks of scientific citations ,
where a node represents a scientific paper , each node has some number of edges pointing to nodes that correspond to cited papers @xcite . in the present context ,
for example , there would be an edge pointing from the node representing this chapter to the node representing the 1965 _ science _ paper of price . while the out - degree of nodes is fairly uniform , as the length of bibliographies is fairly constrained , the in - degree distribution was found to be fat - tailed with the functional form of a power - law , @xmath197 @xcite .
price later proposed a mathematical model for cumulative advantage processes , `` the situation in which success breeds success '' @xcite .
in this model , nodes are added to the network one at a time , and the average out - degree of each node is fixed .
the attachment rule in the model specifies that each new paper will cite existing papers with probability proportional to the number of citations they already have .
thus each incoming node will attach itself with some number of directed edges to the existing network , the exact number of ties being drawn from a distribution , and the nodes these new edges are pointing to will be chosen proportional to their in - degree . in this formulation , however , papers with exactly zero citations can never accrue citations . to overcome this problem , one can either consider the original publication as the first citation so that each paper starts with one citation or , alternatively , add a small constant to the number of citations @xcite . either way , the outcome is that the target nodes are chosen in proportion to their in - degree plus this small positive constant
. a derivation of the resulting in - degree distribution is given by newman @xcite . denoting the average out - degree of a node by @xmath186 and using @xmath198 to denote the small positive constant , the in - degree distribution @xmath199 for large values of @xmath26 has the power - law form @xmath197 , where @xmath200 .
this simple model ( although the derivation of the result is quite involved ) is able to reproduce the empirical citation ( in - degree ) distribution for scientific papers with surprising accuracy given that the model only contains two parameters .
it may seem odd that the model does not incorporate any notion of paper quality , which surely should be an important driver of citations .
here it is important to notice that the model does not make any attempt to predict _ which _ paper becomes popular ( although it can be shown , using the model , that papers published at the inception of a field have a much higher probability to become popular ) .
instead , the model incorporates the quality of papers implicitly , and indeed the number of citations to a paper is frequently seen as an indicator of its quality .
popular papers are also easily discovered , which further feeds their popularity .
the idea of using popularity as a proxy for quality may extend to other areas where resources are scarce ; for example , skilled surgeons are in high demand .
the cumulative advantage model of price @xcite is developed as a modification of the polya urn model , which is used to model a sampling process where each draw from the urn , corresponding to a collection of different types of objects , changes the composition of the urn and thereby changes the probability of drawing an object of any type in the future .
the standard polya urn model consists of an urn containing some number of black and white balls , drawing a ball at random and then returning it to the urn along with a new ball of the same color @xcite .
independently of price , barabasi and albert introduced a similar model in 1999 @xcite .
they examined the degree distributions of an actor collaboration network ( two actors are connected if they are cast in the same movie ) , world wide web ( two web pages are connected if there is a hyperlink from one page to the other ) , and power grid ( two elements ( generators , transformers , substations ) are connected if there is a high - voltage transmission line between them ) , finding that they approximately followed power - law distributions . although the actor collaboration network and the power grid networks are defined much like a projection from a two - mode to a one mode network , a subtle difference between them is that direct interaction between the nodes can be assumed . in other words ,
the nodes can be thought of as directly linked . both of the generic network models in existence at the time , the erds - rnyi and the watts - strogatz models , operated on a fixed set of @xmath16 vertices , and assumed that connections were placed or rewired without any regard to the degrees of the nodes to which they were connected .
the model of barabasi and albert changed both of these aspects .
first , they introduced the notion of network growth , such that at each time step a new node would be added to the network .
second , this new node would connect to the existing network with exactly @xmath142 non - directed edges , and the nodes they attached to were chosen in proportion to their degree .
the probability for the incoming vertex to connect to vertex @xmath1 depends solely on its degree @xmath201 and is given by @xmath202 the model was solved by barabasi and albert using rate equations , which are differential equations for the evolution of node degree over time where both degree and time , as an approximation , are treated as if they were continuous variables @xcite .
more general solutions were provided by krapivsky _
et al . _ also using rate equations @xcite and dorogovtsev _
_ using master equations which , like rate equations , are differential equations for the evolution of node degree , but they ( correctly ) treat degree as a discrete variable while still making the continuous - time approximation for time @xcite . in the master equation approach ,
one writes down an equation for the evolution of the number of nodes of a given degree .
let us use @xmath203 to denote the number of nodes of degree @xmath26 in the network at time @xmath77 , where time is identified with network size , i.e. , time @xmath77 corresponds to the network at the point of its evolution when it consists of @xmath77 nodes .
( the nodes making up the seed network can be usually ignored in the limit as time increases . )
the number @xmath203 can change in two ways : it can either increase as an incoming node attaches itself to a node of degree @xmath27 and thus turns it into a node of degree @xmath26 , or it can decrease as an incoming node attaches itself to a node of degree @xmath26 , turning into a node of degree @xmath204 .
the former situation leads to @xmath205 , and the latter to @xmath206 .
transitions larger than one , e.g. , from @xmath26 to @xmath207 or from @xmath26 to @xmath208 are very unlikely and can be ignored .
the value of @xmath209 increases by one per time step as each incoming node has degree @xmath142 , which also means there are no nodes with degree less than @xmath142 , and hence the equations used to model the evolution of quantities like @xmath203 are not valid for @xmath210 .
the resulting degee distribution has the form @xmath211 which asymptotically behaves as @xmath212 .
the preferential attachment model of barabasi and albert has attracted a tremendous amount of scientific interest in the past few years , and consequently numerous modifications of the model have been introduced .
for example , extensions of the model allow : * ties to appear and disappear between any pairs of vertices ( the original formulation only considers the addition of ties between the incoming vertex and set of vertices already in existence ) .
* vertices to be deleted either uniformly at random or based on their connectivity . *
the attachment probability @xmath213 to be super - linear or sub - linear in degree , or to consist of several terms .
* nodal attributes , such as the _ attractiveness _ ( the propensity with which new ties form with the node ) or _ fitness _ ( the propensity with which established ties remain intact ) of a node , and the attachment probability can incorporate these attributes in addition to degree . * edges to assume weights instead of @xmath214 binary values to codify connection strength between any pair of elements . in the context of physician networks ,
a preferential attachment model could be used to examine the process of new physicians seeking colleagues to ask for advice upon joining a medical organization , such as a hospital . under the preferential attachment hypothesis
, new physicians would be more likely to form ties with and thus seek advice from popular established physicians or physicians in the same cohort ( e.g. , medical school or residency program ) . the class of models known as _ network evolution models _ can be defined via three properties : ( i ) the models incorporate a set of stochastic attachment rules which determine the evolution of the network structure explicitly on a time - step by time - step basis ; ( ii ) the network evolution starts from an empty network consisting of nodes only , or from a small seed network possessing arbitrary structure ; and ( iii ) the models incorporate a stopping criterion , which for growing network models is typically in the form of the network size reaching a predetermined value , and for dynamical ( non - growing ) network models the convergence of network statistics to their asymptotic values .
many network evolution models do not reference intrinsic properties or attributes of nodes , and in this sense they are similar to the various implementations of preferential attachment models that do not postulate node specific fitness or attractiveness .
most network evolution models that are intended to model social networks employ some variants of focal closure and cyclic closure ( see , e.g. , ) .
_ focal closure _ refers to the formation of ties between individuals based on shared foci , which in a medical context could correspond to a group of doctors who practice in a particular hospital ( the focus ) .
the concept of shared foci in network science is analogous to homophily in social network analysis .
more broadly , ties could represent any interest or activity that connects otherwise unlinked individuals .
in contrast , _ cyclic closure _ refers to the idea of forming new ties by navigating and leveraging one s existing social ties , a process that results in a cycle in the underlying network .
because the network is non - directed , the term cycle is used interchangeably with closure .
this differs from when the network is directional and a cycle is a specific form of closure , with transitivity being another form .
_ triadic closure _ , which is the special case of cyclic closure involving just three individuals , refers to the process of getting to know friends of friends , leading to the formation of a closed triad in the non - directed network .
most social networks are expected to ( i ) have skewed and fat - tailed degree distributions , ( ii ) be assortatively mixed ( high - degree individuals are connected to high - degree individuals ) , ( iii ) be highly clustered , and ( iv ) possess the small - world property ( average shortest path lengths are short , or more precisely , scale as @xmath215 ) , and ( v ) exhibit community structure . the models by davidsen _
_ @xcite and marsili _ et al . _
@xcite exemplify dynamic ( non - growing ) network evolution models for social networks .
both have a mechanisms that starts by selecting a node @xmath1 in the network uniformly at random . in the model of davidsen _ et al .
_ , if node @xmath1 has fewer than two connections , it is connected to a randomly chosen node in the network ; otherwise two randomly chosen neighbors of node @xmath1 are connected together . in the model of marsili _ et al .
_ , node @xmath1 ( regardless of its degree ) is connected with probability @xmath216 to a randomly chosen node in the network ; then a second - order neighbor of node @xmath1 , i.e. , a friend s friend , is connected with probability @xmath217 to node @xmath1 .
the first mechanism in each model , the random connection , emulates focal closure , because there are no nodal attributes signifying shared interests .
the point is that the formation of these connections is not driven by the structure of existing connections but , from the point of view of network structure , is purely random .
the second mechanism , the notion of triadic closure , is implemented in slightly different ways across the models .
if these mechanisms were applied indefinitely , the result would be a fully connected network . to avoid this outcome
, the models also delete ties at a constant rate , which makes it possible for network statistics of interest to reach stationary distributions . in the model of davidsen _
_ , tie deletion is accomplished by choosing a node in the network uniformly at random , and then removing all of its ties with some probability ; marsili _ et al . _
accomplish the same phenomenon by selecting a tie uniformly at random , and then deleting it with probability @xmath218 .
growing network evolution models , such as those by vzquez @xcite and toivonen _ et al . _
@xcite , do not usually incorporate link deletion , but instead grow the network to a pre - specified size , which obviates the need for link deletion .
marsili _ et al .
_ use extensive numerical simulations , as well as a master equation approach applied to a mean - field approximation of the model , to explore the impact of varying the probabilities @xmath216 ( global linking ) , @xmath217 ( neighborhood linking ) , and @xmath218 ( link deletion ) for average degree and average clustering coefficient .
consider a situation where the value of @xmath217 ( neighborhood linking ) is increased while keeping the value of @xmath218 ( link deletion ) fixed . at first , for small values of @xmath217 , components with more than two nodes are rare , and the network can be said to be in the sparse phase . upon increasing the value of @xmath217 up to a specific point ,
a large connected component emerges , and the value of the average degree suddenly jumps up .
this point equals @xmath219 and is known as the critical point
it marks the beginning of the dense phase in the phase diagram of the system .
as @xmath217 is increased further , the network becomes more densely connected . reversing the process by slowly decreasing
the value of @xmath217 identifies a range of values from @xmath220 where the largest connected component remains densely connected and the average degree remains high .
only when the value of @xmath217 is decreased below a point denoted by @xmath221 does the network `` collapse '' and re - enter the sparse phase .
this phenomenon , which demonstrates some of the connections between network science and statistical physics , is typical of first - order or discontinuous phase transitions in statistical physics , and it demonstrates how hysteresis , the effect of the system remembering its past state can rise in networked systems . although markov dependence is a special case of hysteresis , its use is generally restricted to probabilistic models whereas hysteresis is typically aligned with nonlinear models of physical phenomena having a continuous state - space . from the social network point of view
this means that the network can remain in a connected phase even if the rate of establishing new connections at the current rate would not be sufficient for getting the network to that phase in the first place . in more practical terms
, this finding implies that it is possible to maintain a highly connected network with a relatively low `` effort '' ( the @xmath217 parameter in the model ) once the network has been established , but that same low level of effort would not be sufficient for establishing the dense phase of network evolution in the first place .
( the analogy in social network analysis is that the threshold for forming a ( e.g. ) friendship is greater than that needed for it to remain intact . )
the model by kumpula _
@xcite , which is another dynamical ( non - growing ) network evolution model for social networks , implements cyclic closure and focal closure ( see figure [ fig : kumpula ] ) in a manner similar to the models of davidsen _ et al . _ and marsili _ et al .
_ , but introduces a minor modification . unlike the previous models which produce binary networks with @xmath222 , this model produced weighted networks with @xmath223 . the main modification deals with the triadic closure step , which here is implemented as a weighted two - step random walk
starting from a randomly chosen node @xmath1 ; this node chooses one of its neighbors @xmath2 with probability @xmath224 , where @xmath225 is the strength of node @xmath1 , i.e. , the sum of the edge weights connecting it to its neighbors .
if node @xmath2 has neighbors other than @xmath1 , such a node @xmath26 will be chosen with probability @xmath226 , where there is a requirement that @xmath227 .
the weights @xmath228 and @xmath229 on the edges just traversed will be increased by a value @xmath230 .
in addition , if there is a link connecting node @xmath1 and node @xmath26 , the weight @xmath231 on that link is similarly increased by @xmath230 ; otherwise a new link is established between node @xmath1 and @xmath26 with @xmath232 . when @xmath233 , there is no clear community structure present , but as the value of @xmath230 is increased , very clear nucleation of communities takes place .
this phenomenon happens because when @xmath234 , a type of positive feedback or memory gets imprinted on the network , which reinforces existing connections , and makes future traversal of those connections more likely .
this is not unlike the models of cumulative advantage or preferential attachment discussed above , but now applies to individual links as opposed to nodes .
if one inspects the community structure produced by the model , most of the strong links appear to be located within communities , whereas links between communities are typically weak .
this type of structural organization is compliant with the so - called weak ties hypothesis , formulated in , which states , in essence , that the stronger the tie connecting two individuals , the higher the fraction of friends they have in common .
_ showed that a large - scale social network constructed from the cell phone communication records of millions of people was in remarkable agreement with the hypothesis only the top 5% of ties in terms of their weight deviated noticeably from the prediction .
the networks produced by the model of kumpula _ et al . _ are clearly reminiscent of observed real - world social networks , and the inclusion of the tuning parameter @xmath230 makes it straightforward to create networks with sparser or denser communities .
the downside is that the addition of weights to the model appears to make it analytically intractable .
, ( b ) @xmath235 , ( c ) @xmath236 , and ( d ) @xmath237 .
figure adapted from kumpula _ et al . _
( 2007 ) . ]
nodal attribute models , in stark contrast to network evolution models , specify nodal attributes for each node , which could be scalar or vector valued .
the probability of linkage between any two nodes is typically an increasing function of the similarity of the nodal attributes of the two nodes in consideration .
this is compatible with the notion of homophily , the tendency for like to attract like .
nodal attribute models can also be interpreted as spatial models , where the idea is that each node has a specific location in a social space .
the models by bogu _
_ @xcite and wong _ et al .
_ @xcite serve as interesting examples .
nodal attribute models do not specify attachment rules at the level of the network , and in some sense can be seen as latent variable models for social network formation .
these types of models have been studied less in the network science literature than network evolution models .
clearly , nodal attribute models have a strong resemblance to models developed and studied in the social network literature that treat dyads as independent conditional on observed attributes of the individuals , other covariates , and various latent variables ( individual - specific random effects in the case of the p@xmath164 model , categorical latent variables in the case of latent class models , continuous latent variables under the latent - space and latent eigenmodels in section [ sec : condindep ] ) .
unlike network science , work on such models in the social network literature has been more prominent than work on network evolution .
a difference in the approach of some nodal attribute models and social network models is that the former may use specific rules for determining whether a tie is expected , such as a threshold function ( in a sense emulating formal decision making ) , whereas the latter rewards values of parameters that make the model most consistent with the observed network(s ) .
many network characteristics are either microscopic or macroscopic in nature ; the value of a microscopic characteristic depends on local network structure only , whereas the value of a macroscopic characteristic depends on the structure of the entire network .
node degree is an example of a microscopic quantity : the degree of a node depends only on the number of its connections .
in contrast , network diameter , the longest of all pairwise shortest paths in the network , can change dramatically by the addition ( or removal ) of even a very small number of links anywhere in the network .
for example , a @xmath26-cycle consists of @xmath26 nodes connected by @xmath26 links such that a cycle is formed with each node connected to precisely two nodes .
the diameter of such a network is @xmath238 , where the floor function @xmath239 maps a real number @xmath240 to the largest previous integer , such that for an even @xmath241 it follows that @xmath242 . for large values of @xmath26 , adding just a few links quickly brings down the value of network diameter .
there is a third , intermediate scale that lies between the microscopic and macroscopic scales which is often known as the _ mesoscopic _ scale .
for example , a @xmath26-clique could justifiably be called a mesoscopic object ( especially if @xmath26 is large ) .
another type of mesoscopic structure is that of a network community , which can be loosely defined as a set of nodes that are densely connected to each other but sparsely connected to other nodes in the network ( but not to the extent of resulting in distinct components ) .
there has been considerable interest especially in the physics literature focusing on how to define and detect such communities , and several review papers cover the existing methods @xcite .
the motivation behind many of these efforts is the idea that communities may correspond to functional units in networks , such as unobserved societal structures .
the examples range from metabolic circuits within cells @xcite to tightly - knit groups of individuals in social networks @xcite .
the interested reader can consult the review articles on community detection methods @xcite for more details .
another application is health care where , for instance , have deduced communities of physicians based on network ties representing their treating the same patients within the same period of time .
the clustering of physicians in communities is shown for one particular hospital referral region ( a health care market encompassing at least one major city where both cardiovascular surgical procedures and neurosurgery are performed ) in the united states ( figure [ fig : commex ] ) .
one potential application of network science methods for community detection is in the area of health education and disease prevention ( e.g. , screening ) . due to limited resources
, it may not be possible to send materials or otherwise directly educate every member of the population .
the partition of individuals into groups would facilitate a possibly more efficient approach whereby the communities are first studied to identify key individuals .
then a few key individuals in each community are trained and advised on mechanisms for helping the intervention to diffuse across the community . a general characteristic of interventions where such an approach might be useful are those where intensive training is required to be effective and where delegation of resources through passing on knowledge or advice is possible .
a number of network community detection methods define communities implicitly via an appropriately chosen quality function .
the underlying idea is that a given network can be divided into a large number of partitions , or subsets of nodes , such that each node belongs to one subset , and each such partition @xmath243 has a scalar - valued quality measure associated with it , denoted by @xmath244 . in principle
one would like to enumerate all possible partitions and compute the value of @xmath245 for each of them , and the network communities would then be identified as the partition ( or possibly partitions ) with the highest quality . in practice , however , the number of possible partitions is exceedingly large even for relatively small networks , and therefore heuristics are needed to optimize the value of @xmath245 .
community detection methods based on quality function optimization therefore have two distinct components , which are the functional form of the quality function @xmath245 , and the heuristic used for navigating a subset of partitions over which @xmath245 is maximized .
the most commonly used optimization based approach to community detection is modularity maximization , where modularity is one possible choice for the quality function @xmath245 ; in statistical terminology , modularity maximization would be regarded as a non - parametric procedure due to the fact that no distributional nor functional form assumptions are relied upon .
there are many variants of modularity , but here the focus is on the original formulation by newman and girvan @xcite .
modularity can be seen as a measure that characterizes the extent of homophily or assortative mixing by class membership , and one way to derive it is by considering the observed and expected numbers of connections between vertices of given classes , where the class of vertex @xmath1 is given by @xmath246 .
the following derivation follows closely that of @xcite , although other derivations , based for example on dynamic processes , are also available .
we start by considering the observed number of edges between vertices of the same class , which is given by @xmath247 , where @xmath248 is the kronecker delta , and the factor @xmath249 prevents double - counting vertex pairs . to obtain the expected number of edges between vertices of the same class , cut every edge in half , resulting in two stubs per edge , and then connect these stubs at random . for a network with @xmath142 edges , there are a total of @xmath250 such stubs .
consider one of the @xmath201 stubs connected to vertex @xmath1 .
this particular stub will be connected at random to vertex @xmath2 of degree @xmath251 with probability @xmath252 , and since vertex @xmath1 has @xmath201 such stubs , the number of expected edges between vertices @xmath1 and @xmath2 is @xmath253 .
the expected number of edges falling between vertices of the same class is now @xmath254 .
the difference between the observed and expected number of within class ties is therefore @xmath255 .
given that the number of edges varies from one network to the next , it is convenient to deal with the fraction of edges as opposed to the number of edges , which is easily obtained by dividing the expression by @xmath142 , resulting in @xmath256 the assignment @xmath243 of nodes into classes that maximizes modularity @xmath257 is taken as the optimal partition and identifies the assignment of nodes into network communities .
note that modularity can be easily generalized from binary networks to weighted networks , in which case @xmath201 stands for the strength ( sum of all adjacent edge weights ) of node @xmath1 , and @xmath142 is the total weight of the edges in the network .
the expression for modularity has an interesting connection to spin models in statistical physics . in a so - called infinite range @xmath258-state potts model ,
each of the @xmath16 particles can be in one of @xmath258 states called spins , and the interaction energy between particles @xmath1 and @xmath2 is @xmath259 if they are in the same state and zero if they are not in different states .
the energy function of the system , known as its hamiltonian , is given by the sum over all of the pairwise interaction energies in the system @xmath260 where @xmath261 indicates the spin of particle @xmath147 and @xmath262 denotes the configuration of all @xmath16 spins . finding the minimum energy state ( the ground state ) of the system corresponds to finding @xmath262 such that @xmath263 is minimized .
the states of the particles ( spins ) correspond to community assignments of nodes in the network problem , and minimizing @xmath263 is mathematically identical to maximizing modularity @xmath257 . in the physical system ,
depending on the interaction energies , the spins seek to align with other spins ( interact ferromagnetically ) or they seek to have different orientations ( interact antiferromagnetically ) . in the community detection problem ,
two nodes seek to be in the same community if they are connected by an edge that is stronger than expected ; otherwise they seek to be in different communities .
this correspondence between the two problems has enabled the application of computational techniques developed for the study of spin systems and other physical systems to be applied to modularity optimization and , more broadly , to the optimization of other quality functions .
simulated annealing , greedy algorithms , and spectral methods serve as examples of these methods .
more details and references are available in community detection review articles @xcite . , where subscripts @xmath1 and @xmath2 are used to index the nodes and subscript @xmath264 is used to index the slices .
each node is coupled to itself in the other slices , and the structure of this coupling , encoded by the @xmath265 tensor , depends on whether the slices correspond to snapshots taken at different times ( time - dependent network ) , to communities detected at different resolution levels ( multiscale network ) , or to a network consisting of multiple types of interactions ( multiplex network ) . for time - dependent and multiscale networks ,
the slice - to - slice coupling extends for each node a tie to itself across neighboring slices only as exemplified for the node in the upper right corner of the slices ; for multiplex networks , the slice - to - slice coupling extends a tie from each node to itself in all the slices as exemplified for the node in the lower left corner .
whatever the form of this coupling , it is applied the same way to each node , although for visual clarity the slice - to - slice couplings are shown just for two nodes .
adapted from mucha _ et al . _
( 2010 ) . ]
although there are several extensions of modularity maximization , only one such generalization is described here .
_ developed a generalized framework of network quality functions that allow the study of community structure of arbitrary multislice networks ( see fig .
[ fig : ms ] ) , which are combinations of individual networks coupled through links that connect each node in one slice to the same node in other slices .
this framework allows studies of community structure in time - dependent , multiscale , and multiplex networks .
much of the work in the area of community detection is motivated by the observation that the behavior of dynamical processes on networks is driven or constrained by their community structure .
the approach of mucha _ et al . _
is based on a reversal of this logic , and it introduces a dynamical process on the network , and the behavior of the dynamical process is used to identify the ( structural ) communities .
the outcome is a quality function @xmath266 \delta(c_{is } , c_{jr}),\ ] ] where @xmath267 encodes the node - to - node couplings within slices and @xmath265 encodes the node - to - node couplings across slices that are usually set to a uniform value @xmath268 ; @xmath269 is the number ( or weight ) of ties within slice @xmath264 and @xmath270 is the weight of all ties in the network , both those located within slices and those placed across slices ; @xmath271 is a resolution parameter that controls the scale of community slices separately for each slice @xmath264 .
the standard modularity quality function uses @xmath246 to denote the community assignment of node @xmath1 , but in the multislice context two indices are needed , giving rise to the @xmath272 terms , where the subscript @xmath1 is used to index the node in question and the subscript @xmath264 to index the slice .
the outcome of minimizing @xmath273 , which can be done with the same heuristics as minimization of the standard modularity @xmath274 , is a matrix @xmath275 that consists of the community assignments @xmath272 of each node in every slice .
the multislice framework can handle any combination of time - dependent , multiscale , and multiplex networks .
for example , the slices in fig .
[ fig : ms ] could correspond , say , to a longitudinal friendship network of a cohort of college students , each slice capturing the offline friendships of the students in each year .
if data on the online friendships of the students were also available , corresponding to a different type of friendship , one could then introduce a second stack of four slices encoding those friendships .
the four offline slices and the four online slices form a multiplex system , and they would be coupled accordingly .
one could further introduce multiple resolution scales , and if one was interested in examining the community structure of the students at three different scales using , say , @xmath276 , this would result in a three - fold replication of the @xmath277 slice array with each of the three layers having a distinct value for @xmath271 .
taken together , this would lead to a three - dimensional @xmath278 array of slices .
cliques are ( usually small ) fully connected subgraphs , and a non - directed @xmath26-clique is a complete subgraph consisting of @xmath26 nodes connected with @xmath279 links . in materials
science the term percolation refers to the movement of fluid through porous materials .
however , in mathematics and statistical physics , the field of percolation theory considers the properties of clusters on regular lattices or random networks , where each edge may be either open or closed , and the clusters correspond to groups of adjacent nodes that are connected by open edges .
the system is said to percolate in the limit of infinite system size if the largest component , held together by open edges , occupies a finite fraction of the nodes .
the method of @xmath26-clique percolation in combines cliques and percolation theory , and it relies on the empirical observation that network communities seem to consist of several small cliques that share many of their nodes with other cliques in the same community . in this framework , cliques can be thought of as the building blocks of communities .
a @xmath26-clique community is then defined as the union of all adjacent @xmath26-cliques , where two @xmath26-cliques are defined to be adjacent if they share @xmath27 nodes .
one can also think about `` rolling '' a @xmath26-clique template from any @xmath26-clique in the graph to any adjacent @xmath26-clique by relocating one of its nodes and keeping the other @xmath27 nodes fixed . a community , defined through the percolation of such a template , then consists of the union of all subgraphs that can be fully explored by rolling a @xmath26-clique template . as @xmath26 becomes larger , the notion of a community becomes more stringent , and values of @xmath280 tend to be most appropriate because larger values become unwieldy .
the special case of @xmath281 reduces to bond ( link ) percolation and @xmath282 reduces to site ( node ) percolation .
the @xmath26-clique percolation algorithm is an example of a local community - finding method .
one obtains a network s global community structure by considering the ensemble of communities obtained by looping over all of its @xmath26-cliques .
some nodes might not belong to any community ( because they are never part of any @xmath26-clique ) , and others can belong to several communities ( if they are located at the interface between two or more communities ) .
the nested nature of communities is recovered by considering different values of @xmath26 , although @xmath26-clique percolation can be too rigid because focusing on cliques typically causes one to overlook other dense modules that are not quite as tightly connected .
the advantage of @xmath26-clique percolation is that it provides a successful way to consider community overlap . allowing the detection of network communities that overlap is especially appealing in the social sciences , as people may belong simultaneously to several communities ( colleagues , family , friends , etc . ) .
however , the case can be made that it is the underlying interactions that are different , and one should not combine interactions that are of fundamentally different types . in statistics , this is analogous to using composite variables or scales that combine multiple items in ( e.g. ) health surveys or questionnaires .
if the nature of the interactions is known , the system might be more appropriately described as a multiplex network , where one tie type encodes professional interactions , another tie type corresponds to personal friendships , and a third tie type captures family memberships . the multislice framework discussed above is able to accommodate memberships in multiple communities as long as distinct interaction types are encoded with distinct ( multiplex ) ties . the latent class models in section [ sec : condindep ] partitions the actors in a network into disjoint groups that can be thought of as communities . the clustering process can be thought of as a search for structural equivalence in that individuals are likely to be included in the same community if the network around them is similar to that of their neighbors .
the criteria for judging the efficacy of the partition of nodes into communities is embedded in the statistical model implied for the network and as such is a balance between all of the terms in the model .
this contrasts a non - model - based objective function such as modularity which focuses on maximizing in some sense the ratio of density of ties within and between communities . to illustrate the difference ,
consider a @xmath26-star .
the greater the value of @xmath26 , the greater the discrepancy in the degree of the actors . therefore ,
if @xmath26-stars occur frequently , the members of the same @xmath26-star are likely to be included in the same group by the latent class model but , due to the difference in degree , are unlikely to be grouped under modularity maximization . however , an advantage of the network science approach is that results are likely to be more robust to model mis - specifications than under the social network approach . in the future it is possible to imagine a bridging of the two approaches to community detection .
for example , a model for the network or the component of the model involving the key determinants of network ties , could be incorporated in the modularity function in ( [ eq : modmax ] ) .
depending on the specification , the result might be a weighted version of modularity in which a higher penalty is incurred if individuals with similar traits or in structurally equivalent positions with respect to @xmath26-stars , triadic closure or other local network configurations are included in different communities than if individuals with different traits are in different communities
. however , to the best of the author s knowledge , such a procedure is not available .
[ sec : discuss ] in this chapter , the dual fields of social networks and network science have been described , with particular focus on sociocentric data . both fields are growing rapidly in methodological results and the breadth of applications to which they are applied . in health applications , social network methods for evaluating whether individuals attributes spread from person - to - person across a population ( social influence ) and for modeling relationship or tie status ( social selection )
have been described .
models of relationship status have not been applied as frequently in health applications , where focus often centers on the patient . however , is a notable exception . due to the ever - growing availability of data , the interest in peer effects , and the need to design support mechanisms ,
the role of social network analysis in health care and medicine is likely to undergo continued growth in the future .
a novel feature of this chapter is the attention given to network science .
although network science is descriptively - inclined and thus is removed from mainstream translational medical research seeking to identify causes of medical outcomes , the increasing availability of complex systems data provides an opportunity for network science to play a more prominent role in medical research in the future .
for example , barabasi and others have created a human disease network by connecting all hereditary diseases that share a disease - causing gene . in other work
, they created a phenotypic disease network ( pdn ) as a map summarizing phenotypic connections between diseases @xcite .
these networks provided important insights into the potential common origins of different diseases , whether diseases progress through cellular functions ( phenotypes ) associated with a single diseased ( mutated ) gene or with other phenotypes , and whether patients affected by diseases that are connected to many other diseases tend to die sooner than those affected by less connected diseases .
such work has the potential to provide insights into many previously untested hypotheses about disease mechanisms .
for example , they may ultimately be helpful in designing `` personalized treatments '' based on the network position held by an individual s combined genetic , proteomic , and phenotypic information .
in addition , they may suggest conditions for which treatments found to be effective on another condition might also be tried .
there are several important topics that have not been discussed , notably including network sampling . in gathering network data , adaptive methods such as link - tracing designs
are often used to identify individuals more likely to know each other and thus to have formed a relationship with other sampled individuals than in a random - probability design .
link - tracing and other related designs are often used to identify hard - to - reach populations @xcite .
however , the sampling probabilities corresponding to link - tracing designs may be difficult to evaluate ( generally requiring the use of simulation ) and it may not be obvious how they should be incorporated in the analysis .
the development of statistical methods that account for the sample design in the analysis of social network data has lagged behind the designs themselves .
however , recently progress has been made on statistical inference for sampled relational network data . in the future it is likely that more bridges will form between the social network and the network science fields with models or methods developed in one field used to solve problems in the other .
furthermore , as these two fields become more entwined , it is likely that they will also become more prominent in the solution to important problems in medicine and health care .
the time and effort of dr .
omalley and dr .
onnela on researching and developing this chapter was supported by nih / nia grant p01 ag031093 ( pi : christakis ) and robert wood johnson award # 58729 ( pi christakis ) .
dr onnela was further supported by nih / niaid grant r01ai051164 ( degruttola ) .
the authors thank mischa haider , brian neelon , and bruce e. landon for reviewing an early draft of the manuscript and providing several useful comments and suggestions .
to help readers familiar with social networks understand the network science component of the chapter and conversely for readers familiar with network science to understand the social network component , the following glossary contains a comprehensive list of terms and definitions . 1 . social network : a collection of actors ( referred to as actors ) and the ( social ) relationships or ties linking them .
2 . relationship , tie : a link or connection between two actors .
dyad : a pair of actors in a network and the relationship(s ) between them ; two relationships per measure for a directed network , one relationship per measure for an undirected network .
triad : a triple of three actors in the network and the relationships between them .
scale or valued relationship : a non - binary relationship between two actors ( e.g. , the level of a trait ) .
we focused on binary relationships in the chapter .
directed network : a network in which the relationship from actor @xmath1 to actor @xmath2 need not be the same as that from actor @xmath2 to actor @xmath1 . 7 .
non - directed network : a network in which the state of the relationship from actor @xmath1 to actor @xmath2 equals the state of the relationship from actor @xmath2 to actor @xmath1 .
sociocentric network data : the complete set of observations on the @xmath283 relationships in a directed network , or @xmath284 relationships in an undirected network , with @xmath241 actors .
collaboration network : a network whose ties represent the actors joint involvement on a task ( e.g. , work on a paper ) or a common experience ( e.g. , treating the same episode of health care for a patient ) . 10 .
bipartite : relationships are only permitted between actors of two different types
unipartite : relationships are permitted between all types of actors
social contagion , social influence , peer effects : terms used to describe the phenomenon whereby an actor s trait changes due to their relationship with other actors and the traits of those actors . 13 .
mutable trait : a characteristic of an actor than can change state .
social selection : the phenomena whereby the relationship status between two actors depends on their characteristics , as occurs with homophily and heterophily . 15 .
homophily : a preference for relationships with actors who have similiar characteristics . popularly referred to as `` birds of a feather flock together . ''
heterophily : a preference for relationships with actors who have different characteristics . popularly referred to as `` opposites attracting . '' 17 . in - degree ,
popularity : the number of actors who initiated a tie with the given actor . 18 .
out - degree , expansiveness , activity : the number of ties the given actor initiates with other actors .
@xmath26-star : a subnetwork in which the focal actor has ties to @xmath26 other actors . 20 .
@xmath26-cycle : a subnetwork in which each actor has degree 2 that can be arranged as a ring ( i.e. , a @xmath26-path through the actors returns to its origin without backtracking .
for example , the ties a - b , b - c , and c - a form a three - cycle . 21 .
@xmath26 degrees of separation : two individuals linked by a @xmath26-path ( @xmath27 intermediary actors ) that are not connected by any path of length @xmath27 or less . 22 . density : the overall tendency of ties to form in the network .
a descriptive measure is given by the number of ties in the network divided by the total number of possible ties .
reciprocity : the phenomena whereby an actor @xmath1 is more likely to have a tie with actor @xmath2 if actor @xmath2 has a tie with actor @xmath1 . only defined for directed networks . 24 .
clustering : the tendency of ties to cluster and form densely connected regions of the network .
closure : the tendency for network configurations to be closed . 26 .
transitivity : the tendency for a tie from individual a to individual b to form if ties from individual a to individual c and from individual c to individual b exist .
a form of triadic closure commonly stated as `` a friend of a friend is a friend . ''
reduces to general triadic closure in an undirected network .
centrality : a dimenionless measure of an actors position in the network .
higher values indicate more central positions .
there are numerous measures of centrality .
four common ones are degree , closeness , betweeness , and eigenvalue centrality .
degree and eigenvalue centrality are extremes in that degree centrality is determined solely from an actor s degree ( it is internally focused ) while eigenvalue centrality is based on the centrality of the actors connected to the focal actor ( it is externally focused ) .
structural balance : a theory which suggests actors seek balance in their relationships ; for example , if a likes b and b likes c then a will endeavor to like c as well to keep the system balanced .
thus , the existence of transitivity is implied by structural balance .
structural equivalence : the network configuration ( arrangement of ties ) around one actor is similar to that of another actor . even though actors may not be connected , they can still be in structurally similar situations .
structural power : an actor in a dominant position in the network .
such an actor may be one in a strategic position , such as the only bridge between otherwise distinct components .
network component : a subset of actors having no ties external to themselves .
graph theory : the mathematical basis under which theoretical results for networks are derived and empirical computations are performed . 33 .
digraph : a graph in which edges can be bidirectional .
unlike social networks , digraphs can contain self - ties .
graphs lie in two - dimensional space
hypergraph : a graph in dimension three or higher .
maximal subset : a set of actors for whom all ties are intact in a binary - network ( i.e. , has density 1.0 ) .
if the set contains @xmath26 actors , the maximal subset is referred to as a @xmath26-clique .
scalar , vector , matrix : terms from linear and abstract algebra .
a scalar is a @xmath285 matrix , a vector is a @xmath286 matrix , and a matrix is @xmath287 , where @xmath288 . 37 .
adjacency matrix : a matrix whose off - diagonal elements contain the value of the relationship from one actor to another .
for example , element @xmath4 contains the relationship from actor @xmath1 to actor @xmath2 .
the diagonal elements are zero by definition .
matrix transpose : the operation whereby element @xmath4 is exchanged with element @xmath289 for all @xmath290 . 39 . row stochastic matrix : a matrix whose rows sum to 1 and contain non - negative elements .
thus , each row represents a probability distribution of a discrete - valued random variable .
random variable : a variable whose value is not known with certainty .
it can relate to an event or time period that is yet to occur , or it can be a quantity whose value is fixed ( i.e. , has occurred ) but is unknown .
parametric : a term used in statistics to describe a model with a specific functional form ( e.g. , linear , quadratic , logarithmic , exponential ) indexed by unknown parameters or an estimation procedure that relies on specification of the complete distribution of the data .
non - parametric : a model or estimation procedure that makes no assumption about the specific form of the relationship between key variables ( e.g. , whether the predictors have linear or additivie effects on the outcome ) and does not rely upon complete specification of the distribution of the data for estimation . 43 .
outcome , dependent variable : the variable considered causally dependent on other variables of interest .
this will typically be a variable whose value is believed to be caused by other variables .
independent , predictor , explanatory variable , covariate : a variable believed to be a cause of the outcome .
contextual variable : a variable evaluated on the neighbors of , or other members of a set containing , the focal actor .
for example , the proportion of females in a neighboring county , the proportion of friends with college degrees .
interaction effect : the extent to which the effect of one variable on the outcome varies across the levels of another variable .
endogenous variable : a variable ( or an effect ) that is internal to a system .
predictors in a regression model that are correlated with the unobserved error are endogeneous ; they are determined by an internal as opposed to an external process . by definition outcome variables
are endogenous .
exogenous variable : a variable ( or an effect ) that is external to the system in that its value is not determined by other variables in the system .
predictors that are independent of the error term in a regression model are exogeneous .
instrumental variable ( iv ) : a variable with a non - null effect on the endogeneous predictor whose causal effect is of interest ( the `` treatment '' ) that has no effect on the outcome other than that through its effect on treatment .
often - used sufficient conditions for the latter are that the iv is ( i ) marginally independent of any unmeasured confounders and ( ii ) conditionally independent of the outcome given the treatment and any unmeasured confounders . in an iv analysis a set of observed predictors may be conditioned on as long as they are not effects of the treatment and the iv assumptions hold conditional on them . while subject to controversy , iv methods are one of the only methods of estimating the true ( causal ) effect of an endogeneous predictor on an outcome .
linear regression model : a model in which the expected value of the outcome ( or dependent variable ) conditional on one or more predictors ( or explanatory variables ) is a linear combination of the predictors ( an additive sum of the predictors multiplied by their regression coefficients ) and an unobserved random error .
longitudinal model : a model that describes variation in the outcome variable over time as a function of the predictors , which may include prior ( i.e. , lagged ) values of the outcome .
observations are typically only available at specific , but not necessarily equally - spaced , times .
longitudinal models make the direction of causality explicit .
therefore , they can distinguish between the association between the predictors and the outcome and the effect of a change in the predictor on the change in the outcome . 52 . cross - sectional model : a model of the relationship between the values of the predictors and outcomes at a given time . because one can not discern the direction of causality , cross - sectional models are more difficult to defend as causal .
stochastic block model : a conditional dyadic independence model in which the density and reciprocity effects differ between blocks defined by attributes of the actors comprising the network .
for example , blocks for gender accomodate different levels of connectedness and reciprocity for men and women .
logistic regression : a member of the exponential family of models that is specific to binary outcomes .
it utilizes a link function that maps expected values of the outcome onto an unrestricted scale to ensure that all predictions from the model are well - defined .
multinomial distribution : a generalization of the binomial distribution to three or more categories .
the sum of the probabilities of each category equals 1 .
56 . exponential random graph model : a model in which the state of the entire network is the dependent variable .
provides a flexible approach to accounting for various forms of dependence in the network . not amenable to causal modeling .
degeneracy : an estimation problem encountered with exponential random graph models in which the fitted model might reproduce observed features of the network on average but each actor draw bears no resemblence to the observed network . often degenerate draws are empty or complete graphs .
latent distance model : a model in which the status of dyads are independent conditional on the positions of the actors , and thus the distance between them , in a latent social space .
latent eigenmodel : a model in which the status of dyads are independent conditional on the product of the ( weighted ) latent positions of the actors in the dyad .
latent variable : an unobserved random variable .
random effects and pure error terms are latent variables . 61 .
latent class : an unobserved categorical random variable .
actors with the same value of the variable are considered to be in the same latent class .
factor analysis : a statistical technique used to decompose the correlation ( or covariance ) matrix of a set of random variables into groups of related items .
generalized estimating equation ( gee ) : a statistical method that corrects estimation errors for dependent observations without necessarily modeling the form of the dependence or specifying the full distribution of the data .
random effect : a parameter for the effect of a unit ( or cluster ) that is drawn from a specified probability distribution . treating the unit effects as random draws from a common probability distribution allows information to be pooled across units for the estimation of each unit - specific parameter .
fixed effect : a parameter in a model that reflects the effect of an actor belonging to a given unit ( or cluster ) . by virtue of modeling the unit effects as unrelated parameters
, no information is shared between units and so estimates are based only on information within the unit .
ordinary least squares : a commonly - used method for estimating the parameters of a regression model .
the objective function is to minimize the squared distance of the fitted model to the observed values of the dependent variable .
maximum likelihood : a method of estimating the parameters of a statistical model that typically embodies parametric assumptions .
the procedure is to seek the values of the parameters that maximize the likelihood function of the data .
likelihood function : an expression that quantifies the total information in the data as a function of model parameters .
markov chain monte carlo : a numerical procedure used to fit bayesian statistical models .
steady state : the state - space distribution of a markov chain describes the long - run proportion of time the random variable being modeled is in each state .
often markov chains iterate through a transient phase in which the current state of the chain depends less and less on the initial state of the chain .
the steady state phase occurs when successive samples have the same distribution ( i.e. , there is no dependence on the initial state ) .
colinearity : the correlation between two predictors after conditioning on the other observed predictors ( if any ) .
when predictors are colinear distinguishing their effects is difficult and the statistical properties of the estimated effects are more sensitive to the validity of the model .
normal distribution : another name for the gaussian distribution .
has a bell - shaped probability density function .
covariance matrix : a matrix in which the @xmath4th element contains the covariance of items @xmath1 and @xmath2 .
absolute or geodesic distance : the total distance along the edges of the network from one actor to another .
cartesian distance : the distance between two points on a two - dimension surface or grid . adheres to pythagorus theorem .
count data : observations made on a variable with the whole numbers ( 0 , 1 , 2 , @xmath291 ) as its state space . 77 .
statistical inference : the process of establishing the level of certainty of knowledge about unknown parameters ( or hypothesis ) from data subject to random variation , such as when observations are measured imperfectly with no systematic bias or a sample from a population of interest is used to estimate population parameters . 78 .
null model : the model of a network statistic typically represents what would be expected if the feature of interest was non - existent ( effect equal to 0 ) or outside the range of interest . 79 .
permutation test : a statistical test of a null hypothesis against an alternative implemented by randomly re - shuffling the labels ( i.e. , the subscripts ) of the observations .
the significance level of the test is evaluated by re - sampling the observed data 50 100 times and computing the proportion of times that the test is rejected . 1 .
network science : the approach developed from 1995 onwards mostly within statistical physics and applied mathematics to study networked systems across many domains ( e.g. , physical , biological , social , etc ) .
usually focuses on very large systems ; hence , theoretical results derived in the thermodynamic limit are good approximations to real - world systems .
2 . thermodynamic limit : in statistical physics refers to the limit obtained for any quantity of interest as system size @xmath16 tends to infinity .
many analytical results within network science are derived in this limit due to analytical tractability .
3 . statistical physics : the branch of physics dealing with many body systems where the particles in the system obey a fix set of rules , such as newtonian mechanics , quantum mechanics , or any other rule set . as the number of bodies ( particles ) in a system grows , it becomes increasingly difficult ( and less informative ) to write down the equations of motion , a set of differential equations that govern the motion of the particles over time , for the system .
however , one can describe these systems probabilistically . the word `` statistical '' is somewhat misleading as there is no statistics in the sense of statistical inference involved ; instead everything proceeds from a set of axioms , suggesting that `` probabilistic '' might be a better term . statistical physics ,
also called statistical mechanics , gives a microscopic explanation to the phenomena that thermodynamics explains phenomenologically .
4 . generative model : most network models within network science belong to this category .
here one specifies the microscopic rules governing , for example , the attachment of new nodes to the existing network structure in models of network growth .
cumulative advantage : a stylized modeling mechanism introduced by price in 1976 to capture phenomena where `` success breeds success . ''
price applied the model to study citation patterns where power - law or power - law like distributions are observed for the distribution of the number of citations and successfully reproduced by the model .
polya urn model : a stylized sampling model in probability theory where the composition of the system , the contents of the urn , changes as a consequence of each draw from the urn
power law : refers to the specific functional form @xmath292 of the distribution of quantity @xmath240 .
also called pareto distribution .
see scale - free network .
preferential attachment : a stylized modeling mechanism introduced by barabasi and albert in 1999 where the probability of a new node to attach itself to an existing node @xmath1 of degree @xmath201 is an increasing function of @xmath201 ; in the case of linear preferential attachment , this probability is directly proportional to @xmath201 . in short ,
the higher the degree of a node , the higher the rate at which it acquires new connections ( increases its degree )
weak ties hypothesis : a hypothesis developed by sociologist mark granovetter in his extremely influential 1973 paper `` the strength of weak ties . ''
the hypothesis , in short , states the following : the stronger the tie connecting persons @xmath5 and @xmath31 , the higher the fraction of friends they have in common .
modularity : modularity is a quality - function used in network community detection , where its value is maximized ( in principle ) over the set of all possible partitions of the network nodes into communities .
standard modularity reads as @xmath293 where @xmath246 is the community assignment of node @xmath1 and @xmath294 is kronecker delta ; other quantities as defined in the text . 11 .
rate equations : rate equations , commonly used to model chemical reactions , are similar to master equations but instead of modeling the count of objects ( e.g. , number of nodes ) in a collection of discrete states ( e.g. , the number of @xmath26-degree nodes @xmath203 for different values of @xmath26 ) , they are used to model the evolution of continuous variables , such as average degree , over time
master equations : widely used in statistical physics , these differential equations model how the state of the system changes from one time point to the next . for example , if @xmath203 denotes the number of nodes of degree @xmath26 , given the model , one can write down the equation for @xmath295 , i.e. , the number of @xmath26-degree nodes at time @xmath296 . 13 . fitness or affinity or attractiveness : a node attribute introduced to incorporate heterogeneity in the node population in a growing network model .
for example , in a model based on preferential attachment , this could represent the inherent ability of a node to attract new edges , a mechanism that is superimposed on standard preferential attachment .
community : a group of nodes in a network that are , in some sense , densely connected to other nodes in the community but sparsely connected to nodes outside the community .
community detection : the set of methods and techniques developed fairly recently for finding communities in a given network ( graph ) .
the number of communities is usually not specified _ a priori _ but , instead , needs to be determined from data .
critical point : the value of a control parameter in a statistical mechanical system where the system exhibits critical behavior : previously localized phenomena now become correlated throughout the system which at this point behaves as one single entity
phase diagram : a diagram displaying the phase ( liquid , gas , etc . ) of the system as one or more thermodynamic control parameters ( temperature , pressure , etc . ) are varied .
phase transition : thermodynamic properties of a system are continuous functions of the thermodynamic parameters within a phase ; phase transitions ( e.g. , liquid to gas ) happen between phases where thermodynamic functions are discontinuous .
network diameter : the longest of the shortest pairwise paths in the network , computed for each dyad ( node pair ) .
hysteresis : the behavior of a system depends not only on its current state but also on its previous state or states . 21 .
quality function : typically a real - valued function with a high - dimensional domain that specifies the `` goodness '' of , say , a given network partitioning .
for example , given the community assignments of @xmath16 nodes , which can be seen as a point in an @xmath16-dimensional hypercube , the standard modularity quality function returns a number indicating how good the given partitioning is . 22 .
dynamic process : any process that unfolds on a network over time according to a set of pre - specified rules , such as epidemic processes , percolation , diffusion , synchronization , etc .
slice : in the context of multi - slice community detection , refers to one graph in a collection of many within the same system , where a slice can capture the structure of a network at a given time ( time - dependent slice ) , at a particular resolution level ( multiscale slice ) , or can encode the structure of a network for one tie type when many are present ( multiplex slice ) . 24 . scale - free network :
network with a power - law ( pareto ) degree distribution .
erds - rnyi model : also known as poisson random graph ( after the fact that the degree distribution in the model follows a poisson distribution ) , bernoulli random graph ( after the fact that each edge corresponds to an outcome of a bernoulli process ) , or the random graph ( as the progenitor of all random graphs ) . starting with a fixed set of @xmath16 nodes , one considers each node pair in turn independently of the other node pairs and connects the nodes with probability @xmath54 .
erds and rnyi first published the model in 1959 , although solomonoff and rapoport published a similar model earlier in 1951 .
watts - strogatz model : a now canonical model by watts and strogatz that was introduced in 1998 .
starting from a regular lattice structure characterized by high clustering and long paths , the model shows how randomly rewiring only a small fraction of edges ( or , alternative , adding a small number of randomly placed edges ) leads to a small - world characterized by high clustering and short paths .
the model is conceptually appealing , and shows how to interpolate , using just one parameter , from a regular lattice structure in one extreme to an erds - rnyi graph in the other .
mean - field approximation : sometimes called the zero - order approximation , this approximation replaces the value of a random variable by its average , thus ignoring any fluctuations ( deviations ) from the average that may actually occur .
this approach is commonly used in statistical physics .
ensemble : a collection of objects , such as networks , that have been generated with the same set of rules , where each object in the ensemble has a certain probability associated with it .
for example , one could consider the ensemble of networks that consists of 6 nodes and 2 edges , each begin equiprobable . | this chapter introduces statistical methods used in the analysis of social networks and in the rapidly evolving parallel - field of network science .
although several instances of social network analysis in health services research have appeared recently , the majority involve only the most basic methods and thus scratch the surface of what might be accomplished .
cutting - edge methods using relevant examples and illustrations in health services research are provided . * keywords * : dyad ; homophily ; induction ; network science ; peer - effect ; relationship ; social network . |
a reliable estimate of the nature and the amount of dust produced by asymptotic giant branch ( agb ) stars proves essential for a number of scientific issues .
these stars are believed to be the dominant stellar sources of dust in the present - day universe and their contribution to dust enrichment can not be neglected even at redshift @xmath4 @xcite . in order to properly include their contribution in chemical evolution models with dust ,
the mass and composition of dust grains released by each star as a function of its mass and metallicity need to be known .
in addition , the corresponding size distribution function allows to compute the extinction properties associated with these grains , which is fundamental information required for correctly interpreting the optical - near infrared properties of high - z quasar and gamma ray burst spectra @xcite .
theoretical modelling of dust formation around agbs has made considerable steps forward in the last years , owing to the pioneering explorations by the heidelberg group ( see e.g. * ? ? ? * and references therein ) , complemented by recent studies @xcite .
the reliability of these models must be tested against the observations , given the many uncertainties affecting the agb phase modelling ( see e.g. * ? ? ?
* ) , and the description of the formation and growth of dust grains in the winds of agbs @xcite . on the observational side ,
the study of evolved stars in the galaxy is hampered by the obscuration determined by the interstellar medium , and by the unknown distances of the objects observed .
this pushed the interest towards other nearby galaxies .
the large magellanic cloud ( lmc ) is an optimum target for this scope .
this stems from the low average reddening ( @xmath5 ) and its proximity , that allows the determination of important stellar properties , such as the absolute magnitudes .
the lmc has been surveyed in the optical by the magellanic clouds photometric survey ( mcps , * ? ? ?
* ) , in the near - ir by the deep near infrared survey of the southern sky ( denis , * ? ? ?
* ) and the two micron all sky survey ( 2mass , * ? ? ?
of particular interest for the study of evolved , dust surrounded stars , is the sage ( surveying the agents of galaxy evolution ) survey obtained by the _
spitzer space telescope _ with the infrared array camera ( irac ; 3.6 , 4.5 , 5.8 and 8.0 @xmath6 m ) and the multiband imaging photometer for _ spitzer _ ( mips ; 24 , 70 and 160 @xmath6 m ) ( see e.g. * ? ? ?
the analysis of evolved stars based on 2mass and irac data lead to a classification of lmc stars into three main categories .
@xcite , based on their analysis of the 2mass color - magnitude diagram ( cmds ) , divided the region above the tip of the red giant branch ( rgb ) into o - rich and c - rich zones ; @xcite selected a group of stars , called `` extreme '' , showing the clear signature of the presence of a dusty circumstellar envelope .
these stars were shown to contribute about 75% of the overall dust from agbs ( riebel et al .
this stimulated an interesting series of papers @xcite aimed at refining such a classification , to interpret the observed cmds by using grids of synthetic spectra for various chemical and physical inputs , the most relevant being the surface chemistry of the star , the size of the grains formed , the borders of the dusty shell , the surface gravity and the effective temperature of the central object . in this letter
we use models of dust formation around agbs , based on full evolutionary computations , to interpret the distribution of lmc extreme stars in the color color and color magnitude diagrams ( ccds and cmds , respectively ) .
our goal is to characterize the extreme stars in terms of the evolutionary phase , and of the amount and type of dust present in their surroundings .
this investigation will be an important benchmark for the studies focused on dust formation around agb stars .
+ 10pt + 10pt the evolutionary sequences have been calculated by means of the aton code for stellar evolution .
the interested reader may find the details of the numerical structure of the code together with the more recent updates in @xcite .
the range of masses investigated is @xmath7 .
more massive stars undergo core collapse by e capture , and do not experience the agb phase . models with @xmath8
achieve carbon ignition in conditions of partial degeneracy , and develop a core made up of oxygen and neon ; their phase of thermal pulses is commonly known as super - agb ( hereafter sagb , * ? ? ?
we assumed a chemical composition typical for lmc stars : @xmath9 , @xmath10 .
we recall here the physical inputs most relevant to the results obtained .
the convective instability is treated according to the full spectrum of turbulence ( fst ) description , developped by @xcite ; in models whose initial mass exceeds @xmath11 , use of the fst model leads to strong hot bottom burning ( hbb , * ? ? ?
* ) conditions , with the activation of an advanced p capture nucleosynthesis at the bottom of the convective envelope @xcite .
mass loss for m stars is modelled following @xcite ; for carbon stars , we use the results by @xcite . in regions unstable to convective motions , nuclear burning and mixing of chemicals
are coupled via a diffusive approach , as described in @xcite ; during the agb phase , convective velocities are allowed to decay exponentially into radiatively stable regions , based on the calibration aimed at reproducing the observed luminosity function of carbon stars , discussed in @xcite .
the wind surrounding agbs is assumed to expand isotropically , and to be accelerated by radiations pressure on dust particles .
the growth of dust grains is described by condensation of gas molecules in the wind impinging on the already formed grains .
the relevant equations can be found , e.g. , in @xcite , and were already used in other investigations by our group @xcite , and also by @xcite .
the spectral energy distribution ( sed ) of the star during the various evolutionary phases was determined from the knowledge of the properties of the central object and of the structure of the dust shell by means of the dusty code @xcite . for m - stars we considered the presence of amorphous silicates ( mainly olivine ) and alumina dust grains , whereas solid amorphous carbon and sic particles were used for c - stars .
iron grains are also accounted for in both cases .
in addition , for a given agb evolutionary time , the location ( and temperature ) of the inner dust region boundary , the density stratification , the relative percentages of the different dust species ( i.e. , sic vs. amc , silicates vs. alumina ) , the optical depth and the dust grain size as calculated self - consistently by our wind agb models ( ventura et al . 2014 and references therein ) are then used as inputs in dusty ( see dellagli et al .
, in preparation ) . finally , to test the reliability of the results obtained , we calculated a few seds with the 2-dust code ( ueta & meixner 2003 ) , run in the spherically symmetric mode .
the location of the various dusty layers and the relative percentages of the dust species present are assumed consistently with the results from the wind modelling .
these tests showed no differences ( @xmath1210% in flux ) among the results obtained with the two codes .
fig . [ fevol ] shows the evolution of models of different mass during the agb phase .
because this work is focused on the most obscured objects , we restrict our analysis to stars that have already reached the c - star stage , or m - stars where hbb is active ; these are the only conditions allowing dust production in meaningful quantities .
we show the variation of the effective temperature , luminosity and size of the grains formed as a function of the optical depth at 10 @xmath6 m ( @xmath13 ) , which quantifies the degree of obscuration of the central star .
it is not possible to define a unique grain size , because particles of various species form , with different dimensions : here we refer to the grains that dominate the dust mass formed , providing most of the acceleration of the wind by radiation pressure , i.e. solid carbon for c - stars and olivine for m - stars @xcite .
the evolution of c - stars is driven by the carbon enrichment at the surface , favoured by the occurrence of third dredge up ( tdu ) .
as the surface of the star becomes more enriched in carbon , the star assumes a more extended configuration @xcite , evolves towards lower effective temperatures ( t@xmath14 ) and loses mass at larger rates @xcite : all these factors favour dust production , mainly under the form of solid carbon , with traces of sic and iron @xcite .
we see in the left panel of fig .
[ fevol ] that the stars become more and more obscured as the external layers cool ; c - stars with the same t@xmath14 have the same degree of obscuration , independent of their previous history .
the most obscured stars , with @xmath15 , are those in the final stages of the agb evolution , when the size of the carbon grains is @xmath16 m @xcite .
the middle panel of fig .
[ fevol ] shows that the luminosity has only a minor effect on the optical depth variation of c - stars ; higher mass stars evolve at larger luminosities , owing to their larger core masses .
we note that these optical depth ( and luminosity ) values are consistent with those derived for the most extreme c - stars in the lmc ( gruendl et al .
2008 ) .
stars with initial mass above @xmath17 experience hbb , that prevents the possibility of reaching the c - star stage ; their evolution is driven by the temperature at the bottom of the envelope , which determines the luminosity , the mass - loss rate and consequently the quantity of dust formed .
as shown in the middle panel of fig .
[ fevol ] , the luminosity ( @xmath18 ) undergoes significant variations during the agb evolution : after the initial increase , due to the growth of the core mass , @xmath18 decreases , owing to the gradual loss of the external mantle .
the phase of maximum dust production ( with the highest @xmath13 ) is in conjunction with the maximum @xmath18 , when the mass loss rate alsoattains its largest values .
note that l in extreme oh / irs ( @xmath19 ; marshall et al .
2004 ) agrees very well with our predictions for the more massive agbs .
stars in the sagb regime , owing to their large core mass , reach the highest values of @xmath20 , that approaches @xmath21 in the @xmath22 model . the right panel of fig .
[ fevol ] shows the tight relationship between the size of the olivine grains ( @xmath23 ) formed and @xmath13 ; @xmath23 changes from @xmath24 m to @xmath25 m as @xmath13 grows from 1 to @xmath26 . in the high - mass domain
the t@xmath14 has a minor effect on @xmath20 , in agreement with the fact that dust production is mostly determined by the extent of the hbb experienced .
the narrow @xmath27 range ( @xmath28 k ) predicted in the more massive agb stars agree quite well with the spectroscopic temperatures derived for such stars in the lmc @xcite and our own galaxy @xcite .
the _ spitzer _ data of extreme stars in the lmc from riebel et al .
( 2012 ) are shown in the two panels of fig .
we also display the smaller group of spectroscopically confirmed c- and o - rich agbs ( and luminous red supergiants , rsgs ) from woods et al .
( 2011 ) as well as the small sample of extreme oh / ir stars from marshall et al .
the position of our models during different stages of the agb phase are also indicated .
in the @xmath29-[8.0]$ ] vs. @xmath1-[4.5]$ ] plane the theoretical models of c - stars define a diagonal sequence , that nicely fits the observed colors .
we interpret the reddest objects , with @xmath1-[4.5 ] > 1.5 $ ] , as the latest evolutionary stages of stars of initial mass @xmath30 , just below the threshold to ignite hbb .
these are the only progenitor masses in which the surface c reaches abundances @xmath31 , with a great production of carbon dust , provoking a strong obscuration of the central star .
we expect that only a few stars populate this region , as confirmed by the observations .
the observed spread in @xmath29-[8.0]$ ] for a given value of @xmath1-[4.5]$ ] is due to the different relative distribution of sic and carbon dust .
models with a high sic / c ratio exhibit a more prominent feature at @xmath32 m , and will be less obscured , particularly in the @xmath1-[4.5]$ ] color .
this is the case of stars with mass @xmath33 , that will populate the upper region in the color - color plane at @xmath1-[4.5 ] < 1.5 $ ] .
strong production of carbon dust limits the formation of sic , owing to the strong acceleration of the wind .
therefore , the stars whose sed exhibits the most pronounced feature of sic are those of initial mass below @xmath34 , in which the surface c abundance keeps below @xmath35 . more massive models ( @xmath36 ) achieve extremely high surface c abundances . in the latest agb phases ,
these models evolve at @xmath1-[4.5 ] > 1.5 $ ] and the sed is characterized by the appearance of a weak absorption sic feature at @xmath37 .
thus , our finding confirms the interpretation / observations of the lmc `` extremely red objects '' by gruendl et al .
( 2008 ) .
c - stars models also fit the observed _ spitzer _ cmd . the observed spread in the @xmath38 m flux
is a luminosity effect ; higher mass stars evolve on more massive cores , thus at larger luminosities ( see middle panel of fig . [ fevol ] ) , and smaller @xmath39 $ ] magnitudes . in both panels of fig.2
we see that the position of the spectroscopically confirmed c - rich agbs by woods et al .
( 2011 ) is nicely reproduced by our models .
the only exception is the group of objects at @xmath1-[4.5 ] < 0 $ ] , representing the c - stars with little ( if any ) dust in their envelopes .
their _ spitzer _
colors and magnitudes are reproduced by the dust - free photosphere models by aringer et al .
( 2009 ) .
agb stars experiencing hbb trace a different path . as shown in fig .
[ fevol ] , these stars evolve at smaller optical depths compared to c - stars , because the extinction coefficient of carbon dust exceeds that of silicates . in the ccd , the theoretical tracks are confined in the region @xmath1-[4.5 ] < 0.5 $ ] .
the slope of the line traced by the theoretical tracks is different from that traced by c - stars ; this is because the strong silicates feature determines a great increase in the @xmath38 m flux , with a considerable increase in the @xmath29-[8.0]$ ] color .
massive agbs with @xmath40 evolve at @xmath41 ( see fig .
[ fevol ] ) , and populate the region at @xmath29-[8.0 ] > 0.5 $ ] . unlike their lower mass counterparts ,
the reddest colors are not reached in the final stages of the agb evolution , but during the phase of maximum luminosity , when the hbb experienced is strongest . in the cmd ,
most of the m - stars populate the regions at @xmath1-[8.0 ] < 1 $ ] , and mix with the c - stars .
the two sequences bifurcate at @xmath39 \sim 8 $ ] , because the sequence of massive agbs , owing to the effects of the silicates feature , trace a steep diagonal line , up to @xmath39 \sim 5 $ ] .
interestingly , the extreme lmc oh / ir stars ( very massive agbs ) are located at the highest [ 8.0 ] magnitudes in good agreement with our massive models ; indeed the four oh / ir stars closer to the model predictions are confirmed very massive rb - rich lmc agbs @xcite .
models of mass above @xmath42 experience strong hbb , and evolve at luminosities above the classic limit for agbs , @xmath43 ( i.e. @xmath44 ) . in short ,
our interpretation of the extreme stars in the lmc is in good agreement with the observations of spectroscopically confirmed c- and o - rich lmc agbs by @xcite .
the only difference is in the identification of high luminosity stars with @xmath39 \sim 5 - 6 $ ] , that we interpret as massive agbs experiencing strong hbb , whereas woods et al .
( 2011 ) identify as rsgs .
our work suggests that rsgs may be distinguished from massive agbs / sagbs in the cmd _
diagram but not in the ccd one ( see fig .
2 ) ; indeed a small fraction of rsgs may be truly massive agbs / sagbs . the o - rich sample by woods et al .
( 2011 ) is nicely reproduced by our massive agb models in the two _ spitzer _ diagrams ( fig .
2 ) , with the exception of the the bluest objects , at @xmath1-[4.5 ] < 0 $ ] that , like their c - rich counterparts , have almost dust - free envelopes .
for the first time , we are able to interpret the _ spitzer
_ observations of extreme stars in the lmc on the basis of evolutionary models of dusty agb stars .
we find that the main diagonal sequence traced by the observations in the @xmath29-[8.0]$ ] vs. @xmath1-[4.5]$ ] diagram is an evolutionary sequence of c - stars that become progressively more obscured as their surface layers dredge - up carbon from the stellar interior .
we identify the reddest objects , with @xmath1-[4.5 ] > 1.5 $ ] , as the descendants of @xmath30 stars in the latest evolutionary phases , when they develop an optically thick circumstellar envelope .
the distribution of stars in the _ spitzer _ ccd is generally determined by the optical depth , but it is also dependent on the relative percentages of carbon and sic dust formed .
stars with a high sic / carbon ratio populate the upper region of the diagram .
this motivates the spread in @xmath29-[8.0]$ ] observed for @xmath1-[4.5 ] < 1.5 $ ] ; the scatter disappears for more obscured objects , because the dust formed is entirely dominated by solid carbon .
massive agbs with @xmath45 experience hbb , which prevents them from becoming c - stars ; in their circumstellar envelopes the formation of silicates occurs .
the strong feature of silicates at @xmath46 m favours a considerable increase in the @xmath47 m flux , thus the slope traced by the evolutionary sequences in the _ spitzer _ ccd is different from c - stars . owing to the small extinction coefficients of silicates , m
- stars do not reach extremely red colors ; most of dusty massive agbs are confined in the region @xmath1-[4.5 ] < 0.5 $ ] , also populated by c - stars . the most massive agbs experience a phase of strong hbb , when the production of silicates is strongly enhanced : both in the _ spitzer _ ccd and cmd these stars will evolve above the main sequence , where most of the stars are detected .
follow - up ( optical and near - ir ) spectroscopic observations of the extreme stars in the lmc will be extremely useful to confirm our analysis .
the authors are indebted to the anonymous referee for the careful reading of the manuscript and for the detailed and relevant comments , that helped to increase the quality of this work .
d.a.g.h . acknowledges support provided by the spanish ministry of economy and competitiveness under grants aya@xmath482011@xmath4827754 and aya@xmath482011@xmath4829060 .
p.v . was supported by prin miur 2011 `` the chemical and dynamical evolution of the milk way and local group galaxies '' ( pi : f. matteucci ) , prot .
2010ly5n2 t . r.s . acknowledges funding from the european research council under the european union s seventh framework programme ( fp/2007 - 2013 ) / erc grant agreement n. 306476 .
99 abia c. , busso m. , gallino r. , dominguez i. , straniero o. , isern j. , 2001 , apj , 559 , 1117 aringer , b. , girardi , l. , nowotny , w. , marigo , p. , lederer , m. t. 2009 , a&a , 503 , 913 blcker t. , 1995 , a&a , 297 , 727 blum r. d. et al . , 2006 ,
aj , 132 , 2034 boyer m. l. et al . , 2011 ,
aj , 142 , 13 canuto v.m.c . , mazzitelli i. , 1991 , apj , 370 , 295 cioni m. r. l. , girardi l. , marigo p. , habing h. j. , 2006 , a&a , 448 , 77 cloutmann , l. , & eoll , j.g .
1976 , apj , 206 , 548 di criscienzo m. , dellagli f. , ventura p. et al .
2013 , mnras , 433 , 313 epchtein n. et al . , 1994 ,
ap&ss , 217 , 3 ferrarotti a.d .
, gail h.p . , 2006 ,
a&a , 553 , 576 gallerani s. , maiolino r. , juarez y. , et al .
, 2010 , a&a , 523 , a85 garca - hernndez d. a. , garca - lario p. , plez b. et al .
2006 , science , 314 , 1751 garca - hernndez d. a. , garca - lario p. , plez b. et al .
2007 , a&a , 462 , 711 garca - hernndez d. a. , manchado a. , lambert d. l. et al .
2009 , apj , 705 , l31 gruendl , r. a. , chu , y .- h . , seale , j. p. et al .
2008 , apj , 688 , l9 herwig f. , 2005 , ar&a , 43 , 435 ivezic z. , nenkova m. , elitzur m , 1999 , user manual for dusty , univ .
kentucky internal rep .
marshall , j. r. , van loon , j. th . , matsuura , m. et al .
2004 , mnras , 355 , 1348 meixner , m. , gordon , k. d. , indebetouw , r. et al .
2006 , aj 132 , 2268 marigo p. , 2002 ,
a&a , 387 , 507 nanni a. , bressan a. , marigo p. , girardi l. 2013 , mnras , 434 , 2390 nanni a. , bressan a. , marigo p. , girardi l. 2014 , mnras , 438 , 2328 renzini a. , voli m. , 1981 , a&a , 94 , 175 riebel , s. , srinivasan s. , sargent , b. , meixner , m. 2012 , apj , 753 , 71 sargent b. a. , srinivasan s. , meixner m. 2011 , apj , 728 , 93 siess , l. , 2006 , a&a , 448 , 717 skrutskie m. f. et al . , 2006 , aj , 131 , 1163 srinivasan s. et al . , 2009 , aj , 137 , 4810 srinivasan s. et al . , 2011 ,
a&a , 532 , a54 ueta t. , meixner m. , 2003 , apj , 586 , 1338 valiante r. , schneider r. , salvadori s. , bianchi s. , 2011 , mnras , 416 , 1916 ventura p. , dantona f. , 2005 , a&a , 431 , 279 ventura p. , dantona f. , 2009 , mnras , 499 , 835 ventura p. , di criscienzo m. , schneider r. et al . 2012a , mnras , 420 , 1442 ventura p. , di criscienzo m. , schneider r. et al . 2012b , mnras , 424 , 2345 ventura p. , dellagli f. , schneider r. , di criscienzo m. et al .
2014 , mnras , 439 , 977 ventura p. , marigo p. , 2009 , mnras , 399 , l54 ventura p. , marigo p. , 2010 ,
mnras , 408 , 2476 wachter a. , winters j. m. , schrder k. p. , sedlmayr e. , 2008 , a&a , 486 , 497 woods p. m. et al . , 2011 , mnras , 411 , 1597 zaritsky d. , harris j. , thompson i. , 1997 , aj , 114 , 1002 | we trace the full evolution of low- and intermediate - mass stars ( @xmath0 ) during the asymptotic giant branch ( agb ) phase in the _ spitzer _ two - color and color - magnitude diagrams .
we follow the formation and growth of dust particles in the circumstellar envelope with an isotropically expanding wind , in which gas molecules impinge upon pre existing seed nuclei , favour their growth .
these models are the first able to identify the main regions in the _ spitzer _ data occupied by agb stars in the large magellanic cloud ( lmc ) . the main diagonal sequence traced by lmc extreme stars in the [ 3.6]-[4.5 ] vs. [ 5.8]-[8.0 ] and [ 3.6]-[8.0 ] vs. [ 8.0 ] planes are nicely fit by carbon stars models ; it results to be an evolutionary sequence with the reddest objects being at the final stages of their agb evolution .
the most extreme stars , with @xmath1-[4.5 ] > 1.5 $ ] and @xmath1-[8.0 ] > 3 $ ] , are @xmath2 stars surrounded by solid carbon grains . in higher mass ( @xmath3 ) models dust formation is driven by the extent of hot bottom burning ( hbb ) - most of the dust formed is in the form of silicates and the maximum obscuration phase by dust particles occurs when the hbb experienced is strongest , before the mass of the envelope is considerably reduced .
[ firstpage ] stars : abundances stars : agb and post - agb .
ism : abundances , dust |
on april 30 , sn 1999by was independently discovered at about @xmath5 by @xcite , and @xcite .
images with the katzman automatic imaging telescope @xcite of the same field provide an upper limit of @xmath6 on april 25.2 ut .
the supernova ( sn ) was found in the sb galaxy ngc 2841 which has been the host of three previous supernovae : sn 1912a , sn 1957a , and sn 1972r @xcite . based on optical spectra ,
sn 1999by was identified as a type ia sn @xcite .
shortly thereafter , @xcite reported that sn 1999by showed stronger than normal 5800 absorption and depressed flux near 4000 , suggesting that sn 1999by would be a significantly subluminous type ia event . according to @xcite ,
sn 1999by reached a maximum light of @xmath7 on ut 1999 may 10.5 , and a maximum in the @xmath8 band of @xmath9 .
the lyon / meudon extragalactic database ( leda ) , gives the heliocentric radial velocity of ngc 2841 , corrected for virgo infall , as @xmath10 km s@xmath11 .
using @xmath12 km s@xmath11 mpc@xmath11 puts the distance to ngc 2841 at 12.5 mpc with a distance modulus of @xmath13 . using this distance modulus and the photometry of @xcite , the absolute peak magnitudes of sn 1999by are @xmath14 and @xmath15 .
thus , sn 1999by is underluminous by roughly 2.5 magnitudes as compared to a typical type ia supernova ( sn ia ) .
detailed analysis of the optical light curves ( lcs ) by @xcite confirm these basic results
. the light curve of sn 1999by shows a very steep post - maximum brightness decline @xmath16 of @xmath17 to @xmath18 .
based on detailed fits of the lcs , @xcite find the interstellar reddening to be @xmath19 which is in agreement with the values for galactic reddening given by ( * ? ? ?
* @xmath20 ) , and ( * ? ? ?
* @xmath21 ) .
recently , @xcite provided detailed optical lcs and spectra and found values consistent with previous data .
these measurements imply that sn 1999by was as underluminous as sn 1991bg , the prototypical example of the subluminous type ia subclass @xcite .
other members include sn 1992k @xcite , sn 1997cn @xcite , and sn 1998de @xcite .
some defining characteristics of the subclass are rapidly declining light curves ( @xmath22 ) , peak magnitudes fainter than normal by 23 mag , and redder colors at maximum light ( @xmath23 @xmath24 0.40.5@xmath25 ) .
one of the currently active areas in sne ia research concerns the nature of these subluminous events .
theoretical interpretations of subluminous sn ia include all three types of explosion mechanism : the centrally triggered detonation of a sub - chandrasekhar mass wd , deflagration and delayed detonations of massive wd and two merging wds ( see sect .
3 & woosley & weaver , 1994 ; hflich et al
. , 1995 ; nugent et al . , 1995 , milne et al . 1999 ) .
the possibility has also been raised that sn1991bg - like sne ia should not be classified with other sne ia at all as they may arise from different progenitors .
sn 1999by is one of the best observed sne ia with data superior to that of sn 1991bg .
in addition to the studies of optical light curves and spectra mentioned above , detailed polarization spectra of sn 1999by have been obtained and analyzed @xcite .
whereas ` normal ' sne ia tend to show little or no polarization ( e.g. * ? ? ?
* ) , this supernova was significantly polarized , up to 0.7% , indicating an overall asphericity of @xmath26 .
this result suggests that there may be a connection between the observed asphericity and the subluminosity in sne ia . in recent years
it has become apparent that infrared spectroscopic observations can be used as valuable tools to determine the chemical structure of sn ia envelopes .
for instance , near - infrared spectra can be used to locate the boundaries between explosive carbon and oxygen burning , or between complete and incomplete silicon burning by measuring , and iron - group lines @xcite . however , the difficulties involved in obtaining high quality infrared spectra of supernovae has limited ir studies to a small handful of objects .
the study of sn 1994d by @xcite is the only one to include both pre- and post - maximum spectra .
since most of their spectra was of the 11.3 region , @xcite concentrated on determining the origin of a feature at 1.05 .
they concluded that this feature might be due to either 1.083 or 1.0926 , but found difficulties with both identifications .
detailed modeling of sn 1994d and sn 1986 g was able to reproduce the basic nir spectral features and their evolution @xcite .
in particular , the identification of the 1.05 feature in sn 1994d as due to could be established ( recently confirmed by @xcite ) , and a broad feature , appearing between @xmath27 and 1.9 was identified as a blend of iron - group elements .
we note that the nature of subluminous sne ia is important for the use of sne ia as distance indicators .
in particular , the calibration of the brightness decline relation depends critically on subluminous sne ia because they significantly increase the required baseline observations . from this perspective
, it may turn out to be critical to determine whether all sne ia form a homogeneous class of objects or not .
however , current limitations of model calculations will not allow us to improve the accuracy of the current estimates for the absolute distance to sn 1999by , and the possible implications of our results on the use of sne ia for cosmology is not the subject of this paper . in this paper
, we present near - infrared spectra of sn 1999by covering the time evolution of the supernova from about four days before to two weeks after maximum light .
detailed models of the sn explosion based on a delayed detonation ( dd ) scenario are used to analyze the data .
the explosion models , light curves , and synthetic spectra are calculated in a self - consistent manner .
given the initial structure of the progenitor and description of the nuclear burning front , the light curves and spectra are calculated from the explosion model without any further tuning .
the observations and data reduction are discussed in 2 . in 3 , general properties of the explosion models , light curves and spectra are discussed for normal and subluminous sne ia within the dd scenario . in 4 , a specific model for this sn is chosen by matching the properties of the predicted light curve to those observed in optical studies of sn 1999by .
synthetic optical and ir spectra for this model are then presented and the latter are compared to the observed ir spectra .
detailed line identifications for the infrared features are provided .
the effects on the predicted spectra of large scale mixing during the deflagration phase also examined . in 5 , we discuss the implications for progenitors and alternative explosion mechanisms .
we close in 6 with a final discussion of the results for sn 1999by , and put our findings in context with other sne ia and different explosion scenarios .
low - dispersion ( r @xmath28 700 ) , near - infrared spectra of sn 1999by from 1.02.4 were obtained with the 2.4 m hiltner telescope at mdm observatory during the nights of 610 may 1999 .
the data were collected using tifkam ( a.k.a .
onis ) , a high - throughput infrared imager and spectrograph with an alladin 512 @xmath29 1024 insb detector .
this instrument can be operated with standard _ j _ , _ h _ , and _
k _ filters for broadband imaging , or with a variety of grisms , blocking filters , and an east - west oriented @xmath30 slit for low ( @xmath31 ) and moderate ( @xmath32 ) resolution spectroscopy .
sn 1999by was observed with three different spectroscopic setups , which covered the 0.961.80 , 1.22.2 , and 1.952.4 wavelength regions .
the observations were broken into multiple 600 s exposures . between each exposure , the supernova light was dithered along the slit to minimize the effect of detector defects and provide first - order background subtraction .
total exposure times vary from spectrum to spectrum and are listed in the log of nir observations in table [ nirlog ] .
wavelength calibration of the spectra was achieved by observing neon , argon , and xenon arc lamps .
the spectra were corrected for telluric absorption by observing a stars and early g dwarfs at similar airmass , chosen from the bright star catalog @xcite . applying the procedure described by @xcite ,
stellar features were removed from the g dwarf spectra by dividing by a normalized solar spectrum @xcite .
the results were used to correct for telluric absorption in the a stars .
hydrogen features in the corrected a star spectra were removed from the raw a star spectra and the results were used to correct the target data for telluric absorption .
[ for further discussion of this procedure see @xcite , and references therein . ]
the instrumental response was calibrated by matching the continuum of the a star telluric standards to the stellar atmosphere models of @xcite .
the first three nights ( 6 - 8 may ) were photometric , and _
j _ , _ h _ , and _
k _ broadband images of sn 1999by and @xcite photometric standards .
the resulting photometry is listed in table [ nirlog ] and was used to set the flux levels of the corresponding spectra . a square bandpass was assumed and the flux levels thus attained are believed accurate to @xmath33 2030% . for may 9 and 10 no flux information was available , but the observed spectral bands had large overlaps , and the relative flux levels were set by matching the data in the overlapping regions .
near - infrared spectra of a star @xmath34 north and @xmath35 west of sn 1999by @xcite were also obtained and reduced in the same manner .
this star was then used as a local relative spectrophotometric standard to reduce spectroscopic data from later observing runs .
spectra of sn 1999by covering 1.01.8 were collected on five more nights during the following two weeks ( 16 may , 18 may , 20 may , 21 may and 24 may ) .
the first three were obtained by p. martini and a. steed using tifkam on the 2.4 m telescope at mdm , and the last two by s. sakai using tifkam on the 2.1 m telescope at the kitt peak national observatory .
figure [ nirspec ] shows a plot of all the nir spectra .
since absolute fluxes were only available for the first three spectra , the data are presented in arbitrary flux units , and they have been shifted vertically for clarity .
regions of very low s / n due to strong telluric absorption have been omitted .
the epochs listed are relative to 10 may 1999 , the date of @xmath36 given by @xcite .
like all the subsequent plots , the data have been corrected to the 638 km s@xmath11 redshift of ngc 2841 @xcite .
there is general agreement that sne ia result from some process involving the combustion of a degenerate white dwarf @xcite . within this general picture ,
three classes of models have been considered : ( 1 ) an explosion of a co - wd , with mass close to the chandrasekhar limit , which accretes mass through roche - lobe overflow from an evolved companion star @xcite .
the explosion is triggered by compressional heating near the wd center .
( 2 ) an explosion of a rotating configuration formed from the merging of two low - mass wds , caused by the loss of angular momentum through gravitational radiation @xcite . ( 3 ) explosion of a low mass co - wd triggered by the detonation of a helium layer @xcite .
only the first two models appear to be viable for normal type ia sne as the third , the sub - chandrasekhar wd model , has been ruled out on the basis of predicted light curves and spectra @xcite .
still , theoretical interpretations of subluminous sn ia include all three types of explosion mechanism @xcite .
the possibility has also been raised that the subluminous sn 1991bg - like supernovae may arise from different progenitors and should not be classified with other sne ia at all . within @xmath37 models
, it is believed that the burning front starts as a subsonic deflagration . however , the time evolution of the burning front is still an open question .
that is , whether the deflagration front burns through the entire wd @xcite , or alternatively , transitions into a supersonic detonation mode as suggested in the delayed detonation ( dd ) model @xcite .
dd models have been found to reproduce the optical and infrared light curves and spectra of ` typical ' sne ia reasonably well ( @xcite , h95 hereafter ; @xcite ) .
in addition , dd models provide a natural explanation for the brightness decline relation @xmath38 observed in the light curves of sne ia @xcite .
this is a consequence of the temperature dependence of the opacity and the lack of a dependence of the explosion energy on the amount of @xmath0 produced ( @xcite ) .
thus , within the delayed detonation scenario , both normal and subluminous sne ia can be explained as variants of a single phenomenon ( h95 , @xcite , @xcite ) .
this makes the dd model an attractive scenario and it is used as the basis for our present analysis . in our models , the explosions lcs and spectra
are all calculated self - consistently . given the initial structure of the progenitor and a description of the nuclear burning front , the light curves and spectra are calculated directly from the explosion model without any additional parameters .
since the predicted observables follow directly from the supernova model , this approach provides a direct link between the observations and the explosion physics and progenitor properties .
finally , we note that detailed analyses of observed sn ia spectra and light curves indicate that mergers @xcite and pure deflagration sne ( such as the w7 model ) could contribute to the population of bright sne ia ( @xcite , hereafter hk96 ; @xcite ) . on the other hand , it is not obvious that subluminous sne ia and the brightness decline relation can be understood within pure deflagration models . in particular , the amount of burning , and consequently the explosion energy , will be strongly correlated with @xmath0 production .
since no detailed studies for deflagration models have been performed , these models can not be ruled out but , as we discuss in 5 , the ir spectra of sn 1999by are hard to reconcile with a pure deflagration . despite recent progress in our understanding of nuclear burning fronts , current 3-d models
are not sufficiently evolved to allow for a consistent treatment of the burning front throughout all phases ( * ? ? ?
* see also 4.3 & 5 ) .
thus detailed models rely on parameterized descriptions guided by 3-d results . within spherical models ,
the transition from deflagration to detonation can be conveniently parameterized by a density @xmath39 which has been found to be the dominating factor for the determining chemical structure , light curves and spectra .
we have calculated explosion models using a one - dimensional radiation - hydro code ( hk96 ) that solves the hydrodynamical equations explicitly by the piecewise parabolic method @xcite .
nuclear burning is taken into account using an extended network of 606 isotopes from n , p to @xmath40 ( * ? ? ?
* and references therein ) .
the propagation of the nuclear burning front is given by the velocity of sound behind the burning front in the case of a detonation wave , and in a parameterized form during the deflagration phase calibrated by detailed 3-d calculations ( e.g. * ? ? ?
we use the parameterization as described in @xcite . for a deflagration front at distance @xmath41 from the center
, we assume that the burning velocity is given by @xmath42 , where @xmath43 and @xmath44 are the laminar and turbulent velocities with @xmath45}\ ] ] with @xmath46 and @xmath47 here @xmath48 is the atwood number , @xmath49 is the characteristic length scale , and @xmath50 and @xmath51 are the densities in front of and behind the burning front , respectively . the quantities @xmath52 and @xmath49
are directly taken from the hydrodynamical model at the location of the burning front and we take @xmath53 .
the transition density is treated as a free parameter . from these explosion models , the subsequent expansion , bolometric and broad band light curves ( lc )
are calculated following the method described by @xcite , and references therein .
the lc - code is the same used for the explosion except that @xmath54-ray transport is included via a monte carlo scheme and nuclear burning is neglected . in order to allow a more consistent treatment of the expansion
, we solve the time - dependent , frequency - averaged radiation moment equations .
the frequency - averaged variable eddington factors and mean opacities are calculated by solving the frequency - dependent transport equations .
about one thousand frequencies ( in one hundred frequency groups ) and about nine hundred depth points are used . at each time step
, we use t(r ) to determine the eddington factors and mean opacities by solving the frequency - dependent radiation transport equation in a co - moving frame and integrate to obtain the frequency - averaged quantities .
the averaged opacities have been calculated assuming local thermodynamic equilibrium ( lte ) .
both the monochromatic and mean opacities are calculated in the narrow line limit .
scattering , photon redistribution , and thermalization terms used in the light curve opacity calculations are taken into account . in previous works ( e.g. @xcite )
, the photon redistribution and thermalization terms have been calibrated for a sample of spectra using the formalism of the equivalent two level approach ( h95 ) . for increased consistency
, we use the same equations and atomic models but solve the rate equations simultaneously with the light curves calculation at about every 100@xmath55 time step , at the expense of some simplifications in the nlte - part compared to h95 . for the opacities we use the narrow line limit , and for the radiation fields we use the solution of the monochromatic radiation transport using @xmath56 frequency groups .
our non - lte code ( h95 , and references therein ) solves the relativistic radiation transport equations in a co - moving frame .
the spectra are computed for various epochs using the chemical , density , and luminosity structure and @xmath54-ray deposition resulting from the light curve coder .
this provides a tight coupling between the explosion model and the radiative transfer .
the effects of instantaneous energy deposition by @xmath54-rays , the stored energy ( in the thermal bath and in ionization ) and the energy loss due to the adiabatic expansion are taken into account .
bound - bound , bound - free and free - free opacities are included in the radiation transport which has been discretized by about @xmath57 frequencies and 97 radial points .
the radiation transport equations are solved consistently with the statistical equations and ionization due to @xmath54-radiation for the most important elements and ions .
typically , between 27 to 137 bound levels are taken into account .
we use ( 27/123/242 ) , ( 43/129/431 ) , ( 20/60/153 ) , ( 35/212/506 ) , ( 41/195/742 ) , ( 62/75/592 ) , ( 137/3120/7293 ) , ( 84/1355/5396 ) , ( 71/865/3064 ) where the first , second and third numbers in brackets denote the number of levels , bound - bound transitions , and number of discrete lines for the radiation transport .
the latter number is larger because nearby levels within multiplets have been merged for the rates .
the neighboring ionization stages have been approximated by simplified atomic models restricted to a few nlte levels + lte levels .
the energy levels and cross sections of bound - bound transitions are taken from @xcite starting at the ground state .
the bound - free cross sections are taken from topbase 0.5 as implemented in straburg @xcite .
collisional transitions are treated in the ` classical ' hydrogen - like approximation @xcite that relates the radiative to the collisional gf - values .
all form factors are set to 1 .
about 10@xmath58 additional lines are included ( out of a line list of @xmath59 ) assuming lte - level populations .
the scattering , photon redistribution , and thermalization terms are computed with an equivalent - two - level formalism that is calibrated using nlte models .
spherical dynamical explosions , light curves , and spectra are calculated for both normal bright and subluminous sne ia . due to the one - dimensional nature of the models , the moment of the transition to a detonation
is a free parameter .
the moment of deflagration - to - detonation transition ( ddt ) is conveniently parameterized by introducing the transition density , @xmath60 , at which ddt happens . within the dd scenario ,
the free model parameters are : 1 ) the chemical structure of the exploding wd , 2 ) its central density @xmath61 at the time of the explosion , 3 ) the description of the deflagration front , and 4 ) the density @xmath39 at which the transition from deflagration to detonation occurs .
note that the density structure of a wd only weakly depends on the temperature because it is highly degenerate . the central density
@xmath61 depends mainly on the history of accretion .
model parameters have been chosen for the progenitor and @xmath61 which allow us to reproduce light curves and spectra of ` typical ' type ia supernovae . in all models
, the structure of the exploding c / o white dwarf is based on a model star with 5 m@xmath62 at the main sequence and solar metallicity . through accretion , this core has been grown close to the chandrasekhar limit ( see model 5p0y23z22 in @xcite ) . at the time of the explosion of the wd , its central density is 2.0@xmath63 g @xmath64 and its mass is close to 1.37@xmath65 .
the transition density @xmath39 has been identified as the main factor which determines the @xmath0 production and , thus the brightness of a sne ia ( h95 ; @xcite ; @xcite ) .
the transition density @xmath39 from deflagration to detonation is varied between 8 and 27 @xmath66 g @xmath64 to span the entire range of brightness in sne ia .
the models are identified by _
5p0z22.ext _ where _ ext _ is the value of @xmath39 in @xmath67 g @xmath64 .
some of the basic quantities are given in table [ modelprop ] . here
we will restrict our discussion to the basic features of dd models which are relevant for the spectral analysis of sn 1999by .
for a more detailed discussion of the hydrodynamical evolution of dd - models , see @xcite , h95 , and references therein . during the deflagration phase ,
the wd is lifted , starts to expand , and after burning about 0.22 to 0.28 @xmath65 , the transition to a detonation is triggered .
figures [ density ] and [ abundance ] show the density , velocity and chemical structures for representative models after a homologous expansion has been achieved .
the final velocity and density structures are rather similar .
the expansion velocities decrease slightly with @xmath39 because a significant amount of oxygen remains unburned and did not contribute to the energy production . in 5p0z22.8 ,
oxygen remained unburned in the outer @xmath68 .
overall , the density is decreasing with radius because burning took place throughout the entire wd except in the very outer layers .
in contrast , pure deflagration model such as w7 @xcite , pulsating delayed detonation @xcite , and merger models @xcite show a significant density bump at the boundary between burned and unburned layers .
iron group elements are produced in the inner layers where density and temperature stay high for a sufficient time during the explosion , whereas intermediate mass elements are produced in the layers above where nuclear statistical equilibrium ( nse ) has no time to set in . in the outer zones , products of explosive oxygen ( e.g. si , s ) and explosive carbon burning ( o , mg , ne ) and some si are seen .
only a very thin layer of unburned c / o remains . for layers with final velocities
@xmath69 km s@xmath11 , the densities in the early stages of the explosion are sufficiently high for electron capture .
this results in the production of @xmath70fe & @xmath71ni whereas in the outer layers only @xmath72ni is produced .
we employ the same description for the deflagration front in all models and , consequently , the very inner structures are almost identical in all models .
a small dip in the ni abundance is present at the ddt because the time scales for burning to nse are comparable to the expansion time scales at densities of @xmath73 g @xmath64 .
this feature can be expected to be smeared out in multi - dimensional calculations @xcite .
after the ddt , a detonation front develops and the burning of fuel is triggered by compression of unburned material .
consequently , burning takes place under higher densities compared to a deflagration .
in general , additional material is burned up to nse and this extends the layers of @xmath72ni for all @xmath74 g @xmath64 . for models with smaller @xmath39 , burning only proceeds up to si despite the compression .
this increases the production of intermediate mass elements ( e.g. si , s ) at the expense of @xmath72ni . in our models , the efficiency for @xmath72ni production drops rapidly as a function of decreasing @xmath39 ( see table [ modelprop ] ) .
we note that for @xmath75 g @xmath64 , nuclear burning stops all together for @xmath76 and , unlike any observed sn , one would see a very slowly expanding c / o envelope heated by little @xmath72ni .
there exists a rather strict lower limit for the absolute brightness of sne ia within the chandrasekhar mass models .
this can be understood as follows .
the absolute brightness of a sne ia is mainly determined by the @xmath0 . in general , @xmath0 is produced both during the deflagration and the detonation phase . for the most subluminous sne
ia however , no @xmath0 is produced during the detonation . within the @xmath37 scenario and dd - models in particular ,
we need to burn a minimum of @xmath77 during the deflagration to achieve the required ` lift / pre - expansion ' of the wd . during the deflagration phase
, iron group elements are produced .
the fraction of @xmath0 depends on the amount of material which undergoes electron capture , i.e. on the central density and , to a smaller extent , on the description of the burning front ( @xcite ; @xcite ) . for a model with @xmath78 g @xmath64
, we find a lower limit of @xmath79 .
recently , @xcite studied the influence of @xmath80 between 1.5 to @xmath81 g @xmath64 on the nuclear burning for similar dd - models with a high transition density which may describe ` normal ' sn ia . in those models , the @xmath0 produced during the deflagration phase varies by @xmath82 which provides an estimate for the possible variation in the minimum @xmath83 production .
we conclude that there exists a lower limit of @xmath84 within dd - models . in principle , 3-d effects may change the lower limit for the @xmath0 production .
however , all current 3-d calculations show that rising blobs are formed at high densities which undergo complete si - burning because the burning time scales to ni shorter than the hydro time scales . in the case of a ddt or a transition to a phase of very fast deflagration , the amount of burning will likely be similar to the 1-d case .
thus , we do not expect a significant change in the minimum @xmath85 production .
optical light curves and spectra of sn ia are not the subject of this paper but we used their general properties for selecting a model for sn1999by . within the dd scenario ,
both normal and subluminous sne ia can be produced ( h95 , @xcite , hkw95 hereafter ; @xcite ) .
the amount of @xmath72ni ( @xmath86 ) depends primarily on @xmath39 ( h95 ; hkw95 ; @xcite ) and to a much lesser extent on the assumed value of the deflagration speed , initial central density of the wd , and initial chemical composition ( ratio of carbon to oxygen ) .
models with smaller transition density give less nickel and hence both lower peak luminosity and lower temperatures ( hkw95 , @xcite ) . in dds ,
almost the entire wd is burned and the total production of nuclear energy is almost constant .
this and the dominance of @xmath39 for determining the @xmath0 production are the basis of why , to first approximation , sne ia appear to be a one - parameter family . the observed @xmath16 relation can be understood as an opacity effect @xcite , namely , the consequence of rapidly dropping opacities at low temperatures @xcite .
less ni means lower temperature and consequently , reduced mean opacities because the emissivity is shifted from the uv toward longer wavelengths with less line blocking .
a more rapidly decreasing photosphere causes a faster release of the stored energy and , as a consequence , steeper declining lcs with decreasing brightness .
the dd models give a natural and physically well - motivated origin of the @xmath16 relation of sne ia ( @xcite ) . the broad band _
b _ and _ v _ light curves are shown in figure [ lcs ] .
some general properties are given in table [ modelprop ] and figure [ dm15 ] .
we note that the decline of the _ b _ and _ v _ lc during the early @xmath87 tail is slightly smaller for the most subluminous models ( e.g. 5p0z22.08 ) compared to the intermediate models ( e.g. 5p0z22.16 ) because , at day 40 , the @xmath88-rays are almost fully trapped in the former whereas , for the latter , the escape probability of @xmath88-rays is already increasing ( hkw95 ) . with varying @xmath39 , we find rise times between 13.9 to 19.0 days , and the maximum brightness @xmath89 spans a range @xmath90 and @xmath91 with @xmath92 between @xmath93 to @xmath94 .
the value of @xmath89 is primarily determined by the transition density .
variations in the progenitors ( main sequence mass and metallicity ) or the central density at the point of ignition causes an spread around @xmath89 of @xmath95 @xcite . as discussed above , the brightness decline relation @xmath16 can be understood as an opacity effect or , more precisely , it is due to the drop of the opacity at low temperatures . therefore , subluminous models are redder and the decline is steeper . in our series of models , @xmath16 ranges between @xmath18 and @xmath96 .
however , @xmath16 is not a strictly linear relation ( fig . [ dm15 ] ) . for bright sne ia ( @xmath97 ) ,
called _ branch - normals _ ( @xcite ) , the linear decline is relatively flat , followed by a steeper decline toward the subluminous models . for normal - luminosity models ,
the photospheric temperature is well above the critical value at which the opacity drops ( @xmath98 ; @xcite ) whereas the opacity drops rapidly in subluminous models soon after maximum light .
detailed b and v light curves of sn 1999by have been collected and published by @xcite and optical spectra around maximum light are available from observations at mcdonald observatory @xcite .
recently , sn 1999by spectra and lcs have been provided by @xcite . within our series of models , we selected the most suitable model based on the optical photometry .
unfortunately , there is a gap in the observed data starting about one week after maximum light , a phase which is most suitable for selecting the appropriate models .
thus , as the main discriminator , we use the complete lcs as reconstructed by @xcite .
sn 1999by showed a decline rate @xmath16 between @xmath99 and @xmath18 ( @xcite ) clearly ruling out all but the most subluminous models with transition densities of @xmath100 and @xmath101 g @xmath64 and maximum brightness of @xmath90 and @xmath102 , respectively ( table [ modelprop ] ) .
both models are consistent with the early light curves given in @xcite i and @xcite ( fig.[lc.obs ] ) . from the models and the early lc
, the time of the explosion can be placed around april 27 @xmath103 1 d , 1999 .
the model post - maximum lc in b is less reliable because it depends very sensitively on the size and time evolution of line blanketing and the thermalization .
@xmath104 is @xmath105 and @xmath106 for 5p0z22.10 and 5p0z22.8 , respectively .
this compares reasonably well with @xmath107 observed for sn1999by @xcite . in the lcs ( fig .
[ lc.obs ] ) , discrepancies of about @xmath108 show up after the time gap in the observations . at these times , the envelope of a subluminous sn ia becomes transparent , and strong emission features emerge ( see below ) .
thus these discrepancies are likely a consequence of discretization errors due to the use use of only 1000 frequency for the lc calculations . according to @xcite , @xmath109 near maximum light on may 10 , 1999 . @xcite give @xmath110 .
we find @xmath111 and @xmath112 for 5p0z22.8 and 5p0z22.10 , respectively . in previous studies ,
we find intrinsic uncertainties in b - v to be @xmath113 for our models at maximum light @xcite .
thus , the interstellar reddening can be expected to be @xmath114 consistent with the values for our galaxy given by @xcite and @xcite and the estimates for the total extinction derived by @xcite . for our models , if we assume @xmath115 , we then derive a distance modulus of @xmath116 ( @xmath117 to 0.35@xmath25 ) .
the first and second error - terms originate from the observational uncertainties plus the brightness range of the models and the reddening correction , respectively .
> from our models , we can derive a distance to ngc 2841 of @xmath118 mpc and @xmath119 mpc if we include and neglect interstellar reddening , respectively .
both from observations @xcite and our models , very small reddening is preferred .
following @xcite , we adopt @xmath20 for all comparisons with the observations .
based on the density , temperature and chemical structure predicted from the explosion model , @xmath120 transport and light curve calculations , we have constructed detailed nlte - spectra for 11 d , 15 d , 22 d and 29 d after the explosion ( days @xmath4 , 0 , + 7 and + 14 after maximum light ) . for the subluminous model with the transition density @xmath121 g @xmath64 ( fig . [ density ] )
, the evolution of the temperature , energy deposition by @xmath54-rays , and rosseland optical depth is given as a function of radius ( fig .
[ structures ] ) . obviously , the total optical depth drops rapidly with time from 98 , 40 , 15 to 6 due to the cooling ; much faster than quadratic as would be expected from the geometric dilution alone ( 98 , 52 , 25 and 12 ) , and thus more rapidly than in normal - bright sne ia @xcite ) . as mentioned above
, the reason is the rapid drop in the mean opacity for @xmath122 which causes the outer layers to become almost transparent .
consequently , the photosphere recedes quickly in mass and even starts to shrink in radius just a few days after the explosion .
the corresponding expansion velocities at the photosphere are approximately 14000 , 10500 , 6500 and 4000 km s@xmath11 at the four epochs computed .
the low photospheric temperatures are due to the small amount of @xmath0 and its low expansion velocity which also causes the local trapping of @xmath54-rays ( fig .
[ structures ] ) .
in contrast to normal sne ia , most of the decay energy is deposited well below the photosphere up to about 2 weeks after maximum light .
this explains , in part , the low ionization stages seen in the ir - spectra ( see below ) .
the location ( and thus the velocity ) of the photosphere determines which layers in the supernova envelope will form the spectral features . before maximum light , the spectra sample the layers of explosive carbon burning which are o , mg , ne and si - rich ( fig .
[ abundance ] ) . up
to 2 weeks after maximum light , spectra are formed within layers of incomplete silicon burning which are si , s and ca - rich .
only thereafter are the layers which are dominated by iron group elements finally exposed .
this is very different from models for normal luminosity sne ia for which the iron rich layers are already exposed near maximum light .
these profound differences are key for our understanding of the optical and ir spectra of subluminous sne ia and the differences compared to normal sne ia .
optical spectra are not the main subject of this paper , and detailed spectra have not been available for this analysis except for maximum light ( fig
. [ mod.opt2 ] , @xcite ) .
however , for completeness , the time sequence of optical spectra is given in figure [ mod.opt ] .
overall , the spectra show an evolution typical for sn ia , ( e.g. sn 1994d , h95 ) . they can be understood as a consequence of the declining temperature and total optical depth with time , and the rapidly dropping temperature .
the doppler shift of the absorption component decreases rapidly because of the receding photosphere . for subluminous models ,
the entire envelope becomes almost transparent and the spectra start to enter the nebular phase at about two weeks after explosion .
this marks the end of the applicability of our atomic models .
some features ( see fig .
[ mod.opt2 ] ) are typical of subluminous sne ia and are a consequence of the low luminosity and photospheric temperature around maximum light . the line at 5800 is strong relative to the line at 6150 , and the flux near 4000 was depressed as in other subluminous sne , and features due to were strong as was the line at 7500 @xcite . as already suggested by @xcite , the strong absorption and the relative strength of the si lines can be understood as a temperature effect . within dd models with @xmath123 progenitors ,
the strong line is a direct consequence of the fact that the spectra at maximum light are formed in massive layers of explosive carbon burning .
while the structures and synthetic spectra of the models 5p0z22.8 and 5p0z22.10 are very similar , 5p0z22.8 is slightly preferred from a consideration of the expansion velocities at the photosphere which are 10500 and 11500 km s@xmath11 , respectively .
therefore , we will will adopt 5p0z22.8 in the following . based on parameterized lte model atmospheres ( synow , @xcite ) ,
a detailed analysis of the doppler shifts of lines has been provided by @xcite .
they found in the entire range between 11,300 down to 6500 km s@xmath11 which is consistent with our hydrodynamical models ( see fig .
[ structures ] ) . near ir - spectra ( 0.9 to 2.5 @xmath124 m ) of 5p0z22.8 at day 11 , 15 , 22 and 29 after explosion are compared to those of sn1999by at may 6 , 10 , 16 , and 24 ( figs .
refir.may6[ir.may24 ] ) .
the spectra are characterized by overlapping p - cygni and absorption lines , and emission features .
unless noted otherwise , the observed wavelengths quoted the emission component . the overall energy distribution in the continuum is determined by bound - free and thomson opacities and the temperature evolution .
in general , the spectra and their time evolution agree reasonably well with the model predictions and all the major features can be identified .
few detailed analyses and line identifications of ir spectra have been performed @xcite and these were limited to spectra of sn 1994d , sn 1998bu and post - maximum spectra of sn1986 g .
to put our analysis given below in context with normal - luminosity sne ia , we will refer to those works as reference points . the observed nir spectra from may 6 and 10 ( figs .
[ ir.may6 ] & [ ir.may10 ] ) are quite similar , and comparison with the model spectra show that both of these spectra are formed in layers of explosive carbon burning .
spectral features can be attributed to , , and .
typically , several thousand overlapping lines contribute to the overall opacity but few strong features leave their mark on the spectrum . some weak features at about 2.1 and 2.38 are due to ( 21369 ,
21432 , 24041 , 24044 , 24125 ) and ( 21920 , 21990 ) .
the strong , broad feature between 1.62 and 1.75 is a blend produced by lines ( 16760 , 16799 , 174119 , 17717 ) and ( 16907 , 16977 , 17183 , 17184 ) .
its blue edge is determined by .
a similar feature is also present in normal luminosity sne ia but does not contribute because the corresponding spectra are formed in layers of explosive oxygen burning .
the strong feature that peaks near 1.35 is due to ( 13650 ) , and weaker features with peaks near 1.23 and 1.43 are a multiplet ( 13692 to 13696 ) and a blend of ( 14542 ) and ( 14454 ) , respectively . a very weak feature in the synthetic spectrum near 1.23 can be attributed to ( 13164 ) , but is well below the noise level of the observed spectra
. features with p - cygni absorption minima at 1.12 , 1.06 , 1.03 and 1.0 are produced by ( 11620 , 11600 ) , ( 11302 , 11286 ) , ( 10914 , 10915 & 10950 ) , ( 10683 & 10691 ) , and ( 10092 ) , respectively .
the multiplet of 1.09 is also prominent in normal luminosity sne ia but the other features are weak and blended with , and .
the observed spectra go down to about 9700 .
however , there are some interesting features at shorter wavelengths which are worth mentioning .
foremost , the model predicts a very strong feature at about 9100 due to ( 9405 ) .
its velocity provides an important constraint on the minimum velocity for the unburned region , and possible mixing .
other features are produced by ( 9344 ) , ( ir - triplet ) , and ( 8446 ) and ( 8246 ) . despite the overall agreement
, some discrepancies remain .
first , the line at 10691 is somewhat too strong , and its doppler shift on may 6 is too large by about 1500 km s@xmath11 .
this feature is formed well above the photosphere ( 13000 km s@xmath11 ) .
this might indicate that sn 1999by has a slightly lower temperature in the outer layers compared to the model , which would cause less excitation of carbon and , might cut off the high velocity contribution to the absorption .
secondly , in the model spectrum there seems to be a lack of emission produced by near 1.18 .
we note that similar problems with si are also commonly present in the optical line at about 6380 for normal sne ia @xcite .
they may be related to the excitation process of si but their origin is not well understood . the spectrum from may 16 ( fig . [ ir.may16 ] ) is formed in layers of explosive oxygen burning .
the lines can be attributed to , , and .
all features due to and seen in the pre - maximum spectra can still be identified .
no lines due to and are present , but strong lines of and appear .
strong features at about 1.14 and 1.1 are due to lines ( 11784 , 11839 & 11849 ) and ( 10821 ) & ( 11800 ) , and the broad feature at 0.91 ( not covered in this spectrum , but included for completeness ) is a blend of ( 9890 ) and ( 10009 ) .
weaker lines ( , 9997 ; , 10455 ) contribute to blends around 1.03 .
some weak lines due to and lines occur throughout the spectrum but the most characteristic feature in post maximum ir - spectra of normal sne ia is missing . starting a few days after maximum , normal sne ia show a wide emission feature between 1.5 and 1.8 produced by thousands of , & lines
. typically , their emission flux is about twice as large as the adjoining continuum .
this feature marks the lower velocity end of the region of complete silicon burning , and it seems to be common in all normal luminosity sne ia such as sn 1986 g , sn 2000br and sn 2000cx ( @xcite ; marion 2001 , priv .
@xcite ) . in 5p0z22.8 , this broad emission feature
is not seen because the photosphere has yet to recede to the layers of complete silicon burning . on may 24th ( fig . [ ir.may24 ] )
this feature finally appears but is very weak .
overall , two weeks after maximum light , the spectrum in dominated by a large number of and lines .
strong features of are around 9600 1.2 @xmath124 m and 1.4 @xmath124 m with strong lines from 9500 to 12443 , and 14000 16700 . in our synthetic spectrum
the , & feature 1.6 to 1.8 @xmath124 m does not show a local minimum at @xmath125 as observed in sn1999by , and as also predicted by normal bright models for sn1986 g @xcite . in our subluminous model ,
the emission near 1.7 comes from the central ni - rich region .
the absence of such emission in the observed spectrum may indicate that the very central region ( @xmath126 ) of the model is too transparent compared sn1999by .
this discrepancy is not critical for measuring the chemical structure of the envelope because this is determined by the blue edge .
the difference in opacity may be caused by missing iron lines in the line list ( 3.1.3 ) , or an underestimation of the excitation by non - thermal electrons because we assume local deposition for the electrons from the @xmath127 decay of @xmath87 .
as we have seen , model 5p0z22.8 reproduces the basic features of both optical and ir spectra , including the time evolution .
this suggests that dd models are a reasonably good description for subluminous sne ia .
unfortunately , we are still missing a handle on an important piece of physics .
the propagation of a detonation front is well understood but the description of the deflagration front and the deflagration to detonation transition ( ddt ) pose problems .
currently , state of the art allows us to follow the front only through the phase of linear instabilities , i.e. only the early part of the deflagration phase .
the resulting structures of these calculations can not account for the observations of typical sne ia ( khokhlov 2001 ) , because a significant fraction of the wd ( @xmath128 ) remains unburned .
nevertheless , these calculations point toward a more general problem , or toward a better understanding of the nature of subluminous sne ia . on a microscopic scale , a deflagration propagates due to heat conduction by electrons .
though the laminar flame speed in sne ia is well known , the front has been found to be rayleigh - taylor ( rt ) unstable , increasing the effective speed of the burning front @xcite .
more recently , significant progress has been made toward a better understanding of the physics of flames .
starting from static wds , hydrodynamic calculations of the deflagration fronts have been performed in 2-d @xcite , and full 3-d @xcite .
the calculations by @xcite demonstrated a complicated morphology for the burning front .
plumes of burned material will fill a significant fraction of the wd , and unburned or partially - burned material can be seen near the center .
thus iron - rich elements will not be confined to the central region as in 1-d models .
@xcite finds that while the expansion of the envelope becomes almost spherical , the inhomogeneous chemical structure will fill about 50 to 70% of the star ( in mass ) .
if a ddt occurs at densities needed to reproduce normal - bright sne ia , most of the unburned fuel in these regions will be burned to iron - group elements during the detonation phase .
this will eliminate the chemical inhomogeneities .
however , in our subluminous model , the chemical structure imposed during the deflagration phase must be expected to survive because no or very little @xmath0 is produced during the detonation . in order to test for this effect
, we have mixed the inner layers of model 5p0z22.8 up to an expansion velocity of 8000 km s@xmath11 .
this brings iron - group elements into the outer regions and mixes intermediate mass elements ( si , s , ca and some o ) into the inner , hotter layers . the layers with expansion velocities less than 8000 km s@xmath11 become visible a few days after maximum light ( see above ) . in figure
[ mix.may16 ] , a comparison is shown between the spectrum of sn1999by on may 16th , 1999 , and the synthetic spectrum about one week after maximum light , and problems with the spectral fit are apparent . as can be expected from the discussion above
, the effective extension of the iron - rich layers to 8000 km s@xmath11 causes the 1.5 to 1.8 blend to become rather strong , and line blanketing due to iron - group elements ( & ) become too strong in the region around 1 to 1.2 .
moreover , at about 0.99 , 1.02 , 1.05 & 1.16 , ( absorption components ) strong blends can be seen due to ( 11302 , 11287 & 11297 ) , ( 11715 , 11745 , 13650 & 13650 ) , ( 9890 , 9931 , 9997 , 11839 & 11745 ) and ( 10455 & 108212 ) .
we have to take into account that our approach assumes microscopic mixing whereas rt instabilities provide large scale inhomogeneities .
if @xmath0 and the intermediate mass elements are separated , direct excitation of intermediate mass elements may be reduced . with respect to the radiation transport
, the main difference is the covering factor .
this will effect line blanketing and thus the line blanketing between 1 to 1.2 @xmath124 m may not pose a problem . however , the emission features between 1.5 to 1.8 @xmath124 m will not go away . the reduced excitation by non - thermal electrons and @xmath54 s in case of macroscopic mixing are important at low optical depths . less excitation may reduce the problem with the o , si , s & ca lines but the problem will no disappear because the corresponding features are formed close to the photosphere where the excitation is thermal in nature . these problems strongly suggest that large scale mixing did not occur in sn 1999by .
this may provide a key for our understanding of the nature of these objects .
as previously discussed , the pre - conditioning of the wd before the explosion may provide a key for understanding the nature of subluminous sne ia .
such pre - conditioning may include the main sequence mass of the progenitor mass and its metallicity @xcite , the accretion history @xcite , or large - scale velocity fields such as turbulence prior to the runaway ( @xcite ) .
up to now , none of these effects have been included in detailed 3-d calculations .
therefore , the suggestions ( below ) require further investigation before taken as explanation . beyond its subluminous nature ,
sn 1999by showed another peculiarity with respect to normal - bright sne ia .
the observed polarization of sn 1999by was rather high ( @xmath129 ) with a well - defined axis of symmetry , whereas normal sne ia show little or no polarization ( @xmath130 ; @xcite ) .
> from a detailed analysis of the polarization spectra , @xcite concluded that the overall geometry of sn 1999by showed a large scale , probably rotational asymmetry of about 20% .
the authors suggested several explanations . either
, sn 1999by may be the result of the explosion of a rapidly rotating wd close to the break - up , or a result of a merger scenario . in
a rapidly rotating wd , large - scale circulations @xcite might influence the deflagration phase of explosive burning , perhaps by breaking up large eddies and preventing their rise .
alternatively , it should be remembered that the chemical structure of the exploding wd depends on the pre - expansion of the wd prior to the detonation phase .
more precisely , we need to burn about @xmath131 to lift the wd in its gravitational potential .
this mass is significantly more than is burned during the smoldering phase ( the phase of slow convective burning just prior to the explosion ) . even in the presence of strong turbulence ,
the amount of burned material is much less ( @xmath132 , @xcite ) . in all models for the ignition ( e.g. @xcite )
, the central temperature rises by non - explosive carbon burning during the last few minutes to hours before the runaway because the heat can not be dispersed .
it may be feasible that large scale velocity fields due to rotation may extend this phase , and significantly increase the amount of burning prior to the explosion .
thus , the pre - expansion in subluminous sne ia may occur directly during the smoldering phase , and the deflagration phase may be even skipped , i.e. a detonation occurs promptly , and we may have a smoldering detonation model ( sd ) . in sd - models ,
no @xmath0 is produced during the phase of pre - expansion and , consequently , this mechanism does not predict a lower limit for @xmath133 and , thus , @xmath89 . in principle , sd - models may be an attractive alternative also for normal - bright sne because they omit the need for a ddt but , from the present understanding of the runaway , in general , extensive burning prior to the runaway can not be expected ( see above ) . from the point of the spectral analysis ,
this option is intriguing because burning during the smoldering phase precedes only to o and the time scales of burning are sufficient to completely homogenize the chemical structure of the wd . finally , there is the question of alternative scenarios for sne ia in the context of subluminous sne ia .
pure deflagrations of @xmath37-wds without any ddt may produce little @xmath134 but , according to current simulations , a large fraction of the wd would remain unburned .
even if this problem were solved in the future , the problem of inhomogeneous chemistry and si - rich layers close to the center would remain .
helium triggered detonations of sub - chandrasekhar mass wds have been suggested in the past . however , their distinct feature is an outer layer of @xmath0 above a low mass layer of explosive oxygen burning , a layer of incomplete si burning , and some @xmath0 close to the center .
this chemical structure is in strong contradiction to the optical spectra and lcs of sn 1999by , and their evolution with time . alternatively , setting aside the problem of triggering the explosion , central detonations in low - mass wds
have been suggested in the past ( e.g. , @xcite ) . to produce little @xmath0 ,
the explosion must occur in wds with central densities @xmath135 g @xmath64 which corresponds to @xmath136 .
in contrast to the dd - models , these models would not show a lower limit for the @xmath0 production and @xmath89 , and they would avoid the problem of mixing during a deflagration phase . nonetheless , we regard these models as unlikely candidates for sn 1999by because there are a couple of severe problems beside the lack of a triggering mechanism for the explosion .
the total mass of the ejecta is lower ( by a factor of 2 ) than is required to keep the line forming region in the layers of explosive carbon and incomplete si - burning up to two and four weeks after the explosion , respectively .
the specific production of nuclear energy does not depend on @xmath137 .
in this case , the diffusion time scales and , thus , the rise times to maximum light are @xmath138 ( e.g. * ? ? ? * ) .
compared to the lcs for a wd with @xmath37 , we expect rise times to be shorter by @xmath139 which is incompatible with sn 1999by observations .
this leaves the merging scenario as an alternative as we may expect a significant asphericity in the explosion if the merging occurs on hydrodynamical time scales .
nevertheless , this alternative is not a very attractive one .
based on 1-d parameterized models , @xcite provided chemical structures for explosions which may resemble merger models . as a general feature ,
these models show a shell - like structure with a massive outer layer of unburned c / o which encloses a thin , low mass layer of explosive carbon burning .
this pattern is not consistent with the ir - spectra for sne 1999by up to maximum light .
however , it should be kept in mind that , besides the limitations of 1-d models for mergers , no chemical structures are available for merging subluminous models . a more severe problem for mergers
may be that merger models tend to ignite prematurely and produce a detonation wave that burns the c / o wd to a o / ne / mg wd prior to the completion of the merger .
thus , merging may result in an accretion induced collapse rather than a sne ia @xcite .
ir - spectra of the subluminous sn1999by have been presented which cover the time evolution from about 4 days before to 2 weeks before maximum light .
this is the first subluminous sn ia ( and arguably the first sn ia ) for which ir spectra with good time coverage are available .
these observations allowed us to determine the chemical structure of the sn envelope .
based on a delayed detonation model , a self - consistent set of hydrodynamic explosions , light curves , and synthetic spectra have been calculated .
this analysis has only two free parameters : the initial structure of the progenitor and the description of the nuclear burning front .
the light curves and spectral evolution follow directly from the explosion model without any further tuning , thus providing a tight link between the model physics and the predicted observables . by varying a single parameter ,
the transition density at which detonation occurs , a set of models has been constructed which spans the observed brightness variation of type ia supernovae .
the absolute maximum brightness depends primarily on the @xmath0 production which , for dd - models , depends mainly on the transition density @xmath39 .
the brightness - decline relation @xmath140 observed in sne ia is also reproduced in these models . in the dd scenario
this relation is a result of the temperature dependence of the opacity , or more precisely , as a consequence of the rapid drop in the opacities for temperatures less than about 7000 to 8000 k. within @xmath37 wd models , a certain amount of burning during the deflagration phase is needed to pre - expand the wd and avoid burning the entire star to @xmath0 .
this implies a lower limit for the @xmath0 mass of about @xmath141 @xmath142 and , consequently , implies a minimum brightness for sne ia within this scenario .
the best model for sn 1999by ( @xmath143 g @xmath64 , selected by matching to the predicted and observed optical light curves ) is close to this minimum ni yield .
the data are consistent with little or no interstellar reddening ( @xmath144 ) , and the derived distance is @xmath3 mpc or @xmath119 mpc if we take the limit for @xmath145 from the models or assume @xmath20 according @xcite for the galaxy . without any further modification , this subluminous model has been used to analyse the ir - spectra from may 6 , may 10 , may 16 and may 24 , 1999 , which correspond to @xmath4 d , 0 d , + 7 d and + 14 d after maximum light . the photosphere ( @xmath146 ) recedes from 14000 , to 10500 , 6500 and 4000 km s@xmath11 .
the observed and theoretical spectra agree reasonably well with respect to the doppler shift of lines and all strong features could be identified . before maximum light ,
the spectra are dominated by products of about explosive carbon burning ( o , mg ) , and si . spectra taken at + 7 d and + 14 d after maximum are dominated by products of incomplete si burning . at about 2 weeks after maximum , the iron - group elements begin to show up
. the long duration of the phases dominated by layers of explosive carbon burning and incomplete si burning implies massive layers of these burning stages that are comparable with our model results .
( @xmath147 and @xmath148 for explosive carbon and incomplete silicon burning , respectively . )
this , together with the @xmath0 mass , argues that sn 1999by was the explosion of a wd at or near the chandrasekhar mass .
finally , we note that the observed ir spectra are at odds with recent 3-d calculations @xcite which predict that large scale chemical inhomogeneities filling 50 70% ( in mass ) of the wd will be formed during the deflagration phase .
when the effect of such inhomogeneous mixing is tested by mixing the inner layers of our sn 1999by model , significant differences appear between the model spectra and the observed data .
this suggests that no significant large - scale mixing took place in sn 1999by .
the lack of observed mixing and the asphericity seen in sn 1999by may be important clues into the nature of subluminous sne ia , and may be related to the reason for a low ddt transition density ( and hence the low luminosity ) .
alternatively , we may have an extended smoldering phase of the wd prior to the explosion which skips the deflagration phase altogether ( see 5 ) .
in either case , the pre - conditioning of the wd prior to the explosion seems to be a key for understanding sne ia .
we would like to thank paul martini and adam steed for aiding us with data acquisition and j. c. wheeler for carefully reading the manuscript and providing helpful suggestions .
p. h. would like to thank nasa for its support by nasa grant nag5 - 7937 . c. l. g. and r. a. f.s research is supported by nsf grant 98 - 76703 .
the calculations for the explosion , light curves and spectra were performed on a beowulf - cluster financed by the john w. cox fund of the department of astronomy at the university of texas .
albrecht , a. , & weller , j. 2000 , , 197 , 6106 aldering , g. , knop , p. , & nugent , p. 2000 , , 119 , 2110 arbour , r. , papenkova , m. , li , w. d. , filippenko , a. v. , & armstrong , m. 1999 , 7156 becker , s. a. , & iben jr . , i. 1979 , , 232 , 831 becker , s. a. , & iben jr . , i. 1980 , , 273 , 111 benz , w. cameron , a. g. w. , press , w. h. & bowers , r. l. 1990 , , 348 , 647 bonanos , a. , garnavich , p. , schlegel , e. , jha , s. , challis , p. , kirshner , r. , hatano , k. , & branch , d. 1999 , 195 , 3806 bowers , e. j. c. , meikle , w. p. s. , geballe , t. r. , walton , n. a. , pinto , p. a. , dhillon , v. s. , howell , s. b. , & harrop - allin , m. k. 1997 , , 290 , 663 brachwitz , f. 2001 , phd thesis , university of basel / switzerland , supervised by f.k .
thielemann branch , d. , romanishin , w. , baron , e. 1996 , 465 , 73 branch , d. 1999 , ara&a 36 , 17 burstein , d. & heiles , c. 1982 , , 87 , 1165 collela , p. , & woodward , p.r .
1984 , j. comp .
phys . , 54 , 174 de vaucouleurs , g. , de vaucouleurs , a. , corwin jr . , h. g. , buta , r. j. paturel , g. , & fouque , p. 1991
, third reference catalogue of bright galaxies , version 3.9 domnguez , i. , & hflich , p. 2000 , , 528 , 854 domnguez , i. , hflich , p. , & straniero , o. 2001 , , in press ( aug .
10 ) erigichi , y. , & mller , e. 1985 , , 147 , 161 filippenko , a. v. et al .
1992 , , 104 , 1543 fisher , a. , branch , d. , hflich , p. , & khokhlov , a. 1995 , , 447 , 73 fisher , a. , branch , d. , hflich , p. , & khokhlov , a. 1998 , , 494 , 47 fisher , a. , branch , d. , hatano , k. , & baron , e. 1999 , , 304 , 679 garcia - senz , d. , & woosley , s. e. 1995 , apj 454 , 895 garnavich , p. , jha , s. , & kirshner , r. 1999 , 7159 garnavich , p. m. , bonanos , a. z. , jha , s. , kirshner , r. p. , schlegel , e. m. , challis , p. , macri , l. m. , hatano , k. , branch , d. , bothun , g. d. , & freedman , w. l. 2001 , , submitted ( astro - ph/0105490 ) gerardy , c. l. , & fesen , r. a. 1999 , 7158 hamuy , m. , phillips , m. m. , maza , j. , suntzeff , n. b. , schommer , r. a. , & avils , r. , 1995 , , 109 , 1 hamuy , m. et al .
1996 , , 112 , 2391 hamuy , m. , phillips , m. m. , schommer , r. a. , suntzeff , n. b. , maza , j. , & avils , r. , 1996 , , 112 , 2398 hanson , m. m. , conti , p. s. , & rieke , m. j. 1996 , , 107 , 281 hanson , m. m. , reike , g. h. , & luhman , k. l. 1998 , , 116 , 1915 hanzl , d. , 1999 , iau - circulars 7156 , 7158 , 7169 , & 7223 hernandez , m. et al .
2000 , , 319 , 223 hatano , k. , branch , d. , lentz , e. j. , baron , e. , filippenko , a. v. , & garnavich , p. 2000 , apj 543l , 49 hoffleit , d , & jaschek , c. 1982 , the bright star catalog , 4th revised edition ( new haven : yale university observatory ) hflich , p. , mueller , e. , khokhlov , a. 1993 , a&a 268 , 570 hflich , p. , khokhlov , a. , & wheeler , j.c .
1993 , a&a , 270 , 223 hflich , p. 1995
, , 443 , 89 [ h95 ] hflich , p. , khokhlov , a. , & wheeler , j.c .
1995 , , 444 , 211 ( hkw95 ) hflich , p. , & khokhlov , a. 1996 , , 457 , 500 ( hk96 ) hflich , p. , khokhlov , a. , wheeler , j. c. , phillips , m. m. , suntzeff , n. b. , & hamuy , m. 1996a , , 472 , l81 hflich , p. ; dominik , c. ; khokhlov , a. ; mller , e. ; & wheeler , j. c. 1996b , @xmath149 texas symposium on relativistic astrophysics , annals of the new york academy of science , 759 , 348 hflich , p. 1997 , in : supernovae and cosmology , eds , labhardt , binggeli & buser , proc . of a colloquium in honor of g.a .
tammann , publ . of the astronomical insitute , p.25 hflich , p. , wheeler , j. c. , & thielemann , f. k. 1998 , , 495 , 617 ( hwt98 ) hflich , p. , nomoto , k. , umeda , h. , & wheeler , j. c. 2000 , , 528 , 590 hflich , p. , & stein , j. 2001 , , submitted ( astro - ph/0104226 ) howell , a. , hflich , p. , wang , l. , & wheeler , j. c. 2001 , , 556 , 302 hoyle , p. , & fowler , w.a .
1960 , , 132 , 565 iben jr , i. 1972 , , 178 , 433 iben jr , i. & tutukov , a.v . 1984 , , 54 , 335 iwamoto , k. , brachwitz , f. , nomoto , k. , kishimoto , n. , umeda , h. , hix , w. r. , & thielemann , f. k. 1999 , , 125 , 439 khokhlov , a. 1991 , , 245 , 114 khokhlov , a. , mller , e. , & hflich , p. 1993
, , 270 , 223 khokhlov , a. 1995 , , 449 , 695 khokhlov , a. 2001 , , in press ( astro - ph/0008463 ) kurucz , r. l. 1993 , atomic data for opacity calculations , cambridge / center for astrophysics , cd kurucz , r. l. 1994 , solar abundance model atmospheres for 0,1,2,4,8 km / s , kurucz cd - rom no .
19 langer , n. , deutschmann , a. , wellstein , s. , & hflich , p. 2000 , , 362 , 1046 lentz , e. j. , baron , e. , branch , d. , hauschildt , p. , & nugent , p. e. 2000 , , 530 , l966 lentz , e. j. , baron , e. , branch , d. , & hausschildt , p. 2001
, , 557 , 2001 leibundgut , b. et al .
1993 , , 105 , 301 li , w .- d . ,
modjaz , m. , king , j. y. , papenkova , m. , johnson , r. a. , friedman , a. , treffers , r. r. , & filippenko , a. v. 1999 , 7126 li , w. -d . ,
filippenko , a. v. , treffers , r. r. , riess , a. g. , hu , j. , & qiu , y. 2000 , astro - ph/0006292 lisewski , a. m. ; hillebrandt , w. ; woosley , s. e. ; niemeyer , j. c. ; kerstein , a. r. 2000 , 537 , 405l livingston , w. , & wallace , l. 1991 july , nso tech .
rept . # 91001 livne , e. 1999 , , 527 , 97 mazzali , p. a. , et al .
1997 , , 284 , 151 meikle , w. p. s. et al .
1996 , , 281 , 263 mendoza , c. , & cunto , w. 1993 , , 275 , l5 mazzali , p. a. , nomoto , k. , cappelaro , e. , nakamura , t. , umeda , h. , & iwamoto , k. 2001 , , in press & astro - ph/0009490 mihalas , d. 1978 , stellar atmospheres , san francisco , freeman milne , p. , the , l .- s . , & leising , m.d . 1999 , 124 , 503 modjaz , et al .
2000 , , submitted , astro - ph/0008012 niemeyer , j. c. , & hillebrandt , w. 1995 , apj , 452 , 779 nomoto , k. , sugimoto , s. , & neo , s. 1976 , apss , 39 , l37 nomoto , k. 1980 , apj , 248 , 798 nomoto , k. 1982 , apj 253 , 798 nomoto , k. , thielemann , f. -k .
, & yokoi , k. 1984 , , 286 , 644 nugent , p. , phillips , m. , baron , e. , branch . , d. , & hauschildt , p. 1995
, apj , 455 , l147 nugent , p. , baron , e. , hauschildt , p. , & branch . , d. 1997 , , 485 , 812 paczyski , b. , 1985 , in : cataclysmic variables and low - mass x - ray binaries , eds .
d.q . lamb , j. patterson , reidel , dordrecht , p.1 papenkova , m. , li , w. d. , & filippenko , a. v. 1999 , 7156 persson , s. e. , murphy , d. c. , krzeminski , w. , roth , m. , & rieke , m. j. 1998 , , 116 , 2475 pinto , p. , eastman , r. g. 2000 , , 530 , 744 phillips , m. m. et al .
1987 , 90 , 592 phillips , m. m. , lira , p. , suntzeff , n. b. , schommer , r. a. , hamuy , m. , & maza , j. 1999 , , 118 , 1766 rudy r.j . ,
lynch d.k . , mazuk s. , venturini c.c . , puetter r.c
, hflich p. 2001 , apj , submitted ruiz - lapuente , p. , jeffery , d. , challis , p. m. , filippenko , a. v. , kirshner , r. p. , ho , l. h. , schmidt , b. p. , sanchez , f. , & canal , r. 1993 , , 365 , 728 saio , h. , & nomoto , k. 1998 , , 500 , 388 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 suntzeff , n. b. , et al . 1999 , , 117 , 1175 toth , i. , & szabo , r. 2000 , , 361 , 63 treffers , r. r. , peng , c. y. , filippenko , a. v. , & richmond , m. w. 1997 , 6627 turatto , m. , piemonte , a. , benetti , s. , cappellaro , e. , mazzali , p. a. , danziger , i. j. , & patat , f. 1998 , , 116 , 2341 thielemann , f. k. , nomoto , k. , & hashimoto , m. 1996 , , 460 , 408 umeda , h. , nomoto , k. , kobayashi , c. , hachisu , i. , & kato , m. 1999 , , 522 , l43 wallace , l. , hinkle , k. , & livingston , w. , 1993 april , nso tech . rept .
# 93001 wang , l. , wheeler , j. c. , & hflich , p. 1997
, , 476 , l27 wang , l. , howell , a. , hflich , p. , & wheeler j.c .
2001 , 550 , 1030 webbink , r. f. 1984 , , 277 , 355 wheeler , j. c. , hflich , p. , harkness , r. p. , & spyromilio , j. 1998 , , 496 , 908 whelan , j. , & iben jr . , i. 1973 , , 186 , 1007 woosley , s. e. , weaver , t. a. , & taam , r. e. 1980 , in : type i supernovae , ed . j. c. wheeler , austin , u. texas , p.96 woosley , s. e. , & weaver , t. a. 1986 , , 24 , 205 woosley , s. e. , & weaver , t. a. 1994 , , 423 , 371 yamaoka , h. , nomoto , k. , shigeyama , t. , & thielemann , f. 1992 , , 393 , 55 ccccc 6 may & -4 d & 0.961.8 & 4@xmath29600 s & gerardy + & & 1.952.4 & 3@xmath29600 s & + 7 may & -3 d & 0.961.8 & 3@xmath29600 s & + & & 1.952.4 & 3@xmath29600 s & + 8 may & -2 d & 0.961.8 & 3@xmath29600 s & + 9 may & -1 d & 0.961.8 & 3@xmath29600 s & + & & 1.22.2 & 3@xmath29600 s &
+ 10 may & 0 d & 0.961.8 & 3@xmath29600 s & + & & 1.22.2 & 3@xmath29600 s & + & & 1.952.4 & 3@xmath29600 s &
+ 16 may & 6 d & 0.961.8 & 3@xmath29600 s & martini & steed + 18 may & 8 d & 0.961.8 & 3@xmath29600 s &
+ 20 may & 10 d & 0.961.8 & 3@xmath29600 s &
+ 21 may & 11 d & 0.961.8 & 3@xmath29600 s & sakai + 24 may & 14 d & 0.961.8 & 3@xmath29600 s & + date & epoch & j & h & k + + 6 may & -4 d & @xmath150 & @xmath151 & @xmath152 + 7 may & -3 d & @xmath150 & @xmath153 & @xmath154 + 8 may & -2 d & @xmath155 & @xmath156 & lccccccc 5p0z22.8 & 8 . & 0.095 & -17.21 & 13.9 & 0.47 & 1.45 + 5p0z22.10 & 10 .
& 0.107 & -17.35 & 14.1 & 0.42 & 1.37 + 5p0z22.12 & 12 . &
0.137 & -17.63 & 14.7 & 0.38 & 1.32 + 5p0z22.14 & 14 . & 0.153 & -17.72 & 14.9 & 0.22 & 1.30 + 5p0z22.16 & 16 . & 0.268 & -18.72 & 15.8 & 0.14 & 1.26 + 5p0z22.18 & 18 . & 0.365 & -18.82 & 16.6 & 0.08 & 1.21 + 5p0z22.20 & 20 . & 0.454 & -18.96 & 17.0 & 0.02 & 1.19 + 5p0z22.23 & 23 . & 0.561 & -19.21 & 18.2 & -0.02 & 1.05 + 5p0z22.25 & 25 . &
0.602 & -19.29 & 18.6 & -0.02 & 1.00 + 5p0z22.27 & 27 .
& 0.629 & -19.35 & 19.0 & -0.03 & 0.91 + as a function of the @xmath0 mass ( upper left ) , @xmath39 ( lower left ) , and @xmath16 ( lower right ) , and @xmath162 as a function of @xmath89 ( upper right ) . for all panels except the lower right , the points correspond to models with @xmath39 of 8 , 10 , 12 , 14 , 16 , 18 , 20 , 23 , 25 and 27 @xmath161 from left to right . in the lower right panel , this order is reversed.,width=309 ] days before maximum light ( 10 days after the explosion ) . between 1.35 and 1.45 , and longward of 2.45 , the s / n in the observed spectrum
is very low due to strong telluric absorption.,title="fig:",width=385 ] days before maximum light ( 10 days after the explosion ) . between 1.35 and 1.45 , and
longward of 2.45 , the s / n in the observed spectrum is very low due to strong telluric absorption.,title="fig:",width=385 ] | near - infrared ( nir ) spectra of the subluminous type ia supernova sn 1999by are presented which cover the time evolution from about 4 days before to 2 weeks after maximum light .
analysis of these data was accomplished through the construction of an extended set of delayed detonation ( dd ) models covering the entire range of normal to subluminous sne ia . the explosion , light curves ( lc ) , and the time evolution of the synthetic spectra were calculated self - consistently for each model with the only free parameters being the initial structure of the white dwarf ( wd ) and the description of the nuclear burning front during the explosion . from these , one model was selected for sn 1999by by matching the synthetic and observed optical light curves , principly the rapid brightness decline .
dd models require a minimum amount of burning during the deflagration phase which implies a lower limit for the @xmath0 mass of about @xmath1 and consequently a lower limit for the sn brightness .
the models which best match the optical light curve of sn 1999by were those with a @xmath0 production close to this theoretical minimum .
the data are consistent with little or no interstellar reddening ( @xmath2 ) and we derive a distance or @xmath3 mpc for sn 1999by in agreement with other estimates . without any modification ,
the synthetic spectra from this subluminous model match reasonably well the observed ir spectra taken on may 6 , may 10 , may 16 and may 24 , 1999 . these dates correspond roughly to @xmath4 d , 0 d , and 6 d and 14 d after maximum light . prior to maximum ,
the nir spectra of sn 1999by are dominated by products of explosive carbon burning ( o , mg ) , and si .
spectra taken after maximum light are dominated by products of incomplete si burning . unlike the behavior of normal type ia sne
, lines from iron - group elements only begin to show up in our last spectrum taken about two weeks after maximum light .
the implied distribution of elements in velocity space agrees well with the dd model predictions for a subluminous sn ia .
regardless of the explosion model , the long duration of the phases dominated by layers of explosive carbon and oxygen burning argues that sn 1999by was the explosion of a white dwarf at or near the chandrasekhar mass . the good agreement between the observations and the models without fine - tuning a large number of free parameters suggests that dd models are a good description of at least subluminous type ia sne .
pure deflagration scenarios or mergers are unlikely and helium - triggered explosions can be ruled out .
however , problems for dd models still remain , as the data seem to be at odds with recent 3-d models of the deflagration phase which predict significant mixing of the inner layers of the white dwarf prior to detonation .
possible solutions include the effects of rapid rotation on the propagation of nuclear flames during the explosive phase of burning , or extensive burning of carbon just prior to the runaway . |
the research field of nanoplasmonics has grown substantially in past few years due to tremendous progress in fabrication techniques and optical characterization @xcite .
while the linear plasmonics is still enjoying significant interest , the field of nonlinear plasmonics @xcite recently began to attract attention @xcite . moreover the new type of nanomaterials using strong coupling of molecular excitons and plasmons has also been extensively discussed@xcite .
an obvious extension of various nonlinear spectroscopy techniques@xcite to the domain of plasmonics is also an ongoing effort @xcite .
free induction decay is a well - documented phenomenon that was first demonstrated in nuclear magnetic resonance ( nmr ) and is also observed in optics @xcite .
a population of two - level emitters can exist in which all emitters have the same _ central _ transition energy but each is detuned by some amount from this central energy due to inhomogeneous broadening , which can result from conditions such as doppler shift in individual gas molecules or variations in electric field from point to point in a solid @xcite .
the macroscopic polarization of an ensemble of emitters is the sum of each individual contribution from every emitter in the ensemble . given that the transition energies of the emitters in the ensemble are described by a distribution
, each individual emitter will oscillate at a frequency that is slightly different from the others after the system is pumped by a strong incident pulse . as a result ,
all of the emitters will oscillate in phase at first , but they will all ( for a realistic number of emitters ) dephase within a characteristic inhomogeneous lifetime and , if left alone , never rephase again @xcite .
however , for times less than the natural lifetime of the emitter , each emitter is still oscillating .
one can in principle invert the dephasing process by applying a second pump in the form of @xmath0 pulse .
the oscillations then all _ run in reverse _ , resulting in a subsequent rephasing . the ensemble polarizes once again , and a photon echo signal is observed .
this technique is widely used in chemistry and is referred to as photon echo spectroscopy @xcite .
inhomogeneous effects due to variations in an emitter s surroundings cause each emitter to oscillate at a slightly different frequency than the others , and photon echo spectroscopy _ removes _ this effect .
any remaining dephasing is irreversible by the echo technique , and is revealed as diminished intensity of the echo @xcite .
for example , as the delay in applying the @xmath0 pulse increases , the natural lifetime of the emitters causes all of their oscillations to decrease , resulting in an echo with lower intensity .
thus a time and frequency structure of a detected photon echo contains all information about a probed system . additionally , the recovery of a signal after dephasing offers prospects for memory storage . in @xcite ,
the optical properties are copied to a spin system whose lifetime is much longer than that of the optical system thus extending the duration of the system s memory .
our major interest in this paper is to test a concept of photon echo in strongly coupled exciton - plasmon systems in order to understand any new features unique to the strong coupling regime .
first , we briefly overview the numerical model used .
next , we consider the application of a @xmath1 pulse to a 1-d ensemble followed by application of a @xmath0 pulse generating a photon echo signal .
the density of molecules in this ensemble is varied and the strength of the echo is considered in terms of the transmission and reflection of the ensemble .
finally , we investigate the photon echo of two exciton - plasmon systems , namely a periodic array of slits and a core - shell nanoparticle , each combined with inhomogeneously broadened molecules .
the time dynamics of molecules is described by the liouville equation @xmath2-i\hbar\hat{\gamma}\hat{\rho } , \label{bloch}\ ] ] where @xmath3 is the single - molecule density matrix , @xmath4 describes relaxation processes , and @xmath5 is the hamiltonian of a single molecule with a dipole moment operator @xmath6 interacting with a local electric field @xmath7 the dynamics of the electric field , @xmath8 , is governed by corresponding maxwell s equations as discussed below .
the field - free hamiltonian for a molecule with energy index k is @xmath9 @xmath10 corresponds to the index of a specific energy in the distribution and @xmath11 is the transition energy of a given molecule .
we consider an ensemble of two - level molecules having a particular distribution of transition energies ( inhomogeneous broadening ( ib ) ) .
a normalized gaussian distribution @xmath12 is used to describe the effect of ib in the ensemble in different geometries schematically depicted in fig .
[ figure1 ] .
specifically @xmath13 where @xmath14 is the normalization constant , @xmath15 is the energy corresponding to the maximum in the gaussian distribution , and @xmath16 characterizes the full width at half - maximum ( fwhm ) @xmath17 of the distribution , which describes how broad a given ib is .
specific values of @xmath17 in this work are `` narrow '' ( @xmath18 ev ) , `` intermediate '' ( @xmath19 ev ) , and `` broad '' ( @xmath20 ev ) .
the distribution is normalized such that @xmath21 schematic setups of systems considered .
a ) an extended 1-d ensemble of interacting molecules with a finite thickness along @xmath22 and infinite in @xmath23 and @xmath24 the incident field propagates in @xmath22 direction and is polarized along @xmath23 .
b ) the periodic array of slits coupled to a thin film of molecules placed on the input side . here
the incident field is propagated from top to bottom and is polarized in horizontal direction .
the signal is detected on both input and output sides .
c ) the core - shell nanoparticle consisting of a metallic core surrounded by a thin shell of interacting molecules.,scaledwidth=90.0% ] the updated expectation value of the dipole moment , @xmath25 , is used to calculate the macroscopic polarization of the molecular ensemble with a number density of @xmath26 , @xmath27 , and then the polarization current , @xmath28 , which is subsequently inserted into maxwell s equations the following set of parameters describing a molecule is used in this paper : the transition dipole moment is @xmath29 debye , the radiationless lifetime of the excited state is @xmath30 ps , and the pure dephasing time is 800 fs .
to model electromagnetic ( em ) interaction between molecules and metal in time and space in addition to the bloch equations we propagate maxwell s equations using the finite - difference time - domain method ( fdtd ) @xcite @xmath31 where @xmath32 is the magnetic field , @xmath33 is the current density , @xmath34 and @xmath35 are the permeability and the permittivity of free space , respectively , @xmath36 is the relative permittivity at a given point ( 1 for vacuum , see below ) . to account for the material dispersion of metal we emply the conventional drude model @xmath37 where @xmath38 is the plasma frequency
, @xmath39 is the phenomenological damping , and @xmath36 is the high - frequency limit of the dielectric function . for silver ,
we use the following parameters @xcite : @xmath40 ev , @xmath41 ev , and @xmath42 . in metal , the dynamics of the current density @xmath33 satisfy the following equation@xcite @xmath43 coupled equations ( [ bloch ] ) , ( [ faraday ] ) , ( [ amperemaxwell ] ) , and ( [ j ] ) are solved numerically using home - built codes .
the numerical convergence is achieved for the grid with @xmath44 nm and a time step of @xmath45 .
it should be noted that inevitable discretization of the transition energy distribution ( [ gaussian ] ) results in unphysical rephasing which must be carefully monitored in all simulations in order to avoid spurious numerical artifacts . in brief , after initial excitation of an ensemble of non - interacting molecules , free induction decay occurs followed by a rephasing after a delay which is proportional to the @xmath46 , where @xmath47 is the discretization step in ( [ gaussian ] ) .
the rephasing is obviously unphysical for a realistic number of molecules ( i.e. continuous transition frequency distribution with the revival time being infinite @xcite ) . in all our simulations we use an energy step @xmath48 no larger than @xmath49 ev , corresponding to a revival time of @xmath50 fs or longer
first we consider non - interacting molecules driven by external incident field , the polarization current @xmath51 is used as an indication of rephasing .. we employ the two - pulse pump - probe photon echo sequence as illustrated in fig .
[ figure2 ] .
@xmath52 corresponds to the time between the maximum of the @xmath1 pulse and the @xmath0 pulse , and @xmath53 corresponds to the time between the maximum of the @xmath0 pulse and the maximum of the echo .
these times are roughly equal as they should be @xcite .
the duration and amplitude of the @xmath1 pulse are such that it drives the density matrix element @xmath54 of a molecule ( whose transition energy corresponds to @xmath15 ) to @xmath55 .
the @xmath0 pulse completely inverts the molecule .
molecules are excited by a @xmath1 pulse polarizing the sample , which then undergoes dephasing resulting in observed free induction decay .
a @xmath0 pulse is then applied , and the maximum of the photon echo signal is observed at time @xmath56 . using these simulations as a test case we verified our numerical procedure and proceed to the case of interacting molecules .
a schematics of the two pulse @xmath1-@xmath0-echo sequence .
the photon echo signal occurs after application of a pump @xmath1 pulse followed by a probe @xmath0 pulse applied to an ensemble of molecules with inhomogeneous broadening.,scaledwidth=60.0% ] the ensemble treated thus far has been a collection of noninteracting molecules . in order to determine the influence of collective effects on the echo amplitude and its form , we employ maxwell - bloch formalism . by solving self - consistently both maxwell s and bloch equations we allow molecules to interact via emission and absorption of em radiation . in the linear regime ( i.e. when the population of molecular excited states is significantly smaller than @xmath30 throughout the entire simulation time ) , the transmission and reflection broaden and flatten as the inhomogeneous broadening is increased .
more importantly , the transmission decreases toward @xmath57 and the reflection increases toward @xmath30 as the density increases beyond @xmath58 m@xmath59 @xcite .
distortion is observed on the edges of the spectral features due to increasing interference between the reflected signals from varying depths of the ensemble . at high molecular concentrations
the reflection becomes quite broad with a plateau near @xmath30 centered at the @xmath15 .
such a behavior is due to the fact that at high densities molecule - molecule interactions result in a redshift of the absorption leading to effective negative electric susceptibility near @xmath15 .
this means that molecules oscillate completely out - of - phase with external field effectively canceling it out , which in turn leads to the high reflection @xcite .
em energy of the photon echo signal ( in arbitrary units ) on the input and output sides of a 1-d ensemble of interacting molecules as a function of the molecular density .
results for the molecular layer of the thickness of @xmath60 nm are shown as circles , triangles correspond to the thickness of @xmath50 nm , and squares correspond to the thickness of @xmath61 nm
. the signal increases with density at lower densities and in several cases drops sharply at higher densities , which is attributed to highly increased reflection .
different panels correspond to different detection sides ( output or input ) and different fwhm of the inhomogeneous broadening distribution ( [ gaussian ] ) .
energy detected on the input side is shown in panels ( a ) ( @xmath62 ev ) , ( b ) ( @xmath63 ev ) , and ( c ) ( @xmath64 ev ) .
the detection on the output side is shown in panels ( d ) ( @xmath62 ev ) , ( e ) ( @xmath63 ev ) , and ( f ) ( @xmath64 ev).,scaledwidth=110.0% ] we apply the @xmath65 pulse sequence ( again tuned to the central frequency of a single molecule at @xmath66 ev , then record the energy of the echo signal ( calculated as a time integral of the electric field intensity ) on both the input ( reflection ) and output ( transmission ) sides .
simulations are performed for several densities , molecular layer thicknesses , and values of inhomogeneous broadening .
the results are shown in fig .
[ figure3 ] . on both sides of the ensemble ,
the echo amplitude increases with the density for lower densities but , in many instances , decreases sharply at higher densities .
this can be understood in terms of how the transmission and reflection of the ensemble vary with density . at higher densities ,
the reflection approaches @xmath30 and the transmission becomes very small as previously discussed .
thus , the driving fields do not make it as far into the ensemble for higher densities making photon echo signal lower as not all molecules contribute to the generation of echo .
another interesting observation is that there is obviously an optimal molecular density , at which the energy of photon echo detected on the output side has a global maximum .
this is explained below . a snapshot of the molecular population as a function of the position is shown in fig . [ figure4 ] .
the pulse travels from left to right , and the ground state population along the ensemble is calculated immediately after the @xmath1 pulse passes through the ensemble .
it is seen that , for a higher density , the ground state population @xmath67 tends toward @xmath30 as we look farther into the ensemble ; the driving fields are smaller and therefore do not drive the molecules through @xmath1 oscillations . with this in mind , it makes sense that the echo on the output side drops off at higher densities : more and more of the ensemble ( toward the output side ) ceases to participate in the echo at higher densities , and the echo from the input side is not able to travel through the ensemble to the output side .
thus , the echo increases with increasing density until reflection increases significantly , disrupting the echo process throughout the ensemble .
the input side is somewhat more complicated , as there is obvious interference between incoming and outgoing reflected waves , and the reflected waves from lesser depths interfere with reflected waves from deeper regions of the ensemble .
furthermore , the ensemble is driven by high fields on resonance , leading to higher dispersion and as a result the low group velocity .
the latter results in high spatially modulated populations which obviously have a significant effect on the generation of photon echoes @xcite .
ground state population along the 1-d ensemble of molecules calculated for @xmath62 ev at the end of the @xmath1 pulse .
the ensemble is @xmath50 nm in length .
density is : ( a ) @xmath68 m@xmath59 , ( b ) @xmath58 m@xmath59,scaledwidth=80.0% ] we now turn our attention to a exciton - plasmon system comprised of a periodic array of slits and molecules with inhomogeneous broadening as schematically depicted in fig .
[ figure1]b .
the bare metallic grating exhibits a set of surface plasmon resonances in the form of enhanced transmission and decreased reflection . for the parameters used we consider the plasmon resonance near @xmath69 ev as our working mode .
we note that the em field distribution corresponding to this mode is highly localized near slits on both input and output sides of the array .
adding resonant molecules ( whose transition energy is @xmath69 ev ) with no inhomogeneous broadening and probing in the linear regime results in rabi splitting of the corresponding plasmon mode into an upper and lower polariton .
however when the molecules exhibit inhomogeneous broadening , the rabi splitting is still present but it is significantly broader as one may expect as not all molecules are as strongly coupled compared to the previous case .
nonetheless anticipated broader spectrum one can actually observe significantly higher rabi splitting .
[ figure5 ] shows transmission spectra for bare slits , the hybrid system with no inhomogeneous broadening , and for the hybrid system whose molecules are inhomogeneously broadened . as seen from fig .
[ figure5 ] , the rabi splitting for the hybrid system with no inhomogeneous broadening reaches @xmath70 mev whereas that for a hybrid system with inhomogeneous broadening is @xmath71 mev .
this is a surprising result .
one would have expected to observe the opposite behavior of the rabi splitting . however
the increase of the splitting in systems with distributed energies is not uncommon .
simulations of optics of exciton - plasmon systems comprised diatomic molecules ( described using the full machinery of a ro - vibrational structure ) and metal reveal similar tendency .
the increase of the exciton - plasmon coupling is observed for the molecules described by two potential energy surfaces with the excited one being a dissociative state , i.e. having a continuum of ro - vibrational states @xcite .
this in part can be explained by the fact that plasmon resonances are usually quite broad and nearly all molecular transition energies @xmath72 contribute coherently to the coupling . linear transmission through periodic array of slits .
slit period is 320 nm , film thickness is @xmath60 nm , slit width is @xmath73 nm , the molecular layer has a thickness of @xmath74 nm , molecular density is @xmath58 m@xmath59 .
solid line indicates bare metal slits , the dashed line is for a hybrid system with no inhomogeneous broadening , and the dotted line corresponds to a hybrid system with inhomogeneous broadening corresponding to @xmath20 ev.,scaledwidth=100.0% ] we apply the @xmath75 - pulse sequence to the exciton - plasmon system and record the transverse component of the electric field on both the input and output side of the system for the extended period of time when the echo is observed .
what is seen is very different from the echo generated by the 1-d ensemble : for this system , an echo with a clear double - peaked structure is observed .
the fourier analysis of the echo signal reveals two peaks as shown in fig .
[ figure6 ] .
we proceed to show that this double - peaked structure has its origin in the interaction between the surface plasmons and the molecules .
echo signal observed for the exciton - plasmon nanosystem comprising periodic array of slits and molecules .
double - peaked photon echo is recorded by calculating the transverse component of electric field , @xmath76 , as a function of time as shown in the upper row .
the corresponding spectra of echo signals are shown in lower row .
the molecular density is @xmath58 m@xmath59 , the thickness of the molecular film is @xmath74 nm , fwhm of the molecular transition energy distribution is 2.06 ev .
a nondispersive spacer of varying thickness is inserted between the film and the slits , and the double - peaked structure gradually reduces to a single peak as the spacer thickness is increased .
the spacer lengths ( from left to right ) are @xmath57 nm , @xmath77 nm , @xmath78 nm , and @xmath79 nm.,scaledwidth=100.0% ] surface plasmons are intense and localized to within nanometers of the metallic structure . in order to ascertain whether our observations of the double - peak echo is influenced by surface plasmons ,
we insert a ( vacuum ) spacer layer of variable thickness between the molecular film and the slits . as the spacer thickness
is increased , the film is moved into progressively weaker plasmon fields leading to a noticeably lower molecule - plasmon coupling . as shown in fig .
[ figure6 ] , the double - peaked structure of the echo gradually disappears as the spacer thickness is increased .
the amount of coupling between plasmons and molecules depends , among other things , on the strength of the plasmon fields .
we therefore expect that if the double - peaked structure of the echo is caused by strong coupling between plasmons and molecules , a spacer layer of increasing size will reduce the echo to a single peak , which is exactly what we observe .
it is however surprising to see how sensitive the double - peak structure is to molecule - plasmon coupling . to further confirm that the double peak structure is the signature of the upper and lower polaritons that are formed as a result of strong coupling between surface plasmons and molecules we tweak the coupling by introducing a spacer between molecules and the array of slits .
an fast fourier transform ( fft ) of each echo signal in fig .
[ figure6 ] indicates two clearly - defined peaks for smaller spacer thickness , and these two peaks reduce to one single peak as the spacer thickness increases .
even though plasmon fields are still quite significant at distances of @xmath78 nm above the metal the spectrum of echo is already reduced to a single resonance at @xmath15 .
these two peaks likely correspond to the upper and lower polariton energies .
however , a precise correspondence between the upper and lower polariton energies and the two peaks in the echo spectra was difficult to obtain for several purely technical reasons . first , the energies of the peaks in the fft of the echo signal were very sensitive to the sampling range , which in turn is affected by the residuals of @xmath0-pulse in the signal .
as the range was changed , the double - peak structure remained roughly the same but the energies of the peaks shifted .
secondly , we sought a relation between the upper and lower polariton amplitudes in the linear spectrum versus the height of the fft peaks , but a larger peak in the linear spectrum does not necessarily correspond to a stronger peak in the fft spectrum as the echo depends on rephasing , and the upper and lower polariton energies may be asymmetrically detuned from the central energy , leading to a weaker rephasing of one component or the other .
finally , spatial inhomogeneities of the fields also complicate the matter in that some molecules may be driven through more or less than the `` correct '' @xmath1 - @xmath0 sequence . if , however , the double - peak echoes are unique to exciton - plasmon systems we should be able to observe similar signals in other systems such as a core - shell nanoparticle , for instance , as depicted in fig . [ figure1]c .
photon echo observed in core - shell nanoparticles .
the core is silver with a radius of @xmath78 nm and the shell with a thickness of @xmath74 nm consists of interacting molecules with inhomogeneous broadening ( @xmath17 = 0.236 ev ) . in order to illustrate the plasmonic origin of the double - peaked echo structure ,
the central energy of the two - level molecular distribution is detuned from the plasmon resonance of @xmath80 ev .
the central energy @xmath15 is a ) @xmath81 ev , b ) @xmath82 ev , c ) @xmath83 ev , and d ) @xmath84 ev .
, scaledwidth=100.0% ] proceeding with the idea that the double peak is caused by surface plasmons , we consider a different hybrid system consisting of a metallic cylinder surrounded by a shell of two - level molecules with inhomogeneous broadening .
the bare silver core shows a plasmon resonance at @xmath80 ev ; the transmission is enhanced and the reflection is reduced . to minimize the interaction between the molecules and the localized surface plasmon
, the central energy of the distribution of molecules is detuned by varying @xmath15 . as shown in fig .
[ figure7 ] , this results in the gradual loss of the double - peak structure in the echo .
detuning was also applied to the hybrid slits system , both by adjusting the molecules central energy as well as adjusting the slit period ( and therefore altering the surface plasmon energy ) . as expected , the double - peak structure disappeared in each instance as detuning was increased .
we performed rigorous numerical studies of photon echo using maxwell - bloch formalism .
it is shown that photon echoes are highly dependent on materials parameters such as a thickness of molecular layer and molecular density .
we have demonstrated that our numerical methods reproduce photon echoes and we have characterized the echo from a one - dimensional ensemble in terms of density , ensemble length , and amount of inhomogeneous broadening .
it is demonstrated that in one - dimensional systems the echo detected on the output side has an optimal molecular concentration at which the transmitted signal has a maximum energy . to the best of our knowledge
it is shown for the first time that applying two - pulse photon echo sequence to exciton - plasmon systems leads to double - peaked echoes .
this is demonstrated by various different means to be of plasmonic origin , with the double - peaked structure resulting from the hybrid modes of the system .
it is also shown that the unique signature of upper and lower polaritons in echo is highly sensitive to exciton - plasmon coupling making it a great tool for ultrafast probes of nanomaterials .
this work is supported by afosr grant no .
fa9550 - 15 - 1 - 0189 .
m.s . is also grateful to the financial support provided by bsf grant no .
2014113 .
18ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty \doibase http://dx.doi.org/10.1016/j.physrep.2004.11.001 [ * * , ( ) ] link:\doibase 10.1364/oe.19.022029 [ * * , ( ) ] http://dx.doi.org/10.1038/nphoton.2012.244 [ * * , ( ) ] link:\doibase 10.1021/acsnano.5b04373 [ * * , ( ) ] http://stacks.iop.org/0034-4885/78/i=1/a=013901 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1021/acsphotonics.5b00587 [ * * , ( ) ] @noop _ _ ( , ) @noop _ _ ( , ) link:\doibase 10.1146/annurev.physchem.49.1.99 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.462686 [ * * , ( ) ] http://dx.doi.org/10.1038/nphoton.2014.219 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physrevb.68.045415 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.163603 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.043835 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.035433 [ * * , ( ) ] @noop `` , '' ( ) | the results of rigorous numerical simulations of photon echoes in exciton - plasmon systems are presented . using a self - consistent model based on coupled maxwell - bloch equations we investigate femtosecond time dynamics of ensembles of interacting molecules and molecular aggregates optically coupled to surface - plasmon supporting materials .
it is shown that observed photon echoes under two pulse pump - probe sequence are highly dependent on various material parameters such as molecular concentration and periodicity .
simulations of photon echoes in exciton - plasmon materials reveal a unique signature of the strong exciton - plasmon coupling , namely a double - peak structure in spectra of recorded echo signals .
this phenomenon is shown to be related to hybrid states ( upper and lower polaritons ) in exciton - plasmon systems under strong coupling conditions .
it is also demonstrated that the double - peak echo is highly sensitive to mild deviations of the coupling from the resonance between molecules and plasmons making it a great tool for ultrafast probes . |
mass distribution at high redshifts is being a hot subject of the large scale structure study . data of various objects at moderate and high redshifts , in particular , clusters of galaxies , are becoming available for probing the formation and evolution of structures and for discriminating among popular dark matter models ( e.g. jing & fang 1994 ; eke , cole & frenk 1996 ; bahcall , fan & cen 1997 ; kitayama & suto 1997 ) . considering that quasars are the most distant among various luminous objects ,
they have also been applied in this approach ( e.g. bi & fang 1997 ) . however , as a mass tracer of the cosmic matter field , quasars are still playing a role different from clusters of galaxies .
the problem stems from so - called `` bias '' .
clusters are a biased tracer of the mass distribution .
the correlation amplitude of clusters is believed to be much higher than that of the underlying matter and increases strongly with the cluster richness ( bahcall & soneira 1983 ) .
this bias is plausibly explained by the mechanism that the observed clusters are identified as massive collapsed halos of the density field ( kaiser 1984 ) .
that is , the bias of clusters is completely determined by the gravitational parameters , like mass @xmath4 and virial radius @xmath5 used to identify the halos . with this approach
, a detailed confrontation can be made between theories and the observations of clusters .
quasars may also be biased tracer of the mass distribution .
recent observations indicate that the correlation amplitude of quasars may also be different from the underlying dark matter ( mo & fang 1993 ; komberg , kravtsov & lukash 1994 ; croom & shanks 1996 ; franca , andreani & cristiani , 1997 ) . however , so far no detailed model is available for the bias of quasars , though their high clustering strength and environment imply that quasars are hosted by massive halos ( see below ) .
because of the lack of such a model , one can not confront the data of quasars with theoretical models in the way as for clusters .
for instance , the abundance of quasars can only be used as an upper or lower limit to certain massive halos ; no detailed comparison between the number densities of quasars and of halos is allowed .
obviously , it is very important to understand what kind of mass halos are associated with the majority of quasars
. such a knowledge will not only enable the observational data of quasars to be powerful tests for theoretical models of galaxy formation but also tell that what type of local environments is responsible for intriguing the nuclear activities of quasars .
like clusters and groups of galaxies , it is generally believed that quasars should be associated with certain type of collapsed dark matter halos . yet , different from identification of clusters , the environment suitable for forming quasars is not merely determined by gravitational parameters , as the hydro processes are also involved .
therefore , the identification of quasar - harboring halos should be given by both gravitational and hydro parameters .
in other words , not all halos with certain @xmath4 and/or velocity dispersion @xmath3 harbor quasars , because certain hydro conditions must be satisfied .
however , considering the hydro processes are local , it is reasonable to assume that the hydro conditions may not be modulated by the density inhomogeneities on scales much larger than the size of the halo @xmath6 . in this case
, the probability for a halo to have a quasar should be the same for all halos of the same kind , without depending on structures larger than @xmath6 .
thus , the relative fractions of quasars with respect to the certain collapsed halos should be the same for all volumes larger than @xmath7 .
consequently , when averaged on scale larger than @xmath6 , the distribution of quasars @xmath8 at redshift @xmath9 should be proportional to that of the considered halos , @xmath10 .
thus , all effects of the hydro processes can be absorbed into a normalization factor " @xmath11 , i.e. @xmath12 , and @xmath11 is less dependent on @xmath9 than @xmath10 .
the bias of quasar distribution with respect to the mass distribution is then dominated by the bias of the selected halos with respect to the mass .
based on this analysis , quasar bias , at least on large scales , may also be only gravitational , depending on the gravitational parameters used for selecting the quasar - suitable halos . accordingly , a possible model for quasar bias should at least satisfy the three conditions .
1 . gravitational environment given by the identified halos is consistent with the observed environment around quasars ; 2 .
the abundance of quasars , @xmath13 , at redshift @xmath9 is proportional to the number density of the identified halos , @xmath14 in a redshift - independent way , i.e. @xmath15 where @xmath11 is a _ z - independent _ constant , 3 .
the amplitude and @xmath9-evolution of the halo - halo correlation function are consistent with the observed correlation function of quasars . in this letter
, we will show within the framework of the cdm cosmogonic theories that such a bias model can indeed be settled following the above - mentioned points .
the details of the points 1 , 2 and 3 will be discussed in the 2 , 3 and 4 , respectively .
a bias model of quasars is just a phenomenological relationship between the cosmic density field and quasars . with this relationship quasars
can be identified from the mass density field . in this sense , bias , in fact , is a model for the environment suitable for the quasar formation .
in what environment are quasars most likely to be formed ?
many observations indicate that quasars are preferentially located in groups of galaxies .
the evidences include the quasar - galaxy covariance function ( yee & green 1987 ) , the galaxy environments around quasars ( ellingson , et al .
1991 ) , clustering of quasars ( bahcall & chokshi 1991 ; mo & fang 1993 ; komberg , kravtsov & lukash 1994 ; croom & shanks 1996 ; franca , andreani & cristiani , 1997 ) , and the civ - associated absorption in high redshift radio - loud quasars ( foltz et al 1988 ) .
recently , an observation of a companion to quasar br1202 - 0725 with high redshift @xmath16 directly shows that the width of their ly@xmath17 emission lines is around 400 km s@xmath18 ( petitjean et al . 1996 . ) .
these observations seemingly point to quasars being identified with the newly collapsed halos with 1-d velocity dispersion like groups , i.e. @xmath19 .
it should be pointed out that this environment condition may not be necessary for low redshift quasars ( smith , boyle & maddox , 1995 ) , because galaxies and clusters underwent a significant evolution at @xmath20 .
but , it is reasonable to assume that the environment with @xmath21 is favored by the formation of quasars at higher redshifts : a ) enough collapsed objects to form the engine of a quasar ; b ) dispersed gas to feed the engine through accretion ; and c ) not _ too _ many proto - galaxies to possibly disrupt the process of quasar formation .
in the next two sections , we shall compare , within the framework of cdm cosmogonic models , the theoretical predictions for such halos with the observed quasar abundance and correlation functions .
for a gaussian initial perturbation , the comoving number density of halos with 1-d velocity dispersion @xmath22 can be calculated with the press & schechter formalism ( 1974 ) as @xmath23 where @xmath24 is the lagrangian radius of the dark halo being considered and @xmath25 almost independent of cosmologies .
@xmath26 is the @xmath27 of the linear density fluctuations at redshift @xmath9 within a top - hat window of radius @xmath24 , and is determined by the initial density spectrum @xmath28 and normalization factor @xmath29 .
the relationship between @xmath22 and @xmath24 is given by ( narayan & white 1988 ) @xmath30 for a universe with the density parameter @xmath31 .
the n - body simulation results ( jing & fang 1994 ) showed @xmath32 which is used for all calculations in this work .
two representative cdm models , i.e. the standard cdm ( scdm ) and flat low - density cdm ( lcdm ) models are employed .
the hubble constant @xmath33 mpc@xmath18 , mass density @xmath31 , cosmological constant @xmath34 and @xmath35 are taken to be @xmath36 and @xmath37 for the scdm and lcdm , respectively .
the models with these parameters seem to provide a good approximation to many observational properties of the universe , especially the abundance of clusters ( which is much related to the topic of this work ) .
the total number density of the collapsed halos with the velocity dispersion greater than a certain value , say @xmath38 , is @xmath39 the birth rate of halos with @xmath40 is @xmath41 .
this birth rate is shown in fig.1 for @xmath42 200 , 400 , and 800 @xmath43 . for each @xmath44 ,
the birth rate in the two models possesses similar shape . at @xmath45 ,
the birth rate keeps steady from @xmath46 to about 2 , and then rapidly drops to zero at @xmath47 for scdm and 0.7 for lcdm .
the peak birth rate however shifts to higher redshift for smaller @xmath44 , and to lower redshift for larger @xmath44 .
as has been discussed in 1 and 2 , each newly collapsed halo with @xmath48 may host @xmath2 quasars in average , and the harboring coefficient @xmath2 is @xmath9-independent . if the mean lifetime of quasars is @xmath49 , which is also @xmath9-independent , the comoving number density of quasars at the epoch of redshift @xmath9 is given by @xmath50 an accurate quasar lifetime @xmath49 is not important only if it is @xmath9-independent and is much less than the age of the universe for redshift @xmath51 . considering that birth rate of the halos is slowly varying with redshift when @xmath52 , we have approximately @xmath53 ( dz / dt)$ ] . since the shape of the birth rate is a strong function of the velocity dispersion ( especially as a function of @xmath9 ) , the abundance of quasars provides a strong test to the bias model proposed here .
fig.2 plots @xmath54 for the scdm and lcdm models .
the normalization constant " @xmath55 are adjusted in order that the theoretical maximum number densities of quasar abundance can fit with the observed one .
the data points are the number density of quasars complete to absolute magnitude @xmath56 ( pei , 1995 , hewett , foltz & chaffee 1993 , schmidt , schneider & gunn 1992 ) . in these observations ,
the density @xmath54 is measured in the einstein - de sitter cosmology ( @xmath57 , @xmath58 and @xmath59 ) .
these data have been corrected to the case of non - zero @xmath34 in the panel of lcdm .
the figure shows the redshift evolution of quasar abundances is fairly well described by the newly collapsed halos with @xmath45 .
both the scdm and lcdm models are in reasonably good agreement with the observed abundance @xmath13 if the _ constant _ ( @xmath9-independent ) parameter @xmath55 is taken to be @xmath60 yr for the scdm model and @xmath61 yr for the lcdm model , though the best fitting value of @xmath22 is slightly lower ( about @xmath62 ) for the scdm model .
however for @xmath63 , and 800 @xmath43 , the predictions can not fit with observed redshift evolution regardless of how to adjust the
normalization constant " @xmath55 .
therefore , the consistency between theoretical and observed redshift evolution of quasar abundance can be achieved only for @xmath64 halos , which is not trivial .
this result strongly supports that the majority of quasars be associated with new born halos with 1-d @xmath65 .
particularly , for an environment with a given mass or velocity dispersion , the luminosities and lifetimes of quasars are still dispersed . however ,
if the distributions of the luminosities and of the lifetimes for this type of halos are not modulated by large scale perturbations , all conclusion here should hold .
according to the analysis of 1 , the bias model for @xmath66 km s@xmath18 halos is available on the scales larger than their typical size which is about @xmath67 @xmath68 mpc .
therefore , on scales larger than 5 @xmath69 mpc the bias of quasar distribution is due mainly to the bias of @xmath70 halos . on such large scales and at high redshifts , the mass distribution is still linear .
therefore , the two - point correlation function of the halos at redshift @xmath9 is given by @xmath71 where @xmath72 is physical radius , @xmath73 is the linear mass correlation function , and the bias factor @xmath74 is given by ( mo & white 1996 ) @xmath75.\ ] ] fig.3a shows the correlation functions for @xmath76 , 400 and 800 @xmath43 halos at @xmath77 .
it shows that for @xmath78 the correlation length @xmath79 defined by @xmath80 is @xmath81 h@xmath18 mpc for the scdm , and @xmath82 h@xmath18 mpc for the lcdm .
observationally , the two - point correlation function of quasars is found to obey the same power law . with @xmath83 taken to be 0.5 , the correlation length @xmath79 at @xmath84
is found to be @xmath85 mpc ( mo & fang 1993 ) , @xmath86 mpc ( komberg , kravtsov & lukash 1994 ) , and @xmath87 mpc ( franca , andreani & cristiani 1997 ) .
actually , these results do not sensitively depend on @xmath83 , because on the length scales of about @xmath88 mpc , the influence of the parameter @xmath89 is small
. we should be careful in comparing the observed data with fig.3a .
generally , the samples of quasars employed for the correlation statistics possess different limit magnitude for different redshifts .
the samples employed for the correlation statistics contain quasars with @xmath90 .
namely , they are not complete in the sense as the data used in fig.2
. the larger @xmath91 may lead to a lower amplitude of the correlations . considering this uncertainty , both scdm and lcdm with @xmath45 halos are acceptable , though a smaller or larger value of @xmath44 may also be tolerated by the clustering observation alone .
the redshift evolution of the halo - halo correlations for scdm and lcdm is plotted in fig.3b .
it is interesting to see that the amplitudes of the correlation functions do not significantly evolve with redshift , having only a slight increase with redshift .
this is because the clustering in the mass distribution of dark matter @xmath73 always increases with time , or decreases with redshift , but the bias factor @xmath92 increases with redshift .
the two effects seem to be balanced by each other , giving a very slowly varying of @xmath93 with @xmath9 .
the @xmath9-evolution of quasar clustering has been studied for more than one decade ( e.g. chu & fang 1987 , shaver 1988 ) and the results are quite scatter . some showed a weak decrease of the correlation amplitudes around redshift 1.5 on scales of @xmath94mpc ( mo & fang 1993 ; komberg , kravtsov & lukash 1994 ; croom & shanks 1996 . ) some concluded no significant @xmath9-dependence from @xmath95 ( zitelli et al .
1992 . ) and some even show very weak increase with redshift . recently , a weak @xmath9-increasing correlation from @xmath96 to @xmath97 is reported ( franca , andreani & cristiani , 1997 . ) obviously , these diverse data can not provide a concrete test on the prediction of fig.3b .
although there is quite a bit uncertainty in the observed redshift evolution , the current result no significant evolution of either @xmath9-increase and @xmath9-decrease is consistent with the developed bias models .
we showed that velocity - dispersion - selected halos are a possible mechanism for the bias of quasars .
the majority of quasars at redshift @xmath0 formed in the environment of new born collapsed halos with 1-d velocity dispersion @xmath98 .
both the harboring coefficient @xmath2 per halo and the lifetime of quasars are @xmath9-independent . with this bias model , the popular structure formation models , like scdm and lcdm ,
can be fairly well reconciled with data of the abundance and correlations of quasars at @xmath99 .
it is interesting to point out that the velocity dispersion identified halos generally do nt have the same mass .
eq.(2 ) shows that for a given @xmath70 , the redshifts the higher , the mass of the halos the smaller . this result has already been recognized in an earlier study , which shows that in order to fit with quasar abundance at high redshift , the mass of the halos has to be smaller than at the lower redshift ( bi & fang 1997 ) . with this model
, one can predict that 1 .
the environment for quasars at redshifts from @xmath100 to 5 should be characterized by a velocity dispersion , @xmath98 ; 2 .
the amplitudes of quasar two - point correlation function at high redshifts do nt significantly evolve with redshifts . in the paper ,
only the models of the scdm and lcdm are considered .
we can expected that with better data of quasars becoming available , the bias model of quasars will play more important role for discriminating among models of structure formations . | we propose that the majority of quasars at redshift @xmath0 formed in the environment of new born collapsed halos with 1-d velocity dispersion @xmath1 .
the harboring coefficient @xmath2 of quasars per halo and the lifetime of quasars depend only on local process , not modulated by the density inhomogeneities on scales larger than the size of the halos .
thus , the bias of quasars on scale larger than the size of these halos is mainly determined by the parameter @xmath3 used for quasar environment identification . with this model , the popular structure formation models , like scdm and lcdm ,
can be fairly well reconciled with the data of quasars , including a. observed feature of the environment for quasars ; b. redshift evolution of quasar abundance ; c. the two - point correlation functions of quasars .
this bias model predicts that the correlation function of quasars does nt significantly evolve , or only slightly increases with redshift . |
a useful tool for manipulating the local magnetic order in artificially structured materials is provided by the possibility to control the magnetization of a nanomagnet by injecting an electrical current .
several such scenarios have been discussed , including the so - called spin torque transfer ( stt ) technique based on the suggestion by slonczewski @xcite and berger @xcite to use an injection current of spin - polarized electrons .
the high current densities needed in this case can easily be achieved in electrical point contacts of submicron size , where densities of the order @xmath4 a/@xmath5 can be reached without significant heating of the material,@xcite but for larger contacts thermal heating can not be avoided . instead , joule heating caused by an ( unpolarized ) current can be used for thermal manipulation of magnetization . in refs . and
, this approach was proposed for varying the strength of the exchange coupling of two strongly ferromagnetic layers ( f ) separated by a weakly ferromagnetic spacer ( f ) . in ref . , the ability of an external magnetic field to change the relative orientation of the magnetization in the outer layers of an f / f / f tri - layer magnetic stack was demonstrated .
the result was that by varying the temperature , and hence varying the strength of the exchange - spring coupling through the spacer layer f , the relative orientation of the magnetization of the outer f - layers could be continuously and reversibly changed from being parallel to being antiparallel .
consequently , joule heating by forcing a dc current through the structure allows an electro - thermal manipulation of the relative magnetization directions .
this kind of dc current - induced manipulation of the magnetization direction was further studied and observed in stacks predicted to have nonlinear n- and s - shaped current - voltage characteristics ( cvc ) in refs . and , where temporal oscillations of the magnetization direction , temperature and electric current in the magnetic stack were also investigated . in this work
we explore the possibility to use joule heating by a current - in - plane ( cip ) dc electrical current to control the spatial distribution of the magnetization directions in an exchange - spring layered structure of the type@xcite sketched in fig .
[ stack ] . we will show that , if the voltage bias exceeds a critical value for which the sign of the differential resistance becomes negative , two coupled magnetic domain walls can spontaneously appear along the current flow ( see fig .
[ t(x ) ] ) at a distance from each other that can be controlled by the bias voltage .
domain formation is of course not a new phenomenon .
the gunn effect , @xcite well known from the physics of semiconductors , is the name given to the spontaneous formation of ( moving ) electric domains in a semiconductor biased in a region of negative differential resistance ( which requires an n - shaped cvc ) .
electric domains in normal metals can also appear and may be due to structural @xcite and magnetic @xcite transitions , a sharp temperature dependence of the resistance at low temperatures and magnetic breakdown,@xcite evaporation @xcite and melting @xcite ( for a review see , e.g. , ref . ) . in all these cases , however , the domain sizes are macroscopically large , typically several cm .
in contrast , we will show here that the n- and s - shaped cvc : s of the magnetic exchange - spring structures suggested in refs . may give rise to magneto - thermo - electric domains with a characteristic size of the order of 10 @xmath2 m .
the spatial distribution of the magnetization in these stacks can vary from one corresponding to a single magnetic domain wall to a spatially periodic magnetization distribution .
the structure of the paper is as follows . in sect .
[ n - shapedivc ] we briefly discuss some important features of the temperature dependence of the magnetization orientation in the exchange - coupled stack sketched in fig . [ stack ] and derive the n - shaped cvc that is the prerequisite for the magneto - thermo - electric instability discussed in sect .
there we show that under certain conditions an instability leads to spatially highly inhomogeneous distributions of the magnetization direction in one of the layers ( layer 2 in fig .
[ stack ] ) , of the temperature , and of the electric field inside the magnetic stack , corresponding to the formation of a stable magneto - thermo - electric domain ( mted ) structure in the stack . in the concluding sect .
[ conclusion ] we summarize the main results of the paper and estimate the values of the relevant parameters that lead to mted formation .
the magnetic stack under consideration has three ferromagnetic layers as shown in fig .
[ stack ] .
the outer two layers ( 0 and 2 ) are strongly ferromagnetic and coupled via the exchange interaction through a weakly ferromagnetic spacer layer ( 1 ) .
the curie temperature @xmath6 of layer 1 is assumed to be lower than the curie temperatures @xmath7 of layers 0 and 2 .
in addition we assume the magnetization direction of layer 0 to be fixed . a static external magnetic field
@xmath8 , directed opposite to the magnetization of layer 0 , is required to be weak enough that at low temperatures @xmath9 the magnetization of layer 2 is kept parallel to the magnetization of layer 0 due to the exchange interaction between them via layer 1 . at @xmath10 and @xmath11 this tri - layer is similar to the spin - flip free layer " widely used in memory device applications.@xcite the stack is incorporated into an external circuit in such a way that a current @xmath12 flows through the cross - section of the layers and @xmath13v\,.\end{aligned}\ ] ] here @xmath14 and @xmath15 are the magnetoresistance and the angle - independent resistance of the stack , @xmath16 is the angle between the magnetization directions of layers 0 and 2 , and @xmath17 is the voltage drop across the stack . is directed antiparallel to the magnetization in layer 0 and a current @xmath12 flows in the plane of the layers ( along the @xmath18-axis ) . ] in ref .
it was shown that a magnetic configuration with parallel orientations of the magnetization in layers 0 , 1 and 2 becomes unstable if the temperature exceeds some critical temperature @xmath19 . the magnetization direction in layer 2 smoothly tilts with an increase of the stack temperature @xmath9 in the temperature interval @xmath20 .
the dependence of the equilibrium tilt angle @xmath16 between the magnetization directions of layers 0 and 2 on @xmath9 and the magnetic field @xmath8 is determined by the equation @xcite @xmath21 where @xmath22 and @xmath23 here @xmath24 and @xmath25 are the widths and the magnetic moments of layers 1 and 2 , respectively ; @xmath26 is the exchange constant , @xmath27 is the exchange energy in layer 1 , @xmath28 is the bohr magneton , @xmath29 is boltzmann s constant , and @xmath30 is the lattice spacing .
@xmath31 is a dimensionless parameter that determines how effective the external magnetic field is to cause the misorientation effect under consideration .
more precisely , it is the ratio between the energy of magnetic layer 2 in the external magnetic field and the energy of the indirect exchange between layers 0 and 2 ( see fig .
[ noflip ] ) . at low temperatures the indirect exchange energy prevails , the parameter @xmath32 and eq .
( [ theta2 ] ) has only one root , @xmath33 , thus a parallel orientation of magnetic moments in layers 0 , 1 and 2 of the stack is thermodynamically stable .
however , at temperature @xmath19 , for which @xmath34 two new solutions , @xmath35 , appear .
the parallel magnetization corresponding to @xmath33 is now unstable , and the direction of the magnetization in region 2 tilts with an increase of temperature in the interval @xmath36 .
the critical temperature @xmath37 of this orientational phase transition is obtained from eq .
( [ d ] ) as@xcite @xmath38 the orientational transition discussed above can be detected by measuring the temperature dependence of the stack magnetoresistance,@xcite @xmath39 $ ] , plotted for a typical case in fig .
[ t(x ) ] .
this temperature dependence is caused by the temperature dependence of the misalignment angle @xmath40 , which is implicitly given by eq .
( [ theta2 ] ) . $ ] , as determined by the @xmath9-dependence according to eq .
( [ theta2 ] ) of the angle @xmath41 between the magnetization directions in layers 0 and 2 of fig . [ stack ] .
the calculations were made for @xmath42 , @xmath43 ; @xmath44 and @xmath6 is the curie temperature of the spacer layer 1 between layers 0 and 2 . ] if the stack is joule heated by a current @xmath12 its temperature @xmath45 is determined by the heat - balance condition @xmath46 where @xmath47 in conjunction with eq .
( [ theta2 ] ) , which determines the temperature dependence of @xmath48 $ ] . here
@xmath17 is the voltage drop across the stack , @xmath49 is the heat flux flowing from the stack and @xmath50 is the total stack magnetoresistance . here and
below we neglect the explicit dependence of the magnetoresistance on @xmath9 since we consider a thin stack in which elastic scattering of electrons is the main mechanism of the stack resistance . on the other hand , we consider the temperature changes caused by the joule heating only in a narrow vicinity of @xmath6 , which is sufficiently lower than both the critical temperatures @xmath7 and the debye temperature . equations ( [ heat ] ) and ( [ theta2 ] )
define the cvc of the stack @xmath51 } , \label{ivc0}\ ] ] where @xmath52 $ ] .
below we will simplify the notation by dropping the subscript eff " and simply write @xmath53 . from eq .
( [ heat ] ) it is clear that the differential conductance of the stack , @xmath54 , is negative if the sample is joule heated to a temperature at which @xmath55\}/dt < 0.\ ] ] therefore , the main properties of the system under consideration are determined by the behavior of the function @xmath56 $ ] plotted for typical parameters in the top panel of fig .
[ chi+x ] .
using eq .
( [ theta2 ] ) one may express the differential conductance in terms of the dependence of the magnetoresistance @xmath57 on the magnetization angle @xmath16 as@xcite @xmath58'}{[r(\theta ) ( 1-{\bar d}\sin{\theta}/\theta)]'}\bigr|_{\theta=\theta(v ) } , \label{diffg}\ ] ] where @xmath59'$ ] means the derivative of the bracketed quantity with respect to @xmath16 , and @xmath60 it follows from eq .
( [ diffg ] ) the differential conductance @xmath54 is negative if @xmath61 in this case the current - voltage characteristics ( cvc ) of the stack is n - shaped as shown in fig .
[ ivc ] . in which the magnetization directions in the layers are homogeneously distributed along the @xmath18-direction ( along the stack ) .
it was calculated for @xmath62 , @xmath63 , @xmath43 ; @xmath64 ; @xmath65 .
the branches @xmath66 and @xmath67 of the cvc correspond to parallel and antiparallel orientations of the stack magnetization , respectively ( the parts @xmath68 and @xmath69 are unstable ) ; the branch @xmath70 corresponds to @xmath71\leq \pi$ ] . ] using eqs .
( [ current ] ) and ( [ diffg ] ) and assuming the magnetoresistance to be of the form@xcite @xmath72 , where @xmath73 one finds that the differential conductance @xmath54 is negative if @xmath74}{(1 + 2r)r(0)+ ( 1-r)^2(1 - 4r)r_+ } \label{d(cip}\ ] ] it follows that a cvc with a negative differential resistance is possible if @xmath75 . in the case that the stack has a negative differential resistance , nonlinear current and magnetization - direction oscillations may spontaneously arise if the stack is incorporated in a voltage biased electrical circuit in series with an inductor .
@xcite in this paper we show that another type of magneto - electrical instability can arise in such a stack if the electrical current flows in the plane of the layers ( cip - configuration ) : a homogeneous distribution of magnetization direction , temperature and electric field along the spring - type magnetic stack becomes unstable and a magneto - thermo - electric domain spontaneous arise in the stack . here and below we
consider the case that the electrical current flowing through the sample is lower than the torque critical current and hence the torque effect is absent.@xcite
in this section we will work in the voltage bias regime , where the resistance of the external circuit into which the magnetic stack is incorporated can be neglected in comparison with that of the stack . in this case , using the known relation between the electric field and the temperature ( see , e.g. , ref . ) and taking into account that the temperature , @xmath76 , being a function of the coordinate @xmath18 along the stack and time @xmath77 , satisfies the continuity equation for the heat flow , one obtains a set of basic equations for the problem , @xmath78;\nonumber \\ & & j(t)\left<\rho \right>=\frac{v}{l}\ , , \label{heateq}\end{aligned}\ ] ] where the @xmath9-dependence of @xmath41 is given by eq .
( [ theta2 ] ) .
furthermore , @xmath79 is the current density , which is independent of @xmath18 due to the condition of local electrical neutrality , @xmath80 is the cross - section area of the stack and @xmath0 is its length , @xmath81 is the heat capacity per unit volume , @xmath82 is the proportionality coefficient between the electric field and the temperature gradient,@xcite @xmath83 is the thermal conductivity , @xmath49 is the heat flux flowing from the stack , @xmath84 is its volume and @xmath85=r_{\rm eff}[\theta]s / l$ ] is the stack magneto - resistivity [ see eq .
( [ reff ] ) ] , the brackets @xmath86 that appear in the last part of eq .
( [ heateq ] ) indicate an average over @xmath18 along the whole stack of length @xmath0 .
the boundary condition needed to solve eq .
( [ heateq ] ) is the continuity of the heat flux at both ends of the stack ( which is coupled to an external circuit with a fixed voltage drop @xmath17 over the stack ) .
we shall not write down any explicit expression for this condition , since it will become clear below that the magneto - thermal domain structure does not depend on the boundary conditions if @xmath0 is sufficiently large . instead ,
for the sake of simplicity , we use the periodic boundary condition @xmath87 . $ ] and the potential " energy @xmath88 calculated for @xmath62 , @xmath63 , @xmath43 , and @xmath89 ; @xmath64 . ]
the set of equations ( [ heateq ] ) always has the steady - state homogeneous solution @xmath90,\nonumber \\ & & j_0=\rho^{-1}[\theta ( t_0)]v / l,\hspace{0.1cm}f(t_0,j_0)=0.\end{aligned}\ ] ] the last equation in ( [ homogeneous ] ) is identical to the energy balance condition ( [ heat ] ) which , together with eq .
( [ theta2 ] ) , determines the n - shaped cvc shown in fig .
[ ivc ] . for @xmath3 ,
i.e. when the stack contains a magneto - thermo - electric domain .
the cvc was calculated for @xmath62 , @xmath91 , @xmath43 ; @xmath64 ; @xmath92 .
the branches @xmath93 and @xmath67 of the ivc correspond to parallel and antiparallel orientations of the stack magnetization , respectively , while the solid branch @xmath70 corresponds to a stable magneto - thermo - electric domain ( mted ) spontaneously formed inside the stack .
the dashed lines indicate sections where the cvc is unstable .
the red vertical arrows indicate the hysteresis loop in the cvc caused by the disappearance and emergence of the mted as the bias voltage is changed .
the thin horizontal line shows the stabilization current @xmath94 . ] as shown in appendix [ instability ] , if the differential conductance @xmath54 is negative the uniform magnetization along the stack is stable only if the stack length @xmath0 is shorter than some critical length @xmath95 , where @xmath96^{'}}}\,\end{aligned}\ ] ] and the derivatives with respect to @xmath16 are evaluated for @xmath97 .
however , when @xmath98 the uniform distribution of the stack magnetization becomes unstable against fluctuations comprising an arbitrary sum of harmonics @xmath99 ( @xmath100 ) with @xmath101 . a fluctuation with @xmath102 or @xmath103 ( uniform fluctuation ) ,
on the other hand , does not destroy the stability of the homogeneous solution ( [ homogeneous ] ) [ see eq .
( [ zeroharmonics ] ) ] .
we note here that the characteristic value of the critical length can be rather short .
using eqs .
( [ lcritical ] ) and ( [ diffg ] ) and the lorentz ratio @xmath104 , where @xmath105 and @xmath106 is the electron charge , one finds that @xmath107 for a realistic experimental situation : @xmath108 , @xmath109 100 k , @xmath110 cm , @xmath111 a/@xmath5 .
therefore , in the range of parameters @xmath112 homogeneous distributions of the magnetization direction , @xmath16 , temperature @xmath113 [ see eq .
( [ theta2 ] ) and eq .
( [ d ] ) ] , and the electric field , @xmath114j\end{aligned}\ ] ] along the system are unstable , and a magneto - thermo - electric domain , moving with a constant velocity @xmath115 may spontaneous arise inside the magnetic stack : @xmath116,\nonumber \\ & & t(x , t)=\nu(x - st , j_d),\hspace{0.1 cm } j_d(v)=\frac{v}{l}\big < \rho[\theta_d ] \big>^{-1}\end{aligned}\ ] ] where the definition of the brackets @xmath86 is the same as in eq .
( [ heateq ] ) , and @xmath117 satisfies the equation of motion of a fictitious particle of mass " @xmath118 governed by a potential force @xmath119 and a friction force proportional to @xmath120 : @xmath121 ( here @xmath18 is the time " and @xmath122 is the coordinate " of the particle ) .
the velocity @xmath115 of the domain is found from the condition that the total change of energy @xmath123 ( see eq .
( [ energychange ] ) ) in the period " @xmath0 vanishes : @xmath124 hence , the velocity of a magneto - thermo - electric domain is @xmath125 for @xmath126 a/@xmath5 , @xmath127 j / k @xmath128 , @xmath109 100 k. in the same manner as for gunn domains in semiconductors @xcite and electric domains in superconductors and normal metals @xcite , the motion of the domain can be stopped by inhomogeneities inside the stack or at its ends .
however , there is a possibility that the domain can move along the sample , periodically disappearing at one end of the sample and then reappearing at the other , so that the result is temporal non - linear electrical oscillations@xcite with period @xmath129 . in the case under consideration
these oscillations involve the magnetization direction @xmath16 , temperature @xmath9 and electrical current @xmath130 .
using the parameter values already introduced one gets @xmath131 mhz for a magnetic stack of length @xmath132 m .
as shown in appendix [ medstructure ] ( see eq . [ xi ] ) , the function @xmath117 satisfies the following equation with a high accuracy : @xmath133 here the constant @xmath134 plays the role of particle " energy , which determines the period of motion " , @xmath135 , of a non - linear oscillator .
its value is found from the condition that @xmath135 is equal to the length @xmath0 of the magnetic stack , that is the magneto - thermo - electric domain @xmath117 as a function of @xmath18 has only one maximum and one minimum along the length of the magnetic stack . in the case
@xmath136 ( where @xmath137 ) , multiple " domains are also possible in principle . however , they are unstable @xcite and are therefore of no interest for us .
the solution of eq .
( [ conserveenergy ] ) , together with eq .
( [ tandtheta ] ) , defines the spacial distributions of temperature and magnetization direction in a magnetic stack with a magneto - thermo - electric domain .
typical examples of such distributions are presented in fig .
[ t(x ) ] , and the magnetization - misorientation angle , @xmath138 , in a magnetic stack containing a magneto - thermo - electric domain , calculated for @xmath62 , @xmath91 , @xmath43 ; @xmath139 , @xmath64 , @xmath92 , @xmath140^{(1/2)}\sim l_c$ ] . ] using eq .
( [ conserveenergy ] ) and the last equation in ( [ mted ] ) one finds the set of equations @xmath141 @xmath142 the solutions @xmath143 and @xmath144 of these equations are , respectively , the energy " of the domain and the current - voltage characteristics of the magnetic stack containing a magneto - thermo - electric domain ( the dynamic cvc ) . in eqs .
( [ period ] ) and ( [ voltage ] ) the limits of integration , @xmath145 and @xmath146 , i.e. , the minimum and maximum temperature of the domain , are obtained as the real roots of the equation @xmath147 .
let us consider the most pronounced case , for which @xmath148 . in this limit ,
as multiple " domains are unstable , one needs to find a solution of eq .
( [ conserveenergy ] ) that has a period much larger than @xmath149 . since the period @xmath135 of the non - linear oscillator
[ see eq .
( [ period ] ) ] diverges logarithmically as @xmath150min@xmath151 ( where @xmath152 are the maxima and minima of the potential " energy , see fig .
[ chi+x ] ) , this condition is satisfied if the energy " @xmath153 of the domain differs from min@xmath154 by an exponentially small amount .
from here it follows that variations of the current inside a narrow interval around @xmath155 , defined by the relation @xmath156 drastically changes the form of the magneto - thermo - electric domain and hence the current voltage characteristics . solving the set of equations eq .
( [ period ] ) and eq .
( [ voltage ] ) in the interval @xmath157 ( where @xmath158 ) and taking into account that the maximal , @xmath146 , and minimal , @xmath145 , temperatures of the domain are close to @xmath159 and @xmath160 , respectively , one find the following implicit form of the dynamic cvc : @xmath161 here @xmath162 , @xmath163^{(1/2)}\sim l_c$ ] and the constants @xmath164 are of the order of @xmath165 while @xmath166 . differentiating both sides of eq .
( [ dynamiccvc ] ) with respect to @xmath167 one finds that the dynamic cvc have vertical tangents at the two points @xmath168 and @xmath169 , where @xmath170 and @xmath171 while @xmath172 it follows from the above equations that the values of the currents @xmath173 and @xmath174 differ from @xmath155 by an amount @xmath175 .
one can see from eq .
( [ dynamiccvc ] ) and eq .
( [ verticaltangents ] ) that for all values of @xmath162 in the interval @xmath176 , except for a small region near the ends of the interval , the current @xmath130 coincides with @xmath155 to an accuracy that is exponential in the parameter @xmath177 .
a change of the bias voltage in this interval does not change the current but it does change the length of the magneto - thermo - electric domain , that is the length of the higher resistive part of it providing the needed voltage drop across the stack at the fixed current value @xmath178 . therefore , a magnetic stack with a magneto - thermo - electric domain inside can work as a high - quality current stabilizer .
near the points @xmath179 and @xmath180 there is a sharp transition from the nearly horizontal segment of the dynamic cvc to the rising segments of the cvc corresponding to a homogeneous state of the magnetic stack ( see fig .
[ ivc+mted ] ) . as a result , there are hysteresis loops in the current - voltage characteristics of the stack of a large enough length @xmath0 . the fact , mentioned above , that the current is nearly independent of the bias voltage in the interval @xmath181 $ ] ( here @xmath182 comes about because in this region the domain structure can be written to exponential accuracy ( i.e. , with an error @xmath183 ) as @xmath184 where the function @xmath185 is a domain - wall type solution of eq .
( [ medeq ] ) at @xmath186 and @xmath187 , the asymptotic behaviour of which is @xmath188 here @xmath189 are the points of deflections of the curve @xmath190 so that @xmath191 is approximately the length of the hot " section of a trapezoidal mted having the maximal temperature @xmath192 and @xmath193 is the length of its cold " section having the minimal temperature @xmath194 ( see fig . [ t(x ) ] )
we have shown that joule heating of the magnetic stack sketched in fig . [ stack ] by a current flowing in the plane of the layers may result in an instability of an initially homogeneous distribution of the magnetization of the stack if the length @xmath0 of the stack in the direction of the current flow is longer than some critical length @xmath95 .
this instability results in the spontaneous appearance of moving domains of magnetization direction in layer 2 of fig .
[ stack ] , @xmath195 , temperature , @xmath196 and electric field , @xmath197 . for the case
@xmath198 the length of the domain is of the order of @xmath95 .
if the length of the stack greatly exceeds the critical length , @xmath3 , the stack is spontaneously divided into two regions , where in one region the magnetization directions in layer 1 and layer 2 of fig .
[ stack ] are parallel to each other , while in the other region they are antiparallel .
the length of the region with antiparallel magnetization orientations ( that is the length of the domain @xmath199 ) is controlled by the bias voltage in the interval @xmath200 ( see fig . [ t(x ) ] ) . in this case
the cvc of a stack containing such a domain has a plateau with hysteresis loops at the ends , as shown in fig .
[ ivc+mted ] .
therefore , the stack can work as a current stabilizer since the current flowing through it has a fixed value @xmath201 to within an exponentially small error @xmath202 ; a change of bias voltage only results in a change of the domain length to provide the needed voltage drop over the stack . for a realistic experimental situation the value of the parameter @xmath203 , which through eq .
( [ r ] ) determines the dependence of the stack resistance on the magnetization - misorientation angle @xmath16 , can be estimated to be of order @xmath204 , while the curie temperature @xmath6 of the spacer layer 1 can be @xmath205 k. using these values and a resistivity of @xmath206 @xmath207 cm , a current density of @xmath111 a/@xmath5 and a specific heat of @xmath127 j / k @xmath128 one finds that the critical length is @xmath208 @xmath2 m .
financial support from the european commission ( fp7-ict - fet proj . no
. 225955 stele ) , the swedish vr , and the korean wcu program funded by mest / nfr ( r31 - 2008 - 000 - 10057 - 0 ) is gratefully acknowledged .
in order to investigate the stability of the spatially homogeneous solution ( [ homogeneous ] ) for temperature , @xmath209 , current density , @xmath186 , and magnetization misalignment angle , @xmath210 , against spatial fluctuations we express these quantities as sums of two terms , @xmath211 where @xmath212 and @xmath173 each is a small correction . inserting eq.([corrections ] ) into eqs .
( [ heateq ] ) and ( [ theta2 ] ) and using the fourier expansion @xmath213 one finds that @xmath214 while the equation for the fourier harmonics of the angle are @xmath215^{'}\big\}_{\theta= \theta_0}\theta^{(0)}_1\end{aligned}\ ] ] if @xmath216 , and @xmath217^{'}\big\}_{\theta= \theta_0}\theta^{(n)}_1;\end{aligned}\ ] ] if @xmath218 . from this result
one sees that if @xmath219_{\theta=\theta_0}^ { ' } < 0\ ] ] [ that is if @xmath220 , see eq .
( [ diffg ] ) ] the uniform magnetization along the stack looses its stability if the stack length exceeds some critical value . setting the right - hand side of eq .
( [ nonzeroharmonics ] ) equal to zero one gets the result ( [ lcritical ] ) for the critical length @xmath95 .
by multiplying both sides of eq .
( [ medeq ] ) by @xmath221 one finds that the quantity @xmath222 plays the role of the total energy " of a fictitious particle , the first term giving the kinetic energy `` and the ' ' potential energy " being defined as @xmath223 ( the dependence of @xmath224 on @xmath9 is shown in fig .
[ chi+x ] ) . the change of energy with time " @xmath18 is caused by the action of the friction " force : @xmath225 as one sees from eq .
( [ medeq ] ) , the ratio of the friction force " to the inertial " term is of the order @xmath226 where @xmath227 is the fermi energy . from here
it follows that the function @xmath228 satisfies an energy " conservation law @xmath229 ( see eq .
( [ totalenergy ] ) ) to within an error @xmath230 , which results in eq .
( [ conserveenergy ] ) . | we investigate the potential to use a magneto - thermo - electric instability that may be induced in a mesoscopic magnetic multi - layer ( f / f / f ) to create and control magnetic superstructures . in the studied multilayer
two strongly ferromagnetic layers ( f ) are coupled through a weakly ferromagnetic spacer ( f ) by an exchange spring " with a temperature dependent spring constant " that can be varied by joule heating caused by an electrical dc current .
we show that in the current - in - plane ( cip ) configuration a distribution of the magnetization , which is homogeneous in the direction of the current flow , is unstable in the presence of an external magnetic field if the length @xmath0 of the sample in this direction exceeds some critical value @xmath1
@xmath2 m .
this spatial instability results in the spontaneous formation of a moving domain of magnetization directions , the length of which can be controlled by the bias voltage in the limit @xmath3 .
furthermore , we show that in such a situation the current - voltage characteristics has a plateau with hysteresis loops at its ends and demonstrate that if biased in the plateau region the studied device functions as an exponentially precise current stabilizer . |
mixed antiferroelectric - ferroelectric systems like pbzr@xmath0ti@xmath1o@xmath2 ( pzt ) , rb@xmath0(nh@xmath3)@xmath1h@xmath4po@xmath3 ( radp ) ( @xmath10 ) attract much attention of scientists because of their unusual properties and phase diagram peculiarities .
these anomalies made pzt material useful for many technical application [ 1 ] . in
particular compositions of pzt enriched by zr ( @xmath11 ) appeared to be useful for application in the novel branches of electronic technique ( see e.g. [ 2 ] , [ 3 ] and ref . therein ) .
theoretical description of the properties and phase diagram nearby antiferroelectric phase transition is hard problem because of complexity of description of antiferroelectric phase itself .
the proximity of free energies of antiferroelectric and ferroelectric state is known to be the characteristic feature of antiferroelectrics [ 4 ] .
this proximity lead to dielectric permittivity anomalies and to the possibility to induce ferroelectric order by application of small enough electric field .
the latter can be the clue of the reasons of ferroelectric phase appearance even at small @xmath12 values .
small content of ferroelectric material in the mixed system can be considered as the source of electric dipoles and induced by them random electric field .
calculation of this field magnitude allowing for dipole - dipole interaction via ferroelectric and nonferroelectric soft modes had shown that mean field is induced by the interaction via ferroelectric soft mode whereas other soft modes increase mean square field [ 5 ] .
the supposition that nonzero internal mean field can destroy antiferroelectric order by the same way as external field made it possible to describe semiquantitively the observed phase diagram [ 5 ] . in present work we performed detailed calculations of phase diagram of mixed antiferroelectric - ferroelectric system for the region of small concentration of ferroelectric material .
the obtained data fitted pretty good measured phase diagram for pzt and radp .
the application of proposed method for calculation of phase diagrams in different mixed ferroelectrics is discussed .
let us begin with the description of antiferroelectric phase . in ising model
free energy @xmath13 can be written as [ 4 ] : @xmath14 \label{e1}\end{aligned}\ ] ] where @xmath15 , effective dipole moment @xmath16 , @xmath17 and @xmath18 are antiferroelectric component susceptibility and lorentz factor respectively , @xmath19 ( @xmath20 are the positions of two - well potential minima ) , @xmath21 is electric field , @xmath22 and @xmath23 are respectively ferroelectric and antiferroelectric transition order parameters ( @xmath24 , @xmath25 , @xmath26 and @xmath27 being mean homogeneous and inhomogeneous displacements respectively ) ; @xmath28 and @xmath29 are respectively antiferroelectric transition and curie temperatures .
note , that in this model mean dipole moment of the unit cell has the form : @xmath30 thus we consider the simplest two - sublattice model of antiferroelectric .
( [ e1 ] ) leads to the following expressions for the order parameters : @xmath31 eqs .
( [ e1])-([e3 ] ) describe antiferroelectric component , i.e. mixed system at @xmath32 .
ferroelectric component corresponds to @xmath33 and its order parameter @xmath34 in ising model has the form [ 4 ] @xmath35 where @xmath36 is mean field , that defined ferroelectric transition temperature , @xmath37 is effective dipole moment of ferroelectric component ions .
addition of small amount of ferroelectric component to antiferroelectric one leads to appearance of random electric field @xmath38 , induced by electric dipoles of this component in the mixed system .
mean field in the mixed system has to include the contribution from ferroelectric component and from ferroelectric order parameter of antiferroelectric component namely @xmath39 this field as the most probable one has to define the position of maximum of random field distribution function @xmath40 .
it is equal to @xmath41 in mean field approximation .
beyond of this approximation one has to take into account random field , i.e. to substitute @xmath42 ( @xmath43 ) or @xmath44 ( @xmath45 ) for @xmath21 in eqs .
( [ e3 ] ) , ( [ e4 ] ) and to perform averaging with random field distribution function .
this yields : @xmath46 @xmath47 @xmath48 one can see that in mean field approximation when @xmath40 has the form of @xmath49-function eqs .
( [ e6 ] ) give eqs .
( [ e3 ] ) at @xmath50 and eq .
( [ e4 ] ) at @xmath33 .
( [ e6 ] ) are general expressions for order parameters dependence on concentrations and characteristics of mixed system components . note , that in the case @xmath51 ( both components of mixed system are ferroelectrics ) the equations for @xmath22 and @xmath34 have the same form as it has to be expected .
let us proceed now to calculations of the transition temperatures @xmath7 ( to antiferroelectric phase ) and @xmath8 ( to ferroelectric phase ) dependence on concentration . for the phase transitions of the second order
these temperatures can be obtained from eqs .
( [ e6 ] ) in the limit of @xmath52 , @xmath53 , @xmath54 .
in such a limit one can simplify the integrands in eqs .
( [ e6 ] ) substituting @xmath55 , @xmath56 and representing the distribution function in the form : @xmath57 here @xmath58 is the distribution function in the case of mean field absence . keeping in mind that @xmath58 is even parity function in aforementioned limit eqs .
( [ e6 ] ) give : @xmath59 \beta i(x)\ ] ] @xmath60 @xmath61 \beta i(x)\ ] ] @xmath62 integral ( [ e9 ] ) has to depend on concentration because of dependence on @xmath63 of the distribution function width .
the solution of the system of eqs .
( [ e8 ] ) yields : @xmath64 i_c(x ) \label{e10a}\ ] ] or in dimensionless variables @xmath65 eq .
( [ e10a ] ) can be rewritten as @xmath66 i_c(x ) \label{e10c}\ ] ] here @xmath67 and @xmath68 are integrals ( [ e9 ] ) at @xmath69 and @xmath70 respectively .
one can see , that the transition temperatures are defined by the corresponding transition temperatures and concentrations of the components of mixed system and random field distribution function characteristics .
equations ( 10 ) define concentrational dependence of transition temperature , i.e. phase diagram of mixed system for small concentration of ferroelectric component .
one can see , that eqs .
( 10 ) give the correct values in mean field approximation : @xmath71 , @xmath72 at @xmath32 and @xmath73 , @xmath74 at @xmath33 .
random field induced by electric dipoles of ferroelectric component of mixed system can destroy completely antiferroelectric phase at some concentration of the dipoles .
this concentration @xmath75 is known to be critical concentration at which @xmath76 , so that antiferroelectric phase exists at @xmath77 only . on
the other hand ferroelectric phase in the mixed system arises at @xmath78 where @xmath79 is critical concentration for ferroelectric phase transition .
it has to be @xmath80 at @xmath81 .
therefore critical concentrations @xmath82 and @xmath79 in the region @xmath83 can be obtained from the equations @xmath76 ( @xmath82 ) and @xmath84 ( @xmath85 ) .
let us limit ourselves by the case @xmath86 which is of importance both for case of pzt and radp on the edges of phase diagram . to perform the calculations on the base of eqs .
( [ e10c ] ) we have to know concentrational dependence of integrals @xmath87 defined by the distribution function @xmath88 . for electric dipoles as random field sources
it can be represented in gaussian form ( see [ 6 ] for details ) : @xmath89 @xmath90 here @xmath91 , @xmath92 and @xmath93 are respectively dielectric permittivity , lattice constant and correlation radius connected with soft ferroelectric mode of antiferroelectric component of mixed system .
parameter @xmath94 takes into account dipole - dipole interaction via antiferroelectric and other nonferroelectric soft modes [ 5 ] . performing changing of variables in integrals @xmath95 ( @xmath96 )
, we obtain that this reduces just to renormalization of parameter @xmath97 in ( [ e12 ] ) , giving @xmath98 .
to map correctly the interval @xmath99 into interval @xmath100 we adopt following parametrization for @xmath101@xmath102 let us begin with critical concentration for antiferroelectric phase transition . at @xmath103 eqs.([e10a ] ) with respect to eqs .
( [ e11 ] ) , ( [ e12 ] ) at @xmath104 @xmath105 yields @xmath106 substitution of eq .
( [ e13 ] ) into the expression for @xmath107 ( see eqs .
( [ e10c ] ) ) gives the following equation for critical concentration @xmath82 : @xmath108 eq .
( [ e14 ] ) has been solved numerically during the calculation of @xmath109 critical concentration for ferroelectric phase transition can be obtained by the same way , but in the limit of @xmath110 , i.e. at @xmath111 , this ratio being larger or close to unity .
since in such a limit integral @xmath112 can be calculated only numerically we shall obtain @xmath79 value in the next section devoted to numerical calculations with two fitting parameters @xmath113 and @xmath114 .
phase diagram .
i.e. concentrational dependence of antiferroelectric and ferroelectric phase transition temeprature was calculated with help of eqs.(10 ) .
integrals @xmath115 were calculated numerically .
the obtained phase diagram is represented in fig.1 in the region of small concentration of ferroelectric component for several values of dimensionless parameters @xmath116 @xmath117 and @xmath118 ( see eq.([e10b ] ) ) .
one can see from fig.1 that antiferroelectric phase transition critical concentration @xmath82 increases with @xmath119 decrease , its values being close to those calculated on the base of eq.([e14 ] ) .
the obtained critical concentration of ferroelectric phase transition several times smaller than that for antiferroelectric transition .
this seems to be the result of proximity of free energies and may be of @xmath120 and @xmath121 which is characteristic feature of antiferroelectrics .
the dimensionless transition temperature to antiferroelectric phase @xmath107 decreases with concentration increase for all the considered parameters , whereas the behaviour of that to ferroelectric phase @xmath122 depends on the choise of the parameters values .
in particular @xmath122 can increase or decrease with increase of electric dipoles concentration ( see fig.1 ) .
physically these types of behaviour depend on distribution function width and ratio ( @xmath118 ) of ferroelectric phase transition temepratures @xmath123 of ferroelectric component of mixed system to that of antiferroelectric component @xmath120 : the smaller the distribution function width and the larger @xmath118 value , the larger ferroelectric phase transition temperature in the mixed system .
let us begin with comparison of obtained data with experiment .
the prominent example of mixed ferro - antiferro system is known to be pbzr@xmath0ti@xmath124o@xmath125 ( @xmath10 ) , where pbzro@xmath125 is antiferroelectric with transition temeprature from cubic paraelectric phase to rhombic antiferroelectric phase at @xmath126 k and pbtio@xmath125 is ferroelectric with transition temeprature from cubic paraelectric phase to tetragonal ferroelectric phase at @xmath127 k. phase diagram of mixed pzt system was investigated in many papers , but mainly at @xmath128 ( see e.g. [ 6 ] and ref . therein ) .
therefore critical concentration @xmath82 is known , and @xmath129 k at @xmath130 .
one can see that the curve for @xmath131 gives @xmath132 k at @xmath133 and @xmath134 .
general form of calculated @xmath135 concentrational dependence looks like experimental curve at @xmath136 ( fig.1 ) . there is a scattering of experimental data for ferroelectric phase transition temperature at small pbtio@xmath125 content ( @xmath137 ) because of influence of some uncontrolled defects and impurities on @xmath138 value .
the most probable behaviour of @xmath139 is its slight increase ( fig.1 ) [ 6 ] .
such behaviour is similar to the calculated curve with @xmath140 and @xmath141 .
another important example of mixed ferro - antiferro system is rb@xmath0(nh@xmath142)@xmath124h@xmath143po@xmath144 ( @xmath10 ) as a representative of hydrogen - bonded material .
it was studied in many details ( see e.g. [ 7 ] ) .
rbh@xmath145po@xmath144 ( rdp ) undergoes a ferroelectric phase transition at @xmath146 k , nh@xmath144h@xmath143po@xmath144 ( adp ) becomes antiferroelectric near @xmath147 k. in mixed radp system antiferroelectric transition temperature @xmath148 decreases with increase of rdp content and @xmath76 k near @xmath149 of rdp admixture , i.e. @xmath150 [ 8 ] .
substitution of this @xmath82 value into eq.([e14 ] ) gives @xmath151 .
one can see from fig.2 that the curve , which corresponds to @xmath152 fits the experimental points .
contrary to pzt there is no data about existence of ferroelectric phase transition at small ( @xmath153 ) content of rdp .
the reason for this difference can be the approximately equal values of @xmath123 and @xmath154 in radp whereas they are different in pzt . because of this theoretical parameter @xmath155 in radp and at @xmath156 @xmath157 curve may be lower than @xmath158 curve and only antiferroelectric phase transition exists .
one can see from fig.2 , that proposed theory explaines the observed behaviour of ferroelectric phase transition decrease with increase of antiferroelectric component concentration and observed critical concentration @xmath79 ( @xmath159 ) .
note that tuneling has to be taken into account for mixed systems of kdp family .
preliminary calculations have shown , that tuneling result into change of fitting parameters values .
therefore the proposed theory describes pretty good phase diagrams of pzt and radp for small @xmath12 values .
note , that only for small @xmath12 ( @xmath160 ) ti and rb ions can be considered as impurities randomly distributed in pbzro@xmath161 and nh@xmath144h@xmath143po@xmath144 matrices so that they can be the sources of random field .
4.2 . the proposed theory can be applied for another types of mixed systems .
for example , mixture of two ferroelectric systems @xmath162 at @xmath11 described with the help of eqs .
( [ e6 ] ) , ( [ e8 ] ) at zero antiferroelectric phase order parameter ( @xmath51 ) . in this case
the equations for @xmath22 and @xmath34 completely the same , i.e. @xmath163 .
in such situation mean field @xmath164 is ferroelectric phase transition temperature of @xmath165 component whereas @xmath166 component defines the form and width of random field distribution function . because the main physical idea of the proposed model is the inhibition and distruction of long range ferroelectric or antiferroelectric order by random electric field of different sources this model can be useful for description to relaxor ferroelectrics like pbmg@xmath167nb@xmath168o@xmath125 ( pmn ) , pbsc@xmath169ta@xmath170o@xmath125 ( pst ) , pb@xmath171la@xmath172zr@xmath173ti@xmath174o@xmath125 ( plzt , @xmath175 ; 0,09 ) .
but the relaxors like pmn , pst are not solid solutions of two component , and in plzt la ions are the impurities .
thus in all these materials there is onlu one ferroelectric component known as burns reference phase ( see [ 9 ] and ref . therein ) , which is pbzr@xmath173ti@xmath174o@xmath125 in the of plzt . the transition temperature from para- to ferroelectric phase of
the reference phase @xmath176 defines the position of random field distribution maxima whereas the substitutional disorder , vacancies of lead and oxygen , impurities like la in plzt define the distribution function width .
namely such a model has been proposed recently [ 10 ] .
it permitted to describe the observed phase diagram of the relaxors , their dynamic and static properties peculiarities ( see e.g. [ 11 ] and ref .
therein ) .
since the model is able to describe the distraction of long range order by any random field sources it can be useful also for description of solid solutions of ferroelectric and dielectric components , e.g. ba(ti@xmath0sn@xmath177)o@xmath125 ( @xmath11 ) . the observed in this material relaxor - like behaviour as well as change of phase transition order [ 12 ] can be the consequence of strong random field influence as it was shown recently [ 11,13 ] .
note that some residual ferroelectric domains were observed in another relaxor plzt ( 9/65/35 ) [ 14 ] .
as the matter of fact , the dipole glass and mixed ferroglass phases ( which appear in the relaxors with temperature lowering ) can be considered as the reentrant phases with respect to this residual ferroelectric phase existing near the burns temperature .
therefore the generality of proposed model gives a possibility to apply it to various mixed systems which are solid solutions of ferro- and antiferroelectrics , two ferroelectrics , ferroelectric and dielectric etc .
as well as to ferroelectric relaxors . | phase diagram of solid solutions of antiferroelectric and ferroelectric materials like pbzr@xmath0ti@xmath1o@xmath2 ( pzt ) , rb@xmath1(nh@xmath3)@xmath0h@xmath4po@xmath5 ( radp ) is calculated at the region @xmath6 . antiferroelectric and ferroelectric order parameters ( which correspond respectively to the difference and sum of two sublattices ions shifts ) were introduced for antiferroelectric host lattice .
small admixture of ferroelectric component was considered as a random electric field source .
the influence of this field on both aforementioned order parameters was calculated .
self consistent averaging with calculated random field distribution function made it possible to obtain the system of equations for order parameters of mixed system .
the antiferroelectric @xmath7 and ferroelectric @xmath8 phase transition temperature concentrational dependence was expressed through the transition temperatures of the composition and members .
the obtained @xmath9 dependence was shown to fit pretty good the observed values for pzt and radp .
the application of the proposed theoretical approach to other types of mixed systems is discussed . 2 |
in @xcite we investigated the time dependent schrdinger - newton equation ( henceforth abbreviated sn equation ) : @xmath0 with a potential term that is the spatial convolution of the absolute - value squared of the wave function with the coulomb potential @xmath1 : @xmath2 our numerical and analytical consideration of the sn equation in @xcite showed that inhibitions of dispersion due to the gravitational self - interaction occur at mass values of about @xmath3u for a given wave packet width of @xmath4 .
this result was contrary to a previous one by salzman and carlip @xcite , where inhibitions of dispersion were claimed to occur at mass values more than six orders of magnitude smaller , namely at @xmath5 . somewhat loosely , this behaviour of a wave packet initially shrinking in width was called a `` collapse '' by salzman and carlip and we adopt this nomenclature .
as we pointed out in the summary of @xcite , more consideration has to be given to the question concerning the application of the sn equation to realistic systems , like molecules . in a recent paper @xcite jskelinen
suggested that the gravitational self - interaction for a molecule should be modelled by the potential of a solid sphere of radius @xmath6 @xmath7 which has to be put into equation ( [ eq : schroedingernewton ] ) instead of the coulomb potential .
jskelinen points out that in our case , for a wave packet width of @xmath4 , the results obtained for this potential deviate from the results for the coulomb potential for masses beyond @xmath8u .
we will next briefly review the rationale behind such a modification and then turn to the quantitative changes it implies for our previous analysis @xcite . for comparison we will also consider the case of a hollow sphere .
in this section we briefly discuss how a one - particle sn equation for the centre of mass is obtained by separation from an @xmath9-particle sn equation , as claimed in @xcite .
consider the @xmath9-particle sn equation @xmath10 \right ] \psi_n(\vec{r}^n ) , \label{eqn : n - particle - sn } \\
v_\text{em}(\vec{r}^n ) & = \sum_{i=1}^n \sum_{j=1}^{i-1 } \frac{q_i q_j}{{\vert \vec{r}_i - \vec{r}_j \vert } } \\ u_\text{g}[\psi_n(\vec{r}^n ) ] & = -g \sum_{i=1}^n \sum_{j=1}^n m_i m_j \int \frac{{\vert \psi_n(\vec{r'}^n ) \vert}^2}{{\vert \vec{r}_i - \vec{r_j ' } \vert } } \ ; { \mathrm{d}}{v'}^n \nonumber\\ & = -g \sum_{i=1}^n \sum_{j=1}^n m_i m_j \int \frac{p_j(\vec{r_j'})}{{\vert \vec{r}_i - \vec{r_j ' } \vert } } \ ; { \mathrm{d}}^3 r_j ' \;.
\label{eqn : ugrav}\end{aligned}\ ] ] here @xmath11 indicates the set of all @xmath9 coordinates @xmath12 , @xmath13 the mass of the @xmath14-th particle and @xmath15 is the laplacian with respect to the @xmath14-th coordinate . by @xmath16
we denote the @xmath14-th marginal distribution that one obtains by integrating @xmath17 over all @xmath18 factors @xmath19 that parametrise the positions of all particles except the @xmath14-th . the intuitive picture underlying the gravitational potential term @xmath20 is as follows : _ each particle is under the influence of a newtonian gravitational potential that is sourced by an active gravitational mass - density to which each particle contributes proportional to its probability density in position space as given by the marginal distribution of the total wave function : _ @xmath21
although each particle lives in its own copy of @xmath19 they all `` feel '' the same matter density @xmath22 .
in particular , the self - interaction of each particle with its own field is included .
the gravitational interaction of all @xmath9 particles is then given by @xmath23 & = -g\ \sum_{i=1}^n \int \frac{m_i \ , \rho^\text{mat}(\vec{x})}{{\vert \vec{r}_i - \vec{x } \vert } } \ ; { \mathrm{d}}^3 x \nonumber \\ & = -g\ \sum_{i=1}^n \sum_{j=1}^n \int \frac{m_i \ , m_j \ , p_j(\vec{x})}{{\vert \vec{r}_i - \vec{x } \vert } } \ ; { \mathrm{d}}^3 x\,.\end{aligned}\ ] ] note that this is the @xmath9-particle sn equation as it was stated by disi @xcite , which is different from that given by jskelinen @xcite . in
@xcite the _ mutual _ gravitational interactions are said to be negligible compared to electromagnetic interactions and only the diagonal terms of @xmath20 are considered .
but in situations where the width of the individual marginal distributions is large compared to the mutual distances of their centres these contributions are of the same order of magnitude as the diagonal terms .
we emphasise that in this treatment there is a _
twofold _ difference between the electromagnetic and the gravitational interaction of the particles .
first , self - interactions of particles with their own _ electromagnetic _ field are not considered .
second , whereas the electromagnetic interaction is local , the gravitational interaction is non - local in the sense explained above . and it is this non - locality that is responsible for a non - trivial contribution of the overall gravitational field to the dynamics of the centre of mass .
we now introduce the centre of mass coordinates @xmath24 and make the separation ansatz @xmath25 , where the prefactor comes from the integral measure and allows us to normalise all wave functions to one .
then @xmath26 only depends on the relative coordinates @xmath27 .
substituting the separation ansatz into the expression ( [ eqn : ugrav ] ) , @xmath20 can be written as @xmath28 where we assumed @xmath9 to be a large number , such that all terms involving the @xmath9-th particle can be neglected .
they give only @xmath29 out of @xmath30 contributions .- th particle .
] let @xmath31 be the marginal distribution of the relative wave function @xmath32 for the @xmath14-th particle and rename the integration variable @xmath33 to @xmath34 . then , in the above approximation , @xmath35 now note that @xmath36 is simply the matter density of the @xmath37-th particle relative to the centre of mass .
we can therefore substitute the inner integral with the gravitational potential of the @xmath37-th particle , which is @xmath38 we then obtain @xmath39 if now we assume that the extent of the molecule is much smaller than the width of the wave function , such that the wave function does not change much over a distance @xmath40 , i.e. @xmath41 , we can neglect the shift by @xmath27 . the sum over all @xmath13
then just yields the total mass @xmath42 and the one particle potentials @xmath43 sum up to yield the full gravitational potential @xmath44 .
we finally get @xmath45 = \int { \vert \psi(\vec{r ' } ) \vert}^2 \ ; \phi(\vec{r } - \vec{r ' } ) \ ; { \mathrm{d}}^3 r ' \,.\ ] ] the sn equation ( [ eqn : n - particle - sn ] ) then separates into the ordinary electromagnetic multi - particle schrdinger equation for the relative coordinates , for which the solution is taken to be given by the present lump of matter , and the sn equation ( [ eq : schroedingernewton ] ) for the centre of mass wave function @xmath46 with the modified gravitational potential @xmath44 for the matter at hand .
+ + we study the evolution of the spherically symmetric sn equation making use of the same numerical methods used in @xcite , simply replacing the potential by the solid - sphere potential ( [ eqn : solid - sphere potential ] ) .
all results refer to an initial gaussian wave packet @xmath47 with a width of @xmath48 .
first we look at the evolution of wave packets if the density of the solid sphere corresponds to the density of either lithium ( 534kg@xmath49 ) or osmium ( 22610kg@xmath49 ) , cf .
@xcite . to illustrate the results ,
the radius within which 90% of the probability density are contained is plotted against time in figure [ fig : plot_r90 ] for several masses .
in addition to the results for the density of lithium ( li ) and osmium ( os ) the results for the sn equation with coulomb potential obtained in @xcite are also plotted
. for all masses considered the modified potential results in an even faster collapsing wave packet .
the shrinking behaviour also sets in for smaller masses , namely for @xmath50u the modified solution shrinks while the sn equation solution is still spreading for this mass . to get a better impression of this behaviour ,
let us take a look at the effect that the radius of the solid - sphere potential has on the mass for which the collapsing behaviour of the wave packet sets in .
we thus ask the following question : given a sphere of radius @xmath6 , what mass @xmath51 do we have to distribute homogeneously `` within '' this sphere . ] for a collapse of the wave packet to take place ? in @xcite we found that for a @xmath52 wave packet a collapse is observed for masses of about @xmath3u and higher . in figure
[ fig : plot_m_r ] this critical mass value is plotted depending on the radius of the solid sphere . for small radii of several nanometers
this critical mass reduces to about two - thirds of the value obtained in @xcite and it stays smaller than this value until the size of the solid sphere exceeds the width of the wave packet significantly . only holds for small radii compared to the extent of the wave packet , the results for larger radii have to be regarded with some scepticism , anyway .
] there seems to be an extremal value somewhere in between 20 and 100 nm .
for which the total energy is zero . ]
the behaviour described in the previous paragraph seems peculiar , as naively one would expect the gravitational field to be the weaker , and therefore the collapse mass to be the higher , the larger the solid - sphere radius gets .
but the ascertained radius - dependence of the collapse mass becomes more comprehensible by an analysis of the total energy value of the sn equation .
we performed a similar analysis in section 3.2 of @xcite where we indicated the total energy for the sn wave packet : @xmath54 where @xmath55 .
@xcite pointed out that for the sn equation a collapse can only be observed if the total energy is positive .
the second time derivative of the second moment @xmath56 is given by @xmath57 for an arbitrary potential @xmath58 .
if the potential fulfils a poisson equation @xmath59 it takes the form @xmath60 which is strictly positive for positive energy .
although this relation does not hold for the modified potentials , the mass value for which the total energy equals zero might nevertheless provide an insight of the qualitative behaviour of the equation .
we thus plot this mass value @xmath61 against the radius @xmath6 of the solid - sphere in figure [ fig : m0 von r ] .
as one can see , this mass value indeed decreases significantly within the width of the wave packet and starts to increase only when the radius of the solid sphere exceeds this width .
thus , qualitatively the behaviour of the zero - total - energy mass - value coincides with the behaviour of the collapse mass . at least for the experimental situation of interferometry with fullerene molecules
, it seems a more realistic model to consider a hollow - sphere potential @xmath62 instead of the solid - sphere potential ( [ eqn : solid - sphere potential ] ) .
let us therefore repeat the analysis from section [ sec : solid - sphere ] for the potential ( [ eqn : hollow - sphere potential ] ) .
the result is plotted in figure [ fig : plot_m_r_hohl ] .
as one can see , the collapse mass does not decrease but remains almost constant as long as the radius of the hollow sphere does not exceed the wave packet width . for larger
radii the collapse mass increases fast . for the coulomb potential ( dashed straight line ) , the solid sphere ( dotted line ) and the hollow sphere ( solid line ) . ]
this can also be understood if , again , we look at the total energy ( see figure [ fig : m0 von r 2 ] ) .
contrary to the solid sphere , @xmath61 for the hollow sphere converges against the sn value for @xmath63 and compared to the solid - sphere value ( dotted line ) it only decreases very weakly before increasing for radius values above the width of the wave packet .
if one does not take the simpler point of view we took in @xcite and models large molecules by a point - particle sn equation , and instead assumes the modified equation given in @xcite to provide a more realistic description for such molecules , our results confirm that there are indeed deviations between both equations , as jskelinen suggested .
however , whereas one naively might expect that a smaller matter density leads to a weaker gravitational interaction and therefore to a behaviour that is closer to the free schrdinger equation , our results show that the converse is true .
as long as the radius of the solid - sphere potential does not significantly exceed the extent of the wave packet , the effect of inhibitions of dispersion is even increased compared to the sn equation .
the analysis of the total energy given in section [ sec : total energy ] provides an insight why this somewhat peculiar behaviour shows up .
such an increase of the effect can not be observed for the hollow - sphere potential , but there is also no significant decrease , as long as the radius of the hollow sphere does not exceed the extent of the wave packet .
the question why there should be a self - gravitational interaction while there is no electromagnetic self - interaction term present in the schrdinger equation remains an issue that should be addressed ( cf .
the discussion in @xcite ) .
this issue arises even more distinctly in the analysis given in section [ sec : separation ] of this paper , where the gravitational interactions are treated differently to the electromagnetic ones .
we conclude that taking into account the finite extent of the source in the way suggested in @xcite does not lead to an attenuation of the effect of inhibitions of dispersion as discussed in @xcite .
the prospects for an experimental verification of self - gravitation of quantum systems in future molecular interferometry experiments remain unchanged .
we gratefully acknowledge funding and support through the centre for quantum engineering and space - time research ( ) at the leibniz university hannover .
salzman p j and carlip s 2006 a possible experimental test of quantized gravity arxiv : gr - qc/0606120 based on the ph.d .
thesis of salzman : `` investigation of the time dependent schrdinger - newton equation '' , univ . of california at davis , 2005 | we modify the time dependent schrdinger - newton equation by using a potential for a solid sphere suggested by jskelinen ( jskelinen 2012 _ phys . rev .
a _ * 86 * 052105 ) as well as a hollow - sphere potential . compared to our recent paper ( giulini and groardt 2011 _ class . quantum grav .
_ * 28 * 195026 ) where a single point - particle , i.e. a coulomb potential , was considered this has been suggested to be a more realistic model for a molecule .
surprisingly , compared to our previous results , inhibitions of dispersion of a gaussian wave packet occur at even smaller masses for the solid - sphere potential , given that the width of the wave packet is not exceeded by the radius of the sphere . |
fractional brownian motion ( fbm ) is one of the most prominent gaussian processes in the probabilistic and statistical literature .
popularized by mandelbrot and van ness @xcite in 1968 , it found various applications in modeling stochastic phenomena in physics , biology , telecommunication and finance among many other fields .
fractional brownian motion is characterized by its self - similarity property , the stationarity of its increments and by its ability to match any prescribed constant local regularity .
mathematically speaking , for any @xmath0 , a fbm with hurst index @xmath1 , denoted by @xmath2 , is a zero mean gaussian process with the covariance function given by @xmath3 = \frac{1}{2 } \left(t^{2h } + s^{2h } - |t - s|^{2h}\right).\ ] ] various representations of fbm can be found in the existing literature ; we refer to @xcite and references therein .
the hurst parameter @xmath4 determines the path properties of the fbm : ( i ) the process @xmath5 is self - similar with index @xmath1 , i.e. @xmath6 in distribution , ( ii ) @xmath5 has hlder continuous paths of any order strictly smaller than @xmath1 , ( iii ) fractional brownian motion has short memory if an only if @xmath7 $ ] .
moreover , fbm presents long range dependance if @xmath1 belongs to @xmath8 .
the statistical estimation of the hurst parameter @xmath1 in the high frequency setting , i.e. the setting of mesh converging to 0 while the interval length remaining fixed , is often performed by using power variation of @xmath9 . recall that a standard power variation of an auxiliary process @xmath10 on the interval @xmath11 $ ] is defined by @xmath12 } \left|y_{\frac{i+1}{n } } - y_{\frac{i}{n } } \right|^p.\ ] ] this type of approach has been investigated in numerous papers ; we refer to e.g. @xcite among many others .
the fact that most of the properties of fbm are governed by the single parameter @xmath1 restricts its application in some situations .
in particular , its hlder exponent remains the same along all its trajectories .
this does not seem to be adapted to describe adequately natural terrains , for instance .
in addition , long range dependence requires @xmath13 , and thus imposes paths smoother than the ones of brownian motion .
multifractional brownian motion ( mbm ) was introduced to overcome these limitations .
several definitions of multifractional brownian motion exist .
the first ones were proposed in @xcite and @xcite .
a more general approach was introduced in @xcite while the most recent definition of mbm ( which contains all the previous ones ) has been given in @xcite .
the latter definition is both more flexible and retains the essence of this class of gaussian processes .
recall first that a fractional brownian field on @xmath14 noted @xmath15 is a gaussian field such that , for any @xmath1 , the process @xmath16 is a fbm with hurst parameter @xmath1 .
a multifractional brownian motion is simply a `` path '' traced on a fractional brownian field .
more precisely , it has been defined in ( * ? ? ?
* definition1.2 . )
as follows : [ def : field ] let @xmath17 be a deterministic function and @xmath18 be a fractional brownian field .
a multifractional brownian motion ( mbm ) with functional parameter @xmath19 is the gaussian process @xmath20 defined by @xmath21 , for all @xmath22 .
define , for any @xmath23 in @xmath24 , the positive real @xmath25 by setting : @xmath26 where @xmath27 denotes the standard gamma function . for any function @xmath28 , it is easy to verify that the process @xmath29 defined by @xmath30 where @xmath31 denotes a complex gaussian measure ) for more details on gaussian complex measures .
] , is a multifractional brownian motion with functional parameter @xmath19 .
intuitively speaking , the multifractional brownian motion behaves _ locally _ as fractional brownian motion , but the functional parameter @xmath19 is time - varying .
moreover , it remains linked to local regularity of @xmath32 , but in a less simple way than in the case of the fbm .
more precisely , if we assume that @xmath19 belongs to the set @xmath33,{\mathbb{r}})$ ] , for some @xmath34 , and is such that @xmath35 } h(t ) \leq h_{\max}:=\max_{t \in [ 0,1 ] } h(t ) < \min\{1 , \eta\},\ ] ] then @xmath36 is the regularity parameter of @xmath32 ( see ( * ? ? ? * corollaries 1,2 and proposition 10 ) ) . in this setting the functional parameter
@xmath19 needs to be estimated locally in order to get a full understanding of the path properties of the multifractional brownian motion @xmath32 .
bardet and surgailis @xcite have proposed to use a local power variation of higher order filters of increments of @xmath32 to estimate the function @xmath19 .
more specifically , they prove the law of large numbers and a central limit theorem for the local estimator of @xmath19 ( i ) based on log - regression of the local quadratic variation , ( ii ) based on a ratio of local quadratic variations . in this paper
we are aiming at the estimation of the parameter @xmath36 , which represents the regularity ( or smoothness ) of the multifractional brownian motion @xmath37 .
for this particular statistical problem the local estimation approach investigated in @xcite appears to be rather inconvenient .
instead our method relies on a ratio statistic , which compares the _ global _ quadratic variation at two different frequencies .
we remark that in general it is impossible to find a global rate @xmath38 such that the normalized power variation @xmath39 converges to a non - trivial limit .
however , ratios of global power variations can very well be useful for statistical inference .
indeed , we will show that under appropriate conditions on the functional parameter @xmath19 , the convergence @xmath40 then a simple log transformation gives a consistent estimator of the global regularity @xmath36 of a mbm . the paper is structured as follows .
section 2 presents the basic distribution properties of the multifractional brownian motion , reviews the estimation methods from @xcite and states the main asymptotic results of the paper .
proofs are given in section 3 .
in @xcite bardet and surgailis deal with a little bit more general processes than multifractional brownian motions . however , in order not to overload the notations we will focus in this paper on the normalized multifractional brownian motion ( i.e. the mbm defined by ) . from now on we will refer to this process as the multifractional brownian motion and denote it by @xmath41 .
we start with the basic properties of the mbm @xmath32 with functional parameter @xmath19 .
its covariance function is given by the expression @xmath42 = \frac { c^2_{h_{t , s}}}{2c_{h(t)}c_{h(s ) } } \left ( { |t|}^{2h_{t , s } } + { |s|}^{2h_{t , s } } - { |t - s|}^{2h_{t , s } } \right),\ ] ] where @xmath43 and @xmath44 has been defined in .
it is easy to check that @xmath45 is a @xmath46-function .
the local behaviour of the multifractional brownian motion is best understood via the relationship @xmath47 where @xmath48 denotes the convergence of finite dimensional distributions . hence , in the neighbourhood of any @xmath49 in @xmath24 , the mbm @xmath32 behaves as fbm with hurst parameter @xmath50 .
this observation is essential for the local estimation of the functional parameter @xmath19 . in the following
we will briefly review the statistical methods of local inference investigated in bardet and surgailis @xcite , which is based on high frequency observations @xmath51 . while the original paper is investigating rather general gaussian models
whose tangent process is a fractional brownian motion , we will specialize their asymptotic results to the framework of multifractional brownian motion .
let us introduce the generalized increments of a process @xmath52 .
consider a vector of coefficients @xmath53 and a natural number @xmath54 such that @xmath55 in this case the vector @xmath56 is called a filter of order @xmath57 .
the generalised increments of @xmath58 associated with filter @xmath59 at stage @xmath60 are defined as @xmath61 standard examples are @xmath62 , @xmath63 ( first order differences ) and @xmath64 , @xmath65 ( second order differences ) . in both cases
we have that @xmath66 .
now , we set @xmath67 and set @xmath68 , \qquad h \in ( 0,1).\ ] ] the function @xmath69 does not depend on @xmath70 and is strictly increasing on the interval @xmath24 . for any @xmath71 , which determines the local bandwidth ,
the ratio type estimator of @xmath50 is defined as @xmath72 here and throughout the paper we denote @xmath73 for any @xmath74 with @xmath75 .
the authors of @xcite only investigate the estimator @xmath76 relative to the filter @xmath77 , which we assume in this subsection from now on .
the consistency and asymptotic normality of the estimator @xmath76 is summarized in the following theorem .
we remark that the condition for the central limit theorem crucially depends on the interplay between the bandwidth parameter @xmath78 and the hlder index @xmath79 of the function @xmath19 .
[ th1 ] ( ( * ? ? ?
* proposition 3 ) ) assume that @xmath19 belongs to @xmath33)$ ] and that condition is satisfied . *
for any @xmath80 and @xmath71 it holds that @xmath81 * when @xmath82 it holds that @xmath83 where the asymptotic variance @xmath84 is defined in ( * ? ? ?
* eq . ( 2.17 ) ) .
the paper @xcite contains the asymptotic theory for a variety of other local estimators of @xmath50 .
we dispense with the detailed exposition of these estimators , since only @xmath76 is somewhat related to our estimation method .
[ rem1 ] nowadays , it is a standard procedure to consider higher order filters for gaussian processes to obtain a central limit theorem for the whole range of hurst parameters .
let us shortly recall some classical asymptotic results , which are usually referred to as breuer - major central limit theorems .
we consider the scaled power variation of a fractional brownian motion @xmath9 with hurst parameter @xmath4 based on first order filter @xmath85 and second order filter @xmath86 : @xmath87 it is well known that , after an appropriate normalization , the statistic @xmath88 exhibits asymptotic normality for @xmath89 $ ] , while it converges to the rosenblatt distribution for @xmath90 . on the other hand , the statistic @xmath91 exhibits asymptotic normality for all @xmath4 .
we refer to @xcite for a detailed exposition .
in this section we will construct a consistent estimator of the global regularity parameter @xmath36 , which has been defined at .
our first condition is on the set @xmath92 , which is necessarily compact since @xmath19 belongs to @xmath33)$ ] .
we assume that this set has the following form @xmath93 \right ) \bigcup \ \left(\overset{m}{\underset{j=1}{\bigcup } } \{x_{j}\}\right ) , \qquad ( q , m ) \in { \mathbb{n}}^2 \setminus ( 0,0),\end{aligned}\ ] ] where @xmath94 and the intervals @xmath95 $ ] are disjoint and such that none of the @xmath96 s belongs to @xmath97 $ ] . depending on whether @xmath98 or @xmath99 , we will need an additional assumption .
below , we denote by @xmath100 ( resp .
@xmath101 ) the @xmath102th left ( resp .
right ) derivative of @xmath19 at point @xmath23 .
+ + + ( @xmath103 ) there exist positive integers @xmath104 such that function @xmath19 is @xmath104 times continuously left and right differentiable at point @xmath105 for @xmath106 such that @xmath107 we remark that since @xmath19 reaches its minimum at points @xmath105 , we necessarily have that @xmath108 and that @xmath109 if @xmath104 is even and @xmath110 if @xmath102 is odd .
now , we proceed with the construction of the consistent estimator of the global regularity parameter @xmath36 based on high frequency observations @xmath111 .
first of all , let us remark that considering the estimator @xmath112 } \widehat{h}_t^{n,\alpha}$ ] , where @xmath76 has been introduced in the previous section , is not a trivial matter since the functional version of theorem [ th1 ] is not available .
instead our statistics relies on the global quadratic variation rather than local estimates . for the mbm @xmath113}$ ] , we introduce the notations @xmath114 our first result determines the limit of @xmath115 / { \mathbb{e}}[v(b^{h};2)^n]$ ] .
[ firstresult ] let @xmath116\rightarrow ( 0,1)$ ] be a deterministic @xmath33)$]-function satisfying and such that the set @xmath117 has the form .
if @xmath99 we also assume that condition ( @xmath103 ) holds . define @xmath118}{{\mathbb{e}}[v(b^{h};2)^n]}.\ ] ] then it holds that @xmath119 the convergence result of proposition [ firstresult ] is rather intuitive when @xmath98 , which means that the minimum of the function @xmath19 is reached on a set of positive lebesgue measure . in this setting it is quite obvious that the statistic @xmath120 is dominated by squared increments @xmath121 for @xmath122 $ ] .
thus , the estimation problem is similar to the estimation of the hurst parameter of a fractional brownian motion @xmath123}$ ] with hurst parameter @xmath124 , for which the convergence at is well known .
when @xmath99 , and hence @xmath125 , the proof of proposition [ firstresult ] becomes much more delicate .
assume for illustration purpose that @xmath99 , @xmath126 , @xmath127 and @xmath128 .
condition ( @xmath103 ) is crucial to determine the precise asymptotic expansion of the quantity @xmath129 $ ] .
the lower and upper bounds in and in the proof show that @xmath130= \text{o}\left ( \frac{n^{1 - 2h_{\min}}}{(\ln n)^{1/p}}\right ) \quad \text{as } n\to + \infty ,
\qquad \text{for } k=1,2.\ ] ] the condition @xmath131 of assumption ( @xmath103 ) is not essential for the proofs . for instance ,
when @xmath132 the expectation @xmath129 $ ] would be dominated by the terms in the small neighbourhood on the right hand side of @xmath23 and the statement of proposition [ firstresult ] can be proved in the same manner .
our main result shows that the statistic @xmath133 and the quantity @xmath134 are asymptotically equivalent in probability .
[ th2 ] assume that @xmath135)$ ] and the set @xmath117 has the form . if @xmath99 we also assume that condition ( @xmath103 ) holds
. then we have the following result : @xmath136 in particular , the following convergence holds : @xmath137 the asymptotic result of theorem [ th2 ] can be extended to more general gaussian processes than the mere multifractional brownian motion .
as it has been discussed in @xcite , when a gaussian process possesses a tangent process @xmath138 at time @xmath49 , we may expect theorem [ th2 ] to hold under certain assumptions on its covariance kernel .
we refer to assumptions @xmath139 and @xmath140 therein for more details on sufficient conditions .
the rate of convergence , or a weak limit theorem , associated with the consistency result at is a more delicate issue . in the setting @xmath99 , which implies that @xmath125 ,
the bias associated with the convergence in may very well dominate the variance of the estimator .
a careful inspection of our proof , and more specifically of statement and remark [ ofjerifrofjerfioerj ] , implies that the bias in proposition [ firstresult ] has a logarithmic rate .
thus , weak limit theorems for the estimator @xmath141 are out of reach in this framework .
when @xmath142 and hence @xmath143 , one may hope to find better rates of convergence for the estimator @xmath141 .
however , we dispense with the exact exposition of this statistical problem .
throughout this section we denote all positive constants by @xmath144 , or @xmath145 if they depend on an external parameter @xmath102 , although they may change from line to line . for @xmath146
we introduce the notation @xmath147 which serves as the first order approximation of the quantity @xmath129 $ ] . applying ( * ? ? ?
* lemma 1 p.@xmath148 ) we conclude that @xmath149 - v_n^{(k ) } \right| \leq c \frac{\ln n}{n^{\eta\wedge 1 } } \sum_{i=0}^{n - k } \left(\frac{i}{n } \right)^{2h(k / n ) } \leq c \frac{\ln n}{n^{2h_{\min } - 1 + \eta\wedge 1}}\ ] ] for any @xmath150 .
we have the inequality @xmath151 - v_n^{(1 ) } | + |{\mathbb{e}}[v(b^{h};2)^n ] - v_n^{(2 ) } |}{v_n^{(2 ) } } + \left| \frac{v_n^{(1 ) } } { v_n^{(2 ) } } - \left(\frac{1}{2}\right)^{2h_{\min } } \right| \nonumber \\[1.5 ex ] \label{def1 } & = : \mu^{(1)}_{n } + \mu^{(2)}_{n}. \end{aligned}\ ] ] we first show that @xmath152 as @xmath153 . when @xmath154 we trivially have @xmath155 . if @xmath156 , we fix @xmath157 . by @xmath158 we denote the lebesgue measure of any measurable set @xmath159 .
we have that @xmath160)\right)>0.\ ] ] thus , there exists @xmath161 such that for all @xmath162 it holds that @xmath163 \ } \geq n\ \text{leb}\left(h^{-1}([h_{\min } , h_{\min } + \epsilon])\right)/2.\ ] ] this implies that @xmath164 } \left(\frac{2}{n } \right)^{2h(i / n ) } \geq c n^{1 - 2(h_{\min } + \epsilon)}.\end{aligned}\ ] ] hence , applying inequality , we conclude that : @xmath165 which proves that @xmath166 , for any @xmath167 small enough .
we first prove that @xmath169 in the case @xmath98 .
assume again that @xmath156 .
first , we observe the lower bound @xmath170 } \left(\frac{k}{n } \right)^{2h(i / n ) } = \left ( \frac{k}{n}\right)^{2h_{\min } } \sum_{l=1}^q { \operatorname{card}}\{i\in \llbracket 0,n - k\rrbracket ; ~ i / n \in [ a_l , b_l ] \ } \nonumber \\[1.5 ex ] \label{lowercase1 } & \geq n \left ( \frac{k}{n}\right)^{2h_{\min } } \sum_{l=1}^q \left(b_l - a_l-\tfrac{2}{n } \right ) . \end{aligned}\ ] ] for the upper bound we fix @xmath171 and consider the decomposition @xmath172 } \left(\frac{k}{n } \right)^{2h(i / n ) } + \sum_{i\in { \llbracket 0,n - k\rrbracket } ; ~h(i / n ) \not \in [ h_{\min } , h_{\min } + \epsilon ] } \left(\frac{k}{n } \right)^{2h(i / n)}.\end{aligned}\ ] ] setting @xmath173\}$ ] , we deduce the assertions @xmath174)\right ) \quad \text{as } n \to \infty , \\[1.5 ex ] \text{leb}\left(h^{-1}([h_{\min } , h_{\min } + \epsilon])\right ) & \to \text{leb}\left(h^{-1}(\{h_{\min}\})\right ) = \sum_{l=1}^q ( b_l - a_l)>0 \quad \text{as } \epsilon \to 0 . \end{aligned}\ ] ] now , we conclude that @xmath175 throughout the proofs we write @xmath176 for @xmath177 and @xmath178 for @xmath179 .
applying inequalities and , we obtain that @xmath180 hence , we deduce that @xmath181)\right ) } \leq { \underset{{n \to + \infty}}{\underline{\lim}}}\frac{v_n^{(1 ) } } { v_n^{(2 ) } } \leq { { \displaystyle}\underset{{n \to + \infty}}{\overline{\lim}}}\frac{v_n^{(1 ) } } { v_n^{(2 ) } } & \leq \frac{2^{-2h_{\min}}\text{leb}\left(h^{-1}([h_{\min } , h_{\min } + \epsilon])\right ) } { \text{leb}\left(h^{-1}(\{h_{\min}\})\right)}. \end{aligned}\ ] ] by letting @xmath167 tend to @xmath182 , we readily deduce taht @xmath169 as @xmath183 . without loss of generality
we assume that @xmath126 and @xmath185 with @xmath186 .
recall that in this setting we assume condition ( @xmath103 ) with @xmath128 .
we let @xmath187 be a positive number such that @xmath188 .
now , there exists a @xmath189 with @xmath190 such that : @xmath191 & \hspace{-17ex}h_{\min } + \frac{1}{p ! } ( y - x)^{p}\ ( h^{(p)}_{r}(x ) - \gamma ) \leq h(y ) \leq h_{\min } + \frac{1}{p ! } ( y - x)^{p } \
( h^{(p)}_{r}(x ) + \gamma),\end{aligned}\ ] ] @xmath192 & h_{\min } + \frac{1}{p ! }
( y - x)^{p}\ ( h^{(p)}_{l}(x ) -(-1)^{p } \gamma ) \leq h(y ) \leq h_{\min } + \frac{1}{p ! } ( y - x)^{p}\ ( h^{(p)}_{l}(x ) + ( -1)^{p } \gamma).\end{aligned}\ ] ] we proceed with the derivation of upper and lower bounds for the quantity @xmath168 .
we start with the decomposition @xmath193 where @xmath194 } \left(\frac{k}{n } \right)^{2h(i / n)};\\ \gamma^{(2)}_{n , k}(\gamma , \epsilon ) & : = \sum_{i\in \llbracket 0,n - k\rrbracket ; ~i / n \in [ x-\epsilon , x ) } \left(\frac{k}{n } \right)^{2h(i /
n ) } ; \\ { \displaystyle}\gamma^{(3)}_{n , k}(\gamma , \epsilon ) & : = \sum_{i\in \llbracket 0,n - k\rrbracket ; ~i / n \in [ x-\epsilon , x+\epsilon]^c } \left(\frac{k}{n } \right)^{2h(i / n)}.\end{aligned}\ ] ] it is clear that @xmath195 , where we have set @xmath196 \}.\ ] ] for the other two quantities , we deduce that @xmath197 with @xmath198 } \left(\frac{k}{n } \right)^{2 ( p!)^{-1 } ( i / n - x)^{p}(h^{(p)}_{r}(x ) + \gamma ) } , \\[1.5 ex ] \underline{\gamma}^{(2)}_{n , k}(\gamma , \epsilon ) & : = \left(\frac{k}{n}\right)^ { 2h_{\min } } \sum_{i\in \llbracket 0,n - k\rrbracket : ~i / n \in [ x-\epsilon , x ) } \left(\frac{k}{n } \right)^{2 ( p!)^{-1 } ( i / n - x)^{p}(h^{(p)}_{l}(x ) + ( -1)^{p } \gamma)}\end{aligned}\ ] ] and @xmath199 and @xmath200 . using and , it is easy to see that , for every @xmath201 : @xmath202 with @xmath203 from we obtain that @xmath204 where @xmath205 \underline{\delta}_{n , k}(\gamma , \epsilon)&:=\underline{\gamma}^{(1)}_{n , k}(\gamma , \epsilon ) + \underline{\gamma}^{(2)}_{n , k}(\gamma , \epsilon ) , \qquad \overline{\delta}_{n , k}(\gamma , \epsilon):=\underline{\delta}_{n , k}(-\gamma , \epsilon ) .
\end{aligned}\ ] ] in view of it is sufficient to show that @xmath206 .
define @xmath207 for any @xmath208 in @xmath209 , we also set @xmath210 } \left(\frac{k}{n } \right)^{a ( i / n - x)^{p } } , \\ \label{edpokezpodk2 } t_{n , k}(b,\epsilon)&:= \sum_{i\in \llbracket 0,n - k\rrbracket : ~i / n \in [ x-\epsilon , x ) } \left(\frac{k}{n } \right)^{b ( i / n - x)^{p}}.\end{aligned}\ ] ] we deduce the identities @xmath211 and @xmath212 .
note moreover that @xmath213 when @xmath102 is even and @xmath214 when @xmath102 is odd .
we therefore assume from now on that @xmath215 when @xmath102 is even and that @xmath216 when @xmath102 is odd . for any @xmath217
, we define @xmath218 since @xmath219 is decreasing on @xmath220 + 1 , [ n(x+\epsilon ) ] \rrbracket$ ] while @xmath221 is increasing if @xmath102 even ( resp . decreasing if @xmath102 odd ) on @xmath222 + 1,[nx ]
\rrbracket$ ] , one can use an integral test for convergence , which provides us with the following upper bounds @xmath223 \label{a2 } \frac{n\left(\int^{\beta'_{n}(b)}_{\alpha'_{n}(b ) } y^{1/p-1 } e^{-y } \
dy-\rho^{(b)}_{n , k}(\epsilon ) \right)}{p ( ( -1)^{p } b \ln(n / k))^{1/p } } & \leq t_{n , k}(b,\epsilon ) \leq \frac{n\left(\int^{\mu'_{n}(b)}_{\tau'_{n}(b ) } y^{1/p-1 } e^{-y } \ dy-\rho^{(b)}_{n ,
k}(\epsilon ) \right)}{p ( ( -1)^{p } b\ln(n / k))^{1/p}}.\end{aligned}\ ] ] here we use the notation @xmath224 + 1}{n}-x\bigg)^{p } , \quad \beta_{n}(a):= a \ln(n / k ) \bigg(\frac{[n(x+\epsilon)]+1}{n}-x\bigg)^{p},\\ \tau_{n}(a)&:= a \ln(n / k ) \bigg(\frac{[nx]}{n}-x\bigg)^{p } , \quad \mu_{n}(a):= a \ln(n / k ) \bigg(\frac{[n(x+\epsilon)]}{n}-x\bigg)^{p}\end{aligned}\ ] ] and @xmath225 + 1}{n})+f^{(b)}_{n , k}(\frac{[nx]}{n})$ ] . furthermore , @xmath226 where we have set @xmath227 - 2}{n}-x\bigg)^{p } , \quad z^{(2)}_{n}(b):= b \ln(n / k ) \bigg(\frac{[n(x- \epsilon)]+1}{n}-x\bigg)^{p},\\ z^{(3)}_{n}(b)&:= b \ln(n
/ k ) \bigg(\frac{[nx]-1}{n}-x\bigg)^{p } , \quad z^{(4)}_{n}(b):= b \ln(n / k ) \bigg(\frac{[n(x- \epsilon)]+2}{n}-x\bigg)^{p}.\end{aligned}\ ] ] in view of the inequalities and , as well as identities and , we then deduce that @xmath228 here we have used the notation @xmath229 since @xmath230 , and also provide us with upper and lower bounds for @xmath231 . finally , we obtain the following lower and upper bounds @xmath232 where @xmath233 denote @xmath234 . recalling the definition of the constants @xmath235 and @xmath236 , a straightforward computation shows that , for any @xmath237 with @xmath238 : @xmath239 \label{zoiejozedijezoidjzeodijzeodijzzz2 } & { { \displaystyle}\underset{{n \to + \infty}}{\overline{\lim}}}|\lambda'_{n , k'}(\gamma , \epsilon)- \lambda_{n , k}(-\gamma , \epsilon)| \leq c\ ( 2 |\gamma| + |1/d_{-\gamma}-1/d_{\gamma } + 1/|d'_{-\gamma}| - 1/|d'_{\gamma}||)\leq c |\gamma|.\end{aligned}\ ] ] starting from , and using and , we see that there exists a positive integer @xmath240 and @xmath241 such that for all @xmath242 @xmath243 finally , inequalities , and imply that there exists a positive integer @xmath244 such that for all @xmath245 : @xmath246 from the previous inequality , @xmath247 and thus we get @xmath248 , which completes the proof . in the first step we will find an upper bound for the covariance function of the increments of @xmath250 .
we define @xmath251 recalling the notation at , we conclude the identity @xmath252 since @xmath253)$ ] and the function @xmath254 defined at is a @xmath46-function , we deduce by an application of taylor expansion @xmath255 denotes the second order derivative in the direction of @xmath256 and @xmath257 , and @xmath258 .
now , we will compute an upper bound for the right side of for @xmath259 .
first , we observe that @xmath260 where @xmath261 we remark that @xmath262 is the covariance kernel of the fractional brownian motion with hurst parameter @xmath4 . since @xmath253)$ ] , @xmath263 and @xmath264 for @xmath265
, we conclude that @xmath266 ^ 2.\ ] ] we concentrate on the second order derivative @xmath267 ; the estimates for the other second order derivatives are obtained similarly .
we have that @xmath268 & + 2\partial_{1 } f(t , s ) \left [ \partial_{1 } g(t , s , h(t)+h(s ) ) + h'(t ) \cdot \partial_{3 } g(t , s , h(t)+h(s))\right ] \\[1.5 ex ] & + f(t , s ) \left [ \partial_{11 } g(t , s , h(t)+h(s ) ) + 2 h'(t ) \cdot \partial_{13 } g(t , s , h(t)+h(s ) ) \right ) .
\\[1.5 ex ] & \hspace{1.75cm}+\left .
h''(t ) \cdot \partial_{3 } g(t , s , h(t)+h(s ) ) + { ( h'(t))}^{2}\cdot \partial_{33 } g(t , s , h(t)+h(s ) ) ) \right].\end{aligned}\ ] ] for the derivatives of the function @xmath269 , we deduce the following estimates @xmath270 + \ln^2 |t - s|\cdot |t - s|^{h(t)+h(s ) } \right),\end{aligned}\ ] ] which hold for @xmath271 $ ] with @xmath272 and
the third inequality holds whenever @xmath273 ( if @xmath274 we simply have @xmath275 ) .
similar formulas and bounds are obtained for other second order derivatives of @xmath276 .
using the boundedness of functions @xmath277 , @xmath19 and its derivatives , together with the above estimates and we obtain the inequality @xmath278 & \leq c n^{-2h_{\min } } \left ( i^{2h_{\min } -2 } + j^{2h_{\min } -2 } + |i - j|^{2h_{\min } -2 }
\right ) , \qquad i , j \geq 1 , |i - j| > 2 .
\nonumber\end{aligned}\ ] ] when @xmath279 we deduce from ( * ? ? ?
* lemma 1 p.@xmath148 ) that @xmath280 we recall the identity @xmath281 for a gaussian vector @xmath282 . by
and we immediately conclude that @xmath283 observing the decomposition @xmath284}{v(b^h;2)^n } - { \mathscr{u}^{\hspace{-0.2ex}h}_{n}}\frac{v(b^h;2)^n - { \mathbb{e}}[v(b^h;2)^n]}{v(b^h;2)^n } \\[1.5 ex ] & + \left({\mathscr{u}^{\hspace{-0.2ex}h}_{n}}- \left ( \frac 12 \right)^{2h_{\min}}\right)\end{aligned}\ ] ] and in view of proposition [ firstresult ] , it is sufficient to show that @xmath285 } \to 0 , \qquad k , l=1,2\end{aligned}\ ] ] to prove theorem [ th2 ] .
we assume again without loss of generality that @xmath99 , @xmath126 and @xmath286 .
using the notations from the previous subsection together with the inequalities and , we deduce the following lower bound , for @xmath70 large enough and for @xmath167 small enough : @xmath287 & \geq \underline{\gamma}^{(1)}_{n , k}(\gamma , \epsilon ) + \underline{\gamma}^{(2)}_{n , k}(\gamma , \epsilon ) \geq c_{\epsilon } \frac{n^{1 - 2h_{\min } } } { ( \ln n)^{1/p}}. \end{aligned}\ ] ] thus , in view of we readily deduce the convergence at for any @xmath288 , which completes the proof of theorem [ th2 ] .
a. ayache , s. cohen , and j. lvy vhel .
the covariance structure of multifractional brownian motion , with application to long range dependence ( extended version ) .
, refereed conference contribution , 2000 . | this paper presents a new estimator of the global regularity index of a multifractional brownian motion .
our estimation method is based upon a ratio statistic , which compares the realized global quadratic variation of a multifractional brownian motion at two different frequencies .
we show that a logarithmic transformation of this statistic converges in probability to the minimum of the hurst function , which is , under weak assumptions , identical to the global regularity index of the path .
* ams subject classification : * 60g15 ; 60g22 62g05 ; 62m09 ; 60g17 |
ultra - relativistic collisions of heavy nuclei are believed to produce nuclear matter at extreme pressure and energy density .
quantum chromodynamics ( qcd ) predicts that a deconfined state of partonic matter which is referred as the quark - gluon plasma ( qgp ) , will be produced in such circumstances .
one of the most striking signatures of qgp formation is the suppression of heavy quarkonia states such as the charmonia ( @xmath3 etc ) in the deconfined phase . till mid-2000
, debye screening was thought to be the only possible mechanism for the anomalous quarkonia suppression in qgp but recent work by laine and collaborators have shown that the effective potential between heavy quarks acquires an imaginary part as well @xcite .
this imaginary part generates the collisional dissociation of bound states by the qgp @xcite .
however , in this paper , we emphasize the debye screening as the dominant mechanism for the quarkonia suppression in the qgp medium .
charmonia suppressions in heavy - ion collisions have experimentally been studied , first at the cern super proton synchrotron ( sps ) by na50 @xcite and na60 experiment @xcite at @xmath4 gev and then at relativistic heavy ion collider ( rhic ) by the phenix experiment at @xmath5 and @xmath6 gev @xcite , and now at large hadron collider ( lhc ) by the cms and alice experiments at @xmath7 tev @xcite . however ,
experimental results involve more puzzling features which defy explanations on the basis of colour - screening in the qgp alone .
for example , we notice significantly less suppression at the mid - rapidity than at forward rapidity ( @xmath8 ) while the medium in the central part of the collision is the most dense and intutively one should expect an opposite behaviour .
another interesting experimental observation is similar @xmath0 suppression at sps and rhic energies for the same number of participants @xcite .
many theoretical efforts have since been made which explore the effects of cold nulear matter ( cnm ) as well as possible coalescence of originally uncorrelated @xmath9 , @xmath10 quark pairs .
however , a complete theoretical interpretation for @xmath0 suppression is still lacking .
many theoretical papers have analyzed the heavy - ion experimental data regarding nuclear modification factor for heavy quarkonia .
for example , @xmath11 data at sps are explained by variety of models @xcite with or without deconfinement phase transition scenario .
recently yunpeng et al .
, @xcite analyzed the @xmath0 suppression data using a transport model calculation involving suppression as well as the regeneration of @xmath0 due to coalescence of @xmath9 and @xmath10 pair produced abundantly in the initial stage of collisions at rhic and lhc energies .
zhen qu et al .
, @xcite used a similar approach to explain the forward rapidity data for @xmath0 at rhic energy .
sharma and vitev @xcite calculated the yields of quarkonia at rhic and lhc as a function of transverse momentum based on heavy quark effective theory in which both colour singlet and colour octet contribituions along with feed down effects from excited states were suitably incorporated .
their model provides good description of the lhc data for central as well as peripheral collisions . but none of the available models could provide a satisfactory and consistent explanation for the suppression at all energies simultaneously .
we have recently modified @xcite the colour - screening model of chu and matsui @xcite by parametrizing pressure instead of energy density since pressure density becomes almost zero at the deconfinment phase transition point .
we have further shown that the variation of survival probability of @xmath0 with respect to participant number obtained from this model compares well with the experimental data ( cnm normalized ) at rhic @xcite . in the present paper , we reinvestigate the colour - screening model by employing a more realistic thermodynamically - consistent quasi - particle prescription for the equation of state ( eos ) of qgp medium .
this highlights a major difference between our present approach and the model of chu , matsui @xcite or mishra et al @xcite because bag model is often considered as a crude eos for qgp .
we have recently demonstrated that various thermodynamical as well as transport quantities obtained from the present quasiparticle model ( qpm ) compare well with the most recent available lattice results @xcite .
in addition , we incorporate here viscous hydrodynamics in our formulation with a constraint that the ratio @xmath12 varies very slowly @xcite with the temperature in the qgp medium .
this assumption helps us in taking @xmath13 to be almost independent of the temperature . feed down from higher resonances
have also been incorporated in the model .
we use our formulation to explain the recent lhc experimental data on @xmath0 suppression along with the rhic and sps data .
we surprisingly find that @xmath0 suppression at mid - rapidity is explained consistentely by our model at sps , rhic and lhc energies in qgp picture alone .
furthermore , we compare results obtained from quasiparticle model ( qpm ) with those from the bag model eos on @xmath0 suppression and find that qpm eos describes the data better and thus lends credibility to the present approach in explaining the @xmath0 suppression .
the main theme of the present formulation is essentially borrowed from our previous model @xcite where we have used the colour screening idea of chu and matsui @xcite .
however , instead of using bag model for qgp , we now use quasiparticle model ( qpm ) as new eos of qgp .
we have assumed that the qgp medium formed during the collision , expands and cools according to the bjorken s boost invariant longitudinal hydrodynamics in mid - rapidity region . employing the conservation of energy - momentum tensor , the rate of the decrease of energy density @xmath14 @xcite is given by @xmath15 where @xmath13 is the shear viscosity of the qgp medium , @xmath16 is the pressure and @xmath17 represents the proper time .
the energy density and pressure are computed by using qpm eos @xcite for qgp . using eq.(1 ) and
the thermodynamical identity @xmath18 , the cooling laws for energy density and pressure in the qpm model can be separately given as ( see appendix a for derivation ) : @xmath19 @xmath20 where @xmath21 , @xmath22 and @xmath23 are constants which can be determined by imposing the initial boundary conditions on energy density and pressure , @xmath24 with @xmath25 as the speed of sound in the medium .
we take @xmath26 at @xmath27 ( initial thermalization time ) and also @xmath28 at @xmath29 ; where @xmath30 is the proper time . consequently , the constants @xmath21 and @xmath22 are related as : @xmath31 , where @xmath32 , @xmath33 and @xmath34 is the reynold s number for qgp .
further : @xmath35 using the other initial condition for @xmath36 at @xmath27 , gives the value of @xmath23 as follows : @xmath37 assuming that the pressure almost vanishes at the transition temperature @xmath38 , we take a pressure profile function in the transverse plane with a transverse distance @xmath39 as @xcite : @xmath40 where the coefficient @xmath41 is yet to be determined , @xmath42 denotes the radius of the cylinderical plasma and it is related to the transverse overlap area @xmath43 as determined by glauber model @xmath44 @xcite . the pressure is thus assumed to be maximum at the central axis but it vanishes at the edge @xmath45 where hadronization first begins .
the exponent @xmath46 depends on the energy deposition mechanism and here we have taken @xmath47 @xcite ; @xmath48 is the unit step - function .
the factor @xmath41 is related to the average initial pressure @xmath49 via @xmath50 the average pressure is determined by the centrality dependent initial average energy density @xmath51 which is further given by bjorken s formula @xcite : @xmath52 here @xmath53 is the transverse energy deposited per unit rapidity .
we use the experimental value of @xmath54 where @xmath55 is the pseudorapidity and then multiply it by a corresponding jacobian factor @xcite in order to obtain @xmath56 for a given number of participants ( @xmath57 ) at a particular center - of - mass energy ( @xmath58 ) . at the initial proper time , @xmath59 in qpm eos @xcite and thus @xmath60 .
since the cooling law for pressure can not be analytically solved for @xmath17 and , therefore , we use a trick to determine the radius of screening region . writing the cooling law of pressure as follows :
@xmath61 where @xmath62 , @xmath63 , @xmath64 and @xmath65 are constants related to @xmath21 , @xmath22 and @xmath23 as : @xmath66 , @xmath67 , @xmath68 and @xmath69 . writing the above equation at the initial time @xmath70 and at screening time @xmath71 we get : @xmath72 and @xmath73 here @xmath74 is the qgp pressure as determined by eos in qpm @xcite .
solving eqs . ( 11 ) and ( 12 ) numerically and equating the screening time @xmath75 to the dilated formation time of quarkonia @xmath76 ( = @xmath77 where @xmath78 is the lorentz factor associated with the transverse motion of the @xmath79 pair ,
@xmath80 gev and @xmath81 is the proper time required for @xmath79 pair in the formation of @xmath0 ) , we can find the radius of the screening region @xmath82 .
the screening region is defined as a region where temperature is more than the dissociation temperature so that the quarkonia formation becomes unlikely inside that region @xcite .
hence the pair will in all probability escape the screening region and form quarkonia only if @xmath83 where @xmath84 is the position vector at which the charm - quark pair is created .
the above kinematic condition takes a simplified form by assuming that @xmath0 is moving with transverse momentum @xmath85 .
thus the above escape condition can be expressed as a trigonometric condition @xcite : @xmath86 where @xmath87 is the angle between the transverse momentum ( @xmath88 ) and the position vector @xmath84 and @xmath89 with @xmath90 . assuming the radial probability distribution for the production of @xmath91 pair in hard collisions at transverse distance @xmath39 as @xmath92 here we take @xmath93 in our calculation as used in ref .
then , in the colour screening scenario , the survival probability for the quarkonia can easily be obtained as @xcite : @xmath94 where the maximum positive angle @xmath95 allowed by eq .
( 13 ) becomes @xcite : @xmath96 generally experimentalists measure the quantity namely @xmath85 integrated nuclear modification factor and , therefore , the theoretical @xmath85 integrated survival probability is of paramount importance in comparing with the experimental results .
it is defined as follows : @xmath97 it has been found that only about @xmath98 of the observed @xmath0 come from hard collisions whereas @xmath99 is from the decay of @xmath1 and remaining @xmath100 from the @xmath2 @xcite .
therefore , the net survival probability of @xmath0 in the presence of qgp medium is : @xmath101 .values of the parameters . [
cols="<,<,<,<,<,<",options="header " , ] in our calculation , we use @xmath102 gev in accordance with the recent lattice qcd results @xcite . similarly initial tempertaure ( @xmath103 ) , pressure density ( @xmath104 ) , entropy density ( @xmath105 ) at proper time @xmath106 along with @xmath107 and @xmath46 at different energies are tabulated in table .
the value of @xmath103 , @xmath104 and @xmath105 are taken in accordance with our qpm results @xcite .
other parameters like masses ( @xmath108 ) , formation time ( @xmath81 ) and two different sets of dissociation temperatures ( @xmath109 ) labelled as set i and set ii for different quarkonia states are given in table .
2 @xcite .
the reason behind the use of two different sets of dissociation temperature will be discussed later .
integrated survival probability of @xmath0 at sps energy , ( @xmath110 ) with respect to @xmath57 in our present model using two different sets of dissociation temperatures ( as given in table .
2 ) . cnm normalized experimental data points are taken from ref .
@xcite.,title="fig : " ] [ fig : fig 2 ] ) with respect to @xmath57 .
cnm normalized experimental data points are taken from ref .
@xcite.,title="fig : " ] [ fig : fig 3 ] integrated survival probability of @xmath0 at lhc energy , ( @xmath111 ) with respect to @xmath57 .
experimental data points are @xmath112 for prompt @xmath0 @xcite normalized by contribution due to cnm effect @xcite.,title="fig : " ] [ fig : fig 4 ] integrated survival probability of @xmath0 at rhic energy , ( @xmath113 ) with respect to @xmath57 in our new model and a comparison with the old model result in which we use bag model as qgp eos along with ideal hydrodynamics .
experimental data points are taken from ref .
@xcite.,title="fig : " ] [ fig : fig 1 ] integrated survival probability of @xmath0 at rhic and lhc energies with respect to @xmath57 calculated within our new model with dissociation temperature of set ii in viscous as well as in ideal hydrodynamics framework.,title="fig : " ] [ fig : fig 5 ]
in some recent lattice calculations @xcite , after constant mode correction , there is no hint for any suppression of @xmath114 and @xmath2 instantly after the transition temperature @xmath115 as pointed out by different studies @xcite .
moreover , a recent calculation based on qcd sum rules shows that @xmath0 quickly dissociates into the continuum and almost disappears completely at temperatures between @xmath116 and @xmath117 @xcite . due to above uncertainty regarding the dissociation temperatures of different quarkonia states , we wish to see the effect of @xmath118 in simultaneous explaining the data of rhic , sps and lhc experiments .
thus we take two different sets of dissociation temperatures for quarkonia states ( labelled as set i and set ii as given in table .
2 ) . fig .
1 presents the variation of @xmath85 integrated survival probability @xmath110 with @xmath57 obtained from our present model for the two sets ( labelled as set i and set ii ) and we have shown comparison with the cnm normalized experimental data @xcite .
moreover , one should keep in mind that from here onwards , we use the same @xmath88 range in all our calculations as used in the related experimental data .
we observe that our new model with set ii yields a reasonable agreement with the experimental data in comparison to the results obtained by the model with set i. furthermore , we find that the changes in @xmath118 values produce more impact on the peripheral collisions rather than on central collisions .
2 demonstrates the variation of @xmath113 with respect to @xmath57 as obtained from our model with set i and set ii , and compare our results with the experimental data again normalized with cnm effect @xcite .
we again observe that the model with dissociation temperatures given by set ii compare well with the data in comparison to the model with set i. at rhic energy the difference in the results of two models is significant because when we increase the dissociation temperatures for @xmath114 and @xmath119 , the energy deposited in the collision at rhic energy is still not sufficient to melt all the @xmath114s and @xmath119s even in the most central collisions and thus they still provide a significant contribution to the survived @xmath0s .
3 shows the variation of @xmath111 with respect to @xmath57 obtained in our model with set i and ii at lhc energy . here
we find that there is almost no difference in the survival probability calculated with set i or ii for all collision centralities .
we get lesser suppression in @xmath0 at lhc in comparison to rhic .
however , one should keep in mind the integrated @xmath88 range here . for lhc , we include quarkonia having @xmath88 upto @xmath120 gev .
we also present the recent experimental data of nuclear modification factor ( @xmath121 ) of prompt @xmath0 at mid - rapidity ( @xmath122 ) obtained by cms collaboration @xcite and normalized by cold nuclear matter ( cnm ) contribution at lhc @xcite .
the quantitative as well as qualitative agreement of our model results with the cnm normalized experimental data appears quite good .
4 shows the variations of @xmath85 integrated survival probability of @xmath0 at rhic energy ( i.e. @xmath123 ) with respect to @xmath57 .
old model represents the results where a bag model eos for qgp along with ideal hydrodynamics @xcite is employed to calculate the @xmath85 integrated survival probability with two sets ( i.e. , set i and set ii ) of dissociation temperatures .
blank square and blank triangular symbols represent the results obtained in our present model with quasi - particle model eos for qgp medium together with the tool of viscous hydrodynamics .
the comparison shows a significant effect of the eos of qgp used in the calculation of survival probability for @xmath0 .
we have also plotted the experimental data @xcite ( as shown by solid circles in fig .
2 ) normalized with cnm effect @xcite .
we find the present model results employing qpm eos for qgp along with viscous hydrodynamics give a reasonably good agreement with the experimental data . again the present model with set ii describes the data more suitably in comparison to all other calculations .
this show the importance of qpm eos for qgp over bag model .
furthermore , in the bag model equation of state @xcite we have used an additional arbitrary parameter @xmath124 in the bjorken s formula for initial energy density .
its value @xmath125 was chosen in such a way that the resulting initial energy density becomes equal to what has been predicted by the self - screened parton cascade model @xcite .
however , in the present model using qpm as eos , the energy density does not require any such parameter . in fig . 5
, we present our new model results for @xmath111 as well as @xmath113 with @xmath57 and also show a comparison with results obtained in our present model when we put shear viscosity ( @xmath13 ) equal to zero ( i.e. , considering ideal hydrodynamic framework ) .
we find that the results in the two cases ( viscous and ideal ) almost overlap on each other for the case of peripheral collisions .
however , @xmath85 integrated suvival probability as calculated from ideal hydrodynamics yields quite significant difference at larger @xmath57 in comparison to the results obtained from viscous hydrodynamics .
this difference becomes more significant at higher energy i.e. , at lhc in comparison to rhic energy .
thus , our study suggests that inclusion of viscous effect in the hydrodynamic evolution is also essential for a complete description of @xmath0 suppression .
in summary , we have presented here a modified colour - screening model for @xmath0 suppression in qgp medium where quasi - particle model ( qpm ) equation of state for qgp and feed down from higher resonance states ( namely , @xmath1 , and @xmath2 ) , also dilated formation time for quarkonia and viscous effects of the qgp medium have been properly taken into account .
furthermore , we assume that the qgp is expanding under bjorken s boost invariant longitudinal expansion .
the above model is thus used to analyze the centrality dependence of the @xmath0 suppression data from sps , rhic and lhc experiments .
we notice that the centrality dependent and @xmath85 integrated survival probability as obtained in the current model accounts well the experimental data in a consistent manner at all available energies .
we have also compared our present model with the previous model results @xcite where bag model eos of qgp has been used .
we notice that present model describes the experimental data better in comparison to earlier model results @xcite .
thus we conclude that our model using colour screening picture describes all the features of @xmath0 suppression data simultaneously at sps , rhic and lhc energies with the same set ( i.e. , set ii ) of dissociation temperatures for different quarkonia states .
further extension of this work is underway to explain the complete rapidity dependence of the data for charmonium and bottomonium resonances in order to draw a firm conclusion regarding the overall level of suppression arising from qgp picture alone .
one of the author ( m. mishra ) is grateful to the department of science and technology ( dst ) , new delhi for financial assistance as fast - track young scientist project .
m. mishra also acknowledges the financial support from summer research fellowship programme , of indian academy of sciences , bangalore .
k. srivastava thanks the university grant commission , new delhi for financial support .
[ [ section ] ] in order to determine the explicit proper time ( @xmath17 ) dependence of temperature ( @xmath126 ) , we first start with the entropy equation for ( 1 + 1)-dimensional expansion in ( 3 + 1)-dimensional space given as follows [ * prd 41 * , 2903 ( 1990 ) ] : where @xmath130 is the entropy density of the system and @xmath34 is the reynold s number defined as : @xmath131 here @xmath13 and @xmath132 are ahear and bulk viscosities of the medium .
( a.1 ) can be rewritten as : we now obtain the entropy curve using qpm as eos for qgp at zero baryon chemical potential ( see fig .
we then derive an empirical relation for entropy density as @xmath134 which fits the curve shown above .
the values of @xmath127 and @xmath128 as obtained from the fit are @xmath135 mev@xmath136 and @xmath137 , respectively . now substituting @xmath130 in eq .
( a.3 ) and treating @xmath34 as almost independent of @xmath17 , we get the following relation : now using @xmath162 from qpm , we get : @xmath163 substituting @xmath164 , and @xmath165 where @xmath166 , we get : @xmath167 or @xmath168\epsilon = -\frac{4\eta}{3e^{z}}\eqno ( a.24)\ ] ] auxiliary equation for the above differential eq .
( a.24 ) as @xmath169 the roots of auxiliary equation is @xmath170 and @xmath171 .
thus the final cooling law for energy density from eq .
( a.24 ) can be given as follows : @xmath172 again starting from energy density rate equation : @xmath173 it can be rewritten as : @xmath174 using @xmath175 from qpm in eq .
( a.28 ) , we get : @xmath176 now , using eq .
( a.26 ) , we get the final differential equation for the cooling law of pressure : @xmath177 to solve this above partial differential equation , we need to determine its integrating factor ( i.f . ) which comes out to be @xmath178 .
finally the solution of eq .
( a.30 ) after using this i.f .
is : @xmath179 where @xmath24 . using the initial time boundary condition @xmath180 at @xmath144 in eq .
( a.26 ) , we get the first constraint relation as follows : @xmath181 second boundary condition @xmath28 at @xmath29 ( corresponds to temperature @xmath182 ) in eq .
( a.26 ) , yields a second constraint as follows : @xmath183 subtracting eq .
( a.33 ) from eq .
( a.32 ) , we get @xmath184 as : @xmath185 where @xmath30 is calculated from eq .
similarly @xmath186 is obtained from the following relation : @xmath187 finally , using initial condition @xmath188 at @xmath144 in eq .
( a.31 ) , we can get the following expression for @xmath189 : @xmath190 99 t. matsui and h. satz , phys .
lett . * 178 * , 416 ( 1986 ) .
m. laine , o. philipsen , m. tassler , p. romatschke , jhep * 0703 * , 054(2007 ) .
m. laine , o. philipsen , m. tassler , jhep * 0709 * , 066 ( 2007 ) .
m. laine , nucl .
phys . * a 820 * , 25c ( 2009 ) .
m. c. abreu et al . , ( na50 collaboration ) , phys . lett . * b 477 * , 28 ( 2000 ) ; b. alessandro et al . , ( na50 collaboration ) , eur
j. * c 39 * , 335 ( 2005 ) .
r. arnaldi et al .
( na60 collaboration ) , phys .
lett . * 99 * , 132302 ( 2007 ) ; r. arnaldi ( na60 collaboration ) , presentation at the ect * workshop on `` heavy quarkonia production in heavy - ion collisions '' , trent ( italy ) , may 25 - 29 ( 2009 ) .
a. adare et al . , ( phenix collaboration ) , phys .
* 98 * , 232301 ( 2007 ) ; arxiv:1208.2251v1[nucl - ex ] ( 2012 ) .
the cms collaboration , arxiv:1201.5069v1[nucl - ex ] ( 2012 ) .
k. j. eskola , b. mller and x. n. wang , phys . lett .
* b 374 * , 20 ( 1996 ) . the alice collaboration , phys .
lett . * 109 * , 072301 ( 2012 ) .
j. p. blaizot and j.y ollitrault , phys .
rev . lett . * 77 * , 1703 ( 1996 ) .
j. p. blaizot , p. m. dinh and j.y .
ollitrault , phys .
rev . lett . * 85 * , 4012 ( 2000 ) .
a. capella , e. g. ferreiro and a. b. kaidalov , phys .
lett . * 85 * , 2080 ( 2000 ) .
a. k. chaudhuri , phys.rev .
* c 64 * , 054903 ( 2001 ) ; phys.lett .
* b 527 * , 80 ( 2002 ) .
a. k. chaudhuri , phys .
lett . * 88 * , 232302 ( 2002 ) .
a. k. chaudhuri , phys . rev .
* c 66 * , 021902 ( 2002 ) .
a. k. chaudhuri and partha pratim bhaduri , arxiv:1202.3291[nucl - th ] ( 2012 ) , phys .
c ( to be published ) .
l. grandchamp , r. rapp and g. e. brown , phys .
lett . * 92 * , 212301 ( 2004 ) .
yunpeng liu et al . , j. phys . *
g 37 * , 075110 ( 2010 ) .
zhen qu et al . , nucl
. phys . * a 830 * , 335c ( 2009 ) .
rishi sharma and ivan vitev , arxiv:1203.0329v1 [ hep - ph ] ( 2012 ) .
m. mishra , c. p. singh , v. j. menon and ritesh kumar dubey , phys .
* b 656 * , 45 ( 2007 ) ; m. mishra , c. p. singh and v. j. menon , proc . of qm , indian .
j. physics * 85 * , 849 ( 2011 ) .
chu and t. matsui , phys .
d 37 * , 1851 ( 1988 ) .
k. srivastava , s. k. tiwari and c. p. singh , phys .
d 82 * , 014023 ( 2010 ) .
p. k. srivastava and c. p. singh , phys .
d 85 * , 114016 ( 2012 ) .
d. a. teaney , int .
* e 21 * , 38 ( 2009 ) .
b. alver et al . , arxiv:0805.4411 [ nucl - ex ] ( 2008 ) ; see the tglaubermc page on hepforge(http://www.hepforge.org / downloads / tglaubermc ) . s.s .
adler et al . , ( phenix collaboration ) , phys . rev . * c 71 * , 034908 ( 2005 ) .
j. d. bjorken , phys . rev .
* d 27 * , 140 ( 1983 ) . the cms collaboration , phys .
109 * , 152303 ( 2012 ) .
h. satz , j. phys . *
g 32 * , 25 ( 2006 ) ( r ) .
t. gunji , h. hamagaki , t. hatsuda and t. hirano , phys .
* c 76 * , 051901(r ) ( 2007 ) ; t. gunji ( for phenix coll . ) , j. phys . * g 34 * , s749 ( 2007 ) ; m. j. leitch , j. phys .
* g 34 * , s453 ( 2007 ) .
t umeda , h ohno and k kanaya ( for the hot - qcd collaboration ) , j. phys . *
g 35 * , 104157 ( 2008 ) .
t. umeda , phys . rev . * d 75 * , 094502 ( 2007 ) ; t. umeda , poslat * 2007 * , 233 ( 2007 )
. s. datta , f. karsch , p. petreczky and i. wetzorke , phys .
d 69 * , 094507 ( 2004 ) ; h. ohno et al .
, phys . rev . *
d 84 * , 094504 ( 2011 ) .
g. aarts , c. allton , m. b. oktay , m. peardon and j. i. skullerud , phys . rev .
* d 76 * , 094513 ( 2007 ) .
a. jakovac , p. petreczky , k. petrov and a. velytsky , phys . rev . *
d 75 * , 014506 ( 2007 ) .
m. b. oktay et al . , poslat * 2007 * , 227 ( 2007 ) .
p. gubler , k. morita and m. oka , phys . rev . lett . * 107 * , 092003 ( 2011 ) .
a. rakotozafindrabe , e. g. ferreiro , f. fleuret , j. p. lansberg , n. matagne , nucl . phys . * a 855 * , 327 ( 2011 ) .
p. kovtun , d. son , and a. starinets , phys .
* 94 * , 111601 ( 2005 ) . | we present a modified colour screening model for @xmath0 suppression in the quark - gluon plasma ( qgp ) using quasi - particle model ( qpm ) as equation of state ( eos ) .
other theoretical ingredients incorporated in the model are feed - down from higher resonances namely , @xmath1 , and @xmath2 , dilated formation time for quarkonia and viscous effects of the qgp medium .
assuming further that the qgp is expanding with bjorken s hydrodynamical expansion , the present model is used to analyze the centrality dependence of the @xmath0 suppression in mid - rapidity region and compare it with the data obtained from sps , rhic and lhc experiments .
we find that the centrality dependence of the data for the survival probability at all energies is well reproduced by our model .
we further compare our model predictions with the results obtained from the bag model eos for qgp which has usually been used earlier in all such calculations .
+ pacs numbers : 12.38.mh , 12.38.gc , 25.75.nq , 24.10.pa |
probing the statistical properties of the large - scale structure of the universe has a great importance in studying the origin of our universe .
recent galaxy surveys have been revealing the statistical properties of distribution of galaxies on very large scales .
the number density of galaxies are not necessarily proportional to the density of mass , and this ambiguity is known as galaxy biasing problem ( kaiser 1984 ; davis et al .
1985 ; bardeen et al . 1986 ) . since galaxies of different types have different clustering properties ( e.g. , dressler 1980 ; davis & geller 1976 ; giovanelli , haynes , & chincarini 1986 ; santiago & strauss 1992 ; loveday et al .
1996 ; hermit et al . 1996 ; guzzo et al .
1997 ) , not all types of galaxies can simultaneously be unbiased tracers of mass .
this ambiguity is undesirable in extracting cosmological information from the data of galaxy distribution .
the simplest model for the galaxy bias is the _ local _ , _ linear _ bias . in this simple model , the number density field of galaxy with a fixed smoothing scale @xmath1
is assumed to be proportional to density field of mass with a same smoothing scale : @xmath2 where @xmath3 and @xmath4 are density contrast of galaxy and of mass , respectively , with a fixed smoothing length @xmath1 .
this model is viable if ( a ) @xmath5 is dependent only on @xmath6 and ( b ) density contrast of mass is sufficiently small .
the latter condition ( b ) is achieved by considering large scales on which density fluctuations are small enough so that only linear term becomes significant .
the linear coefficient @xmath0 is called as bias parameter .
the bias parameter is often assumed to be a constant , although it can depend on @xmath1 in general .
the effect of nonlinearity should be taken into account when we are interested in nonlinear scales .
on weakly nonlinear scale , this affect the estimation of higher order statistics ( e.g. , fry & gaztaaga 1993 ) .
although the latter condition is reasonably considered to be valid if we are interested in linear scales , the former condition ( a ) is not trivial so far .
the non - triviality of condition ( a ) leads us to a concept of stochastic bias which is recently argued ( dekel & lahav 1998 ; pen 1998 ; tegmark & peebles 1998 ; tegmark & blomley 1998 ; taruya , koyama & soda 1998 ; taruya & soda 1998 ; blanton et al .
1998 ) . in the stochastic biasing scheme ,
@xmath5 is not supposed to be determined solely by @xmath6 , but the scatter in @xmath3-@xmath4 relation is taken into account . in linear regime in which the density contrast of mass @xmath4 is small enough , and is approximated by random gaussian field , the two - point statistics fully characterize the statistics of the scatter . in literatures , the bias parameter @xmath0 and the dimensionless cross correlation @xmath7 are used to characterize the linear stochastic biasing scheme : @xmath8 where @xmath9 and @xmath10 are rms density fluctuations of mass and galaxies , respectively .
the bias parameter @xmath0 in the equation ( [ eq1 ] ) is a generalization of the bias parameter in linear deterministic biasing scheme of equation ( [ eq0 ] ) . in deterministic case ,
the cross correlation @xmath7 is always unity . from schwarz inequality
, @xmath7 can not exceed unity and @xmath11 means that biasing is deterministic , @xmath12 .
thus , the deviation from @xmath11 measures the stochasticity .
if the smoothing scale @xmath1 is large enough so that @xmath4 and @xmath3 are considered as bivariate gaussian field , these three parameters @xmath0 , @xmath7 and @xmath13 contain all the statistical information about the stochastic biasing . in literatures ,
these parameters are sometimes considered as free parameters to be determined by observation . however , if we could know the process of galaxy formation in detail , the parameters @xmath0 and @xmath7 would be derived from some fundamental physical processes .
this is because the bias and its stochasticity come from our ignorance of the galaxy formation .
therefore , there should be some theoretical constraint between @xmath0 and @xmath7 . at first sight
, it seems difficult to find any constraint as we do not exactly know the process of galaxy formation .
this is true especially on small scales where the nonlinear and nonlocal characters of galaxy formation plays an important role on the statistics of galaxy distributions .
this problem is one of the most important issues in astrophysics and much numerical and analytical work is needed ( e.g. , rees & ostriker 1977 ; white & frenk 1991 ; cen & ostriker 1992 ; mo & white 1996 ) .
however , on large scales , such undesirable characters can be expected to be small .
scherrer & weinberg ( 1998 ) , based on the local biasing scheme , showed that the stochasticity actually vanishes on large scales and galaxy autocorrelation function behaves exactly as in deterministic biasing scheme .
dekel & lahav ( 1999 ) also imply the same property based on a specific simple model . in this paper , based on a general nonlocal method
, we show that stochasticity in fourier space asymptotically vanishes on linear scales under a certain condition , explicitly deriving the relation between the stochastic parameters @xmath0 , @xmath7 and the nonlinear , nonlocal functional form of galaxy formation .
we will show the first coefficient of generalized wiener - hermite functional , which is defined below , of the nonlocal , nonlinear relations of galaxy formation will contribute to the galaxy statistics on large scales , if that coefficient does not vanish . in the derivation , we can use the technique developed in matsubara ( 1995 ) , in which the diagrammatic methods for the calculation of general nonlocal biasing are introduced .
the diagrammatics are useful especially when the non - gaussianity and/or higher - order correlations are interested in . in this paper , however , we derive the result without employing diagrammatics for self - consistency and for the simplicity of the problem .
the diagrammatics make it easier to generalize the present results to higher - order statistics . in 2
, we revisit the mathematical methods for nonlinear , and nonlocal bias , which have been developed by matsubara ( 1995 ) .
then we derive the relation between the stochastic bias parameters and the nonlocality of biasing . in 3 , we examine three types of biasing schemes , i.e. , the local lagrangian bias models , the cooperative model , and the peak model , according to the result of 2 . in 4 , we discuss the results and present the conclusions .
in the present paradigm , the distribution of galaxies is determined by initial density fluctuations . whether a galaxy has formed at some place or not should be fully determined by the initial fluctuations . in this sense , the galaxy formation is deterministic , although it nonlinearly , nonlocally , and possibly chaotically depends on initial density fluctuations . in the stochastic biasing scheme ,
this complex features of nonlinearity and nonlocality of the galaxy formation are expressed by phenomenological scatter of local @xmath3@xmath4 relation .
thus , in principle , stochasticity can be determined by nonlinear and nonlocal deterministic processes of galaxy formation . since both the present mass density field @xmath14 and the density field of galaxies
@xmath15 are determined by initial density fluctuations @xmath16 , they are expressed by functionals @xmath17 and @xmath18 . in the following , instead of the initial fluctuations @xmath16 , we will alternatively use linearly extrapolated density fluctuations @xmath19 , where @xmath20 is a linear growth rate .
the variable @xmath21 is simply a linear extrapolation of the evolution of density contrast , regardless of whether or not the small scale fluctuations are actually in linear regime .
the introduction of the linearly extrapolated field is just for convenience and it simply represents the initial density fluctuations . in this notation , one can write the relations as @xmath22,{{\mbox{\boldmath $ x$}}}),\qquad { \rho_{\rm g}}({{\mbox{\boldmath $ x$ } } } ) = f_{\rm g}([{\delta_{\rm l}}],{{\mbox{\boldmath $ x$ } } } ) , \label{eq2}\end{aligned}\ ] ] where we introduce the notation @xmath23 $ ] which means that @xmath17 and @xmath18 are nonlocal ( and , of course , nonlinear ) functionals of @xmath21 .
these functionals also depend on the position , @xmath24 .
it is sometimes more convenient to use the density contrast of @xmath14 and @xmath15 : @xmath25,{{\mbox{\boldmath $ x$ } } } ) } { \langle f_{\rm m}([{\delta_{\rm l}}],{{\mbox{\boldmath $ x$}}})\rangle } - 1,\qquad { \delta_{\rm g}}({{\mbox{\boldmath $ x$ } } } ) = \frac{f_{\rm g}([{\delta_{\rm l}}],{{\mbox{\boldmath $ x$ } } } ) } { \langle f_{\rm g}([{\delta_{\rm l}}],{{\mbox{\boldmath $ x$}}})\rangle } - 1
. \label{eq2 - 1}\end{aligned}\ ] ] one may want to expand these functional by taylor series of @xmath21 , as in usual perturbative approach .
however , this approach would be useful only if the nonlinearities on _ all _ scales were small , and/or galaxy formation process were linear or quasi - linear .
in reality , however , the galaxy formation process is a highly nonlinear process , and it depends on the nonlinear behavior of mass density field in a very complex way .
thus , we can not use taylor expansion in our treatment below .
alternatively , we will find below that the expansion of these functional by orthogonal functionals , i.e. , generalized wiener - hermite functionals , is useful for our purpose .
the @xmath26-th order generalized wiener - hermite functional @xmath27 , which is introduced in matsubara ( 1995 ) for the first time , is defined by @xmath28 \nonumber \\ & & \qquad\qquad \times\,\frac{(-1)^n \delta^n}{\delta { \delta_{\rm l}}({{\mbox{\boldmath $ x$}}}_1 ) \cdots \delta { \delta_{\rm l}}({{\mbox{\boldmath $ x$}}}_n ) } \exp\left [ -\frac12\int d^3x d^3y { \delta_{\rm l}}({{\mbox{\boldmath $ x$ } } } ) { \xi_{\rm l}}^{-1}({{\mbox{\boldmath $ x$}}},{{\mbox{\boldmath $ y$ } } } ) { \delta_{\rm l}}({{\mbox{\boldmath $ y$ } } } ) \right ] , \label{eq3}\end{aligned}\ ] ] where @xmath29 is a functional derivative with regard to @xmath21 , and @xmath30 is a two - point correlation function of linear density field , @xmath31 .
the inverse of correlation function , @xmath32 , is defined formally by @xmath33 the original wiener - hermite functional is the special case when the correlation function is identical to dirac delta function , @xmath34 . if @xmath21 has only finite degrees of freedom instead of functional of space , the generalized wiener - hermite functionals reduce to generalized wiener - hermite polynomials which are seen in the literature ( appel & friet 1926 ) . for the application to modern cosmology , the generalized wiener - hermite functionals are more appropriate than the others .
we expand the functionals @xmath17 and @xmath18 by these functionals .
it is useful to define the following functional which is an infinite dimensional linear combination of @xmath27 : @xmath35 which we also call as generalized wiener - hermite functional .
the explicit form of the first three generalized wiener - hermite functionals are @xmath36.\end{aligned}\ ] ] the generalized wiener - hermite functionals have the orthogonality relation : @xmath37 , \label{eq6}\end{aligned}\ ] ] where @xmath38 denote the terms of symmetric permutation of the preceding term with respect to the indices . in this case , there are @xmath39 terms in total in the right hand side of equation ( [ eq6 ] ) .
the gaussian averaging @xmath40 denote the averaging by linear density field @xmath21 by the random gaussian probability functional @xmath41 $ ] : @xmath42 \cdots { \cal p}[{\delta_{\rm l } } ] , \label{eq7}\end{aligned}\ ] ] where @xmath43 = { \cal n } \exp \left [ -\frac12 \int d^3x d^3y { \delta_{\rm l}}({{\mbox{\boldmath $ x$ } } } ) { \xi_{\rm l}}^{-1}({{\mbox{\boldmath $ x$}}},{{\mbox{\boldmath $ y$ } } } ) { \delta_{\rm l}}({{\mbox{\boldmath $ y$ } } } ) \right ] .
\label{eq8}\ ] ] the formal normalization constant @xmath44 is given by @xmath45 \exp \left [ -\frac12 \int d^3x d^3y { \delta_{\rm l}}({{\mbox{\boldmath $ x$ } } } ) { \xi_{\rm l}}^{-1}({{\mbox{\boldmath $ x$}}},{{\mbox{\boldmath $ y$ } } } ) { \delta_{\rm l}}({{\mbox{\boldmath $ y$ } } } ) \right ] \right\}^{-1}. \label{eq9}\ ] ] because the degrees of freedom is infinite , @xmath46 formally diverges , but the proper regularization is always possible by discretizing the three dimensional continuum space .
one can prove the orthogonality equation ( [ eq6 ] ) , just generalizing the proof of orthogonality of simple hermite polynomial , which is well known .
we consider the functionals of equation ( [ eq5 ] ) as base functionals for the expansion of the functional of mass density field , and of galaxy density field : @xmath47 ) = \sum_{n=1}^\infty \frac1{n ! } \int d^3x_1 \cdots d^3x_n k_{\rm a}^{(n)}({{\mbox{\boldmath $ r$}}}-{{\mbox{\boldmath $ x$}}}_1,\cdots,{{\mbox{\boldmath $ r$}}}-{{\mbox{\boldmath $ x$}}}_n ) { \cal h}_{(n)}({{\mbox{\boldmath $ x$}}}_1,\ldots,{{\mbox{\boldmath $ x$}}}_n ) , \label{eq10}\ ] ]
where a = m or g. the reason why @xmath48 term is not appeared in the above expression is that @xmath49 and @xmath50 for @xmath51 [ set @xmath52 in equation ( [ eq6 ] ) ] .
this expansion is complete because the kernel @xmath53 is uniquely given by @xmath54)\right\rangle \label{eq11 } \\ & = & \left\langle \frac{\delta^n \delta_{\rm a}({{\mbox{\boldmath $ r$}}},[{\delta_{\rm l } } ] ) } { \delta{\delta_{\rm l}}({{\mbox{\boldmath $ x$}}}_1)\cdots \delta{\delta_{\rm l}}({{\mbox{\boldmath $ x$}}}_n ) } \right\rangle .
\label{eq12}\end{aligned}\ ] ] according to the expansion ( [ eq10 ] ) and the orthogonality relation ( [ eq6 ] ) , the two - point auto - correlation function of matter @xmath55 and of galaxies @xmath56 and the cross correlation function @xmath57 are given by @xmath58 where a , b = m , g , and @xmath59 .
note that this expansion is valid as long as @xmath21 is a random gaussian field .
if the initial density field is non - gaussian , there are additional terms in equation ( [ eq14 ] ) which depend on initial higher - order correlation functions [ see matsubara ( 1995 ) for detail ] . both the nonlocal kernels @xmath60 and @xmath61 do not depend on @xmath62 .
thus , on large scales with respect to the separation @xmath62 , we can approximate equation ( [ eq14 ] ) by only considering lower order terms of @xmath30 , provided that the kernel @xmath63 does not have broad profile .
if the kernel falls off slowly on large scales , we can not truncate the expansion ( [ eq14 ] ) . in the rest of this paper
, we assume the lowest order term in the expansion ( [ eq14 ] ) actually dominates the higher - order terms . before proceeding to the analysis of the lowest order approximation
, we consider the cases in which this assumption breaks down .
imagine , for example , that @xmath64 , and @xmath65 on large scales
. then the @xmath26-th order term in the integral in expansion ( [ eq14 ] ) is approximately given by @xmath66 the last expression is derived by the fact that the integral is the form of convolution and the ( 3d ) fourier transform of @xmath67 is proportional to @xmath68 ( peebles 1980 ) . as seen by this expression , @xmath69 should be negative to ensure the equation ( [ eq14.5 ] ) actually falls off on large scales . for example
, if @xmath70 falls off as @xmath71 , and @xmath72 , then equation ( [ eq14.5 ] ) does not fall off anymore . on the other hand , if @xmath73 , then equation ( [ eq14.5 ] ) falls off rapidly enough so that the higher order terms can be neglected . in the latter case ,
the spatial integration of @xmath53 is finite .
it is a natural assumption that the spatial integration of @xmath53 is finite .
if it is infinite , as seen from equation ( [ eq10 ] ) , the information of density fluctuations of infinitely distant places affect the galaxy formation as much as , or more than the fluctuations of nearer places , which is unlikely in reality .
we assume @xmath73 in the rest of this paper .
in other words , we assume that fourier transform of higher - order kernels @xmath74 are finite in the limit @xmath75 .
note that the expansion ( [ eq14 ] ) is essentially different from usual perturbative approach by taylor expansion of density contrast itself .
instead , we employ the orthogonal expansion for galaxy and mass density fields , and the resulting expression , ( [ eq14 ] ) , can be interpreted as an asymptotic expansion by correlation function , @xmath30 .
thus , we only assume the smallness of correlation function @xmath30 on large scales , whether the density contrast on small scale is large or not .
since our argument in this section is extremely formal , it would be instructive to calculate an example .
consider a model in which @xmath76 and @xmath4 is calculated up to second order in perturbation theory [ e.g. , fry ( 1984 ) ] : @xmath77 where @xmath78 e^{i { { \mbox{\boldmath $ k$}}}_1 \cdot { { \mbox{\boldmath $ x$}}}_1 + i { { \mbox{\boldmath $ k$}}}_2 \cdot { { \mbox{\boldmath $ x$}}}_2 } , \label{eq - a2}\end{aligned}\ ] ] and we assume the einstein - de sitter universe for simplicity . then , according to equation ( [ eq12 ] ) , we can calculate the kernels as @xmath79 since @xmath80 drops off as @xmath81 on large scales , the discussion at the end of the previous subsection suggests that @xmath26-th order term in equation ( [ eq14 ] ) actually drops off as @xmath82 on large scales . in the following
, we are interested in large scales and consider only the lowest order approximation of equation ( [ eq14 ] ) : @xmath83 assuming that this term does not vanish and that the higher - order terms are negligible . since this expression has the form of convolution , it becomes just products in fourier space . from statistical isotropy ,
the fourier transform of @xmath84 , denoted as @xmath85 , is a function of the absolute value of the wave vector @xmath86 : @xmath87 where @xmath88 is the linear power spectrum as the true spectrum , there is possibly a constant term in addition to the above equation , which comes from the small scale inaccuracy of expression ( [ eq15 ] ) [ see scherrer & weinberg ( 1998 ) ; dekel & lahav ( 1998 ) ] .
however , in the following , we consider @xmath89 as a merely mathematical quantity which represents just the fourier transform of @xmath90 of the equation ( [ eq15 ] ) . ] .
the linear - scale power spectrum of mass is simply given by @xmath91 .
this means @xmath92 on linear scales .
thus , denoting @xmath93 , @xmath94 these equations are valid as long as @xmath95 . if @xmath96 , equation ( [ eq16 ] ) vanishes and is no longer the lowest order in the expansion of equation ( [ eq14 ] ) . in such case
, we should consider the higher - order terms . in equations ( [ eq17])([eq19 ] ) , @xmath97 can be identified to the linear bias parameter in fourier space , and is given by , from fourier transform of equation ( [ eq12 ] ) , @xmath98 where @xmath99 is a fourier transform of linear density fluctuations , @xmath100 .
these simple equations ( [ eq17])([eq20 ] ) are the primary results of this paper .
these equations show , as long as @xmath101 , that there are no residual cross correlation in fourier space in linear regime ( except for the constant term which comes from the small scale behavior of correlation function ) and that the bias parameter in fourier space is @xmath102 which is scale - dependent and is related to nonlinearity and nonlocality of galaxy formation through equation ( [ eq20 ] ) .
this means that the fourier - mode stochasticity which arise from the nonlinearity and nonlocality of the galaxy formation vanishes in linear regime .
thus , the cross correlation @xmath103 in fourier space should approach to unity in large - scale limit , as long as the galaxy formation is such that @xmath101 .
the case , @xmath96 can happen in special cases .
it can happen , for example , when the large - scale linear power is completely erased by some peculiar form of nonlocal biasing , and mode - mode coupling from nonlinear scales dominates on linear scales .
it also can happen when the biasing is represented by an even function of @xmath21 , e.g. , purely quadratic , @xmath104 , or quartic , @xmath105 , etc .
we assume the galaxy formation does not have such special form and satisfies the condition @xmath101 in the rest of this paper .
the vanishing stochasticity is an important constraint on large scales .
at this point , the naive introduction of stochasticity @xmath7 in linear or quasi - linear regime , as is sometimes done in literatures , should be cautious . on large scales
, @xmath7 could not be freely adjusted , but would be close to unity .
this fact is already noticed in the paper by dekel & lahav ( 1998 ) , in which only a specific , simple model is considered .
the same conclusion is derived by scherrer & weinberg ( 1998 ) based on local galaxy formation scheme .
our conclusions apply not only to local schemes , but also to nonlocal schemes of galaxy formation .
the difference between local and nonlocal scheme is that local bias generates constant bias factor , while nonlocal bias generates scale - dependent one .
if there is a non - unity value of @xmath7 in large - scale limit , it means that there is some exotic stochasticity which is not relevant to the initial density fluctuations , or that @xmath106 .
as all the structures in the universe are supposed to be formed from initial density fluctuations , there is no specific reason to introduce such kind of exotic processes , at least in the framework of the present standard theory of structure formation in the universe . of course , in nonlinear regime where the approximation of equation ( [ eq15 ] ) breaks down ,
the stochasticity in fourier space arises by mode - mode coupling .
since the dynamics of such nonlinear regime is complex enough to trace analytically , the concept of stochastic bias is useful especially in this regime . in the nonlinear regime
, one should be aware that the nonlinearity of galaxy - density relation also dominate and that only parameters @xmath0 and @xmath7 are not sufficient to characterize the biasing properties ( dekel & lahav 1998 ) . the equations ( [ eq17])([eq19 ] ) is valid on scales larger than a nonlinear scale @xmath107 , because these equations are derived from lowest order approximation , which assumes @xmath108 is small .
if the nonlocality of galaxy formation is small and @xmath109 is localized on some scales smaller than @xmath110 , then its fourier transform , @xmath97 , will be constant for @xmath111 . if @xmath112 , the bias factor is constant in the validity region of the equations .
this is equivalent to the purely local galaxy formation .
in contrast , if @xmath113 , there appears the scale - dependence of bias parameter besides mode - mode coupling .
this scale - dependence comes from the nonlocality of the galaxy formation .
the @xmath114-dependence of @xmath102 thus is not negligible on scales below @xmath110 and its behavior can be describable by equation ( [ eq20 ] ) on scales above @xmath107 .
even when there is no stochasticity in fourier space , it may still appear in real space on linear scales because of the scale - dependence of bias parameter , besides mode - mode coupling . actually , the galaxy - density cross correlation and galaxy correlation for a smoothed field with linear smoothing length @xmath1 are given by @xmath115 where @xmath116 is a fourier transform of a smoothing function .
this expression and equation ( [ eq1 ] ) explicitly show the scale dependence and stochasticity of the biasing in real space .
we define the following notation for k - space averaging for an arbitrary function @xmath117 : @xmath118 then , equation ( [ eq1 ] ) reduces to , simply , @xmath119 to obtain more insight on this scale - dependence and stochasticity , let us expand the bias parameter @xmath97 in terms of @xmath120 , where @xmath110 is the scale of nonlocality of galaxy formation as above ( terms of odd power of @xmath114 do not appear for reflection symmetry ) : @xmath121 with this expansion , equation ( [ eq23 - 2 ] ) is expanded in a straightforward manner , and the results are @xmath122 ( r _ * k_{\rm g})^{-4 } + { \cal o}(r _ * k_{\rm g})^{-6 } , \label{eq24 } \\ & & r(r ) = 1 - \frac92 \gamma^2(1-\gamma^2 ) \left ( \frac{b_{\rm f}^{(1)}}{b_{\rm f}^{(0 ) } } \right)^2 ( r _ * k_{\rm g})^{-4 } + { \cal o}(r _ * k_{\rm g})^{-6}. \label{eq25}\end{aligned}\ ] ] in this expression , spectral parameters @xmath123 and @xmath124 ( bardeen et al . 1986 , bbks , hereafter ) are given by @xmath125 where notation of the equation ( [ eq23 - 1 ] ) is applied .
the parameter @xmath123 is of order unity , and @xmath124 is of order @xmath1 .
for example , if the power spectrum has the form of power - law , @xmath126 , and the smoothing function is gaussian , @xmath127 , these parameters are given by @xmath128 and @xmath129 .
equations ( [ eq24 ] ) and ( [ eq25 ] ) give the expression for stochastic bias parameter from nonlocality of galaxy formation .
these equations represent the minimal stochasticity which is inevitable when bias is scale dependent .
we have neglected the mode coupling near nonlinear scales .
the scale - dependence and stochasticity appears on the scale of the nonlocality , @xmath110 . on scales larger than nonlocality
, such scale - dependence and stochasticity disappears .
especially , stochasticity rapidly vanishes for the lack of @xmath130 term in equation ( [ eq25 ] ) .
these results do not depend on specific details of galaxy formation as long as @xmath95 , and higher - order terms are negligible .
the information of galaxy formation involves only through two parameters , @xmath131 and @xmath132 , up to the order @xmath133 , or three parameters @xmath134 , @xmath132 and @xmath135 up to the order @xmath136 .
these parameters are related to galaxy formation through equation ( [ eq20 ] ) which should be calculable if we could know the details of galaxy formation process .
otherwise , they can be considered as free parameters to be fitted by observation , instead of fitting functions @xmath137 and @xmath138 , which have the infinite degrees of freedom .
so far the argument is quite formal . in this section ,
we consider specific models of galaxy formation , i.e. , the lagrangian local biasing models , the cooperative model , and the peak model , as simple examples .
although the quantitative correspondence of these models and the actual galaxy formation still needs investigation , these examples can give qualitative aspects on the nonlocal galaxy formation .
in the lagrangian local galaxy formation models , the number density of galaxies is a local function of a smoothed linear density field . the smoothed density field , however , is a nonlocal function of linear density field , @xmath21 .
thus , in some sense , local galaxy formation models fall in the category of nonlocal models , with particularly simple functional form : @xmath139 where @xmath140 is an usual one - variable function of the smoothed linear density field .
the smoothed linear density field , @xmath141 , is defined by @xmath142 where @xmath143 is a smoothing kernel function with smoothing length @xmath144 . with this particularly simple form
, the bias parameter in fourier space , given by equation ( [ eq20 ] ) , reduces to @xmath145 where @xmath146 is a fourier transform of the smoothing kernel @xmath143 .
all the other higher - order kernels @xmath53 ( @xmath147 ) vanish .
this result is essentially the same which is derived previously in real space ( szalay 1988 ; coles 1993 ) : @xmath148 since @xmath149 . in usual nonlocal biasing models , the smoothing length @xmath144 is taken to be in nonlinear regime . as we consider linear scales , @xmath150 and
there is no scale - dependence on @xmath97 and thus there is no stochasticity in real space .
this result is consistent with the work by scherrer & weinberg ( 1998 ) , who showed the local models produce constant bias factor and there is negligible stochasticity on large scales .
examples of local models are density - threshold bias , @xmath151 , where @xmath152 is a step function ( kaiser 1984 ; jensen & szalay 1986 ) , weighted bias , @xmath153 ( catelan et al .
1994 ) , cen - ostriker bias , @xmath154 ^ 2\}$ ] ( cen & ostriker 1992 ) , etc . in nonlinear regime
, there is really scale - dependence on @xmath97 ( mann , peacock , & heavens 1998 ; narayanan , berlind , & weinberg 1998 ) .
recently , the mo & white model ( mo & white 1996 ) for the clustering of dark matter halos , which is an extension of press - schechter formalism ( press & schechter 1974 ; cole & kaiser 1989 ; bond et al . 1991 ) , is interested in ( e.g. , catelan et al .
1998 ; catelan , matarrese & porciani 1998 ; jing 1998 ; sheth & lemson 1998 ) .
this model is also an example of the lagrangian local biasing model , because the halo density contrast @xmath155 in their model is determined by linear density field @xmath156 which is smoothed on scale @xmath157 .
next , we consider cooperative galaxy formation model introduced by bower et al .
( 1993 ) , in which they showed that a large - scale ( @xmath158 ) , but weak modulation of galaxy luminosities can reconcile the discrepancy between the scdm power spectrum and apm galaxy data ( see also babul & white 1991 ) . as a simple example , they identify sites for galaxy formation as places where density contrast @xmath159 satisfies the following relation @xmath160 where @xmath161 is the density contrast smoothed on a scale @xmath162 , which represent the large - scale modulation , and @xmath163 is a constant which is called as `` the modulation coefficient '' .
this simple model is mathematically equivalent to the standard density - threshold bias model , but for the new field defined by @xmath164 it is easy to see that this new field is just a new density field with a smoothing function @xmath165 , where @xmath146 and @xmath166 are smoothing function for @xmath159 and @xmath161 , respectively .
the bias parameter for the cooperative model in fourier space is similar to equation ( [ eq42 ] ) : @xmath167 \label{eqc03}\\ & = & \sqrt{\frac{2}{\pi } } \frac{e^{-{\nu'}^2/2 } } { { \rm erfc}\left(\nu'/\sqrt{2}\right ) } \left [ w_{\rm s}(ks ) + \kappa w_{\rm mod}(k r_{\rm mod } ) \right ] , \label{eqc04}\end{aligned}\ ] ] where @xmath168 .
this expression and equation ( [ eq23 - 2 ] ) enable us to evaluate the stochastic parameters .
in figure 1 , we plot the scale dependence of stochastic parameters of the cooperative model . in this figure
, we assume a cdm power spectrum of bbks @xmath169^{-1/2 } , \label{eq37}\\ & & \qquad\qquad\qquad q = \frac{k}{\gamma h { \rm mpc}^{-1}},\end{aligned}\ ] ] with primordial spectral index @xmath170 and shape parameter @xmath171 and linear amplitude @xmath172 which corresponds to @xmath173 .
the top - hat smoothing function is applied for smoothing @xmath1 , while the gaussian smoothing function is applied for smoothing @xmath144 and @xmath162 : @xmath174 and we set @xmath175 , @xmath176 and @xmath177 .
the modulation coefficient is adjusted so as to produce the 2.5% rms modulation of the threshold , according to bower et al .
( 1993 ) , i.e. , @xmath178 .
this required taking @xmath179 , @xmath180 , @xmath181 for @xmath182 , @xmath183 , @xmath184 , respectively and bower et al.s are due to different fitting formula for power spectrum . ] .
as one can see from the figure , there appears the scale - dependence of bias parameter @xmath137 on scales of @xmath162 so that the galaxy clustering on large scales are enhanced by cooperative bias .
this fact is a main motivation for bower et al .
( 1993 ) to introduce the cooperative model .
the stochastic parameter @xmath138 is very close to unity except on the modulation scale , where there is weak stochasticity due to scale dependence of bias parameter .
next example of nonlocal galaxy formation is the peak model . in the peak model , the sites for galaxy formation
are identified as high peaks of initial density field with a fixed smoothing length ( see bbks ) . treating the constraint properly for density peaks
is difficult but there are several approximations . in this paper , we approximate the density peaks by density extrema ( otto , politzer & wise 1986 ; cline et al . 1987 ; catelan et al .
the number density of density extrema above threshold @xmath185 is given by @xmath186 where @xmath187 is a smoothed linear density field with smoothing length @xmath144 .
density extrema are identical to density peaks above some moderate threshold where almost all density extrema would be density peaks . from the general consideration in the previous section , it is straightforward to obtain parameters @xmath137 and @xmath138 of stochastic bias for this model . from equation ( [ eq20 ] ) and after tedious calculation of this model are finite in large - scale limit .
] , we obtain @xmath188 w_{\rm s}(ks ) , \label{eq29}\end{aligned}\ ] ] where @xmath189 is a fourier transform of smoothing function for @xmath141 .
other quantities in this expression are defined by @xmath190 hermite polynomial @xmath191 are defined with the normalization , @xmath192 equation ( [ eq29 ] ) explicitly show the scale - dependence of bias parameter in fourier space .
the scale of nonlocality corresponds to @xmath193 , which is of order of smoothing length @xmath144 for obtaining density peaks .
the stochastic parameters @xmath137 , @xmath138 in linear regime for density peaks are derived from the equation ( [ eq23 - 2 ] ) .
assuming @xmath194 , the result is @xmath195 where spectral indices of various kind are defined as follows : @xmath196 equations ( [ eq33 ] ) and ( [ eq34 ] ) describe stochastic parameters in real space .
it is known that if we take both high threshold limit , @xmath197 , and large scale limit , @xmath198 , the correlation function of the peak model reduces to that of linear bias , @xmath199 ( kaiser 1984 ; bbks 1986 ) .
this property is easily confirmed from equation ( [ eq33 ] ) , where @xmath200 for @xmath198 . in figure 2
, these equations are plotted for @xmath201 with @xmath202 , where we assume cdm power spectrum of bbks , equation ( [ eq37 ] ) , with primordial spectral index @xmath170 and shape parameter @xmath203 and linear amplitude @xmath204 which corresponds to @xmath173 .
the top - hat smoothing function is applied for smoothing @xmath1 , while the gaussian smoothing function is applied for smoothing @xmath144 : @xmath205 the large smoothing lengths , @xmath206 do not correspond to galaxy formation , but rather they correspond to cluster formation , since the cluster of galaxies are density peaks of large smoothing length .
as seen in the figure , once the smoothing scale @xmath1 exceeds the smoothing length @xmath144 , which is the scale of galaxy or cluster formation in this model , the parameter @xmath7 of stochasticity rapidly converges to unity , which means there is no stochasticity above that scale . as
stochasticity vanishes , the bias parameter @xmath0 converges to a constant on large scales .
in this paper we explicitly derive the stochasticity parameters of the bias in linear regime from the nonlocality of the galaxy formation . by using the generalized wiener - hermite functionals
, we can derive the two - point correlation on linear scales which is valid even if the galaxy formation process itself is both nonlinear and nonlocal .
this is in contrast to the usual taylor expansion which can not treat the strongly nonlinear features of galaxy formation .
wiener - hermite functionals are orthogonal functionals and we do not have to assume the smallness of @xmath21 itself , and even do not have to know the exact nonlinear evolution of density contrast , @xmath4 .
instead , we assume only the smallness of correlation function on large scales .
we show that the stochasticity in fourier space does not exist in linear regime ( except for the constant term which comes from the small scale behavior of correlation function ) , and that the biasing parameter in fourier space @xmath97 is given by @xmath207 .
this conclusion is true as long as the galaxy formation process satisfies the relation , @xmath101 , and higher - order kernels @xmath208 ( @xmath147 ) do not increase with scales when @xmath209 .
this property in fourier space is simply because the galaxy - galaxy and galaxy - mass correlation functions can be expressed as convolutions of mass correlation function at lowest order of the expansion by @xmath30 [ equation ( [ eq15 ] ) ] .
a local model of galaxy formation has a constant bias factor , while the nonlocal model has a scale - dependent one , besides mode - mode coupling . in the linear regime where mode - mode coupling is negligible , stochasticity in real space
comes simply from the scale - dependence of the biasing when the galaxy formation is nonlocal and @xmath101 .
thus , naive introduction of stochasticity in fourier mode in the linear regime should be avoided .
one can not introduce stochasticity in fourier mode in the linear regime simply because of the lack of knowledge about galaxy formation .
if there is any stochasticity in fourier mode in the linear regime , it means that there are exotic process in the galaxy formation which does not come solely from the initial density field and such process should be correlated on linear scales , unless galaxy formation process has a special form to satisfy @xmath96 .
such kind of exotic process is not likely , at least in the present framework of the standard theory of structure formation in the universe .
we should note our analyses are restricted to the linear regime . in the nonlinear regime
, there are mode coupling from both nonlinearity of density evolution and nonlinearity of galaxy formation and it makes the fourier mode stochastic . in the region where stochasticity is prominent , the nonlinear density evolution , which is difficult to track analytically , is also prominent , so that the phenomenological approach of stochastic bias should be effectively applied in nonlinear regime ( dekel & lahav 1998 ) . in strongly nonlinear regime ,
phenomenological approach by the ( hyper ) extended perturbation theory ( colombi et al .
1997 ; scoccimarro & frieman 1998 ) can shed light on how nonlinearity makes cross correlation @xmath7 deviate from unity .
in this theory , the higher order cumulants of mass density field is given by @xmath210 , where @xmath211 is a constant predicted by tree - level perturbation theory , and @xmath212 .
although this theory contains an extrapolation of weakly nonlinear result to strongly nonlinear regime , it phenomenologically describe the numerical results . in strongly nonlinear regime ,
a mere averaging and cumulant are approximately equivalent in this ansatz : @xmath213 .
thus , if @xmath214 , one can obtain @xmath215 and @xmath216 . finally , one has @xmath217 .
this value of @xmath7 depends on the scale , and departs significantly from unity on small scales .
the conclusion that the stochasticity is weak on linear scales is good news for determining the redshift distortion parameter @xmath218 on linear scales from a redshift survey .
the linear redshift distortion of power spectrum is given by , in the plane - parallel limit ( kaiser 1987 ; pen 1998 ) , @xmath219 p_{\rm g}(k ) , \label{eq39}\end{aligned}\ ] ] where @xmath220 is redshift - space power spectrum of galaxies , and @xmath221 is a direction cosine of the angle between the wave vector @xmath114 and the line of sight [ see hamilton ( 1992 ) for an expression for two - point correlation function and szalay , matsubara & landy ( 1998 ) for its generalization to non - plane - parallel case ] .
since we see @xmath222 on linear scales except some special cases , we do not need to fit @xmath7 from the observation when we use only the linear redshift distortion .
however , the previous analyses so far usually assume the bias parameter @xmath0 as a scale - independent constant .
this assumption is justified if the scale of nonlocality of galaxy formation is actually below the linear scale .
if it is not , the scale - dependence of @xmath223 should also be determined by observation ( or by theories , if possible ) .
the forthcoming large - scale redshift surveys will reveal the galaxy distribution especially on linear scales , on which we have not had sufficient data so far .
as shown in this paper , the linear clustering properties are analytically tractable even when the galaxy formation itself is a too complex phenomenon to analytically track .
the exploration of linear - scale galaxy distribution can overcome our ignorance of detailed galaxy formation processes , and will give a great insight on the primordial features of our universe .
appel , p. , & de friet , j. k. 1926 ; fonctions hypergomtriques et hypersphriques , polynmes dhermite;paris;ganthier - villars ; babul , a. , & white , s. d. m. 1991;mnras;253;31p ; bardeen , j. m. , bond , j. r. , kaiser , n. & szalay , a. s. 1986;apj;304;15 ; blanton , m. , cen , r. , ostriker , j. p. , & strauss , m. a. 1998;apj;submitted;astro - ph/9807029 ; bond , j. r. , cole , s , efstathiou , g. , & kaiser , n. 1991;apj;379;440 ; bower , r. g. , coles , p. , frenk , c. s. & white , s. d. m. 1993;apj;405;403 ; catelan , p. , coles , p. , matarrese , s. , & moscardini , l. 1994;mnras;268;966 ; catelan , p. , lucchin , f. , & matarrese , s. 1988;phys .
lett.;61;267 ; catelan , p. , lucchin , f. , matarrese , s. , & porciani , c. 1998;mnras;297;692 ; catelan , p. , matarrese , s. , & porciani , c. 1998;apj;502;l1 ; cen , r. y. , & ostriker , j. p. 1992;apj;399;l113 ; cline , j. m. , politzer , h. d. , rey , s .- j . ,
& wise , m. b. 1987;commun .
phys.;112;217 ; cole , s. , & kaiser , n. 1989;mnras;237;1127 ; coles , p. 1993;mnras;262;1065
; colombi , s. , bernardeau , f. , bouchet , f. r. , & hernquist , l. 1997;mnras;287;241 ; davis , m. , & geller , m. j. 1976;apj;208;13 ; davis , m. , efstathiou , g. , frenk , c. s. & white , s. d. m. 1985;apj;292;371 ; dressler , a. 1980;apj;236;351 ; dekel , a. , & lahav , o. 1998;preprint;astro - ph/9806193 ; fry , j. n. 1984;apj;279;499 ; fry , j. n. & gaztaaga , e. 1993;apj;413;447 ; giovanelli , r. , haynes , m. p. , & chincarini , g. l. 1986;apj;300;77 ; guzzo , l. , strauss , m. a. , fisher , k. b. , giovanelli , r. , & haynes , m. p. 1997;apj;489;37
; hamilton , a. j. s. 1992;apj;385;l5 ; hermit , s. , santiago , b. x. , lahav , o. , strauss , m. a. , davis , m. , dressler , a. , & huchra , j. p. 1996;mnras;283;709
; jensen , l. g. , & szalay , a. s. 1986;apj;305;l5 ; jing , y .- p .
1998;apj;503;l9 ; kaiser , n. 1984;apj;284;l9 ; kaiser , n. 1987;mnras;227;1 ; loveday , j. , efstathiou , g. , maddox , s. j. , & peterson , b. a. 1996;apj;468;1 ; matsubara , t. 1995;apjs;101;1 ; mann , r. g. , peacock , j. a. , & heavens , a. f. 1998;mnras;293;209 ; mo , h. j. , & white , s. d. m. 1996;mnras;282;347 ; narayanan , v. k. , berlind , a. a. , & weinberg d. h. 1998;apj;submitted;astro - ph/9812002 ; otto , s. , politzer , h. d. , & wise , m. b. 1986 ; phys .
lett.;56;1878 ; peebles , p. j. e. 1980;the large - scale structure of the universe;princeton university press;princeton ; pen , u. 1998;apj;504;601 ; press , w. h. , & schechter , p. 1974;apj;187;425 ; rees , m. j. & ostriker , j. p. 1977;mnras;179;541
; santiago , b. x. , & strauss , m. a. 1992;apj;387;9 ; scherrer , r. j. & weinberg , d. h. 1998;apj;504;607 ; scoccimarro , r. , & frieman , j. a. 1998;astro - ph/9811184 ; ; szalay , a. s. 1988;apj;333;21 ; szalay , a. s. , matsubara , t. & landy , s. d. 1998;apj;498;l1 ; taruya , a. , koyama , k. , & soda , j. 1998;apj;in press;astro - ph/9807005 ; taruya , a. , & soda , j. 1998;apj;submitted;astro - ph/9809204 ; tegmark , m. , & blomley , b. c. 1998;apjl;submitted;astro - ph/9809324 ; tegmark , m. , & peebles , p. j. e. 1998;apj;500;l79 ; white , s. d. m. , & frenk , c. s. 1991;apj;379;52 ; | if one wants to represent the galaxy number density at some point in terms of only the mass density at the same point , there appears the stochasticity in such a relation , which is referred to as `` stochastic bias '' .
the stochasticity is there because the galaxy number density is not merely a local function of a mass density field , but it is a nonlocal functional , instead .
thus , the phenomenological stochasticity of the bias should be accounted for by nonlocal features of galaxy formation processes .
based on mathematical arguments , we show that there are simple relations between biasing and nonlocality on linear scales of density fluctuations , and that the stochasticity in fourier space does not exist on linear scales under a certain condition , even if the galaxy formation itself is a complex nonlinear and nonlocal precess .
the stochasticity in real space , however , arise from the scale - dependence of bias parameter , @xmath0 . as examples ,
we derive the stochastic bias parameters of simple nonlocal models of galaxy formation , i.e. , the local lagrangian bias models , the cooperative model , and the peak model .
we show that the stochasticity in real space is also weak , except on the scales of nonlocality of the galaxy formation .
therefore , we do not have to worry too much about the stochasticity on linear scales , especially in fourier space , even if we do not know the details of galaxy formation process . |
compared to previous satellites , swift has made a large difference on two main accounts .
first , the sensitivity of the burst alert detector ( bat , in the range 20 - 150 kev ) is a factor @xmath1 5 higher than for the corresponding instruments in the predecessor cgro - batse , bepposax and hete-2 .
second , swift can slew in less then 100 seconds in the direction determined by the bat instrument , positioning its much higher angular resolution x - ray ( xrt ) and uv - optical ( uvot ) detectors on the burst @xcite as of december 2005 , at an average rate of 2 bursts detected per week , over 100 bursts had been detected by bat , of which 90% were followed promptly with the xrt within @xmath2 s from the trigger , and about half within 100 s @xcite , while @xmath3 were detected with the uvot @xcite . of these , over 23 resulted in redshift determinations .
ten short grb were detected , of which five had detected x - ray afterglows , three had optical , and one had a radio afterglow , and five had a redshift determination .
the new observations brings the total redshift determinations to over 50 since 1997 when bepposax enabled the first one .
the redshifts based on swift have a median @xmath4 , which is a factor @xmath5 higher than the median of those previously culled via bepposax and hete-2 , @xcite .
this can be ascribed to the higher sensitivity of bat and the prompt accurate positions from xrt and uvot , making possible ground - based detection at a stage when the afterglow is much brighter .
the highest swift - enabled redshift so far is in grb 050904 , obtained with subaru , @xmath6 @xcite , the second highest being grb 050814 at @xmath7 , whereas the previous beppo - sax era record was @xmath8 .
the relative paucity of uvot detections versus xrt detections may be ascribed in part to this higher median redshift , and in part to the higher dust extinction at the implied shorter rest - frame wavelenghts for a given observed frequency @xcite , although additional effects may be at work too .
the bat light curves show that in some of the bursts which fall in the long " category ( @xmath9 s ) faint soft gamma - ray tails can be followed which extend the duration by a factor up to two beyond what batse could have detected @xcite .
a rich trove of information on the burst and afterglow physics has come from detailed xrt light curves , starting on average 100 seconds after the trigger , together with the corresponding bat light curves and spectra .
this suggests a canonical x - ray afterglow @xcite with one or more of the following:1 ) an initial steep decay @xmath10 with a temporal index @xmath11 , and an energy spectrum @xmath12 with energy spectral index @xmath13 ( or photon number index @xmath14 ) , extending up to a time @xmath15;2 ) a flatter decay @xmath16 with @xmath17 and energy index @xmath18 , at times @xmath19;3 ) a normal " decay @xmath20 with @xmath21 and @xmath18 ( generally unchanged the previous stage ) , up to a time @xmath22 , or in some cases longer;4 ) in some cases , a steeper decay @xmath23 with @xmath24 , after @xmath25;5 ) in about half the afterglows , one or more x - ray flares are observed , sometimes starting as early as 100 s after trigger , and sometimes as late as @xmath26 .
the energy in these flares ranges from a percent up to a value comparable to the prompt emission ( in grb 050502b ) .
the rise and decay times of these flares is unusually steep , depending on the reference time @xmath27 , behaving as @xmath28 with @xmath29 , and energy indices which can be also steeper than during the smooth decay portions .
the flux level after the flare usually decays to the value extrapolated from the value before the flare rise .
another major advance achieved by swift was the detection of the long burst grb 050904 , which broke through the astrophysically and psychologically important redshift barrier of @xmath30 .
this burst was very bright , both in its prompt @xmath31-ray emission ( @xmath32 erg ) and in its x - ray afterglow .
prompt ground - based optical / ir upper limits and a j - band detection suggested a photometric redshift @xmath33 @xcite .
spectroscopic confirmation with the 8.2 m subaru telescope gave a @xmath6 @xcite .
there are several striking features to this burst .
one is the enormous x - ray brightness , exceeding for a full day the x - ray brightness of the most distant x - ray quasar know to - date , sdss j0130 + 0524 , by up to a factor @xmath34 in the first minutes @xcite .
the implications as a tool for probing the igm are thought - provoking .
another feature is the extremely variable x - ray light curve , showing many large amplitude flares extending up to at least a day .
a third exciting feature is the report of a brief , very bright ir flash @xcite , comparable in brightness to the famous @xmath35 optical flash in grb 990123 .
the third major advance from swift was the discovery and localization of short grb afterglows .
as of december 2005 nine short bursts had been localized by swift , while in the same period hete-2 discovered two , and one was identified with the ipn network . in five of the swift short bursts an x - ray afterglow was measured and followed up , with grb 050709 , 050724 and 051221a showing an optical afterglow , and 050724 also a radio afterglow , while 040924 had an optical afterglow but not an x - ray one @xcite .
these are the first afterglows detected for short bursts .
also , for the first time , host galaxies were identified for these short bursts , which in four cases are early type ( ellipticals ) and in two cases are irregular galaxies .
the redshifts of four of them are in the range @xmath36 , while another one was initially given as @xmath37 but more recently has been reported as @xmath38 @xcite .
the median @xmath39 is @xmath40 that of the long bursts .
there is no evidence for significant star formation in any of these host environments , which corresponds to what one would expect for neutron star mergers or neutron star - black hole mergers , the most often discussed progenitor candidates ( it would also be compatible with other progenitors involving old compact stars ) .
the first short burst seen by swift , grb 05059b , was a low luminosity ( @xmath41 erg ) burst with a simple power - law x - ray afterglow which could only be followed for @xmath42 s @xcite .
the third one , grb 050724 , was brighter , @xmath43 erg , and could be followed in x - rays for at least @xmath34 s @xcite .
the remarkable thing about this burst s x - ray afterglow is that it resembles the typical x - ray light curves described above for long grb except for the lack of a slow - decay phase , and for the short prompt emission which places it the the category of short bursts , as well as the elliptical host galaxy candidate .
it also has x - ray flares , at 100 s and another one at @xmath44 s. the first flare has the same fluence as the prompt emission , while the late flare has @xmath45 of that .
the interpretation of these pose interesting challenges , as discussed below .
the afterglow is expected to become important after a time t_ag= max [ ( 3/4 ) ( r_dec/2c ^2)(1+z ) , t ] = max [ 10 ^ 2 ( e_52/n_0)^1/3 _ 2 ^ -8/3(1+z)|s , t ] , [ eq : tag ] where the deceleration time is @xmath46 and @xmath47 is the duration of the prompt outflow , @xmath48 marking the beginning of the self - similar blast wave regime . denoting the frequency and time dependence of the afterglow spectral energy flux as @xmath49 ,
the late x - ray afterglow phases ( 3 ) and ( 4 ) described above are similar to those known previously from beppo - sax .
( for a review of this earlier behavior and its modeling see e.g. @xcite ) .
the normal " decay phase ( 3 ) , with temporal decay indices @xmath50 and spectral energy indices @xmath51 , is what is expected from the evolution of the forward shock in the blandford - mckee self - similar late time regime , under the assumption of synchrotron emission . the late steep decay decay phase ( 4 ) of [ sec : obs ] , occasionally seen in swift bursts , is naturally explained as a jet break , when the decrease of the ejecta lorentz factor leads to the light - cone angle becoming larger than the jet angular extent , @xmath52 ( e.g. @xcite ) .
it is noteworthy , however , that this final steepening has been seen in less than @xmath45 of the swift afterglows , and then with reasonable confidence mainly in x - rays .
the corresponding optical light curve breaks have been few , and not well constrained .
this is unlike the case with the @xmath53 beppo - sax bursts , for which an achromatic break was reported in the optical @xcite , while in some of the rare cases where an x - ray or radio break was reported it occurred at a different time @xcite .
the relative paucity of optical breaks in swift afterglows may be an observational selection effect due to the larger median redshift , and hence fainter and redder optical afterglow at the same observer epoch , as well as perhaps reluctance to commit large telescope time on more frequently reported bursts ( an average , roughly , of 2/month with beppo - sax versus 2/week with swift ) . among the new early afterglow features detected by swift ,
the steep initial decay phase @xmath54 in x - rays of the long grb afterglows is one of the most puzzling .
there could be several possible reasons for this .
the most immediate of these would be the cooling following cessation of the prompt emission ( internal shocks or dissipation ) .
if the comoving magnetic field in the emission region is random [ or transverse ] , the flux per unit frequency along the line of sight in a given energy band , as a function of the electron energy index @xmath55 , decays as @xmath56 with @xmath57 $ ] in the slow cooling regime , where @xmath58 , and it decays as @xmath59 $ ] in the fast cooling regime where @xmath60 , i.e. for the standard @xmath61 this would be @xmath62 $ ] in the slow cooling or @xmath63 $ ] in the fast cooling regime , for random [ transverse ] fields @xcite . in some bursts
this may be the explanation , but in others the time and spectral indices do not correspond well .
currently the most widely considered explanation for the fast decay , either in the initial phase ( 1 ) or in the steep flares , attributes it to the off - axis emission from regions at @xmath64 ( the curvature effect , or high latitude emission @xcite . in this case , after the line of sight gamma - rays have ceased , the off - axis emission observed from @xmath65 is @xmath66 smaller than that from the line of sight . integrating over the equal arrival time region ,
this flux ratio becomes @xmath67 .
since the emission from @xmath68 arrives @xmath69 later than from @xmath70 , the observer sees the flux falling as @xmath71 , if the flux were frequency independent . for a source - frame flux @xmath72 ,
the observed flux per unit frequency varies then as f_(t - t_0)^-2- [ eq : hilat ] i.e. @xmath73 . this
high latitude " radiation , which for observers outside the line cone at @xmath74 would appear as prompt @xmath31-ray emission from dissipation at radius @xmath75 , appears to observers along the line of sight ( inside the light cone ) to arrive delayed by @xmath76 relative to the trigger time , and its spectrum is softened by the doppler factor @xmath77 into the x - ray observer band . for the initial prompt decay ,
the onset of the afterglow ( e.g. phases 2 or 3 ) , which also come from the line of sight , may overlap in time with the delayed high latitude emission . in equation ( [ eq : hilat ] )
@xmath27 can be taken as the trigger time , or some value comparable or less than by equation ( [ eq : tag ] ) .
this can be used to constrain the prompt emission radius @xcite .
when @xmath78 , the emission can have an admixture of high latitude and afterglow , and since the afterglow has a steeper spectrum than the high latitude ( which has a prompt spectrum ) , one can have steeper decays @xcite .
values of @xmath27 closer to the onset of the decay also lead to steeper slopes .
structured jets , when viewed on - beam produce essentially the same slopes as homogeneous jets , while off - beam observing can lead to shallower slopes @xcite . for the flares , if their origin is assumed to be internal ( e.g. some form of late internal shock or dissipation ) the value of @xmath27 is just before the flare , e.g the observer time at which the internal dissipation starts to be observable @xcite .
this interpretation appears , so far , compatible with most of the swift afterglows @xcite .
alternatively , the initial fast decay could be due to the emission of a cocoon of exhaust gas @xcite , where the temporal and spectral index are explained through an approximately power - law behavior of escape times and spectral modification of multiply scattered photons .
the fast decay may also be due to the reverse shock emission , if inverse compton up - scatters primarily synchrotron optical photons into the x - ray range .
the decay starts after the reverse shock has crossed the ejecta and electrons are no longer accelerated , and may have both a line of sight and an off - axis component @xcite .
this poses strong constraints on the compton - y parameter , and can not explain decays much steeper than @xmath79 , or @xmath80 if the off - axis contribution dominates .
models involving bullets , whose origin , acceleration and survivability is unexplained , could give a prompt decay index @xmath81 @xcite , but imply a bremsstrahlung energy index @xmath82 which is not observed in the fast decay , and require fine - tuning .
finally , a patchy shell model , where the lorentz factor is highly variable in angle , would produce emission with @xmath83 .
thus , such mechanisms may explain the more gradual decays , but not the more extreme @xmath84 values encountered in some cases . the slow decay portion of the x - ray light curves ( @xmath85 ) , ubiquitously detected by swift , is not entirely new , having been detected in a few cases by bepposax .
this , as well as the appearance of wiggles and flares in the x - ray light curves after several hours were the motivation for the refreshed shock " scenario @xcite .
refreshed shocks can flatten the afterglow light curve for hours or days , even if the ejecta is all emitted promptly at @xmath86 , but with a range of lorentz factors , say @xmath87 , where the lower @xmath88 shells arrive much later to the foremost fast shells which have already been decelerated .
thus , for an external medium of density @xmath89 and a prompt injection where the lorentz factor spread relative to ejecta mass and energy is @xmath90 , @xmath91 , the forward shock flux temporal decay is given by @xcite = [ ( g-4)(1+s)+(24 -7 g + sg)]/[2(7+s-2 g ) ] .
[ eq : shallow ] it needs to be emphasized that in this model all the ejection can be prompt ( e.g. over the duration @xmath92 of the gamma ray emission ) but the low @xmath88 portions arrive at ( and refresh ) the forward shock at late times , which can range from hours to days .
i.e. , it is not the central engine which is active late , but its effects are seen late .
fits of such refreshed shocks to observed shallow decay phases in swift bursts @xcite lead to a @xmath88 distribution which is a broken power law , extending above and below a peak around @xmath93 .
another version of refreshed shocks , on the other hand , does envisage central engine activity extending for long periods of time , e.g. @xmath94 day ( in contrast to the @xmath94 minutes engine activity in the model above ) .
such long - lived activity may be due to continued fall - back into the central black hole @xcite or a magnetar wind @xcite .
one characteristic of both types of refreshed models is that after the refreshed shocks stop and the usual decay resumes , the flux level shows a step - up relative to the previous level , since new energy has been injected . from current analyses ,
the refreshed shock model is generally able to explain the flatter temporal x - ray slopes seen by swift , both when it is seen to join smoothly on the prompt emission ( i.e. without an initial steep decay phase ) or when seen after an initial steep decay .
questions remain concerning the interpretation of the fluence ratio in the shallow x - ray afterglow and the prompt gamma - ray emission , which can reach @xmath95 @xcite .
this requires a higher radiative efficiency in the prompt gamma - ray emission than in the x - ray afterglow .
one might speculate that this might be achieved if the prompt outflow is poynting - dominated .
alternatively , a more efficient afterglow might emit more of its energy in other bands , e.g. in gev , or ir . or @xcite a previous mass ejection might have emptied a cavity into which the ejecta moves , leading to greater efficiency at later times , or otherwise the energy fraction going into the electrons increases @xmath96 .
refreshed shocks can also explain some of the x - ray flares whose rise and decay slopes are not too steep .
however , this model encounters difficulties with the very steep flares with rise or decay indices @xmath97 , such as inferred from the giant flare of grb 0500502b @xcite around 300 s after the trigger .
also , the flux level increase in this flare is a factor @xmath98 above the smooth afterglow before and after it , implying a comparable energy excess in the low versus high @xmath88 material .
an explanation based on inverse compton scattering in the reverse shock @xcite can explain a single flare at the beginning of the afterglow , with not too steep decay . for multiple flares , models invoking encountering a lumpy external medium have generic difficulties explaining steep rises and decays @xcite , although extremely dense , sharp - edged lumps , if they exist , might satisfy the steepness @xcite .
currently the more widely considered model for the flares ascribes them to late central engine activity @xcite .
the strongest argument in favor of this is that the energy budget is more easily satisfied , and the fast rise / decay is straightforward to explain . in such a model the flare energy can be comparable to the prompt emission , the fast rise comes naturally from the short time variability leading to internal shocks ( or to rapid reconnection ) , while the rapid decay may be due to the high latitude emission following the flare , with @xmath27 reset to the beginning of each flare ( see further discussion in @xcite ) .
however , some flares are well modeled by refreshed forward shocks , while in others this is clearly ruled out and a central engine origin is better suited @xcite . aside from the phenomenological desirability based on energetics and timescales ,
a central engine origin is conceivable , within certain time ranges , based on numerical models of the core collapse origin in long bursts .
these are interpreted as being due to core collapse of a massive stellar progenitor , where continued infall into fast rotating cores can continue for a long time @xcite .
however , large flares with a fluence which is a sizable fraction of the prompt emission occurring hours later remain difficult to understand .
it has been argued that gravitational instabilities in the infalling debris torus can lead to lumpy accretion @xcite .
alternatively , if the accreting debris torus is dominated by mhd effects , magnetic instabilities can lead to extended , highly time variable accretion @xcite .
swift , and in smaller numbers hete-2 , have provided the first bona fide short burst x - ray afterglows followed up starting @xmath99 s after the trigger , leading to localizations and redshifts .
in the first of these , grb 050509b @xcite the extrapolation of the prompt bat emission into the x - ray range , and the xrt light curve from 100 s to about 1000 s ( after which only upper limits exist , even with chandra , due to the faintness of the burst ) can be fitted with a single power law of @xmath100 , or separately as @xmath101 and @xmath102 .
the x - ray coverage was sparse due to orbital constraints , the number of x - ray photons being small , and no optical transient was identified , probably due to the faintness of the source . an optical host was however identified @xcite , an irregular galaxy at @xmath103 ( and the observations ruled out any supernova association ) . on the other hand ,
grb 050724 was relatively bright , and besides x - rays , it also yielded both a decaying optical and a radio afterglow @xcite .
this burst , together with the greater part of other short bursts , is associated with an elliptical host galaxy .
it also had a low - luminosity soft gamma - ray extension of the short hard gamma - ray component ( which would have been missed by batse ) , and it had an interesting x - ray afterglow extending beyond @xmath34 s @xcite .
the soft gamma - ray extension , lasting up to 200 s , when extrapolated to the x - ray range overlaps well with the beginning of the xrt afterglow , which between 100 and 300 s has @xmath104 , followed by a much steeper drop @xmath105 out to @xmath106 , then a more moderate decay @xmath107 .
an unexpected feature is a strong flare peaking at @xmath108 s , whose energy is 10% of the prompt emission , while its amplitude is a 10 times increase over the preceding slow decay . with about a half dozen of reasonably identified objects ,
the distribution of shorts bursts in redshift space and among host galaxy types , including fewer spiral / irregulars and more ellipticals , is typical of old population progenitors , such as neutron star binaries or black hole - neutron star binaries @xcite .
the main challenges posed by the short burst afterglows are the relatively long , soft tail of the prompt emission , and the strength and late occurrence of the flares
. a possible explanation for the extended long soft tails ( @xmath109 ) may be that the compact binary progenitor is a black hole - neutron star system @xcite , for which analytical and numerical arguments ( @xcite , and references therein ) suggest that the disruption and swallowing by the black hole may lead to a complex and more extended accretion rate than for double neutron stars .
the flares , for which the simplest interpretation might be as refreshed shocks ( which would be compatible with a short engine duration @xmath110 s , for a lorentz factor distribution ) , requires the energy in the slow material to be at least ten times as energetic as the fast material responsible for the prompt emission , for the grb 050724 flare at @xmath111 s. the rise and decay times are moderate enough for this interpretation .
another interpretation might be an accretion - induced collapse of a white dwarf in a binary , leading to a flare when the fireball created by the collapse hits the companion @xcite , which might explain moderate energy one - time flares .
however , for repeated , energetic flares , as also in the long bursts , the total energetics are easier to satisfy if one postulates late central engine activity ( lasting at least half a day ) , containing @xmath45 of the prompt fluence @xcite . a possible way to produce this might be temporary choking up of an mhd outflow @xcite ( c.f .
@xcite ) , which might also imply a linear polarization of the x - ray flare @xcite .
such mhd effects could plausibly also explain the initial @xmath99 s soft tail .
however , a justification for substantial @xmath112 s features remains so far on tentative grounds .
the similarity of the x - ray afterglow light curve with those of long bursts is , in itself , an argument in favor of the prevalent view that the afterglows of both long and short bursts can be described by the same paradigm , independently of any difference in the progenitors .
this impression is reinforced by the fact that the x - ray light curve temporal slope is , on average , that expected from the usual forward shock afterglow model , and that in two short bursts ( so far ) there is evidence for what appears to be a jet break @xcite .
however , while similar to zeroth order , the first order differences are revealing : the average isotropic energy is factor @xmath99 smaller , while the average jet opening angle ( based on two breaks ) is a factor @xmath5 larger @xcite . using the standard afterglow theory ,
the bulk lorentz factor decay can be expressed through @xmath113 , where @xmath114 , @xmath115 is the external density in units of @xmath116 , and @xmath117 is the isotropic equivalent energy in units of @xmath118 ergs .
if the jet break occurs at @xmath119 the jet opening angle and the total jet energy @xmath120 are _
j= 9^o ( n_o / e_50)^1/8 t_d , br^3/8 , e_j = _ j^2 e ~10 ^ 49 n_o^1/4 ( e_50 t_d , br)^3/4 erg .
[ eq : short ] for the first two well studied afterglows grb 050709 and grb 050724 , together with the standard afterglow expressions for the flux level as a function of time before and after the break , this leads to fits @xcite which are not completely determined , allowing for grb050709 either a very low or a moderately low external density , and for grb050724 a moderately low to large external density .
the main uncertainty is in the jet break time , which is poorly sampled , and so far mainly in x - rays .
as brighter short bursts are occasionally detected , e.g. grb 051221a , the chances of tighter constraints on jet breaks should increase .
optical / uv afterglows have been detected with the swift uvot telescope in roughly half the bursts for which an x - ray afterglow was seen . for a more detailed discussion of the uvot afterglow observations
see @xcite .
of particular interest is the ongoing discussion on whether dark grb " are really optically deficient , or the result of observational bias @xcite .
another puzzle is the report of a bimodal intrinsic brightness distribution in the rest - frame r - band @xcite .
this suggests possibly the existence of two different classes of long bursts , or at least two different types of environments .
compared to a few years ago , a much larger role is being played by ground - based robotic optical follow - ups , due to the increased rate of several arc - second x - ray alerts from xrt , and the larger number of robotic telescopes brought on - line in the last years .
for the most part , these detections have yielded optical decays in the @xmath121 few 100 s range , initial brightness @xmath122 and temporal decay slopes @xmath123 previously associated with the evolution of a forward shock @xcite , while in a few cases a prompt optical detection was achieved in the first 12 - 25 s. the most exciting prompt robotic ir detection ( and optical non - detection ) is that of grb 050904 @xcite .
this object , at the unprecedented high redshift of @xmath6 @xcite , has an x - ray brightness exceeding for a day that of the brightest x - ray quasars @xcite , and its o / ir brightness in the first 500 s ( observer time ) was comparable to that of the extremely bright ( @xmath35 ) optical flash in grb 990123 , with a similarly steep time - decay slope @xmath124 @xcite . such prompt , bright and steeply decaying optical emission is expected from the reverse shock as it crosses the ejecta , marking the start of the afterglow @xcite .
however , aside from the two glaring examples of 990123 and 05094 , in the last six years there have been less than a score of other prompt optical flashes , typically with more modest initial brightnesses @xmath125 .
there are a number of possible reasons for this paucity of optically bright flashes , if ascribed to reverse shock emission .
one is the absence or weakness of a reverse shock , e.g. if the ejecta is highly magnetized @xcite .
a moderately magnetized ejecta is in fact favored for some prompt flashes @xcite .
alternatively , the deceleration might occur in the thick - shell regime ( @xmath126 .
( [ eq : tag ] ) , which can result in the reverse shock being relativistic , boosting the optical reverse shock spectrum into the uv @xcite .
another possibility , for a high comoving luminosity , is copious pair formation in the ejecta , causing the reverse shock spectrum to peak in the ir @xcite . since both grb 990123 and grb 050904 had @xmath127 erg , among the top few percent of all bursts , the latter is a distinct possibility , compatible with the fact that the prompt flash in grb 050904 was bright in the ir i - band but not in the optical . on the other hand
, the redshift @xmath6 of this burst , and a ly-@xmath128 cutoff at @xmath129 nm would also ensure this ( and grb 990123 , at @xmath130 , was detected in the v - band ) .
however , the observations in these two objects but a suppression in lower @xmath131 objects appears compatible with having a relativistic ( thick shell ) reverse shock with pair formation . | the rapid follow - up of gamma - ray burst ( grb ) afterglows made possible by the multi - wavelength satellite _ swift _ , launched in november 2004 , has put under a microscope the grb early post - burst behavior , this is leading to a significant reappraisal and expansion of the standard view of the grb early afterglow behavior , and its connection to the prompt gamma - ray emission .
in addition to opening up the previously poorly known behavior on minutes to hours timescales , two other new pieces in the grb puzzle being filled in are the the discovery and follow - up of short grb afterglows , and the opening up of the @xmath0 redshift range .
we review some of the current theoretical interpretations of these new phenomena .
address = dept . of astronomy & astrophysics and
dept . of physics , pennsylvania state university ,
university park , pa 16802 , usa |
recently , a new breed of topic models , dubbed counting grids ( cg ) @xcite , has been shown to have advantages in unsupervised learning over previous topic models , while at the same time providing a natural representation for visualization and user interface design @xcite .
cg models are _ generative _ models based on a grid of word distributions , which can best be thought of as the grounds for a massive venn diagram of documents .
the intersections among multiple documents ( bags of words ) create little intersection units with a very small number of words in them ( or rather , a very sparse distribution of the words ) .
the grid arrangement of these sparse distributions , which we will refer to here as _ microtopics _ , facilitates fast cumulative sum based inference and learning algorithms that chop up the documents into much smaller constitutive pieces than what traditional topic models typically do .
for example , fig .
[ fig : fig0 ] shows a small part of such a grid with a few representative words with greatest probability from each microtopic .
each of the science magazine abstracts used to train this grid is assumed to have been generated from a group of microtopics found in a single 4 @xmath0 4 window with equal weight given to all component microtopics .
thus , each microtopic can be 16 times sparser than the set of documents grouped into the window .
a document may share a window with another very similar document , but it is also mapped so that it only partially overlaps with a window that is the source for a set of slightly less related documents .
the varying window overlap literally results in a varying overlap in document themes .
this modeling assumption results in a trained grid where nearby microtopics tend to be related to each other as they are often used together to generate a document .
consider , e.g. , the lower right 4@xmath04 window in fig .
[ fig : fig0 ] . the word distributions in these 16 cells are such that a variety of science papers on evidence of ancient life on earth could be generated by sampling words from there .
( note that each cell , though of very low entropy , contains a distribution over the entire vocabulary . ) in the posterior distribution , this window is by far the most likely source for an article on a bizarre microorganism that produced nitrogen in cretaceous oceans . in the 4@xmath04 window
two cells to the left of this example we find mapped a variety of articles on even more ancient events on earth , e.g. on how sulfur isotopes reveal a deep mantle storage of ancient crust .
but there we also start to see words which increase the fit for articles that describe similar events on other planets .
further movement to the left gets us away from the earth and into astronomy . to demonstrate the refinement of the microtopics compared to topics from a typical topic model , the color labeling of the grid was created so as to reflect the kullback-leibler ( kl ) divergence of the individual microtopics to the topics trained on the same data through latent dirichlet allocation ( lda ) .
the lda topics , hand - labeled after unsupervised training , correspond to fairly broad topics , while the cg represents the data as a group of slowly evolving microtopics .
for example , all the yellow coded microtopics map to the `` physics '' lda topic , but they occupy a contiguous area in which from left to right the focus slowly shifts from electromagnetism and particle physics to material science . furthermore , it is interesting to see the microtopics that occupy the boundaries between coarser topics that lda model found , capturing the links among astronomy , physics and biology .
it is immediately evident that the 2d cgs can have great use in data visualization , though the model can be trained for arbitrary dimensionality @xcite .
these models combine topic modeling and data embedding ideas in a way that facilitates intuitive regularization controls and allows creation of much larger sets of organized sparse topics .
furthermore , they lend them selves to elegant visualization and browsing strategies , and we encourage the reader to see the example http://research.microsoft.com/en-us/um/people/jojic/cgbrowser.zip
. however , the existing em algorithm for cg learning is prone to local minima problems which occasionally lead to under performance @xcite .
in addition , no direct testing of the microtopic coherence has been performed to date , which makes it unclear if they are meaningful outside their windowed grouping .
after all , a variety of sophisticated topic models have been developed and tested by the research community , but lda seems to still beat them often in practice .
e.g. , [ 16,17 ] raise doubts that various reported perplexity improvements over the basic lda model are meaningful as they are sensitive to smoothing constants in the model , and also fail to translate to improvements in human judgement of topic quality . in fact , lda usually outperforms more complex models on tasks that involve human judgement , which may be the main reason why practitioners of data science prefer this basic model to others @xcite .
here we develop hierarchical versions of cg models , which in our experiments produced embeddings of considerably higher quality .
we show that layering into deeper architectures primarily aids in avoiding bad local minima , rather than increasing representational capacity : the trained hierarchical model can be collapsed into an original counting grid form but with a much higher likelihood compared to the grids fit to the same data using em with random restarts .
the better data fit then translates into quantitatively better summaries of the data , as shown in numerical experiments as well as human evaluations of microtopics obtained through crowdsourcing .
[ fig : fig0 ]
[ [ the - ccg - grids ] ] the ( c)cg grids @xcite : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the basic counting grid @xmath1 @xcite is a set of distributions on the @xmath2-dimensional toroidal discrete grid @xmath3 indexed by @xmath4 . the grids in this paper are bi - dimensional and typically from @xmath5 to @xmath6 in size .
the index @xmath7 indexes a particular word in the vocabulary @xmath8 $ ] .
thus , @xmath9 is the probability of the word @xmath7 at the @xmath2-dimensional discrete location @xmath10 , and @xmath11 at every location on the grid .
the model generates bags of words , each represented by a list of words @xmath12 with each word @xmath13 taking an integer value between @xmath14 and @xmath15 .
the modeling assumption in the basic cg model is that each bag is generated from the distributions in a single window @xmath16 of a preset size , e.g. , @xmath17 .
a bag can be generated by first picking a window at a @xmath2-dimensional location @xmath18 , denoted as @xmath19 , then generating each of the @xmath20 words by sampling a location @xmath21 for a particular microtopic @xmath22 uniformly within the window , and finally by sampling from that microtopic .
because the conditional distribution @xmath23 is a preset uniform distribution over the grid locations inside the window placed at location @xmath18 , the variable @xmath21 can be summed out@xcite , and the generation can directly use the grouped histograms @xmath24 where @xmath25 is the area of the window , e.g. 25 when 5@xmath05 windows are used .
in other words , the position of the window @xmath18 in the grid is a latent variable given which we can write the probability of the bag as @xmath26 as the grid is toroidal , a window can start at any position and there is as many @xmath27 distributions as there are @xmath28 distributions .
the former will have a considerably higher entropy as they are averages of many @xmath28 distributions .
although the basic cg model is essentially a simple mixture assuming the existence of a single source ( one window ) for all the features in one bag , it can have a very large number of ( highly related ) choices @xmath27 to choose from .
topic models @xcite , on the other hand , are admixtures that capture word co - occurrence statistics by using a much smaller number of topics that can be more freely combined to explain a single document .
componential counting grids ( ccg ) @xcite combine these ideas , allowing multiple groups of broader topics @xmath27 to be mixed to explain a single document .
the entropic @xmath27 distributions are still made of sparse microtopics @xmath28 in the same way as in cg so that the ccg model can have a much larger number of topics than an lda model without overtraining .
more precisely , each word @xmath13 can be generated from a different window , placed at location @xmath29 , but the choice of the window follows the same prior distributions @xmath30 for all words . within the window at location @xmath29
the word comes from a particular grid location @xmath31 , and from that grid distribution the word is assumed to have been generated .
the probability of a bag is now @xmath32 in a well - fit ccg model , each data point has an inferred @xmath33 distribution that usually hits multiple places in the grid , while in a cg , each data point tends to have a rather peaky posterior location distribution because the model is a mixture .
both models can be learned efficiently using the em algorithm because the inference of the hidden variables , as well as updates of @xmath28 and @xmath27 can be performed using summed area tables @xcite , and are thus considerably faster than most of the sophisticated sampling procedures used to train other topic models
. an intriguing property of these models is that even on a @xmath34 grid with @xmath35 microtopics @xmath28 and just as many grouped topics @xmath27 , there is no room for too many independent groups . with a window size @xmath36 , for example , we can place only @xmath37 windows without overlap , and the remaining windows are overlapping the pieces of these 16 . the ratio between grid and window size is referred to as the _ capacity _ of the model , and the training set size necessary to avoid overtraining the model only needs to be 1 - 2 orders of magnitude above the capacity number .
thus a grid of 1024 microtopics may very well be trainable with thousands of data points , rather than 100s of thousands that traditional topic models usually require for that many topics .
[ fig : dig ] [ [ raw - image - embedding - using - ccgs ] ] raw image embedding using ( c)cgs : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in previous applications of cg models to computer vision , images were represented as spatially disordered bags of features .
we experimented with embedding raw images with full spatial information preserved , and we present this here as we feel that the image data helps in illuminating the benefits of hierarchical learning .
an image described by a full intensity function @xmath38 could be considered as a set of words , each word being an image location @xmath39 . for a @xmath40 image
, we have a vocabulary of size @xmath41 .
the number of repetitions of word @xmath42 is then set to be proportional to the intensity i(x , y ) .
( in case of color images , the number of features is simply tripled with each color channel treated in this way ) . in other words , an unwrapped image is considered to be a word ( location ) histogram . @xmath28 and @xmath27 distributions
can then also be seen as images , as they provide weights for different image locations .
if we tile the image representations of these distributions we get additional insight into cgs as an embedding method .
[ fig : dig ] shows a portion of a @xmath43 grid trained on 2000 mnist digits assuming a @xmath44 window averaging . to illustrate the generative model , in c )
we show the partial window sums for two overlapping windows over @xmath28 .
the green and blue areas form a window that generates a version of digit 3 , which can be seen at the top left of this portion of the @xmath27 grid ( panel b ) ) .
the blue and red , on the other hand , combine into a window that represent a digit 2 at the position ( 3,3 ) in panel b ) .
partial sums for green , blue and red areas are shown in c ) and these partial sums , color coded and overlapped are also illustrated in d ) .
careful observation of b ) or the full grid in the appendix , demonstrates the slow deformation of digits from one to another in the @xmath27 distributions .
the appendix has additional examples of image dataset embedding , including rendered 3d head models and images of bold eagles retrieved by internet search .
the cg @xmath28 distributions shown here look like little strokes , while @xmath27 distributions are full digits .
the ccg model , on the other hand , combines multiple @xmath27 distributions to represent a single image , and so @xmath27 looks like a grid of strokes fig .
[ fig : digits2]a , while @xmath28 distributions are even sparser .
[ [ hierarchical - grids ] ] hierarchical grids : + + + + + + + + + + + + + + + + + + + by learning a model in which microtopics join forces with their neighbors to explain the data , ( c-)cg models tend to exhibit high degrees of relatedness of nearby topics . as we slowly move away from one microtopic , the meaning of the topics we go over gradually shifts to related narrowly defined topics as illustrated by fig . [
fig : fig0 ] ; this makes these grids attractive to hci applications .
but this also means that simple learning algorithms can be prone to local minima , as random initializations of the em learning sometimes result in grouping certain related topics into large chunks , and sometime breaking these same chunks into multiple ones with more potential for suboptimal microtopics along boundaries .
to illustrate this , in fig .
[ fig : digits2]a we show a @xmath43 grid of strokes @xmath27 ( eq . [ eq : h ] ) learned from 2000 mnist digits using a ccg model assuming a @xmath45 window averaging .
nearby features @xmath27 are highly related to each other as they are the result of adding up features in overlapping windows over @xmath28 ( which is not shown ) .
ccg is an admixture model , and so each digit indexed by @xmath46 has a relatively rich posterior distribution @xmath47 over the locations in the grid that point to different strokes @xmath27 .
fig : digits2 ] , we show one of the main principal components of variation in @xmath48 as an image of the size of the grid . for three peaks there , we also show @xmath27-features at those locations .
the combination of these three sparse features creates a longer contiguous stroke , which indicates that this longer stroke is often found in the data .
thus , the separation of these features across three distant parts of the map is likely a result of a local minimum in basic em training . to transfer this reasoning to text models ,
consider the 5th cell in the first row in fig .
[ fig : fig0 ] with words hiv , aids , and the blue cell in the middle of the last column with words selection , adaptive .
the separation of these two things in faraway locations may very well be a result of a local minimum , which could be detected if location posteriors exhibit correlation .
this illustration points to an idea on how to build better models .
the distribution over locations @xmath18 that a data point @xmath46 maps to ( a posteriori ) could be considered a new representation of the data point ( digit in this case ) , with the mapped grid locations considered as features , and the posterior probabilities for these locations considered as feature counts .
thus another layer of a generative model can be added to generate the locations in the grid below , fig .
[ fig : gm]c - d
. it is particularly useful to use another microtopic grid model as this added layer , because of the inherent relatedness of the nearby locations in the grid
. the layer above can thus be either another admixture grid model ( componential counting grid - ccg ) , or a mixture ( cg ) , and this layering can be continued to create a deep model .
as cg is a mixture model , it terminates the layering : its posterior distributions are peaky and thus uncorrelated .
however , an arbitrary number of ccgs can be stacked on top of each other in this manner , terminating on top with a cg layer to form a hierarchical cg ( hcg ) model , or terminating in a ccg layer to form a hierarchical ccg ( hccg ) model . in each layer , the pointers to features below are grouped , which should result in creating a contiguous longer stroke as discussed above in a grid cell that contains a combination of pointers to the lower layers . for the sake of brevity
, we only derive the hcg learning algorithm with a single intermediate ccg layer . the extension to hccg and higher order hierarchies
is reported in the appendix .
variational inference and learning procedure for counting grid - based models utilizes cumulative sums and is only slower than training an individual ( c)cg layer by a factor proportional to the number of layers . the graphical model for hcg
is shown in fig .
[ fig : gm]c , where location variables pointing to grids in different layers have the same name , @xmath18 but carry a disambiguating superscript . _ to avoid superscripts in the equations below , we renamed the cg s location variable from @xmath49 to @xmath50 and dropped the superscript
@xmath51 in the layer above_. the bottom ccg layer follows @xmath52
@xmath53 the latter is a pre - set distribution over the grid locations , uniform inside @xmath54 . instead of the prior @xmath55 the locations
are generated from a top layer cg , indexed by @xmath50 ( @xmath56 in the figure ) , @xmath57 this equation also shows that the lower - levels grid locations act as observations in the higher level .
we use the fully factorized variational posterior @xmath58 to write the negative free energy @xmath59 bounding the non - constant part of the loglikelihood of the data as @xmath60 we maximize @xmath59 with the em algorithm which iterates e- and m - steps until convergence .
e : @xmath61 the m step re - estimates the model parameters using these updatedposteriors : @xmath62 \nonumber \\
\pi_{{\tiny cg},{{\bf i}}}({{\bf l } } ) \hspace{-0.3 cm } & \propto & \hspace{-0.1 cm } \hat{\pi}_{{\tiny cg},{{\bf i}}}({{\bf l } } ) \cdot \sum_{t , n } q^t(\ell_n = { { \bf l}})\hspace{-0.cm}\cdot\hspace{-0.2 cm } \sum_{{{\bf k}}| { { \bf i}}\in w_{{\bf k } } } \hspace{-0.2cm}\frac { q^t ( k_n = { { \bf i}})}{\hat{h}_{{\tiny cg},{{\bf i}}}({{\bf l } } ) } \nonumber \end{aligned}\ ] ] where the last ( cg ) update is performed analogous with @xcite .
interestingly , training these hierarchical models stage by stage , reminiscent of deep models where such incremental learning was practically useful @xcite .
+ although it has been shown that a deep neural network can be compressed into a shallow broader one through post training @xcite , the stacked ( c-)cg models can be collapsed mathematically . in this sense we can view hcg and hccg as _ hierarchical learning algorithms _ for cg and ccg , which are easier to visualize than deeper models .
for example , for hcg in fig .
[ fig : gm]c - d , it is straightforward to see that the following grid defined over the original features @xmath63 , @xmath64 can be used as a single layer grid that describes the same data distribution as the two - layer model are the grouped microtopics in the window @xmath65 - eq .
[ eq : h ] ] . however , the grids estimated from the hierarchical models should be more compact as the scattered groups of features are progressively merged in each new layer .
_ learning in hierarchical models is thus more gradual and results in better local maxima , and we show below that the results are far superior to regular em learning of the collapsed cg or ccg models . _
in all the experiments we used models with two extra layers , although , in some experiments , we found that three levels worked slightly better . in general , the optimal number of layers will depend on the particular application .
[ [ likelihood - comparison ] ] likelihood comparison : + + + + + + + + + + + + + + + + + + + + + + in the first experiment we compared the local maxima on models learned using the ( full ) mnist data set .
the two layer hcg model was first pre - trained stage - wise as , e.g. , @xcite , by training the higher level on the posterior distribution from the lower level as the input .
then , the model was refined by further variational em training .
the procedure is repeated 20 times with different random initializations to produce twenty hierarchical models . as discussed above
, these models can be collapsed to a cg model by integrating out intermediate layers ( [ eq : collapse ] ) .
these models were then compared with twenty models learned by directly learning cg models through previously published standard em learning algorithm starting from twenty random initializations . despite being collapsible to the same mathematical form ,
the hcg models consistently produced higher likelihood than the cg models directly learned using the standard method .
_ in fact , each cg model created by collapsing one of the learned hcg models had log likelihood at least two standard deviations above the highest log likelihood learned by basic em ( p - value @xmath66 ) . _
both learning approaches used the computation time equivalent to 1000 iterations of standard em , which was more than enough for convergence .
[ [ document - classification ] ] document classification : + + + + + + + + + + + + + + + + + + + + + + + + next we ran test to see if the increased likelihood obtainable with a better learning algorithm translates into increased quality of representation when posterior distributions for individual text documents are considered as features in classification tasks .
we considered the 20-newsgroup dataset ( 20n ) and the mastercook dataset ( mc ) composed by 4000 recipes divided in 15 classes .
previous work @xcite reduced 20-newsgroup dataset into subsets with varying similarities and we considered the hardest subset composed by posts from the very similar newsgroups ` comp.os.ms-windows ` , ` comp.windows.x ` and ` comp.graphics ` .
we considered the same complexities as in @xcite , using 10-fold cross validation and classified test document using maximum likelihood .
results for both datasets are shown in tab . [
tab : doccl ] .
.document classification .
when bold , hierarchical grids outperformed the basic grids with statistical significance ( hcg p - value @xmath672.01e-4 , hccg p - values @xmath68 1e-3 ) .
`` linsvm '' stands for linear support vector machines which we reported as baseline . [ cols="<,^,^,^,^,^",options="header " , ] [ tab : doccl ] [ [ evaluation - of - microtopic - quality - using - quantitative - measures - related - to - the - use - in - visualization - and - indexing ] ] evaluation of microtopic quality using quantitative measures related to the use in visualization and indexing : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we evaluated the coherence and the clarity of the microtopics comparing the collapsed ( 2 layers ) hierarchical grids - hcg and hccg with regular grids @xcite , latent dirichlet allocation ( lda ) @xcite , the correlated topic model ( ctm ) @xcite which allows to learn a large set of correlated topics and few non - parametric topic models @xcite .
+ generative models are often evaluated in terms of perplexity .
however different models , even different learning algorithms applied to the same model , are very difficult to compare @xcite and better perplexity does not always indicate better quality of topics as judged by human evaluators @xcite . on the other hand , the subjective evaluation of topic quality
is highly related to measures that have to do with data indexing , e.g. quality of word combinations when used for information retrieval .
thus we start with a novel evaluation procedure for topic models which is strongly related to information indexing and then show that we obtain similar evaluation results when we use human judgement . in the following experiments
, we considered a corpus @xmath69 composed of science magazine reports and scientific articles from the last 20 years .
this is a very diverse corpus similar to the one used in @xcite . as preprocessing step , we removed stop - words and applied the porters stemmer algorithm @xcite .
we considered grids of size @xmath70 and @xmath43 fixing the window size to @xmath71 .
( previous literature showed that counting grids are only sensitive to the ratio between grid and window area , as long as windows are sufficiently big . ) we varied number of topics for lda and ctm in @xmath72 .
for each complexity we trained 5 models starting with different random initializations and we averaged the results . in each repetition
, we considered a random third of this corpus , for total of roughly @xmath73 documents , @xmath74 different words and more than @xmath75 tokens
. to evaluate ( micro)topics , we repetitively sampled _
k_-tuples of words and checked for consistency , diversity and clarity of the indexed content . in the following , we describe the procedure used for evaluating grids .
an equivalent procedure was used to evaluate other topic models for comparison .
+ to pick a tuple @xmath76 of @xmath77 words , we sampled a grid location @xmath78 .
then , we repetitively sampled the microtopic @xmath79 to obtain the words in the tuple @xmath80 .
we did not allow repetitions of words in the tuple .
we considered @xmath81 different @xmath82-tuples , not allowing repeated tuples .
+ then we checked for consistency , diversity and clarity of content indexed by each tuple .
the * consistency * is quantified in terms of the average number of documents from the dataset that contained _ all _ words in @xmath76 .
the * diversity * of indexed content is illustrated through the cumulative graph of acquired unique documents as more and more @xmath77-tuples are sampled and used to retrieve documents containing them .
as this last curve depends on the sample order , we further repeated the process 5 times for a total of 25k different samples .
finally the * clarity * @xcite , measures the ambiguity of a query with respect to a collection of documents and it has been used to identify ineffective queries , on average , without relevance information .
formally , the query clarity is measured as the entropy between the n - tuple and the language model @xmath83 ( unigram distributions ) as @xmath84 where @xmath85 .
we estimated the likelihood of an individual document model generating the tuple @xmath86 and obtain @xmath87 using uniform prior probabilities for documents that contains a word in the tuple , and a zero prior for the rest .
finally , to estimate @xmath88 we employed montecarlo sampling .
+ results are illustrated in fig.[fig : results ] and should be appreciated by looking at all three measures together , as some can be over - optimized at the expense of others .
the diversity curve that consistently grows as more tuples are sampled indicates that the sampled tuples belong to different subsets of the data , and are thus discriminative in segmenting the data into different clusters .
the average tuple consistency , on the other hand , demonstrates that the sampled tuples do occur in large chunks of the data , demonstrating that the induced clusters are of significant size .
the clarity measure shows that the clusters made of texts retrieved using different tuples have clear differentiation from the rest of the dataset in usage of all the words in the dictionary .
we report results for the @xmath89 grids and the best result of lda and ctm which peaked respectively at 80 and 60 topics .
results for other grid sizes can be found in the additional material ; they are stable across complexities with slightly better performances for larger grids .
+ all grid models show good consistency of words selected as they are optimized so that documents words map into overlapping windows . through positioning and intersection of many related documents
the words end up being arranged in a fine - grained manner so as to reflect their higher - order co - occurrence statistics .
hierarchical learning greatly improved the results despite the fact that hccg and hcg can be reduced to ( c)cgs through marginalization ( [ eq : collapse ] ) .
+ overall hccg strongly outperformed all the methods , especially with a total gain of 0.5 bits on clarity , which is around third of the score for lda / ctm . despite allowing for correlated topics that enable ctm to learn larger topic models , ctm trails lda in these graphs as topics were over expanded .
we also considered non - parametric topic models such as `` dilan '' @xcite and the hierarchical dirichlet process @xcite but their best results were poor and we did not reported them in the figure .
to get an idea , both models only indexed 25% of the content after 5000 2-tuples samples and had a clarity lower of 0.7 - 1.2 bits than other topic models .
[ fig : intrusion ] [ [ human - judgments - of - topic - coherence ] ] human judgments of topic coherence : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we next tested the quality of the inferred topics . topic coherence is often measured based on co - occurrence of the top @xmath90 words per topic . while good as a quick sanity check of a single learned model , when this measure is used to compare models , it will favor models that lock onto top themes and distribute the rest of the words in the tails of the topic distributions .
the lda models usually have a large drop off in topic coherence when the number of topics is increased to force the model to address more correlations in the data . indeed ,
using this measure , lda topics outperform cg topics in case of small models . but
as the number of topics grows , the microtopics trained by hcg significantly outperform both lda and cg ( see the appendix ) .
a more interesting measure of topic quality , which not only depends on individual topic coherence but also on meaningful separation of different topics , requires human evaluation of _ word intrusions_. in a word intrusion task @xcite , six randomly ordered words are presented to a human subject who then guesses which word is an outlier . in the original procedure a target topic is randomly selected and then the five words with _
highest _ probability are picked .
then , an intruder is added to this set .
it is selected at random from the low probability words of the target topic that have high probability in some other topic . finally the six words are shuffled and presented to the subject .
if the target topic shows a lack of coherence or distinction from the intruding topic , the subject will often fail to correctly identify the intruder .
this task is again geared towards only getting the top words right in a topic model and ignoring the rest of the distribution , which makes it unsuitable to comparison with microtopic models which attempt to extract much more correlation from the data .
thus instead of picking the top words from each topic , we sampled the words from the target topic to create the in - group . after sampling the location of a microtopic from the grid @xmath78 , we picked three randomly chosen words from @xmath79 or from the small groups of microtopics in the window of size 2@xmath02 , and 3@xmath03 around @xmath78 ( the latter is equivalent to computing the window distributions @xmath27 using windows of smaller size than the ones used in training and should give us the indication if the granularity assumed in the window size was exaggerated : if it is then averaging of nearby topics should significantly reduce the noise due to forced topic splitting ) .
for each of these groups we choose the intruder word using the standard procedure .
if in this harder task humans can identify intruders better for microtopic models than for lda models , this would indicate that the microtopics are not simply random subsamples of broader topics captured in @xmath27 and similar in entropy to lda topics .
they would be a meaningful breakup of broad topics into finer ones .
we compared lda ( known to performed better than ctm on intrusion tasks @xcite ) , hcg , and hccg , on randomly crawled 10k wikipedia articles and used amazon mechanical turk ( 24000 completed tasks from 345 different people ) .
the trained grids were of size 32 @xmath0 32 and the windows 5 @xmath0 5 .
the optimal lda size was chosen using likelihood crossvalidation over the range of complexities as in the previous experiments ( the peak performance there was at 80 topics ) .
results are shown in fig.[fig : intrusion ] as a function of the euclidean distance on the grid of the intruder word from the topic .
hccg outperformed lda ( p - values for the 3 tasks 1.20e-11 , 1.88e-5 , 2.97e-05 ) and hcg (
p - values for the 3 tasks 3.97e-18 , 1.01e-11 , 3.14e-19 ) indicating that learning microtopics is possible with a good algorithm .
overall , _ users were able to solve correctly 71% of hccg problems and only 58% of lda problems_. interestingly , the performance of hccg and hcg does not seem to depend on the distance of the intruder word : even picking intruder word from a very close location rather than from a far away one lead to no additional confusion for the user .
this shows that hccg chops up the data into meaningful microtopics which are then combined into a large number of groups @xmath27 that do not over broaden the scope .
hccg and hcg also outperformed respectively cg and ccg ( see the appendix ) .
[ [ learning - to - separate - mixed - digits . ] ] learning to separate mixed digits .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + finally , we show that an hccg model can be used to perform a task that eludes most unsupervised _ and _ supervised models .
we created a set of 10000 @xmath91 images , each containing two different mnist digits overlapped , fig .
[ fig : digitmix ] .
we trained an hccg model consisting of five @xmath89 layers on this data stagewise by feeding @xmath92 from one layer to the next .
windows of size 5 @xmath0 5 were used in all layers . from layer to layer , the new representations of the image consist of growing combinations of low level features @xmath27 from the bottom layer ( sparseness of which is similar to fig .
[ fig : digits2]a ) .
the hierarchical grouping is further encouraged by simply smoothing @xmath93 with a @xmath94 gaussian kernel with deviation of 0.75 , before feeding it to the next layer ( this is motivated by the fact that nearby features in @xmath27 are related and so if two distant locations should be grouped , so should those locations neighbors ) . once the model is collapsed to a single hccg grid the components no longer look like short strokes but like whole digits , mostly free of overlap : the model has learned to approximately separate the images into constitutive digits .
reasoning on overlapping digits even eludes deep neural networks trained in a supervised manner , but here we did not use the information about which two digits are present in each of the training images .
we show that with new learning algorithms based on a hierarchy of ccg models , possibly terminated on the top with a cg , it is possible to learn large grids of sparse related microtopics from relatively small datasets .
these microtopics correspond to intersections of multiple documents , and are considerably narrower than what traditional topic models can achieve without overtraining on the same data . yet , these microtopics are well formed , as both the numerical measures of consistency , diversity and clarity and the user study on 345 mechanical turkers show .
another approach to capturing sparse intersections of broader topics is through product of expert models , e.g. rbms @xcite , which consist of relatively broad topics but model the data through intersections rather than admixing .
rbms are also often stacked into deep structures . in future work
it would be interesting to compare these models , though the tasks we used here would have to be somewhat changed to focus on the intersection modeling , rather than the topic coherence ( as this is not what rbm topics are optimized for ) .
hccg and hcg models have a clear advantage in that it is easy to visualize how the data is represented , which is useful both to end users in hci applications , and to machine learning experts during model development and debugging .
another parallel between the stacks of ccgs and other deep models is that the uniform connectivity of units is directly enforced through window constraints , rather than encouraged by dropout .
finally , in this specific context we illustrate a broader phenomenon that requires more methodical and broader treatment by the machine learning community .
a more complex ( deeper ) model showed here large advantages in terms of training likelihood , but these advantages were _ not _ due to the expanded parameter space , because the resulting model is equivalent to a collapsed single layer model .
rather than being a reflection of increased representational abilities of the model , better likelihoods were thus the result of better fitting algorithm that consists of training a deep model ( and then collapsing it into a simpler but equivalent parameterization ) .
similar phenomena are likely regularly encountered elsewhere in machine learning , but not always recognized as such , as in the absence of the full knowledge of the extrema of the fitting criterion , an increase in performance is often inappropriately ascribed to better modeling rather than better model fitting . | the counting grid is a grid of _ microtopics _ , sparse word / feature distributions .
the generative model associated with the grid does not use these microtopics individually , but in predefined groups which can only be ( ad)mixed as such .
each allowed group corresponds to one of all possible overlapping rectangular windows into the grid .
the capacity of the model is controlled by the ratio of the grid size and the window size .
this paper builds upon the basic counting grid model and it shows that hierarchical reasoning helps avoid bad local minima , produces better classification accuracy and , most interestingly , allows for extraction of large numbers of coherent microtopics even from small datasets .
we evaluate this in terms of consistency , diversity and clarity of the indexed content , as well as in a user study on word intrusion tasks .
we demonstrate that these models work well as a technique for embedding raw images and discuss interesting parallels between hierarchical cg models and other deep architectures . |
following the pioneering work of @xcite , many studies have confirmed that tidal disturbance following galaxy interactions is an efficient trigger of star formation in galaxies ( e.g. * ? ? ?
* ; * ? ? ?
such star formation frequently takes the form of centrally - concentrated nuclear starbursts @xcite , fuelled by the central concentrations of molecular gas found to occur naturally in simulations of highly - disturbed systems @xcite .
the strength of the link between starbursts and interactions was highlighted by the finding that almost all of the ` ultra - luminous infrared galaxies ' ( ulirgs ) display signs of interactions or mergers @xcite , and by correlations between galaxy - galaxy separations and starburst strength @xcite .
even minor mergers with low - mass companions have been shown through simulations to result in significant nuclear star formation activity @xcite .
several early studies of nuclear starbursts suggested that this star formation might require a top heavy initial mass function ( imf ) , preferentially producing high mass stars @xcite .
there is theoretical support for this suggestion , with simulations showing that an imf weighted to high - mass stars naturally arises in high - density regions , due to feedback processes heating the gas . in a recent study ,
@xcite have demonstrated that such regions should have a high - mass stellar fraction at least 1.7 times larger , and possibly much more , than lower density , more quiescent regions . however , the observational evidence for this variation has to date proved controversial ( see @xcite for a recent review ) .
some studies have found indirect evidence for top - heavy imfs with , for example , @xcite concluding that the nearby starburst galaxy m82 requires an imf biased to high mass stars to explain its emission line ratios and total luminosity .
similar techniques have been used for ngc 3256 ( an ongoing merger with a ` super - starburst ' ) which have again shown indications of a modified imf with an excess of high mass stars @xcite .
@xcite showed that , in order to reproduce the observed colour - luminosity relation of elliptical galaxies , an imf much flatter than that of @xcite needed to be adopted .
@xcite had to employ a top heavy imf for the starbursts powering the distant population of highly luminous submillimetre galaxies in order to explain the number counts of these systems . finally , @xcite studied nine interacting galaxies from the @xmath2 survey and found that highly disturbed systems showed a strongly enhanced infrared luminosity compared to that expected from the x - ray emission , again suggesting the need for a top - heavy imf .
more direct evidence of a variation in imf has been found for the resolved stellar population of the young arches cluster in the galactic centre .
@xcite all find evidence for stellar mass functions weighted towards high - mass stars in this cluster or the general galactic centre region .
such mass functions are parametrized as an imf that is either much flatter than that found by @xcite , or having a higher mass turnover than is found in the function for field stars .
one possible tracer of the imf that has not been fully exploited to date is the relative numbers of core - collapse supernovae ( ccsne ) of different types .
their short progenitor lifetimes and high luminosities make them powerful indicators of recent or ongoing star formation , and indeed they provide the only direct tracer of recent star formation within unresolved stellar populations .
recent advances in the understanding of supernovae and their progenitors raise the possibility that they can provide information on the initial mass function of a young stellar population .
theoretical models of single star progenitors predict that snii should have lower mass progenitors than snib or snic @xcite .
this has received observational support from studies of the strength of association with h@xmath1 emission @xcite , confirming that snii have the lowest mass cc progenitors , but additionally indicating that the snic have still higher mass progenitors than the snib .
the existence of this ii - ib - ic progenitor mass sequence allows information on the imf of the stellar population in the sn environments to be derived from the relative numbers of type ii , ib and ic supernovae .
@xcite studied the distribution of sne events in 32 interacting systems containing 12 known core collapse sne .
they found that the radial distribution of these core collapse events showed a higher concentration towards the nuclear regions of the interacting galaxies when compared to isolated galaxies .
this confirmed the enhanced star formation around the central regions of the systems , but the sample was too small to analyse the separate types of ccsne .
this paper will therefore use a larger sample of local ccsne to explore the imf in nuclear starbursts , resulting from galaxy disturbance , by studying the ratio of type ii / ibc sne in both disturbed and undisturbed host galaxies . throughout this paper , we use ` ibc ' to encompass all sne with classifications of ib , ic or ib / c . the structure of the paper is as follows : in section [ sec : sample ] we will define and discuss the sample used throughout this work . section [ sec : results ] will describe the results on the radial distributions , for disturbed and undisturbed hosts and looking separately at type ii and ibc sne . in section [ sec : disc ]
we discuss the possible interpretations of our results , in terms of metallicity , extinction and imf effects .
finally , section [ sec : conc ] contains a summary of our conclusions .
the sample used in this work consists of 140 local ( recession velocity @xmath36000 km / s ) spiral galaxies , hosts to 178 ccsne ( 110 snii and 68 snibc ) , for which we have h@xmath1 and @xmath0-band observations from the liverpool telescope ( lt ) and isaac newton telescope ( int ) .
( some galaxies do not have usable images in either h@xmath1 or @xmath0-band and have been omitted from the corresponding plots and statistics ; see tables 1 & 2 in the online material ) .
this is the same dataset as was used by @xcite with a small number of subsequent observations .
sne classified as type iib are not included in this sample as they are thought to be transitional objects between snii and snib , with substantially larger progenitor masses ( @xmath425 m@xmath5 ; @xcite ) than typical snii .
a comparison performed on january 21st 2010 with all ccsne host galaxies within the same recession velocity limit in the iau sn catalogue , and where the sne have accurate classifications and positions , showed this sample to be @xmath434% complete for snibc and @xmath418% complete for snii .
the classification of host galaxies as disturbed is purely by visual inspection by the authors and thus is subjective .
galaxies which show signs of tidal tails , definite interaction , double nuclei or strong asymmetry have therefore been classed as disturbed .
the total sample of ccsne is dominated by snii ( 62% of the total ) . when the sample is constrained only to supernovae which lie in disturbed hosts ( 64 ccsne )
this falls to 56% snii , compared to 65% snii in the non - disturbed hosts .
for each of the ccsne in our sample we have calculated the fr(@xmath0 ) and fr(h@xmath1 ) statistics used , and explained fully , in @xcite .
briefly , these represent the fractions of galaxy emission , in the @xmath0-band and h@xmath1 respectively , that lie within the circle or ellipse which contains the sn
. thus fr(@xmath0)@xmath60.0 corresponds to a supernova at the central @xmath0-band peak of the galaxy emission , or closer to this peak than any h@xmath1 emission , in the case of fr(h@xmath1 ) ; whilst [email protected] implies an extreme outlying sn .
if the emission is statistically a good tracer of the parent population of supernovae , the fr values should have a flat distribution with a mean value of 0.5 .
figures 1 and 2 show the distributions of fr(@xmath0 ) values for the ccsne in the present sample , for the undisturbed and disturbed galaxies respectively . in all histograms shown in this paper ,
the upper plot represents the ccsne sample , the middle the type ii sne and the lower snibc . looking first at the overall distributions of ccsne
, there is a clear difference between the disturbed and undisturbed subsets , in the sense that the disturbed galaxies have substantially more ccsne occurring in their central regions , with low fr(@xmath0 ) values .
for example , 36 of the 58 ccsne in the disturbed sample occur within the central 50% of the @xmath0-band light , 62% of the total , compared with 50 out of 112 ( 45% ) in the undisturbed galaxies .
a kolmogorov - smirnov ( ks ) test shows that the chance of the two total ccsne distributions being drawn from the same parent distribution is [email protected] .
thus there is evidence at the 2@xmath7 level that galaxy disturbance correlates with centrally - enhanced star formation and hence the production of an increased central fraction of ccsne .
the most striking aspect of figure 2 is the types of sne that make up this central excess in the disturbed galaxies .
remarkably , given that snibc only comprise 38% of the overall ccsn sample ( 68/178 ) , all 5 of the ccsne coming from the central 10% of the disturbed host galaxy light , and 11 of the 13 coming from the central 20% of the light , are of type ibc . a ks test of the fr(@xmath0 ) distributions for the disturbed galaxy subsample finds [email protected] , indicating a very low probability that the snibc and snii fr(@xmath0 ) values are drawn from the same parent distribution .
the mean values of fr(@xmath0 ) are 0.31 ( 95% confidence limits 0.200.42 ) for the snibc in the disturbed galaxies , compared with 0.51 ( 0.440.59 ) for the snii in the disturbed galaxies .
this is the main observational result from this paper ; the ccsne occurring in the central regions of disturbed galaxies are heavily weighted towards types ib , ic and ib / c .
we will discuss possible interpretations of this in section 4 .
some further statistical tests were also performed on the ccsn distributions shown in figures 1 & 2 .
figure 1 shows that even in the undisturbed galaxies , there is some evidence for a larger fraction of snibc in the central regions , principally due to a central ` hole ' in the radial distribution of snii .
a ks test applied to the snibc and snii distributions shown in figure 1 shows this difference to be only marginal , p=0.082 , and hence clearly less marked than for the disturbed galaxies ; disturbance does seem to play a part in the central concentration of the snibc .
this point was further explored by comparing the snibc distributions for undisturbed and disturbed galaxies , i.e. figure 1 vs. figure 2 ; this did indicate the snibc in disturbed galaxies to be more centrally concentrated , with a ks p value of 0.06 , again of marginal significance .
the mean snibc fr(@xmath0 ) value is 0.48 ( 0.380.57 ) for the snibc in the undisturbed galaxies , again to be compared with 0.31 ( 0.200.42 ) already quoted for the disturbed galaxies . finally for figure 2 , it might be asked whether there is evidence for a _ suppression _ of snibc fraction in the outer regions of these galaxies . however , given the current sample size this can not be determined with any significance .
for example we find 6 snibc in the outer 50% of the light distributions of the disturbed galaxies , but with only 22 ccsne in total from these regions , this is not significantly below the expectation value of 8.4 , based on the snibc / snii ratio for the full sample .
figures 3 & 4 show the distributions of supernova locations relative to the h@xmath1 distributions of their host galaxies .
overall these show the same patterns as figures 1 & 2 , but they do enable one specific issue to be addressed : are the snibc more centrally concentrated than the h@xmath1 light , which is presumably a good tracer of the youngest stellar population ?
figure 4 shows that there is some evidence for this ; the central 10% of the h@xmath1 emission in the disturbed galaxies gives rise to 7 of the 22 snibc in these galaxies .
the mean fr(h@xmath1 ) value for the snibc in disturbed galaxies is 0.33 ( 0.210.45 ) , so this population does seem to be more centrally concentrated than the h@xmath1 emission .
this is not true for the snibc in the undisturbed galaxies , or for the snii in either of the galaxy subsets ; all of these distributions have mean fr(h@xmath1 ) values consistent with 0.5 .
@xcite found a central excess of snibc in a sn - host galaxy sample .
this work has found that this central excess is exaggerated in galaxies which appear disturbed .
a more centrally located distribution of snibc has been suggested previously ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , though previous studies often suffered from low number statistics .
@xcite also find snibc to be more centrally located than snii , however in conflict to our results they do not find the central excess of snibc clearly seen in our data .
an important difference between our work and most other studies in the literature is that our method implicitly normalizes the tests to the measured distributions of different stellar populations ; other studies use distances normalized to isophotal radii .
most of these results have been interpreted as an increase in metallicity of the snibc progenitors , although @xcite also make the suggestion of a shallower imf within the central regions .
studies conducted into active and star - forming galaxies @xcite and seyfert galaxies @xcite have also noted marginal evidence for an increased fraction of both ccsne and specifically snibc within these galaxies when compared to ` normal ' ones .
there are various observational biases which may affect our analysis .
@xcite found a bias in supernova samples , in the sense that it is more difficult to detect sne in the inner regions of distant galaxies .
the sample is also subject to any bias contained within the object selection found in the asiago @xcite and iau sn catalogues . for the asiago and crimea searches , @xcite estimated the number of sne lost due to overexposure combined with the shaw effect , which for the velocity range of our sample is @xmath435% .
another source of bias is the loss of sne in the central regions of galaxies through the large amount of dust obscuration which has been investigated through near infrared studies ( e.g. * ? ? ?
* ; * ? ? ?
such biases should affect all sn types , although the intrinsically fainter sniip @xcite may be rather more likely to be lost through these effects . however , if our results are correct and snibc are more centrally concentrated than snii then recovering all of the lost central sne would lead to an even more exaggerated excess .
one possible source of error is our eyeball classifications of host galaxy disturbance . in future
we plan to quantify this through near - ir observations and use of objective measures of asymmetry @xcite .
however , we are quite confident in our disturbance classifications ; images of 12 of our ` disturbed ' galaxies with centrally - located sne are shown in figure 5 , confirming that this is not a ` normal ' group of galaxies .
the high central excess of snibc in the central regions of the disturbed host galaxies found in this work is difficult to explain in terms of effects other than an imf biased towards high - mass stars .
a possible alternative explanation is the effect of metallicity , given that @xcite find that the ratio of snibc / snii increases with both local and global metallicity .
looking at the absolute magnitudes of the host galaxies , we do indeed find that the disturbed galaxies are somewhat more luminous than the undisturbed galaxies ( ks probability of 0.07 that they are drawn from the same parent distribution ) , by almost 0.4 mag in the mean , which might imply a somewhat higher mean metallicity in the disturbed galaxies .
however , this does not seem to be driving the result we find . splitting the total sample ( disturbed and undisturbed ) by absolute magnitude
, we find no significant differences in the fr(@xmath0 ) distributions of bright and faint galaxies . splitting into bright and faint halves ,
the ks probability is 0.998 ( complete consistency ) , whereas when the bright third is compared with the faintest two - thirds ( to better match the disturbed / undisturbed split ) , p is 0.276 , but in the sense that the bright galaxies have a slight bias towards _ high _ fr(@xmath0 ) values . in any case , the expected metallicity bias resulting from a difference of 0.4 in galaxy absolute magnitude is very small .
the mass - metallicity relation of @xcite for galaxies of a few times 10@xmath8 m@xmath5 predicts a corresponding change of only @xmath40.025 dex in log(o / h ) , highly unlikely to cause any significant effects .
finally , in interacting systems the central metallicity is lowered by the in - fall of unenriched gas @xcite .
this would therefore act in the opposite sense to the result we find .
it should also be noted that whilst a study of gas - phase metallicities of the local environments of ccsne ( anderson et al .
mnras submitted ) finds a trend favouring snibc in high - metallicity regions , even the highest metallicity environments seem to host a significant fraction of snii .
it is also possible that stellar rotation ( e.g. * ? ? ?
* ) and binarity ( e.g. * ? ? ?
* ) could contribute to this effect .
it is not clear why the binary fraction should be higher within the disturbed galaxy sample , but it should be noted that the increased densities within these nuclear starburst regions could lead to more massive and denser clusters , within which processes such as stellar mergers and binary interactions would be more prevalent ( e.g. * ? ? ?
* ) . to conclude
, our preferred explanation of this central excess of snibc is that the central regions of these disturbed galaxies are hosting starbursts with initial mass functions biased to high stellar masses . given the small numbers of sne involved , the uncertain mass limits corresponding to progenitors of different sn types , and the likely role of binarity in determining sn type , it is hard to quantify the implications of this result for the imf .
however , an illustrative calculation can be performed as follows . under the assumption of a salpeter imf , and the ( admittedly simplistic )
assumptions that ccsne arise from single stars with masses between 8 and 80 m@xmath9 and that mass alone determines sn type , the relative numbers of snii and ibc in the outer regions of undisturbed galaxies ( 2.3:1 ) indicate a transition at about 18 m@xmath9 .
the apparent inversion of this ratio for the central regions of the disturbed galaxies ( 0.18:1 ) , if interpreted purely as a change in imf slope , appears to require a positive index in the imf slope ( formally @xmath10 + 0.95 cf .
1.35 , assuming the transition mass is unchanged at 18 m@xmath9 ) . however , we emphasize that this is purely illustrative ; all of the assumptions are likely to be in error at some level , and binarity and metallicity effects may play some part in the changes we find .
we note here that we are modifying the conclusions of @xcite in that it is hard to interpret the previously found central snibc excess purely in terms of metallicity effects .
however , we note that there is still a marginal central excess of snibc in the undisturbed galaxy sample , indicating some effect of metallicity .
quantifying the relative sizes of the different effects will be the focus of future studies . finally , it is interesting to note that with current research indicating a connection between gamma ray bursts ( grbs ) and type ic sne @xcite , recent studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have found that grb host galaxies show an over - abundance of merging or interacting galaxies compared to other star - forming hosts .
we have analysed the spatial distribution of 178 ccsne within a sample of host galaxies with recession velocities less than 6000 km / s .
host galaxies were classified by eye according to whether they show disturbance due to strong tidal interactions or mergers .
the main results are as follows : * ccsne of all types show a strong degree of central concentration in the disturbed galaxies , probably as a result of nuclear starbursts in these galaxies . *
this central excess is dominated by snibc .
* the snibc in disturbed galaxies are more centrally concentrated than the h@xmath1 emission . * the snibc excess can not easily be explained in terms of metallicity effects , extinction , or central incompleteness of sne . *
our preferred explanation of the snibc excess is that the central regions of the disturbed galaxies are dominated by nuclear starbursts with imfs biased towards high mass stars , although metallicity , binarity and stellar rotation may also play a role .
this paper has made use of data provided by the central bureau for astronomical telegrams .
we would like to acknowledge members of staff at the astrophysics research institute , in particular sue percival and david bersier for their helpful comments and discussion .
the authors would also like to thank the referee for a constructive and helpful report .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
the isaac newton telescope is operated on the island of la palma by the isaac newton group in the spanish observatorio del roque de los muchachos of the instituto de astrofsica de canarias .
the liverpool telescope is operated on the island of la palma by liverpool john moores university in the spanish observatorio del roque de los muchachos of the instituto de astrofsica de canarias with financial support from the uk science and technology facilities council .
smh would like to acknowledge stfc for a research studentship .
+ host&sn&sn type&fr(@xmath0)&fr(h@xmath1 ) + ngc493&1971s&iip&0.605&0.570 + ngc918&2009js&iip&0.703&- + ngc941&2005ad&ii&0.831&0.864 + ngc991&1984l&ib&0.498&0.401 + ngc1035&1990e&iip&0.272&0.363 + ngc1058&1961v&ii&0.968&0.931 + ngc1058&1969l&iip&1.000&1.000 + ngc1058&2007gr&ib / c&0.421&- + ngc1073&1962l&ic&0.754&0.518 + ngc1087&1995v&ii&0.368&0.497 + ngc1187&1982r&ib&0.695&0.760 + ngc1187&2007y&ib&0.981&1.000 + mcg-01 - 09 - 24&2002ei&iip&0.195&0.195 + ngc1343&2008dv&ic&0.195&0.134 + ugc2906&2008im&ib&0.682&- + ugc2971&2003ig&ic&0.176&0.108 + ic381&2001ef&ic&0.082&0.052 + ngc1832&2004gq&ib&0.672&0.328 + ngc1832&2009kr&ii&0.489&- + ic2152&2004ep&ii&0.461&0.560 + ugc3804&2002a&iin&0.419&0.253 + ngc2551&2003hr&ii&0.914&1.000 + ngc2596&2003bp&ib&0.486&0.362 + ugc4436&2004ak&ii&0.887&0.882 + ngc2726&1995f&ic&0.037&0.050 + ngc2742&2003z&iip&0.675&0.736 + ngc2715&1987m&ic&0.129&0.044 + ugc4904&2006jc&ib / c&0.332&0.525 + ngc2841&1972r&ib&0.855&0.904 + ngc2906&2005ip&ii&0.399&0.528 + ugc5249&1989c&iip&0.017&0.058 + ngc3074&1965n&iip&0.110&0.059 + ngc3074&2002cp&ib / c&0.936&0.961 + ngc3147&2006gi&ib&0.984&0.991 + ngc3184&1921b&ii&0.856&0.954 + ngc3184&1937f&iip&0.808&0.930 + ngc3184&1999gi&iip&0.276&0.112 + ngc3198&1966j&ib&0.898&0.963 + ngc3198&1999bw&iin&0.745&0.755 + ngc3240&2001m&ic&0.323&0.251 + ngc3294&1990h&iip&0.156&0.125 + ngc3340&2005o&ib&0.322&0.305 + host&sn&sn type&fr(@xmath0)&fr(h@xmath1 ) + ngc3340&2007fp&ii&0.170&0.125 + ngc3430&2004ez&ii&0.788&0.833 + ngc3437&2004bm&ic&0.073&0.076 + ngc3451&1997dn&ii&0.872&0.946 + ngc3504&2001ac&iin&0.826&0.992 + ngc3512&2001fv&iip&0.669&0.689 + ngc3556&1969b&iip&0.197&0.494 + ngc3631&1964a&ii&0.915&0.992 + ngc3631&1965l&iip&0.622&0.658 + ngc3631&1996bu&iin&0.923&0.993 + ngc3655&2002ji&ib / c&0.709&0.957 + ngc3683&2004c&ic&0.532&0.545 + ugc6517&2006lv&ib / c&0.480&- + ngc3756&1975t&iip&0.846&0.856 + ngc3810&2000ew&ic&0.261&0.147 +
ngc3810&1997dq&ib / c&0.774&0.734 + ngc3949&2000db&ii&0.364&0.253 + ngc3963&1997ei&ic&0.197&0.053 + ngc4030&2007aa&ii&0.942&0.828 + ngc4041&1994w&iin&0.491&0.541 + ngc4051&1983i&ic&0.498&0.473 + ngc4051&2003ie&ii&0.838&0.885 + ic758&1999bg&iip&0.669&0.657 + ngc4136&1941c&ii&0.880&0.882 + ngc4210&2002ho&ic&0.146&0.051 + ngc4242&2002bu&iin&0.896&0.930 + ngc4303&1926a&iil&0.607&0.736 + ngc4303&1961i&ii&0.697&0.877 + ngc4303&1964f&ii&0.189&0.106 + ngc4303&1999gn&iip&0.418&0.429 + ngc4303&2006ov&iip&0.418&0.429 + ngc4303&2008in&iip&0.845&- + ngc4369&2005kl&ic&0.271&0.540 + ngc4384&2000de&ib&0.087&0.140 + ngc4451&1985g&iip&0.138&0.212 + ngc4559&1941a&iil&0.208&0.131 + ugc7848&2006bv&iin&0.579&- + ngc4666&1965h&iip&0.324&0.198 + ngc4708&2003ef&ii&0.335&0.352 + ngc4725&1940b&iip&0.675&0.802 + ngc4900&1999br&iip&0.786&0.932 + ngc4961&2005az&ic&0.426&- + ngc4981&2007c&ib&0.320&0.236 + ngc5012&1997eg&iin&0.503&0.449 + ngc5033&1950c&ib / c&0.972&1.000 + ngc5033&1985l&iil&0.585&0.571 + ngc5334&2003gm&iin&0.480&- + ngc5371&1994y&iin&0.355&0.212 + ngc5468&2002ed&iip&0.811&0.791 + host&sn&sn
type&fr(@xmath0)&fr(h@xmath1 ) + ngc5559&2001co&ib / c&0.618&0.497 + ngc5584&1996aq&ic&0.178&0.086 + ngc5630&2005dp&ii&0.534&0.590 + ngc5630&2006am&iin&0.604&0.617 + ngc5673&1996cc&ii&0.924&0.934 + ngc5668&2004g&ii&0.657&0.595 + ngc5775&1996ae&iin&0.757&0.671 + ngc5806&2004dg&iip&0.484&0.378 + ngc5850&1987b&iin&0.995&- + ngc5879&1954c&ii&0.615&0.511 + ngc5921&2001x&iip&0.579&0.369 + ngc6118&2004dk&ib&0.673&0.626 + ngc6207&2004a&iip&0.729&0.660 + ugc10862&2004ao&ib&-&0.215 + ngc6643&2008ij&iip&0.519&0.620 + ngc6643&2008bo&ib&0.451&0.510 + ngc6700&2002cw&ib&-&0.642 + ngc6946&2004et&ii&0.975&- + ngc6951&1999el&iin&0.320&0.259 + ugc11861&1995ag&ii&0.343&0.170 + ugc11861&1997db&ii&0.633&0.396 + ugc12160&1995x&ii&0.564&0.484 + ugc12182&2006fp&iin&1.000&1.000 + + host&sn&sn type&fr(@xmath0)&fr(h@xmath1 ) + ngc895&2003id&ic&0.524&- + ugc2984&2002jz&ic&0.091&0.099 + ngc1614&1996d&ic&0.275&- + ngc1637&1999em&iip&0.276&0.268 + ic391&2001b&ib&0.062&0.060 + ngc1961&2001is&ib&-&0.749 + ngc2207&1999ec&ib&-&0.521 + ngc2207&2003h&ib&-&0.259 + ngc2146&2005v&ib / c&0.033&0.091 + eso492-g2&2005lr&ic&-&0.005 + ugc3829&2001ej&ib&0.152&0.391 + ngc2276&1968v&ii&0.699&0.790 + ngc2276&2005dl&ii&0.247&0.099 + ngc2276&1993x&ii&0.899&0.619 + ngc2532&1999gb&iin&0.485&0.443 + ngc2532&2002hn&ic&0.023&0.011 + ngc2604&2002ce&ii&0.381&0.560 + ngc2782&1994ak&iin&0.725&0.977 + ngc2993&2003ao&iip&0.456&0.784 + ngc3169&1984e&iil&0.684&0.731 + ngc3310&1991n&ic&0.268&0.277 + ngc3323&2004bs&ib&0.191&0.119 + ngc3323&2005kk&ii&0.766&0.875 + ngc3367&1992c&ii&0.689&0.687 + ngc3367&2007am&ii&0.302&0.314 + ngc3627&1973r&iip&0.471&0.566 + ngc3627&1997bs&iin&0.362&0.348 + ngc3627&2009hd&ii&0.496&- + ngc3690&1993g&iil&0.464&0.744 + ngc3690&1998t&ib&0.056&0.056 + ngc3690&1999d&ii&0.560&0.849 + ngc3786&1999bu&ic&0.180&0.522 + ngc3811&1971k&iip&0.809&0.900 + ic2973&1991a&ic&0.742&0.588 + ngc4038&2004gt&ib / c&0.834&0.991 + ngc4088&1991g&iip&0.466&0.453 + ngc4088&2009dd&ii&0.100&- + ngc4141&2008x&iip&0.194&0.085 + ngc4141&2009e&iip&0.594&0.491 + ngc4254&1967h&ii&0.664&0.648 + ngc4254&1972q&iip&0.811&0.791 + ngc4254&1986i&iip&0.334&0.318 + ngc4273&2008n&iip&0.300&0.316 + ngc4273&1936a&iip&0.569&0.598 + ngc4490&1982f&iip&0.277&0.202 + host&sn&sn type&fr(@xmath0)&fr(h@xmath1 ) + ngc4568&1990b&ic&0.302&- + ngc4568&2004cc&ic&0.158&- + ngc4618&1985f&ib&0.121&0.087 + ngc4615&1987f&iin&0.489&0.333 + ngc4688&1966b&iil&0.571&0.454 + ngc4691&1997x&ic&0.171&0.472 + ngc5000&2003el&ic&0.482&0.476 + ngc5021&1996ak&ii&0.619&0.659 + mcg-04 - 32 - 07&2003am&ii&0.211&0.339 + ngc5194&1994i&ic&-&0.122 + ngc5194&2005cs&iip&-&0.222 + ngc5395&2000cr&ic&0.538&0.549 + ngc5480&1988l&ib&0.230&0.369 + ngc5682&2005ci&ii&0.204&0.191 + ngc7479&1990u&ic&0.603&0.488 + ngc7479&2009jf&ib&0.764&- + ngc7537&2002gd&ii&0.759&0.685 + ngc7714&2007fo&ib&0.377&- + ugc12846&2007od&iip&0.945&1.000 + , c. l. , mazzali , p. a. , prochaska , j. , cappellaro , e. , panaitescu , a. , berger , e. , van putten , m. , van den heuvel , e. p. j. , young , p. , hungerford , a. , rockefeller , g. , yoon , s. , podsiadlowski , p. , nomoto , k. , chevalier , r. , schmidt , b. , & kulkarni , s. 2007 , pasp , 119 , 1211 , t. , genzel , r. , martins , f. , nayakshin , s. , beloborodov , a. m. , levin , y. , trippe , s. , eisenhauer , f. , ott , t. , gillessen , s. , abuter , r. , cuadra , j. , alexander , t. , & sternberg , a. 2006 , apj , 643 , 1011 | we compare the radial locations of 178 core - collapse supernovae to the @xmath0-band and h@xmath1 light distributions of their host galaxies .
when the galaxies are split into ` disturbed ' and ` undisturbed ' categories , a striking difference emerges .
the disturbed galaxies have a central excess of core - collapse supernovae , and this excess is almost completely dominated by supernovae of types ib , ic and ib / c , whereas type ii supernovae dominate in all other environments .
the difference can not easily be explained by metallicity or extinction effects , and thus we propose that this is direct evidence for a stellar initial mass function that is strongly weighted towards high mass stars , specifically in the central regions of disturbed galaxies . _
keywords _ : galaxies : interactions galaxies : ism supernovae : general |
silicon nitride - based waveguides form a promising , cmos - compatible platform in integrated photonics research @xcite . especially , stoichiometric silicon nitride ( si@xmath2n@xmath3 ) deposited using low - pressure chemical vapor deposition ( lpcvd ) offers extremely low intrinsic losses and superior reproducibility .
further advantages of this platform are a broad transparency ranging from the visible @xcite to the mid - infrared , a high index contrast and absence of two - photon absorption in the near - infrared , including all the telecommunication bands . due to their low propagation losses in the c - band as well as their low coupling losses by using spot - size converters si@xmath2n@xmath3
waveguides are favored in the field of microwave photonics @xcite and for novel types ( glass - semiconductor ) lasers with record low spectral bandwidth @xcite .
furthermore , ultra - low propagation losses in the c - band of 0.32 db / m have been demonstrated with weakly confined modes using 40 nm thin si@xmath2n@xmath3 waveguides @xcite .
the low confinement of the mode , however , has the drawback of increased bending losses and , as a consequence , limit the density of devices in integrated circuits .
further applications of si@xmath2n@xmath3 waveguides are , e.g. , bio - chemical applications such as optical coherence tomography @xcite and lab - on - a - chip devices due to the compatibility of si@xmath2n@xmath3 waveguides with microfluidic channels @xcite .
furthermore , si@xmath2n@xmath3 waveguides are of high interest for nonlinear integrated photonics @xcite due to their high kerr index @xcite , while supporting highly confined modes due to their high index contrast and lacking nonlinear losses in the near - infrared .
nonlinear effects such as supercontinuum generation @xcite and parametric frequency comb generation @xcite have been demonstrated .
for the latter , huge potential in nonlinear optical signal processing was shown by transmitting a data stream of 1.44 tbit / s @xcite .
it is desirable to select the waveguides such that the pump wavelength for broadband wavelength conversion lies in the spectral region where laser sources are readily available , i.e. 1000 to 2000 nm .
however , to obtain phase matching for broadband wavelength conversion , the dispersion of the waveguide must be engineered such that pump wavelength is in the anomalous dispersion regime , while also being close to the zero dispersion wavelength ( zdw ) . to achieve high modal confinement and to shift the zdw to the near - infrared ,
the thickness of the silicon nitride waveguide core must be increased substantially from the values that are typically used ( @xmath4 nm ) . to obtain the dispersion of the waveguides we simulated the effective refractive index @xmath5 for various wavelengths using a finite element solver ( comsol multiphysics ) that takes both the dispersion of the materials and , as well , the dispersion induced by the geometry of the waveguide core into account . from this the dispersion parameter , @xmath6 @xcite , can be calculated , where @xmath7 is the vacuum wavelength of the light and @xmath8 the speed of light .
this is shown in fig.[fig:0 ] , where we plot @xmath9 as a function of the wavelength for waveguides with a fixed width @xmath10 0.8 @xmath11 m and three different heights ranging from 0.8 to 1.2 @xmath11 m . as a comparison
, the dispersion parameter for bulk si@xmath2n@xmath3 is shown as well .
while bulk si@xmath2n@xmath3 does not have a zdw in the wavelength range from 1000 to 2000 nm , the zdw for the waveguides is calculated to range from 1200 nm ( @xmath12 @xmath11 m , red ) to 1560 nm ( @xmath13 @xmath11 m , green ) by changing the thickness @xmath14 and , hence , providing phase matching for wavelengths around 1550 nm .
calculated dispersion parameter , @xmath9 , for bulk si@xmath2n@xmath3 ( black ) and for quasi - tm modes ( primarily polarized along the y - direction ) of waveguides with a fixed width of 0.8 @xmath11 m and a height of 0.8 @xmath11 m ( red ) , 1.0 @xmath11 m ( blue ) , and 1.2 @xmath11 m ( green ) .
while bulk si@xmath2n@xmath3 does not exhibit a zdw at wavelengths between 1000 nm and 2000 nm , the zdw of the waveguides are 1200 nm ( @xmath15 0.8 @xmath11 m ) , 1420 nm ( @xmath15 1.0 @xmath11 m ) , and 1560 nm ( @xmath15 1.2 @xmath11m).,height=181 ] however , realizing si@xmath2n@xmath3 waveguides with sufficient thickness is problematic .
typically , a si@xmath2n@xmath3 layer is deposited on a thermally oxidized silicon substrate via lpcvd , and waveguide cores are etched from this si@xmath2n@xmath3 layer .
this layer has a limited thickness due to the increasing tensile stress that develops during the deposition of the si@xmath2n@xmath3 with increasing layer thickness .
ultimately , this stress leads to the formation of cracks when the thickness is larger than about 400 nm @xcite . recently ,
improved methods for manufacturing low - loss thick si@xmath2n@xmath3 waveguides have been introduced @xcite , however , they still suffer from cracks in the si@xmath2n@xmath3 layer @xcite , limiting the yield and reproducibility of device fabrication . in this paper
we report on the fabrication of the first silicon nitride waveguides with thicknesses greater than 800 nm in a reproducibly crack - free manner . instead of depositing a thick layer of si@xmath2n@xmath3 , followed by etching the waveguide cores from this layer , we first etch trenches in the silicon dioxide cladding that is formed by thermally oxidizing the top surface of the silicon substrate .
the trenches are then filled with si@xmath2n@xmath3 using lpcvd and the resulting waveguide cores are crack - free for depths of up to @xmath16 1.2 @xmath11 m . note , that this method , in general , provides waveguides with an aspect ratio , @xmath17 , unlike commonly used si@xmath2n@xmath3 waveguides . as a result ,
the width of the waveguide core is the critical parameter that controls stress - induced cracks , which is related to the thickness of the si@xmath2n@xmath3 layers .
waveguides fabricated in this way are referred to as triplex@xmath18@xcite .
details of this process are given in section 2 .
we then discuss in section 3 the characterization of several thick waveguides that are manufactured according to this novel procedure and end with a discussion and conclusion in section 4 .
in this method , the waveguide devices are manufactured by filling predefined trenches with si@xmath0n@xmath1 using lpcvd .
the individual fabrication steps are shown in fig .
[ fig:2 ] . in the first step , as shown in fig .
[ fig:2](a ) , reactive ion etching ( rie ) is used to etch trenches with a width @xmath19 and a depth @xmath14 into the thermally oxidized top surface ( oxide depth 8 @xmath11 m ) of a silicon wafer . in the next step ,
a 250 nm thick layer of stoichiometric si@xmath0n@xmath1 is deposited . during this deposition process
, the si@xmath0n@xmath1 layer grows equally thick from the sidewalls and bottom of the trenches , as illustrated in fig .
[ fig:2](b ) . to release stress from the material
, the deposition is performed in steps and wafers are allowed to cool down to room temperature between depositions @xcite .
as the deposition continues , the trench fills with si@xmath0n@xmath1 until the the two layers at the side walls merge in the middle of the trench , such that the trench is completely filled , as shown in fig .
[ fig:2](c ) . for the manufactured waveguides no more
than two deposition steps were required .
the final waveguide core is obtained by removing the remaining si@xmath0n@xmath1 from the top of the wafer using chemical - assisted mechanical polishing and dry etching , as shown in fig .
[ fig:2](d ) . to ensure that etching is stopped when the si@xmath0n@xmath1 from the top of the wafer is removed an etch rate monitor is used .
if required , this last processing step can also be applied between deposition steps , to thin the si@xmath2n@xmath3 layer on top of the wafer to prevent cracks . to complete the waveguide cladding , first a layer of tetraethyl orthosilicate ( si(oc@xmath20h@xmath21)@xmath1 , teos )
is deposited by lpcvd , which will form sio@xmath22 after annealing .
second , a layer of sio@xmath22 is grown to a thickness of 8 @xmath11 m by plasma - enhanced chemical vapor deposition ( pecvd ) ( both not shown in fig .
[ fig:2 ] ) . to convert the teos layer into sio@xmath22 and to reduce the absorption in the near infra - red caused by dangling si - h bonds , the wafers are annealed after deposition at 1150 @xmath23c . using this procedure ,
several triplex - based si@xmath0n@xmath1 waveguides were manufactured with their width , @xmath19 , ranging from 0.6 to 0.9 @xmath11 m , and depths , @xmath14 , of 0.8 , 1.0 , and 1.2 @xmath11 m . to ensure that these waveguides have low propagation losses , it is important that the trenches have a low surface roughness at the bottom as well as at the side walls , since rougher surfaces will result in higher scattering losses for propagating light .
a scanning electron microscope ( sem ) picture of an etched trench is shown in fig .
[ fig:1](a ) for a width of @xmath240.8 @xmath11 m and a depth of @xmath151.0 @xmath11 m . note that the vertical direction of all the sem pictures in fig .
[ fig:1 ] appears to be smaller because the pictures were taken under an angle .
figure [ fig:1](a ) also shows a slight positive taper , meaning that the width of the trench increases from bottom to top , which ensures a conformal deposition . without tapering , the deposition is not fully conformal , and a gap between the silicon nitride and the trench wall appears , which increases the propagation losses as a consequence .
sem pictures of two completed waveguide cores are shown in fig .
[ fig:1](b ) and [ fig:1](c ) with a cross section of @xmath24 0.83 @xmath11 m by @xmath15 0.8 @xmath11 m and @xmath24 0.9 @xmath11 m by @xmath15 1.2 @xmath11 m , respectively , with and without a top cladding , respectively . to increase the contrast between the si@xmath0n@xmath1 and the sio@xmath22 in the sem pictures ,
the samples were etched , which resulted in the si@xmath0n@xmath1 waveguide core sticking out from the sio@xmath22 because of a difference in etching rates of the materials . in fig .
[ fig:1](b ) it can be seen that a homogeneous waveguide core with @xmath25 @xmath11 m and @xmath26 @xmath11 m is formed using this manufacturing method . by comparing the shape of the trench as shown in fig .
[ fig:1](a ) and the waveguides in figs .
[ fig:1](b ) and [ fig:1](c ) , the stress induced by the deposition of silicon nitride is evident .
the large difference in material stress between the sio@xmath20 and si@xmath0n@xmath1 layers leads to a change in both the shape of the waveguide as well as an inwards bend of the side walls .
the rounded shape of the waveguides leads to a small ( @xmath2710 nm ) blue - shift in the zdw compared to the rectangular shape assumed in fig.[fig:0 ] .
however , the variations in waveguide dimensions , given by fabrication tolerances , cause larger shifts . in the case of the deepest etching depth ( @xmath28 @xmath11 m )
, it is observed that the change of the waveguide shape is even stronger . here , the tapering of the trench was insufficient to compensate for the sidewall bending induced by the tensile stress of the si@xmath0n@xmath1 layer .
this effects the final closure of the trench when the deposited si@xmath0n@xmath1 , which grows from the two side walls , meet in the middle . as a result
, the trench is not filled entirely with si@xmath0n@xmath1 as can be seen in fig .
[ fig:1](c ) . in future
, we will compensate for this effect by applying a stronger positive taper to the sidewall of the trenches . during the fabrication ,
a large number of waveguides and waveguide circuits were obtained with the described technique covering a total of six wafers ( 100 mm diameter ) and not a single crack was observed in the si@xmath0n@xmath1 layers .
this clearly indicates that , using our novel approach , crack - free si@xmath0n@xmath1 waveguides with a width of up to 0.9 @xmath11 m can be fabricated .
the propagation losses are an essential characteristic for integrated waveguides , since they are critical for the design of devices such as micro - resonators . by measuring the transmission of a broadband light source through the waveguides ,
the propagation losses were determined .
the broadband light from a superluminescent diode with a center wavelength of 1560 nm and a spectral bandwidth at full width at half maximum of 45 nm was used .
the light was injected and collected using lensed fibers ( 2 @xmath11 m spot size at 1550 nm ) , while the input polarization was controlled to excite the quasi - tm mode which is primarily polarized along the y - direction ( see fig . [ fig:2](d ) ) of the waveguides using an in - line optical fiber polarization controller .
the transmitted power was measured for three etching depths , 0.8 @xmath11 m , 1.0 @xmath11 m , and 1.2 @xmath11 m , and for four different propagation lengths , using spiral waveguides with lengths of 2.30 cm , 3.89 cm , 6.90 cm , and 10.48 cm and of various widths .
the transmission as a function of waveguide length is shown in fig .
[ fig:3](a ) , and from this data , the propagation loss coefficient is determined to be 0.37 db / cm ( @xmath12 @xmath11 m , blue ) , 0.45 db / cm ( @xmath29 @xmath11 m , red ) , and 1.37 db / cm ( @xmath13 @xmath11 m , black ) , while @xmath19 ranged from 0.7 to 0.9 @xmath11 m .
no variation in the propagation loss was observed for the range of waveguide widths investigated .
the propagation losses are believed to be mainly due to scattering losses since absorptive losses have been reported to amount to values of below 0.055 db / cm in si@xmath0n@xmath1 waveguides @xcite .
we address the comparably high losses for a @xmath14 of 1.2 @xmath11 m to higher scattering losses resulting from the small gap in the center of the si@xmath0n@xmath1 core , as shown in fig .
[ fig:1](c ) .
another important performance parameter is the bending loss , as this indicates how densely devices can be packed on a wafer . in our measurements we found that the bending losses are so low that we could not experimentally distinguish these from the differences in coupling efficiencies from waveguide to waveguide . to quantify bending losses for thick si@xmath0n@xmath1 waveguides , we simulated them using a finite element solver ( fielddesigner , phoenix bv ) .
the simulated bending losses for a quasi - tm mode at a wavelength of 1550 nm are shown in fig .
[ fig:3](b ) as a function of the radius of the bend for waveguide cross sections of @xmath26 @xmath11 m and 1.2 @xmath11 m and @xmath30 @xmath11 m and 0.8 @xmath11 m , respectively . as a result of the high confinement ,
the bending losses of the guided mode are lower for larger waveguide cross sections . as shown in fig .
[ fig:3](b ) , e.g. , the losses for a bending radius of 20 @xmath11 m decrease from 2.9 db / cm for the smallest waveguide cross section , @xmath31 @xmath11 m and @xmath12 @xmath11 m ( red ) , to as low as @xmath32 db / cm for @xmath33 @xmath11 m and @xmath13 @xmath11 m ( green ) .
the bending losses are negligible for the bigger cross sections when compared to the propagation losses .
the simulated bending loss for waveguides with high modal confinement and @xmath34 @xmath11 m ( blue ) is only 0.7 db / m for a bending radius of 20 @xmath11 m . in comparison , to obtain a similar bending loss from a waveguide with low modal confinement @xcite , the bending radius must be as large as 2 mm or more .
consequently , the density of waveguides circuits on a wafer can be increased by orders of magnitude using waveguides with a high confinement , which leads to a superior efficiency in the fabrication of si@xmath0n@xmath1 waveguides .
in conclusion , we have shown a novel way to fabricate triplex - based si@xmath0n@xmath1 waveguides with a width of up 0.9 @xmath11 m , while having a depth of up to 1.2 @xmath11 m .
this method is promising to achieve a high yield compared to previously reported approaches to fabricate si@xmath0n@xmath1 waveguides of similar cross - section dimensions .
the complete area of six wafers with a diameter of 100 mm was used to produce waveguide structures , while not a single stress - induced crack appeared during or after fabrication .
the propagation losses are measured to lie below 0.4 db / cm for waveguides with a depth of 0.8 @xmath11 m , while the bending losses were simulated to be below 0.01 db / cm for a bending radius of 20 @xmath11 m , both for a quasi - tm mode at a wavelength of 1550 nm .
the measured coupling efficiency and observed variation in coupling efficiency can be improved by implementing spot - size converters at either end of the waveguide .
the spot size converter expands the mode field diameter , which makes the coupling easier and less sensitive for imperfections in the flatness of the end facet .
this research is supported by the dutch technology foundation stw , which is part of the netherlands organisation for scientific research ( nwo ) , and which is partly funded by the ministry of economic affairs s. romero - garca , f. merget , f. zhong , h. finkelstein , and j. witzens , `` silicon nitride cmos - compatible platform for integrated photonics applications at visible wavelengths , '' * 21*(12 ) , 1403614046 ( 2013 ) . c. g. h. roeloffzen , l. zhuang , c. taddei , a. leinse , r. g. heideman , p. w. l. van dijk , r. m. oldenbeuving , d. a. i. marpaung , m. burla , and k .- j .
boller , `` silicon nitride microwave photonic circuits , '' * 21*(19 ) , 2293722961 ( 2013 ) .
d. marpaung , b. morrison , r. pant , c. roeloffzen , a. leinse , m. hoekman , r. heideman , and b. j. eggleton , `` si@xmath2n@xmath3 ring resonator - based microwave photonic notch filter with an ultrahigh peak rejection , '' * 21*(20 ) , 2328623294 ( 2013 ) .
r. m. oldenbeuving , e. j. klein , h. l. offerhaus , c. j. lee , h. song , and k .- j .
boller , `` 25 khz narrow spectral bandwidth of a wavelength tunable diode laser with a short waveguide - based external cavity , '' laser phys .
lett . * 10*(1 ) , 015804 ( 2013 ) .
y. fan , r. m. oldenbeuving , e. j. klein , c. j. lee , h. song , m. r. h. khan , h. l. offerhaus , p. j. m. van der slot , and k .- j .
boller , `` a hybrid semiconductor - glass waveguide laser , '' proc .
spie 9135 , 91351b ( 2014 ) .
j. f. bauters , m. j. r. heck , d. john , d. dai , m .- c .
tien , j. s. barton , a. leinse , r. g. heideman , d. j. blumenthal , and j. e. bowers , `` ultra - low - loss high - aspect - ratio si@xmath2n@xmath3 waveguides , '' * 19*(4 ) , 31633174 ( 2011 ) .
g. yurtsever , b. povaay , a. alex , b. zabihian , w. drexler , and r. baets , `` photonic integrated mach - zehnder interferometer with an on - chip reference arm for optical coherence tomography , '' biomed .
express * 5*(4 ) , 10501061 ( 2014 ) .
a. ymeti , j. s. kanger , j. greve , g. a. j. besselink , p. v. lambeck , r. wijn , and r. g. heideman , `` integration of microfluidics with a four - channel integrated optical young interferometer immunosensor , '' biosens . bioelectron . * 20*(7 ) , 14171421 ( 2005 ) . j. s. levy , a. gondarenko , m. a. foster , a. c. turner - foster , a. l. gaeta , and m. lipson , `` cmos - compatible multiple - wavelength oscillator for on - chip optical interconnects , '' nature photon . *
4*(1 ) , 3740 ( 2009 ) .
j. pfeifle , v. brasch , m. lauermann , y. yu , d. wegner , t. herr , k. hartinger , p. schindler , j. li , d. hillerkuss , r. schmogrow , c. weimann , r. holzwarth , w. freude , j. leuthold , t. j. kippenberg , and c. koos , `` coherent terabit communications with microresonator kerr frequency combs , '' nature photon . * 8*(5 ) , 375380 ( 2014 ) . | in this paper we present a novel fabrication technique for silicon nitride ( si@xmath0n@xmath1 ) waveguides with a thickness of up to 900 nm , which are suitable for nonlinear optical applications .
the fabrication method is based on etching trenches in thermally oxidized silicon and filling the trenches with si@xmath0n@xmath1 . using this technique no stress - induced cracks in the si@xmath0n@xmath1 layer
were observed resulting in a high yield of devices on the wafer .
the propagation losses of the obtained waveguides were measured to be as low as 0.4 db / cm at a wavelength of around 1550 nm . |
the light - curve modulation of rrab stars ( blazhko effect ) is one of the most outstanding problems of stellar pulsation theory , which has remained unsolved despite continuous observational and theoretical efforts .
recent , accurate , ground based and space observations @xcite reveal more and more details of the snapshots of the modulations properties .
however , contrary to the remarkable stability of the light curves of non - modulated rrab stars , the modulation properties of blazhko variables may show significant changes on time - scales of years and decades e.g. in the cases of rr gem @xcite , rr lyr @xcite and xz cyg @xcite .
the only way to study these long - term changes is to combine archival data with modern observations .
globular clusters , which have been already observed regularly on a time base longer than a century , are ideal targets for following the variations of the modulation properties of blazhko stars , provided that the modulation is strong enough to be detected in photographic data
. an accurate knowledge of the blazhko - star population in globular clusters also yields an important information for determining the incidence of this phenomenon .
recent surveys show that about half the rrab stars exhibit light - curve modulation in some samples @xcite , while previous estimates gave only 1030 per cent .
m5 , one of the closest globular clusters with about 100 rr lyrae stars having been continuously observed during the 20@xmath2 century , is one of the few possible targets for these studies .
blazhko stars are quite numerous in m5 based on previous photographic and ccd investigations .
already ( * ? ? ?
* hereafter oo41 ) noted that the light curves of v2 , v4 , v14 , v18 , v27 , v52 , v63 and v72 in m5 were not stable ; they showed irregular or rw dra - type variability .
blazhko effect was also identified by ( * ? ? ? * hereafter kk71 ) in v2 , v14 , v18 , v58 and v63 .
using photographic observations , @xcite determined blazhko modulation periods for v2 , v14 and v63 . based on their ccd observations , ( * ? ? ?
* hereafter r96 ) and ( * ? ? ?
* hereafter k00 ) noted light - curve variability of v2 , v4 , v5 , v8 , v14 , v27 , v56 , v65 , v89 and v97 .
a comprehensive study of the blazhko stars in m5 is , however , still missing . in (
* paper i ) we collected all the photometric data of m5 variables in order to study the period changes of rr lyrae stars .
these data are utilized here to detect and study blazhko variables .
values of v1 ( top panel ) derived for the entire photometric data .
period analysis of the @xmath3 data results in a 512 d periodicity of the phase variation .
the bottom panel shows the @xmath3 data folded with this modulation period.,width=313 ] , of different segments of v2 s light curve are shown .
the top ( @xmath4 ) , the middle ( @xmath5 ) and the bottom ( @xmath6 ) panels refer to data for the 19711986 , 19561965 and 19341955 intervals , respectively . in panel @xmath4 frequencies at @xmath7 cd@xmath8 appear , corresponding to a modulation with a 136-d period . in 19561965 this modulation
could also be detected ; however , the period was a bit shorter at that time , 130 d. some modulation features also appear in the earliest data ( panel @xmath6 ) , a signal at @xmath9 is the highest signal in the entire residual spectrum in the @xmath10 cd@xmath8 frequency range .
the one - cycle - per - year alias peak of this frequency at @xmath11 is at a separation that is twice the separation of the modulation frequencies appearing in panels @xmath4 and @xmath5 .
, width=309 ] and @xmath12 , respectively .
modulation frequencies appear at @xmath13 , @xmath14 and @xmath15 , @xmath16 cd@xmath8 , corresponding to an @xmath17 d modulation period .
, width=309 ] light curve of v5 .
the different parts of the data are plotted with different symbols as indicated by the key labels .
the prefixes r and k refer to the r96 and k00 data , respectively .
, width=257 ] are shown .
modulation peaks with @xmath18 cd@xmath8 ( @xmath19 d ) separation appear in both data sets .
a yearly alias solution with 0.0183 cd@xmath8 ( 55 d ) is also possible.,width=309 ] and @xmath12 , for this data set .
modulation frequencies appear in these panels at @xmath13 , @xmath14 and @xmath20 , @xmath21 cd@xmath8 , which corresponds to a 47.8-d modulation period .
bottom panel shows the residual amplitude spectrum at @xmath22 for another subset of the data , between jd 4233068 and jd 2436762 .
the modulation peak appearing in this data set corresponds to a significantly different modulation period , @xmath23 d.,width=313 ]
the homogenized ( see section 2.3 of paper i ) light curves of the rrab stars in m5 were carefully examined to identify variables showing light - curve instability . due to crowding problems , this was not possible for all the rrab stars in our sample .
variables lying in too crowded areas and/or having close bright companions could not be measured accurately enough to detect light - curve variability without doubt .
this was especially true if no ccd observation of the variable was available .
from the 65 rrab stars for which phase variation diagrams were constructed in paper i , this was the case for 15 stars ( v6 , v13 , v17 , v25 , v26 , v36 , v37 , v54 , v74 , v83 , v85 , v90 , v91 , v92 and v96 ) .
these stars were denoted by the letter ` @xmath24 ' in the ` remarks ' column of table 4 in paper i. the combined photographic and ccd @xmath25 light curves of the remaining 50 rrab stars were studied using different methods in order to detect light - curve variations and to determine modulation periods if it was possible .
when ccd @xmath1 observations from different sources were also available , these data were investigated separately .
different segments of the data were checked and analysed . if light - curve variability was suspected in any part of the data , fourier analyses of the light curves and the maximum brightness / maximum timing data
were also applied .
the maximum brightness / phase data of globular - cluster variables are , however , very sparse ( one - two data per season if any ) , and give useable information only in very few cases ( e.g. in v19 ) .
we searched for modulation - frequency signals in the vicinity of the low - order pulsation frequency components ( @xmath26 ) corresponding to the same modulation frequency ( @xmath27 ) . the modulation period ( @xmath28 )
was then determined by a non - linear process that fitted the pulsation and some of the modulation frequencies ( @xmath29 and @xmath30 ) and their amplitudes and phases simultaneously to the data .
a blazhko classification was assigned if at least one of the following criteria was fulfilled : * changes in the light curve s amplitude , shape and/or phase were detected , * at least three modulation frequencies , corresponding to the same modulation period , appeared around the frequencies of the pulsation components in the residual spectrum of one subset of the light - curve data , * frequencies , corresponding to the same modulation period , appeared in the residual spectra of different subsets of the observations .
we note , however , that without accurate , extended , ccd photometric observations the results obtained from inhomogeneous , photographic data have to be regarded as somewhat uncertain .
even though the available ccd observations of the m5 variables show clear evidence of light - curve modulation in some cases , they do not allow the determination of the blazhko period for any of the variables .
another problem is that the photographic data are seriously affected by daily , monthly and yearly aliases , which may result in an erroneous blazhko period .
if a modulation period could be determined , its reality was visually checked using animations of the light curves in different phases of the modulation . the animated light curves of the m5 blazhko variables and the data sets that were used are available at the url : http + : www.konkoly.hu/24/publications/m5 .
we identified 20 blazhko candidates among the 50 rrab variables that were studied .
the properties of the individual stars are discussed below .
* v1 * there is a slight difference between the shape of minima observed by r96 and k00 .
the maximum brightness in the r96 data varies between 14.41 and 14.45-mag .
the star is far from the centre , so it can be measured accurately on the photographic plates .
the residual spectrum of the combined photographic data between jd 2441447 and jd 2449104 ( 19721993 ) shows a large - amplitude ( 0.05-mag ) signal at @xmath31 cd@xmath8 .
we found 522 d as the probable period of the modulation for this time interval .
the modulation is dominated by @xmath32 min phase oscillations .
this result is supported by the analysis of the variation of the seasonal @xmath3 values of the entire photometric data ( 19351997 ) as shown in fig .
[ v01oc ] .
phase oscillations with a period of 512 d fit the observed 0.02-d variation of the seasonal @xmath3 data reasonably well .
the ccd data do not contradict this modulation period .
the pulsation period of v1 remained constant during the time base of more than 100 years of the observations . * v2 * the star is well separated .
the pulsation period did not show any systematic changes during the hundred years of the observations , but irregular period changes of the order of @xmath33 d were detected .
ccd data from k00 show different descending branches .
already oo41 noted differences between two maxima .
a modulation period of 132.38 d was determined by @xcite .
we have found clear evidence of the modulation in three segments of the combined photographic data . between 1971 and 1986
, v2 exhibited a very strong modulation with a 136.5-d period .
the pulsation amplitude varied between 0.6 and 1.3-mag ; the detected changes in the maximum light were 0.6-mag and 75-min in brightness and phase , respectively .
the scarce observations from 1966 and 1968 showed a large - amplitude ( 1.1-mag ) light curve , without any evidence of modulation . between 1956 and 1965
the modulation properties were similar to those found for the 19711986 interval , but due to the scarceness of the data , this result is less certain .
we also found some indication of the modulation in the data from 1934 to 1955 but with half the period that was found in the other data segments .
the pulsation amplitude did not change significantly ; it was around 0.8-mag in this part of the data .
any amplitude modulation larger than 0.1-mag was not present , but the amplitude of the phase modulation of the light curve was as large as @xmath3460 min .
the residual spectra of v2 in the vicinity of the main pulsation frequency are shown for different segments of the data in fig .
[ v02sp ] .
frequency , corresponding to a 39.4-d periodicity of the amplitude variation .
the bottom panel shows the maximum brightness data folded with this modulation period.,width=317 ] light curve of v24 .
the different parts of the data are plotted with different symbols as indicated by the key labels .
the prefixes r and k refer to the r96 and k00 data , respectively.,width=272 ] light curve of v27 .
the different parts of the data are plotted with different symbols as indicated by the key labels .
the prefixes r and k refer to the r96 and k00 data , respectively.,width=272 ] * v4 * this is the shortest - period rrab star in m5 , close to the cluster centre , in a relatively crowded region .
contamination of close companion stars makes the photographic photometry noisy .
oo41 noted rw dra - type variation of the light curve . the ccd light curve published in r96 indicates phase variation of the rising branch . the pulsation amplitude is larger in r96 than in the k00 data .
the pulsation period had been steadily increasing prior to 1970 , then a rapid period jump of about 0.00005 d occurred .
we searched for a modulation period in the combined photographic data prior to and after the period jump . between jd 2442553 and jd 2449104 we found clear signals of modulation with a 110-d period ( see fig .
[ v04sp ] ) .
the light curve was dominantly phase modulated .
the light - curve modulation before 1970 was not so evident , however , a 107-d solution was possible .
* v5 * the photographic data are very noisy because of crowding .
r96 noted some evidence of light - curve variability .
the r96 light curve has somewhat larger amplitude than the k00 one .
based on the combined ccd v data shown in fig [ v5 ] , v5 is a possible balzhko star .
the pulsation period was decreasing during the time span of the observations with one of the largest , steady , period - change rates observed in m5 . *
v8 * the light - curve variability in the phases of the descending branch and at the base of the ascending branch was already noticed by r96 .
differences around minimum light are also indicated by the k00 data .
the amplitude of the ccd @xmath1 light curves from different observations varies between 0.85 and 1.0-mag .
the star is well separated , therefore these variations are regarded as real .
the residual signals appear at similar , 0.01550.0150 cd@xmath8 separations from @xmath22 in the 19711977 and 19781993 data , indicating that the modulation period is about 64 d ( fig . [ v08sp ] ) . we failed to detect periodic light - curve modulation in the photographic data before 1971 .
the pulsation period has been steadily increasing but the period increase is not linear .
, the instantaneous periods and the fourier parameters of 12 ( 25 years long ) subsets , shown in fig . [ v29lc ] , of the photographic light curve of v29 are plotted .
the time spans of the data subsets are indicated by horizontal lines in the top of the first panel .
the @xmath3 shows cyclic fluctuations with about 6000-d period .
the consequent changes in the pulsation period are in the 0.00003-d range .
the fourier parameters of the light curves of the data subsets ( @xmath35 ) show larger variations than their errors could explain , indicating that the light curve of v29 is not stable.,width=317 ] and @xmath12 , respectively .
modulation frequencies appear at @xmath13 , @xmath14 and @xmath20 , @xmath36 cd@xmath8 , corresponding to an @xmath37 d modulation period .
a yearly alias solution corresponds to a modulation frequency of @xmath38 cd@xmath8 ( 621 d ) . ]
light curve of v38 .
the different parts of the data are plotted with different symbols as indicated by the key labels .
the prefixes r and k refer to the r96 and k00 data , respectively.,width=279 ] light curve of v52 .
the different parts of the data are plotted with different symbols as indicated by the key labels .
the prefixes r and k refer to the r96 and k00 data , respectively .
light - curve variations at minimum and maximum light are very probable according to the ccd data .
, width=279 ] and @xmath12 , respectively .
modulation frequencies are apparent at @xmath13 , @xmath14 and @xmath20 , @xmath39 cd@xmath8 , corresponding to an @xmath40 d modulation period . , width=309 ] and @xmath12 for the 19711993 period and around @xmath22 for the 19521963 period .
although the largest - amplitude signals appear at different frequencies in the three residual spectra , the frequency at @xmath41 cd@xmath8 separation has significant amplitudes in each spectra .
this frequency is identified as @xmath27 ( 170 d ) for v56 , however , the one - cycle - per - year alias solution corresponding to @xmath42 cd@xmath8 can not be excluded for sure . ,
width=309 ] ) of the photographic data of v58 for the periods jd 24409852449104 ( top panel ) and jd 24330682439979 ( bottom panel ) are shown .
modulation with a 0.017 cd@xmath8 frequency ( 59 d period ) appears in both data sets.,width=313 ] [ v63sp ] ) of photographic data of v63 for the periods
jd 24409852449104 ( top panel ) and jd 24320002439979 ( bottom panel ) are shown .
modulation with a 0.0215 cd@xmath8 frequency ( 46 d ) appears in both data sets.,title="fig:",width=332 ] light curve of v65 .
the different parts of the data are plotted with different symbols as indicated by the key labels .
the prefixes r and k refer to the r96 and k00 data , respectively .
, width=279 ] ( phase shift ) data of v72 are shown .
the @xmath43 d ( @xmath44 hour ) scatter of the data is much larger than the uncertainty of the @xmath3 values .
such rapid , random period variation of an rrab star is unlikely , therefore the significant changes in the phases of the light curve are most probably caused by blazhko effect . ,
width=325 ] * v14 * @xcite determined a 75-d blazhko - modulation period for v14 .
the ccd ( r96 , k00 ) light curves show strong amplitude modulation .
the maximum - brightness variation is 0.7-mag in the @xmath1 band .
the pulsation - period variation is strong and irregular , which makes the analysis of the light curve difficult .
the star can be easily measured ; the photographic data are not biased .
a 47.8-d ( 0.021 cd@xmath8 ) modulation period was determined for the 19711981 interval ( see upper panel in fig .
[ v14sp ] ) .
although strong light - curve modulation was present in the other parts of the photographic data also , a modulation period close to 48 d was not found in any other subset .
a 63.8-d modulation period fits the 19491959 data reasonably well .
the residual spectrum of this data set is shown in the bottom panel of fig .
[ v14sp ] .
the ccd data do not contradict either the 50 or the 64-d modulation period .
* v18 * the star is far from the centre , but a close companion ( iv 75 ; * ? ? ?
* ) at 6 " distance with brightness similar to v18 at minimum light may contaminate the photometric data of low - resolution observations .
oo41 noticed considerable changes in the shape and amplitude of the light curve , which were attributed to probable blazhko effect in kk71 .
the pulsation - period variation of the star is very complex as can be seen in fig .
[ v18oc ] .
the total range of the period variation is larger than 0.0001 d and abrupt period changes occur on time - scales of several thousands of days .
the seasonal light curves of all the photographic and @xmath25 band ccd data are plotted in fig .
[ v18.1 ] .
the observations from different sources are denoted by different symbols .
the pulsation light curve of v18 varies significantly from one season to the next , while the scatter of the light curves in the individual panels does not exceed what is expected from the inhomogeneous , photographic data .
the @xmath1 band ccd light curves of @xcite , @xcite and k00 ( not plotted in fig . [ v18.1 ] ) also show significant variations .
the amplitude of the pulsation was smaller than 0.5-mag in 1977 and 1979 , while no light variability was detected in 1943 and 1984 at all .
there are three stars close to v18 : iv74 , iv75 and iv76 with 14.60 , 15.82 and 17.19 @xmath25 magnitude , respectively @xcite .
the mean magnitude of v18 measured in the 1984 las campanas observations is , however , 15.35-mag .
this brightness value is close to the mean brightness of v18 in any other season , but it is significantly different from the magnitudes of the neighbouring constant stars .
therefore , we exclude the possibility that wrong identification of v18 can account for the constancy of its light curve in these years .
as the seasonal light curves of v18 do not show large inhomogeneity , the modulation cycle has to be relatively long , about hundreds of days .
however , we failed to determine any modulation period when analysing the light curves of different segments of the data .
the phase modulation can not be separated from the irregular period changes ( note the different periods of each column in fig .
[ v18.1 ] ) .
the residual spectrum of any data subset , which is long enough for analysis , is dominated by signals arising from the phase variation and/or the period change .
the analysis of the amplitude variation using the maximum brightness data does not give any definite solution for the modulation period , either .
the maximum brightness
maximum phase plots of the seasonal light curves also do not help to determine the modulation period ( see fig [ v18egg ] ) .
we conclude that the modulation properties and the pulsation period of v18 vary probably on a similar time - scale , making the determination of the modulation period from the available data impossible .
* v19 * light - curve modulation is probable according to the ccd @xmath1 data ( * ? ? ?
the pulsation - period variation is complex , but it can be fitted with a smooth cubic function for the period between 1952 and 1987 as shown in fig .
[ v19oc ] .
if this part of the combined observations is transformed in time to eliminate the period variation as described in paper i , the modified data show clear evidence of an amplitude modulation with a @xmath45 d period .
[ v19max ] shows the amplitude spectrum of the maximum brightness data between 1952 and 1987 , and the data phased with the largest - amplitude frequency , 0.0254 cd@xmath8 ( 39.4 d ) .
the star is far from the centre with no close companion affecting the photometry .
* v24 * the minima of the r96 and k00 ccd @xmath1 light curves differ conspicuously in phase and magnitude as shown in fig .
it clearly indicates that the light curve of v24 is not stable .
no maximum light was measured in r96 .
the star lies in a crowded region , the photographic data are seriously defective .
the pulsation - period variation is irregular , because abrupt and continuous period changes both occur .
* v27 * the ccd data shown in fig . [ v27 ] confirm the blazhko effect of v27 , which has already been recognized by oo41 .
because of crowding problems and irregular fluctuations of the pulsation period , no conclusive result for the modulation period can be drawn from the photographic data . )
dominate the spectrum .
the panels @xmath5 , @xmath6 and @xmath24 enlarge the residual spectrum at around the pulsation frequencies .
the modulation frequencies with 0.0008 cd@xmath8 separation , corresponding to a 1200-d period of the modulation , at @xmath13 , @xmath14 , @xmath15 , @xmath20 and @xmath46 are the dominant frequencies , although yearly alias solutions with 0.0035 cd@xmath8 or 0.0019 cd@xmath8 separations are also possible .
the bottom panel shows the residual spectrum at @xmath22 of another subset of the data ( 19551968 ) .
the signal at @xmath47 cd@xmath8 or at @xmath48 cd@xmath8 separation is the largest - amplitude signal of the entire residual spectrum in this part of the data , too.,width=325 ] * v29 * a separate star , far from the cluster centre .
the ccd light curve ( k00 ) has a somewhat anomalous shape and small amplitude .
the anomalous shape of the light curve was already recognised by oo41 .
v29 is the second shortest - period rrab star in m5 ( @xmath49 d ) .
light curves belonging to the 12 subsets of the photographic observations are shown in fig .
[ v29lc ] .
variation especially on the upper part of the ascending branch can be suspected .
the pulsation period of v29 decreased by 0.00006 d during the century - long time interval of the observations as shown in fig . 1 of paper i. during the second half of the observations ,
an @xmath50 d fluctuation was superimposed on the gradual period decrease .
this is illustrated in the top panel of fig .
[ v29 ] , which is a plot of the @xmath3 data from 1946 to 1993 .
though no significant light - curve variation is detected , the fourier parameters of the 23 year subsets of the photographic data show variations larger than their uncertainties would indicate ( fig .
[ v29 ] ) .
no clear correlation between the variations of the fourier parameters is evident ; however , the changes in the ligh - curve parameters of blazhko stars can be quite complex in some cases ( see e.g. fig . 8 in @xcite and fig . 12 in @xcite ) .
therefore , we suppose that v29 is probably a blazhko variable , with modulation either with a period very close to one year or on a very long ( @xmath50 d ) time - scale . * v30 * the ccd @xmath1 amplitude of k00 is 0.05-mag smaller than that in the @xcite data .
the star is far from the centre .
maximum - brightness variation in the photographic and ccd @xmath25 @xcite data between 1971 and 1993 is also probable .
the fourier spectrum of this data set indicates a modulation frequency of 0.0011 cd@xmath8 ( 900 d ) , however , a yearly alias solution ( 621 d ) is equally possible ( see fig .
[ v30sp ] ) .
the pulsation period of v30 was constant during the century - long time base of the observations .
* v38 * the light curves are different in the r96 and k00 data as shown in fig .
companions at 5``10 '' affect some of the photometric data .
because of irregular fluctuations in the pulsation period , we failed to find the period of the modulation from the photographic data . the pulsation period varied irregularly with a @xmath51 d amplitude .
* v52 * very faint stars are close to v52 , a brighter star is at 12 " distance .
the light curve of v52 is somewhat anomalous , with smaller amplitude and less steep rising branch than that other rr lyrae stars with similar period have .
the pulsation - period variation is complex ; abrupt and continuous changes of the period are detected .
oo41 classified the star as irregular .
the ccd data shown in fig . [ v52lc ] indicate light - curve variability .
the fourier spectrum of the residual photographic data between 19741993 show modulation - frequency components at @xmath13 , @xmath14 and @xmath20 with @xmath39 cd@xmath8 ( 95 d ) ( fig .
[ v52sp ] ) .
the modulation is the most prominent in the phase and steepness of the rising branch .
* v56 * the ccd light curves of r96 and k00 have different shapes and amplitudes .
the maximum - brightness variation is about 0.2-mag in @xmath1 band
. a very faint close companion does not affect the photometry seriously .
the photographic data show strong maximum - brightness variation , too .
modulation - frequency components corresponding to the same , a 115-d modulation period appear in the residual spectra of both the 19711993 and 19521963 photographic data , however , a 170-d alias solution can not be excluded ( fig .
[ v56sp ] ) .
the ccd data do not contradict the 115-d modulation period .
the period changes indicate long - term , cyclic variation . * v58 * a separate star , far from the cluster centre .
no ccd observation is available .
@xcite classified v58 as irregular , while kk71 noted its blazhko effect .
the photographic data indicate modulation with a 59-d period ( fig . [ v58sp ] ) .
strong phase and amplitude modulations are detected .
the pulsation period of v58 has monotonously decreased by 0.00002 d during the one - hundred - years of the observations .
* v63 * the star is far from the centre , there is no special source of photometric distortion .
rw dra - type variability was recognized by oo41 , and @xcite determined a 146.8-d blazhko period of the amplitude variation .
the ccd data from k00 cover only the maximum and descending branch of the light curve with @xmath52 mag differences .
modulation with a 46-d period can be detected in two different parts of the photographic observations ( fig .
[ v63sp ] ) .
strong amplitude and phase modulations of the light curve are evident in both data sets .
the pulsation period has been steadily increasing , but the instantaneous periods show large scatter probably connected to the blazhko modulation of the star .
* v65 * a close ( 3 " ) companion with brightness similar to v65 at minimum light makes the photographic data unsuitable for detecting light - curve variability .
the ccd data show light - curve variation both at maximum and minimum light ( fig .
[ v65 ] ) .
the range of the pulsation period variation was about 0.000025 d ; continuous and rapid changes were observed .
* v72 * a separate star at large distance from the cluster centre . oo41 classified the star as irregular .
no ccd observation of v72 is available .
the pulsation period increased 0.00014 d abruptly sometime between 1917 and 1934 , when no observation was obtained . since 1934 , the pulsation period has not changed systematically , but variations of about 0.05 d were detected in the yearly @xmath3 data ( fig . [ v72oc ] ) .
the fourier spectra of the residuals of two different segments of the data ( 19771991 and 19551968 ) have the largest - amplitude signals close to the main pulsation frequencies , which can be attributed to either pulsation - period variation or long - period blazhko modulation of the light curve .
the dominant signals are , however , at the same ( @xmath53 cd@xmath8 ) separation from @xmath22 in both residual spectra as shown in fig .
therefore , we conclude that some of the scatter in the @xmath3 of v72 is the consequence of light - curve modulation with about a 1200-d period . * v97 * the ccd light curve ( r96 ) shows larger than 0.3-mag amplitude variation , and indicates that the period of the modulation is longer than 50 d. photographic data are confined to the oo41 and the konkoly 1-m ( paper i ) observations . though crowding affects the photographic data seriously , a modulation period of about 211-d ( or its yearly alias corresponding to 499 d ) is presumed in the konkoly data . due to the sparseness of the observations , no definite conclusion on the pulsation - period variation of the star can be drawn .
we searched for the blazhko effect in the collected photometric data of 50 rrab stars in m5 . as a result
, we found clear evidence of the blazhko effect in 18 cases , and , with less certainty , in two additional cases .
however , small - amplitude modulation can not be excluded in the remaining 30 rrab stars either , especially in the ones that lack ccd observations .
therefore , the detected 40 per cent incidence rate is a lower limit of the true occurrence of blazhko stars in m5 .
this is in good agreement with the results of the recent , high - accuracy surveys of rrab stars : 47 and 40 per cent incidence rates of modulated light curves have been found in a sample of short - period ( @xmath54 d ) variables @xcite and in the _ kepler _ field @xcite , respectively .
the modulation periods could be determined for 13 stars .
the results are summarized in table [ bltab ] .
the variable , the time interval of the data used to identify the modulation , the pulsation and the modulation periods , the amplitudes of the amplitude and the phase modulations , the intensity - averaged mean @xmath1 brightness ( see paper i ) and the type of the pulsation - period change are given in the columns .
the formal errors of the modulation periods are 0.10.2 and 12 d for variables with blazhko periods shorther and longer than about 100 d , respectively . however , the incompleteness of the data sampling resulted in alias solutions which were equally acceptable in some cases .
the alias solutions corresponded to modulation periods that differed by tens / hundreds of days .
if the selection of the modulation period was somewhat ambiguous , this was noted in the previous section , and the other possible modulation periods were also given .
the modulation properties are simply read from the folded light curves .
thus they give only rough estimates of the amplitudes of the amplitude and phase modulations , which might , in fact , be even larger for some of the blazhko stars .
the mean @xmath1 magnitudes are derived from the ccd data of a narrow modulation - phase bin .
they are estimated to be accurate within about 0.010.03 mag .
the uncertainties of the mean @xmath1 magnitudes arise from zero - point errors / differences in the ccd data , from the incomplete phase coverage and from the possible changes in the mean brightness in the different phases of the blazhko cycle ( @xmath55 mag ) .
the period distribution of the blazhko variables is different from the distribution of the total sample of rrab stars in m5 .
the blazhko effect favours shorter periods .
the mean period of all of the rrab stars is 0.546 d , but for the blazhko variables it is only 0.504 d. in the sample of rrab variables with periods shorter than 0.55 d and photometric data suitable for detecting light - curve variability , the occurrence rate is as high as 60 per cent .
this is not a unique property of m5 ; @xcite already recognized that ` there may be a tendency for such variables to occur at slightly shorter periods than the class as a whole . ' in fig .
[ sum ] the mean @xmath1 magnitudes and the @xmath1 amplitudes of m5 rr lyrae stars are plotted against pulsation period and the @xmath0 colour index .
data are taken from table 4 in paper i. the observed largest amplitudes of blazhko stars are plotted .
the periods of the first - overtone variables are fundamentalized for direct comparison .
the blazhko stars occupy the short - period range of rrab stars , and are adjacent to the first - overtone rrc variables .
their @xmath0 colours overlap with those of the rrc stars .
in fact , among the eleven bluest rrab stars with mean @xmath0 colour indices falling in the regime occupied by the rrc stars , all but one ( v83 ) are identified as blazhko variables . the poor quality of the photometry of the one , exceptional star ( denoted by asterisk in fig . [ sum ] ) , does not allow detection of any light - curve variability of this star .
llccccccc & time interval [ jd ] & & & @xmath56&period change + & & & pulsation&modulation&(b ) [ mag ] & phase [ min ] & & + v1 & & 2441447 2449104 & 0.521787 & 521.9 & 0.1 & 15 & 15.103 & stable + v2 & a & 2440985 2446592 & 0.526260 & 136.5 & 0.6 & 75 & 15.093 & irreg .
+ & b & 2435629 2438968 & 0.526277 & 129.7 & 0.6 & 75 & & + & c & 2427540 2435310 & 0.526262 & 66.7 & @xmath570.1 & 60 & & + v4 & a & 2442598 2449104 & 0.449618 & 110.1 & 0.1 & 45 & 15.047 & irreg .
+ & b & 2434122 2438193 & 0.449671 & 106.9 : & 0.3 : & 30 : & & + ( v5 : ) & & 2448399 2450664 & 0.545851 & & @xmath570.1(v ) & - & 15.107&decreasing + v8 & a & 2443631 2449104 & 0.546243 & 64.5 : & @xmath570.1 & 15 & 15.085&irreg.@xmath58 + & b & 2440985 2443286 & 0.546237 & 66.5 : & 0.2 & - & & + v14&a & 2440985 2444767 & 0.487188 & 47.8 & 0.5 & 50 & 15.080 & irreg .
+ & b & 2433068 2436762 & 0.487159 & 63.8 & 0.5 & 50 & & + v18 & & 2411148 2444767 & 0.464060 & @xmath59500 & 1.5 & 60 & 15.075&irreg .
+ v19 & & 2434122 2447298 & 0.469987 & 39.4 & 0.4 & - & 15.088 & irreg .
+ v24 & & 2448399 2450664 & 0.478445 & & 0.1@xmath60(v ) & 100@xmath60&15.068 & irreg . + v27 & & 2431259 2449104 & 0.470312 & & 0.3(v ) & - & 14.995 & irreg .
+ ( v29 : ) & & 2448399 2450664 & 0.451427 & & 0.1 : & & 15.127 & decreasing@xmath61 + v30 & & 2440985 2448745 & 0.592176 & 901/621 & 0.2 & - & 15.067 & stable + v38 & & 2440985 2444767 & 0.470419 & & 0.15(v ) & & 15.090 & irreg .
+ v52 & & 2442126 2449104 & 0.501518 & 94.5 & 0.2 : & 45 & 14.976&irreg .
+ v56&a & 2440985 2449104 & 0.534695 & 115.1 & 0.5 & 30 & 15.116&irreg .
+ & b & 2434122 2438193 & 0.534690 & 114.2 & 0.6 & 45 & & + v58&a & 2440985 2449104 & 0.491249 & 59.3 & 0.4 & 30 & & decreasing + & b & 2433068 2439979 & 0.491254 & 58.9 & 0.4 & 30 & & + v63&a & 2440985 2449104 & 0.497684 & 46.4 & 0.4 & 30 & 15.06 : & increasing@xmath61 + & b & 2432000 2439979 & 0.497680 & 46.6 & 0.3 & 30 & & + v65 & & 2448399 2450664 & 0.480671 & & 0.2@xmath60(v ) & & 15.085 & irreg .
+ v72&a & 2443225 2448394 & 0.562276 & 1199.7 & 0.1 & 80& & irreg .
+ & b & 2435251 2439979 & 0.562270 & 1230.8 & 0.1 & 80 & & + v97 & & 2442598 2449104 & 0.544623 & 211:/499 : & @xmath590.4 & @xmath5960&15.042 & insuff .
data + + + another interesting feature of the location of the blazhko stars is that they tend to have faint mean - brightness values .
the mean values of the brightness of the blazhko and the regular rrab stars are @xmath62 and @xmath63 mag , respectively .
no similar result has been previously shown . however , in most studies of globular - cluster blazhko stars , the mean brightness has been determined without any distinction of the pulsation / modulation phase coverage of the data .
these mean magnitudes may be seriously affected by the incomplete light - curve coverage that is typical in photometric observations of globular - cluster variables . in our previous studies of the field blazhko variables ( e.g. in * ? ? ?
* ) it was shown that the mean brightness determined using a complete pulsation light curve at only one blazhko phase does not differ more than @xmath64 mag from the true mean brightness .
the mean - brightness values of blazhko variables are determined in this way here , so they can be regarded as relatively accurate .
these specific features of the blazhko stars raise the possibility that the onset of the blazhko phenomenon is induced by evolution .
one may speculate that the pulsation of rrab stars evolving blueward becomes unstable against the blazhko effect ( whatever it is ) just preceding the mode switch to first - overtone pulsation . a possible relation between the blazhko and
the double - mode stars has been already proposed by @xcite . however , detailed modelling of the mode switch from fundamental to overtone mode is still lacking . reviewing the variable populations of the globular clusters rich in rr lyr stars , one finds that they are either rich in blazhko stars ( ngc362 , ngc 3201 , m5 , @xmath65 cen ) or in double - mode variables ( ic4499 , m15 ) or in both ( m3 , m68 ) .
blazhko phenomenon is quite common among the galactic - field rrab stars @xcite , but double - mode pulsators are rare .
the occurrence of double mode and/or blazhko stars in the different stellar populations , however , does not show clear correlation with any of their properties e.g. metallicity , hb - type , oosterhoff - type .
while blazhko stars appear in all the globular clusters with large rr lyrae population , variables showing fundamental - mode and first - overtone pulsation simultaneously are known basically only in m3 , m15 , m68 , ic4499 .
searching for the properties that foster the appearance of double mode or blazhko modulation would be an important issue in disclosing any possible connection between these phenomena .
there is even a unique variable , v79 @xcite in m3 , which shares the properties of both the double - mode and the blazhko phenomena .
v79 was a double - mode pulsator for 15 years , between 1992 and 2007 .
however , observations made in 20082009 showed it to be pulsating in the fundamental mode with a 65-d modulation period .
thus this is a good example of a star that confirms nemec s ( 1985 ) proposal .
in paper i it was already shown that variables with irregular changes in their pulsation period all show the blazhko effect , provided they have photometric data suitable for detecting light - curve variability . from the 20 blazhko variables listed in table [ bltab ] ,
two do not show any period change , one has an increasing period with irregular fluctuations superimposed , one has a period that increases , but not at a uniform rate , two have periods that decrease steadily , one has strong irregularity superimposed on its gradual period decrease and 12 show nothing but irregular period fluctuations .
( the remaining star does not have enough observations for characterizing its period - change behaviour . ) clearly , it can be concluded that the occurrence of the blazhko modulation is usually accompanied by irregular fluctuations of the pulsation period and vice versa .
we note that @xcite proposed a model that could account for both the irregular period changes and the blazhko effect .
he attributed them to hydromagnetic effects in the atmosphere of the pulsating star .
the amplitude of the modulation of v18 is so large that sometimes it suppresses the pulsation totally .
a similar example is v442 her @xcite with pulsation amplitude varying between 1.4 and 0.2-mag in the @xmath1 band .
modulation periods could not be derived for either star . in both stars , significant changes of
the light - curve shape and amplitude occur from one season to the next , while the seasonal light curves , spanning @xmath66 days , are surprisingly homogeneous .
the pulsation - period variations are extremely large ; random period changes of the order of @xmath67 d are detected in both stars .
another globular - cluster variable that exhibits light - curve changes from one season to the next is the short - period rrab star , v15 , in the oosterhoff - type i cluster m4 @xcite .
it seems that in these special cases the blazhko effect manifests itself in strong , random , abrupt changes of the light curves , lacking any clear periodic behaviour .
this type of the modulation definitely can not be explained by the appearence of non - radial - mode oscillations with frequencies close to the pulsation frequencies .
other blazhko stars with a special character are v29 and v72 .
these stars show hardly any amplitude variation , but their pulsation phase vary considerably on time - scales of some hundreds to thousands of days .
the galactic - field analogue of these stars is rs boo @xcite , with @xmath68 mag amplitude and @xmath69 min phase modulations during its 530-d modulation cycle .
the long - term pulsation - period variation of each of these stars is relatively simple : it is steadily decreasing for v29 , fits two constant periods differing by @xmath70 d for v72 , and has increased throughout the 110 years time base of the observations for rs boo . for v29 and v72 some fluctuations are superimposed on the global trend of their period variations ( see also paper i ) .
the modulations of these blazhko stars are dominated by cyclic or periodic phase variations .
an important issue would be to determine why in some blazhko stars the modulation is dominated by amplitude variations , while in others it is dominated by phase oscillations .
significantly different frequencies of the modulation have been derived for different epochs for two blazhko stars ( v2 and v14 ) , which can not be explained by continuous changes of the modulation period .
the ratios of the detected modulation periods are close to 2:1 and 4:3 for v2 and v14 , respectively .
another case of strong multiperiodicity of the modulation has been documented in ( * ? ? ?
* cz lacertae ) .
the modulation frequencies detected in cz lac were very close to 3:4 and 4:5 resonances .
rich modulation frequency patterns with multiplets of the same modulation frequency were revealed in extended , accurate observations of some blazhko stars ( mw lyr and v1127 aql in * ? ? ?
* ; * ? ? ?
* ) . in these cases the dominant modulation corresponded to the shortest frequency component of the multiplets
however , the examples of v2 , v14 and cz lac show that in some blazhko stars , from time to time , different components of a hidden modulation frequency multiplet may emerge .
if these modulations correspond to elements of series of equidistant - spacing multiplets , their frequency ratios are close to fractions of small integer numbers .
this indicates that the observed modulation - frequency values may not always correspond to the ` base frequency ' , i.e. the shortest frequency component of a multiplet .
one must be cautious in interpreting the data . using a large sample of blazhko variables , including stars from the galactic field , globular clusters , the galactic bulge , the lmc and the sagittarius - dwarf galaxy
, @xcite showed that the larger the pulsation frequency of a blazhko star is , the larger its modulation frequency can be .
the modulation versus pulsation frequencies of the m5 blazhko variables are plotted in fig .
[ mod ] , and it appears that a similar trend might occur in m5 .
however , the ranges of the observed pulsation and modulation frequencies are much smaller in m5 than in the sample investigated by @xcite .
most probably , if a relation between the pulsation and the modulation frequencies indeed exists , its actual form strongly depends on the global parameters of the stellar population studied ( e.g , luminosity , chemical composition , age , etc . ) .
lccccrrr & @xmath71 [ d ] & @xmath72 [ @xmath33d ] & @xmath73 [ d ] & @xmath74 [ d ] & @xmath75 & @xmath76[@xmath77 & @xmath78 + rr gem@xmath79 & 0.3973 & 4.0 & 7.2 & 0.07 & 96 & 0.2 & 5.26 + dm cyg@xmath80 & 0.4199 & 0.5 & 10.6 & 0.07 & -554 & -1.4 & -21.94 + rw dra@xmath81 & 0.4429 & 11.0 & 41.8 & 0.5 & -48 & -0.5 & -0.51 + m5 v4@xmath82 & 0.4496 & 5.3 & 110 & 3.2 & -250 & -6.0 & -1.04 + xz cyg@xmath83 & 0.4665 & 15.0 & 58 & 1.0 & -54 & -0.7 & -0.43 + rv uma@xmath84 & 0.4681 & 0.5 & 90 & 0.5 & -519&-10.0 & -2.70 + xz dra@xmath85 & 0.4765 & 5.0 & 76 & 4.0 & 505 & 8.0 & 3.19 + m5 v58@xmath82 & 0.4913 & 0.5 & 59 & 0.4 & -666 & -8.0 & -5.53 + m5 v63@xmath82 & 0.4977 & 0.4 & 46.5&0.2 & -535 & -5.0 & -5.73 + m5 v2@xmath82 & 0.5263 & 1.7 & 133 & 6.8 & -1582&-40.0 & -6.26 + m5 v56@xmath82 & 0.5347 & 0.5 & 115 & 0.9 & 840 & 18.0 & 3.92 + m5 v8@xmath82 & 0.5462 & 0.6 & 65.5&2.0 & -2790 & -33.0 & -23.3 + m5 v72@xmath82 & 0.5623 & 0.6 & 1200&31 & -2414&-520.0 & -1.12 + + + the long time - base of the observations of the m5 variables also made it possible to determine the modulation properties at different epochs for some variables .
significant changes in the modulation properties were already detected in rr lyr @xcite , rr gem @xcite , xz cyg @xcite and cz lac @xcite . in these stars the strengths of the amplitude and the phase modulations as well as the pulsation and the modulation periods vary .
these changes are probably connected either to multi - periodicity of the modulation or to abrupt and continuous changes of the pulsation period . in table
[ bllong ] we summarize the detected changes in the pulsation and the modulation periods for all those variables , which have data at least for two different epochs .
we note , however , that in those cases when the observed period changes have been determined for several epochs ( e.g. xz dra , xz cyg , rw dra , rv uma ) , no strict relation between the pulsation- and the modulation - period changes of the variables holds .
the connection between the period changes reflects some tendency rather than an exact relation .
therefore , results obtained from two - epoch data have to be taken with caution . the columns of table [ bllong ] give the pulsation and the modulation periods , the absolute values of their changes calculated from the periods given in table [ bltab ] or taken from the literature , the normalized relative period variations , and the relative period and frequency variations for seven m5 and six field rrab stars , listed in order of increasing pulsation period .
examining these data we failed to find any regular pattern , but some general trends are evident .
first , it seems that in general , there is an anticorrelation between the directions of the pulsation and the modulation period changes , i.e. a negative sign of @xmath86 is more common than a positive one .
secondly , it can be noticed that the relative period variations ( @xmath87 ) and their normalized values tend to have larger absolute values for variables with longer pulsation periods than for shorter period ones .
the pulsation - period variation of one of the blazhko stars , v56 , can be described with a long , periodic / cyclic variation .
this gives a unique possibility to test the binary origin of the long - period @xmath3 variation .
if this were caused by the light - time effect the pulsation and modulation periods should have to show similar variation , because @xmath88 is the function of the elements of the orbit for any periodic signal @xcite .
the observed normalised relative period variation of the pulsation and the modulation periods of v56 is , however , much larger than 1 , it is 840 contradicting the orbital origin of the cyclic , long - period variation .
however , keeping in mind that random variations in the modulation periods may override the supposed synchronous variations of the pulsation and the modulation periods , this result is not in fact conclusive .
although blazhko stars are usually numerous in globular clusters , which are rich in rr lyrae stars , we hardly know anything about their properties because of a lack of data suitable for studying them .
results have been restricted to deriving the blazhko periods , but even they have been successfully obtained only in a few cases , e.g. in m3 @xcite . in this paper the rrab population of m5 are reviewed using all the available photometric data of the variables , in order to identify the blazhko stars and to study their behaviour .
modulation periods were determined for 13 out of the 20 blazhko stars that were identified .
the long time base of the observations also made it possible to record changes in the observed properties of some of the blazhko stars .
the most important results of this study can be summarized as follows : * blazhko stars tend to be located in the short - period , low - luminosity region of the fundamental - mode instability strip .
their @xmath0s overlap with the domain occupied by the first - overtone pulsators . *
the pulsation - period variation of most of the blazhko stars is slightly or significantly irregular .
* there is no clear connection between the pulsation and the modulation period changes of the blazhko stars , both parallel and antiparallel changes are detected .
finally , we note that the possibilities of studying globular - cluster blazhko stars are still unexploited .
although these studies would need relatively large observational efforts , the homogeneity of such samples would be helpful for determining what properties of the rr lyrae stars predestine them to become blazhko variables .
the constructive , helpful comments of the referee , katrien kolenberg are much appreciated . the financial support of otka grant k-068626 is acknowledged . c.
clement thanks the natural science and engineering council of canada for financial support .
h. thanks the ` lendlet ' program of the hungarian academy of sciences for supporting his work .
99 arp , h. , 1962 , apj , 135 , 311 bailey , s. i. , 1917 , harvard ann . , 78 , 103 benk , j. , bakos , g. , nuspl , j. , 2006 , mnras , 372 , 1657 benk , j. et al , 2010 , mnras , in press blazhko , s. , 1907 , an , 175 , 325 chadid , m. et al . , 2009 , aipc , 1170 , 235 chadid , m. et al . , 2010 , a&a , 510 , 39 clementini , g. , merighi , r. , pasquini , l. cacciari , c. , gouiffes , c. , 1994 , mnras , 267 , 83 cohen , j.g . ,
gordon , g.a . , 1987 , apj , 318 , 215 cohen , j.g . ,
matthews , k. , 1992 , pasp , 104 , 1205 coutts , ch.m . ,
sawyer hogg , h. , 1969 , publ .
david dunlap observatory 3(1 ) , 1 coutts , ch .
m. , 1971 , in strohmeier , w. ed , new directions and new frontiers in variable star research , proc .
no.15 , verff .
der remeis - sternwarte bamberg ix.(nr.100 . ) , 238 goranskij , v. p. , 1976
, per . zvezdy pril . 2 , 323 goranskij , v. p. , 1980a
1096 , 3 goranskij , v. p. , 1980b
1111 , 4 goranskij , v. , clement , c. m. , thompson , m. , 2010 , in sterken , c. , samus , n. and szabados , l. eds , variable stars , the galactic halo and galaxy formation , published by sternberg astronomical institute of moscow university , russia .
, p.115 hurta , zs . ,
jurcsik , j. , szeidl , b. , sdor , .
, 2008 , aj , 135 , 957 jurcsik , j. , benk , j. m. , & szeidl , b. , 2002 , a&a , 396 , 539 jurcsik , j. , szeidl , b. , nagy , a. , sdor , . , 2005 , acta astr .
55 , 303 jurcsik , j. et al . , 2006 ,
aj , 132 , 61 jurcsik , j. et al . , 2008a , mnras , 391 , 164 jurcsik , j. et al . ,
2008b , mnras , 393 , 1553 jurcsik , j. et al .
, 2009a , mnras , 397 , 350 jurcsik , j. et al .
, 2009b , mnras , 400 , 1006 kaluzny , j. , olech , a. , thompson , i. , pych , w. , krzeminski , w. , schwarzenberg - czerny , a. , 2000 , a&as , 143 , 215 ( k00 ) kolenberg , k. et al . , 2010 , apj , 713 , l198 kukarkin , b.v . , kukarkina , n.p . 1971 , perem
. zvezdy pril . , 1 ( no.1 ) , 1 ( kk71 ) lacluyz , a. et al . , 2004 ,
aj , 127 , 16 nagy , a. , 1998 , a&a , 339 , 440 nemec , j. m. , 1985 , aj , 90 , 204 oosterhoff , p. th .
, 1941 , leiden ann . ,
17 no.4 ( oo41 ) preston , g. w. , 1964 , ann .
a&a , vol 2 . ,
23 reid , i. n. , 1996 , mnras , 278 , 367 ( r96 ) schmidt , e. g. , lee , k. m. , 2000 , pasp , 112 , 1262 shapley , h. , 1927 , harvard bull .
, 851 , 15 sdor , . , szeidl , b. jurcsik , j. , 2007 , a&a , 469 , 1033 sdor , .
et al . , 2010 , mnras , in press storm , j. , carney , b.w . , beck , j.a . , 1991 ,
pasp , 103 , 1264 stothers , r. , 1980 , pasp , 92 , 475 szab , r. et al . , 2010 , mnras in press szeidl , b. , 1976 , in fitch , w. s. ed , multiple periodic variable stars , iau coll .
, assl , vol .
60 , p. 133
szeidl , b. , olh , k. , barlai , k. , szabados , l. , 2001 , communications from the konkoly observatory , no .
102 szeidl , b. , hurta , zs .
, jurcsik , j. , clement , c. , lovas , m. , 2010 , mnras , submitted ( paper i ) | the light curves of 50 rrab ( rr0 ) stars in m5 collected in paper i are investigated to detect blazhko modulation .
18 blazhko stars are identified , and modulation is suspected in two additional cases .
the mean pulsation period of blazhko stars is 0.04 d shorter than the mean period of the entire rrab sample in m5 . among the rrab stars with period shorter than 0.55
d the incidence rate of the modulation is as high as 60 per cent .
the mean @xmath0 colours of blazhko stars overlap with the colours of first overtone rrc ( rr1 ) pulsators .
the mean @xmath1 magnitudes of blazhko stars are on the average 0.05-mag fainter than those of the rrab stars with stable light curves .
blazhko stars tend to be situated close to the zero - age horizontal branch at the blue edge of the fundamental - mode instability strip in m5 .
we speculate that this specific location hints that the blazhko effect may have an evolutionary connection with the mode switch from the fundamental to the overtone - mode pulsation .
[ firstpage ] stars : horizontal branch stars : oscillations stars : variables : rr lyr globular clusters : individual : m5 . |
the original formulation of the stellar population concept ( @xcite@xcite ) was driven by the realization that the ages and metallicities of galactic stars were related to their kinematics . while baade s identification of the young , metal - rich disk stars as pop i , and the old , metal - poor halo stars as pop ii was a crucial starting point , observations in the last half century have identified both old and metal - rich stellar populations associated with the inner parts of spheroids ( e.g. , @xcite@xcite ; @xcite@xcite ) and young and metal - poor stellar populations associated with low - luminosity star - forming galaxies ( e.g. , @xcite@xcite ; @xcite@xcite ) and with the outer parts of luminous disks ( e.g. , @xcite@xcite ; @xcite@xcite ; @xcite@xcite ) .
understanding young , metal - poor stellar populations is an essential task in modern astrophysics , both because it allows us a fuller understanding of star formation and stellar evolution as a function of metallicity , and because it offers us potential insights into the nature of star formation in the early , low - metallicity universe .
our understanding of stellar populations has received an enormous boost from our ability to obtain arcsecond ( or better ) resolution images in the ultraviolet ( uv ) and near infrared ( nir ) . in the nir ,
the last decade has shown the complimentary power of large ground - based surveys ( @xcite@xcite ; @xcite@xcite ) and higher - resolution _ hubble space telescope _ ( hst ) near infrared camera and multiobject spectrograph ( nicmos ) data . in the uv
, we are now seeing the benefits of combining insights from the large , but moderate - resolution , galex ( @xcite@xcite ) database with high - resolution imaging from hst ( e.g. , @xcite@xcite ) .
imaging at hst resolution provides the opportunity to study resolved stellar populations out to distances of @xmath110 mpc . broad wavelength coverage ( uv through nir ) allows us to investigate the full history of star formation in nearby galaxies .
ngc 1311 is a very nearby ( @xmath0 mpc ) , but little - studied late - type ( sbm ) galaxy .
@xcite(@xcite ) identify it as a member of the 14@xmath514 association , a loose group dominated by the luminous spiral ngc 1313 .
table 1 summarizes the basic properties of the system .
ngc 1311 was a target in two broad - band hst snapshot surveys ( go programs 9124 and 9824 ; @xcite@xcite , @xcite@xcite , @xcite@xcite , and 2 below ) . as a result ,
a set of broad - band images spanning a wide wavelength interval ( 0.31.6@xmath6 m ) at sub - arcsecond resolution now exists for this galaxy . as ngc 1311 is quite nearby , its bright star clusters and luminous individual stars are detected as discrete sources .
we can thus probe the spatially resolved star - formation history of ngc 1311 by studying the broad - band spectral energy distributions of its star clusters , its individual stars , and its unresolved light .
our results for the star clusters are presented in @xcite(@xcite ) .
this paper is concerned with the individual luminous stars .
we shall address the unresolved light in a future publication . in 2
we briefly summarize the hst observational data .
we present the observed properties of the luminous stars in 3 , and our analysis of these observations in 4 .
we summarize our conclusions , and discuss issues for further research in 5 .
the data for this study are a set of uv , optical and nir images obtained with the wide - field and planetary camera 2 ( wfpc2 ) and nicmos on board hst .
we have wfpc2 images taken through the f300w , f606w and f814w filters , and a nicmos image taken with the nic3 camera through the f160w filter .
a summary of the observations is given in table 2 .
the f300w image was obtained as part of the hst program go-9124 `` mid - uv snapshot survey of nearby irregulars : galaxy structure and evolution benchmark '' ( r. windhorst pi ) . for this program ,
the centers of the target galaxies were placed on the wf3 chip .
details of the observing and reduction procedures for these data are given in @xcite(@xcite ) .
the f160w image was obtained as part of the hst program go-9824 `` nic3 snaps of nearby galaxies imaged in the mid - uv : the remarkable cool stellar population in late - type galaxies '' ( r. windhorst pi ) .
details of the observing and reduction procedures for these data are given in @xcite(@xcite ) .
the archival wfpc2 f606w and f814w images of ngc 1311 were obtained as part of the hst program go-9162 `` local galaxy flows and the local mass density '' ( r. tully pi ) . the wfpc2 wide - field camera ( wfc ) spatial sampling is @xmath7 per pixel .
the planetary camera ( pc ) spatial sampling is @xmath8 per pixel .
no stars are detected in the f300w pc image , but we do use data from the pc in f606w and f814w .
the nic3 spatial sampling is @xmath9 per pixel .
we show images of the central 42@xmath10 of ngc 1311 in the four observed bands in figure 1 .
we used hstphot(@xcite@xcite ) to extract psf - fitted stellar photometry from the wfpc2 images .
hstphot is designed to perform stellar photometry on wfpc2 images , including aperture corrections , charge - transfer efficiency ( cte ) corrections , and zero - point calibrations .
the zero - points to the vegamag system are thus updated from those of @xcite(@xcite ) . for the nicmos image
, we first applied the non - linearity correction determined by @xcite(@xcite ) .
this refines the zero - point calibration from the 2004 june standard ( @xcite@xcite ) .
we then extracted stellar photometry using the version of daophot ( @xcite@xcite ) embedded in the xvista image analysis package ( @xcite@xcite ) .
for the nicmos photometry , we had to determine the aperture correction manually by measuring the asymptotic count rates for two bright , isolated stars in the observed nic3 image mosaic .
the correction from a 2-pixel radius to infinite aperture is [email protected] magnitudes .
all of our photometry is calibrated to the vegamag system .
we detect a total of 4369 stars in at least two bands .
table 3 shows our photometry for the first ten objects in the sample .
the full table is presented in the on - line edition of this paper .
column 1 gives a running i d number ( in order of increasing right ascension ) .
columns 2 and 3 give the j2000.0 right ascension and declination determined from the wfpc2 wcs header information .
columns 4 and 7 give the wfc chip on which the star is found in the f300w and f606w / f814w observations , respectively .
columns 5 and 6 give the @xmath12 and @xmath13 pixel positions of the star in the f300w image , column 8 and 9 give the f606w / f814w pixel positions , and column 10 and 11 give the nicmos pixel positions .
columns 12 through 15 give the measured magnitudes ( top row ) and errors ( bottom row ) in f300w ( @xmath14 ) , f606w ( @xmath15 ) , f814w ( @xmath16 ) and f160w ( @xmath17 ) . columns 16 through 21 give the measured colors ( top row ) and errors ( bottom row ) in @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , and @xmath23 respectively .
we conducted a series of artificial star tests on our images to determine both the completeness and accuracy of our photometry . for the wfpc2 images
, we used the artificial star capability of hstphot . for the nic3 image
, we used the version of daophot embedded in the xvista image analysis package .
figure 2 and table 4 show our completeness fraction as a function of magnitude in all four bands . our photometric accuracy is shown in figure 3 .
table 5 shows the mean offset and dispersion as a function of magnitude in all four bands .
our 50% completeness limits , in magnitudes , are 22.2 ( f300w ) , 25.7 ( f606w ) , 24.4 ( f814w ) and 20.2 ( f160w ) . at these limits ,
the artificial star tests have offsets that are statistically equivalent to zero , and dispersions of 0.1 to 0.2 mag .
figure 4 shows a suite of color - magnitude diagrams ( cmds ) for the resolved stellar populations of ngc 1311 .
the long - dashed lines show the 50% completeness levels for the relevant bands .
the star clusters studied by @xcite(@xcite ) are not plotted .
the @xmath15 vs. @xmath21 cmd shows two plumes of luminous stars , separated by about a magnitude in color .
the dashed boxes in fig .
4a show the two regions of the cmd that we use to select the stars in these two plumes ( the blue - plume and red - plume stars , hereafter ) .
inspection of fig .
4 allows us to make several qualitative statements about the resolved stellar populations .
first , the blue - plume stars appear to be a mix of hot main - sequence ( ms ) stars and post - main sequence blue supergiants ( bsgs ) , based on where they lie in fig .
second , only the turn - off stars of the hot main - sequence are bright enough in @xmath16 to be unambiguously detected
. the main - sequence stars are bluer on average in @xmath18 than the bsgs ( see fig .
we use this to refine our classification of blue - plume stars as either bsgs ( more luminous and redder than the diagonal line in fig .
4b ) or ms stars ( less luminous and bluer than this line ) .
third , the red - plume stars are red supergiants ( rsgs ) .
they are nearly all undetected in @xmath14 , but show a clear vertical plume structure in fig .
4d . in fig .
4bcd , the bsgs are circled and the rsgs are crossed .
we note that the tip of the red giant branch ( rgb ) of an old stellar population is at an absolute magnitude of @xmath24 , which corresponds to an apparent magnitude of @xmath25 for ngc 1311 .
this is much fainter than the magnitudes of the red - plume stars in fig .
we show a selection of color - color diagrams in figure 5 , chosen to show both the bsgs and rsgs . in all cases ,
the bsgs are well isolated from the other detections , and the rsgs are mixed with red stars from neither plume .
these are the faint red stars in fig . 4d , and appear to be evolved stars from populations older than the bright rsg stars . as we have data in a somewhat unusual set of filters ( even for hst observations
, it seems prudent to compare our data to stellar population models before attempting any more detailed interpretation .
we compare our stellar photometry to isochrones derived from the models of @xcite(@xcite ) . in order to plot the isochrones with the photometry
, we adopt the distance estimate for ngc 1311 of @xcite(@xcite ) , who quote @xmath26 mpc based on the magnitude of the tip of the red giant branch .
there appear to be no metallicity measurements of ngc 1311 in the literature , although @xcite(@xcite ) report [ ] / h@xmath27 , consistent with a moderately low metallicity .
for the adopted distance , we find that the isochrones with @xmath28 provide the best fit to the stellar photometry .
the more metal - poor isochrones have main sequences that are substantially bluer than observed in all available colors .
the isochrones more metal - rich than @xmath29 do not predict stars luminous enough in the uv or blue enough in colors including f300w compared to the data . while crowding and/or binarity could artificially inflate the brightness of some fraction of our stars ,
our artificial star tests indicate that this is not a significant problem at the bright end of our photometry .
we adopt a metallicity of @xmath28 below . in figures 6&7
, we show the suite of cmds and color - color diagrams from figs .
4&5 , with the @xcite(@xcite ) @xmath28 isochrones overlayed .
the most luminous stars in ngc 1311 have absolute visual magnitudes of about @xmath30 .
based on the isochrone comparison , the cmds show a mix of young ( @xmath3110 myr old ) ms stars , and supergiants with ages ranging from @xmath110 myr up to @xmath11 gyr .
there is some evidence of a small population of stars younger than 10 myr ( see fig .
the isochrones are broadly consistent with the stellar distribution in the color - color diagrams , with a few exceptions .
there are stars , including a few of the bsgs , that appear anomolously red in colors involving @xmath17 .
inspection of the images shows that the stars in question typically have faint companions visible in the f606w and f814w images . at the poorer sampling of the f160w image
, these companions are blended with the brighter stars .
if the companion stars are red , this will create exactly the offset we observe in the color - color plots .
there are a few stars , not selected as supergiants , that have colors that are very red in colors involving f606w and redder filters .
these stars are too faint in @xmath15 , given their magnitudes in @xmath16 and @xmath17 , to match the isochrone models .
we speculate that these stars may be dust enshrouded asymptotic giant branch ( agb ) stars .
there is also the possibility that a few objects are unresolved background extremely red objects ( e.g. , @xcite@xcite ) .
the isochrone matching reinforces the qualitative impressions from the cmds and color - color plots .
the blue - plume stars are a mix of hot ms stars with ages of @xmath3110 myr and bsgs with ages from @xmath110 up to @xmath140 myr .
the red - plume stars are red supergiants with ages from @xmath110 up to @xmath1100 myr .
the hot main sequence stars are generally well - matched by the same 10 myr isochrone that describes the youngest of the bsgs .
the distribution of ms and bsg stars in fig .
6abc , as well as the @xmath32 emission reported by @xcite , are evidence for star formation in the last @xmath110 myr , but at a rate much lower than 10 myr ago .
we thus see evidence for a burst of star formation about 10 myr ago superposed on what had been lower - level , but ongoing star formation over ( at least ) the previous 100 myr or so .
the use of the ratio of blue to red supergiants ( b / r ratio hereafter ) as an astrophysical diagnostic of stellar populations was first suggested by @xcite(@xcite ) .
the b / r ratio turns out to depend on ( at least ) the metallicity and age of a stellar population .
@xcite(@xcite ) and @xcite(@xcite ) provide recent work on the relevant theory , while @xcite(@xcite ) and @xcite(@xcite ) present recent observational results in the context of nearby low - luminosity star - forming galaxies . the boxes in fig .
4a enclose 280 blue - plume and 190 red - plume stars ( 295 and 200 , after correcting for incompleteness ) .
however , as noted in 3.2 , not all the blue - plume stars identified in fig .
4a are bsgs .
also , as pointed out by @xcite(@xcite ) , the blue and red supergiant plumes in the cmds of nearby galaxies will , in general , be due to stars of a range of ages .
as , even for a fixed metallicity , the b / r ratio is a function of age , some care is required to choose fair samples of stars to evaluate the b / r ratio .
we have small samples to work with , so we choose to define three age - based samples , using the isochrones in fig .
6a , with rough ages of 10 myr , 18 myr and 32 myr .
it is well known that models of rsgs do not produce stars as red as are observed ( see @xcite@xcite and references therein ) .
thus we include , as rsgs of a given age , those stars that lie to the red of the isochrone red loops at a given luminosity . in table 6
we show the resulting age - based samples of bsgs and rsgs , and the b / r ratios for these samples .
the quoted uncertainties in table 6 are from poisson statistics . at a given age
, we have somewhat smaller b / r ratios than @xcite(@xcite ) found for sextans a. this is consistent with a higher metallicity for ngc 1311 than for sextans a. the sextans a metallicity , from both region spectroscopy ( @xcite@xcite ) and cmd fitting ( @xcite@xcite ) , is @xmath33 , compared to our adopted metallicity for ngc 1311 of @xmath34 .
this conclusion is reinforced by the [ ] / h@xmath35 ratios of the two galaxies reported by @xcite .
in addition to the bsg and rsg samples selected above , we define a sample of hot main sequence ( ms ) stars by taking all stars fainter than @xmath36 mag , and blueward of a line extending from ( @xmath372.6 , 20.7 ) to ( 0 , 22.8 ) in fig .
we show the distribution of all three samples ( ms , bsg , rsg ) on the f300w wf3@xmath5wf4 image in figure 8 .
we note that there are a handful of sources that are not assigned to any of the three groups .
the brightest of these are the star clusters discussed in @xcite .
the fainter sources are only detected in @xmath14 .
these are likely to be ms stars , but as we do not have colors for them , we do not classify them as such .
there are two regions with strong concentrations of ms and bsg stars .
the boxes in fig .
1a enclose these two regions . in figure 9
, we show cmds of the stars in the boxes and in the rest of the galaxy .
nearly half the recent star formation in ngc 1311 has occured in these two small regions , with the remainder broadly distributed throughout the system .
the total area of the two regions is about 0.15 kpc@xmath38 for our adopted distance , or about 2% of the surface area of ngc 1311 .
these regions are also the locations of the strong h@xmath35 emission shown in @xcite .
the regions of enhanced recent star formation are about 10@xmath39 across .
this corresponds to roughly 200 pc for our adopted distance .
velocity - dispersion measurements in late - type galaxies , for either stars or gas , are sparse but have characteristic values of about 10 to 20 km s@xmath40 ( e.g. , @xcite@xcite ; @xcite ; @xcite ) .
if we assume that the velocity dispersion of the stars in the concentrations is similar to this , the stars will travel roughly 100 to 200 pc in 10 myr .
thus the sizes of the concentrations of young stars are consistent with the ages implied by the cmds .
we can also compare these regions to galactic ob associations .
ob associations are unbound groups of young stars that appear to be dispersing with internal velocity disperions of @xmath415 km s@xmath40 ( e.g. , @xcite@xcite ; @xcite@xcite ; @xcite@xcite ) .
this gives a slightly longer lifetime , but one that is still in range of 10s of myr .
the distribution of star cluster ages in ngc 1311 is consistent with a bursting mode of cluster formation , with active episodes of age @xmath110 myr , @xmath1100 myr and @xmath421 gyr ( @xcite@xcite ) .
the resolved stars tell us a complimentary , but different story : they do not show any compelling evidence for substantial periods of quiescence in the last @xmath1100 myr .
our cmds show a clear population of stars with ages @xmath110 myr .
these are the ms and bsg stars seen in all panels of fig .
there is evidence of low - level on - going star formation over the last 10 myr ( see fig .
the distribution of bsgs and rsgs in the cmds tell us that there was on - going star formation in ngc 1311 from @xmath110 myr ago back to at least @xmath1100 myr ago . although we are mainly concerned with the young stellar populations , we note that fig .
6d and the work of @xcite argue that star formation occured in a more or less steady fashion back to ages of at least @xmath11 gyr .
we can make a somewhat more quantitative assessment of the recent star formation history from the luminosity function ( lf ) of the core helium burning stars , as was first pointed out by @xcite . in figure 10
, we show the bsg differential lfs of the bsg and rsg stars , with age points from the models of @xcite . to convert the lfs into a star formation history we need to specify an initial mass function ( imf ) and a stellar evolution model .
we use the salpeter imf @xcite and the padua stellar evolution models ( @xcite@xcite ) , and apply them following the precepts of @xcite .
figure 11 shows the resulting star formation history in the age range 1040 myr from the bsgs .
the error bars in fig .
11 represent statistical uncertainties from the star - counts only .
there are additional systematic uncertainties due to the adopted imf and stellar evolution model
. these will affect the absolute normalization of the star formation history , but should not substantially affect either its shape or its timescale .
the most luminous bsgs have ages of @xmath110 myr . at ages above @xmath140 myr
, it becomes impossible to isolate the bsgs on the cmds . over the time period we can address , the rate of star formation appears to have been rising steadily , ending in a strong enhancement @xmath110 myr ago that corresponds to the most recent epoch of cluster formation .
there are no bsgs more luminous than @xmath43 .
this indicates that star formation did not continue at the rate seen @xmath110 myr ago . in principle
, this is quantifiable by studying the ms lf .
the upper ms stars are most easily isolated in the @xmath14 vs. ( @xmath44 cmd ( see figs . 4b , 6b ) .
the 50% completeness limit is at @xmath45 mag , corresponding to an absolute magnitude of @xmath46 .
this means our data only probe the upper @xmath11 magnitude of the ms lf .
we applied the iterative technique of @xcite to the completeness - corrected ms lf of ngc 1311 , again assuming a salpeter imf and the padua models .
the result is that the upper ms stars provide evidence for star formation from 2 to 2.5 myr ago .
the absolute star formation rate has large statistical uncertainty due to the small number of stars and the large completeness corrections .
but there has clearly been residual ongoing star formation in ngc 1311 since the enhancement in star formation @xmath110 myr ago .
we can extend the star formation history back to @xmath1100 myr by taking advantage of the rsgs .
we caution the reader that theoretical modelling of the evolution of rsgs is on a less firm footing than that of bsgs ( @xcite@xcite ) .
our resulting star formation history will therefore be subject to additional systematic uncertainty for ages larger than 40 myr .
however , we procede , again using a salpeter imf and the padua models .
the resulting star formation history is shown in logarithmic form in figure 12 . over the last @xmath1100 myr
, the star formation rate in ngc 1311 appears to have been gradually increasing , peaking with the enhancement @xmath110 myr ago .
we see no evidence for any gap in star formation over the last @xmath1100 myr .
these data are more consistent with a star formation history described as gasping or breathing ( e.g. , @xcite@xcite ; @xcite@xcite ; @xcite@xcite ) than bursting . in @xcite
we identified a small population of star clusters in ngc 1311 .
eight of these clusters have sufficient multiwavelength photometry to use for age estimation .
these ages break into three groups .
two clusters have ages of @xmath3110 myr .
these clusters would appear to be associated with the enhanced episode of star formation @xmath110 myr ago that is revealed by the bsgs ( see fig .
we note that the total mass in stars formed in the 10 myr enhancement is about an order of magnitude larger than the mass estimated by @xcite for the two young clusters , there are no clusters with ages in the interval from 10 to 50 myr .
the bsgs show that there was low - level on - going star formation throughout this interval , but no bound clusters survive from it . as we note in 4.2 ,
the bsgs are concentrated in two regions with properties similar to galactic ob associations .
that is , the stars in that age range are spatially clumped , but they do not inhabit gravitationally bound clusters .
four clusters have ages between 50 myr and 130 myr . the rsgs show evidence for a gradually increasing star formation rate toward the younger / more recent part of this range .
how this relates to the cluster formation in this epoch may be discernable by a more extensive analysis of the hst advanced camera for surveys ( acs ) data than was reported by @xcite .
however , we note that the total mass in intermediate - age clusters ( @xcite@xcite ) is roughly equal to the total mass in intermediate - age field stars based on the rsg lf .
thus the star formation of @xmath1100 myr ago was efficient at producing clusters dense enough to remain bound over @xmath1100 myr timescales , whereas the 10 myr enhancement appears to have produced mainly unbound stellar associations
. fits to isochrones indicate a metallicity of @xmath2 ( @xmath47 \approx -0.7 $ ] ) , consistent with [ ] /h@xmath35 = 0.08 ( @xcite@xcite ) .
given its absolute magnitude ( see table 1 ) , this places ngc 1311 in the midst of the magnitude - metallicity relation for late - type dwarf galaxies ( e.g. , @xcite@xcite ) .
we show this in figure 13 , where we plot b - band absolute magnitudes and nebular oxygen abudances for star forming dwarf galaxies from a number of recent studies ( @xcite@xcite ; @xcite@xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite@xcite ) . to place ngc 1311 on the figure , we convert the linear metallicity ( @xmath34 ) to logarithmic notation , assuming @xmath48 .
this gives @xmath47 = -0.68 $ ] .
if we further assume that @xmath49 = 0 $ ] for ngc 1311 , and take @xmath50 from @xcite , this gives us an estimate of @xmath51 for ngc 1311 . this is shown in fig .
10 as a large diamond . as noted in 4.1 ,
our metallicity estimate is also consistent with the observed b / r ratios for supergiants as sorted by age .
ngc 1311 is a nearby ( @xmath0 mpc ) late - type ( sbm ) galaxy .
we have analysed stellar photometry from hst images of this galaxy covering a wavelength range from 0.31.6@xmath6 m .
we detect hot ms stars in the uv and optical images , providing strong evidence for star formation as recently as @xmath12 myr ago .
we also detect bsg and rsg stars that provide evidence of ongoing star formation over the last @xmath1100 myr .
our nir data reveal populations of stars as old as @xmath421 gyr , in agreement with the work of @xcite .
the field stars of ngc 1311 show a less burst - like age distribution than do the star clusters in ngc 1311 ( @xcite@xcite ) .
about half of the recent star formation , as traced by the ms and bsg stars , is confined to two regions about 200 pc across in the bright central part of the system . assuming a velocity dispersion for the stars that is typical of galactic ob associations , the crossing time is on the order of the age of 10 myr determined from the cmds .
the metallicity of @xmath2 that follows from the isochrone fits and the b / r ratio places ngc 1311 in the midst of the metallicity - luminosity relation for low - luminosity late - type galaxies .
our results for ngc 1311 are consistent with those found for the general population of late - type , low - luminosity galaxies ( e.g. , @xcite@xcite ; @xcite@xcite ; @xcite@xcite ; @xcite@xcite ; @xcite@xcite ; @xcite@xcite ; @xcite@xcite ; @xcite@xcite ; @xcite@xcite ) .
the arising consensus is that such galaxies form stars over their entire histories .
the rate of star formation may change with time , but only rarely does it cease entirely .
the long - term star forming histories of low - luminosity galaxies are best studied with deep long - wavelength optical ( @xmath52 ) or nir ( @xmath53 ) images .
our uv and optical data allow us to explore the star forming history of ngc 1311 over the last @xmath1100 myr .
our next step is a study of the unresolved light of ngc 1311 with our combined wfpc2/nicmos hst data .
as we have observations in four wavelengths from 3000 to 1.6@xmath54 , we will sample stellar populations from ages of @xmath110 myr to @xmath4210 gyr .
nearby star - forming galaxies are known to have unresolved uv light ( e.g. , @xcite@xcite ) at hst resolution .
the stellar photometry we present in this paper provides information on individual ms stars as late as early b - type stars .
late b - type and a - type stars will contribute significant diffuse uv light that traces the distribution of stars of @xmath1100 myr ages .
the ancient stellar populations , dominating the nir light , appear ubiquitous even in very late - type galaxies ( e.g. , @xcite@xcite ; @xcite@xcite ) .
the combination of our cluster and stellar photometry with an analysis of the unresolved light should give us a clearer picture of the star formation history in this system , and a fuller understanding of the process of star formation in low - luminosity late - type galaxies in general .
pbe would like to thank the department of physics & astronomy at minnesota state university for support from a faculty research grant during this project , and to the school of earth & space exploration at arizona state university for their hospitality during his sabbatical visit .
we thank andy dolphin for answering the first author s hstphot questions with patience .
this research has made use of nasa s astrophysics data system bibliographic services , and the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
this work was supported in part by nasa hubble space telescope grants hst - go-09124 * and hst - go-09824 * awarded by the space telescope science institute , which is operated by aura for nasa under contract nas 5 - 26555 .
koribalski , b. s. , staveley - smith , l. , kilborn , v. a. , ryder , s. d. , kraan - korteweg , r. c. , ryan - weber , e. v. , ekers , r. d. , jerjen , h. , henning , p. a. , putman , m. e. , zwaan , m. a. , de blok , w. j. g. , calabretta , m. r. , disney , m. j. , minchin , r. f. , bhathal , r. , boyce , p. j. , drinkwater , m. j. , freeman , k. c. , gibson , b. k. , green , a. j. , haynes , r. f. , juraszek , s. , kesteven , m. j. , knezek , p. m. , mader , s. , marquarding , m. , meyer , m. , mould , j. r. , oosterloo , t. , obrien , j. , price , r. m. , sadler , e. m. , schrder , a. , stewart , i. m. , stootman , f. , waugh , m. , warren , b. e. , webster , r. l. , & wright , a. e. 2004 , , 128 , 16 martin , d. c. , fanson , j. , schiminovich , d. , morrissey , p. , friedman , p. g. , barlow , t. a. , conrow , t. , grange , r. , jelinsky , p. n. , milliard , b. , siegmund , o. h. w. , bianchi , l. , byun , y .- i .
, donas , j. , forster , k. , heckman , t. m. , lee , y .- w . , madore , b. f. , malina , r. f. , neff , s. g. ; rich , r. m. , small , t. , surber , f. , szalay , a. s. , welsh , b. , wyder , t. k. 2005 , , 619 , l1 .
meurer , g. r. , hanish , d. j. , ferguson , h. c. , knezek , p. m. , kilborn , v. a. , putman , m. e. , smith , r. c. , koribalski , b. , meyer , m. , oey , m. s. , ryan - weber , e. v. , zwann , m. a. , heckman , t. m. , kennicutt , r. c. , jr . ,
lee , j. c. , webster , r. l. , bland - hawthorn , j. , dopita , m. a. , freeman , k. c. , doyle , m. t. , drinkwater , m. j. , staveley - smith , l. , & werk , j. 2006 , , 165 , 307 sharina , m. e. , karachentsev , i. d. , dolphin , a. e. , karachentseva , v. e. , tully , r. b. , karataeva , g. m. , makarov , d. i. , makarova , l. n. , sakai , s. , shaya , e. j. , nikolaev , e. y. , & kuznetsov , a. n. 2008 , , 384 , 1544 .
skrutskie , 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 , , 131 , 1163 wilson , g. , huang , j .- s . , fazio , g. g. , yan , r. , koekemoer , a. m. , salim , s. , faber , s. m. , lotz , j. , willmer , c. n. a. , davis , m. , coil , a. l. , newman , j. a. , conselice , c. j. , papovich , c. , ashby , m. l. n. , barmby , p. , willner , s. p. , ivison , r. , miyazaki .
s. , & rigopoulou , d. 2007 , , 660 , l59 windhorst , r. a. , taylor , v. a. , jansen , r. a. , odewahn s. c. , chiarenza , c. a. t. , de grijs , r. , de jong , r. s. , frogel , j. a. , eskridge , p. b. , gallagher , j. s. , conselice , c. , hibbard , j. e. , matthews , l. d. , mackenty , j. & oconnell , r. w. 2002 , , 143 , 113 lcc @xmath55 & [email protected] & 1 + @xmath56 & [email protected] & 1 + @xmath57 & [email protected] & 1 + @xmath58 & [email protected] & 1 + @xmath59 & 568@xmath115 km / sec & 2 + @xmath60 & 0.07 & 3 + @xmath61 & 0.021 & 3 + d & [email protected] mpc & 4 + @xmath62 & @xmath3715.5 & 1,4 + @xmath63 & 15.4 jykm / sec & 2 + @xmath64 & 0.46 @xmath65 & 1,2,4 + @xmath66 [ ] ) & 33@xmath112 & 5 + @xmath67@xmath68 & 0.08 & 5 + cclcc u6dw7101 m & 9124 & wfpc2/f300w & 27 - 08 - 2001 & 300 + u6dw7102 m & 9124 & wfpc2/f300w & 27 - 08 - 2001 & 300 + u6g22403 m & 9162 & wfpc2/f606w & 22 - 09 - 2001 & 300 + u6g22404 m & 9162 & wfpc2/f606w & 22 - 09 - 2001 & 300 + u6g22401 m & 9162 & wfpc2/f814w & 21 - 09 - 2001 & 300 + u6g22402 m & 9162 & wfpc2/f814w & 21 - 09 - 2001 & 300 + n8ou36010 & 9824 & nic3/f160w & 28 - 11 - 2003 & 512 + cccccccccccrrrrcccccc 1 & 3:20:00.637 & -52:12:00.83 & & & & wf4 & 548.55 & 664.68 & & & & 22.58 & 21.41 & & & & & 1.17 & & + & & & & & & & & & & & & 0.02 & 0.02 & & & & & 0.03 & & + 2 & 3:20:01.006 & -52:12:09.91 & & & & wf4 & 645.57 & 650.82 & & & & 22.03 & 20.93 & & & & & 1.11 & & + & & & & & & & & & & & & 0.02 & 0.02 & & & & & 0.02 & & + 3 & 3:20:02.103 & -52:12:04.99 & & & & wf4 & 618.56 & 540.61 & & & & 23.47 & 22.35 & & & & & 1.12 & & + & & & & & & & & & & & & 0.05 & 0.05 & & & & & 0.07 & & + 4 & 3:20:02.197 & -52:12:05.62 & & & & wf4 & 626.68 & 533.60 & & & & 23.18 & 22.08 & & & & & 1.10 & & + & & & & & & & & & & & & 0.03 & 0.03 & & & & & 0.04 & & + 5 & 3:20:02.265 & -52:12:03.02 & & & & wf4 & 602.26 & 521.75 & & & & 21.84 & 20.55 & & & & & 1.28 & & + & & & & & & & & & & & & 0.01 & 0.01 & & & & & 0.02 & & + 6 & 3:20:02.273 & -52:12:00.11 & & & & wf4 & 573.71 & 515.00 & & & & 22.47 & 21.31 & & & & & 1.16 & & + & & & & & & & & & & & & 0.02 & 0.02 & & & & & 0.03 & & + 7 & 3:20:02.284 & -52:12:01.41 & & & & wf4 & 586.65 & 516.49 & & & & 23.77 & 23.84 & & & & & @xmath370.08 & & + & & & & & & & & & & & & 0.04 & 0.11 & & & & & @xmath690.12 & & + 8 & 3:20:02.389 & -52:11:54.74 & & & & wf4 & 525.16 & 489.27 & & & & 24.51 & 23.92 & & & & & 0.60 & & + & & & & & & & & & & & & 0.08 & 0.14 & & & & & 0.16 & & + 9 & 3:20:02.405 & -52:11:59.42 & & & & wf4 & 569.45 & 501.47 & & & & 23.47 & 22.47 & & & & & 1.00 & & + & & & & & & & & & & & & 0.04 & 0.04 & & & & & 0.05 & & + 10 & 3:20:02.452 & -52:11:45.89 & & & & wf4 & 437.16 & 468.12 & & & & 23.48 & 22.53 & & & & & 0.95 & & + & & & & & & & & & & & & 0.04 & 0.04 & & & & & 0.06 & & + ccccc 17.0 & & & & 0.99 + 17.5 & & & 0.97 & 0.90 + 18.0 & & & 0.97 & 0.92 + 18.5 & 0.97 & & 0.97 & 0.91 + 19.0 & 0.97 & 0.96 & 0.97 & 0.82 + 19.5 & 0.97 & 0.97 & 0.97 & 0.65 + 20.0 & 0.97 & 0.97 & 0.95 & 0.54 + 20.5 & 0.96 & 0.97 & 0.96 & 0.37 + 21.0 & 0.96 & 0.96 & 0.96 & 0.24 + 21.5 & 0.92 & 0.96 & 0.97 & 0.16 + 22.0 & 0.58 & 0.96 & 0.97 & + 22.5 & 0.17 & 0.96 & 0.96 & + 23.0 & 0.01 & 0.96 & 0.95 & + 23.5 & & 0.96 & 0.94 & + 24.0 & & 0.95 & 0.83 & + 24.5 & & 0.94 & 0.38 & + 25.0 & & 0.88 & 0.06 & + 25.5 & & 0.61 & 0.01 & + 26.0 & & 0.18 & & + 26.5 & & 0.03 & & + 27.0 & & 0.01 & & + ccccccccc 17.0 & & & & & @xmath370.028 & 0.230 & 0.008 & 0.014 + 17.5 & & & & & @xmath370.021 & 0.203 & 0.014 & 0.030 + 18.0 & & & & & @xmath370.004 & 0.075 & 0.021 & 0.037 + 18.5 & & & @xmath370.005 & 0.094 & @xmath370.002 & 0.012 & 0.031 & 0.053 + 19.0 & @xmath370.008 & 0.025 & @xmath370.001 & 0.010 & @xmath370.002 & 0.013 & 0.060 & 0.115 + 19.5 & @xmath370.011 & 0.036 & @xmath370.001 & 0.011 & @xmath370.003 & 0.016 & 0.088 & 0.157 + 20.0 & @xmath370.015 & 0.043 & @xmath370.001 & 0.013 & @xmath370.005 & 0.022 & 0.120 & 0.202 + 20.5 & @xmath370.021 & 0.061 & @xmath370.002 & 0.016 & @xmath370.007 & 0.026 & 0.195 & 0.295 + 21.0 & @xmath370.027 & 0.081 & @xmath370.003 & 0.017 & @xmath370.010 & 0.034 & 0.268 & 0.426 + 21.5 & @xmath370.024 & 0.109 & @xmath370.004 & 0.022 & @xmath370.013 & 0.050 & 0.313 & 0.570 + 22.0 & @xmath370.009 & 0.144 & @xmath370.008 & 0.033 & @xmath370.022 & 0.061 + 22.5 & @xmath690.112 & 0.238 & @xmath370.007 & 0.043 & @xmath370.028 & 0.091 + 23.0 & @xmath690.358 & 0.463 & @xmath370.011 & 0.066 & @xmath370.044 & 0.144 + 23.5 & @xmath690.899 & 0.996 & @xmath370.013 & 0.097 & @xmath370.043 & 0.177 + 24.0 & & & @xmath370.025 & 0.120 & @xmath690.054 & 0.216 + 24.5 & & & @xmath370.023 & 0.169 & @xmath690.391 & 0.508 + 25.0 & & & @xmath690.010 & 0.217 + 25.5 & & & @xmath690.196 & 0.352 + 26.0 & & & @xmath690.728 & 0.871 + 26.5 & & & @xmath691.699 & 1.842 + 27.0 & & & @xmath692.524 & 2.645 + ccccccc 10 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 18 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 32 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + | we have extracted psf - fitted stellar photometry from near - ultraviolet , optical and near - infrared images , obtained with the _ hubble space telescope _ , of the nearby ( @xmath0 mpc ) sbm galaxy ngc 1311 .
the ultraviolet and optical data reveal a population of hot main sequence stars with ages of 210 myr .
we also find populations of blue supergiants with ages between 10 and 40 myr and red supergiants with ages between 10 and 100 myr .
our near - infrared data shows evidence of star formation going back @xmath11 gyr , in agreement with previous work .
fits to isochrones indicate a metallicity of @xmath2 .
the ratio of blue to red supergiants is consistent with this metallicity .
this indicates that ngc 1311 follows the well - known luminosity - metallicity relation for late - type dwarf galaxies .
about half of the hot main sequence stars and blue supergiants are found in two regions in the inner part of ngc 1311 .
these two regions are each about 200 pc across , and thus have crossing times roughly equal to the 10 myr age we find for the dominant young population .
the luminosity functions of the supergiants indicate a slowly rising star formation rate ( of @xmath110@xmath3 ) from @xmath1100 myr ago until @xmath115 myr ago , followed by a strong enhancement ( to @xmath110@xmath4 ) at @xmath110 myr ago .
we see no compelling evidence for gaps in the star - forming history of ngc 1311 over the last 100 myr , and , with lower significance , none over the last gyr .
this argues against a bursting mode , and in favor of a gasping or breathing mode for the recent star - formation history . |
the geant4 toolkit @xcite includes a collection of models and packages for hadronic physics which are applicable to various particle transport problems @xcite . in geant4
, final state generation is separated from the access and use of cross sections and from tracking .
it is therefore possible and desirable to have independent , alternative physics models . in order to optimize these models and to identify their most appropriate application ,
detailed comparisons between experimental data and model predictions are required .
the cascade energy region from the reaction threshold to incident hadron energies of a few gev is problematic for all existing monte carlo packages . in order to verify and further develop geant4 models in this energy range , a verification suite has been created .
the stages of the verification , including experimental data handling , event generation and comparison to data are described here .
results of the comparison are also presented .
the verification suite is generic , as it is based on an abstract interface to a final state hadronic interaction generator .
this interface can be found in level 2 of the geant4 hadronic physics framework @xcite .
it focuses on quantities relevant to the lhc radiation environment and spallation source targets .
the modular structure of geant4 allows the generation of single events with a known incident particle energy and an explicitly defined hadronic final state generator .
the kinematics of secondaries produced in the interaction are analyzed and the resulting angular , momentum , energy , and baryon number spectra are stored in histograms .
the energy - momentum balance can be controlled as well .
the histograms are compared to published measurements of the differential and double differential cross sections , @xmath0 , @xmath1 , @xmath2 , and the invariant cross sections , @xmath3 .
the cross section contributing to the @xmath4-th bin of the histogram are given by @xmath5 where @xmath6 is the total cross section for the interaction being tested , @xmath7 is the number of simulated events in the sample , and @xmath8 is the number of times bin @xmath4 is incremented during the run .
each bin represents a small region of phase space such as @xmath9 , @xmath10 , or @xmath11 , into which the secondary particle goes after being produced in the interaction .
the double differention cross section is estimatated as @xmath12 the verification suite is organized into a number of test cases , each defined by a unique incident beam energy and a single target nucleus .
all data files , macro files , kumac files and results for each case are stored in a separate subdirectory .
the aida abstract interface @xcite is used in the suite for histogram creation and filling .
the output is done in ascii and hbook formats .
the verification is done by comparing simulation results with experimental data mainly from the exfor database @xcite .
only data with absolute measurements of the differential cross sections are utilized in the suite .
the data downloaded from the database are re - formatted in order to provide intermediate files acceptable for paw analysis . in some cases re - binning of the data
was performed during this process .
this was required mainly for the low energy part of the spectra .
if , within a given test case , only double differential cross section data are available , the single differential cross sections are obtained by numerical integration .
initial verifications have been performed for neutrons and pions produced by protons incident upon various targets . for these test cases
the initial proton energies were all below 1 gev . in this energy region
the inclusive reaction channel @xmath13 has been studied experimentally for many years .
the secondary neutrons can be identified and their energies can be measured with good precision by using time - of - light techniques . because this reaction is important for many applications , a significant number of test cases have been created for it ( table [ t_pn ] ) .
.neutron production by incident protons .
[ cols="^,^",options="header " , ] [ t_ppi ] the user interface is implemented by macro files which allow the specification of various parameters of the verification .
these include : * initial particle and its energy , * beam energy spread , * target nucleus , * hadronic interaction generator , * generator options , and * histogram types and bins .
both linear and logarithmic binning are available .
each verification test case contains a script which executes the macro files .
hence the user can perform the complete verification process by issuing a single command .
the results are stored in hbook format and can be processed in paw by prepared kumac files .
the verification suite has been used extensively during the release phase of the new geant4 bertini cascade and binary cascade packages @xcite .
the bertini cascade model is a classical cascade code which was described in detail at another presentation at this conference @xcite .
the binary cascade introduces a new approach to cascade calculations .
the interaction is modeled exclusively on binary scattering between reaction participants and nucleons .
the nucleus is described by a detailed 3-dimensional model in which the nucleons are explicitly positioned in phase space .
free hadron - hadron elastic and reaction cross sections are used to define the collisions .
propagation of the particles in the nuclear field is done by numerically solving the the equation of motion .
the cascade begins with a projectile and a description of the nucleus , and terminates when the both average and maximum energy of all particles within the nuclear boundary are below a given threshold .
the remaining nuclear fragment is treated by pre - equilibrium decay and de - excitation models @xcite .
the single and the double differential inclusive neutron spectra are very sensitive to the physics model used in the cascade code . as shown in fig .
[ fig1 ] , the binary cascade reproduces the data rather well for all targets , except for energies below 50 mev where neutron evaporation is important .
in contrast , the bertini cascade ( fig.[fig2 ] ) does well below 50 mev for all but the lightest nuclei . above 50 mev
only the general trend of the data is reproduced , while the overall normalization improves as @xmath14 increases .
[ fig3 ] shows forward scattering in which the proton transfers its energy to one target neutron . here
the binary cascade describes the data above 50 mev .
the bertini cascade ( figs.[fig4 ] ) does less well , significantly underestimating the data at higher neutron energies .
this discrepancy is reduced for heavier nuclei .
additional verification test cases indicate that the binary cascade approach is reasonably accurate for other angles and energies as well . fig .
[ fig5 ] shows comparisons for aluminum at forward and backward neutron angles at energies of 113 , 256 , 585 and 800 mev . at 113 and 256
mev , agreement at all angles is good . at 585 and 800
mev the backward angle spectra are not well - reproduced .
the same set of plots is shown in fig .
[ fig6 ] for iron , and in fig .
[ fig7 ] for lead . for both targets the same trends apply as
were observed in aluminum . ] ] ] ] another useful test of the cascade codes is to look at the double differential cross sections for the production of @xmath15 and @xmath16 .
experimental cross sections show that @xmath15 production by protons is significantly larger than that for @xmath16 .
this feature is well - reproduced by the binary cascade as shown in fig.[fig8 ] .
the trend of the cross section versus energy is also reproduced , although the overall normalization is underestimated by a factor of 2 - 3 for carbon , aluminum and nickel .
similar plots for @xmath15 and @xmath16 are shown for an incident proton energy of 730 mev .
figs.[fig9 ] and [ fig10 ] show @xmath16 and @xmath15 at forward and backward angles from aluminum , while figs .
[ fig11 ] and [ fig12 ] show @xmath16 and @xmath15 from copper . in all these cases
the observed ratio of @xmath15 to @xmath16 production is reproduced . as was the case for 597 mev incident protons ( fig .
[ fig8 ] ) , the trend of cross section versus energy is also reproduced .
however at 730 mev it is seen that the binary cascade does an increasingly poor job of reproducing the overall normalization as the pion angle increases . while the existing verification suite has already been used extensively ,
further development is required .
so far the validity of the cascade codes has only been demonstrated for incident protons at the low end of their energy ranges . in order to test the full range of energies and particles ,
many more test cases are required , including : * incident proton energies up to 15 gev , * incident pion energies up to 15 gev . * incident neutrons , and * proton inclusive spectra .
a verification suite for hadronic interaction models in the cascade energy region has been designed and implemented .
this suite has been used both to develop the bertini and binary cascade model codes for the geant4 toolkit and to compare their predictions with data from thin target scattering experiments .
the suite has proved to be an efficient tool for model validation which can be used for the development and testing of other hadronic interaction models .
the suite is also being expanded to include tests for more types of incident particles and higher energies .
+ 99 geant4 ( s. agostinelli et al . ) , geant4 : a simulation toolkit , slac - pub-9350 , aug 2002 .
86pp . will be published in _ nucl .
instrum .
a_. j. p. wellisch , geant4 hadronic physics status and validation for large hep detectors .
talk at this conference momt002 . | a geant4 hadronic process verification suite has been designed to test and optimize geant4 hadronic models in the cascade energy range .
it focuses on quantities relevant to the lhc radiation environment and spallation source targets .
the general structure of the suite is presented , including the user interface , stages of verification , management of experimental data , event generation , and comparison of results to data .
verification results for the newly released binary cascade and bertini cascade models are presented . |
gp com ( @xmath7g6129 ) is a high proper motion star @xcite with an optical spectrum dominated by broad hei emission lines @xcite .
warner and smak found rapid photometric variability , similar to the flickering observed in cataclysmic variable stars and suggested that , like them , it is a binary star , although they found no convincing periodicity .
nather , robinson & stover ( 1981 , hereafter nrs ) found a narrow emission - line component which varied in radial velocity on a period of @xmath8 , providing strong support for the binary nature of gp com .
the variable component moved between the two outer peaks of the triple - peaked emission lines , similar to phenomena observed in hydrogen - dominated cataclysmic variable stars ( e.g. u gem ) .
thus , by analogy , nrs identified the moving component with the region where the gas stream hits the accretion disc . the short period and absence of hydrogen
qualifies gp com as one of the am cvn group of accreting double - degenerate systems ; it is in fact the longest period of these systems and stands out from the crowd as the only one with a spectrum strongly in emission and as the only one to show an ` s'-wave . the most prominent element other than helium is nitrogen in the form of nv in iue @xcite and hst @xcite spectra and ni in spectra covering the r- and i - bands @xcite
. the absence of hydrogen and the strong helium and nitrogen emission are consistent with our seeing material from the core of the star that has undergone hydrogen burning and cno - cycle processing of most of the carbon and oxygen into nitrogen .
this is as expected given the very short period of gp com .
the emission lines from gp com show the characteristic double - peaked , broad profiles of emission from an accretion disc .
however , unlike hydrogen - dominated cataclysmic variables with such profiles , gp com also sports a narrow component at the centre of each emission line which we will refer to as the `` central spike '' .
the origin of this component is unclear .
nrs could find no radial velocity variability of the central spike and suggested that it might come from a nebula surrounding gp com , by analogy with old novae .
however a subsequent search failed to find any nebula @xcite . on the other hand the mass donor star in gp
com is probably of such a low mass that the spike could originate on the mass accretor and still show little radial velocity variability .
we look at this component again in this paper .
another remarkable feature of gp com was seen in hst spectra @xcite in which flaring was seen in the emission lines , representing a factor of 5 change in flux from minimum to maximum .
this is probably driven by x - ray variability @xcite which might also influence the optical lines .
we look for this in the time - resolved data we present here .
we begin with a description of the observations and an analysis of orbital variability in our data .
we used the @xmath9 m isaac newton telescope on the island of la palma to take 414 spectra on three nights from the 21st to the 23rd april 1988 .
the spectra were taken with the intermediate dispersion spectrograph ( ids ) with an image photon counting system ( ipcs ) detector , and cover the range @xmath10 to @xmath11 in 2040 pixels with a full width half maximum ( fwhm ) resolution of 2 pixels .
apart from the first six exposures of 120 seconds each , all exposures were 100 seconds long , taken through a @xmath12 by 50 arcsecond slit aligned to a position angle of @xmath13 to capture simultaneous spectra of the variable gp com and a nearby comparison star .
arc spectra were taken every half hour or so to track flexure in the spectrograph .
spectra of the comparison star and the flux standards bd@xmath14 and feige 34 were taken to provide flux calibration .
the sky background was estimated by interpolation from uncontaminated sky regions on each side of the object spectra .
after subtraction of the background , the spectra were extracted by summation perpendicular to the dispersion .
flat field corrections were applied at the same time .
we fitted 13th order polynomials to the average arc spectrum for each night .
the two lowest order terms were then refitted for the individual arc spectra .
the root mean square ( rms ) of the fits was typically @xmath15 , equivalent to @xmath16 or @xmath17th of a pixel at our dispersion .
the wavelength scales for gp com were then interpolated between neighbouring arc spectra .
this procedure eliminates the effects of flexure which amounted to @xmath18 during the night .
finally the flux standard spectra were used to correct for the instrumental sensitivity and the comparison star spectra were used to correct for slit losses .
the average of our spectra has already been discussed in marsh et al . .
for convenient reference , we show it once more in fig . [
fig : avspec ] .
the spectrum confirms the earlier investigations of burbidge & strittmatter and nrs .
the only strong lines detected are those of hei and heii , and these are broad and triple - peaked , a structure seen most clearly in the hei 5015 line .
smak and nrs suggest that the lines are made up of a broad , double - peaked profile from an accretion disc plus a separate narrow component near the centre of the line .
the variation in the relative strength of the central `` spike '' from line to line supports this hypothesis .
double - peaked profiles are well known from the hydrogen - dominated cataclysmic variables @xcite , but the central spike is unique to gp com . the first reliable detection of a periodicity in gp com was made by nrs who discovered a narrow emission - line component ( in addition to the central spike ) which varied in velocity with a period of @xmath19 .
the velocity of the component varied from @xmath20 to @xmath21 taking it between the two outer peaks of the lines .
such ` s'-waves are well - known from other cvs , and are associated with the region where the gas stream from the secondary star hits the disc .
if this is the case in gp com , then @xmath8 is also the orbital period of the binary .
the ` s'-wave is also present in our data . in the upper panel of fig .
[ fig : trail ] we show the phase binned trailed spectrum of gp com , and in the lower panel we show the result of subtracting the average spectrum from this .
fifty phase bins were equally spaced around the cycle and they are repeated twice for clarity .
the ` s'-wave is clear in all the lines , while there is little other variability .
[ fig : trail ] also shows the lack of radial velocity variation in the central spike .
thus it is the ` s'-wave that gives the best way of measuring the orbital period in gp com . in order to measure the period quantitatively
we undertook a multi - gaussian fit which was made to all the spectra and all lines at once , representing the ` s'-wave as a single gaussian with the same sinusoidal radial velocity curve for all the lines .
the period was one of the free parameters of the fit , and is determined by the need to keep the ` s'-wave in step over the three nights of our run .
we obtained a period of @xmath22 days or @xmath23 .
this differs from nrs s value by @xmath24 , but given uncertainties caused by the round - off implicit in nrs s value , we doubt this is significant .
there is no clear indicator of the conjunction phase for gp com , and thus we chose an arbitrary zero point close to the middle of our observations to compute phases .
thus we use the following ephemeris for all phases in this paper : @xmath25 this is the ephemeris used in constructing fig .
[ fig : trail ] . we will discuss a possible constraint on the conjunction phase in section [ spike ] . the sub - class of cataclysmic variables called the su uma stars are known for quasi - periodic flaring behaviour during outburst which occurs on a period somewhat longer than the orbital period .
these flares , usually known as `` superhumps '' , are thought to be the result of a distortion of the outer disc induced by the 3:1 resonance @xcite between the disc and binary orbit .
the distortion only occurs if the 3:1 resonance radius is inside the disc and this requires the mass ratio @xmath26 ( @xmath27 ) to be less than about @xmath28 .
this neatly explains why the su uma stars are all at the short period end of the cataclysmic variable period distribution because then the secondary star is of relatively low mass .
the restriction on mass ratio is easily satisfied by the even shorter period gp com , and so it is of interest to see whether it shows any superhump - like behaviour . since our spectra
are slit - corrected we can attempt this .
we computed light curves over two regions avoiding the emission lines .
the first of these covered 4250 to 4350 and 4510 to 4640 while the second covered 4750 to 4880 and 5080 to 5190 .
periodograms @xcite of these light curves are shown in fig .
[ fig : perg ] .
these do indeed show some evidence for a signal on a slightly longer period than the spectroscopic one ; the difference visible in the right - hand panels of fig .
[ fig : perg ] corresponds to a difference of about @xmath29 cycles over the 2 day baseline of our observations , which is easily measurable .
thus our data are at least consistent with the possibility of superhump - like behaviour in gp com . however , given that the signal in the continuum near the spectroscopic period is only one of several peaks of similar power , the evidence can only be regarded as suggestive .
it should be noted that superhumps in su uma stars are only seen during some outbursts , and so it would not be surprising if gp com failed to show them .
the emission lines in gp com show the broad , doppler - shifted profiles produced by accretion discs .
doppler tomography provides a powerful method of interpreting such profiles .
we imagine that the profiles are built from the sum of many ` s'-waves which have radial velocity @xmath30 at orbital phase @xmath31 .
the strength of each ` s'-wave is plotted with coordinates @xmath32 , @xmath33 to build up an image .
the coordinates are chosen for consistency with a coordinate system in which the @xmath34 axis points from the primary to the secondary star and the @xmath35 axis points in the direction of the secondary star s motion .
a detailed description of the computation of the images is given in marsh & horne .
images of the seven lines are presented in fig .
[ fig : dopp ] .
these images were computed in three sets , the first covering hei 4387 and 4471 , the second heii 4686 and hei 4713 and the third hei 4921 , 5015 and 5047 . in
this way blending between lines could be accounted for .
the fits to the lines were very good with reduced @xmath36 of order @xmath37 .
all lines show a similar structure with a central spot , matching the central spikes of figs .
[ fig : avspec ] and [ fig : trail ] , an offset spot equivalent to the ` s'-wave and a smooth ring - shaped background for the rest of the disc .
we have scaled the lines so that the ring from the disc is seen at about the same level in each case which brings out the marked variation in the relative strengths of the central spike and bright - spot components .
the outer ring visible in heii 4686 is a consequence of its blending with hei 4713 ( which has a similar but less obvious artefact ) .
it is the inner ring that corresponds to the edge of the heii 4686 emission region and it can be seen to be markedly larger than its equivalents in the hei lines .
this indicates that the heii 4686 emission does not cover all of the disc , a point we will return to in section [ sec : flare ] when we discuss flaring behaviour in gp com .
the spot equivalent to the ` s'-wave appears to have some structure in several of the images of fig .
[ fig : dopp ] , most obviously in hei 4387 . on examining the data more closely ( fig .
[ fig : trail ] ) , it turns out that the ` s'-wave is significantly non - sinusoidal , a phenomenon also seen in the strong ` s'-wave of wz sge ( this is best seen by looking at fig .
[ fig : trail ] sideways ) . in order to assess the departure from a sinusoid
, we returned to the multi - gaussian fits , forcing the velocity of the s - wave to be the same in all lines .
we first fitted the ` s'-wave velocity as before when determining the period but now , rather than just a simple sinusoid , we fitted a fourier series of the form : @xmath38 with this as a starting point , we then measured velocities individually from 50 phase - folded spectra ( the signal - to - noise being too low to attempt this on the raw data ) . the results and fourier
fit are plotted in fig . [
fig : swave ] .
the parameters of the fourier fit are listed in table [ tab : fourier ] .
.parameters of the fourier fit to the s - wave in gp com [ cols= " < , > , < , > , < , > , < " , ] [ tab : spike ] the last entry denoted by `` all '' was derived from a simultaneous fit to all lines in which they were forced to have the same @xmath32 and @xmath33 .
individually the signal - to - noise in the lines is too marginal to be confident of a detection , but the simultaneous fit to all of them , which gives @xmath39 , @xmath40 , is statistically significant .
however , it should be remembered that this measurement was made against a background of disc and ` s'-wave and that the semi - amplitude is only of order 1/10@xmath41 of the fwhm of the spike .
nevertheless the detection supports the accretor origin for the central spike . after making a small correction for noise - induced bias ,
our measurement is equivalent to an amplitude of @xmath4 .
this can be compared to nrs s measurement of @xmath42 based upon the emission line wings .
the values are consistent , although unfortunately nrs do not specify the phasing of their measurement .
we have not tried to measure the amplitude from the line wings as they are so broad that the measurement of such small amplitudes can not be relied upon , as nrs warn .
if the spike truly does reflect the motion of the accretor , and the ` s'-wave originates from the gas stream / disc impact region , then their relative phases and amplitudes depend upon the mass ratio .
the connection is illustrated in fig .
[ fig : roche ] which shows that the s - wave parameters listed in table [ tab : swave ] ( plotted as the cluster of points in upper - right quadrant of each panel ) are consistent with the measured position of the spike ( which lies on the straight , dashed lines ) for a mass ratio @xmath43 if the ` s'-wave represents the velocity of the gas stream or @xmath44 if it represents the velocity of the disc along the path of the stream . in the former case
the disc would extend to 75% of the way to the inner lagrangian point , whereas in the latter it would only need reach @xmath45% of the way .
both mass ratios and disc radii are reasonable for gp com , and provide further circumstantial evidence that the spike may indeed be from the accretor .
if this interpretation is correct , then inferior conjunction of the secondary star ( the expected phase of eclipse of the disc , if there was one ) , occurs at orbital phase @xmath46 on our ephemeris .
the same method was used to produce the predicted stream and disc ` s'-waves plotted in fig .
[ fig : swave ] .
so far we have mainly considered variations with orbital phase .
however , strong flaring behaviour has been observed in hst data which does not appear to be related to orbital phase @xcite . in the hst data the flux of nv 1240
was seen to increase by a factor of up to five during three flares which occurred in the 13 hour observing interval , while the continuum only increased by @xmath47% .
it is likely that this variation is driven by x - ray variability through photo - ionisation , and indeed , variations have been seen in rosat data on gp com @xcite .
if so , we can expect similar effects at optical wavelengths . that this is so is demonstrated in fig .
[ fig : lig ] which shows the light curves of the continuum and emission lines of gp com during our run .
there is significant , correlated variability in all components , and it has a larger amplitude in the lines than the continuum .
we can derive a spectrum for the flaring component using the method employed by marsh et al . .
in this the spectra are modelled as the sum of a spectrum representing the mean plus multiples of another spectrum representing the flaring component . both spectra and the multipliers ( which are different for each of the 414 spectra ) are optimised by minimising @xmath36 .
the constant and flare components derived from our data are shown in the top two panels of fig .
[ fig : flare ] .
it has to be remembered that our spectra were taken through a narrow slit , and although they were corrected for slit losses , it is likely that some part of the `` flaring '' could be artificially induced .
this would have the effect of making the flare spectrum look like the mean spectrum .
therefore as a further check we repeated the computation of the flare spectrum after subtracting the continuua of the spectra .
this allows the variations in the lines to determine the flare spectrum rather than the lines and continuum and should weaken the influence of poor slit corrections , although not eliminate them entirely .
the result of this is shown in the lower panel of fig .
[ fig : flare ] and probably gives the most accurate representation of the line profiles .
expanded views of four of the line profiles are shown in fig .
[ fig : flare_prof ] .
looking first at the upper two panels confirms the behaviour seen in the light curves of fig .
[ fig : lig ] .
the emission lines are relatively stronger than the continuum in the flare compared to the mean .
this is exactly as found in the hst data @xcite , although not to such a marked extent ( the possible contamination due to poor slit loss correction should be remembered however ) .
the line ratios in the two spectra are very similar with the exception of heii 4686 which is stronger in the flare component ; with hindsight the heii 4686/hei 4713 light curve can be seen to be the most modulated in fig [ fig : lig ] .
this supports photo - ionisation as the driver of the flares .
finally there are substantial differences in the line profiles ( fig .
[ fig : flare_prof ] ) .
the profiles in the flare spectrum have broader wings than they do in the mean spectrum , indicative of a larger contribution from the inner disc and they also have a weaker central spike component compared to the mean .
however , it is important to note that the flare spectrum does appear to have some central spike component , although the caveat about slit loss contamination is worth repeating .
the weakening of the central spike leads to profiles with more pronounced double - peaks .
most remarkable of all is the heii 4686 profile which in the flare spectrum has a blue - shifted peak at around @xmath48 ( the blending with hei 4713 rendering it impossible to be sure of the red - shifted peak ) .
if interpreted as a standard profile from an accretion disc , this suggests that the heii 4686 emission region only extends out to about 1/4 of the radius of the disc ( assuming that the hei lines come from all of it ) .
this peak is consistent with the radius of the inner ring of the heii 4686 map of fig .
[ fig : dopp ] .
the persistent narrow emission at the centre of all the lines is a puzzling feature of gp com .
the nearest equivalent we know of is the zero velocity emission observed in ip peg and ss cyg @xcite , but this was only seen during outburst .
the parameters in table [ tab : spike ] provide several constraints upon its origin .
the amplitude of any radial velocity variations is low .
the spike is seen best in hei 4713 , hei 5015 and heii 4686 and these lines indicate a narrow fwhm of @xmath49 .
we take these lines as representative since the ambiguity about whether flux is from the spike or from the disc is likely to cause much larger systematic errors for the other lines . taken together ,
the small width and the radial velocity amplitude allow only a few possible sites of origin which are ( a ) the accreting white dwarf , ( b ) a nebula or ( c ) a wind or jet . any other part of the binary moves too fast ( e.g. the donor star ) or would produce too broad a profile ( e.g. the disc ) .
all the above sites have their own problems however .
a wind origin faces difficulties since the winds observed in normal cataclysmic variables are too highly ionised to produce hei emission and have large velocities characteristic of the escape velocity of the white dwarf @xcite .
it is hard to see how expansion velocities of order @xmath50 arise in such circumstances .
moreover a wind origin has no good explanation for the variations in the systemic velocities , most obvious in hei 5015 . the main point in favour of
the accreting star is our detection of radial velocity variation , albeit of low amplitude . as we showed in section
[ sec : spike ] the measured amplitude and phase of the spike are in accord with the ` s'-wave amplitude and phase . however , once more the systemic velocities are hard to understand , although the primary star has the possibilities of gravitational , pressure @xcite and zeeman shifts .
it is encouraging that the most discrepant line , hei 5015 , is the only singlet line in our wavelength range , but there does not seem to be enough known about these lines for us to say anything more .
a nebula origin copes best with the systemic velocities .
hei 5015 is the line most sensitive to optical depth effects . under case a conditions the hei 5015/4922 ratio should be of order @xmath51 whereas under case b it increases to about @xmath9 @xcite .
thus a correlation between the velocity of the emitting material and its density could be the reason behind the hei 5015 anomaly .
the nebula , if it exists , would have to be asymmetric , which is no great obstacle .
there are however other problems with a nebula origin .
first , it is not resolved .
a nebula expanding at @xmath50 would reach a size of 1 arcsecond in only 10 years if gp com lies at the upper limit of @xmath52 deduced by marsh et al . .
we find no evidence for any extended emission in our long slit spectra and estimate that an extension of order 1 arcsecond should have been detected ; stover found that what appeared to be extended emission around gp com is in fact a group of stars .
perhaps the most significant problems are the radial velocity and flux variability that we have detected
. the nebula explanation can only survive these if they are dismissed as systematic artefacts .
given the small amplitude of the radial velocity variations , it is hard to rule out this possibility . however , while we can not be sure that poor slit loss correction has not artificially produced the spike component in the flare spectra , our feeling is that it is too strong for this to be the case .
luckily , it should be possible to improve upon this with better data in the future .
the evolution that leads to a system such as gp com has been discussed in several papers @xcite . at any particular orbital period
there are two possibile configurations . in one case
the donor may be a helium star of relatively high mass and luminosity . in this case
the orbital period decreases with time .
alternatively , at a later stage after the system has passed its minimum period ( of order @xmath53 ) , the donor adopts a semi - degenerate structure of very low mass and luminosity . in this state as the donor loses mass it increases in radius and the orbital period increases as well .
it is unable to reach a truly degenerate structure because its thermal timescale becomes so long .
the latter case is almost certainly the one that applies to gp com since there is no sign of the donor star .
our favoured mass ratio of @xmath54 also supports this scenario .
based upon savonije et al.s work , warner deduces that @xmath55 where the orbital period @xmath56 is measured in hours .
this gives a mass of order @xmath57 for the donor star .
given an accretor mass of order @xmath58 , this is acceptably close to the mass ratio we deduce , and confirms the status of gp com . using the relation given by warner for a fully - degenerate secondary ( his equation 9.43 ) leads to @xmath59 , still consistent with our mass ratio if the accretor is more like @xmath60 .
for @xmath61 , @xmath62 and @xmath63 , we calculate @xmath64 , a reasonable value
. uncertainties in @xmath26 and @xmath65 are too large to deduce any useful value of the inclination however , and we just note again that the non - sinusoidal ` s'-wave behaviour suggests that it might be high , although there is no sign of an eclipse .
the flaring we have found at optical wavelengths , and seen more dramatically still at uv wavelengths , is another unusual feature of gp com .
it strongly suggests that the inner disc is variable in a way not seen in the systems that look most like gp com in terms of their optical spectra , the quiescent dwarf novae ( leaving aside the absence of hydrogen ! ) .
gp com has never been seen to show outbursts , and there has been discussion as to whether this is because it has a very long inter - outburst interval or because it is in a steady - state of very low mass transfer rate @xcite .
the former possibility is quite reasonable given the decades - long intervals between the outbursts of the short - period dwarf novae known as the wz sge stars . on the other hand , it has long been realised that , if the thermal instability model of dwarf nova outbursts is correct , a disc composed largely of helium may behave very differently from one dominated by hydrogen @xcite .
the turning points on the thermal equilibrium curve plotted as @xmath66 versus @xmath67 ( the `` s - curve '' ) are at a higher temperature and surface density for helium compared to hydrogen @xcite . as a result , it is more likely that the entire disc can be on the lower branch of the s - curve , and thus the system will not undergo outbursts .
we suggest that the high - level of x - ray to optical flux in gp com @xcite and the flaring behaviour support the latter possibility .
that is , matter accretes at all radii in the disc rather than accumulating in the outer disc as it is believed to do in quiescent dwarf novae . in this state
the disc can be optically thin @xcite , which explains the strong emission lines displayed by gp com , in contrast to the rest of the am cvn group .
of course , some other instability must be occurring in the inner disc to explain the flaring ; its cause remains to be determined .
we have analysed time - resolved spectrophotometry of the double - degenerate binary gp com .
we confirm the presence of the ` s'-wave feature found by nrs and refine its period to @xmath23 .
we have detected a small radial velocity variation with a semi - amplitude of @xmath4 in the sharp component at the centre of the emission lines , which may indicate that it comes from the accreting primary star .
the amplitude and phase are consistent with such an origin together with the measured parameters of the ` s'-wave if the mass ratio @xmath5 is of order @xmath6 .
this is roughly as expected if gp com is now increasing in period and has a near - degenerate donor star .
however , the systemic velocity of the narrow component , which varies from line to line , remains to be explained .
while the orbital variations that we find are consistent with earlier data , we have also discovered erratic flaring which we believe to be analogous to similar but more obvious behaviour seen at uv wavelengths .
we suggest that this supports models in which gp com s disc is on the lower branch of the thermal instability curve and does not undergo the global outbursts followed by intervals of mass accumulation as occur in dwarf novae .
trm was supported by a pparc advanced fellowship during the course of part of this work .
the data reduction and analysis were carried out on the southampton node of the uk starlink computer network . | we present and analyse time - resolved @xmath0-band spectra of the double - degenerate binary star , gp com .
the spectra confirm the presence and period ( @xmath1 ) of the ` s'-wave feature found by nather , robinson and stover .
gp com is erratically variable at x - ray and uv wavelengths .
we have found the equivalent variability in our data , which , as also seen in uv data , is mostly confined to the emission lines .
the heii 4686 changes by the largest amount , consistent with x - ray driven photo - ionisation .
the flaring part of the line profiles is broader than the average , as expected if they are dominated by the inner disc .
the heii 4686 profile is especially remarkable in that its blue - shifted peak is @xmath2 from line centre compared to @xmath3 for the hei lines ( the red - shifted peak is blended with hei 4713 ) .
we deduce that heii 4686 emission is confined to the inner 1/4 of the disc .
we suggest that the activity of the inner disc indicates that accretion is significant ( and unstable ) there , in contrast to quiescent dwarf novae , in support of models in which gp com is in a ( quasi-)steady - state of low mass transfer rate .
gp com shows triple - peaked lines profiles which consist of the usual double - peaked profiles from a disc plus a narrow component at line centre .
the latter has previously been ascribed to emission from a nebula , although none could be found in direct images .
however , we find evidence for both radial velocity and flux variability in this component , inconsistent with a nebula origin . the radial velocity amplitude of @xmath4 and its phase relative to the ` s'-wave are consistent with an origin on the accreting white dwarf , if the mass ratio , @xmath5 , is of order @xmath6 , as expected on evolutionary grounds .
however this explanation is still not satisfactory as the systemic velocity of the narrow component shows significant variation from line to line , and we have no explanation for this . accretion , accretion discs novae , cataclysmic variables stars : individual : gp com |
a calorimeter was constructed as part of the instrumentation of the jefferson lab ( jlab ) hall a experiment e99 - 114 , `` exclusive compton scattering on the proton '' @xcite , the schematic layout for which is shown in fig . [ fig : layout ] .
the study of elastic photon scattering provides important information about nucleon structure , which is complementary to that obtained from elastic electron scattering @xcite .
experimental data on the real compton scattering ( rcs ) process at large photon energies and large scattering angles are rather scarce , due mainly to the absence of high luminosity facilities with suitable high - resolution photon detectors .
such data are however crucial , as the basic mechanism of the rcs reaction is the subject of active debate @xcite . the only data available before the jlab e99 - 114 experiment were obtained at cornell about 30 years ago @xcite .
[ tbh ] the construction of the cebaf ( continuous electron beam accelerator facility ) accelerator has led to an extension of many experiments with electron and photon beams in the gev energy range and much improved precision .
this is the result of a number of fundamental improvements to the electron beam , including a 100% duty cycle , low emittance and high polarization , in addition to new dedicated target and detector systems .
the cebaf duty factor provides an improvement of a factor of 15 compared to the best duty factor of a beam extracted from a synchrotron , at a similar instantaneous rate in the detectors . in 1994
work began on the development of a technique for an rcs experiment at jlab , leading in 1997 to the instigation of a large - scale prototyping effort .
the results of the subsequent test runs in 1998 and 1999 @xcite provided sufficient information for the final design of the apparatus presented in the present article .
the fully realized physics experiment took place in 2002 ( see fig .
[ fig : layout ] ) at a photon - nucleon luminosity which was a factor of 1300 higher than in the previous cornell experiment .
the experimental technique involves utilizing a mixed electron - photon beam which is incident on a liquid hydrogen target and passes to a beam dump .
the scattered photons are detected in the calorimeter , while the recoiling protons are detected in a high resolution magnetic spectrometer ( hrs - l ) . a magnet between the hydrogen target and the calorimeter deflects the scattered electrons , which then allows for clean separation between compton scattering and elastic e - p scattering events .
the data acquisition electronics ( daq ) is shielded by a 4 inch thick concrete wall from the beam dump and the target .
figure [ fig : picture ] shows a photograph of the experimental set - up with the calorimeter in the center .
the experiment relied on a proton - photon time coincidence and an accurate measurement of the proton - photon kinematic correlation for event selection .
the improvement in the event rate over the previous measurement was achieved through the use of a mixed electron - photon beam , which in turn required a veto detector in front of the calorimeter or the magnetic deflection of the scattered electron @xcite . in order to ensure redundancy and cross - checking ,
both a veto and deflection magnet were designed and built .
the fact that a clean photon beam was not required meant that the photon radiator could be situated very close to the hydrogen target , leading to a much reduced background near the beam line and a dramatic reduction of the photon beam size .
this small beam size in combination with the large dispersion in the hrs - l proton detector system @xcite resulted in very good momentum and angle resolution for the recoiling proton without the need for a tracking detector near the target , where the background rate is high .
good energy and coordinate resolutions were key features of the photon detector design goals , both of which were significantly improved in the jlab experiment as compared to the cornell one . an energy resolution of at least 10% is required to separate cleanly rcs events from electron bremsstrahlung and neutral pion events . in order to separate further the background from neural pion photo - production , which is the dominant component of the high - energy background in this measurement , a high angular resolution between proton and photon detectors is crucial .
this was achieved on the photon side by constructing a highly segmented calorimeter of 704 channels .
the rcs experiment was the first instance of a calorimeter being operated at an effective electron - nucleon luminosity of @xmath1 @xmath2/s @xcite ( a 40 @xmath3a electron beam on a 6% cu radiator upstream of a 15 cm long liquid hydrogen target ) .
it was observed in the test runs that the counting rate in the calorimeter fell rapidly as the threshold level was increased , which presented an opportunity to maintain a relatively low trigger rate even at high luminosity .
however , on - line use of the calorimeter signal required a set of summing electronics and careful equalizing and monitoring of the individual channel outputs during the experiment .
as the rcs experiment represented the first use of such a calorimeter at very high luminosity , a detailed study of the calorimeter performance throughout the course of the experiment has been conducted .
this includes a study of the relationship between luminosity , trigger rate , energy resolution and adc pedestal widths . an observed fall - off in energy resolution as the experiment progressed allowed for characterization of radiation damage sustained by the lead - glass blocks .
it was possible to mitigate this radiation damage after the experiment by annealing , with both uv curing and heating proving effective .
we begin by discussing the various components which make up the calorimeter and the methods used in their construction .
this is followed by a description of veto hodoscopes which were used for particle identification purposes .
an overview of the high - voltage and data acquisition systems is then presented , followed , finally , by a discussion on the performance of the calorimeter in the unique high - luminosity environment during the rcs experiment .
the concepts and technology associated with a fine - granularity lead - glass erenkov electromagnetic calorimeter ( gams ) were developed by yu .
prokoshkin and collaborators at the institute of high energy physics ( ihep ) in serpukhov , russia @xcite .
the gams type concept has since been employed for detection of high - energy electrons and photons in several experiments at jlab , ihep , cern , fnal and desy ( see for example @xcite ) .
many of the design features of the calorimeter presented in this article are similar to those of serpukhov . a schematic showing
the overall design of the rcs calorimeter can be seen in fig .
[ fig : side - view ] .
the main components are : * the lead - glass blocks ; * a light - tight box containing the photomultiplier tubes ( pmts ) ; * a gain - monitoring system ; * a doubly - segmented veto hodoscopes ; * the front - end electronics ; * an elevated platform ; * a lifting frame .
the calorimeter frame hosts a matrix of 22@xmath432 lead - glass blocks together with their associated pmts and high voltage ( hv ) dividers . immediately in front of the lead - glass blocks is a sheet of ultraviolet - transmitting ( uvt ) lucite , which is used to distribute calibration light pulses for gain - monitoring purposes uniformly among all 704 blocks .
the light - tight box provides protection of the pmts from ambient light and contains an air - cooling system as well as the hv and signal cable systems .
two veto hodoscopes , operating as erenkov counters with uvt lucite as a radiator , are located in front of the calorimeter .
the front - end electronics located a few feet behind the detector were assembled in three relay racks .
they are comprised of 38 analog summers , trigger logic and patch panels .
the elevated platform was needed to bring the calorimeter to the level of the beam line , while the lifting frame was used to re - position the calorimeter in the experimental hall by means of an overhead crane .
this procedure , which took on average around two hours , was performed more than 25 times during the course of the experiment .
[ ht ] the main frame of the calorimeter is made of 10 inch wide steel c - channels .
a thick flat aluminum plate was bolted to the bottom of the frame , with a second plate installed vertically and aligned to 90@xmath5 with respect to the first one by means of alignment screws ( see fig . [
fig : front ] ) .
another set of screws , mounted inside and at the top of the main frame on the opposite side of the vertical alignment plate , was used to compress all gaps between the lead - glass modules and to fix their positions .
the load was applied to the lead - glass blocks through 1 inch @xmath4 1 inch @xmath6 0.5 inch plastic plates and a 0.125 inch rubber pad . in order to further assist block alignment , 1 inch wide stainless steel strips of 0.004 inch thickness running from top to bottom of the frame
were inserted between every two columns of the lead - glass modules .
all pmts and hv dividers are located inside a light - tight box , as shown in fig .
[ fig : cooling ] . as the current on each hv divider is 1 ma , simultaneous operation of all pmts would , without cooling , lead to a temperature rise inside the box of around 50 - 70@xmath5c .
an air - cooling system was developed to prevent the pmts from overheating , and to aid the stable operation of the calorimeter .
the air supply was provided by two parallel oil - less regenerative blowers of r4110 - 2 type , which are capable of supplying air at a maximum pressure of 52 inches water and a maximum flow of 92 cfm .
the air is directed toward the hv divider via vertical collector tubes and numerous outlets .
when the value on any one of the temperature sensors installed in several positions inside the box exceeds a preset limit , the hv on the pmts is turned off by an interlock system .
the air line is equipped with a flow switch of type fst-321-spdt which was included in the interlock system .
the average temperature inside the box during the entire experimental run did not exceed the preset limit of 55@xmath7c .
a simple and reliable cabling system is one of the key features of multichannel detectors , with easy access to the pmts and hv dividers for installation and repair being one of the key features .
the cabling system includes : * 1 foot long hv and signal pig - tails soldered to the hv divider ; * patch panels for lemo and hv connectors ; * 10 feet long cables from those patch panels to the front - end electronics and the hv distribution boxes ; * the hv distribution boxes themselves ; * bnc - bnc patch panels for the outputs of the front - end modules ; * bnc - bnc patch panels on the daq side for the analog signals ; * bnc - lemo patch panels on the daq side for the veto - counter lines .
figure [ fig : cabling ] shows the cabling arrangement inside the pmt box .
the patch panels , which are custom - built and mounted on the air supply tubes , have the ability to swing to the side in order to allow access to the pmts and the hv dividers .
the box has two moving doors , the opening of which leads to activation of an interlock system connected to the hv supply . in order to reduce the diameter of the cable bundle from the pmt box , rg-174 cable ( diameter 0.1 inch )
was used for the pmt signals , and a twisted pair for the hv connection ( two individually insulated inner 26 awg conductors with an overall flame - retardant pvc jacket , part number 001 - 21803 from the general wire product company ) .
the box patch panels used for the hv lines each convert 24 of the above twisted pairs ( single hv line ) to the multi - wire hv cables ( the part 001 - 21798 made by general wire product ) , which run to the hv power supply units located in the shielded area near daq racks .
the basic components of the segmented calorimeter are the tf-1 lead - glass blocks and the feu 84 - 3 pmts . in the 1980s the yerevan physics institute ( yerphi ) purchased a consignment of tf-1 lead - glass blocks of 4 cm @xmath84 cm@xmath640 cm and feu 84 - 3 pmts of 34 mm diameter ( with an active photo - cathode diameter of 25 mm ) for the construction of a calorimeter to be used in several experiments at the yerphi synchrotron . in january of
1998 the rcs experiment at jlab was approved and soon after these calorimeter components were shipped from yerevan to jlab .
this represented the yerphi contribution to the experiment , as the properties of the tf-1 lead - glass met the requirements of the experiment in terms of photon / electron detection with reasonable energy and position resolution and radiation hardness .
the properties of tf-1 lead - glass @xcite are given in table [ tab : tf1 ] .
.important properties of tf-1 lead - glass . [
cols="<,^",options="header " , ] [ tab : resolution ] at high luminosity the energy resolution degrades due to fluctuations of the base line ( pedestal width ) and the inclusion of more accidental hits during the adc gate period .
however , for the 9-block cluster size used in the data analysis the contribution of the base line fluctuations to the energy resolution is just 1 - 2% .
the measured widths of adc pedestals confirmed the results of monte carlo simulations and test runs that the radiation background is three times higher with the 6% cu radiator upstream of the target than without it .
the resolution obtained from e - p calibration runs was corrected for the drift of the gains so it could be attributed directly to the effect of lead glass radiation damage .
it degraded over the course of the experiment from 5.5% ( for a 1 gev photon energy ) at the start to larger than 10% by the end .
it was estimated that this corresponds to a final accumulated radiation dose of about 3 - 10 krad , which is in agreement with the known level of radiation hardness of the tf-1 lead glass @xcite .
this observed radiation dose corresponds to a 500 hour experiment with a 15 cm lh2 target and 50 @xmath3a beam .
the front face of the calorimeter during the experiment was protected by plastic material with an effective thickness of 10 g/@xmath2 . for the majority of the time
the calorimeter was located at a distance of 5 - 8 m and an angle of 40 - 50@xmath5 with respect to the electron beam direction .
the transparency of 20 lead - glass blocks was measured after the experiment , the results of which are shown in fig .
[ fig : rad_glass ] .
this plot shows the relative transmission through 4 cm of glass in the direction transverse to the block length at different locations .
the values were normalized to the transmission through similar lead - glass blocks which were not used in the experiment .
the transmission measurement was done with a blue led ( @xmath10 of 430 nm ) and a hamamatsu photo - diode ( 1226 - 44 ) .
a uv technique was developed and used in order to cure radiation damage .
the uv light was produced by a 10 kw total power 55-inch long lamp , which was installed vertically at a distance of 45 inches from the calorimeter face and a quartz plate ( c55quartz ) was used as an infrared filter .
the intensity of the uv light at the face of the lead - glass blocks was found to be 75 mw/@xmath2 by using a uvx digital radiometer . in situ
uv irradiation without disassembly of the lead - glass stack was performed over an 18 hour period .
all pmts were removed before irradiation to ensure the safety of the photo - cathode .
the resultant improvement in transparency can be seen in fig .
[ fig : rad_glass ] . an alternative but equally effective method to restore the lead - glass transparency , which involved heating of the lead - glass blocks to 250@xmath5c for several hours , was also tested .
the net effect of heating on the transparency of the lead - glass was similar to the uv curing results . in summary , operation of the calorimeter at high luminosity , particularly when the electron beam was incident on the bremsstrahlung radiator , led to a degradation in energy resolution due to fluctuations in the base - line and a higher accidental rate within the adc gate period .
for typical clusters this effect was found to be around a percent or two . by far
the largest contributor to the observed degradation in resolution was radiation damage sustained by the lead - glass blocks , which led to the resolution being a factor of two larger at the end of the experiment .
the resulting estimates of the total accumulated dose were consistent with expectations for this type of lead - glass .
finally , it was found that both uv curing and heating of the lead - glass were successful in annealing this damage .
the design of a segmented electromagnetic calorimeter which was used in the jlab rcs experiment has been described . the performance of the calorimeter in an unprecedented high luminosity , high background environment has been discussed .
good energy and position resolution enabled a successful measurement of the rcs process over a wide range of kinematics .
we acknowledge the rcs collaborators who helped to operate the detector and the jlab technical staff for providing outstanding support , and specially d. hayes , t. hartlove , t. hunyady , and s. mayilyan for help in the construction of the lead - glass modules .
we appreciate s. corneliussen s careful reading of the manuscript and his valuable suggestions .
this work was supported in part by the national science foundation in grants for the university of illinois university and by doe contract de - ac05 - 84er40150 under which the southeastern universities research association ( sura ) operates the thomas jefferson national accelerator facility for the united states department of energy . 00 c. hyde - wright , a. nathan , and b. wojtsekhowski , spokespersons , jlab experiment e99 - 114 .
charles hyde - wright and kees de jager _ ann.rev.nucl.part.sci . _ * 54 * , 217 ( 2004 ) .
radyushkin , _ phys .
* d 58 * , 114008 ( 1998 ) .
huang , p. kroll , t. morii , _ eur .
j. _ * c 23 * , 301 ( 2002 ) ; _ erratum ibid .
_ , * c 31 * , 279 ( 2003 ) .
r. thompson , a. pang , ch .- r .
ji , _ phys .
* d 73 * , 054023 ( 2006 ) .
shupe _ et al .
_ , _ phys .
_ * d 19 * , 1929 ( 1979 ) . e. chudakov _ et al .
_ , study of hall a photon spectrometer .
hall a internal report , 1998 .
j. alcorn _ et al .
_ , _ nucl .
_ * a 522 * , ( 2004 ) 294 .
hamilton _ et al .
_ , _ phys . rev
. lett . _ * 94 * , 242001 ( 2005 ) .
a. danagoulian _ et al .
_ , _ phys .
lett . _ * 98 * , 152001 ( 2007 ) . yu.d .
prokoshkin _ et al .
_ , _ nucl .
meth . _ * a 248 * , 86102 ( 1986 ) .
balatz _ et al .
_ , _ nucl . instr .
meth . _ * a 545 * , 114 ( 2005 ) .
astvatsaturov _ et al .
_ , _ nucl .
* 107 * , 105 ( 1973 ) .
crittenden _ et al .
meth . _ * a 387 * , 377 ( 1997 ) .
v. popov _ et al .
_ , proceedings of ieee 2001 nuclear science symposium ( nss ) and medical imaging conference ( mic ) .
valentine ieee ( 2001 ) p. 634 - 637 . v. popov , _ nucl
meth . _ * a 505 * , 316 ( 2003 ) .
et al . _ , coda : a scalable , distributed data acquisition system , in : proceedings of the real time 1993 conference , p. 296 ; e. jastrzembski _ et al .
_ , the jefferson lab trigger supervisor system , 11th ieee npss real time 1999 conference , jlab - tn-99 - 13 , 1999 .
inyakin _ et al .
_ , _ nucl .
* 215 * , 103 ( 1983 ) . | a lead - glass hodoscope calorimeter that was constructed for use in the jefferson lab real compton scattering experiment is described .
the detector provides a measurement of the coordinates and the energy of scattered photons in the gev energy range with resolutions of 5 mm and 6%/@xmath0}$ ] .
features of both the detector design and its performance in the high luminosity environment during the experiment are presented .
calorimeters , erenkov detectors 29.40vj , 29.40.ka |
the o4f star @xmath1 puppis has for decades been at the cutting edge of research regarding early type stars because of its optical and uv brightness . in addition , because of the low interstellar attenuation , has been the prime x - ray target to study soft x - ray emission from o stars .
researchers have looked to this source to settle outstanding controversies about the physical location , quantity , and nature of the hot x - ray emitting plasma on ob stars . in this paper
we present the highest resolution x - ray spectrum of ever measured and explore the physical mechanism of hot star x - ray production .
prior to the discovery of x - ray emission from o stars ( harnden et al .
1979 ; seward et al .
1979 ) , cassinelli & olson ( 1979 ) postulated thin base coronal zones as a source for x - rays to explain the observed uv superionization ( lamers & morton 1976 ) by way of the auger effect .
the coronal models ( cassinelli & olson 1979 ; waldron 1984 ) predicted an x - ray absorption at the oxygen k - shell edge larger than that observed with sss ( cassinelli & swank 1983 ) and ( corcoran 1993 ) . these observations , as well as detailed modeling of the superionization profiles ( macfarlane 1993 ) and lack of detection of the iron `` green line '' ( baade & lucy 1987 ) in caused the coronal model to fall out of favor .
increasingly , a consensus formed around a wind - shock picture in which a series of shocks , perhaps related to the line - force instability ( lucy 1982 ; owocki , castor , & rybicki 1988 ) , causes hot , x - ray emitting gas to be distributed _ throughout _ the dense stellar wind of and other ob stars .
wind shock models developed by lucy & white ( 1980 ) , feldmeier ( 1997 ) and others , consistently failed to predict the high levels of x - ray emission observed in the brightest o stars like , leading to the suggestion that perturbations somehow form and propagate up from the photosphere into the wind and drive stronger shocks ( feldmeier 1995 ; feldmeier , puls , & pauldrach 1997 ) .
broad - band x - ray observations of ( corcoran et al .
1993 ; hillier et al .
1993 ) indicate that some wind attenuation is affecting the soft x - ray flux .
however with the advent of and we can apply , for the first time , diagnostic emission line ratios and measure line profiles to determine the locations and doppler velocities of x - ray sources in the stellar wind of .
we obtained a 67 ks high energy transmission grating spectrometer ( hetgs ) observation of from 2000 mar 28 , 13@xmath2 31@xmath3 ut to mar 29 , 09@xmath2 12@xmath3 .
the standard pipeline tools were used to reprocess the data with the most recent calibration files available .
line emission is clearly evident in both the high - energy grating ( heg ) and the medium - energy grating ( meg ) .
the combined @xmath4 order , background - subtracted , hetgs spectra are shown in figure 1 .
triads of he - like ions known as the ( forbidden , intercombination , and resonance ) lines ( , , , , and ) are seen , in addition to isolated emission lines and numerous l - shell lines of iron , especially .
traditionally , he - like ion @xmath5 line ratios have been used to derive electron densities of x - ray line emitting regions since the populations of the 2@xmath6p levels are controlled by collisional excitations from the 2@xmath6s levels ( gabriel & jordan 1969 ) . however , when a strong external uv source is present
, the excitation 2@xmath6s @xmath7 2@xmath6p is predominantly radiative and this means that the @xmath5 ratio is no longer a density diagnostic , but rather it can be used to determine the strength of the uv radiation field ( blumenthal 1972 ) .
the uv wavelengths associated with the radiative excitation are in the range from @xmath8 to 1650 , where o star uv fluxes are very strong . for s photospheric flux
we use a kurucz ( 1993 ) model , assuming = 42000k . since the 2@xmath6s @xmath7 2@xmath6p wavelengths are not on the wien side of the spectrum , the model atmosphere fluxes are relatively well determined .
waldron & cassinelli ( 2001 ) demonstrated how the observed @xmath5 line ratios can be used to derive the line formation radii from the geometric dilution of the photospheric radiation field in ( 09.5ia ) .
kahn ( 2001 ) used a similar approach to derive line formation radii from their _ rgs _ spectrum of .
our hetg spectrum has better resolution at essentially all he - like ion wavelengths and greater sensitivity to the short - wavelength lines ( e.g. , si and s ) than the spectrum . this high resolution is very important for o stars since their x - ray lines show strong broadening .
the observed he - like @xmath5 line ratios are listed in table 1 . * table 1 : @xmath0 ratio measurements * llllll & & & & & + & @xmath9&@xmath10&@xmath11 & @xmath12 0.1 & @xmath12 0.1 + _ r / i _ & @xmath13&@xmath14 & @xmath15&@xmath16&@xmath17 0.6 + @xmath18 & 0.81 & 3.06&3.26&5.77&3.56 + @xmath19total flux of the three lines in @xmath20 erg @xmath21 s@xmath22 figure 2a shows the radii derived from our measurements of the he - like @xmath5 ratios .
the radii of line formation are seen to vary from ion to ion , with lines from high nuclear charge ions ( si , s ) forming at small radii , and lower nuclear charge ion lines ( o , ne , mg ) forming at larger radii .
one might expect the @xmath0 line complexes for each element to be formed over a range of radii and not just in a thin shell .
however , since the line emission is proportional to @xmath23 , the observed emission in each line complex should be dominated by the densest regions ( the smallest radius ) from which the line radiation can escape . for the @xmath0 formation locations
correlates with the associated radial optical depth unity radius , @xmath24 .
the optical depth is measured through the wind , using the continuum opacity at the wavelength of each triad of @xmath0 lines . any line radiation that originates much below @xmath24
could not be seen by an external observer .
the wavelength dependence of @xmath24 for is shown in figure 2b using a @xmath25-law with @xmath26 0.75 as the velocity law , the wind absorption cross sections in waldron ( 1998 ) , and the stellar parameters of lamers & leitherer ( 1993 ) ( @xmath27 = 16 , = @xmath28 , and = 2200 ) .
an inspection of figure 2 leads to the following conclusions : first , the he - like line formation radii are all consistent with their associated @xmath24 values , as expected from the above argument .
second , the @xmath5 ratio of indicates that hot plasma exists below 1.2 where the nominal wind velocity is @xmath12 600 .
most radiative instability simulations do not show strong shocks this close to the star ( feldmeier 1995 , owocki , castor , & rybicki 1988 ) .
however , using a velocity jump even half as large as the local wind velocity ( @xmath29 ) to calculate the shock temperature from @xmath30 , we obtain a post - shock temperature of 1 mk ; too low to produce .
thus , anomalously strong shock jumps are required deep in the wind .
a similar conclusion was reached by waldron & cassinelli ( 2001 ) , but for the highest ion stage observed was instead of the seen here .
the most notable feature observed in our hetg spectrum of is the clear presence of blue - shifted x - ray line centroids in all strong lines .
this is in contrast with previous o star observations ( schulz 2000 ; waldron & cassinelli 2001 ) where the lines were broad , symmetric and un - shifted .
although our line widths are comparable to other o - star observations , with hwhm of @xmath31 1000 , our observation is the first detection of blue - shifted and blueward skewed x - ray line profiles in an o star .
figure 3 displays six x - ray lines arranged in order of wavelength .
the results for the line fits are given in table 2 .
there is an interesting progression that holds for these lines : the centroid shifts are generally larger for the broader lines , indicating a connection between radial location , wind absorption , and the doppler broadening of the line emission . in table 4 of their xmm study of , kahn ( 2001 ) list blueshifts of the lyman-@xmath32 lines of and .
the blueshifts agree with ours to within 2 sigma , and some of the discrepancy can be attributed to the absolute wavelength calibration uncertainties of the instruments ( @xmath31 100 km s@xmath22 for the hetg ( chandra x - ray center ( cxc ) 2001 ) , see den herder 2001 for the rgs ) .
the xmm results show the same trend between shift and broadening .
the line is an exception to the trend in table 2 because of its small centroid shift .
it is different morphologically from the other lines ( also noted by kahn ( 2001 ) ) .
it has a roughly flat - topped profile so we use a box fit instead of a gaussian to estimate the values for table 2 .
* table 2 : line profile widths and centroid shifts * [ cols= " < , < , < , < , < , < " , ] there is a blueward skewness observed in many of the lines which is not attributable to the hetgs line response function ( cxc 2001 ) .
the blue sides of these lines tend to be steeper , while the red sides have a shallower slope .
in general , the shape of these lines indicates wind broadening combined with attenuation as would arise from line formation in the marginally thick region near @xmath24 .
macfarlane ( 1991 ) , ignace ( 2001 ) , and owocki & cohen ( 2001 ) demonstrate that emission line profiles can provide important information about the spatial distribution of hot plasma within a stellar wind based both on the velocity - dependence of the intrinsic emission and also on the continuum attenuation across the line .
assuming isotropic emission , the red - shifted emission from the back side of the wind is suppressed by the continuum opacity along the line of sight , while the blue - shifted emission from the near side is less so .
this leads to skewed , almost triangular - shaped lines in an optically thick wind . in the hetg spectrum ,
waldron & cassinelli ( 2001 ) found that an optically thin model was needed to provide a reasonable fit because the lines were symmetric and un - shifted .
this fit required that the assumed mass loss rate for be greatly reduced .
for we have the odd situation that the blueward shift and asymmetric line profiles agree with what has long been expected for hot stars , but this is the first time that it has been so clearly observed in a spectrum for any star .
we find that a fit can be made to the line profiles without making a significant reduction in the assumed mass loss rate in contrast with the results .
the line profile is different from those of the other ions as can be seen in figure 3 and table 2 .
the doppler half - width of this line is much greater than any of the other x - ray lines .
this velocity of 1570 can be compared with the terminal speed of 2200 determined from uv wind lines .
a flat - topped line profile such as this is expected from a fast moving shell source suffering little or no wind attenuation ( macfarlane 1991 ) .
therefore , in contrast with the other observed lines , the nvii profile is probably formed well above its continuum optical depth unity radius .
the hetg emission line spectrum of , with its unprecedented resolution ( in excess of @xmath33 for some lines ) and sensitivity over a wide range of wavelengths , allows us to draw quantitative conclusions about the nature of the x - ray source on this prototypical hot star .
most importantly , we find that the wind plays an important , observable role in determining the x - ray line profiles of , in contrast to results for other stars
. we can begin to constrain the physical processes which lead to the production of the hot , x - ray emitting plasma on this star .
taking all the he - like ions from oxygen through sulfur into account , the correlation between the formation radius of the @xmath0 lines and the radius at optical depth unity , @xmath24 , we find evidence for a spatially distributed source of x - rays throughout the expanding wind .
there are two lines of particular interest in the development of future shock models : , and .
the shock jump required to produce appears inconsistent with the expected local wind conditions , requiring a post - shock flow velocity of approximately zero .
although this is not impossible for a shock model to reproduce , it places strong constraints on future modeling efforts .
the line line may have a shape distinct from the other lines in our spectrum because it is formed in the outer regions of the wind where there is no further acceleration of the x - ray source region and little overlying wind attenuation .
this hetg data set provides the most detailed and complete picture to date of an extended distribution of hot plasma embedded within a strong , optically thick stellar wind .
baade , d. & lucy , l. b. 1987 , , 178 , 213 blumenthal , g. r. , drake , g. w. f. , & tucker , w. h. 1972 , , 172 , 205 cassinelli , j. p. & olson , g. l. 1979 , , 229 , 304 cassinelli , j. p. & swank , j. h. 1983 , , 271 , 681 chandra x - ray center ( cxc ) 2000 , chandra proposers observatory guide , version 3.0 , td 403.00.003 ( cambridge , ma : cxc ) , 180 , 172 corcoran , m. f. et al .
1993 , , 412 , 792 den herder , j. w. et al .
2001 , , 365 , l7 feldmeier , a. 1995 , , 299 , 523 feldmeier , a. , kudritzki , r. p. , palsa , r. , pauldrach , a. w. a. , & puls , j. 1997 , , 320 , 899 feldmeier , a. , puls , j. , & pauldrach , a. w. a. 1997 , , 322 , 878 gabriel , a. h. & jordan , c. 1969 , , 145 , 241 harnden , f. r. et al .
1979 , , 234 , l51 hillier , d. j. , kudritzki , r. p. , pauldrach , a. w. , baade , d. , cassinelli , j. p. , puls , j. , & schmitt , j. h. m. m. 1993 , , 276 , 117 ignace , r. 2001 , apjl , 549,119 kahn , s. m. , leutenegger , m. a. , cottam , j. , rauw , g. , vreux , j .-
m . , den boggende , a. j. f. , mewe , r. , & gdel , m. 2001 , , 365 , l312 kurucz , r. l. 1993 , in peculiar versus normal phenomena in a - type and related stars .
, iau colloq .
m. m.dworetsky , f. castelli , & r. faraggiana ( san francisco : asp ) , 87 lamers , h. j. g. l. m. & leitherer , c. 1993 , , 412 , 771 lamers , h. j. g. l. m. & morton , d. c. 1976 , , 32 , 715 lucy , l. b. 1982 , , 255 , 286 lucy , l. b. & white , r. l. 1980 , , 241 , 300 macfarlane , j. j. , cassinelli , j. p. , welsh , b. y. , vedder , p. w. , vallerga , j. v. , & waldron , w. l. 1991 , , 380 , 564 macfarlane , j. j. , waldron , w. l. , corcoran , m. f. , wolff , m. j. , wang , p. , & cassinelli , j. p. 1993
, , 419 , 813 morrison , r. & mccammon , d. 1983 , , 270 , 119 owocki , s. p. , castor , j. i. , & rybicki , g. b. 1988 , , 335 , 914 owocki , s. p. , cohen , d. h. 2001 apj , submitted , preprint ( astro - ph/0101294 ) schulz , n. s. , canizares , c. r. , huenemoerder , d. , & lee , j. c. 2000 , , 545 , l135 seward , f. d. , forman , w. r. , giacconi , r. , griffiths , r. e. , harnden , f. r. , jones , c. , & pye , j. p. 1979
, , 234 , l55 waldron , w. l. 1984 , , 282 , 256 waldron , w. l. , corcoran , m. f. , drake , s. a. , & smale , a. p. 1998
, , 118 , 217 waldron , w.l.,cassinelli , j. p. 2001 apj , 548 , l45 | we report on a 67 ks hetg observation of the optically brightest early o - star , ( o4 f ) . many resolved x - ray lines
are seen in the spectra over a wavelength range of 5 to 25 . has sufficient spectral resolution to study the velocity structure of isolated x - ray line profiles , and to distinguish the individual forbidden , intercombination , and resonance ( @xmath0 ) emission lines in several he - like ions even where the individual components are strongly doppler broadened .
in contrast with x - ray line profiles in other hot stars , shows blue - shifted and skewed line profiles , providing the clearest and most direct evidence that the x - ray sources are embedded in the stellar wind .
the broader the line , the greater the blueward centroid shift tends to be .
the line at 24.78 is a special case , showing a flat - topped profile .
this indicates it is formed in regions beyond most of the wind attenuation .
the sensitivity of the he - like ion @xmath0 lines to a strong uv radiation field is used to derive the radial distances at which lines of , , , , and originate .
the formation radii correspond well with continuum optical depth unity at the wavelength of each line complex , indicating that the x - ray line emission is distributed throughout the stellar wind .
however , the emission lines form deeper in the wind than expected from standard wind shock models . |
we have now convincing evidence that the three active neutrinos of the standard model have different masses and they mix with each other .
the results of the superk@xcite , sno@xcite and kamland @xcite experiments have now narrow the @xmath0 mass pattern to one of three possibilities inverted mass hierarchy ( imh ) normal mass hierarchy ( nmh ) almost degenerate masses the first two cases are depicted in fig . 1 and we concentrate on them
. the notations in the figure are standard . however , the overall mass scale is not known .
we only have upper bounds of @xmath1ev from wmap @xcite and @xmath2ev from tritium end point experiments @xcite .
if neutrinos have majorana masses then a better limit of @xmath3 ev comes from neutrinoless double beta decay @xmath4 experiments @xcite .
this latter result is model dependent and is subject to uncertainties in nuclear matrix element calculations as well as the implicit assumption that neutrino mass is the dominant effect leading to the decay .
there are many excellent reviews on these are other questions we list the most recent ones in @xcite . as with other fermions when neutrinos are massive their mass eigenstates and weak eigenstates do not coincide . in general
they are related via a mixing matrix @xmath5 @xcite .
analysis of the data yield u_pmns&= & ( ccc u_e1&u_e2&u_e3 + u_1&u_2&u_3 + u_1&u_2&u_3 ) + & = & ( ccc0.79 - 0.86&0.50 - 0.61&.0 - 0.16 + 0.24 - 0.52&0.44 - 0.69&0.63 - 0.79 + 0.26 - 0.52&0.47 - 0.71&0.60 - 0.77 ) .
[ eq : mns ] in a gauge theory @xmath5 is a product of two mixing matrices : u_pmns = u_l u _ where @xmath6 diagonalizes the charged lepton mass matrix and @xmath7 does the same for neutrinos .
often one chooses the mass eigenbasis for the charged leptons and @xmath5 is just the neutrino mixing matrix .
furthermore , eq.([eq : mns ] ) is a low energy solution .
it is typical that gauge models are formulated at some high energy scale and reach the above via renormalization group running . whether this has a large effect is highly model dependent . in most of our discussions
we assume that this is not large .
now we return to the mixing matrix .
it is clear that , unlike the down quark mixing , neutrino mixing are bilarge and the recent sno data also rule out bi - maximal .
one can reconstruct the neutrino mass matrix in the weak basis @xmath8 via @xmath9 where @xmath10 is the diagonal mass matrix with eigenvalues @xmath11 .
it is @xmath12 that is of the greatest interest to theorists . in general @xmath13
are complex numbers for majorana neutrinos . customarily one phase is put into @xmath5 and two are left in @xmath10 . with the values of mass differences depicted in fig . 1 and eq . ( [ eq : mns ] ) the leading mass patterns are given by imh _ ~m_0 ( 1 & & + & 1/2&1/2 + & 1/2&1/2 ) _ ~m_0 ( & 1&1 + 1 & & + 1 & & ) nmh _
~m_0 ( & & + & 1/2&-1/2 + & -1/2&1/2 ) where @xmath14 is the unknown overall mass scale and @xmath15 denotes some small number .
the challenge to theorists is to construct viable models that give rise to one of the above structures . at the same time one
gets a more or less natural understanding of why the overall neutrino mass scale is so much smaller than the charged fermions ; i.e. what physics sets the scale of neutrino masses .
the strong evidence of neutrino oscillations and mixing has certainly shattered the concept of separately conserved electron , muon or tau lepton number ; however , the question of whether total lepton number is conserved is still open . the usual way of stating this is in terms of whether neutrinos are dirac or majorana particle . if neutrinos were majorana then they are own anti - particle and lepton number violating reactions can be expected to proceed .
otherwise they are dirac particles .
experimentally this is a very difficult question to answer .
the favorite process is @xmath4 decays .
if one such decay were observed then it will be clear evidence that lepton number is not conserved and it will natural to conclude that neutrinos are majorana or at least a majorana mass term is not forbidden
. on the other hand to prove that neutrinos are dirac particles will be very difficult .
the nature of neutrinos has deep theoretical significance is modern particle physics .
the active neutrino in the sm is a left - handed weyl particle ; i.e. is a helicity eigenstate . due to gauge invariance
it has no bare mass term . to give it a mass one
usually introduces a right - handed fermion @xmath16 which is a sm singlet .
whether the @xmath16 exists and what is its nature are two of the most fundamental question in neutrino physics . certainly if @xmath16 exists how many are there ?
what is its mass ? is it dirac or majorana particle ? in the sm this is put in by hand and so the same is true @xmath17 gut models .
it is customary to use at least one to generate masses for some of the active neutrinos and three if we believe in some family symmetry . in @xmath18 gut models
@xmath16 naturally exists since the fundamental representation has 16 fermions which is just the right number to accommodate each sm family plus a @xmath16 .
here it is natural to have three of them .
there are many theoretical reasons to pursue gut models and understand neutrino masses in this context ( see @xcite for an up to date review ) ; moreover , we wish to study this in a wider setting .
as is well known @xmath16 being a singlet can have a majorana mass term , @xmath19 as well as a dirac mass term coupling given by @xmath20 where @xmath21 is the sm higgs boson and @xmath22 is the yukawa coupling .
after symmetry breaking the neutrino masses for one family is given by ( |_l^c & |n_r^c ) ( m_&m_d + m_d^t & m_r ) ( _
l + n_r ) [ eq : nummatrix ] for three families each entry in the above is a @xmath23 matrix . in the sm @xmath24 and @xmath25 gev . if @xmath16 is a dirac fermion then @xmath26 and we have dirac masses for the active neutrinos .
this is the simplest extension of the sm . in this case
it is very difficult to understand the smallness of the neutrino masses and extreme fine tuning of the yukawa coupling @xmath27 to get it below the experimental bound . in gut models
it is natural to take @xmath28 gev and we have the seesaw mechanism for generating neutrino masses .
recently new perspectives in neutrino physics arise from the so call brane world scenario . in the simplest form
this makes use of the possibility that there are more then 4 dimensions . the number of extra dimensions @xmath29 is between 1 and 7 as hinted by string theory .
they are taken to be spatial and can be relatively large @xcite which enables one to solve the hierarchy problem .
fields with sm quantum numbers are confined on a @xmath30 hyper surface whereas sm singlets such as the graviton can propagate in the full @xmath31 dimensional bulk .
the extra dimensions are compactified in tori of radii @xmath32 where @xmath33 . introducing @xmath16 as a bulk field
was done in @xcite .
as is well known compactification leads to 4d kaluza - klein(kk ) excitations of @xmath16 with masses @xmath34 where we have taken all radii to be equal .
the zero mode can now couple to the active neutrino on the brane and the sm higgs in the usual way .
if @xmath16 is taken to be dirac then a small dirac neutrino mass can be generated given by m_d= v=245 without excessive fine tuning of @xmath22 .
this is due to the volume dilution factor of @xmath35 where @xmath36 is the fundamental scale for the theory we can obtain @xmath37ev .
even with this simple model and only one family there is an important consequence for neutrino oscillation .
the kk excitations form an infinite tower of sterile neutrinos .
one would expect an effective active - sterile oscillation @xcitein the superk data .
however , there is no evidence in the data .
this may indicate that the mixing between the active neutrinos and the kk modes are small and/or @xmath38 can not be taken too large .
other effects in nuclear and astrophysics of this simple model are discussed in @xcite .
we have seen that small dirac neutrino masses are quite natural in extra dimensional model ; however , to obtain the hierarchy indicated by the data is more difficult .
one way is to make use of family symmetry as in 4d models . another solution is purely extra dimensional which can also incorporate the masses of quarks and charged leptons .
this is the split fermion scenario @xcite where the left - handed and right - handed fields are localized at different positions in the extra dimensional space .
for example , the @xmath39 doublet is at @xmath40 and @xmath41 is at @xmath42 where @xmath43 denotes the extra dimension .
this is repeated for other fermions .
the gauge and higgs fields propagate in the bulk .
the chiral fermions are also bulk fields and their zero modes , denoted by @xmath44 , are given gaussian profiles in @xmath43 .
for a fermion located at @xmath45 its wave function is given by _
i(x , z)~1 ^14 ^ 12 _ i(x ) e^- . where @xmath46 is a free parameter that is taken to be universal for all fermions for simplicity .
the product of two fermion fields can be approximately replaced by _
i(x , z)_j(x , z ) ~(- ) ( z-|z_ij)|_i(x)_j(x ) where @xmath47 is their average positions and @xmath48 . after integrating out @xmath43
it is the overlap of the two different chiralities of a fermion that determines its mass .
explicitly it is given by @xmath49 .
the origin of the large mass for fermion is due to the proximity of its left and right - handed components in the extra dimension .
similarly , if the chiral components of a fermion is far apart in @xmath43 it will be light .
hence , one does not need to fine tune yukawa couplings but instead the positions @xmath45 are used to fit the data .
it is found that realistic quark and lepton mass matrices and an acceptable ckm matrix can be be found @xcite . in this scenario
the fermion masses becomes a geographical problem in the extra dimension .
stating it differently we need a mechanism that puts the fermions in their correct positions instead of putting it in by hand . as a result , although it is non - trivial to be able to reproduce the observed masses and mixing , these models suffer from a lack of predictive power .
as can be seen from esq.([eq : nummatrix ] ) that neutrinos will be massless in the sm and can have only a majoring mass via radiative effects .
this is first done in @xcite within the sm gauger group by adding a @xmath50 singlet but @xmath51 charged scalar field .
the model also requires an additional doublet hogs field @xmath52 . due to fermi statistics
@xmath53 can only couple to leptons of two different families .
the feynman diagram generating neutrino mass is given below the resulting mass matrix is symmetric and takes the form _ = m_o ( 0&a&b + a&0&c + b&c&0 ) [ eq : zeemass ] where @xmath14 is an overall mass scale set by the mass of @xmath53 .
in eq.([eq : zeemass ] ) @xmath54 are functions of lepton masses and higgs masses .
the characteristics of the zee mass matrix are the vanishing diagonal elements .
this structure leads to bi - maximal mixing @xcite and does not agree with the latest sno data @xcite . furthermore the model suffers from being arbitrary .
however , the mass matrix is intriguing and can serve as the leading order to the correct mass matrix .
to do so one has to find a natural way of filling in some of the zero s in the zee mass matrix .
this can be achieved in 5d unification models which possess gauge symmetries in the bulk .
the components of the the 5d fields that form representations of the symmetry can have different boundary values on the 4d brane .
this reduces the symmetry that particles resides on the brane sees .
this is known as orbifold symmetry breaking ( osb)mechanism . to appreciate this a little more we consider a 5d theory with the extra dimension compactified in a circle ; i.e @xmath55 $ ] . if we impose a parity of @xmath56 such that a field @xmath57 transform as p : z - z pa(z)=a(z ) . clearly @xmath40 and @xmath58 are fixed points and the background geometry is @xmath59 .
now we impose a second parity by defining @xmath60 such that p^ : z^-z^ the fixed points are at @xmath40 and @xmath61 and the geometry is @xmath62 .
the fields can have different parities under @xmath63 and @xmath64 .
in particular the fourier decomposition of a bulk field into their kk modes that conform to the parities assignments are listed in table i. [ cols="^,^,^,^,^",options="header " , ] table i. kk decomposition of a bulk field @xmath57 with parities @xmath65 the zero modes have @xmath66 parity and are identified as sm fields . as an example consider @xmath17 with the higgs field in the @xmath67 representation .
we can assign @xmath68 parities to the @xmath69 components and different parities to the @xmath70 color higgs components .
the former is just the sm higgs field and couples to sm matter fields which are placed on the @xmath40 brane whereas the colored higgs are kk excitations and will have masses @xmath71 .
they latter will be heavy if the compactification radius @xmath38 is small and of the order of the unification scale .
this provide a new natural solution to the doublet - triplet splitting problem in @xmath17 gut @xcite . for a review of orbifold gut models
see @xcite . with this technique we can now shape new tools for investigating the fermion mass problem . in ref.(@xcite
) and ( @xcite ) this has been applied to study neutrino mass generation without the benefit of @xmath16 fields for @xmath70 and @xmath17 unification models .
the first model is particularly simple and illustrates the clearly physics involved and extends previous work on this model @xcite .
the @xmath72 symmetry is in the bulk and acts on a background geometry of @xmath73 .
it only unifies the @xmath74 symmetry to @xmath75 and has obvious built in lepton number number violation when the leptons are placed in the fundamental @xmath76 i.e. @xmath77 .
the unified value of the weak mixing angle is well known to be given by @xmath78 @xcite and one can expect unification to occur around @xmath79 after renormalization group considerations .
the lepton triplet is placed on the @xmath40 brane and @xmath70 is broken by orbifolding the gauge fields and not by the higgs mechanism which is reserved for sm symmetry breaking .
we require a bulk higgs field in the @xmath80 denoted by @xmath81 and a symmetric antisextet to give realistic charged lepton masses .
we denote fields in by @xmath82 . due to the requirement of @xmath83 invariance
we also introduce a second higgs triplet @xmath84 for the construction of the necessary coupling . in this model the parities @xmath63 and @xmath64 are represented by @xmath85 matrices : p= ( ccc 1&0 & 0 + 0 & 1 & 0 + 0 & 0 & 1 ) , p^= ( ccc 1&0 & 0 + 0 & 1 & 0 + 0 & 0 & -1 ) .
the parities of the components of higgs fields are engineered according to phenomenological needs and are _
3= ( c _ 3 ^ -(++ ) + _ 3 ^ 0(++ ) + h_3^+(+- ) ) , _3= ( c _
3^-(+- ) + _ 3^0(+- ) + h_3^+(++ ) ) , [ p3 ] and [ p6 ] _ 6= ( ccc ^++_\{11}(+- ) & ^+_\{12}(+- ) & ^0_\{13}(++ ) + ^+_\{12}(+- ) & ^0_\{22}(+- ) & ^-_\{23}(++ ) + ^0_\{31}(++ ) & ^-_\{32}(++ ) & ^_\{33}(+- ) ) . with these ingredients
1 loop neutrino masses can be generated via the dominant contribution comes from [ fig : loopsu3]a and leads to a zee - like neutrino mass matrix .
the next two diagrams give the necessary correction to this .
in contrast to 4d models these terms exist as part of the 5d model .
we do not have to put them in by hand .
two parameters central to all extra dimension models are the compactification radius @xmath38 and the cutoff scale @xmath86 at which the field theory becomes strong
. a lower limit on @xmath87tev can be obtained from the non observation of kk modes and @xmath88 . without fine tuning of yukawas other than that required to get the charged lepton masses and using @xmath89 2 tev one can obtain @xmath90 which gives a good fit to the superk and sno data .
this solution is of the imh type and interestingly predicts a detectable neutrinoless double beta decay rate in the next round of experiments .
furthermore , the model has very rich phenomenology at the tev scale which is the compactification scale here .
in particular double charged gauge boson are predicted for this model . which can be searched for in linear colliders .
rare decays such as @xmath91 and similar @xmath92 decays are also sensitive probes of the model @xcite . in the @xmath70 model it is known that quarks can not be accommodated and has to be placed at the other fixed point .
it may be more natural to place quarks and leptons on the same footing as in the @xmath17 model .
to generate neutrino masses with only the fifteen sm fermions per family requires the use of either @xmath93 or @xmath94 bulk higgs field .
the procedure is similar to the case of @xmath70 exploiting osb .
the addition of exotic higgs exacerbates the gauge unification problem in @xmath17 .
this can be solved by additional fermions ( see @xcite ) . after the dust settles unification
is found to take place at @xmath95 gev .
an interesting result emerges for this model .
it is more natural to get the nmh with the @xmath96 higgs and imh solutions can only be found with @xmath97 higgs fields .
we have reviewed the construction of neutrino mass models in extra dimensions .
the possibility of large extra dimension ; i.e @xmath98 has resulted in gaining very different perspectives in model building compared similar exercises in 4d .
these models also have different low energy phenomenology than their 4d counterparts . in the dirac neutrino case one
expects a sterile component in neutrino oscillation pattern . in the radiative mass mechanism
we expect kk modes of gauge bosons that couples to bileptons .
the 4d models on the other hand have only exotic scalar particles .
these models has a serious down side .
thus far all of them has many free parameters and is no better or worse than the 4d models .
we also encounter very deep theoretical questions .
what physics determines the compactification scale remains unanswered . for the orbifold models the origin of parities of the various fields is a mystery . due to the lack of space we have not discuss neutrino mass in models that implement the seesaw mechanism in extra dimensions .
this is studied for the @xmath18 case in @xcite . omitted
are also discussions of neutrino masses in models with warp extra dimensions @xcite .
we will only mention that this has been explored in the case of dirac neutrinos @xcite and @xcite . within the context of unification models
the @xmath17 case was studied in @xcite .
the many beautiful neutrino experiments have given us the first hint of physics beyond the sm .
the bilarge neutrino mixing came as real surprise to theorists .
they have stimulated us to explore scenarios which we have not ventured before .
we can look forward for more data from the ongoing neutrino experiments and perhaps even more unexpected results .
happily there will be more data to come both from high energy experiments with the anticipated turning on of lhc as well as many precision measurements at lower energies . before the new phyiscs becomes clear it behooves us to keep an open mind and be alert of new signatures .
i thank prof c.w .
kim , dr e.j .
park and the organizers of icfp03 for their hospitality and for a stimulating conference .
i am grateful to dr.w.f chang for many discussions and for teaching me much during our collaborations .
99 y. fukuda , superk collaboaration , q.r .
ahmad , sno collaboration , k. eguchi , kamland collaboration , c.l .
bennet , wmap collaboration , v.m .
lobashov , + c. weinheimer , h.v .
klapdor - kleingrothaus et al , mod .
* 16 * ( 2001 ) 2409 s.m .
bilenky , s. pascoli , and s.t .
petcov , + j.d .
vergados , phys .
c361 * ( 2001 ) 1 + f. feruglio , a. strumia , and f. vissani , b. pontecorvo , jetp * 7 * ( 1958 ) 172 and * 26 * ( 1968 ) 984 + z. maki , m. nakagawa , and s. sakata , s.f .
king hepph 0310204 n. arkani - hamed , s .
dimopoulos , and g. dvali .
+ i. antoniadis , n. arkani - hamed , s. dimopoulos , and g. dvali , + i. antoniadis , 317 k.r .
dienes , e. dudas , and t. gherghetta , + n. arkani - hamed , s. dimopoulos , g. dvali , and j. march - russell , g. dvali and a.yu .
smirnov , r.n .
mohapatra and a. perez - lorenzana , + a. lukas , p. ramond , a. ramanino , and g.g .
ross , + k.r .
dienes and i. sarcevic , + c.s .
lam and j.n .
ng , g. mclaughlin and j.n .
ng , , + _ ibid _ * b493 * ( 2000 ) 88 + g. mclaughlin and j.n .
ng , n. arkani - hamed and m. schmaltz , + n. arkani - hamed , y. grossman , and m. schmaltz , e.a .
mirabelli and m. schmaltz , + g.c .
branco , a. de gouvea , and m.n .
rebelo , + g. barenboim , g.c .
branco , a. de gouvea , and m.n .
reblo , + w.f .
chang , i.l . ho , and j.n .
ng , + w.f . chang and j.n .
ng , + j.m .
frere , g.moreau , and e.nezri hepph 0309218 a. zee , ; c. jarlskog , m.matsuda , s.skadhauge , and m. tanimoto , + p.h .
frampton and s.l .
glashow , + d.a .
dicus , h - j he , and j.n .
ng , y. kawamura , for a review see l.j .
hall and y. nomura , ann . phys .
* 206 * ( 2003 ) 132 .
chang , w.f .
chang , and j.n .
ng , w.f . chang and j.n .
. hall and y. nomura , s. weinberg , w.f chang and j.n .
ng in preparation l. randall and r. sundrum , y. grossman and m. neubert , g. mclaughlin and j.n .
ng , for a review see s. raby hepph 0308052 for a review see s.j . huber and q. shafi , j. phys .
* g29 * ( 2003 ) 1807 | neutrinos play a crucial role in many areas of physics from particle physics at very short distances to astrophysics and cosmology .
it is a long held believe that they are good probes of physics at the gut scale .
recent developments have made it clear that they can also be of fundamental importance for the physics of extra dimensions if these exist .
here we pedagogically review the construction of neutrino mass models in extra dimensions within the brane scenarios .
these models are usually nontrivial generalization of their four dimensional counter parts .
we describe the theoretical tools that have been forged and the new perspectives gained in this rapidly developing area . in particular we discuss the issues involve with building models without the use of right - handed singlets .
it is very difficult to directly test the origin of neutrino masses in the different models be it in four or more five dimensions .
we point out that different models give very different indirect signatures at the tev region and precision measurements . * neutrino mass models in extra dimensions * + john n.ng + triumf theory group , 4004 wesbrook mall + vancouver , b.c .
+ canada v6 t 2a3 + |
if macromolecules such as colloids are immersed in a solvent of smaller particles , the difference in size makes it useful to describe this mixture in terms of effective interactions by integrating out the degrees of freedom of the solvent .
the resulting interactions between particles of the remaining larger component are often referred to as depletion interactions .
such depletion forces have been studied in detail both theoretically @xcite and experimentally @xcite .
most theoretical approaches are based on a brute - force approach in which the solutes are frozen in a given configuration and thereby turned into an external field for the solvent @xcite . from the inhomogeneous structure of the solvent in the external field due to two fixed solute particles
one can calculate the solvent mediated effective force acting on the solutes .
this brute - force approach turns out to be very time consuming because for each separation and orientation , for which one wants to determine the depletion force , the inhomogeneous solvent distribution has to be calculated anew .
the insertion approach to depletion potentials @xcite differs in character in that there only one solute particle is fixed .
the advantage is that the calculation of the solvent density profile in the external field of a single solute is usually much simpler and hence computational less demanding .
the potential of the depletion force can be determined from the solvent density profile close to one solute by inserting the second solute into the system using the potential distribution theorem @xcite . however , for the insertion step a theoretical description of a mixture consisting of solute and solvent particles is required .
so far , most studies of depletion forces have been focused on rather simple geometries , such as the force between a big sphere and a planar wall or between two big spheres in a solvent of small spheres . for these symmetric systems
the depletion potential depends on the sphere - wall or the sphere - sphere separation as the only parameter characterizing the configuration .
in addition , for these geometries the use of the aforementioned insertion approach is facilitated by the availability of reliable density functional theories for hard - sphere mixtures . in colloidal mixtures one
generally encounters more complex particle shapes both of the solute and solvent particles .
however , the understanding of depletion potentials for non - spherical objects is still rudimentary .
the corresponding calculations are much more challenging because in these cases the depletion potential depends not only on the separation between the solutes but also on their relative orientation .
a first extension beyond the mixture of spheres is the case of spherical solutes immersed in a solvent of non - spherical particles .
depletion agents such as thin rods @xcite or infinitely thin platelets @xcite can generate big depletion effects even at rather low solvent concentrations @xcite .
the depletion potential in these cases can be calculated in the limit of small solvent densities so that correlations among the depletion agents are small .
similar in spirit is also the calculation of the depletion force between spheres in a solvent of a liquid - crystal in its nematic phase @xcite , for which the strong correlations between particles of the liquid crystal are taken into account effectively by reducing their orientational degrees of freedom .
the strength of the depletion interaction is then estimated by excluded volume calculations , following the ideas of asakura and oosawa @xcite . a second , more complicated situation is the one studied in ref .
@xcite , where one spherocylinder immersed in a spherical solvent close to a planar wall was considered .
these calculations employed the insertion approach .
the resulting depletion potential depends not only on the separation of the solute from the wall but also on its orientation .
hence in addition to the depletion force an entropic torque acts on the solute .
there are recent studies of depletion forces between two spherocylinders in a solvent of spheres which overcome the asakura - oosawa approximation @xcite . using a three - dimensional integral - equation theory ,
kinoshita @xcite showed that the corresponding depletion potential displays a rich behavior and depends sensitively on the path along which the spherocylinders approach each other .
similar findings were reported in a simulation study , in which the depletion potential was determined by the acceptance ratio method @xcite . here
we extend the dft insertion approach to depletion potentials @xcite to the case of non - spherical objects . in sec . [
sec : theory ] we recall the basic theory and highlight how the geometry of the solutes can be accounted for .
we test the new elements of the theory in sec .
[ sec : test ] .
as an application we discuss the depletion potential between two ellipsoids in sec .
[ sec : application ] .
we conclude in sec .
[ sec : conclusion ] .
we follow the versatile and successful approach to calculating depletion potentials between two objects , @xmath0 and @xmath1 , immersed in a solvent within the framework of density function theory ( dft ) @xcite , which is referred to as the insertion approach . to this end
we fix one of the objects , say @xmath0 , at the origin at a given orientation so that it acts as an external potential for the solvent particles . in response to this external potential , the _ s_olvent particles aquire an inhomogeneous equilibrium number density distribution @xmath2 . in the fluid phase , the case we are interested in here , @xmath2 shares the spatial symmetry with that of object @xmath0 . if object @xmath0 is a sphere , the density distribution @xmath2 possesses also spherical symmetry . in a second step
we insert the second object , denoted by @xmath1 , into the inhomogeneous solvent at position @xmath3 and with relative orientation @xmath4 . as a result of this insertion the grand potential @xmath5 of the system changes .
the depletion potential is given by @xcite @xmath6 which can be re - written in terms of the one - body direct correlation function @xmath7 @xcite : @xmath8 for the numerical calculation of @xmath9 two challenges have to be overcome : ( i ) the _ accurate _ calculation of the density profile @xmath2 of solvent particles around an object of complex shape , and ( ii ) the insertion of a non - spherical object into an inhomogeneous solvent of spheres . since both of these problems require non - standard approaches , in the following we shall pay special attention to them .
although our approach is flexible and can treat also soft solvent - solvent and solute - solvent interactions , in the following we shall restrict our considerations to the case of hard - core interactions .
the solvent is represented by a hard - sphere fluid characterized by its radius @xmath10 and bulk density @xmath11 or bulk packing fraction @xmath12 .
the solute - solvent interaction is infinitely repulsive in the case of overlap and zero otherwise .
concerning the first issue of calculating the density profile around non - spherical objects we apply the recently suggested and successfully tested curvature expansion of the density profile @xcite . to this end
we introduce normal coordinates . any point @xmath3 outside the fixed object @xmath0
can be reached from @xmath13 as the point closest to @xmath3 on the parallel surface of the object , where the density profile @xmath2 vanishes discontinuously .
this particular surface is special to the case of hard - core solute - solvent interaction ; however , it is possible to employ any parallel surface as long as all the calculations are in line with this definition of @xmath13 . for the vector connecting the points @xmath13 and @xmath3
one has @xmath14 , where @xmath15 is the unit vector normal to the parallel surface at point @xmath13 and @xmath16 is the normal distance . at point
@xmath13 the parallel surface exhibits two principle radii of curvature @xmath17 and @xmath18 leading to the dimensionless mean and gaussian curvatures @xmath19 and @xmath20 , respectively . as an ansatz for the density profile we employ @xcite @xmath21 which factorizes the local geometry of object @xmath0 at position @xmath13 , specified by the local mean and gaussian curvature @xmath22 and @xmath23 , respectively , and the structure of the solvent via the coefficient functions @xmath24 with @xmath25 .
this separation of geometry and the coefficient functions allows one to infer @xmath24 from simple geometries with high symmetry , such as a fluid close to planar , spherical , or cylindrical walls .
the functions @xmath24 have been determined in ref .
@xcite . there
the direct comparison between a density profile predicted by the curvature expansion , based on these coefficient functions and obtained within dft , with results from monte - carlo simulations has demonstrated the high accuracy of this approach .
note that the curvature expansion of the density profile , eq .
( [ cerho ] ) , implicitly assumes that the curvature of the fixed object varies smoothly on its the surface .
a sharp edge would be problematic because the mean curvature jumps from zero to a non - vanishing value at the edge . on the surface of a spherocylinder both the mean and the gaussian curvature vary discontinuously where the spherical cap meets the cylinder .
also this discontinuity of the curvatures can not be captured fully by the curvature expansion .
the contact density @xmath26 of a fluid of hard spheres with radius @xmath10 and packing fraction @xmath27 at a hard spherocylinder with length @xmath28 and radius @xmath29 so that the contact density occurs at @xmath30 .
the path on the parallel surface of contact of the spherocylinder is parameterized by @xmath31 , as is indicated by the dotted line in the inset .
note that the inset is not drawn to scale .
according to the curvature expansion the contact value @xmath26 jumps at @xmath32 , where the cylindrical and the spherical parts meet , as shown by the full line .
the symbols denote data from monte carlo simulations for @xmath27 . ] in order to analyze the reliability of the curvature expansion we have performed a monte carlo simulation of a hard - sphere fluid exposed to a spherocylinder . as parameters we have chosen @xmath27 for the packing fraction of the fluid , @xmath28 for the length of the cylinder , and @xmath30 for the radius of the parallel surface at which the density profile discontinuously drops to zero . in this test we focus on the contact density @xmath26 as the most sensitive quantity , where @xmath31 parameterizes a path along the surface as depicted in the inset of fig .
[ fig : compare ] . the density profile
away from contact decays towards the bulk density and we verified that that the effects of the discontinuity of the curvature decreases with increasing normal distance from the spherocylinder .
the dimensionless curvatures on the cylindrical part of the surface are @xmath33 and @xmath34 , while they are @xmath35 and @xmath36 on the spherical caps .
accordingly , the curvature expansion predicts a jump in the contact value of the density profile where the spherical caps meet the cylinder .
the result of the curvature expansion is plotted as the full line in fig .
[ fig : compare ] .
note that in eq .
( [ cerho ] ) only the three terms @xmath37 , @xmath38 , and @xmath39 have a non - vanishing contact value and hence contribute to @xmath26 @xcite .
in contrast to the jump of the contact density between constant values as predicted by the curvature expansion , in the computer simulations we find a smooth , slightly oscillatory transition between the contact densities at the cylindrical and the spherical part of the spherocylinder . in fig . [
fig : compare ] the simulation data for @xmath26 are shown as symbols .
interestingly , the spatial region of deviation between the simulation data and the prediction of the curvature expansion is relatively narrow .
although the density profile @xmath2 , as given by eq .
( [ cerho ] ) , is in principle sufficient for the calculations we intend to perform , it is numerically more efficient to exploit the particular form of fundamental - measure theory ( fmt ) functionals @xcite . within fmt
the one - body direct correlation function is given by @xmath40 where @xmath41 is the excess free energy density and @xmath42 depends on @xmath2 in a complicated , nonlinear way .
the convolution product is denoted as @xmath43 .
instead of calculating @xmath44 from @xmath2 directly , we argue that @xmath45 can be equivalently expanded in terms of powers of @xmath46 and @xmath47 and therefore can be written as @xmath48 one can adopt the point of view that curvature expansions such as those given in eqs .
( [ cerho ] ) and ( [ cepsi ] ) are merely approximations of the functions @xmath2 or @xmath49 that take the shape of the external potential into account in an efficient way . within this line of arguments
there is nothing special about the density distribution or any other function entering the dft .
even auxiliary functions such as the weighted densities @xmath50 and @xmath45 have a curvature expansion ensuring that the output of the dft calculation has the form of eq .
( [ cerho ] ) .
a more systematic , albeit more involved , point of view , which we present here only as a sketch , starts with an approach similar in spirit to the one that leads to eq .
( [ cerho ] ) @xcite as a suitable form for the density profile . here
, however , we analyze the weighted densities , defined as @xmath51 close to planar , spherical , and cylindrical walls .
our results suggest that analogous to the density profile [ eq . ( [ cerho ] ) ] also the weighted densities can be expanded in terms of powers of the curvatures @xmath46 and @xmath47 . as in ref .
@xcite we can determine uniquely the coefficient functions @xmath52 for @xmath53 up to third order in the inverse of the radii of curvatures inferred from planar , spherical , and cylindrical geometries . in order to obtain higher order contributions one would have to consider more complex wall shapes .
the numerical accuracy of weighted densities calculated at more complex walls is , however , unsatisfactory and practically prevents a reliable decomposition into coefficient functions . since we are interested in calculating the depletion potential between two _ big _ non - spherical objects immersed in a solvent of _ small _ spheres , the coefficient functions which we have determined are sufficient . for our approach to be quantitatively reliable , curvatures of the surface of the non - spherical objects should always be sufficiently small .
we therefore use as an ansatz for the weighted densities , following ref .
@xcite , @xmath54 which we can insert into the free - energy density @xmath41 or its derivative with respect to @xmath55 , i.e. , @xmath44 .
both @xmath41 and @xmath44 are highly nonlinear functions of @xmath55 . however , by taylor expanding @xmath41 or @xmath44 into powers of @xmath55 and by re - arranging terms one can see immediately that the curvature expansions of @xmath44 follow directly from eq .
( [ cena ] ) .
we now turn to the second part of the present problem , i.e. , the calculation of the insertion free energy [ eq .
( [ dep ] ) ] of the non - spherical object @xmath1 .
note that the change in the grand potential due to the insertion of object @xmath1 into a _ homogeneous _
bulk fluid at @xmath56 can be described by using the morphometric approach @xcite . in the bulk , the change in grand potential of the system can not depend on the orientation of the inserted object , which simplifies the problem somewhat .
furthermore , it was shown that the problems even simplify further due to the separation of the geometry and the shape independent thermodynamical coefficients . in the morphometric approach
, the insertion free energy of object @xmath1 in a bulk fluid can be written as @xcite @xmath57 where @xmath58 , @xmath59 , @xmath60 , and @xmath61 are the pressure , the planar wall surface tension , and two bending rigidities , respectively , which depend on the state of the bulk fluid and the interaction between the fluid and object @xmath1 .
these coefficients can be obtained in simple geometries .
the corresponding geometrical measures @xmath62 , @xmath63 , @xmath64 , and @xmath65 describing the shape of object @xmath1 are the volume , the surface area , and the integrated ( over the surface area ) mean and gaussian curvature , respectively . in order to calculate the convolutions in eq .
( [ cb1 ] ) the weight functions @xmath66 for a non - spherical object are required . for this problem
we employ rosenfeld s formulation of fundamental measure theory generalized to convex hard bodies @xcite . there are four scalar weight functions : @xmath67 which defines the volume @xmath62 of object @xmath1 , @xmath68 which defines the surface area @xmath63 of @xmath1 , @xmath69 which defines the integrated ( over the surface ) mean curvature @xmath64 of @xmath1 , and @xmath70 which defines the integrated ( over the surface ) gaussian curvature or euler characteristics @xmath65 of object @xmath1 . besides the scalar weight functions , which represent the geometrical properties of @xmath1 , there are two additional vector - like weight functions which are required for the de - convolution of the mayer-@xmath71 function describing the interaction between non - spherical particles .
the vector - like weight functions are given by @xmath72 where @xmath73 is the unit vector of the surface normal at point @xmath3 , and @xmath74 in a bulk system the vector - like weighted densities @xmath75 and @xmath76 vanish .
these weight functions have been employed successfully in the calculation of the depletion potential between a hard spherocylinder and a planar hard wall @xcite . with eqs .
( [ dep ] ) and ( [ cb1 ] ) we can now calculate the depletion potential @xmath77 . from the knowledge of the depletion potential @xmath77 it is possible to determine the entropic force and the entropic torque @xcite acting on object @xmath1 with orientation @xmath4 at a given position @xmath3 . if object @xmath1 is translated by an infinitesimal vector @xmath78 , while keeping its orientation fixed , the depletion potential changes by @xmath79 , which defines the depletion force @xmath80 the torque acting on the object @xmath1 can be calculated by a similar consideration , rotating object @xmath1 by an infinitesimal angle @xmath81 .
the direction of @xmath82 is parallel to the axis of rotation and its modulus specifies the angle of rotation . if one keeps the center of @xmath1 fixed at @xmath3 and performs a rotation by @xmath82 , the depletion potential changes by @xmath83 , so that @xmath84 .
therefore we can write the entropic torque as @xcite @xmath85
depletion interaction between an oblate ( a ) [ prolate ( b ) ] ellipsoid with half - axes @xmath86 [ @xmath87 and a sphere with @xmath88 .
the center of the inserted sphere approaches the north pole of the fixed ellipsoid normal to the surface .
both particles touch if @xmath89 .
the symbols denote results obtained by approach ( 1 ) , for which we employ fmt for convex objects , and the lines denote corresponding results from approach ( 2 ) , for which we use the curvature expansion .
as both approaches involve approximations of very different nature , the excellent agreement is very likely to be due to the fact that the systematic error is very small in both approaches .
we verified this observation also for ellipsoids with different half - axes .
for these data the depletion agent is a fluid of hard spheres with radius @xmath10 modeled via the white - bear version of fmt .
the bulk packing fraction is @xmath27 .
note that the insets here and in the following figures are not drawn to scale . ]
before we apply the above formalism to the calculation of the depletion potential between two non - spherical objects , we perform a test that enables us to estimate the errors introduced into the numerical calculation through the approximations we have made .
one source of error is the use of the curvature expansion of the density profile @xmath2 [ eq . ( [ cerho ] ) ] and of the derivatives @xmath45 of the excess free - energy density [ eq . ( [ cepsi ] ) ] .
it introduces an approximation because we have to truncate the expansion after the third order in the inverse radii of curvature .
another source of error is the application of the fmt weight functions for non - spherical objects , which introduces a different approximation .
the fact that these two approximations are very distinct in nature allows us to perform a stringent numerical test . to this end
we calculate the depletion potential between one big sphere and one big ellipsoid in a solvent of small spheres . in an _ exact _ treatment
the depletion potential between these two objects would depend only on their relative position and orientation . in the numerical implementation of the insertion method we can follow two different routes which make independent use of the different approximations .
we can choose to fix either the sphere and insert the ellipsoid or fix the ellipsoid and insert the sphere . by the choice of particle fixed at the origin we decide about the symmetry of the external potential and how to calculate the density distribution of the solvent .
if we choose to fix the big sphere the calculation of the density profile @xmath90 is straightforward and , using the spherical symmetry of the problem , it was established that the results agree extremely well with , e.g. , monte carlo simulations . along this route , the main approximation for the calculation of @xmath77 stems from using the weight functions for the ellipsoid .
note that the resulting depletion potential @xmath91 depends on both the orientation @xmath92 of the ellipsoid relative to the vector * r * connecting the centers of the ellipsoid and the sphere and the distance @xmath93 between the ellipsoid and the sphere . for this test
, however , we fix the orientation and consider the approach between the ellipsoid and the sphere along the surface normal of the sphere for the chosen orientation of the ellipsoid . in fig .
[ fig : test ] we show the depletion potential between a big sphere , denoted as object @xmath0 , with radius @xmath94 and an oblate ellipsoid with half - axes @xmath95 ( a ) , and between a sphere with radius @xmath94 and a prolate ellipsoid with half axes ( @xmath96 ( b ) . in both cases
the solvent is a fluid of small spheres with a packing fraction @xmath27 , which we model by the white bear version of fmt @xcite .
the insets in fig .
[ fig : test ] depict the orientation between the ellipsoids and the sphere chosen in the calculation .
the symbols denote the results obtained via the first route , corresponding to a fixed sphere . along the second route
, we fix the ellipsoid at the origin so that it acts as an external potential for the solvent spheres .
now we employ the curvature expansion [ eq .
( [ cepsi ] ) ] in order to evaluate the derivatives of the excess free energy density @xmath45 .
the weight functions we need in order to describe the insertion of the big sphere in eq .
( [ cb1 ] ) are well tested and are known to be accurate @xcite .
the results for the depletion potentials along the same paths between the sphere and the oblate and prolate ellipsoid obtained from this route are shown in fig .
[ fig : test ] as lines .
we find that the results obtained from both routes agree extremely well , which provides confidence in the reliability of the approximations and the numerical approach . only in the case of the prolate ellipsoid , for which the curvatures are considerably higher than in the case of the oblate ellipsoid , we find some deviations between the two routes close to the first potential barrier . however , these deviations are very small . close to contact between the sphere and the ellipsoids , one can employ arguments based on considerations about the overlap of excluded volumes @xcite . in the case of high curvature one
expects that the contact value of the depletion potential is considerably reduced compared to cases of low curvature .
this expectation is confirmed by our results .
in addition to the contact value , we find that the amplitude of the oscillations of the depletion potential for the oblate ellipsoid [ fig . [
fig : test](a ) ] is larger than the one for the prolate ellipsoid [ fig .
[ fig : test](b ) ] .
we conclude from the results of this test that both the curvature expansion of the functions @xmath45 and the insertion of a non - spherical particle into an inhomogeneous solvent of small spheres work reliably .
we can now turn to the calculation of the depletion potential between _ two _ non - spherical , convex objects . in order to accomplish this
we have to combine both steps mentioned and tested above .
first we fix one of the two non - spherical objects and thus turn it into an external potential for the solvent of small spheres .
the structure of the resulting inhomogeneous solvent density distribution at a given bulk density is captured by the curvature expansion given by eq .
( [ cepsi ] ) . in the second step
we employ the insertion of a non - spherical object .
depletion interaction between two oblate ellipsoids with half axes @xmath86 .
the inserted particle approaches the north pole of the first ellipsoid perpendicular to its surface .
@xmath97 denotes the angle between the z - axis and the large half - axis of the inserted ellipsoid ( see the inset ) . if both ellipsoids are aligned ( as drawn in the inset ) , i.e. , @xmath98 , the minimal separation of the centers is @xmath99 , while for @xmath100 it is @xmath101 ; @xmath102 and @xmath103 . the positions of the extrema and the zeros are basically independent of @xmath97 .
the solvent is a fluid of hard spheres with radius @xmath10 and bulk packing fraction @xmath27 and is modeled via the white - bear density functional . ] in order to illustrate our approach we calculate the depletion potential between two hard oblate ellipsoids with half - axes @xmath95 in a solvent of hard spheres with radius @xmath10 and bulk packing fraction @xmath27 .
the resulting depletion potential depends on both the relative orientation @xmath92 of the two ellipsoids and the vector @xmath3 connecting their centers and thus depends on 6 variables . out of this high - dimensional parameter space
we select a few examples of paths along which we study the behavior of the depletion potential . in the first example ,
the center of the inserted ellipsoid approaches the north pole of the fixed particle along the surface normal , as shown in the inset of fig .
[ fig : twooblates ] . for this path
we vary also the relative orientation @xmath92 between the particles . due to the symmetry of this configuration the orientation between the oblate ellipsoids
can be expressed in terms of a single angle , which we denote by @xmath97 . if we choose @xmath98 ,
the ellipsoids are parallel and the minimal separation of their centers is @xmath99 . at contact
the overlap of excluded volume is larger than for any other value of @xmath97 .
hence the contact value of the depletion potential is most negative . for our choice of parameters we find @xmath104 ( see fig . [
fig : twooblates ] ) .
in addition to the strong attraction close to contact , the depletion potential displays a pronounced oscillatory structure away from contact .
this oscillatory structure of the depletion potential reflects mainly the properties of the solvent .
the structure of the hard - sphere solvent normal to the surface displays oscillatory , exponentially decaying packing effects .
we have shown previously @xcite that beyond the first maximum the depletion potential between two spheres or between a sphere and a planar wall decays in a closely related , exponentially damped oscillatory fashion .
the decay length and the wavelength of the oscillations in the depletion potential are exactly the same as those in the decay of the bulk pair correlation function .
only the amplitude of this decay and the phase of the oscillations depend on the shape and the orientation of the two solutes .
depletion potential for the same setup as in fig .
[ fig : twooblates ] with fixed @xmath98 .
the angle @xmath105 characterizes the straight line along which the center of the inserted ellipsoid is moved away from its contact position at the north poles ( see the inset ) .
the abscissa measures the normal ( i.e. , minimal ) distance @xmath16 of the center of the inserted ellipsoid from the surface of the fixed one . for small @xmath105
the depletion potentials almost coincide due to the oblate shape @xmath86 of the fixed ellipsoid . ] as we decrease the value of @xmath97 from 90@xmath106 to 0 , thereby changing the relative orientation from parallel to normal , we find that the contact value as well as the potential barrier at @xmath107 and the amplitude of the oscillations decrease monotonically ; the positions of the extrema and of the zeros basically do not vary with @xmath97 . in the second example the ellipsoids always approach each others north poles in a parallel configuration .
the corresponding path of the centers is a straight line forming an angle @xmath105 with the fixed direction of the small axes ( see the inset of fig .
[ fig : oblatesangle ] ) . in fig .
[ fig : oblatesangle ] , by construction the path for @xmath108 is identical to the path with @xmath98 in fig .
[ fig : twooblates ] . upon increasing @xmath105 , at first
the resulting depletion potential changes only slightly but for angles @xmath109 it is possible to observe a clear decrease in the wavelength of oscillation , which is most pronounced in the case of @xmath110 .
this , however , is not a contradiction to the aforementioned universality of the oscillatory decay of the depletion potential , because this path cuts through the three - dimensional oscillatory structure of the solvent , which is organized _ normal _ to the surface of the fixed solute , at the angle @xmath105 .
as pointed out earlier , for non - spherical objects it is also possible to keep the distance between their centers constant and change the relative orientation . from the corresponding change in the depletion potential
we can obtain the entropic torque @xcite acting on the inserted particle .
for the same geometrical setup as in fig .
[ fig : twooblates ] we calculate the torque @xmath111 for center to center separations @xmath112 , 12 , 14 , and 16 . due to the symmetry of this setup
the entropic torque acting on the inserted ellipsoid , relative to its center , is parallel to the rotation @xmath92 , i.e. , @xmath113 , with @xmath114 and @xmath115 the symmetry of the problem leads to @xmath116 for @xmath100 and @xmath117 . a positive value of the torque acts on the inserted ellipsoid as to increase the angle @xmath97 ( rotating it towards an orientation parallel to the fixed ellipsoid ) , while a negative value of @xmath111 leads to a decrease of @xmath97 ( rotating it towards an orientation normal to the fixed ellipsoid ) . some typical examples of the torque as a function of @xmath97 for various values of @xmath118 are shown in fig .
[ fig : oblatetorque ] . for small values of @xmath118
the amplitude of the torque is largest ; however , the hard - core interaction prohibits small values of @xmath97 due to geometrical constraints . for larger separations @xmath118 between the ellipsoids , the accessible range of values of @xmath97 increases until finally the inserted ellipsoid can rotate freely without encountering the fixed ellipsoid . due to the symmetry of the problem
the torque vanishes for @xmath98 as well as for @xmath100 , provided this orientation is accessible .
entropic torque between two oblate ellipsoids with half - axes @xmath86 immersed in a solvent of hard spheres with radius @xmath10 as sketched in the inset .
the angle @xmath97 parameterizes the rotation of the inserted ellipsoid and @xmath118 denotes the distance between the centers .
the ellipsoids are aligned such that the vector connecting their centers runs through the north poles of the fixed ( lower ) ellipsoid . for this setup
the entropic torque @xmath111 is given by @xmath119 @xcite . if @xmath120 , the inserted ( upper ) ellipsoid can be rotated as indicated in the inset and experiences an entropic torque .
positive values of @xmath111 mean that the ellipsoid is pushed towards larger values of @xmath97 .
the curves end if the two ellipsoids overlap due to the geometrical constraint .
note that the visible discontinuities in the first derivative of the curves for @xmath121 and @xmath122 are not artifacts .
they occur if both ellipsoids come so close that just a single small sphere with radius @xmath10 fits in between their surfaces . for @xmath123
the distance between the ellipsoids is large enough so that this effect does not occur . ]
if for any relative orientation of the ellipsoids the minimal distance between the surfaces equals the diameter @xmath124 of a solvent spheres , the torque exhibits a cusp , as can be seen in the cases of @xmath125 . for @xmath126
this orientation occurs at @xmath98 .
accordingly , for sufficiently large @xmath118 no such cusps occur . all these feature of the torque ( see fig . [
fig : oblatetorque ] ) are similar to those of the entropic torque acting on a spherocylinder close to a planar wall @xcite . from these results
, one can speculate how two freely floating ellipsoids most likely would approach each other .
similar to the case of a spherocylinder close to a planar wall @xcite we find that the ellipsoids prefer to be parallel once they touch each other . this configuration corresponds to @xmath98 in fig .
[ fig : twooblates ] . in order to be able to overcome potential barriers while approaching ,
it is easier for the ellipsoids to approach with a relative orientation different from the parallel configuration . by first seeking contact in regions of high curvature , and then by adapting the orientation such that regions of low curvature get in contact with each other , potential barriers , which have to be overcome , are significantly reduced and the minimum of the depletion potential can be reached .
this interpretation agrees with observations reported in refs .
@xcite where the depletion potential between two spherocylinder was studied .
we have extended our density functional theory approach for calculating depletion potentials @xcite to the effective interactions between generally shaped convex particles with surface curvatures which vary smoothly . as in the previously studied @xcite geometrically simple case of two spherical particles ,
it is most efficient to carry out the calculation in two steps . in the first step
we describe the structure of the solvent close to a fixed object . we have shown that a curvature expansion of the density profile , or equivalently of the derivatives @xmath44 of the free energy density [ eq . ( [ cepsi ] ) ] provides an efficient tool . by employing the curvature expansion we implicitly assume that the mean and gaussian curvatures of the surface of the solute vary smoothly across its surface .
discontinuities in the curvatures , such as those observed at sharp edges or close to the spherical caps of a spherocylinder can not be captured fully by the curvature expansion in its present form ( see fig . [
fig : compare ] ) . in a second step
we insert the second particle into the inhomogeneous solvent . to this end
we employ an extension of fundamental measure theory to non - spherical particles @xcite , which was already used in the study of a spherocylinder close to a planar wall @xcite . using this approach
we have studied the depletion potential between one ellipsoid and a big sphere immersed in a hard - sphere solvent . in this case
we could test the accuracy of our approach by carrying out the calculations in two different ways , which employ the various approximations of our approach independently .
we have found excellent agreement between the two routes ( see fig .
[ fig : test ] ) which provides confidence for the scheme used .
furthermore , we have studied the depletion potential between two equally sized oblate ellipsoids . for this case
we have illustrated the potential of our approach . from the resulting depletion potential ( see figs . [
fig : twooblates ] and [ fig : oblatesangle ] ) one can calculate the entropic force acting on the centers of the solutes as well as the entropic torque ( fig . [ fig : oblatetorque ] ) .
this provides a picture for the likely pathway of how two freely floating ellipsoids approach each other under the action of entropic forces besides the application to colloidal mixtures of non - spherical objects and spheres , this approach should prove useful to studying biological inspired model key and lock systems @xcite , for which depletion interactions between non - spherical objects and geometrically structured substrates are considered .
the non - spherical objects and the substrates display a perfect geometrical match , similar to biological macromolecules ( key ) which can form chemical bonds with a cavity ( lock ) only if they are sufficiently close together . the issue is how the key can be guided into the lock using a robust , chemically unspecific mechanism .
our present analysis provides first steps towards addressing this issue .
99 s. asakura and f. oosawa , j. chem . phys .
* 22 * , 1255 ( 1954 ) .
s. asakura and f. oosawa , j. polym .
sci . * 33 * , 183 ( 1958 ) .
t. biben , p. bladon , and d. frenkel , j. phys . : condens .
matter * 8 * , 10799 ( 1996 ) .
r. dickman , p. attard , and v. simonian , j. chem . phys . * 107 * , 205 ( 1997 ) .
r. roth , b. gtzelmann , and s. dietrich , phys .
lett . * 83 * , 448 ( 1999 ) .
r. roth , r. evans , and s. dietrich , phys .
e * 62 * , 5360 ( 2000 ) .
d. rudhardt , c. bechinger , and p. leiderer , phys .
lett . * 82 * , 1330 ( 1998 ) . c. bechinger , d. rudhardt , p. leiderer , r. roth , and s. dietrich , phys .
lett . * 83 * , 3960 ( 1999 ) .
l. helden , r. roth , g.h .
koenderink , p. leiderer , and c. bechinger , phys .
lett . * 90 * , 048301 ( 2003 ) .
m. kinoshita , j. chem .
phys . * 116 * 3493 ( 2002 ) .
m. kinoshita , chem .
. lett . * 387 * , 47 ( 2004 ) .
m. kinoshita , chem .
sci . * 61 * , 2150 ( 2006 ) .
henderson , mol .
* 50 * , 741 ( 1983 ) .
y. mao , m.e .
cates , and h.n.w .
lekkerkerker , phys .
75 * , 4548 ( 1995 ) .
y. mao , m.e .
cates , and h.n.w .
lekkerkerker , j. chem .
phys . * 106 * , 3721 ( 1997 ) .
k. yaman , c. jeppesen , and c.m .
marques c m , europhys . lett . * 42 * , 221 ( 1998 ) .
r. roth , j. phys .
: condens .
matter * 15 * , s277 ( 2003 ) .
oversteegen and h.n.w .
lekkerkerker , phys , rev .
e * 68 * , 021404 ( 2003 ) .
l. harnau and s. dietrich , phys .
e * 69 * , 051501 ( 2004 ) .
l. harnau and s. dietrich , in _ soft matter _ , edited by g.gompper and m. schick ( wiley , berlin , 2006 ) , vol . 3 , to appear .
p. van der schoot , j. chem . phys .
* 112 * , 9132 ( 2000 ) .
p. van der schoot , j. chem . phys . * 113 * , 3931 ( 2000 ) .
r. roth , r. van roij , d. andrienko , k.r .
mecke , and s. dietrich , phys .
89 * , 088301 ( 2002 ) .
w. li and h.r .
ma , euro .
j. e * 16 * , 225 ( 2005 ) .
knig , p. bryk , k. mecke , and r. roth , europhys .
lett . * 69 * , 832 ( 2005 ) .
knig , r. roth , and k. mecke , phys .
* 93 * , 160601 ( 2004 ) .
y. rosenfeld , phys rev .
lett . * 63 * , 980 ( 1989 ) .
y. rosenfeld , phys .
e * 50 * , r3318 ( 1994 ) .
y. rosenfeld , mol . phys . * 86 * , 637 ( 1995 ) .
r. roth , r. evans , a. lang , and g. kahl , j. phys . : condens . matter * 14 * , 12063 ( 2002 ) .
yu and j. wu , j. chem .
phys . * 117 * , 10156 ( 2002 ) . | we extend the insertion approach for calculating depletion potentials to the case of non - spherical solutes . instead of a brute - force calculation we suggest to employ the recently developed curvature expansion of density profiles close to complexly shaped walls .
the approximations introduced in the calculation by the use of the curvature expansion and of weight functions for non - spherical objects can be tested independently . as an application for our approach
we calculate and discuss the depletion potential between two hard oblate ellipsoids in a solvent of hard spheres . for this system
we calculate the entropic force and torque acting on the objects . |
monodisperse colloidal suspensions are rare in nature . whether these colloidal particles are artificially prepared or not , in the best case the particle - size distribution is a narrow distribution around an average size .
it is obvious that such a polydisperse nature will influence the physics and properties of these systems , and one could imagine that by the use of truly asymmetric mixtures it is possible to create systems with a behavior that can not be described by the simple monodisperse - like approximation .
conceptually we can distinguish between two types of polydispersity .
one of them , which we could refer to as ` intrinsic polydispersity ' , arises from the fact that the particles present in the system are different by construction ( in size , charge , or any other feature ) and their characteristics are not changed by the interaction with other particles .
this kind of systems are like multicomponent mixtures in which , at least in principle , the composition can be externally imposed . the new phenomenology that we can expect from this systems has its origin in fractionations into phases with different compositions ( constrained by particle conservation ) and their coupling with other transitions already present in the monodisperse system @xcite . the second kind of polydispersity can be found in self - assembling systems @xcite ( surfactants forming micelles , monomers forming chains , vesicles , etc ) .
the aggregates present in these systems can be identified as the particles , each with different size , shape , conformation , etc ; the difference with the intrinsically polydisperse systems being that the composition is determined by the chemical equilibrium between the constituents of the aggregates . as a consequence
the equilibrium is not constrained by conservation of the number of particles and therefore no fractionation is to be expected . in principle
the system can compensate losses of entropy by adjusting its composition .
there are , however , other constraints in the system ( the number of small constituents , for instance ) and these may induce new kinds of transitions characterized by the appearance of one or a few macroscopic aggregates . as such can be cataloged phenomena as the appearance of lamellar and columnar phases in surfactant solutions @xcite , emulsification failure in micro - emulsions or long chain formation in polymer solutions @xcite .
a study of polydisperse mixtures of either type is far from trivial . in the case of intrinsic polydispersity
there is the experimental problem of how to fabricate colloidal particles according to a given particle - size distribution .
although in simulations this seems to be somewhat better under control , one easily could run into the problem of finite size effects due to an insufficient or inadequate sampling of particle sizes .
this is not the case in self - assembling systems , although experimentally , their polydispersity can not be easily characterized .
theoretical descriptions of polydisperse systems are mainly based on a small set of moments of the particle - size distribution and it is assumed that mixtures with the same set of moments show a corresponding behavior .
this seems to be a rather successful approach , although it is not obvious that could still be applied to very asymmetric mixtures . in two previous articles we analyzed the behavior of idealized versions of polydisperse mixtures .
the system was assumed to be composed of @xmath0 spherical aggregates which only interact via hard - core repulsion .
the chemical equilibrium of the underlying constituents is accounted for by allowing particles to exchange size in such a way that the total volume @xcite ( compact aggregates ) or surface @xcite ( surfactant micellar membranes ) of the particles remains fixed at all times and the number of particles is constant . under the influence of the applied pressure the particle - size distribution of these systems changes . in the case of the constrained surface ,
the particle - size distribution changed from an exponentially decaying function in the low density limit to a single peaked distribution in the denser liquid phase .
for sufficiently high pressures this system can form a number of different mechanically stable crystals , a monodisperse face - centered - cubic crystal , or a bidisperse @xmath1 or @xmath2 crystal .
for the system with the constraint on the total volume of the particles , it was found by theory and simulations that at a rather low volume fraction of approximately @xmath3 the system could no longer be described by percus - yevick type of equation of state , due to the formation of macroscopically large particles . in the present work
we will show that this phenomenon is a true phase transition and provide a self - consistent theory for this sort of system .
the nature of this transition has recently been studied in ideal systems @xcite and found to be connected to bose - einstein condensation , and from this work we can conclude that interaction only changes the details , not the essential features .
the connection of the present model with bose - einstein condensation was earlier suggested by d. frenkel @xcite .
the remainder of this paper is organized as follows . in sec .
[ sec : mc ] we will show the results of computer simulations we have performed , and explain some of the details involved .
these results are compared with a theory based on the percus - yevick equation of state in sec .
[ sec : py ] . as we show in sec .
[ sec : bmcsl ] the treatment based on the heuristic bmcsl equation of state , leads to a slightly better theoretical description . in sec .
[ sec : discussion ] we finish with a brief discussion on some of the issues raised in this paper .
the system under consideration here is one formed by @xmath0 spherical particles with different sizes .
these particles are only interacting via a hard - core repulsion , and hence temperature is not a relevant parameter .
we allow , however , pairs of particles to exchange volume under the constraint that the total volume occupied by all particles remains fixed . as a consequence
the system has the freedom to explore different size - distribution functions and relax to the `` optimal '' one . in the previous work @xcite we found by computer simulations that beyond a relative small pressure or density , some of the particles became macroscopically large . from theoretical arguments an upper bound to the pressure
was obtained for fixed volume fraction , above which a single size distribution is unstable .
however , simulations indicated a discrepancy with this result .
the formation of macroscopic particles ( in simulations of the size of the simulation volume ) beyond a critical value @xcite is in fact comparable to bose - einstein condensation , or alternatively as a sort of sedimentation . in order to describe and simulate this phenomenon in our system in a correct manner
, one has to eliminate the finite size effects that originate from the relative small number ( typically of the order 1000 ) of particles that is being used in computer simulations .
this could easily be solved by increasing the number of particles by one or several orders of magnitude , but this is seldom a preferred option .
moreover , in this particular case there exists a much more elegant solution to the problem .
the introduction of huge particles in a sea of smaller ones has two effects .
the first is the reduction of the available volume that can be occupied by the smaller ones , and the second is the introduction of a hard wall formed by the big particles .
it turns out that it is the second effect that caused the discrepancy between the simulation results and the upper bound of the equation of state @xcite .
the natural solution is therefore to eliminate this particular effect , which can be done by extracting the big particle(s ) out of the simulation box .
we will assume that there is only a single big particle ( aggregate ) with a volume @xmath4 , but it can easily be extended to include several aggregates , albeit their number with respect to the simulation is not relevant . the total volume that can be occupied by the @xmath0 particles plus the aggregate , is denoted by @xmath5 , where @xmath6 is the average diameter cubed of the particles and @xmath7 defines a natural length scale .
the choice of @xmath7 in this formulation is based on the infinite dilute system for which the size of the aggregate is zero .
the volume that is accessible to the particles is denoted by @xmath8 .
but since we have extracted the aggregate out of the system the total thermodynamical volume of the system is given by @xmath9 . on this system we have performed monte carlo ( mc ) simulations using the isothermal - isobaric constant ( @xmath10 ) ensemble . for a detailed description on monte carlo simulations we refer the reader to ref .
@xcite ; here we will only list the main features . as positional order will play no role , we will assume the simulation volume @xmath8 to be cubic and we will be using four different types of moves . the first and simplest move is randomly selecting a particle and displacing it isotropically .
the second type of move was introduced before @xcite .
it selects randomly a pair of particles and attempts to exchange a finite amount of volume , by adding this to one , and subtracting it from the other particle .
hereafter their new volumes are used in order to obtain the appropriate diameters . obviously ,
if this process led to a negative volume , the move could not be accepted .
apart from that , both moves are accepted provided that no overlap is caused by these changes . moreover ,
the amount of displacements and volume exchanges are drawn homogeneously from an interval , of which the size is continously adjusted in order to obtain an average acceptance per move between 35 - 50% . in the third move
we allow the simulation box to shrink or expand isotropically , in order for the system to equilibrate with respect to the applied pressure .
this is in principle also a standard move in mc simulations using the @xmath10 ensemble .
however , in this case it is required to modify the acceptance criterion of this move .
this is caused by the fact that the thermodynamical volume @xmath11 and the volume of the simulation box @xmath8 , although related , are not identical .
the weight of any non - overlapping configuration in the @xmath10-ensemble is proportional to @xmath12 , where @xmath13 the inverse temperature .
in addition to that , there is in mc - simulations a correction factor @xmath14 due to the use of scaled coordinates in order to facilitate the volume moves and periodic boundary conditions . as a result the volume move , if no overlaps are produced , is accepted with the probability @xmath15),\ ] ] where the labels @xmath16 and @xmath17 refer to the new and old configuration respectively .
the fourth type of move is the exchange of volume between a particle and the aggregate .
although this seems to be equivalent to the volume exchange between particles , there is a subtle difference .
provided that no overlap is produced and both volumes remain positive this move is allowed . however , if the aggregate changes its volume , this results in a change of the available volume @xmath8 for the @xmath0 particles , while the thermodynamic volume @xmath11 remains fixed . as a result
this move needs to be accepted with probability @xmath18).\ ] ] we have performed a series of simulations on this system with @xmath19 particles . in each simulation
run we used @xmath20 sweeps , where in each sweep on average we try @xmath0 particle moves , @xmath21 volume exchanges , 1 volume move , and 5 exchanges of volume between the aggregate and one of the normal particles .
average quantities where obtained from a single run after equilibrium was reached .
the time required to reach equilibrium strongly depends on the initial configuration . in simulations where an equilibrium configuration of a slightly higher or lower applied pressure is used
, usually one run is sufficient .
if started from a monodisperse simple cubic lattice , one or several runs are required , depending on the initial choice of the size of the aggregate and volume of the simulation box .
the results of the equation of state are shown in fig .
[ fig : eos ] , where the the reduced pressure @xmath22 is plotted as a function of the volume fraction @xmath23 .
the results obtained from expansion and compression runs , as well as started from monodisperse systems , all lead , within the estimated errors , to the same results . near the volume fraction 0.26 , where the critical value was found @xcite , a change in slope can be observed .
a natural order parameter for this system is the relative amount of fixed total volume of the particles found in the aggregate @xmath24 and is shown in fig .
[ fig : order ] .
comparison with simulations on systems with less particles and the absence of hysteresis in the equation of state suggests the possibility of a continuous phase transition .
the simulation results below the critical value are consistent with those obtained previously @xcite . above this value
the results are different .
the presence of the aggregate seems to prevent the formation of other macroscopic particles . note that the difference is not so much the fact that there is a aggregate , but the way it is treated as an external buffer , thus eliminating the surface effects .
in order to describe the system under consideration theoretically , one should realize that the presence of a thermodynamic volume @xmath11 as well as an available volume @xmath8 for the particles should carefully be taken into account .
the system can be described by a polydisperse system of @xmath0 hard spheres in a volume @xmath8 , which is in contact with an external bath , because in addition to the exchange of volume between the particles themselves , this can also be done with this bath or aggregate .
we will describe the polydisperse system within the percus - yevick approximation of a polydisperse hard - sphere mixture @xcite , from which the equation of state yields @xmath25 where @xmath26 is the @xmath27th moment of the particle - size distribution function @xmath28 here @xmath29 is the number density defined with respect to the available volume , @xmath30 the volume fraction , and @xmath31 is the continuous particle - size distribution function , which due to the type of interaction , is a function of the particle volume @xmath32 . the helmholtz free energy functional @xmath33 of this system , however , is a function of the thermodynamic volume @xmath11 @xmath34 where the excess free energy per particle is given by @xmath35 in order to obtain the equilibrium distribution , @xmath31 , we need to minimize this free energy functional , taking into account that there are two constraints which need to be satisfied .
this is solved by adding the following two terms to the free energy functional @xmath36 where @xmath37 and @xmath38 are lagrange multipliers .
the first term is due to the requirement that the particle - size distribution function should be normalized to unity , and the second is that the combined volume of all particles and the aggregate is fixed to be @xmath39 .
the equilibrium particle - size distribution can now be obtained by minimizing the free energy functional with respect to @xmath31 and the free parameter @xmath40 .
stationarity with respect to changes in @xmath31 leads to @xmath41 from which we can immediately derive the functional form of the function @xmath31 that will minimize the free energy @xmath42 where the coefficients @xmath43 are determined by @xmath44 @xmath45 the values of @xmath46 and @xmath47 are fixed by the constraint on the total combined volume of the particles and the aggregate , and the normalization of the particle size - distribution , respectively .
formally , however , the coefficient @xmath46 is given by @xmath48 in addition to these equations we have to minimize with respect to the volume of the aggregate . note that the @xmath49 depend on the value of @xmath40 through the number density .
therefore we obtain @xmath50 this result does not automatically imply that @xmath51 .
the reason is that we have to minimize the free energy with respect to the volume of the aggregate under the additional condition that the aggregate has a non - negative volume . as a consequence this can result in the possibility that the minimum is found for @xmath52 and the stationarity equation ( [ eq : stat ] ) not being satisfied
this leads to a natural splitting of the behavior of this system . for low pressures the volume of the aggregate will be zero , and
the particle size - distribution is the exponential of a third order polynomial in the particle diameter .
on increasing the pressure the coefficient @xmath46 , which can only be non - positive in order to allow a proper normalization , goes to zero . in this region
the behavior is as explained before @xcite .
according to the theoretical description in the percus - yevick approximation , the coefficient @xmath46 becomes zero at volume fraction @xmath53 pressure @xmath54 . at this point the value @xmath52 not only minimizes the free energy , but also satisfies the stationarity equation ( [ eq : stat ] ) . for pressures larger than this value the stationarity equation can always be satisfied with a positive volume for the aggregate and
hence the coefficient @xmath46 remains zero .
it turns out that this point is the location of a continuous transition . beyond this point the system relaxes by transferring volume to the aggregate , and effectively scaling the polydisperse system to a smaller size .
if we denote by a script @xmath55 the values of the critical system , this can be shown in the following way .
if we replace the coefficients @xmath56 in the particle size - distribution by @xmath57 , where @xmath58 is a positive number , and fix @xmath47 by the normalization to unity , it follows directly from the special form ( [ eq : w ] ) of @xmath31 that the moments of the size - distribution are given by @xmath59 .
this results in @xmath60 and provided that @xmath58 is determined by @xmath61 this is a solution of eq .
( [ eq : alpha1 ] ) . applying this procedure to
@xmath62 leads to the same expression for @xmath58 .
therefore the scaled size - distribution function leads to a minimum free energy , albeit for a different applied pressure .
the pressure equation of state beyond the critical point also becomes fairly simple now .
if we rewrite the pressure ( [ eq : py ] ) using the solutions ( [ eq : alpha1 ] ) and ( [ eq : alpha2 ] ) , we obtain @xmath63 where we used that if @xmath51 the functional form of @xmath31 leads to the identity @xmath64 . we can also derive a simple expression for the order parameter @xmath24 of the system @xmath65 the curves of the equation of state and the order parameter are shown in figs .
[ fig : eos ] and [ fig : order ] respectively , and show a perfect agreement with the simulation results . in fig .
[ fig : eos ] the dotted line represents the equation of state below the transition , i.e. eq . ( [ eq : py ] ) , while the solid line is the one above as given by eq .
( [ eq : py - high ] ) .
the later one is extended to the zero density limit in order to visualize the difference between both branches . from these results two additional illustrative properties
can be derived for the system beyond the point of the transition . using that the particle size - distribution only changes according to a dimensional scaling or from the order parameter ( [ eq : order ] ) , we find that @xmath66 this simply states that the system becomes scale invariant , with a constant pressure if it is defined in terms of the average particle volume .
the other result is that @xmath67 , which can be identified with the local volume fraction of the particles , remains fixed , i.e. @xmath68 .
these results can be used in order to determine the point of transition from simulations .
this is illustrated in fig .
[ fig : transition ] , where @xmath69 as a function of @xmath70 , and the applied pressure @xmath71 as a function of @xmath67 are shown , as measured in our mc - simulations . according to these simulations
the transition is found at a volume fraction @xmath72 and pressure @xmath73 .
the prediction of the particle size - distribution @xmath31 is compared with the simulation results in fig.[fig : sizedist ] . by plotting @xmath74 as a function of @xmath75 , the distributions above the transition
coincide and agree perfectly with the predicted curve from percus - yevick .
the discrepancy at higher values for @xmath75 is merely a consequence of an insufficient sampling of the largest particles in the system . for @xmath76
, there is the additional effect that this is close to the transition , where the system tends to probe larger particles .
it turns out that within the percus - yevick approximation the system shows a continuous phase transition , which is characterized by the non - zero volume of a aggregate .
in fact it shows some similarities to bose - einstein condensation , because what actually happens is that the interaction of microscopic particles leads to the formation of a macroscopic aggregate @xcite .
it is a well known fact that the percus - yevick equation of state in a monodisperse hard sphere system can be improved by combining it with the virial equation of state .
this results in the carnahan - starling equation of state @xcite .
a similar approach can be made for polydisperse sphere mixtures , and the resulting heuristic equation of state is due to boublk , mansoori , carnahan , starling , and leland ( bmcsl ) @xcite @xmath77 a straightforward integration of this equation of state shows that the excess free energy up to a constant is given by @xcite @xmath78 the additional constant is independent of the volume , but could in principle depend on the moments of the size - distribution , without changing the equation of state
. such a constant would be relevant for this particular model , since the size - distribution of the system can be changed , and therefore an additional change in free energy can be produced . from general applicability of the free energy functional to different models ,
however , it can be shown that this contribution necessarily vanishes .
we now can apply the same procedure as was used in the case of the percus - yevick approximation .
the functional derivative of the free energy with respect to @xmath31 leads to the same functional form ( [ eq : w ] ) as before , but the stationarity equations on the coefficients @xmath43 need to be modified @xmath79 @xmath80 @xmath81.\end{aligned}\ ] ] the derivative of the free energy with respect to the volume of the aggregate leads again to @xmath82 the second term on the right hand side of eq .
( [ eq : alpha3-cs ] ) is , except in the zero density limit , always positive . as a consequence eq .
( [ eq : stat - cs ] ) can never be satisfied , because such a solution would lead to a positive @xmath46 and hence a non - normalizable size - distribution function .
the minimum free energy for given volume fraction will therefore not be a solution of all the stationarity equations simultaneously .
since the stationarity equations ( [ eq : alpha1-cs ] ) - ( [ eq : alpha3-cs ] ) are directly related to the equations for chemical equilibrium between the different particle sizes , they need to be fulfilled . if we fixed the volume of the aggregate @xmath40 , these equations would be sufficient to obtain the unique distribution function that minimizes the free energy . in fig .
[ fig : free ] we have plotted the free energy per volume for several fixed values of @xmath40 ( solid curves ) . on these curves
the stationarity equations are satisfied , hence the value of @xmath40 that leads to the minimum free energy corresponds to the equilibrium situation .
these curves are bounded by an lower and upper limit of the volume fraction .
the lower limit corresponds to the case where @xmath83 and the upper limit corresponds to the case where @xmath51 . outside this interval
no self - consistent solution of the stationarity equations can be obtained .
similar to what we found in the percus - yevick approach , we can scale the particle size - distribution function such that we obtain a solution to the stationarity equations for a different volume fraction by replacing @xmath56 by @xmath84 and fixing the normalization constant . the free energy per volume for several of these scaled solutions is also shown in fig . [
fig : free ] as dotted curves .
both sets of curves are similar , but use a different parameter to describe them .
the equilibrium situation of our model is the one that corresponds to the lowest free energy for given volume fraction and will correspond to the envelope of the set of curves in fig . [
fig : free ] .
one can show that we have two regions , which is illustrated by this figure .
below the transition the equilibrium situation corresponds to the curve for which @xmath52 , i.e. without a macroscopic aggregate . above the transition
we will follow the branch of a scaled solution , emerging from the end point of this curve determined by @xmath51 .
the transition point is therefore obtained by self - consistently solving the case for @xmath85 and leads to a volume fraction @xmath86 and pressure @xmath87 .
actually the same reasoning could have been used for the percus - yevick approach .
but since for that case all stationarity equations could be solved simultaneously , there was no need to follow this line of argumentation .
the numerical predictions of the transition point for the bmcsl approach are slightly better in agreement with the simulation results than the results from percus - yevick .
the difference between both predictions is however too small in order for it to be visible in any of the figures . in principle
also the size - distribution function can be used to distinguish both theories , however , in order to do so the number of particles should be increased several orders of magnitudes in order to obtain sufficient statistics for the larger particles and to make the difference in the tail of the distribution visible .
we have demonstrated how to obtain a self - consistent theory for a polydisperse hard sphere system , where particles are able to change their size by exchanging volume . within the percus - yevick approximation
this system shows a continuous phase transition , in which the microscopic interactions lead to the formation of a macroscopic aggregate , a behavior similar to bose - einstein condensation .
the theoretical results have been compared with a new series of computer simulations and show an excellent agreement , because in the new treatment we have eliminated the finite size effect that previously led to discrepancies between theory and simulation .
this peculiar system , which at best can be considered to be an extremely idealized version of spherical micelles , exposes an unusual behavior in the bmcsl - approach .
the equilibrium state of the system is one which does not satisfies the stationarity equations but is determined by the boundary conditions on the problem , i.e. a normalizable particle size - distribution .
apart from that the difference between both approximations is minor and in order to determine which one is better , simulations would be required with several orders of magnitude more particles .
r.b . acknowledges the financial support of the eu through the marie curie individual fellowship program ( contract no .
hpmf - ct-1999 - 00100 ) .
j.a.c . acknowledges financial support of ministerio de ciencia y tecnologa ( spain ) through the project no .
bfm2000 - 0004 . | in a previous paper ( j. zhang _ et al . _ , j. chem .
phys . * 110 * , 5318 ( 1999 ) ) we introduced a model for polydisperse hard sphere mixtures that is able to adjust its particle - size distribution . here
we give the explanation of the questions that arose in the previous description and present a consistent theory of the phase transition in this system , based on the percus - yevick equation of state .
the transition is continuous , and like bose - einstein condensation a macroscopic aggregate is formed due to the microscopic interactions .
a bmcsl - like treatment leads to the same conclusion with slightly more accurate predictions . 2 |
at present , neutrino oscillations , together with cosmological evidence for dark matter and dark energy , present the only evidence for physics beyond the standard model ( sm ) .
thus , models accommodating neutrino masses become an important component in the search for new physics and it is therefore fundamental to seek experimental answers to questions such as whether neutrinos are dirac or majorana fermions . in this context ,
one of the most promising types of experiments is that of searching for neutrinoless double beta decay ( @xmath0 decay ) , in which a peak at the endpoint of the beta radiation energy is searched for , see @xcite for reviews .
since this process is lepton number violating , its observation would imply that neutrinos are majorana fermions @xcite .
present bounds on @xmath0 decay for different nuclei have been set by the heidelberg - moscow @xcite , igex @xcite , cuoricino @xcite , nemo @xcite , elegant @xcite , solotvina @xcite and dama @xcite collaborations , as well as geochemical measurements @xcite . in the next few years , experiments like gerda @xcite , exo @xcite , sno+
@xcite and cuore @xcite will search for a @xmath0 decay signal with unprecedented sensitivity .
furthermore , there are several other experiments proposed for the upcoming future , which include candles @xcite , super - nemo @xcite , majorana @xcite , next @xcite , carvel @xcite , cobra @xcite , dcba @xcite , moon @xcite and xmass @xcite .
apart from the ongoing searches , there is a claim for a @xmath0 decay signal from ( part of ) the heidelberg - moscow collaboration @xcite .
however , this measurement is in apparent conflict with other bounds on neutrino masses , in particular those coming from the influence of neutrino masses in cosmology @xcite . among the more popular models for neutrino masses , we find the different types of seesaw mechanisms . in the type - i
seesaw @xcite , the sm is extended by the addition of extra fermion singlets .
the majorana masses of such singlets do not violate gauge invariance and are presently not constrained .
however , a very heavy mass is usually assumed and a yukawa coupling between the singlets and the sm neutrino fields of the same order of magnitude as the sm yukawas is introduced .
such a scenario provides a natural realization of small neutrino masses even if other possibilities , _
e.g. _ , approximate lepton number conservation @xcite , exist . in the current literature on @xmath0 decay , it is common to study the effects of the sm neutrinos or the extra states introduced to account for their masses independently ( see , _
e.g. _ , refs .
@xcite ) . however , the interplay between both contributions , when combined , displays interesting phenomenology such as cancellations in certain regimes that is otherwise lost . by considering these contributions independent , the presence of some extra component to the majorana neutrino mass different from the extra states and some degree of fine tuning is implicitly assumed . in the present work
, we will compute the nuclear matrix element ( nme ) involved in the @xmath0 decay rate without any assumption on the neutrino mass mediating the process , describing in detail the approximations involved .
the uncertainties associated to each of these approximations are also discussed , so as to estimate the total uncertainty on the final nme .
the results of this computation are publicly available at ref .
we will also discuss in detail the interplay between the contributions of the sm neutrinos and the extra states in seesaw models and discuss under which conditions they can be considered independent . in doing
so we will cover the full parameter space and deduce what implications can actually be inferred on the models from observations .
finally , we will also comment on how the contribution of extra states with different mass scales could allow for a large @xmath0 decay rate even in presence of more stringent bounds on neutrino masses , such as those derived from cosmology .
in particular , it will be shown that the heidelberg - moscow claim requires a tuning of only about 50 % amongst the extra contributions in order to be compatible with the present cosmology bounds .
the remainder of this work is organized as follows : in sec .
[ sec : nmecomp ] , we review the physics and assumptions used to compute the nmes and perform these calculations in order to get their values as functions of the mass of the exchanged fermions .
next , in sec .
[ sec : pheno ] , we discuss the general phenomenology of the @xmath0 process and the approximations that are usually made in @xmath0 decay analyses .
the realizations of the @xmath0 decay signal in the different types of seesaw models are treated in sec .
[ sec : models ] before we summarize and give our conclusions in sec .
[ sec : summary ] .
the @xmath0 decay is mediated by the weak hamiltonian : @xmath2 where @xmath3 is the leptonic current , which consists of the electron coupled to left handed electron neutrino .
the neutrino can be written as a linear combination of the light and heavy mass eigenstates @xmath4 given by the mixing matrix @xmath5 : @xmath6 on the other hand , the hadronic ( nuclear ) current @xmath7 can be obtained phenomenologically by imposing symmetry requirements to the more general combination that can be built with the available lorentz vectors @xmath8 , @xmath9 and @xmath10 , the neutron and proton four - momenta and the spin matrices , respectively .
we also need to assume the impulse approximation , _
i.e. _ , that nucleons in nuclei can be treated as free when dealing with the weak interaction . then imposing lorentz , parity and time - reversal invariance
the nuclear current is given by and @xmath11 , for which there is no experimental evidence , will be ignored . ]
@xmath12\psi , \label{eq : j_mu}\ ] ] where @xmath13 is the transferred momentum from hadrons to leptons , @xmath14 is the nucleon mass , @xmath15 represents a nucleon field and @xmath16 is the isospin lowering operator , _
i.e. _ , it turns a neutron into a proton .
the form factors , @xmath17 , @xmath18 , @xmath19 and @xmath20 , are real functions of the lorentz scalar @xmath21 .
their values at zero - momentum transfer are known as the vector , magnetic , axial and pseudoscalar coupling constants , respectively .
note that in single @xmath22 and two - neutrino @xmath23 decays only the vector and axial terms are usually considered , due to the small transferred momenta ( @xmath24 mev ) .
the magnetic and pseudoscalar couplings can be written in terms of the vector and axial ones by assuming the conserved vector current ( cvc ) and the partially conserved axial current ( pcac ) hypotheses @xcite .
the cvc hypothesis also implies that @xmath25 in the nuclear medium .
we now take the non relativistic approximation to the hadronic current . if terms are kept up to @xmath26 ( @xmath27 mev as will be discussed below ) and the energy transfer between nucleons is neglected ( @xmath28 ) , we are left with , being proportional to @xmath29 instead of @xmath30 .
however , their leading contribution is suppressed an extra order of magnitude because of their odd - parity character , which requires electron @xmath31-waves . hence these terms will be neglected .
their contribution will be discussed when referring to the electron @xmath32-wave approximation . ]
@xmath33 \delta(\mathbf{x}-\mathbf{r}_n ) , \label{eq : j_norel}\ ] ] where @xmath34 hence , we have a sum over all @xmath35 nucleons of the nucleus , whose coordinates are denoted by @xmath36 .
note that nucleon operators present in the nucleon fields from now on will be included in the nuclear wavefunctions . the momentum dependence of the couplings is usually parametrized by the standard dipolar form @xcite , and takes into account that nucleons are not point particles but finite size bodies , _
i.e. _ , the nucleon structure . since in @xmath0 decay the neutrino is being exchanged in t - channel and the outgoing electrons have essentially the same energy , the energy exchange can be neglected and thus @xmath37 .
then , the form factors look like @xmath38 where @xmath39 is the pion mass and @xmath40 and @xmath41 denote the proton and neutron anomalous magnetic moments , respectively .
the values of the cutoffs of the vector and axial nucleon form factors , @xmath42 gev and @xmath43 gev , are taken from experimental observations @xcite .
their effect is to weaken the couplings for large momentum transfers , _ i.e. _ , they avoid contributions arising from nucleons being too close to one another .
these form factors are commonly denoted as the finite nuclear size ( fns ) terms . with this hamiltonian
the rate of the @xmath0 decay can be calculated by using the second order fermi s golden rule @xcite : @xmath44 where the transition amplitude is given by @xmath45 the operator @xmath46 is included to fulfill antisymmetry for the electrons , whose energies are denoted by @xmath47 , @xmath48 .
the energy of the virtual neutrino @xmath49 is denoted by @xmath50 , and @xmath51 is the energy of the virtual intermediate nuclear state @xmath52 . the leptonic part of the numerator in eq .
( [ eq : ronu ] ) can be written more explicitly as @xmath53 this term turns out to be proportional to the neutrino masses @xmath54 because of the left handed character of both leptonic currents .
the transition amplitude is then @xmath55 the axial coupling @xmath56 , the electron mass @xmath57 and the nuclear radius @xmath58 have been introduced for convenience and to make the second line in eq .
( [ eq : ronu_2 ] ) dimensionless .
for the nuclear radius we have taken @xmath59 fm . using energy conservation
, we have also rewritten the denominator using the new parameter @xmath60 which gives the relative energy of the ( virtual ) state of the intermediate nucleus with respect to the mean energy of the initial and final states .
typical values of this parameter for the different decays are @xmath61 mev @xcite . in the following , two approximations will be made : * closure approximation .
* approximation of @xmath62 final states and electrons emitted in @xmath32-wave .
the first of these takes advantage of the high momentum of the virtual neutrino @xmath63 mev .
this nuclear scale comes from the integral over the transferred momentum of eq .
( [ eq : ronu_2 ] ) . for light neutrinos ,
the integrand of this equation is approximately proportional to @xmath64 in radial coordinates .
hence , momenta @xmath65 10 mev will be disfavoured in the transition .
in addition , if we recall the form of the fns terms in eq .
( [ eq : g_q ] ) , we see that transferred momenta above the cutoffs @xmath66 gev will be suppressed as well . on the other hand ,
the leading contribution of the exponential term will arise when @xmath67 .
since nucleons are typically few fermis ( femtometers ) apart in nuclei , the nuclear wavefunctions will select the preferred momentum for the virtual neutrino to be @xmath63 mev .
thus , we would expect it to be the typical virtual neutrino momentum of the decay .
this result has been confirmed by explicit calculation @xcite . in the case of heavy neutrinos , the transition operator will now have stronger preference for larger momenta @xmath68 , since @xmath69 appears in the denominator in eq .
( [ eq : ronu_2 ] ) .
therefore , the tendency of the nuclear interaction for @xmath63 mev can be overcome resulting in large transferred momenta that would imply internucleonic distances much shorter than @xmath70 fm .
again , such a distance is very suppressed by taking the fns terms into account , _
i.e. _ , the nucleon structure information . hence , in these cases
, the reduction due to the fns effects is very large so that we end up with a @xmath71 value of a few hundreds of mev at most , which is also the expected value for a process taking place between nucleons in nuclei . in any case
, the term @xmath72 , which can amount up to a couple of mev and vanishes on average , can be safely neglected .
moreover , the intermediate state energies @xmath73 , which can differ from one another by a few mev , can also be replaced by an average value @xmath74 .
thus , only a common parameter @xmath75 is required @xcite . with the removal of the dependence on the actual energy of the intermediate states
, it follows that they are no longer needed in the calculation , since the closure relation can be applied .
thus , the hadronic part in eq .
( [ eq : ronu_2 ] ) is now @xmath76 this closure approximation has been shown to be correct to more than 90 % , using the quasiparticle random phase approximation method @xcite , to be presented in sec .
[ sec : wf ] .
as for the limitation of our study to transitions to @xmath77 final states , and to cases where electrons are emitted in @xmath32-wave , corrections are expected to be of the order of 1 % at most . in the case of @xmath31-waves , they are suppressed to @xmath32-waves by an order of magnitude at least @xcite . moreover , since they have odd - parity , they need odd - parity terms in the current to couple to a @xmath78 final state and these only appear at @xmath79 , see note on eq .
( [ eq : j_norel ] ) .
thus , in the end we have a contribution of the order of 1 % of the leading ones .
on the other hand , since all the final nuclei of @xmath0 decay have even number of both protons and neutrons , their ground states are always @xmath77 .
any transition to an excited final state will be suppressed by a phase space factor , which in this case is approximately proportional to @xmath80 , where @xmath81 is the energy available for the decay . considering this factor ,
the only other low - lying final states of interest are @xmath82 excited states .
moreover , apart from the phase space suppression , these final states also need electron @xmath31-waves due to angular momentum coupling , since two electron @xmath32-waves can only couple to angular momentum 0 or 1 . hence , these transitions can also be safely neglected . within these approximations
, the transition amplitude can be written as @xmath83 where the index contractions and the @xmath84 and @xmath85 integrations have been performed .
the operator @xmath86 is the result of the product of the nuclear currents given in eqs .
( [ eq : j_norel ] ) and ( [ eq : jota ] ) , and has fermi , gamow - teller and tensor contributions : @xmath87 with the tensor operator @xmath88 .
the functions @xmath89 can be labeled according to the terms of the hadronic current [ see eq .
( [ eq : j_mu ] ) ] from which they originate : @xmath90 their explicit form can be found in ref . @xcite .
it was in this work that the importance of the non leading terms [ _ i.e. _ , all but @xmath91 and @xmath92 in eq .
( [ eq : h_xx ] ) ] was first shown .
they are referred to as higher order components of the nuclear current ( hoc ) .
notice that , since these terms are of orders @xmath93 and @xmath94 in the current , and @xmath95 come at order @xmath94 .
the reason to keep these second order terms is the enhancement of the coupling constants @xmath96 and @xmath97 due to the factors @xmath98 and @xmath99 , respectively , see eq .
( [ eq : g_q ] ) . in each of these terms the corresponding factor appears squared . ]
their contribution will be enhanced for larger transferred momentum , _ i.e. _ , for heavy neutrinos .
it is easy to insert @xmath100 into eq .
( [ eq : anch_onu ] ) and derive an expression for the @xmath0 decay rate : @xmath101 here , @xmath102 is a well known kinematic factor , and comes essentially from the leptonic degrees of freedom .
it can be written explicitly as @xmath103^{4}m_{e}^{2}}{64\pi^{5 } r^{2 } \ln2}\int f_{0}(z,\varepsilon_{1})f_{0}(z,\varepsilon_{2 } ) q_{1}q_{2}\varepsilon_{1}\varepsilon_{2}\delta(\varepsilon_{1}+\varepsilon_{2}+e_{f}-e_{i } ) d\varepsilon_{1}d\varepsilon_{2}d\left(\mathbf{\hat{q}_{1}\cdot\hat{q}_{2}}\right ) , \label{eq : g01}\end{aligned}\ ] ] where @xmath104 are the so - called fermi functions , with @xmath105 the proton number and @xmath106 the electron momenta .
the fermi functions also depend on the nuclear radius @xmath58 and their explicit form is @xmath107 the quantity @xmath108 is the nuclear matrix element , which takes into account the initial and final nuclear wavefunctions and the transition operator .
this operator originates from both the nuclear currents and the virtual neutrino .
the nme is given by @xmath109 when the integral over @xmath30 is performed , we obtain @xmath110 where @xmath111 is the distance between the decaying neutrons and the @xmath112 are the so - called neutrino potentials . before the radial integration over @xmath71 they are given by @xmath113 where @xmath114 are the spherical bessel functions .
as can be seen from eq .
( [ eq : nme ] ) , a key ingredient in the calculation of the nmes are the wavefunctions of the initial and final nuclei .
this is a complicated nuclear structure problem which can not be solved in the complete space , _ i.e. _ , taking into account all neutrons and protons of the corresponding nucleus in all their possible configurations .
thus , truncated valence spaces and effective interactions are used to solve the nuclear many body problem . as a consequence of this
, a fully consistent treatment would demand regularizing the @xmath0 decay operator in eq .
( [ eq : nme ] ) using the same prescription as for the bare nuclear interaction @xcite .
this has only been performed very recently @xcite . instead
, we will simplify the problem keeping the bare @xmath0 decay operator and including new correlations [ called short range correlations ( src ) ] in the calculation via a general prescription .
the findings of refs .
@xcite and @xcite show that for light neutrinos , the effect of these correlations is rather moderate once fns terms have been taken into account , of the order of 5% correction to the nme . in order to implement these
src one needs to assume some prescription and the most commonly used are either a jastrow - type function @xcite or a unitary correlation operator method ( ucom ) transformation @xcite .
the nme is thus transformed as @xmath115 where @xmath5 is either a jastrow - type function or a ucom transformation .
the actual parametrizations can be found in refs .
@xcite and @xcite , respectively
. in our calculations have used the ucom prescription , even though similar results are expected within the jastrow approach . note that these src terms , which do not have much importance for light neutrinos , will be more relevant for heavy ones , which require shorter distances between the decaying nucleons . mainly two different methods are used to obtain the nmes for the @xmath0 decay , the quasiparticle random phase approximation ( qrpa ) @xcite and the interacting shell model ( ism ) @xcite .
the qrpa includes relatively large valence spaces but is not able to comprise all the possible configurations . on the other hand ,
the ism is limited to smaller configuration spaces , but all possible correlations within the space can be included .
when comparing the nmes obtained by both methods a clear disagreement is found , the ism values being about 1.5 - 2 times smaller than qrpa ones @xcite .
this disagreement is not overcome when taking into account the estimated errors of both calculations and it is under discussion whether the difference is mainly due to the lack of correlations of qrpa calculations ( which reduce the value of the nme ) , the small valence space used in the ism ( to be discussed in section [ sec : error ] ) or both @xcite .
however , it must be stressed that due to the theoretical effort made over the last years this disagreement is now much less severe than it was five years ago , and studies using recently available experimental information for the decay of @xmath116ge suggest that the present situation can be improved @xcite .
this applies to nmes obtained with light neutrino exchange . in the heavy neutrino case ,
since the available qrpa results are rather outdated @xcite , it is probably not meaningful to compare them with those of the present work .
however , the same relation between ism and qrpa results of the light neutrino case is to be expected , since the difference between these methods lies on the calculation of the wavefunctions , and therefore it should not be very much dependent on changes on the transition operator , for whom both methods give an equivalent description @xcite . in this work
, we have used ism nuclear wavefunctions . following the considerations of sec .
[ sec : nme ] , we have performed calculations for the @xmath54 dependent @xmath0 decay nmes of the emitters @xmath117ca , @xmath116ge , @xmath118se , @xmath119sn , @xmath120te and @xmath121xe , using the ism coupled code nathan @xcite , ideally adapted for the calculation of @xmath77 states .
full diagonalizations are accomplished within different valence spaces and effective interactions .
for instance , the decay of @xmath117ca is studied in the @xmath122 major shell , where the kb3 interaction @xcite is employed . for the case of @xmath116ge and @xmath118se ,
the valence space consisting on the 1@xmath123 , 0@xmath124 , 1@xmath125 and 0@xmath126 orbits is diagonalized using the gcn28.50 interaction @xcite .
finally the 0@xmath127 , 1@xmath128 , 1@xmath129 , 2@xmath130 and 0@xmath131 valence space and the gcn50.82 interaction @xcite are used in the decays of @xmath119sn , @xmath120te and @xmath121xe .
the results of the computation of the nme as a function of the neutrino mass are shown in fig .
[ fig : nme ] , which is in agreement with the findings of refs .
@xcite and @xcite .
as can be seen from this figure , the dependence of the matrix element on the nuclei is mild and all the curves show a similar behaviour .
we will discuss qualitatively the observed dependence of the nme on the neutrino mass as well as the phenomenology associated to the different mass regimes in the next sections .
, scaledwidth=70.0% ] as we have pointed out several times throughout the present section , since the computation of the nmes can be rather challenging , a number of approximations need to be made . in consequence
, the results obtained will have some uncertainties , which we will now estimate .
first of all , when dealing with the ism , we have to worry about the valence space and the effective interaction used to obtain the nuclear wavefunctions .
the effect of having larger valence spaces was analyzed in ref .
@xcite , with the result that the nmes increased , within a conservative analysis , by @xmath132 .
the same number was obtained in a qrpa calculation when it was quantified the effect of the orbits absent in a ism valence space @xcite . of the three valence spaces employed in the present work ( see sec . [
sec : wf ] ) , only that corresponding to the @xmath117ca decay was not explored in ref .
@xcite , so that we can consider the former @xmath132 increase as a general estimate of the uncertainty due to the configuration space . as for the nuclear interaction , a moderate dependence of @xmath133 was found both in refs .
@xcite and @xcite .
since they study nuclei in different regions and we can consider all the effective interactions employed of similar accuracy , we will also take this figure as general . until recently ,
another considerable source of uncertainty were the src .
however , it now seems that their contribution is rather small and that proper ucom or jastrow - type parametrizations can take these terms very well into account , with a precision of @xmath134 @xcite . for its part ,
the variation in the nme due to different but reasonable values of the cutoffs appearing in the fns terms @xcite is also very small , less than 5 % .
in addition , due to the fact that the nmes have been calculated using the closure approximation , we have to include an additional error of @xmath135 to our results , as suggested by qrpa calculations @xcite .
this is in agreement with the very soft variation that is seen in the nme as the parameter @xmath136 is modified .
furthermore , we quantify the possible effect of missing terms in the operator ( next order terms in the current , nucleon recoil , @xmath31-wave emitted electrons ) by an additional uncertainty of less than 5 % in our results .
it is currently under discussion whether the axial coupling should be quenched or not @xcite . in the ism calculations presented here we take @xmath137 , _
i.e. _ , we do not quench it , contrary to what is required by the single @xmath22 and two - neutrino @xmath23 decays , where a pure gamow - teller operator appears and its value has to be quenched to @xmath138 .
however , in the @xmath0 decay case the operator is more involved due to the extra radial dependence introduced by the virtual neutrino . moreover , in the case of the pure gamow - teller @xmath139 channel , it is not dominant in the @xmath0 process , and depending on its relative sign , quenching it may result even in an enhancement of the nme .
until this issue is explored in more detail , we will take the most conservative option , allowing for a full quenching of the axial coupling and also for the quenching only of the @xmath139 channel . under these assumptions , taking into account that the fermi part of the nme accounts for 10 - 15% of the full nme in our ism calculations and is never quenched , we estimate a @xmath140 error due to this effect .
notice that the uncertainty in the valence space only moves the estimate up , the effect of axial quenching essentially moves it down , while the remaining contributions are expected to be gaussian - distributed . even though some of these errors may be correlated in a rather complicated way , as a first approximation we will take them as independent .
altogether , adding every contribution in quadrature we expect an overall uncertainty in the final nme of @xmath141 .
the above analysis applies to the case of light neutrino exchange . for heavy neutrinos
, the nmes get very dependent on the src and fns treatments . in this case , a similar study to the one above gives a 15 - 20% for src uncertainties and
10% for fns ones , which would lead to a final @xmath142 uncertainty .
however , one should take this number cautiously until the accuracy of the fns approach for such heavy exchanged particles is firmly established @xcite .
as an example , we will consider the case of the @xmath116ge decay . in ref .
@xcite the nme was obtained with different effective interactions and src , obtaining the interval @xmath143 .
if we take into account the further uncertainties of the valence space , the fns , the closure approximation , the next order hadronic current terms and the @xmath144 quenching , we end up with @xmath145 .
this result will be used in sec .
[ sec : models ] to derive bounds on the neutrino masses from the @xmath0 decay process .
according to eq . ( [ eq : t-1_rep ] ) , the contribution of a single neutrino to the amplitude of @xmath0 decay is given by @xmath146 where @xmath147 is the mass of the propagating neutrino and @xmath148 is the nuclear matrix element that characterizes the process and depends on the nucleus that undergoes the @xmath0 transition .
figure [ fig : nme ] shows two distinct regions where the behaviour of the nme as a function of the neutrino mass changes from almost constant up to @xmath149 mev to decreasing quadratically as the neutrino mass increases beyond @xmath150 mev .
this behaviour is easily understood : the neutrino can be characterized as _ light _ if @xmath151 or _ heavy _ if @xmath152 , which would mean that the neutrino propagator in the nme would be dominated by @xmath21 or @xmath153 , respectively , where @xmath31 is the momentum exchanged in the process . as already mentioned in sec .
[ sec : nme ] , in the @xmath0 decay @xmath154 .
we will therefore define two regimes : * * the light neutrino regime : * for @xmath155 mev , where the neutrino propagator is @xmath156 and hence , the nme is maximum in this regime and is almost independent of the neutrino mass : @xmath157 .
\label{eq : lightexp}\ ] ] * * the heavy neutrino regime : * for @xmath158 mev where the nme decreases as @xmath159 providing an extra suppression to its contribution to the @xmath0 decay amplitude because of the neutrino propagator : @xmath160 in principle , one could expect a more involved behaviour or even a resonance if @xmath161 .
however , since the @xmath0 transition does not occur through an @xmath32-channel type diagram , the characteristic momentum transfer has @xmath162 with @xmath163 .
thus , the nuclear matrix element does not exhibit a resonant mass and the transition between the light and heavy neutrino mass regimes is relatively smooth ( _ c.f .
_ , fig .
[ fig : nme ] ) .
we will therefore not define a third transition region between the heavy and light regimes , as is sometimes done in the literature , since there is no new phenomenology associated to it .
the usual bound derived from the @xmath0 process in the literature is obtained summing over the active neutrinos , implicitly neglecting the contribution of extra degrees of freedom . with this assumption
the only contribution comes from the neutrinos in the light regime , with @xmath164 , which results in @xmath165 where @xmath166 is the well - known expression used for the `` effective @xmath0 decay neutrino mass '' : @xmath167 where @xmath147 are the masses of the neutrino mass eigenstates , @xmath168 , @xmath169 , @xmath170 are the neutrino mixing angles and @xmath171 are combinations of the majorana and dirac phases .
it is important to note that this expression holds only when the sm neutrinos dominate the @xmath0 process .
however , if the sm is not extended , the majorana mass required for the @xmath0 transition is forbidden .
as we will discuss , all extensions of the sm that induce a majorana mass for the sm neutrinos imply the inclusion of extra degrees of freedom that can contribute to the @xmath0 process and should be added to the sm decay amplitude .
allowed regions by neutrino oscillation data at @xmath172 for @xmath173 as a function of the mass of the lightest neutrino @xmath174 , for normal ( blue ) and inverted ( red ) neutrino mass hierarchies .
the cosmologically disfavoured neutrino masses from cosmology are also shown ( green band ) as well as the heidelberg - moscow claim for @xmath0 decay ( yellow band).,scaledwidth=70.0% ] under the assumption that only the light neutrinos give a significant contribution to this process , eq .
( [ eq : mbb ] ) can be combined with the present constraints on neutrino masses and mixings @xcite to derive fig .
[ fig : nh - ih ] , where the allowed value of @xmath173 as a function of the mass of the lightest neutrino @xmath174 is shown . however , as will be discussed in sec .
[ sec : models ] , this assumption is not always valid and fig .
[ fig : nh - ih ] does not always provide an accurate description of the @xmath0 decay . in the same figure we also show the heidelberg - moscow claim for @xmath0 decay and the bound on the mass of the lightest neutrino from cosmology arguments @xcite .
notice that the bounds from cosmology apply to the sm active neutrinos only . for extra sterile neutrinos
, the cosmology bounds would depend on their mixing with the active ones , which would determine their abundance .
indeed , sterile neutrinos with masses @xmath175 kev larger than the cosmology bound are actually considered good candidates for warm dark matter @xcite . as can be seen from the figure
, there is a tension between the heidelberg - moscow claim , the contribution to @xmath0 from sm neutrinos and the present bounds on their mass from cosmology . in secs . [
sec : heavylight ] and [ sec : mix ] we will discuss possible solutions to this tension .
in this section we analyze the contributions to @xmath0 decay of the different mechanisms that lead to majorana neutrino masses and can therefore induce the required lepton number violation for the process .
we will discuss here the tree - level realizations of the weinberg @xmath176 effective operator : @xmath177 here , @xmath178 denotes the sm higgs field , which breaks the electroweak ( ew ) symmetry after acquiring its vacuum expectation value ( vev ) @xmath179 , @xmath180 is the scale of new physics that gives rise to the operator and we have used the definition @xmath181 .
this is the only @xmath176 operator that can be built from the sm particle content respecting both gauge and lorentz invariance @xcite .
since the low - energy effects of physics beyond the sm can be encoded in an expansion of effective operators of @xmath182 suppressed by inverse powers of @xmath183 , it is a natural expectation that this sole @xmath176 operator will be the least suppressed one .
it is then very suggestive that one of the few evidences we have for physics beyond the sm is the existence of small , but non vanishing , neutrino masses . indeed , after the higgs field develops its vev , the operator of eq .
( [ eq : d=5 ] ) induces a majorana mass term for the sm neutrinos @xmath184 , suppressed by the scale @xmath180 .
there are three different extensions of the sm particle content that lead to the operator of eq .
( [ eq : d=5 ] ) after the extra mediators have been integrated out .
they are known as `` seesaw '' mechanisms of type - i @xcite , where the heavy particles are fermion singlets , type - ii @xcite , where scalars triplets are included , and type - iii @xcite , where the sm is extended by fermion triplets .
all these extra degrees of freedom , required to induce the majorana nature of the sm neutrinos , can also contribute to the @xmath0 process .
the contribution of neutral fermions , such as the singlets and triplets added in the type - i and iii seesaws , to the @xmath0 decay rate is depicted on the left side of fig .
[ fig : diag ] ; in particular this includes that of the light active neutrinos . the contribution of the scalar triplet of the type - ii seesaw ( see , _ e.g. _ , ref .
@xcite ) is depicted on the right side of fig .
[ fig : diag ] . in both diagrams the @xmath185 lines can also be exchanged for the physical singly - charged scalar present in the type - ii seesaw . in principle , the contributions of the light active neutrinos and those of the extra degrees of freedom that are introduced should be combined .
this is especially so in the case of the type - i seesaw , since important cancellations @xcite are present in certain regimes @xcite that are missed if the constraints are placed separately .
if they are assumed to be independent , an extra contribution to the neutrino mass beyond those extra states is implicitly assumed .
moreover , taking into account the relations between the high- and low - energy parameters allows the derivation of stronger bounds on the former through the active neutrino contribution .
it is then important not to neglect it , since the naive constraints stemming directly from the contribution of the extra degrees of freedom are generally much weaker .
( 160,110)(0,0 ) ( 8,10)[r]@xmath186 ( 152,10)[l]@xmath31 ( 10,10)(150,10 ) ( 8,100)[r]@xmath186 ( 152,100)[l]@xmath31 ( 10,100)(150,100 ) ( 78,27)[br]@xmath185 ( 60,10)(100,40)34 ( 78,83)[tr]@xmath185 ( 60,100)(100,70)-34 ( 98,55)[r]@xmath187 ( 100,40)(100,70 ) ( 152,40)[l]@xmath188 ( 100,40)(150,40 ) ( 152,70)[l]@xmath188 ( 100,70)(150,70 ) ( 160,110)(0,0 ) ( 8,10)[r]@xmath186 ( 152,10)[l]@xmath31 ( 10,10)(150,10 ) ( 8,100)[r]@xmath186 ( 152,100)[l]@xmath31 ( 10,100)(150,100 ) ( 53,35)[br]@xmath185 ( 40,10)(70,55)34 ( 53,75)[tr]@xmath185 ( 40,100)(70,55)-34 ( 90,57)[b]@xmath189 ( 70,55)(110,55)5 ( 125,55)(15,90,270 ) ( 152,40)[l]@xmath188 ( 125,40)(150,40 ) ( 152,70)[l]@xmath188 ( 125,70)(150,70 ) in this section we will discuss the phenomenology of @xmath0 decay when extending the standard model with fermion gauge singlets , _
i.e. _ , right handed neutrinos @xmath190 .
the standard model lagrangian is then extended as @xmath191 after the higgs develops its vev the neutrino mass matrix is @xmath192 this mass matrix can be diagonalized by a unitary mixing matrix @xmath5 : @xmath193 we therefore have @xmath186 neutrino mass eigenstates with masses @xmath147 and mixings @xmath194 with the electron . out of these , at least three mass eigenstates must be very light and form the main components of the active neutrinos , whose number is measured by the invisible decay width of the @xmath105 @xcite . on the other hand , the masses of the extra states are not determined .
we thus have two contributions to the amplitude of @xmath0 decay , one from the light active neutrinos and another from the extra degrees of freedom : @xmath195 where we have used capital letters to denote the mass index of the mostly sterile states and lowercase letters for that of the mostly active states .
depending on whether the extra mass eigenstates fall in the light or heavy neutrino mass regimes we can further split their respective contributions to the amplitude : @xmath196 we can now distinguish three cases exhibiting very different phenomenologies depending on the mass regime of the extra mass eigenstates : in this scenario all the mass eigenstates are lighter than @xmath150 mev . in principle
, this does not allow to explain the smallness of neutrino masses through the naive seesaw mechanism , _
i.e. _ , with @xmath197 yukawa couplings . however , since the value of the parameter @xmath198 in eq .
( [ eq : seesawi ] ) is not restricted and it is technically natural for it to be small ( for vanishing majorana mass term the @xmath199 symmetry is recovered ) , we believe this is a possibility worth exploring even if less appealing than the canonical type - i seesaw scenario . ca decay , but the same quadratic dependence is seen for the remaining nuclei studied .
the value of the parameter is @xmath200 mev@xmath201 .
[ fig : gim],scaledwidth=70.0% ] notice that , if all neutrinos belong the the light regime , eq .
( [ diag ] ) implies @xmath202 since the left - left entry of the mass matrix in eq .
( [ eq : massmat ] ) vanishes ( it is forbidden by the sm gauge symmetry ) .
thus @xmath203 indeed , for neutrino masses in the light regime , the nuclear matrix elements are basically independent of the neutrino mass ( see fig .
[ fig : nme ] ) @xmath204 and therefore the rate of @xmath0 decay in eq .
( [ eq : zero ] ) is very suppressed when all mass eigenstates are lighter than @xmath205 mev .
this works in a way similar to the gim suppression in flavour violating processes @xcite .
indeed , a process converting from flavour @xmath206 to flavour @xmath22 is gim suppressed due to the unitarity relation @xmath207 , so that only the mass of the propagating particle does not make the cancellation exact and a suppression of @xmath208 is found , where @xmath209 is the mass squared difference between the different propagating particles . in our @xmath0 process
we have eq . ( [ eq : gimii ] ) operating a similar cancellation and again it is the different neutrino masses in the nme of eq .
( [ eq : zero ] ) that would prevent a full cancellation leading to a suppression of @xmath210 with @xmath163 [ see the second term in eq .
( [ eq : lightexp ] ) and , _
e.g. _ , ref .
the @xmath210 dependence of the @xmath211 term that drives the contribution to @xmath0 in eq .
( [ eq : zero ] ) is depicted in fig .
[ fig : gim ] .
as expected , the two contributions cancel up to a factor @xmath212 with @xmath163 and deviations from this behaviour start to be non negligible for @xmath213 mev . decay in @xmath120te , with a 90 % cl half - life @xcite .
we have assumed the extra neutrinos are non hierarchical and show the bounds as a function of their common mass .
we compare the case in which the contribution from the sm neutrinos is properly taken into account ( striped area ) to that in which only the extra contribution is considered ( above the red line ) .
[ fig : u_bounds],scaledwidth=70.0% ] we want to emphasize that phenomenological analyses ( see , _
e.g. _ , refs .
@xcite and @xcite ) that use the non observation of @xmath0 decay to derive bounds on the mixing of an extra light neutrino with mass around @xmath205 mev or below neglecting the contribution of sm neutrinos , implicitly assume that the cancellation described in eq .
( [ eq : zero ] ) does not take place .
this would be the case if some extra contribution to the neutrino masses is present and eq .
( [ eq : gimii ] ) is consequently modified .
examples of this situation , which imply some degree of cancellation between the different contributions to neutrino masses , will be discussed in secs .
[ sec : heavylight ] and [ sec : mix ] . in the absence of these extra contributions , the remaining leading term , gim suppressed as @xmath210 ,
can then be used to derive the corrected bound on the mixing of the extra state . in fig .
[ fig : u_bounds ] we show this bound using the constraints on @xmath0 decay in @xmath120te from cuoricino , with a 90 % cl half - life @xcite and compare it to the one that would be obtained neglecting this cancellation and considering only the extra neutrino contribution .
it is clearly seen that the bound indeed weakens when @xmath214 mev . in particular , for @xmath215 mev , the @xmath211 suppression in fig .
[ fig : gim ] becomes almost @xmath216 , so that the @xmath0 decay rate would be suppressed by six orders of magnitude and for @xmath214 kev the bound on the mixing from fig .
[ fig : u_bounds ] becomes meaningless since @xmath217 .
clearly , if the extra states that accommodate neutrino masses are all in this regime the @xmath0 decay becomes experimentally inaccessible even if neutrinos are majorana particles .
an important sub - case is when @xmath218 with three extra neutrinos , which corresponds to the case of dirac neutrinos . in this scenario ,
the cancellation is perfect since the left and right handed states are exactly degenerate , which corresponds to the fact that there is no lepton number violation for dirac neutrinos . in order to derive the constraints of fig .
[ fig : u_bounds ] we have assumed as an example that all the extra states are degenerate in mass ( or that there is only one extra state ) .
however , a similar gim - like cancellation would also be present when assuming a different hierarchy as long as all the extra states are in the light regime .
the extra contribution in the light regime can only dominate over the light active one and avoid the gim - like cancellation if the model is extended with other heavier states .
this situation also implies a certain degree of fine - tuning between the extra contributions and will be discussed in detail in secs .
[ sec : heavylight ] and [ sec : mix ] .
this is the simplest situation and the one implicitly assumed when using eq .
( [ eq : mbb ] ) or fig .
[ fig : nh - ih ] to forecast the rate of @xmath0 decay .
it also corresponds to the canonical implementation of the type - i seesaw mechanism , where the majorana mass of the right handed neutrinos in eq .
( [ eq : seesawi ] ) is assumed to be very large @xmath219 so that the smallness of the active neutrino masses is naturally explained and extra heavy eigenstates are obtained . since the extra states would fall in the heavy neutrino mass regime their nuclear matrix elements are very suppressed : @xmath220 ( see fig .
[ fig : nme ] ) .
furthermore , eq . ( [ diag ] ) implies that @xmath221 so that the @xmath0 decay amplitude is @xmath222 the contribution from the light active neutrinos thus dominates and fig .
[ fig : nh - ih ] provides an accurate prediction for the @xmath0 transition rate .
notice that , using the contribution of the extra states @xmath223 to the @xmath0 process in order to derive a bound on their mixing @xmath224 would lead to rather weak constraints since it is very subleading ( see , _ e.g. _ , refs .
@xcite ) . on the other hand , eq .
( [ eq : heavyfromlight ] ) can be used to express the dominant light neutrino contribution as a function of the heavy parameters : @xmath225 .
this expression then allows to derive a much stronger constraint on @xmath224 @xcite .
this constraint is also shown in fig .
[ fig : u_bounds ] when @xmath226 mev and compared to that when only the contribution of the heavy neutrinos is considered . in order to derive the constraints of fig .
[ fig : u_bounds ] we have assumed as an example that all the extra states are degenerate in mass ( or that there is only one extra state ) .
however , a similar behaviour is found when a different hierarchy is considered as long as all the extra states are in the heavy regime . in this scenario
we would have the full contribution to @xmath0 decay of eq .
( [ eq : typeiampl ] ) and the constraint from eq .
( [ diag ] ) : @xmath227 as discussed above , for the neutrinos in the heavy regime , the nme receives an extra suppression to their contribution to the @xmath0 decay rate and the leading terms stem from the light states : @xmath228 however , in this case the gim - like cancellation is prevented since the heavy contribution is suppressed .
this scenario thus offers the richest phenomenology .
in particular , it is possible to satisfy eq .
( [ eq : gimiii ] ) even in a situation where @xmath229 by canceling the contribution of the extra heavy states against that of the extra light ones while keeping the light neutrino masses small .
this implies a certain level of fine - tuning since extra sterile neutrinos in both the heavy and light regimes are necessary and some degree of cancellation between their respective contributions is required in order to keep the neutrino masses small .
however , in such a situation , the contribution of the light extra states could dominate over that of the active and induce a rate for the @xmath0 process larger than the one forecasted from fig .
[ fig : nh - ih ] .
this could thus be a possible solution to an eventual discrepancy between a positive result in @xmath0 decay and a negative result in the searches for neutrino masses in cosmology .
indeed , the bounds from cosmology apply to the active sm neutrinos only . as an example
, we will here consider the heidelberg - moscow claim for a positive @xmath0 decay signal @xcite .
the accommodation of this signal through only sm neutrinos [ see eq .
( [ eq : mbb ] ) ] would require @xmath230 at @xmath231 , where the allowed numbers have been obtained with the ism results of sec .
[ sec : error ] following the rather conservative procedure described in ref .
almost all the error bar comes from the theoretical error of the nme , which is much larger than the one associated to the experimental claim . as can be seen in fig .
[ fig : nh - ih ] , the interpretation of this claim as light active sm neutrinos is very disfavoured ( see , _
e.g. _ , ref .
@xcite ) by the constraints from cosmology and neutrino oscillation data .
however , this signal could be accommodated in a model with heavier neutrinos ( which are not bounded by cosmology ) mediating the process .
indeed , following eq .
( [ eq : tuning ] ) , we could reinterpret the result as @xmath232 this larger contribution to @xmath0 decay would not be in conflict with neutrino masses if the extra heavy and light neutrino contributions cancel each other in eq .
( [ eq : gimiii ] ) .
the level of cancellation required to accommodate the the heidelberg - moscow claim with sterile neutrinos is actually only at the @xmath233 level , since the light active neutrinos only fail to explain it by about a factor two , given the bounds we have on their masses from cosmology . in fig .
[ fig : tuning ] we show the degree of cancellation that would be required to accommodate an eventual stronger bound on the mass of the lightest neutrino @xmath174 from cosmology @xcite with a discovery of @xmath0 decay requiring a given @xmath234 when interpreted as the contribution of the sm neutrinos alone .
the contours for no tension between the mass bound and @xmath173 are shown together with the situations that require cancellations at the level of @xmath235 , @xmath236 , @xmath237 and @xmath238 .
the dot represents the present situation of the heidelberg - moscow claim that requires a mild @xmath235 cancellation to avoid conflict with the cosmological bounds on @xmath174 .
notice that , since a minimum size of @xmath239 ev is guaranteed for an inverted hierarchy , the required level of cancellation never exceeds the @xmath236 in this situation . on the other hand ,
if the neutrino mass hierarchy is found to be normal , tunings up to @xmath240 would be necessary to reconcile a discovery of @xmath234 on the same level as the present heidelberg - moscow claim with an eventual bound of @xmath241 ev .
decay with size @xmath234 with a bound on the lightest neutrino mass @xmath174 from future cosmology or @xmath22 decay experiments .
the solid and dashed lines are for normal and inverted mass hierarchy respectively .
the red , yellow , green , blue and black contours are for cancellations to the level of @xmath238 , @xmath237 , @xmath236 , @xmath235 and no cancellation respectively .
the dot represents the present heidelberg - moscow claim for @xmath234 and the cosmology bound on @xmath174 and requires a tuning of @xmath233 in order to be explained through sterile neutrinos .
[ fig : tuning],scaledwidth=70.0% ] in the type - ii seesaw , the standard model content is expanded by the addition of a scalar @xmath242 triplet with hypercharge @xmath243 ( where the hypercharge is defined such that @xmath244 ) : @xmath245 the scalar triplet couples to the lepton doublet through the yukawa terms : @xmath246 where @xmath247 is the second pauli matrix .
furthermore , the scalar triplet @xmath189 has a coupling @xmath136 to a pair of higgs fields , so that it gets an induced vev after ew symmetry breaking : @xmath248 , where @xmath249 is the mass of the scalar triplet .
the vev of the triplet then induces a majorana neutrino mass @xmath250 in eq .
( [ eq : seesawii ] ) .
this in turn implies that , at low energies , eq .
( [ eq : gimi ] ) is modified to @xmath251 making @xmath252 the analogous contribution to the one of the heavy neutrinos in the type - i seesaw .
the @xmath0 process can be mediated both by the neutrinos and by the scalar triplet ( see fig . [
fig : diag ] ) .
however , such charged scalars would have been produced at tevatron for masses below 100 gev @xcite .
thus , the contribution of the charged scalar is suppressed with respect to the neutrino one .
the diagram in the right side of fig .
[ fig : diag ] replaces the neutrino propagator by a scalar propagator and thus its amplitude is suppressed by a factor @xmath253 with respect to the sm neutrino contribution .
the diagram in the left side of fig .
[ fig : diag ] in which one of the @xmath185 bosons is replaced by the physical charged scalar are also suppressed .
the scalar is an admixture of the charged components of the higgs doublet and the scalar triplet and its coupling to the quarks is therefore proportional to the quark mass .
thus , the amplitude of these contributions turn out to be suppressed by a factor @xmath254 , where @xmath255 is the mass of either the up or down quark . therefore , in this scenario , as in the type - i seesaw with all extra states heavy , the light active neutrino contribution dominates and the usual description of @xmath0 decay in fig .
[ fig : nh - ih ] applies . in the type - iii
seesaw models the standard model is expanded by fermion @xmath242 triplets with zero hypercharge : @xmath256 the fermion triplets couple to the sm lepton doublets and the higgs field through the yukawa terms and have majorana mass terms of their own : @xmath257 the @xmath0 decay phenomenology of the type - iii seesaw is then completely analogous to that of the type - i with the neutral component of the triplet playing the role of the right handed neutrino , except that , since the triplet also has charged components , stringent lower bounds on its mass exist and in practice only the heavy mass eigenstate regime is available .
the situation then reduces to the one for the type - ii seesaw instead , _
i.e. _ , the same @xmath0 phenomenology applies , with the replacement @xmath258 it is interesting to note that the same phenomenology that could stem from a type - i seesaw with both heavy and light eigenstates can also arise from a type - ii or iii seesaw in combination with type - i sterile neutrinos in the light regime .
indeed , adding a type - ii or iii contribution to neutrino masses @xmath259 as in eqs .
( [ eq : massii ] ) and ( [ eq : massiii ] ) , eq .
( [ eq : massmat ] ) would instead read @xmath260 this in turn implies that eq .
( [ eq : gimii ] ) is modified to @xmath261 thus , it is possible to have a dominant contribution to @xmath0 decay from the extra light sterile neutrinos if @xmath262 , while eq .
( [ eq : nogim ] ) and the smallness of neutrino masses is respected by a cancellation between @xmath263 and @xmath264 .
the level of the cancellation required also corresponds to the one depicted in fig .
[ fig : tuning ] .
indeed , the amplitude of @xmath0 decay would now be @xmath265 and , as an example , the heidelberg - moscow claim can be interpreted as @xmath266 in this context .
we have discussed the general phenomenology of neutrinoless double beta decay ( @xmath0 decay ) in different types of seesaw models . in particular ,
we have focused on the contributions of the extra degrees of freedom in different mass regimes without assuming preference to a particular mass scale . in order to do this , we computed the nuclear matrix element ( nme ) involved in the decay amplitudes as function of the mass of the mediating field , detailing all the assumptions performed in each step and estimating the final error due to the approximations taken to be at most 30 % for light neutrinos and around 40 % for heavy neutrinos .
the results of this computation are publicly available at ref . @xcite . in particular
, the behaviour of the nme is found to be that which can be expected from the propagator @xmath267 , where @xmath268 mev@xmath269 is the typical momentum transfer between the nucleons .
thus , the nmes are essentially constant for @xmath270 mev ( light regime ) and decrease as @xmath271 for @xmath272 mev ( heavy regime ) .
the transition region around 100 mev is smooth and no significantly new phenomenology takes place at this regime .
in our discussion we have seen that , for the type - i seesaw , a number of possibilities exist . in the case where all the masses of the extra fermion singlets are in the heavy regime ,
the contribution of these states to @xmath0 decay is negligible and difficult to constrain directly .
however , the contribution from the light left handed neutrinos can be rewritten in terms of the masses and mixings of the heavy states , which results in quite stringent bounds on their mixings . on the contrary , if the extra fermion singlet states are instead in the light regime , a gim - like cancellation with the left handed neutrino contributions occurs .
in this situation , neutrinos are majorana particles , but the @xmath0 decay rate becomes unobservable , being suppressed by at least six orders of magnitude if the mass of the extra states is below 1 mev .
thus , considering the extra states alone and neglecting the sm neutrino contribution will result in bounds on their mixings which are generally too strong .
such bounds only apply when considering extra contributions to neutrino masses beyond the light extra states in order to prevent the cancellation from taking place .
dirac neutrinos , where the majorana mass term of the extra fermion states is zero , are a special case of the situation with only light extra states . in this scenario ,
the right and left handed states are exactly degenerate and the cancellation is exact .
if there are extra states in both the light and heavy regimes , then the main contribution to the @xmath0 transition could come from the light extra states , although some fine - tuning is necessary . as such , this could be a way to reconcile a large @xmath0 decay rate ( _ e.g. _ , the heidelberg - moscow claim ) with more stringent cosmological bounds with a mild cancellation of about 50 % .
as for the other types of seesaws , current bounds from accelerator experiments place the extra degrees of freedom in the heavy regime .
this effectively reduces the situation to that which appears for the type - i seesaw with only heavy extra states .
however , in mixed seesaw models , the situation can instead resemble that of the type - i with states in both regimes and thus be used to reconcile large @xmath0 decay rates with cosmological bounds . in conclusion , the contribution to @xmath0 decay from the light active neutrinos can be forecasted by combining present and future neutrino oscillation data on the neutrino mixing and mass hierarchy with probes of the absolute neutrino mass scale such as cosmology .
these predictions can be compared to future @xmath0 decay searches so as to gain information on the origin and nature of the neutrino masses . in this comparison
, we can distinguish the following scenarios : * * the @xmath1 process is observed to be in agreement with the forecasted rates .
* this indicates that the light active neutrinos dominate the @xmath0 decay rate .
since new degrees of freedom are in any event required to give the light neutrinos majorana masses , this implies that there is necessarily new physics above the nuclear scale , so that its contribution is suppressed . * * the @xmath1 process is observed to be smaller than the forecasted rates .
* this means that there is a partial cancellation between the active and extra neutrino contributions .
sterile neutrinos around the nuclear scale are then necessary .
a higher mass would imply too big a suppression through their nme to show any sizable cancellation , while too small masses would make the gim - like cancellation exact . * * the @xmath1 process is observed to be larger than the forecasted rates . * in this situation the light active neutrinos can not dominate the @xmath0 decay rate .
extra sterile neutrinos , lighter or around the nuclear scale , could have a significant contribution and reconcile the observations .
however , the gim - like cancellation between both contributions has to be avoided .
this implies either a cancellation between extra neutrinos both above and below the nuclear scale ( see sec . [
sec : heavylight ] ) or between the extra neutrinos and a type - ii or iii seesaw contribution ( see sec .
[ sec : mix ] ) .
this is the case that would correspond to a confirmation of the heidelberg - moscow claim . * * the @xmath1 process is not observed but was forecasted . *
while this could imply that neutrinos are dirac and not majorana particles , it can also be the case that neutrinos are majorana but extra sterile neutrinos below the nuclear scale are present .
thus , the gim - like cancellation takes place and the @xmath0 decay rate becomes unobservable . *
* the @xmath1 process is not observed and was not forecasted . *
this is the most pessimistic scenario since it is impossible to draw any conclusion on the nature and origin of neutrino masses .
we are specially indebted to andrea donini for carefully reading through this paper and providing valuable comments .
we also acknowledge very interesting and fruitful discussions with carla biggio , belen gavela , pilar hernandez , alessandro mirizzi , ann nelson , alfredo poves and achim schwenk .
this work was supported by the european union through the european commission marie curie actions framework programme 7 intra - european fellowship : neutrino evolution [ m.b . ] and the european commission framework programme 07 design study euro@xmath273 , project 212372 [ j.l.p ] ; the spanish ministry of education and science through the cup consolider - ingenio 2010 , project csd2008 - 0037 [ j.l.p . ] and grants fpa2009 - 09017 [ j.l.p ] and fpa2009 - 13377 [ j.m . ] ; the dfg through cluster of excellence `` origin and structure of the universe '' [ e.f.m . ] and grant sfb 634 [ j.m . ] ; the helmholtz association through the helmholtz alliance program , contract ha216/emmi `` extremes of density and temperature : cosmic matter in the laboratory '' [ j.m . ] ; and the comunidad autnoma de madrid through project hephacos - cm ( s2009esp-1473 ) [ j.l.p . , j.m . | we study the general phenomenology of neutrinoless double beta decay in seesaw models .
in particular , we focus on the dependence of the neutrinoless double beta decay rate on the mass of the extra states introduced to account for the majorana masses of light neutrinos . for this purpose ,
we compute the nuclear matrix elements as functions of the mass of the mediating fermions and estimate the associated uncertainties .
we then discuss what can be inferred on the seesaw model parameters in the different mass regimes and clarify how the contribution of the light neutrinos should always be taken into account when deriving bounds on the extra parameters .
conversely , the extra states can also have a significant impact , canceling the standard model neutrino contribution for masses lighter than the nuclear scale and leading to unobservable neutrinoless double beta decay amplitudes even if neutrinos are majorana particles .
in particular , the decay rate is reduced by at least six orders of magnitude for masses of the extra states below 1 mev in absence of extra contributions .
we also discuss how seesaw models could reconcile large rates of neutrinoless double beta decay with more stringent cosmological bounds on neutrino masses . |
the evolution of uv and ir luminosity densities can be understood as a decrease by a factor of @xmath1 6 of the global star formation density from the epoch corresponding to redshift z=1 to the present epoch ( lilly et al .
1996 ; flores et al .
these luminosity densities evolve similarly and contribute equally to the star formation in that redshift range , if one assumes that the shapes of the uv and ir luminosity functions to be similar to the local ones ( hammer 1999 ) .
integration of the derived global star formation implies that between half and two thirds of the present - day stars have been formed since z=1 .
this is somewhat in contradiction with the primordial collapse scenario in which the most massive galaxies were formed at much earlier epochs , at redshifts larger than @xmath1 3 ( e.g. renzini and cimatti 1999 ) .
the latter scenario is supported by the apparent non - evolution in the number density of large disks ( lilly et al .
1998 ) , and by its success in reproducing galactic disk evolution with chemical evolution models normalized to local galaxies ( boissier and prantzos 2000 ) .
+ the study of the luminosity density evolution relies on observations of luminous galaxies with @xmath8 -20 in the optical , and with @xmath9 2 @xmath10 @xmath11 @xmath7 in the ir .
the two main galaxy populations responsible for the observed decline in the star formation density are , in decreasing importance , the luminous ir galaxies ( flores et al .
1999 ) and the luminous compact galaxies ( lcg ) ( guzman et al . 1997 ; lilly et al .
the former are generally large and massive galaxies found mostly in interacting systems ( hammer , 1999 ) , and which show star formation rates ( sfr ) exceeding 50 @xmath12 .
they represent @xmath1 4% of the field galaxy population .
the latter are much more numerous and contribute to more than 40% of the uv luminosity density at z@xmath1 1 ( guzman et al .
1997 ; hammer & flores 1998 ) .
they constitute the most rapidly evolving galaxy population seen in the uv .
the properties of these lcgs are still largely unknown .
they are sometimes so compact that their nucleus is barely resolved in the wfpc2 hst images , which prevents their classification on the basis of their luminosity profile .
+ + koo & kron ( 1981 ) first drew attention to these compact galaxies during their search for faint qsos on the basis of colors and compactness .
they identified a sparse population ( the surface density is @xmath1 30 sources per square degree ) of very blue compact galaxies with z @xmath13 0.7 and exhibiting narrow emission lines . using the keck hires spectrograph , koo et al .
( 1995 ) demonstrated that the emission lines of lcgs have a gaussian profile , and suggested that they may be progenitors of local spheroidal galaxies through fading by @xmath1 4 to 5 magnitudes .
more recently , guzman et al . ( 1997 ) and phillips et al . (
1997 ) have identified a large population of emission - line compact galaxies in the flanking fields around the hubble deep field ( hdf , williams et al .
1996 ) at z @xmath13 1 , comprising @xmath1 20% of the field galaxy population .
they found the [ oii]@xmath143727 emission - line velocity widths of the compact galaxies to range between 35 and 150 km s@xmath15 ( phillips et al .
1997 ) , comparable to those of local dwarf galaxies , although the compact galaxies are 10 to 100 hundred times more luminous ( up to a few @xmath16 ) .
the similarities between the hdf compact galaxies and those from koo and kron ( 1981 ) have led guzman ( 1999 ) to suggest that blue compact galaxies at intermediate redshifts may be the progenitors of local spheroidal or irregular galaxies through fading by @xmath13 4 magnitudes .
if true , according to guzman ( 1999 ) , these should have extremely low m / l ratio , and this scenario would imply a severe downwards revision of the amount of star formation at z@xmath1 1 , since the baryonic content in irregulars ( or spheroidals ) is very small ( fukugita et al .
+ + we develop in this paper a different scenario for the nature of the brightest lcgs , which are major contributors to the observed decline of star formation density since z=1 . in our picture , lcgs are not the progenitors of present - day spheroidal and irregular galaxies , but rather of today s spiral bulges which built up gradually over time by a series of mergers of smaller units .
there have been several previous observations that have hinted at that scenario , but their meaning was not realized at the time .
thus , phillips and al .
( 1997 ) found that a large fraction of galaxies selected by their compactness appeared to be small but bright spirals . even adding a color criterion is not sufficient to distinguish between old and young stellar populations : schade et al .
( 1999 ) found that fully a third of the distant ellipticals they studied show star formation and colors as blue as those of blue compact galaxies .
the merging scenario comes naturally as the merging rate was about ten times higher at z=1 than it is today ( le fvre et al .
moreover , guzman ( 1999 ) and hammer ( 1999 ) have noticed that a large fraction of compact galaxies selected from the hdf and canada - france redshift survey ( cfrs ) fields are in interacting systems , with the presence of complex and distorted features surrounding their cores .
finally , the lcgs in the cfrs fields have near - ir and visible colors very similar to those of other luminous non - dwarf galaxies , with values apparently not consistent with those produced by a strong and short burst of star formation in a low - mass object ( hammer et al .
we present next new observational data which support and sharpen the above scenario .
+ in the following , @xmath17=@xmath18/50 and if not specified we adopt @xmath17=1 and @xmath19=0.5 .
our lcg sample was selected from two galaxy fields studied in the cfrs : the cfrs 0300 + 00 ( hammer et al . 1995 ) and the cfrs 2230 + 00 ( lilly et al .
1995 ) fields . to define our sample ,
we apply 3 selection criteria .
first is the compactness criterion , as determined from deep hst images ( see brinchmann et al .
1998 ) , or from deep ground - based images obtained by the cfrs team , with an image quality better than 0``.8 fwhm and a pixel size of 0.207 '' .
it is based on either a small size : @xmath20 @xmath13 5 @xmath21 kpc , where @xmath20 is the half - light radius and @xmath17 the hubble constant in units of 50 km s@xmath15 mpc@xmath15 or on light concentration as defined by the parameter : @xmath22 i = @xmath23 - @xmath24 @xmath13 0.75 mag ( hammer and flores 1998 ) .
+ + imagery from hst has been analysed with sextractor software package ( bertin and arnouts , 1996 ) .
the detection threshold was set to3.5 @xmath25 and a 35 pixel annulus was used for sampling the background .
the minimum contrast parameter for deblending was set to 0.005 , and for each source with an identified companion at more than 2 `` ( 18 kpc at z=0.5 ) , an image with a mask superimposed on the companion has been generated .
the background level at the mask location has been set using the background mean level , and the mask size was set from the derived kron radius of the companion ( bertin and arnouts 1996 ) .
this procedure has the advantage of eliminating systematic biases against compact sources with companions beyond 2 '' . for hst images ,
15 apertures with diameters varying from 0``.2 to 3''.0 with increasing steps of 0``.2 , were used to derive aperture photometry .
the half - light radius has been derived assuming that the isophotal magnitude generated by sextractor was a good approximation to the total magnitude of the source .
calculation of the light concentration parameter was done after interpolating to the luminosities enclosed within the 5 kpc and 15 kpc radii .
ground - based images were analyzed in a similar way , but a 20-pixel annulus was used for background sampling .
another problem with images obtained from the ground come from background fluctuations . for the majority of the objects , aperture magnitudes are increasing smoothly with aperture radius .
those that show a magnitude decrease larger than the measurement errors have been excluded from the sample .
+ + the above selection criteria are equivalent for a face - on disk or a spheroidal component with a r@xmath26 profile since for a @xmath20 @xmath13 5 kpc source , @xmath23 is a good approximation of the total source luminosity .
figure 1 demonstrates this equivalence for sources imaged with the hst .
the bottom panel shows that the light concentrations calculated from hst and cfht images for the same object correlate well , the correlation coefficient being 0.73 .
we note however that the average value of @xmath22 i from ground images is larger by 0.1 magnitude as compared to that from the hst images .
this can be easily explained by the broadening of the profile due to seeing conditions , and we have applied a slight empirical correction to the values derived from the ground .
the correlation between ground and hst data is found to be better for the concentration parameter ( upper panel ) than for the half - light radius .
we interpret this to be related to the uncertainties in the calculation of the isophotal magnitudes for ground - based images , because background fluctuations affect them more than aperture magnitudes . as a result
, we have decided to adopt the concentration parameter as the selection criterion for compact galaxies .
examination of images of the selected compact galaxies shows that the compactness selection criterion has also picked out a few highly inclined disks with a large major axis extent .
these were removed from the sample using the axis ratio calculated by sextractor .
one can also notice from figure 1 that @xmath1 20% of the sources are found to be compact on the basis of their hst photometry , although they appear to be extended from the ground .
examination of their images show that they are mostly sources with nearby companions ( within 3 '' ) for which the deblending procedure applied to ground - based images has failed .
in other words , our procedure applied to ground - based images appears to be more restrictive than when applied to higher spatial resolution hst images . a selection from ground images
will inevitably generate a sample of galaxies slightly biased against galaxies with nearby companions .
+ + second , we apply a luminosity and redshift criterion . only galaxies with @xmath27(b ) @xmath13 -20 , and in the redshift range from 0.45 to 0.8 are included .
the luminosity cut - off is motivated by the fact that the reported evolution of the rest - frame uv luminosity density is based on observations of galaxies with @xmath27(b ) @xmath13 -20 ( lilly et al .
1996 ) , while the redshift range is the one in which the cfrs is considered complete for luminous galaxies ( see hammer et al .
third , we include only lcgs with known [ oii]@xmath143727 emission ( @xmath28)$ ] @xmath29 15 ) , because we wish to compare the gas properties to those of the underlying stellar population . applying these three selection criteria to the 2 cfrs fields
yields a sample of 59 galaxies , of which we have chosen a small representative subsample of 14 lcgs for a first investigation .
they represent @xmath1 23% of the @xmath27(b ) @xmath13 -20 galaxy population in the redshift range defined above . their global properties , including b and k luminosities and compactness and concentration parameters ,
are given in table 1 .
their apparent magnitudes range from i = 20.5 mag to 22 mag , and they possess colors spanning the entire morphological sequence from irr to sbc galaxies .
they are very similar in their properties to the brightest @xmath1 25% of the blue compact galaxies selected by guzman et al .
( 1997 ) in the hubble deep field at 0.1@[email protected] . however most of the brightest compact galaxies of guzman et al .
have higher redshifts than the lcgs considered here , which complicates a direct comparison .
spectrophotometric observations of the 14 lcgs were obtained during one night ( 1999 august 10 - 11 ) with the european southern observatory 8 m vlt / ut1 using the fors1/r600 and i600 spectrograph at a resolution of 5 and covering the wavelength range from 5500 to 9200 .
this ensures the coverage from [ oii]@xmath143727 to [ oiii]@xmath145007 in the rest - frame spectrum for a large fraction of the targets .
additional exposures have been done through service observing mode , during july and august 2000 .
each galaxy was observed through a 1 `` slit ( figure 2 ) in 5 to 7 exposures of @xmath1 40 mn each , using the ' ' shift and add " technique to ensure proper removal of cosmic rays and of possible ccd defaults .
spectra were extracted and wavelength - calibrated using the iraf package .
flux calibration was done using 15 min exposures of three different photometric standard stars . to ensure the reliability of the data ,
all spectrum extractions as well as lines measurements using the splot program were performed independently by two of us ( f.h . and n.g . ) . to disentangle the continuum properties from those of the emission lines ,
we have adopted the following specific procedure for all spectra .
we applied a two - step smoothing ( first by 7 pixels and then by 15 pixels ) to the continuum , while excluding from the smoothing the well identified and mostly unresolved emission lines to keep their original spectral resolution .
this smoothing procedure has the advantage of not affecting the shape of the continuum and the narrow emission lines , while revealing the more prominent absorption features . for the latter
the nominal spectral resolution is @xmath1 10 11 , comparable to that of the stellar cluster spectra used as templates in our spectral synthesis work to be described later ( section 4.2 ) .
figure 3 presents the 14 resulting rest - frame spectra , where the location of strong sky emission lines ( at 5577 , 6300 and 6364 ) as well as sky absorption lines ( at 6874 , 7590 and 7630 ) have been indicated . excluding these night sky lines , the smoothed continuum of all spectra shows a s / n per pixel element always larger than 5 . for each spectrum ,
the locations of the most prominent lines are marked and their identifications are given in the last panel . + + the @xmath30 , @xmath31 and @xmath32 balmer emission lines were found to be systematically contaminated by absorption lines . to disentangle the absorption from the emission , we have synthesized the stellar continuum by a combination of star spectra .
stellar energy distributions were taken from the library of jacoby et al .
the stars have solar metallicity which is appropriate for our objects as shown later .
we used a combination of stars with spectral type varying from b to k , and optimized the fit to the continuum for the important spectral absorption lines , such as the high order balmer lines ( h10 to h8 ) , the caii k line and the g band ( figure 4 , top panel ) .
we impose the additional constraint that the resulting synthesized stellar continuum not be redder than the object continuum , since the spectra of lcgs have simultaneously strong absorption features characteristic of old stellar populations and blue u - v colors characteristic of young ones .
the synthesized stellar continuum is then subtracted from the observed spectrum to yield a pure emission spectrum uncontaminated by absorption ( figure 4 , bottom panel ) which we can use to measure the balmer decrement and derive extinction . for 12 lcg spectra with both h@xmath33 and h@xmath34 lines ,
the residual emission lines have been measured using the splot package , and the measurement errors estimated by trying several combinations of the stellar templates . in most cases ,
the balmer emission lines are seen to be either blue or red - shifted with respect to the center of the absorption lines , which minimize their contamination by underlying stellar absorption , and hence the measurement error for the gas extinction .
emission line widths were estimated after a careful examination of the actual spectral resolution because seeing conditions were varying from 0``.65 to 1''.0 fwhm , and the object sizes were comparable to the slit size .
the instrumental resolution is around 5 ( fwhm ) , as measured from absorption lines of the calibration stars and unresolved sky lines .
each line measurement was checked against the closest sky emission line and velocity widths were derived assuming a gaussian profile .
all spectra but one include the [ oiii]@xmath144959 and 5007 emission lines . following guzman et al .
( 1997 ) , velocity widths derived from this set of lines are averages , weighted by the quality of individual measurements .
they are given in table 2 .
ten of the 14 lcgs possess k band photometry ( hammer et al .
1997 ) along with b , v and i photometry . at the reference redshift of z
= 0.5182 , the central wavelength of the observed i band is exactly that of the rest frame v band , while the observed v band matches well the rest frame u band centered at 3622 . at this redshift
the observed k band represents a filter centered around 1.416 @xmath35 m .
observed colors have been derived within an aperture of 3 `` diameter .
calculations of rest - frame colors has been done by systematic fits of the observed v , i and k photometric points , using bruzual and charlot ( 2000 ) models with exponentially decreasing sfr characterized by e - folding times @xmath36 of 0 , 0.1 gyr , 0.5 gyr , 1.2 gyr , 4 gyr and infinity .
the fits have been performed without extinction correction , and reproduce well the colors of all objects except for cfrs 03.1349 , cfrs 03.1540 and cfrs 22.1064 .
the latter objects have observed ( i - k ) colors which are too red even when compared with those of a 15 gyr old stellar population . because the 14 selected lcgs have redshifts close to the reference one , the model - dependent corrections applied to the rest - frame colors ( u - v ) and ( v - 1.4 @xmath35 m ) are always negligible or smaller than the measurements errors ( i.e. @xmath3 0.05 mag ) .
thus , our derived rest - frame colors are very accurate , with only a second - order dependency on the model used to derive them .
+ + the resulting color distribution in figure 5 illustrates the following point : our sample which was constructed using a selection criterion on the emission lines includes galaxies with ( u - v ) colors ranging from those of irr to those of sbc , just as the sample of guzman et al .
on the other hand , our lcgs display a large dispersion of the ( v - 1.4 @xmath35 m ) color , with 7 of them showing colors much too red compared to the predictions of the models .
model uncertainties of the ( v - k ) color which is very similar to the ( v - 1.4@xmath35 m ) color can not account for this effect .
these very red colors can be explained if the presently observed burst is superimposed on a red old stellar population , if there is significant extinction , if the stellar population has over - solar abundances , or if there is a combination of the above .
for the reddest objects ( cfrs 03.1242 , cfrs 03.1349 , cfrs 03.1540 and cfrs 22.1064 ) , a combination of these effects is likely at work .
interestingly , these four extremely red lcgs all possess a bright companion within 4 '' or 36 kpc ( figure 2 ) .
these companions are not affecting the k band aperture photometry , except maybe for 03.1242 .
the cfrs 0300 + 00 field was also observed by iso to a limit of 350 @xmath35jy ( 4@xmath25 ) at 15@xmath35 m and by the vla to a limit of 90@xmath35jy ( 5@xmath25 ) at 21 cm .
two very red lcgs , cfrs03.1349 , 03.1540 in our sample have been detected at both mid - ir and radio wavelengths .
+ + the dispersion in the ( v - 1.4 @xmath35 m ) color justifies _ a posteriori _ our adopted name for these objects , luminous compact galaxies ( and not blue lcgs ) , since only their ( u - v ) colors are blue . the lower luminosity compact galaxies of guzman et al .
are significantly bluer in ( v - 1.4@xmath35 m ) as compared to the lcgs in our sample ( guzman , private communication ) .
+ all the lcg spectra show strong and broad absorption lines .
most prominent are the balmer lines , the ca i 4227 and ca ii h and k lines , the cn band at 4182 , the ch band at 4301 , and the mg i line at 5173 ( figure 3 ) .
a special mention should also be made for the important iron absorption line system in lcg spectra , including fe i 4037 , 4264 , 4384 and 4435 , and sometimes fe i 4004 , 4595 and 4645 .
fei lines are even present below 3900 .
this suggests that a large fraction of stars in lcgs have solar abundances or greater , and that sn ia have already released iron - peak elements , giving an age of about 1 gyr or more for the bulk of the stars ( see e.g. mcwilliams 1997 ) . in some lcg spectra the cn band affects the location and shape of the h8 and h9 lines ( see for example the spectrum of 03.1540 ) , which also suggests the presence of high metal abundance stars .
table 3 gives the equivalent widths of the most important absorption lines .
the continuum has been synthesized using a grid of stellar cluster spectra with different ages and metallicities , and following the procedure outlined by bica & alloin ( 1986a and b , herafter baa and bab ) . to test our computer routines ,
we have first performed a continuum fit of the stellar cluster spectra , and check our results against those of baa , before applying the procedure on lcg spectra .
our measurement errors are slightly lower than those of bab s table 5 ( from 1 to 2 ) .
we found a correlation between the equivalent widths of the metal absorption lines and the @xmath37m@xmath38 color index ( figure 6 ) .
this is to be expected if the latter is mainly sensitive to age and metallicity effects .
a few objects depart from the correlation .
these are the reddest lcgs , cfrs03.1349 and 22.1064 , for which we suspect extinction to modify the @xmath37m@xmath38 color index .
figure 6 also shows that the properties of most lcgs are consistent with their stellar populations being composed of a 5 gyr old population with solar abundance , mixed with a young ( @xmath130.1 gyr ) population , the latter being responsible for the dilution of the absorption lines .
the bica & alloin population synthesis method allows us to perform a two - parameter analysis which takes into account both age and metallicity effects .
because galaxy interactions are important in lcgs , we prefer not to adopt any assumption concerning the abundances of the young population relative to those of the old population , as done by bica ( 1988 ) in his analysis of e / s0 and spiral nuclei . on the other hand , to simplify the analysis , we have made the plausible assumption that , in a small volume with intense star formation , the gas is chemically homogeneous at least during the last @xmath39 years .
ten absorption lines have been used to perform the fit .
they include h 10 ( 3798 ) , h8 ( 3889 ) , ca ii k ( 3933 ) , fei ( 4037 ) , cn ( 4182 ) , cai ( 4227 ) , fei ( 4264 ) , the g band ( 4301 ) , fei ( 4384 ) and mgi ( 5175 ) . for most lcgs , this results in 4 to 7 lines available in the spectral range , and not contaminated by sky emission or absorption .
five star cluster populations with ages 5 @xmath10 @xmath40 , @xmath40 , 5 @xmath10 @xmath39 , @xmath41 and @xmath42 yr have been used as templates .
we have chosen not to use globular cluster spectra ( age = 15 @xmath10 @xmath40 ) because at z@xmath1 0.5 , lcgs emitted their light not more than 6 - 7 gyr ago .
the metal abundance of the older population was allowed to be one of the following values : @xmath43 = - 1.5 , - 1 , - 0.5 , 0 and 0.6 . the same metal abundance has been assumed for the @xmath40 and 5 @xmath10 @xmath39 yr old populations ( @xmath43= -1 , -0.5 , 0 and 0.6 ) , and for the @xmath39 and @xmath41 yr old populations ( @xmath44 = -0.5 , 0 and 0.6 ) .
our goal here is not to fully disentangle the metallicity from age effects for each lcg ( our spectra do not have the required s / n to accomplish that ) , but rather to have a first rough estimate of its stellar population properties .
in many cases , lcgs show large equivalent widths of fei 4384 and mgi 5175 and relatively small equivalent widths of caii k lines .
the latter have been fitted thanks to the line dilution in the blue by the younger population which has a hii spectrum .
the fit has been done as follows : we add first a new star cluster population to the mixture in incremental steps of 5% in the light contribution .
if the objects are not fitted , we reduced the step to 2% . in some cases , the individual iron line equivalent widths were not well reproduced by the models
this is not too surprising given the complexity of the fe line system : even for well - known stars , fe / h is not known to better than a factor of 2 or 3 ( wheeler , sneden & truran 1989 ) .
for the 6 lcgs with non - satisfactory fits of individual fe lines , we have fitted only the sum of the three fei rather than each of them individually .
table 4 presents the result of the fits .
@xmath45@xmath46 values are calculated as averages of the sum of the squares of the differences between model and lcg spectrum divided by the squares of the measurement errors .
the latter are taken to be the quadratic sum of the bab errors and our errors .
we also restrict the parameter space to be explored by imposing the plausible constraint that the ( 3728 - 5175 ) color ( approximately the ( u - v ) color ) of the model has to be bluer than that of the lcg .
the contribution to the v light of each population in the synthesis model is given in table 4 , in bold characters for the major contributor .
most lcgs are dominated by populations having ages ranging from 1 to 5 gyr .
we have tested this result by attempting to fit the lcg absorption lines by limiting the age range to values lower than or equal to 5@xmath10@xmath39 yr : for all lcgs but cfrs 22.0637 and cfrs 22.0919 , the @xmath45@xmath46 value of the fit is well in excess of 1 . except for cfrs 22.0637 and cfrs 22.0919
, the stellar populations in lcgs invariably include an important contribution ( @xmath29 35% ) stars older than 1 gyr and with metal abundance @xmath43 @xmath29 -0.5 . for 9 of them
, we find an important contribution of @xmath47 1gyr stars with @xmath43 @xmath29 0 .
we have tested the robustness of the latter result by running models where the metal abundance has been constrained to have @xmath43 values lower than or equal than -0.5 .
we found that for 7 lcgs , cfrs 03.0523 , cfrs 03.0570 , cfrs 03.0595 , cfrs 03.1242 , cfrs 22.0344 , cfrs 22.0429 and cfrs 22.0619 , there is no possibility to fit their absorption line spectra without an important contribution of solar or oversolar stellar population .
special mention should be made of cfrs 03.1242 for which we could not find a satisfactory fit .
the reason is that it shows very strong absorption lines ( table 3 ) , stronger than the most metal - rich star clusters in our library .
four lcgs , cfrs 03.1540 , cfrs 22.0637 , cfrs 22.0919 and cfrs 22.1064 , apparently require low - abundance populations to fit their absorption line spectra
. however our modelling was unsuccessful to provide physical solutions for cfrs 03.1540 and cfrs 22.1064 , for which no ( or in very small amounts ) young population responsible of their emission lines have been identified .
any young component added would dilute the absorption lines , and would lead to an old stellar component with a solar metallicity .
while our spectral synthesis work allows us to put constraints on the ages and metallicities of the stellar populations in lcgs , the reader should not take the values given in table 4 as being the actual stellar population distributions in lcgs .
there are several caveats which limit the results of our adopted spectral synthesis procedure .
first , there is more than one solution with @xmath45@xmath46 values lower than 1 , and table 4 only displays the solution with the minimum value .
these appear to be true minima as they do not change when the incremental step for adding the light contribution of a stellar population is changed .
second , the fit of a galaxy population using only a few stellar cluster templates can smooth over some important events in the star formation history of a galaxy in which star formation is sustained over long periods of time .
we believe this effect to be especially important in lcgs within interacting systems .
the complex nature of the star formation history and the resulting young stellar population could then explain its absence in cfrs03.1540 and cfrs 22.1064 .
it can also be the reason why our modeling of cfrs 03.1242 proves to be unsuccessful .
third , and more importantly , our ability in disentangling age from metal abundance effects is limited by the relatively small number of absorption lines .
for example , some lcg spectra do not cover the blue region and do not include the important lines below 4000 .
this would bias the synthesis method against finding young stellar populations which would dilute the uv absorption lines .
this effect can be seen in table 4 for cfrs 03.1540 and cfrs 22.1064 for which the best - fit model found no young population , although their spectra show emission lines .
what can be firmly said about the metal abundance and age of lcgs ?
first , all lcgs , but cfrs 22.0637 and cfrs 22.0919 , are dominated by 1 to 5 gyr old population , characterized in most cases by solar to over - solar abundances . to test the robustness of this result ,
we have attempted to fit the 12 lcgs with an old stellar population with the added constraint that the contribution of that old population does not exceed 25% .
we found that half of the lcgs spectra could not be fitted , their @xmath45@xmath46 being invariably greater than 1 .
as for the remaining lcgs , they could be only fitted by assuming a young population with the highest metal abundance available in our stellar library ( @xmath43 = 0.6 ) .
this is not realistic when compared to the gas abundance values discussed in the next section .
second , the best fits to the lcgs spectra always give a mix of stellar populations with metal abundance covering a wide range from undersolar ( generally characterizing the youngest population ) to solar or oversolar values ( characterizing the oldest population ) .
this is illustrated in figure 6 , which shows that the most important lines within a lcg can not all be fitted with a single metallicity .
as noted above , some models of lcg spectra show no young population , in evident contrast with their emission line spectra . adding
a young stellar component would inevitably enhance the metal abundance of the oldest stellar component , making our conclusion that most lcgs are dominated by old and solar metallicity stellar population even stronger .
for the 12 spectra with both @xmath30 and @xmath31 detected , we have been able to derive @xmath48 extinction values , after correcting the balmer emission lines for stellar absorption using the results of our spectral synthesis work ( section 3 ) .
as shown in table 5 and figure 7 ( left panel ) , the derived extinctions can have values much larger than those currently found in present - day irregular galaxies which have an average @xmath48 of @xmath1 0.57 ( gallagher et al .
these large extinctions imply that the sfr derived from the 2800 luminosity of lcgs is greatly underestimated .
we have calculated for each lcg the dereddened @xmath30 line flux and derived the corresponding sfr using a salpeter imf ( kennicutt 1998 ) and assuming the standard ratio between @xmath49 and @xmath30 , case b and a gas temperature of 10000 k. figure 7 ( right panel ) compares the sfrs so calculated to those derived using the 2800 continuum , assuming the same imf ( see kennicutt 1998 ) . even allowing for errors due mainly to uncertainties in the extinction estimates ,
it is clear that the uv luminosity underestimates the actual sfr in lcgs by factors ranging from 2 to 30 , with an average of 14 and a median of 11 .
the largest underestimates are for cfrs 03.0523 , cfrs 03.1540 and cfrs 03.1349 , which are both ir - luminous galaxies . from their ir and radio energy distributions , flores & hammer ( 2000 , in preparation ) derive sfrs respectively of 87 , 135 and 94 @xmath2 .
these values are in agreement with those derived from the extinction corrected balmer lines , allowing for the error bars which have large upper values . as another consistency check ,
we have applied the extinction factor derived from the @xmath50 ratio to correct the sfr values at 2800 ( see table 5 ) .
extinction - corrected @xmath51 values are always consistent with extinction - corrected @xmath52 values within the error bars . the o / h abundance ratio for the gas can be estimated from the quantity @xmath53 = @xmath543727 + [ oiii]4959,5007 ) /h\beta$ ] ( edmunds and pagel 1984 ) , although the relationship is degenerate and there are two o / h values for a given @xmath53 .
the calculated @xmath53 for the 12 lcgs with the necessary lines range from 0.5 to 1 ( table 6 ) , in good agreement with the values obtained for similar objects at lower redshifts by kobulnicky and zaritsky ( 1999 ) .
flux measurements of the faint [ oiii]@xmath144363 line allow to lift the degeneracy for two objects ( cfrs22.0619 and cfrs22.0919 ) . for those ,
we derive gas temperatures of 12500 and 13000 k , close to the average upper value for the rest of the sample ( table 6 ) . using figure 8 of kobulnicky et al ( 1999 ) , it was possible to set for each line ratio @xmath53 , an upper and a lower value of o / h , with the help of the ionisation ratio @xmath554959,5007 /[oii]3727 $ ] .
we found that all lcgs have gas abundances @xmath56 in the range from 0.25 to 0.4 , with the exception of cfrs 22.0919 , to be discussed later , which shows a very low o / h . adopting a statistical correction of + 0.1 dex on the o / h abundance as suggested by kobulnicky et al .
( 1999 ) for physical parameters derived from global galaxy spectra , our data are finally consistent with a median metal abundance @xmath56 = 0.45 , similar to the value derived by kobulnicky and zaritsky ( 1999 ) for their objects .
thus , gas in lcgs displays oxygen abundances intermediate between those of local spirals ( @xmath56 @xmath1 1 ) and those of irregular dwarf galaxies ( @xmath56 @xmath1 0.3 ) .
we have obtained vlt spectra of 14 lcgs selected by their compactness and luminosity in the rest - frame blue .
the blue luminosities of lcgs range from 0.35 to 1.4 @xmath16 with an average of 0.70 @xmath16 .
they are also bright in the near - ir , with a rest - frame k - band luminosity ranging from 0.1 to 2.8 @xmath16 and an average of 0.86 @xmath16 , assuming @xmath57(ab)=-22.55 for @xmath58 ( glazebrook et al , 1995 ) .
guzman ( 1999 ) has noticed that lcgs are also compact at near - ir wavelengths .
here we assume that the k - light samples the stellar mass ( charlot 1998 ) , with a mass - to - ir light ratio very close to unity .
this follows charlot ( 1996 ) who has investigated this ratio for a solar abundance population with various ages , appropriate for lcgs .
a salpeter initial mass function ( imf ) has been adopted , as it allows models to reproduce well the evolution of the global ir and b luminosity densities from z = 0 to z = 1 , to which the lcgs contribute importantly . on the other hand , lilly et al .
( 1996 ) found that the use of a scalo imf in modeling the z = 0 1 galaxy population produces too many long - lived low - mass stars .
also photometric properties of local disks suggest a salpeter rather than a scalo imf ( kennicutt et al .
1994 ) . in any case ,
the adopted 0.22 dex uncertainty based on charlot ( 1996 ) s study of the conversion of k luminosity to mass likely accounts for the major uncertainties related to the imf .
lcgs have stellar masses ranging from @xmath59 to 2.5 @xmath60 @xmath61 , with an average of 7 @xmath62 @xmath61 ( see table 1 ) .
k - band luminosities derived from the 1.4@xmath35 m luminosities are robust as the ( 1.4@xmath35m - k ) color shows small variations during the period from @xmath39 years ( it has the value -0.32 mag ) to @xmath59 years ( its value is -0.15 mag ) after the beginning of star formation .
uncertainties in these derivations are small compared to the large uncertainty involved in the conversion of the k - light into mass ( @xmath1 0.22 dex in figure 8) . from the way they have been selected and their properties , the lcgs studied here are very similar to the 25% most luminous blue compact galaxies identified by guzman et al ( 1997 ) in the hubble deep field ( hdf ) .
they are the galaxies responsible for the decrease in the uv luminosity density since the epoch corresponding to z=1 ( lilly et al .
1998 ) . with the noticeable exception of cfrs 22.0919 and cfrs 22.0637 , lcgs
all display strong metallic absorption lines and have ( u - v ) colors similar to those of late - type galaxies , in the range sbc to irr .
these properties can be reproduced by a combination of an old star population with age @xmath29 @xmath40 yr with a younger population with age @xmath13 5 @xmath10 @xmath39 yr ( see section 4.2 ) .
it is important to note that the best spectral synthesis fits are generally obtained with an older stellar population with abundances larger than or equal to solar values , while the younger component has solar or sub - solar abundances .
the latter component diminishes the observed strength of the absorption lines due to the older stellar population , as well as making the @xmath63 color bluer than that of a sbc galaxy .
analysis of the emission lines indicates that the ionized gas in lcgs also displays oxygen abundances in the range from 0.3 to 0.5 the solar value , i.e. the metal abundance of the gas is similar to that of the younger stellar population .
given these observations , what can we say about the nature of lcgs ?
the presence of a moderately old stellar population with solar abundances or larger , stellar masses near that of the milky way , large sfrs similar to or only slightly below those of luminous ir galaxies , all argue against a scenario in which lcgs are progenitors of today s spheroidal or irregular dwarf galaxies as discussed by koo et al .
( 1995 ) and guzman et al . (
that hypothesis is based mainly on one key observation : that of the small velocity widths of the emission lines of lcgs ( table 2 )
. these authors argue that the velocity widths are representative of virialized motions of the gas and since the velocity widths are similar to those of dwarf galaxies , they are dwarf progenitors . however an examination of the lcg spectra ( figure 3 ) , shows that the balmer emission lines are often red - shifted relatively to the balmer absorption , suggesting that more complex mechanisms may be responsible of the narrowness of their widths .
we also note that , in some cases , emission lines have asymetric profiles . in the case of cfrs 03.0523 , they show two peaks separated by @xmath1 470 km s@xmath15 , which are likely to be associated with each of the two merging components ( gruel et al .
2001 , in preparation ) .
a special comment should be made about cfrs 22.0637 and cfrs 22.0919 which , as we have noted before , appear to be different from other lcgs and resemble more hii galaxies .
in particular , cfrs 22.0919 has a young low - metallicity stellar population as well as a low gas oxygen abundance , it shows a moderate gas extinction , and a large velocity difference ( 800 km s@xmath15 ) between the emission and absorption lines .
gas outflow can be responsible for the blue - shifted balmer absorption lines , so that cfrs 22.0919 may be just the type of dwarf progenitor object discussed by koo et al .
( 1995 ) and guzman et al .
( 1997 ) . but
objects such as cfrs 22.0919 constitute more the exception than the rule in our lcg sample .
the stellar mass as well as the light concentration of other lcgs is more reminiscent of those of bulges of today s sab to sbc spirals or even ellipticals .
could lcgs be the progenitors of bulges of present - day massive galaxies ?
the bulk of their stars displays a large range in metal abundance , as in the milky way s bulge . taking into account their large extinctions , lcgs are found to be still forming stars at a very vigorous rate , with a median sfr equal to 40 @xmath2 . for 8 lcgs with the necessary data ,
the median characteristic formation time scale for the bulk of the stars derived from the emission lines is @xmath1 1.8 gyr ( figure 8) .
this value is consistent with that obtained by the absorption line analysis , but is much larger than the characteristic time of @xmath1 0.1 gyr for the monolithic scenario of bulge formation ( elmegreen 1999 ) .
thus , the latter scenario does not appear to be the appropriate one for explaining the observed properties of the majority of lcgs .
which mechanism can then explain the transformation of lcgs to present - day spirals ?
an important hint can be obtained from hst images .
these show that an important fraction , if not all , of the lcgs possess a low - surface - brightness component which surrounds their core ( figure 2 , see also guzman , 1999 or hammer , 1999 ) .
moreover , a substantial fraction of them ( 6/14 ) are seen to have bright companions within a radius of 40 kpc .
the lcgs with companions show the largest sfrs in the sample , and they are also the reddest in ( v-1.4@xmath35 m ) .
the three galaxies in the sample which have been detected by iso ( flores et al .
2000 , in preparation ) are among them .
it is thus plausible to think that gas tidally pulled out from the companion and falling into a lcg can feed the star formation within it .
this would naturally explain the metal abundance difference between the gas and the oldest stars .
a good example of a lcg where such a mechanism may be at work is cfrs 03.1540 : its low - surface - brightness companion acts as a reservoir of low abundance gas to sustain the star formation in it during several @xmath39 yr .
cfrs 03.1540 also shows a well - developped disk
it was classified as a disk galaxy by lilly et al .
( 1998 ) and star formation is probably occurring in both bulge and disk .
the tadpole galaxies described by brinchmann et al .
( 1998 ) which would have been selected as lcgs by our criteria , can be interpreted in this context as undergoing the first stage of the formation of a disk around the bulge .
thus , by observing lcgs , we are witnessing an important phase of galaxy formation : the building of the bulge and the beginning of the disk formation of @xmath16 spirals , where galaxy interactions play a major role .
the scenario developped here has to account for the narrow emission line widths reported by phillips et al .
( 1997 ) and also seen in several of our spectra . in normal galaxies , the bulge is dynamically relaxed , and the gas emission likely samples most of the velocity field , i.e. the emission line widths should be indicative of the whole galaxy potential .
weedman ( 1983 ) reported however unusually narrow emission line widths in starburst nuclei .
lehnert and heckman ( 1996 ) studying a sample of ir - selected starburst galaxies , have shown that the starburst occurs preferentially in the inner region , so that the gas does not sample fully the solid body part of the rotation curve .
they found several objects with emission line velocity widths of less than 50 km s@xmath15 in systems with rotational velocities ranging from 100 to 200 km s@xmath15 .
since they have sfrs and hence ir luminosities , equal or slightly below those of iso galaxies ( flores et al .
1999 ) , we conclude that our lcgs are very similar to some lehnert & heckman galaxies , and that the narrow emission line widths are not indicative of the entire potential of the lcgs and lead to a systematic underestimate of their true masses .
the present study thus suggests that lcgs are progenitors of today s massive ( non - dwarf ) galaxies , which have formed the bulk of their stars at z @xmath3 1 .
we note that : + 1 ) the gas supplied to lcgs from interacting companions is enough to sustain their star formation during several @xmath39 yr , until completion of the bulge stellar content as well as to feed a gradual formation of the disk .
+ 2 ) the characteristic time for the formation of the bulk of lcg stars is relatively small , averaging to only @xmath1 1.8 gyr , i.e. most of the stars formed at z @xmath3 1 .
+ because lcgs represent @xmath1 20% of the field galaxies , the fraction of stars in massive galaxies formed at relatively recent epochs could have reached 20% , and even more if the star formation in interacting systems of large ( non - compact ) galaxies detected by iso ( flores et al .
1999 ; hammer , 1999 ) is taken into account .
this estimate is much higher than the one obtained by brinchmann & ellis ( 2000 ) , who find the fraction to be negligible .
the main reason for the discrepancy comes from the way the sfr is derived .
brinchmann and ellis ( 2000 ) calculate the sfr from the [ oii]3727 flux , while we use the extinction - corrected @xmath49 emission , assuming the standard ratio between @xmath49 and @xmath30 , case b and t = 10000 k. our derived sfrs can be a factor 10 or more greater than those derived by brinchmann & ellis ( 2000 ) .
this can be seen by comparing their figure 3 to our figure 8 . using the same sample as that of brinchmann & ellis ( 2000 ) , hammer and flores ( 1998 )
have shown that sfrs derived from either [ oii]3727 or 2800 luminosities do not correlate with those obtained from @xmath49 luminosities , mainly because of extinction effects . given their large extinction and @xmath49-derived sfrs , it is likely that the ir luminosities of lcgs are similar to or only slightly below those of the luminous ir galaxies detected by iso in the same redshift range ( flores et al .
because they represent 20% of the field galaxies in the redshift range from 0.4 to 0.8 , the contribution of the lcgs to the global sfr density needs to be reevaluated .
such a reevaluation will probably lead to an increase in the fraction of the sfr density related to photons reprocessed by dust , i.e. there should be an evolution in the shape of the ir luminosity function , due to the addition of new populations such as lcgs around @xmath5=@xmath64 .
this subject can not be addressed properly with our present small sample .
we shall discuss it in a future paper with a much enlarged sample .
spectroscopic observations of a small sample of 14 luminous compact galaxies ( lcg ) with the vlt have provided new and important information on a crucial stage in the formation of galaxies , that of bulges in massive ( non - dwarf ) spiral galaxies .
the lcg spectra are characterized by the following properties : lcgs with redshifts between 0.4 and 0.8 are undergoing intense bursts of star formation , with sfrs averaging to @xmath1 40 @xmath67 .
because of their large extinctions , they are expected to be luminous in the ir , at a luminosity level just below the ir luminous galaxies detected by iso in the same redshift range ( flores et al .
1999 ) . indeed 3 of the 14 lcgs discussed in this paper are detected by iso , implying higher extinction for them .
\3 ) the majority of lcgs show low - surface - brightness components around their high - surface - brightness cores ( the bulges ) .
most are in interacting systems , and the tidally pulled gas from the companions is likely to sustain star formation in the lcgs at a high rate for several hundreds myr , resulting in the completion of the bulge formation and the beginning of the disk formation .
the characteristic time for forming the bulk of the stars in lcgs is rather small , only @xmath1 1.6 gyr on average .
it is comparable to the age of the oldest stars in lcgs derived from the analysis of their absorption lines .
this means that a significant fraction ( several tens of percent ) of present day s massive galaxies could have been formed at relatively recent epochs , at z @xmath13 1 .
this is consistent with the cosmic star formation history which predicts that an important fraction of the stellar mass was formed in the z = 0 1 redshift range ( hammer 1999 ) .
the importance role played by galaxy interactions in the formation of bulges of spiral galaxies is also consistent with the hierarchical theory of galaxy formation . on the other hand
, the scenario outlined here may appear to be in contradiction with the results of lilly et al .
( 1998 ) who found that the number density of large spirals at z @xmath1 1 is comparable to that today .
however the lilly et al .
results are based on only @xmath1 40 morphologically selected large spirals in the redshift range from 0.2 to 1 , and their large one @xmath25 error bar can easily accomodate a density evolution of spiral disks of @xmath1 20% as found here . moreover , some of the large spirals in lilly et al .
( 1998) sample are in interacting systems and are strong infrared emittors , such as cfrs 03.1540 discussed here , or other examples discussed by flores et al .
presumably , these large interacting disks will merge leading to a density evolution . if the 14 lcgs described here are representative of the whole lcg population at z @xmath680.4 , then their contribution to the global star formation density should be revised upwards by factors as large as 7 10 .
that contribution could then be as high or higher than that of the luminous ir galaxies detected by iso ( flores et al .
the latter are more massive systems and formed stars more rapidly than lcgs , but they constitute a much sparser population of galaxies , comprising only a few percent of the total population .
most lcgs at z @xmath1 0.5 would have bolometric ir luminosities not far below the iso sensitivity limit , and we predict that they will be easily detected in large numbers by sirtf . finally we reiterate that our interpretation of the nature of lcgs is quite different from that proposed by koo et al .
( 1995 ) , guzman et al .
( 1997 ) and guzman ( 1999 ) for the majority of compact galaxies found in the hubble deep field and flanking fields .
this is probably because most of the guzman et al galaxies have intrinsically lower luminosities than lcgs .
we believe lcgs to be 10 to 100 times more massive than dwarf galaxies , to be forming stars at large rates , so they can not be the progenitors of local spheroidal or irregular systems through fading .
corresponding to the brightest 25% of the compact galaxies selected by guzman et al ( 1997 ) , the lcgs selected in cfrs fields are mostly evolved starbursts probably similar to those classified as such in the hdf . due to their large derived sfrs ,
lcgs are the main contributors to the large sfr density observed at z=0.5 to 1 .
in future papers ( gruel et al . in preparation ) we will investigate the velocity fields and dynamics of lcgs , using data of superior quality and based on a larger sample .
we will also be studying the possible revision of the star formation history and of the ir luminosity density caused by the lcg population .
we are grateful to c. balkowski , f. combes , r. guzman , d. kunth and c. vanderriest for useful discussions and advices .
we are especially indebted to pascale jablonka who introduced us to the methodology of fitting absorption line spectra by star cluster synthesis .
we thank the eso allocation time committee for the one night attributed to this program at ut1 .
f.h . is very grateful to claire moutou and thomas szeifert for their help at paranal and their patience .
we thank the data flow operations team for the fast delivery of the data .
we thank the cnrs / conicyt for financial support of our collaboration .
t.x.t . is grateful for the hospitality of the departement dastronomie extragalactique et de cosmologie at the observatoire de meudon and the institut dastrophysique during his sabbatical leave .
he thanks the partial financial support of the centre national de la recherche scientifique , of the universit of paris vii and of a sesquicentennial fellowship from the university of virginia .
bertin , e. , & arnouts , s. 1996 , , 117 , 393 bica , e. , & alloin , d. 1986a , , 162 , 21 bica , e. , & alloin , d. 1986b , a&as , 66 , 171 bica , e. 1988 , , 195 , 76 boissier , s. , & prantzos , n. 2000 , , submitted brinchmann , j. , et al .
1998 , , 499 112 brinchmann , j. , & ellis , r. 2000 , , 536 , 77 bruzual , g. , & charlot , s. 2000 , in preparation charlot , s. , 1996 , in the universe at high - z : large scale structures and cosmic microwave background , ed .
martinez gonzales & j. sanz ( berlin : springer ) , 53 charlot , s. , 1998 , in the next generation space telescope : science drivers and technological challenges , 34th liege astrophysics colloquium , 135 ( astro - ph/9810408 ) charlot , s. , worthey , g. , & bressan , a. 1996 , , 457 , 625 edmunds , m. , & pagel , b. , 1984 , , 211 , 507 elmegreen , b. , 1999 , , 517 , 103 flores , h. et al .
1999 , , 517 , 148 fukugita , m. , hogan , c. , & peebles , p.j.e . , 1998 , , 503 , 518 gallagher , j. , bushouse , h. , & hunter , d. , 1989 , , 97 , 700 glazebrook , k. , peacock , j. , miller , l. , collins , c. , 1995 , 275 , 169 guzman , r. , gallego , j. , koo , d.c . , phillips , a.c . , lowenthal , j.d . , faber , s.m . ,
illingworth , g.d .
, & vogt , n.p .
1997 , , 489 , 559 guzman , r. , 1999 , in proceedings of the xixth moriond conference on `` building galaxies : from the primordial universe to the present '' , ed .
f. hammer , t.x .
thuan , v. cayatte , b. guiderdoni & j.t.t .
van ( paris : editions frontires ) , hammer , f. , crampton , d. , le fvre , o. , & lilly , s.j .
1995 , , 455 , 88 hammer f. , et al .
1997 , , 480 , 59 .
hammer f. , & flores h , 1998 , in proceedings of the xviiith moriond conference on `` dwarfs galaxies and cosmology '' , ed .
thuan , c. balkowski , v. cayatte & j.t.t .
van ( paris : editions frontires ) , 353 ( astro - ph/9806184 ) hammer , f. 1999 , in asp conf .
197 , `` dynamics of galaxies : from the early universe to the present '' eds .
f. combes , g. mamon and v. charmandaris ( san francisco : asp ) , in press jacoby , g. , hunter , d. , & christian , c. , 1984 , , 56 , 257 kennicutt , r. , tamblyn , p. , congdon , c. , 1994 , , 435 , 22 kennicutt , r. 1998 , , 36 , 189 koo , d. , & kron , r. 1981 , , 105 , 107 koo , d. , guzman , r. , faber , s. , illingworth , g.d . , bershady , m.a . ,
kron , r.g .
, & takamiya , m. 1995 , , 440 , 49 kobulnicky , h. , & zaritsky , d. , 1999 , , 511 , 118 kobulnicky , h. , kennicutt , r. , & pizagno , j. 1999 , , in press le fvre et al .
2000 , , 311 , 565 lehnert , m. & heckman , t. , 1996 , 472 , 546 lilly , s.j . ,
hammer , f. , le fvre , o. , & crampton , d. 1995 , , 455 , 75 lilly s. , le fvre o. , hammer f. , & crampton , d. 1996 , , 460 , l1 lilly , s.j .
1998 , , 500 , 75 mcwilliams , a. , 1997 , 35 , 503 phillips , a. , guzman , r. , gallego , j. , koo , d. , lowenthal , j.d . ,
vogt , n.p . ,
faber , s.m . , & illingworth , g.d .
1997 , , 489 , 543 renzini , a. , & cimatti , a. , 1999 , astro - ph/9910162 schade , d. , et al .
1999 , , 525 , 31 weedman , d. 1983 , , 266 , 479 wheeler , j. , sneden , c. , truran , j. , 1989 27 , 279 williams , r. , et al , 1996 , , 112 , 1335 ccrrrrrrrrrc + cfrs & z & m@xmath69 & @xmath70 & err & m@xmath71 & @xmath72 & err & @xmath73@xmath74 & @xmath22i@xmath75 & @xmath22i@xmath76 & log(r@xmath77 + + 03.0442 & 0.47807 & -20.30 & 0.436 & 0.064 & -21.06 & 0.254 & 0.084 & 2.2 & 0.68 & & + 03.0523 & 0.65355 & -21.20 & 1.000 & 0.147 & -22.24 & 0.752 & 0.194 & 6.5 & 0.43 & 0.46 & 0.58 + 03.0570 & 0.64682 & -20.38 & 0.470 & 0.117 & & & & & 0.30 & & + 03.0595 & 0.60442 & -20.78 & 0.680 & 0.119 & -22.09 & 0.655 & 0.157 & 5.6 & 0.31 & 0.54 & 0.65 + 03.0645 & 0.52737 & -20.77 & 0.673 & 0.049 & -22.20 & 0.724 & 0.113 & 6.3 & 0.62 & 0.64 & 0.69 + 03.1242 & 0.76786 & -21.21 & 1.009 & 0.223 & -22.99 & 1.500 & 0.317 & 12.9 & 0.73 & & + 03.1349 & 0.61640 & -21.43 & 1.236 & 0.114 & -23.67 & 2.805 & 0.206 & 24.2 & 0.73 & 0.35 & 0.52 + 03.1540 & 0.68931 & -21.52 & 1.343 & 0.124 & -22.83 & 1.294 & 0.274 & 11.2 & 9999 & 0.56 & 0.67 + 22.0344 & 0.51680 & -20.30 & 0.437 & 0.048 & & & & & 0.62 & & + 22.0429 & 0.62433 & -20.29 & 0.433 & 0.083 & -22.13 & 0.679 & 0.144 & 5.9 & 0.57 & & + 22.0619 & 0.46706 & -20.08 & 0.356 & 0.082 & & & & & 0.75 & & + 22.0637 & 0.54188 & -21.26 & 1.057 & 0.107 & -21.83 & 0.515 & 0.099 & 4.4 & 0.61 & & + 22.0919 & 0.47144 & -20.22 & 0.406 & 0.034 & & & & & 9999 & 0.37 & 0.42 + 22.1064 & 0.53685 & -20.06 & 0.350 & 0.074 & -22.43 & 0.895 & 0.214 & 7.7 & 0.21 & &
+ + ccrcrcrcr + cfrs & fwhm@xmath80 & q@xmath78 & fwhm@xmath81 & q@xmath78 & fwhm@xmath82 & q@xmath78 & fwhm@xmath74 & q@xmath78 + + 03.0442 & & 9 & 50 & 3 & 50 & 1 & 50 & 1 + 03.0523 & 115 & 2 & 150 & 2 & 180 & 1 & 180 & 1 + 03.0570 & 50 & 3 & 79.5 & 3 & 130 & 1 & 130 & 1 + 03.0595 & 54 & 4 & 150 & 2 & 50 & 2 & 50 & 2 + 03.0645 & 50 & 2 & 50 & 2 & 55 & 2 & 55 & 2 + 03.1242 & 135 & 3 & 81 & 3 & & 9 & 108 & 2 + 03.1349 & 252 & 2 & 60 & 2 & 250 & 2 & 250 & 1 + 03.1540 & 50 & 3 & 50 & 3 & 60 & 2 & 60 & 1 + 22.0344 & & 9 & 50 & 2 & 50 & 2 & 50 & 2 + 22.0429 & 52 & 4 & 50 & 3 & 50 & 2 & 50 & 2 + 22.0619 & 50 & 3 & 50 & 2 & 50 & 1 & 50 & 1 + 22.0637 & 89 & 4 & 50 & 2 & 50 & 1 & 50 & 1 + 22.0919 & & 9 & 50 & 1 & 50 & 1 & 50 & 1 + 22.1064 & & 9 & 76 & 2 & 82 & 2 & 80 & 1 + + ccccccccccc + cfrs & h10 & h8 & caiik & fei & cn & cai & fei & g & fei & mgi+mgh + + @xmath14 & 3795 & 3890 & 3933 & 4137 & 4182 & 4229 & 4264 & 4301 & 4392 & 5176 + 03.0442 & 9999.0 & 9999.0 & 9999.0 & 1.825 & 4.121 & 2.232 & 9999.0 & 9999.0 & 5.500 & 9999.0 + 03.0523 & 9999.0 & 5.262 & 8.413 & 0.878 & 9999.0 & 1.537 & 1.972 & 2.372 & 8.273 & 9999.0 + 03.0570 & 11.172 & 9999.0 & 13.042 & 3.825 & 9999.0 & 6.507 & 8.937 & 6.227 & 10.147 & 9999.0 + 03.0595 & 9999.0 & 10.948 & 9999.0 & 2.843 & 7.734 & 2.543 & 9999.0 & 9999.0 & 6.808 & 9999.0 + 03.0645 & 8.185 & 9999.0 & 7.354 & 9999.0 & 9999.0 & 1.462 & 3.613 & 5.967 & 5.868 & 9999.0 + 03.1242 & 6.647 & 9999.0 & 6.758 & 6.176 & 14.247 & 7.236 & 10.440 & 9999.0 & 13.837 & 9999.0 + 03.1349 & 2.672 & 9999.0 & 9999.0 & 2.031 & 4.828 & 3.384 & 9999.0 & 3.662 & 3.640 & 9999.0 + 03.1540 & 9999.0 & 7.031 & 5.142 & 2.660 & 2.510 & 3.945 & 3.836 & 6.328 & 7.587 & 9999.0 + 22.0344 & 9999.0 & 9999.0 & 9999.0 & 9999.0 & 9999.0 & 3.206 & 3.202 & 7.177 & 9.421 & 8.239 + 22.0429 & 4.091 & 9999.0 & 9999.0 & 2.591 & 5.497 & 9999.0 & 0.314 & 3.806 & 9.394 & 9999.0 + 22.0619 & 9999.0 & 3.185 & 5.668 & 0.898 & 5.983 & 3.679 & 4.410 & 9999.0 & 6.773 & 9999.0 + 22.0637 & 2.596 & 4.662 & 2.309 & 9999.0 & 1.107 & 0.691 & 0.880 & 0.833 & 9999.0 & 9999.0 + 22.0919 & 9999.0 & 9999.0 & 9999.0 & 3.414 & 1.569 & 1.751 & 9999.0 & 9999.0 & 3.373 &
4.118 + 22.1064 & 9999.0 & 9999.0 & 9999.0 & 9999.0 & 3.782 & 4.684 & 4.969 & 3.799 & 3.081 & 6.119 + + cccccccccc + cfrs & @xmath83 & n@xmath74 & age & @xmath84 & @xmath85 & @xmath86&@xmath87 & @xmath88 & @xmath89 + 03.0442 & 0.029 & 4 & @xmath56@xmath78 & -1.0 & -0.5 & & -0.5 & -0.5 & + & & & @xmath90 & 15 & * 75 & 0 & 5 & 5 & 0 + 03.0523 & 0.250 & 5 & @xmath56 & & 0.6 & & & -0.5 & + & & & @xmath90 & 0 & * 50 & 0 & 0 & * 50 & 0 + 03.0570 & 0.361 & 5 & @xmath56 & 0.6 & 0.6 & & -0.5 & & + & & & @xmath90 & 10 & * 80 & 0 & 10 & 0 & 0 + 03.0595 & 0.229 & 5 & @xmath56 & -1.5 & 0.6 & & & -0.5 & + & & & @xmath90 & 25 & * 60 & 0 & 0 & 15 & 0 + 03.0645 & 0.471 & 6 & @xmath56 & -1.5 & 0.6 & & & -0.5 & 0.0 + & & & @xmath90 & 35 & * 50 & 0 & 0 & 10 & 5 + 03.1242 & 1.530 & 6 & @xmath56 & & 0.6 & & & -0.5 & + & & & @xmath90 & 0 & * 95 & 0 & 0 & 5 & 0 + 03.1349 & 0.286 & 5 & @xmath56 & -1.5 & 0.6 & & & & 0.0 + & & & @xmath90 & * 50 & 30 & 0 & 0 & 0 & 20 + 03.1540 & 0.915 & 6 & @xmath56 & -1.5 & -0.5 & & & & + & & & @xmath90 & 10 & * 90 & 0 & 0 & 0 & 0 + + * * * * * * * * * crrrrccl + cfrs & @xmath91 & err & a@xmath92 & err & sfr@xmath93 & sfr@xmath94 & err + + 03.0442 & 2.64 & 0.52 & 1.53 & @xmath95 & 1.52 & 11.54 & @xmath96 + 03.0523 & 2.68 & 0.24 & 1.65 & @xmath97 & 4.20 & 107.35@xmath79 & @xmath98 + 03.0570 & 2.18 & 0.43 & 0.15 & @xmath99 & 1.80 & 3.27 & @xmath100 + 03.0595 & 2.45 & 0.48 & 1.00 & @xmath101 & 2.00 & 18.70 & @xmath102 + 03.0645 & 2.64 & 0.17 & 1.54 & @xmath103 & 3.33 & 37.76 & @xmath104 + 03.1242 & & & & & & & + 03.1349 & 2.98 & 0.45 & 2.42 & @xmath105 & 3.79 & 134.38@xmath79 & @xmath106 + 03.1540 & 3.04 & 0.36 & 2.56 & @xmath107 & 4.39 & 88.50@xmath79 & @xmath108 + 22.0344 & 2.76 & 0.22 & 1.86 & @xmath109 & 1.67 & 27.57 & @xmath110 + 22.0429 & & & & & & & + 22.0619 & 2.29 & 0.22 & 0.51 & @xmath111 & 1.04 & 9.62 & @xmath112 + 22.0637 & 2.37 & 0.19 & 0.76 & @xmath109 & 3.69 & 5.25 & @xmath113 + 22.0919 & 2.25 & 0.08 & 0.38 & @xmath114 & 2.41 & 8.42 & @xmath115 + 22.1064 & 2.78 & 0.19 & 1.91 & @xmath116 & 1.61 & 37.66 & @xmath117 + + notes : + @xmath74 sfr in [ m@xmath118 , without correction for extinction + @xmath78 sfr in [ m@xmath118 corrected from extinction calculated from the @xmath30/@xmath31 ratio + @xmath79 also detected by iso ( flores et al , 2000 , in preparation ) .
+ ccrcrcrrrr + cfrs & @xmath119 & r@xmath120 & oiii ratio@xmath79 & t@xmath121[k ] & e(t)@xmath122 & @xmath123 & @xmath124 & @xmath125 & err + + 03.0442 & 1.10 & 0.22 & 65 & 15000 & 3 & 8.2 & 8.4 & 8.3 & 0.1 + 03.0523 & 0.76 & 0.57 & 233 & 11000 & 3 & 8.0 & 8.5 & 8.3 & 0.2 + 03.0570 & 0.69 & 0.76 & 149 & 12000 & 3 & 7.9 & 8.7 & 8.3 & 0.5 + 03.0595 & 0.31 & 0.30 & 50 & 18000 & 3 & 7.6 & 8.9 & 8.3 & 0.6 + 03.0645 & 0.79 & 0.56 & 111 & 12500 & 3 & 8.3 & 8.55 & 8.4 & 0.15 + 03.1349 & 0.60 & 0.59 & 174 & 11500 & 3 & 8.0 & 8.8 & 8.4 & 0.4 + 03.1540 & 0.57 & 0.05 & 30 & 22000 & 3 & 8.2 & 8.9 & 8.5 & 0.3 + 22.0344 & 0.67 & 0.31 & 119 & 12500 & 3 & 8.0 & 8.8 & 8.4 & 0.4 + 22.0619 & 0.87 & 1.54 & 116 & 12500 & 2 & 8.3 & 8.2 & 8.2 & 0.05 + 22.0637 & 0.72 & 0.58 & 75 & 14500 & 3 & 8.2 & 8.6 & 8.4 & 0.2 + 22.0919 & 0.86 & 33.24 & 105 & 13000 & 1 & 7.8 & 8.8 & 7.8 & 0.05 + 22.1064 & 0.73 & 0.97 & 40 & 21000 & 3 & 7.85 & 8.8 & 8.3 & 0.5 + + notes : + @xmath74 see text .
+ @xmath78 ionization + @xmath79 @xmath555007+[oiii]4959/[oiii]4363 $ ] ratio + @xmath122 quality factor : 1 excellent , 2 estimated , 3 temperature limit + @xmath126 o / h : 1 lower value and 2 higher value ( cf .
kobulnicky , 1999 ) + | vlt spectra of 14 luminous compact galaxies ( lcgs ) reveal strong metallic absorption line sytems as well as narrow and intense emission lines .
their gas extinction is found to be large ( a@xmath0 @xmath1 1.5 mag ) leading to an upward revision of their star formation rate ( sfr ) to an average value of @xmath1 40 @xmath2 .
large extinction values are also supported by the large rate of detection in one field observed by iso .
gas metal abundances in lcgs have about half the solar value .
lcg absorption spectra can be synthesized with a mix of a few gyr old and relatively metal - rich ( generally solar to over - solar values ) stellar population and a younger stellar population ( @xmath3 5@xmath4 years ) having a metal abundance similar to that of the gas .
+ we argue that lcgs are the progenitors of present - day spiral bulges .
lcgs have masses and light concentrations similar to those of present - day bulges .
they could have been formed entirely during a period of a few gyr prior to the epoch of their observations if the star formation has been sustained at the observed rate . as in present - day galactic bulges ,
lcg stars show a wide range of abundances .
thus , observing lcgs allows us to directly witness an important stage in the formation of a massive galaxy , the building of the bulge prior to that of the disk .
the gas needed to feed the observed star formation is likely to be falling in from the outskirts of the galaxy , being tidally pulled out from interacting companion galaxies .
an infall scenario naturally explains the gas metal abundance which is generally lower than that of the older stellar component .
at least for the strongest star - forming lcgs , there is clear imaging evidence for the presence of companions .
some lcgs also show evidence for the beginning of a disk formation .
+ if the above scenario holds for most lcgs , we estimate that at least 20% of present - day spiral galaxies have formed the bulk of their stars at relatively recent epochs , during the last 8 - 9 gyr , at redshifts less than @xmath1 1 .
since they are heavily extincted , we predict their ir luminosities to be relatively large , around @xmath5=@xmath6 @xmath7 , i.e. near or slightly below the luminosities of the galaxies detected by iso in the same redshift range . taking into account the integrated ir luminosity of the lcg galaxy population can lead to a significant upward revision of the cosmic sfr density in the redshift range from 0.5 to 1 . |
the discovery that , according to quantum field theory in its general curved spacetime version , black holes should radiate away their energy @xcite , has had profound impact on both our understanding of the interplay between gravitation and quantum physics teaching us for instance that the laws of black hole mechanics are in fact the laws of thermodynamics when applied to situations involving black holes and in contributing to our realization that there is much that we still need to understand @xcite . regarding the latter we are referring , of course , to what is commonly known as the `` black hole information paradox '' .
there have been many attempts to address this issue on the basis of proposed theories incorporating quantum treatments of gravitation , and it is fair to say that none of those seem to offer truly satisfactory resolution . for
a review see for instance @xcite .
in fact , there is even a controversy as to whether there is or there is not , a paradox or some open issue that needs confronting . in @xcite
, this question has been discussed and clarified .
the basic issue that seems to underlie the various postures in this respect is associated to the view that one takes regarding the nature of the singularity that is generically found deep in the black hole interior .
if one views this singularity as a fundamental boundary of spacetime , there is in fact no paradox whatsoever , as one can say that information either is registered on " or else escapes through " that boundary . on the other hand ,
if one views , as do most researchers working in the various approaches to quantum gravity , the singularity as something that must be ultimately cured " by an appropriate quantum theory of gravitation , in the sense of replacing it by something amenable to treatment by such theories and not as any kind of essential boundary ( for instance the proposal within loop quantum gravity discussed in @xcite ) , one must explain how to reconcile the unitarity of quantum mechanical evolution ( a feature that among other things requires reversibility and thus the preservation of information ) with the thermal nature of the hawking radiation . without any reasonable reconciliation of the divergent conclusions
one would be entitled to describe the situation as a paradox . in order to explore the most explicit version of the problem it is customary to consider a black hole formed by the gravitational collapse of a large lump of matter ( with mass of the order of , say , a few solar masses ) characterized by a pure quantum mechanical state .
the problem one faces then is to reconcile the purity of the initial state with the thermal nature of the hawking radiation .
the issue has been studied extensively ( see for instance the nice reviews @xcite , or the works @xcite ) as it is considered one of the major challenges of contemporary theoretical physics the approaches that have been considered in the search for such a reconciliation seem to be relatively limited
. faced with the fundamental assumption ( * ) : _ the validity of quantum field theory in curved space - times , at least , in regions where curvature is far from the planckian regime _ , these approaches essentially represent variations of the following ideas : \1 ) somehow , during the late time " part of the black hole evaporation , the hawking radiation is not truly thermal , and is in fact highly correlated with the early - time radiation ( which must remain thermal due to ( * ) ) , so that the full state of the radiation field is pure .
\2 ) the black hole evaporation is not complete , and modifications associated with quantum gravity lead to the formation of a stable remnant .
such a remnant would have to be in a highly entangled state with the emitted radiation so that the full state of the ( radiation field + remnant ) , is pure .
\3 ) the hawking radiation is thermal all the way until the eventual evaporation of the black hole but the information somehow crosses the region where the singularity would have been found , and that is now described in terms of the quantum gravity theory .
alternative proposals might involve some combinations of the three proposals above .
however it seems clear that , at least one of them should be able to account for the fate of most of the information .
that is , if none of them can account for anything more than a very small fraction thereof , then the three alternatives together will not be able to account for more than a slightly larger fraction of the full information that needs to be recovered , if the process is to be compatible with the unitarity of quantum mechanics .
the fact is that each of these 3 alternatives have serious drawbacks .
\1 ) this idea , which is generally framed within the context of the so called black hole complementarity proposals @xcite , has been the subject of recent detailed studies which show , based on the so called monogamy of quantum entanglement , that one of the consequences of such entanglement ( even forgetting for the moment the question of how would such correlations be generated ) would be the formation of firewalls " @xcite ( or regions of divergent energy momentum of the quantum field ) around the black hole horizon .
\2 ) here the issue is that one would be postulating the existence of peculiar kinds of objects , the remnants , typically with a mass of few times the planck mass , which must have an enormous number of internal states , essentially as many as those of a large star .
that is because the full state of the ( radiation field + remnant ) must be pure while the reduced density matrix characterizing the radiation field is thermal , and has an energy content of few solar masses .
\3 ) this alternative seems to be favored by researchers working in loop quantum gravity , and has been considered in some detail in @xcite . here , there are two issues that need to be clarified .
first , one needs to explain precisely how the information crosses the quantum gravity region that replaces the classical singularity , in particular given that in the lqg context , that region seems to be characterized by signature changes in the metric @xcite .
secondly , there seems to an even more problematic aspect of this proposal : the fact that , after the complete evaporation of the black hole , the information missing in the thermal radiation would have to be encoded in the quantum gravity degrees of freedom ( dof ) , which however would have an essentially vanishing associated energy . that is , the quantum gravity dof would have to be entangled with the hawking radiation in such a way that the complete state of the quantum gravity sector plus the quantum matter field sector , would be pure , and yet the energy would be essentially all in the radiated quanta of the field .
we must note that there have been other proposals such as those considered in @xcite but we feel it is fair to say that none of these has gained any kind of universal acceptability within the community interested in the issues , as each faces some difficulties of its own .
we want to explore a possible resolution of the paradox , by assuming that qg would indeed replace the singularity by something else , suitably described in terms of the fundamental dof of such a theory , but that quantum theory would have to be modified along the lines of the proposals put forward to address the measurement problem " by treating collapse of the wave function as a physical process , occurring spontaneously and independently of observers or measuring devices " , and that the corresponding modification is such , as suggested in @xcite , that essentially all the initial information is actually lost .
the first aspect we must note about the general proposal is that its setting is within the general context of semi - classical gravity .
that is , a scheme where the gravitational degrees of freedom are treated using a classical spacetime metric , while the matter degrees of freedom are treated using the formalism of quantum field theory in curved space - times@xcite .
the first reaction of many people towards this is to cite the paper @xcite which supposedly rules out the viability of semi - classical gravity . here
we must first point out the various caveats raised about such a conclusion in @xcite and note , in particular , that the work in @xcite centered mainly in the consideration of a formulation in which quantum theory did not involve any sort of collapse of the quantum state , a situation that contrasts explicitly with what we will be focusing on .
the other point noted in @xcite is that if one wants to consider semi - classical gravity together with a version of quantum theory involving the collapse of the quantum state , one faces the problem that the semi - classical einstein equation can not hold during a collapse simply because einstein s tensor is by construction divergenceless while the expectation value of the energy momentum tensor would generically have a non - vanishing divergence .
the point is that one can view semi - classical gravity , not as a fundamental theory , but as providing a suitably approximated description in limited circumstances , something akin to say the hydrodynamical description of a fluid which , as we know , corresponds only to the description of something that at a deeper level needs to be described in terms of molecules moving and interacting among themselves in rather complex ways . following the analogy
, we view the metric description of gravity and the characterization of the matter sector using quantum field theory ( and connected to gravity via einstein s semi - classical equations ) just as an approximated description of limited validity .
in fact this is a point of view that has been explored in the cosmological setting to deal with certain difficulties that arise in the inflationary cosmological account for the emergence of the seeds of cosmic structure @xcite .
the introduction of dynamical collapse within the general framework can be treated in a scheme where one allows an instantaneous violation of the equations , in association with the collapse of the quantum state taking place on a given spatial hypersurface , and in analogy with israel s matching conditions @xcite requiring continuity of the metric across such a hypersurface .
the details of that formalism were first described in @xcite .
we will not discuss these issues further here as they have been thoroughly treated in the above reference and also in the previous works by some of us on the black hole information problem @xcite .
here it is worth pausing to reconsider in more detail certain aspects of the discussion around the issues of energy content .
the setting of the discussion is that of black holes in asymptotically flat space - times . for these space - times
we have a well defined notion of adm mass which is taken as the covariant measure of the energy content of the spacetime , and the quantity that is conserved in the sense that the evaluation of the adm mass gives the same number when computed using any cauchy hypersurface @xmath0 ( which in the extended spacetime ends at @xmath1 ) .
moreover as the spacetime extensions also include the regions @xmath2 and @xmath3 i.e asymptotic future and past null infinity respectively ) one can use the notion of bondi mass associated with any hypersurface @xmath4 ending at a section @xmath5 , ( that is , @xmath4 , together with the segment of @xmath2 starting at @xmath1 and ending at @xmath6 would be a cauchy hypersurface ) .
the point is that the bondi mass at @xmath6 should be equal to the initial adm mass of the spacetime minus the amount of energy that has been radiated to the segment of @xmath2 starting at @xmath1 and ending at @xmath6 .
we must now clarify in what sense are we going to be using the notions of adm mass and bondi mass as being associated with cauchy hypersurfaces @xmath0 and partial cauchy hypersurfaces @xmath4 .
let us concentrate for a moment on the spacetime that lies well to the past of the singularity ( or the would be singularity that presumably is cured by qg ) .
the point is that , although formally the expression for say the adm mass is associated with an integral at infinity " the behavior of the metric variables at infinity is conditioned by the energy momentum content associated with the matter fields by the einstein equation . in other words
we can compute the adm mass using the cauchy data for the gravitational sector on @xmath0 , data which are tied to the energy momentum of the matter fields through the hamiltonian and momentum constraints . in
that regard we might want to understand how the various components of the matter field contribute to these constraints in each one of the hypersurfaces in question .
we can say , for instance , that associated with the initial set up , we have a cauchy hypersurface @xmath7 ( see fig .
[ penrose4d ] ) where we have a large lump of matter with a spacetime that is only very mildly curved and characterized by a pure quantum mechanical state of the matter fields , the adm mass is @xmath8 which , as we noted , is of the order of a few @xmath9 . in that case
we would say that the energy of the spacetime , is represented almost completely by that encoded in the energy momentum on the matter fields . at relatively late times , but still at the past of the singularity , we might consider a cauchy hypersurface @xmath10 , that starts at @xmath1 stays close to @xmath2 and finally enters the horizon and ends at the center of the gravitationally collapsed lump of matter which at this stage is well within the black hole horizon . alternatively , we might consider deforming @xmath10 into a hypersurface that ends on the section @xmath11 which we call the hypersurface @xmath12 ( this hypersurface is , of course , not a cauchy hypersurface ) .
if we want now , to account for the energy content in terms of data on @xmath10 , as represented by @xmath13 ( which should be equal to @xmath8 ) , we would have to say that there is a very important component of the energy content , corresponding to the energy momentum tensor of the outgoing hawking radiation , located in the part of @xmath10 which lies on the region exterior to the horizon , while the energy contained in the original lump of matter has been red - shifted by the gravitational potential associated with the black hole , and at the same time there is a negative contribution to the energy content associated with the in - falling counterpart of the hawking radiation , which might be considered as also lying in the proximity of the intersection of the event horizon with @xmath10 .
the situation is depicted , for the realistic 4 dimensional case in figure [ penrose4d ] , and for the 2 dimensional cghs model in [ cghs ] . in terms of @xmath12
we would say that we need to account for the relatively small value of the bondi mass at its endpoint @xmath11 , a value that is obtained by subtracting from @xmath8 the energy carried away by the hawking radiation that has reached @xmath2 to the past of @xmath6 .
that small value of the bondi mass would , in turn , be accounted for , in terms of the data on @xmath12 , as resulting from the red shifted energy of the original lump of matter , and the negative contribution associated with the in - falling counterpart of the hawking radiation .
we note that the situation on @xmath12 , as far as energy is concerned is very similar to that on any hypersurface characterizing the situation well after the evaporation of the black hole such as @xmath14 in the accompanying figure .
now , we might in a similar way , want to consider the fate of the information in the picture above .
that is we want to consider how is the information , that was present in the quantum state characterizing the initial state of system at @xmath15 , accounted for , in terms of the quantum state characterizing the system on @xmath10 ?
the point is that , by deforming @xmath16 into @xmath17 ( where @xmath18 is the part of @xmath2 to the past of @xmath6 ) , we can describe the state , as an entangled state , which on @xmath19 is just the hawking thermal radiation , and on @xmath20 is also a highly mixed density matrix , but such that the complete state is pure .
the point that we want to make here is that the above situation seems to be afflicted by the same troublesome aspects which were raised in the context of the alternative 3 ) above .
that is , the state of the system on @xmath20 is one with an enormous number of degrees of freedom and yet a very small value of the energy .
it seems therefore that if we have an explanation for the loss of information in the black hole evaporation that relies on losses associated only with the qg region ( i.e. losses that , in the gr language , would be described as produced by the singularity ) we would still face the uncomfortable aspects that lead to the rejection of alternative 3 ) above , but this time associated with the situation prior to the singularity ( i.e. for instance the situation on @xmath12 ) .
we think that in addressing the problem via the introduction of modifications of quantum theory that last problem is dramatically ameliorated .
the notion that one could learn to live with information being lost in association with the evaporation of black holes , has been considered in some detail in @xcite , where the earlier arguments @xcite indicating that such proposals would necessarily involve large violations of known conservation laws or dramatic violations of causal behavior have been dispelled . in that analysis
, however , the resulting picture seems to be that all the information loss occurs at the singularity , i.e at the region that would be described in non - metric terms in a quantum theory of gravitation , while , in the regions where the metric description would be appropriate , one would have exact quantum mechanical unitary evolution .
our view is that such an approach offers a less unified view of physics than the one we are advancing , and that , as a result , it might be more vulnerable to questions of self consistency .
for instance , if we accept that there are violations to the quantum mechanical unitary evolution , but that those only occur in connection with black hole evaporation , we might have problems , with an ultimate quantum gravity theory , that can be expected to include the possibility of processes involving virtual black holes .
in other words , we might have to face up to the expected result of such a theory indicating that all physical processes must involve contributions from all possible intermediate states according to a path integral formulation of the process , and that those intermediate states would involve also virtual black holes , which in turn would have associated violations of unitarity .
as first discussed in @xcite , this kind of problem seems less likely to arise in a more unified version we are considering , where the violation of unitarity is an integral aspect of the fundamental physical laws as envisaged in the various proposals for modifications of quantum theory that have been made in the context of the search for a resolution of the so called measurement problem " @xcite - @xcite . it is worthwhile reminding the reader that the so called measurement problem in quantum theory is tied to the interpretational difficulties that arise when one does not want to introduce , in the treatment , some artificial classical/ quantum cut ( sometimes presented as a macro/ micro physics cut ) and instead , one wants to consider that everything , including potential observers and measuring apparatuses , should be treated in a quantum mechanical language .
we direct the reader to the works in @xcite for a good overview , to @xcite for a more extensive collection of postures , or to @xcite for a very clear recent analysis .
the relevance of this issue to the problem at hand can be seen from the fact that quantum theory calls for purely unitary evolution only when one is dealing with a completely isolated system in which all degrees of freedom are treated quantum mechanically .
many proposals to deal with the general interpretational difficulties of quantum theory , and in particular with the measurement problem have been considered since the inception of the theory @xcite and there is also a good body of literature devoted to the problems of many of these proposals @xcite .
we want to focus on the dynamical reduction theories which involve a modification of quantum dynamics involving spontaneous reduction of the quantum state .
these kind of proposals are commonly known as collapse theories " and have a rather long tradition . for recent reviews
see @xcite .
recently various relativistic versions of spontaneous dynamical collapse theories have been put forward @xcite , @xcite .
we could not end this introduction without acknowledging the strong inspiration that we have drawn from r. penrose s discussions connecting foundational aspects of quantum theory to ideas about the nature of quantum gravity @xcite .
in fact in a very early analysis @xcite r. penrose noted that if one wanted to obtain a self - consistent picture of a situation involving thermodynamical equilibrium that included black holes one would need to have a theory of quantum mechanics that incorporated some violation of unitarity in ordinary conditions ( involving no black holes ) .
we view that analysis as providing further support for our approach in contrast to those where violation of unitarity is only associated with the singularity in evaporating black holes .
some of penrose s recent works @xcite on such issues , are also relevant , in a broader sense , to our general views underlying this proposal .
we will present here an concrete version of the above approach based on the theory developed in @xcite .
the article is organized as follows : in section ii we present a brief description of the cghs 2 dimensional model of black hole formation and evaporation , in section iii we present a relativistic model of dynamical collapse , section iv describes the general setting in which we will put together the two elements previously described , and in section v we will use them to describe the evolution of the quantum state of the mater field thus accounting for the loss of information . in section vi we discuss some subtle points regarding the energetic aspects of the proposal and we end in section vii with the general conclusions indicating what has been achieved and what would need be left as issues for further research .
we have added two appendices for the interested reader convenience : appendix a discusses in detail the foliation independence of the proposal , exhibiting its general covariance , and appendix b presents in some detail the manner in which the delicate issue regarding the expectation of unbounded energy creation is resolved by the introduction of the _ pointer field_.
the two dimensional model , first introduced by callan - giddings - harvey - strominger ( cghs ) @xcite involving black hole formation is a very convenient toy model for the study of issues related to the formation and evaporation of two dimensional black holes .
we now review the basic features of this model . for more details
we refer the reader to @xcite .
the cghs action is @xmath21-\frac{1}{2}(\nabla f)^2\right ] } , \nonumber \end{aligned}\ ] ] where @xmath22 is the ricci scalar for the metric @xmath23 , @xmath24 is the dilaton field , considered in this model as part of the gravity sector , @xmath25 is a cosmological constant and @xmath26 is a scalar field , representing matter . the solution corresponding to the cghs model is shown in fig .
[ cg ] .
it corresponds to a null shell of matter collapsing gravitationally along the world line @xmath27 and leading to the formation of a black hole . for @xmath28 ,
this solution is known as the dilaton vacuum ( region i and i ) .
the metric is found to be @xmath29 which is flat , whereas for @xmath30 the solution is described by the black hole metric ( region ii , iii ) represented by @xmath31 where @xmath32 . here ( null ) kruskal - type coordinates ( @xmath33 ) are useful to describe the global structure of the spacetime . on the other hand , for physical studies involving quantum field theory ( qft ) in curved spacetime , it is convenient to use special coordinates for the various regions . in the dilation vacuum region ,
the natural coordinates are @xmath34 , and thus the metric can be expressed as @xmath35 with @xmath36 . in the bh exterior ( region
ii ) , a natural set of coordinates is provided by @xmath37 , so that the metric in this region is @xmath38 with @xmath39 and @xmath40 in order to exhibit the asymptotic flatness , we express the bh metric in schwarzschild - like coordinates ( @xmath41 ) which are defined through the implicit formulas @xmath42 so that we get @xmath43 .
the temporal and spatial kruskal coordinates @xmath44 can be related to schwarzschild - like time @xmath45 and space @xmath46 coordinates , through @xmath47 and @xmath48 .
now we consider the quantum treatment of the matter field @xmath26 .
we will consider the null past asymptotic regions @xmath49 and @xmath50 as the _ in _ region and the black hole ( exterior and interior ) region as the asymptotic _ out _ region . in the _ in _ region ,
the field operator can be expanded as @xmath51 where @xmath52 , and the basis of functions ( modes ) are : @xmath53 and @xmath54 with @xmath55 .
the superscripts @xmath22 and @xmath56 refer to the right and left moving modes respectively .
these modes define the bases of field quantization and thus the right _ in _ vacuum ( @xmath57 ) and the left _ in _ vacuum ( @xmath58 ) whose tensor product ( @xmath59 ) defines our _ in _ vacuum .
as is well known , one might also proceed to the construction of the field theory in terms of modes that are natural in the _ out _ region by expanding the field operator @xmath60 in terms of the complete set of modes having support both outside ( exterior ) and inside ( interior ) the event horizon . once more we can write the field operator in the form @xmath61 where @xmath62 in the above we have used the convention whereby modes and operators with and without tildes correspond to the regions inside and outside the horizon , respectively . for the mode functions in the exterior to the horizon we use : @xmath63 and @xmath64 similarly we can choose the set of modes in the black hole interior , ensuring that the basis of modes in the _ out _ region is complete .
the left moving modes are kept the same as before ( since these modes travel from the black hole exterior to the interior ) , while for the right moving mode we take : @xmath65 .
following @xcite , we now replace the above delocalized plane wave modes by a complete orthonormal set of discrete wave packets modes , given by @xmath66 where the integers @xmath67 and @xmath68 .
these wave packets are naturally peaked about @xmath69 with width @xmath70 respectively .
the next step in our analysis is to consider the bogolubov transformations . in our case
, the relevant non - trivial one refers to the right moving sector , and the corresponding transformation from _ in _ to _ exterior _ modes is what accounts for the hawking radiation .
we note that the initial state , corresponding to the vacuum for the right moving modes and the left moving pulse forming the black hole @xmath71 can be written as : @xmath72 where particle states @xmath73 consist of arbitrary but _ finite _ number of particles , @xmath74 is a normalization constant , and the coefficients @xmath75 s are determined using the bogolubov transformations .
their explicit expressions can be seen in @xcite .
it is well known that , if one ignores the degrees of freedom of the quantum field lying in the black hole interior , and describes just the exterior dof of freedom , one ends up ( partially ) describing the state in terms of a density matrix .
that is , one obtains the reduced density matrix by tracing over the interior degrees of freedom ( dof ) , and in this case one ends up , with a density matrix corresponding to a thermal state .
note , at this point this density matrix represents , in the language of @xcite an _ improper _ mixture , as it arises after ignoring part of the system which as a whole is in a pure state .
we will therefore say that what we obtain at this point is an _ improper thermal state_. see discussion in section @xcite for a more exhaustive discussion and analysis of this issue .
it is also discussed in previous works @xcite that the task of accounting for the information loss in black hole evaporation within the approach we are considering requires among other things showing how , as the result of the dynamics , one ends up with a _ proper thermal state _
( i.e. , one that describes an actual mixed state that is not a partial description of a pure state ) starting with an _ initial pure state_.
for the purpose of presenting relativistic collapse models in generality we employ the interaction picture @xcite in which the quantum state of matter @xmath76 is assigned to a space - like hypersurface @xmath77 . as we advance the hypersurface @xmath77 to the future via some arbitrary foliation of spacetime , the state changes according to @xmath78 where @xmath79 is the interaction hamiltonian density .
the functional derivative is defined as @xmath80 where @xmath81 is the invariant spacetime volume enclosed by @xmath77 and @xmath82 with @xmath83 ( meaning that no point in @xmath77 is to the future of @xmath82 ) .
covariance requires that @xmath84 = 0 $ ] for spacelike separated @xmath85 and @xmath86 .
this guarantees that advancing of the hypersurface across the points @xmath85 and @xmath86 is independent of the order in which this is done , and more generally it guarantees foliation independence of the state development .
the solution to ( [ tom0 ] ) can be written as @xmath87 satisfies @xmath88}{\delta\sigma(x ) } = \hat{\cal h}_{\rm int}(x)\hat{u}[\sigma,\sigma_0],\end{aligned}\ ] ] with initial condition @xmath89 = 1 $ ] .
this can be formally solved to give @xmath90 = t\exp\left [ -i \int_{\sigma_1}^{\sigma_2 } \hat{\cal h}_{\rm int}(x ) dv\right].\end{aligned}\ ] ] where @xmath91 is the time ordering operator , and @xmath92 .
from time to time we suppose that the state undergoes a discrete collapse event associated to a spacetime point @xmath85 .
when the hypersurface @xmath77 crosses the point @xmath85 , the state ceases for an instant to satisfy equation ( [ tom0 ] ) and instead changes according to the rule @xmath93 is the collapse operator at @xmath85 and @xmath94 is a random variable which corresponds to the collapse outcome .
one normally assumes that there is a fixed probability of a collapse event occurring in any incremental spacetime region of invariant volume .
this results in collapse events which have a poisson distribution with density @xmath95 , in any unit volume of spacetime .
this distribution of collapse events in spacetime is covariantly defined and makes no reference to any preferred foliation .
the collapse operators must satisfy the completeness condition @xmath96 this allows us to define the probability density for the outcome @xmath94 , for a collapse event on the state @xmath76 at point @xmath85 , by @xmath97 the completeness condition ensures that ( [ prob ] ) is normalized .
this formula corresponds to the standard formula for the quantum probability of a generalized measurement with measurement operator @xmath98 .
the collapse outcomes thus occur with standard quantum probability . in appendix [
ax : folindep ] we demonstrate that if the following microcausality conditions hold , @xmath99= 0 , \label{com2}\end{aligned}\ ] ] and @xmath100= 0 , \label{com3}\end{aligned}\ ] ] for spacelike separated @xmath85 and @xmath86 , then ( i ) given a poisson distributed set of collapse locations @xmath101 ( with labels @xmath102 , which give an arbitrary total ordering which respects the causal ordering of the spacetime ) occurring between hypersurfaces @xmath103 and @xmath104 , and a compete set of collapse outcomes at these locations @xmath105 , the state dynamics leads to an unambiguous and foliation - independent change of state between @xmath103 and @xmath104 ; and ( ii ) the probability rule specifies the joint probability of complete sets of collapse outcomes @xmath106 independently of spacetime foliation , given only the state on the initial surface @xmath103 .
the joint probability density for the set of outcomes @xmath105 can be determined from ( [ prob ] ) by repeatedly making use of the definition of conditional probability , and is given by @xmath107 where @xmath108 depends on @xmath109 as @xmath110 and where the choice of foliation @xmath111 is arbitrary and corresponds to the arbitrary total ordering of @xmath112 . at this point
one can take the view that the resulting state histories with respect to different foliations are merely different descriptions of the same events @xcite .
alternatively , one can regard the collapse outcomes as the primitives of the theory from which the quantum state histories are derived .
the covariant form of the collapse dynamics together with the absence of any foliation dependence result in an adequate framework for a relativistic collapse model . to realize such a model we must propose a form for a collapse operator @xmath113 which satisfies the above requirements .
we begin by choosing @xmath114 where @xmath115 is an , as yet unspecified hermitian operator , and @xmath116 is a new fundamental parameter .
this collapse operator describes a quasi projection of the state of the system onto an approximate eigenstate of @xmath115 about the point @xmath94 meaning that , if the state previous to the collapse event was represented in terms of eigenstates of @xmath115 , the collapse effect is to diminish the relative amplitude of eigenstates whose eigenvalues are far from @xmath94 with respect to those that have eigenvalues close to @xmath94 . the effect of many such collapses is to drive the state towards a @xmath115-eigenstate .
this collapse operator automatically satisfies the completeness condition .
the microcausality conditions are satisfied if @xmath117 = 0 \quad\text{and}\quad \left[\hat{b}(x ) , \hat{\cal h}_{\rm int}(y)\right ] = 0,\end{aligned}\ ] ] for space - like @xmath85 and @xmath86 .
we therefore propose that , for a theory of a scalar field such as the one we are considering in this work , @xmath118 where @xmath119 is the scalar field operator .
this meets the above conditions for any interaction hamiltonian given as a function of @xmath119 .
however , with this choice we face an immediate problem .
if we calculate the average energy change in the field as a result of a collapse event we find @xmath120|\psi_{\sigma}\rangle}{\langle\psi_{\sigma}|\psi_{\sigma}\rangle } = \frac{1}{2\zeta^2}\delta^{d-1}(0)\langle|\hat{f}(x)|^2\rangle,\end{aligned}\ ] ] for a @xmath121 dimensional spacetime where @xmath122 is the hamiltonian operator for the scalar field , and the final expression is the first order term in the large @xmath116 expansion .
this expression is infinite for a continuum spacetime .
this could be ameliorated by a spacetime with fundamental discreteness ( which should not enter in conflict with special relativity @xcite ) . with discreteness length scale
@xmath123 we could approximate @xmath124 , and might then , by appropriate choices for the parameters of theory , be able to construct a model in which the collapse of massive objects is sufficiently rapid whilst the average energy increase is sufficiently small to satisfy experimental lower bounds @xcite .
( there are three parameters in this model : @xmath116 ; the discreteness length scale @xmath123 ; and the spacetime density of collapse events @xmath95 , which could possibly be taken to correspond to the effective density of spacetime points , reducing the number of parameters to two . )
alternatively we propose the use of a new field to mediate the collapse process with the effect of preventing infinite energy increase .
this construction is outlined in appendix [ ax : aux ] where the effective collapse process satisfied by the scalar field is derived . in either the discrete space model or the auxiliary field model ,
the end result is a collapse model which drives the scalar field towards eigenstates of the operator @xmath125 . as in any event
this is the end result , we will be making free use of it throughout this paper . thus from here on we
will mostly ignore the details of precisely how we deal with the problem of energy increase . to understand the collapse rate we introduce the density matrix representation @xmath126 a collapse event at point @xmath85 on the surface @xmath77 converts the pure state into another pure state with a smaller uncertainty in @xmath127 .
however as the specific state is stochastically determined , it is convenient to pass to a description in terms of ensembles .
that is we consider the statistical mixture representing the ensemble of a large number of identical systems characterized by the same state just before the collapse event , and their collective change , just after such an event .
this is thus described by : @xmath128 } = \int dz \hat{l}_x(z)\hat\rho_{\sigma}\hat{l}_x(z).\end{aligned}\ ] ] this equation describes how the pure state at any stage is transformed into an ensemble of possible resultant states , each element of which results from a particular value of the as yet unknown collapse outcome .
the change in the statistical density matrix operator characterizing the ensemble is then , @xmath129\right],\end{aligned}\ ] ] in the large @xmath116 limit . if we choose a foliation parametrized by @xmath45 , with lapse function @xmath130 and spatial metric on the timeslices @xmath131 , and assume that there is a spacetime collapse density of @xmath95 then we can write @xmath132-\int d^{d-1 } x n\sqrt{h } \frac{\mu}{8\zeta^2 } \left[|\hat{f}(x)|^2,[|\hat{f}(x)|^2,\hat\rho_{t}]\right ] , \label{phicoll}\end{aligned}\ ] ] where @xmath133 stands for the determinant of the components of the metric @xmath131 in the coordinates @xmath134 .
the first term corresponds to the unitary dynamics of the interaction hamiltonian , which would vanish in the case of a free field theory such as the one we are considering .
it is convenient at this point to consider the evolution in terms of a basis of instantaneous field eigenstates for the hypersurface @xmath135 ( corresponding to a leaf of the foliation , @xmath136 _ constant_. that is @xmath137 are field eigenstates on the hypesurface @xmath135 ( i.e. states which satisfy @xmath138 , @xmath139 ) . such states form a complete basis of states for each value of @xmath45 .
it thus follows that , @xmath140\langle f |\hat\rho_t|f'\rangle,\end{aligned}\ ] ] @xmath141 = \int d^{d-1 } x n\sqrt{h } \frac{\mu}{8\zeta^2 } \left[|f(x)|^2-|f'(x)|^2\right]^2 .
\label{rate}\end{aligned}\ ] ] the coupling parameter @xmath142 is usually taken as a constant but as first suggested in @xcite we will assume it is a local function of curvature scalars . for concreteness
we take @xmath143 where @xmath144 is an increasing function of its argument , @xmath145 , and @xmath146 is the weyl tensor for the spacetime metric @xmath147 .
this feature ensures not only that the collapse effects will be much larger in regions of high curvature than in regions where the spacetime is close to flat but it might also be used to ensure that the completely flat regions where among other things the matter content corresponds to the vacuum , the effects of collapse disappear completely . in the two dimensional setting of the cghs model ,
the weyl curvature is zero and so as a substitute we take @xmath148 to be an increasing function of the scalar curvature @xmath22 .
the upshot is that the particular relativistic collapse model determined by the proposal ( [ propose ] ) leads to collapse in the @xmath149 state basis at a rate given by ( [ rate ] ) .
the collapse process will not , however , lead to a precise field eigen - states , simply because the collapse is only assumed to narrow the uncertainty , and the free dynamics of the field , will cause dispersion of the field state in competition with the collapse . in fact , what the result of eq .
( [ rate ] ) shows , is that that states with different @xmath150 are distinguished , rather than states with different @xmath26 .
this would mean , in principle , that at the end of the collapse process .
we would be left with states having a relatively well defined value of @xmath151 but possibly different values of @xmath26 .
we do not think this will be a problem , because once the unitary dynamics and the interactions are taken into account , this kind of situation would be very unstable : any kind of unitary process which distinguished @xmath152 from say @xmath153 , would lead to differences which would be subsequently distinguished by the collapse process .
in fact , a more realistic analysis , where backreaction effects would have to be considered , indicates that energetics will strongly disfavor field configurations with large spatio - temporal fluctuations in the phase .
this follows from the fact such configurations will have a relatively large energy momentum tensor , and thus a large spacetime curvature , and as a consequence , they will be subjected to an increased collapse rate .
the ensuing randomness in the dynamics will only decrease when a configuration with a rather smooth @xmath26 is arrived at .
this is analogous to the effect considered in @xcite .
thus it is natural to expect that ultimately the collapse will be to the @xmath154 basis . on the other hand , as we will be assuming that the collapse rate increases with curvature in an unbounded fashion , we can expect that collapse effects will accumulate and dominate over any dispersion in the high curvature region near to the black hole singularity , more precisely as the quantum gravity region is approached , and thus we will assume that the collapse process leads to a field eigenstate on hypersurfaces that are close enough to that region as an idealization . finally , the particular choice @xmath155 can be justified by demonstrating that this reduces to the well established csl model @xcite , @xcite , in the non relativistic limit of a massive complex scalar field of mass @xmath156 ( the non relativistic limit of a real scalar field or a massless field is less obvious ) .
this can be seen from the well know correspondence @xmath157 the collapse basis is then @xmath158 where @xmath159 is the number density of non relativistic particles . up to a spatial smearing function , this is the collapse basis for the csl model .
the smearing is introduced either by spacetime discreteness or the use of an auxiliary field to mediate the collapse process ( see appendix [ ax : aux ] ) .
the situation we want to consider is that corresponding to the formation of a black hole by the gravitational collapse of an initial matter distribution characterized by a pure quantum state @xmath160 describing a relatively localized excitation of the field @xmath161 .
the spacetime is supposed to be described by a manifold @xmath162 with a metric @xmath23 defined on @xmath162 except for a compact set @xmath163 corresponding to the region where a full quantum gravity treatment is required and that is taken to just surround the location of the classical singularity .
this characterizes the formation and evaporation of an essentially schwarzschild black hole , supplemented by the region @xmath163 that is not susceptible to a metric characterization and where a full quantum theory of gravity is needed to provide a suitable description .
we assume that @xmath164 is a compact boundary surrounding the quantum gravity region , which , by assumption , corresponds to that region where otherwise ( i.e , in the absence of a radical modification of gr due to qg effects ) we would have encountered the black hole singularity .
we will further make some relatively mild ( and rather common ) assumptions about quantum gravity .
\i ) the first assumption , which we have already mentioned , is that qg will cure the singularities of general relativity , however in doing so it will require that there would be regions where the standard metric characterization of spacetime does not apply .
this is what in our case was referred as the set @xmath163
\ii ) we will assume that quantum gravity does not lead at the effective level to dramatic violations of basic conservation laws such as energy or momentum .
\iii ) we will assume that the spacetime region that results at the other side of the qg region is a reasonable and rather simple spacetime .
with these assumptions we can already make some simple predictions about the nature of the full spacetime .
given that by assumption the effects of the collapse dynamics will be strong only in the region with high curvature , and more explicitly in the regions where the value of @xmath165 ( @xmath22 in the two dimensional models ) is large , the dynamics characterizing the early evolution of our initial pulse of matter will be essentially the same as that found in the standard accounts of black hole formation and evaporation : the pulse will contract due to its own gravitational pull , and as shown by birkoff s theorem the exterior region will be described by the schwarzschild metric ; the pulse will eventually cross the corresponding schwarzschild radius , and generate a killing horizon for the exterior time - like killing field @xmath166 . the early exterior region and even the region to the interior of the killing horizon but close to it at early times are regions of small curvature and thus the picture based on standard quantum field theory in curved spacetime that leads to hawking radiation will remain unchanged .
this by itself indicates that essentially all the initial adm mass of the spacetime would be radiated in the form of hawking radiation and will reach @xmath167 ( asymptotic null infinity ) .
next let us consider the spacetime that emerges at the other side of the singularity .
given that essentially all the initial energy has been radiated to @xmath167 and in light of assumptions ii ) above the resulting spacetime should correspond to one associated with a vanishing mass ( this would be the bondi mass corresponding to a spacetime hypersurface lying to the future of region @xmath163 and intersecting @xmath167 in a segment to the future of that containing the hawking flux ) .
this conclusion , together with assumption iii ) indicates that this spacetime region should be a simple vacuum spacetime which we take for simplicity to correspond to a flat minkowski region .
let us now focus on the state of the quantum field @xmath168 .
the initial state , as we indicated , corresponds to the _ in _ vacuum except for a pulse of matter falling under its own gravity and leading to the formation of a black hole .
the state can be represented in the first qft construction in section ii as : @xmath169 where the sum is over the sets of occupation number for all modes @xmath170 ( which indicates that the mode @xmath171 is excited by @xmath172 quanta ) , @xmath173 is the total energy of the state according to the notion of energy associated with the asymptotic region , @xmath174 is the hawking thermal coefficient , and @xmath175 is a normalization constant . at these late times
the excitations associated with the in - falling pulse are all located in the region interior to the killing horizon so that we can write the state ( [ initial state ] ) simply as : @xmath176 where the part in parenthesis corresponds to the black hole interior region and the rest to the exterior .
the point of writing things in this manner is to underscore the fact that both the collapse dynamics and the changes in the state associated with quantum gravity will only affect the modes in the black hole interior region . in the case of the collapse dynamics
this follows from the assumption that the collapse parameter @xmath177 is strongly dependent of curvature and thus its effects will only be relevant in regions of high curvature .
as we have explained , one of the assumptions that underlies the present approach to deal with the information question during the hawking evaporation of the black hole is that the collapse dynamics , although valid everywhere , deviate most strongly from the unitary evolution of standard quantum theory in the regions where curvature becomes large . in the two dimensional context , this is achieved by assuming the that parameter @xmath148 controlling the strength of the modifications is a function of the scalar curvature @xmath22 .
thus the changes to the quantum state of the system result mainly from the nontrivial evolution occurring in the region interior to the black hole horizon , and to the future of the matter shell .
for simplicity we will therefore ignore the modification of the quantum state of the field resulting from the dynamics in the exterior region and the flat region before the matter shell and focus only in the effects of the collapse dynamics in the interior of the black hole lying to the future .
we call this _ the collapse region_. with these considerations we take the initial state at a hypersurface lying well to the past of the collapse region ( for instance on @xmath15 in fig .
[ cghs ] ) .
this can be expressed in terms of the corresponding density matrix as : @xmath178 we can now simplify things using the basis of eigen - states of collapse operators which we will refer to as _ the collapse basis_. we thus rewrite the above density matrix in the form @xmath179 next we use the fact that , in the collapse region , especially in the late part thereof , the collapse dynamics becomes extremely strong and effective and thus drives the state of the system to one of the eigen - states of the collapse operators .
this allows us to write the state representing an ensemble of systems initially prepared in the same state ( [ initial state ] ) , at any hypersurface @xmath10 lying just before the would - be classical singularity or more precisely the quantum gravity region(see fig .
[ cghs ] ) , after the complete collapse process has taken place as , @xmath180 finally , we need to consider the system as it emerges on the other side of the quantum gravity region , i.e the state describing the ensemble after the would - be classical singularity .
as we have discussed in the introduction we assume that quantum gravity would resolve the singularity and lead on the other side of it , to some reasonable spacetime and state of the quantum fields .
we now consider the characterization of the system on a hypersurface lying just to the future of the would - be classical singularity .
such a hypersurface would not be a cauchy hupersurface as it would intersect @xmath2 rather than @xmath1 .
as such one can partially characterize the state of fields on it by the value of the bondi mass .
it is clear , as have argued in the the introduction , that if we assume that quantum gravity does not lead to large violations of energy and momentum conservation laws , the only possible value for this bondi mass would have to be the mass of the initial matter shell minus the energy emitted as hawking radiation , which is present to the past of the singularity on @xmath2 .
this remaining mass will thus have to be very small .
the task for quantum gravity is to turn the internal state , post singularity into a straightforward low energy state . for simplicity
assume that it is the vacuum @xmath181 being chosen by the collapse process .
this means that the final state characterizing the ensemble of systems ( on @xmath104 ) should be of the form : @xmath182 that is , the system has evolved from an initially pure state to state representing the proper thermal state of radiation on the early part of @xmath2 and the vacuum state afterwards .
one of the most serious challenges one faces when attempting to construct relativistic models of spontaneous dynamical reduction of the wave function , either of the discrete or continuous kind , is their intrinsic tendency to predict the violation of energy conservation by infinite amounts : the problem is resolved in the non - relativistic setting where one can easily control the magnitude of that kind of effect , by relying on suitable spatial smearings of the collapse operators , usually taken to be the position operators for the individual particles that make up the system .
when passing to a relativistic context the tendency is for energy violation to become unbounded unless special care is used in the construction of the theory to ensure it does not .
this issue becomes relevant in the present context at two places .
first and foremost at the point where one wants to consider the back reaction of the spacetime metric to the changes in the quantum state of the field @xmath183 induced by the collapse dynamics
. the second place where the issue appears is the point where one considers the role of the quantum gravity region . in a previous treatment
the argument was that , provided that quantum gravity did not result in large violations of energy conservation one can expect the state after the quantum gravity region to correspond energetically to the content of the region just before the would be singularity , and that this region would have almost vanishing energy content being made up of the positive energy contribution of the collapsing matter shell and the negative energy contribution of the in - falling counterpart to the hawking flux .
we would face a serious problem with this argument if the region just before the would - be singularity could contain an arbitrarily large amount of energy as a result of the unboundedness of the violation of energy conservation brought about by the collapse dynamics .
in that case we would not be able to reasonably argue for the step ( [ post - singularity - vacuum ] ) .
there are various schemes whereby this issue can be tackled : \1 ) we might consider a fundamental discreteness of spacetime ( which however as discussed in@xcite should not be tied to violations of special relativity ) .
\2 ) we might adjust the choice of collapse operators and provide a sensible spacetime smearing scheme for them that relies on the energy momentum of the matter fields or on the geometric structure of the curved spacetime . in this context
is it worth noting that when one considers that the parameter controlling the strength or intrinsic rate of the collapse dynamics depends on the spatial curvature ( i.e. the ricci scalar @xmath184 in 2 dimensions and something like the weyl tensor , through @xmath185 , in the more realistic 4 dimensional case ) one might assume that in flat space - times the collapse rate actually vanishes removing most concerns about the stability of the vacuum in these theories . in that case one would adopt the position that the collapse associated with individual particles in the non - relativistic quantum mechanical context is actually derived from the small deformation of flat spacetime associated to that same particle .
that is , one would consider that the particle s energy momentum curves the spacetime and this in turn turns - on the quantum collapse dynamics .
this is only a rough idea at this point but one that certainly seems worthy of further exploration .
\3 ) we might rely on the effective smearing provided by the use of the auxiliary pointer field as a way to introduce the smearing procedure without seriously affecting the simplicity of the treatment as discussed in the appendix [ ax : aux ] below .
we have studied the possibility of accounting for the information loss in the processes of formation and hawking evaporation of a black hole through the explicit use of a relativistic version of a dynamical reduction theory . in previous works
it has been argued that the consideration of theories involving departure from the standard schrdinger unitary dynamics offers a promising path to dealing with what many researchers in the community considered as one of the most challenging paradoxes of modern theoretical physics .
those works were carried out using a non - relativistic version of dynamical reduction theories known as continuous spontaneous localization , and one of the open issues in those treatments was whether similar results could be obtained relying on fully relativistic settings .
the present work provides a positive answer in the form of proof of existence of a relativistic approach that leads essentially to the same results as those of the previous non - relativistic treatments .
however it is clear that we are not yet in the possession of a fully satisfactory scheme . for
that we need to consider in detail the issues related to energy production and its possible back reaction effects .
furthermore eventually one would like to consider the issue of uniqueness and completeness in the sense of determining the collapse operators valid for a general setting that reduces to the appropriate ones ( i.e smeared particle position operators ) in the non - relativistic situations ( the ones treated by the standard csl or grw theories ) , and finding the dependence of the parameters such as @xmath148 on the spacetime curvature .
db is supported by the templeton world charity foundation .
skm is an international research fellow of the japan society for the promotion of science .
ds acknowledges partial financial support from dgapa - unam project ig100316 and by conacyt project 101712 .
consider two collapse events occurring at spacelike separated points @xmath85 and @xmath86 with @xmath186 ( meaning that the points @xmath85 and @xmath86 are not to the past of @xmath103 and not to the future of @xmath104 ) .
an explicit choice of foliation places @xmath85 and @xmath86 in a sequence .
suppose that @xmath85 occurs first on surface @xmath10 and @xmath86 occurs second on surface @xmath14 with @xmath187 .
we therefore have @xmath188 which follows from ( [ prob ] ) by making use of the definition of conditional probability .
now suppose instead that we choose an alternate foliation in which the collapse event at @xmath86 occurs first on surface @xmath12 and @xmath85 occurs second on surface @xmath189 with @xmath190 .
now @xmath191 we now show that given the conditions @xmath99= 0 , \label{acom2}\end{aligned}\ ] ] and @xmath100= 0 , \label{acom3}\end{aligned}\ ] ] for spacelike separated @xmath85 and @xmath86 , then @xmath192 . in order to do this
we define a surface @xmath193 on which both points @xmath85 and @xmath86 are found .
we can then write @xmath194\hat{u}[\sigma_2,\sigma_{xy}]\hat{l}_y(z_y)\hat{l}_x(z_x)\hat{u}[\sigma_{xy},\sigma_1 ] \hat{u}[\sigma_1,\sigma_i]|\psi_{\sigma_i}\rangle \nonumber\\ & = \hat{u}[\sigma_f,\sigma_{xy}]\hat{l}_x(z_x)\hat{l}_y(z_y)\hat{u}[\sigma_{xy},\sigma_i]|\psi_{\sigma_i}\rangle.\end{aligned}\ ] ] the second line uses the fact that @xmath195\right ] = 0,\end{aligned}\ ] ] if @xmath85 is found on both @xmath77 and @xmath82 , which follows from ( [ acom3 ] ) .
the third line follows from ( [ acom2 ] ) .
next we define the surface @xmath196 on which both points @xmath85 and @xmath86 are found and such that @xmath197 , along with a further surface @xmath198 , also containing @xmath85 and @xmath86 and satisfying @xmath199 ; @xmath200 ; and @xmath201 .
we then have @xmath202\hat{l}_x(z_x)\hat{l}_y(z_y)\hat{u}[\sigma''_{xy},\sigma_i]|\psi_{\sigma_i}\rangle \nonumber\\ & = \hat{u}[\sigma_f,\sigma'_{xy}]\hat{u}[\sigma'_{xy},\sigma''_{xy } ] \hat{l}_x(z_x)\hat{l}_y(z_y)\hat{u}[\sigma''_{xy},\sigma_i]|\psi_{\sigma_i } \rangle \nonumber\\ & = \hat{u}[\sigma_f,\sigma'_{xy}]\hat{l}_x(z_x)\hat{l}_y(z_y)\hat{u}[\sigma'_{xy},\sigma_i]|\psi_{\sigma_i}\rangle \nonumber\\ & = \hat{u}[\sigma_f,\sigma'_2]\hat{u}[\sigma'_2,\sigma'_{xy}]\hat{l}_x(z_x)\hat{l}_y(z_y)\hat{u}[\sigma'_{xy},\sigma'_1 ] \hat{u}[\sigma'_1,\sigma_i]|\psi_{\sigma_i}\rangle \nonumber\\ & = \hat{u}[\sigma_f,\sigma'_2]\hat{l}_x(z_x)\hat{u}[\sigma'_2,\sigma'_{xy}]\hat{u}[\sigma'_{xy},\sigma'_1]\hat{l}_y(z_y)\hat{u}[\sigma'_1,\sigma_i]|\psi_{\sigma_i}\rangle \nonumber\\ & = |\psi'_{\sigma_f}\rangle.\end{aligned}\ ] ] using this result it follows from ( [ prob2a ] ) and ( [ prob2b ] ) that the probability density for the pair of collapse outcomes @xmath94 and @xmath203 is independent of the choice of foliation .
iteration of the above procedure for further collapses demonstrates foliation independence of the complete set of collapse outcomes @xmath106 occurring at the set of collapse locations @xmath204 between any @xmath103 and @xmath104 .
a way to understand the infinite energy increase of the collapse dynamics described in section [ covcoll ] is to notice that each collapse on the quantum state occurs at a single point on the spacetime .
this results in sharp spatio - temporal discontinuities in the state of the field , and hence a large energy increase . in order to prevent this
the collapse should happen smoothly .
for a quantum field this means that whenever a local collapse occurs it should act over some spacetime region rather than at an infinitesimal space time point .
this requires some form of smeared interaction . in order to facilitate this we use a new type of relativistic quantum field which we call the _ pointer field _ as introduced in @xcite
this field has an independent degree of freedom at each space time point
. we will denote it by @xmath205 ( not to be confused with the matter field @xmath60 ) .
the commutation properties of the pointer field are as follows : @xmath206 = \frac{1}{\sqrt{g(x)}}\delta^4(x - x ' ) ; \quad \left[\hat\psi(x),\hat\psi(x')\right ] = 0.\end{aligned}\ ] ] notice that the dirac delta extends over the whole space time , not just over a hyper surface . given these annihilation and creation operators we can define a smeared field operator @xmath207.\end{aligned}\ ] ] and a smeared number density operator which will be the collapse basis of equation ( [ l ] ) @xmath208 the smearing functions @xmath209 and @xmath210
are assumed to be defined in terms of the ( fixed ) space time properties such as local curvature .
they should each satisfy certain properties of smoothness and finiteness under integration to be determined by their consequences for the relativistic collapse theory .
we assume that there is an interaction between ordinary matter fields and the pointer field of the form @xmath211 where @xmath212 is a coupling parameter , which , as we will see below , affects the effective collapse rate of the matter field .
the state @xmath76 now includes the state of both matter field @xmath154 and the pointer field @xmath205 .
the required micro causality conditions ( [ com2 ] ) and ( [ com3 ] ) can be used to begin to constrain the form of @xmath209 and @xmath210 . in general
it follows from the above commutation properties that @xmath213 = 0 ; \quad \left[\hat{b}(x),\hat{b}(x')\right ] = 0 , \label{com0}\end{aligned}\ ] ] for all space time points @xmath85 and @xmath214 .
it also follows that for space like separated @xmath85 and @xmath214 , @xmath215 = 0 , \label{com}\end{aligned}\ ] ] provided that the domain of @xmath216 only includes points where @xmath86 is inside the future light cone of @xmath85 , and the domain of @xmath217 only includes points where @xmath86 is inside the past light cone of @xmath85 .
this commutation property result in ( [ com3 ] ) . as
a concrete proposal for @xmath209 on a spacetime with metric @xmath218 we will take : @xmath219 where the integration measure @xmath220 along the geodesic is the differential of @xmath221 the invariant line element , rather than volume element where @xmath222 is some positive integer , @xmath223 is a suitably chosen dimensional parameter , @xmath224 is the characteristic function of @xmath225 , the chronological future of @xmath85 , that is @xmath226 iff @xmath227 and vanishes otherwise , @xmath228 is the causal geodesic connecting @xmath85 and @xmath86 ( which we will assumed to be unique in a convex normal neighborhood of @xmath85 . for points outside this region we can replace the prescription to one where we replace the integral along a single geodesic by the sum of integrals over all such geodesics . ] ) , @xmath229 is the tangent to the geodesic @xmath230 by proper time , and @xmath231 is the bell tensor of the spacetime metric @xmath147 .
we note that as the @xmath229 are future directed time - like vectors the integrand in the above equation is positive semi - definite .
that is @xmath232 .
in fact , generically this quantity will vanish only along the principal null directions of the weyl tensor , which as we know @xcite , are in general just a discrete set of directions ( 4 in 4 spacetime dimensions )
. it would be only when such null directions coincide with a tangent @xmath229 along the full null geodesic connecting @xmath85 and @xmath233 that such an integral would vanish .
it is therefore , only in those very unusual cases ( where such a @xmath234 exists ) , that for points @xmath235 that approach the points @xmath236 , that the integral @xmath237 might fail to be bounded from below by a positive number . otherwise , in the generic situations , the functions @xmath238 will rapidly decrease as the point @xmath86 gets further from @xmath85 even along the directions that approach those in the
null cone " corresponding to the boundary of the chronological future of @xmath85 : @xmath239 . to understand how the collapse mechanism works consider first the interaction term @xmath245 .
this has the effect of coupling the state of the @xmath154 field to the state of the @xmath205 field .
an excited matter field will lead to an excitation of the pointer field in the local region determined by @xmath209 . for a matter field in a superposition of different @xmath154 states
this interaction will lead to an entanglement between the different @xmath154 states and different states of the pointer field .
the state of the pointer field is the one that is now subjected directly to the collapse dynamics of sction iii .
the state of the system is an element of the product space between the hilbert space of the matter fields and that of the pointer field .
the action of the collapse operator leads to a collapse of the pointer field in the ( smeared ) number density basis .
they act as quasi projectors onto a number density state with a central value determined by the random choice @xmath94 .
this causes the pointer field to collapse towards a state of definite number density .
since the matter field is entangled with the pointer field , the collapse of the pointer field induces a collapse of the matter field state in the field state basis .
now we derive an effective characterization of the resulting collapse dynamics involving just the quantum field @xmath183 . starting with the fully covariant model we will trace out the the pointer field leaving an approximate dynamical equation for the @xmath154 state .
we start by choosing a particular foliation parametrized by a time coordinate @xmath45 , and write the spacetime metric in terms of the standard 3 + 1 decomposition as : @xmath246 where @xmath247 is the lapse function and @xmath248 are the components of the shift vector characterizing the foliation and @xmath249 are the components of the induced metric on the corresponding spatial hyper surface .
the equation describing the interaction between the scalar field and the pointer field is then @xmath250 where for now we ignore the collapses .
the state is always pure and so the density matrix can be written as @xmath251 we further assume that the pointer field is always in an approximate vacuum state .
] so that we can write @xmath252 this assumption requires that the coupling parameter @xmath212 is weak and that the pointer field collapses are able to resolve very small differences in number density so that the collapse occurs even after a very weak interaction between @xmath154 and @xmath205 .
this requires choosing @xmath116 to be sufficiently small .
the master equation describing the development of the density matrix is @xmath253,\end{aligned}\ ] ] where the interaction hamiltonian is given by @xmath254 we can use the born approximation , valid for weak interactions , to find an approximate solution to the master equation @xmath255.\end{aligned}\ ] ] this can then be inserted back into the master equation @xmath256 - \int_0^t dt ' \left[\hat{h}_t,\left[\hat{h}_{t'},\hat\rho_t\right]\right].\end{aligned}\ ] ] next we take the partial trace over the pointer field degrees of freedom .
to do this we use @xmath257 = \nu \int d^3 x n\sqrt{h } \left[|\hat{f}(x)|^2,\hat\rho^{f}_0\right ] { \rm tr}_{\psi}\left[\hat{a}(x)\hat\rho^{a}_0\right],\end{aligned}\ ] ] and @xmath258\right ] = \nu^2 \int d^3 x n\sqrt{h}\int d^3 x ' n'\sqrt{h ' } \left[|\hat{f}(x)|^2,\left[|\hat{f}(x')|^2,\hat\rho^{f}_t\right]\right ] { \rm tr}_{\psi}\left[\hat{a}(x)\hat{a}(x')\hat\rho^{a}_t\right].\end{aligned}\ ] ] now assume that the function @xmath216 can be reasonably well approximated by a delta function @xmath259 where @xmath260 is a spacetime dependent scale factor .
together with the assumption that the pointer field density matrix is approximately vacuum this results in @xmath261 = 0,\end{aligned}\ ] ] and @xmath262 = \frac{\eta^2(x)}{\sqrt{g(x)}}\delta^4(x - x').\end{aligned}\ ] ] putting all this together we find the master equation for the scalar field to be of the form @xmath263\right ] .
\label{masterxi}\end{aligned}\ ] ] this is of precisely the same form as equation ( [ phicoll ] ) once we identify @xmath264 with @xmath265 .
this form of the master equation predicts infinite energy increase .
this is tempered by choosing a form for @xmath216 which is not a delta function and ( [ masterxi ] ) should be considered an idealized limit for describing collapse . in section [ covcoll ]
we demonstrated that a relativistic master equation of this form reduces to the non relativistic csl model .
the correspondence can be made more precise by choosing a suitable frame of reference defined by the coordinates @xmath266 in which @xmath216 [ invariantly defined according to ( [ eq : sa ] ) ] takes the approximate form , @xmath267 where we assume that we have tuned the choice of the parameter @xmath223 so as to ensure that , in ordinary laboratory situations , where the spacetime is almost flat ( except for the curvature induced by the few particles involved ) , @xmath268 reduces approximately to the standard csl smearing function .
@xmath269 with @xmath270 the grw length scale .
the resulting non relativistic csl model is well known and produces finite energy increase which can be kept suitably small by appropriate choice of the parameters .
l. susskind , l. thorlacius and j. uglum , `` the stretched horizon and black hole com- plementarity , '' phys .
d 48 , 3743 ( 1993 ) [ hep - th/9306069 ] . c. r. stephens , g. t hooft and b. f. whiting , `` black hole evaporation without infor- mation loss , '' class .
11 , 621 ( 1994 ) [ gr - qc/9310006 ] .
l. parker and d. toms , `` quantum field theory in curved spacetime , '' cambridge university press ( 2007 ) ; r. m. wald , `` quantum field theory in curved spacetime and black hole thermodynamics , '' the university of chicago press ( 1994 ) .
a. perez , h. sahlmman and d. sudarsky , on the quantum mechanical origin of the seeds of cosmic structure , " classical and quantum gravity * 23 * , 2317 ( 2006 ) ; d. sudarsky , shortcomings in the understanding of why cosmological perturbations look classical , " international journal of modern physics d * 20 * , 509 ( 2011 ) [ arxiv:0906.0315 [ gr - qc ] ] ; s. j. landau , c. g. scoccola and d. sudarsky , cosmological constraints on nonstandard inflationary quantum collapse models , " physics review d * 85 * , 123001 ( 2012 ) [ arxiv:1112.1830 [ astro-ph.co ] ] ; g. len garca , s. j. landau and d. sudarsky , quantum origin of the primordial fluctuation spectrum and its statistics , " physics review d * 88 * , 023526 ( 2013 ) [ arxiv:1107.3054 [ astro-ph.co ] ] ; p. caate , p. pearle and d. sudarsky , csl quantum origin of the primordial fluctuation , " physics review d * 87 * , 104024 ( 2013 ) [ arxiv:1211.3463[gr - qc ] ]
. s. k. modak , l. ortz , i. pea and d. sudarsky , `` non - paradoxical loss of information in black hole evaporation in a quantum collapse model , '' phys .
d * 91 * , no .
12 , 124009 ( 2015 ) [ arxiv:1408.3062 [ gr - qc ] ] .
g. ghirardi , a. rimini , t. weber , a model for a unified quantum description of macroscopic and microscopic systems , " in a. l. accardi ( ed . ) quantum probability and applications , p. p. 223
- 232 , springer , heidelberg ( 1985 ) .
d. albert , quantum mechanics and experience ( harvard university press , 1992 ) , chapters 4 and 5 ; j. bell , quantum mechanics for cosmologists " , in quantum gravity ii , oxford university press , 1981 ; d. home , conceptual foundations of quantum physics : an overview from modern perspectives ( plenum , 1997 ) .
chapter 2 ; e. wigner , `` the problem of measurement , '' am .
j. of physics * 31 * , 6 ( 1963 ) ; a. lagget , `` macroscopic quantum systems and the quantum theory of measurement , '' prog .
. suppl . * 69 * , 80 ( 1980 ) ; j. s. bell , speakable and unspeakable in quantum mechanics , " ( cambridge university press 1987 ) ; j. s. bell , against measurement , " phys .
world * 3 * , 33 ( 1990 ) . for reviews about the various approaches to the measurement problem in quantum mechanics
see , for instance , the classical reference m. jammer , `` philosophy of quantum mechanics .
the interpretations of quantum mechanics in historical perspective , '' ( john wiley and sons , new york 1974 ) ; a. peres , quantum theory : concepts and methods " ( kluwer , academic publishers , 1993 ) ; r. omnes , the interpretation of quantum mechanics , " ( princeton university press 1994 ) ; and the more specific critiques s. l. adler why decoherence has not solved the measurement problem : a response to pw anderson , " stud .
mod . phys . * 34 * , 135 - 142 ( 2003 ) [ arxiv : quant - ph/0112095 ] ; a. elby , why modal interpretations of quantum mechanics do nt solve the measurement problem , found . of phys .
* 6 * , 5 - 19 ( 1993 ) .
d. durr , s. goldstein , and n. zangh , `` bohmian mechanics and the meaning of the wave function , '' in cohen , r. s. , horne , m. , and stachel , j. , eds . , experimental metaphysics quantum mechanical studies for abner shimony , volume one ; boston studies in the philosophy of science 193 , ( kluwer academic publishers , 1997 ) ; j. s. bell , on the impossible pilot wave " , foundations of physics 12 ( 1982 ) , pp .
989 - 99 ; d. wallace , the emergent multiverse .
oxford university press , 2012 ; c. fuchs and a. peres , quantum theory needs no interpretation `` '' .
physics today 53(3 ) ( 2000 ) , pp .
70 - 71 ; o. lombardi and d. dieks , modal interpretations of quantum mechanics " , the stanford encyclopedia of philosophy , 2014 ; e. joos et al , decoherence and the appearance of a classical world in quantum theory , 2nd edition ( springer , 2003 ) ; w. zurek , decoherence and the transition from quantum to classical , " phys .
44 , no . 10 , 1991 . a. kent ,
against many - worlds interpretations " , online at http://xxx.arxiv.org/abs/gr-qc/9703089 ; h. brown , and d. wallace , solving the measurement problem : de broglie - bohm loses out to everett " .
foundations of physics 35 ( 2005 ) , pp.517 - 540 ; j. bub , interpreting the quantum world ( cambridge , 1997 ) , chapter 8 , pp .
212 - 236 .
( rather critical discussion of the decoherence - based approaches ) .
r. penrose , `` the emperor s new mind , '' ( oxford university press 1989 ) ; r. penrose , `` on gravity s role in quantum state reduction , '' in physics meets philosophy at the planck scale , callender , c. ( ed . )
( 2001 ) .
r. penrose , `` on the gravitization of quantum mechanics 1 : quantum state reduction , '' found .
phys . * 44 * , 557 - 575 ( 2014 ) ; r. penrose , `` on the gravitization of quantum mechanics 2 : conformal cyclic cosmology , '' found .
* 44 * , 873 - 890 ( 2014 ) . c. g. callan ,
s. b. giddings , j. a. harvey , and a. strominger , evanescent black holes , " phys .
d * 45 * , r1005 ( 1992 ) .
a. fabbri and j. navarro - salas , modeling black hole evaporation ( imperial college press , london 2005 ) .
j. collins , a. perez , d. sudarsky , l. urrutia , and h. vucetich , lorentz invariance in quantum gravity : a new fine tunning problem ? , " _ phys .
lett . _ * 93 * , 191301 , ( 2004 ) ; c. rovelli and s. speziale , reconcile planck - scale discreteness and the lorentz - fitzgerald contraction , _ phys .
* d 67 * , 064019 ( 2003 ) [ arxiv : gr - qc/0205108 ] ; f. dowker and r. sorkin , `` quantum gravity phenomenology , lorentz invariance and discreteness , '' [ arxiv : gr - qc/0311055 ] . | we study a proposal for the resolution of the black hole information puzzle within the context of modified versions of quantum theory involving spontaneous reduction of the quantum state .
the theories of this kind , which were developed in order to address the so called _ measurement problem _ in quantum theory have , in the past , been framed in a non - relativistic setting and in that form they were previously applied to the black hole information problem . here , and
for the first time , we show in a simple toy model , a treatment of the problem within a fully relativistic setting .
we also discuss the issues that the present analysis leaves as open problems to be dealt with in future refinements of the present approach . |
our local cosmological neighborhood is made of baryons , while fraction of antimatter , presumably of astrophysical origin , is vanishingly small .
so the observations indicate that the universe is 100% baryo - asymmetric , at least locally .
the baryon asymmetry ( of the universe ) , bau , can not be explained in the frameworks of the standard model ( sm ) of particle physics . alongside with evidence for dark matter and dark energy
it is considered as unambiguous proof of the existence of new physics beyond sm .
many quite different extensions of the sm and various scenarios for the bau generation were suggested in the literature , for a review see e.g. refs .
typically , consideration is restricted to the models where the universe is asymmetric globally .
this is the simplest possibility .
however , it is not excluded that the real universe may be globally symmetric .
it may consist of domains of matter and antimatter , and if the domains are sufficiently large and far away , they may escape observational constraints on matter - antimatter annihilation at the domain boundaries
. in the simplest version of the scenario the distance to the nearest domain of antimatter should be close to the present day cosmological horizon @xcite .
corresponding particle physics models , leading to the universe creation with abundant antimatter domains were suggested and developed in the past . while being more involved , the models of this type also suffer from the inherent problem a domain wall problem @xcite .
indeed , bau can be generated only if @xmath1 violation is sufficiently strong , beyond the sm capabilities .
in addition , in globally symmetric universe @xmath1 should have different signs in different domains .
this non - trivial pattern of @xmath1 violation could be provided by a dedicated physical field , one way or another .
therefore , unavoidably , domains with different @xmath1 phase would be separated by domain walls with unacceptably high energy density , in conflict with observations .
there is only one way out of this restriction .
namely , domains with different sign ( and possibly strength ) of @xmath1 violation should exist only in the universe past and should disappear by now , all together with domain walls , though their effects in the form of matter and antimatter objects would survive to the present day .
one class of models where this can be achieved has been suggested in refs .
the main idea behind is a possibility of an unusual symmetry behavior at high temperatures .
it is well known that a symmetry , which is broken in vacuum , at high temperatures tends to be restored .
but in general , the situation is not that simple and straightforward .
it is also possible that a symmetry is broken only in a particular range of temperatures , i.e. it is restored at the highest as well as at the lowest temperatures , for the particular models and details see @xcite .
this is just what is needed for a matter - antimatter domain generation without domain wall problem .
however , if a model is based on the unusual symmetry behavior at high temperatures , then the size of domains will be too small from the cosmological point of view .
such models are still interesting , because they could provide a local excess of antimatter , which would be large compared to the capabilities of astrophysical sources , if an excess @xcite turns out to be real . but cosmologically large and separated domains of antimatter can not be created by this mechanism .
cosmology with domains of matter and antimatter was discussed in the lectures @xcite , where a list of relevant references can be found . as argued in ref .
@xcite , domain walls could be also eliminated if the vacua in the model were not exactly degenerate . in this case
the higher energy vacuum would be `` swallowed '' by the lower energy one if the energy difference is sufficiently high . in the present paper
we suggest another scenario of unusual symmetry behavior .
now this happens during inflationary stage in the universe evolution .
domains with different sign of @xmath1 disappear by now also , so the domain wall problem is absent .
however , they appeared during inflation and survived at the baryogenesis epoch , therefore , cosmologically large domains of matter and antimatter could be created .
in the suggested model the difference between matter and antimatter is generated by pseudoscalar field @xmath2 which interacts with inflaton field @xmath3 .
we assume the following lagrangian : @xmath4 where @xmath5 here the metric tensor enters into kinetic terms in the usual way , and @xmath6 are some constant parameters , with @xmath7 having dimension of mass , @xmath8 being dimensionless .
we do not include the quartic term @xmath9 in the lagrangian , though it possibly may lead to some interesting consequences .
this case will be studied elsewhere .
the dimensionless function @xmath10 is chosen in the way that it is non - zero only when @xmath3 is close to some constant value @xmath11 . in this paper
we choose it as gaussian function @xmath12 , \label{v - of - phi}\ ] ] though other forms may be possible .
the plot of function @xmath10 ( see fig . [
fig : v ] ) is a bell - shaped curve .
parameters @xmath11 and @xmath13 have dimension of mass and indicate position of the `` bell '' center and characteristic width of the `` bell '' , respectively . function @xmath10 .
the parameters are @xmath14 and @xmath15 .
field @xmath3 is measured in units of @xmath16 and @xmath10 is dimensionless . ]
the equations of motion have the form : @xmath17 where @xmath18 is the hubble parameter , @xmath19 is the cosmological scale factor which enters into the flrw metric as @xmath20 it is assumed here that fields @xmath3 and @xmath2 depend only on time , @xmath21 and @xmath22 .
the hubble parameter is expressed through the energy density @xmath23 as @xmath24 where @xmath25 gev is the planck mass .
the interaction introduced above leads to the following scenario . during inflation the magnitude of the inflaton field @xmath3 decreases and when it reaches vicinity of @xmath26 is larger than @xmath26 . ]
two minima appear in the potential @xmath27 at constant @xmath3 , so the point @xmath28 becomes local maximum ( see fig .
[ fig : u - phi - chi ] ) . [ cols="^,^,^ " , ] in fig .
[ fig : chi ] one can see how the field @xmath2 rolls down to the minimum of the potential @xmath29 and oscillates there .
then it goes back to the old minimum at @xmath30 , but even at the end of inflation @xmath2 remains an order of magnitude larger than it was at the beginning ( @xmath31 at @xmath32 , while @xmath33 ) . at the moment @xmath34 the field @xmath2 crosses zero and starts slow oscillations with very low frequency , which could be much lower than the characteristic rate of baryogenesis .
so we see that parameters ( [ eq : parameters ] ) are chosen in a proper way . in this model
the influence of @xmath2 on the process of inflation is insignificant , thus the domain size grows exponentially .
however , as we have mentioned at the end of the previous section , it is also important to know how the regions where @xmath28 expand during inflation .
the hypothetical situation is possible when the areas where field is zero expand exponentially faster than the areas where field is non - zero .
for example , the model where inflation never stops near the domain walls , i.e. near @xmath35 , was suggested in @xcite .
also the idea of never ending topological inflation where @xmath36 is forced to stay near the maximum of its potential @xmath35 was proposed in @xcite .
however , our situation is quite different .
numerical calculation shows that if @xmath28 the hubble parameter remains almost the same as in the case of non - zero @xmath2 , and the size of region where @xmath28 is approximately two times less than the domain size .
therefore , our model predicts the domains of cosmological size with the distances between them of the same order of magnitude .
as it is commonly known , the stage of inflation is followed by the stage of ( re)heating , during which the very heavy @xmath37 particles can be produced through decay of inflaton field ( see e.g. book @xcite ) or from vacuum fluctuations in gravitational field @xcite .
let us consider the situation when the inflaton field @xmath3 interacts with @xmath37 particles ( for example , scalar bosons ) through the coupling @xmath38 . as is shown in refs .
@xcite , the particle production can be strongly enhanced due to parametric resonance . for an unsuppressed production
it is essential that the resonance is broad .
the corresponding condition has the form : @xmath39 here @xmath40 is the inflaton mass , and @xmath41 is the magnitude of the inflaton field at the end of inflation in our model .
one can consider @xmath42 also as the initial amplitude of the inflaton oscillations .
the condition ( [ broadres ] ) is true for the wide range of coupling constant , @xmath43 . because of resonance ( [ broadres ] )
, very heavy @xmath37-bosons with masses even of gut scale @xmath44 gev can be produced under quite reasonable assumption @xmath45 @xcite . moreover in this case of comparatively large coupling constant @xmath45 the decay of the inflaton field occurs very rapidly , during only one or few oscillations .
let us assume that the produced @xmath37-bosons in turn decay into fermions , for example , into quark - quark and antiquark - antilepton pairs , @xmath46 and @xmath47 , respectively .
if the corresponding coupling constants are large enough , the @xmath37-bosons decay very quickly . therefore , the field @xmath48 may remain non - zero yet to the moment when @xmath37-bosons have been completely decayed , indeed @xmath49 gev at @xmath50 ( see fig . [
fig : chi ] ) .
the field @xmath2 is real and pseudoscalar and interacts with the produced fermions as ( cf .
( [ l - ferm ] ) ) @xmath51 where @xmath52 and @xmath53 denote the fermion flavor ( sum over repetitive indices is assumed ) , @xmath54 , @xmath55 . from the hermicity of the interaction it follows that @xmath56 is real for @xmath57 , and @xmath58 for @xmath59 .
since @xmath2 is supposed to be electrically neutral , it interacts with quarks with the same electric charge , so @xmath52 and @xmath53 either run over @xmath60 or @xmath61 and there are no cross terms . the lagrangian of free fermions can be written as follows @xmath62 therefore , the sum of @xmath63 and @xmath64 can be presented in matrix form as @xmath65 here @xmath66 is a non - hermitian matrix , whereas @xmath67 and @xmath68 are hermitian ones .
using simultaneously two unitary transformations @xmath69 and @xmath70 one can always diagonalize the mass matrix in ( [ l - ferm - free - chi ] ) .
therefore , the interaction of fermions with pseudoscalar field @xmath2 can be `` rotated away '' ( the elements of transformation matrices @xmath71 and @xmath72 must depend on the magnitude of @xmath2 in this case ) , and the lagrangian ( [ l - ferm - free - chi ] ) takes the simple form @xmath73 where @xmath74 and @xmath75 are the mass eigenstates and correspondingly @xmath76 is diagonal matrix with real diagonal elements . quite similarly one can transform into mass term the interaction of fermions with any scalar @xmath77 and pseudoscalar @xmath78 fields .
however , the interaction of fermions with vector ( gauge ) boson @xmath37 remains the same under these transformations : @xmath79 here @xmath80 and @xmath81 are matrices of coupling constants in mass eigenstate basis , so @xmath74 describes @xmath82-th sort of fermion with definite mass . the constants @xmath83 are complex in general case , and if there are at least three species of fermions , one can not rotate away simultaneously all phases in complex matrices @xmath84 @xcite .
the complexity of the coupling constants means that @xmath1 is violated in the @xmath37-boson decays @xcite .
the magnitude of this @xmath1 violation depends on the value of the field @xmath2 through the matrices @xmath85 and coupling constants @xmath84 . since @xmath2 is essentially non - zero after the end of inflation and during baryogenesis
, the @xmath1-odd effects can be large enough .
we assume also that gauge interactions involve fermions with certain chirality , see ( [ vecint ] ) , and thus these interactions break @xmath0-invariance . @xmath0 and @xmath1 violation is one of the necessary sakharov conditions of generation of baryon asymmetry @xcite .
the another one is baryon number violation , so one needs to assume also that in the decays of @xmath37-bosons the baryon number is not conserved .
let @xmath86 be the baryon asymmetry generated in the decay of one @xmath37-boson .
then it can be easily demonstrated @xcite that the ratio of the baryon number density to the entropy density is estimated as @xmath87 where @xmath88 and @xmath89 are typical coupling constants of @xmath37-boson with fermions and inflaton , respectively , @xmath90 is mass scale of the theory .
it is quite reasonable to believe that @xmath91 and @xmath92 , so one has @xmath93 .
thus , to get observed value @xmath94 it is sufficient to have only @xmath95 .
such small @xmath86 seems to be easily produced in the decay of @xmath37-boson .
therefore , the observed baryon asymmetry of the universe can be generated in the decays of @xmath37-bosons , without fine tuning of parameters of the theory .
we have to stress that the baryogenesis should proceed after the end of inflation ( as it has been already mentioned above ) . otherwise the baryon asymmetry would be strongly diluted by the universe ( re)heating .
when inflation was over , the field @xmath2 evolved down to @xmath96 and so @xmath2 was not equal to constant @xmath97 but had considerably smaller and time dependent value .
however , since at this stage @xmath48 changed very slowly one can repeat the above arguments with adiabatically evolving @xmath98 .
if baryogenesis proceeded faster than @xmath2 evolved , then the effective amplitude of @xmath1 violation might be considered as a constant and the corresponding mass matrix could be diagonalized in the same way as it has been done above .
the phase of the mass matrix would be unsuppressed if @xmath99 was close to the value of the bare quark mass . evidently in this case
the imaginary part of the mass matrix would be of the same order of magnitude as the real part .
we have presented here a model of baryogenesis which may lead to baryo - symmetric universe with cosmologically large domains of matter and antimatter , avoiding the domain wall problem .
the model satisfies three sakharov criteria for successful baryogenesis : non - conservation of the baryon number , deviation from thermal equilibrium , and @xmath0 and @xmath1 violation . however , the latter is different from the normally exploited one . breaking of charge symmetry
is induced by a non - zero amplitude of a scalar field @xmath2 , which slowly relaxed down to equilibrium , much slower than the process of baryogenesis goes on . in classification of different types of @xmath1 violation which might be operative in cosmology this type
is called dynamical one @xcite . later , after the baryon asymmetry was developed , @xmath2 evolved down to the equilibrium point @xmath96 and thus domain walls disappeared rather early in the universe .
inflation is an essential ingredient of the scenario .
a coupling to the inflaton field was introduced on purpose to generate a non - zero value of @xmath2 and to keep it non - zero during baryogenesis .
the model allows for successful baryo - symmetric cosmology without yet being drawn into astronomical controversy .
we acknowledge support of the russian federation government grant no.11.g34.31.0047 . s.g .
is partially supported under the grants rfbr no.14 - 02 - 00995 and nsh-3830.2014.2 .
is also supported by mk-4234.2015.2 . and by dynasty foundation .
v.a.rubakov and m.e.shaposhnikov , _ electroweak baryon number nonconservation in the early universe and in high - energy collisions _ ,
* 39 * ( 1996 ) 461 [ usp
nauk * 166 * ( 1996 ) 493 ] [ hep - ph/9603208 ] .
f.w.stecker , _ the matter - antimatter asymmetry of the universe _ , in the proceedings of 14th rencontres de blois on matter - antimatter asymmetry , chateau de blois , france , june 17 - 22 , 2002 [ hep - ph/0207323 ] .
a.d.dolgov , _ cosmological matter antimatter asymmetry and antimatter in the universe _ , in the proceedings of 14th rencontres de blois on matter - antimatter asymmetry , chateau de blois , france , june 17 - 22 , 2002 [ hep - ph/0211260 ] .
a.d.sakharov , _ violation of cp invariance , c asymmetry and baryon asymmetry of the universe _ , jetp lett . * 5 * ( 1967 ) 24 [ pisma zh .
. fiz . * 5 * ( 1967 ) 32 ] , sov .
* 34 * ( 1991 ) 392 [ usp .
nauk * 161 * ( 1991 ) 61 ] .
a.d.dolgov , _ cp violation in cosmology _ , in the proceedings of 163rd course of international school of physics enrico fermi : cp violation : from quarks to leptons , varenna , italy , july 19 - 29 , 2005 [ hep - ph/0511213 ] . | we present a model of spontaneous ( or dynamical ) @xmath0 and @xmath1 violation where it is possible to generate domains of matter and antimatter separated by cosmologically large distances .
such @xmath0(@xmath1 ) violation existed only in the early universe and later it disappeared with the only trace of generated baryonic and/or antibaryonic domains .
so the problem of domain walls in this model does not exist .
these features are achieved through a postulated form of interaction between inflaton and a new scalar field , realizing short time @xmath0(@xmath1 ) violation . |
interaction in a low - dimensional system does not merely result in strong renormalization of physical quantities , but can in many cases drive the system into completely new phases with peculiar properties . for a two - dimensional ( 2d ) electron gas in a perpendicular magnetic field , the interaction effects are especially important because of landau level quantization .
when electrons are entirely restricted to the lowest landau level by a large magnetic field , electron - electron interaction completely dominates the properties of the system as the electron kinetic energy is quenched to an unimportant constant .
one of the most interesting phenomena in this strongly - correlated system is the quantum hall ( qh ) effect , which has attracted a great deal of experimental and theoretical interest during the last fifteen years@xcite .
recent advances in materials growth techniques have made it possible to fabricate high - quality double - layer two - dimensional electron systems with the electrons confined to two parallel planes separated by a distance comparable to that between electrons within a plane . with the introduction of this layer degree of freedom ,
many qualitatively new effects due entirely to interlayer correlations appear@xcite .
many new qh phases in double - layer systems become real possibilities because of the increased degree of freedom and the complicated interplay among interlayer tunneling energy , zeeman energy , and electron - electron coulomb interaction energy . in this paper
, we present the details of our earlier theoretical investigations@xcite of the possible qh phases in a double - layer system at a _ total _ landau level filling factor @xmath3 , where @xmath1 is a filling factor at which an isolated single layer system forms a fully spin polarized incompressible qh state ( _ e.g .
@xmath4 , etc .
_ ) we will discuss three distinct ground states , and the nature of the zero or finite temperature transitions / crossovers between them : * a fully polarized ferromagnetic ( fpf ) qh state in which the spins in each layer are aligned parallel to the magnetic field .
this state is adiabatically connected to the ground state of well separated layers , each forming a polarized qh state at filling fraction @xmath1 .
we will denote this fpf state also as the fm ( for `` ferromagnetic '' ) state . *
a spin singlet ( ss ) state , which can be visualized crudely as consisting of singlet pairs of electrons in opposite layers .
alternatively , at @xmath2 , we will discuss the hartree - fock picture of spin up and spin down electrons fully occupying single - particle states which are symmetric in the layer `` pseudospin '' index ; hence the singlet state will also be referred to as sym . in the limit of a vanishing tunneling matrix element between the layers ,
this state is simply the pseudospin polarized state of refs for both spin up and spin down electrons separately . throughout
, we will consider the case of a non - vanishing tunneling matrix element : in this case the pseudospin polarization is chosen by the phase of the tunneling amplitude , and not spontaneously .
none of the phase transitions we consider here require a vanishing tunneling matrix element ; on the contrary , changes in the value of the tunneling matrix element can drive the quantum transitions . *
a canted state ( c ) in which the average spin moments in the layers have an antiferromagnetic correlation in the plane perpendicular to the magnetic field , and a ferromagnetic correlation parallel to the magnetic field .
both ferromagnetic and antiferromagnetic moments can vary continuously at zero temperature as parameters are varied .
the planar antiferromagnetic ordering breaks spin rotation symmetry about the magnetic field axis : as a consequence there is a gapless , linearly dispersing , goldstone collective mode in its excitation spectrum and a kosterlitz - thouless transition at a finite temperature .
the c phase is the canted antiferromagnetic phase ( caf ) discussed in our earlier short publications @xcite .
we will use two distinct and complementary approaches to understand these phases .
the first is a mean - field hartree - fock calculation : this applies only for integer values of @xmath1 , but has the advantage of working with a precise microscopic hamiltonian involving only parameters which are directly known experimentally .
the second is a phenomenological , quantum field - theoretic formulation which applies for general @xmath1 , and allows us to more precisely understand the consequences of thermal and quantum fluctuations .
we will now discuss some of the results of these two approaches in turn . in the hartree - fock approximation@xcite , we are able to show that the canted antiferromagnetic ( c ) phase is the energetically favored ground state for @xmath5 at intermediate layer separations for systems with small zeeman energy , and that the phase transitions from the c to the fm or sym phases are continuous .
we evaluate at @xmath5 the intersubband spin density wave ( sdw ) dispersions of all phases in the time - dependent hartree - fock approximation @xcite and show that , as the precursor of the phase transitions , the collective intersubband sdw mode softens at the phase boundaries of the fm and sym phases to the c phase .
the sdw becomes the linearly dispersing goldstone mode in the c phase , and the temperature of the kosterlitz - thouless transition is obtained by evaluating its effective spin - stiffness in the hartree - fock approximation .
in addition we present results on the stability energetics of the various phases , the antiferromagnetic order parameter , the phase diagram , the collective intersubband sdw excitation dispersions , and the specific heat .
the @xmath5 hartree - fock results may also be qualitatively applicable to the case of @xmath6 if the landau level mixing is ignored ( the landau level mixing may not be negligible at @xmath6 , though . ) on the other hand , the situation at @xmath7 is very different from the situation at @xmath5 , since the inter - landau level excitation energies are comparable to the cyclotron energy ; our results do not apply at @xmath7 .
the microscopic hartree - fock analysis obviously does not apply to a situation where the average filling factor @xmath1 in each layer is fractional ( e.g. @xmath8 ) with each isolated layer supporting a spin polarized laughlin fractional qhe state ; such a many - body state will not appear in any mean - field decoupling of the hamiltonian .
however , an essential property of the phases we are discussing is that they all have a gap towards charged excitations , and the transitions between them are driven by changes in the nature of the mean spin polarizations , and of the spin excitations .
this suggests that it may be possible to develop a more general effective theory which focuses on the spin excitations alone .
we will present such a theory in section [ nls ] : it turns out to be the @xmath9 quantum non - linear sigma model in the presence of a magnetic field . for the case @xmath5
, we are able to use our earlier hartree - fock computations to precisely obtain all the renormalized parameters which universally determine the low temperature properties of the non - linear sigma model ; for other values of @xmath1 , including the fractional cases , these parameters remain as phenomenological inputs
. we will present the phase diagram of the sigma model , and describe the nature of the finite temperature crossovers above the various phases in some details .
in particular , we will obtain explicit predictions for the temperature dependence of the line shape of the inelastic light scattering spectrum .
we note that our findings from the two approaches are consistent with recent inelastic light scattering measurement@xcite , where a remarkable ( and temperature dependent ) softening of the long wavelength intersubband sdw mode in a @xmath5 double - layer system is observed .
we hope that our other explicit theoretical results may be tested in future experiments .
the experimental situation will be discussed in section [ exp ] .
this paper is organized as follows .
the results of the hartree - fock theory are presented in section [ hft ] . in section [ grou ] ,
we study the ground state properties of the @xmath5 double - layer system in a self - consistent mean - field approximation .
the intersubband sdw excitations in the time - dependent hartree - fock approximation and associated mode softening are studied in section [ sdw ] .
the thermodynamic properties are discussed in section [ kt ] , and some further discussion , along with an assessment of the validity of the calculation , appear in sections [ multi ] and [ earlier ] . in a long and self - contained section [ nls ]
we give our non - linear @xmath10 model effective field theoretic description for a generic @xmath3 situation .
comparison of our theory with recent light scattering experiments is discussed in section [ exp ] .
a short summary in section [ sum ] concludes this paper .
we note that the readers who are interested only in microscopic hartree - fock theory could skip section [ nls ] , and the readers who are interested only in our long wavelength effective field theory could skip section [ hft ] .
we have taken care in writing the two parts of our work , namely the microscopic hartree - fock calculation for @xmath5 ( section [ hft ] ) and the non - linear @xmath10 model description for @xmath3 ( section [ nls ] ) as two separate self - contained pieces which can be read reasonably independent of each other if so desired .
we begin by writing down the explicit microscopic hamiltonian of a double layer quantum hall system . within the lowest landau level ,
the single particle eigenstates may be denoted by @xmath11 , where @xmath12 is the intra - landau - level index in the lowest landau level , @xmath13 is the pseudospin index which labels the symmetric and antisymmetric subbands , and the spin index @xmath14 labels @xmath15 and @xmath16 spins.@xcite the hamiltonian of the double - layer system is @xmath17 where the non - interacting hamiltonian is @xmath18 where the pseudospin splitting @xmath19 is the tunneling - induced symmetric - antisymmetric energy separation , the spin splitting @xmath20 is the zeeman energy , and @xmath21 is electron creation ( annihilation ) operator .
the coulomb interaction hamiltonian @xmath22 is @xmath23 where @xmath24 is the area of the system , @xmath25 is the magnetic length .
the non - zero coulomb potential matrix elements are @xmath26 and @xmath27 , with @xmath28 $ ] , where @xmath29 and @xmath30 are the intralayer and interlayer coulomb interaction potentials , respectively .
( the finite well - thickness corrections can be taken into consideration by including appropriate form factors@xcite . )
the following subsections will examine various properties of @xmath31 at @xmath5 by mean - field and rpa - like treatments of the interactions in @xmath22 . in this subsection , we investigate the ground state properties of @xmath31 , and obtain the three phases discussed in the introduction . performing hartree - fock pairing of ( [ equ : hi ] )
, one obtains the mean - field interaction hamiltonian as @xmath32 where @xmath33 , which depends on the electronic state being sought through the expectation value @xmath34 .
we self - consistently search for the symmetry broken states where , in addition to @xmath35 , the possibility that @xmath36 is also allowed .
because of the complete landau level degeneracy , the hartree - fock hamiltonian @xmath37 in a uniform state is a @xmath38 matrix , representing the dimension of the subspace associated with the spin and layer degrees of freedom .
it thus has four eigenenergies @xmath39 and four eigenstates @xmath40 ( @xmath41 ) , which are obtained as shown below . in the non - interacting base @xmath42 ,
@xmath43 becomes @xmath44 where @xmath45 where @xmath46 and @xmath47 are associated with the hartree - fock eigenstates @xmath40 which need to be obtained self - consistently , @xmath48 are electron occupation numbers @xmath49 , and @xmath50 .
the off - diagonal matrix elements @xmath51 represent the possibility of the broken symmetry ( @xmath52 ) mentioned above . by diagonalizing the hartree - fock hamiltonian @xmath43 of eq .
( [ equ : hm1 ] ) , one obtains the eigenstates @xmath53 and the eigenenergies @xmath54 eqns .
( [ equ : hm1 ] ) to ( [ equ : e10 ] ) form the complete self - consistent hartree - fock equations which need to be solved numerically .
in fact , the only quantities to be determined in this self - consistent manner are the two parameters @xmath46 and @xmath47 , which , in turn , uniquely define the eigenstates through eq .
( [ equ : eig1 ] )
. the eigenenergies always satisfy @xmath55 ( @xmath56 ) , so the ground state at @xmath5 is given by @xmath57 , where @xmath58 is the vacuum state .
the ground state energy is given by @xmath59 .
there are several sets of @xmath46 and @xmath47 which make eq .
( [ equ : eig1 ] ) the self - consistent solutions to the mean - field hartree - fock equations .
one is @xmath60 and @xmath61 , which corresponds to the symmetric ( sym ) state .
another is @xmath60 and @xmath62 , which corresponds to the spin polarized ferromagnetic ( fm ) state .
these two are the spin - ferromagnets ( fm ) or layer pseudospin-`ferromagnets ' @xcite ( sym ) whose existence is naturally expected in the presence of finite zeeman and tunneling energies .
more interesting is that , for @xmath63 , there exists a solution at intermediate interlayer separations with @xmath64 .
as we shall see shortly , this new state possesses a canted antiferromagnetic ordering ( the c phase ) , _ i.e. _ an interlayer inplane antiferromagnetic spin ordering with the inplane spin magnetic moment in each layer being equal in magnitude and opposite of each other .
the energies of these different states are shown in fig.[fig1 ] .
it is clear from this figure that the energetically favored ground state is the sym state at small interlayer separations , the c state at intermediate separations , and the fm state at large interlayer separations .
the @xmath5 double - layer qh system thus undergoes two quantum phase transitions as the layer separation is increased from @xmath65 to @xmath66 at a fixed magnetic field . to show the antiferromagnetic spin correlations , we rearrange the eigenstates as @xmath67 where @xmath68 , electron spin configurations in the left ( right ) layer in the eigenstate @xmath40 , are @xmath69 , and satisfy @xmath70 .
we immediately obtain the canted antiferromagnetic spin order as @xmath71 where @xmath72 is the electron spin operator in the left ( right ) layer , and @xmath73 denotes the spin alignment direction within the two dimensional plane .
this canted interlayer antiferromagnetic spin ordering is shown schematically in fig .
[ fig41 ] .
note that the total spin magnetic moment still points in the direction of the magnetic field as required by symmetry .
it is obvious that this antiferromagnetic order breaks the @xmath74 symmetry associated with the spin - rotational invariance of the system .
its consequences on the low temperature thermodynamic properties will be discussed later .
the numerical result of this order parameter @xmath75 is shown in fig.[fig2 ] .
one can see that when zeeman energy @xmath20 is increased , the range of the layer separations where the canted antiferromagnetic state exists shrinks in favor of the ferromagnetic state , as the zeeman energy obviously favors the spin polarized state .
it is clear that the phase transition is continuous .
the phase diagram , shown in fig .
[ fig3 ] , can be constructed from this mean - field approximation .
the states @xmath76 and @xmath77 are occupied in the fm phase , @xmath76 and @xmath78 are occupied in the sym phase , and the c phase interpolates between them .
the sym phase exists for @xmath63 and @xmath79 , the c phase exists for @xmath63 and @xmath80 , and the fm phase exists for either @xmath81 or @xmath82 .
the fm phase is favored when @xmath20 is increased , while the sym phase is favored when @xmath19 is increased . in the next subsection , the same phase diagram will be obtained by studying the softening of the intersubband sdw excitations in the time dependent hartree - fock approximation . in this subsection
we have studied the ground state properties of @xmath5 double - layer qh systems in a mean - field hartree - fock approximation and showed the existence of three stable qh phases .
the most interesting observation is the existence of a canted antiferromagnetic phase , with a broken spin rotation symmetry , in between the symmetric and the ferromagnetic phases . in this section ,
we study collective intersubband sdw spectrum of @xmath5 double - layer qh systems in the time - dependent hartree - fock approximation.@xcite these excitations involve flipping both spin and pseudospin of the electron and are the lowest energy excitations at @xmath5 .
the phase instability is studied by investigating the softening of the collective intersubband sdw excitations .
the results obtained in this section are in complete quantitative agreement with the results obtained from the ground state studies in the previous section , as , of course , they should be if the calculations are done correctly . in the absence of interaction , the two branches of the intersubband sdw excitations which correspond to transitions @xmath83 and @xmath84 have excitation energies @xmath85 , where @xmath19 and @xmath20 are interlayer tunneling and zeeman energies , respectively .
interaction renormalizes the excitation energies in two ways .
one is a self - energy correction to the polarizability due to the loss of exchange energy when an electron is excited to a higher but empty level , which raises the excitation energies .
the other is the vertex correction to the polarizability due to an excitonic attraction between the electron excited to the higher level and the hole it leaves behind , which lowers the excitation energies . in diagrammatic perturbation theories , the effect of the exchange energy on the excitation energies is accounted for by including the corresponding self - energy in electron greens functions , and the effect of the excitonic attraction is represented by vertex corrections .
the self - energy and the vertex correction must be consistent with each other obeying the ward identity . the direct hartree term does not influence the sdw excitations because coulomb interaction is spin - rotationally invariant .
since the coulomb interaction potentials are subband - index dependent , they may introduce mode - coupling between the two branches of the intersubband sdw excitations .
this mode - coupling pushes down the frequency of the low - lying excitation and hence helps mode softening . the intersubband sdw excitation spectra
are obtained as the poles of the retarded intersubband spin - density response function @xmath86 \rangle , \label{equ : x1}\ ] ] where the intersubband sdw operator is defined as @xmath87 @xmath88 recovers to familiar forms @xmath89 in the spin polarized state ( @xmath60 and @xmath62 ) , and @xmath90 in the symmetric state ( @xmath91 ) .
@xmath92 is evaluated in the time - dependent hartree - fock approximation , @xcite which we adapt to double - layer systems and , for simplicity , we ignore all the higher landau levels . as argued earlier , this should be a good approximation for our problem . in this approximation ,
one includes the single - loop self - energy and the ladder vertex diagrams in the theory , which satisfies the ward identities .
this time - dependent hartree - fock approximation , therefore , corresponds to solving the vertex equation shown in fig.[fig4 ] , where the electron propagators are the self - consistent hartree - fock green s functions obtained from the mean - field approximation discussed in the previous section .
due to the fact that the coulomb interaction is frequency independent and that the landau levels are completely degenerate , the integral vertex equation can be transformed into an algebraic matrix equation.@xcite the matrices can be further block diagonalized into @xmath38 matrices , from which the poles of the spin - density response function can be ( almost ) analytically calculated .
combining eqns .
( [ equ : x1 ] ) and ( [ equ : ro1 ] ) , one obtains the spin density response function in the matsubara frequencies @xcite @xmath93 where @xmath94 where @xmath95 , @xmath96 is the green s function corresponding to the self - consistent hartree - fock eigenstate @xmath97 and eigenenergy @xmath98 given in eqns .
( [ equ : eig1 ] ) and ( [ equ : e10 ] ) , respectively .
the ladder diagram vertex function is @xmath99l_o^2/2}e^{ip_xq_yl_o^2 } \\ & & \times d_{i^\prime\lambda^\prime}\gamma_{i^\prime\lambda^\prime } ( q , i\omega,\alpha^\prime ) \langle i\lambda ; i^\prime-\lambda^\prime|v(p_x,\alpha-\alpha^\prime)| i-\lambda ; i^\prime\lambda^\prime\rangle\nonumber , \end{aligned}\ ] ] where the interaction matrix element is @xmath100 \left((s^l_{i_1\lambda_1})^\dagger s^l_{i_4\lambda_4}\right ) \left((s^l_{i_2\lambda_2})^\dagger s^l_{i_3\lambda_3}\right ) \nonumber \\ & & \times\left[v_+(q)\delta_{i_2i_3}+v_-(q)(1-\delta_{i_2i_3})\right ] , \label{equ : v01}\end{aligned}\ ] ] where @xmath101 is the electron spin states given in eq .
( [ equ : eig2 ] ) . to solve the vertex equation , we perform the following fourier transformations @xcite @xmath102 and @xmath103 where @xmath104 and @xmath105 , as given by eq .
( [ equ : v01 ] ) . after an analytical continuation ,
one obtains @xmath106 where @xmath107 @xmath108 and @xmath109 are @xmath110 matrices , with @xmath111 , the number of magnetic flux passing through the system , and @xmath112 .
@xmath113 and @xmath114 are @xmath38 matrices , with @xmath115 , and @xmath116 defined in eq .. ( [ equ : v02 ] ) . the intersubband sdw dispersion @xmath117 , which occurs as the pole of the retarded spin density response function @xmath118 , is the solution to @xmath119 .
after a lengthy but straightforward algebraic manipulation , the two intersubband sdw dispersions @xmath120 are obtained as @xmath121 ^ 2 + 4b^2(a+c)(a - c)}\ , \label{equ : w13}\end{aligned}\ ] ] where @xmath122 , @xmath123 , and @xmath124 , with @xmath125 the intersubband sdw dispersions in both the canted antiferromagnetic qh phase ( c ) and the normal qh phases ( fm or sym ) can be obtained from the above expression by incorporating appropriate values of @xmath46 and @xmath47 . in the following ,
we show @xmath120 only at zero temperature for the sake of simplicity , although the formalism applies equally at finite temperatures . in fig .
[ fig5 ] , we show the dispersion of the intersubband sdw above the fm ground state . as mentioned earlier , these two intersubband sdw modes @xmath126 correspond respectively to transitions @xmath127 , approaching asymptotic values @xmath128 , where @xmath129 are the non - interacting excitation energies and @xmath130 is the exchange energy of the electron in the ground state .
mode coupling , which pushes down @xmath131 and hence helps mode softening , is most visible at @xmath132 . at zero layer separation , mode - coupling disappears , and we recover previously known results.@xcite in fig .
[ fig6 ] , we show the intersubband sdw dispersion above the sym state .
the results are qualitatively similar to those in fig .
[ fig5 ] , except that there is no mode coupling in the symmetric state because coulomb interaction is spin independent .
the important thing to be noticed is that the long wavelength collective excitations are gapped in both the symmetric phase and the spin polarized phase .
however , the mode softening does occur at the phase boundaries , as we show below . to illustrate the phase instability , we show , in fig.[fig8 ] , the lower - energy branch of the intersubband sdws at @xmath133 as a function of interlayer tunneling .
we see that @xmath134 indeed softens when approaching the phase boundaries from both the symmetric phase and the spin polarized phase , and remains zero inside the canted antiferromagnetic phase .
the canted antiferromagnetic order parameter , calculated in the previous section , is also shown in fig.[fig8 ] for comparison purpose .
we notice that the phase boundaries determined from these two independent approaches agree completely , as shown in the figure .
the softening of the collective mode and the appearance of the antiferromagnetic order parameter implies that we have discovered a quantum phase transition in double - layer qh systems . in fig.[fig7 ] , the collective intersubband sdw dispersions in the canted antiferromagnetic qh state are shown .
the first thing to be noticed is that the lower energy branch @xmath131 is a gapless mode .
the existence of such a gapless goldstone mode is due directly to the canted antiferromagnetic spin ordering which spontaneously breaks the spin - rotational symmetry of the hamiltonian .
this goldstone mode is found to be linear in the long wavelength limit , consistent with the fact that it describes antiferromagnetic fluctuations .
the existence of the gapless excitation in the canted antiferromagnetic phase implies that some thermodynamic quantities , such as specific heat , have power - law temperature dependence in the canted antiferromagnetic phase in contrast to their exponential temperature dependence in the normal ( symmetric or ferromagnetic ) phases .
simple expressions governing the phase boundaries can be derived from the mode softening .
the boundary between the symmetric ( sym ) phase and the canted antiferromagnetic ( c ) phase is found to satisfy the following equation @xmath135 where @xmath136 . it should be noted that , for any given @xmath19 , the critical layer separation at this boundary is considerably smaller than the critical layer separation where the charge density excitation in the @xmath137 state becomes soft.@xcite the reason for this is the absence of hartree contribution to the sdw excitations .
the boundary between the spin polarized ( fm ) phase and the canted antiferromagnetic ( c ) phase is found to satisfy @xmath138 the simplicity of eqns .
( [ equ : bd1 ] ) and ( [ equ : bd2 ] ) makes the phase diagram easy to construct .
it is worthwhile to note that the phase boundaries are determined by only three energy scales @xmath20 , @xmath19 , and @xmath139 in spite of the fact that the hamiltonian is determined by four independent energy scales @xmath20 , @xmath19 , and @xmath140 , of which the inter- and intra - layer interactions @xmath140 are in fact continuous functions of wavelength @xmath141 .
this unexpected dependence of the phase diagram ( fig .
[ fig21 ] ) on just three energy scales which are entirely determined by the magnetic field , the sample parameters ( i.e. inter - layer separation , well width , etc . ) , and the tunneling strength , is a specific result of the hartree - fock approximation .
the zero temperature phase diagram can thus be expressed as a function of two independent dimensionless variables @xmath142 and @xmath143 , as shown in fig .
[ fig21 ] .
this phase diagram applies to all double - layer quantum hall systems at @xmath5 which may have any values of zeeman energy , tunneling energy , layer separation , layer - thickness , etc .
we believe , however , that this remarkable scaling in the phase diagram ( which enables us to reduce an infinite number of @xmath144 versus @xmath145 diagrams for various values of @xmath20 , of which examples are shown in fig .
[ fig3 ] , to just one phase diagram shown in fig .
[ fig21 ] ) remains approximately valid , although the relative size of various phases in the universal phase diagram of fig .
[ fig21 ] may very well be quantitatively not particularly accurate .
we also mention here that this phase diagram is topologically identical to that of a @xmath146dimensional quantum @xmath9 nonlinear @xmath10-model in a magnetic field,@xcite as discussed in section [ nls ] of this paper . in this subsection
, we have studied the collective intersubband sdw excitations for @xmath5 double - layer qh systems in the time dependent hartree - fock approximation .
we have presented the dispersions of the collective sdw excitations in both the normal qh phases ( fm and sym ) and in the canted antiferromagnetic qh phase , and investigated the mode softening which signals the phase instabilities .
we have rederived the same phase diagram as that obtained in the previous section , and obtained analytic equations for the two phase boundaries separating the new canted antiferromagnetic phase from the normal fm and sym phases . in this subsection
, we discuss some thermodynamic properties of @xmath5 double - layer systems which arise from the spontaneous symmetry - breaking associated with the breaking of @xmath74 planar spin rotational symmetry in the canted antiferromagnetic quantum hall phase .
there should be a finite temperature kosterlitz - thouless transition in the canted antiferromagnetic phase , since the spin - rotational @xmath74 symmetry is broken .
below the critical temperature , the system supports a linear goldstone mode , which gives rise to a power - law temperature dependence for the specific heat . above the critical temperature
the @xmath74 symmetry is restored and the system is paramagnetic .
these properties are , in principle , experimentally observable and provide direct ways to test our theory .
we can estimate the kosterlitz - thouless transition temperature for our problem in the following manner . in the canted antiferromagnetic phase , the low temperature thermodynamics
is governed by long wavelength phase fluctuations of the order parameter .
let @xmath147 , where @xmath148 is the ground state of the canted antiferromagnetic phase , and @xmath149 , with @xmath150 as the spin operator of the j - th electron and @xmath151 is the ( magnetic field ) direction normal to the two dimensional plane . in the long wavelength limit ,
one obtains @xmath152 with @xmath153d q , \label{equ : j1}\ ] ] where @xmath154 is the magnetic length and @xmath155 ( @xmath156 ) is intralayer ( interlayer ) coulomb potential . for future convenience
, we have written the stiffness as an explicit function of @xmath20 , which arises from the dependence of the angles @xmath157 on the zeeman splitting . the effective planar @xmath158 model defined by eq .
( [ equ : e1 ] undergoes a kosterlitz - thouless phase transition @xcite at approximately @xmath159 .
finite temperature spin - wave and vortex - antivortex polarizations reduce the transition temperature to approximately @xcite @xmath160 these finite temperature renormalizations can be much larger in the vicinity of the c n and c sym phase boundaries : the expression ( [ equ : t1 ] ) can then no longer be used , and we will discuss modifications near these boundaries later in section [ nls ] .
since we know @xmath161 exactly within the microscopic hartree - fock approximation , the kosterlitz - thouless transition temperature can be easily determined for our problem . in fig .
[ fig33 ] , we show the calculated kosterlitz - thouless critical temperature in @xmath5 double - layer quantum hall systems within the mean - field hartree - fock approximation ( i.e. from eq .
( [ equ : j1 ] ) for @xmath161 ) .
the phase transition exists only in the canted antiferromagnetic quantum hall phase .
the critical temperature vanishes at the phase boundaries as the symmetry - breaking order parameter drops continuously to zero as the phase boundaries are approached from within the canted antiferromagnetic phase .
we notice that the calculated kosterlitz - thouless temperature ( @xmath162 ) is well within the experimentally accessible regime for typical algaas / gaas based double - layer systems .
the effective spin - stiffness @xmath161 given in eq .
( [ equ : j1 ] ) is obtained in the mean - field hartree - fock approximation , i.e. using the results from sections [ grou ] and [ sdw ] , where quantum fluctuation effects are not included .
the results in fig .
[ fig33 ] should thus be regarded as the upper bound for the kosterlitz - thouless critical temperature .
we emphasize that the kosterlitz - thouless transition discussed here is present even in the presence of interlayer tunneling ( in fact , the presence of finite interlayer tunneling is essential to stabilize the canted antiferromagnetic phase , as described in the last two sections ) , unlike the case associated with the pseudospin transition @xcite at @xmath163 ( @xmath164 odd integers ) where interlayer tunneling suppresses kosterlitz - thouless transition . below the kosterlitz - thouless transition temperature , the specific heat in the antiferromagnetic phase has qualitatively different temperature dependence from those of the normal quantum hall phases .
this is of practical significance since it is possible to experimentally measure the specific heat of quantum hall systems.@xcite at low temperatures , the main contribution to the specific heat comes from long wavelength low energy intersubband sdws . with their dispersions calculated in each of the quantum hall phases , the heat capacity is easily obtained : @xmath165 , where @xmath166 is the energy of the low - lying intersubband sdw excitation and @xmath167 is its bose occupation factor .
the results are shown in fig .
[ fig34 ] .
it is clear that the specific heat has an activated behavior in the normal quantum hall phases because of the existence of excitation gap in its spin wave spectra , and a quadratic power - law temperature dependence in the canted antiferromagnetic phase because of the existence of the linear goldstone mode in the symmetry broken phase .
the spin stiffness goes to zero discontinuously at @xmath168 , and for @xmath169 we have the usual disordered @xmath170@xmath171 phase of the kosterlitz - thouless transition .
our analysis so far has obtained solutions for the fm , sym and c phases obtained by varying the parameters @xmath20 , @xmath19 , @xmath145 in the hamiltonian ( see figs .
[ fig3 ] and [ fig21 ] ) which modify the relative strengths of the zeeman energy , the tunneling energy , and the coulomb interaction energy , respectively .
generically , these phases separated by phase boundaries representing second - order quantum transitions .
however , there are also special quantum multicritical points in figs [ fig3 ] and [ fig21 ] whose physical significance we will now discuss .
the first quantum multicritical point is apparent in fig .
[ fig3 ] where the fm , c and sym phases come together at a single point on the abscissa corresponding to vanishing interlayer separation ( @xmath65 ) .
it is easily noted from fig .
[ fig3 ] that this quantum multicritical point is in fact defined by @xmath172 which is equivalent to the conditions @xmath173 using the definition of @xmath140 given immediately following eq .
( [ equ : hi ] ) in section [ grou ] .
the simple physical reason for the vanishing of @xmath174 along the @xmath65 line is that the intra- and interlayer coulomb interactions are identical in the limit of vanishing interlayer separation @xmath145 .
note also that the vanishing of @xmath175 ( and consequently of @xmath139 ) pushes the quantum multicritical point to an infinite value of the abscissa ( @xmath176 ) in the scaled universal diagram given in fig .
[ fig21] in fig .
[ fig21 ] the two phase boundaries separating the three phases approach each other asymptotically as @xmath176 and @xmath177 at the multicritical point .
note that the condition @xmath178 for the quantum multicritical point is a particularly interesting criterion because , in the absence of our predicted canted antiferromagnetic phase ( i.e. if the @xmath5 double - layer qh systems allowed only the ferromagnetic and the symmetric phases , as was assumed in the literature before our work ) , the condition of the equality of the zeeman splitting and the symmetric - antisymmetric gap ( i.e. @xmath178 ) is precisely the single particle _
level crossing _
criterion where , at @xmath5 , one would make a transition from the ferromagnetic phases where the two up - spin symmetric and antisymmetric levels are occupied and the down - spin levels are empty for @xmath81 to the symmetric ( spin singlet ) phase where the spin - up and spin - down symmetric subbands are occupied ( and the antisymmetric levels are empty ) for @xmath63 .
what our theory definitely predicts is that such a simplistic one particle level crossing picture ( which appears to be obvious intuitively ) _ does not occur _ in a double - layer qh system at @xmath5 instead coulomb interaction breaks the @xmath179 spin rotational symmetry and drives the system into an intervening antiferromagnetic phase where spin and pseudospin levels are intrinsically mixed .
the fact that the intuitively expected level crossing phenomenon ( at @xmath178 ) has never been observed @xcite in a @xmath5 double - layer qh system in spite of systematic efforts @xcite is , in our opinion , rather strong indirect evidence in support of our phase diagram .
the second multicritical point becomes apparent only in the universal phase diagram shown in fig .
[ fig21 ] ( and can be inferred implicitly from the trend that can be seen in the phase diagrams shown in fig .
[ fig3 ] ) .
its existence is a consequence of the intriguing finding that our antiferromagnetic state , in fact , persists all the way to @xmath180 ( as can be clearly seen in fig .
[ fig21 ] where a finite region of the antiferromagnetic state exists along the @xmath180 line ) where the spin - polarized ferromagnetic phase no longer exists , and the antiferromagnetic phase is separated from the spin singlet phase by a multicritical point ( m ) defined by the condition @xmath181 thus the critical line defining the phase boundary between the antiferromagnetic and the symmetric phases for @xmath182 ends at a critical point ( m ) for @xmath180 .
it is evident that in the absence of any zeeman energy ( @xmath180 ) the spin magnetic moment in each layer lies completely in the 2d plane of the electron gas where they must be equal and opposite in the two layers .
therefore , the @xmath180 antiferromagnetic phase of fig .
[ fig21 ] is _ not _ a _ canted _ phase , but is a purely nel phase ( n ) ; indeed the hamiltonian has full @xmath179 spin rotation symmetry for @xmath183 , and spin moments in the n phase can point in any two anti - parallel directions .
the n , c , and sym phases meet at the multicritical point m. this multicritical point will take on special significance in our effective field theoretical formulation in the next section .
let us also note that the existence of this purely nel qh antiferromagnet at @xmath5 double - layer system may not be just a theoretical curiosity because it is possible to obtain vanishing zeeman splitting in a gaas double - layer system in a finite magnetic field situation by applying external pressure which under suitable conditions could lead to the vanishing of the effective gyromagnetic ratio ( the @xmath184-factor ) due to band structure effects . before concluding this section , and going on to the effective field theoretic description of the double - layer qh system
, we will discuss the relationship of our results to some earlier work on double layer systems .
we will also use this opportunity to comment on the validity of the hartree fock approximation in our and earlier work .
most earlier studies,@xcite however , have focused on @xmath137 ( with some work@xcite on @xmath185 ) .
although the @xmath137 and the @xmath5 qh systems exhibit some similarities such as the softening of their low energy collective excitations under certain conditions , there are important distinctions between them . at @xmath137 , the spin degree of freedom is normally frozen out by the external magnetic field .
the relevant low energy excitations in the @xmath137 qh state are therefore intersubband charge - density - wave excitations , and the properties of the system are determined by the interplay between interlayer tunneling energy and coulomb interaction energy . in this sense , the @xmath137 system is in fact a single - layer system with a layer pseudospin - dependent interaction.@xcite at @xmath5 , both the spin degree of freedom and the layer degree of freedom are relevant , and the low energy excitations are intersubband sdw excitations .
consequently , the properties of the system are determined by the interplay among tunneling energy , zeeman energy , and coulomb interaction energy . because of the increased degree of freedom , the system has more ways to optimize the total energy , and new states which are not possible at @xmath137 become possible at @xmath5 . the symmetric qh state is energetically favored at small layer separations because it optimizes the tunneling energy ; the spin polarized qh state is favored at large layer separations because it optimizes the coulomb interaction energy ; the canted antiferromagnetic state is energetically favored at intermediate layer separations .
the reason for this is that the canted antiferromagnetic state tends to simultaneously optimize both the tunneling energy and the coulomb interaction energy , which prevails at intermediate layer separations where the tunneling energy and the coulomb interaction energy are equally important .
both the canted antiferromagnetic state and the symmetric state exist only for systems with small enough zeeman energy , as the zeeman energy clearly favors the spin polarized state .
another important distinction between the @xmath137 systems and the @xmath5 systems is that although at @xmath137 the mode softening destroys the qh effect@xcite , and there is no reliable description of the electronic state in the non - qh phase because beyond the critical layer separation the system becomes effectively a pair of isolated layers with compressible half - filled landau level states ; in contrast , at @xmath5 , the qh effect prevails at all phases because there is always a charge gap in both layers ( even as @xmath186 ) , and we have good understanding of the ground state and the low energy excitations in each phase due to the existence of incompressible filled landau levels . nevertheless , the mode softening and the associated phase transitions at @xmath5 are likely to be observable through inelastic light scattering experiments@xcite and thermodynamic measurements . our work has studied @xmath5 double - layer systems by numerically solving the self - consistent mean - field equations@xcite , and obtained collective excitation dispersions using many - body diagrammatic techniques . @xcite both approaches are , however , based on the hartree - fock approximation . in single - layer integer qh systems ,
calculations @xcite in the hartree - fock approximation agree well with experiments.@xcite in double - layer systems , the hartree - fock approximation is less accurate because coulomb interaction potential is layer - index dependent .
nevertheless , we expect that the hartree - fock approximation remains a reasonably good description for a double - layer system at @xmath5 , since the hartree - fock ground state , which is non - degenerate and separated in energy from higher levels , is a good approximation for the real many - body ground state at @xmath5 due to the existence of incompressible filled landau level states with charge excitation gaps at any layer separations .
we want to especially emphasize the difference in the validity of the hartree - fock approximation between @xmath137 and @xmath5 .
the approximation is valid at @xmath137 only at small layer separations and fails completely beyond a critical layer separation where the system becomes effectively a pair of isolated layers with compressible half - filled landau level state in each layer . at @xmath5 ,
incompressible states with filled landau levels exist at any layer separations .
in particular , there is still one filled landau level in each layer at @xmath66 .
this fact , namely the existence of an incompressible energy gap at all layer separations , ensures that the hartree - fock approximation , upon which our calculations are based , is a reasonable formalism at @xmath5 regardless of the value of the layer separation .
the hartree - fock analysis used in the previous sections has the advantage of working with a realistic microscopic hamiltonian and of making definite quantitative predictions for experimental observables in realistic samples . in this section
, we will present an alternative analysis based upon a continuum effective quantum field theory for the low - lying spin excitations of a double layer quantum hall system .
we will find that the global phase diagrams obtained in the two approaches are very similar , and are , in fact , topologically identical , and that detailed additional predictions for the temperature dependence of various observables can be made by a combination of the two methods . in particular , some advantages offered by the continuum approach are : * it will become clear from the analysis below that there are two basic ingredients necessary to obtain the phases of fig [ fig21 ] : two well separated layers form fully polarized ferromagnets with a gap towards charged excitations ( i.e. an incompressible qh effect gap ) , and the primary coupling between the layers is an antiferromagnetic exchange ( i.e. a superexchange ) interaction . as such , we expect a similar phase diagram to apply not only at filling @xmath187 , but also at any @xmath188 , where @xmath1 is any filling fraction where a single layer has a charge gap , and is fully polarized . in particular , this criterion is satisfied at @xmath189 , @xmath164 an odd integer , where each layer forms a polarized laughlin fractional quantum hall state .
the hartree - fock analysis clearly can not be applied for @xmath190 , as the single layer charge gap appears only after inclusion of the non - trivial correlations implicit in the laughlin state .
* the hartree - fock theory significantly overestimates the energy of the spin - unpolarized symmetric ( sym ) or the spin singlet ( ss ) state , as we will refer to it in this section .
spin up and spin down electrons are simply placed into the same orbitals which are symmetric in the layer index .
this is costly in coulomb energy as there are no correlations in the layer positions of the up and down spin electrons .
it is clearly more advantageous to form a spin singlet states between electrons which are localized in opposite layers .
the non - linear sigma model continuum field theory to be discussed below does this in a natural way . from now on in this section we refer to this symmetric or the spin singlet phase as the ss to emphasize its correlated singlet nature . *
a number of quantum - critical points have been uncovered in the hartree - fock analysis .
there is the @xmath191 quantum critical point between the spin singlet ( ss or sym ) and the nel ( n ) phases , and a critical line between the ss and the canted antiferromagnetic ( c ) phases .
our continuum approach will obtain the critical theory for these transitions , and we will find that they have dynamic critical exponents @xcite @xmath192 and @xmath193 respectively . there is also a second critical line between the c and the fully spin polarized ferromagnetic ( fm ) phases : this transition has @xmath193 and will be discussed only in passing , as the critical theory is rather similar to one of the models discussed in detail in ref .
* the continuum theory offers not only provides us the zero temperature quantum phase diagram but also a streamlined approach to the study of properties at non - zero temperature , especially in the vicinity of the quantum critical points where effects of fluctuations can not be neglected .
the price one pays is that in general the parameters defining the effective field theory are quantitatively unknown and can only be calculated from a microscopic theory such as the hartree - fock theory of the previous sections .
we motivate our formulation of the continuum theory by consideration of the physics of two well - separated identical layers at @xmath194 .
more specifically , the layer separation , @xmath145 , is much larger than the magnetic length , @xmath195
. then the two layers ( labeled 1,2 ) are approximately decoupled , and each separately has filling fraction @xmath196 .
their ground states will be the familiar laughlin states for @xmath190 , or a fully filled lowest landau level at @xmath197 , both of which are incompressible states with large energy gap to all charged excitations . these states are also fully spin polarized ; the spin polarization is induced not just by the zeeman coupling to the external magnetic field , but also by the significantly larger intralayer ferromagnetic exchange @xcite .
the low - lying excitations in each layer are spin waves which have a small excitation gap given precisely by the zeeman energy @xmath198 . for small @xmath184 , a complete description @xcite of the low energy excitations of each layer
can be given in terms of an action for unit vector fields @xmath199 ( @xmath200 ) representing the orientation of the ferromagnetic orders .
spin waves are small fluctuations of @xmath199 about an ordered state , while charged quasiparticles are skyrmion @xcite textures of @xmath199 .
the effective action describing the two layers is @xcite ( in units with @xmath201 ) @xmath202 + { \cal l}_f [ \vec{n}_2 ]
\right ) \nonumber \\ & & { \cal l}_f[\vec{n } ] \equiv im_0 \vec{a}(\vec{n } ) \cdot \frac{\partial \vec{n}}{\partial \tau } + \frac{\rho_s^0}{2 } \left ( \nabla_x \vec{n } \right)^2 - m_0 \delta_z n_z \label{action0}\end{aligned}\ ] ] here @xmath203 is the magnetization density per layer , with @xmath154 the magnetic length .
the spin stiffness of each well separated layer is represented by @xmath204 ; for @xmath197 , we have the exact result @xcite @xmath205 , while for @xmath190 numerical estimates of @xmath204 are given in ref @xcite .
the term involving @xmath206 accounts for the berry phase accumulated under time evolution of the spins ; here @xmath206 is any functional of @xmath207 which satisfies @xmath208 this berry phase term also has a `` dual '' interpretation in the picture in which @xmath209 is viewed as an action for skyrmions @xcite : it represents the coupling of the skyrmion current to a `` magnetic field '' of strength @xmath210 .
now imagine reducing the value of @xmath145 to couple the two layers . as there is a charge gap in each layer
, we can neglect all charge transfer processes , and focus solely on spin exchange .
because of the strong repulsive interactions within each layer , we expect by an extension of the familiar arguments made in the context of the hubbard model that there will be an _ antiferromagnetic _ superexchange coupling between the layers
. this can also be inferred easily by considering the leading effect of interlayer tunneling and pauli principle , which immediately provides a superexchange coupling between the layers .
the complete double layer action is therefore @xmath211 + { \cal l}_f [ \vec{n}_2 ] + j
\vec{n}_1 \cdot \vec{n}_2 \right ) \label{action1}\ ] ] the value of the interlayer exchange , @xmath212 , is not known precisely ; we expect that it is of order @xmath213 where @xmath19 is the tunneling matrix element ( see eq .
( [ equ : h1 ] ) , for example ) between the layers , and @xmath214 is the coulomb repulsion energy .
in addition to the imprecisely known @xmath212 , the present approach also requires knowledge of the nature of the short distance cutoff at lengths of order @xmath215 beyond which present continuum approach can not be applied .
we will show later that our ignorance of these quantities can be reduced entirely to uncertainties in the value of a certain renormalized energy scale .
this energy scale can be either measured directly in an experiment , or computed by any microscopic theory such as the hartree - fock approach ( appropriate at @xmath5 ) described in sections [ grou][kt ] .
apart , from this single energy scale , however , all of the predictions of the present effective field theoretical approach will be quantitative and precise .
some potentially important terms have been omitted from @xmath216 and our analytic computations : the hopf term which endows the skyrmions with fractional statistics , and the long - range coulomb interaction between the skyrmions .
we believe this is permissible because of the charge gap .
further @xcite , as the layers are antiferromagnetically correlated , skyrmions in one layer will be correlated with anti - skyrmions in the other , and this neutralizes the leading contribution of both terms .
this latter argument should continue to hold even if the charge gap were to vanish at a quantum critical point ( the charge gap remains non - zero at the quantum critical points in both our present calculations ) .
note also that no new term is necessary to induce charge transfer between the layers : a hedgehog / anti - hedgehog pair in the two layers corresponds to an event transferring skyrmion number between them .
such spacetime singularities are absent in the strict continuum limit but appear when a short - distance regularization is introduced . for completeness
, we note that the purely @xmath207 field formulation becomes incomplete for @xmath190 and larger @xmath184 , as the spin zero laughlin quasiparticles can become the lowest energy charged excitations .
these should , in principle , be accounted for by a separate complex scalar field .
however , these can also be neglected for the same reasons presented above for non - zero spin charged excitations .
we now manipulate the effective action into a form more suitable for our subsequent analysis .
we solve the constraints @xmath200 by representing @xmath217 where @xmath207 and @xmath218 are vectors satisfying @xmath219 note that this representation is so far exact .
next , we insert ( [ nl ] ) into @xmath220 .
because the layers are antiferromagnetically correlated we expect that @xmath218 will not be too large , and it is therefore permissible to expand the resulting action to quadratic order in @xmath218 .
this is clearly an approximation : in appendix [ twospin ] we examine a model solvable hamiltonian by the same method in order to assess the damage done by neglecting terms higher order in @xmath218we find that this procedure obtains the low energy spectrum correctly but introduces some spurious states at higher energies . to quadratic order in @xmath218 , @xmath220 takes the form @xmath221 , \label{action1a}\ ] ] where @xmath151 is a unit vector in the direction of the magnetic field .
now we integrate out @xmath218 while maintaining the constraint @xmath222 by adding an additional term to the energy @xmath223 and then taking the limit @xmath224 .
this yields the following effective action for the antiferromagnetic order parameter @xmath207 @xmath225 \label{action2}\ ] ] where @xmath226 this is precisely the action of the 2 + 1 dimensional quantum @xmath9 non - linear sigma model in a field @xmath227 coupling to the conserved global @xmath9 charge.@xcite it is expected to apply to double - layer quantum hall systems with @xmath194 at length scales larger than @xmath228 .
the remainder of this section consists of a detailed analysis of the properties of @xmath229 .
the techniques and some results have already been presented earlier in refs : we shall present here a unified treatment with a special emphasis on dynamical properties at non - zero temperature .
we begin in section [ mft ] by developing a simple mean - field phase diagram of @xmath229 .
this section will summarize the results of the application of the mean field theory of ref to the action @xmath229 .
formulation of the mean field theory requires some short distance regularization , and we choose to place the continuum theory on a square lattice in the spatial directions , with a lattice spacing @xmath230 ; a continuum formulation is maintained along the time direction .
the resulting action is equivalent to the following lattice quantum rotor hamiltonian @xmath231 where the coupling constants in @xmath232 are @xmath233 the hamiltonian is expressed in terms of operators @xmath234 which represent the orientation of the rotors , and @xmath235 which are the rotor angular momenta .
the operators on different sites commute , while those on a single site obey the commutation relations ( dropping the site indices ) @xmath236 & = & i \epsilon_{\alpha\beta\gamma } \hat{l}_{\gamma } \nonumber \\ \left [ \hat{l}_{\alpha } , \hat{n}_{\beta } \right ] & = & i \epsilon_{\alpha\beta\gamma } \hat{n}_{\gamma } \nonumber \\ \left [ \hat{n}_{\alpha } , \hat{n}_{\beta } \right ] & = & 0\end{aligned}\ ] ] we will describe the properties of @xmath232 by choosing the best among the mean field hamiltonians given by @xcite @xmath237 here @xmath238 ( @xmath239 ) is the lattice co - ordination number , and @xmath240 is a variational parameter to be chosen so that the expectation value of @xmath232 in the ground state of @xmath241 is as low as possible ; by the usual argument , this is expected to happen when @xmath242 . as in ref , we numerically diagonalized @xmath241 by truncating its spectrum at some large angular momentum , and then optimized the value of @xmath240 .
the resulting phase diagram is shown in fig [ mftfig ] .
we discuss the properties of the various phases in turn : each rotor is in its non - degenerate @xmath243 state , @xmath244 , and there is a gap to all excitations .
the ground state is a spin singlet , and is therefore unaffected by variations in the value of @xmath20 .
again , @xmath244 , each rotor now has azimuthal angular momentum @xmath246 and this value remains pinned as various parameters are varied .
each rotor is also in precisely the state with @xmath247 , although this latter feature is a special property of mean field theory which will not survive fluctuations . of these phases ,
only the @xmath248 case is actually allowed for the double layer action @xmath220 , and it is clearly the fm phase of fig [ fig21 ] .
the other phases are an artifact of the approximations made in mapping @xmath220 to @xmath229 : this should be clear from the discussion in appendix [ twospin ] where we show that expanding in powers of @xmath218 introduce spurious higher angular momenta states .
these states have @xmath249 and varying continuously as the parameters are varied ; we have @xmath250 , @xmath251 and @xmath252 . from ( [ nl ] )
, this implies that the two layers have opposite spin polarizations in the @xmath253 plane .
the two layers also have an identical ferromagnetic polarization , given by @xmath254 which is oriented along the @xmath255 direction .
this ferromagnetic moment varies continuously as parameters are varied , and vanishes when @xmath256 .
so for general @xmath257 this state is canted ( c ) , while for @xmath180 it is a pure nel ( n ) antiferromagnet .
the c phase has a single linearly dispersing spin wave mode in the @xmath253 plane , while the n phase has two spin waves @xcite . in the remainder of this section
, we will present a detailed theory of the universal properties of the system in the vicinity of the multicritical point m. this is the same quantum multicritical point ( m ) which exists in the universal hartree - fock phase diagram of fig .
[ fig21 ] where the n phase ( along the @xmath180 line ) , the c phase and the sym ( ss ) phase come together at @xmath258 .
we point out in this context that the other distinct multicritical point of the hartree - fock theory where the canted antiferromagnetic phase , the ferromagnetic phase , and the symmetric phase coexist ( the point on the abscissa defined by @xmath65 and @xmath178 in fig .
[ fig3 ] ) is not accessible within the effective field theory due to the long wavelength restriction @xmath259 .
( we mention that in our notations @xmath20 in the hartree - fock theory corresponds to just @xmath227 in the field theory due to our choice of units . )
note that the c , n , and ss phases meet at m , and so we will also discuss the universal second - order transitions between them .
we will not discuss the nature of the second - order transitions between the @xmath260 and @xmath261 phases : very closely related transitions , in the same universality class , have been discussed in some detail in ref . a first study of the properties in the vicinity of the point m has appeared in ref using a large @xmath108 expansion in a non - linear sigma model with @xmath108 component fields .
the issues of interest here are more conveniently obtained using a recently developed expansion @xcite in spatial dimensionality , @xmath145 , in powers of @xmath262 .
the latter approach is expressed in terms of a soft - spin field theory , and we therefore begin with a soft - spin version of the non - linear sigma model @xmath229 : @xmath263 \label{hphi}\end{aligned}\ ] ] here @xmath264 is the soft - spin field which measures the staggered moment of the two layers .
we have taken the magnetic field to point in the @xmath255 direction .
we will also be interested in the uniform ferromagnetic moment density of the system , @xmath265 , and this is given by @xmath266 where @xmath267 is the free energy density associated with the action @xmath268 .
we have introduced two new coupling constants , @xmath269 and @xmath270 in @xmath271 ; these are related to the coupling @xmath272 of @xmath229 , and its short - cutoff @xmath273 .
we will not specify the precise values of these parameters here , as they merely appear at intermediate stages of our computation , and not in our final results .
let us first discuss the mean field properties of @xmath268 , obtained simply by minimizing the action while ignoring all spatial and time dependence of @xmath274 . for @xmath275
, the ground state has @xmath276 , and is therefore in the quantum paramagnetic ss phase .
for @xmath277 , the ground state has @xmath278 and in the @xmath253 plane .
this is the c phase and the fields have the expectation values @xmath279 or any rotation of @xmath274 in the @xmath253 plane .
notice that @xmath265 vanishes for @xmath180 , and therefore the line @xmath280 , @xmath180 is the n phase .
the resulting mean field phase diagram is shown in fig [ landau ] .
notice that the vicinities of the points m are very similar in figs [ mftfig ] and [ landau ] .
the quantum critical point m is at @xmath180 , @xmath281 , and it is clear from the lorentz - invariant structure of @xmath268 at @xmath191 that this point has dynamic exponent @xmath192 .
rotations of the order parameter @xmath274 in the @xmath253 plane have associated with them a stiffness @xmath282 given by @xmath283 this is the same stiffness which was computed in section [ kt ] and eqn [ equ : j1 ] in the hartree - fock theory .
we now include the effects of fluctuations at one loop .
we will quote results for the dynamic longitudinal and transverse susceptibilities of the @xmath274 field which are measured in light scattering . recall that in terms of the spin polarizations of the two layers @xmath284 , @xmath285 , we have @xmath286 . we define ( with @xmath287 as the temperature @xmath288 in our units in this section ) @xmath289 and @xmath290 we can use the methods of ref to compute the one loop values of these susceptibilities in the ss phase ( this is the phase with no broken symmetry ) in the vicinity of the point m ; we obtain @xmath291 here the quantity @xmath292 is an observable defined by @xmath293 the value of @xmath292 should either be measured experimentally , or computed by a detailed microscopic calculation like the hartree - fock theory discussed earlier in the paper .
we will express all our results for @xmath294 completely in terms of universal functions of parameters @xmath292 , @xmath287 and @xmath227 ( so that the microscopic couplings @xmath269 and @xmath270 do not appear anywhere in our results . ) clearly , in the mean field theory @xmath295 ; at one loop order , we have @xmath296 , where the exponent @xmath297 . we need a separate experimental observable to measure the deviation of the system from the point @xmath265 at @xmath191 for @xmath280 .
a convenient choice , also used in refs is the spin stiffness .
we therefore define @xmath298 all our results for @xmath280 will be expressed in terms of @xmath299 .
again @xmath300 , and the actual value of @xmath299 should be measured experimentally or computed in hartree - fock or microscopic numerical studies of the double - layer hamiltonian . before closing this subsection , we draw attention to the fact that there are two phase boundaries that terminate at the point m : the ss to c transition and the n to c transition . in the vicinity of these transitions
the response functions computed near the critical point m should turn into _ reduced _ scaling functions @xcite characteristic of the respective phase transitions . in the following subsections ,
we discuss simplified versions of the action @xmath268 which can be used to compute these reduced scaling functions .
in this region we can neglect @xmath302 fluctuations and focus only on the @xmath303 which is undergoing bose condensation .
further , it can also be shown that the second - order time derivative in @xmath268 can be dropped . making these approximations , and defining @xmath304
we see that @xmath268 reduces to @xmath305 . \label{hpsi}\ ] ] this action has been previously studied in some detail @xcite : it has a @xmath193 quantum critical point at @xmath306 , and we will use the existing results later . thus the ss - c transition is a line of @xmath193 critical points terminating in @xmath192 critical end - point m. both the @xmath108 and @xmath261 phases are ordered , and it is sufficient to simply focus on static , thermal , orientational fluctuations of the order parameter .
we therefore quench the magnitude fluctuations of @xmath274 and return to the fixed length vector @xmath207 .
the effective action for static @xmath207 fluctuations can be deduced from @xmath268 to be @xmath308 .
\label{hn}\ ] ] as noted earlier , @xmath299 is the spin stiffness of the nel state , fully renormalized by quantum fluctuations .
the anisotropy @xmath309 to lowest order in @xmath270 , and we expect @xmath310 more generally . the action @xmath311 has been studied in ref , and we will use their results in the following subsection . a number of new phenomena occur at non - zero temperatures , and these are addressed in a relatively straightforward manner using the present continuum effective field theory approach . * there is a broken @xmath253 symmetry in the c phase , and therefore a non - zero temperature ( @xmath168 ) at which this order disappears in a kosterlitz - thouless transition .
an estimates @xmath168 was given earlier ( sections [ kt ] and fig .
[ fig33 ] ) in the hartree - fock theory which is valid when the system is well away from one of the @xmath312 phase boundaries of the c phase in figs [ mftfig ] , and [ landau ] .
we expect @xmath168 to vanish continuously as the system in the c phase approaches the @xmath312 boundaries to the n or the ss phase : there is nonzero temperature phase transition above the n or the ss ground state .
we discuss below the behavior of @xmath168 near the c - n and c - ss @xmath312 phase boundaries . near the point m , @xmath168 is determined completely and universally by the two energy scales which measure the deviation of the ground state from m. so for @xmath294 we expect @xmath313 where @xmath314 is a fully universal function ; because the ss - c phase boundary occurs precisely at @xmath315 , we have @xmath316 similarly for @xmath280 we have @xmath317 where @xmath318 is also a universal function .
clearly the two functions should agree at @xmath281 , and therefore we have @xmath319 ; actually it is possible to say much more for @xmath320 we expect that @xmath168 is a smooth and analytic as a function of @xmath269 through @xmath281 , and so using the dependencies of @xmath321 and @xmath299 on @xmath269 , it is possible to express @xmath322 as analytic continuations of each other .
we will give explicit expressions for @xmath322 to leading order in the expansion in @xmath262 below . *
the one - loop @xmath312 results for the ss phase ( [ chiteq0 ] ) predict infinitely sharp absorption peaks in @xmath323 at @xmath324 , and in @xmath325 at @xmath326 . as the ss phase has a spin gap , we expect these infinitely sharp peaks to survive at higher orders in the perturbation theory at @xmath312 . for @xmath327 two
qualitatively new features will arise .
first , thermal damping will lead to a broadening of the peaks .
second , the peak positions will themselves become temperature dependent .
we will describe these processes below in the vicinity of the point m , where both the broadening and the @xmath287-dependent shifts are quite significant .
deep inside the ss phase , well away from the m point , these @xmath287-dependencies are exponentially activated , and therefore much weaker .
we will restrict our results for the most part to the paramagnetic phase , although results in the magnetically ordered phases can be obtained by very similar methods .
this means that we are working at @xmath327 above the ss phase , and at @xmath328 above the c phase , all within the vicinity of the point m. the results are obtained using methods discussed in some detail in ref : the only change is that the zeeman splitting @xmath20 has to be included in the propagators for the @xmath329 fields , and this modifies the values of the matsubara frequency summations in the loop diagrams by replacing an energy @xmath330 by @xmath331
. the reader may also consult appendix d of ref where a simpler derivation of just the one loop results of ref is given .
the non - zero @xmath287 generalization of ( [ chiteq0 ] ) takes the form @xmath332 here @xmath333 and @xmath334 depend implicitly upon the energy scales @xmath287 , @xmath335 , and @xmath292 ( @xmath299 ) for @xmath294 ( @xmath280 ) in a manner we shall describe below to lowest order in @xmath336 . clearly , the `` masses '' @xmath333 represent the peak absorption frequency , while @xmath334 are the absorptive pieces which lead to a @xmath287-dependent broadening of the line .
first we describe the behavior of @xmath337 .
closely related results can be obtained for @xmath280 . in this case , the masses are universal functions of @xmath299 , @xmath335 and @xmath287 . however
, considerable ambiguity arises in the @xmath336 expansion for the result because @xmath299 does not simply have the dimensions of energy for all @xmath145 . the appropriate scaling variable @xcite is @xmath358 , and it is necessary to keep the full @xmath359 power , rather than expand it in powers of @xmath336 in order to make the engineering dimensions of the results come our correct .
this then leads to ambiguities as to precisely which numerical factors should be raised to the power @xmath359 and which to @xmath360 . a convenient choice , which leads to the most compact expressions is to define @xmath361 where we have written the general expression for the @xmath362-component order parameter : in the present case @xmath363 .
the factor @xmath364 is a phase - space factor and is given by @xmath365 $ ] ( this factor was inadvertently omitted in ref ) .
notice that @xmath366 has the dimensions of energy in @xmath367 ( which is of interest here ) .
the value of @xmath366 , however must be regarded as subject to large systematic corrections , in view of the ambiguities noted above . using the methods and results of ref for @xmath280 , and expressing them in terms of @xmath366 , we find that the results ( [ resm ] ) still hold , but ( [ resr ] ) are replaced by @xmath368 + \frac{\epsilon t^2}{11 } \left [ 3 { g}\left(- \frac{\widetilde{\rho}_s^2 } { 2 t^2 } , 0\right ) + 2 { g}\left(- \frac { \widetilde{\rho}_s^2 } { 2 t^2 } , \frac{\delta_z } { t } \right ) \right ] \nonumber \\ r_{\perp } = & & - \frac{\widetilde{\rho}_s^2 } { 2 } \left [ 1 -\frac{\epsilon}{22 } + \frac{5\epsilon}{11 } \ln \frac{t}{\widetilde{\rho}_s } \right ] + \frac{\epsilon t^2}{11 } \left [ { g}\left(- \frac { \widetilde{\rho}_s^2 } { 2 t^2 } , 0\right ) + 4 { g}\left(- \frac{\widetilde{\rho}_s^2 } { 2 t^2 } , \frac{\delta_z } { t } \right ) \right].\end{aligned}\ ] ] notice that @xmath369 is now needed for negative values of @xmath370 . despite appearances , the expression ( [ resg ] ) actually also holds for @xmath371one simply uses the identity @xmath372 when the square root becomes purely imaginary . indeed , it is not difficult to show that the expression in ( [ resg ] ) is actually analytic for all real @xmath373 provided @xmath374 .
we can use the same stability condition used for @xmath294 to now obtain the leading order @xmath336-expansion result for @xmath168 : @xmath375 which is of the scaling form ( [ tkt2 ] ) .
the @xmath336 expansion fails when @xmath376 where the system approaches the c - n phase boundary ; here , we use the effective action @xmath311 of section [ nc ] , and results for it in ref to obtain @xmath377 finally , we obtain the damping coefficients @xmath378 .
this requires evaluation of two - loop diagrams and the results are extremely lengthy .
we will be satisfied here by simply quoting the results valid for @xmath379 , @xmath380 which were obtained in ref : @xmath381 where @xmath382 is the dilogarithm function @xmath383 the effective field theory for the double - layer qh system at a filling factor of @xmath384 ( with @xmath164 an odd integer ) that we develop above is entirely built on the effective action @xmath385 , defined by eq .
( [ action2 ] ) . in particular
, we make use of the fact that this effective action for our problem is identical to the action of the @xmath386 dimensional @xmath9 quantum non - linear @xmath10 model @xcite with the additional feature of an external magnetic field coupled to the conserved global @xmath9 charge .
once this precise mapping of our effective action to that of the @xmath386 dimensional @xmath9 quantum nonlinear @xmath10 model becomes explicit , the rest of the results derived in sections [ nls]a c follow naturally .
the question now arises about the correspondence between our effective field theory results in this section and the microscopic hartree - fock results ( for @xmath5 ) described in sections [ grou][kt ] .
it is to be noted that both the microscopic hartree - fock theory ( sections [ grou][kt ] ) and the effective nonlinear @xmath10 model field theory predict the same number of zero temperature quantum phases , namely the fully spin polarized ferromagnetic , the canted antiferromagnetic , the nel , and the symmetric spin singlet phase , for the double - layer qh system at @xmath5 .
( the effective field theory , in addition , enables us to predict that the double - layer system at all fillings @xmath3 , where @xmath387 with @xmath164 odd , has these four phases with the spin singlet phase in the general case being a non - trivial correlated ss phase rather than just the pseudospin - symmetric spin - antisymmetric sym phase of the @xmath5 hartree - fock theory . ) it should also be noted that both the hartree - fock theory and the effective field theory predict the existence of a finite temperature kosterlitz - thouless phase transition in the canted antiferromagnetic phase with the planar antiferromagnetic ordering disappearing above the kosterlitz - thouless transition temperature .
the underlying physics in both the theories is that the system is essentially an @xmath170@xmath171 antiferromagnet in the layer ( i.e. in the plane perpendicular to the magnetic field direction ) in the new canted phase . on a more quantitative level
it is easy to show that both theories predict the same topology of the zero temperature quantum phase diagram .
this is demonstrated in fig .
[ fig15 ] where we have redrawn the hartree - fock phase diagram ( fig .
[ fig15]a ) of fig .
[ fig21 ] inverting abscissa ( from @xmath388 to @xmath389 ) and have somewhat reconfigured the effective field theory phase diagram ( fig .
[ fig15]b ) from fig .
[ mftfig ] by keeping only the @xmath390 phase and by modifying the relative size of the various phases ( which are arbitrary within the effective field theory ) . using the definitions @xmath391 from eq .
( [ equ : tt1 ] ) to define the abscissa , the effective field theory phase diagram in @xmath392 space ( fig .
[ fig15]b ) can be seen to be identical topologically to the quantitatively calculated hartree - fock phase diagram ( for @xmath5 ) in the @xmath393 space ( fig .
[ fig21 ] ) .
note that , in addition to the identical topology involving four distinct quantum phases as shown in fig .
[ fig15 ] of the two phase diagrams with the effective coupling parameter @xmath272 of the field theory ( the abscissa in fig .
[ fig15]b ) being proportional to the parameter @xmath388 ( the abscissa in fig .
[ fig15]a ) of the hartree - fock theory ( which is expected , because @xmath394 with @xmath212 being the interlayer superexchange coupling ) and the ordinate ( @xmath395 ) being the same in both figs .
[ fig15]a and [ fig15]b , the multicritical point m on the zero magnetic field line shows up in both phase diagrams . at the ( zero temperature ) quantum multicritical point m , the canted , the spin singlet , and the nel phase
( the other distinct multicritical point of the hartree - fock theory , which is apparent on the abscissa of fig . [ fig21 ] where @xmath396 and @xmath65 , where the ferromagnetic , the canted and the symmetric phase coexist is not accessible within the effective field theory because of its long wavelength approximation , and can not be seen in fig .
[ fig15]a as it is pushed to the point @xmath397 , @xmath398 where the two hartree - fock phase boundaries of fig .
[ fig15]a come together . )
it is , therefore , obvious that , except for very small values of @xmath145 ( where the effective field theory which applies only when @xmath259 ) , the quantum phase diagrams predicted by the two theories are topologically identical .
finally , we can actually estimate the @xmath5 double - layer kosterlitz - thouless transition temperature , @xmath168 of section [ nls]c , in the effective field theory by using the microscopic parameters obtained within the hartree - fock theory .
this calculation @xcite , where one incorporate the calculated hartree - fock parameters for @xmath292 and @xmath399 in eq .
( [ tcrho ] ) , leads @xcite to an estimated effective field theory @xmath400 which is somewhat larger than the critical temperature @xmath168 ( eq .
( [ equ : t1 ] ) ) estimated within the long wavelength mean field hartree - fock treatment of section [ kt ] . in general , we believe the @xmath336expansion leads substantial overestimates of transition temperatures because it does not properly account for the low - dimensional vortex effects responsible for the transition .
in this section , we discuss some recent double - layer @xmath5 inelastic light scattering experiments whose findings are consistent with our theoretical results .
a detailed quantitative comparison between our theory and the experiment requires an accurate knowledge of the temperature dependence of the related experimentally relevant response functions as the system undergoes a finite temperature phase transition at @xmath168 .
such a quantitative description is at present lacking , and therefore we restrict ourselves mostly to a qualitative discussion . in a recent inelastic light scattering experiment,@xcite the long wavelength @xmath401 mode of the intersubband sdw triplet ( see fig .
[ fig16 ] for schematic details of the various possible sdw modes in the system ) , which corresponds to transition @xmath402 , is measured for @xmath5 double - layer quantum hall systems .
the double - layer samples used in the experiment are by design in the canted antiferromagnetic phase according to our zero temperature hartree - fock phase diagram , _ i.e. , _ the ground state of the experimental system is the canted antiferromagnetic quantum hall state ( see figs .
[ fig3 ] , [ fig21 ] , and [ fig15 ] for the location of the experimental sample in our theoretical diagram ) .
the experiment @xcite shows two important and striking features : one is that there is a threshold temperature ( @xmath403 ) below which the @xmath401 mode becomes unobservable as it seems to lose all spectral weight , the other feature is that the excitation energy @xmath401 approaches the zeeman energy @xmath20 when the threshold temperature is approached from the above , _
i.e. _ @xmath404 .
we argue below that these experimental observations are completely consistent with our predicted kosterlitz - thouless transition in the canted antiferromagnetic phase being the observed experimental transition at @xmath168 .
first , we notice that the @xmath401 mode , which involves a no - spin - flip transition with @xmath405 , has a maximum spectral weight in the symmetric phase , where there are as many spin - up ( down ) empty states as there are spin - up ( down ) electrons .
the spectral weight of the @xmath401 mode is identically zero in the ferromagnetic phase , where all spin - up states are occupied and all empty states are spin - down , and hence the @xmath401 mode ( which does not involve any spin flip ) is forbidden
. the spectral weight of the @xmath401 mode should be nonzero but small in the antiferromagnetic phase .
this is because the canted antiferromagnetic phase lies between the symmetric phase and the ferromagnetic phase in the phase diagram and its spin - flip dynamics should thus be intermediate .
more over , the canted antiferromagnetic phase is not an eigenstate for either spin or pseudospin , so the small spectral weight of the @xmath401 mode is shared by many allowed transitions , spreading the mode intensity over these transitions and thus making the spectral weight of each transition even smaller .
it is thus plausible to regard the observed disappearance of the @xmath401 mode at the threshold temperature as the transition to the canted antiferromagnetic phase at lower temperatures ( where the spectral intensity for the @xmath401 mode becomes very small ) . above the transition temperature
the system is essentially a disordered planar @xmath170-@xmath171 magnet , and thus behaves like a paramagnet whose sdw properties should be very similar to the paramagnetic spin - singlet symmetric phase .
next , we notice that , in the symmetric phase , the excitation energies of the intersubband sdw triplet have the following simple relationship @xmath406 this expression can be derived explicitly , using either the diagrammatic time - dependent hartree - fock approximation or the single - mode approximation .
it is a direct consequence of the fact that coulomb interaction is spin independent .
the above relationship bears a clear physical meaning : @xmath407 means that @xmath408 , _
i.e. _ mode softening ( see fig .
[ fig16 ] ) .
thus , the experimental observation that @xmath401 approaches the zeeman energy as the threshold temperature is reached from above suggests that there is mode softening ( @xmath409 ) at the phase boundary , as predicted by the computations of the @xmath287-dependent peak positions in section [ nonzerot ] .
finally , we note that the critical temperature ( the threshold temperature ) in the experiment @xcite is @xmath410 , which is reasonably close to our calculated kosterlitz thouless critical temperature @xmath411 in the hartree - fock theory ( eq . ( [ equ : t1 ] ) ) using the actual experimental sample parameters . @xcite this discrepancy between the experiment and the hartree - fock theory is small when compared with the energy scale of coulomb interaction , which is about @xmath412 in this particular sample .
in addition , quantum fluctuations neglected in the hartree - fock theory should lower the calculated critical temperature and reduce this discrepancy . from the above discussions , we conclude that our theoretical predictions are consistent with the recent light scattering experimental results .
the most dramatic aspect of the experimental observations which give us confidence in believing that the experiment is really seeing the canted antiferromagnetic phase are ( i ) the unambiguous observation of a mode softening ( _ i.e. _ @xmath407 implying @xmath408 ) ; ( ii ) the observed temperature dependence indicating a finite temperature phase transition ; ( iii ) the location of the experimental sample in our calculated phase diagram and ( iv ) the @xmath407 collapse being observed precisely at @xmath5 . while the recent inelastic light scattering experiment @xcite provides , in our opinion , rather compelling evidence in favor of there being a finite temperature ( kosterlitz - thouless ) transition in the @xmath5 double - layer system with the low temperature phase being the canted antiferromagnetic phase ( by virtue of the vanishing of the @xmath413 mode at the phase boundary , as discussed in section [ sdw ] of this article ) , a complete verification of our theory awaits further more conclusive and direct experimental measurements , especially heat capacity measurements which should shown ( fig . [ fig34 ] ) power law temperature dependence in the canted phase due to the existence of the goldstone mode and exponential temperature dependence in the two normal phases due to the existence of gaps in the excitation spectra , would be particularly well - suited in verifying our phase diagram .
the direct observation of a gapless goldstone mode ( fig .
[ fig7 ] ) in the inelastic light scattering measurement in the ( low temperature ) canted phase would also be rather definitive in establishing the existence of the canted phase . in this context
we mention that the sdw softening indicating a phase transition to the canted phase is a long wavelength instability , and therefore optical spectroscopy @xcite may also be useful in studying our proposed @xmath5 double - layer phase diagram . both of these proposed direct experiments are fraught with considerable ( experimental ) difficulties , however .
electronic heat capacity measurements in quantum well structures are notoriously difficult by virtue of the extremely small magnitude of the ( 2d ) electronic heat capacity compared with the background ( lattice ) contribution . as for the direct experimental observation of the goldstone mode ,
the experimental inelastic light scattering spectroscopy is severely restricted by the selection rules inherent in the resonant light scattering spectroscopy , and at this stage it is unclear whether the problems associated with the selection rules would allow to directly observe the goldstone mode .
one striking difference between the physics of @xmath5 double - layer system and the corresponding @xmath137 situation is the existence of a charge gap in the @xmath5 case for all values of @xmath145 and @xmath19 : the system is always incompressible ( in all its quantum phases including the canted phase ) . thus the quantized hall effect exists throughout our phase diagram unlike in the corresponding @xmath137 situation .
@xcite the existence / nonexistence of the qh effect , which has been useful in mapping out the @xmath137 double - layer phase diagram @xcite would not work in our problem in a direct sense .
we do , however , speculate that the activation energy ( _ i.e. _ the effective value of the incompressible charge gap ) for the @xmath3 double - layer qh effect may very well show observable structure at our calculated phase boundaries even though all the phases ( ferromagnetic , canted , symmetric ) would exhibit @xmath3 qh effect . we suggest systematic experimental investigations of @xmath3 double - layer ( @xmath387 with @xmath414 ) qh activation energies by tuning @xmath415 , and @xmath145 to look for signatures of our proposed zero and finite temperature phase transitions . in this context
we point out that there is already some experimental evidence @xcite that the naive @xmath178 level crossing in @xmath5 double - layer qh systems does not exist ( as our theory proposes and clearly demonstrates in our calculated phase diagrams ) .
the experimental observation @xcite has been that the naive @xmath5 level crossing phenomenon ( at @xmath178 ) between ferromagnetic and symmetric phases , which would exist in the absence of our intervening canted phase , if it happens at all in double - layer systems , must happen at magnetic fields _ much lower _ than that satisfying @xmath178 condition .
this is , of course , exactly what our phase diagram ( see fig .
[ fig3 ] ) ) predicts nothing interesting happens for finite @xmath145 at @xmath178 or for that matter even for @xmath416 at @xmath417 in fig .
[ fig3]a for example the system remains in the fully spin polarized ferromagnetic phase and the naive expectation of a level crossing transition to the symmetric phase simply does not occur . in this sense ,
our phase diagram for the @xmath5 double - layer system may have already been verified in 1990 !
@xcite further experiments along this line at @xmath3 double - layer systems would be useful .
in summary , we have studied both zero and finite temperature properties of the @xmath5 double - layer qh systems within the framework of hartree - fock approximation .
we show that , in addition to the fully polarized state adiabatically connected to the well separated layer state , there are two other double - layer quantum hall phases : the first is a spin singlet , and the second is characterized by a finite interlayer inplane canted antiferromagnetic spin ordering .
the transition between the different quantum hall phases is continuous , and is signaled by the softening of collective intersubband spin density excitations . because of the broken @xmath74 symmetry in the canted antiferromagnetic phase , the system has a finite temperature kosterlitz - thouless transition ( @xmath418 ) . below the critical temperature ,
the canted antiferromagnetic phase supports a linear goldstone mode . above
, the system is essentially a paramagnet similar to the symmetric phase .
our findings are consistent with recent light scattering spectroscopic experimental results .
we present detailed results of our study , including the stability energetics of various phases , the antiferromagnetic order parameter in the canted phase , the phase diagram , the collective excitation dispersions , the specific heat , and the kosterlitz - thouless critical temperature , and suggest various experiments which could , in principle , probe the rich double - layer phase diagram predicted by our theory .
in addition to the microscopic @xmath5 hartree - fock theory , we have developed a rather general long wavelength effective field theory for the @xmath3 , where @xmath387 with @xmath164 an odd integer , double - layer system .
the essential inputs for this effective field theory are the existence of charge gaps in the two layers and an effective interlayer antiferromagnetic ( superexchange ) interaction . by mapping the effective action for this problem to that of an @xmath9 quantum nonlinear @xmath10 model
, we have been able to show that the qualitative phase diagram calculated within the hartree - fock theory for @xmath5 is actually generically valid ( topologically ) for any @xmath3 ( with @xmath419 ) double - layer system with the symmetric phase of the hartree - fock calculation being replaced by a highly non - trivial correlated spin singlet phase ( of which the @xmath5 symmetric phase is a rather trivial example ) .
thus , there could be rather non - trivial canted ( and perhaps even nel , if one can apply sufficient external pressure to produce vanishing gyromagnetic ratio ) antiferromagnets at , for example , @xmath420 in a double - layer system , where each single fully spin polarized laughlin state spontaneously develops in - plane antiferromagnetic spin ordering .
observation of the canted or the spin - singlet phase in a @xmath420 double - layer qh system would significantly enrich the many - body strong correlation physics associated with qh systems .
we conclude by pointing out that , although we have confined ourselves in this article to the @xmath384 case , with @xmath164 an odd integer , it is obvious that the physics we are considering here applies , in principle , to all double - layer qh systems with @xmath3 where a single layer at filling @xmath1 forms a fully _ spin polarized _
incompressible qh state with a charge gap .
thus , the same physics as at @xmath5 should apply , in principle , at @xmath6 ( but _ not _ at @xmath421 where the charge gap is the cyclotron gap not @xmath422 . ) in double - layer system . in principle , however , our approximations which neglect all ( orbital ) landau level couplings become progressively worse at higher landau levels . in this respect ,
it is very gratifying that the experimental light scattering measurements @xcite find qualitatively similar ( but quantitatively much suppressed ) behavior at @xmath6 as at @xmath5 , but the @xmath7 situation is qualitatively different .
the authors thank dr .
a. pinczuk and dr .
v. pellegrini for helpful discussions on the experimental data .
the work at university of maryland is supported by the u.s .- onr and the u.s .- aro .
the work at yale university is supported by nsf grant no dmr 96 - 23181 .
here we will assess the validity of the mapping from the action @xmath220 in ( [ action1 ] ) to @xmath229 in ( [ action2 ] ) by examining a simple toy model of two spins .
we consider the hamiltonian @xmath423 where @xmath424 are two quantum spins of spin @xmath425 .
the energy spectrum of @xmath426 is clearly @xmath427 where @xmath428 is an overall constant we shall not be interested in .
let us attempt to obtain this result using the coherent state path integral .
first , we transcribe @xmath232 into the effective action @xmath429 \label{app1}\ ] ] where @xmath200 .
notice that this is the analog of the action @xmath220 in ( [ action1 ] ) with only the spatial gradient spin stiffness terms now being absent . now insert the parameterization ( [ nl ] ) into ( [ app1 ] ) , and expand to quadratic order in @xmath218 .
the neglect of terms higher order in @xmath218 is the only approximation being made here .
this gives us the analog of ( [ action1a ] ) @xmath430 \label{app2}\ ] ] now we integrate out @xmath218 as described above ( [ action2 ] ) to obtain @xmath431 where recall that the functional integral is over the unit vector field @xmath432 satisfying @xmath433 for all @xmath434 .
this last form of @xmath435 is the effective action for a quantum rotor in a field @xmath436 .
this action is equivalent to the hamiltonian @xmath437 where @xmath438 is the rotor angular momentum operator .
the eigenvalues of @xmath439 are easily seen to be identical to those of @xmath232 in ( [ app0 ] ) with one simple difference .
the allowed values of @xmath440 now extend over _ all _ non - negative integers .
thus the only effect of dropping terms higher order in @xmath218 in the functional analysis is that the upper bound @xmath441 has disappeared .
this only introduces additional states at relatively high energies and is therefore not expected to be of importance in our study of the low energy properties of @xmath229 . , edited by r.e .
prange and s.m .
girvin ( springer - verlag , new york , 1987 ) ; _ perspectives in quantum hall effects _ , edited by s. das sarma and a. pinczuk ( wiley , new york , 1997 ) .
murphy j.p .
eisenstein , g.s .
boebinger , l.n .
pfeiffer , and k.w .
west , phys .
* 72 * , 728 ( 1994 ) k. yang , k. moon , l. zheng , a.h .
macdonald , s.m .
girvin , d. yoshioka , and shou - cheng zhang , phys .
. lett . * 72 * , 732 ( 1994 ) ; x.g .
wen and a. zee , phys .
lett . * 69 * , 1811 ( 1992 ) ; z.f .
ezawa and a. iwazaki , phys .
b * 47 * , 7295 ( 1993 )
. k. moon , h. mori , kun yang , s.m .
girvin , a.h .
macdonald , l. zheng , d. yoshioka , and shou - cheng zhang , phys .
b * 51 * , 5138 ( 1995 ) . y.w .
suen , l.w .
engel , m.b .
santos , m. shayegan , and d.c .
tsui , phys .
lett . * 68 * , 1379 ( 1992 ) ; j.p .
eisenstein , g.s .
boebinger , l.n .
pfeiffer , k.w .
west , and song he , _ ibid . _ * 68 * , 1383 ( 1992 ) ; d. yoshioka , a.h .
macdonald , and s.m .
girvin , phys .
b * 39 * , 1932 ( 1989 ) ; song he , s. das sarma , and x.c . xie , _ ibid .
_ * 47 * , 4394 ( 1993 ) .
macdonald , p.m. platzman , and g.s .
boebinger , phys .
lett . * 65 * , 775 ( 1990 ) ; l. brey , _ ibid .
_ * 65 * , 903 ( 1990 ) ; h.a .
fertig , phys .
b * 40 * , 1087 ( 1989 ) ; x.m . chen and j.j .
quinn , _ ibid . _ * 45 * , 11054 ( 1992 ) ; g.s .
boebinger , h.w .
jiang , l.n .
pfeiffer , and k.w .
west , phys .
lett . * 64 * , 1793 ( 1990 ) . for notational clarity ,
we use @xmath442 for the subband index , @xmath443 ( @xmath444 ) for the layer index , @xmath10 for the electron spin , @xmath12 for the intra - landau level index , and @xmath445 ( @xmath41 and @xmath446 ) for the four hartree - fock eigenstates .
j. tobochnik and g.v .
chester , phys .
b * 20 * , 3761 ( 1979 ) ; j. fernandez , m.f .
ferreira , and j. stankiewicz , _ ibid . _ * 34 * , 292 ( 1986 ) ; r. gupta , j. delapp , g.g .
batrouni , g.c .
fox , c.f .
baillie , and j. apostolakis , phys .
lett . * 61 * , 1996 ( 1988 ) .
s. sachdev , phys .
b * 55 * , 142 ( 1997 ) ; a factor of @xmath364 has been inadvertently omitted from the expressions for @xmath447 and @xmath448 in this paper .
this omission also infected eqn ( 6 ) in ref .
the present paper contains corrected expressions . | we present details of earlier studies ( zheng _ et al _ , phys .
rev . lett . * 78 * , 310 ( 1997 ) and das sarma _ et al _ , _ ibid _ * 79 * , 917 ( 1997 ) ) and additional new results on double - layer quantum hall systems at a total filling @xmath0 , where a single layer at filling @xmath1 forms a ferromagnetic , fully spin - polarized , gapped incompressible quantum hall state . for the case
@xmath2 , a detailed hartree - fock analysis is carried out on a realistic , microscopic hamiltonian .
apart from the state continuously connected to the ground state of two well separated layers , we find two double - layer quantum hall phases : one with a finite interlayer antiferromagnetic spin ordering in the plane orthogonal to the applied field ( the ` canted ' state ) , and the other a spin singlet .
the quantum transitions between the various quantum hall states are continuous , and are signaled by the softening of collective intersubband spin density excitations . for the case of general @xmath1 ,
closely related results are obtained by a semi - phenomenological continuum quantum field theory description of the low - lying spin excitations using a non - linear sigma model .
because of its broken symmetry , the canted phase supports a linearly dispersing goldstone mode and has a finite temperature kosterlitz - thouless transition .
we present results on the form of the phase diagram , the magnitude of the canted order parameter , the collective excitation dispersions , the specific heat , the form of the dynamic light scattering spectrum at finite temperature , and the kosterlitz - thouless critical temperature .
our findings are consistent with recent experimental results . |
vandermonde matrices are well known .
they have a number of interesting properties and play a role in ( lagrange ) interpolation problems , partial fraction expansions , and finding solutions to linear ordinary differential equations , to mention just a few applications .
usually , one takes these matrices square , @xmath0 say , in which case the @xmath1-th column is given by @xmath2 , where we write @xmath3 . if all the @xmath4 ( @xmath5 ) are different , the vandermonde matrix is non - singular , otherwise not .
the latter case obviously takes place when all @xmath4 are the same , @xmath6 say , in which case one could speak of a confluent vandermonde matrix .
non - singularity is obtained if one considers the matrix @xmath7 whose @xmath1-th column ( @xmath5 ) is given by the @xmath8-th derivative @xmath9 , or by @xmath18 . in the latter case , one has @xmath19 .
a slightly more general situation is obtained , when one considers the in general non - square @xmath20 matrix @xmath7 , with @xmath1-th column @xmath9 , @xmath21 . in this case one
has that @xmath7 has rank equal to @xmath22 and for @xmath23 , the kernel of @xmath7 is the @xmath24-dimensional subspace of @xmath25 consisting of the vectors whose first @xmath26 elements are equal to zero .
note that the building elements of the matrix @xmath27 are the column vectors @xmath28 and a number of its derivatives .
the observations above can be generalized in many directions . in the present paper we opt for one of them , in which we will consider generalizations of the confluent vandermonde matrix @xmath7 by considering matrices obtained by using as building blocks the matrices @xmath29 , with @xmath11 as above and @xmath12 , together with its derivatives @xmath30 with @xmath31 .
note that @xmath32 if @xmath33 .
a special case of what follows is obtained by considering the matrix @xmath34 in a recent paper @xcite , the kernel of the matrix @xmath35 ( or rather , a matrix obtained by a permutation of the columns of @xmath35 ) has been studied for the case @xmath36 and @xmath37 .
more precisely , the two aims of the cited paper were to find a right inverse and a parametrization of the kernel of the matrix @xmath35 .
the two cases @xmath38 or @xmath39 have been analyzed in detail .
for these two cases two algorithms have been proposed to construct the kernel and the rank and the dimension of the kernel have been computed .
this was of relevance for the characterization of a matrix polynomial equation having non - unique solutions .
the origin of this problem was of a statistical nature and lied in properties of the asymptotic fisher information matrix for estimating the parameters of an armax process and a related stein equation .
the results on the parametrization of the kernel obtained in @xcite turned out to be unnecessary complicated , partly due an apparently irrelevant distinction between different cases , and not transparent . in the present paper
we reconsider the problem by embedding it into a much more general approach to obtaining properties of matrix polynomials that can be viewed as generalizations of a confluent vandermonde matrix ( in a single variable ) as indicated above .
the solution , which is shown to exhibit a very simple and elegant structure , to the original problem of describing the kernel under consideration , now follows as a byproduct of the current analysis . these are special cases of the unified situations of corollary [ cor : rankaz ] and proposition [ prop : kerkz ] of the present paper . in this way we generalize in the present paper the results of @xcite .
another approach to find a basis for the kernel has been followed in @xcite , which is also subsumed by proposition [ prop : kerkz ] .
the results of @xcite are closer in spirit to those of the present paper than the results in @xcite .
both cited papers contain some examples of right inverses of @xmath35 , that are special cases of what will be obtained in the present paper .
moreover , we will generalize the results for @xmath35 to results for the matrix @xmath40 that is defined by @xmath41 whose blocks we denote @xmath42 , @xmath43 , @xmath44 , so @xmath45 .
we will also consider the related matrix @xmath46 whose blocks are @xmath47 .
a special case occurs for the choices of the parameters @xmath33 , @xmath48 and @xmath38
. then we write @xmath49 and @xmath50 , ordinary ( normalized ) @xmath0 confluent vandermonde matrices .
the results in the next sections ( almost ) reduce to trivialities for ordinary confluent vandermonde matrices , so the contribution of the present paper originates with allowing the parameters @xmath51 , @xmath52 , @xmath26 and @xmath53 to be arbitrary .
classical vandermonde matrices and confluent vandermonde matrices have often been studied in the literature , see @xcite for definitions .
more recent papers often focussed on finding formulas for their inverses and on efficient numerical procedures to compute them , a somewhat random choice of references are the papers @xcite .
nevertheless it seems that the generalization @xmath40 of the confluent vandermonde matrix is , to the best of our knowledge , unknown in the literature .
derivatives of a function @xmath54 ( often matrix valued ) are denoted by @xmath55 or by @xmath56 .
the variable @xmath6 is in principle @xmath57-valued .
if @xmath58 is an @xmath59 matrix , for notational convenience we adopt the convention to label its elements @xmath60 with @xmath61 and @xmath62 .
we will see shortly why this a convenient convention .
entries of a matrix are also indicated both by superindices , according to what is typographically most appropriate .
when dealing with matrix polynomials @xmath63 , in proving results we usually suppress the dependence on @xmath6 and simply write @xmath58 .
if @xmath64 is a block matrix , we sometimes use subindexes to indicate its constituting blocks , but more often denote its blocks by superindices , so we write @xmath65 or @xmath66 in the latter case . also single superindices are used and we will come across notations like @xmath67 .
these should not be confused with powers of @xmath58 .
the meaning of @xmath67 will always be clear from the context . throughout the paper @xmath68 and @xmath52
are fixed positive integers , although often certain relations among them are supposed ( e.g. @xmath69 ) .
given the integers @xmath68 and @xmath52 and the variable @xmath6 , we consider a number of matrix polynomials .
the first are @xmath70 and @xmath71 . with the above labelling convention , valid throughout the paper , we have for the elements of @xmath11 the expression @xmath72 for @xmath73 .
next to these we consider the matrix polynomials @xmath74 we also consider the @xmath75 matrix polynomial @xmath76 which is defined by its @xmath14-blocks @xmath77 for @xmath78 , @xmath44 .
the size of @xmath40 is equal to @xmath79 .
if @xmath48 , we retrieve @xmath35 .
note that @xmath40 has a block hankel structure , even for the case @xmath80 .
along with @xmath40 we consider the matrix @xmath46 that has a block structure with blocks @xmath81 for @xmath43 and @xmath82 .
note that @xmath46 also has dimensions @xmath83 and that @xmath46 reduces to @xmath84 for @xmath48 . unlike @xmath40
, @xmath46 does nt exhibit a block hankel structure .
the following obvious relation holds .
@xmath85 where @xmath86 is the diagonal matrix with entries @xmath87 , @xmath43 and @xmath88 likewise .
let us now introduce the matrices @xmath89 ( of size @xmath0 ) and @xmath90 ( of size @xmath91 ) by @xmath92 and @xmath93 along with the matrices @xmath89 and @xmath90 , we introduce the matrices @xmath94 with @xmath95 and @xmath96 given by @xmath97 with @xmath98 we have the obvious relations @xmath99 where @xmath100 and @xmath101 are diagonal matrices , similarly defined as @xmath86 .
one easily verifies that @xmath102 has elements @xmath103 and by a simple computation that @xmath104 .
similar properties hold for @xmath96 . later on
we will also come across the matrices @xmath105 and @xmath106 , which have the same structure as @xmath89 and @xmath90 and only differ in size ( unless @xmath107 and @xmath108 ) . finally we introduce the often used square shift matrices @xmath109 ( @xmath110 arbitrary ) , defined by its @xmath14-elements @xmath111 ( kronecker deltas ) . other matrices will be introduced along the way .
+ the structure of the remainder of the paper is as follows . after having fixed some notation and other conventions , in section [ section : ca ] we shall introduce an auxiliary matrix polynomial @xmath112 that is instrumental in deriving properties of @xmath40 and @xmath46 . among them are factorizations , which will be treated in section [ section : factorization ] .
these lead to establishing the rank of @xmath40 , which is shown to be independent of @xmath6 .
a major issue in the present paper is to find a simple and transparent parametrization of the kernels of the matrices @xmath40 and @xmath46 .
this is done first in section [ section : kerneln0 ] for the special case of @xmath84 .
the results of that section will be used in section [ section : kerneln ] to characterize the kernel of @xmath46 .
the results of section [ section : intermezzo ] form the cornerstone of finding right inverses of @xmath40 , especially those that are not dependent on @xmath6 , which is a main topic of the final section [ section : rightinverse ] .
this section is devoted to the matrix @xmath112 , to be defined below , that is instrumental in deriving properties of @xmath40 and @xmath46 .
in particular it is used for obtaining useful factorizations in section [ section : factorization ] .
the matrix @xmath112 , also of block hankel type and of size @xmath83 is defined by its blocks @xmath113 ( here @xmath114 is used as a super index ) of size @xmath115 , for @xmath116 , @xmath44 .
the matrices @xmath117 for @xmath118 are the @xmath115 quasi - symmetric matrix polynomials ( this matrix polynomial is only square for @xmath36 ) given by their elements ( @xmath119 ) @xmath120 note that @xmath110 in @xmath117 is used as a super index and in @xmath121 as a power .
the block hankel matrix @xmath112 of size @xmath75 is then defined by its blocks @xmath122 which are equal to @xmath123 , for @xmath43 and @xmath44 .
one easily verifies that ( with @xmath124 ) for @xmath125 it holds that @xmath126 in matrix notation this relation becomes @xmath127 let @xmath128
. then we can rewrite ( [ eq : az1 ] ) as @xmath129 or as @xmath130 by induction one easily proves that the recursion ( [ eq : a11 ] ) leads to the explicit expressions , involving powers of @xmath131 and @xmath132 , @xmath133 and @xmath134 likewise one shows that ( [ eq : a21 ] ) leads to @xmath135 on the other hand , the recursion ( [ eq : az1 ] ) has solution @xmath136 it also follows that @xmath137 look back at the @xmath117 as defined in ( [ eq : akz ] ) .
one computes for @xmath118 @xmath138 henceforth we shall write @xmath67 instead of @xmath139 . note that @xmath140 , which is in general not square , has only nonzero elements on one ` anti - diagonal ' , the one with @xmath14-elements having sum @xmath141 .
we have the following result .
[ prop : arec ] the matrices @xmath142 satisfy the recursion for @xmath118 @xmath143 and hence @xmath144
morever , we also have the polynomial expansion @xmath145 the recursion follows by taking @xmath146 in ( [ eq : az1 ] ) .
the expansion follows from proposition [ prop : arec ] and equation ( [ eq : sol ] ) .
two more matrices will be introduced next .
first , let @xmath147 be the @xmath75 block - matrix whose @xmath14-th block ( of size @xmath115 ) is given by @xmath148 for @xmath43 and @xmath44 .
second , we introduce the matrix @xmath149 of size @xmath150 with blocks @xmath151 for @xmath152 and zero else , @xmath153 . since @xmath149 is block lower diagonal with identity matrix as diagonal blocks , it follows that @xmath149 is invertible and that its inverse has @xmath14-block equal to @xmath154 .
likewise one defines the matrix @xmath155 of size @xmath156 .
[ thm : azfactor1 ] the factorization @xmath157 holds true . for @xmath146 and with @xmath158 and @xmath159 and @xmath160 ,
this becomes @xmath161 moreover , @xmath162 , when both matrices are square .
we compute the @xmath14-block on the right hand side of ( [ eq : lal1 ] ) . using the definitions of @xmath149 , @xmath155 and @xmath147 , we obtain ( see the explanation below ) @xmath163 in the third equality above we used that @xmath164 and @xmath132 commute , in the fourth equality we used ( [ eq : ak01 ] ) with an appropriate change of the indices , whereas the last equality similarly follows from ( [ eq : akl1 ] ) .
the relation between the determinants follows from @xmath165 .
let @xmath166 be the @xmath1-th standard column basis vector of @xmath167 , @xmath168 and let @xmath169 be the @xmath1-th standard column basis vector of @xmath170 , @xmath171 . for convenience of notation , we put @xmath172 for @xmath173 and @xmath174 for @xmath175 . with this convention
, we always have for example @xmath176 .
[ thm : ao1 ] it holds that the block @xmath177 , the rank of @xmath147 is equal to @xmath178 . if @xmath147 is a fat matrix , i.e. @xmath179 , then @xmath147 has full ( row)rank iff @xmath180 and @xmath181 . if @xmath147 is a tall matrix ,
so @xmath182 , then it has full ( column ) rank iff @xmath180 and @xmath37 . in the special case
that @xmath147 is square , so @xmath183 , we get that @xmath147 has full rank , and it is then invertible , iff @xmath184 and @xmath38 , in which case @xmath147 is a matrix of size @xmath185 . in this case
we have for the inverse @xmath186 and @xmath187 .
specializing even more to @xmath36 , we get @xmath188 . since @xmath189 and @xmath148
, it follows that @xmath190 .
the rank of @xmath147 is equal to the rank of @xmath191 , which is easy to compute .
we get for its @xmath14-th block @xmath192 , where @xmath193 .
it follows that @xmath194 .
furthermore we have @xmath195 for @xmath196 and @xmath197 iff @xmath198 .
we conclude that @xmath191 is block diagonal , where the diagonal @xmath199-blocks are equal to @xmath200 for @xmath198 and zero otherwise .
the number of nonzero diagonal blocks is equal to @xmath201 , hence the rank of @xmath191 is equal to @xmath178 .
assume that @xmath147 is fat and that @xmath180 and @xmath181 .
then the rank of @xmath147 equals @xmath202 , the number of rows of @xmath147 .
for the converse statement we assume that @xmath203 ( the case @xmath204 can be treated similarly ) .
in this case the rank becomes @xmath205 , which is strictly less than @xmath206 , the number of rows of @xmath147 , which then has rank deficiency .
the dual statements for a tall @xmath147 follow by symmetry .
assume next that the matrix @xmath147 is square .
it has full rank iff the two sets of conditions for the tall and fat case hold simultaneously , which yields the assertion on invertibility .
assume then that @xmath184 and @xmath38 . by the computations in the first part of the proof we see that the diagonal blocks of @xmath191 are all equal to @xmath207 .
since there are now @xmath53 of these blocks , we obtain that @xmath191 is the @xmath185 identity matrix . to compute the determinant for this case
, we observe that the columns of @xmath147 consist of all the basis vectors of @xmath208 , but in a permuted order .
therefore , the determinant is equal to plus or minus one . to establish the value of the sign , we compute the number of inversions of the permutation .
this turns out to be equal to @xmath209 , which results in a determinant equal to @xmath210 iff this number is even , and @xmath211 in the other case .
if @xmath212 , then @xmath213 . a way of computing the number of inversions is to write the order of the numbering of the column basis vectors in rectangular array
decomposing every number @xmath214 in a unique way as @xmath215 with @xmath216 and @xmath217 , we can identify every such @xmath218 with a pair @xmath219 .
an inversion @xmath220 occurs when @xmath221 , but in the permuted order @xmath218 is preceded by @xmath222 . for every @xmath218
the number of inversions @xmath220 is the number of elements in the rectangle strictly to the south - west of @xmath218 in the rectangular array .
so , if @xmath218 corresponds to @xmath219 then the number of inversions @xmath220 is equal to @xmath223 . summing these numbers for @xmath224 and @xmath225 yields the total number of inversions . in the case
were @xmath147 is square and invertible , it is the permutation matrix having the property @xmath226 , for any @xmath227 .
[ cor : rankaz ] the rank of the matrix @xmath112 is for all @xmath228 equal to @xmath178 . if @xmath183 , then @xmath112 is square and invertible and @xmath229 .
the assertion on the rank follows from theorems [ thm : azfactor1 ] and [ thm : ao1 ] upon noting that the matrices @xmath149 and @xmath155 are invertible . since @xmath165 , also the assertion about the determinant follows .
in the present section we obtain a factorization of the matrix polynomial @xmath40 , from which a factorization of the matrix @xmath46 follows as a simple corollary . in the next proposition
we use the kronecker symbol @xmath230 to denote tensor products .
[ prop : factorm ] the factorization @xmath231 holds true .
moreover , @xmath76 and @xmath232 have the same rank equal to @xmath233 .
computation of the product @xmath234 by using the definition of @xmath232 as block matrix , yields a matrix that consists of blocks @xmath235 using equation ( [ eq : a ] ) , we get that this expression reduces to @xmath236 since @xmath237 for @xmath173 and @xmath238 for @xmath239 .
recall that the blocks @xmath76 consist of the derivatives @xmath240 .
the product rule for differentiation then yields ( [ eq : m ] ) .
the statement concerning the rank immediately follows from theorem [ thm : ao1 ] , since @xmath89 and @xmath90 are invertible .
the factorization in proposition [ prop : factorm ] exhibits a nice form of symmetry
. therefore it would be nice if also the matrix @xmath64 could be factorized in some symmetric way .
there does nt seem to be an easy way to do this .
consider a truly symmetric case , for instance @xmath241 . in this case
@xmath64 is non - singular and we have @xmath242 a first reasonable factorization in this symmetric case would be of the form @xmath243 , which would imply that @xmath64 is positive definite .
one easily sees that this is not the case . as a next attempt
, one could try to use the singular value decomposition of @xmath64 , but the eigenvalues of @xmath64 are not particularly nice , so this look as a dead end too . instead of the matrices @xmath89 and @xmath90
, one can also use the matrices @xmath102 ( see ( [ eq : utilde ] ) ) and @xmath96 ( see ( [ eq : wtilde ] ) ) to get a factorization of @xmath40 .
one then has to replace the matrix @xmath64 with @xmath244 , whose blocks @xmath245 are equal to @xmath246 , where the matrices @xmath247 are specified by their elements @xmath248 one then gets @xmath249 which can be proved in the same way as ( [ eq : m ] ) , or by using this identity and the relation @xmath250 next we derive a factorization of the matrix @xmath46 .
let the matrix @xmath251 be defined by its @xmath252 blocks @xmath253 , for @xmath78 , @xmath44 , where the matrices @xmath254 have elements @xmath255 for @xmath256 , @xmath257 .
[ prop : factorn ] the factorization @xmath258 holds true .
the matrices @xmath64 and @xmath251 are related through @xmath259 further relations between @xmath64 , @xmath244 and @xmath251 are @xmath260 the proof of the factorization ( [ eq : mt ] ) is similar to the proof of ( [ eq : m ] ) .
relation ( [ eq : aahat ] ) follows by an elementary computation .
likewise one proves ( [ eq : aa1 ] ) and ( [ eq : aa2 ] )
. additionally we have the following factorizations of @xmath40 and @xmath46 .
@xmath261 first we show an auxiliary result .
let @xmath262 .
then @xmath263 indeed , by direct computation starting from the definition of @xmath149 , one finds that the @xmath14-block @xmath264 equals @xmath265 for @xmath266 .
since @xmath267 , the result follows .
we now proceed to prove the indentities .
use ( [ eq : lal1 ] ) , both for arbitrary @xmath6 and @xmath146 , and equation ( [ eq : pmuq ] ) to get ( [ eq : aza0 ] ) . to obtain ( [ eq : mzm0 ] ) , in the same vain we use ( [ eq : m ] ) twice
. then ( [ eq : mzaz ] ) follows upon combining ( [ eq : aza0 ] ) and ( [ eq : m ] ) .
finally , ( [ eq : nzaz ] ) follows from ( [ eq : mzaz ] ) and ( [ eq : mdnd ] ) and ( [ eq : nzn0 ] ) follows from ( [ eq : mzm0 ] ) and ( [ eq : mdnd ] ) .
in the situation where @xmath40 and @xmath46 are square , we are able to compute their determinants .
[ cor : dets ] let @xmath183 . then all relevant matrices are square and we have the following expressions for the determinants of @xmath40 and @xmath46 .
@xmath268 with @xmath269 given in corollary [ cor : rankaz ] .
hence the matrices @xmath40 and @xmath46 are unimodular .
the matrix @xmath270 has determinant one , which gives the first equation in view of ( [ eq : mt ] ) .
use ( [ eq : mm ] ) , ( [ eq : dad ] ) and its consequence @xmath271 to get the formula for @xmath272 .
then the last equation follows from ( [ eq : aahat ] ) . in the next sections we investigate and characterize the kernel of the matrix @xmath46 for different values of the parameters .
it turns out that characterizing the kernel of @xmath46 yields more elegant results than characterizing the kernel of @xmath40 . on the other hand
, a factorization of @xmath40 leads to more elegant expressions than factorizations of @xmath46 .
of course , in view of ( [ eq : mdnd ] ) , results for one of the two can easily be transformed into results for the other .
we will focus on the kernel of @xmath46 only , simple because the obtained results have a more elegant appearance . as it turns out ,
it is instrumental to consider the kernel of @xmath84 ( which is @xmath46 for @xmath48 ) first , since the results obtained for this case , serve as building blocks for the kernel of @xmath46 .
we will investigate the kernel of @xmath84 for different values of @xmath52 , @xmath26 and @xmath53 .
this is in different notation the problem alluded to in the introduction and earlier investigated in @xcite for a special case .
notice that @xmath84 is @xmath273-dimensional .
furthermore for @xmath37 all derivatives @xmath274 ( @xmath275 ) are nonzero and moreover linearly independent vectors , whereas @xmath276 for @xmath277 .
it follows that the set @xmath278 has dimension equal to @xmath279 .
it is then an easy exercise to directly see that the rank of @xmath84 is also equal to @xmath279 .
of course , this also follows from proposition [ prop : factorn ] . in order to characterize the kernel of @xmath84
, we have to discern three different cases .
these are @xmath37 , @xmath280 and @xmath281 , each case being treated in a separate section . in all cases , our aim is to find _ simple , transparant _ parametrizations of the kernels
. let @xmath282 be the @xmath283 matrix given by @xmath284 observe the trivial but crucial property , that the columns of @xmath282 span the @xmath285-dimensional null space of @xmath286 .
+ next we consider the matrix @xmath287 that has the following block structure .
it consists of matrices @xmath288 ( @xmath289 , @xmath290 ) where each @xmath288 has size @xmath283 and is given by @xmath291 whereas all the other blocks are equal to zero .
for @xmath33 the matrix @xmath287 is empty and by convention we say that the columns of @xmath287 in this case span the vector space @xmath292 .
note that @xmath287 has dimensions @xmath293 .
+ the matrix @xmath287 looks as follows , where we suppress the dependence on @xmath6 and omit zero blocks .
@xmath294 where @xmath295 is the @xmath52-dimensional unit matrix and @xmath296 the @xmath52-dimensional shift matrix , with elements @xmath297 .
[ prop : kerkz ] let @xmath37 .
the @xmath298 matrix @xmath287 has rank @xmath299 and is such that @xmath300 .
in other words , the columns of @xmath287 form a basis for the kernel of @xmath84 . pick the @xmath301-th block column of @xmath287 , @xmath302 say ( @xmath303 ) .
we compute @xmath304 . for @xmath305 ,
this reduces to @xmath306 , which is zero , since @xmath307 .
then , for @xmath308 we have @xmath309 .
hence , for @xmath308 we get ( using that @xmath310 for @xmath311 , @xmath312 hence @xmath313 , so all columns of @xmath287 belong to @xmath314 .
since we know that the rank of @xmath84 is equal to @xmath52 , we get that @xmath315 , which equals the number of columns of @xmath287 . since @xmath287 is upper triangular with the full rank matrices @xmath282 on the block - diagonal , it has full rank . therefore , the columns of @xmath287 exactly span the null space of @xmath84 .
a nice feature of the matrix @xmath287 in proposition [ prop : kerkz ] is that it is a matrix polynomial of degree @xmath316 , only whereas the matrix polynomial @xmath84 has degree @xmath317 .
next we consider what happens if @xmath319 .
note that this case is void if @xmath33 .
henceforth we assume that @xmath320 .
let the matrix @xmath321 have elements @xmath322 for @xmath323 and zero else .
with @xmath324 the @xmath53-dimensional shift matrix , we have the compact expression @xmath325 , since @xmath326 for @xmath327 .
we now give some auxiliary results .
[ lemma : g ] let @xmath328 be as above and let @xmath329 be the shifted @xmath91 identity matrix , @xmath297 .
it holds that @xmath330 and the matrices @xmath331 and @xmath332 commute for all @xmath333 .
one also has the following two equivalent properties @xmath334 moreover , @xmath335 to prove ( [ eq : g1 ] ) , we recall that @xmath336 for @xmath175 .
hence @xmath337 then ( [ eq : g2 ] ) simply follows by differentiation , which immediately yields ( [ eq : g3 ] ) , from which ( [ eq : g4 ] ) also follows . we next prove ( [ eq : w1 ] ) for @xmath338 . this is then equivalent to @xmath339 , which can be verified by elementary calculations .
we proceed by induction and assume that ( [ eq : w1 ] ) holds
. then we have @xmath340 this proves ( [ eq : w1 ] ) , which is easily seen to be equivalent to ( [ eq : w2 ] ) .
equation ( [ eq : wgk ] ) follows by induction , using ( [ eq : w1 ] ) .
in fact , above statements about @xmath328 also follow from the parallel ones concerning @xmath286 and the relation @xmath341 indeed , since the matrices @xmath328 and @xmath329 commute , one has for instance @xmath342 which together with similar relations leads to ( [ eq : g3 ] ) .
[ lemma : mug ] one has for @xmath308 @xmath343 for @xmath344 , this reduces to @xmath345 more generally , one has for @xmath346 @xmath347 recall that @xmath10 .
below we use ( [ eq : w2 ] ) to compute @xmath348 equation ( [ eq : mgq ] ) follows from ( [ eq : mg ] ) , because @xmath349 for @xmath344 .
finally , ( [ eq : mgpq ] ) follows by iteration of ( [ eq : mgq ] ) .
we consider the matrix polynomial @xmath350 which for the present case has the following structure .
for @xmath351 it has blocks @xmath352 of size @xmath283 such that @xmath353 and @xmath354 , @xmath282 as before . for @xmath355
the blocks are @xmath356 and @xmath357 , all of size @xmath91 . finally , we have that @xmath358 .
all other blocks are equal to zero .
notice that @xmath350 is of size @xmath359 .
one easily verifies that @xmath350 has full column rank .
+ the matrix @xmath350 looks as follows , where again we suppress the dependence on @xmath6 and omit zero blocks . @xmath360
a compact expression for @xmath361 is @xmath362 where @xmath363 is the first standard basis vector of @xmath364 and @xmath365 the last standard basis vector of @xmath167 .
[ prop : kerkz1 ] let @xmath366 .
the @xmath367 matrix @xmath350 is such that @xmath368 . in other words ,
the columns of @xmath350 form a basis for the kernel of @xmath84 .
next we proceed as in the proof of proposition [ prop : kerkz ] .
we pick the @xmath301-block column @xmath369 . if @xmath370 , then @xmath353 and @xmath354 , whereas all other @xmath352 are zero .
the computation of @xmath371 is then exactly as in the previous proof .
next we consider the case where @xmath372 , which is quite different .
again we pick the @xmath301-th block column of @xmath350 .
recall the definition of the @xmath352 for this case .
we get , @xmath373 in view of equation ( [ eq : mgq ] ) .
this shows that @xmath350 belongs to the kernel of @xmath84 . since @xmath84 has rank @xmath26 , the dimension of the kernel is equal to @xmath374 , which is equal to the rank of @xmath350 . hence the columns of @xmath350 span this kernel .
post - multiplying the matrix @xmath350 by @xmath375 we obtain @xmath376 an alternative matrix whose columns span @xmath377 . for this case ,
the kernel of @xmath84 is closely related to what we have obtained in the previous case .
[ prop : kerkz2 ] let @xmath378 .
consider the matrix @xmath379 of ( [ eq : kbar ] ) in the special case that @xmath380 , then @xmath379 has @xmath381 columns .
we have that @xmath382 , which is now @xmath374-dimensional , is the product of the space spanned by the columns of @xmath379 and @xmath383 . since the highest power of @xmath6 that appears in @xmath84 is @xmath317 , we have that the matrices @xmath384 are identically zero if @xmath385 , whereas the matrix @xmath386 is the same as the matrix @xmath84 for the case @xmath380 .
the assertion follows by application of proposition [ prop : kerkz1 ] .
we have seen that we had to distinguish three different cases to describe the kernel of @xmath84 .
the same distinction has to be made in the present section .
first we introduce some more notation . for the matrix @xmath287 of ( [ eq : defk ] )
we now write @xmath387 and for @xmath388 , we define @xmath389 in the same way as @xmath287 , but now with @xmath390 replaced with @xmath391 , a @xmath392 matrix having the same structure as the original @xmath282 of section [ section : kern0 ] , so @xmath393 , if @xmath394 and @xmath395 , if @xmath396 . in particular @xmath397 . formally , for @xmath398 , we have @xmath399 whereas @xmath391 is the empty matrix for @xmath400 .
by @xmath401 for @xmath402 , we denote the product @xmath403 of size @xmath404 , whereas we take @xmath401 the zero matrix for @xmath405 . for @xmath406 we put @xmath407 and @xmath408 . with this convention , we have @xmath409 , @xmath410 .
let @xmath411 and @xmath412 . for @xmath308 one
has @xmath413 since @xmath307 , also @xmath414 for all @xmath308 .
it then follows that @xmath415 , since all higher order derivatives of @xmath390 vanish . using @xmath416
, we arrive at ( [ eq : wf ] ) . similarly , one obtains ( [ eq : mff ] ) .
[ thm : kernel1 ] the kernel of @xmath46 is spanned by the columns of @xmath401 and its dimension is equal to @xmath417 . hence the rank of @xmath46 is equal to @xmath418 .
in particular the matrix @xmath419 has full rank iff @xmath405 .
the arguments used in the proof of proposition [ prop : kerkz ] can also be applied to this more general case .
let for @xmath118 @xmath420 proposition [ prop : kerkz ] yields @xmath421 , and therefore @xmath422 .
consider now @xmath125 .
as before , @xmath302 denotes the @xmath301-th block - column of @xmath423 .
then , from ( [ eq : mff ] ) it follows that @xmath424 for @xmath308 we get @xmath425 where we used ( [ eq : mff ] ) again .
it follows that for @xmath125 @xmath426 where @xmath427 which is a matrix of size @xmath428 . invoking proposition [ prop : kerkz ] again
, we obtain for @xmath429 @xmath430 and hence @xmath431 .
assume that @xmath432 .
to have @xmath433 , we need @xmath434 for @xmath435 .
this now follows by iteration of ( [ eq : mbark ] ) .
in fact , by induction , one can show @xmath436 where @xmath437 for @xmath438 one has @xmath439 .
since each of the matrices @xmath389 for @xmath440 has full rank , which is equal to @xmath441 , we get that @xmath401 has rank equal to @xmath442 .
all assertions for @xmath402 now follow . on the other hand , for @xmath405 ,
the matrix @xmath46 has rank equal to @xmath443 , and therefore has a zero kernel .
the matrix @xmath444 for @xmath445 , which is of size @xmath446 turns out to be upper block - triangular .
consider for this case the product @xmath447 .
then one has for @xmath448 the @xmath14-block @xmath449 which can easily be shown by induction .
next we extend the result of proposition [ prop : kerkz1 ] to obtain the kernel of the matrix @xmath46 for the present case .
the approach that we follow is the same as the one leading to theorem [ thm : kernel1 ] . to obtain our results , we need to introduce additional notation .
let @xmath451 denote the upper left block of @xmath328 having size @xmath452 for @xmath453 .
so @xmath454 we also need the matrices @xmath455 , and @xmath456 , for @xmath457 .
[ lemma : gjik ] it holds that @xmath458 for @xmath308 and for @xmath459 one has @xmath460 .
the first assertion follows from the decomposition @xmath461 where @xmath462 is the last column of @xmath463 . for the proof of the second assertion
we need the following property of the shift matrix @xmath464 ( @xmath110 according to the context ) : @xmath465 , where @xmath466 since any @xmath467 is of the form @xmath468 ( with @xmath329 of size @xmath469 , see lemma [ lemma : g ] ) , we have @xmath470
. then we compute @xmath471 [ lemma : ggg ] for @xmath472 , it holds that @xmath473 using the definition of @xmath474 , the equality @xmath475 immediately follows from ( [ eq : g3 ] ) .
the second equality ( [ eq : ggg2 ] ) is obviously true for @xmath476 .
we use induction .
let @xmath472 and assume that @xmath477 . then , using lemma [ lemma : gjik ] , @xmath478 . to prove ( [ eq : ggg3 ] ) , we use ( [ eq : ggg2 ] ) and lemma [ lemma : gjik ] to write @xmath479 [ lemma : mqpm ]
let @xmath480 and @xmath481 be as in section [ section : nlq ] .
it holds that @xmath482 moreover , for @xmath118 it holds that @xmath483 we need the following observation . for a row vector @xmath218 of appropriate length and a scalar @xmath222 , one has @xmath484 because @xmath485 we now show ( [ eq : wgkj1 ] ) .
it is obviously true for @xmath305 .
assume it holds for some @xmath486 .
we get , using ( [ eq : ggg3 ] ) and ( [ eq : gxy ] ) @xmath487 equation ( [ eq : wgkj2 ] ) follows by combining ( [ eq : wgkj1 ] ) and ( [ eq : w2 ] ) , whereas ( [ eq : mgkj ] ) is an immediate consequence of ( [ eq : wgkj1 ] ) . next we compute , using ( [ eq : mgkj ] ) and ( [ eq : mgpq ] ) , @xmath488 which yields ( [ eq : mqq-1 ] ) .
let for @xmath489 the matrix @xmath490 be given by @xmath491 here @xmath492 has @xmath26 diagonal entries @xmath493 and @xmath494 diagonal entries @xmath495 .
hence @xmath496 has dimensions @xmath497 .
a compact expression of @xmath496 is as follows .
let @xmath365 be the last standard basis vector of @xmath167 , @xmath363 the first basis vector of @xmath364 , and @xmath498 the shift matrix of size @xmath0 .
then , similar to ( [ eq : bark ] ) , @xmath499 note that the matrices @xmath493 are empty for @xmath500 , @xmath501 and that @xmath502 is empty for @xmath175 .
hence we define @xmath503 a matrix of size @xmath504 , whereas for @xmath175 we define @xmath505 a matrix of size @xmath506 .
+ in what follows , we need the matrices @xmath507 , where the matrices @xmath508 have been introduced in ( [ eq : kbarj ] ) , ( [ eq : kbarr-1 ] ) , ( [ eq : kbarr+ ] ) .
then @xmath509 is of size @xmath510 for @xmath511 and of size @xmath512 for @xmath513 .
note that @xmath514 is always of full column rank .
the next lemma extends equation ( [ eq : nkj ] ) , obtained for the case @xmath37 .
[ lemma : fundamental ] let @xmath515 . for @xmath516 one
has @xmath517 where @xmath518 and @xmath519 are explicitly given by @xmath520 for @xmath521 it holds that @xmath522 .
the case @xmath523 has been verified in the proof of proposition [ prop : kerkz1 ] .
let therefore @xmath125 . to prove that the assertion holds true for @xmath524
, we assume the right hand side of formula ( [ eq : nki ] ) to be valid for @xmath525 and proceed by induction . to that end
we multiply it by @xmath526 and verify the answer . as before ,
we denote the @xmath301-th block column of @xmath526 by @xmath527 , for @xmath44 .
we will discern the four cases @xmath305 , @xmath528 , @xmath529 and @xmath530 .
let @xmath305 .
then the product @xmath531 becomes @xmath532 .
the analogue of ( [ eq : mff ] ) , with @xmath533 and @xmath534 substituted for @xmath535 and @xmath390 , yields that this equals @xmath536 , as should be the case .
let @xmath537 .
one obtains @xmath538 the analogue of equation ( [ eq : mff ] ) yields @xmath539 .
hence the right hand side of ( [ eq : nki1 ] ) becomes @xmath540 invoking the analog of ( [ eq : mff ] ) again , we can rewrite this as @xmath541 , a typical block of @xmath542 , as required .
next we consider the more involved case @xmath529 . in this case
the block column @xmath543 has entry @xmath544 on the @xmath545st row ( see lemma [ lemma : ggg ] ) and @xmath546 on the @xmath26-th row .
hence we get @xmath547 using ( [ eq : mgkj ] ) we obtain @xmath548 . in view of ( [ eq : mgpq ] )
, it holds that @xmath549 .
hence we van rewrite the right hand side of ( [ eq : nkij ] ) as @xmath550 which is equal to @xmath551 the first block of @xmath552 , as was to be shown .
finally we treat the case @xmath553 .
the block columns @xmath527 have @xmath554 at the @xmath555st row and @xmath546 at the @xmath301th row .
hence , we obtain @xmath556 since @xmath557 , we apply ( [ eq : mgpq ] ) to get @xmath558 , and the right hand side of ( [ eq : nkij2 ] ) reduces to @xmath559 a typical block of @xmath552 , as desired .
this settles the proof of the validity of equation ( [ eq : nki ] ) .
[ thm : kernel2 ] it holds that @xmath522 , for @xmath521 . for @xmath445 , the matrix @xmath560 is of size @xmath561 and has full rank , equal to @xmath562 .
if @xmath405 , @xmath560 is of size @xmath563 and has full rank , equal to @xmath564 .
summarizing , the kernel of @xmath46 is @xmath565-dimensional and spanned by the columns of @xmath560 .
the rank of @xmath46 is equal to @xmath566 and therefore @xmath46 never has full column rank .
we show that @xmath522 for @xmath521 , for which it is clearly sufficient to show that @xmath567 .
starting point is equation ( [ eq : nki ] ) for @xmath568 .
we have @xmath569 we multiply this equation with the block columns @xmath570 and , as above , we discern the case @xmath305 , @xmath537 , @xmath529 and @xmath571 . for @xmath305
we get @xmath572 , whereas for @xmath537 one computes @xmath573 in view of an analogue of ( [ eq : mff ] ) . for @xmath529
, we obtain @xmath574 we can now use equation ( [ eq : mqq-1 ] ) and ( [ eq : mgq ] ) to get @xmath575 hence , the right hand side of ( [ eq : gkk ] ) is zero .
next we consider the case @xmath576 .
we then get , parallel to ( [ eq : nkij2 ] ) , @xmath577 which is zero , in view of equation ( [ eq : mqq-1 ] ) . to show that @xmath578 , one has to show that @xmath579 , for all @xmath580 , but this has implicitly been shown above .
the other statements in the theorem have already been addressed before .
the theorem is proved .
we follow the approach leading to proposition [ prop : kerkz2 ] .
we observe that the matrix @xmath46 for @xmath378 can be decomposed as @xmath581 where @xmath582 is the
@xmath46 matrix " for the case @xmath380 , since all derivatives of @xmath17 of order higher than @xmath317 vanish .
let @xmath583 be the @xmath560 matrix for the case @xmath380 .
put @xmath584 where @xmath546 is the identity matrix of order @xmath585 .
if @xmath432 , then @xmath583 is of size @xmath586 , and if @xmath405 , then it has size @xmath587 .
then @xmath588 has size @xmath589 for @xmath432 and has size @xmath590 for @xmath405 . in short , @xmath588 has dimensions @xmath591 .
[ thm : kernel3 ] let @xmath281 .
the kernel of the matrix @xmath46 is spanned by the columns of @xmath588 , has dimension @xmath592 if @xmath432 and dimension
@xmath593 if @xmath405 .
so @xmath594 .
similar to the proof of proposition [ prop : kerkz2 ] , using the results of theorem [ thm : kernel2 ] for the case @xmath380 .
the results of this section will be used in section [ section : rightinverse ] , where we want to find ( special ) right inverses of the matrix @xmath35 .
we focus on the matrix @xmath596 , the first block row of @xmath64 , the matrix defined in section [ section : ca ] .
one directly sees that the rank of @xmath597 is equal to @xmath22 , although it also follows from theorem [ thm : ao1 ] with @xmath48 .
hence @xmath597 has full rank iff @xmath598 .
we introduce the matrix @xmath599 consisting of the @xmath600 blocks @xmath601 as follows .
@xmath602 where each @xmath601 has elements @xmath603 for @xmath171 , @xmath604 .
[ lemma : a0b ] let @xmath181 .
then @xmath597 has full row rank and @xmath605 . in other words
, @xmath606 is a right inverse of @xmath597 .
we have to compute the @xmath14-elements of @xmath607 . using the definitions of the matrices @xmath67 and
@xmath601 that only have nonzero entries on corresponding anti - diagonals , we see that @xmath608 is a diagonal matrix . hence we only have to consider the @xmath199-entries of @xmath609 .
note that @xmath610 for @xmath277 .
one obtains @xmath611 to compute the latter summation , we write it as @xmath612 by definition of the @xmath613-function @xmath614 . using the well - known fact that this can be computed in terms of @xmath615-functions ( @xmath616 ) we obtain @xmath617 it follows that @xmath618 .
we need some additional properties .
[ lemma : bj ] it holds that @xmath619 , if @xmath277 or @xmath620 . for the case @xmath621 ( which requires @xmath622 ) only the last row of this matrix is nonzero .
in fact , this row is equal to @xmath623 , with the convention that @xmath166 denotes the @xmath1-th basis vector of @xmath167 ( @xmath168 ) .
we compute the @xmath14-element of @xmath624 .
for @xmath171 and @xmath604 it is equal to @xmath625 clearly , for @xmath626 and @xmath627 , the last expression in the display equals zero , as is the case for @xmath628 and @xmath629 .
we next consider the two remaining cases , the first being @xmath511 and @xmath630 .
since we only have to consider @xmath631 , we get @xmath632 , which has to be nonnegative , so @xmath633 .
but then the binomial coefficient is equal to zero .
the remaining case is @xmath628 .
then we only have to consider @xmath634 , the other values of @xmath301 again give zero . note
that this implies that @xmath327 is necessary to get a nonzero outcome , whereas we already know that also @xmath635 is necessary .
hence nonzero elements in the last row of @xmath636 can only occur if @xmath637 .
under this last condition we find @xmath638 .
hence the bottom row of @xmath636 equals @xmath639 , which is equal to @xmath640 , for @xmath641 .
[ remark : bjnot ] here is an example where @xmath636 as defined in lemma [ lemma : bj ] is not equal to zero .
take @xmath642 and @xmath643 .
then @xmath644 and @xmath645 .
[ prop : gconstant ] define @xmath646 by @xmath647 .
then @xmath648 is a constant mapping , @xmath649 , under the condition @xmath277 or @xmath650 .
first we prove the following auxiliary results .
one has @xmath651 equation ( [ eq : uprime ] ) follows from the definition of @xmath102 and the elementary identity @xmath652 .
equation ( [ eq : wprime ] ) can be proved similarly .
we now compute @xmath653 according to equations ( [ eq : uprime ] ) and ( [ eq : wprime ] ) . putting @xmath654 and @xmath655
, we see that @xmath648 satisfies the linear differential equation @xmath656 this equation has a unique solution and we claim that it is given by the constant function as asserted . to that end
we check @xmath657 by lemma [ lemma : bj ] , since @xmath277 or @xmath650 .
furthermore , we have @xmath658 , since @xmath659 .
equation ( [ eq : lde ] ) has as the general solution @xmath660 where the exponentials can be computed as finite sums , since @xmath661 and @xmath132 are nilpotent .
elementary computations yield for instance that the @xmath14-element of @xmath662 is given by @xmath663 for @xmath152 and zero otherwise .
hence we obtain @xmath664 , which is in agreement with the definition of @xmath665 .
an example of a solution that is not constant is obtained for @xmath33 and @xmath643 . for the case @xmath429 one
finds directly from the definition of @xmath665 that @xmath666 and @xmath667 in view of remark [ remark : bjnot ] .
this is in agreement with equation ( [ eq : hgh ] ) , whose right hand side is equal to @xmath668
we return to one of our original aims , finding a right inverse of the @xmath670 matrix @xmath671 .
recall from theorem [ thm : ao1 ] and proposition [ prop : factorm ] that @xmath35 has rank equal to @xmath279 .
hence the matrix is of full rank iff @xmath181 .
equations like @xmath672 will in general not have a solution @xmath673 for a given @xmath674 , if @xmath675 .
in fact , we are interested in solutions @xmath673 that are independent of @xmath6 .
it is easy to see that such solutions only exist if @xmath676 and then @xmath677 . the uninteresting case @xmath204 will therefore be ignored and the standing assumption in the remainder of this section is @xmath181 . under this assumption , there are two subcases to discern , @xmath678 and @xmath679 . [ prop : mki ] assume that @xmath678 and @xmath181 .
let @xmath207 be the @xmath26-dimensional unit matrix .
there exists a constant ( not depending on @xmath6 ) matrix @xmath680 such that @xmath681 for all @xmath6 .
the equation @xmath672 for @xmath682 then has the constant solution @xmath683 .
the constant matrix @xmath684 is unique iff @xmath38 . in all cases
one can take @xmath685 , with @xmath606 as in ( [ eq : b ] ) . in this proof
we simply write @xmath546 for @xmath207 .
suppose that we have found a constant matrix @xmath684 with the desired property @xmath686 by differentiation of ( [ eq : mk0 ] ) @xmath110 times , with @xmath687 , we obtain , recall the definition of @xmath76 with @xmath184 , that @xmath688 we note that now @xmath40 is of size @xmath689 and has rank equal to @xmath690 . hence @xmath40 has a right inverse , @xmath691 say , and a true inverse in the case that @xmath38 , see e.g. corollary [ cor : dets ] .
it follows that @xmath684 should be the first block - column of @xmath692 .
next we use the factorization ( [ eq : m ] ) and note that also @xmath64 has a right inverse , @xmath693 say .
then @xmath694 hence , we can choose @xmath695 which means that @xmath684 is the first block - column of @xmath691 , so @xmath696 where @xmath697 is a right inverse of the matrix @xmath698 , since @xmath698 is the first block - row of @xmath64 .
but , a right inverse of @xmath698 is in proposition [ lemma : a0b ] shown to be @xmath606 .
therefore , we can now explicitly pose our candidate for @xmath684 , @xmath699 where @xmath606 as defined in ( [ eq : b ] ) .
hence we have to show that + ( 1 ) the matrix @xmath684 in ( [ eq : kz ] ) , in fact does nt depend on @xmath6 , + ( 2 ) @xmath700 . using matrices introduced in section [ section : factorization ]
, we write @xmath701 decomposing @xmath684 as @xmath702 where each block @xmath703 ( @xmath704 ) is of size @xmath600 , we get @xmath705 hence we see that @xmath706 , which was in proposition [ prop : gconstant ] shown to be constant and equal to @xmath707 , if we have @xmath650 or @xmath277 .
obviously , this is true of @xmath708 , but for @xmath709 , we have @xmath710 by assumption .
this proves claim ( 1 ) .
since @xmath684 is constant in @xmath6 , we can take @xmath146 in ( [ eq : kz ] ) .
for the second one we have @xmath711 in view of lemma [ lemma : a0b ] . finally , if @xmath38 , then @xmath40 is invertible , which implies that @xmath684 is the unique constant matrix solving @xmath669 , since in this case equation ( [ eq : mk ] ) has a unique solution .
the special choice @xmath712 in the proof of proposition [ prop : mki ] is rather crucial in finding a right inverse of @xmath35 that does nt depend on @xmath6 .
we illustrate this with the following example .
our point of departure is equation ( [ eq : c ] ) with @xmath48 . recalling ( [ eq : aa0 ] ) , we can take @xmath713 , with @xmath714 any right inverse of @xmath147 .
we choose @xmath715 and compute @xmath716 because @xmath717 . since @xmath718 ( theorem [ thm : ao1 ] ) and @xmath719 , for the @xmath110-th block @xmath720 of @xmath721 we get @xmath722 ( @xmath723 . we conclude that @xmath724 . in order to see that this may result in a right inverse of @xmath35 that depends on @xmath6 , we choose @xmath725 .
the conditions of proposition [ prop : mki ] are then satisfied .
we have @xmath726 and it follows from the above that @xmath727 we close this remark by noting that there also other right inverses of @xmath35 , depending on @xmath6 , but still having a simple structure .
an example ( essentially taken from @xcite , where it was only given for @xmath728 in a slightly different situation ) is @xmath729 where @xmath730 is the first basis vector of @xmath170 .
this follows from the easy to verify identity @xmath731 the assertion of proposition [ prop : mki ] is not true if @xmath679 ( the second subcase ) .
indeed , in the proof of this proposition we used the fact that all @xmath703 are indeed constant matrices , under the condition @xmath678 .
if this is not the case , @xmath679 , the matrices @xmath703 for @xmath732 are not constant , in view of proposition [ prop : gconstant ] .
let us give an example to illustrate this .
consider the case @xmath733 and @xmath33 . then @xmath734 and the equation @xmath735 has the _ unique _ on @xmath6 depending solution @xmath736 .
+ we now treat the case @xmath737 in more detail . to that end
we need the following auxiliary result .
[ lemma : kerm0 m ] the subspace of the kernel of @xmath35 that consists of vectors that are constant in @xmath6 , i.e. the intersection @xmath738 , is @xmath739-dimensional .
this subspace is equal to the kernel of @xmath40 with @xmath184 , which is the same for all @xmath6 and hence can be parametrized free of @xmath6 .
the first observation is that a vector @xmath218 in @xmath740 that does nt depend on @xmath6 also satisfies @xmath741 for arbitrary @xmath51 , in particular for @xmath184 .
the case @xmath37 follows from theorem [ thm : kernel1 ] , since in this case the kernel of @xmath40 is the null space for all @xmath405 .
let then @xmath23 .
let @xmath218 be a column vector consisting of @xmath53-dimensional sub - vectors @xmath742 that do nt depend on @xmath6 .
recalling that @xmath743 consists of a row of blocks @xmath744 , we have @xmath745 the vectors @xmath746 are zero for @xmath277 and otherwise linear independent .
hence , to have the above sum equal to zero is equivalent to @xmath747 the equation for arbitrary @xmath748 can be differentiated to get @xmath749 which , subtracting from the equation for @xmath750 yields @xmath751 valid for @xmath752 . the only constant solutions to these equations are the zero vectors , so @xmath753 .
on the other hand , for @xmath754 we keep the equation @xmath755 now we relabel the unknowns by setting @xmath756 to get @xmath757 we differentiate this equation @xmath301 times , with @xmath406 to get @xmath758 where @xmath222 is obtained by stacking the @xmath759 . the first block - column in the above matrix
is @xmath760 , the second can be written as @xmath761 , up to the last one equal to @xmath762 .
hence the above system of equations can be compactly written as @xmath763 let @xmath764 be the diagonal matrix with elements @xmath765 .
a simple computation shows that @xmath766 , and therefore @xmath767 .
writing @xmath768 , we can rewrite the last equation in @xmath222 as @xmath769 since @xmath770 is invertible for any @xmath6 , this reduces to @xmath771 since the coefficient matrix has full row rank equal to @xmath53 , its kernel has dimension @xmath772 .
actually , this kernel is spanned by the columns of the @xmath773 matrix @xmath774 this proves the claim .
the result of lemma [ lemma : kerm0 m ] is also valid for @xmath775 .
this can be seen from equation ( [ eq : wy ] ) . indeed ,
if @xmath775 one has to extend the coefficient matrix with additional rows , all involving derivatives @xmath776 , with @xmath327 .
but these are all equal to zero .
one might think that the assertion of the lemma can alternatively be proven by explicitly computing the matrix @xmath560 for @xmath184 ( noting that @xmath777 represents the kernel of @xmath40 in view of ( [ eq : mdnd ] ) ) and showing that it is not depending on @xmath6 .
it turns out that this idea is false , as shown by the following simple example .
let @xmath778 , @xmath779 .
we compute @xmath560 and show that it is not free of @xmath6 . according to the results of section [ section:2ndcase ] we find for @xmath780 , @xmath781 and @xmath782 the following .
@xmath783 @xmath784 and the product @xmath785 yields the kernel of @xmath40 spanned by the columns of @xmath786 we see that the last column of @xmath787 has a term @xmath788 in the fourth row , hence this parametrization of @xmath789 is not the one we are looking for .
the reparametrization of @xmath789 given by @xmath790 with @xmath791 yields @xmath792 , with @xmath793 which is free of @xmath6 .
the procedure in the proof of lemma [ lemma : kerm0 m ] ( in this case @xmath794 , since @xmath778 ) yields a kernel of @xmath40 spanned by column vectors that are not depending on @xmath6 .
the result of that procedure is the matrix @xmath795 above . according to proposition [ prop : mki ] a constant solution @xmath684 exists .
any other constant right inverse @xmath799 is such that the @xmath26 columns of @xmath800 belongs to the kernel of @xmath743 and hence to the kernel of @xmath801 for @xmath184 . in view of lemma [ lemma : kerm0 m ] ,
a basis of this kernel can be obtained can be obtained by choosing @xmath593 linearly independent vectors . applying this result to each the columns of @xmath800
, we obtain the result . to illustrate the fact that the constant solution @xmath684 of proposition [ prop : mki2 ] is in general not unique , we consider the case @xmath778 , @xmath802 .
then @xmath803 and all constant solutions are given by @xmath804 with @xmath805 , which form an affine space of dimension @xmath806 .
+ of course all right inverses of @xmath84 , also those that depend on @xmath6 , are given by a much larger affine subspace .
assume @xmath181 and let @xmath807 be any right inverse of @xmath84 .
then any matrix @xmath808 , with @xmath809 a matrix whose columns belong to @xmath377 is a right inverse . since , @xmath810 , the affine subspace of these right inverses has dimension @xmath811 .
+ the natural extension of the equation @xmath797 is @xmath812 , with @xmath40 of order @xmath83 and @xmath813 the identity matrix of order @xmath202 . the matrix @xmath40 has rank equal to @xmath178 and therefore has full row rank if and only if @xmath181 and @xmath180 .
hence , under the latter condition , and only then , a right inverse exists , and the equation @xmath812 has a solution .
this equation can be decomposed as @xmath814 where every @xmath815 has size @xmath816 .
parallelling our previous aim , also here one could be interested in finding solutions @xmath684 that are constant in @xmath6 .
for @xmath817 we are in the previous situation , since a constant @xmath817 satisfying @xmath818 , also satisfies @xmath819 for all @xmath125 . the situation for
the the other @xmath820 is different .
consider for example @xmath821 .
it should satisfy @xmath822 and @xmath823 .
however , this is impossible for a @xmath821 that is constant in @xmath6 , since differentiating @xmath822 yields @xmath824 .
we conclude that the equation @xmath812 for @xmath825 has no constant solutions .
nonconstant solutions are for instance moore - penrose inverses .
these can be obtained by using the moore - penrose inverse of the matrix @xmath147 .
it follows from the proof of theorem [ thm : ao1 ] that for @xmath180 and @xmath181 , the matrix @xmath191 is the identity matrix .
hence @xmath826 is a right inverse of @xmath147 . using then theorem [ thm : azfactor1 ] and proposition [ prop : factorm ]
we obtain that @xmath827is a right inverse of @xmath40 .
the inverses @xmath828 and @xmath829 can be computed easily , since one has for instance @xmath830 and @xmath104 .
the inverses @xmath831 and @xmath832 can be computed in view of the formulas just above theorem [ thm : azfactor1 ] . since @xmath833 ,
one obtains that its @xmath14-block ( @xmath152 ) is given by @xmath834 .
summarizing , we have | vandermonde matrices are well known .
they have a number of interesting properties and play a role in ( lagrange ) interpolation problems , partial fraction expansions , and finding solutions to linear ordinary differential equations , to mention just a few applications .
usually , one takes these matrices square , @xmath0 say , in which case the @xmath1-th column is given by @xmath2 , where we write @xmath3 . if all the @xmath4 ( @xmath5 ) are different , the vandermonde matrix is non - singular , otherwise not .
the latter case obviously takes place when all @xmath4 are the same , @xmath6 say , in which case one could speak of a confluent vandermonde matrix .
non - singularity is obtained if one considers the matrix @xmath7 whose @xmath1-th column ( @xmath5 ) is given by the @xmath8-th derivative @xmath9 .
we will consider generalizations of the confluent vandermonde matrix @xmath7 by considering matrices obtained by using as building blocks the matrices @xmath10 , with @xmath11 as above and @xmath12 , together with its derivatives @xmath13 .
specifically , we will look at matrices whose @xmath14-th block is given by @xmath15 , where the indices @xmath16 by convention have initial value zero .
these in general non - square matrices exhibit a block - hankel structure .
we will answer a number of elementary questions for this matrix .
what is the rank ?
what is the null - space ? can the latter be parametrized in a simple way ? does it depend on @xmath6 ? what are left or right inverses ?
it turns out that answers can be obtained by factorizing the matrix into a product of other matrix polynomials having a simple structure .
the answers depend on the size of the matrix @xmath17 and the number of derivatives @xmath13 that is involved .
the results are obtained by mostly elementary methods , no specific knowledge of the theory of matrix polynomials is needed .
+ _ keywords : _ hankel matrix , confluent vandermonde matrix , matrix polynomial + _ ams classification : _
15a09 , 15a23 , 15a24 , 15b99 a block hankel generalized confluent vandermonde matrix + andr klein and peter spreij |
in @xcite the authors introduce the concepts of orientation for a @xmath0hypergraph , transitivity for an oriented @xmath0hypergraph , and define the class of _ comparability @xmath0hypergraphs _ as the class of non oriented @xmath0hypergraphs , which can be transitively oriented ( precise definitions are provided in section [ sec : pre ] ) .
these @xmath0hypergraphs are a natural generalization of ( simple ) comparability graphs ( graphs which can be transitively oriented or , equivalently , graphs associated to a partially ordered set ) .
comparability graphs are well known to be perfect graphs .
a graph is said to be _ perfect _ if all of its induced subgraphs have chromatic number equal to their clique number .
this concept was introduced by claude berge in 1961 @xcite .
in essence it means that a graph is perfect if for every of its induced subgraphs the chromatic number is as low as possible in terms of its clique number .
thus , it is natural to ask whether or not comparability @xmath0hypergraphs are perfect in this sense .
hypergraphs have been studied in relation to perfection in @xcite , and @xcite .
however , the precise concept of perfection for hypergraphs remains imprecise , to the best of our knowledge .
aiming to find a suitable definition of perfection in hypergraphs , we study the relationship between the chromatic number and the clique number of comparability @xmath0hypergraphs .
we define a _ @xmath0hypergraph _ @xmath1 as usual , @xmath2 where @xmath3 is the set of _ vertices _ of @xmath1 , and @xmath4@xmath5 is the set of _ edges_. the _ chromatic number _ , @xmath6 , is defined as the minimum @xmath7 , such that @xmath3 can be partitioned into @xmath7 parts , called _ color classes _ , in such a way that no edge of @xmath1 is monochromatic , in other words , no edge is contained in a single color class . the _ clique number _ , @xmath8 , of a @xmath0hypergraph @xmath1 is the largest cardinality of a subset of @xmath3 inducing a complete @xmath0hypergraph .
given that for any complete @xmath0hypergraph on @xmath9 vertices,@xmath10 , we have that @xmath11 , then for any @xmath0hypergraph the following equation holds : @xmath12 in this paper we study three important subclasses of comparability @xmath0hypergraphs which show three different behaviors in relation to ( 1 ) .
firstly , we exhibit a family of comparability @xmath0hypergraphs for which the difference , @xmath13 , is arbitrarily large .
secondly , we exhibit an interesting subclass of comparability @xmath0hypergraphs , named cyclic permutation @xmath0hypergraphs ( the analogues of permutation graphs ) , such that their chromatic number is bounded by a ( linear ) function of its clique number .
finally , we exhibit another interesting subclass of comparability @xmath0hypergraphs namely , the ones associated to a family of intervals in the circle . for these hypergraphs the chromatic number is as low as it can be in respect to their clique , that is , equality holds in ( 1 ) .
the paper is organized as follows : in section [ sec : pre ] we state the required definitions and preliminary results necessary to prove our main theorems , our main results are stated in section [ sec : thm ] and the proofs are located in the remaining sections .
let @xmath14 be any set of order @xmath9 .
a _ linear ordering _ of @xmath14 is a bijection @xmath15 . a _
cyclic ordering _ of @xmath14 is an equivalent class of the set of linear orderings with respect to the _ cyclic equivalence relation _ defined as : @xmath16 , if and only if there exists @xmath17 , such that @xmath18 for every @xmath19 where @xmath20 is taken modulo @xmath9 . for the remainder of this paper
we will denote each cyclic ordering , @xmath21 $ ] , in cyclic permutation notation , @xmath22 .
for example , there are two different cyclic orderings of @xmath23 , namely @xmath24 and @xmath25 , where @xmath26 and @xmath27 . given a @xmath0hypergraph @xmath1 , an _ orientation _ of @xmath1
is an assignment of exactly one of the two possible cyclic orderings to each of its edges .
an orientation of a @xmath0hypergraph is called an _ oriented @xmath0hypergraph _ , and we denote the oriented edges by @xmath28 . [ ex ] let @xmath2 be a @xmath0hypergraph with @xmath29 and @xmath30 , then a possible orientation of @xmath1 could be @xmath31 obtaining the oriented @xmath0hypergraph depicted in figure [ fig1 ] .
+ it is usual to associate transitive oriented graphs to partial ( linear ) orders .
similarly , we can associate to partial cyclic orders ( a ternary relation which is cyclic , asymmetric and transitive ) transitive oriented @xmath0hypergraphs in the following manner : an oriented @xmath0hypergraph @xmath1 is said to be _ transitive _ , if whenever @xmath32 and @xmath33 then @xmath34 ( this implies also @xmath35 ) .
now it is natural to define and study the following class of @xmath0-hypergraphs .
[ comphyp ] a non - oriented @xmath0hypergraph is called a _
comparability @xmath0hypergraph _ if it admits a transitive orientation .-
hypergraphs " however we believe that it should be simply called comparability @xmath0-hypergraphs according to the classical concept of comparability graphs . ]
the oriented @xmath0hypergraph defined in example [ ex ] is not transitive , however , its underlaying @xmath0hypergraph @xmath1 is a comparability @xmath0hypergraph since it can be transitively oriented ; take for instance @xmath36 .
in contrast , a @xmath0hypergraph with four vertices and three edges is not a comparability @xmath0hypergraph .
an oriented @xmath0hypergraph containing all possible @xmath0edges is called a _ @xmath0hypertournament_. as in the case of oriented graphs , there is a unique transitive oriented @xmath0hypertournament with @xmath9 vertices that we will denote by @xmath37 @xcite . in this paper
we study the chromatic number of two important subclasses of comparability @xmath0hypergraphs , namely the ones associated to a cyclic permutation , and the ones associated to a finite family of closed intervals in the circle .
we will now proceed to define both classes .
a _ cyclic permutation _ is a cyclic ordering of the set @xmath38 .
this is , an equivalence class @xmath21 $ ] of the set of bijections @xmath39 , in respect to the cyclic equivalence relation .
let @xmath21 $ ] be a cyclic permutation .
three elements @xmath40 , with @xmath41 , are said to be in _ clockwise order _ in respect to @xmath21 $ ] , if there is @xmath42 $ ] , such that @xmath43 .
otherwise the elements @xmath44 are said to be in _ counter - clockwise order _ with respect to @xmath21 $ ] .
the _ oriented @xmath0hypergraph @xmath45}$ ] associated to a cyclic permutation _ @xmath46 $ ] is the @xmath0hypergraph with vertex set @xmath47})=\{1,2, ...
,n\}$ ] where the edges are the triplets @xmath48 with @xmath49 , which are in clockwise order in respect to @xmath46 $ ] , and where the edge orientations are induced by @xmath46 $ ] .
it is not difficult to check that a @xmath0hypergraphs associated to a cyclic permutations is indeed transitive , and thus its underlying @xmath0hypergraph is a comparability @xmath0hypergraph .
consider the identity cyclic permutation @xmath50 and its reversed cyclic permutation @xmath51 , then their respective associated oriented @xmath0hypergraphs are the transitive @xmath0hypertournament @xmath37 and the null @xmath0hypergraph on @xmath9 vertices respectively . a non - oriented @xmath0hypergraph is called a _ cyclic permutation @xmath0hypergraph _
, if it can be oriented in such a way that the resulting oriented hypergraph is isomorphic to the one associated to a cyclic permutation .
emulating a classical result of pnueli , lempel and even @xcite , the authors of @xcite provided a characterization of the class of cyclic permutation @xmath0hypergraphs in terms of comparability @xmath0hypergraphs .
[ thm : gmmo ] @xmath1 is a cyclic permutation @xmath0hypergraph , if and only if @xmath1 and its complement @xmath52 are comparability @xmath0hypergraphs .
a simple graph @xmath53 is called an _ interval graph _
, if there exists a finite set of closed intervals in the real line , whose intersection graph is @xmath53 .
interval graphs , as well as its complements are well known to be perfect graphs . in a similar way , we will study @xmath0hypergraphs arising from finite sets of closed intervals in the circle @xmath54 .
@xmath0hypergraph @xmath55 associated to a finite family of closed intervals in the circle @xmath56 _ , is the @xmath0hypergraph with vertex set @xmath57 whose edges are the triplets of vertices with the property that their corresponding intervals are pairwise disjoint . for instance , if @xmath56 is a family of @xmath9 pairwise disjoint closed intervals in the circle , then @xmath55 is the complete @xmath0hypergraph on @xmath9 vertices .
a @xmath0hypergraph @xmath1 is called a _
circle interval @xmath0hypergraph _ , if there exists a set of closed intervals in the circle for which the associated @xmath0hypergraph is isomorphic to @xmath1 .
it is not difficult to prove that any circle interval @xmath0hypergraph @xmath1 is a comparability @xmath0hypergraph , since the natural orientation of @xmath1 , given by the clockwise order in which each of the intervals appears in @xmath54 , is transitive .
recall that comparability simple graphs are always perfect .
we will show three classes of comparability @xmath0hypergraphs exhibiting three different types of behaviors among the relationship between @xmath58 and @xmath59 , thus proving that comparability @xmath0hypergraphs are far from perfection .
first we will show that within the class of comparability @xmath0hypergraphs the difference @xmath60 is not bounded .
[ thm : main1 ] for each positive integers @xmath61 and @xmath7 , such that @xmath62 , there exists a comparability @xmath0hypergraph with clique number @xmath61 and chromatic number at least @xmath7 . in view of the above result it is an interesting problem to study families of comparability @xmath0hypergraphs for which the chromatic number of each member in the family is bounded by a function of its clique number .
[ thm : main2 ] let @xmath1 be a cyclic permutation @xmath0hypergraph , then @xmath63 .
furthermore , this bound is tight .
finally , we show an infinite family of comparability @xmath0hypergraphs for which the the chromatic number is as small as it can be in terms of the clique number .
[ thm : main3 ] let @xmath1 be a circle interval @xmath0-hypergraph , then @xmath64 .
in this section we will prove theorem [ thm : main1 ] by constructing a comparability @xmath0hypergraph with fixed clique number and arbitrarily large chromatic number .
we will use a construction given in @xcite and prove that the @xmath0hypergraphs obtained can be transitively oriented .
let @xmath1 be a @xmath0hypergraph of order @xmath9 , with vertex set @xmath3 and edge set @xmath65 .
we define @xmath66 as follows : * @xmath67 where @xmath68 and @xmath69 * @xmath70 where : + @xmath71 is the set of triplets in @xmath72 whose projection ( on to h ) is an edge of @xmath1 , that is , + @xmath73 , if and only if @xmath74 and @xmath75 .
+ @xmath76 is formed by the pairs of @xmath72 whose projection ( on to h ) is a single vertex , together with the vertex @xmath61 , that is , @xmath77 if and only if @xmath78 .
note that @xmath79 , @xmath80 , @xmath81 , and that the subhypergraphs induced by @xmath82 and @xmath83 are both copies of @xmath1 furthermore : @xmath84 see @xcite for a complete proof of this fact .
next we will show that , if @xmath1 is a comparability @xmath0hypergraph then @xmath85 is also a comparability @xmath0hypergraph .
let @xmath28 be a transitive orientation of @xmath1 .
we will define a natural transitive orientation @xmath86 using the following rules : let @xmath87 .
a. if @xmath88 , say @xmath89 , @xmath90 , and @xmath91 , then orient @xmath92 simply as @xmath93 . b. if @xmath94 , say @xmath95 , @xmath96 , @xmath97 where @xmath98 , and @xmath75 , then orient @xmath99 as induced by the orientation of @xmath100 , i.e. @xmath101 if and only if @xmath102 . @xmath86 is transitive .
_ proof _ [ proof of claim 1 ] we have to check that the definition of transitivity is satisfied by every pair of edges .
notice first , that if @xmath92 and @xmath103 are two edges in @xmath104 , and one of them is in @xmath76 then @xmath105 , so there is no conflict between the orientations of @xmath92 and @xmath103 .
suppose @xmath106 are such that @xmath107 then , by the previous observation , both @xmath92 and @xmath108 are in @xmath71 .
first suppose @xmath109 , @xmath110 , then the orientation satisfies the transitivity condition.secondly , suppose @xmath109 , @xmath111 , @xmath112 , @xmath113 , @xmath114 and @xmath115 where @xmath116 for @xmath117 .
by the orientation rules @xmath118 and @xmath119 . by the definition of transitivity of @xmath1 it follows that @xmath120 , implying that @xmath121 , thus the definition of transitivity is satisfied .
claim 1 proves that @xmath104 is a comparability @xmath0hypergraph .
+ * proof of theorem [ thm : main1 ] .
* start with @xmath1 a comparability @xmath0hypergraph with clique number @xmath61 and chromatic number @xmath122 ( take for instance the complete @xmath0hypergraph with @xmath61 vertices ) .
define @xmath123 where @xmath124 .
then , using the facts discussed in the previous paragraphs , for any @xmath7 , @xmath125 is a comparability @xmath0hypergraph with clique number @xmath61 and chromatic number at least @xmath126 .
in this section we prove the upper bound of theorem [ thm : main2 ] using an interesting class of @xmath0hypergraphs called even @xmath0hypergraphs .
a @xmath0hypergraph @xmath1 is said to be _ even _ , if every four vertex subset of @xmath3 induces an even number of edges .
the above definition arises naturally when studying the class of @xmath0hypergraphs associated to cyclic permutations .
in fact , the class of cyclic permutation @xmath0hypergraphs is an important subclass of the class of even @xmath0hypergraphs . to see this recall that by theorem [ thm : gmmo ] a cyclic permutation @xmath0hypergraph @xmath1 is such that both @xmath1 and its complement @xmath52 can be transitively oriented .
consequently , any subset of four vertices in @xmath3 induces @xmath127 , @xmath128 , @xmath129 or @xmath130 edges in @xmath1 as well as in @xmath52 .
therefore , [ rem : even ] any cyclic permutation @xmath0hypergraph is an even @xmath0hypergraph .
next we prove two lemmas which relate the chromatic number ( respectively , the clique number ) of an even @xmath0hypergraph @xmath1 with the chromatic number ( respectively , the clique number ) of a simple graph associated to @xmath1 .
let @xmath1 be a @xmath0hypergraph and @xmath131 ; we define the _ link _ of @xmath132 as the simple graph @xmath133 , where @xmath134 , and @xmath135 .
[ lem : link_chi ] let @xmath1 be an even @xmath0hypergraph , then @xmath136 for all @xmath131 .
_ proof _ for a fixed @xmath131 , let @xmath137 and @xmath138 be a proper coloring .
we extend this coloring to a coloring of @xmath1 by defining @xmath139 as @xmath140 for @xmath141 , and @xmath142 for @xmath143 . in order to prove that @xmath144 it remains to prove that @xmath145 is a proper coloring of @xmath1 , that is , there are no edges of @xmath1 whose vertices have the same color .
let @xmath146 be the partition induced by the coloring defined in the previous paragrph .
suppose , by contradiction , that @xmath147 is such that @xmath148 for some @xmath149 .
firstly assume that @xmath150 , then the pair @xmath151 is an edge of @xmath133 , which is a contradiction since the coloring restricted to @xmath152 is proper .
assume now that @xmath153 , and consider the set of vertices @xmath154 , which should induce an even number of edges of @xmath1 .
thus , at least one pair in @xmath155 is an edge of @xmath133 , which is a contradiction for the same reason .
hence , @xmath145 is a proper @xmath7coloring of @xmath1 and the statement follows true .
[ lem : link_omega ] let @xmath1 be an even @xmath0hypergraph then @xmath156 for all @xmath131 .
_ proof _ set @xmath157 , and let @xmath158 be a clique of @xmath133 , then the subhypergraph of @xmath1 induced by the vertex set @xmath159 is a clique of @xmath1 due to the eveness of @xmath1 .
now we are ready to prove the upper bound of theorem [ thm : main2 ] .
* proof of theorem [ thm : main2 ] ( upper bound ) * : let @xmath1 be a cyclic permutation @xmath0hypergraph . by observation [ rem : even ] , @xmath1 is an even @xmath0hypergraph .
hence , by lemma [ lem : link_chi ] and lemma [ lem : link_omega ] we obtain @xmath136 and @xmath160 respectively . to conclude the proof it remains to argue that the graph @xmath133 is a perfect graph , thus @xmath161 and the statement holds true . to see that @xmath133 is a perfect graph we will prove that it is a comparability graph ( a graph that can be transitively oriented )
we claim that a transitive orientation of @xmath1 naturally induces a transitive orientation of @xmath133 .
let @xmath28 be a transitive orientation of @xmath1 .
now consider the orientation of @xmath133 defined as follows : for @xmath162 let @xmath163 , if @xmath164 . by the definition of transitivity for @xmath1 this orientation of @xmath133 is transitive .
this is , if @xmath165 and @xmath166 are arcs of @xmath133 then @xmath167 and @xmath168 , therefore @xmath169 and so @xmath170 .
in this section we prove the tightness of theorem [ thm : main2 ] by defining a special family of cyclic permutation @xmath0hypergraphs which we call winding permutation @xmath0-hypergraphs . before proceeding we provide some terminology .
for @xmath171 we denote as @xmath172 $ ] the _ cyclic sub - permutation _ of @xmath21 $ ] induced by @xmath173 , that is , @xmath174 .
note that @xmath172 $ ] is a an equivalent class of the set of mappings ( not necesarly surjective ) @xmath175 in respect to the cyclic equivalence relation .
if a cyclic sub - permutation @xmath172 $ ] is such that there is @xmath176 $ ] with @xmath177 then we say that @xmath172 $ ] is a _ clockwise increasing _ cyclic sub - permutation of @xmath21 $ ] , if there is @xmath178 $ ] with @xmath179 we say that @xmath172 $ ] is a _ clockwise decreasing _ cyclic sub - permutation of @xmath21 $ ] . with the aid of this definition we will remark that : [ obs : omwgachi ] let @xmath46 $ ] be a cyclic permutation then : 1 .
@xmath180})$ ] is the length of the longest clockwise increasing cyclic sub - permutation of @xmath21 $ ] .
@xmath181})$ ] is the minimum number of clockwise decreasing cyclic sub - permutations of @xmath21 $ ] , which cover @xmath21 $ ] , that is , the minimum number k such that there is a partition @xmath182= s_1 \cup s_2\ , \ldots \cup s_k $ ] which satisfies that each @xmath183 $ ] is a clockwise decreasing cyclic sub - permutation .
for example , consider the cyclic permutation @xmath0hypergraph @xmath184}$ ] where @xmath21=(5\,2\,6\,3\,7\,4\,1)$ ] .
then , @xmath185})=4 $ ] since @xmath186 is a clockwise increasing cyclic sub - permutation of @xmath21 $ ] and there are no clockwise increasing cyclic sub - permutation of @xmath21 $ ] with length @xmath187 .
also @xmath188})=3 $ ] since the clockwise decreasing cyclic sub - permutations @xmath189 , @xmath190 and @xmath191 cover @xmath21 $ ] , and this can not be done with just two clockwise decreasing cyclic sub - permutations of @xmath21 $ ] . for positive integers @xmath192 and @xmath193
we define _ the winding permutation @xmath0hypergraph _ , @xmath194 , as the @xmath0hypergraph with @xmath195 vertices associated to the cyclic permutation @xmath196 $ ] defined as follows
. for @xmath19 let : @xmath197 for example , let @xmath198 and @xmath199 then we have @xmath200 vertices , and @xmath201 is the @xmath0hypergraph associated to the permutation @xmath202 .
given @xmath192 and @xmath193 we distinguish two sets of useful cyclic sub - permutations of @xmath196 $ ] , namely those induced by the sets @xmath203 where @xmath204 , and those induced by the sets @xmath205 where @xmath206 .
for instance , in the previous example we have : @xmath207=(10\,\ , 6\,\ , 2)$ ] , @xmath208=(11\,\ , 7\,\ , 3)$ ] , @xmath209=(12\,\ , 8\,\ , 4)$ ] , @xmath210=(13\,\ , 9\,\ , 5)$ ] @xmath211=(10\,\ , 11\,\,12 \,\,13 \,\,1)$ ] , @xmath212=(6\,\ , 7\,\,8 \,\,9)$ ] , @xmath213=(2\,\ , 3\,\,4 \,\,5)$ ] by definition , it follows that : [ obs : ab ] every cycle sub - permutation of the form @xmath214 $ ] is a clockwise decreasing cycle sub - permutation , while every cycle sub - permutation of the form @xmath215 $ ] is a clockwise increasing cycle sub - permutation . in the next two lemmas we compute the clique and chromatic numbers of @xmath216 .
[ lem : clique ] for any integers @xmath192 and @xmath193 , the clique number @xmath217 .
_ proof _ note that @xmath218 is an increasing clockwise cyclic sub - permutation of @xmath219 $ ] thus , by observation [ obs : omwgachi ] , @xmath220 . in order to prove that @xmath221 , let @xmath222 be a maximal clockwise increasing cyclic sub - permutation of @xmath219 $ ] .
first note that , by observation [ obs : ab ] , it follows that @xmath223 for all @xmath224 .
moreover , if @xmath225 for some @xmath226 then @xmath227 for all @xmath228 and @xmath229 . on the other hand , if @xmath230 for some @xmath231 , then @xmath227 for every @xmath232 . in both cases it
we must have @xmath233 completing the proof .
[ lem : chi ] for any integers @xmath192 and @xmath193 , the chromatic number @xmath234 .
_ proof _ in this proof we will denote @xmath21=[\phi_{q , r}]$ ] , for convenience . in order to show that @xmath235 we will exhibit a set of @xmath236 clockwise decreasing cyclic sub - permutations which cover @xmath21 $ ] .
take the cyclic sub - permutation @xmath237 $ ] induced by the fist @xmath238 elements : @xmath239 , then take the cyclic sub - permutation @xmath240 $ ] induced by the next @xmath238 elements : @xmath241 , and continue until the last cycle sub - permutation @xmath242 $ ] where @xmath243 .
note that @xmath244 for @xmath245 and @xmath246 where @xmath247 .
for instance , in the previous example with @xmath198 and @xmath199 we have @xmath248 , @xmath249 , and the obtained partition is : @xmath237=(10\,\ , 6\,\ , 2\,\ , 11)$ ] , @xmath240=(7\,\ , 3\,\ , 12\,\ , 8)$ ] , @xmath250=(4\,\ , 13\,\ , 9\,\ , 5)$ ] and @xmath251=(1)$ ] we claim that for all @xmath252 the cycle sub - permutation @xmath253 $ ] is clockwise decreasing .
let @xmath254 be fixed .
we first note that @xmath255 , intersects at most two consecutive sets of @xmath256 . therefore , by observation [ obs : ab ] , @xmath253 $ ] is made up of two clockwise decreasing subsequences .
hence , it remains to show that @xmath257 . to see this , note that both @xmath258 and @xmath259 are congruent to @xmath260 modulo @xmath193 .
thus , by using the definition of @xmath261 , we obtain @xmath262 , and @xmath263 , being @xmath264 as desired .
recall that we set @xmath243 . in order to prove that @xmath265 we will proceed by contradiction . suppose that @xmath266 is a partition of \{1,2, ...
,n } inducing a set of clockwise decreasing cyclic sub - permutations @xmath267 $ ] , @xmath268 $ ] , ... , @xmath269 $ ] , which covers @xmath21 $ ] .
then , by the pigeonhole principle , for some @xmath270 it must be true that @xmath271 . since , by observation [ obs : ab ] , it follows that @xmath272 for all @xmath273 , and moreover , if @xmath274 for some @xmath275 then @xmath276 for all @xmath277 , which is a contradiction .
this completes the proof .
* proof of theorem [ thm : main2 ] ( tightness ) * : for any given @xmath278 we will exhibit a cyclic permutation @xmath0hypergraph @xmath1 , with @xmath279 and @xmath280 .
take @xmath281 with @xmath282 . by lemma [ lem : clique ] and lemma [ lem : chi ] we know that @xmath279 and @xmath283 where @xmath9 is the order of @xmath1 and satisfies @xmath195 .
consequently , @xmath284 , which can be written as @xmath285 .
it remains to observe that @xmath286 by hypothesis , then @xmath280 as claimed .
recall that a circle interval @xmath0-hypergraph @xmath55 associated to a finite family @xmath287 of closed intervals in the circle , is a @xmath0hypergraph with vertex set @xmath57 and whose edges are the triplets of vertices whose corresponding intervals are pairwise disjoint . in this section
we will prove theorem [ thm : main3 ] , which claims that @xmath288 .
note that a sub - family of intervals @xmath289 corresponds to a clique of @xmath55 if and only if @xmath290 is a pairwise disjoint subset of intervals of @xmath56 .
thus , @xmath291 is the cardinality of the largest pairwise disjoint subset of intervals in @xmath56 .
* proof of theorem [ thm : main3 ] * : as we already know @xmath292 then it suffices to show that @xmath293 .
we proceed by induction on the size of @xmath291 .
let @xmath294 then the @xmath0-hypegraph @xmath55 has an empty set of edges , thus @xmath295 . for the remainder of the induction argument we will deal with the odd case and the even case separately .
assume that @xmath296 .
let @xmath297 be a pairwise disjoint set of intervals .
without loss of generality , we might assume that there is no other interval of @xmath287 contained in @xmath298 for every @xmath299 .
denote by @xmath300 the set of intervals of @xmath287 that intersect @xmath301 .
note that , as no interval of @xmath287 is contained in @xmath301 , necessarily out of every three intervals in @xmath300 two of them intersect .
thus , @xmath300 corresponds to a independent set of @xmath55 .
next observe that , @xmath302 ( consisting of all intervals in @xmath54 that do not intersect @xmath301 ) , may be regarded as a set of intervals in the real line , assume with out lost of generality that they are ordered form left to right as @xmath303 .
as @xmath296 , the maximum number of pairwise disjoint intervals of @xmath302 is @xmath304 . consider @xmath305 the left hand side of each of this intervals .
for @xmath306 , let @xmath307 be the subset of intervals of @xmath308 that intersects @xmath309 for each @xmath310 .
clearly , each @xmath307 is an independent set because it satisfies that out of every three intervals in @xmath307 , two intersect .
consequently , @xmath55 contains @xmath311 independent sets and thus , @xmath312 .
now assume that @xmath313 .
let @xmath314 be a pairwise disjoint set of intervals .
without loss of generality , suppose that the distance between @xmath315 and @xmath316 is the smallest possible between all choices of pairwise disjoint subsets of @xmath287 .
this implies two facts ; firstly , that the collection , @xmath300 of intervals of @xmath287 which intersect @xmath317 satisfies that out of every three intervals in @xmath300 , two intersect , thus @xmath300 is an independent set of the hypergraph @xmath55 ; secondly that the collection of intervals @xmath302 may be regarded as a set of intervals in the real line .
here we also have that the maximum number of pairwise disjoint intervals of @xmath302 is @xmath318 , and as before assume that they are ordered from left to right as @xmath319 and each one of them do not contain completely any other interval , consider @xmath320 the left hand side of each one of this intervals .
let @xmath307 be the subset of intervals of @xmath302 that intersects @xmath309 for each @xmath321 .
clearly , each @xmath307 is independent because it satisfies that out of every three intervals in @xmath307 , two intersect .
consequently @xmath322 .
this , concludes the proof of the theorem .
clearly any sub - hypergraph of a circle interval @xmath0hypergraph is a circle interval @xmath0hypergraph , thus theorem [ thm : main3 ] implies that any circle interval @xmath0hypergraph , as well as all its induced sub - hypergraphs satisfy @xmath323 which , in the classic context of perfection would mean that this hypergraphs are perfect " .
initially we were aiming to find perfection " on comparability @xmath0hypergraphs as a natural generalization of comparability graphs , fact that turns out to be false do to theorem [ thm : main1 ] and [ thm : main2 ] , however it remains an interesting question what does perfection in hypergraphs means .
natalia garcia - colin , amanada montejano , luis montejano and deborah oliveros .
_ transitive oriented 3 hypergraphs of cyclic orders _ order , a journal on the theory of ordered sets and its applications , vol.30 , ( 13 ) , 869875 , ( 2013 ) .
johana luviano , amanda montejano , luis montejano and deborah oliveros .
_ mycielski type constructions for hypergraphs associated with fractional colorings _ bolletn de la sociedad matemtica mexicana . to appear ( 2013 ) | beginning with the concepts of orientation for a @xmath0hypergraph and transitivity for an oriented @xmath0hypergraph , it is natural to study the class of comparability @xmath0hypergraphs ( those that can be transitively oriented ) . in this work
we show three different behaviors in respect to the relationship between the chromatic number and the clique number of a comparability @xmath0hypergraph , this is in contrast with the fact that a comparability simple graph is a perfect graph .
perfection in @xmath0hypergraphs , transitivity in @xmath0hypergraphs , comparability @xmath0hypergraphs . |
many natural phenomena have shown self - similarity or fractal properties in diverse areas .
it is well known that time series in neuroimaging such as electroencephalogram , magnetoencephalogram , and functional mri have fractal properties @xcite .
most of fractal time series can be well modeled as long memory processes .
a long memory process generally has scale - invariant properties in low frequencies ; for example , the scale - dependent wavelet variance has asymptotically power law relationship in low - frequency scales , which implies its fractal property @xcite .
such a property also has been demonstrated by achard _
_ @xcite by exploiting the taylor expansion of wavelet variances ; regrettably , they did not provide the detailed proof . here , we show the detailed procedure of achard _
a long memory process can be defined as follows as suggested by moulines _
_ @xcite ; that is , a real - valued discrete process @xmath0 is said to have long memory if it is covariance stationary and have the spectral density @xmath1 for @xmath2 where @xmath3 is a non - negative symmetric function and is bounded on @xmath4 .
[ theorem : wv ] let @xmath5 is the wavelet coefficient of the time series @xmath6 at the @xmath7-th scale and the @xmath8-th location , and let @xmath9 be the squared gain function of the wavelet filter such that @xmath10 then , @xmath11 satisfies @xmath12 achard _ et al . _ @xcite made its taylor expansion to estimate the long memory parameter of @xmath13 as shown in the corollary [ corollary : awv ] .
we give the detailed proof of the corollary [ corollary : awv ] .
[ corollary : awv ] let @xmath14 .
then , @xmath15 where @xmath16 @xmath17 it is well known we have the following taylor series @xmath18 from and , @xmath19 let us define four integrals : @xmath20 @xmath21 @xmath22 @xmath23 since if @xmath24 for @xmath25 then @xmath26 , @xmath27 @xmath28 from and the integrals - , rcl[eq : s(f)-int - sol ] ( j)&=&22^j+1_2/2^j+1 ^ 2/ 2^j ( s(f ) ) df + & = & 2^j+2(1-d ) + & = & 2^j+2(1-d ) + note that @xmath29 and @xmath30 thus , rcl[eq : s(f)-int - sol2 ] ( j)&=&2^j+2(1-d ) + & = & 2^j+2(1-d ) + where rcl[eq : s(f)-int - sol3 ] a_2&:= & ( + d/12 ) m_12 + & = & ( 2)^2 ( + ) . + then , we can have rcl[eq : s(f)-int - sol4 ] ( j)&=&2 ^ 2dj + & = & 2 ^ 2dj + where @xmath31 in the corollary , as the scale @xmath32 increases , the wavelet variance @xmath33 converges ; in other words , @xmath34 it implies that we can simply estimate the long memory parameter @xmath35 by linear regression method if we have the estimates of wavelet variances .
we verified achard _ et al . _
s proof on asymptotic properties of wavelet variances of a long memory process .
unfortunately , achard _
s taylor expansion of wavelet variance is not perfectly consistent with our result ; while @xmath36 in our result , they computed as @xmath37 . on the other hand ,
the assumption on short memory such that @xmath38 loses generality of their proof since there exist more general classes of short memory ; indeed , moulines _
_ proved that the asymptotics of wavelet variances hold when the short memory @xmath3 belongs to the function set @xmath39 defined as the set of even non - negative functions @xmath40 on @xmath41 $ ] such that @xmath42 nevertheless , their method is relatively simpler than other mathematical proofs .
moreover , we will be able to apply their method to investigate the asymptotic properties of multivariate long memory processes as achard _
et al . _ already attempted for bivariate long memory processes @xcite .
e. moulines , f. roueff , m. taqqu , _ on the spectral density of the wavelet coefficients of long - memory time series with application to the log - regression estimation of the memory parameter _ ,
journal of time series analysis . 28 ( 2007 ) 155187 . | a long memory process has self - similarity or scale - invariant properties in low frequencies .
we prove that the log of the scale - dependent wavelet variance for a long memory process is asymptotically proportional to scales by using the taylor expansion of wavelet variances .
memory process , taylor expansion |
ct cha introduced in the 65th name - list of variable stars by ( * ? ? ?
* kholopov et al . ( 1981 ) ) , was originally found by ( * ? ? ?
* henize & mendoza ( 1973 ) ) , often called hm9 , an emission - line star in chamaeleon exhibiting variations in its h@xmath3line from plate to plate and partial veiling ( * ? ? ?
* ( rydgren 1980 ) ) .
while the star was first classified as t tauri star by ( * ? ? ?
* whittet et al .
( 1987 ) ) it was later on found to be a classical t tauri star by ( * ? ? ?
* weintraub ( 1990 ) ) and ( * ? ? ?
* gauvin & strom ( 1992 ) ) from iras data , before ( * ? ? ?
* natta et al . ( 2000 ) ) could find evidence for a silicate feature disk with @xmath4 and @xmath5 using iso data .
the variations of the h@xmath3line could later be interpreted when ( * ? ? ?
* hartmann et al . ( 1998 ) ) measured a mass accretion rate of logm@xmath6 .
additional variations of infrared ( * ? ? ?
* ( glass 1979 ) ) and optical photometry can possibly be explained by surface features on ct cha and a possible rotation period of 9.86 days found by ( * ? ? ? * batalha et al .
( 1998 ) ) .
all additional properties of the k7 ( * ? ? ?
* ( gregorio hetem et al . 1988 ) ) star ct cha , such as e.g. its age estimated to be 3 myr by ( * ? ? ?
* feigelson et al . ( 1993 ) ) or 0.9 myr by ( * ? ? ?
* natta et al . ( 2000 ) ) as well as its equivalent width of the lithium absorption line of w@xmath7(li)= [email protected] ( * ? ? ?
* ( guenther et al . 2007 ) ) and its radial velocity of [email protected] km / s ( * ? ? ?
* ( joergens 2006 ) ) and proper motion ( table[tab1 ] ) are consistent with a very young member of the cha i star - forming cloud with an age of 2@xmath82myr and a medium radial velocity of [email protected] km / s ( * ? ? ?
* ( melo et al .
we observed ct cha in two epochs in february 2006 and in march 2007 with the european southern observatory ( eso ) very large telescope ( vlt ) at cerro paranal in the programmes with ids 076.c-0292(a ) & 078.c-0535(a ) .
all observations were done using the adaptive optics ( ao ) instrument naos - conica ( naco , ( * ? ? ? * lenzen et al .
2003 ) & ( * ? ? ?
* rousset et al . 2003 ) ) . in all cases
the s13 camera ( @xmath113 mas / pixel ) and the double - correlated read - out mode were used . for the raw data reduction we subtracted a mean dark from all science frames and the flatfield frames , then divided by the normalized dark - subtracted flatfield and subtracted the mean background by using eso _ eclipse / jitter_. in all three images a companion candidate is found 2.67 northwest of ct cha ( fig .
[ fig1 ] ) corresponding to @xmath1440 au at a distance of 165@xmath830 pc estimated from the combination of data by ( * ? ? ?
* bertout et al . ( 1999 ) ) and ( * ? ? ?
* whittet et al . ( 1997 ) ) for cha i members . to check for common proper motion of the tentative companion of ct cha we used the proper motion ( here pm ) of the star published in the literature ( table [ tab1 ] ) .
we use the weighted mean proper motion for checking , whether the two objects show common proper motion below .
we calibrated the naco data using the wide binary star hip 73357 for our two measurements in 2006 and 2007 .
the astrometric calibration for our images was done relative , hence the error bars of separation and position angle given in table [ tab2 ] include possible orbital motion of the binary inbetween our observations as well as their measurement uncertainties . to determine the positions of both components we constructed a reference psf from both objects .
thus , we obtained an appropriate reference psf for each single image . by using idl / starfinder we scaled and shifted the reference psf simultaneously to both components in each of our individual images by minimizing the residuals . as a result we obtained positions ( see table [ tab2 ] ) and relative photometry ( see table [ tab3 ] ) including realistic error estimates for each object by averaging the results of all single images taken within each epoch . from separation in figure [ fig2 ] , we can exclude by 3.4@xmath2 that ct cha b is a non - moving background object .
due to the location of the companion ( northwest of a ) and the proper motion ( towards northwest ) , the position angle information does not give much additional significance ( 1.0@xmath2 deviation from the background hypothesis ) , hence resulting in a combined significance of 3.7@xmath2 .
for a k7 star and a sub - stellar companion ( see below ) and the given separation ( at @xmath9 pc ) , the orbital period is @xmath10 yrs , so that the maximum change in separation due to orbital motion ( for circular edge - on orbit ) is @xmath11 mas / yr or @xmath12 pix / yr ( @xmath13/yr in pa for pole - on orbit ) . neither in separation nor in position angle ( measured from north over east to south ) , being @xmath14 , signs of orbital motion could be detected given the short epoch difference ( @xmath11 year ) .
we have found a new co - moving companion candidate of ct cha and estimated its spectral type from photometry to be m5 l5 . after this result
we have taken integral field spectra of the object with the vlt instrument sinfoni and could determine temperature and surface gravity of the object by comparison with synthetic spectra by ( * ? ? ?
* brott & hauschildt ( 2005 ) ) . from comparison with the eclipsing binary brown dwarf 2m0535 ( * ? ? ?
* ( stassun et al . 2006 ) ) we conclude that the object has less mass than 36 m@xmath15 , similar to the gq lup companion found by ( * ? ? ?
* neuhuser et al .
( 2005 ) ) .
our best mass estimate from the spectral modelling gives a few jupiter masses , so that the companion can be a planet imaged directly , hence has to be seen as planetary mass companion candidate , hence called b ( small letter for planets and planet candidates ) .
a detailed discussion of the spectra and further informations can be found in ( * ? ? ?
* schmidt et al .
( submitted to a&a ) ) . | we have searched for close and faint companions around t tauri stars in the chamaeleon star forming region . two epochs of direct imaging data were taken with the vlt adaptive optics instrument naco in february 2006 and march 2007 in ks band for the classical t tauri star ct cha together with a hipparcos binary for astrometric calibration . moreover
a j band image was taken in march 2007 to get color information .
we found ct cha to have a very faint companion ( ks@xmath0=14.6 mag ) of @xmath12.67 separation corresponding to @xmath1440au .
we show that ct cha a and the faint object form a common proper motion pair and that the companion is not a non - moving background object ( with 4@xmath2 significance ) . |
in the recent years a renewed interest on the glass transition phenomenon has motivated extensive experimental and theoretical works ( see @xcite and reference therein ) . on the theoretical side , new descriptions of the glass transition has been developed : they emphasized either the dynamic ( as the mode coupling theory ( mct ) of gtze @xcite or the coupled oscillators model of ngai and tsang @xcite ) or the thermodynamic ( as the first principle computation based on a replica formulation of @xcite or the inherent structure formalism computation of @xcite ) aspects of the transition itself .
a common feature of all these theories is they have been developed for `` model systems '' , often monoatomic models .
the comparison of the theoretical results with the real experiment are , therefore , complicated by the trivial observation that in the real world the glass forming systems are made out of `` molecular systems '' . as a consequence in the current literature
there is a large debate on the applicability of the theoretical predictions to the experimental outcome and the more stringent tests of the theories come from molecular dynamics ( md ) works . as an example , it is highly debated in literature the origin , in molecular glassformer , of _ secondary _ relaxations that can be observed by several experimental techniques beside the well - known microscopic and structural dynamics ( see , among others , @xcite and reference therein ) .
for instance , _ fast _ relaxations ( i. e. in the @xmath0 s range ) have been observed in several glasses and it is not yet clear if their origin is related to the molecular center of mass motion , as mct would explain in terms of @xmath1-process , or rather to rotational or intramolecular dynamics .
it is clear the crucial role that md simulations can play to solve this specific problem . if it is possible to build a
_ realistic _
model able to take into account the internal degrees of freedom , as well as the translational dynamics , computer simulations allow one to access any observable quantity of the system , also those not directly measurable by present experimental techniques . such possibility , together with the physical intuition , could allow the identification of the microscopic mechanisms underlying the different observed relaxation processes . in this paper
we want to address the problem to set up a `` realistic '' potential for a glass model system capable to account for the internal molecular degrees of freedom . among the glassforming molecular liquids characterized by an extremely rich dynamical behavior , the organic _ fragile _
@xcite glass former _ ortho - terphenyl _
( otp ) ( @xmath2 k , @xmath3 k @xcite , @xmath4 k ) has received much attention from both experimental and numerical simulation points of view .
the structure of otp molecule , shown in fig .
[ otp ] , is known from neutron @xcite and x - ray @xcite diffraction studies ; in condensed state the otp molecules are bound together only by van der waals forces , which resemble the lennard - jones ones often used by most theories and computer simulations aiming to study the glass transition problem . due to its structural complexity , if we would like to obtain reliable results to be compared with experimental data , we need to take into account not only the translations of the molecular center of masses and the rotations of the molecules as a whole , but also intramolecular motions like stretching along the molecular bonds , tilt of the bonds , rotation of the side rings with respect to the central one and so on . in other words , we need to describe the dynamics of the liquid at _ atomic _ level . on the contrary , in order to set up a computational scheme which is affordable in not - too - long time with the nowadays computer capability , we need the _ simplest _ model potential able to capture the relevant features of the dynamical behavior of the real system . in the literature numerical studies of otp have been proposed making use of different techniques ranging from _ harmonic lattice dynamics _
@xcite to molecular dynamics simulations on atomic level based on a general force field provided by the standard program alchemy iii @xcite . nevertheless , at our knowledge ,
only two studies based on molecular dynamics simulations of _ molecular models _ of otp have been proposed so far : _ i ) lewis et al . _
@xcite represent the molecule like a three - sites complex , each site playing the role of a whole phenyl ring , without internal dynamics and an intermolecular interaction of the lennard - jones ( lj ) type .
this model take in account only the dynamical behavior associated with the translations of the molecular center of masses and with the rotations of the molecules as a whole ; _ ii ) kudchadkar and wiest _ @xcite propose a more realistic model with the `` true '' structure of the molecule .
the intermolecular interaction is of lj type and , as internal degrees of freedom , only the rotational dynamics of the side rings with respect to the central one is taken into account .
these internal degrees of freedom are effectively the most relevant , nevertheless the authors parameterize the potential in such a way that the side rings are , at equilibrium , in a configuration that corresponds to a saddle point in the molecular energy surface .
the model potential we are going to introduce results to be much more efficient in mimicking the complexity of the dynamical behavior of the real system .
the paper is organized as follows : in the next section we introduce the intramolecular model potential ; in section [ force constants ] we explain how we calculated the force constants in order to reproduce a realistic isolated molecule vibrational spectrum . in section [ computation ]
we present some computational details and in section [ mct ] we remind some of the main well - established predictions of the ideal mode coupling theory used in section [ results ] to test the center of mass dynamical behavior ; in section [ results ] we discuss our md simulation results mainly with regard to the study of the diffusion and self - dynamic properties .
the last section contains an overall discussion and the conclusions .
in our model the otp molecule is constituted by three _ rigid _ hexagons ( phenyl rings ) of side @xmath5 nm connected as shown in fig . [ otp ] , i. e. two adjacent vertices of the _ parent _ ( central ) ring are bonded to one vertex of the two _ side _ rings by bonds whose length , at equilibrium , is @xmath6 nm . in our scheme ,
each vertex of the hexagons is thought to be occupied by a fictious atom of mass @xmath7 a.m.u .
representing a carbon - hydrogen pair ( c - h ) .
the choice of such a fictious atom , with its renormalized mass , greatly simplify the computer simulation but presents some drawbacks : _
i ) _ in the real molecule the two couples of carbon atoms connecting the rings are not bonded to hydrogen atoms , while in our model we consider all the 18 vertices having the same mass @xmath8 so that the total molecular mass is overestimated ( 234 rather than 230 a.m.u ) and _ ii ) _ the moments of inertia of the model rings are smaller than the real ones the hydrogen atoms being too close to the ring center .
nevertheless , we expect only minor effects on the overall dynamics from the previous simplification .
the three rings of a given molecule interact among themselves by an _
intramolecular _ potential , such potential being chosen _
i _ ) to preserve the molecule from `` dissociation '' ; _ ii _ ) to give the correct relative equilibrium positions for the three rings ; _ iii _ ) to represent the real intramolecular vibrational spectrum as close as possible . the interaction among different molecules , actually among the rings pertaining to different molecules , is accounted for by a site - site pairwise additive potential energy of the ( 6 - 12 ) lennard - jones type , each site being one of the hexagons vertices . to sum up
, the total interaction potential energy is written as the sum of an _ intermolecular _ and an _ intramolecular _ term : @xmath9 the first term can be written explicitly as : @xmath10 where @xmath11 is the position of the @xmath12-th atom ( @xmath12=1 6 ) in the @xmath13-th ring ( @xmath13=1 3 , hereafter @xmath13=1 indicates the parent ring ) belonging to the @xmath14-th molecule ( @xmath14=1 n ) , and @xmath15 .
\nonumber\ ] ] the optimal choice of the two intermolecular force parameters , @xmath16 and @xmath17 , will be discussed later . in principle
, the _ intramolecular _ interaction potential can be expressed in terms of the degrees of freedom describing the center of mass positions ( @xmath18 , @xmath13=2,3 ) and orientations ( e.g. the set of eulerian angles ) of the side rings with respect to the parent ring . however ,
for computational purposes , it is simpler to express the intramolecular potential in terms of orthonormal unit vectors attached to each ring or better in terms of quantities built from these vectors . with reference to fig .
[ geometry ] the sets of unit vectors for each ring @xmath19 are defined as @xmath20 i.e. @xmath21 and @xmath22 are orthogonal unit vectors in the ring plane , while @xmath23 is the normal to that plane .
the unit vectors that are parallel to the ring - ring bonds at equilibrium , are given by @xmath24 the positions of the four carbon atoms that link the three rings , i.e
@xmath25 and @xmath26 in the parent ring and @xmath27 and @xmath28 in the side rings , are given by , with respect to their ring centers @xmath18 , @xmath29 , @xmath30 , @xmath31 and @xmath32 .
finally , it is useful to define four further interaction sites two pertaining to the parent ring , @xmath33 and @xmath34 and two pertaining to the side rings ( 2 and 3 respectively ) , @xmath35 and @xmath36 . here @xmath37 , so that at the equilibrium position @xmath38 and @xmath39 . the variable we have introduced will be used to simply express the different contributions to the intramolecular potential in the next subsections .
the fluctuations of the distances among the centers of mass of the three rings are accounted for by introducing three `` springs '' .
the parent - side ring stretching implies the elongation of a c - c bond , which is expected to have a high stiffness .
the corresponding potential in harmonic approximation is written as : @xmath40^{2}+\\ & & c_{1 } \left [ \vert { \bar p}_{1(3)}-{\bar p}_{3}\vert -l_b \right ] ^{2 } \nonumber\end{aligned}\ ] ] on the other hand , no direct chemical bond is present between the two side rings , and we expect a less stiff spring for the fluctuation of the distance between the side ring centers .
we model this interaction by : @xmath41^{2}\ ] ] the determination of the force constants @xmath42 and @xmath43 , as well as the others we are going to introduce , will be discussed later . in the otp crystal structure@xcite , the bond angles @xmath44 and are 123.6@xmath45 and 123.0@xmath45 respectively , while the angles @xmath46 and @xmath47 are 118.4@xmath45 and 117.4@xmath45 ( see fig . [ geometry ] ) further , in the isolated molecule , the ring - ring bonds are forced out of the plane of the parent ring so that the dihedral angle is 5.2@xmath45 .
this lack of planarity is due to the little asymmetry introduced by the difference between a carbon bonded to an hydrogen and a carbon bonded to a carbon of another ring . in our model , all these angles at the equilibrium are put equal to 120@xmath45 and @xmath48 equal to @xmath49 .
we model the restoring forces for these angles by using the scalar product of the unit vectors @xmath50 and @xmath51 ( as well as that of @xmath52 and @xmath53 ) .
being , at equilibrium , the quadratic term in the small oscillation approximation is given by : @xmath54 however , this term is not enough to ensure the co - planarity of the vectors @xmath50 and @xmath55 with the parent ring ; to force the rings towards the co - planar equilibrium condition we make use of the `` sites '' @xmath56 and @xmath57 ( as well as @xmath58 and @xmath59 ) introducing between them a spring of vanishing equilibrium length : @xmath60 inside the intramolecular dynamics , we expect that a crucial role is played by the rotation of the side rings planes around the ring - ring bonds@xcite in interfering or modifying the intermolecular relaxation processes .
the relevant variables describing this motion are the two angles @xmath61 between the normals to the side ring and parent ring planes . in the crystalline structure ,
the two side rings have slightly different angles , 42.4@xmath45 and 62.0@xmath45 @xcite .
however we model the disordered condensed phases ( liquid and glass ) by using as equilibrium angle values those of the isolated molecule , i.e. 54.0@xmath45 , as discussed below .
we have to remark that the isolated molecule symmetry implies two iso - energetic configurations separated by a finite barrier as is qualitatively illustrated in fig . [ map ] .
we performed an _ ab - initio _ calculation of the single molecule potential energy surface as a function of @xmath62 and @xmath63 with all the other internal degrees of freedom fixed to their equilibrium values .
such calculation consists in the minimization of the hartree - fock energy over the gaussian basis set 3 - 21 g .
for each atomic species the inner shell is made up of 3 gaussians while the valence shell is a linear combination of 2 gaussians orbital plus a 1 gaussian orbital ; the minimization was carried out by the standard package gaussian 94 . in fig .
[ map ] a contour plot of the hartree - fock potential energy is shown as a function of the rotation angles .
it is important to quote that this map has only a semi - quantitative meaning , the structure of the whole molecule being not re - optimized during the scan of @xmath62 and @xmath63 angles . however , a careful study , performed reoptimizing the whole molecular structure , has been done around the saddle point @xmath64 and around the two equivalent minima that turn out to be at @xmath65 and @xmath66 . from this calculation the barrier height has been estimated to be @xmath67 and , from this value , it is possible to envisage the nature of the rotational motion at the temperatures of interest : the two side rings can pivot in phase around the bonds crossing from one minimum to the other degenerate one . moreover they can perform librational out - of - phase motions of ( approximatively ) harmonic type . in order to represent this potential surface
we express the in - phase rotation of the two side rings with a high - order ( 6th ) polynomium and the out - of - phase rotation in the harmonic approximation . for this purpose
we use as primary variables the scalar products @xmath68 and @xmath69 ( their equilibrium value being @xmath70 ) and to disentangle the in - phase from the out - of - phase motion , we introduce the two variables : @xmath71 the final form of the `` internal rotation '' potential will be @xmath72 with @xmath73 the parameters @xmath74 , @xmath75 and @xmath76 describing the in - phase rotation potential are derived according to the following procedure . in the harmonic approximation , in proximity of the equilibrium position @xmath77 for the scalar products , it must hold @xmath78 and taking into account that the barrier height must be equal to @xmath79 , we have the following conditions : @xmath80 implying @xmath81
we can finally write our internal model potential like @xmath82 this potential is parametrized by the set of six free coefficients @xmath83 whose actual value can be tuned in order to obtain a _
realistic free molecule vibrational spectrum_. indeed , in the small oscillations approximation , we can determine the values of the coefficients @xmath83 by diagonalizing the dynamical matrix and fitting the resulting eigenfrequencies @xmath84 to the _ lowest _ frequencies @xmath85 obtained by an hartree - fock calculation of the vibrational frequencies in the electronic ground state of the isolated molecule .
we have @xmath86 eigenvalues for the dynamics of three rings ( @xmath87 ) but actually only @xmath88 eigenvalues @xmath84 are non vanishing , the other being associated with the translations rotations of the molecule as a whole . more explicitly , the set @xmath83 is obtained by minimizing , by the standard levenberg - marquardt algorithm@xcite , the error function @xmath89 where , for a given set of @xmath83 , the quantities @xmath84 are the solution of the eigenvalues problem @xcite @xmath90 here @xmath91 and @xmath92 are the hessian matrixes of the second partial derivatives of the potential and kinetic energy respectively , with respect translational and rotational degrees of freedom of the three rings . in tab .
[ coeff ] are reported the values of the coefficients @xmath83 determined by the minimization together with the values for @xmath93 derived from eqs .
( [ bval ] ) .
the corresponding set of frequencies is shown in tab . [ frequencies ] where @xmath94 are the frequencies derived from a computer simulation on a system of isolated molecules at low temperature .
these frequencies have been identified via the peaks of the spectrum of the velocity autocorrelation function .
other model details are reported on ref .
we have studied a system composed by 108 molecules ( 324 rings , 1944 interaction centers ) ; the sample is large enough to neglect finite size effects on the investigated properties with reasonable computation times .
the values of the parameters entering the site - site lennard - jones interactions were determined by preliminary simulations : the value of @xmath16 were firstly determined by comparing the computed and the experimental @xcite self - diffusion coefficients versus temperature in the range @xmath95 k @xmath96 k ; the value of @xmath17 was estimated by tuning the static structure factor so to place the first maximum in the right position @xcite .
successive iterations led to @xmath97 k and @xmath98 nm . to speed up the calculation of the intermolecular potential and to assure all the torques to be estimated in a consistent way
, the cut off for l - j interactions is applied between rings centers , i.e. the sites are considered not interacting when they pertain to rings whose center distance is larger than @xmath99 . at this stage a remark about the rotational potential expressed in eq .
( [ torsion ] ) is needed : such potential presents a serious drawback since the intrinsic ambiguity due to the parity of the scalar products involved can force the side rings in the wrong positions
. this problem can be bypassed introducing a new term in the intramolecular potential that gives zero contribution to the energy when the molecule is close to the equilibrium position .
if we define the product @xmath100 , and we put @xmath101 ( the actual value of @xmath102 is irrelevant , it must only be able to force the rings in the right way ) we can write @xmath103 this term has been activated only during the preparation of the initial configuration ( i. e. at high temperature ) ; during the thermalized evolution of the system we expect molecules do not drive too much away from their equilibrium configuration and then this term to be always equal to zero . to integrate the equations of motion we have treated each ring as a separate rigid body , identified by the position of its center of mass @xmath104 and by its orientation expressed in terms of quaternions @xmath105 @xcite .
the standard verlet leap - frog algorithm @xcite has been used to integrate the translational motion while , for the most difficult orientational part , the refined algorithm due to ruocco and sampoli @xcite has been employed ; such choices allow a very stable integration with a relatively long time step .
the rotational dynamical problem can be written as @xmath106 where @xmath107 is the angular momentum of ring @xmath13 in the molecule @xmath14 , @xmath108 the torque acting on it and @xmath109 is the inverse of the inertia tensor in quaternion coordinates .
the expression of the torques is simplified as the rotational part @xmath110 of the intramolecular potential energy has been expressed in terms of suitable scalar products written in the general form @xmath111 .
therefore the value of the torques @xmath112 and @xmath113 associated to the degree of freedom corresponding to the angle @xmath114 can then be evaluated by @xcite @xmath115 the leap - frog algorithm for the rotational motion reported can be written in the form @xcite @xmath116 the crucial point is that the dependence of the matrix @xmath117 on the angular variables implies the need of using a time step @xmath118 to perform the numerical integration appearing in eq .
( [ rot_alg ] ) , smaller then @xmath119 used for the cms integration . in turn , the value of @xmath119 is limited by the highest vibrational frequency of about @xmath120 . in our case
, the chosen values of @xmath121 fs and @xmath122 are found to be sufficiently small to reduce the fluctuations of the total energy to a negligible fraction of the kinetic energy .
it is important to note that the use of the smaller @xmath118 does not increase significantly the cpu computational time since the time consuming part in each md step is the calculation of the forces and torques which are kept fixed during the integration of eq .
( [ rotation ] ) .
we considered a wide temperature range spanning the liquid phase and reaching the region close to @xmath123 as shown in tab .
[ protocol ] in which the whole set of simulation times is reported . during the different temperature runs , the size of the cubic box has been rescaled in order to keep the system at the experimental density which , for @xmath124 , can be fitted by the polynomium @xcite @xmath125 where @xmath126 is in @xmath127 and @xmath128 in k. at each temperature @xmath128 we have organized the simulations following this scheme : * the system was coupled for a time @xmath129 , chosen in a somehow arbitrary way , to a stochastic heat bath , i.e. , the velocities of the rings were replaced following a logarithmic pattern with the velocities drawn from a boltzmann distribution corresponding to such temperature ; * at this point the system was at the desired temperature and we let it to perform a _
microcanonical _ time evolution ( constant energy ) for a period @xmath130 comparable with the experimental structural relaxation time @xmath131 at the same temperature ; in such a way we expect every slow degree of freedom of the system to be correctly thermalized and we control that there was no drift in temperature and the degree of energy conservation ( fluctuations are always less than @xmath132 of the kinetic energy ) .
* we considered the final system configuration obtained in this way as a good equilibrium starting state for a molecular dynamics trajectory . at each temperature
we perform three different runs : the first one @xmath133 ps long with a @xmath134 ps configuration saving time has been used to compute the small time dynamical behavior of the system ; a second one @xmath135 ps long with a @xmath136 ps saving time has been used for some intermediate frequency analysis .
the last one , with length @xmath137 and saving time @xmath138 dependent on temperature , has been used to calculate static quantities and the long time behavior of the system .
all the calculations have been performed on a cluster of four @xmath139-cpu with a frequency of 500 mhz ; every nanosecond of simulated dynamics needed approximately 24 hours of cpu - time .
a great improvement to our understanding of the glassy state of matter has come from the extension of the theoretical building of the mode coupling theory ( mct ) @xcite developed for the _ equilibrium _ description of the dynamics of simple , i.e. monatomic , liquid to the study of the glassy state .
although such theory is the only microscopic approach to the glass transition leading to many predictions on the experimental data , it is still at the center of a strong debate and some questions stay open .
infact , even if the real range of validity of mct for the study of molecular liquids has been cleared in the last years ( see , among others , refs .
@xcite ) , some experimental results , like the presence of the so - called knee characterizing the low - frequency behavior of the light scattering susceptibility @xcite or the presence of a cusp in the non - ergodicity parameter @xcite , seem to contradict fundamental predictions of the idealized version of mct . in this section
we sum up the main predictions of the so called _ ideal _ mct where it is hypothesized a complete dynamical freezing and the so called `` thermally activated hopping '' processes are neglected ; such predictions will be compared with our simulation data . in the ideal mct
the glass formation is interpreted as a _ dynamical transition _ from an ergodic to a non - ergodic behavior at a cross - over temperature @xmath140 .
mct provides a self - consistent dynamical treatment @xcite for the density correlation function of an isotropic system @xmath141 where @xmath142 is the number of the particles , @xmath143 , @xmath144 and @xmath145 is the position of particle @xmath14 at time @xmath146 .
mct proposes a particular _ ansatz _ for the memory kernel in the related integro - differential generalized langevin equation , such kernel is coupling non - linearly the density fluctuations with one another .
if the coupling increases upon lowering the temperature , the resulting dynamical feedback leads to a progressive slowing down of the density fluctuations until they become completely frozen at the critical temperature @xmath123 .
the ideal mct describes the behavior as much as the temperature approaches @xmath140 , i.e. the parameter @xmath147 is small ( however real comparisons have to be made for @xmath17 not too small , in contrast to the case of scaling laws in phase transitions ) . for temperature @xmath148 , @xmath149 is characterized by two step decays taking place at different time scales and the theory gives specific predictions for such different time regions . the first one , the _ @xmath1-process _
region , is centered around a time @xmath150 which is predicted to scale like @xmath151 with @xmath152 and to be bounded in the interval @xmath153 where @xmath154 is the time scale of the microscopic dynamics and @xmath155 is the structural rearrangement time scale . in this region
the _ factorization property _ holds , in the sense that the density correlation function can be written as @xmath156 where @xmath157 is the _ non - ergodicity parameter _ ( i.e. _ debye - waller factor _ for collective correlators or _ mssbauer - lamb factor _ for single - particle correlators ) , @xmath158 is an amplitude independent of temperature and time and the @xmath159 in @xmath160 corresponds to time larger or smaller with respect to @xmath161 .
so , the time dependence of the correlation functions is all embedded in the _ q - independent _ function @xmath160 , namely spatial and temporal correlations result to be completely independent .
@xmath162 is asymptotically expressed by two power laws , respectively the _ critical decay _ and the _ von schweidler law _
@xcite , characterized by the _ temperature and momentum independent _ exponents @xmath163 and @xmath164 @xmath165 here @xmath163 is the same exponent of the power divergence of @xmath150 at @xmath123 and it is related to the exponent @xmath164 ( @xmath166 ) via the equation @xmath167 where @xmath168 is the gamma function .
the second time region is the so - called _ @xmath139-region _ where the second decaying step takes place .
this region is connected to the collective structural relaxations and asymptotically the theory predicts the validity of the well known _
time - temperature superposition principle _ ; it states that , on time scales of same order of magnitude of @xmath131 , the following scaling law holds at every temperature @xmath128 @xmath169 in other words , the correlation functions of any observables at different temperatures can be collapsed into a master curve when the time is scaled with @xmath170 .
moreover mct predicts that this master curve can be fitted by a kolrausch - william - watts function ( _ stretched exponential _ ) @xmath171 the @xmath139 time scale @xmath155 depends on temperature trough a power law of the form @xmath172 where the _ q - independent _ exponent @xmath173 is related to the power exponents @xmath163 and @xmath164 of the @xmath1-region by the relation @xmath174 the inverse of the diffusion constant @xmath175 is predicted to scale like @xmath155 @xcite and consequently it follows eq .
[ [ powerlaw ] ] . up to now ,
all dynamical results reviewed are universal in the sense that they are predicted to hold for the correlators of every observable with non - zero overlap with density ; in particular this is true for both the one - particle and the collective density correlation functions .
nevertheless important differences are predicted to hold for the q - dependence in these two cases : in the former case @xmath157 and @xmath158 depend smoothly on @xmath176 , in the latter one they oscillate respectively in phase and out of phase with the static structure factor @xmath177 .
moreover , @xmath178 is predicted to be a smooth function of @xmath176 in the one - particle case ; at variance , it shows pronounced oscillations in phase with @xmath177 in the collective case .
in this subsection some thermodynamical time - independent results are shown like potential , kinetic , and total energy ( see tab .
[ thermotab ] and fig .
[ thermo ] ) .
the interest of these results is clarified by the following argument : in computer simulations dealing with the glass transition it is possible to define a temperature often named @xmath179 @xcite at which one - time quantities show some sort of discontinuity
. such discontinuity , whose position depends on the thermal history of the system , represents the thermodynamical point at which the system undergoes a glass transition on the time scale of the computer simulation , falling out of equilibrium .
it is clear from fig .
[ thermo ] that no discontinuity is present , i.e. @xmath179 of our simulations is less then the lowest temperature studied and then we have good chance to have well - thermalized results .
nevertheless , whether or not the system is in equilibrium can be checked only _ a - posteriori _ by comparing the total simulation time with the measured relaxation time . from the linearity of @xmath180
we deduce a specific heat @xmath181 constant in the temperature range investigated and equal to @xmath182 ; such value must be compared with the experimental value of @xmath183 @xcite .
it is possible to explain the inconsistency between the two values keeping in mind that our md value is a classical ( non - quantic ) result and , more important , we are neglecting many ( @xmath184 ) degrees of freedom concerning the deformations of the phenyl rings .
in general the static structure of a fluid is well described by the _ pair distribution function _ @xcite : @xmath185 in computer simulations @xcite , we can identify the distances @xmath186 with different quantities . in fig . [ gdr ]
we report some @xmath187 s at @xmath188 k where we have considered as @xmath186 the distances between the carbon atoms belonging to different rings ( a ) and between the center of masses of rings ( b ) and molecules ( c ) ; both total ( solid line ) and intermolecular ( dashed line ) contributions are shown in order to separate the internal molecular structure and the mean structural organization of the whole bulk sample . in fig . [ gdr ]
( b ) a two peak structure is present : the first sharp peak is placed at @xmath189 nm corresponding to the mean distance between rings belonging to the same molecule ; the second one , of intermolecular origin , is placed at @xmath190 nm .
it is worth noting that such distance is less than the greatest intramolecular c - c distance ( @xmath191 nm ) .
moreover the molecular centers of mass @xmath187 also shows a large value on distances less than @xmath192 nm giving the evidence of a strong packing of the molecules .
all these features appear to be approximately temperature independent .
such packing depends strongly on the orientational internal configuration of the molecules namely on the positions of the two side rings with respect to the parent one ; the computation of the probability distribution of the scalar products among the versors @xmath193 , @xmath194 and @xmath195 , introduced in sect .
[ molecular model ] , is somehow instructive in this sense . in fig .
[ distrang ] the distribution functions for the quantities @xmath196 , @xmath197 , @xmath198 are shown .
the first two distributions are practically temperature independent and give us only informations on the correctness of the simulated geometry : they are sharply peaked on the the correct equilibrium positions of about @xmath199 and @xmath200 respectively . at variance with the distribution of @xmath201 and of @xmath198 that are symmetric around @xmath202 ,
the distribution of @xmath196 does not present the symmetric peak on negative values so that we can argue that the auxiliary term @xmath203 worked correctly .
the most interesting distribution is the third one in which the peak intensity ( the peak is correctly placed at @xmath204 ) is higher the lower the temperature , as shown in fig .
[ distrnorm2 ] , indicating therefore that the correspondent degree of freedom ( the in - phase motion of @xmath198 ) is more and more frozen on its equilibrium value with decreasing the temperature .
we have seen that in the isolated molecule the rotational motion of the two side rings can be separated in two contributions : an out - of - phase harmonic libration and an in - phase pivoting around the bonds which permits rings to cross from one equilibrium position to the other degenerate one .
it is clear from the structure of the distribution function in proximity of @xmath205 shown in fig .
[ distrnorm ] that the time needed for the transition from a minimum to the other one will be longer lowering the temperature ; moreover fig . [ distrnorm ] can be considered as a restatement of the energy map shown in fig
. [ map ] , being the intensity of the maximum in zero a measure of the transition probability between the two minima .
such fenomenology will be clarified in future communications where we will study the relaxation processes associated to the angular degrees of freedom .
the space fourier transform of @xmath187 is the _
static structure factor_. in a poliatomic system this quantity is defined as @xmath206 where the coefficients @xmath207 are the _ scattering lengths _ in principle different for each species involved . the @xmath177 has been determined experimentally for otp by neutron scattering @xcite , and the following main features have been observed : _ i _ ) in contrast to atomic systems its main peak is splitted in two sub - peaks placed around @xmath208 nm@xmath209 and @xmath210
nm@xmath209 _ ii _ ) in the @xmath211 region , lowering the temperature , a reduction of scattering intensity is observed due to the decrease of the isothermal compressibility @xmath212 ( @xmath213 ) _ iii _ ) increasing the density , a slight shift of the peak position to higher @xmath176 values is observed _ iv _ ) on decreasing temperature , the height of the peak around @xmath210 nm@xmath209 increases while the intensity of the peak at @xmath208 nm@xmath209 remains nearly unaffected except for a slight reduction mostly connected to the decrease in @xmath212 . in fig .
[ sk_cm_rng ] we show our results for the structure factors calculated assuming as scattering centers the molecules and the rings centers of mass with @xmath214 ; every point is an average on all the independent miller indices corresponding to the given @xmath176 .
it is much more interesting to make a comparison among the md results and the experimental data obtained by neutron ( fig .
[ essekexp ] ( a ) ) and to test what is expected for x - ray scattering ( fig .
[ essekexp ] ( b ) ) . in evaluating @xmath177 by computer simulation for a comparison with neutron data
, we have to take in account the contribution due to both carbon and hydrogen atoms ; @xmath215 atoms are not considered in our dynamics , nevertheless it is possible to place them in fixed positions on the line extending from the center of the ring trough a carbon atom at fixed @xmath216 distance computed to be @xmath217 nm . in this case
we would have to consider different scattering lengths for the two species , @xmath218 and @xmath219 ; nevertheless they are both positive and about of the same magnitude so that the product @xmath220 in eq .
( [ essek ] ) is an ineffective constant . in fig .
[ essekexp ] ( a ) the calculated @xmath177 at @xmath221 is shown and compared with the data of ref .
@xcite at @xmath222 k ; in this paper the authors show their results in terms of the _ coherent scattering cross section _ @xmath223 measured in @xmath224 which is proportional to our @xmath177 . in order to compare the two results we renormalized the experimental data in such a way the values of the two curves coincide at large @xmath176 .
the high - q region of the calculated @xmath177 appears to be in excellent agreement with the experiment but no double peak structure is present at low momenta .
in particular the md calculated first peak presents a small bump at about @xmath86 nm@xmath209 ; this is better seen in fig .
[ sk_err ] where we show the small-@xmath176 part of @xmath177 calculated at @xmath225 together with the error bars estimated by means of the statistical fluctuation of the data .
the noise can not allow us to determine the correct structure of the main peak .
it is worth noting , moreover , that at this low temperature the characteristic relaxation time is of order @xmath226 ns , so that , considering a simulation run @xmath227 ns long , we have only about @xmath227 really independent system configurations . in order to calculate the simulated @xmath177 as it is expected by x - ray scattering ,
we consider only the carbon atoms ; also in this case no double - peak structure is observed in the data but a clear prepeak appears at a q - value less then the @xmath176 of the first maximum , being the high - q behavior similar to the neutron case , as shown in fig .
[ essekexp ] ( b ) .
an important quantity to consider in the study of the dynamics of our system at a microscopic level is the _
mean squared displacement _ ( msd ) defined as @xmath228 where @xmath229 is the position of the center of mass of the ring @xmath13 in the molecule @xmath14 at time @xmath146 ; from the msd is possible to determine the _ self - diffusion coefficient _ @xmath230 via the einstein relation @xmath231 the temperature dependence of the msd is shown in fig .
[ erre2 ] ; each curve follows the usual _ cage - effect _ scenario . at small time
( less than @xmath232 ) they present the @xmath233 behavior corresponding to the ballistic motion ; at long time the diffusive linear time dependence of eq .
( [ einstein ] ) is found . at intermediate times
a small region is present where msd stays almost constant and whose duration increases with decreasing temperature ; on these time scales molecules are trapped in cages builded up by their neighbors , and they can only vibrate in these limited region , the length of the plateau being a measure of the mean life - time of the cages .
the calculated values of the self - diffusion coefficient are shown in tab .
[ diff_tab ] and plotted in fig [ diff_fit ] ( open circles ) as a function of temperature , together with the power - law temperature dependence ( solid line ) predicted by the mct @xmath234 a three parameters fit to these data has been performed obtaining the following values @xmath235 in the same figure we also show the experimental data ( full squares ) @xcite that are well represented by eq .
( [ power ] ) ( dashed line ) with the values @xmath236 and @xmath237 k. it is clear from these values and from fig .
[ diff_fit ] that a discrepancy is present among the lower temperatures diffusive behavior of the simulated and real system respectively ; this is most likely due to the fact that we have tuned the value of the l - j potential parameters @xmath16 and @xmath17 in order to reproduce the high - t diffusion properties of the real system .
however it is worth noting that it is possible to reproduce quite well the experimental results on the whole investigated temperature range shifting the molecular dynamics points at temperatures @xmath238 k above their true values . in other words we have to assume that our actual thermodynamic point is shifted with respect to the real one ; from now on , whenever we will compare our molecular dynamics results with the experimental ones , our calculated points will be shifted of @xmath133 k above the measured temperature and the competing temperatures will be indicated as @xmath239 . on this grounds from the previous study of the self diffusion properties of our model
we obtain @xmath240 to be compared with the experimental value @xmath241 . a different way to determine the parameters entering in the power law reported in eq .
[ power ] is possible , even if not independent from the previous one ; it is based on the study of the _ non - gaussian _
parameter @xmath242 defined as @xcite : @xmath243 where the mean square displacement @xmath244 and @xmath245 are , respectively , the second and forth momenta of the _ van hove self - correlation function _ @xmath246 the parameter @xmath242 quantifies the degree of non - gaussianity of @xmath247 in space as a function of time and it is normalized in such a way that , if @xmath248 was a gaussian function in space at a given time @xmath249 , we would have @xmath250 .
the time dependence of @xmath251 at all temperatures investigated is shown in fig .
[ alpha2 ] .
we are not interested here in the specific time dependence of such function but only in the fact that @xmath252 , the position of the maximum of @xmath242 , has the power dependence on @xmath128 similar to that of eq .
( [ power ] ) @xcite ( see fig .
[ alpha_fit ] ) .
a fit to this quantity performed in the same way as before gives us the values @xmath253 compatible with the values determined by the temperature dependence of @xmath254 , even if the error bars are larger in this case .
comparisons of the _ coherent _ ( collective ) and _ incoherent _ ( self ) density fluctuations dynamics data measured by different techniques ( neutron time - of - flight and backscattering spectroscopy , photon correlation spectroscopy , depolarized raman and rayleigh - brillouin light scattering ) with the main predictions of mct have been reported in literature with great details @xcite . in this section
we will study the single particle density fluctuation dynamics of our model and we will compare our results with the experimental results mainly contained in refs .
@xcite and with the mct predictions .
the single particle dynamics of the model is embedded in the _ incoherent self - intermediate scattering function _ defined as @xmath255}\rangle\ ] ] where , again , @xmath229 is the position of the center of mass of the ring @xmath13 in the molecule @xmath14 at time @xmath146 . at every temperature considered two sets of configurations , produced with two different storing time as described in section [ computation ] ,
have been used to reconstruct the whole curve .
we considered the @xmath128-dependence of @xmath256 at the two momentum values @xmath257 nm@xmath209 corresponding to the first and second peaks of the static structure factor , averaging on values of @xmath176 falling in the interval @xmath258 with @xmath259 nm@xmath209 .
finally , we spanned at @xmath188 k the whole interesting @xmath176-space in the interval @xmath260 nm@xmath209 ( averaging on the values of @xmath176 falling in the same interval @xmath261 wide ) . in fig .
[ f_self ] we show @xmath262 for nearly all temperatures investigated at @xmath263 nm@xmath209 ; all the curves decay to zero i.e. the length of all the simulations allows the fluctuations to become completely uncorrelated .
we are in a `` good '' thermodynamical equilibrium at every temperature , at least on the space scales corresponding to the inverse of @xmath264 . at temperatures
lower then @xmath265 k the relaxation follows clearly the predicted two step pattern : on microscopic time - scales the correlation is quadratic in time , this time scale being the one on which the intramolecular vibrations happen ; on intermediate time scales we observe the formation of a plateau , whose height is the non - ergodicity parameter @xmath157 and whose length in time is comparable to the one of the plateau in the msd @xmath244 . on long time scales we observe the structural relaxation in the form of a stretched exponential . at the highest temperatures no double pattern is visible anymore and only a nearly exponential relaxation can be recognized . a stretched exponential fit ( see eq .
( [ stretched ] ) ) on the structural time scale ( @xmath139-process ) gives us the temperature dependence of the three free parameters @xmath178,@xmath131 and @xmath157 .
the parameter @xmath178 ( circles ) is shown in fig .
[ beta_err ] ; it appears to be nearly @xmath128-independent for temperature lower than @xmath266 k and its mean value @xmath267 ( dashed line ) has to be compared with the experimental value @xmath268 . for temperature in the higher region it tends toward the value @xmath269 ( errors are clearly much more greater ) ; such behavior is due to the fact that in this temperature region it is no longer possible to sharply separate the long - time relaxation region from the microscopic short - time one .
the study of the temperature dependence of the non - ergodicity parameter @xmath157 in the interesting region is not possible due to our limited temperature range which do not permit to observe the expected low - temperature ( @xmath270 ) harmonic debye - like behavior , and the onset of the anomalous decrease of @xmath157 with increasing @xmath128 for @xmath271 .
our data suggest us @xmath272 k and the mean value @xmath273 ( dot - dashed line ) agrees with the experimental value determined at @xmath274 k shown in fig .
[ fit1_q ] .
it is worth to test the power law temperature dependence eq .
( [ powerlaw ] ) for the relaxation time @xmath155 ; the calculated relaxation times ( circles ) shifted of @xmath133 k with respect to the measured temperatures , as explained above , are plotted in fig .
[ tau_err ] together with the experimental ( full triangles ) _ shear viscosity _ @xmath275 data ( @xmath275 is expected to be proportional to @xmath155 ) of ref . @xcite and the theoretical fitted curve ( solid line ) of parameters @xmath276 to be compared with the experimental results of ref .
@xcite @xmath277 k and @xmath278 .
these values are compatible , within the statistical error , with the values calculated from the diffusion data so that we can conclude that the diffusive behavior and the self - dynamics of our model follow the same _ critical _ power law with @xmath279 k and @xmath280 .
also the values of @xmath131 for @xmath281 nm@xmath209 ( squares ) corresponding to the second peak of the static structure function are reported with the theoretical curve .
a fit has been performed ( dashed line ) only on the prefactor keeping fixed the values of the other two parameters in order to show that these data also are compatible with the same power law . the crucial observation here
is that the values of the two parameters @xmath140 and @xmath173 are effectively q - independent and they can be considered universal for our model , as predicted by the mct . the relaxation time @xmath131 can be also used to test the time - temperature superposition principle eq .
( [ superposition ] ) . in fig .
[ rescaling ] the curves are shown in function of the rescaled time @xmath282 and it is clearly seen that all the curves tend to collapse on the same master curve as predicted by the theory .
we now quantify the q - dependence of the self dynamics long - time behavior of the system at @xmath188 k. in fig .
[ f_self_q ] are reported the curves @xmath262 for values of @xmath283 nm@xmath209 with @xmath284 ; the choice of the temperature value @xmath188 k has been due to the need of `` well - thermalized '' results in a large range of q. also in these data is well defined the two step behavior and we can calculate the long time stretched exponential fit parameters ; the resulting values are shown in fig .
[ fit1_q ] and fig .
[ fit2_q ] . in fig .
[ fit1_q ] the values of @xmath178 ( left side ) and @xmath157 ( right side ) are shown .
@xmath178 ( open circles ) appears to be a smooth function of @xmath176 and it tends , for large values of @xmath176 , to the experimental ( full circles ) evaluated value of @xmath285 .
such behavior is quite general ( see , for instance , ref .
@xcite ) and can be easily explained by the following argument @xcite : for large values of @xmath176 , corresponding to length scales of the same order of magnitude of the cages dimension , the dynamics becomes slower and slower approaching the cage dynamics described trough the von schweidler exponent @xmath164 . at variance , in the opposite limit of small @xmath176 , we consider a diffusive dynamics on large distances ; at such length scales the decay of the self - density fluctuations is of the usual purely exponential form corresponding to @xmath286 ( see fig .
[ gauss_approx ] ) . at this stage
, however , we have not a reasonable explanation of the disagreement with the experimental data . in the right side of fig .
[ fit1_q ] the q - dependence of the non - ergodicity parameter @xmath157 is also shown as calculated from the short time limit of the @xmath139-process ( squares ) , from the long - time part ( triangles ) of the @xmath1-region ( as we will see below ) and from the experimental data ( full squares ) @xcite ; it seems clear a good agreement between our values and the experimental results .
the data appear to be monotonic decreasing as increasing @xmath176 and this dependence is expected to be approximately gaussian ; in fig . [ fit1_q ] a gaussian fit ( solid line ) in the form @xmath287 with @xmath288 nm@xmath209 to molecular dynamics @xmath139-region data is also shown .
it is clear that the @xmath176-range considered here is too limited to really decide on the validity of this functional form ( a linear approximation would work well too ) , a good estimate of the error bars lacking in this case . in fig .
[ fit2_q ] the @xmath176-dependence of @xmath289 is shown ( circles ) together with the experimental data @xcite(full circles ) .
molecular dynamics points have been rescaled by a factor @xmath290 to take into account the fact , as discussed above , that our system temperature is @xmath133 k higher the real one . the correct square law behavior at low-@xmath176 @xmath291 ( see eq .
[ gaussapprox ] ) is also shown as a solid line ; here @xmath254 is the self diffusion coefficient and @xmath292 nm@xmath293/ps .
finally , in fig . [ tau_d ] , two products of some calculated quantity , expected to be constant , are shown ; on the left side the statement @xmath294 is proven for @xmath295 ( good at highest temperature ) .
on the right side we show the product @xmath296 evaluated at @xmath188 k which is nearly constant up to @xmath297 as expected , being this value approximately the crossover point among the correct quadratic behavior and the asymptotically linear regime as we found above .
let us now probe the mct conclusions about the @xmath1-_region _ which is predicted to follow the power laws of eq .
( [ abpower ] ) : we show here the results at fixed temperature @xmath188 k for values of @xmath298 nm@xmath209 with @xmath299 .
we fit all the curves by two power functions on the time ranges @xmath300 and @xmath301 : @xmath302 where @xmath303 $ ] ( ps ) and @xmath304 $ ] ( ps ) ; some selected fits are shown in fig .
[ beta_resc_q ] and they seems to work quite well .
some observations are needed on the following analysis : a great uncertainty stems from the choice of the fit range ( i.e. from the choice of @xmath300 and @xmath301 ) due to the consideration of a cross - over region between two process not sharply separated in time ; moreover it can not be excluded an analogous problem between the microscopic region and the critical decay region characterized by the exponent @xmath163 .
such difficulty implies a great uncertainty on the determination of @xmath157 which is supposed to be considered as the long - time limit of the @xmath1-process and the short - time behavior of the @xmath139-relaxation . at this point
it is clear that , lacking a careful error analysis on the data points , such fits can only state that a parameterization of the data in the form of eq .
( [ abfit ] ) is possible which is consistent with the theory predictions @xcite , the current values of the determined parameters being considered only from a qualitative point of view .
nevertheless the values of such fitting parameters , shown in fig .
[ beta_fit_q ] and fig .
[ h_c1_c2 ] , appear to be in good agreement with the values obtained from the experimental data . in fig .
[ beta_fit_q ] are shown ( left ) the values of the power exponents @xmath163 ( squares ) and @xmath164 ( circles ) and of the exponent @xmath173 ( triangles ) calculated by means of eq .
( [ abgamma ] ) ( right side ) ; the mean values are @xmath305 , @xmath306 , @xmath307 respectively , to be compared with the experimental determined values @xmath308 ( dot - dashed line ) , @xmath309 ( dashed line ) , @xmath310 ( solid line ) @xcite .
it is clear from these results that one of the main prediction of mct , namely the @xmath176-independence of @xmath163 and @xmath164 is verified in the limit of the error fluctuations .
moreover it is important to note that the parameter @xmath173 , given by eq .
( [ abgamma ] ) , remains constant , as expected , being its mean value @xmath307 clearly compatible with the value @xmath311 determined from experimental data @xcite ; nevertheless this value overestimates the value @xmath312 at @xmath313 nm@xmath209 previously determined by the fit to the @xmath139-region . furthermore @xmath314 is always less then the determined value of the large-@xmath176 value of the stretching parameter @xmath315 ( see fig .
[ fit1_q ] ) verifying an other mct prediction , namely @xmath316 . from eqs .
( [ betaregion ] ) ( [ abpower ] ) ( [ abfit ] ) the two parameters @xmath317 and @xmath318 result to be proportional to @xmath158 , the proportionality constants being dependent on @xmath17 , @xmath161 , @xmath163 and @xmath164 . from @xmath319 k
we have @xmath320 while we choose as a good estimate of @xmath161 the intersection point of the two power laws of fig .
[ beta_resc_q ] , obtaining @xmath321 ps ; if we put @xmath322 and @xmath323 we finally obtain the factors @xmath324 and @xmath325 for @xmath42 and @xmath43 respectively .
unfortunately these value are not able to correctly rescale our data on the experimental results , the correct values being @xmath192 and @xmath326 as shown in fig .
[ h_c1_c2 ] ; this result could be expected considering the great uncertainties on the parameter values used for the estimate .
nevertheless , simulated data agree quite well with the experimental points at low-@xmath176 presenting a strong bending toward a constant value in the region @xmath327 nm@xmath209 . to complete the picture of the self - motion in our model , we test the validity of the _ gaussian approximation _ to @xmath256 in the limit of small momentum @xmath176 . the first order term of the expansion of @xmath256 in powers of @xmath328 gives @xcite @xmath329 in fig .
[ gauss_approx ] some curves at @xmath265 k and different values of @xmath176 are shown together with the corresponding approximations ; such approximation seems to work quite good and it becomes worse on increasing @xmath176 as expected . at the end , we show in fig .
[ tau_global ] all the time scales related to centers of mass motion investigated up to now as a function of temperature .
full circles and squares indicate respectively the experimental structural relaxation time @xmath330 , following the vogel - fulcher law , and the experimental inverse self - diffusion coefficient @xmath331 multiplied by a factor @xmath332 cm@xmath293 in order to superimpose to @xmath330 .
the open symbols are used to represent the molecular dynamics results : @xmath334 ( diamonds ) is the relaxation time of the one - particle dynamics at @xmath313 nm@xmath209 multiplied by a scale factor @xmath335 and @xmath336 ( triangles up ) is the inverse of the diffusion coefficient rescaled by the same factor @xmath332 cm@xmath293 we used for the experimental points .
all the molecular dynamics points have been shifted of @xmath133 k with respect to the measured temperatures ; it is quite clear that both experimental and molecular dynamics data points collapse on a well defined master curve .
our model is , at least , a good model for centers of mass dynamics of otp .
in this paper we have introduced a new interaction potential model capable to describe the intramolecular dynamics of the fragile glass former otp ; such model appears to be much more efficient with respect to the ones introduced so far in the sense that it represents a much more better compromise between the resulting computing needing and its capability to mimic all the complexity of the dynamical behavior of the real system .
it takes into account not only the translational and rotational dynamics of the molecules as a whole but also the stretching along the molecular bonds , the tilt of the bonds , the rotations of the side rings with respect to the parent ring .
it is tuned in such a way to reproduce the isolated molecule vibrational spectrum . in this way ,
most likely , we have introduced the degrees of freedom whose interplay causes the complex dynamical behavior of the real system .
we have , then , presented the results of molecular dynamics computer simulations of such model ; we mainly studied the static structure of a bulk sample , the self diffusion properties and the self part of the density - density correlation functions .
the static structure factor simulated in such a way is compared with the experimental measures and shows a good agreement with the neutron scattering data except for the very low momenta behavior ( due , probably , to the finite size of our system ) .
moreover we have no evidence of the splitting of the main peak in two subpeak placed at @xmath337 nm@xmath209 , only the first one being clearly visible .
this luck may be due mainly to the temperature range investigated ( the intensity of the second sub - peak increases with lowering temperature ) .
the self - diffusion properties of the system have been investigated trough the mean squared displacement and the self diffusion coefficient temperature behavior ; comparisons with experimental self - diffusion data give a very good agreement , showing the evidence of compatible critical dynamics behavior in approaching the instability temperature @xmath140 of mct which is here find to be @xmath338 to be compared with the experimental value @xmath241 determined by a mct analysis of the dynamics of the density fluctuations , a discrepancy that is most to likely ascribed to the intermolecular lj potential parameters @xmath339 that have been tuned in the temperature region close to @xmath188 k. moreover we considered the critical temperature dependence of the so - called non - gaussianity parameter @xmath251 obtaining compatible values for the power law parameters .
the self dynamics of the density fluctuations has been studied in great detail on the whole accessible time window , spanning the range from time scale of the order of few fs to times of order of some ns ; moreover its dependence on temperature and momenta has been investigated .
all the correlation curves calculated show the typical two step behavior predicted by mct , the first one on short time being associated to the so - called `` microscopic '' processes , i.e. the vibrational motion of molecules in the cage built up by their neighbors ; the second one being associated to the @xmath139-process which controls the structural rearrangements of molecules on long time scale .
the critical dynamics on @xmath139 time scale approaching the correspondent @xmath140 is in good agreement with experimental findings ; indeed the estimate values for our model of the exponent @xmath312 must be compared with the experimental value @xmath340 .
the q - dependence at @xmath188 k in the momentum range @xmath341 nm@xmath209 has been analyzed in terms of a stretched exponential fit ; the values of the determined parameters are in good agreement with the ones calculated by fitting the experimental points and with the mct expected behavior . to summarize , in the present work we have shown that our model for otp fluid is mimicking rather well the center of mass dynamical features of the real system , giving results in most cases fully compatible with the experimental findings .
it is clear that its skill to help us in the understanding of the most exotic dynamic features of the real systems and , in particular , of the relevance of the internal degrees of freedom on the translational dynamics , has not been cleared in this paper ; problems like the collective density fluctuation behavior , the rotational dynamics , the origin of the unusual fast relaxational dynamics and many others , will be addressed in future works @xcite . the authors wish to thank w. gtze for a critical reading of the manuscript , l. bernardini and g. giuliani of the `` parco tecnologico dabruzzo '' for technical support , g. monaco and f. sciortino for some useful discussions .
w. gtze , in _ liquids , freezing and the glass transition _ , edited by j. p. hansen , d. levesque and j. zinn - justin ( north - holland , amsterdam , 1991 ) ; w. gtze and l. sjrgen , rep .
. phys . * 55 * , 241 ( 1992 ) ; w. gtze , j. phys . : condens
. matter , * 11 * , a1 ( 1999 ) .
r. schilling , in _ disorder effects on relaxational processes _ edited by a. richert and a. blumen ( springer verlag , 1994 ) ; w. kob , in _ experimental and theoretical approaches to supercool ed liquids : advances and novel applications _ edited by j. fourkas _ et al .
_ ( acs books , washington , 1997 ) .
g. wahnstrm and l. j. lewis , physica a * 201 * , 150 ( 1993 ) ; l. j. lewis and g. wahnstrm , solid state comm . * 86 * , 295 ( 1993 ) ; l. j. lewis and g. wahnstrm , j. of non - crystalline solids * 172 - 174 * , 69 ( 1994 ) ; l. j. lewis and g. wahnstrm , phys .
e * 50 * , 3865 ( 1994 ) . c. m. roland , k. l. ngai and l. j. lewis , j. chem .
phys . * 103 * , 4632 ( 1995 ) ; c. m. roland and k. l. ngai , j. chem . phys . * 106 * , 1187 ( 1997 ) ; f. sciortino and p. tartaglia , j. phys . : condens
. matter , * 11 * , a261 ( 1999 ) . 1
. values of the internal potential coefficients determined by a least square minimization of the error function , eq .
( [ chiquadro ] ) .
the coefficients @xmath343 are the force constants associated to the different intramolecular potential terms ; the @xmath344 , derived from the @xmath345 via the eq .
( [ bval ] ) , describe the form of the in - phase rotational potential according to eq .
( [ vr1 ] ) . [ coeff ] 2 .
free molecule vibrational frequencies : in the second column are reported the values @xmath346 determined by a harthree - fock calculation of the ground state of the isolated molecule ( we consider only the @xmath347 lowest eigenvalues ) ; in the third column are shown the frequencies @xmath94 corresponding to the different peaks of the spectrum determined from the atomic velocity autocorrelation function calculated by means of a preliminary molecular dynamics simulation of the isolated molecule ( @xmath348 ) at @xmath349 k. [ frequencies ] 3 .
simulation runs details : at each temperature @xmath128 , the system is coupled for a time @xmath129 to a stochastic heat - bath , then it evolvs freely , at constant energy , for a time @xmath130 . at the end of this process , we consider the final configuration to be in a good equilibrium state and we start a trajectory @xmath137 ps long , saving a system configuration every @xmath138 ps . [ protocol ] 4 . some thermodynamical results : temperatures @xmath128 effectively measured , total energy @xmath350 , total potential energy @xmath351 and kinetic energy @xmath128 . [ thermotab ] 5 . temperature dependence of the molecular dynamics self - diffusion coefficient @xmath254 .
[ diff_tab ] 1 . molecular structure of otp ( @xmath352 ) ; it is constituted by three phenyl rings , the two side rings being attached to the parent ( i.e. central ) ring by covalent bonds . [ otp ] 2 .
model geometry : each phenyl ring is represented by a rigid hexagon of side @xmath5 nm and the equilibrium bond lenght is @xmath353 nm . @xmath354 , @xmath355 , @xmath356 represent the origins of the reference frames fixed with the rings ; @xmath357 , @xmath358 , @xmath359 , @xmath360 are the vectors parallel to the ring bonds ; @xmath361 and @xmath362 are two versors identifying the parent ring plane ; @xmath363 , @xmath364 , @xmath25 , @xmath26 are the positions of the carbon atoms bonding together the rings ; @xmath365 , @xmath366 , @xmath57 , @xmath367 are four interaction sites introduced to force the rings towards the co - planar equilibrium condition ; the symbols @xmath368 and @xmath369 have been introduced to identify the angles @xmath46 and @xmath47 .
[ geometry ] 3 .
rotational energy : each point has been determined by an _ ab - initio _ hartree - fock calculation of the potential energy surface as a function of the rotational angles @xmath370 and @xmath371 ( expressed in degrees ) with all the other degrees of freedom fixed to their equilibrium values ; the energy is expressed in hartree units ( 1 hartree=27.2ev ) .
we quote by a the isolines ranging from @xmath372 to @xmath373 mhartree with an increment @xmath374 mhartree , by b the ones from @xmath375 to @xmath376 mhartree with @xmath377 mhartree and by c those from @xmath306 to @xmath226 mhartree with @xmath378 mhartree .
this figure has to be considered only from a semi - quantitative point of view as explained in the text .
[ map ] 4 .
temperature dependence of the energies tabulated in tab .
[ thermotab ] , @xmath350 ( circles ) @xmath351 ( squares ) and @xmath128 ( triangles up ) together with the internal @xmath379 ( triangles down ) and the intermolecular lennard - jones @xmath380 ( diamonds ) contributions to @xmath351 .
[ thermo ] 5 .
pair static distribution functions at @xmath188 k calculated on atoms ( a ) , ring centers of mass ( b ) and molecular centers of mass ( c ) ; full lines represent the total contribution of both intramolecular and intermolecular distances , dashed lines only the intermolecular contribution .
[ gdr ] 6 .
static distribution functions of the scalar products @xmath197 ( triangles ) , @xmath196 ( circles ) , and @xmath198 ( squares ) evaluated at @xmath381 k. [ distrang ] 7 .
intensity of the peaks of the static distribution function of @xmath198 at @xmath382 k ( from bottom to top ) .
a blow - up of this figure around @xmath202 is shown in fig .
[ distrnorm ] .
[ distrnorm2 ] 8 .
temperature dependence of the static distribution function of @xmath198 near the saddle point position @xmath205 at @xmath383 k ( from top to bottom ) .
[ distrnorm ] 9 .
static structure factors at @xmath384 calculated on on the molecular ( solid line ) and ring centers of mass ( dashed line ) ; each @xmath176 point is the average over all the independent miller indeces corresponding to it .
[ sk_cm_rng ] 10 . : comparison among molecular dynamics structure factor ( solid line ) calculated taking in account both carbon and hydrogen atoms as scattering centers and experimental structure factor ( dashed line ) measured by neutron scattering ( from @xcite ) .
_ bottom _ : molecular dynamics structure factor ( solid line ) calculated taking in account only carbon atoms ; this should be the correct result to be compared with experimental structure factor measured by x - ray scattering .
[ essekexp ] 11 .
enlargement of the low - momenta region of @xmath177 calculated at @xmath381 k : the error bars are estimated from the fluctuation of the single configuration @xmath177 s .
[ sk_err ] 12 .
temperature dependence of the mean square displacement @xmath385 calculated on ring centers of mass at all temperature investigated except @xmath386 k ( higher temperature on top ) .
inset : linear scale plot of the mean square displacement at some selected temperatures ( open symbols ) together with the long time linear behavior ( dashed lines ) .
[ erre2 ] 13 .
temperature dependence of molecular dynamics ( open circles ) and experimental ( full triangles ) diffusion coefficients together with the power law fits in the form of eq .
[ power ] ( solid and dashed lines respectively ) ; we show also the md data shifted of @xmath133 k ( open squares ) as explained in the text .
[ diff_fit ] 14 .
time dependence of the non - gaussian parameter @xmath242 for all temperature investigated ( lower temperatures on top ) .
[ alpha2 ] 15 .
power law fit of the temperature dependence of the position @xmath387 of the maximum of the non - gaussian parameter .
[ alpha_fit ] 16 .
temperature dependence of @xmath262 calculated at @xmath313 nm@xmath209 for all temperatures investigated except @xmath388 k ( lower temperatures on top ) .
[ f_self ] 17 .
temperature dependence of the stretching parameter @xmath178 ( circles ) and of the non - ergodicity parameter @xmath389 ( squares ) ; the horizontal lines indicate the mean values of @xmath390 ( dashed line ) and @xmath389 ( dot - dashed line ) .
[ beta_err ] 18 .
temperature dependence of @xmath131 at @xmath337 nm@xmath209 ( circles and squares respectively ) together with the power law fits with @xmath391 k and @xmath392 ( solid and dashed lines respectively ) ; also the experimental shear viscosity @xmath393 data ( full triangles ) are reported ( see @xcite and reference therein ) multiplied by a factor @xmath335 ps / poise .
molecular dynamics results have been shifted of @xmath133 k with respect to the measured temperatures , as explained in text .
[ tau_err ] 19 .
@xmath256 at @xmath313 nm@xmath209 rescaled to @xmath282 ; all the curves verify the time - temperature superposition principle .
[ rescaling ] 20 .
q - dependence of @xmath262 at @xmath188 k for @xmath394 nm@xmath209 with @xmath395 ( from top to bottom ) .
[ f_self_q ] 21 .
q - dependence of the stretching and non - ergodicity parameters : _ left _ : molecular dynamics ( circles ) and experimental ( filled circles , from @xcite ) values of the coefficient @xmath390 as determined by the kww fits ; _ right _ : experimental values ( filled squares ) of the non - ergodicity parameter @xcite together with the molecular dynamics results as determined by a mct analysis of both @xmath139 ( squares ) and @xmath1 ( triangles ) regions and the gaussian fit ( solid line ) to @xmath139-region results with @xmath396 nm@xmath397 .
[ fit1_q ] 22 .
q - dependence of the inverse relaxation time @xmath289 ; molecular dynamics data ( open circles ) have been multiplied by a factor @xmath398 in order to overlap the experimental @xcite data ( full circles ) , as explained in the text .
the solid line is the correct small-@xmath176 behavior @xmath291 where @xmath292 nm@xmath293/ps . [ fit2_q ] 23 . :
temperature dependence of the product @xmath399 at @xmath400 nm@xmath209;_right _ : q - dependence at @xmath188 k of @xmath401 .
these quantity are expected to be constant .
[ tau_d ] 24 .
power laws in the @xmath1-region for @xmath402 nm@xmath209 ; critical decay of exponent @xmath163 ( dot - dashed lines ) and von scweidler law of exponent @xmath164 ( dashed lines ) have been reported with simulation points .
[ beta_resc_q ] 25 .
q - dependence of the power laws exponents : _ left _ : the experimental values ( filled circles ) for @xmath314 are plotted together with the molecular dynamics values for @xmath314 ( open circles ) and @xmath403 ( open squares ) ; the mean values for our results are shown by the dashed and dott - dashed line respectively .
_ right _ : our values for the q - dependence of the exponent @xmath173 determined by the relation @xmath404 .
[ beta_fit_q ] 26 .
q - dependence of the experimental ( full circles ) coefficient @xmath405 of eq .
[ betaregion ] and of our fitting parameters @xmath42 ( triangles ) @xmath43 ( squares ) ; note that our results have been rescaled by a factor @xmath192 and @xmath406 respectively in order to superimpose to experimental data .
[ h_c1_c2 ] 27 .
gaussian approximation to @xmath262 at @xmath265 k for @xmath407 nm@xmath209 ( from top to bottom ) .
[ gauss_approx ] 28 .
master plot of the temperature dependence of all the centers of mass time scales discussed : we have used full symbols for experimental results and open symbols for molecular dynamics data .
molecular dynamics points collapse exactly on the master curve identified by the experimental data if they are multiplied by a scale factor ( taking into account the momentum dependencies of the relaxation time @xmath334 and the correct dimensionality of the diffusion coefficients ) and the corresponding temperatures are shifted of @xmath133 k above the measured ones , as discussed in the text . in particular @xmath408 ( open diamonds )
is multiplied by a factor @xmath335 , the self diffusion coefficients @xmath409 ( full squares ) and @xmath410 ( open triangles ) are rescaled by a factor @xmath332 cm@xmath293 .
[ tau_global ] | we present a realistic model of the fragile glass former orthoterphenyl and the results of extensive molecular dynamics simulations in which we investigated its basic static and dynamic properties . in this model
the internal molecular interactions between the three rigid phenyl rings are described by a set of force constants , including harmonic and anharmonic terms ; the interactions among different molecules are described by lennard - jones site - site potentials .
self - diffusion properties are discussed in detail together with the temperature and momentum dependencies of the self - intermediate scattering function .
the simulation data are compared with existing experimental results and with the main predictions of the mode coupling theory . |
with the advent of new data from deep optical , near - infrared ( nir ) , mid - infrared ( mir ) and x - ray surveys , a considerable interest has been evolved to study low mass young stellar objects ( ysos ) as well as to study the evolution of circumstellar disks around the ysos associated with young star clusters ( e.g. , prisinzano et al .
2011 , 2012 , guarcello et al .
2012 , pandey et al . 2008 , 2013 , haisch et al .
2001 , sicilia - aguilar et al .
2006 , carpenter et al .
2006 , chauhan et al .
2009 , 2011 ) .
most of the disk evolution studies available in the literature are based on the nir/ mir excess which traces the evolution of inner accretion disks and it is found that in majority of the cases the inner disk does not last for more than @xmath2 6 myr ( see e.g. , haish et al .
2001 ) .
most of the pre - main - sequence ( pms ) stars are t - tauri stars ( ttss ) and have masses @xmath4 3 m@xmath5 .
the ttss are further divided in to weakline ttss ( wttss ) and classical ttss ( cttss ) .
the h@xmath6 and nir excess signatures in cttss indicate the existence of a well - developed circumstellar disk actively interacting with the central star .
strong h@xmath6 emission ( ew @xmath7 10 ) in cttss is attributed to the magnetospheric accretion of the innermost disk matter on to the central star ( edwards et al .
1994 ; hartmann , hewett & calvet 1994 ; muzerolle , calvet & hartmann 2001 , and references therein ) . on the other hand ,
the weak h@xmath6 emission ( ew @xmath8 10 ) in wttss , which lack disks ( or , at least inner disks ) , is believed to originate from their chromospheric activity ( e.g. walter et al .
1988 ; martin 1998 ) . since both types of ttss ( cttss and wttss ) are strong x - ray emitters , the x - ray observations of active star - forming regions and young galactic clusters play an important role in the studies of the star formation process and the properties of young pms stars ( e.g. , feigelson & decampli 1981 , guarcello et al .
2012 and references therein ) .
in fact , the level of x - ray emission in pms stars , which is higher than that of field main - sequence ( ms ) stars , provides a very efficient mean of selecting stars associated with star - forming regions .
however , the studies based on h@xmath6 emission may fail to detect part of the wtts population , which can easily be identified in x - rays . several x - ray studies of low - mass pms stars have been carried out which revealed very strong x - ray activity , exceeding the solar levels by several orders of magnitude ( e.g. , feigelson & montmerle 1999 ; feigelson et al . 2002 ; favata & micela 2003 ; gdel 2004 ; caramazza et al 2012 ) . the influence of a circumstellar disk , and particularly the influence of accretion on x - ray activity has been of special interest .
x - ray studies of pms stars in star - forming regions have yielded contradictory results ( see e.g. , stelzer & neuhuser 2001 , feigelson et al .
2002 , flaccomio et al .
2003 , preibisch et al .
2005 , telleschi et al .
2007 , guarcello at al . 2012 ) .
the ttss show variation in their brightness in all wavelengths , from x - ray to infrared and have highly elevated levels of x - ray activity , however the relation between rotation and x - ray activity in ttss remained unclear until recently .
stelzer & neuhuser ( 2001 ) have found that in the case of taurus - auriga region , pleiades and hyades , the rotation and x - ray emission are clearly correlated in the sense that faster rotators are more luminous .
however , majority of the results do not find any correlation between rotation and x - ray activity ( e.g. , flaccomio et al .
2003 , rebull et al .
2006 , alexander & preibisch 2012 ) . a number of data sets dealing with the observational relationship between rotation and magnetic activity for young stellar clusters with ages as young as 30 myr ( e.g. , prosser et al . 1996 ; stauffer et al .
1997 ; jeffries et al .
2011 ) are available , however the data is still lacking for younger clusters having ages @xmath4 5 myr . in view of the contradictory results
as discussed above regarding the effect of the disk on the x - ray luminosity of cttss and wttss as well as the lack of observational studies on the relationship between rotation and the x - ray properties , we feel that the young cluster ngc 1893 having an age of @xmath2 2 - 4 myr and located at a distance of 3.25 kpc ( sharma et al .
2007 , pandey et al .
2013 , hereafter pa13 ) , is very well - suited in this respect .
the cluster , located at the center of the aur ob2 association , is associated with the region ic 410 .
the cluster contains at least five o - type stars along with a rich population of pms stars ( see e.g. pa13 ) .
several photometric and low resolution spectroscopic studies of the region have been carried out ( for details see sharma et al .
2007 , prisinzano et al . 2011 : hereafter p11 ) which manifest that star formation is still ongoing in the region and therefore ngc 1893 is an ideal object to study the x - ray properties of the pms stars .
caramazza et al .
( 2012 ) analyzed the x - ray properties of the class ii and class iii sources identified by p11 in the ngc 1893 region .
they found that class iii stars appear intrinsically more x - ray luminous than class ii stars and concluded that this effect may be due to the presence of the magnetically connected disk itself or to the ongoing accretion from the disk to the star .
however , in pa13 we have shown that the sample of the class ii and class iii ysos in ngc 1893 region identified by p11 on the basis of _ spitzer _ mir data was contaminated by the field stars .
chauhan et al .
( 2011 ) have also found a significant amount of contamination in the sample of ysos identified by koenig et al .
( 2008 ) on the basis of _ spitzer _ data in the w5 e region .
pa13 and chauhan et al . (
2011 ) used nir @xmath9 color - color ( nir - cc ) diagram to avoid the contamination due to field star population in the yso sample selected on the basis of _ spitzer _ data .
hence it is considered worthwhile to re - analyze the x - ray properties of class ii and class iii sources and the relation between rotation and x - ray activity in ttss using the sample of the ysos identified by pa13 and the rotation period of the ysos determined by lata et al .
( 2012 , 2013 ) .
on the basis of optical , nir , mir and x - ray data , p11 identified 1034 and 442 class ii and class iii ysos , respectively .
however , in our previous study ( pa13 ) we have shown that the sample of ysos by p11 is significantly contaminated by the non - pms stars .
pa13 have removed contamination in the sample of p11 using the nir - cc diagram .
they found 367 and 246 class ii and class iii sources , respectively with optical counterparts .
the @xmath10 color - magnitude diagram ( cmd ) of the 367 and 246 class ii and class iii sources ( figure 14 of pa13 ) reveals that the contamination due to field stars is greatly reduced , however still we noticed a few sources classified as class ii sources by p11 below the 5 myr pms isochrone . to further avoid the contamination due to background stars , we have considered only those sources as the pms objects associated with ngc 1893 which are located above the 5 myr isochrone ( cf .
figure 14 of pa13 ) .
this selection yields 298 and 240 class ii and class iii sources , respectively .
this sample is used to study the x - ray properties of the ysos in the ngc 1893 region .
nir - cc diagram of class ii and class iii sources used in the present study is shown in fig .
1 , where the thin and thick dashed curves represent the unreddened ms and giant branches ( bessell @xmath11 brett 1988 ) , respectively . the dotted line indicates the locus of intrinsic cttss ( meyer et al .
1997 ) . the curves and the data are in the california institute of technology ( cit ) photometric system .
the parallel dashed lines are the reddening vectors drawn from the tip of the giant branch ( left reddening line ) , from the base of the ms branch ( middle reddening line ) and from the base of the tip of the intrinsic cttss line ( `` right reddening line '' ) .
the extinction ratios @xmath12 and @xmath13 have been adopted from cohen et al .
following ojha et al .
( 2004 ) we classify nir - cc diagram into three regions .
the sources lying in the ` f ' region ( located between the left and middle reddening lines ) could be either field stars ( ms stars , giants ) or class iii and class ii sources with small nir excesses .
the ` t ' region , located between the middle and right reddening lines contains mostly cttss ( class ii objects ) .
there may be an overlap in nir colors of herbig ae / be stars and cttss in the ` t ' region ( hillenbrand et al .
the ` p ' region is redward of the ` t ' region and the sources lying in this region are likely class i objects .
the age and mass of each yso have been estimated by pa13 ( cf . their table 1 ) using the @xmath14 cmd .
the effect of random errors due to photometric errors and reddening estimation in determination of ages and masses were estimated by propagating the random errors to their observed estimation by assuming normal error distribution and using the monte - carlo simulations . the x - ray data has been taken from caramazza et al .
x - ray luminosity distribution of class ii ( cttss ) and class iii ( wttss ) sources used in the present study , for the mass range [email protected] , is shown in fig .
2 which shows a similar distribution for both the samples .
the bias due to low mass incompleteness was checked by plotting the distribution for the mass range [email protected] and we again found the same distribution for both the samples indicating that the class ii and class iii sources have indistinguishable x - ray properties .
our previous study ( pa13 ) using the present sample has also shown that the cttss and wttss are coeval and have indistinguishable optical properties .
in fig . 3 we plot the x - ray luminosity , @xmath0 , as a function of the stellar mass .
the cttss ( class ii ) and wttss ( class iii ) are plotted with filled and open circles , respectively . an increasing trend with a significant scatter in the @xmath0 with stellar mass is apparent for the entire sample of the ttss .
a linear regression fit to the function log(@xmath0 ) = @xmath15 , in the mass range [email protected] , yields @xmath16 = 30.71@xmath17 0.04 , 30.68 @xmath17 0.06 and 30.74 @xmath17 0.05 and @xmath18 = 1.16 @xmath17 0.10 , 1.14 @xmath17 0.15 and 1.13 @xmath17 0.13 , respectively , for the whole sample of the ttss , cttss and wttss .
however , a careful inspection of fig .
3 reveals that the lower envelope of the distribution in the mass range 0.2 @xmath4 m/@xmath19 @xmath4 0.35 is rather flat .
the sources around lower envelope of 0.2 @xmath4 m/@xmath19 @xmath4 0.35 have log(@xmath0 @xmath20 29.7 ) .
this may indicate a bias in the sample in the sense that only x - ray luminous low mass sources might have been detected .
a linear regression fit in the mass range [email protected] yields @xmath16 = 30.69@xmath17 0.05 , 30.65 @xmath17 0.08 and 30.72 @xmath17 0.06 and @xmath18 = 1.05 @xmath17 0.16 , 0.99 @xmath17 0.28 and 1.07 @xmath17 0.20 , respectively , for the whole sample of the ttss , cttss and wttss .
the results for two mass ranges are almost same within error .
the sample of class ii sources ( cttss ) has a significant number of sources with upper limit . to account for upper limits of nondetections we used asurv package as described by preibisch and feigelson ( 2005 ) .
the values of the slope @xmath18 in the mass range [email protected] , comes out to be @xmath18 = 0.88@xmath17 0.11 , 0.51 @xmath17 0.20 and 1.13 @xmath17 0.13 , respectively , whereas the values of @xmath16 are found to be 30.73@xmath17 0.04 , 30.71 @xmath17 0.07 and 30.74 @xmath17 0.05 , respectively , for the whole sample of ttss , cttss and wttss .
the correltion is found to be @xmath21 with a probability of correlation greater than 99.9% in all the cases . in the case of taurus molecular cloud ( tmc ) ,
telleschi et al . (
2007 ) have found @xmath16 = 30.33 @xmath17 0.06 ; 30.13 @xmath17 0.09 ; 30.57 @xmath17 0.09 and @xmath18 = 1.69 @xmath17 0.11 ; 1.70 @xmath17 0.20 ; 1.78 @xmath17 0.17 , respectively for the whole ttss , cttss and wttss . in the case of onc ,
preibisch et al .
( 2005 ) also found a linear regression for the whole ttss sample ( m @xmath4 2 ) with @xmath16 = 30.37 @xmath17 0.06 and @xmath18 = 1.44 @xmath17 0.10 , whereas feigelson et al .
( 2003 ) reported @xmath16 = 30.33 and @xmath18 = 1.5 for the ttss samples ( m @xmath4 3 ) in the onc .
the value of power - law slope @xmath18 for the ttss in the chamaeleon star - forming region in the mass range 0.6 - 2 is reported to be @xmath18 = 3.6 @xmath17 0.6 ( feigelson et al . 1993 ) and in the very young stellar cluster ic 348 , the slope @xmath18 in the mass range 0.1 - 2 is found to be 1.97 @xmath17 0.2 ( preibisch & zinnecker 2002 ) . in the case of ngc 2264 ,
the linear fit for the detected sources yields @xmath16 = 30.6 @xmath17 0.3 and @xmath18 = 0.8 @xmath17 0.1 ( dahm et al . 2007 ) .
however , taking into account the x - ray non - detections the slope is found to be steeper ( @xmath18 = 1.5 @xmath17 0.1 ) .
the values of the power - law slopes reported in the case of tmc , onc , ic 348 and chamaeleon star - forming regions are higher than that obtained in the present study for ngc 1893 .
one possible reason for the difference could be the bias in the sample , i.e. the present sample may have only x - ray luminous low mass stars , however we have shown above that the slope @xmath18 does not seem to be affected by the bias in the sample . recently
, guarcello et al . ( 2012 ) for class iii stars having mass m @xmath4 0.8 in ngc 6611 have found @xmath22 and @xmath23 , whereas the distribution of stars more massive than 0.8 m@xmath5 is flatter with a slope of 0.4 @xmath17 0.2 .
the intercept of class iii sources at 1 m@xmath5 ( log@xmath24 = 0 ) has comparable value to that of class ii sources indicating that the presence of circumstellar disks has no influence on the x - ray emission .
this result is in agreement with that reported by feigelson et al .
( 2002 ) and is in contradiction with those reported by stelzer & neuhuser ( 2001 ) , preibisch et al .
( 2005 ) and telleschi et al .
caramazza et al .
( 2012 ) on the basis of x - ray luminosity functions ( xlfs ) have concluded that cttss in ngc 1893 are globally less x - ray active than the wttss . as we have already shown in our previous study ( pa13 ) that the sample of class ii sources used by p11 is strongly contaminated by field population , the xlf of class ii sources by caramazza et al .
( 2012 ) may not be representing the true class ii xlf .
the age distribution of identified ysos in ngc 1893 reveals a non - coeval star formation in the cluster ( e.g. pa13 ) .
keeping the caveats in mind ( e.g. age determination is model dependent , systematic errors due to presence of unresolved binaries etc . )
as discussed by preibisch & feigelson ( 2005 ) as well as probable bias in the sample as discussed in sec 3.1 , we present evolution of x - ray emission in cttss and wttss in the mass range @xmath25 in fig .
4 . although the x - ray emission as discussed above , is correlated with mass , we could not study the evolution of x - ray emission for mass stratified sample because of limited data sample .
the x - ray luminosity for both cttss and wttss can be seen to decrease systematically with age ( in the age range @xmath3 0.4 to 5 myr ) .
a linear regression fit to the function log@xmath26 = a + b*log@xmath27 yields @xmath16 = 35.26 @xmath17 0.45 , 35.06 @xmath17 0.63 and 35.29 @xmath17 0.65 ; and @xmath18 = -0.81 @xmath17 0.07 , -0.78 @xmath17 0.10 and -0.81 @xmath17 0.11 , respectively for the whole sample of the ttss , cttss and wttss .
all the slopes are same within the errors .
we have also computed linear regression using the asurv package and found @xmath16 = 34.00 @xmath17 0.50 , 32.31 @xmath17 0.81 and 35.29 @xmath17 0.65 ; and @xmath18 = -0.58 @xmath17 0.08 , -0.29 @xmath17 0.13 and -0.81 @xmath17 0.11 , respectively , for the whole sample of the ttss , cttss and wttss .
the slopes obtained in the case of ngc 1893 are slightly steeper in comparison to those ( -0.2 to - 0.5)reported by preibisch & feigelson ( 2005 ) and telleschi et al .
( -0.36 @xmath17 011 ; 2007 ) .
this difference in the slope could be more significant in the case of an unbiased sample , where we expect more low luminous low mass x - ray sources , which would give more higher value for the slope . if x - ray luminosities of accreting pms stars are systematically lower than non - accreting pms stars ( e.g. flaccomio et al .
2003 ; preibisch et al .
2005 ) and if cttss evolve to the wttss , one might expect that in the case of cttss @xmath0 should increase with age rather than decrease .
however , the evolution of cttss up to @xmath3 5 myr ( fig .
4 ) does not show any sign of increase in x - ray luminosity .
this result seems to contradict the notion that cttss evolve to wttss .
rebull et al .
( 2006 ) have pointed out that two processes associated with disk accretion can affect the observed x - ray properties of solar - type pms stars . in the first case local heating of the stellar photosphere at the `` footprints '' of magnetospheric funnel
flows , where highly supersonic ( @xmath28 @xmath7 200 km s@xmath29 ) gas lands on the stellar surface ( e.g. , kastner et al .
2002 ; schmitt et al .
2005 ) , thus producing shocks and local heating , and as a result produces soft x - rays .
this process would qualitatively produce an increase in x - ray emission above coronal activity levels as disk accretion rates increase . in the second case absorption of kev coronal x - ray emission by gas associated with accretion - driven outflows ( walter & kuhi 1981 )
would produce a decrease of x - ray emission with increasing accretion rate , since the column density , and hence the x - ray optical depth of out - flowing material , is expected to increase with increasing accretion rate
. there has been several attempts to study the relationships between circumstellar disks , disk accretion , and x - ray emission , however the results are contradictory ( see e.g. rebull et al .
2006 and references therein ) .
kastner et al . ( 2002 ) and schmitt et al .
( 2005 ) found evidence of enhanced x - ray emission associated with transient increases in disk accretion rate .
this fact indicates that at least some x - ray emission may be associated with accretion as opposed to coronal activity .
majority of the results , however , available in the literature suggest that there is no evidence of excess x - ray emission for accreting pms stars .
rather , there seems to be anti - correlation between the presence of an accretion disk and total x - ray luminosity ( e.g. , rebull at al .
2006 ) .
walter & kuhi ( 1981 ) reported an inverse correlation between soft x - ray flux and h@xmath6 emission for cttss . assuming that the observed equivalent width of h@xmath6 serves as a surrogate for mass outflow rate ,
walter & kuhi ( 1981 ) argued that the apparent decrease of x - ray emission with increasing h@xmath6 equivalent width reflected the effects of x - ray absorption by outflowing gas .
flaccomio et al .
( 2003 ) also found that a larger fraction of stars with active accretion signatures have fainter @xmath0 values than non - accretors .
however , dahm et al . (
2007 ) , in the case of ngc 2264 , do not find any correlation between h@xmath6 equivalent width and x - ray luminosity of ttss .
the nir excess ( e.g. ( h - k ) excess ) is a useful indicator for the presence of disk , hence can be used to look for correlations of x - ray emission with the presence / absence of accretion disks for which there is no direct evidence of accretion . on the basis of nir excess stars , feigelson et al .
( 2002 , 2003 ) , rebull et al . (
2006 ) , dahm et al .
( 2007 ) have concluded that the presence or absence of infrared excess , has no discernible effect on the distribution of x - ray luminosities .
to study the effect of disk on the x - ray emission we define nir excess @xmath30 as the horizontal displacement from the left reddening vector of ` f ' region in the nir - cc diagram as described in fig .
1 . in fig .
5 we plot @xmath0 of cttss ( filled circles ) and wttss ( open circles ) as a function of nir excess @xmath30 .
although the scatter is large , however fig .
5 manifests that cttss having relatively larger @xmath31 0.4 - 0.5 mag have smaller value of log @xmath0 ( @xmath2 30 ) , whereas the cttss having @xmath30 @xmath32 0.1 mag have log @xmath0 values up to @xmath2 31.5 .
it is interesting to point out that the upper bound for the class ii sources ( cttss ) for @xmath33 appears to decrease with increasing value of @xmath30
. a similar trend has been noticed by rebull at al .
( 2006 ) in the @xmath34 excess versus log @xmath0 plot of ngc 2264 .
this may be indicative of that the sources having relatively large nir excess have relatively lower @xmath0 values .
however , feigelson et al .
( 2003 ) on the basis of @xmath35 excess in the case of orion nebula cluster do not find any correlation between x - ray emission and the presence of disk .
the distribution of wttss in fig .
5 does not show any obvious correlation .
since @xmath36-band data is not available for these ttss , it is not possible to estimate the mass accretion rate for the cttss studied here .
sicilia - aguilar et al .
( 2005 ) have presented nir data and mass accretion rate for a few ttss of young cluster tr 37 .
the data reveals a positive correlation between nir excess @xmath30 and mass accretion rate .
assuming that the same correlation exists in the case of ngc 1893 , we can presume that the mass accretion rates are inversely correlated with x - ray luminosity .
both wttss and cttss show variation in their brightness .
these variations are found to occur at all wavelengths , from x - ray to infrared .
the variability time - scale of ttss ranges from a few minutes to years ( appenzeller & mundt 1989 ) .
several mechanisms such as circumstellar gas and dust ( remnant of parent molecular cloud ) , accretion and magnetic field have been suggested to explain the photometric variations ( herbst et al .
the variations in the brightness of ttss are most probably due to the presence of spots ( cool or hot ) on stellar surface and circumstellar disk .
the cool spots on the surface of the stars are produced by the emergence of stellar magnetic fields on the photosphere , and are thus indicators of magnetic activity .
the cool spots on the photosphere rotate with stars hence are responsible for brightness variation in wttss .
these objects are found to be fast rotators as they have either thin or no circumstellar disk .
the hot spots on the surface of young stars are the consequence of accretion process ( lynden - bell & pringle 1974 ; koenigl 1991 ; shu et al .
1994 ) . irregular or non - periodic variations are produced because of changes in the accretion rate .
the time - scales of varying brightness range from hours to years .
the hotspots cover a smaller fraction of the stellar surface but higher temperature causes larger brightness variations ( carpenter , hillenbrand & skrutskie 2001 ) .
the accreting cttss show a complex behaviour in their optical and nir light curves ( scholz et al .
2009 ) .
although the ysos in all evolutionary stages , from class i protostars to ttss , have highly ( @xmath2 10@xmath37 - 10@xmath38 times ) elevated levels of x - ray activity , the physical origin of this x - ray activity remains poorly understood . however , there is strong evidence that in most ttss the x - ray emission originates from magnetically confined coronal plasma ( e.g. , preibisch et al .
2005 ) , it is still not clear about the dynamo processes that create the required magnetic fields .
it is well established that the level of the magnetic activity , i.e. , the strength of the x - ray emission in ms stars is mainly determined by their rotation rate ( e.g. , gdel , guinan , & skinner 1997 ; gaidos 1998 ; pallavicini et al .
1981 ; pizzolato et al .
the relation between rotation and x - ray activity in ttss remained unclear until recently , since in most studies of star - forming regions the number of x - ray detected ttss with known rotation periods was too small to draw any conclusion . the chandra orion ultradeep project ( coup , getman et al .
2005 ) and the xmm - newton extended survey of the taurus molecular cloud ( xest , gdel et al . 2007 ) provided very sensitive x - ray data sets for large samples of ttss .
these data revealed that the ttss in orion as well as in taurus do not follow the relation between rotation period and x - ray luminosity for ms stars ( preibisch et al .
2005 ; briggs et al . 2007 ) .
in addition the ttss spin up during the first @xmath2 10 - 30 myr ( herbst & mundt 2005 ) does not lead to an increase in the x - ray luminosity .
this places doubt on the solar - like dynamo activity scenario for ttss ( alexander & preibisch , 2012 ) . in fig .
6 we plot x - ray luminosity l@xmath39 as a function of rotation period .
the data of rotation period for 38 and 10 ttss have been taken from lata et al .
( 2012 ) and lata et al .
( 2013 , in preparation ) , respectively .
the plot manifests that the ttss in ngc 1893 do not follow the well established x - ray activity - rotation relation as in the case of main - sequence stars , i.e. increase in x - ray activity with decreasing rotation period .
the least - squares fit to the data points in fig .
6 yields a slope of @xmath40 which indicates that the x - ray activity for the ttss in ngc 1893 seems to be independent of the rotation period .
flaccomio et al .
( 2003 ) found for the first time that ttss in onc having rotation period show higher x - ray luminosities than ttss do not have known rotational periods .
the same has been found in the case of ngc 2264 ( rebull et al .
7 shows x - ray luminosity distribution of ttss having rotation periods as well as ttss without data for rotation periods .
the ttss with rotation periods show a uniform distribution for the x - ray luminosity whereas remaining tts sample peaks at log @xmath41 30 .
on the basis of a comprehensive multi - wavelength study of the star forming region ngc 1893 , we continued our attempt to understand the properties of the stellar content in the young cluster ngc 1893 .
this paper is focused on understanding the x - ray properties of ttss in the ngc 1893 region .
the main results of the present work are as follow : \1 ) we found a correlation between x - ray luminosity , @xmath0 , and the stellar mass ( in the range [email protected] ) of ttss of ngc 1893 , as already has been reported by several authors , however the values of the power - law slopes ( @xmath2 3.6 - 1.4 ) reported in the case of tmc , onc , ic 348 and chameleon star forming regions are higher than that obtained in the present study ( @xmath42 ) for ngc 1893 .
the slope ( @xmath43 ) of the distribution of class iii sources in the case of ngc 1893 is found to be comparable to that reported in the case of ngc 6611 .
\2 ) the intercept of class iii sources at 1 ( log@xmath24 = 0 ) in the present sample has comparable value to that of class ii sources indicating that the presence of circumstellar disks has no influence on the x - ray emission .
this result is in agreement with that reported e.g. , by feigelson et al .
( 2002 ) and is in contradiction with those reported by stelzer & neuhuser ( 2001 ) , preibisch et al .
( 2005 ) and telleschi et al .
( 2007 ) . on the basis of xlfs caramazza et al .
( 2012 ) have concluded that cttss in ngc 1893 are globally less x - ray active than the wttss .
however , we have to keep in mind that pa13 have already shown that the sample of class ii sources by caramazza et al .
( 2012 ) is strongly contaminated by field population .
\3 ) the x - ray luminosity for both cttss and wttss is found to decrease systematically with age ( in the range @xmath3 0.4 myr - 5 myr ) .
the decrease of the x - ray luminosity ( slope @xmath2 -0.6 ) in the case of ngc 1893 is faster than observed in the case other star forming regions ( e.g. , onc : slope -0.2 - 0.5 , preibisch & feigelson 2005 ; tmc : @xmath44 @xmath17 @xmath45 , telleschi et al .
2007 ) .
\4 ) it is found that the upper bound of the distribution of @xmath30 vs log @xmath0 for the class ii sources ( cttss ) having @xmath30 @xmath46 @xmath47 appears to decrease with increasing value of @xmath30 .
a similar trend has been noticed by rebull at al .
( 2006 ) in the ( i - k ) excess versus log @xmath0 plot of ngc 2264 .
this may be indicative of that the sources having relatively large nir excess have relatively lower @xmath0 values .
the x - ray properties of the class ii and class iii sources of the present sample are rather same because a significant number of class ii sources have weak nir excess ( cf .
present result suggests that cttss having higher nir excess ( an indication of the presence of inner disk ) could have lower x - ray luminosity because of higher extinction due to x - ray absorption by circumstellar disks .
\5 ) ttss in ngc 1893 like ttss in other young clusters ( e.g. , in onc : flaccomio et al .
2003 , ngc 2264 : rebull et al .
2006 , ic 348 : alexander & preibisch 2012 ) do not follow the well established x - ray activity - rotation relation as in the case of main - sequence stars .
present result is in contrast with stelzer & neuha.user ( 2001 ; taurus - auriga region , pleiades and hyades region ) and preibisch et al .
( 2005 ; orion region ) where an anti - correlation and correlation , respectively , was reported between the x - ray luminosity and rotation period .
akp is thankful to prof .
b. soonthornthum director , narit for encouraging the collaboration between aries ( india ) and narit ( thailand ) .
this publication makes use of the data products from the 2mass , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology .
99 alexander f. , preibisch t. , 2012 , a&a , 539 , 64 appenzeller i. , mundt r. , 1989 , a&ar , 1 , 291 bessell m.s . , brett j.m .
, 1988 , pasp , 100 , 1134 briggs , k. r. et al . , 2007 , a&a , 468 , 413 caramazza m. , micela g. , prisinzano l. , sciortino s. , damiani f. , favata f. , stauffer j. r. , vallenari a. , wolk s. j. , 2012 , a&a , 539 , 74 carpenter j. m. , hillenbrand l. a. , skrutskie m. f. , 2001 , aj , 121 , 3160 carpenter j. m. , mamajek e. e. , hillenbrand l. a. , meyer m. r. , 2006 , apj , 651 , l49 chauhan n. , pandey a. k. , ogura k. , ojha d. k. , bhatt b. c. , ghosh s. k. , rawat p. s. , 2009 , mnras , 396 , 964 chauhan n. , pandey a. k. , ogura k. , jose , j. , ojha d. k. , samal , m. r. , mito , h. , 2011 , mnras , 415 , 1202 cohen j.g . , frogel j.a . ,
persson s.e .
, ellias j.h . , 1981 , apj , 249 , 481 dahm s. e. , simon t. , proszkow e. m. , patten b. m. , 2007 , aj , 134 , 999 edwards s. , hartigan p. , ghandour l. , andrulis c. , 1994 , aj , 108 , 1056 favata f. , micela g. 2003 , space sci .
, 108 , 577 feigelson e. d. , broos p. , gaffney j. a. , iii , garmire g. , hillenbrand l. a. , pravdo s. h. , townsley l. , tsuboi y. , 2002 , apj , 574 , 258 feigelson e. d. , casanova s. , montmerle t. , guibert j. , 1993 , apj , 416 , 623 feigelson e.d . ,
decampli w.m . , 1981 , apj , 243 , 89 feigelson e. d. , gaffney iii j. a. , garmire g. , hillenbrand l. a. , townsley l. , 2003 , apj , 584 , 911 feigelson e. d. , montmerle t. , 1999 , ara&a , 37 , 363 flaccomio e. , damiani f. , micela g. , sciortino s. , harnden f. r. , jr .
, murray s. s. , wolk s. j. , 2003 , apj , 582 , 382 gaidos , e. j. , 1998 , pasp , 110 , 1259 getman , k. v. et al . , 2005 , apjs , 160 , 319 guarcello , m. g. , caramazza , m. , micela , g. , sciortino , s. , drake , j. j. , prisinzano , l. , 2012 , apjs , 753 , 117 gdel m. , 2004 , a&ar , 12 , 71 gdel m. , guinan , e. f. , skinner , s. l. , 1997 , apj , 483 , 947 gdel m. , et al . , 2007 , a&a , 468 , 353 haisch k.e . , lada e.a . , lada c.j . , 2001 , aj , 121 , 2065 hartmann l. , hewett r. , calvet n. , 1994 , apj , 426 , 669 herbst w. , herbst d. k. , grossman e. j. , weinstein d. , 1994 , aj , 108 , 1906 herbst w. , mundt r. , 2005 , apj , 633 , 967 hillenbrand l. a. , strom s. e. , vrba f. j. , keene j. , 1992 , apj , 397 , 613 jeffries r. d. , jackson , r. j. , briggs k. r. , evans p. a. , pye , j. p. , 2011 ,
mnras , 411 , 2099 kastner j. h. , huenemoerder d. p. , schulz n. s. , canizares c. r. , weintraub d. a. , 2002 , apj , 567 , 434 koenig x.p .
, allen l.e . , gutermuth r a. , hora j.l . ,
brunt c.m . , muzerolle j. 2008 , apj , 688 , 1142 koenigl a. , 1991 , apj , 370 , 39 lata s. , pandey a. k. , chen w. p. , maheswar g. , chauhan n. , 2012 , mnras , 427 , 1449 lynden bell d. , pringle j. e. , 1974 , mnras , 168 , 603 martn e. l. , 1998 , aj , 115 , 351 meyer m. , calvet n. , hillenbrand , l. a. , 1997 , aj , 114 , 288 muzerolle j. , calvet n. , hartmann l. , 2001 , apj , 550 , 994 ojha d. k. et al . , 2004 , apj , 608 , 797 pallavicini , r. , golub , l. , rosner , r. , vaiana , g. s. , ayres , t. , linsky , j. l. , 1981 , apj , 248 , 279 pandey a.k . , sharma s. , ogura k. , ojha d.k . , chen w.p .
, bhatt b.c . , ghosh s.k .
2008 , mnras , 383 , 1241 pandey , a. k. , samal , m. r. , chauhan , n. , eswaraiah , c. , pandey , j. c. , chen , w. p. , ojha , d. k. , 2013 , na , 19 , 1 , ( * pa13 * ) pizzolato , n. , maggio , a. , micela , g. , sciortino , s. , ventura , p. 2003
, a&a , 397 , 147 preibisch t. , et al .
, 2005 , apjs , 160 , 401 preibisch t. , feigelson e. d. , 2005 , apjs , 160 , 390 preibisch t. , zinnecker h. , 2002 , aj , 123 , 1613 prisinzano . l. , sanz - forcada .
j. , g. micela .
g. , caramazza .
m. , guarcello , m. g. , sciortino , s. , & testi , l. , 2011 , a&a , 527 , 77 , ( * p11 * ) prisinzano .
l. , g. micela .
g. , sciortino , s. , affer , l. , damiani , f. , 2012 , a&a , 546 , 9 , prosser c. f. , randich s. , stauffer j. r. , schmitt j. h. m. m. , simon t. , 1996 , aj , 112 , 570 rebull l. m. , stauffer j. r. , ramirez s. v. , flaccomio e. , sciortino s. , micela g. , strom s. e. , wolff s. c. , 2006 , aj , 131 , 2934 schmitt j. h. m. m. , robrade j. , ness j .- u .
, favata f. , stelzer b. , 2005 , a&a , 432 , l35 scholz a. , eisl ffel j. , mundt r. , 2009 , mnras , 400 , 1548 sharma s. , pandey a.k .
, ojha d.k .
, chen w.p . , ghosh s.k . ,
bhatt b.c .
, maheswar g. , sagar r. 2007 , mnras , 380 , 1141 shu f. h. , najita j. , ruden s. p. , lizano s. , 1994 , apj , 429 , 797 sicilia - aguilar a. , hartmann l. w. , harnandez j. , briceno c. , calvet n. 2005 , aj , 130 sicilia - aguilar a. , hartmann l. w. , frsz g. , henning t. , dullemond c. , brandner w. , 2006 , aj , 132 , 2135 stauffer j. r. , hartmann l. w. , prosser c. f. , randich s. , balachandran s. , patten b. m. , simon t. , giampapa m. , 1997 , apj , 479 , 776 stelzer b. , neuhuser r. , 2001 , a&a , 377 , 538 telleschi a. , gdel m. , briggs k. r. , audard m. , palla f. , 2007 , a&a , 468 , 425 walter f. m. , brown a. , matthieu r. d. , meyer p. c. , vrba f. j. , 1988 , aj , 96 , 297 walter f.m . , kuhi l.v .
, 1981 , apj , 250 , 254 | continuing the attempt to understand the properties of the stellar content in the young cluster ngc 1893 we have carried out a comprehensive multi - wavelength study of the region .
the present study focuses on the x - ray properties of t - tauri stars ( ttss ) in the ngc 1893 region .
we found a correlation between the x - ray luminosity , @xmath0 , and the stellar mass ( in the range [email protected] ) of ttss in the ngc 1893 region , similar to those reported in some other young clusters , however the value of the power - law slope obtained in the present study ( @xmath2 0.9 ) for ngc 1893 is smaller than those ( @xmath21.4 - 3.6 ) reported in the case of tmc , onc , ic 348 and chameleon star forming regions .
however , the slope in the case of class iii sources ( weak line ttss ) is found to be comparable to that reported in the case of ngc 6611 ( @xmath2 1.1 ) .
it is found that the presence of circumstellar disks has no influence on the x - ray emission .
the x - ray luminosity for both cttss and wttss is found to decrease systematically with age ( in the range @xmath3 0.4 myr - 5 myr ) .
the decrease of the x - ray luminosity of ttss ( slope @xmath2 -0.6 ) in the case of ngc 1893 seems to be faster than observed in the case of other star - forming regions ( slope -0.2 to -0.5 ) .
there is indication that the sources having relatively large nir excess have relatively lower @xmath0 values .
ttss in ngc 1893 do not follow the well established x - ray activity - rotation relation as in the case of main - sequence stars .
open clusters and associations - individual : ngc 1893 - pre - main - sequence stars : x - ray properties |
the existence of the gravitino in the particle spectrum of locally supersymmetric theories leads to problems in the standard cosmological scenario .
one manifestation of this so - called cosmological gravitino problem appears in scenarios where the baryon asymmetry is generated via thermal leptogenesis .
the late decays of the abundantly produced gravitino are in conflict with the successful predictions of big bang nucleosynthesis .
a similar problem even persists in scenarios where the gravitino is the lightest superpartner and thus a candidate for the dark matter . in this case
late decays of the next - to - lightest superpartner can spoil the predictions of big bang nucleosynthesis . among several other possibilities a very attractive solution to this problem are theories with a slight violation of r - parity @xcite .
theories with bilinear r - parity violation only violate lepton number and are thus not affected by proton stability constraints .
the size of the r - parity violating couplings are , however , constrained by cosmological considerations : they must be large enough to allow for the decay of the next - to - lightest superpartner before the time of big bang nucleosynthesis and they must be small enough that the baryon asymmetry of the universe is not washed out by these additional interactions .
interestingly , the allowed window for the couplings leads to a gravitino lifetime exceeding the age of the universe by several orders of magnitude , thus making the unstable gravitino a viable dark matter candidate @xcite .
an intriguing feature of this scenario is that it exhibits a rich phenomenology although the gravitino usually is thought to be one of the most elusive particles .
for instance , the next - to - lightest superpartner decays mainly via the small r - parity violating couplings and is thus long - lived on collider scales @xcite . in addition , the decay of gravitino dark matter in the galactic halo can lead to observable signals in the spectra of cosmic rays as we will discuss in the following .
the gravitino can decay into several two - body final states : @xmath0 and @xmath1 .
we calculate analytically the branching ratios for these decay channels , while the spectra of stable final state particles
_ i.e. _ photons , electrons , protons , deuterons and neutrinos are obtained by a pythia simulation of the gravitino decay @xcite .
the purpose of this conference contribution is to briefly report on the phenomenological results obtained in @xcite .
in recent years several anomalies in the spectra of cosmic rays have been reported that can not be explained by astrophysical secondary production and could be a hint for a signal from dark matter annihilation or decay . indeed , the decay of gravitino dark matter leads to a hard spectrum of positrons ( and electrons ) that depending on the gravitino mass and lifetime could explain the rise in the positron fraction observed by the pamela experiment @xcite .
this contribution could also be observed as a subdominant component in the absolute electron flux ( see figure [ positronelectron ] ) .
positron fraction ( _ left _ ) and electron spectrum ( _ right _ ) expected from gravitino dark matter decays compared to data and the expected astrophysical background .
shaded bands show the uncertainty due to charged cosmic - ray propagation through the galaxy.,title="fig:",scaledwidth=49.0% ] positron fraction ( _ left _ ) and electron spectrum ( _ right _ ) expected from gravitino dark matter decays compared to data and the expected astrophysical background .
shaded bands show the uncertainty due to charged cosmic - ray propagation through the galaxy.,title="fig:",scaledwidth=49.0% ] however , gravitino decays unavoidably produce associated fluxes of antiprotons and gamma rays that are only compatible with observations if the gravitino lifetime is larger than needed for an explanation of the pamela data ( see figure [ antiprotongamma ] ) .
therefore , a gravitino dark matter explanation of the pamela data is ruled out and additional astrophysical sources like pulsars are required to explain the observed positron fraction .
antiproton spectrum ( _ left _ ) and gamma - ray spectrum ( _ right _ ) expected from gravitino dark matter decays compared to data and the expected astrophysical background . from @xcite.,title="fig:",scaledwidth=49.0% ] antiproton spectrum ( _ left _ ) and gamma - ray spectrum ( _ right _ ) expected from gravitino dark matter decays compared to data and the expected astrophysical background . from @xcite.,title="fig:",scaledwidth=49.0% ] also antideuterons can be formed in the fragmentation of gravitino decay products .
the expected signal may exceed the astrophysical background especially for lower gravitino masses and can be tested with forthcoming experiments like ams-02 and the gaps balloon , thereby making antideuterons a valuable tool to probe light gravitinos ( see figure [ antideuteron ] ) .
neutrino spectrum expected from gravitino dark matter decays compared to data and the expected background of atmospheric neutrinos .
figure taken from @xcite . ]
neutrino spectrum expected from gravitino dark matter decays compared to data and the expected background of atmospheric neutrinos .
figure taken from @xcite . ] another valuable channel are neutrinos that like gamma rays provide angular information that can be used to discriminate gravitino decay from dark matter annihilation @xcite .
an intriguing feature of the expected signal is the line at the end of the spectrum that persists in contrast to gamma rays also for large gravitino masses ( see figure [ neutrino ] ) . in neutrino telescopes like icecube neutrinos
are mainly detected via the cherenkov light of neutrino - induced upward through - going muons .
although the rather poor energy resolution of this channel obscures the expected spectral features , a statistical analysis might reveal a departure from the featureless spectrum of atmospheric neutrinos ( see figure [ muons ] ) .
event topology of neutrino - induced upward through - going muons ( _ left _ ) , muon spectrum with finite energy resolution ( _ middle _ ) and statistical significance of the muon signal ( _ right _ ) . figures taken from @xcite.,title="fig:",scaledwidth=29.0% ] event topology of neutrino - induced upward through - going muons ( _ left _ ) , muon spectrum with finite energy resolution ( _ middle _ ) and statistical significance of the muon signal ( _ right _ ) .
figures taken from @xcite.,title="fig:",scaledwidth=34.0% ] event topology of neutrino - induced upward through - going muons ( _ left _ ) , muon spectrum with finite energy resolution ( _ middle _ ) and statistical significance of the muon signal ( _ right _ ) .
figures taken from @xcite.,title="fig:",scaledwidth=34.0% ] we find that a new detection channel , namely neutrino - induced showers , has two advantages compared to the muon signal : the dominant background of atmospheric muon neutrinos contributes less to this channel and its better energy resolution allows for a better discrimination of spectral features ( see figure [ showers ] ) . although there are currently still problems to discriminate the shower signal from short muon tracks we conclude that this channel is the most promising for a detection of neutrinos from gravitino decays .
event topology of neutrino - induced showers ( _ left _ ) , spectrum of showers with finite energy resolution ( _ middle _ ) and statistical significance of the shower signal ( _ right _ ) . from @xcite.,title="fig:",scaledwidth=29.0% ] event topology of neutrino - induced showers ( _ left _ ) , spectrum of showers with finite energy resolution ( _ middle _ ) and statistical significance of the shower signal ( _ right _ ) . from @xcite.,title="fig:",scaledwidth=34.0% ] event topology of neutrino - induced showers ( _ left _ ) , spectrum of showers with finite energy resolution ( _ middle _ ) and statistical significance of the shower signal ( _ right _ ) . from @xcite.,title="fig:",scaledwidth=34.0% ] the non - observation of gravitino decay signals in the spectra of various cosmic - ray species allows to estimate lower limits on the gravitino lifetime ( see figure [ lifetime ] ) . for light gravitinos
the strongest limits come from gamma - ray line searches , while at larger masses the antiproton and diffuse gamma - ray limits are the most important .
for very heavy gravitinos also neutrino bounds become significant .
we expect that forthcoming antideuteron observations and searches based on neutrino - induced showers allow to improve these bounds in the future . estimated limits on the gravitino lifetime from gamma - ray observations ( _ left _ ) , observations of charged cosmic rays ( _ middle _ ) and forthcoming observations of upward through - going muons at icecube ( _ right _ ) .
figures taken from @xcite.,title="fig:",scaledwidth=37.0% ] estimated limits on the gravitino lifetime from gamma - ray observations ( _ left _ ) , observations of charged cosmic rays ( _ middle _ ) and forthcoming observations of upward through - going muons at icecube ( _ right _ ) .
figures taken from @xcite.,title="fig:",scaledwidth=37.0% ] estimated limits on the gravitino lifetime from gamma - ray observations ( _ left _ ) , observations of charged cosmic rays ( _ middle _ ) and forthcoming observations of upward through - going muons at icecube ( _ right _ ) .
figures taken from @xcite.,title="fig:",scaledwidth=24.4% ]
unstable gravitino dark matter with a lifetime exceeding the age of the universe by many orders of magnitude is well motivated from cosmology .
although gravitino decays produce cosmic - ray spectra that could explain the pamela excess , this possibility is ruled out due to strong constraints on the associated production of antiprotons and gamma rays .
forthcoming antideuteron searches will be a valuable channel to search for signals of light gravitinos while neutrino telescopes like icecube have the capability to probe heavy gravitinos in particular if new detection channels like shower events are employed . in conclusion , we find that the gravitino lifetime can be effectively constrained over a large range of gravitino masses using a multi - messenger approach . for an exhaustive discussion of the presented results see @xcite .
i am grateful to laura covi and gilles vertongen for useful discussions and collaboration on part of the presented work .
i acknowledge the support of the dfg within the collaborative research center 676 , the marie curie itn `` unilhc '' under grant number pitn - ga-2009 - 237920 and the grant hephacos s2009/esp-1473 of the comunidad de madrid .
9 buchmller w _ et al .
_ 2007 _ j. high energy phys .
_ jhep03(2007)037 ( _ preprint _ hep - ph/0702184 ) takayama f and yamaguchi m 2000 _ _ phys . lett .
_ _ b * 485 * 38892 ( _ preprint _ hep - ph/0005214 ) bobrovskyi s _
_ 2010 _ j. high energy phys .
_ jhep10(2010)061 ( _ preprint _ arxiv:1007.5007 [ hep - ph ] ) grefe m 2011 _ unstable gravitino dark matter
prospects for indirect and direct detection _ + desy - thesis-2011 - 039 ( _ preprint _ arxiv:1111.6779 [ hep - ph ] ) adriani o _
[ pamela collaboration ] 2009 _ _ nature _ _ * 458 * 6079 ( _ preprint _ arxiv:0810.4995 [ astro - ph ] ) covi l _
_ 2010 _ j. cosmol .
. _ jcap04(2010)017 ( _ preprint _ arxiv:0912.3521 [ hep - ph ] ) | the gravitino in models with a small violation of r - parity is a well - motivated decaying dark matter candidate that leads to a cosmological scenario that is consistent with big bang nucleosynthesis and thermal leptogenesis .
the gravitino lifetime is cosmologically long - lived since its decays are suppressed by the planck - scale as well as the small r - parity violating parameter .
we discuss the signals in different cosmic - ray species coming from the decay of gravitino dark matter , namely gamma rays , positrons , antiprotons , antideuterons and neutrinos .
comparison to cosmic - ray data can be used to constrain the parameters of the model . |
the rapid development of technology , the distance elimination through communication networks and competitiveness have influenced the field of education .
tele - education is a new form of education which is the result of the above factors . in a few words , tele - education aims at education through distance , without the simultaneous presence of the trainers and trainees and the use of new technology such as networks , multimedia etc .
it is a form of education which is adjusted to the needs of each trainee and where time and distance do nt matter . on the other side
, there is an endless ability of expansion and there is no restriction in the number of the participants .
there are many tele - education platforms either synchronous or asynchronous see for example @xcite , @xcite .
nevertheless , there are a few platforms that can support the evaluation of trainees automatically .
the s.a.t.e.p .
application was designed as an autonomous platform for synchronous/ asynchronous tele - education through which the administrator just inserts the various files and questions and the platform produces the tests randomly .
the application was created under the philisophy of open source software , which gives the potential to anyone having the basic knowledge of programming languages , to adjust the platform to his needs with no cost see also @xcite and @xcite .
we make use of html , javascript , mysql , apache server , php and css .
the html was used for the creation of the web site design@xcite .
javascript was used mainly for the control of the elements existing in the forms and generally whenever there is a dynamic content for the local process of the data ( client - side ) @xcite .
mysql was used for the design of database .
mysql is a very powerful tool for the management of database , through which is given the potential to insert , modify and delete tables , data and relations among tables @xcite . through mysql server ,
is ensured that only authorized users have access to it and that a number of users can work simultaneously on the same database .
apache is an open source web server @xcite .
it s highly adaptive with advanced attributes . through css
were created the menus of the web site , for the formatting of the various elements and also for their behavior in certain events . finally , php is a server side script language which is available for all operating systems @xcite . for the creation of the source code of the web site
were also used open source programs such as cssed and tswebeditor .
the s.a.t.e.p .
application is used in the `` web programming language '' course of the m.sc .
applied informatics course in kozani .
in the s.a.t.e.p . application three types of users are supported .
there are the guest users , who do nt have access to the system , the registered users who do have access to the platform and finally administrators .
the guest users can only see the name of lecture files , apply for registration to the platform and finally communicate with the administrator . the system has been designed in a way such that not everyone can access the platform . only the administrators have the ability to allow the users to have access .
when a user wants to have access to the system he fills his personal details in a form .
the application makes use of a captcha textbox such that to deny access to bots that make automatic registrations .
also there is a check for duplicates not only for the personal details but also for the e - mail . if all these informations are unique the administrator grants the user the ability to use the platform .
also , at this point if a user wants to recover his password , he can complete his username and the new password will be sent to his e - mail address .
in addition there is a second group of users , the registered users . in this group
they can view and edit some of their account information , can read the lecture files , do the lecture tests in order to test their knowledge and communicate with other users or the trainer through a chat module .finally they can take exams that are defined by the admin . as shown in figure 3 , a registered user can edit his username , e - mail and password .
the registration index , the name and the last name can not be edited by the user , but only from the administrator .
one of the most important module of the application is the tests and exams procedure .
all the questions appear in random order from a pool a large number of questions that are exist in the database .
so if someone takes the first lecture test more than once , he might not even have one common question .
furthermore , if two users take the same test simultaneously , either the questions do nt appear in the same order or they are nt the same .
moreover , in multiple choice questions the answers appear in random order , with the first one always selected .
the check of the answers is done automatically and the results are stored in the database . the registered user can communicate with the administrator via a contact form .
the day / time , name and e - mail of the user are completed automatically by the system and all he has to do is to compose the message .
the third and last group of users are the administrators .
they can insert , modify and delete users , lectures , as well as all the types of questions such as gap filling and/or multiple choice questions .
they can also communicate with other users with chat or email , define the date of the tests and exams , as well as their duration .
after the date/ time duration is defined , the users are informed automatically via email .
when the users decide to be members of the platform and apply for , they do nt immediately obtain access .
the administrator must authenticate that they are really community `` members '' and then either grant access to them or delete them from the system minimizing the possibility of `` bots '' scripts . moreover , the administrator is able to view all the registered users and their personal details , but he can edit only some of them such as their register number , last name and name .
in addition , he has the potential to search for registered users based on their register number , name , last name , username or e - mail by inserting a string of characters .
the results are displayed based on the category chosen . in the `` delete users '' section , the users are displayed in order based on their register number and the administrator is able to delete them by choosing the corresponding checkbox .
the administrator has also the ability to insert , delete and edit the lectures . due to the fact that the lectures that can be taken in each course vary , the administrator has the ability to add lectures adjusted to the needs of each course .
the opposite procedure is the deletion of the lectures . during that ,
not only the lecture files but also the questions related to the specific lecture are deleted .
furthermore , the administrator is able to display , upload and delete files of each lecture ( figure 12 )
. he can also select the lecture whose files need to be processed . in figure 13 the name , size and file type of the lecture files are shown . in the column `` action '' the administrator can select
which files he wishes to delete from a specific lecture . by pressing the
`` submit '' button , the files that have been marked as `` delete '' will be erased from the database .
another important module of the application that the administrator has access is the questions category .
the questions in the platform are divided in multiple choice questions and gap filling questions . during the insertion of the questions according to the type of question the form of the fields changes .
for instance , if the type of question that has been chosen by the administrator is a gap filling question , the result will be as shown in figure 14 .
the administrator is asked to fill in both the question and the right answer , but also to choose in which lecture the specific question belongs . in case that the administrator chooses to insert a multiple choice question , the menu is adjusted as shown in figure 15 .
the administrator is asked to choose in which lecture the question belongs and fill in the fields of the question , the right answer and at least of one wrong answer .
the administrator can also delete one or more questions using a menu . during the view of gap - filling questions for deletion
, the questions are ordered according to the lecture .
the fields of the question , the answer and the lecture in which they belong are displayed , as well as the checkbox that is chosen for question deletion .
the administrator can change the page size , he can also search for some question by entering a string in the textbox .
the search is executed either there is a full search string or a part of the word we are looking for . the same procedure also applies to multiple choice and gap - filling questions as well . during processing the administrator can make changes to the questions and to the answers . administrator also has the ability to send massive e - mails to all registered users . finally , the administrator sets the dates and the duration of the tests / exams .
the users are informed via e - mail for any changes automatically .
mysql is the most well - known database management system @xcite .
it is an open source tool and its major advantage is that it is constantly improved in regular time intervals . as a result mysql has evolved into a fast and extremely powerful database .
it performs all functions like storing , sorting , searching and recall of data in an efficient way .
moreover , it uses sql , the standard worldwide query language .
finally , it is distributed free .
the database which is used in the application has been designed to use the storage engine innodb , so that the restrictions of the foreign keys can be created .
it consists of twelve tables , eight of which are the most important as they are connected to each other with relations .
its complete diagram is as follows : the table register consists of : field am ( type : integer ) which is the primary key . the fields name , surname , username , password , email and department are varchar type . in this table
the records which represent the users applications in order to access the system , are stored .
the table admins consists of : the field idadm ( type : integer ) which is the primary key .
the fields name , surname , username , password , email and department are varchar type . in this table all the administrators of the system , are stored .
the table misc consists of fields idmisc ( type : integer ) which is the primary key , time ( type : time , format : hh : mm : ss ) and date ( type : date , format : mmyyyydd ) . in this table
is stored the date and the time that the final and lecture tests will take place .
the table history consists of fields id_gen ( type : integer ) which is the primary key , am ( type : integer ) , date ( type : date , format : mmyyyydd ) and percent ( type : float ) . in this table
are stored all the final tests that the user gave , with every date and percentage he achieved each time . in the table lectures are stored only the lectures of the lessons .
it consists of idd ( type : integer ) which is the primary key and lecture ( type : varchar ) . in table lecture_files
are stored the file paths that are related with lectures .
it consists of i d ( type : integer ) , idd ( type : integer ) , name ( type : varchar ) , type ( type : varchar ) and size ( type : integer ) . the primary key is i d whereas idd is the foreign key which defines in which lecture the file of the lesson is referred . in the filed name is stored the full path of the file . in the field type
is stored the type of the file whereas in the field size its size . in tables
multiple_questions and filling_questions are stored multiple choice questions and filling questions respectively .
the fields of multiple_questions table are ide ( type : integer ) , idd ( type : integer ) whereas question , ra , wa1 , wa2 , wa3 are varchar type .
ide is the primary key whereas idd is a foreign key which defines in which lecture each multiple choice question , belongs .
the field question contains the question , ra contains the right answer and in the rest three fields are the wrong answers .
the fields of the table filling_questions are idf ( type : integer ) , idd ( type : integer ) , question ( type : varchar ) and answer ( type : varchar ) .
idf is the primary key whereas idd is a foreign key which defines in which lecture each filling question , belongs . in the field questions
the question is stored , whereas in the field answer the right answer .
each of user_mult_test and user_fill_test tables consist of two fields , where both of these fields are keys .
fields idum ( type : integer ) and iduf ( type : integer ) are the primary keys respectively .
these two tables are intermediate stops for the creation of the final table user_complete_test that is shown below .
their presence is necessary and the reason is , that for the various tests which the system provides the number of questions for each type is different . in the final test of the system for instance are chosen twenty multiple choice questions and ten filling questions . supposing we want to create a test with ten multiple choice and five gapfilling questions .
the questions are grouped in the user_multr_test and user_field_test tables and each group is put as a foreign key in the user_complete_test_table . in the table users are stored the registered users of the system .
it consists of the fields am ( type : integer ) which is the primary key and name , surname , username , password , email , department which are varchar .
the fields are equally important with fields of the register table . in the final user_complete_test table
are stored all the tests that the user gave , either they refer to lecture tests or final tests .
its fields are idum ( type : integer ) , iduf ( type : integer ) , am ( type : integer ) , date ( type : date , format : mmyyyydd ) and percent ( type : float ) .
the table does nt have a primary key of its own , but three foreign keys ( idum , iduf , am ) are its primary . in the field
am is stored the record number of the student to whom the test corresponds . in the fields
idum and iduf are stored the multiple choice question and filling groups that have been chosen for the specific am test . in the field date
is stored the date that the test is held and in the field percent the percentage that the student achieved .
in the present paper the s.a.t.e.p . :
synchronous - asynchronous tele - education platform for educational purposes is developed and can be easily parametrized to meet the needs for any educational inst .
the administrator can upload the files that the user intends to read , but also inserts the questions of each lecture .
the trainee can read / download the files of each lesson and also test his knowledge on what he has studied through a series of questions / tests .
the automatic mix of the exam questions and their automatic correction as well , makes the current assignment different from the rest of the content management systems ( cms ) . unlike cms
, the current platform is ready to be used and does nt need the addition of extra plugin .
html , xhtml , and css : complete , gary b. shelly and denise m. woods , 2010 .
php solutions : dynamic web design made easy , david powers , 2010 .
high performance mysql : optimization , backups , and replication , baron schwartz & peter zaitsev & vadim tkachenko , 2012 .
apache : the definitive guide 3rd edition , ben laurie and peter laurie , oreilly , 2003 .
e - airquality : a dynamic web based application for evaluating the air quality index for the city of kozani , hellas , 15th panhellenic conference on informatics ( pci ) , sept .
30 2011-oct . 2 2011 , pp .
171 - 174 , 2011 .
a.q.m.e.i.s . : air quality meteorological and enviromental information system in western macedonia , hellas , arxiv:1406.0975 , 2014 .
tele - learning : the challenge for the third millennium ( ifip advances in information and communication technology ) , edited by don passey and mike kendal , 2011 . international perspectives on tele - education and virtual learning environments , edited by graham orange and dave hobs , 2010 . | s.a.t.e.p . :
synchronous - asynchronous tele - education platform is a software application for educational purposes , with a lot of parametrizing ( configuration ) features written entirely from scratch .
it aims at the training and examination of computer skills , a platform that can be adjusted to the needs of each lesson . in the application
the trainer and the administrator can define the number of the lectures and upload files for each one of them .
furthermore , he can insert , modify and delete questions which are used for evaluation tests but also for the trainees examinations .
the trainee can read / download the files of each lesson and also test his knowledge on what he has studied through a series of questions / tests . a chat module where registered users as well as system administrator can discuss and solve questions
is also developed .
web based application ; mysql ; php ; open source software;tele - education ; |
physicists and engineers often need to find methods for confining a field as much as possible with a minimum of knowledge of the system at hand . in the case of monochromatic waves
, it has been found that phase and amplitude wave engineering is of crucial importance for subwavelength focusing of waves@xcite . in the case of pulsed waves ,
the situation is more complex , mainly due to the fact that the superposition of many spectral components in most cases tend to broaden the focal spot .
theoretically , one may hope to be able to focus strongly confined pulses as in ref .
@xcite , but the aperture field and medium required remains elusive . however , there are a few promising pathways , e.g. , using rainbow - colored apertures which combine to a single white - colored focus@xcite or diffractive optical elements@xcite .
well developed theories and numerical algorithms are available for studying monochromatic and polychromatic waves in focal regions@xcite , but they do not provide direct insight into the reverse engineering problem and may not provide a fast solution or be optimal with the minimal amount of information given . of particular interest here is the emerging field of evanescent wave focusing , where up to now mainly monochromatic waves have been studied@xcite
. however , also for waves containing several frequency components it is necessary to have straightforward methods which allow backtracking of the signal in a manner such that the initial field resulting in a strong focus is found . in the present work we detail a method for finding the intensity distribution near the focal region using phase and amplitude compensation . in order to achieve this goal
, we pinpoint a certain distribution at focus and then try to calculate the required aperture field .
moreover , we show an analogy between our compensation method and the transformation of frequencies between inertial , relativistic coordinate systems .
the method of phase and amplitude compensation has been described in a number of studies ( see e.g. ref.@xcite and references therein ) , but was more recently re - introduced to study focusing of evanescent waves@xcite .
the procedure is most easily demonstrated for scalar monochromatic waves satisfying the helmholtz equation @xmath0 where @xmath1 is the wavenumber , @xmath2 is the angular frequency and @xmath3 is the velocity of light in the medium under consideration . using the angular spectrum representation ,
the field can be written as @xmath4 where the angular spectrum is given by @xmath5 where @xmath6 here @xmath7 represent the homogenous plane waves , whereas @xmath8 correspond to inhomogeneous or evanescent plane waves that propagate in the @xmath9 direction and decay in the z - direction .
the trick is now to express the angular spectrum as @xmath10 where we must again require @xmath11 to be bandlimited ( i.e. nonzero only in a range @xmath12 ) and converges sufficiently quickly such that the solution for the field does not possess any unphysical divergency problem .
the conditions for this was discussed in ref .
@xcite , and will not be repeated here .
the field at the focal point ( @xmath13 ) is now given by @xmath14 i.e. a fourier transform of the aperture function @xmath15 resulting in a finite resolution at the focal point .
as an example , assume now that @xmath16 for @xmath17 ( @xmath18 ) and zero elsewhere .
then we have @xmath19 in order to find the field @xmath20 we must evaluate @xmath21 in ref.@xcite it was found that in the case of paraxial waves ( @xmath22 ) the field @xmath20 is approximately a quadratic phase function , which has been described in detail in standard textbooks on optics@xcite .
thus , in the paraxial approximation the phase compensation method directly leads us to the well - known phase profile for a lens giving the most strongly focused profile@xcite .
recently , the phase compensation method has been used to design evanescent waves in focal regions@xcite . the simplest wave profile exhibiting a focal region
is found by letting @xmath23 if @xmath12 and zero elsewhere .
here @xmath24 and @xmath25 is a constant .
the field can then be expressed as @xmath26 \,\,\ , .
\label{perf}\ ] ] the intensity at the focus is therefore just @xmath27 whereas the aperture field oscillates rapidly within an envelope such that the intensity is @xmath28 . notice that the aperture field is much stronger than the field in the focal region , i.e. @xmath29 .
however , we also see that the intensity distribution is more confined at the focal region than anywhere else . in fig . [
f1](a ) the intensity of a monochromatic , evanescent wave is displayed with @xmath30 ( arbitrary units ) , @xmath31 ( arbitrary units ) and @xmath32 . the solid line shows the field in focus , whereas the dashed line shows the fields at @xmath33 .
this illustrates our point , namely that we may obtain a more spatially confined intensity distribution ( although the magnitude is considerably smaller than that at the aperture ) .
if we assume that @xmath34 , a possible criterium for focusing could be obtained by requiring that the aperture intensity ( at @xmath13 ) envelope at @xmath35 , corresponding to the position with half the maximum intensity , should be wider than the focused intensity distribution of half - width @xmath36 .
thus , by requiring @xmath37 , we ensure that the field is focused . at the same time
, the field should not diverge at the aperture , and it is clear that @xmath38 can not be too large .
interestingly , the compensation method described above is not limited to designing waves in focal regions described by the helmholtz equation .
it can also be used to construct solutions to other differential equations , such as the diffusion equation , and we will here look at one example .
imagine that we generate a spatial magnetic field distribution by positioning a system of co - aligned current - carrying wires in a homogeneous medium of frequency - independent conductivity @xmath39 and permittivity @xmath40 .
each of the wires carry a low frequency current where the frequencies of any temporal waveform fulfills @xmath41 . in absence of sources
the resulting electric ( or magnetic ) field can be described by the diffusion equation on the form@xcite @xmath42 where @xmath43 is the diffusion coefficient .
we now want to describe the field in the vicinity of the wires , but not including any sources such that eq .
[ adiff ] is valid .
the field can then be written on the following general form @xmath44 where @xmath45 is the fourier amplitude .
usually , a solution of eq .
[ hhh ] will represent a diffusion process where the field delocalizes with increasing time@xcite .
however , we are here interested in finding a field @xmath20 with an optimally localized field @xmath46 ( @xmath47 ) . to see how one may design such an initial field ,
let us assume that @xmath48 is a constant for @xmath49 and zero for @xmath50 .
thus , the field is given by @xmath51 the real value of eq .
[ hh1 ] at @xmath52 is given by @xmath53 whereas at @xmath54 it is @xmath55 by selecting @xmath56 such that @xmath57 , we see that the leading contribution to @xmath58 , which is an oscillatory function with period @xmath59 .
thus , the field at @xmath52 is strongly delocalized , but with sharp local fluctuations , as seen in fig .
[ f1 ] b ) . on the other hand
, @xmath46 is strongly localized with a central lobe @xmath60 , but is a factor @xmath61 smaller than the field at @xmath52 . a sharp localization is therefore obtained at the expense of an exponential reduction in field amplitude .
this is rather similar to the behavior we observe for monochromatic evanescent waves , although the details are different .
in addition to describing electromagnetic fields , the diffusion equation may also describe concentration or temperature ( but now with a different diffusion coefficient ) . in those cases
it should be mentioned that we do not consider absolute temperature or concentrations ( as required by our boundary conditions ) , but rather modulations about an equilibrium .
thus , the negative values seen in fig .
[ f1 ] b ) do not mean absolute negative temperature and concentration , they are just signatures of the variations about the mean temperature or concentration seen when designing an initial profile required to bring the diffusive waves to a focus .
an interesting question is whether the method above can be applied directly to polychromatic waves ( e.g. pulsed waves ) satisfying the wave equation @xmath62 in the angular spectrum representation , the field can be expressed as @xmath63 where @xmath64 and @xmath65 in order to bring the polychromatic waves to focus we require cross - compensation of the phase , which amounts to setting @xmath66 , i.e. the spatial part of the phase is compensated by the temporal part at @xmath67 .
such a requirement can only be fulfilled if @xmath2 depends on the spatial frequencies ( i.e. direction of each plane wave ) , where the angular frequency is given by @xmath68 . here
@xmath69 and @xmath70 .
the fact that the frequency of the spatial wave vector depends on the direction was utilized in ref . @xcite to combine a rainbow spectrum to a single , focused spot .
conceptually the idea presented here is somewhat similar , although we use an entirely different approach to achieve the goal .
note that we must distinguish between the two cases @xmath71 and @xmath72 .
the case @xmath71 results in superluminal , localized waveforms which have been discussed extensively in the literature@xcite . here
@xmath73 is real and given by @xmath74 ( @xmath75 ) .
the scalar field is then given by @xmath76 a similar representation was found in ref .
@xcite describing one - dimensional nondiffracting pulses , but here we have arrived at eq . [ e ] from an entirely different perspective based on requirement of specific field profile resulting from phase compensation . of greater interest in the current study is the case @xmath72 , since that represents an evanescent wave solution similar to that seen in the previous section .
now @xmath77 is purely imaginary , such that @xmath78 as an example , we may approximate the angular spectrum as @xmath79 for @xmath80 and zero elsewhere . here
we also assume that @xmath24 , such that only evanescent waves are excited with @xmath81 and @xmath82 .
this evanescent - wave approximation therefore requires @xmath83 , and the field will be @xmath84 \,\,\ , .
\label{evantime}\ ] ] note that eq .
[ evantime ] is identical to eq.[perf ] if we set @xmath85 , and is therefore just a translation of the evanescent wave along the optical axis with the intensity distribution similar to that of fig .
since we require @xmath86 , which is fulfilled if we , as an example , set @xmath87 , the half - width of the intensity distribution at the traveling focal line is narrower than @xmath88 .
the intensity at @xmath33 grows as @xmath89 .
however , when @xmath90 our approximate theory for evanescent waves does not longer hold in the vicinity of @xmath33 ( this follows from the considerations above ; see also refs .
@xcite for a discussion about diverging solutions ) .
the fact that the evanescent wave does not change at the focal line @xmath91 is surprising , given the condition @xmath92 above .
however , we also note that @xmath93 such that the increasing energy of the field at the aperture is used to keep the field at the focal line @xmath91 unchanged in magnitude as it propagates outwards .
the technical implementation of focusing of polychromatic waves is in effect similar to that of monochromatic waves , but with two new important features : a ) the focusing must take place in a spatially dispersive media , and b ) the field at the aperture must grow with time ( within the approximation given above ) . an interesting analogy occurs if one compares the problem of designing an aperture field that generate optimal focus with that of transformation of frequencies between inertial , relativistic coordinate systems@xcite . to see this , consider a plane wave of the form @xmath94 .
any wavepacket is a weighted sum over such plane waves in spatial and frequency coordinates . in order to bring the wave packet to focus
we require that every spectral component is at focus a certain distance @xmath95 from the aperture at a given time @xmath96 .
this can be done by setting @xmath97 , such that the phase is exactly compensated at @xmath98 .
it should be noted , as seen above , that such a requirement may give rise to strong localization at other positions as well . in any case , such a cross - compensation requirement can only be fulfilled if @xmath2 depends on the spatial frequencies ( i.e. direction of each plane wave ) , where the frequency is given by @xmath99 \left [ \left ( \frac{k_x z_1 } { \phi ( k_x ) } \right ) ^{2 } -1 \right ] } \,\,\ , .
\label{timec}\ ] ] now consider the problem of transforming plane waves between inertial , relativistic coordinate systems .
that is , consider an observer moving at a speed @xmath3 relative to a fixed frame . according to einstein
s special theory of relativity , the moving observer will detect a frequency @xmath100 given by @xmath101 , where @xmath102@xcite . from the above it may be inferred that this expression corresponds to @xmath103 if we make the associations @xmath104 and @xmath105 , i.e. the phase factor @xmath106 must be associated with the frequency measured in the moving frame . in the special case considered in this study @xmath107 , which corresponds to @xmath108 . for evanescent waves
we then immediately find the spatial dispersion relationship given above . for propagating waves we may set @xmath109 , where @xmath110 is the real angle at which a specific plane wave makes with the direction of motion .
then we must have @xmath111 and therefore @xmath112 , in agreement with the observations made above .
in conclusion , we have suggested a compensation method for designing aperture fields giving rise to strongly confined waves .
the idea is to first try to compensate the phase or amplitude such that only transverse spatial frequencies are left in the angular spectrum representation at a given region in space , thus allowing a strongly focused wave to form here .
next we calculate the aperture field required to obtain such a phase or amplitude compensation .
the method has been employed to study diffusive , monochromatic and polychromatic waves , and shown to give new insight into the problems at hand .
the method here can probably also be applied to other wave systems where phase or amplitude compensation is beneficial . in ( a ) the intensity of a monochromatic , evanescent waves
is displayed with @xmath30 ( arbitrary units ) , @xmath31 ( arbitrary units ) and @xmath32 . in ( b ) the diffusive waves ( a ) with @xmath126 ( arbitrary units ) , @xmath127 ( arbitrary units ) and @xmath128 are displayed .
the solid lines show the field in focus , whereas the dashed lines show the fields at @xmath52 for diffusive waves and @xmath33 for monochromatic evanescent waves.,width=453 ] | a general method for focusing of waves , based on phase and amplitude compensation , is applied to monochromatic , polychromatic and diffusive waves .
monochromatic waves may form spatially localized waves in free space , whereas polychromatic waves form non - decaying traveling evanescent modes confined to subwavelength regions in media where the frequency depends on the wave vector .
we suggest an analogy between the phase compensation method and the transformation of frequencies between inertial , relativistic coordinate systems . |
in the last sesto meeting on galaxy formation and evolution in 1996 , some fundamental questions concerning how do galaxies build up and evolve in a cosmological frame have been established .
probably one of the most interesting question is how do galaxies retain memory about the physics of their formation , memory that deals with their mass aggregation histories ( mahs ) in a hierarchical cosmogony and with the origin of their angular momentum .
this memory may represents a powerful link between the present features of the galaxies and the nature of the dark matter ( dm ) as well as the power spectrum and the statistical properties of its density fluctuation field ( avila - reese & firmani 1997 ) .
the facts that most disk galaxies have flat rotation curves , exponential surface brightness profiles , star formation ( sf ) histories , colors and gas fractions that correlate with the morphological type , and show a tight correlation between the luminosity and the circular velocity , point out to a common physical origin for many of their present features .
much progress has been carried out from the observational and theoretical points of views , particularly since a self - consistent scenario for galaxy formation that integrates cosmology and astronomy has appeared .
however , not much progress has been attained concerning the connection between the internal physics of galaxies and the overall cosmological frame .
the current theoretical approaches on galaxy formation can be formulated on the ground of three main methods : ( 1 ) the _ numerical simulations _ , ( 2 ) the _ semianalytical _ models , and ( 3 ) the _ analytical _ models . the numerical simulations ( e.g. , yepes 1997 and references therein ) are the most direct way to study the complex evolution of the density fluctuations , but also , due to technical and numerical limitations , at least when including the gas hydrodynamics , this is currently the less predictive and most expensive method .
the semianalytical models ( kauffmann , white & guiderdoni 1993 ; cole et al .
1994 ; somerville & primack 1998 , and more references therein ) are useful as a method to constrain the space of parameters which resume some ignorance about several physical processes of galaxy formation and evolution , however , they do not deal with the internal metabolism of the galactic disks .
the analytical method ( mo , mao & white 1998 ; see also dalcanton , spergel , & summer 1997 ) is the most economical approach to study the global properties of galaxy populations at different redshifts .
nevertheless , this method does not deal with the evolution and physics of individual galaxies .
we shall develop a method that allows to calculate the internal structure and physics of the disk galaxies that forms in the evolving dm halos , using for this aim simple self - consistent physical models implemented in a numerical code .
the scenario for which this _ seminumerical _ method was developed is that of an _ inside - out _ galaxy formation where the rate of gas accretion is given by the rate of the hierarchical mass aggregation , and where the primordial angular momentum is acquired by the dm halo during its linear gravitational regime by tidal torques .
we start from the hydrodynamics of a gas and a stellar disk with axial symmetry ( firmani & tutukov 1994 ; firmani , hernndez & gallagher 1996 ) .
the disk gas gravitational instabilities induce the sf which is self - regulated by the sn energy feedback on the ism .
a simple stellar population synthesis model is considered to account for the luminosity evolution .
a secular bulge formation is introduced applying a local gravitational stability criterion to the stellar disk .
the galactic evolutionary models are inserted in a cosmological background : the structure and evolution of the dm halos which surround the disks and the mass accretion rates over them are calculated from initial conditions defined by the cosmological model , which is specified by the mass fractions of the dm species ( cold @xmath1 , hot @xmath2 ) , the vacuum energy ( @xmath3 ) , and the baryon matter ( @xmath4 ) , and by the value of the hubble constant ( @xmath5 ) . with the aim to study general behaviors
we shall use the @xmath6 model normalized to @xmath7 , since this model is very frequently found in the literature . in some cases
we also shall present results for some of the most representative cosmologies : the @xmath8 ( @xmath9 , @xmath10 , @xmath11 ) , the @xmath12 ( @xmath9 , @xmath11 ) , the @xmath13 ( @xmath14 , @xmath15 , @xmath16 , @xmath17 ) , and the @xmath6 ( @xmath18 , @xmath17 ) . where the value of @xmath4 is not specified it was taken equal to @xmath19 .
a gaussian statistical distribution with a power spectrum taken from sugiyama ( 1996 ) and normalized to the cobe data ( except for the @xmath6 normalized to @xmath7 ) characterizes the primordial density fluctuation field .
the mah of a given halo is obtained by a monte - carlo method applied on the conditional probability of the density fluctuation field ( lacey & cole 1993 ) . in this work
we are interested in disk galaxies that have not suffered during their evolution a major merger able to destroy the disk .
then we have excluded from our sample haloes that in some time have collided with an other halo with a mass greater than half of its mass at that time .
starting from the linear evolution of a density fluctuation and assuming spherical symmetry we obtain the virialized halo density profile that results from the non - linear evolution of the density fluctuation . to calculate this we use a statistical approach based on the adiabatic invariants ( avila - reese , firmani & hernndez 1998 , afh )
. the non - radial component of the kinetic energy is calibrated on the base of the n - body simulations .
the angular momentum of each shell is calculated assuming for the virialized structure a value of the spin parameter @xmath20 constant in time ( in agreement with the linear theory of angular momentum acquirement , white 1994 ) .
we have assumed an average value of @xmath21 with a lognormal distribution .
once a mass shell is incorporated into the halo , its baryon mass fraction falls onto the disk .
we assume that the disk - halo feedback is negligible and that the cooling flow does retain just a negligible amount of gas within the halo .
this is consistent with the hypothesis that the sn energy feedback is mainly localized into the disk in agreement with the observational evidence .
we have used a simple parametrized gas cooling mechanism into the halo in order to test the sensitivity of the results to this very complex process .
the gas of the shell is distributed across the disk assuming rigid rotation of the shell , in agreement with the zeldovich approximation , and assuming detailed angular momentum conservation during the gas dissipative collapse .
adiabatic invariants are used to calculate the halo gravitational contraction produced by the increasing gravitational field due to the disk mass growth .
for a fixed present - day mass we obtain a large sample of possible mahs ( see fig . 1 in firmani & avila - reese 1998 ) which after the virialization give rise to a _ wide range _ of density profiles .
the average mah produces a density profile similar to that obtained in n - body simulations by navarro , frenk and white ( 1997 ) .
however , our method gives in a natural way the statistical deviations from the average mah , which lead to a _
rich variety of halo structures _ ( afh ) .
the properties of luminous galaxies may depend on the features of these evolving halos because of luminous galaxies form within them .
the surface density profiles of the disks in centrifugal equilibrium formed within the dm halos with the gradually infalling gas shells in solid body rotation are nearly exponential over several scale lengths ( fig .
models with high angular momentum have extended disks with low surface brightness . according to dalcanton , summer & spergel ( 1997 )
, these models would correspond to the low surface brightness ( lsb ) galaxies .
observations show systematic blue radial color gradients in spiral galaxies ( e.g. , de jong 1996a ) .
our models show a similar trend ( fig .
1b ) , although some excessive gradient appears compared to the observations .
this problem may be due to angular momentum redistribution during the gas infalling or to some inaccuracy of the simple population synthesis technique that works in our code .
the former one is particularly a very complex problem that we shall explore in future works .
the rotation curves obtained for the average mah and @xmath21 for the masses @xmath22 , @xmath23 and @xmath24 and for the cosmological models taken into account here are nearly flat ( fig .
this explains the cosmological nature of the conspiracy between baryon and dark matter in the flat profile of the rotation curves .
the decomposition of the rotation curve for the model with @xmath25 ( @xmath8 ) is presented in fig.3a , where the halo gravitational contraction due to the disk is shown comparing the halo component _ with _ and _ without _ this contraction .
we note here the problem already pointed out in burkert ( 1995 ) ( see also flores & primack 1994 , moore 1994 ) : the gravitational contribution of the halo is dominant until the center .
even if the observational methods to decompose a rotation curve are affected by a large uncertainty , some physical process may be being misundertood here .
recent high - resolution numerical results obtained by kravtsov et al .
( 1998 ) might be showing that the problem is not too serious , and , comparing with our results , suggest that the merging process and the slope of the power spectrum at the scale in consideration are the responsible of giving rise to shallow cores in the dm halos .
if we artificially shallow the central regions of our predicted density profiles in order to be in agreement with the rotation curves of the dwarf and lsb galaxies , then the final rotation curve decomposition results more realistic ( fig.3b ) . in fig . 4 and 5 we compare some global properties obtained for the @xmath6 model normalized to @xmath7 ( for the other models the situation is nearly the same ) with the observations . in fig .
4a and 4b the gas fraction is displayed _ vs. _ @xmath26 color and the central surface brightness , respectively .
the data were taken from a compilation given in mcgaugh and de blok ( 1996 ) and include lsb galaxies . in fig . 5a and 5b the disk - to - bulge ( b / d )
ratio is displayed _ vs. _ the same variables of fig .
it is surprising that the b / d ratios predicted by the models using a simple gravitational instability criterion are in agreement with those inferred from observations ( de jong 1996b ) .
our models reproduce the observational trends across the hubble sequence : _ the redder and more concentrated the disks , the smaller are the gas fractions , and the larger are the b / d ratios_. the models seem to follow a biparametrical sequence , where the two parameters are the color index b - v and the central surface brightness .
these two properties are mainly determined by the mah and by the value of @xmath20 , respectively .
the intensive global properties are almost independent on the third initial fundamental factor , the mass , which is important in defining the extensive properties ( luminosity , radius , etc . ) . in fig .
6 we plot the stellar disk mass @xmath27 _ vs. _ @xmath28 obtained from our simulations for the different cosmological models .
the observed i - band tf relation was transformed to the @xmath29 relation using the mass - to - luminosity radio @xmath30 ( afh98 ) .
for the h - band tfr we have used @xmath31 this mass - to - luminosity ratio is obtained from direct observational estimates in the solar neighborhood ( thronson & greenhouse 1989 ) .
the agreement in the slope as well as in the zero - point ( except for the cobe - normalized scdm model ) between the theoretical ( cosmological ) and empirical @xmath29 relations is excellent , pointing out this to a _ cosmological origin _ for the i- or h- band tf relation .
the tf relation represents a fossil of the primordial density fluctuation field . this structural relation of galaxies
is mainly imprinted by the power spectrum of fluctuations , that for the cdm models at galactic scales produces a variance rather independent on mass , which implies virialized objects whose average densities are almost independent on mass @xmath32 , @xmath33 , or , that is the same , using @xmath34 , whose masses scale as @xmath35 , i.e. since @xmath36 , approximately @xmath37 .
the mass - velocity relation for the dm haloes indeed scales as the velocity to 3.3 - 3.4 ( afh ; see also navarro et al .
1997 ) ; this relation remains almost unaltered for the luminous disks which form within the dm halos .
we have explored the robustness of this result with respect to the baryon fraction and the infall dissipative processes that retain the gas in the halo . in both cases
we find that the tf relation is almost not affected , mainly due to the conspiracy of the gravitational contraction of the dark halo : if the fraction of mass which goes to the disk decreases ( increases ) , then the final rotation velocity also decreases ( increases ) in such a way that the mass - velocity relation remains nearly the same one .
different approaches based on numerical simulations arrive to similar conclusions ( elizondo et al . 1998 ) . regarding the scatter of the tf relation we find that it is produced by the dispersion of the mahs for a given mass and by the statistical distribution of the spin parameter @xmath38 .
as an extreme case we assume that they are independent one from another . for the @xmath39 , @xmath12 and @xmath13 models we predict a total scatter of 0.4 - 0.5 mag that is in marginal agreement with the observations .
it is important to have in mind that in our models , at difference of the observational works , the lsb galaxies are included . for the @xmath6 models ( cobe or @xmath7 normalized ) the scatter is rather high , larger than 0.6 mag .
our models show that the scatter of the tf relation correlates with some galaxy properties . in fig .
7 it is shown how for a given mass the maximum rotation velocity does correlate with the b - v color of the galaxy .
the observational data seem to follow the same trend . in the b - band
we obtain a tf relation with slopes roughly of 3.5 for the different cosmological models .
the observational data give slopes between 2 and 3 .
this discrepancy may be due to the extinction in galaxies and its dependence with mass .
wang & heckman ( 1996 ) , for a large sample of galaxies , have found that the dust opacity increases with the luminosity . using the opacity - b luminosity relation
they give we have `` decorrected '' the b - band luminosities of our models .
after this the slope of the tf relation in the range @xmath40 is well approximated by a line with slope @xmath412.7 , i.e. the predicted tf relation now agrees with the observational estimates ( see fig . 8) .
the seminumerical approach allows us to follow the evolution of an individual galaxy , then we are able to predict how a given galaxy does appear at different redshifts .
9 shows the behavior of @xmath42vs .
@xmath43 . here
we see the difficulty of @xmath13 to produce galaxies with high rotation velocity in the redshift range proper of the damped @xmath44 absorbers ( e.g. , klypin et al .
1995 ) . in fig .
10 we show our predictions for the evolution of the tf zero - point in the h and b bands ( the slopes remain almost constant ) .
the zero - point of the h - band tf relation decreases with @xmath43 because galaxies after reaching their maximum rotation velocity continue aggregating matter without changing this velocity . in the b - band , because the b - luminosity increases with @xmath43 , some compensation happens in such a way that the zero - point of the tf relation remains almost constant as deep field observations have recently shown ( vogt et al .
( 1997 ) ; see fig .
11 displays the disk scale radius _ vs. _ ( 1+z ) .
an inside - out evolution is evident . regarding the sf history we find that it is controlled by the gas accretion ( determined by the mah ) and by the disk surface density ( determined by @xmath20 ) .
for the average mahs and @xmath45 the sf rates in the @xmath8 and @xmath12 cosmological models attain a maximum at @xmath46 depending on the mass .
this rate is approximately 2 - 3 times larger than the present one , a result that is in marginal agreement with deep field observations of a selected population of large spiral galaxies ( lilly et al .
models with low @xmath45 and/or violent mahs attain maximum sf rates that are 3 - 6 times larger than at the present epoch .
the @xmath6 models show an earlier and higher peak in the sf rate . in the case of the @xmath13 model the sf rate peaks at redshifts smaller than 0.5 and it remains almost constant up to @xmath47 .
it is important to remark that our evolutionary tracks concern basically to isolated galaxies , where the environment may supply any amount of gas to a galaxy according to its gravitational field .
our results are not representative of the average conditions of the universe neither of more complex situations as the case of galaxies in clusters .
more work is planned in the future on this direction .
\1 ) the _ seminumerical _ models presented in this work support the viability of an inside - out disk galaxy formation scenario , where the rate of gas accretion to the disk is dictated mainly by the cosmological ( hierarchical ) mass aggregation rate , and where the sf occurs in a stationary self - regulated regime _ within _ the disk .
2 ) the properties , correlations and evolution features of the models are tightly related to three main factors determined by the initial cosmological conditions : the mah , the angular momentum and the mass .
3 ) the infrared tf relation manifest itself as a clear imprint of the power spectrum of fluctuations .
the gravitational pull of the luminous matter on the dark halo makes this relation robust with respect to intermediate processes ( cooling , feedback ) . in the b - band
the luminosity dependent extinction gives rises to a decrement of the tf relation slope ; this may explains the `` color '' tf relation of disk galaxies .
4 ) the scatter of the tf relation for most of the models is in marginal agreement with those derived from the observations .
5 ) the slopes of the tf relations do not change up to intermediate redshifts .
while in the h - band the zero - point decreases , in the b - band it remains almost constant .
6 ) concerning the predictive abilities of different cosmologies with respect to galaxy formation and evolution : i ) @xmath8 , @xmath12 cosmologies are able to predict many of the galaxy features up to intermediate redshift .
ii ) @xmath6 is ruled out because is unable to predict the tully - fisher relation , and gives a too high scatter for this relation .
iii ) @xmath13 is marginal because predicts a too late galaxy formation and sf history . | making use of a seminumerical method we develop a scenario of disk galaxy formation and evolution in the framework of inflationary cold dark matter ( cdm ) cosmologies . within the virializing dark matter halos , disks in centrifugal equilibrium are built - up and their galactic evolution is followed through an approach which considers the gravitational interactions among the galaxy components , the turbulence and energy balance of the ism , the star formation ( sf ) process due to disk gravitational instabilities , the stellar evolution and the secular formation of a bulge .
we find that the main properties and correlations of disk galaxies are determined by the mass , the hierarchical mass aggregation history and the primordial angular momentum .
the models follow the same trends across the hubble sequence than the observed galaxies .
the predicted tf relation is in good agreement with the observations except for the standart cdm . while the slope of this relation remains almost constant up to intermediate redshifts , its zero - point decreases in the h - band and slightly increases in the b - band .
a maximum in the sf rate for most of the models is attained at @xmath0 . |
physics in low dimensional systems has long provided a rich source of fascinating and often unexpected phenomena . with
the steady progress of experimental methods in ultra - cold gases , effective one - dimensional systems are beginning to be realised in the laboratory @xcite .
the integrability of certain many - body problems in one dimension @xcite provides an opportunity to reliably examine many - body quantum physics beyond mean - field theory .
it is becoming clear that many condensed matter theories of phase transitions and collective excitations depend on the number of degrees of freedom within the system @xcite .
in this paper we focus on a system of identical bosons tightly confined in a ring trap such that the system can be considered to be purely one - dimensional ( but with three - dimensional scattering ) .
we are primarily interested in obtaining the excitation spectrum of the system .
the motivation for this work was provided by the prediction of a quantum phase transition in the regime of attractive interactions . as the interaction strength for a fixed number of particles becomes more negative
, zero temperature quantum fluctuations eventually cause the gas to form a soliton - like localised state @xcite .
quantum solitons in 1d bose gases were first predicted and observed with photons in optical fibres @xcite , and more recently soliton - like behaviour has been observed in systems of massive particles with attractive interactions in quasi-1d @xcite and 3d geometries @xcite .
the quantum phase transition in the 1d toroidal geometry was initially identified using a mean - field approach by kavoulakis @xcite and kanamoto _ et .
. further work by kanamoto _
_ has taken a quantum many - body approach using exact diagonalization by truncating the hilbert space to three ( or five ) single - particle states for up to 200 atoms @xcite .
they found a number of interesting physical features in the region of the phase transition , in particular evidence of symmetry - breaking as the difference in energy between the ground and first excited stated tended to zero and scaled as @xmath0 .
whilst the results obtained by kanamoto _
et al . _
@xcite remain qualitatively correct at low temperature and weak interactions , the bethe ansatz provides us with an exact solution for all excited states of the system at all values of interaction strength . the ground state quasi - momenta and energy for the one - dimensional bose gas has previously been calculated using the bethe ansatz by sakmann
_ et al . _
. however , we emphasize that for experimentally realistic temperatures for this system excited states will be involved in both the static and dynamic properties of the system . in this paper
we extend the work of sakman _
et al . _
@xcite to calculate the low - lying excitation spectrum of the finite one - dimensional bose gas as a function of interaction strength for up to @xmath1 particles .
this is a non - trivial result due to the complicated behaviour of the quasi - momenta in the complex plane in the attractive case ( see section [ quasi ] ) . to the best of our knowledge
this work is the first example of calculations made on excited states of this system ( with more than 3 particles ) without any truncation of the hilbert space and without restricting oneself to the limiting _ string _ solutions .
the string solutions arise in the case where the interatomic interactions are sufficiently attractive , or alternatively in the zero density limit .
the problem reduces to one solved by mcguire @xcite and elaborated upon in references @xcite .
the point of interest in our numerical solutions is the deviations of the quasi - momenta from these previously known solutions .
our present work extends our understanding of quantum solitons from the large boson numbers in optical fibres towards the much smaller numbers possible in atomic systems .
related work has been performed by oelkers and links @xcite , who concentrate on a toroidal lattice governed by the bose - hubbard hamiltonian with periodic boundary conditions .
physically one expects the results of their calculations to tend toward those of the continuum as the number of lattice sites increases .
the bose - hubbard hamiltonian is , however , non - integrable and it has been shown by seel _
_ @xcite that the spin-1/2 xxz heisenberg chain ( which is integrable ) maps onto the 1d bose gas model in the continuum limit .
other examples of lattice hamiltonians that have the 1d bose gas in the continuum limit can be found in @xcite .
the structure of this paper is as follows . in section [ model ]
we review the work of lieb and liniger @xcite in order to correctly pose the problem . in section [ quasisect ]
we discuss our method of solution and present the results , whilst also discussing its connection to earlier work by mcguire @xcite . in section [ exc_spect ] we present the excitation spectrum of the system obtained from our calculations , and in section [ comparison ] we show a comparison of our work to that of the truncated hilbert space approach of reference @xcite .
finally we calculate the excitation spectrum for the repulsive case in section [ sec : repulsive ] , before concluding in section [ sec : conclusions ] .
we are considering a system of @xmath2 bosons , each with a mass @xmath3 , on a ring of radius @xmath4 , in the limiting case of the radial confinement _ freezing _ out all the transverse degrees of freedom of the atoms .
the interactions are assumed to be short range , contact interactions , modelled using the delta function . with these assumptions
the system will be described by the hamiltonian ( in first quantised formalism ) @xmath5 where length is measured in units of @xmath4 and energy is measured in units of @xmath6 ( @xmath7 is the angular coordinate of the @xmath8th particle ) .
the parameter @xmath9 is related to the s - wave scattering length by eq . , and quantifies the strength of the two body interactions , @xmath10 describes an attractive gas while @xmath11 describes a repulsive gas .
since our spatial universe is of course three dimensional it is worthwhile to note the physical limits of applicability of this model . assuming harmonic confinement , the length scale of the dimensions tranverse to the ring are @xmath12 and @xmath13 .
another important length scale is the s - wave scattering length , @xmath14 , that represents the strength of two body interactions .
the healing length of the gas @xmath15 , where @xmath16 is the three dimensional density , reveals how sensitive the gas is to irregular features in the potential . for the system in which we are interested , it is possible to define a _
local _ set of cartesian coordinates such that @xmath17 , where the @xmath18 axis runs tangential to ring .
essentially this amounts to a torus of radius @xmath4 and cross - sectional area @xmath19 satisfying the condition @xmath20 . in this way
the energy level spacings for the tranverse ( @xmath21 and @xmath22 ) degrees of freedom supersede all other energy scales within the system and the system is essentially one dimensional .
therefore we can assume a ground harmonic oscillator state in the transverse dimensions , and integrate them out to obtain an effective one dimensional interaction @xcite @xmath23 where @xmath3 is the mass of a single particle .
some interesting work has recently been done by parola et .
@xcite on the quasi - one dimensional limit which indicated a transverse collapse ( in addition to the angular collapse ) at a certain critical interaction strength . in this work they took into closer consideration the effects of the excluded two dimensions .
current experimental realisations of such ring traps @xcite are still far from the quasi-1d regime @xcite . given the hamiltonian , schrdinger s equation is @xmath24 where @xmath16 is the label for different eigenstates .
equation can be solved exactly via the bethe ansatz @xcite . due to the symmetry under permutation of particle coordinates , the region over which we need to integrate eq .
can be restricted to @xmath25 .
lieb and liniger noted that in this region the delta function vanishes everywhere except along the boundaries where @xmath26 , hence the problem could be cast in a different manner by writing the interactions as a _
boundary condition _ on the restricted region rather than an explicit term in the hamiltonian @xcite .
the new problem was to find the appropriate solution to @xmath27 where eq .
is the previously mentioned boundary condition arising from the interaction .
furthermore the boundary conditions pertaining to the ring geometry of the system , i.e. @xmath28 , must also be recast onto the new region , vis [ bc ] @xmath29 the bethe ansatz is employed as a means of integrating eq . .
the ( unnormalized ) wave function in the restricted region is given the form @xmath30 where @xmath31 is some permutation of the integers @xmath32 , @xmath33 is the @xmath34th element in @xmath31 and the sum runs over all the @xmath35 distinct permutations of @xmath31 .
equation is one form of the bethe ansatz and was originally used as a means of integrating spin chain hamiltonians @xcite .
the coefficients of each permutation , @xmath36 , can be determined ( up to some arbitrary phase ) by the interaction boundary condition eq . .
we choose the convention @xmath37 and for any other permutation @xmath31 , @xmath36 is found by constructing @xmath31 out of @xmath38 by transposing adjacent elements .
then @xmath36 is a product of the terms @xmath39 where @xmath40 and @xmath41 are the adjacent elements being transposed , with @xmath40 to the left of @xmath41 _ before _ the transposition .
the scalar quantities @xmath42 , known as the _ quasi - momenta _ ( and sometimes the rapidities ) , are determined by the ring geometry of the system , i.e. , by the boundary conditions in eqs . .
these quasi - momenta furnish important quantities for the system such as the energy eigenvalues and the total momentum , however it should be stressed that ( other than the case of no interactions @xmath43 ) the individual quasi - momenta do not have a direct relationship to the single particle momentum states . the latter must be obtained from the single particle reduced density operator , which can be obtained from the many body wave function .
substitution of eq . into the boundary conditions and ,
yield the following set of @xmath2 simultaneous equations to determine the quasi - momenta as functions of @xmath9 @xmath44 the purpose of this work is to explore solutions of these equations to reveal key features of the system such as the excitation spectrum of the gas as a function of the interaction strength .
analytical solutions of eq . have previously been found for up to three particles @xcite , however to our knowledge the only numerical calculations have been for the ground state energy @xcite .
once eqs .
have been solved it is true that , in principle at least , one has obtained all neccessary information to obtain the wave function for the system .
however in practice , even for a modest number of particles such as ten , the wave function itself will involve a summation over @xmath45 terms , a somewhat cumbersome object .
however , for certain physical quantities , simple analytical expressions involving only the quasi - momenta can easily be derived , for instance the energy of the @xmath16th eigenstate is given by @xmath46 and likewise the total momentum by @xmath47 we will utilise equation in section [ exc_spect ] to obtain the excitation spectrum .
the behaviour of the bose gas with attractive interactions has received far less scrutiny than its repulsive counterpart .
originally lieb and liniger did not consider this regime to be of physical relevance @xcite , however mcguire showed @xcite that once an @xmath2-particle bound state had formed , the ground state energy of the system scales as @xmath48 this posed problems for the solution in the thermodynamic limit where @xmath49 , @xmath50 and @xmath51 constant .
in contrast for the repulsive case it is found @xmath52 , and hence an expression for the energy density could be found @xcite .
however @xmath53 diverges in the attractive regime .
it is perhaps worth noting that the limit of an extremely dilute gas is free from these divergences , provided one took the limit as @xmath54 constant .
another possibility is to consider the limit @xmath50 while @xmath2 remains constant @xcite .
as previously mentioned the quantum phase transition to a soliton - like state provides added interest for the attractive gas .
this transition spontaneously breaks the translational symmetry of the system even at zero temperature .
we discuss this point further in section [ quasi ] and see how the eigenstates obtained from the bethe ansatz maintain the translational symmetry of the hamiltonian by forming a continuous superposition of localised states around the ring .
we have included in appendix [ meanfieldc0 ] a simple derivation ( following the work of kavoulakis @xcite ) of the critical interaction strength ( @xmath55 ) at which the phase transition is predicted to occur . [ quasisect ] to find the quasi - momenta @xmath56 that fully characterise the exact solution , we must numerically solve eqs . .
for @xmath2 particles this gives @xmath2 simultaneous nonlinear equations .
the nature of most root - finding algorithms is such that it requires a reasonable initial guess for the location of the roots .
we have developed a suitable procedure to solve these equations for a range of interaction strengths @xmath9 starting from the known solutions for the ground and excited states at @xmath43 , which are given by all @xmath57 being an integer . to find the solution for non - zero @xmath9 ,
we first choose a value for @xmath9 close to zero , and use the ideal gas solution for the state we are interested in as the initial guess for the root - finding algorithm .
we make use of the built - in root - finder ` fsolve ` in the software package matlab on a standard desktop pc . by choosing a small enough @xmath9
the initial guess for the quasi - momenta are close to the actual solution , and the root - finding algorithm converges relatively rapidly .
once the first solution close to @xmath43 is found , @xmath9 is decreased in small increments @xmath58 and an initial guess for the quasi - momenta for the new value of @xmath9 is based on a smooth extrapolation from the previous values . typically the extrapolation is based only on the two previous values of @xmath9 and is thus linear .
difficulties arise in this method when a particular quasi - momenta at an interaction strength @xmath9 differs greatly from that at @xmath59 .
this problem arises particularly when @xmath9 is close to the critical interaction strength @xmath60 and we discuss how we deal with the problem later
. the actual size of @xmath58 will be dependent on how large @xmath2 is .
the scaling behaviour of the critical interaction strength goes as @xmath61 and we find that the efficiency of the algorithm is improved by using @xmath58 as small as @xmath62 .
it is possible to use a larger value of @xmath58 but we have found that smaller step sizes give better initial predictions for the roots for the next step and results in faster convergence of the root - finding algorithm .
we have not found it necessary to make use of high precision arithmetic used by sakmann _
@xcite in finding our solutions . in their work
they make use of a particular transformation [ see eq .
( 33 ) in ref .
@xcite ] popularised in the seminal work of lieb and liniger @xcite .
the transformation simplifies the numerics in the repulsive case by _ spreading _ the transformed quasi - momenta into numbers which can be easily distinguished by machine precision .
the same is not true in the case of attractive interactions . as pointed out by sakmann _
et al . _ , working with the transformed quasi - momenta in the attractive case requires numerical precision of approximately @xmath63 . in this work
we solve eqs . directly .
we set the tolerance of the ` fsolve ` algorithm to iterate until eqs .
are solved such that the left hand side equals the right hand side to an absolute accuracy of @xmath64 .
once the value of @xmath65 has been increase to be point that the quasi - momenta are within @xmath64 of the string solution the algorithm terminates .
we present the results of these calculations in two different ways .
figures [ sixpartkatt ] , [ attractive_limit1 ] , and [ attractive_limit2 ] show the continuous evolution of the real and imaginary parts of the quasi - momenta , for a number of excited states as @xmath9 becomes more negative .
figure [ sixpartkatt ] shows eight excited states for @xmath66 , whereas figs .
[ attractive_limit1 ] and [ attractive_limit2 ] show two representative excited states for @xmath1 .
figures [ comp1 ] , and [ comp2 ] show snapshots of the distribution of the quasimomenta in the complex plane for @xmath1 .
our calculations elucidate the _ deviations _ from the string solutions at finite density @xcite .
we see from fig .
[ sixpartkatt](b ) and ( e ) that as the real parts of @xmath57 and @xmath67 become equal the imaginary parts _ bifurcate _ , and vice - versa .
these splittings of the quasi - momenta provide additional challenges for the root - finding algorithm . because the splitting often occurs sharply , the initial guess for the quasi - momenta found from the previous value of @xmath9 can be sufficiently inaccurate that the algorithm does not converge . thus if one finds that the root - finding algorithm is not converging at a particular value of @xmath9 then this could be an indication that two equal quasi - momenta are beginning to split .
when this occurs there is always a degeneracy between the quasi - momenta as to which one _ goes up _ and which one _ goes down _ ( this degeneracy is of no _ physical _ consequence , it is merely a mathematical hurdle ) .
the numerical fluctuations of the values of the quasi - momenta is usually somewhere close to machine precision , @xmath68 , and are difficult to detect . in order to observe the splitting we found it was necessary to manually alter the initial guess for the quasi - momenta at these points , forcing one to _ go up _ and one to _ go down _ ( again see fig .
[ sixpartkatt ] ( b ) and ( e ) ) .
particles as as a function of the interaction strengnth @xmath9 .
( a ) and ( b ) show the real and imaginary parts respectively of the quasi - momenta .
the blue solid lines are the roots calculated using the bethe ansatz , whereas the red dashed lines show the string solutions given by eqs . which give the asymptotic behaviour of the quasi - momenta in the attractive limit .
the two solutions are in good agreement above the critical interaction strength , validating the essential properties of the _ bound states _ described in references @xcite .
the inset shows the region where the quasi - momenta _ bifurcate _ causing numerical difficulties.,width=264 ] particles as as a function of the interaction strengnth @xmath9 .
( a ) and ( b ) show the real and imaginary parts respectively of the quasi - momenta .
the blue solid lines are the roots calculated using the bethe ansatz , whereas the red dashed lines show the string solutions given by eqs .
with @xmath69 the inset shows where the quasi - momenta _ bifurcate _ for this case.,width=264 ] there is no fundamental limit to the excited states we can reach with this procedure , and similar graphs can be produced for up to approximately @xmath1 particles ( see fig . [ attractive_limit1 ] and [ attractive_limit2 ] as examples of the @xmath1 particle gas ) . in principle
this work could be extended to larger numbers of particles , and in fact we have found the ground state for up to @xmath70 . however , the _ bifurcations _ in the excited states require some time to locate and we have had to deal with them manually .
this means that the algorithm begins to take a considerable amount of time for @xmath2 significantly larger than twenty .
roots of eq . for @xmath10 have a far more complicated behaviour than in the more commonly studied repulsive regime .
apart from the obvious reason that the quasi - momenta can be complex ( as opposed to strictly real for the repulsive gas @xcite ) , there also exists the change in behaviour at the critical interaction strength @xmath71 . for some time
it has been suspected that the roots for large enough attraction would take values on _ strings _ in the complex plane which correspond to some bound state of the system @xcite .
that is , the wave function would represent some kind of localised state @xcite such as @xmath72 and the corresponding quasi - momenta of this state would be @xmath73 where @xmath74 is the total momentum of the state .
it is worthwhile to note that although we refer to the wave function in eq . as a _ localised state _
it is not localised to any specific point .
rather it has a localised pair correlation @xcite @xmath75 where @xmath76 ( @xmath77 ) are bosonic field operators which annihilate ( create ) a particle at position @xmath78 .
thus measurements of the density of an ensemble will always yield a localised projection , whilst the system is not localised _
prior _ to the measurement .
one can think of the system prior to the measurement as being in a ( macroscopic ) superposition of localisation at _ every _ point on the ring .
this form for the quasi - momenta given by eqs .
is commonly referred to in the literature as a _ string _
solution @xcite .
it is often assumed that multiple strings may exist for one system ( @xcite and references therein ) , with each string corresponding to a soliton of different momentum .
this choice of quasi - momenta also seems intuitively reasonable and gives the result , mandated by mcguire @xcite for the total energy of the bound state @xmath79 the degree to which the system is localised by these string solutions can be quantified by the absolute values of the imaginary parts of the quasi - momenta .
thus , in eqs .
, as @xmath65 increases , so does the degree of localisation ( as one would naively expect ) .
however , it is easy to show that in the ideal limit , @xmath80 , the quasi - momenta are exactly the single particle momentum states . thus if one accepts the quasi - momenta given by eqs .
, then one is left with chronic discontinuities in the behaviour of the quasi - momenta as a function of @xmath65 .
our results give direct evidence that the two limits are bridged via eqs . .
we also find other families of excited states , with multiple bound states forming in the limit of strong attraction . these show agreement with the truncated diagonalization results discussed later in section [ comparison ] .
these families of excited states can be interpreted as the formation of multiple bound states .
it is very interesting to observe this intermediate behaviour between the ideal gas and strongly attractive gas since this is where soliton formation occurs at zero temperature . in fig .
[ sixpartkatt ] we see the behaviour of the quasi - momenta for a system of @xmath66 particles in the eight lowest energy states ordered via their energies at @xmath43 . as @xmath65 becomes larger , energy level crossings are observed as in fig .
[ sixpartbetheatt ] .
the quasi - momenta begin at @xmath43 at their respective single particle momentum states , and as @xmath9 becomes more negative the quasi - momenta tend toward a string solution as predicted .
the initial behaviour of the imaginary part of the quasi - momenta goes like @xmath81 for @xmath82 , but for @xmath83 when the system is in a localised state the low - lying eigenstates have quasi - momenta corresponding to the string solutions eqs . .
this is verified again for the @xmath1 particle gas in fig .
[ attractive_limit1 ] where the solid blue lines show the quasi - momenta as a function of @xmath84 and the red dashed lines show the string solution eqs . , and
both are in agreement in the strongly attractive limit .
however , we also find excited states that do not agree with eqs . , but instead form multiple bound states .
this results in a decrease in the degree of localisation . in these cases
we find the behaviour of the quasi - momenta at large @xmath65 to be given by @xmath85 we interpret this solution physically as an @xmath86 particle bound state with momentum @xmath87 and an @xmath88 particle bound state with momentum @xmath89 .
evidence of these states can be seen in fig .
[ attractive_limit2 ] where the solid blue lines show the quasi - momenta as a function of @xmath84 , whereas the red dashed lines show the solutions given by eqs . with @xmath69 , @xmath90 and @xmath91 .
in fig . [ sixpartkatt](d ) and ( f ) the same kind of behaviour is seen for the @xmath66 particle gas .
the behaviour is perhaps clearer in figs .
[ comp1](c ) and [ comp2](c ) where one can see the two distinct strings in the complex plane .
one can derive a simple expression for the energy of these states ( similar to eq . ) , @xmath92 this trend continues such that three or more bound states of atoms occur within the system .
the evidence for this is most easily seen by the grouping of the quasi - momenta in the complex plane [ figs .
[ comp1](c ) and [ comp2](c ) ] .
thus for this system once @xmath93 and localisation has occurred we observe several different families of solutions .
the first is the ground state of the system , corresponding to a single , stationary @xmath2-particle soliton .
then there are the elementary excitations of this state whereby the soliton has some ( integer valued ) total momentum about the ring .
furthermore there exist the higher order excitations in which multiple solitons form around the ring .
equally these multiple solitons can have elementary excitations of their own , corresponding to some integer valued total momentum .
we can now calculate the excited states of the system via the bethe ansatz , using eq . and the numerically determined roots of eqs .
( the quasi - momenta ) .
we show a comparison between the truncated diagonalization approach of kanamoto _
et al . _
@xcite in section [ comparison ] ( see figs .
[ betheexactcompare ] and [ betheexactcompare2 ] ) .
we plot the excitation spectrum for @xmath66 particles in fig .
[ sixpartbetheatt ] and @xmath1 particles in fig .
[ energy20 ] .
the point at which mean - field theory predicts the quantum phase transition , @xmath71 ( see appendix [ meanfieldc0 ] ) , is indicated on all figures with a vertical dot - dashed line .
we observe the _ sharpening _ of the cross - over regime as @xmath2 increases as expected for a phase transition in the limit of large @xmath2 .
particle system obtained via the bethe ansatz .
the solid red lines show the formation of a single bound state .
the green dashed lines show the formation of two bound states .
the blue dot - dash lines show the formation of three separate bound states .
the qualitative change in the behaviour of the eigenstates at the phase transition point becomes more visible than for the two particle case ( see fig .
[ betheexactcompare]).,width=302 ] particle system obtained via the bethe ansatz .
the solid red lines show the formation of a single bound state .
the green dashed lines show the formation of two bound states .
the blue dot - dash lines show the formation of three separate bound states.,width=302 ] once the size of the system gets up to around @xmath1 , the difference in the behaviour of the states for @xmath82 and @xmath83 is clear . for @xmath83 transition (
once the bound state has formed ) , the different families of solutions to the bethe eqs .
( discussed in the previous section ) become distinguishable through their separation from the ground state . the solid red lines in figs .
[ sixpartbetheatt ] and [ energy20 ] show the behaviour of the single bound states found by mcguire and others @xcite given by eqs .
, the green dashed lines show the family of solutions given by eqs . which are made up of two separate bound states .
finally the blue dot - dashed lines show the existence of three separate bound states within the gas .
particle system .
the vertical dash - dotted line marks the mean field crititical interaction strength @xmath71 derived in appendix [ meanfieldc0 ] .
( a ) single particle momentum state cut - off of @xmath94 .
( b ) single particle momentum state cut - off of @xmath95 , and so on , up to ( f ) which has a single particle momentum cut - off of @xmath96.,width=340 ] particle system .
( a ) truncation at @xmath97 ( single particle momentum states @xmath98 ) .
( b)truncation at @xmath99 ( single particle momentum states @xmath100 ) .
, width=302 ] the excitation spectrum of the attractive gas has been numerically studied via truncation of the hilbert space in previous work @xcite . the key element in the approach
is to exclude all single particle momentum states beyond a particular threshold .
the results of these calculations are of course exact at zero interactions , but deviate from the exact solution as the interaction strength , @xmath65 , is increased .
the degree of deviation can of course be decreased by increasing the threshold momentum state , however the dimension of the hilbert space will eventually become unmanageably large in doing so .
an example of this process is shown in fig .
[ betheexactcompare ] for a two particle system where it is possible to make the cut - off relatively high without having the hilbert space dimension expand beyond a practical number .
the results indicate a good qualitative description from a very small number of momentum states @xcite .
for example if @xmath101 , a momentum cutoff of @xmath94 is reasonable up to and even beyond ( @xmath102 ) the critical point .
this was supported by calculations made by kavoulakis @xcite who used a mean field approximation with a suitably chosen wave function to examine the ground state and low - lying states near the phase transition .
the truncated diagonalization procedure we use follows that of kanamoto _
et al . _
the first quantised hamiltonian in eq .
is rewritten in second quantised formalism as , @xmath103,\label{secondham}\end{aligned}\ ] ] where @xmath76 is the bosonic field operator that annihilates a boson at coordinate @xmath78 and obeys the usual commutation rules and the periodic boundary conditions of the ring geometry .
this field operator is then approximated by its truncated expansion in terms of single particle states @xmath104 where @xmath105 is the single particle eigenstate with angular momentum @xmath34 ( see appendix [ singleparticle ] ) , @xmath106 annihilates a boson with angular momentum @xmath34 and @xmath107 is the single particle momentum state cut - off . in this manner , for finite @xmath2 and finite @xmath107 , @xmath108 is a finite matrix .
furthermore , @xmath108 ( when written in the appropriately ordered basis set ) will be block diagonal , as it can be divided into total momentum subspaces which can be individually diagonalised .
if we use @xmath3 to denote the total number of single particle momentum states used in the expansion ( i.e. @xmath109 ) then the total dimension of the hilbert space will be the binomial coefficient @xmath110 .
this hilbert space will split up into @xmath111 ( total momentum ) subspaces .
the largest subspace will be the zero momentum subspace .
for example if we were to consider a gas of @xmath1 particles with the momentum cut - off at @xmath95 ( @xmath99 ) , then the total dimension of the truncated hilbert space would be 10626 .
however this splits up into 81 different total momentum subspaces , the largest of which has dimension 318 , which can be exactly diagonalised on a standard desktop pc .
table [ hilbertdiminfo ] gives an indication of the scale of computation required for the truncated diagonalization procedure . in fig .
[ betheexactcompare2 ] we show a comparison between excitation spectrums obtained from the truncated diagonalization procedure and the bethe ansatz for the @xmath1 particle gas .
we observe that for @xmath97 ( single particle momentum cut - off at @xmath94 ) the truncated diagonalization over - estimates the energy difference between ground and excited states in the cross - over regime , but under - estimates the difference in the regime where the bound states have formed . for @xmath99 ( single particle momentum cut - off at @xmath95 ) the agreement between the two methods in the cross - over regime is vastly improved , however the truncated diagonalization still underestimates the difference between ground and excited states in the bound state regime .
.the size of the hilbert space and subspaces for selected particle numbers and momentum cut - offs .
this gives an idea of the size of computation required for the truncated diagonalisation procedure . [ cols="^,^,^,^,^ " , ]
the root - finding algorithm can easily be applied to the case of repulsive attractions .
a proof exists showing that all roots of eqs .
lie on the real axis @xcite .
this means that we avoid the problems associated with the bifurcations which appeared in the attractive case and the situation is therefore significantly easier . in this case
the limiting behaviour as @xmath112 of the quasi - momenta is given by the single particle momentum states of an ideal fermi gas @xcite .
this limit is known as the tonks - girardeau gas and has been experimentally observed @xcite . in this work
we use our numerical root - finding algorithm to bridge from the ideal gas to the tonks - girardeau gas .
we do so for both the ground state and the low - lying excited states for finite numbers of particles .
the ground state has appeared previously in the work of sakmann _ et al . _
the thermodynamic limit has been addressed in the homogeneous case @xcite and the trapped case @xcite ( and references therein ) .
the root - finding algorithm is essentially unchanged from the case of attractive interactions .
it is worth mentioning however that , since there is no phase transition in the repulsive gas the step size @xmath58 can be increased without sacrificing a reasonable initial guess . in our calculations we use @xmath113 for @xmath114 , then @xmath115 for @xmath116 , then @xmath117 for @xmath118 , then @xmath119 for @xmath120 and finally @xmath121 for @xmath122 . in fig .
[ repulsivek ] we see how the quasi - momenta behave as a function of the interaction strength @xmath9 . the single particle excitations of the ideal _ bosons _ at @xmath43 result in single particle excitations of the ideal _ fermions _ at @xmath123 .
the concept of particle - hole excitations for the strongly repulsive gas ( see references @xcite ) is clear in fig .
[ repulsivek ] .
we see low - lying excited state quasi - momenta are similar to that for the ground state , but with a small number of quasi - momenta excited , leaving behind _
it is possible to push the algorithm for the repulsive case to much higher atom numbers than with the attractive case .
this is due to the absence of the exotic features in the quasi - momenta spectrum ( see for example the splitting occurring in the insets of figs [ attractive_limit1 ] and [ attractive_limit2 ] ) . in fig .
[ largen ] we show the ground state quasi - momenta for a repulsive gas with @xmath124 .
particle repulsive gas.,width=325 ] as for the attractive gas , the excitation spectrum can be found using eq . .
[ repulsiveenergy ] we show the energies of the low lying excited states for @xmath11 .
the red dashed lines show the result obtained from truncating the hilbert space down to five single particle states and diagonalising the hamiltonian ( see section [ comparison ] ) .
the truncation procedure is accurate for small enough values of @xmath9 ( see the inset of fig .
[ repulsiveenergy ] where @xmath9 runs from zero to one ) , but fails as @xmath9 increases beyond one .
the dotted line shows the ground state energy of a @xmath1 particle fermi gas .
particle gas with repulsive interactions @xmath11 ( solid black lines . the red dashed lines were obtained by truncating the hilbert space down to five single particle momentum states .
the inset shows the region of @xmath125 where the truncated diagonalization is accurate .
the dotted lines shows the ground state energy of a free fermi gas , and the so - called tonks girardeau ground state asymptotes to this value.,width=325 ]
we have studied the behaviour of the roots of the bethe ansatz eqs . for the 1d bose gas on a ring as a function of the interaction strength for both the attractive case ( up to @xmath1 particles ) and the repulsive case .
we used these roots to determine the exact many - body energy spectrum of the one - dimensional bose gas with @xmath126-function interactions in a 1d ring trap for small numbers of particles .
we compared our results to those obtained via the approximate method of truncated diagonalization , to which we found good qualitative agreement and reasonable quantitative agreement for small enough interaction strength .
we have quantitatively addressed the issue of when it is reasonable to assume a bound state solution to eqs . , and have quantified the deviations from the string solutions that occur for a finite density gas , thus establishing the point at which the string solutions become valid .
furthermore we found evidence for the existence of multiple , independent _ solitons _ existing on the ring .
these excitations had been hypothesised in previous work , however to the best of our knowledge this is the first direct evidence of their existence .
we have described the analytical expressions for the quasi - momenta of these new families of string solutions in eqs . ) and an expression for the energies of these states .
it may be interesting in the future to make use of the solutions found in this paper to calculate correlation functions for the one - dimensional bose gas .
this work was supported by the australian research council centre of excellence for quantum - atom optics and the queensland state government .
the authors wish to acknowledge the useful discussions with all members of the university of queensland node .
ags wishes to particuarly thank chris foster and david barry for technical help with computational tasks and j. s. caux and norman oelkers for useful discussions .
the single particle hamiltonian is @xmath127 and the single particle energy eigenvalues are @xmath128 with corresponding eigenstates are @xmath129
we include a derivation of the mean - field critical point of localisation using the same approach as kavoulakis @xcite .
considering the hamiltonian we define a mean field @xmath130 and approximate the system as having the many body wave function @xmath131 .
we next expand @xmath130 in terms of single particle states given by eqs .
, however we truncate all momentum states greater than one , @xmath132 we then assume @xmath133 and from the symmetry @xmath134 , and finally the normalisation condition @xmath135 .
the energy per particle @xmath136 is then found to be @xmath137,\label{energyperparticle}\end{aligned}\ ] ] where @xmath138 .
setting @xmath139 , then @xmath136 is minimised by choosing @xmath140 , the normalisation condition further simplifies eq . to @xmath141 from eq .
it is straightforward to see that the approach one must take in order to minimise @xmath136 will depend on whether @xmath142 ( in which case the gas is delocalised ) or @xmath143 ( in which case the gas is localised ) .
thus there exists a critical interaction strength of @xmath144 | the exactly solvable lieb - liniger model of interacting bosons in one - dimension has attracted renewed interest as current experiments with ultra - cold atoms begin to probe this regime . here
we numerically solve the equations arising from the bethe ansatz solution for the exact many - body wave function in a finite - size system of up to twenty particles for attractive interactions .
we discuss the novel features of the solutions , and how they deviate from the well - known string solutions [ h. b. thacker , rev .
mod .
phys . * 53 * , 253 ( 1981 ) ] at finite densities .
we present excited state string solutions in the limit of strong interactions and discuss their physical interpretation , as well as the characteristics of the quantum phase transition that occurs as a function of interaction strength in the mean - field limit .
finally we compare our results to those of exact diagonalization of the many - body hamiltonian in a truncated basis . we also present excited state solutions and the excitation spectrum for the repulsive 1d bose gas on a ring . |
the effective nucleon - nucleon ( nn ) interaction is expected , on physical grounds , to be intrinsically non - local .
the quark structure of nucleons implies that the potential between two interacting nucleons can not be expressed only in terms of the center of mass degree of freedom , at least for distances smaller than twice the nucleon radius .
even a description of the nn interaction in terms of one - boson exchange processes introduces some degree of non - locality @xcite . for practical purposes , however , it is convenient to have a local ( energy independent ) nn interaction , to be used in numerical applications .
the modeling of the nn interaction by a local potential , with possible guidance from meson exchange processes , has been extremely successful in reproducing the nn experimental phase shifts and deuteron properties @xcite .
this approach has developed for many years and has reached a high degree of sophistication .
it has been established by now that these local nn potentials , which are essentially phase equivalent , give slightly different results for the three- and four - nucleon systems .
furthermore , in general they underestimate the binding energy and do not give the correct values for some polarization observables .
this drawback has been usually overcome by introducing three - body forces , which can be again justified by the non - elementary nature of the nucleon .
actually three - body forces can be generated also by meson - exchange processes of higher order , where the nucleon is excited in the intermediate states of the interaction processes or meson - meson couplings are introduced @xcite .
both non - locality and three - body forces can be considered to have a common origin , the inner structure of the nucleon .
however , one can not fully reduce one to another , since non - locality has definite effects already at the two - body level , e.g. a non - local potential can not be equivalent , in general , to an energy independent local potential .
furthermore , it is not clear to what extent the off - shell behavior of a non - local potential , i.e. the off shell @xmath1-matrix , can be simulated by the presence of three - body forces .
another question where non - locality could play a role is the mechanism of saturation in nuclear matter .
in fact , it turns out that in order to get saturation with a local ( energy independent ) interaction , it is essential to include a strong repulsive core , as also suggested by the behavior of the nn phase shifts .
again , three - body forces are necessary to tune the saturation point close to the empirical one .
it is surely conceivable that a certain degree of non - locality could simulate the strong repulsion at short distance , which is one of the main features of the nn interaction , and maybe reduce the need of three - body forces .
it appears , therefore , of great interest to investigate to what extent non - locality alone could enable to get saturation , possibly close to the empirical one , and at the same time to describe three- and four - body systems correctly .
recently , in a series of papers by doleschall and collaborators @xcite , a new realistic non - local nn potential was proposed , which is able to describe accurately both the nn phase shifts and the properties of three - nucleon systems ( @xmath2 and @xmath3 ) , i.e. their binding energies and radii .
also the alpha particle appears to be reasonably well described @xcite .
the need of three - body forces in this case seems to be absent , or completely simulated by the non - locality .
this nn interaction introduces non - locality at short distance ( @xmath4 @xmath5 ) for the s - wave channels , and possibly also for the p - waves ones @xcite .
the rest of the potential is taken as the argonne v@xmath0 @xcite , while the non - local part is purely phenomenological . in this paper
we address the issue of the saturation properties of non - local potentials . to this purpose
we present nuclear matter calculations within the bethe - brueckner - goldstone method extended up to three hole - line contributions , which is known to converge well and to give an accurate nuclear matter saturation curve . as the main representative of non - local nn interaction we take the potential of ref .
we compare the corresponding nuclear matter equation of state ( eos ) with those obtained starting from the charge - dependent ( cd ) bonn @xcite and the v@xmath0 potentials .
the former can be considered the most recent and accurate interaction based on the one - boson exchange model and , therefore , it contains some degree of non - locality .
the latter is a convenient reference local potential for the comparison with doleschall s , which , as we stated above , coincides with the v@xmath0 in the @xmath6 channels .
additionally , other modern local potentials , like argonne v@xmath7 @xcite or paris potential @xcite , give similar results , close enough to v@xmath0 to be considered equivalent for the present purpose @xcite .
the study of phase - shift equivalent nn interactions and their predictions of nuclear matter saturation properties has a long history .
for example in refs .
@xcite unitarily equivalent interactions which differ in the short range part of the @xmath8 channel were shown to provide rather different values of the saturation energy and density .
more recently , the predictions of modern nn potentials were studied in refs .
@xcite . here
the role of non - locality in the binding energy of nuclear matter was also partially discussed , using the cd bonn and the nijmegen - i @xcite potentials , which were shown to provide a stronger binding than local potentials such as the v@xmath0 . however , both these potentials underestimate the binding energy of three- and four - body systems @xcite , at variance with doleschall s interaction @xcite .
therefore when using the latter we can , in principle , expect different results also for nuclear matter calculations .
it is well known that the saturation points obtained within the brueckner approximation for the gap choice and different nn interactions tend to lie along a band @xcite , the celebrated coester band " , in the energy - density plane .
the position inside the band for a given nn interaction turns out to be correlated with the d - state probability in the deuteron .
the high - energy , high - density part of the band is usually associated with low values of this probability . however , as shown in ref .
@xcite , this trend does not hold any more if three hole - line contributions are included , as we do in the present paper . in this case
the results are quite insensitive to the d - wave probability value and actually close to each other , at least for local potentials or quasi - local potentials like the paris one .
the same conclusion holds true if the continuous choice for the single particle potential is used in the brueckner approximation @xcite .
the paper is organized in the following way . in section [
s : pot ] we shortly study the free space properties of the doleschall , cd bonn and v@xmath0 potentials . in section [ s : eos ] , after introducing the bbg formalism , we study the eos of nuclear matter obtained with doleschall s potential , addressing , in particular , the issue of the convergence of the hole - line expansion in this case .
we then compare this eos with those obtained starting from the two other potentials .
finally in section [ s : concl ] we draw our conclusions .
( solid line ) , cd bonn ( dotted ) potentials and of doleshall s is potential of ref .
@xcite ( dot - dashed , labeled as nl ) , for the np interaction in the @xmath8 ( upper panel ) , @xmath9 ( middle ) and in the coupling @xmath9-@xmath10 ( lower panel ) channels .
the on - shell @xmath11 @xmath12-matrix elements are also shown ( symbols ) , for laboratory energies ranging from @xmath13 to @xmath14 mev , in steps of 50 mev . ] before studying infinite nuclear matter and in order to better understand the role played by the different nn interactions in the nuclear matter calculations of next section , it is useful to briefly compare the properties and characteristics of the potentials we are going to employ , on the basis also of some established results obtained in the literature @xcite the general matrix elements in the momentum representation of a nn interaction @xmath15 in a given two - body channel @xmath16 can be written v^c_ll(k , k ) = 2 r^2 dr r^2 dr j_l(kr ) v^c_ll(r , r ) j_l(kr ) [ eq : v ] where @xmath17 are the initial and final orbital angular momenta in the channel , and @xmath18 the corresponding spherical bessel functions . for a local potential @xmath19 . in the momentum representation
the matrix elements @xmath20 of eq .
( [ eq : v ] ) are in any case a function of two variables , the initial and final relative momenta @xmath21 and @xmath22 .
the non - locality is incorporated in a non trivial way in the analytical properties of the matrix elements of the potentials , as well as of the corresponding @xmath1-matrix , or of the real @xmath12-matrix , which is used here for later convenience .
the latter describes the free space nn scattering and is obtained by solving the lippmann - schwinger equation , which for two nucleons interacting with energy @xmath23 in the center of mass frame can be written as : & & r^c_ll(k , k;e ) = v^c_ll(k , k ) + [ eq : lippmann ] + & & + , where the symbol @xmath24 denotes the principal value integral and @xmath25 is the free nucleon mass . for realistic potentials the on - shell @xmath12-matrix , @xmath11 with @xmath26 ,
is fixed , to a large extent , by the fitting of the experimental nn phase shifts .
the potentials we consider in the present paper were fit to reproduce the nijmegen phase shifts @xcite for energies up to @xmath14 mev in the laboratory frame , and therefore give rise to practically identical on - shell values of the @xmath12-matrix in this energy range .
however , due to their totally different analytical structure , a quite different behavior of the bare nn matrix elements of eq .
( [ eq : v ] ) can be expected , especially when non - local potentials are compared with local ones .
this is illustrated in fig .
[ fig : diag ] , where the bare nn diagonal matrix elements and the corresponding on - shell @xmath12-matrix for the local v@xmath0 and the non - local cd bonn are compared with those obtained from the non - local potential indicated as is in ref .
@xcite and hereafter referred to as nl . here
the dominant s - wave channels are considered , for which the main non - locality has been introduced in ref .
one can see that , despite the expected agreement of the various interactions on the @xmath12-matrix , the behavior of the bare nn matrix elements is quite different . at small momenta
we see that in the @xmath8 channel the two non - local potentials are rather close to each other , while they differ significantly from the v@xmath0 this is still true in the @xmath9 channel , where , however , the difference between nl and cd bonn is increased .
it can be also seen that at increasing non - locality , i.e. going from the cd bonn to the nl potential , the deviation form the local potential v@xmath0 is increasing .
this is in agreement with refs .
@xcite , where it was shown that including non - locality makes the attraction provided by the integral term of eq .
( [ eq : lippmann ] ) weaker , thus requiring a less repulsive diagonal matrix element of @xmath27 , in order to obtain the same on - shell @xmath12-matrix .
in particular it is interesting to note that the diagonal nl potential is always rather close to the corresponding on - shell @xmath12-matrix elements , especially for the @xmath9 channel . in the coupling channel @xmath9-@xmath10
we see that the three potentials are very close to each other at small momenta , with nl deviating from the others at intermediate @xmath28 .
considering the large momentum behavior , we observe that in all cases the cd bonn potentials exhibits a much larger tail .
since the diagonal matrix elements of the bare potentials are already so different from each other , is not surprising then that the off - shell behavior of both the bare @xmath27 and the @xmath12-matrix is also different , as illustrated in figs .
[ fig : half1s0 ] and [ fig : half3s1 ] , where the half off - shell matrix elements of the bare potentials and of the @xmath12-matrix are shown for three choices of the incident energy .
the behavior of the bare potentials ( left panels in the figures ) shows essentially the same features we have already observed for the diagonal matrix elements .
the curves representing the @xmath12-matrix ( right panels ) are closer to each other , but still sensible differences are observed , especially when the results corresponding to the v@xmath0 are compared with the two other non - local potentials . finally in fig . [
fig : half3sd1 ] we also show the half off - shell @xmath27 and @xmath12 @xmath29-@xmath30 matrix elements . here
only the tensor part of the nn interaction can contribute , which was shown in previous works to be strongly affected by the inclusion of non - locality @xcite , and which is known to play an important role in nuclear structure calculations .
we see that indeed , the inclusion of non - locality reduces the strength of the interaction and , for the nl case , also the range of momentum values where the matrix elements are significantly different from zero .
the microscopic calculations of nuclear matter equation of state , i.e. the energy per particle as a function of density , on the basis of realistic interactions has a long history .
theoretical and numerical methods , accurate enough for our considerations , are nowadays available
. we will follow the bethe - brueckner - goldstone scheme , which has been proved @xcite to be reliable in a wide range of density . in this expansion the original bare nn interaction @xmath31
is systematically replaced by the @xmath32-matrix , which satisfies the integral equation g [ ; ] = v + _ k_a ,
k_b v |k_a k_bq k_a k_b| - e(k_a ) - e(k_b ) g [ ; ] , [ eq : bg ] where @xmath33 is the single particle energy and @xmath34 the so - called pauli operator which projects both intermediate single particle states above the fermi surface .
the diagrams in terms of the @xmath32-matrix are then ordered according to the numbers of hole - lines they contain .
the introduction of the self - consistent single particle potential @xmath35 is an essential ingredient of the method u(k ) = _ k<k_f k k | g ( e(k)+e(k))| k k_a [ eq : u ] since @xmath36 , eqs . ( [ eq : bg],[eq : u ] ) include a self - consistent procedure which determines @xmath35 .
in particular we use the continuous choice for the potential @xmath35 , i.e. the definition of eq .
( [ eq : u ] ) is extended to all momenta @xmath21 , below and above the fermi momentum @xmath37 .
more details can be found in review papers or in textbooks @xcite .
we extend our calculations up to the three - hole line level of approximation , which has been shown to be accurate in refs .
@xcite for a variety of nn potentials , including the argonne v@xmath7 and v@xmath0 and the paris interactions . however , since doleschall s nl potential is used here for the first time in bbg calculations of nuclear matter , before making comparisons of different eos , we briefly study the accuracy of the bbg expansion in this case .
this is done in fig .
[ fig : bind ] , where the contributions to the binding potential energy of the two hole - line ( brueckner ) and three hole - line diagrams are reported . to keep the calculations reliable
, the latter has to be substantially smaller than the former .
two different choices for the nl potential are considered : the curves labeled as nl1 have been obtained using doleschall s is potential of ref .
@xcite only in the @xmath8 channel , while for the curves labeled as nl2 the non - local potential has been included also in the @xmath9-@xmath10 channel . in the remaining channels
the v@xmath0 interaction has been used .
channel only , and nl2 ( triangles ) , which has the non - locality both in the @xmath8 and in the @xmath38 channels .
empty symbols correspond to the two hole - line calculations , while full symbols represent the full two + three hole - line results . ]
the three hole - line contributions turn out to be slightly larger than in the case of local potentials ( see @xcite ) .
however these corrections always remain within @xmath39-@xmath40% for the nl1 case and within @xmath41-@xmath42% for nl2 , thus indicating that our calculations up to three hole - lines can be considered accurate .
it is interesting to note that with the non - local potentials the corrections to the bhf curves are positive , contrary to the case of the v@xmath0 , where , in the region @xmath43 @xmath44 the three hole - line contributions tend to make the eos more attractive .
having established the validity of our calculations , we can now turn our attention to the study of the saturation properties of nuclear matter . to better understand the role that the different potentials we studied in section [ s : pot ] play in the nuclear matter eos it is useful to rewrite the bethe - goldstone equation ( [ eq : bg ] ) in the channel @xmath16 , & & k l | g^c ( , ) |
k l= v^c_l l(k , k ) [ eq : bg_c ] + & & + , where , following standard procedures @xcite , we have introduced the angle averaged expressions of the pauli operator , @xmath45 , and of the energy , @xmath46 , in the denominator of the integral term .
( [ eq : bg_c ] ) closely resembles the lippmann - schwinger equation ( [ eq : lippmann ] ) , the difference being due to the presence , in the former , of the pauli blocking operator and of a different energy denominator .
both these terms produce a quenching of the integral term contribution in the bethe - goldstone equation with respect to the lippmann - schwinger case .
therefore the behavior of the g - matrix is expected to be rather similar to that of the r - matrix , with this quenching effect being weaker for the non - local potentials , for which the integral term is already small . .
the upper panel corresponds to the @xmath8 channel , the middle panel to the @xmath47 channels , while the lower panel shows the total potential binding energy .
all curves are calculated at the 2 hole - line level . ] since the non - locality in the nl potential is introduced in the @xmath8 and in the coupled @xmath9-@xmath10 channels , we begin by studying the contributions of these channels to the potential binding energy per particle .
this is done , at the brueckner level , in fig .
[ fig : bind_can ] , where the results corresponding to the non - local cd bonn ( empty dots / dashed line ) and nl potentials ( triangles / dot - dashed ) are compared with the v@xmath0 ( crosses / solid ) . in agreement with the discussion in section [ s :
pot ] , where we compared bare potentials and r - matrix elements , we observe that in the @xmath8 channel the cd bonn and nl potential contributions are very close to each other and that they provide additional binding , with respect to the v@xmath0 case .
when we consider the @xmath9 and @xmath10 channels , again we see that the non - local potentials provide more binding , but in this case the nl curve is significantly lower than the cd bonn one .
since the contributions from all other channels are very similar for the three potentials ( with differences of the order of 1 mev at the largest densities considered here ) , the differences observed in the @xmath8 and @xmath9 and @xmath10 channels determine the behavior of the total binding potential energy , shown in the bottom panel of fig .
[ fig : bind_can ] .
( mev ) of eq .
( [ eq : u ] ) for @xmath48 ( upper panel ) , @xmath49 ( middle ) and @xmath50 @xmath51 ( lower ) and for the v@xmath0 ( solid line ) , non - local ( dot - dashed ) and cd bonn ( dashed ) potentials . ] in order to get further insight on the origin of the differences in the potential binding energy , in fig .
[ fig : upot ] we also show the potential @xmath35 for three different choices of the density , namely @xmath48 , @xmath49 and @xmath50 @xmath5 . the most important contribution to
the total binding energy comes from the low @xmath21 region , where the nl potential is , indeed , lower than the others . at large @xmath21
the behavior of the cd bonn curve differs from the others , which is not surprising considering the different high momentum tail of the potential @xmath27 shown , for example , in fig .
[ fig : diag ] .
however this behavior does not seem to affect the final result for the binding energy .
finally , the saturation curve of nuclear matter is shown in figure [ fig : sat ] , where , in order also to make contact with the existing literature , _ e.g. _
@xcite , in the upper panel we show results at the brueckner level , while in the lower panel we include contributions at the three - hole line level .
again , as a reference curve we take the one obtained using the argonne v@xmath0 interaction ( solid line ) : as it is well established , the saturation density it gives is slightly larger than the empirical one , and three - body forces can be introduced to correct for this drawback @xcite .
this is just the issue we want to address , and therefore all the calculations will be presented without the introduction of three - body forces , in order to see to what extent the non - locality can affect this result and in which direction .
the discrepancy of the calculated saturation point from the empirical one can be taken as a measure of the amount of three - body forces which is needed .
also relativistic effects can shift the saturation point , as described within the dirac - brueckner ( db ) approach @xcite .
however , it has been shown in ref .
@xcite that the relativistic effects introduced in db calculations can be described in terms of non - relativistic three - body forces due to virtual creation of nucleon - antinucleon pairs .
let us now consider the other two interactions , which display different degrees of non - locality .
the cd bonn at the brueckner level has been already studied in ref .
@xcite and more recently in ref .
the corresponding saturation curve reported in fig .
[ fig : sat ] is in agreement with previous studies .
the saturation density and energy are substantially larger than for v@xmath0 .
the introduction of three hole - line contributions slightly improves the situation , since the three hole - line diagrams give an overall additional repulsion , which however is much smaller than the two hole - line contribution .
as shown in the bottom panel of fig .
[ fig : sat ] , the binding energy decreases , but still remaining too large with respect to the empirical value , while the saturation density is essentially unchanged .
all that indicates that , in order to get the correct saturation point , one needs an amount of three - body force larger than in the v@xmath0 case .
this is somehow at variance with the results for few nucleon systems , where the cd bonn seems to produce binding energies closer to the experimental ones @xcite .
the doleschall s potential moves further in this direction . as in fig.[fig : bind ] we consider again the nl1 and nl2 cases . and cd bonn potentials and for the non - local interactions nl1 , which includes the non - locality in the @xmath8 channel only , and nl2 , where the non - locality is included both in the @xmath8 and in the @xmath38 channels .
the upper panel shows results at the two hole line level , while in the lower panel three hole - line contributions are included . ] the saturation density , as one can see in fig .
[ fig : sat ] , is larger than in the v@xmath0 case already for the nl1 curve and it becomes abnormally large when both @xmath8 and @xmath29-@xmath30 channels include non - locality ( nl2 ) , with the energy per particle going down to about -26 mev .
the very large strength for the three - body forces needed in this case , for shifting the saturation curve close to the phenomenological one , is in sharp contrast with the excellent results obtained @xcite for few body systems , where virtually no three - body forces are needed .
it looks that any increase of the non - locality would improve the fitting of the binding for the few body , but would shift the saturation point to higher density and binding energy .
we have analyzed the effects of the non - locality of the nn interaction on the saturation curve of nuclear matter by considering a set of interactions with different degree of non - locality . while the introduction of non - locality is able to improve the agreement with phenomenology in few - body systems , where it can reduce or even eliminate the need of three - body forces , in nuclear matter we observe a parallel shift of the saturation point in the wrong direction , away from the empirical saturation point .
the presence of non - locality seems to be unable to solve one of the fundamental problem in the many - body description of nuclear systems , i.e. the ( in)consistency of the three - body forces needed in few - body nuclear systems and in nuclear matter .
we thank p. doleschall and r. machleidt for providing the subroutines for the non - local and , respectively , the cd bonn potentials .
we would also like to thank p. schuck for illuminating discussions on the problem of non - locality , which stimulated the present work .
99 r. machleidt , advances in nuclear physics , vol .
19 , 189 ( 1989 ) , j.w .
negele and e. vogt eds .
, plenum ny .
r. b. wiringa , v. g. j. stoks and r. schiavilla , phys .
c * 51 * , 38 ( 1995 ) .
r. b. wiringa , r. a. smith and t. l. ainsworth , phys .
c * 29 * , 1207 ( 1984 ) .
a. lejeune , p. grange , m. martzolff , and j. cugnon , nucl .
a453 * , 189 ( 1986 ) .
p. doleschall and i. borbely , phys .
c * 62 * , 054004 ( 2000 ) .
p. doleschall , i. borbely , z. papp and w. plessas , phys .
c * 67 * , 064005 ( 2003 ) .
p. doleschall , phys .
c * 69 * , 054001 ( 2004 ) .
r. lazauskas and j. carbonell , phys .
c * 70 * , 044002 ( 2004 ) .
r. machleidt , phys .
c * 63 * , 024001 ( 2001 ) .
m. lacombe , b. loiseau , j. m. richard , r. vinh mau , j. cote , p. pires and r. de tourreil , phys .
c * 21 * , 861 ( 1980 ) .
see for example _ nuclear methods and the nuclear equation of state _ , edited by m. baldo , world scientific , singapore,1999 .
f. coester , s. cohen , b. day and c.m .
vincent , phys .
c * 1 * , 769 ( 1970 ) .
p. u. sauer , ann .
phys * 80 * , 242 ( 1973 ) .
l. engvik , m. hjorth - jensen , r. machleidt , h. muther and a. polls , nucl .
a * 627 * , 85 ( 1997 ) .
r. machleidt , `` nuclear forces and nuclear structure , '' arxiv : nucl - th/9809069 .
r. machleidt , f. sammarruca and y. song , phys .
c * 53 * , 1483 ( 1996 ) .
v. g. j. stoks , r. a. m. klomp , c. p. f. terheggen and j. j. de swart , phys .
c * 49 * , 2950 ( 1994 ) .
day , phys .
* 47 * , 226 ( 1981 ) .
a. nogga , h. kamada and w. gloeckle , phys .
lett . * 85 * , 944 ( 2000 ) .
v. g. j. stoks , r. a. m. klomp , m. c. m. rentmeester and j. j. de swart , phys .
c * 48 * , 792 ( 1993 ) .
m. baldo , a. fiasconaro , h. q. song , g. giansiracusa and u. lombardo , phys .
c * 65 * , 017303 ( 2002 ) . h. q. song , m. baldo , g. giansiracusa , and u. lombardo , phys .
lett . * 81 * , 1584 ( 1998 ) .
h. q. song , m. baldo , g. giansiracusa , and u. lombardo , phys .
b473 * , 1 ( 2000 ) ; x. r. zhou , g. f. burgio , u. lombardo , h .- j .
schulze , and w. zuo , phys .
* c69 * , 018801 ( 2004 ) .
brown , w. weise , g. baym and j. speth , _ comm .
nucl . part .
* 17 * , 39 ( 1987 ) .
p. bozek and p. czerski , acta phys .
b * 34 * , 2759 ( 2003 ) .
gad , eur .
j. a * 22 * , 405 ( 2004 ) . | we study the possible relationship between the saturation properties of nuclear matter and the inclusion of non - locality in the nucleon - nucleon interaction . to this purpose
we compute the saturation curve of nuclear matter within the bethe - brueckner - goldstone theory using a recently proposed realistic non - local potential , and compare it with the corresponding curves obtained with a purely local realistic interaction ( argonne v@xmath0 ) and the most recent version of the one - boson exchange potential ( cd bonn ) .
we find that the inclusion of non - locality in the two - nucleon bare interaction strongly affects saturation , but it is unable to provide a consistent description of few - body nuclear systems and nuclear matter . |
spectroscopic surveys of radio sources at flux density levels much lower than those of the 3c catalogue and its various revisions ( e.g. laing , riley & longair 1983 , hereafter lrl ) are essential to provide coverage of the redshift - radio luminosity plane .
such coverage is required to establish the cosmic evolution of these sources both in number density and physical properties such as size and structure . over the past two decades , several flux - limited samples selected at low radio frequency ( and hence directly comparable with lrl in being selected primarily on the basis of unbeamed steep spectrum radio lobe emission ) have been defined and imaged in the radio and optical . only recently , however , have redshifts have been obtained for a significant fraction of the radio sources .
this paper describes spectroscopic observations of the 38-mhz 8c north ecliptic cap ( nec ) sample of lacy , rawlings & warner ( 1992 ; hereafter paper i ) and lacy et al .
1993 , 1999a ( hereafter papers ii and iii respectively ) , and the closely - related 7c - iii sample .
this latter sample was selected at a frequency of 151 mhz using the 7c survey data of lacy et al .
( 1995 ) with a flux limit chosen to maximise overlap with the 8c sample .
this sample is defined in section 2 , and will be described in more detail in a future paper ( lacy et al .
a further paper will discuss spectroscopy of the 6c sample of eales ( 1985 ) , and details of the 7c - i and 7c - ii samples are presented in willott ( 1998 , and papers in preparation ) .
the properties of these samples are summarised in table 1 , and the coverage of the radio - luminosity redshift plane is illustrated in , e.g. fig .
1 of blundell , rawlings & willott ( 1999 ) .
samples such as these are very valuable for studying the cosmic evolution of the radio source population , in particular the question of the evolution of source size , @xmath5 , with redshift . as redshift and radio luminosity
are always strongly correlated within any flux - limited sample , partial correlation coefficients need to be calculated to investigate whether the true correlation is with redshift or luminosity .
studies based on low - frequency selected samples typically show a weak correlation with redshift only , parameterised as @xmath6 where @xmath7 ( e.g. the 6c sample , eales 1985 ; neeser et al .
1995 , and the 7c - i and 7c - ii samples ; blundell , rawlings & willott 1999 ) , although samples selected at higher frequencies seem to have stronger dependences , with @xmath8 ( oort , katgert & windhorst 1987 ; kapahi 1989 ) combined with a radio luminosity dependence @xmath9 with @xmath10 . in paper ii we performed a partial rank correlation on the 8c - nec sample using photometric redshifts .
this showed that the fundamental correlation was between size and redshift , but having only redshifts based on @xmath11-band magnitudes we were forced to only consider @xmath12 objects and to accept the uncertainties associated with photometric redshifts .
we also investigated the four - way correlation coefficients of redshift , luminosity , size and spectral index , finding that spectral index at 2ghz in the rest frame was most strongly correlated with redshift , and any correlation with luminosity was weak . since then , independent work on the 6c , 7c - i and 7c - ii surveys has been carried out by blundell et al . ( 1999 ) with results that are broadly consistent with the neeser et al .
result and the results of paper ii .
they find , however , a significant correlation of spectral index with luminosity at a rest - frame frequency of 1 ghz , although a correlation with redshift is seen at higher frequencies .
in this paper we use the redshifts for the 7c - iii and 8c - nec samples to study the size - redshift relation further .
we assume @xmath13 and an einstein de sitter ( @xmath14 , @xmath15 ) cosmology unless otherwise stated .
positions are all epoch b1950.0 .
spectral indices , both radio and optical are defined in the sense that @xmath16 . | we present the results of optical spectroscopy of two flux - density limited samples of radio sources selected at frequencies of 38 and 151 mhz in the same region around the north ecliptic cap , the 8c - nec and 7c - iii samples respectively .
both samples are selected at flux density levels @xmath0 times fainter than samples based on the 3c catalogue .
they are amongst the first low - frequency selected samples with no spectral or angular size selection for which almost complete redshift information has been obtained , and contain many of the lowest - luminosity @xmath1 radio galaxies so far discovered .
they will therefore provide a valuable resource for understanding the cosmic evolution of radio sources and their hosts and environments .
the 151-mhz 7c - iii sample is selected to have @xmath2jy and is the more spectroscopically complete ; out of 54 radio sources fairly reliable redshifts have been obtained for 44 objects .
the 8c sample has a flux limit of @xmath3jy and contains 58 sources of which 46 have fairly reliable redshifts .
we discuss possible biases in the observed redshift distribution , and some interesting individual objects , including a number of cases of probable gravitational lensing . using the 8c - nec and 7c - iii samples in conjunction
, we form the first sample selected on low - frequency flux in the rest - frame of the source , rather than the usual selection on flux density in the observed frame .
this allows us to remove the bias associated with an increasing rest - frame selection frequency with redshift .
we investigate the difference this selection makes to correlations of radio source properties with redshift and luminosity by comparing the results from traditional flux - density selection with our new method .
we show in particular that flux - density based selection leads to an overestimate of the steepness of the correlation of radio source size with redshift .
galaxies:@xmath4active galaxies:@xmath4evolution radio continuum:@xmath4galaxies gravitational lensing |
i discuss two related topics .
the first is an approximate evaluation of the qcd coupling at zero momentum , @xmath3 , using the functional schrdinger equation ( fse ) .
the second is the beginning of a program , not completed , for relating chiral symmetry breakdown ( csb ) properties for both quarks and ( hypothetical ) adjoint dirac fermions in qcd to @xmath3 .
this program can be looked on as using known csb results ( from the lattice ) to constrain the allowed range of @xmath3 or as using a theoretically - determined value of @xmath3 to predict the csb properties .
the present accuracy of some preliminary investigations of these issues is not high enough to be definitive , but it does suggest that @xmath3 in the range [email protected] is consistent with known properties of csb .
the fse has long been of interest to me @xcite ; more recently , motivated by a talk @xcite of s. olenk at the 2006 oberwlz symposium , i wrote a paper last year @xcite , one conclusion of which is an estimate of the @xmath5 strong coupling constant at zero ( or small ) momentum , @xmath6 .
this estimate is somewhat unusual since this estimate is based on one - loop gluon gap equations and lattice numerics in @xmath7 , not @xmath5 , plus some theory that attempts to relate @xmath7 qcd to @xmath5 qcd through the fse .
the @xmath7 studies @xcite vary somewhat , but in my view ( ref .
@xcite argues that the one - loop gap - equation results for @xmath0 are too small ) the most reliable value for the ratio @xmath1 is @xmath8 ( for gauge group @xmath9 ) , where @xmath0 is the dynamical gluon mass and @xmath10 is the @xmath7 gauge coupling .
fse theoretical estimates suggest @xmath11 .
there are , of course , other estimates of @xmath3 : phenomenology sensitive to infrared properties of qcd gives @xmath12 @xcite .
other pinch - technique calculations @xcite suggest @xmath13 , just as i use here .
recently , lattice simulations @xcite have been reported for an interesting question with a long lineage @xcite : how does csb work in qcd with adjoint fermions instead of the usual fundamental - representation quarks ?
these authors works state that there is csb for quenched adjoint fermions not only at temperatures below the deconfinement phase transition but also up to a temperature exceeding the deconfinement phase transition .
in contrast , for quarks there is no csb above the deconfinement phase transition , consistent with lattice findings @xcite that center vortices , the standard confinement machinery of today , are both necessary and sufficient for csb with quarks .
adjoint fermions , blind to the long - range parts of center vortices , are not confined and the argument is , therefore , that there are different mechanisms for csb for fundamental and adjoint fermions .
i do not claim to have definitively answered this question , but present numerical estimates based both on estimates of @xmath3 and on fermion gap equations that have appropriate kernels for small momentum suggest an answer that , given fairly large uncertainties , seems to accord with present - day knowledge from the lattice .
what i report here is at best the beginning of a program of refining our quantitative understanding of qcd by comparing estimates of @xmath3 with lattice and theoretical studies of adjoint csb .
theoretical papers from long ago @xcite argue that various confinement mechanisms produce csb for quarks .
( later i will give a brief update of some of these ideas , based on the role of a condensate of center vortices and their close relatives , nexuses , in confinement and csb ) .
these confining mechanisms depend only indirectly on @xmath3 . in principle
it could be ( although the lattice data say otherwise ) that there would be csb for quarks even with no confinement , a possibility that does depend on @xmath3 .
generally , csb for quarks and adjoint fermions should be sensitive to three couplings : ( 1 ) the standard qcd coupling @xmath14 at zero momentum transfer ; ( 2 ) the critical coupling @xmath15 above which csb occurs for fundamental - representation fermions ( quarks ) as found from a gap equation that does _ not _ contain confinement effects ; and ( 3 ) @xmath16 , the same coupling for ( quenched ) adjoint fermions . because these last two couplings differ from the first only by casimir factors , the critical couplings for gap equations are inversely proportional to @xmath17 , the quadratic casimir eigenvalue for the fermions in the gap equation , and so for qcd , @xmath18 .
however , just knowing this is not enough to settle the issue of whether csb can or can not take place through a standard gap equation for quarks or adjoint fermions ; we need not just the ratios but also the values of the critical couplings to compare with estimates from other qcd models of the couplings
. there are in principle three possibilities : @xmath19 in the first case only quarks can show csb , and confinement is necessary for quark csb . in the second case ( which is favored
both from the lattice data and from the estimates i give here ) there is csb for both adjoint and fundamental fermions .
confinement is again necessary for quark csb , whose transition temperature must be rather near that of the deconfinement transition ; the gap equation is largely irrelevant , althouogh it may account for some separation in these two transition temperatures ..
csb for adjoint fermions comes solely from non - confining effects as summarized in a gap equation , and adjoint - fermion csb may or may not extend above the deconfinement transition . in the third case ,
there is csb for both kinds of fermions , and this ought to persist even if confiners such as center vortices are removed from the lattice simulations .
since there is more than a factor of two between adjoint and fundamental critical couplings , the inequality in the second case is a fairly broad one and great accuracy is not needed to single out this case .
fermion gap equations have a long history , beginning with the jbw equation @xcite , and nearly all previous work that does not address confinement issues makes three approximations : 1 ) gluons are massless ; 2 ) landau gauge is used ; 3 ) because to one - loop order in this gauge vertex corrections are not ultraviolet - divergent , vertex corrections are ignored .
non - perturbative phenomena of low - energy qcd require a more careful treatment that i will sketch here .
in the first place , infrared slavery implies dynamic gluon mass generation , which can be only studied effectively in the gauge - invariant pinch technique @xcite . in the second place , because fermion mass generation is an infrared effect there may be important low - energy fermion - gluon vertex corrections .
finally , the possibility of such corrections requires a more careful study of gauge invariance of the gap equation . to some extent
these issues have been addressed before @xcite , but not in a context particularly useful here . in the present work
i include a dynamical gluon mass in the gluon propagator ; give a sketch of the derivation of a gauge - invariant gap equation using the pinch technique ( which seems not to have been addressed in detail before ) ; and use the gauge technique to infer approximate low - momentum vertex corrections that satisfy the correct ward identities of the pinch technique .
the simplest application of these principles , all that i report here , yields a gap equation much like the jbw @xcite equation except that the gluon is massive .
the work reported here is still very much in progress , and needs considerable sharpening .
however , i believe it is already at the point where one can qualitatively see and explain the differences between csb for quarks and for adjoint fermions .
this work has already been published @xcite so i will be brief here . the vacuum wave functional of qcd is a gauge - invariant functional that i write in the form : @xmath20 in which @xmath21 is a real gauge - invariant @xmath7 effective action ( a factor of two because @xmath22 is the weight function for constructing vacuum expectation values ) .
it is constructed to satisfy @xmath23 with hamiltonian @xmath24+v.\ ] ] the functional @xmath25 has infinitely many terms : @xmath26 gauge invariance requires @xcite that @xmath27 be conserved , and any two successive terms in the expansion are related by ghost - free ward identities . even the simplest of these terms for a free gauge theory is not familiar as an effective action , because it has a square root : @xmath28 here @xmath29 is the usual transverse projector : @xmath30 my approach to the fse ( and later to the fermion gap equations ) begins with the fact that qcd , because of infrared slavery , undergoes dynamical gluon mass generation @xcite .
the fse must display this fact ; how does it do so ?
it is easy to describe in a toy model , an abelian gauge theory with gauge - invariant mass @xmath31 put in by hand .
the hamiltonian is : @xmath32\}\ ] ] the corresponding @xmath25 that exactly satisfies this hamiltonian is : @xmath33 once i add the mass , a nonlocality appears from the transverse projector . but this is easily remedied , by introducing a scalar field in the mass term : and project out its gauge - independent part by integrating over all gauge transformations of the gauge potential . ]
@xmath34 ^ 2\ ] ] and functionally integrating over not only the gauge potentials but also over @xmath35 when constructing vacuum expectation values .
this is entirely equivalent to using the non - local transverse projector .
note that the mass term by itself satisfies a fse with the @xmath36 term missing from the hamiltonian of eq .
( [ abelham ] ) .
since we are interested in infrared effects , it is reasonable to make an expansion in inverse powers of the mass @xmath0 ( or equivalently in powers of the gradient operator ) .
the leading term is @xmath37 and is just the mass term @xmath38 itself .
however , a naive expansion runs into a little bit of trouble , as we see by expanding the square root in the exact abelian solution : @xmath39 p_{ij}a_j.\ ] ] ( observe that the second and succeeding terms in the expansion are local ; in fact , the second term is , up to a factor @xmath40 , the usual @xmath36 term . )
if only these two terms are kept , the field has mass @xmath41 and not @xmath0 .
higher - order terms not written must correct for this discrepancy .
of course , no such expansion in @xmath40 is necessary for the abelian case , but it is for the non - abelian case , and requires @xcite a somewhat better approximation to the square root which is usable for momenta whose components are comparable in size to @xmath0 .
how does this go for the non - abelian case ?
i have argued for decades @xcite that one describes locally gauge - invariant gauge - boson mass generation through a gauged non - linear sigma model ( gnsm ) , analogous to the local mass action @xmath42 above . to simplify the notation i use the anti - hermitean matrix gauge potential @xmath43 where the @xmath44 are the standard gell - mann matrices , and covariant derivative @xmath45 introduce a unitary matrix @xmath46 , with the gauge transformation properties @xmath47 when the gauge potential is transformed by : @xmath48
then the locally gauge - invariant gnsm mass term is : @xmath49 ^ 2.\ ] ] the non - covariant derivative @xmath50 is the non - abelian generalization of the abelian scalar @xmath51 ; in fact , the gnsm action can be written as : @xmath52 ^ 2.\ ] ] one can , just as in the abelian case , eliminate @xmath46 through its equations of motion ( that is , minimize @xmath53 ) , which are ( after some non - trivial algebra ) : @xmath54=0.\ ] ] the perturbative solution has infinitely many terms , of which a few are @xcite : @xmath55+\frac{1}{2}[\partial \cdot a,\frac{1 } { \nabla^2}\partial \cdot a]+\cdots \right \}\ ] ] the linear term simply generates the transverse projector i have already used in the abelian case . in addition , there are non - perturbative solutions relevant for center vortices .
i claim that this gnsm mass term is the leading term in the @xmath40 expansion of an effective action that capture the leading non - perturbative effect of infrared slavery , which is dynamic gluon mass generation .
( it also captures the structure of massless poles in the pinch - technique schwinger - dyson equation yielding the mass dynamically . )
it is a good candidate for the leading term of the effective action @xmath25 because it is gauge - invariant ; indeed , just as in the abelian case , it comes from projecting out the non - abelian gauge invariant from the simple @xmath56 mass term .
it is almost evident without calculation that the next - leading term should be the usual yang - mills term , which is just the non - abelian gauge completion of the abelian term shown in eq .
( [ sqroot ] ) : @xmath57 ^ 2-\frac{1}{4g^2 m } \int tr g_{ij}^2+\dots\ ] ] where @xmath58 is the usual yang - mills field strength .
the normalization of the second term follows from the fact that the quadratic term in @xmath25 is just the abelian action of eq .
( [ sqroot ] ) , one copy for each gauge boson .
but just as in the abelian case this wrongly yields a free - field mass of @xmath41 instead of @xmath0 .
i have proposed @xcite to cure this approximately by choosing a renormalization factor @xmath59 that best approximates the actual square root operator of the fse with a two - term expansion ; the result for the approximate two - term action @xmath60 is : @xmath61 ^ 2\}+\frac{z}{mg^2 } \int d^3x trg_{ij}^2+\mathcal{o}(m^{-3})\ ] ] with @xmath62 .
the final step is to compare this to the standard @xmath7 form of gauge - invariant massive qcd for the given mass @xmath0 , which is : @xmath63
^ 2\right \}.\ ] ] here the @xmath7 coupling @xmath10 has the dimensions of mass . comparing the two forms of @xmath64 yields : @xmath65 several authors @xcite have given either lattice or theoretical estimates of the dimensionless ratio @xmath66 for gauge group @xmath9 with @xmath67 .
the results are quite consistent with an average value of @xmath68 .
one can roughly convert the no - quark coupling in eq .
( [ compare2 ] ) to three light flavors by multiplying the right side of eq .
( [ compare2 ] ) by 11/9 , and the resulting value of the strong coupling at zero momentum is : @xmath69 this estimate and other quite similar pinch - technique estimates @xcite , combined with phenomenology @xcite that gives somewhat higher values , suggests that @xmath70 .
i will now see how this range of values fits into fermion gap equations for csb .
the first step is to understand the difference between gap equations that purport to show the effects of confinement and those that do not .
i will not do that in any detail here , but simply draw a few conclusions from the fact that confinement comes from the long - range pure - gauge parts of center vortices , which are quantum solitons of an effective action of the type given in eq .
( [ cand3 ] ) or its @xmath5 extension . because of complications having to do with integrating over center - vortex collective coordinates , it is easiest to present the argument in @xmath71 , where i will use not the familiar action of @xmath71 yang - mills theory but rather the effective action , with a mass term , of eq .
( [ cand3 ] ) in two dimensions @xcite .
( i have argued @xcite that this is ( possibly up to an overall factor ) the correct action for the @xmath72 fse vacuum wave functional , and that leaving out the mass term can not be right ; the reason is that without the mass term wilson loops of all representations show an area law , while with it only @xmath73-ality @xmath74 0 representations are confined , which is correct for @xmath72 .
but all this is irrelevant to the present argument . )
the simplest center vortex is a soliton solution to the equations of motion of the effective @xmath71 action : @xmath75 here @xmath76 is a generator of the cartan subalgebra of @xmath9 , normalized so that @xmath77 $ ] is an element of the center , and @xmath78 are free propagators of mass @xmath79 .
the vector @xmath80 is a collective coordinate for translations , and i do not indicate collective coordinates for group rotations .
a gluon propagator can be defined by integrating the product of two soliton potentials over their common collective coordinates : @xmath81}(-2)tr(q_j^2)a_i(x - a;k)a_j(y - a;k).\ ] ] this propagator has a long - range part coming from the @xmath82 term , and the remainder is short - range . the full gluon propagator , with both terms , can be used in the gap equation for the pinch - technique fermion propagator .
this gap equation is derived from an s - matrix element , or equivalently from some complicated functional of a wilson loop .
the long - range pure - gauge parts are detected by their linkage with the wilson loop . if a vortex is inside the loop it gives a non - trivial center element ; otherwise it gives unity .
since there are no non - trivial elements for @xmath73-ality zero representations , such as the adjoint , the adjoint fermion is completely blind to them and sees only the short - range parts .
so for quarks the gap equation should be described with a gluon propagator containing the long - range pure - gauge terms , which gives the long - range gluon propagator : @xmath83 where @xmath84 is the gauge propagator of @xmath71 qcd .
this propagator , singular at large distances , not only confines quarks , it breaks csb .
( generally in @xmath85 dimensions the propagator behaves like @xmath86 , which i used in @xmath5 in an earlier discussion @xcite of confinement and csb . )
the remaining short - range part couples to all fermions with strength proportional to the quadratic casimir @xmath17 , and has range @xmath40 .
in fact , these short - range soliton parts must sum to a standard massive gauge propagator of the form ( omitting the group labels ) : @xmath87 i have written this propagator in the form given by the pinch technique , where the physical part is both gauge - invariant ( unlike conventional propagators ) and transverse .
the last term on the right is a necessary but inert term depending on the chosen gauge that can not enter any pinch - technique physical prediction . in particular
, it must cancel out in pinch - technique fermion gap equations .
in this section i briefly mention the older gap equations , which are oriented toward ultraviolet behavior .
then i go onto newer equations that treat the infrared regime of qcd more accurately , including a quick discussion of a pinch - technique gap equation .
more details on these newer gap equations will be published elsewhere .
the history of gap equations , from the johnson - baker - willey ( jbw ) equation of the sixties @xcite to work of the nineties , can be traced from various specializations of an approximate gap equation for the csb - breaking running fermion mass @xmath88 : @xmath89 an often - studied variant , and the only one i will consider explicitly , drops the non - linear fermion mass terms in the denominator of the fermion propagator on the right side of the equation , replacing @xmath90 by @xmath91 .
trouble arises with any linearized gap equation that has a massless gluon propagator , because it is impossible to have a finite fermion mass @xmath92 at zero momentum .
removing this problem by keeping the full non - linear equation with a massless gluon and a fermion mass in the denominator is not really a good solution , since the only mass scale would have to come from the running charge , which is essentially constant in the low - momentum regime where the non - linear fermion mass is needed as an infrared cutoff .
in fact the fundamental infrared cutoff is the gluon mass @xmath0 and i expect that @xmath93 .
jbw used the linearized equation for qed , where @xmath94 , the charge does not run , the photon mass @xmath0 vanishes , and @xmath31 is the electron mass , supposed to be generated spontaneously in massless qed .
the idea was to show that this equation leads to a running mass vanishing at large momentum , hence requiring no bare mass counterterm . when this equation is used for qcd
, @xmath17 is the quadratic casimir eigenvalue , @xmath0 is the gluon mass , and @xmath95 is the running qcd charge .
( in principle the dynamical gluon mass @xmath96 must run too @xcite , but since the most important effect of this mass is in the infrared region i will not include such running . )
the master gap equation is a simplified form of the schwinger - dyson equations for the fermion propagator @xmath97 , which has the form : @xmath98.\ ] ] as defined , the fermion mass " is essentially gauge - invariant , at least in the sense that its ultraviolet anomalous dimension is gauge - invariant . in practice , most workers specialize to the landau gauge because large - momentum radiative corrections to the fermion - gluon vertex are absent in one - loop order , so it is argued that in this gauge it should be a decent approximation to ignore vertex corrections and set ( using the qed ward identity ) @xmath99 for all momenta .
i will also use landau gauge and assume @xmath100 does not change much with momentum , although this requires a few words of justification for the pinch technique / gauge technique approach given later .
the linearized fermion gap equation has the generic matrix form @xmath101 where @xmath102 is the kernel , derived from the single skeleton graph for the inverse fermion propagator .
there is always a chiral - symmetry - preserving solution @xmath103 , but we seek csb - breaking solutions with @xmath104 . if the kernel is a well - behaved ( finite - dimensional , bounded ) matrix it is clear that csb can only occur if @xmath105 is sufficiently large ; otherwise the determinant @xmath106 will not vanish . actually , @xmath102 ( from , for example , eq . ( [ jbw ] ) )
is not that well - behaved , but in the equations we use there is a critical coupling @xmath107 marking the boundary between csb and chiral symmetry preservation ; most students of gap equations give rather similar values for this critical coupling .
there are two forms of criticality : the first is based on a differential equation derived from the gap equation
. its solution may be well - behaved for sufficiently small coupling , and then show unphysical features , such as alternation of signs of the running fermion mass , for larger coupling .
the second is based on the original linearized integral equation , which imposes a boundary condition equivalent to consistency between the left - hand and right - hand sides of the gap equation evaluated at zero momentum .
consistency fails if the coupling is too small . whether either or both criticality criteria hold depends on the gap equation , as i will show by an explicit example .
now i give some simple special cases , the first of which is well - known , of the master gap equation of eq .
( [ jbw ] ) .
the first is the original jbw equation used to study possible dynamical generation of the electron mass in qed .
the jbw equation sets @xmath105 to a constant and has no mass terms in the propagators on the right side . with the aid of @xmath108 the jbw equation (
now for qcd , so @xmath17 is reinstated ) becomes the differential equation @xmath109 where @xmath110 and the primes indicate derivatives with respect to @xmath111 .
there are two linearly - independent solutions : @xmath112^{1/2}\}\ ] ] for small coupling the @xmath113 solution decreases roughly at large momentum like @xmath114 and is commonly called infrared - dominated ; the @xmath115 solution corresponds to the original jbw solution , which falls off very slowly at large momentum and is called ultraviolet - dominated .
although we may term one of the solutions infrared - dominated , this does not mean that it can be used for small momenta ; in fact , both @xmath116 diverge .
the infrared - dominated solution is only useful in the ultraviolet .
these solutions @xmath117 are not appropriate for finding the ultraviolet behavior in qcd , where the charge runs at large momentum .
this issue was first clarified by lane @xcite , who showed that the master gap equation with a running charge and no masses exactly captures the ultraviolet behavior of the fermion running mass . at large momentum
the gluon mass @xmath0 can be dropped in the denominators , and the running coupling has its usual behavior ( at leading order ) @xmath118 where @xmath119 is the leading term in the @xmath120-function ( @xmath121 ) and @xmath122 is the qcd mass . with the aid of
the renormalization group lane showed that the ultraviolet behavior appropriate to csb is : @xmath123
i will recover this behavior later in a jbw - like equation with both a gluon mass and a running charge .
the simple jbw equation , with a non - running charge evaluated at zero momentum , has a critical coupling at @xmath124 , corresponding to a critical coupling @xmath107 of value @xmath125 for couplings larger than critical the exponents @xmath126 become complex , with @xmath127 , and the asymptotic solutions both decay and oscillate , for example , like @xmath128 $ ] .
there is certainly no reason to accept as physical a running fermion mass that alternates in sign .
i have already noted that for a well - behaved kernel @xmath105 must exceed a certain value for csb to take place , yet for the simple jbw equation criticality marks the onset of apparently unphysical behavior .
so is criticality in the jbw equation at all related to csb ?
the general answer is yes , although csb may require a somewhat different value from @xmath107 .
there are in fact a few reasons to believe that @xmath107 above is indeed close to the true critical coupling for csb .
for example , miransky @xcite and his collaborators @xcite have studied the positronium bethe - salpeter equation and find that if @xmath129 tachyonic levels appear .
they relate these to vacuum rearrangment and scale - breaking phenomena , of which csb is an example , and compare the oscillatory behavior of equations like the jbw equation with supercritical coupling to the quantum - mechanical problem of fall into the center or to the behavior of solutions to the dirac equation with supercritical qed charge @xmath130 .
in fact , the jbw differential equation is nothing but the radial schrdinger equation at zero angular momentum and energy for an attractive potential @xmath131 : @xmath132 this potential shows fall into the center if @xmath133 , which is just the critical coupling given above .
the qcd critical coupling of refs .
@xcite would be @xmath134 , not much different from the jbw critical value .
others ( see @xcite and references therein ) claim that the jbw value is the critical coupling for the pion bethe - salpeter equation to admit a massless pion .
so i will assume that a critical coupling deduced from the gap equation is close to , if somewhat below , the critical coupling above which there is csb .
the renormalization group can not say anything about the behavior of @xmath88 at low momentum , where ( among other things ) it becomes necessary to include the effects of the gluon mass @xmath0 not only on the propagator but also on the running coupling .
i and others ( see @xcite , which has many references to other works ) claim that in qcd there is a quasi - conformal infrared regime where the running charge @xmath95 is only slowly changing with momentum .
long ago it was argued @xcite that a decent approximation to the running charge at both low ( euclidean ) and high momentum , with the right sort of two - gluon threshhold , is : @xmath135 this quasi - conformal coupling runs very slowly at @xmath136 .
because higher - order terms are neglected this expression can not be more than perhaps 10% accurate . in the ultraviolet region a wide range of values of @xmath122 does not change the coupling very much , but in the infrared regime dimensional transmutation has taken place , with the zero - momentum coupling determined by the ratio @xmath137 .
the reader can verify that the choice @xmath138gev , @xmath139 gev is within 10% of a recent evaluation from data @xcite of the strong coupling at the masses of the @xmath140 and @xmath59 , and ( for three light flavors ) gives @xmath141 .
now consider the linear gap equation keeping the gluon mass , but the running charge is replaced by the fixed - point value @xmath2 .
there is no simple differential equation , but one can do the angular integrals .
the resulting integral equation is : @xmath142 where @xmath143^{1/2}}.\ ] ] these equations can only be solved numerically , but they are closely related to a solvable differential equation with a dominating kernel @xmath144 , such that @xmath145 .
the new kernel is : @xmath146.\ ] ] the new kernel @xmath144 is exactly equal to @xmath102 for @xmath147 ( or @xmath148 ) , and @xmath149 are asymptotically the same at large momentum . when both momenta are non - zero , @xmath144 is greater , by a maximum factor of about 1.3 at @xmath150
. it should therefore be that the critical value @xmath107 for @xmath102 is greater than that for @xmath144 .
i expect a further ( modest ) increase in the critical coupling because if i had included a properly - running charge in the equation with kernel @xmath102 it would further reduce this compared to the presently - considered kernel @xmath144 with a fixed charge .
the approximate version of the original ( linearized ) master gap equation , using the kernel @xmath144 and fixed charge , is : @xmath151 where @xmath152 , a positive number between zero and one , measures the discrepancy between using the kernel @xmath153 and the kernel @xmath154 .
one should think of @xmath152 as roughly measuring some sort of momentum average of the form @xmath155}\rangle.\ ] ] i estimate roughly that @xmath156 .
the differential equation emerging from eq .
( [ jbwir ] ) is exactly the same as eq .
( [ jbw2 ] ) with no gluon mass , except that the independent variable is changed from @xmath111 to @xmath157 and @xmath158 of eq .
( [ lambda ] ) is changed to @xmath159 : @xmath160 the solutions are : @xmath161^{1/2}\}\ ] ] and , unlike the massless - gluon solutions of eq .
( [ jbwsol ] ) , these are finite at @xmath162 .
criticality occurs now at @xmath163 , or @xmath164 as mentioned above , there is another criticality criterion .
i consider evaluating the integral in eq .
( [ jbwir ] ) at zero momentum : @xmath165 this sets a boundary condition on the linear combination of solutions @xmath117 of the differential equation , from which the solution of the integral equation must be formed
. one might expect that if @xmath158 is too small this criterion can never be satisfied , so the differential equation can be satisfied but not the corresponding integral equation .
however , because both @xmath166 and @xmath167 are finite at the origin and integrable over the kernel @xmath144 at infinity , it is in fact always possible to find a linear combination that satisfies eq .
( [ zeromom ] ) .
the problem is to find two numbers @xmath168 such that @xmath169 and @xmath170.\ ] ] these 2@xmath1712 linear equations are soluble for @xmath168 except for at most one value of @xmath159 .
( in the case at hand , @xmath172 ; even the limit @xmath163 exists . ) so the criterion of eq .
( [ zeromom ] ) is not useful whenever the two solutions to the differential equation are both finite at the origin and integrable over the kernel at infinity .
the final change to be made in the gap equation is to replace @xmath2 by @xmath95 in the @xmath144-equation given as eq .
( [ jbwir ] ) . as long as the momentum dependence of the running charge is on the integration variable @xmath91 alone
, a differential equation can be found for any @xmath173-dependence .
this differential equation is : @xmath174 here @xmath175 accounts for the average difference between @xmath102 and @xmath144 but not for the running charge ( _ cf . _
( [ zave ] ) ) .
i estimate @xmath176 . using the running charge of eq .
( [ corn82 ] ) gives the equation : @xmath177}=0\ ] ] where @xmath80 is lane s constant , from eq .
( [ lane ] ) . to my knowledge
this is not a form of any standard differential equation , but it reduces to one in two cases . for small momentum if @xmath111 is dropped compared to @xmath178 , the result is eq .
( [ jbwmass ] ) , already solved .
an equation that is infrared - finite and asymptotically - exact for large momentum results from replacing @xmath179 by @xmath180 , which yields a confluent hypergeometric equation . unfortunately , this is not very accurate for small momentum , but since it is asymptotically - exact for large momentum i will consider it briefly .
replace the running charge by : @xmath181},\ ] ] as one should insist , @xmath2 is unchanged by this modification , and the large - momentum behavior is insensitive to @xmath0 .
in addition to the correction factor @xmath175 , the last term of eq .
( [ jbwrun2 ] ) should be multiplied by a factor @xmath182 that attempts to correct for this mutilation of the running charge .
think of @xmath183 as an average of the type : @xmath184}{\ln [ ( p^2 + 4m^2)/\lambda^2 } ] .\ ] ] now change to the new independent variable @xmath185.\ ] ] with this factor , eq . ( [ jbwrun2 ] ) becomes : @xmath186 where dots indicate @xmath187 derivatives .
a simple estimate suggests that @xmath188 , so the product @xmath189 is essentially unity within the accuracy to which i aspire , and i drop this product .
equation ( [ teq ] ) is a confluent hypergeometric equation ( see , for example , @xcite ) .
the two linearly - independent solutions corresponding to whittaker functions have different asymptotic behaviors . at large momentum
the infrared - dominated solution goes like @xmath190\ ] ] which is just the behavior found by lane @xcite .
the ultraviolet - dominated solution goes like @xmath191.\ ] ] both of these solutions are finite at @xmath162 and both are integrable over the kernel ( including the running charge ) , so the same situation arises as with the closely - related eq .
( [ jbwir ] ) : except for at most one value of @xmath80 , there is always a linear combination of whittaker functions that satisfies the zero - momentum consistency condition analogous to eq .
( [ zeromom ] ) that holds when the charge does not run .
our other criterion for criticality is the onset of zeros in @xmath88 . the confluent hypergeometric equation , eq .
( [ teqn ] ) , shows critical behavior at @xmath192not in the sense of singularities in the solution , but , as for the original jbw equation , if @xmath193 zeros of the mass function @xmath88 set in @xcite .
the criticality condition @xmath194 is not really dynamical ; it depends only on the particle spectrum and gauge group ( see eq .
( [ lane ] ) ) . in qcd with three light flavors i find for quarks that @xmath195 , but for ( quenched ) adjoint fermions , @xmath196 .
it appears unlikely that the modified jbw equation could lead to csb for quarks , but it might well do so for adjoint fermions , in view of all the approximations and uncertainties of our development .
i can , for what it is worth , convert the condition @xmath194 into a criterion for the critical coupling , combining eqs .
( [ lane],[corn82 ] ) by eliminating the @xmath120-function coefficient @xmath119 .
the result is : @xmath197 this estimate differs from that coming from the ultraviolet jbw equation by the factor in parentheses .
this factor plausibly varies from 1 ( at @xmath198 ) to 2 ( at @xmath199 ) , given uncertainties in both @xmath0 and the effective value of @xmath122 that works best for a one - loop approximation .
my final infrared gap equation is based on the gauge technique ( ref .
@xcite and references therein ) , in which one solves " the ward identity for the fermion - gluon vertex in terms of the fermion propagator .
the gauge technique is combined with the pinch technique , which is a way ( already mentioned ) of finding off - shell schwinger - dyson equations , such as the fermion gap equation , that are locally gauge - invariant .
these are not , of course , the usual schwinger - dyson equations ; pinch - technique proper vertices and self - energies contain contributions from parts of graphs naively unrelated to the proper vertex under study . as a result ,
the ward identities of a non - abelian gauge theory are modified ; they are just the naive ward identities of a qed - like theory .
the pinch technique begins by setting up an on - shall s - matrix elements containing the off - shell green s function of interest see fig .
[ fig1 - 4 ] . the on - shell quark - gluon amplitude for finding the pinch - technique quark proper self - energy .
the last graph ( h ) is an example of a new kind of graph emerging from pinching .
, height=384 ] parts of numerators of some of these graphs contain longitudinal momenta which , when they strike an elementary vertex , trigger an elementary ward identity of the form @xmath200 when an inverse propagator hits an extermal quark line it annihilates it , and the inverse propagator with momentum of an adjoining internal quark line simply removes that propagator , replacing it by unity . the resulting pinch " yields graphs such as fig .
[ fig1 - 4](h ) .
a pinch can change part of a vertex graph to a propagator graph , resulting in new terms that must be added to the conventional fermion propagator .
since the whole s - matrix element is gauge - invariant , it turns out that the new fermion propagator is gauge - invariant ; gauge - dependent terms in the original feynman graphs are cancelled by these other pinch terms .
it is already known @xcite that this procedure yields , in an @xmath201 gauge , a gauge - invariant gluon propagator of the type : @xmath202 where @xmath203 is completely gauge - invariant ( independent of @xmath204 ) and the @xmath204 term receives no radiative corrections ( other than wave - function renormalization ) .
not only do the longitudinal terms @xmath205 cancel out of the s - matrix , they also cancel out of the fermion pinch - technique propagator .
( because of this cancellation proper vertices obey naive ghost - free ward identities . )
although i have illustrated the pinch technique only to one loop in the figure , it is possible to extend it to all orders and to non - perturbative phenomena ( see @xcite and references therein ) . for the gap equation
the essential non - perturbative phenomenon is a gluon mass , and i use @xmath206 . as is well - known @xcite , the pinch technique leads to schwinger - dyson equations identical to those of the background - field method in the feynman gauge ( @xmath207 ) .
however , one can formulate the pinch technique in any gauge @xcite , in the sense that ghost - free ward identities and certain other structural elements of schwinger - dyson equations important for the pinch technique in the feynman gauge still hold .
i will , in pilaftsis @xcite sense of the pinch technique , use landau gauge as all other workers do , because ultraviolet corrections to the vertex are unimportant to one loop in this gauge . because the wasd identity relating the vertex to the fermion propagator has no ghost terms , the same is true ( as in qed ) for the coefficient @xmath100 of @xmath208 in the fermion propagator [ see eq .
( [ fprop ] ) ] . in the infrared regime
@xmath100 should , like the running charge , not run much .
so i expect @xmath209 over a large momentum range , and will set it to unity .
( with the gauge technique vertex , any constant @xmath100 cancels out from the gap equation . )
the gauge - technique solution for a pinch - technique form factor is an infrared - valid approximation to the form factor that is asymptotically exact as the gluon momentum vanishes , and its ultraviolet inaccuracies can be compensated for @xcite , although i will not attempt that here . because its ward identity is ghost - free , just as in qed , all necessary formulas can be read off with straightforward modifications from the qed work in @xcite .
one pinch technique / gauge technique improper form factor is ( omitting irrelevant group factors and indices ) : @xmath210\ ] ] and obeys the qed - like ward identity @xmath211 eq .
( [ imp ] ) is not a unique choice for the form factor , but it has the advantage of being an identity for a free massive theory , with @xmath212 and @xmath213 .
all other choices for the pinch - technique form factor , as well as corrections needed for the ultraviolet behavior at loop level , are identically conserved and therefore differ from eq .
( [ imp ] ) only by terms that vanish with at least one more power of @xmath214 at small values of @xmath214 .
the schwinger - dyson ( sd ) equation is : @xmath215 and with the gauge - technique form factor it is _ linear _ in the fermion propagator . in the pinch technique
the gauge - boson propagator , which shows dynamical mass generation , has the form given in eq .
( [ gluonpropagator ] ) , except that the gauge - dependent term is dropped . to study csb
i take @xmath216 and extract the coefficients of @xmath217 in the sd equation ; this yields the gap equation . because the form factor of eq .
( [ imp ] ) is identically equal to that of a free massive field theory and because the running fermion mass is not running very fast in the infrared , the resulting pinch - technique equation at small momentum ( including small integration momentum @xmath173 ) is really the same as i started with in eq .
( [ jbw ] ) , except that for small momenta i ignore the running of masses and the coupling .
it is plausible that a nearly - correct modification of the infrared pinch - technique equation that accounts for ultraviolet effects is simply to let the masses and coupling run
cases i have already reviewed or analyzed . actually , it is somewhat more complicated than this , but the final analysis is much too elaborate to discuss here .
the gauge technique incorporates infrared - important vertex corrections and the pinch technique yields a unique gauge - invariant gap equation in an arbitrary gauge for the underlying feynman graphs . at the present rather simple level of approximation to the pinch technique / gauge technique gap equation , the deep - infrared gap equation , with plausible ultraviolet corrections ,
is the same as found by many others over decades by ignoring vertex corrections and working in the landau gauge ( of course , self - consistent at one - loop level ) .
the full finite - temperature ( @xmath218 ) gap equation , especially for the pinch technique , has a number of complications that i have not dealt with as of this writing , but will save for a more detailed work .
these include separating the gluon propagator into space - space , space - time , and time - time components , which are no longer related by lorentz invariance and which have two mass scales , the magnetic and electric ( debye ) .
this complicates the application of pinch - technique cancellation mechanisms @xcite .
i will only present here the crudest initial attempts at understanding finite - temperature csb with massive - gluon exchange , using what i call the superconductor approximation , because it is in the same spirit as used in the original bcs paper on superconductivity .
it amounts to saying that the gluon propagator is relatively unchanged by temperature effects as long as the temperature is not too large compared to the phase - transition temperature @xmath219 mev .
the reason is that the gluon already has a large @xmath220 mass of some 600 mev , which is not changed drastically by thermal effects at @xmath221 .
however , the fermion mass steadily decreases from its @xmath220 value , eventually vanishing at the csb phase transition .
the difference between fermionic and gluonic dynamical mass generation at finite @xmath218 is that the effective coupling strength for fermions is decreased as @xmath218 increases , by a factor something like @xmath222 where @xmath223 and @xmath224 is a characteristic energy scale ; ultimately , the fermionic mass has to vanish as the coupling diminishes .
but the gluonic mass is increased by a factor roughly @xmath225 ( so the mass grows like @xmath218 at large @xmath218 ) . in the superconductor approximation
i convert the original zero - temperature gap equation , eq .
( [ jbw ] ) , to finite temperature by the usual replacement of the integral over the ( euclidean ) time momentum @xmath226 by a discrete frequency sum : @xmath227.\ ] ] as mentioned above it is not quite correct simply to make this substitution in an equation such as ( [ jbw ] ) where the vector - propagator kinematics have been worked out at zero temperature .
a further approximation is to average certain momentum - dependent quantities that vary fairly slowly by averages , which allows us to make contact with an already - studied zero-@xmath218 equation , eq .
( [ jbwir ] ) .
the resulting gap equation is then eq . ( [ jbwir ] ) with the modified zero - momentum but finite-@xmath218 coupling : @xmath228 with the finite-@xmath218 coupling determined from the zero - momentum form of eq .
( [ corn82 ] ) by using a plausible form for the temperature dependence of the finite-@xmath218 gluon mass : @xmath229\}^{-1};\;\;m(t)= m(t=0)\coth ( \frac{\beta \omega_g}{2})\ ] ] where i choose for the average gluon frequency @xmath230 so as to give a simple but fairly accurate high-@xmath218 limit ( _ cf _ the value used in connection with the fse of @xmath231 , with @xmath232 ) .
the fermion frequency @xmath233 is of the form : @xmath234 at the level of the simple approximations used here , it is not possible to predict with any accuracy the actual ratio of the temperature @xmath235 , at which adjoint chiral symmetry breaking is restored , to the usual deconfinement transition temperature @xmath236 .
however , after some not very interesting numerics involving plausible ranges of unknown and approximated quantities , and using the superconductor approximation given here , it appears possible that @xmath237 . if this possibility survives more detailed numerics then adjoint chiral symmetry breaking survives deconfinement , as the lattice simulations show .
of course , there is much more to learn about qcd in general from adjoint csb at finite temperature beyond our interests here , as the literature shows @xcite .
i hope to take up these questions in more detail later .
i use the fse to estimate , using @xmath7 dynamical mass calculations , the usual @xmath5 strong coupling at zero momentum : @xmath13 ( for three light flavors ) , a value in accord with other estimates using the pinch - technique schwinger - dyson equations .
this is somewhat smaller than phenomenological estimates of around 0.75 .
i then consider modern versions of the fermion gap equation both for quarks and for adjoint fermions in qcd made modern by adding dynamical gluon mass effects and by consideration of the gauge - invariant pinch techniqe / gap technique gap equation .
the final results , although not impressively accurate , suggest that confinement is essential to break csb for quarks , but that standard fermion gap physics can explain csb for adjoint fermions , for some ( uncalculated ) range of temperatures , possibly reaching above the deconfinement temperature .
these results are consistent with present - day lattice simulations .
99 j. m. cornwall , phys .
d * 38 * , 656 ( 1988 ) .
j. greensite and s. olejnik , `` confinement from gauge invariance in 2 + 1 dimensions , '' arxiv : hep - lat/0610073 . j. m. cornwall , phys . rev .
d * 76 * , 025012 ( 2007 ) .
j. m. cornwall , confinement and infrared properties of yang - mills theory " , invited talk at the us - japan seminar on geometric models of the elementary particles , osaka , june , 1976 ( available at the spires website www.slac.stanford.edu/spires/hep ) . j. m. cornwall , _ proceedings of the french - american seminar on theoretical aspects of quantum chromodynamics _ , marseille , france , 1981 , ed . j. w. dash ( centre de physique thorique report no .
cpt-81/p-1345 , marseille , 1982 ; preprint available at the spires website www.slac.stanford.edu/spires/hep ) .
j m. cornwall , phys .
d * 26 * , 1453 ( 1982 ) .
j. m. cornwall and j. papavassiliou , phys .
d * 40 * , 3474 ( 1989 ) .
d. binosi and j. papavassiliou , phys .
d * 66 * , 111901(r ) ( 2002 ) .
a. c. aguilar , d. binosi and j. papavassiliou , phys .
d * 78 * , 025010 ( 2008 ) .
a. c. aguilar , d. binosi and j. papavassiliou , arxiv:0810.2333 [ hep - ph ] .
see s. j. brodsky , `` the conformal template and new perspectives for quantum chromodynamics , '' arxiv : hep - ph/0703109 , and references therein . j. m. cornwall , w. s. hou and j. e. king , phys .
b * 153 * , 173 ( 1985 ) .
j. m. cornwall and b. yan , phys .
d * 53 * , 4638 ( 1996 ) .
g. alexanian and v. p. nair , phys .
b * 352 * , 435 ( 1995 ) .
w. buchmuller and o. philipsen , phys .
b * 397 * , 112 ( 1997 ) .
j. m. cornwall , phys .
d * 57 * , 3694 ( 1998 ) .
f. eberlein , phys .
b * 439 * , 130 ( 1998 ) ; nucl.phys .
b * 550 * , 303 ( 1999 ) .
f. karsch , t. neuhaus , a. patkos and j. rank , nucl .
phys.b * 474 * , 217 ( 1996 ) .
u. m. heller , f. karsch and j. rank , phys .
d * 57 * , 1438 ( 1998 ) and references therein .
a. cucchieri , f. karsch and p. petreczky , phys . rev .
d * 64 * , 036001 ( 2001 ) .
a. nakamura , t. saito and s. sakai , phys .
d * 69 * , 014506 ( 2004 ) .
e. g. s. luna , arxiv : hep - ph/0609149 ; a. a. natale , arxiv : hep - ph/0610256 and references therein ; a. c. aguilar , a. mihara and a. a. natale , phys .
d * 65 * , 054011 ( 2002 ) .
m. garcia perez , a. gonzalez - arroyo and a. sastre , pos * lat2007 * , 328 ( 2007 ) .
j. engels , s. holtmann and t. schulze , pos * lat2005 * , 148 ( 2006 ) . f. basile , a. pelissetto and e. vicari , jhep * 0502 * , 044 ( 2005 ) . f. karsch and m. lutgemeier , nucl
. phys .
b * 550 * , 449 ( 1999 ) .
p. de forcrand and m. delia , phys .
lett . * 82 * , 4582 ( 1999 ) .
j. m. cornwall , phys . rev .
d * 22 * , 1452 ( 1980 ) .
j. f. donoghue and k. johnson , phys .
d * 21 * , 1975 ( 1980 ) .
a. casher , phys .
b * 83 * , 395 ( 1979 ) .
k. johnson , m. baker and r. willey , phys .
rev . * 136 * , b1111 ( 1964 ) .
b. j. haeri , phys . rev .
d * 38 * , 3799 ( 1988 ) .
b. haeri and m. b. haeri , phys .
d * 43 * , 3732 ( 1991 ) . j. papavassiliou and j. m. cornwall , phys .
d * 44 * , 1285 ( 1991 ) .
j. m. cornwall , phys .
d * 10 * , 500 ( 1974 ) .
j. m. cornwall , phys .
d * 57 * , 7589 ( 1998 ) .
k. d. lane , phys .
d * 10 * , 2605 ( 1974 ) .
v. a. miransky , phys .
b * 91 * , 421 ( 1980 ) .
i. fomin , v. p. gusynin and v. a. miransky , phys .
b * 78 * , 136 ( 1978 ) .
s. j. brodsky and g. f. de teramond , arxiv:0707.3859 [ hep - ph ] .
m. davier , s. descotes - genon , a. hocker , b. malaescu and z. zhang , arxiv:0803.0979 [ hep - ph ] .
e. t. whittaker and g. n. watson ,
_ a course of modern analysis _
" ( cambridge university press , cambridge , 1962 ) .
f. g. tricomi , math . z. * 52 * , 669 ( 1950 ) .
j. m. cornwall and w. s. hou , phys .
d * 34 * , 585 ( 1986 ) .
d. binosi and j. papavassiliou , phys .
d * 65 * , 085003 ( 2002 ) .
d. binosi and j. papavassiliou , arxiv:0805.3994 [ hep - ph ] .
a. pilaftsis , nucl .
b * 487 * , 467 ( 1997 ) .
j. e. king , phys .
d * 27 * , 1821 ( 1983 ) .
k. higashijima , phys . rev .
d * 29 * , 1228 ( 1984 ) . | the functional schrdinger equation ( fse ) for qcd gives a unique perspective on generation of a gluon mass @xmath0 , as required for center vortices .
the fse , which yields a special d=3 gauge action , combined with lattice calculations strictly in d=3 give a value for the dimensionless ratio of d=3 coupling to mass @xmath1 .
from this we infer a reasonably accurate value for the d=4 running coupling @xmath2 in the region of low momentum where it is nearly constant .
the result , consistent with other estimates , is too low to drive chiral symmetry breaking ( csb ) for quarks in a standard gap equation that has no explicit confinement effects .
we recall and improve on old work showing that confinement implies csb for quarks , and consider csb for test ( that is , quenched ) dirac fermions in the adjoint representation . here
the previously - found value of @xmath2 is large enough to drive csb in a gap equation , which we relate to the presence of center vortices ( non - confining , for the adjoint ) and nexuses that drive fermionic zero modes .
we discuss the extension of adjoint csb to finite temperature .
* center vortices , the functional schrdinger equation , and csb * + talk at the symposium _ approaches to quantum chromodynamics _
, oberwlz , september 2008 + john m. cornwall + department of physics and astronomy , + university of california , los angeles ca 90095 |
thanks to the development of gravitational wave astronomy , the possibility of studying and understanding new features of the universe is becoming a reality . in this regard
, there is an ongoing effort to learn about all the physical aspects of one of the main sources of gravitational waves ( gw ) for the future space - based laser interferometer space antenna ( lisa)@xcite : extreme - mass - ratio inspirals ( emris ) .
emris are formed when a massive black hole ( mbh ) located at a galactic centre , with masses in the range @xmath0 , captures a stellar - mass compact object ( sco ) , with masses in the range @xmath1 ( white dwarf , neutron star or a stellar black hole ) .
the sco inspirals towards the mbh following a long series of highly eccentric orbits that shrink due to the loss of energy and angular momentum through the emission of gws .
this inspiral is produced by _
effects related to the action of the sco s gravitational field onto its own trajectory . in order to model emri systems and their gw emission in a way that we can produce useful gw templates for analyzing the data stream produced by lisa observations we need to consider the details of the gravitational backreaction responsible of the inspiral .
this is a very challenging task for theorists .
nevertheless , the problem can be simplified due to the extreme mass ratios of these systems , in the range @xmath2 .
this allows us to describe them in the framework of bh perturbation where the sco is modeled as an accelerated point - like mass in the mbh background and the gravitational backreaction is pictured as the action of a local force , the _ self - force_. for the purposes of our work , we can further simplify the problem by studying an analogous emri system which consist on a scalar charged point particle , @xmath3 , orbiting around a schwarzschild mbh ( see , e.g. @xcite ) that fixes the spacetime geometry . in this simplified model , the inspiral proceeds due to the emission of a scalar field , @xmath4 , generated by the sco motion and which affects the sco trajectory through the action of a self - force given by the gradient of the scalar field : @xmath5 where @xmath6 and @xmath7 denote the sco s trajectory and unit four - velocity respectively .
we use this simplified model as a test bed to illustrate the techniques that we have developed for self - force computations . in the case of non - rotating mbh ,
the spherical symmetry of the bh spacetime provides additional simplifications of the problem .
in particular , the scalar field can be decomposed into scalar spherical harmonics , @xmath8 , which are decoupled between them . on the other hand
, the point - like description of the sco produces divergences in the retarded scalar field and hence we need to regularize it . in order to do so
, we use the _ mode sum _ regularization scheme @xcite , which provides an analytical expression for the singular contribution of the retarded field , @xmath9 , at the particle location . in this way , by subtracting it from the full retarded field , @xmath4 , we obtain a smooth and differentiable field at the particle location , @xmath10 . then , the meaningful expression for the self - force is : @xmath11 thus , what we need is a ( numerical ) technique to compute the full retarded field , @xmath4 , and to use the mode sum scheme to obtain the regularized self - force . in what follows we report on work
we have carried out recently on the development of a new accurate and efficient technique for self - force computations of eccentric emris @xcite .
this work is the extension of a previous one @xcite , where these techniques were introduced for the case of circular emris .
our computational scheme consists in diving the computational domain in a number of subdomains in such a way that the sco ( the particle ) is always located at the interface between two subdomains .
this has two main advantages : ( i ) we avoid introducing a spatial scale to resolve the point particle .
( ii ) the equations for the scalar field , which nominally have singular source terms , become homogeneous equations at each subdomain .
these equations , assuming that appropriate initial date is precribed , lead to smooth solutions .
this fact translates into good convergence properties of the numerical method used to implement this method . because of these properties we call this formalism the _ particle - without - particle _ ( pwp ) formulation .
we evolve the individual wave - type equations at each subdomain by using time domain methods , which perform well for eccentric orbits .
this technique was already implemented for the case of circular orbits and has also been presented at the previous lisa symposium @xcite .
we have recently extended the method to make computations also in the case of eccentric orbits @xcite . in what follows
we summarize the main results of this work .
we start by describing the multidomain structure of our pwp formulation ( see figure [ multidomain ] ) .
once we have expanded the scalar field in spherical harmonics , each mode satisfies an independent ( not coupled to the other modes ) 1 + 1 wave type equation of the regge - wheeler type ( with the potential associated to a scalar field ) .
then , the spatial domain is one - dimensional : @xmath12 $ ] , where @xmath13 is the radial tortoise coordinate , @xmath14 corresponds to the truncation in the direction to the mbh horizon and @xmath15 corresponds to the truncation in the direction to spatial infinity . in order to describe the subdomain communication
let us consider for the moment a splitting of the computational domain into two regions or subdomains , one to the left and one to the right of the particle : @xmath16 = [ { r^{\ast}}_{{{\mbox{\tiny h}}}},{r^{\ast}_{p}}]\cup [ { r^{\ast}_{p}},{r^{\ast}}_{{{\mbox{\tiny i}}}}]$ ] .
each of these regions can be in turn divided into more subdomains ( see figure [ multidomain ] ) : @xmath17 , where @xmath18 $ ] with @xmath19 . in the eccentric case
@xcite , we need to use a coordinate system in which the particle is comoving with the interface of two such domains , or in other words , the coordinates @xmath20 and @xmath21 associated with the particle must be time dependent .
in this setup , we use reduction the wave - like equation to a first - order system by promoting the time and radial derivatives to new variables .
then , these variables , @xmath22 , can also be split as follows : @xmath23 where @xmath24 denotes the heaviside step function .
we can define the jump across the particle s location in an arbitrary quantity as : @xmath25^{}_{p } } = \mathop{\lim } \limits_{{r^{\ast}}\to { r^{\ast}}_{p}}\lambda^{}_{+}(t , { r^{\ast}})- \mathop{\lim } \limits_{{r^{\ast}}\to { r^{\ast}}_{p}}\lambda^{}_{-}(t , { r^{\ast}})$ ] .
we can obtain expressions for the jumps by inserting ( [ globalsolution ] ) into the field equations ( see @xcite and @xcite )
. using these junction conditions we communicate the solutions of the homogenous equations at each subdomain .
the numerical method that we use to solve our equations is the pseudospectral collocation ( psc ) method @xcite .
we discretize each subdomain by using a _ lobatto - chebyshev _ grid .
then , in the psc method the variables have a _
spectral _ representation in terms of an expansion in chebyshev polynomials , @xmath26 ( @xmath27 $ ] , @xmath28 ) , and a _ physical _ representation in terms of the values of the variables at the collocation points : @xmath29 where @xmath30 is the mapping between the spectral and physical domains ( the time dependence appears only in the eccentric case and is the way in which we keep the particle at the interface between two subdomains ) , @xmath31 are the spectral coefficients , and @xmath32 are cardinal functions associated to the _ lobatto - chebyshev _
grid @xcite .
one can change from one representation to the other by means either of matrices or fast fourier transforms ( as we do in our implementation ) .
an important feature of the psc method is that it provides exponential convergence for smooth functions , which is the case of our solutions after applying the pwp formulation .
this is illustrated in figure [ convergencewithparticle ] , where we show some convergence plots .
, of the variable @xmath33 .
we show the results for the harmonic modes @xmath34 ( left column ) and @xmath35 ( right column ) and for the orbital parameters @xmath36 @xmath37 @xmath38 ( top row ) and @xmath36 @xmath37 @xmath39 ( bottom row ) , where here @xmath40 denotes the eccentricity and @xmath41 the _
semilatus rectum_. we observe spectral convergence ( straight line ) until we reach machine roundoff error ( plateau ) .
the data for these plots has been obtained from the subdomain at the right of the particle .
[ convergencewithparticle],title="fig:",scaledwidth=45.0% ] , of the variable @xmath33 .
we show the results for the harmonic modes @xmath34 ( left column ) and @xmath35 ( right column ) and for the orbital parameters @xmath36 @xmath37 @xmath38 ( top row ) and @xmath36 @xmath37 @xmath39 ( bottom row ) , where here @xmath40 denotes the eccentricity and @xmath41 the _ semilatus rectum_. we observe spectral convergence ( straight line ) until we reach machine roundoff error ( plateau ) .
the data for these plots has been obtained from the subdomain at the right of the particle .
[ convergencewithparticle],title="fig:",scaledwidth=45.0% ] , of the variable @xmath33 .
we show the results for the harmonic modes @xmath34 ( left column ) and @xmath35 ( right column ) and for the orbital parameters @xmath36 @xmath37 @xmath38 ( top row ) and @xmath36 @xmath37 @xmath39 ( bottom row ) , where here @xmath40 denotes the eccentricity and @xmath41 the _ semilatus rectum_. we observe spectral convergence ( straight line ) until we reach machine roundoff error ( plateau ) .
the data for these plots has been obtained from the subdomain at the right of the particle .
[ convergencewithparticle],title="fig:",scaledwidth=45.0% ] , of the variable @xmath33 .
we show the results for the harmonic modes @xmath34 ( left column ) and @xmath35 ( right column ) and for the orbital parameters @xmath36 @xmath37 @xmath38 ( top row ) and @xmath36 @xmath37 @xmath39 ( bottom row ) , where here @xmath40 denotes the eccentricity and @xmath41 the _
semilatus rectum_. we observe spectral convergence ( straight line ) until we reach machine roundoff error ( plateau )
. the data for these plots has been obtained from the subdomain at the right of the particle .
[ convergencewithparticle],title="fig:",scaledwidth=45.0% ] once we have performed the spatial discretization using the psc method , we obtain a set of ordinary differential equations that can be evolved using the method of lines .
the particular implementation that we use involves a runge - kutta 4 solver .
the evolution must include the communication between subdomains that we have discussed above .
we use two different numerical techniques to communicate the subdomains : ( i ) the _ penalty method _ , where the system is driven dynamically to satify the junction conditions , and ( ii ) the _ direct communication of characteristic fields _ , where the junction conditions are imposed by communicating the characteristic fields of the first - order system of partial differential equations . to illustrate the ability of our method to resolve the field and its derivatives around the particle we show some snapshots of the evolution in figure [ scalar_variables ] .
until now we have introduced the foundations of our method and the particular numerical techniques that we use in order to implement it . we have shown that the pwp formulation lead to smooth solutions at each computational subdomain , which in turn exploits the exponential convergence of the psc method as shown in figure [ convergencewithparticle ] .
we have also shown how well this method can resolve the field and its derivatives around the particle location ( see figure [ scalar_variables ] ) .
in particular , we can see how the jumps in the time and radial derivatives of the scalar field modes @xmath42 are well resolved ( see the inset plots in figure [ scalar_variables ] where we zoom in the area near the particle ) .
finally , we have computed the values of the gradient of the regularized components of the gradient of the scalar field , @xmath43 , which contain information equivalent to the self - force ( see eq .
( [ self_force ] ) ) , which is the projection orthogonal to the particle four - velocity of @xmath43 . in figure [ self_force_evolution ]
we show the evolution of the components of the regularized field obtained by truncating the sum over harmonic modes in such a way that we only include modes with @xmath44 .
this figure shows the evolution of the regularized field for two different orbits , which in terms of the eccentricity and semilatus rectum are : ( i ) @xmath45 and ( ii ) @xmath46 . in order to quote some numerical results ,
we have computed the values of the self - force components using the method of the comunication of the characteristic field described in sec .
[ pwpformulation ] .
we obtain : @xmath47 , @xmath48 and @xmath49 for the eccentric orbit ( i ) and @xmath50 , @xmath51 and @xmath52 for the eccentric orbit ( ii ) .
all our calculations have used between 8 and 39 subdomains and 50 collocation points per domain .
the average time for a full self - force calculation ( which involves the calculation of 231 harmonic modes ) in a computer with two quad - core intel xeon processors at 2.8 ghz is always in the range 20 - 30 minutes .
these calculations can be further optimized by distributing the subdomains and collocation points so that the resolution is adapted further to our physical problem , and is this is the subject of ongoing work .
the calculations can be easily parallelized , either by spreading the computational calculations for individual modes or for individual subdomains . in the last case
the parallelization is not trivial as we need to pass the relevant information for subdomain communication .
looking to the future , we are currently working on extending these techniques to the gravitational case , where the challenge comes from the fact that each harmonic mode is described by a set of coupled 1 + 1 wave type equations , but where the pwp formulation can be used as it has been described here .
pcm is supported by a predoctoral fpu fellowship of the spanish ministry of science and innovation ( micinn ) .
cfs acknowledges support from the ramn y cajal programme of the ministry of education and science of spain and by a marie curie international reintegration grant ( mirg - ct-2007 - 205005/phy ) within the 7th european community framework programme . financial support from the contracts esp2007 - 61712 ( mec ) and fis2008 - 06078-c03 - 01 ( ministry of science and innovation of spain )
is gratefully acknowledged .
the authors have used the resources of the centre de supercomputaci de catalunya ( cesca ) and the centro de supercomputacin de galicia ( cesga ; project number icts-2009 - 40 ) .
1 url # 1#1urlprefix[2][]#2 | the gravitational - wave signals emitted by extreme - mass - ratio inspirals will be hidden in the instrumental lisa noise and the foreground noise produced by galactic binaries in the lisa band .
then , we need accurate gravitational - wave templates to extract these signals from the noise and obtain the relevant physical parameters .
this means that in the modeling of these systems we have to take into account how the orbit of the stellar - mass compact object is modified by the action of its own gravitational field .
this effect can be described as the action of a local force , the _ self - force_. we present a time - domain technique to compute the self - force for geodesic eccentric orbits around a non - rotating massive black hole . to illustrate the method we have applied it to a testbed model consisting of scalar charged particle orbiting a non - dynamical black hole .
a key feature of our method is that it does not introduce a small scale associated with the stellar - mass compact object .
this is achieved by using a multidomain framework where the particle is located at the interface between two subdomains . in this way
, we just have to evolve homogeneous wave - like equations with smooth solutions that have to be communicated across the subdomain boundaries using appropriate junction conditions . the numerical technique that we use to implement this scheme
is the pseudospectral collocation method .
we show the suitability of this technique for the modeling of extreme - mass - ratio inspirals and show that it can provide accurate results for the self - force . |
the development of structure in the universe was well advanced at high redshifts .
quasars have been detected nearly to @xmath2 , and the most distant galaxies to even greater distances ( weymann 1998 ) .
the forest in the spectra of high redshift qsos shows that the intergalactic medium ( igm ) itself developed extensive nonlinear structures by early times . still unknown , however , is the epoch during which the first generation of stars was formed , and the time scale over which the transition from a neutral universe to one that is almost completely reionized took place .
while numerical simulations have shown that the igm is expected to fragment into structures at early times in cold dark matter ( cdm ) cosmogonies ( cen 1994 ; zhang 1995 ; hernquist 1996 ; zhang 1998 ) , the same simulations are much less able to predict the efficiency with which the first gravitationally collapsed objects lit up the universe at the end of the `` dark age '' ( rees 1998 ) .
the nature of the first dominant ionizing sources is still unknown , and although ir observations will permit even higher redshift galaxies to be observed , detections become increasingly difficult because of the surface brightness dimming from cosmological expansion .
madau , meiksin , & rees ( 1997 , hereafter mmr ) demonstrated that radiation sources prior to the epoch of reionization may be detected indirectly through their impact on the surrounding neutral igm and the resulting emission or absorption against the cosmic microwave background ( cmb ) at the frequency corresponding to the redshifted 21cm line .
while its energy density is estimated to be about two orders of magnitude lower than the cmb , the 21cm spectral signature will display angular structure as well as structure in redshift space due to inhomogeneities in the gas density field ( hogan & rees 1979 ; scott & rees 1990 ) , hydrogen ionized fraction , and spin temperature ( mmr ) .
because of the smoothness of the cmb sky , fluctuations in the 21cm radiation will dominate the cmb fluctuations by about 2 orders of magnitude on several arcmin scales .
radio observations at meter wavelengths have therefore the potential to tell us how ( as well as when ) the primordial gas was reheated and reionized : if the reheating was due to ultraluminous but sparsely distributed sources like qsos , or by fainter but more uniformly distributed objects like an abundant population of pregalactic stars . in this paper
we discuss possible ways in which the next generation of radio telescopes could provide the first observations of a neutral igm at a level of a few tenths of @xmath3jy arcmin@xmath4 . the proposed _ square kilometer array _
( _ ska _ ) would provide an enhancement of about two orders of magnitude in sensitivity relative to present
day radio telescopes ( taylor & braun 1999 ) .
we discuss the technical requirements necessary to detect the 21cm signature from the igm and describe a possible design for a special purpose telescope .
we emphasize that within existing technological limits , radio astronomy could effectively open up much of the universe to a direct study of the reheating and reionization epochs .
the ground state hyperfine levels of hydrogen will very quickly reach thermal equilibrium with the cmb . in order to detect the igm in either emission or absorption ,
it is crucial that an effective mechanism be present to decouple the hydrogen spin states from the cmb .
following mmr , we briefly review below the mechanisms that determine the level populations of the hyperfine state of neutral hydrogen in the intergalactic medium ( see also meiksin 1999 ) .
the emission or absorption of 21cm radiation from a neutral igm is governed by the hydrogen spin temperature @xmath5 , defined by @xmath6 where @xmath7 and @xmath8 are the singlet and triplet @xmath9 hyperfine levels , @xmath10k , @xmath11 is the frequency of the 21cm transition , @xmath12 is planck s constant , and @xmath13 is boltzmann s constant . in the presence of only the cmb radiation with @xmath14k
, the spin states will reach thermal equilibrium with the cmb on a timescale of @xmath15yr ( @xmath16s@xmath17 is the spontaneous decay rate of the hyperfine transition of atomic hydrogen ) , and intergalactic will produce neither an absorption nor an emission signature .
a mechanism is required that decouples @xmath5 and @xmath18 , e.g. by coupling the spin temperature instead to the kinetic temperature @xmath19 of the gas itself .
two mechanisms are available , collisions between hydrogen atoms ( purcell & field 1956 ) and scattering by photons ( wouthuysen 1952 ; field 1958 ) .
the collision induced coupling between the spin and kinetic temperatures is dominated by the spin
exchange process between the colliding hydrogen atoms .
the rate , however , is too small for realistic igm densities at the redshifts of interest , although collisions may be important in dense regions , @xmath20^{-2}$ ] ( mmr ) . in the low density igm ,
the dominant mechanism is the scattering of continuum uv photons redshifted by the hubble expansion into local photons .
the many scatterings mix the hyperfine levels of neutral hydrogen in its ground state via intermediate transitions to the @xmath21 state , the wouthuysen field process .
an atom initially in the @xmath9 singlet state absorbs a photon that puts it in an @xmath22 state , allowing it to return to the triplet @xmath9 state by a spontaneous decay .
as the neutral igm is highly opaque to resonant scattering , the shape of the continuum radiation spectrum near will follow a boltzmann distribution with a temperature given by the kinetic temperature of the igm ( field 1959 ) . in this case
the spin temperature of neutral hydrogen is a weighted mean between the matter and cmb temperatures , in eq .
( [ eq : tspin ] ) .
the radio emission may make an important contribution in the vicinity of a radio loud quasar ( bahcall & ekers 1969 ) , and would itself permit the igm to be detected in 21cm radiation .
this may be an important process if the emission of ionizing radiation is more tightly beamed than the radio continuum emission , allowing the gas to remain neutral near the quasar after it has switched on ( mmr ) . ]
@xmath23 where @xmath24 is the pumping efficiency . here
@xmath25 is the indirect de excitation rate of the triplet @xmath9 state via absorption of a photon to the @xmath22 level , and @xmath26 is assumed . if @xmath27 is large , @xmath28 , signifying equilibrium with the matter .
a consideration of the net transition rates between the various hyperfine @xmath9 and @xmath22 levels above shows that the @xmath29 transition rate via scattering is related to the total rate @xmath30 by @xmath31 ( field 1958 ) .
this relation and equation ( [ eq : tspin ] ) are derived in the appendix . in the limit @xmath32 ,
the fractional deviation in a steady state of the spin temperature from the temperature of the cmb is @xmath33 there exists then a critical value of @xmath30 which , if greatly exceeded , would drive @xmath34 .
this `` thermalization '' rate is ( mmr ) @xmath35 consider a patch of neutral hydrogen with an angular diameter on the sky larger than the beamwidth of the telescope , and a radial velocity width broader than the bandwidth . in an einstein de sitter universe with @xmath36 and baryon density parameter @xmath37 , the optical depth of the patch at @xmath38 cm for a spin temperature @xmath39 is @xmath40 the optical depth will typically be much less than unity .
since the brightness temperature through the igm is @xmath41 , the differential antenna temperature observed at earth between this region and the cmb will be @xmath42 if the intergalactic gas has been significantly preheated , @xmath5 will be much larger than @xmath18 , and the igm will be observed in emission at a level that is independent of the exact value of @xmath5 . by contrast , when @xmath43 ( negligible preheating ) , the gas will appear in absorption , and @xmath44 will be a factor @xmath45 larger than in emission , so that it becomes relatively easier to detect intergalactic ( scott & rees 1990 ) .
the presence of a sufficient flux of photons will thus render the neutral igm `` visible . '' without heating sources , the adiabatic expansion of the universe will lower the kinetic temperature of the igm well below that of the cmb , and the igm will be detectable through its absorption . if there are sources of radiation that preheat the igm , it may instead be possible to detect the igm in emission .
the energetic demand for heating the igm above the cmb temperature is meager , only @xmath46 ev per particle at @xmath47 .
consequently , even relatively inefficient heating mechanisms may be important warming sources well before the igm has been reionized .
possible preheating sources are soft x rays from an early generation of quasars or thermal bremsstrahlung emission produced by the ionized gas in the collapsed halos of young galaxies ( mmr ) . while photons emitted by a quasar just shortward of the photoelectric edge are absorbed at the ionization front , photons of much shorter wavelength will be able to propagate into the neutral igm much further .
most of the photoelectric heating of the igm by a qso is accomplished by soft x
the timescale required for the radiation near the light front to heat the intergalactic gas to a temperature above that of the cmb is typically 10% of the hubble time .
the region produced by a qso will therefore be preceded by a warming front .
an additional heating source is the photon scattering itself ( mmr ) . the heating rate by the recoil of scattered photons is @xmath48 where @xmath30 is the scattering rate per hydrogen atom , and @xmath49 is the mass of a hydrogen atom . in the case of excitation at the thermalization rate @xmath50 , equation ( [ eq : edot ] ) becomes @xmath51 the characteristic timescale for heating the medium above the cmb temperature via resonant scattering at this rate is @xmath52 about 20% of the hubble time at @xmath53 .
this result has an important consequence : if a substantial population of sources of uv radiation turn on quite rapidly at ( say ) @xmath47 , there exists a finite interval of time during which photons couple the spin temperature to the kinetic temperature of the igm before heating the gas above the cmb ( reionization occurs at even later times ) . a window
is then created in redshift space where a large fraction of intergalactic gas may be observed at @xmath54mhz in _ absorption _ against the cmb .
we will show next how this transient phase may imprint a strong signature on the radio sky .
numerical @xmath55body / hydrodynamics simulations of structure formation within the framework of cdm dominated cosmologies ( cen 1994 ; zhang 1995 ; hernquist 1996 ; zhang 1998 ) have recently provided a definite picture for the topology of the igm at @xmath56 , one of an interconnected network of sheets and filaments with virialized systems located at their points of intersection . in between the filaments
are underdense regions .
this `` cosmic web '' appears to be a generic feature of cdm ( pogosyan 1998 ) , a pattern imprinted in the matter fluctuations at decoupling and sharpened by gravity . in a uv photoionizing background , modestly overdense filaments ( @xmath57 ) will give rise to the forest seen in the spectra of high redshift quasars .
as discussed below , radio observations at meter wavelengths on several arcmin scales may probe the cosmic web at times prior to the epoch of reionization .
because of the clumpiness of the igm , the 21cm optical depth of equation ( [ eq : tau ] ) will vary for different lines of sight . to estimate the _ rms _ level of brightness temperature fluctuations at @xmath38 cm , we shall consider three different cosmological models with parameters suggested by a variety of recent observations : an open model ( ocdm , @xmath58 , @xmath59 , @xmath9 ) , a flat model ( @xmath60cdm , @xmath61 , @xmath62 , @xmath63 , @xmath9 ) , and an einstein de sitter tilted model ( tcdm , @xmath64 , @xmath65 , @xmath66 , @xmath67 ) . in all cases
the amplitude of the power spectrum or , equivalently , the value of the _ rms _ mass fluctuation in a @xmath68 mpc sphere , @xmath69 , has been fixed in order to reproduce the observed abundance of rich galaxy clusters in the local universe ( e.g. eke 1998 ) .
the tilted model has been designed to also match _ cobe _ measurements of the cmb on large scales .
our choice of an open cdm cosmology ( with a normalization which is 2 _ rms _ away from the value observed by _ cobe _ ) is guided by the growing evidence in favour of a low value of @xmath70 from cluster studies ( e.g. carlberg 1997 ) .
the @xmath60dominated model matches both the cluster constraint on small scales and _ cobe _ constraint on large .
it is also able to account for most known observational constraints , ranging from globular clusters ages to cmb anisotropies ( ostriker & steinhardt 1995 ) to recent analyses of the type ia sne hubble diagram ( perlmutter 1998 ; riess 1998 ) , and just fits within the constraints imposed by gravitational lensing ( kochanek 1996 ; cooray 1999 ) . in all cosmologies the baryon density is @xmath71 ( burles & tytler 1998 ) .
the model parameters are sumarized in table [ cosmo ] .
l l l l l l tcdm & @xmath72 & @xmath73 & @xmath74 & @xmath75 & @xmath76ocdm & @xmath77 & @xmath73 & @xmath78 & @xmath79 & @xmath80 @xmath60cdm & @xmath81 & @xmath82 & @xmath83 & @xmath72 & @xmath84 as an illustrative example , consider a scenario in which sources of photons are in sufficient abundance throughout the universe to preheat the igm to a temperature well above that of the cmb , and to couple the spin temperature to the kinetic temperature of the intergalactic gas everywhere . while the average 21cm signal is two orders of magnitude lower than that of the cmb ( and will be swamped by the much stronger non thermal backgrounds that dominate the radio sky at meter wavelengths and that must be removed ) , its fluctuations relative to the mean of the surveyed area induced in this example by density inhomogeneities only will greatly exceed those of the cmb .
brightness temperature fluctuations will be present both in frequency and in angle across the sky . at high redshifts , and with a spatial resolution of @xmath85 arcmin
, a patchwork of radio emission will be produced by regions with overdensities still in the linear regime and by voids . from equations ( [ eq : tau ] ) and ( [ eq : dt ] ) , the _ rms _ temperature fluctuation relative to the mean , neglecting angular correlations in the projected igm density , is @xmath86 ^{1/2 } } , \label{eq : deltat}\ ] ] where @xmath87 is the curvature contribution to the present density parameter .
this corresponds to a _
rms _ fluctuation in the observed flux of @xmath88^{1/2}}. \label{emission}\ ] ] here @xmath89 is the variance of the density field in a volume corresponding to a given bandwidth [ @xmath90 and angular size @xmath91 .
the former corresponds to a comoving length @xmath92 mpc , where @xmath93 is the hubble constant and @xmath94 ghz is the observation frequency , the latter to a comoving radius @xmath95 , where @xmath96 is the angular diameter distance . on the scales of interest here
, one can use the linearly evolved power spectrum @xmath97 ( bardeen 1986 ) and a filter function that is a cylinder of radius @xmath98 and height @xmath99 , to compute the density variance : @xmath100 where @xmath101 is the bessel function of order one , @xmath102 is the linear growth factor ( e.g. peebles 1993 ) , and @xmath103 takes into account the effect of the peculiar velocities in compressing the same emitting volume into a narrower bandwidth ( as the hubble flow is reduced in overdense regions , see kaiser 1987 ) . in this regime
it is also legitimate to assume that the baryons follow the dark matter , since segregation effects are important only on smaller scales .
we have verified using @xmath55body simulations the validity of the linear approximation and the neglect of angular correlations on scales of @xmath85 arcmin .
figure 1 shows the fluctuation levels at 150 mhz ( @xmath104 ) as a function of beam size for the models listed in table 1 . at a fixed bandwidth of 1 mhz
, @xmath105 increases with decreasing angular scale from about 1 to 10 mk as @xmath89 increases with decreasing linear scale .
since the growth of density fluctuations ceases early on in an open universe ( and the power spectrum is normalized to the abundance of clusters today ) , the signal at a given @xmath91 is much larger in ocdm than in tcdm at high redshift . in a @xmath60cdm universe
the level of fluctuations is comparable with ocdm .
the range of density fluctuations that would be detectable at the @xmath106 level by a projected facility like the _ ska _ is shown in figure 2 , as a function of beam size and frequency bandwidth , and for several assumed integration times .
the detection thresholds are scaled according to a theoretical _ rms _ continuum noise in a 80 mhz bandwidth at 150 mhz of 64 njy over an 8 hour integration , allowing for only 21 of the 30 dishes in a compact array configuration ( taylor & braun 1999 ) , @xmath107 the central compact array in the ` straw man ' _ ska _ design proposal has a diameter of @xmath108 km in order to achieve 110 arcsec resolution . a more appropriate resolution for measuring fluctuations in the igm intensity
is @xmath109 arcmin .
the compact array would have then to be scaled down to @xmath110 to be optimized for such an experiment .
the dashed lines show curves of constant _ rms _ fluctuation of the signal in the corresponding beam .
each angular resolution observation corresponds to a different compact array diameter @xmath111 ( in other words , the array is being optimally resized to match its resolution to the desired angular scale ) .
because the received flux is proportional to the solid angle of the beam , the signal decreases with decreasing angular scale until it falls below the detection threshold , indicated by the solid lines . with 100 hours of observation in a tcdm model
, 2 @xmath3jy fluctuations will be detectable only on scales exceeding 3 arcmin .
in ocdm , and similarly in @xmath60cdm , a signal of @xmath112jy may be detected at significantly higher resolution with the same integration time . figure 3
show radio maps at @xmath104 for tcdm and ocdm , assuming a beamwidth of 2 arcmin and a frequency window of 1 mhz around @xmath113 mhz . here
a collisionless @xmath55body simulation with 64@xmath114 particles has been performed with hydra ( couchman , thomas , & pearce 1995 ) . the simulation box size is @xmath115 comoving mpc , corresponding to 17 ( 11 ) arcmin in tcdm ( ocdm ) .
the baryons are assumed to trace the dark matter distribution without any biasing . from a visual inspection of the two simulated images , as well as from figure 2
, it seems possible that observations with a modified ( much less spread out ) _ ska _ design may be used to reconstruct the matter density field at redshifts between the epoch probed by galaxy surveys and recombination , on scales as small as @xmath116 @xmath117 comoving mpc , i.e. masses in the range between @xmath118 and @xmath119 . in the real universe , the observed radio pattern on the sky depend on a combination of the underlying gaussian density field and the distribution and luminosities of the sources that cause the preheating and reionization of the igm . in this and the following sections we shall discuss the detectability of the first radiation sources in the universe through their impact on the surrounding neutral igm . in hierarchical clustering theories for the origin of cosmic structure ,
massive objects grow ` bottom up ' low mass perturbations collapse early , then merge into progressively larger systems . in such a scenario ,
one naturally expects the high redshift universe to contain small scale substructures , systems with a virial temperature well below @xmath120k .
the resultant virialized systems will remain as neutral gas clouds unless they can cool due to molecular hydrogen , in which case they may form pop iii stars .
these stars will produce a background of uv continuum photons near the frequency which will immediately escape into the igm , while at the same time ionizing the gas in their vicinity .
photons will propagate into uncollapsed , largely neutral regions of the igm where the kinetic temperature in the absence of preheating will be @xmath121 ( couchman 1985 ) , well below @xmath18 because of adiabatic cooling during cosmic expansion .
if pop iii sources are in sufficient abundance throughout the universe , the flux will couple the spin temperature to the kinetic temperature , and @xmath5 will be pulled below @xmath18 everywhere .
after about @xmath122 yr ( see eq .
[ eq : theat ] ) , the same photons responsible for the coupling will warm the igm to a temperature well above that of the cmb
. eventually a sufficient number of photons above 1 ryd may be able to penetrate all of the surrounding neutral gas in the cloud , reach the external intergalactic gas , and create expanding intergalactic regions .
the crucial assumption we make here is that the radiation field switches on at a redshift @xmath123 following the formation of the first stars , and reaches the thermalization value @xmath50 at @xmath124 on a timescale much shorter than the hubble time at that epoch .
while in hierarchical models of structure formation the spatial number density of halos of a given mass grows initially exponentially ( press & schechter 1974 ) , the rise time of the uv radiation background will be modulated by gasdynamical processes . in the model of gnedin & ostriker ( 1997 ) , for instance ,
a metagalactic flux is rapidly produced in the range @xmath125 . in this case coupling between @xmath5 and @xmath126 is almost instantaneous at @xmath127 , and the igm is suddenly detectable in absorption . for @xmath128 ,
photons begin to heat the igm and the kinetic temperature increases at the rate @xmath129 until @xmath130 and the igm is now detectable in emission . here , @xmath131 is the mean molecular weight for a neutral gas with a fractional abundance by mass of hydrogen equal to 0.75 .
such a chain of events will leave a strong imprint on the cmb , and mark the epoch of reheating by the first generation of stars . at 150 mhz ( @xmath132 ) , the brightness temperature ( shown in figure 4 ) shows an absorption feature of 40 mk , with a width of about 10 mhz .
such a depression in the cmb is much easier to observe compared with brightness fluctuations , as the isotropic signal may be recorded with a larger beam area . assuming @xmath133 mk as a fiducial estimate , the signal to noise ratio for the _ ska _ is @xmath134 where @xmath135 is the frequency resolution
, @xmath136 is the beam area in arcmin@xmath137 , and @xmath138 is the observation time . while the strength of this feature depends mainly on the timescale @xmath139 over which the continuum ly@xmath0 background field reaches the thermalization value , we have checked that the absorption dip is well above the detection limit for all timescales @xmath140 , where @xmath141 is the then hubble time .
note that similar sharp signatures in the radio background at meter wavelengths could also be produced at the reionization epoch , as recently discussed by gnedin & ostriker ( 1997 ) and shaver ( 1999 ) .
very little is actually known observationally about the nature of the first bound objects and the thermal state of the universe at early epochs .
if ionizing sources are uniformly distributed , like an abundant population of pregalactic stars , the ionization and thermal state of the igm will be the same everywhere at any given epoch , with the neutral fraction decreasing rapidly with cosmic time . as discussed above
, this may be the case in cdm dominated cosmologies , since bound objects sufficiently massive ( @xmath142 ) to make stars form at high redshift .
on the other hand , reionization may also occur in a highly inhomogeneous fashion , as widely separated but very luminous sources of photoionizing radiation such as qsos , present at the time the igm is largely neutral , generate expanding regions on mpc scales : the universe will be divided into an ionized phase whose filling factor increases with time , and an ever shrinking neutral phase . if the ionizing sources are randomly distributed , the regions will be spatially isolated at early epochs .
in such a scenario , 21cm emission on mpc scales will be produced in the quasar neighborhood as the medium surrounding it is heated by soft x
rays from the qso itself ( mmr ) .
figure [ fig6 ] shows the ( differential ) radio map resulting from a qso ` sphere of influence ' soon after it turns on at @xmath104 ( tcdm ) . the kinetic temperature profile and radiation field around the quasar
were computed including finite light travel time effects and heating by secondary electrons collisionally produced by primary electrons photo ejected by the soft x rays ( mmr ) .
the specific luminosity of the qso was taken to be proportional to @xmath143 , normalized to a production rate of ionizing photons of @xmath144 .
the igm was assumed uniform with @xmath145 and @xmath146 .
the resulting radial temperature profiles were then superimposed on the surrounding density fluctuations as computed using hydra .
the visual effect is due to the convolution of the spin temperature profile with the ( linearly ) perturbed density field around the quasar . outside the bubble
, there is an inner thin shell of neutral gas where the igm is heated to @xmath147 as the hyperfine levels are mixed by continuum photons from the qso , and a much larger external shell where @xmath148 because of adiabatic cooling . at larger distances from the quasar the coupling strength is weakening as @xmath149 , and @xmath150 .
as the warming front produced by the quasar expands , a growing amount of the surrounding igm is unveiled both in emission and in absorption . the size and intensity of the detectable 21cm region
depend on the qso luminosity and age . in the figure the signal ranges from @xmath151jy to @xmath152jy per beam .
note that , although the quasar was placed in a high density region in the corner of the simulation volume , the figure can equally be viewed as the emission due to heating by a beam of soft x rays from the quasar with an opening angle of @xmath153 . thus imaging
the gas surrounding a quasar in 21cm emission could provide a direct means of measuring the opening angle of quasar emission .
the requirements for observing the epoch of igm reheating and reionization are a narrow bandwidth ( @xmath154 1 mhz ) , to achieve adequate resolution in velocity space at the relevant redshift , and an angular resolution of a few arcminutes , the typical scale of the expected density fluctuations .
the straw man proposal for the _ ska _ suggests a large diameter for its central compact array in order to achieve a resolution of @xmath155 arcsec .
while gas density fluctuations would persist to smaller scales in the igm , figure 2 shows that the associated signal would drop well below the threshold for a practical detection . for a successful experiment
it is necessary to significantly decrease the diameter of the compact array .
the current design for the _ ska _ ( and the _ gmrt _ ) has too widely distributed ` sub
apertures ' for a program oriented toward detecting arcmin scale structures .
an alternative possibility is to build a compact , special purpose radio telescope dedicated to searching for the expected reheating and reionization signals at low frequencies . a single dish with a diameter of 200 m ( smaller than arecibo )
would resolve scales of 30 arcminutes , corresponding to @xmath156 comoving mpc .
this is comparable to the largest scale expected for coherent primordial density fluctuations ( pogosyan 1998 ) .
figure 1 shows that the expected brightness temperature fluctuations on this scale will be in excess of 1 mk .
rms _ brightness temperature is given by the radiometer equation as @xmath157 ( burke & graham
smith 1997 ) for a bandwidth @xmath158 and integration time @xmath138 .
the system temperature @xmath159 is sky limited to 200 k in the coldest directions through the galaxy ( burke & graham
if a density fluctuation filled the beam , it would be detectable at the @xmath106 level in a 300 hour integration .
since such structures tend to be filamentary , a more typical covering factor may be @xmath160 , increasing the required integration time by a factor of 10 .
this may be offset by increasing the bandwidth .
the velocity width of a @xmath161 comoving mpc filament at @xmath104 corresponds to a frequency width at 150 mhz of about 8 mhz .
a bandwidth of 5 mhz would then require an integration time of 500 hours .
such a survey operating over several years would enable a substantial range in redshift to be probed .
the history of the universe between the epoch of recombination and the birth of bright sources of radiation by @xmath162 is currently shrowded in mystery : the absence of radiation sources leaves this period a `` dark age . ''
yet spectra of qsos show not only metal absorption features in their spectra , the signature of ongoing and past star formation , but evidence for non linear density fluctuations on large scales as well , a `` cosmic web '' in the form of the forest . we have shown that it may be possible to detect both the development of the cosmic web and the epoch of the first radiation sources , whether stars or qsos , using 21cm measurements from intergalactic gas .
prior to reionization , the radio signal will display emission and/or absorption structures both in angle across the sky and in redshift .
fluctuations in 21cm emission will result from non uniformities in the gas density and from non uniformities in the distribution of radiation sources required to decouple the spin temperature from the cmb temperature . with an instrument like the _ square kilometre array _ , such fluctuations would be detectable on arminute scales , and so reveal whether the reheating is due to galaxies or qsos .
absorption measurements may permit the epoch of the first stellar sources to be determined .
the continuum radiation from an early generation of stars , redshifted to , will couple the spin temperature to the low kinetic temperature of the igm for a transitory interval before heating the gas above the cmb temperature .
the result is a detectable absorption signature against the cmb that would flag the first epoch of massive star formation in the universe .
if reheating were due to ultraluminous but sparsely distributed sources ( e.g. , the first bright qsos ) , the resultant patches would have a larger scale than any gravitationally induced inhomogeneities . because of the very large scales involved , there is no difficulty in principle in achieving a sufficiently narrow bandwidth and an adequate angular resolution with the _ ska _ or _ gmrt _ ; the main limitations are the low flux levels and the radio background subtraction .
alternatively , a special purpose telescope dedicated to scanning the sky in frequency along cold lines of sight through the galaxy may be used to detect the epochs of reheating .
subsequent high resolution measurements with the _ ska _ would provide a means of identifying the nature of the first radiation sources and so shedding new light on the thermal and reionization history of the universe . support for this work was provided by nasa through atp grant nag54236 ( pm , am , and pt ) , and by the royal society ( mjr ) . we thank the anonymous referee for clarifying the technical requirements needed for this project .
the results in this paper made use of the hydra @xmath55body code ( couchman , thomas & pearce 1995 ) .
bahcall , j. n. , & ekers , r. d. 1969 , , 157 , 1055 bardeen , j. m. , bond , j. r. , kaiser , n. , & szalay , a. s. 1986 , apj , 304 , 15 bethe , h. a. , & salpeter , e. e. 1957 , quantum mechanics of one and two electron atoms , ( new york : academic press ) burles , s. , & tytler , d. 1998 , apj , 507 , 732 burke , b. f. , & graham smith , f. 1997 , an introduction to radio astronomy ( cambridge : cambridge univ .
press ) burles , s. , & tytler , d. 1998 , apj , 507 , 732 cen , r. , miralda
escud , j. , ostriker , j. p. , & rauch , m. 1994 , apj 437 , l9 carlberg , r. g , morris , s. l. , yee , h. k. c. , & ellingson , e. 1997 , apj 497 , l19 cooray , a. r. , quashnock , j. m. , & miller , m. c. 1999 , apj , 511 , 562 couchman , h. m. p. 1985 , mnras , 214 , 137 couchman , h. m. p. , thomas , p. a. , & pearce , f. r. 1995 , , 452 , 797 eke , v. r. , cole , s. , frenk , c. s. , & henry , j. p. , 1998 ,
mnras , 298 , 1145 field , g. b. 1958 , proc .
i.r.e . , 46 , 240 .
1959 , , 129 , 551 gnedin , n. y. , ostriker , j. p. 1997
, apj , 486 , 581 hernquist , l. , katz , n. , weinberg , d. h. , miralda
escud , j. 1996 , apj 457 , l51 hogan , c. j. , & rees , m. j. 1979 , mnras , 188 , 791 kaiser , n. 1987 , mnras , 227 , 1 kochanek , c. s. 1996 , apj 466 , 638 madau , p. , meiksin , a. , rees , m. j. 1997 , apj , 475 , 429 ( mmr ) meiksin , a. 1999 , in perspectives in radio astronomy : scientific imperatives at cm and m wavelengths , ed . m. p. van haarlem & j. m. van der hulst ( dwingeloo : nfra ) , in press ostriker , j. p. , & steinhardt , p. j. 1995 , nature , 377 , 600 peebles , p. j. e. 1993 , principles of physical cosmology ( princeton : princeton univ . press ) perlmutter , s. , et al .
1998 , nature , 391 , 51 pogosyan , d. , bond , j. r. , kofman , l. , & wadsley , j. 1998 , preprint ( astro ph/9810072 ) press , w. h. , & schechter , p. 1974
, , 187 , 425 purcell , e. m. , & field , g. b. 1956 , , 124 , 542 rees , m. j. 1998 , in the next generation space telescope : science drivers and technological challenges ( noordwijk : esa pub . ) , 5 riess , a. , et al .
1998 , aj , 116 , 1009 scott , d. , & rees , m. j. 1990 , mnras , 247 , 510 shaver , p. a. , windhorst , r. a. , madau , p. , & de bruyn , a. g. 1999 , a&a , 345 , 380 taylor , a. r. , & braun , r. 1999 , science with the square kilometre array weymann , r. j. , stern , d. , bunker , a. , spinrad , h. , chaffee , f. h. , thompson , r. i. , & storrie lombardi , l. j. 1998 , , 505 , l95 wouthuysen , s. a. 1952 , aj , 57 , 31 zhang , y. , anninos , p. , & norman , m. l. 1995 , , 453 , l57 zhang , y. , meiksin , a. , anninos , p. , & norman , m. l. 1998 , , 495 , 63
electric dipole radiation ( photons ) can induce a spin transition through spin orbit coupling .
the relevant quantum number is the total angular momentum @xmath163 , where @xmath164 is the proton spin and @xmath165 is the total angular momentum of the electron , @xmath166 .
there are six hyperfine states in total that contribute to the transition .
only four of these , the @xmath9 singlet @xmath167 and triplet @xmath168 states ( the notation is @xmath169 ) , and the two triplet @xmath22 states @xmath170 and @xmath171 contribute to the excitation of the 21cm line by scattering .
the selection rule @xmath172 permits the transitions @xmath173 and @xmath174 , and so effectively @xmath175 occurs via one of the @xmath22 states . denoting the occupation number of @xmath167 by @xmath7 and that of @xmath168 by @xmath8 , the rate equation for @xmath7 is @xmath176 where the radiation intensity at the 21cm frequency has been expressed in terms of the antenna temperature @xmath177 , @xmath178 .
the ratio @xmath179 is the number of 21cm photons per mode . here ,
@xmath180 and @xmath181 are the effective excitation and de excitation rates due to scattering .
these may be related to the total scattering rate of photons , @xmath30 , as shown below .
it is convenient to label the @xmath22 levels as states @xmath182 , from lowest energy to highest .
then the external radiation intensity at the frequency @xmath183 corresponding to the @xmath184 transition may be expressed as the antenna temperature @xmath185 , as above .
the temperature corresponding to the energy difference itself , @xmath186 , is expressed as @xmath187 .
if @xmath188 denotes the spontaneous decay rate for the transition @xmath189 , then @xmath180 and @xmath181 are given by @xmath190 and @xmath191 the ratios @xmath192 , where @xmath193 is the total spontaneous decay rate , may be solved for using a sum rule for the transitions .
this states that the sum of all transitions of given @xmath194 to all the @xmath195 levels ( summed over @xmath196 and @xmath197 ) for a given @xmath195 is proportional to @xmath198 ( e.g. bethe & salpeter 1957 ) .
there are four sets of downward transitions to @xmath199 , @xmath200 , corresponding to the decay intensities @xmath201 , @xmath202 , @xmath203 , @xmath204 , @xmath205 , @xmath206 , @xmath207 , and @xmath208 .
these give @xmath209 similarly , there are four sets of upward transitions to @xmath210 , @xmath200 or @xmath211 , giving the intensities @xmath206 , @xmath208 , @xmath205 , @xmath207 , @xmath202 , @xmath204 , @xmath201 , and @xmath203 . these ratios are @xmath212 using @xmath213 , where @xmath214 is the total decay intensity summed over all the hyperfine transitions , and @xmath215 is the statistical weight of level @xmath216 ( @xmath217 is the sum of the weights of the upper levels ) , gives @xmath218 , @xmath219 , @xmath220 , and @xmath221 .
we then obtain @xmath222 ( @xmath223 ) , and @xmath224 . statistical equilibrium for @xmath225
requires @xmath226 .
solving the rate equation ( [ eq : rate ] ) for @xmath227 gives equation ( [ eq : tspin ] ) , where @xmath228 , and where @xmath26 , @xmath229 , and @xmath230 have been assumed . | the emission of 21cm radiation from a neutral intergalactic medium ( igm ) at high redshift is discussed in connection with the thermal and ionization history of the universe . the physical mechanisms that make such radiation detectable against the cosmic microwave background include ly@xmath0 coupling of the hydrogen spin temperature to the kinetic temperature of the gas and preheating of the igm by the first generation of stars and quasars .
three different signatures are investigated in detail : ( _ a _ ) the fluctuations in the redshifted 21cm emission induced by the gas density inhomogeneities that develop at early times in cold dark matter ( cdm ) dominated cosmologies ; ( _ b _ ) the sharp absorption feature in the radio sky due to the rapid rise of the ly@xmath0 continuum background that marks the birth of the first uv sources in the universe ; and ( _ c _ ) the 21cm emission and absorption shells that are generated on several mpc scales around the first bright quasars .
future radio observations with projected facilities like the _ giant metrewave radio telescope _ and the _ square kilometer array _ may shed light on the power spectrum of density fluctuations at @xmath1 , and map the end of the dark ages " , i.e. the transition from the post
recombination universe to one populated with radiation sources . #
1to 0pt#1 |
the first sources of light that ended the cosmic dark ages are expected to have reionized the intergalactic medium and polluted it with metals ( see a review by , e.g. , @xcite ) . to use the metal enrichment as a potential probe of the reionization epoch , recent studies have concentrated on the neutral oxygen ( oi ) produced by the first massive stars @xcite .
the ionization potential of oi is only 0.2 ev higher than that of hydrogen ( hi ) , and the two species are in charge exchange equilibrium .
therefore , oxygen is likely to be highly ionized in regions of the igm where hydrogen is ionized .
however , the recombination time for oxygen ( as for hydrogen ) is shorter than the hubble time at @xmath13 , indicating that oxygen can be neutral even in regions where h has been ionized but where short lived ionizing sources have turned off , allowing the region to recombine @xcite . indeed , @xcite show that such fossil hii regions can occupy @xmath14 of the volume of the igm prior to reionization . while the metal pollution of these regions is poorly understood , it is feasible that they contain significant amount of oxygen with a neutral fraction close to unity .
previous work has proposed to exploit the scattering of uv photons by oi , and the corresponding absorption features in the spectra of quasars the oi forest @xcite , analogous to the lower redshift hi lyman @xmath15 forest ( e.g. , @xcite ) . in the case of hi ,
another interesting signature from the high redshift igm is the 21 cm hyperfine structure line , which is made detectable by uv pumping by the lyman @xmath15 background ( the so called wouthuysen - field effect ; see @xcite or @xcite for a recent review ) .
here we examine whether a similar effect occurs for oi . here
we demonstrate that there exist oi lines with the required features : the fine structure lines of the electronic ground state , @xmath16 and @xmath17 m , can be pumped by @xmath9 photons produced by the first stars or black holes , via the balmer @xmath15 line of oi .
we compute the spectral distortion in the cmb created by this effect , and find that it can be as high as @xmath18 if the oi metallicity at @xmath19 , just prior to reionization , is @xmath20 .
this distortion could be detectable with future cmb experiments , and opens the possibility of performing tomography of the reionization epoch using this effect . in combination with 21 cm studies , it could yield direct measurements of the abundance and spatial distribution of metals in the high redshift igm .
the basic criteria for a line @xmath21 of a metal species or their ions to produce an effect analogous to the wouthuysen
field pumping of the hi 21 cm line , are as follows : ( i ) abundant metals in the igm in the required ionization state ; ( ii ) the @xmath22 transition at a frequency suitable for detection , with einstein coefficients small enough to allow the line to deviate from equilibrium with the cmb ; ( iii ) the upper and lower states should be connected via allowed transitions to another state ( hereafter called state 2 ) , so that they can be `` pumped '' in a two step process ; ( iv ) a large enough background flux at the wavelength @xmath23 corresponding to the @xmath24 and @xmath25 transitions .
in particular , the last criterion imposes the constraint @xmath26 on the wavelength of the uv pumping photons before reionization , since neutral hi depletes the background at shorter wavelengths @xcite . we selected oxygen for our study because it can be relatively abundant at high redshifts and because the similarity of its ionization potential to that of hydrogen ( we leave a systematic search through other metal lines to future work ) .
the ground state of neutral oxygen is split into the three fine structure levels of the outermost electrons in the @xmath27 shell , hereafter denoted as @xmath28 , @xmath29 , and @xmath30 .
we found that among the electronic states of oxygen , these are the only one satisfying all of the criteria above and which photons would be observed in the cmb frequency band .
the states @xmath28@xmath29 and @xmath29@xmath30 are connected by magnetic dipolar transitions [ oi ] @xmath31 at 63.2 @xmath32 m and @xmath33 at 145.5 @xmath32 m respectively and the states @xmath28@xmath30 are connected by an electric quadrupolar transition @xmath34 at 44.1 @xmath32 m .
we shall consider only lines involving the @xmath28 state @xmath35 as the lower level for the transition , since the cmb ambient field is practically unable to populate the @xmath36 state , and the 145.5 @xmath32 m line will be suppressed .
since these transitions are forbidden , the spontaneous emission einstein coefficients ( @xmath37 , @xmath38 s@xmath39 , @xmath40 ) are much smaller than the typical values for electric dipole transitions ( @xmath41 ) , a property shared by the hydrogen 21 cm line with @xmath42 s@xmath39 .
the states @xmath28 and @xmath43 are connected to the excited electronic state @xmath44 in the @xmath45 shell ( hereafter denoted by @xmath46 ) by means of the absorption of an oi balmer-@xmath15 photon with wavelength @xmath47 . a schematic level diagram is shown in figure [ fig : sketch ] . in the absence of a uv field , oi is in thermal equilibrium with the cmb and , as shown in @xcite , resonant scattering is more important than collisionally induced emission , except at extremely large over - densities .
however , if the first stars or black holes generated a uv background at 1300 , then the relative populations of the fine structure states 0 and 1 are modified by the two step `` pumping '' transitions @xmath48 and @xmath49 . for simplicity
we start by considering only one fine structure transition and only later address the combined effect . in what follows we use the indices @xmath50 to denote the two levels in the fundamental state with @xmath51 referring to the lowest one .
for example for the 63.2.1 @xmath32 m transition ( @xmath50 ) corresponds to ( @xmath52 ) ( see fig [ fig : sketch ] ) .
the steady state solution for the level population reads @xmath53}\ , .
\label{eq : sss1}\ ] ] this equation defines the spin temperature @xmath54 . here
, @xmath55 is the equivalent temperature of the @xmath56 transition ( @xmath57k with @xmath58 the boltzmann constant ) , @xmath59 , @xmath60 and @xmath61 are the einstein coefficients for spontaneous emission , induced emission and absorption , respectively , @xmath62 and @xmath63 denote the uv de - excitation and excitation rates , @xmath64 denotes the specific intensity at the resonant frequency of the @xmath56 transition , and @xmath65 , @xmath66 , @xmath67 are the degeneracy factors .
it is useful to define the uv color temperature , @xmath68 in terms of the ratio of the uv de - excitation and excitation rates , @xmath69\,,\ ] ] which represents the spin temperature reached in the limit of strong pumping .
following @xcite , these pumping rates can be written as : @xmath70 here @xmath71 , and the subscript @xmath72 in the right parentheses denotes evaluation at @xmath73 , i.e. at the frequency of the transition connecting states @xmath74 and @xmath75 .
the sum over @xmath76 takes into account the five possible downward transitions from the upper state @xmath46 , listed in table ( [ tab : trans ] ) .
.possible oi transitions between @xmath27 states and + the @xmath45 @xmath44 level@xmath77 . [
cols="^,^,^ " , ] \(a ) taken from nist url site : http://physics.nist.gov/physrefdata/asd/lines_form.html .
+ [ tab : trans ] from the definition of the uv excitation and de - excitation rates , it is easy to see that their ratio is proportional to the ratio of the quantity @xmath78 evaluated at two slightly different frequencies : @xmath79 and @xmath80
. it follows then that the color temperature can be approximated by @xmath81 where @xmath82 is proportional to the number of photons with frequency @xmath83 , and @xmath84 is the logarithmic slope of the background spectrum near the oi balmer @xmath15 line .
we have assumed that @xmath85 .
if @xmath86 is small and negative then the pumping mechanism will be more effective in the transition @xmath87 than in @xmath88 , and as a result the level @xmath51 will be relatively depopulated .
in particular , the @xmath89 photons will be seen in emission if @xmath90 , which will generally be the case since @xmath91k @xmath92 .
the amplitude and shape of the uv photon field at 1300 at the beginning of reionization is uncertain . unlike at the hi ly @xmath15 frequency of 1215 , however , the high@xmath2 igm should be optically thin at 1300 ( see discussions of the shape of the soft uv background in @xcite )
. therefore , the relevant background at redshift @xmath2 should reflect the intrinsic spectrum of the dominant sources over the @xmath93 redshift range @xmath94 scattering for these sources should be negligible @xcite , while the contribution from yet more distant sources will be blocked , because the photons redshifting down to 1300 by redshift @xmath2 will have passed through the hi lyman @xmath95 resonance ( or higher lines ) , and they will have been broken down into lower energy photons . ] . over the 0.2% fractional difference between @xmath79 and @xmath80 , we then expect a relative decrement of @xmath96 of @xmath97 throughout @xmath98 , or a decrement of order @xmath99 for sources with @xmath100 . for our calculations ,
we adopted a fiducial choice of @xmath99 , but in practice , our results are insensitive to this value ( see below ) .
given the specific intensity @xmath64 , the steady state solution for the level population provides the following expression for the spin temperature @xmath54 : @xmath101 } { \exp{(-t_{\star}/t_{uv } ) } + \frac{a_{ji}}{p_{ji}^{uv}}\left(i_{\nu}c^2 /2h\nu^3\right)_{\nu_{ji } } } \right\ } \label{eq : tspin}\ ] ] if at a given redshift @xmath102 the presence of dust can be neglected , @xmath64 at @xmath98 is given by the cmb , @xmath103)$ ] . in the limit of @xmath104 , in equation ( [ eq : tspin ] ) we have @xmath105 , whereas if @xmath106 then @xmath107 .
as long as the slope of the background at 1300 is such that @xmath108 , equation ( [ eq : tl ] ) implies that @xmath109 $ ] , so that @xmath54 must be bracketed between @xmath110 $ ] and @xmath111 . in the limit of @xmath112 , and
thus @xmath113\simeq 1 $ ] , and @xmath104 ( both of these conditions are satisfied for the 63.2@xmath32 m line ) , the departure of @xmath54 from @xmath114 becomes @xmath115 i.e. it depends linearly on the uv flux at 1300 and the dependence on @xmath111 is negligible .
the left panel of figure [ fig : ts_y ] shows @xmath116 as a function of the uv flux for @xmath117 ( solid line ) and @xmath118 ( dashed line ) for the case of the 63.2 @xmath32 m [ 0 @xmath119 1a ] line .
while @xmath116 deviates only one part in @xmath120 from @xmath121 , this small temperature difference nevertheless modifies the population of the levels 0 and 1 . -
[ fig : fig1 ]
the distortion in the cmb spectrum can be parameterized by the @xmath5parameter , @xmath122 , where @xmath123 is the deviation in surface brightness from the planck spectrum with the unmodified cmb temperature . before computing the distortion using the spin temperature above , it is useful to obtain an order of magnitude estimate . because the optical depth of the igm for the uv photons in our case is small ( we find @xmath124 at redshift @xmath125 for @xmath126 ) , multiple scatterings will be rare .
( note that this is different from the case of the hi 21 cm , where the lyman@xmath15 photons responsible for the pumping have a large optical depth and scatter multiple times . ) in this case , roughly a fraction @xmath127 of every 1300 photon produced by stars will scatter off an oi atom and produce an `` excess '' @xmath128 photon . just prior to reionization , assuming that a few uv photons have been produced per h atom ( corresponding to a background flux of @xmath129 erg s@xmath39 @xmath130 sr@xmath39 hz@xmath39 ) , the distortion should be roughly @xmath131 , where @xmath132 is the baryon to photon ratio , and @xmath133 represents the fraction of cmb photon in the frequency band corresponding to the redshift range @xmath134 .
to compute the distortion of the cmb at the redshifted @xmath98 frequency more precisely , we use the radiative transfer equation @xmath135 , \label{eq : jnu}\ ] ] where @xmath136 denotes the cosmological line element along the line of sight . since the line profile @xmath137 is narrow , we have ignored the effect of cosmological expansion . after taking the oi abundance in the igm to be @xmath138 times the solar value @xcite
, we find the optical depth for the @xmath139 and @xmath140 transitions @xmath141 to be very small , ( @xmath142 . in this equation
we have approximated the line profile function @xmath143 by a dirac delta , and @xmath144 is the fraction of oxygen atoms in the @xmath35 level ground level , ( @xmath145 to a good approximation ) . assuming a constant uv background ( justified by the fact that @xmath146 ) , the net change to the cmb due to the presence of oi can be written as an integral along the line of sight of the emissivity , resulting in @xmath147 @xmath148 .
\label{eq : deltainu}\ ] ] the corresponding cmb distortion at the observed frequency @xmath149 is given by @xmath150)$ ] . in our case , @xmath151 , and the distortion parameter can be written in the the simple form @xmath152 $ ] , with @xmath153 $ ] . for the @xmath0 m line @xmath154 but @xmath155 ,
thus we conclude that @xmath156 in this case , the distortion parameter @xmath5 is simply the relative deviation of @xmath157 with respect to @xmath121 multiplied by the optical depth times the ratio @xmath158 .
the right panel of figure [ fig : ts_y ] shows the amplitude of the distortion introduced by the oi 63.2 @xmath32 m transition at @xmath159 and @xmath125 as a function of the uv background intensity , over a range expected to be relevant for the epoch just prior to reionization .
note that , apart from the @xmath160 factor , eq . ( [ eq : ysimple ] )
is identical to that found for the hi 21 cm spin temperature . since in our case
, for a metallicity of @xmath161 the relative deviations of @xmath157 with respect to @xmath121 are typically @xmath162 smaller than in the hi 21 cm scenario , and since @xmath163 , our @xmath5distortion is @xmath164 smaller than in the 21 cm case . for the 44.1 @xmath32 m ( @xmath165 )
transition , the amplitudes of @xmath5 are a factor @xmath166 larger , but the relevant frequencies are more contaminated by dust and thus more difficult to detect , as we comment next .
our basic result above is that the oi 63.2@xmath32 m ( and 44.1@xmath32 m ) transition could be seen as deviations in the present day cmb black - body spectrum of the order of up to @xmath167 at around @xmath168 ghz ( 481 ghz corresponds to z=10 and 160 ghz to z=30 , for the 63.2@xmath32 m ) .
the firas experiment has already obtained constraints on the deviation of the cmb spectrum from a black - body .
we find that in the range of frequencies corresponding to @xmath169 for the 63.2@xmath32 m line , the firas data @xcite constrains @xmath170 at 1-@xmath171 level , ( @xmath172 at the 3-@xmath171 level ) .
considering the expected background uv flux at these redshifts we obtain a firas constraints of 5 ( 40 ) solar metallicity at @xmath173 .
although this constraint is not yet at an interesting level , given the measurements of metallicities at @xmath174 ( e.g. , @xcite ) , it is the first _ direct _ constraint on the metallicity of the igm in the universe at @xmath175 .
the upper limit on the metallicity scales linearly with the constraint on @xmath5 .
for example , by comparing different , accurately calibrated planck hfi channels constraints @xmath176 could be imposed .
an experiment able to reduce the firas @xmath5 uncertainty ( e.g. , @xcite ) by two to three orders of magnitude could impose interesting constraints .
the main limitation to the measurement is foreground dust emission from the galaxy and from ir galaxies , however clean patches of the sky ( to remove the dust emission from the galaxy ) and dust observations at higher frequencies ( to remove the ir galaxies ) could be used for this purpose .
of course , even if the required s / n is achieved in a future observation , there could be other cosmological effects ( such as decaying particles ; see @xcite for a brief discussion ) producing distortion at the feeble @xmath177 level .
it remains to be seen , then , whether the oi distortion can be disentangled from these .
one further consideration is that clustering of the sources will increase the signal greatly making it more easily detectable .
the signal should trace the clustering properties of the fossil hii regions during reionization and should dominate over the primary cmb angular power spectrum in the damping tail .
the clustering properties of the signal should be qualitatively similar to those of the ostriker - vishniac effect but will a peculiar frequency dependence : the signal will be non - zero only for frequencies corresponding to the the @xmath178 m transition at redshifts with non - negligible oi abundance ( and uv background ) .
this will be presented elsewhere .
our technique complements future experiments that will detect the h21 cm hyperfine transition in absorption because it is sensitive to different systematics and operates at different wavelengths
. this would be highly valuable , since hi 21 cm + oi measurements in combination can probe the spatial distribution of metallicity directly .
zh acknowledges partial support by nasa through grant nng04gi88 g , by the nsf through grant ast-0307291 , and by the hungarian ministry of education through a gyrgy bksy fellowship .
chm , lv and rj acknowledge support by nasa grant adp04 - 0093 , and nsf grant pire-0507768 .
we thank a. miller , p. oh , d. spergel and j. miralda - escud for discussions . | we consider the pumping of the @xmath0 m fine structure line of neutral oi in the high redshift intergalactic medium ( igm ) , in analogy with the wouthuysen
field effect for the 21 cm line of cosmic hi .
we show that the soft uv background at @xmath1 can affect the population levels , and if a significant fraction of the igm volume is filled with `` fossil hii regions '' containing neutral oi , then this can produce a non negligible spectral distortion in the cosmic microwave background ( cmb ) .
oi from redshift @xmath2 is seen in emission at @xmath3 m , and between @xmath4 produces a mean spectral distortion of the cmb with a @xmath5parameter of @xmath6 , where @xmath7 is the mean metallicity of the igm and @xmath8 is the uv background at @xmath9 in units of @xmath10 erg / s / hz/@xmath11/sr .
because o is in charge exchange equilibrium with h , a measurement of this signature can trace the metallicity at the end of the dark ages , prior to the completion of cosmic reionization and is complementary to cosmological 21 cm studies . while future cmb experiments , such as planck could constrain the metallicity to the @xmath12 level , specifically designed experiments
could potentially achieve a detection .
fluctuations of the distortion on small angular scale may also be detectable .
epsf # 1= # 1 |
recently quantum computation and quantum communication have attracted much attention , because the use of quantum mechanical phenomena can offer unusual efficiency in computation and communication .
we have to protect quantum states from environmental noise in quantum computation and some methods in quantum communication , such as the quantum superdense coding @xcite . the quantum error - correcting codes ( or quantum codes ) independently proposed by shor @xcite and
steane @xcite is one of techniques for protecting quantum states .
recently it was recognized that construction of quantum codes is connected with the problem of finding a linear space over a finite field with certain properties ( see theorem [ th1 ] ) . in this paper
we propose a method of constructing such linear spaces from an algebraic curve .
let us explain quantum codes and their connection with linear spaces over finite fields .
we begin with the notion of @xmath0-error correction .
let @xmath1 be a @xmath2-dimensional complex linear space , where @xmath2 is a prime power , and suppose that @xmath1 represents a physical system of interest .
a quantum code @xmath3 is a @xmath4-dimensional subspace of @xmath5 .
when we want to protect a quantum state in @xmath6 , we encode @xmath7 into a state in @xmath3 .
so we encode a quantum state of @xmath8 particles into that of @xmath9 particles .
suppose that we send @xmath10 and receive @xmath11 .
a quantum code @xmath3 is said to be _ @xmath0-error - correcting _ if we can decode @xmath7 from @xmath12 provided that at least the states of @xmath13 particles in @xmath12 are left unchanged from @xmath7 . since
a change of a quantum state is continuous , the notion of @xmath0-error correction seems nonsense at first glance @xcite .
this notion can be justified as follows : in general the decoding process of a quantum code does not decode perfectly the transmitted quantum state from a received one .
however , the decoded state and a transmitted state become closer as @xmath0 increases provided that the quantum channel used is memoryless as a @xmath2-ary channel ( * ? ? ?
* section 7.4 ) , @xcite . a quantitative relation between the closeness of states
, the noisiness of a channel , and @xmath0 can be found in @xcite .
in @xcite it is shown that one can make the decoded state arbitrary close to the transmitted state by increasing the code length provided that the ratio @xmath14 is fixed and is sufficiently large compared with the noisiness of the channel .
this is a major motivation for studying long codes as in the classical coding theory ( * ? ? ?
* section 4.3 ) .
when the code length is small , we can find good quantum codes by examining all possible codes by a computer .
however , when the code length is large , we need some systematic construction method for producing good quantum codes . in this paper
we propose such a method .
we are now able to state the connection between quantum codes and finite fields , which is obtained for the binary code ( @xmath15 ) by calderbank et al .
@xcite and generalized to the nonbinary case by ashikhmin and knill @xcite .
let @xmath16 be the finite field with @xmath2 elements . for vectors @xmath17 , ,
@xmath18 and @xmath19 , , @xmath20 we define the standard symplectic form ( or alternating form ) by @xmath21 a linear space with a nondegenerate symplectic form is called a symplectic space . for a vector @xmath22 define the weight of @xmath23 by @xmath24 for a subspace @xmath25 we define @xmath26 that is , the orthogonal space of @xmath27 with respect to ( [ symplectic ] ) . [ th1]@xcite if there is an @xmath28-dimensional subspace @xmath25 such that @xmath29 , then we can construct a @xmath30-error - correcting quantum code @xmath31 of dimension @xmath4 , where @xmath32 the quantum code @xmath3 constructed by this method is called a _ stabilizer code _ and is proposed independently by gottesman @xcite and calderbank et al .
the nonbinary generalization is due to knill @xcite and rains @xcite .
the stabilizer code is the most general algebraic construction of quantum codes proposed so far .
the value @xmath33 is called the _ minimum distance _ of a stabilizer code . a stabilizer code with minimum distance @xmath34 encoding @xmath8 particles into @xmath9 particles is called an @xmath35 $ ] code
. rains ( * ? ? ? * , remarks ) observed that a @xmath36-ary stabilizer code is a tensor product of a @xmath37-ary stabilizer code and a @xmath38-ary one if @xmath37 and @xmath38 are relatively prime .
so we restrict ourselves to stabilizer codes for quantum systems of prime power dimension .
this paper is organized as follows : in section [ sec2 ] we propose a construction method of quantum stabilizer codes from algebraic curves and discuss decoding process of the constructed codes . in section [ sec31 ]
we construct an asymptotically good sequence of quantum codes as an example of the proposed construction method . in section [ sec32 ]
we improved the construction of asymptotically good sequence in @xcite as a byproduct of the construction in section [ sec31 ] , and compare the sequences in section [ sec3 ] with the known asymptotically good sequences @xcite in figure [ fig1 ] .
we shall use the formalism of algebraic function fields instead of algebraic curves .
notations used are exactly the same as those in stichtenoth s textbook @xcite .
[ prop2 ] let @xmath39 be an algebraic function field of one variable , @xmath40 an automorphism of order @xmath41 of @xmath42 not moving elements in @xmath16 , and @xmath43 ,
, @xmath44 pairwise distinct places of degree one such that @xmath45 for all @xmath46 , @xmath47 , , @xmath9 .
let us introduce a condition on a differential @xmath48 : @xmath49 the existence of such @xmath48 is guaranteed by the strong approximation theorem of discrete valuations ( * ? ? ?
* theorem i.6.4 ) .
further assume that we have a divisor @xmath50 such that @xmath51 , @xmath52 .
define @xmath53 let @xmath54 where @xmath48 is as eq .
( [ etacondition ] ) .
then we have @xmath55 . in the following argument , @xmath17 , , @xmath18 and @xmath19 , , @xmath56 . by proposition vii.3.3 in @xcite and the assumption on @xmath40 and @xmath50
, we have @xmath57 the assertion is proved as follows : @xmath58 [ coro3 ] notations as in proposition [ prop2 ] .
assume further that @xmath59 .
then we can construct an @xmath35 $ ] quantum code @xmath3 , where @xmath60 for the minimum distance @xmath34 of @xmath3 , we have @xmath61 theorem [ th1 ] and proposition [ prop2 ] show that we can construct a quantum code from @xmath62 because @xmath63 . by theorem [ th1 ]
we have @xmath64 .
theorem ii.2.2 of @xcite asserts @xmath65 which shows eq .
( [ eqk ] ) .
we shall prove eq .
( [ eqd ] ) .
suppose that @xmath66 , , @xmath67 for @xmath68 .
then there exists a set @xmath69 ,
, @xmath70 such that @xmath71 , which implies @xmath72 .
since @xmath73 , we have @xmath74 the above construction provides good codes only when @xmath2 is large as the classical algebraic geometry codes . for small @xmath2 , we construct a @xmath2-ary quantum code from a @xmath75-ary one by [ th : ak ] let @xmath76 be a positive integer , @xmath77 , , @xmath78 an @xmath16-basis of @xmath79 .
define @xmath16-linear maps @xmath80 sending @xmath81 ,
, @xmath82 to @xmath83 , and @xmath84 sending @xmath81 , , @xmath82 to @xmath85 where @xmath86 is an @xmath87 matrix defined by @xmath88 with the trace function @xmath89 from @xmath79 to @xmath16 . for @xmath90 ,
let @xmath91 , , @xmath92 , @xmath93 , , @xmath94 , , @xmath95 . if @xmath96 for @xmath97 , then @xmath98 .
we also have @xmath99 .
the decoding process of a quantum stabilizer code is usually proceeded as follows @xcite : one measures each observable corresponding to a generator of the stabilizer group of the code , then determines which unitary operator on @xmath5 should be applied to the received state . in the determination of the operator from measurement outcomes we have to solve the following problem .
[ decpro ] let @xmath27 be an @xmath28-dimensional subspace of @xmath100 such that @xmath96 .
given @xmath101 ,
, @xmath102 , find a vector @xmath103 having the minimum weight ( [ eq : weight ] ) in the set @xmath104 for @xmath105 ,
, @xmath106 , where @xmath107 , , @xmath108 is a basis of @xmath109 .
if there exists a vector @xmath110 such that @xmath111 for @xmath105 , , @xmath112 and that @xmath113 where @xmath107 , , @xmath108 is a basis of @xmath55 constructed in corollary [ coro3 ] , then we can efficiently find @xmath103 from @xmath101 ,
, @xmath114 as follows .
the algorithm of farrn @xcite efficiently finds the unique vector @xmath23 having the minimum _ hamming _ weight @xmath115 in the set @xmath116 for @xmath105 ,
, @xmath106 from given @xmath101 , , @xmath114 , provided that @xmath117 , where @xmath118 is the standard inner product of @xmath23 and @xmath119 and @xmath107 , , @xmath108 is a basis of @xmath120 .
let @xmath121 , , @xmath122 and @xmath123 , , @xmath124 , @xmath125 , , @xmath126
. then @xmath127 .
since @xmath128 , eq .
( [ eq : weighte0 ] ) implies @xmath129 and the algorithm of farrn finds @xmath130 from @xmath101 , , @xmath114 correctly .
we can easily find @xmath103 from @xmath130 .
we shall consider the garcia - stichtenoth function field @xmath139 over @xmath140 with @xmath141 .
let @xmath142 . since the galois group of @xmath143 is isomorphic to the additive group of @xmath144 ( * ? ? ?
* proposition 1.1 ( i ) ) , there exists an automorphism @xmath145 of order @xmath41 .
let @xmath146 , and @xmath147 .
the zero divisors of @xmath148 consist of @xmath149 places of degree one ( * ? ? ?
* section 3 ) .
let @xmath150 be the fixed field of @xmath40 .
let @xmath3 be a zero of @xmath148 .
there exists a zero @xmath151 of @xmath148 such that @xmath152 and @xmath153 . since @xmath154 is galois , by theorem iii.7.1 of @xcite we have @xmath155 .
therefore we can write the zero divisor of @xmath148 as @xmath156 such that @xmath157 for all @xmath158 .
let @xmath159 . by proposition vii.1.2 of @xcite
, @xmath48 satisfies the condition ( [ etacondition ] ) .
let @xmath160 , and @xmath161 the unique pole of @xmath162 in @xmath139 .
we have @xmath163 the different exponent of @xmath164 is even at every place of @xmath139 ( see the text below lemma 2.9 of @xcite ) .
hence the discrete valuation of @xmath165 is even at every place of @xmath139 by remark iv.3.7 of @xcite .
observe that @xmath166 @xcite .
therefore the valuation of the divisor @xmath167 is an even integer at every place of @xmath139 .
define @xmath168 .
we have @xmath169 where @xmath170 is the genus of @xmath171 .
let @xmath172 be a nonnegative integer .
since @xmath173 , it satisfies the condition on @xmath50 in proposition [ prop2 ] .
let @xmath174 . since @xmath175 , @xmath176 . by corollary [ coro3 ] we can construct an @xmath177 $ ] quantum stabilizer code with @xmath178
let @xmath179 be a real number such that @xmath180 , and set @xmath172 to @xmath181 . by theorem
[ th : ak ] we can construct a sequence of @xmath182 $ ] binary quantum stabilizer codes with @xmath183 because @xmath184 converges as @xmath185 , which led the wrong value of @xmath186 .
the author apologize for misleading the reader .
] to @xmath187 as @xmath188 @xcite .
simple calculation shows @xmath189 . by choosing an appropriate value @xmath76 for every @xmath190
, we can construct a sequence of @xmath136 $ ] binary quantum codes with @xmath191 where @xmath192 the function @xmath193 is plotted in figure [ fig1 ] . in @xcite ashikhmin
et al . constructed an asymptotically good sequence of binary quantum codes from self - orthogonal classical algebraic geometry codes . in their construction
they do not use at least @xmath194 points on the curve ( see remark below theorem 4 of @xcite ) , where @xmath194 is the genus of the curve . by (
* proposition vii.1.2 ) we have @xmath195 for the code @xmath196 over @xmath140 constructed in the proof of proposition [ prop : sequence ] , where @xmath197 is the dual code of @xmath196 with respect to the standard inner product . in the construction of @xmath196
we asymptotically use all the points on the curve .
therefore we can construct a better sequence of binary quantum codes if we use @xmath196 in the construction of ashikhmin et al .
let us calculate the asymptotic parameters of the sequence constructed by the method of ashikhmin et al . with @xmath196 .
let @xmath198 ( @xmath199 ) be the code length of @xmath196 as a classical linear code .
we have @xmath200 and the minimum hamming distance of @xmath196 is not less than @xmath201 by the construction of binary quantum codes in @xcite , from the inclusion of classical codes @xmath202 for @xmath203 , we can construct @xmath204 $ ] binary quantum codes with @xmath205 since @xmath206 @xcite , by setting @xmath207 we have @xmath208 where @xmath209 it is clear that eq .
( [ improvedalt ] ) is larger than eq .
( 21 ) of @xcite .
ashikhmin et al .
@xcite constructed an asymptotically good sequence of binary quantum codes from algebraic curves .
chen et al .
@xcite constructed a sequence based on the idea in @xcite better than @xcite in certain range of parameters .
their sequences and the sequences in this section are compared in figure [ fig1 ] .
note that chen @xcite also proposed the same construction of quantum codes as @xcite . | the stabilizer code is the most general algebraic construction of quantum error - correcting codes proposed so far .
a stabilizer code can be constructed from a self - orthogonal subspace of a symplectic space over a finite field .
we propose a construction method of such a self - orthogonal space using an algebraic curve .
by using the proposed method we construct an asymptotically good sequence of binary stabilizer codes . as a byproduct we improve the ashikhmin - litsyn - tsfasman bound of quantum codes .
the main results in this paper can be understood without knowledge of quantum mechanics . |
comparisons between galaxy populations throughout the redshift range of @xmath6 indicate that the red sequence has roughly doubled in mass during this timeframe , while the mass of the blue cloud is unchanged @xcite .
the red sequence gains mass from the quenching of blue galaxies , while the corresponding loss in blue cloud mass is balanced by the ongoing star formation within blue cloud galaxies .
this evolution is accompanied by an order of magnitude decrease in the cosmic star formation rate ( e.g. , * ? ? ?
* ) , a transition from disc - dominated to bulge - dominated galaxies @xcite , a decrease in the galaxy merger rate ( e.g. , * ? ? ?
* ) , and the hierarchical buildup of massive galaxies . galaxy - galaxy interactions and mergers are thought to contribute to this evolution by triggering the formation of new stars , by quenching star formation in gas - rich galaxies , and by moving galaxies up the red sequence via dry mergers @xcite .
close encounters are also thought to play a role in producing a wide variety of transient astrophysical phenomena , such as quasars @xcite , submillimetre galaxies @xcite , luminous infrared galaxies @xcite , and ultra - luminous infrared galaxies ( ulirgs ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite provided the first clear evidence of enhanced star formation in interacting galaxies , by comparing the optical colours of morphologically peculiar galaxies to normal galaxies . in recent years
, numerous lines of evidence have confirmed this finding .
the level of enhancement is typically about a factor of two ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but varies depending on the types of galaxies involved ( e.g. , massive galaxies vs. starforming galaxies ) , how the interactions are identified ( e.g. , visual classifications vs. presence of a close companion ) , how advanced the interaction / merger is , and the method used to measure the star formation ( e.g. , h@xmath7 equivalent width vs. far infrared emission ) . a number of recent studies of close galaxy pairs have demonstrated that the enhancement in star formation increases as pair separation decreases @xcite . studying this problem in reverse reveals that galaxies which are undergoing strong star formation have an increased likelihood of having a close companion ( e.g. , * ? ? ?
* ; * ? ? ?
there is also evidence that recent ( rather than ongoing ) star formation is enhanced in interacting galaxies , from studies of e+a galaxies @xcite , the absorption - line spectra of early type galaxies @xcite , and the recent star formation histories of ulirgs @xcite .
nevertheless , many questions remain .
assessment of the relative roles of gas - rich versus gas - poor galaxies in interactions and mergers has yielded some conflicting results , due in part to the methods used to identify these systems and detect their star formation .
interacting galaxies which are identified based on morphological signs of interactions are strongly biased towards systems with large gas fractions , and these tidal disturbances remain visible for longer in gas - rich systems @xcite . this issue can be avoided by identifying interacting systems via the presence of close companions .
however , many such close pair studies employ star formation rate ( sfr ) indicators which are primarily sensitive to gas - rich galaxies ( e.g. , those which use nebular emission lines ) ; low levels of enhanced star formation in gas - poor systems may therefore be overlooked .
moreover , these same sfr indicators are sensitive to relatively short - lived _ ongoing _ star formation , and therefore may only be able to identify signs of triggered star formation in systems which are seen very shortly after close passages .
finally , a perennial problem with studies of interacting galaxies is the question of `` nature versus nurture '' . that is , if one detects differences between interacting and non - interacting galaxies , it is difficult to distinguish between interaction - induced effects ( such as triggered star formation ) and pre - interaction differences ( e.g. , if interacting and non - interacting galaxies reside in different environments ) .
one relatively obvious way forward is to analyze the optical colours of galaxies in close pairs , and compare them to a fair sample of non - paired galaxies .
optical colours can be measured for all types of galaxies , and provide a clear method of distinguishing between gas rich and gas poor galaxies , due to the well - established bi - modality of galaxy colours ( e.g. , * ? ? ?
in addition , induced star formation alters the colours of galaxies , and on timescales which are considerably longer than those of the starbursts themselves .
however , while there have been a number of previous studies of the optical colours of galaxies in interacting / merging galaxies , they have yielded some conflicting results .
the earliest studies of close galaxy pairs generally found that the colours of galaxies in close pairs are similar to field galaxies ( see * ? ? ? * and references therein ) , although these studies were limited to small samples of galaxy pairs , often without redshifts . in recent years , large redshift surveys
have greatly increased the yield of close spectroscopic galaxy pairs , allowing differences to emerge .
@xcite find that galaxies in close pairs are bluer than galaxies in their parent sample .
other studies report an excess of both extremely blue and extremely red galaxies in close pairs @xcite and visually identified mergers @xcite .
interpretation of these and other results is complicated by the limited size of many close pair samples , deficiencies in the quality and size of the control samples , uncertainties about the quality of colours measured in these crowded systems , and the fact that some studies do not directly probe colour changes as a function of pair separation .
optical colours can also provide further insight into the degree to which triggered star formation is centrally concentrated .
simulations indicate that strong interactions can cause the infall of gas onto the central regions of galaxies , triggering star formation @xcite .
this process may contribute to the growth of bulges @xcite , even if the interactions do not lead to mergers .
several lines of observational evidence indicate that interaction - induced star formation tends to be centrally concentrated .
@xcite and @xcite use optical colours to infer that star formation is enhanced in the centres of interacting and merging galaxies , while @xcite use both optical colours and h@xmath7 equivalent widths to reach the same conclusion in a sample of close galaxy pairs .
@xcite use bulge versus disc colours to infer evidence of enhanced star formation in the bulges , but not the discs , of galaxies in close pairs .
@xcite find that the surface brightness profiles of toomre - sequence galaxies are consistent with the presence of newly formed stars in the centres of these merging galaxies , while @xcite find a central excess of core collapse supernovae in the cores of disturbed galaxies .
evidence for the infall of gas onto the centres of galaxies comes from an offset of the luminosity - metallicity relation and mass - metallicity relation to lower metallicities in close galaxy pairs @xcite , and a higher proportion of strongly disturbed systems in lower metallicity galaxies @xcite , and is consistent with predictions from the simulations of @xcite . however , there are also clear examples of galaxy - galaxy interactions which trigger off - centre star formation @xcite or galaxy wide star formation @xcite .
furthermore , @xcite find no excess of central star formation in a sample of galaxies with close companions , despite the overall sfrs of these galaxies being nearly twice as high as galaxies in their control sample .
we aim to shed new light on these issues by measuring the @xmath8 colours of galaxies in sloan digital sky survey ( sdss ) close pairs , and comparing them with a control sample of non - paired galaxies that are matched in both stellar mass and redshift .
in addition , by comparing with both wide separation pairs and close projected pairs ( i.e. , interlopers ) , we wish to tease apart colours differences which are due to interactions from those which result from environmental differences or poor photometry .
finally , by comparing global colours to central ( fibre ) colours , we will investigate the degree to which induced star formation is centrally concentrated .
compared with earlier studies of the colours of galaxies in close pairs , our study is unparalleled in terms of the size of the pairs sample , the quality of the photometry , the combination of global and central colours , the size and robustness of the control sample , the comparison with close projected pairs , and the use of colour offsets .
we describe the selection of our pairs and control samples in section [ secsample ] , along with our measurements of global and central colours . in section [ seccolour ] ,
we present the overall distributions of global colours in paired galaxies , along with their dependence on pair separations and relative velocities . in section [ secclass ] ,
we divide our sample into four subsets ( red sequence , blue cloud , extremely red , and extremely blue ) , and explore how the fractions and colours of galaxies in these subsets depend on projected separation .
we then assess the degree to which these trends are related to central ( rather than global ) colours by analyzing fibre colours ( [ secfibre ] ) .
we introduce a new measure called colour offset in section [ secoffset ] , and relate our findings to predictions from a simple starburst model ( section [ secsb99 ] ) .
we finish with our conclusions in section [ secconclusions ] .
we adopt a concordance cosmology of @xmath9 , @xmath10 , and @xmath11 mpc@xmath12 throughout the paper .
we wish to analyze the optical colours of galaxies in close pairs , and to compare them with a control sample of galaxies which do not have close companions .
in addition , we will compare close and wide pairs , in order to be certain that any effects attributed to ongoing interactions / mergers decline at wider separations , as would be expected in this scenario .
finally , we will compare close physical pairs to close projected pairs , in order to ensure that our findings are not adversely affected by poor photometry due to crowding . in this section
, we describe our initial acquisition of galaxies from the sdss data release 7 ( hereafter dr7 ) of @xcite , along with our selection of galaxies in pairs , and the creation of an unbiased control sample .
studies of close pairs of galaxies have benefitted greatly from the advent of large redshift surveys .
the availability of redshifts for both members of a close pair reduces the contamination due to unrelated foreground / background companions , and allows one to compare intrinsic galaxy properties as a function of projected physical separation and relative line - of - sight velocity .
we therefore begin by requiring all galaxies in our sample to have secure spectroscopic redshifts from the sdss dr7 ; specifically , we select galaxies from the specphoto table which have zconf @xmath13 0.7 ( i.e. , all redshifts are at least 70% secure ) .
we further limit our analysis to the sdss main galaxy sample @xcite , by requiring extinction - corrected petrosian apparent magnitudes of @xmath14 @xmath15 17.77 .
we impose an additional limit of @xmath16 in order to avoid the unreliable deblending of large galaxies , which can lead to single galaxies being misclassified as close pairs , triples , etc .
we impose a minimum redshift of 0.01 to ensure that redshifts are primarily cosmological , and we impose a maximum redshift of 0.2 to avoid the regime of high incompleteness and poor spatial resolution .
we also require all objects to be classified as galaxies both photometrically ( specphoto.type=3 ) and spectroscopically ( specphoto.specclass=2 ) . finally , studies of galaxy pairs benefit from knowledge of the luminosity or mass ratio of every pair .
we therefore also require every galaxy to be included in the mpa - jhu dr7 stellar mass catalogue .
these stellar masses were measured using fits to sdss @xmath17 photometry @xcite , rather than using spectral features @xcite , although in general these mass estimates agree very well . together , these criteria yield a catalogue of 615,196 galaxies .
we identify a sample of galaxy pairs following the general approach of @xcite . for each galaxy in the catalogue
described above , we identify the closest companion satisfying the following criteria : 1 .
projected physical separation of @xmath0 2
. line - of - sight rest - frame velocity difference of @xmath1 3 .
stellar mass ratio ( companion mass divided by host mass ) of @xmath18 mass ratio @xmath19 if several companions are found for a given host galaxy , the companion with the smallest @xmath20 is selected .
if no companions are found , the galaxy is designated as a potential control galaxy ( see below ) . finally , following @xcite
, we randomly remove 67.5% of galaxies which are in pairs with angular separations greater than 55@xmath21 , in order to compensate for the fact that pairs with smaller angular separations would otherwise be underrepresented in our sample . this small scale spectroscopic incompleteness results from the well known sdss fibre collision constraint @xcite , whereby one can not simultaneously acquire spectra for two galaxies within 55@xmath21 .
thankfully , plenty of these pairs are nevertheless present in sdss , due to overlap between adjacent plates and the use of two or more plates in some regions .
@xcite use the spectroscopic incompleteness measurements of @xcite to estimate that 67.5% of pairs at angular separations below 55@xmath21 are missed as a result of these fibre collisions .
we therefore exclude the same fraction of galaxies in pairs with separations @xmath13 55 @xmath21 .
application of all of these criteria to the catalogue described in section [ catalog ] yields a sample of 22,777 galaxies with a companion .
hereafter , we refer to these galaxies as `` paired galaxies '' . in order to ascertain
how interactions / mergers affect galaxy properties , we wish to compare our sample of paired galaxies with a sample of non - interacting galaxies .
however , if one simply selects galaxies at random from the remaining ( non - paired ) galaxies in the catalogue , the resulting distribution of stellar masses and redshifts will be significantly different from the pairs sample @xcite , due to the pair selection criteria , sdss fibre collisions , etc . given that many observed properties of galaxies are redshift dependent , and that intrinsic galaxy properties are known to correlate with stellar mass ( e.g. , * ? ? ?
* ) , we therefore create a control sample that is matched to the pairs sample in both stellar mass and redshift . unlike @xcite , we do not attempt to explicitly match our control sample to the environment of paired galaxies ; however , we do compare the environments of paired and control galaxies in section [ secrfrac ] , and we investigate the dependence of our results on environment in section [ secoffset ] .
the control sample is created by first finding the best simultaneous match in redshift and stellar mass for each paired galaxy .
this yields an initial control sample which is equal in size to the paired sample , with nearly identical distributions of redshift and stellar mass .
we confirm this agreement by carrying out kolmogorov - smirnov ( k - s ) tests on the two distributions , in which low significance levels would imply significantly different cumulative distribution functions ; we instead find significance levels of 100.0000% in each case .
we therefore increase the size of our control sample by repeating this exercise until the distributions in redshift and stellar mass of the control sample start to diverge significantly from those of the paired sample . with this approach , we are able to create a control sample that is 13 times larger than the pairs sample .
figure [ fighistzm ] confirms that the resulting redshift and stellar mass histograms appear to be in excellent agreement with one another , and this impression is verified by k - s test results of 64% and 69% respectively ( these values would have dropped to 11% and 27% respectively if we had included a fourteenth iteration of the matching procedure ) . the fact that the control sample is much larger than the pairs sample means that the control sample will make a negligible contribution to the statistical errors in our results .
we tag each control galaxy with the properties of the paired galaxy it was matched to : specifically , the projected separation @xmath20 , the rest - frame velocity difference @xmath22 , the stellar mass ratio , and the unique identifier ( sdss objid ) of the paired galaxy .
this will enable us to compare individual paired galaxies with their associated control samples , or to compare subsets of the paired sample ( selected based on any combination of pair properties ) with the corresponding control galaxies .
both approaches will prove to be very powerful in revealing differences in the colours of paired and control galaxies .
there are several important advantages of this approach .
the first is that we are able to account for changes in redshift and/or stellar mass as a function of pair separation , @xmath22 , etc .
this can be contrasted with other published studies in which the properties of control galaxies are averaged over the full sample , rather than being computed as a function of pair separation , etc .
secondly , we can compare any individual paired galaxy to its own control sample , thereby providing an additional tool for assessing which paired galaxies are most different from their controls . in the analysis and interpretation that follows later in this paper , the reader should keep the following two points in mind .
first , any differences between galaxies in the paired and control samples can not be due to differences in stellar mass or redshift ; therefore , these differences may tell us something fundamental about how interacting galaxies differ from non - interacting galaxies . secondly ,
if the properties of control sample galaxies are found to vary with pair properties ( e.g. , pair separation ) , this is likely due to variation in the stellar mass and/or redshift distribution of galaxies in the pairs sample .
while both types of information are in principle useful , the latter must be treated with caution , since changes in the stellar mass and/or redshift distribution may be the result of selection effects in the pairs sample ( e.g. , redshift - dependent selection effects , incompleteness in the stellar mass catalog , spectroscopic incompleteness , etc . ) . as a result
, we will focus primarily on differences between the pairs and control samples .
in addition to satisfying the basic requirements of our pair and control sample algorithms , we now further require that all galaxies in our sample have high quality global ( integrated ) rest - frame @xmath8 colours measured by @xcite .
these colours were computed using the galaxy image 2d ( gim2d ) software of @xcite , using simultaneous @xmath23 and @xmath24 band fits to the sdss images .
all colours are corrected for galactic extinction , and converted to rest - frame quantities using the @xmath25-correction software of @xcite .
@xcite demonstrate that these fits , which were carried out using improved background subtractions and segmentation maps , provide robust colour measurements for galaxies which have close companions .
this allows us to avoid known challenges with the photometry of crowded systems ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in particular , we require each galaxy in the pairs and control sample to satisfy the following criteria : 1 . successful gim2d simultaneous @xmath23- and @xmath24-band fit from @xcite 2 . rest - frame @xmath8 colour error less than 0.1 mag 3 .
the fibre colour predicted by the gim2d model fit must be within 0.1 mag of the observed sdss fibre colour ( fibre colour ) '' parameter defined and reported in @xcite must be within 0.1 mag of the control sample mean . ]
visual inspection confirms that the object is a distinct galaxy ( inspection is complete only for pairs with @xmath26 ) overall , 94% of galaxies in our preliminary paired and control galaxy samples satisfy all of these criteria .
this yields final samples of 21,347 paired galaxies and 261,023 control galaxies , with an average of 12.23 control galaxies for each paired galaxy .
we note that these samples are substantially larger than those of @xcite , since their pairs sample is restricted to dr4 galaxies with @xmath27 , and their control sample is only four times larger than their pairs sample ( due to their additional requirement for a match on local density ) .
the @xcite catalogue contains global colours for each galaxy in our paired and control galaxy samples , measured using bulge+disc decomposition .
these colours are representative of the galaxies as a whole , and should be sensitive to induced star formation that is either global or centrally concentrated
. however , the sdss database also provides measurements of fibre colours , which are measured within the central 3 arcseconds of each galaxy .
we have converted these to rest - frame fibre colours using the extinction corrections from the sdss database and the @xmath28corrections described in the preceding section .
these fibre colours provide a probe of the central star formation in each galaxy , although the degree to which these colours are non - global depends on the covering fraction ( cf ) of the fibres . moreover ,
if we wish to compare the fibre colours of paired and control galaxies , we must ensure that they have similar covering fraction distributions . in figure [ fighistcf ] , we present @xmath29band cf histograms of paired and control galaxies .
we find broadly similar distributions , with a small offset ( @xmath30 ) towards larger cf s for paired galaxies .
while these samples were not matched on cf , reasonable agreement is to be expected , given that each paired galaxy is matched in both redshift and stellar mass to its control galaxies .
the offset is in the direction expected , given that @xcite report that galaxies in a subset of this pairs sample have somewhat higher bulge fractions than control galaxies .
most importantly , the offset in cf between paired and control galaxies is sufficiently small that it should have a negligible influence on the resulting colours , allowing us to make a fair comparison of fibre colours between paired and control sample galaxies .
we also note that the median cf of paired galaxies is 29% , and 96% of paired galaxies have covering fractions of less than 50% .
therefore , these fibre colours do in fact provide a good probe of central colours .
moreover , unlike the global colours of @xcite , the fibre colours are model independent , and they are not affected by deblending problems due to close neighbouring galaxies or stars .
therefore , fibre colours also provide an independent check on the trends reported in this study .
armed with secure measurements of global and fibre colours , we now set out to find and understand any differences between the colours of paired and control galaxies .
our paired galaxy sample spans a line of sight rest - frame velocity difference of @xmath31 .
pairs with @xmath5 can not be physically close together , as the required peculiar velocities for this scenario are unphysical .
we therefore will refer to these systems as _ projected pairs _ ( interlopers ) , and will use them as a sanity check ; specifically , we would expect any real differences between paired and control galaxies to disappear in the projected pair sample . in order to focus on the effects of galaxy interactions
, we wish to impose a maximum @xmath22 on our non - projected pairs sample .
the primary motivation is to minimize chance superposition of non - interacting galaxies within relatively dense environments such as groups or clusters .
@xcite show that there is a correlation between @xmath22 and projected galaxy density , in that pairs with higher @xmath22 lie in regions of higher density ( within their sample of pairs with @xmath32 500 ) .
to guide our choice of a maximum @xmath22 , we plot in figure [ figgrdv ] the mean global colours of galaxies in pairs as a function of @xmath22 , considering three subsets of the pairs sample : close pairs ( @xmath2 ) , intermediate separation pairs ( @xmath33 ) and wide pairs ( @xmath34 ) .
we find that pairs with @xmath35 have mean colours which are quite red with respect to both low and high velocity pairs .
this is true for close , intermediate , and wide separation pairs , indicating that this trend is unlikely to be associated with galaxy interactions / mergers .
instead , we interpret these redder colours as being due to the higher densities probed by these relative velocities @xcite , and the fact that galaxy colour and local density are correlated .
this is consistent with the relatively high proportion of late type galaxies in low velocity pairs ( @xmath3 ) reported by @xcite .
while these higher velocity pairs are certain to include some ongoing interactions and eventual mergers ( and in fact clear signs of interactions are seen in the images of some systems ) , we elect to avoid the expected high superposition rate in this regime by imposing a maximum @xmath22 of 200 , which comfortably avoids these higher velocity environments while allowing us to retain a sizeable sample of pairs . for comparison ,
other close pair studies have imposed less restrictive @xmath22 limits ranging from @xmath36 @xcite , to 500 @xcite and up to 1000 @xcite .
we now proceed to compare the colours of galaxies in low velocity pairs ( @xmath3 ) with their control galaxies . in figure
[ figgrmeannew ] , we plot mean global colours as a function of @xmath20 , for both paired and control galaxies . compared with the strong dependence of mean colour on @xmath22 that was seen in figure [ figgrdv ] , we find that the global colours of low velocity pairs vary much less with @xmath20 .
the mean colour of galaxies in pairs decreases smoothly as pair separation decreases , with galaxies in the closest pairs being on average about 0.01 mag bluer than galaxies in the widest pairs .
no significant change in the mean colours of the associated control galaxies is seen over this range in @xmath20 .
the closest pairs have global colours which are equivalent to their controls , whereas the widest pairs are @xmath37 mag redder on average . based on the lack of a dependence of mean colour on @xmath20 , one might be tempted to conclude that ( a ) the colours of galaxies are generally unaffected by galaxy - galaxy interactions or ( b ) that few of the galaxies in close pairs are actually undergoing interactions . however , as has been shown elsewhere ( e.g. , * ? ? ?
* ) , _ mean _ colours alone are not very sensitive to changes within galaxy populations .
for example , if there are excesses of both extremely red and extremely blue galaxies within samples of interacting / merging galaxies , as has been reported elsewhere ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , these effects might cancel out , at least in part , when computing mean colours . with this in mind
, we now compare the colour distributions of paired and control galaxies .
we begin by plotting histograms of global colours in figure [ figgrhist ] , for close pairs , wide pairs , and projected pairs .
overall , the distributions of paired galaxy colours are broadly similar to those of the associated controls , with a distinct red sequence and a more extended distribution of blue galaxies ( the `` blue cloud '' ) .
this is the aforementioned colour bimodality , within which the relative proportion of red versus blue galaxies has been found to depend on environment and luminosity @xcite .
however , we find a small but significant deficit in galaxies with intermediate global colours ( @xmath38 ) for galaxies in close and wide pairs with @xmath32 200 ( relative to their control samples ) , and an excess of galaxies which are relatively red ( on the redward half of the red sequence peak ) ] .
these differences are not seen in the projected pairs sample .
in addition , a small but significant population of extremely blue galaxies ( @xmath39 ) is seen in the close pairs sample , but is greatly diminished in the wide pairs sample , and non - existent in the projected pairs sample .
this lends support to the notion that these extremely blue galaxies may be directly associated with galaxy - galaxy interactions .
we will return to this intriguing sub - population in section [ seceber ] .
in order to further examine these differences between paired and control sample galaxies , we now divide the sample into four subsets based on colour and absolute magnitude . figure [ figcmdcat ] illustrates this division of the sample into extremely red galaxies , red sequence galaxies , blue cloud galaxies , and extremely blue galaxies .
the division between red sequence and blue cloud galaxies corresponds to a line with slope @xmath40 which passes through @xmath8 = 0.65 at @xmath41 .
the slope provides a good fit to the colour magnitude relation seen on figure [ figcmdcat ] , and the intercept was selected by examining the colour histograms of figure [ figgrhist ] . throughout the remainder of this paper
, this division is used to distinguish between red and blue galaxies .
we also identify subsets of extremely red and extremely blue galaxies .
the criterion of @xmath42 0.9 at @xmath41 for extremely red galaxies applies to the reddest 1% of galaxies in projected pairs , and is sufficiently red that galaxies are unlikely to have been scattered from the red sequence ( recall from [ secgim2d ] that all galaxies are required to have @xmath8 colour errors of @xmath43 0.1 mag ) .
this threshold is slightly less strict than the @xmath8 = 0.95 cut used by @xcite .
our threshold for extremely blue galaxies corresponds to @xmath8 = 0.3 at @xmath41 , and applies to the bluest @xmath441% of galaxies in the projected pairs sample .
this threshold is notably stricter than the extremely blue cut of @xmath8 = 0.4 employed by @xcite .
@xcite find that rising sfrs are needed to produce colours bluer than @xmath8 = 0.3 .
the fraction of galaxies which are classified as red ( either red sequence or extremely red ) , hereafter called the red fraction , is plotted versus projected separation in the lower panel of figure [ figbluered ] .
the red fraction of paired galaxies is consistently larger than the associated control sample at all separations , although this difference may decline at smaller separations .
these findings are consistent with the dependence of mean colours on separation reported in section [ secgrrp ] , and with the excess of red sequence galaxies ( and corresponding deficit of blue cloud galaxies ) described in section [ secgrhist ] .
we note that other studies have also reported that galaxies in pairs are significantly redder than galaxies without nearby companions ( e.g. , * ? ? ?
* ) , with correspondingly higher bulge fractions than their isolated counterparts @xcite . the most obvious cause of this difference would be if pairs reside in higher density environments , since the red fraction is known to increase with density @xcite .
@xcite show that gas - poor pairs reside preferentially in higher density environments , and that this is due primarily to the colour - density relation .
@xcite find that paired galaxies in simulations occupy higher - mass haloes than isolated galaxies , and predict that this should lead to mean @xmath8 colours which are about 0.05 mag redder than field galaxies .
the fact that we find a smaller difference than this ( @xmath45 mag in the mean ; [ secgrrp ] ) is likely due to the fact that our control sample is matched to the pairs in both stellar mass and redshift , thereby providing a fairer comparison than random field galaxies .
nevertheless , we can test this hypothesis directly by using the @xcite measurements of projected local density ( @xmath46 ) as a probe of environment .
@xmath46 is computed using the distances to the fourth and fifth nearest neighbours within 1000 .
these measurements have recently been updated to include dr7 galaxies .
the suitability of these measurements for close pair studies is addressed by @xcite .
the @xcite requirement for redshifts to lie within the range 0.010 - 0.085 means that these measurements are available for only 57% of the galaxies in our full pairs sample .
however , as our control sample is matched in redshift to the pair sample , we are still able to make a fair comparison between the local densities of paired and control galaxies . in figure [ figdensity ] , we provide histograms of @xmath46 for paired and control galaxies . as with figure [ figgrhist ]
, we separate our pairs sample into close pairs , wide pairs , and projected pairs .
we find that galaxies in close and wide pairs are skewed to higher local densities than their associated control galaxies , with this difference being nearly two times larger in wide pairs than in close pairs .
no significant difference is seen in the @xmath46 distributions of projected pairs and their control galaxies .
these findings are consistent with the hypothesis that the higher fraction of red galaxies in close and wide pairs is due to these systems residing in higher density environments than their control galaxies , and the inference that more of these pairs result from chance superpositions in overdense regions @xcite .
moreover , the fact that this excess is more pronounced for wide pairs than close pairs is consistent with the expectation that chance superpositions within groups and clusters should be more common in wide pairs than in close pairs @xcite .
we now proceed to assess the colours of red and blue galaxies separately .
the middle and upper panels of figure [ figbluered ] show the mean colours of galaxies classified as blue ( blue cloud or extremely blue ) and red ( red sequence or extremely red ) respectively . in both cases
we see clear differences between paired and control galaxies . blue galaxies in pairs are bluer than blue galaxies in the control sample at all separations , with a typical offset decreasing from @xmath44 0.03 mag for close pairs ( @xmath2 ) to 0.02 mag for wide pairs ( @xmath4 ) .
in contrast , red galaxies in pairs are _ redder _ than red control galaxies by @xmath44 0.01 mag at the closest separations , with this relatively small offset becoming negligible beyond @xmath44 25 .
the fact that these trends go in opposite directions ( blueward and redward ) means that they will reduce any associated trends in mean colour or red fraction , which is consistent with our findings in sections [ secgrrp ] and [ secrfrac ] . in section [ secgrhist ] and figure [ figgrhist ] , we noted a population of extremely blue galaxies which are present in the close pair sample but nearly absent in the wide pair sample ( and non - existent in the projected pair sample ) .
this is reminiscent of the excess populations of extremely blue galaxies reported in several close pair studies @xcite , although we do not detect an obvious population of of extremely red galaxies as found in these same studies . to explore this further
, we first plot the fraction of galaxies in our @xmath3 pairs which are extremely blue in the lower left panel of figure [ figeber ] .
we find a clear excess of extremely blue galaxies , rising from @xmath44 3% at wide separations to @xmath44 7% at small separations .
a similar but less pronounced trend is seen in the control sample , implying that part of the trend in paired galaxies is due to a change in the mix of stellar mass and/or redshift with separation .
we repeat these measurements for our sample of projected pairs ( @xmath47 ) in the middle left panel of figure [ figeber ] , finding no rise as @xmath20 decreases ) than low velocity pairs ( @xmath32 200 ) , since correlated pairs are much more common than chance superpositions .
this explains why the error bars on the projected pairs sample are larger , and why these data do not extend as far inwards in @xmath20 . ] .
this essentially rules out the possibility that the rise towards small @xmath20 seen for true close pairs could be due to poor photometry in crowded pairs .
finally , we also compute the ratio of the extremely blue fractions in pairs versus control ( upper left hand panel of figure [ figeber ] ) , in order to see how the rise in the extremely blue fraction of pairs compares with the rise for the control sample .
contrary to what is observed for projected pairs ( dashed line ) , the ratio for low velocity pairs ( solid line ) is significantly higher than unity throughout the full range of @xmath20 , rising from about 1.5x for wide pairs to about 2x for the closest pairs . in the right hand panels of figure [ figeber ]
, we apply the same approach to the extremely red galaxies .
we find a much lower fraction of galaxies in this category , rising from @xmath48 for wide pairs to @xmath49 for the closest pairs .
the control sample remains constant at about @xmath50 for all separations .
the ratio of pairs to control exhibits a gradual rise in extremely red galaxies towards the smallest pair separations , reaching a factor of 2.5x at @xmath51 .
however , a comparable rise is also seen in the projected pairs sample , leading us to believe that this rise may be due to residual problems with the photometry of crowded systems . together , our results are qualitatively consistent with the high fractions of extremely blue galaxies in pairs reported by @xcite , @xcite , and @xcite , but appear to be at odds with the substantial fractions of extremely red galaxies described in these same studies .
for example , @xcite find extremely red fractions ranging from @xmath52 to @xmath53 for three different environmental classes , using a _
stricter _ criterion ( @xmath54 ) than we do ( @xmath55 at @xmath41 ) .
we have found that a substantial number of galaxies in pairs are classified as extremely blue ( up to @xmath56 in the closest pairs ) , whereas very few are classified as extremely red ( @xmath57 in the closest pairs ) .
the widest pairs also contain roughly 50% more extremely blue galaxies than the control sample , but no excess of extremely red galaxies .
moreover , the trends seen in the extremely blue ( red ) fraction are absent ( present ) in the projected pairs sample .
this implies that poor photometry can not explain the extremely blue population , but may explain the extremely red population .
@xcite provide a clear demonstration that the standard sdss pipeline does a poor job of galaxy photometry for closely separated pairs , whereas their recomputed gim2d global colours are much more secure . to directly assess the effects of using these different measurements of @xmath8 colours
, we compute the extremely blue and extremely red fractions in figure [ figebercomp ] , using our gim2d global colours ( bottom row ) , sdss petrosian colours ( middle row ) , and sdss modelmag colours ( top row ) .
exactly the same set of paired and control galaxies is used in each case .
the rise in the extremely blue fraction of close pairs towards small separations is seen with all three colour indices .
conversely , a large increase in the extremely red fraction of close pairs towards small separations is seen only with sdss petrosian and modelmag colours , reaching 6% and 8% respectively .
this provides compelling evidence that poor photometry is in fact largely responsible for the large extremely red fractions seen in close pairs when using photometry directly from sdss , and that the re - computed colours used in our analysis are effective in removing nearly all of these anomalously red systems .
related factors which may contribute to the lower extremely red fractions found in our study include our imposition of a bright apparent magnitude limit of @xmath16 ( which allows us to avoid the brightest galaxies where deblending is particularly problematic ) and our use of a lower relative velocity threshold than most other studies ( which preferentially avoids pairs in more crowded regions ) .
we caution that the small rise in our gim2d extremely red fraction at small @xmath20 could be due to residual effects from poor photometry in crowded systems .
more importantly , it seems likely that published reports of large fractions of extremely red galaxies in close pairs or merging galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) are the result of poor sdss photometry , rather than dust obscuration or other physical effects associated with induced star formation .
further insight into the nature of these extremely blue and extremely red galaxies can be gleaned by visual inspection of their images , as shown in figure [ figbluestreddest ] .
clear morphological signs of interactions are seen within both sets of galaxies .
there are obvious indications of dust in some of the extremely red galaxies .
some of this dust could have been stirred up as a result of galaxy interactions ( e.g. , * ? ? ?
however , in some cases , the dust is associated with edge - on discs ; the ensuing reddening might be expected to be on the order of 0.1 mag in @xmath8 @xcite .
nevertheless , edge - on discs would be expected to be present in equal measure in the control sample , so they are unlikely to be responsible for differences between paired and control galaxies .
finally , we note that these images provide vivid evidence of the well - known holmberg effect @xcite , in that galaxies within individual pairs tend to have colours which are similar to one another ( i.e. , there are few red - blue pairs ) .
in section [ secfibredata ] , we described our computation of rest - frame fibre colours , and the suitability of these colours as a probe of central ( rather than global ) star formation . in this section ,
we begin by analyzing the distribution of fibre colours in close , wide and projected pairs .
we then investigate the differences between fibre and global colours . in figure [ figgrfhist ] , we compare the fibre colours of paired and control galaxies , for close , wide and projected pairs .
we find a small but significant excess of extremely blue galaxies in close pairs , which is barely detectable in wide pairs and absent in projected pairs . on the other hand
, we find no excess of paired galaxies with extremely red fibre colours .
this is consistent with our hypothesis in [ seceberphot ] that residual problems with the photometry of crowded systems is responsible for the small population of galaxies with extremely red global colours , since this effect should be much smaller when using fibre colours .
we also find a significant deficit of galaxies in close pairs with intermediate fibre colours ( 0.5 @xmath58 0.75 ) , and a corresponding excess of galaxies on the red half of the red sequence .
this effect is smaller than was seen using global colours , particularly for wide pairs .
we again attribute this difference between paired and control galaxies to the higher density environments of paired galaxies . in the preceding subsection
, we compared the distributions of fibre colours in close , wide and projected pairs to those found using global colours . while useful
, this approach does not tell us how fibre and global colours compare on a galaxy - by - galaxy basis . if a subset of interacting galaxies experience centrally - concentrated bursts of star formation , we might expect to find evidence of this effect in the colour gradients of these galaxies . in this section
, we will use the difference between fibre and global colour as a probe of this effect . in figure [ figfiboffall ] ,
we plot the difference between fibre and global colour versus global colour for four subsets of the pairs sample : close low velocity pairs ( upper left panel ) , wide low velocity pairs ( lower left panel ) , close projected pairs ( upper right panel ) , and wide projected pairs ( lower right panel )
. overall , fibre colours are redder than global colours ; i.e. , @xmath8 ( fibre ) - @xmath8 ( global ) is positive .
this is as expected , given that bulges are generally redder than discs .
the difference between fibre and global colours is largest for galaxies of intermediate colour .
the most obvious explanation is that galaxies of intermediate colour are the least likely to be either bulge - dominated or disc - dominated galaxies . in any case , we are primarily interested in how this index differs between paired and control galaxies .
figure [ figfiboffall ] shows that , for red galaxies , there is good agreement between the fibre - global colours of paired and control galaxies , for both wide and close pairs .
conversely , for blue galaxies , close low velocity pairs have fibre colours which are typically 0.03 mag bluer than those of their associated control galaxies ( upper left panel ) , with this difference dropping to 0.015 mag for wide low velocity pairs ( lower left panel ) .
this difference is absent in the projected pairs sample ( right hand panels ) , confirming that crowding can not be responsible for the offset .
this is a highly significant effect , and is our strongest indication yet that centrally triggered star formation is taking place in some of these systems .
figure [ figcentralsb ] gives a particularly striking example of a system in which morphological signs of an interaction accompany a relatively blue central colour .
the fact that the mean offset is largest in close pairs , but still present in wide pairs , implies that the effects of this star formation on galaxy colours diminishes as galaxies move apart after close encounters , but that the effects persist long enough that they are still present in some widely separated pairs .
we explore this scenario further in section [ secsb99 ] .
the increase in extremely blue galaxies in close pairs , combined with the relatively high _ red _ fraction of galaxies in close and wide pairs , demonstrates that galaxies in pairs can be either bluer or redder than their control counterparts , for different reasons .
the former appears to be caused by ongoing galaxy - galaxy interactions , whereas the latter can be attributed to the higher density environment occupied by paired galaxies , and may therefore be true for pre - interaction galaxy pairs too . in principle , these competing effects mean that some of the colour differences between paired and control galaxies will cancel out when comparing their colour distributions .
that is , only the _ net changes _ in the colour distributions will be detected .
however , it may be possible to uncover more of the underlying colour differences by comparing , _ on an individual basis _ , the colour of every paired galaxy with the @xmath44 12 galaxies in its associated control sample .
this method has the potential to detect much more of the underlying differences in galaxy colours , and may even allow us to identify which galaxies have had their colours changed the most as a result of ongoing / recent interactions . to this end , we compute the _ colour offset _ for every paired galaxy in the sample .
we define colour offset as the colour of the paired galaxy minus the mean colour of its associated ( @xmath44 12 ) control galaxies .
we also compute the colour offset of every control galaxy , by computing the difference between its colour and the mean colour of the remaining @xmath59 associated control galaxies .
we then compute the difference between the colour offsets of paired and control galaxies ( hereafter , we will refer to this quantity as @xmath60 ) .
if paired galaxy colours are drawn at random from the same parent population as control galaxies , we would expect to find @xmath60 to be zero , on average .
figure [ figgroffsettrend ] plots @xmath60 as a function of projected separation for low and high velocity pairs , and treats blue and red galaxies separately . the left hand column of this figure refers to global colours , and the right hand column refers to fibre colours .
for blue galaxies , a highly significant negative ( blueward ) fibre colour offset is found in low velocity pairs , changing smoothly from about 0.02 mag for wide pairs to about 0.075 mag for the closest pairs . at the smallest separations ,
this difference is significant at the 11@xmath61 level .
this trend is much more modest in global colours , reaching @xmath44 0.02 mag for the closest pairs .
this trend is absent in the high velocity ( projected ) pair sample , demonstrating that crowding errors can not be responsible for this effect .
images of blue galaxies in very close pairs ( @xmath62 ) which have the greatest blueward offsets are presented in figure [ figoffblue ] .
many of these systems exhibit clear morphological signs of interactions , along with indications of relatively blue central colours .
this demonstrates that our colour offset parameter does in fact appear to be effective at identifying systems with atypical star forming properties ( note that more than half of the galaxies in figure [ figoffblue ] do not qualify as extremely blue , and therefore do not stand out when using colour rather than colour offset ) .
together , these results are consistent with the presence of central induced star formation which is strongest in the closest pairs and weaker ( but still present ) in wide pairs .
a much weaker trend is found for red galaxies in low velocity pairs , with an increase ( reddening ) in @xmath60 of up to 0.015 mag for the closest pair separations .
the size of this effect is the same in fibre and global colours , implying that the effect is global .
however , it appears from figure [ figgroffsettrend ] that a comparable trend may also be present in the projected pairs sample ; therefore , we are unable to rule out the possibility that crowding errors are responsible for this trend . in a related study , @xcite found a small bluing of the bulges , and not the discs , of galaxies in close low velocity pairs .
this effect was seen only at low densities , and was interpreted as evidence of central triggered star formation .
given the striking blueward fibre offsets seen in the upper right panel of figure [ figgroffsettrend ] , we now investigate the dependence of this effect on projected local density ( @xmath46 ) .
measurements of @xmath46 are available for 57% of our paired galaxies , as described in section [ secrfrac ] .
we subdivide our sample into three equal bins ( tertiles ) of @xmath46 , and present the global and fibre colour offsets of each in figure [ figgrdensity ] .
this figure indicates clearly that the blueward fibre offsets at small separations are driven by blue galaxies residing in low and medium density environments , though a small blueward fibre offset is also detected in the highest density tertile .
small blueward offsets in _
global _ colours ( left hand column of figure [ figgrdensity ] ) are seen at small separations at low and medium densities , but not at high density .
the fact that the total ( blue+red ) population closely traces the red population in the high density regime and traces the blue population in the low density regime is consistent with the well known colour - density relation , whereby the red fraction increases with density ( see [ secrfrac ] ) .
compared with figure 9 of @xcite , we find a larger and more significant difference between the colours of paired and control galaxies at small separations , and we find that this difference extends further into the medium- to high - density regime .
we attribute this added sensitivity to triggered star formation to several factors : ( 1 ) our dr7 pairs sample is larger than the dr4 sample of @xcite , ( 2 ) we have approximately 3 times as many control galaxies per paired galaxy as @xcite , ( 3 ) colour offsets are sensitive to colour differences on a galaxy - by - galaxy basis , and ( 4 ) treating blue and red galaxies separately is effective in isolating the effects of triggered star formation to the ( blue ) population of galaxies which is most susceptible to this process . one significant difference between our study and several other close pair studies ( e.g. , * ?
* ; * ? ? ?
* ; * ? ? ?
* ) is that we do not attempt to match our control sample to the projected local densities of paired galaxies .
as demonstrated in section [ secrfrac ] , our resulting control sample is skewed to lower densities than the pairs sample .
this has some implications for the interpretation of differences between paired and control galaxies in our study .
we therefore investigate this issue by regenerating figure [ figgroffsettrend ] using a control sample which _ is _ matched on local density .
this revised control sample is generated using the same methodology as outlined in section [ seccontrol ] , but now matching simultaneously on redshift , stellar mass and @xmath46 .
however , since it is more difficult to find matches for paired galaxies in the highest density environments , we restrict our analysis to galaxies with log@xmath63 ( this excludes 12% of galaxies in our paired sample ) .
we are able to find 3 control galaxies per paired galaxy .
the results are given in figure [ figgrdensityc ] .
comparison with figure [ figgroffsettrend ] reveals very similar trends in colours offsets . with a density - matched control sample
, there is a slight blueward shift in the global and fibre colour offsets of blue and red galaxies .
this small shift may result from the removal of galaxies in the highest density regime , since we know from figure [ figgrdensity ] that galaxies in the highest density environments exhibit the smallest blueward colour offsets .
another factor which may contribute to this shift is the fact that our revised control sample is no longer biased towards lower densities than paired galaxies ; this would be expected to further accentuate the blueward colour offsets in pairs due to induced star formation .
we conclude that the main results of our study are unchanged if a density - matched control sample is used .
the interpretation of the colours of interacting galaxies can be aided by employing model starbursts which predict how colours evolve during and after a starburst .
we create a simplistic galaxy+starburst model by starting with two pre - existing galaxies ( a red sequence galaxy and a blue cloud galaxy ) and superimposing a model starburst upon each .
the goal here is to see how the colour of each galaxy changes with time as a result of the superimposed starburst . for the colours of the starburst itself
, we use an instantaneous starburst99 model starburst @xcite , with @xmath64 , @xmath65 , and @xmath66 , converting from model @xmath67 colours to sdss @xmath8 .
we monitor these colours for 1 gyr following the starburst . for the pre - existing blue cloud galaxy ,
we use a starburst99 model galaxy with a continuous star formation rate of @xmath68/year , and add the instantaneous starburst after 10 gyr .
this galaxy has a @xmath8 colour of 0.43 immediately preceding the starburst , and is therefore representative of the blue cloud galaxies seen in figure [ figcmdcat ] . for the pre - existing red sequence galaxy ,
we begin with an instantaneous starburst , and let it age for 10 gyr , yielding a galaxy colour of 0.83 , which lies within the red tail of the red sequence seen in figure [ figcmdcat ] .
we then add the model starburst and monitor how the galaxy s colour changes .
the evolution in colours of these galaxy+starburst models is shown in figure [ figgrt ] , for burst strengths ( by stellar mass ) of 10% , 20% , and 30% .
this figure shows that a 20% starburst within a pre - existing red sequence galaxy will cause the galaxy to become bluer by @xmath69 mag , with this offset persisting for @xmath70 myr before gradual reddening begins .
even at 1 gyr after the starburst , this galaxy is considerably bluer than it was initially .
conversely , for a starburst that occurs in a blue cloud galaxy , the galaxy will become @xmath71 mag bluer for about 400 myr , and will return to its pre - starburst colour after 1 gyr . to compare with our results from the previous section
, we note that the difference in colour offset between paired and control galaxies ( see figure [ figgroffsettrend ] ) should be analogous to the offsets between our model starburst galaxies and their pre - existing counterparts ( figure [ figgrt ] ) , if all galaxies in our close pair sample were undergoing interactions with induced star formation . in reality , of course , not all of the galaxies in close pairs can be undergoing interactions , as some will not be close in three dimensions ( i.e. , interlopers ) , some will be approaching one another and therefore will not yet have had a close encounter , and others may be undergoing interactions without triggered star formation .
therefore , we would expect the mean colour offsets in our close pairs sample to be smaller than our model predictions , perhaps by a factor of a few . in figure [ figgroffsettrend ]
, we found that blue galaxies in close pairs have mean global colour offsets which are @xmath44 0.02 mag bluer than their control galaxies .
this is smaller than the @xmath44 0.05 mag offset predicted for a 20% starburst in a blue galaxy ( figure [ figgrt ] ) , but consistent within the hypothesis that 40% of galaxies in close pairs are experiencing induced starbursts of this strength .
we find a much higher offset of 0.075 mag in fibre ( rather than global ) colours .
this makes sense if the fractional starbursts in the central regions of these galaxies are substantially higher than 20% , as would be the case if most of the induced star formation is centrally concentrated , as implied by figure [ figfiboffall ] . finally , the blueward offsets we detect decrease markedly going from close to wide pairs .
this is consistent with the aging and subsequent reddening of our models in figure [ figgrt ] , and implies that we are seeing starbursts age as galaxies in close pairs separate following close peri - centre encounters .
the fact that this offset is still visible at the largest separations probed ( 80 ) is consistent with a simple calculation showing that a pair of galaxies separating at 200 in the plane of the sky will take @xmath70 myr to reach @xmath72 . for red galaxies
, we find a small _ redward _ offset in close pairs ( figure [ figgroffsettrend ] ) , which may be due to residual errors with the photometry .
we see no indications of the potentially large blue colour offsets that are predicted by our model galaxy+starburst for red galaxies .
this appears to indicate that starbursts of order 10 - 30% are not commonly found in red galaxies . while it is obvious that sufficiently large offsets would move red sequence galaxies into the blue cloud , the absence of even a small blueward offset for typical red galaxies appears to rule out this scenario
this finding is not unexpected , however , given that red sequence galaxies are generally depleted in gas , and therefore less capable of triggered star formation .
while deep imaging reveals that gas - poor ( early type ) galaxies show frequent signs of tidal disturbances , these signs of interactions are not generally accompanied by star formation @xcite .
@xcite find a correlation between the specific sfr and pair separation for galaxies in blue pairs , but not in red pairs , confirming this interpretation .
moreover , while few red mergers contain significant amounts of dust @xcite , @xcite find that , in intermediate to high density environments ( within which many of our close pairs lie ) , galaxies which have had most of their star formation suppressed tend to have their remaining star formation obscured . in summary
, it appears that red sequence galaxies in close pairs exhibit little in terms of induced star formation in the optical , with few ( if any ) unobscured starbursts with large burst strengths .
we have compiled a large , well - defined sample of 21,347 sdss dr7 galaxies in pairs with @xmath0 , @xmath1 , and stellar mass ratios between 0.1 and 10 .
we have measured high quality @xmath8 global colours for each galaxy , and have acquired their central ( fibre ) colours .
we have also created a very large control sample which is matched to the pairs sample in stellar mass and redshift , with @xmath73 control galaxies associated with every paired galaxy . by comparing galaxies in close pairs with their control samples , with wider separation pairs , and with close projected pairs ( i.e. , interlopers ) ,
we have been able to distinguish trends which are associated with interactions from those which are due to differences in environment or those which result from photometric errors due to crowding .
our findings can be summarized as follows : 1 .
60% of galaxies in close and wide pairs are classified as red , compared with 56% of control galaxies .
these paired galaxies are found in higher density environments than their controls .
we interpret these results as an indication that galaxies which are involved in interactions are preferentially red before the interactions start , due to the older stellar populations which are present in higher density environments .
galaxy - galaxy interactions make blue galaxies in close pairs ( @xmath2 and @xmath32 200 ) bluer by an average of 0.075 mag in fibre colours , and 0.02 mag in global colours .
these colour offsets are diminished but still detectable out to pair separations of at least 80 , and are strongest at low and medium projected local densities .
the fraction of extremely blue galaxies rises from about 3% for wide pairs to 8% for close pairs .
the use of projected pairs ( interlopers ) and alternate colour measurements confirm that this effect is not due to photometric errors .
galaxy - galaxy interactions appear to have little ( and perhaps no ) effect on the optical colours of red galaxies in pairs .
the slight reddening we detect at small separations ( up to 0.015 mag in global and fibre colours ) could be due to dust obscuration , but the rapid decline with pair separation and the absence of any difference between global and fibre colours instead suggest that residual problems with crowded - field photometry may be responsible .
5 . unlike previous studies of close pairs
, we find no significant excess ( @xmath74 ) of extremely red galaxies in close pairs .
we demonstrate that this is due to our improved photometry in crowded systems , given that we do find a strong excess ( @xmath75 ) if we replace our gim2d colours with petrosian or modelmag colours from the sdss database .
6 . at a fixed global colour
, blue cloud galaxies in close pairs have bluer fibre colours than control galaxies , with this difference decreasing as pair separation increases .
no such difference is found for red galaxies at any separations .
our simple starburst+galaxy model predicts that a 20% induced starburst should make a blue cloud ( red sequence ) galaxy bluer by about 0.05 mag ( 0.15 mag ) , and should persist for @xmath44 400 myr before starting to diminish .
our observed colour offsets indicate that starbursts such as this are in fact found in blue cloud galaxies in pairs , but are absent in red sequence galaxies in pairs . together
, these results provide further evidence that gas - rich galaxies in close pairs undergo induced star formation during close peri - centre passages , with the starburst then aging as the galaxies move apart from one another .
fibre colours confirm that this star formation is centrally concentrated , and measurements of projected local density show that this process occurs primarily in low- to medium - density environments .
we find no evidence from optical colours for such induced star formation in red sequence galaxies , thereby confirming that any such star formation is likely to be obscured .
we refute earlier claims of a substantial excess of extremely red galaxies in close pairs .
we thank the anonymous referee for insightful comments which significantly improved this paper .
we are grateful to anja von der linden and the mpa / jhu group for access to their data products and catalogues ( maintained by jarle brinchmann at http://www.mpa-garching.mpg.de/sdss/ ) .
we thank ivan baldry for making his dr7 projected local density measurements publicly available .
sle and drp gratefully acknowledge the receipt of nserc discovery grants which funded this research .
the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
abazajian , k. n. , et al .
2009 , , 182 , 543 abbas , u. , & sheth , r. k. 2006 , , 372 , 1749 alonso , m. s. , tissera , p. b. , coldwell , g. , & lambas , d. g. 2004 , , 352 , 1081 baldry , i. k. , glazebrook , k. , brinkmann , j. , ivezi , . , lupton , r. h. , nichol , r. c. , & szalay , a. s. 2004 , , 600 , 681 baldry , i. k. , balogh , m. l. , bower , r. g. , glazebrook , k. , nichol , r. c. , bamford , s. p. , & budavari , t. 2006 , , 373 , 469 balogh , m. l. , baldry , i. k. , nichol , r. , miller , c. , bower , r. , & glazebrook , k. 2004 , , 615 , l101 barton , e. j. , geller , m. j. , & kenyon , s. j. 2000 , , 530 , 660 barton gillespie , e. , geller , m. j. , & kenyon , s. j. 2003 , , 582 , 668 barton , e. j. , arnold , j. a. , zentner , a. r. , bullock , j. s. , & wechsler , r. h. 2007 , , 671 , 1538 bergvall , n. , laurikainen , e. , & aalto , s. 2003 , , 405 , 31 blanton , m. r. , & roweis , s. 2007 , , 133 , 734 brown , m. j. i. , et al .
2009 , , 703 , 150 chen , y. , lowenthal , j. d. , & yun , m. s. 2010 , , 712 , 1385 conselice , c. j. , chapman , s. c. , & windhorst , r. a. 2003 , , 596 , l5 cooper , m. c. , et al .
2006 , , 370 , 198 cox , t. j. , jonsson , p. , primack , j. r. , & somerville , r. s. 2006 , , 373 , 1013 darg , d. w. , et al .
2010 , , 401 , 1552 dasyra , k. m. , yan , l. , helou , g. , surace , j. , sajina , a. , & colbert , j. 2008 , , 680 , 232 de propris , r. , liske , j. , driver , s. p. , allen , p. d. , & cross , n. j. g. 2005 , , 130 , 1516 de propris , r. , conselice , c. j. , liske , j. , driver , s. p. , patton , d. r. , graham , a. w. , & allen , p. d. 2007 , , 666 , 212 deng , x .- f .
, he , j .- z . ,
jiang , p. , song , j. , & tang , x .- x .
2008 , , 677 , 1040 di matteo , p. , combes , f. , melchior , a .-
, & semelin , b. 2007 , , 468 , 61 di matteo , p. , bournaud , f. , martig , m. , combes , f. , melchior , a .-
l . , & semelin , b. 2008 , , 492 , 31 ellison , s. l. , patton , d. r. , simard , l. , & mcconnachie , a. w. 2008 , , 135 , 1877 ellison , s. l. , simard , l. , cowan , n. b. , baldry , i. k. , patton , d. r. , & mcconnachie , a. w. 2009 , , 396 , 1257 ellison , s. l. , patton , d. r. , simard , l. , mcconnachie , a. w. , baldry , i. k. , & mendel , j. t. 2010 , , 407 , 1514 faber , s. m. , et al .
2007 , , 665 , 265 gallazzi , a. , et al .
2009 , , 690 , 1883 geller , m. j. , kenyon , s. j. , barton , e. j. , jarrett , t. h. , & kewley , l. j. 2006 , , 132 , 2243 goto , t. , yagi , m. , & yamauchi , c. 2008 , , 391 , 700 green , p. j. , myers , a. d. , barkhouse , w. a. , mulchaey , j. s. , bennert , v. n. , cox , t. j. , & aldcroft , t. l. 2010 , , 710 , 1578 habergham , s. m. , anderson , j. p. , & james , p. a. 2010 , , 717 , 342 heiderman , a. , et al .
2009 , , 705 , 1433 holmberg , e. 1958 , meddelanden fran lunds astronomiska observatorium serie ii , 136 , 1 hopkins , p. f. , lidz , a. , hernquist , l. , coil , a. l. , myers , a. d. , cox , t. j. , & spergel , d. n. 2007 , , 662 , 110 hou , l. g. , wu , x .- b .
, & han , j. l. 2009 , , 704 , 789 inami , h. , et al . 2010 , , 140 , 63 kannappan , s. j. , guie , j. m. , & baker , a. j. 2009 , , 138 , 579 kauffmann , g. , et al .
2003 , , 341 , 33 kaviraj , s. 2010 , , 406 , 382 kewley , l. j. , geller , m. j. , & barton , e. j. 2006 , , 131 , 2004 knapen , j. h. , & james , p. a. 2009 , , 698 , 1437 lambas , d. g. , tissera , p. b. , alonso , m. s. , & coldwell , g. 2003 , , 346 , 1189 larson , r. b. , & tinsley , b. m. 1978 , , 219 , 46 leitherer , c. , et al . 1999 , , 123 , 3 li , c. , kauffmann , g. , heckman , t. m. , jing , y. p. , & white , s. d. m. 2008 , , 385 , 1903 lin , l. , et al . 2008 , , 681 , 232 lin , l. , et al .
2010 , , 718 , 1158 lpez - sanjuan , c. , balcells , m. , prez - gonzlez , p. g. , barro , g. , garca - dab , c. e. , gallego , j. , & zamorano , j. 2010 , , 710 , 1170 lotz , j. m. , jonsson , p. , cox , t. j. , & primack , j. r. 2010 , , 404 , 590 madau , p. , ferguson , h. c. , dickinson , m. e. , giavalisco , m. , steidel , c. c. , & fruchter , a. 1996 , , 283 , 1388 martin , d. c. , et al .
2007 , , 173 , 342 masjedi , m. , et al .
2006 , , 644 , 54 masters , k. l. , et al .
2010 , , 404 , 792 mihos , j. c. , & hernquist , l. 1994 , , 425 , l13 nikolic , b. , cullen , h. , & alexander , p. 2004
, , 355 , 874 nolan , l. a. , raychaudhury , s. , & kabn , a. 2007 , , 375 , 381 oesch , p. a. , et al .
2010 , , 714 , l47 owers , m. s. , blake , c. , couch , w. j. , pracy , m. b. , & bekki , k. 2007 , , 381 , 494 park , c. , & choi , y .- y .
2009 , , 691 , 1828 patton , d. r. , pritchet , c. j. , yee , h. k. c. , ellingson , e. , & carlberg , r. g. 1997 , , 475 , 29 patton , d. r. , carlberg , r. g. , marzke , r. o. , pritchet , c. j. , da costa , l. n. & pellegrini , p. s. 2000 , , 536 , 153 patton , d. r. , et al .
2002 , , 565 , 208 patton , d. r. , grant , j. k. , simard , l. , pritchet , c. j. , carlberg , r. g. , & borne , k. d. 2005 , , 130 , 2043 patton , d. r. , & atfield , j. e. 2008 , , 685 , 235 perez , m. j. , tissera , p. b. , scannapieco , c. , lambas , d. g. , & de rossi , m. e. 2006 , , 459 , 361 perez , j. , tissera , p. , & blaizot , j. 2009a , , 397 , 748 perez , j. , tissera , p. , padilla , n. , alonso , m. s. , & lambas , d. g. 2009b , , 399 , 1157 pracy , m. b. , kuntschner , h. , couch , w. j. , blake , c. , bekki , k. , & briggs , f. 2009 , , 396 , 1349 robaina , a. r. , et al .
2009 , , 704 , 324 rogers , b. , ferreras , i. , kaviraj , s. , pasquali , a. , & sarzi , m. 2009 , , 399 , 2172 rossa , j. , laine , s. , van der marel , r. p. , mihos , j. c. , hibbard , j. e. , bker , t. , & zabludoff , a. i. 2007 , , 134 , 2124 ruhland , c. , bell , e. f. , huler , b. , taylor , e. n. , barden , m. , & mcintosh , d. h. 2009 , , 695 , 1058 rupke , d. s. n. , kewley , l. j. , & barnes , j. e. 2010 , , 710 , l156 salim , s. , et al .
2007 , , 173 , 267 schiminovich , d. , et al .
2007 , , 173 , 315 shi , y. , rieke , g. , lotz , j. , & perez - gonzalez , p. g. 2009 , , 697 , 1764 simard , l. , et al .
2002 , , 142 , 1 simard , l. , patton , d. r. , ellison , s. l. , j. trevor mendel , & mcconnachie , a. w. 2010 , , submitted skelton , r. e. , bell , e. f. , & somerville , r. s. 2009 , , 699 , l9 sol alonso , m. , lambas , d. g. , tissera , p. , & coldwell , g. 2006 , , 367 , 1029 sol alonso , m. , michel - dansac , l. , & lambas , d. g. 2010 , , 514 , a57 strauss , m. a. , et al .
2002 , , 124 , 1810 tacconi , l. j. , et al .
2008 , , 680 , 246 tal , t. , van dokkum , p. g. , nelan , j. , & bezanson , r. 2009 , , 138 , 1417 treister , e. , natarajan , p. , sanders , d. b. , urry , c. m. , schawinski , k. , & kartaltepe , j. 2010 , science , 328 , 600 wang , j. l. , xia , x. y. , mao , s. , cao , c. , wu , h. , & deng , z. g. 2006 , , 649 , 722 west , a. a. , garcia - appadoo , d. a. , dalcanton , j. j. , disney , m. j. , rockosi , c. m. , & ivezi , .
2009 , , 138 , 796 whitaker , k. e. , & van dokkum , p. g. 2008 , , 676 , l105 wild , v. , walcher , c. j. , johansson , p. h. , tresse , l. , charlot , s. , pollo , a. , le fvre , o. , & de ravel , l. 2009 , , 395 , 144 woods , d. f. , geller , m. j. , & barton , e. j. 2006 , , 132 , 197 woods , d. f. , & geller , m. j. 2007 , , 134 , 527 woods , d. f. , geller , m. j. , kurtz , m. j. , westra , e. , fabricant , d. g. , & dellantonio , i. 2010 , , 139 , 1857 yamauchi , c. , yagi , m. , & goto , t. 2008 , , 390 , 383 rodrguez zaurn , j. , tadhunter , c. n. , & gonzlez delgado , r. m. 2010 , , 403 , 1317 zhang , h .- x . , gao , y. , & kong , x. 2010 , , 401 , 1839 | we have assembled a large , high quality catalogue of galaxy colours from the sloan digital sky survey data release 7 , and have identified 21,347 galaxies in pairs spanning a range of projected separations ( @xmath0 ) , relative velocities ( @xmath1 , which includes projected pairs that are essential for quality control ) , and stellar mass ratios ( from 1:10 to 10:1 ) .
we find that the red fraction of galaxies in pairs is higher than that of a control sample matched in stellar mass and redshift , and demonstrate that this difference is likely due to the fact that galaxy pairs reside in higher density environments than non - paired galaxies .
we detect clear signs of interaction - induced star formation within the blue galaxies in pairs , as evidenced by a higher fraction of extremely blue galaxies , along with blueward offsets between the colours of paired versus control galaxies .
these signs are strongest in close pairs ( @xmath2 and @xmath3 ) , diminish for more widely separated pairs ( @xmath4 and @xmath3 ) and disappear for close projected pairs ( @xmath2 and @xmath5 ) .
these effects are also stronger in central ( fibre ) colours than in global colours , and are found primarily in low- to medium - density environments .
conversely , no such trends are seen in red galaxies , apart from a small reddening at small separations which may result from residual errors with photometry in crowded fields .
when interpreted in conjunction with a simple model of induced starbursts , these results are consistent with a scenario in which close peri - centre passages trigger induced star formation in the centres of galaxies which are sufficiently gas rich , after which time the galaxies gradually redden as they separate and their starbursts age .
galaxies : evolution , galaxies : interactions , galaxies : photometry |
ieee 802.11a / b / g / n based wireless local area networks ( lans ) in `` infrastructure mode '' are very common in many places . in this paper
, we are concerned with an analytical model for evaluating the performance of tcp - controlled downloads in a wlan .
`` tcp '' is the transmission control protocol , which is regarded as the workhorse of the internet ; numerous applications , including web browsing , file transfer , and secure e - commerce , rely on tcp as the transport protocol .
the system we consider is shown in figure [ fig : ap_sta ] .
a detailed analysis of the aggregate throughput of tcp - controlled file downloads for a `` single rate '' access point ( ap ) is given in kuriakose et al .
@xcite ; in this , all stas are assumed to be associated with the ap at the same rate . in practice , many data rates are possible and hence considering multiple rates is important .
the aggregate download throughput is evaluated for the _ two rates _ case in krusheel and kuri @xcite . in this paper
, we consider an arbitrary but finite number @xmath4 of possible rates of association between stations ( sta ) and a single ap .
we are motivated to study an analytical model because of the improved understanding that it leads to , and the useful insights that it can provide .
closed - form expressions or numerical calculation procedures are helpful because other features and capabilities can be built upon them .
one possible application , which we are studying now , is to utilize the results reported here in devising a better ap - sta association policy .
our approach is to model the number of stas with tcp acknowledgements ( acks ) in their medium access control ( mac ) queues as an embedded discrete time markov chain ( dtmc ) , embedded at the instants of successful transmission events .
we consider a successful transmission from the ap as a reward .
this leads to viewing the aggregate tcp throughput in the framework of renewal reward theory given in kumar @xcite .
almost the entire existing literature considers a _ single _ rate of association only .
this is rather limiting , because in practice , it is extremely likely that a wlan will have stas associated at a number of rates allowed by the technology ( for example , one of 6 mbps , 12 mbps , 18 mbps , 24 mbps , 30 mbps , 36 mbps , 48 mbps or 54 mbps in 802.11 g , and one of 1 mbps , 2 mbps , 5.5 mbps or 11 mbps in 802.11b ) . a first step towards considering multiple association rates
was taken in krusheel and kuri @xcite , but there , only 2 possible association rates were considered . in this paper , _
any _ number of association rates is allowed . because of this , our model is applicable to any variant of wlan technology , for example : 802.11a / b / g / n .
the contributions of this paper are as follows .
we present a model for analyzing the performance of tcp - controlled file transfers with @xmath5 rates of association .
this generalizes earlier work .
secondly , our model incorporates tcp - specific aspects like `` delayed acks ; '' this is a technique to reduce the frequency of tcp ack generation by a tcp receiver . in most implementations ,
a tcp receiver generates a tcp ack for every @xmath6 tcp packet received ; our model is general , and considers that one tcp ack is generated for every @xmath3 tcp packets .
our analytical results are in excellent agreement with simulations , with the discrepancy being less than @xmath7% in all cases .
the paper is organized as follows : in section [ sec : related_work ] , related works are discussed . in section
[ sec : system_model ] , we state the assumptions in first part and then present our analysis . in section [ sec : evaluation ] , we present performance evaluation results . in section [ sec : discussion ]
we discuss the results .
finally , the paper is concluded in section [ sec : conclusion ] .
the literature on throughput modelling in a wlan can be classified into several groups depending on the approach . in the first group , all wlan entities ( stas and the ap )
are assumed to be _ saturated _ , _ i.e. _ , each entity is backlogged permanently .
bianchi @xcite , kumar et al .
@xcite and cali et al .
@xcite consider this saturated traffic model .
however , our interest is in modelling aggregate _
throughput , and the saturated traffic model does not capture the situation well . to see why unsaturated traffic makes a difference , we consider figure [ fig : unsaturated ] .
the left part shows a saturated traffic scenario , where all wlan entities have packets to transmit ; therefore , @xmath8 entities contend for the channel .
the right part shows the situation with tcp in the picture .
essentially , for many tcp connections , the entire window of packets sits at the ap , leaving the corresponding stas with nothing to send .
this means that the number of contending wlan entities is much smaller as mentioned in kuriakose et al .
this indicates why approaches relying on a model with saturated nodes are inadequate .
the second group considers tcp traffic .
kuriakose et al .
@xcite propose a model for tcp - controlled file downloads in a _ single rate _ wlan ; _ i.e. , _ one in which all stas are associated with a single ap at the _ same _ rate .
bruno et al .
@xcite , @xcite @xcite , @xcite , @xcite , and vendictis et al .
@xcite generalize this and analyze tcp - controlled file uploads as well as downloads ; however , it is assumed again that all stas are associated at the same rate .
similarly , yu et al .
@xcite provide an analysis for a given number of stas and a maximum tcp receive window size by using the well - known @xmath9-persistent model proposed in cali et al .
as noted above , these papers analyze tcp - controlled file transfers ( in some cases udp traffic is allowed as well ) but limit themselves to a single rate of association . in the third category ,
bharadwaj et al .
@xcite consider _ finite _ ap buffers , in contrast to the previous two , where ap buffers were assumed to be infinite ; however , the single rate assumption is retained .
the three groups mentioned above focus on _ long _ file transfers , where the tcp sender is assumed to have a file that is infinite in size .
miorandi et al .
@xcite model a different situation motivated by web browsing over a wlan . a queuing model
is proposed to compute the mean session delay for short - lived tcp flows .
the impact of maximum tcp congestion window size on this delay is studied as well . even though a fair amount of work modelling tcp - controlled transfers has been done , we are unaware of any work that allows _ multiple _ ap - sta association rates . clearly , this is the situation observed most often in practice , where the distance between the ap and a sta governs the rate of association . in this paper , we consider an arbitrary ( but finite ) number of rates of association between stas and the ap ; to the best of our knowledge , this is the first paper to consider this general model .
we consider @xmath10 stations associated with an ap as shown in figure [ fig : ap_sta ] .
all the nodes contend for the channel using the dcf mechanism as given in ieee 802.11a / b / g / n .
the stations are associated with the ap at @xmath4 different physical rates ( @xmath11 stas at rate @xmath0 , @xmath12 stas at rate @xmath1 , @xmath13 @xmath14 stas at rate @xmath2 ) .
we assume that there are no link errors .
this is not merely a simplifying assumption ; the `` auto rate fallback '' mechanism , implemented widely in stas and aps , is intended to ensure that we have an _ error - free but lower rate _
channel rather than a higher rate but error - prone channel .
thus , our assumption of no link errors is consistent with the usual mode of wlan operation .
packets in the medium are lost only due to collisions .
each station has a single tcp connection to download long files from the server and all tcp connections have equal window sizes .
the ap transmits tcp packets for the stations and the stations return tcp - ack packets .
further , we assume that the ap uses the rts - cts mechanism while sending packets to stations and stations use basic access to send ack packets ( rts : request to send , cts : clear to send ; these are control packets that reserve the wireless medium for the subsequent long data packet ) . in ieee
802.11 wlans , the parameter determines whether the rts - cts exchange will precede a packet transmission . in most operational wlans ,
the rts threshold is set such that tcp data packets are larger than the rts threshold , and hence sent after rts - cts exchange , while tcp ack packets are smaller than the rts threshold , and hence sent without rts - cts exchange ( the latter is referred to as `` basic access '' ) . upon reception of @xmath3 data packets
, a sta generates an ack packet and it is enqueued at the mac layer for transmission .
we assume that all nodes have sufficiently large buffers , so that packets are not lost due to buffer overflow .
also , tcp timeouts do not occur .
tcp start - up transients are ignored by considering all connections to be in congestion avoidance . for long file transfers ( which we are considering in this paper ) , this is a reasonable assumption because the initial start - up phase of a tcp connection lasts for a time that is completely negligible compared to the connection lifetime ; the tcp connection moves quickly to the congestion avoidance phase and remains there .
the value of rtt is very small , since files are downloaded from a server located on the lan as shown in figure [ fig : ap_sta ] .
thus , several tcp connections exist simultaneously and all stas with tcp ack packets , and the ap ( which is full of tcp data packets for the stas ) , contend for the channel .
since no preference is given to the ap , and it has to serve all stas , the ap becomes a bottleneck , and it is modelled as being backlogged permanently .
the aggregate throughput of the ap is shared equally among all @xmath10 stations .
let @xmath15 be the number of stations associated with the ap at the physical transmission rate @xmath16 , where @xmath17 @xmath18 with @xmath19 .
given that the ap wins the channel , the conditional probability that it sends a tcp data packet to a station at rate @xmath20 is @xmath21 .
we assume that @xmath22 is large .
our results will show that @xmath23 or @xmath24 , @xmath25 ( with @xmath26 ) suffices for the analysis to be applicable .
figure [ fig : channel_activity ] shows a possible sample path of the events on the wlan channel .
the random epochs @xmath27 indicate the end of the @xmath28 successful transmission from either the ap or one of the stations .
we observe that most stas have empty mac queues , because , in order for many stas to have tcp - ack packets , the ap must have had a long run of successes and this is unlikely because no special preference is given to the ap .
so when the ap succeeds in transmitting , the packet is likely to be for a sta with an empty mac queue . at epoch
@xmath29 , let @xmath30 be the number of stations at rate @xmath20 , ready with an ack .
let @xmath31 be the number of nonempty stas .
if there are @xmath32 nonempty stas and a nonempty ap , each nonempty wlan entity attempts to transmit with probability @xmath33 as in kumar et al .
@xcite , where @xmath34 is the attempt probability with @xmath8 _ saturated _ entities .
it can be seen that @xmath35 evolves as a discrete time markov chain ( dtmc ) over the epochs @xmath27 .
this allows us to consider @xmath36 as a markov renewal sequence , and @xmath37 as a semi - markov process .
stations associated with the ap at @xmath38 different data rates . ]
we have a multidimensional dtmc which is shown in figure [ fig : markovchain ] ; transition probabilities are indicated as well .
a sta generates a tcp ack after receiving @xmath3 tcp packets ; this is incorporated in our model in the following way . when the dtmc state is @xmath39 , there are @xmath40 backlogged wlan entities ( including the ap ) ; so , the probability that a station at rate @xmath16 wins the channel is @xmath41 , since each entity is equally likely to win the contention .
further , given that a sta at rate @xmath16 wins the channel , the conditional probability that it will generate a tcp - ack is @xmath42 . by inspection
, we can say that the dtmc is irreducible ; further , the detailed balanced equation holds for a properly chosen set of equilibrium probabilities . the detailed balance equation ( dbe )
is @xmath43 here @xmath44 , @xmath45 is the stationary distribution of the dtmc . from the set of equations given in and @xmath46 the stationary distribution is @xmath47 to obtain the throughput
, we use markov regenerative analysis , culminating in the renewal reward theorem given in wolff @xcite , kumar @xcite . for a given state @xmath48 , successive entries into state @xmath49 form renewal epochs . to obtain the mean time between successive entries ( the mean renewal cycle lengths )
, we obtain the mean sojourn time in the state @xmath48 .
let @xmath50 be the sojourn time in a state @xmath51 .
conditioning on various events ( idle slot , collision or successful transmission ) that can happen in the next time slot , the following expression for the mean cycle length can be written down : + @xmath52 in the above expression , @xmath53 is the probability of the slot being idle , @xmath54 is the probability that the ap wins the contention and transmits the data packet at rate @xmath55 and @xmath56 is the probability that a sta associated at rate @xmath16 wins the channel ( `` s '' in the suffix stands for `` success '' ) .
correspondingly , the conditional expected sojourn times in state @xmath57 , given the events are , respectively , @xmath58 , @xmath59 and @xmath60 .
detailed expressions for these quantities are provided in the appendix .
the @xmath61 and @xmath62 terms on the right side of correspond to collision events .
the third term in arises when the ap transmits a tcp data packet to a station at rate @xmath20 and some other stations are involved in a collision ; in other words , the third term captures the situations in which the ap is involved in a collision .
the fifth term in captures collision events in which the ap is _ not _ involved ; we have a sta transmitting a tcp ack packet to the ap at rate @xmath20 and one or more other stas transmitting simultaneously .
the various probabilities have been obtained by using the attempt probability @xmath63 , when there are @xmath64 contending nodes . from equation
we have @xmath65 @xmath66 we are interested in finding the long run time average of successful transmissions from the ap .
we obtain this by applying the renewal reward theorem of wolf @xcite . to get the mean renewal cycle length
, we can use the mean sojourn time given in equation and use theorem 5.3 in @xcite .
the mean reward in a cycle can be obtained as follows .
a reward of 1 is earned when the ap transmits a tcp data packet successfully by winning the channel .
the probability of the ap winning the channel is @xmath67 .
similarly , a reward of 0 is earned with probability @xmath68 .
therefore , the expected reward is @xmath67 . putting all this together
, the aggregate tcp throughput can be calculated as @xcite + @xmath69
to verify the accuracy of the model , we performed experiments using the qualnet 4.5 network simulator @xcite .
we considered 802.11b physical data rates : 1mbps , 2mbps , 5.5mbps and 11mbps ; higher rates correspond to smaller distances between the stas and the ap . in table
[ table : aps_stas ] , results are given for a few cases of this multirate scenario .
for example , the first row considers a total of 10 stas associated with the ap , out of which 2 stas are associated at 11 mbps and 2 mbps respectively , while 3 stas are associated at 5.5 mbps and 1 mbps , respectively .
the values of @xmath70 in equation are calculated by using the number of stas associated with the ap .
for example , in the first row in table [ table : aps_stas ] , @xmath71 ( for 11 mbps ) is @xmath72 . | we consider several wlan stations associated at rates @xmath0 , @xmath1 , ... , @xmath2 with an access point .
each station ( sta ) is downloading a long file from a local server , located on the lan to which the access point ( ap ) is attached , using tcp .
we assume that a tcp ack will be produced after the reception of @xmath3 packets at an sta .
we model these simultaneous tcp - controlled transfers using a semi - markov process .
our analytical approach leads to a procedure to compute aggregate download , as well as per - sta throughputs , numerically , and the results match simulations very well .
wlan , association , access points , infrastructure mode . |
researchers currently studying active galactic nuclei agree on a few key premises .
the central engine is powered by a supermassive black hole , and the optical through ultraviolet continuum emission is generated by optically thick material in an accretion disk within the central parsec down to a few gravitational radii .
important specifics such as the thickness of the disk as a function of radius are still being hashed out , but the overall picture is consistent ; it is hard to avoid disk formation given the persistence of angular momentum . moving to larger radii , ionized gas moving at high velocities covers approximately @xmath0% of the sky and gives rise to broad ( 1000s km s@xmath1 ) resonance and semi - forbidden emission lines that characterize type 1 quasars and seyfert galaxies . in many luminous quasars ,
the broad uv emission lines show signatures of winds .
the most obvious are the p cygni profiles seen in the high ionization resonance lines of broad absorption line ( bal ) quasars ( e.g. , @xcite ) .
more subtle are the blue - shifts and asymmetries seen in the same lines in other luminous quasars ( e.g. , @xcite ) .
the wind manifest in these lines is understood to be driven by radiation pressure on ions at uv resonance transitions . for
the wind to be launched in the vicinity of the broad - line region , the quasar continuum must not only have significant power in the uv the source of the line - driving photons but not too much power in the x - ray .
x - rays will strip atoms such as c and o of their electrons , thus dropping the absorption cross section dramatically .
overionized gas can therefore only be driven by thomson scattering , which is much less efficient than line driving ( e.g. , @xcite ) . the elegant picture of murray et al .
@xcite that unifies the broad emission lines seen in all quasars with the bal wind directly observed in only @xmath2% ( e.g. , @xcite ) by generating broad emission lines in an equatorial disk wind is compelling , but has not been widely adopted in other models .
broad - line `` clouds '' are still evoked as the source of the broad - line emission , though the smoothness of the broad - line profiles and the requirement of some means of pressure - confinement to prevent the putative clouds from shredding remain persistently unsolved problems in the cloud picture ( e.g. , @xcite ) .
one structure that is even more uncertain is the so - called torus .
the inner radius of the torus is set by the dust sublimation radius , beyond which some fraction of the accretion disk continuum is absorbed by grains and reradiated .
operationally , the torus must account for the @xmath3 of the integrated radiant quasar power that comes out in the near - to - mid - infrared .
furthermore , it must obscure the central continuum and broad - line region in a significant fraction of type 2 objects . while a static toroidal structure ( whether smooth or clumpy ) beyond the dust sublimation radius will serve these purposes , such a cold ( 1001500 k ) structure with the dust mass implied by the mid - infrared luminosity would quickly become gravitationally unstable and collapse .
this is an old problem , which was addressed by knigl & kartje @xcite with a magneto - hydrodynamically launched wind , shaped also by radiation pressure , on dust .
this picture was expanded by our group ( @xcite ; hereafter k12 ) by incorporating dust - radiation driving into the magneto - hydrodynamic ( mhd ) wind models of everett @xcite , and then illuminating the wind with the central continuum to generate a model spectral energy distribution ( sed ) . while the details of the shape of the output continuum still require some refining to match those observed , the overall power and general shape generated by our ` fiducial ' model is promising .
this model uses as inputs the empirical sdss composite quasar continuum @xcite , a black hole mass of @xmath4 @xmath5 , an eddington ratio of @xmath6 , and a column density at the base of the wind of @xmath7 @xmath8 .
we assume a milky way interstellar - medium dust distribution . in this article , we expand on the results presented in k12 by exploring how key input model properties affect the structure of the wind .
following blandford & payne @xcite and knigl & kartje @xcite , we model the torus as a self - similar dusty wind driven by mhd forces and radiation pressure .
the mhd radiative wind code has been described in more detail previously @xcite , and we give only a brief outline here .
first , a purely mhd - driven wind solution is calculated given a set of initial conditions input by the user .
next , cloudy ( version 06.02.09b , as described by @xcite ) is called to calculate the radiation field and dust opacity in the wind , allowing for the model to determine the radiative acceleration of the wind due to the central accretion disk continuum .
a new mhd solution is calculated for the wind , now taking into account the radiative acceleration , and the process is iterated until the wind converges to a stable profile .
operationally , our assumption of self - similarity means that several important quantities , including magnetic field strength and mass density , scale with spherical radius ; this simplifies the calculation of the solution to the mhd equations considerably .
furthermore , it allows for the incorporation of radiation pressure from the disk by simply reducing the effective gravitational potential .
importantly , it allows us to use one streamline as an approximate template for the entire wind , simply scaling appropriately inwards or outwards to find properties of the specific region of the wind under consideration .
the final , converged mhd radiative wind model therefore provides the wind structure necessary to calculate the mass density of the wind as a function of _ r _ , the radial distance from the axis of rotation of the disk , and _ z _ , the vertical distance from the plane of the disk .
the model supplies the density along each streamline as a function of the mass density at the base of the streamline and of polar angle .
the density of the wind at any arbitrary point is then relatively easy to calculate as a consequence of the self - similarity assumption . because we know the shape of the wind streamlines , we can use the shape of a single streamline to trace back from an arbitrary point to the disk - launching radius for any wind streamline .
from there , we use our knowledge of the number density of particles on the disk at the inner launch radius and the fact that it is assumed to be proportional to @xmath9 @xcite to find the number density at the launch radius of that particular streamline .
we convert from mass density to number density by assuming that the average particle has the mass of a proton , thus allowing us to find the number density at our specified point .
in figure [ fig1 ] we present a color map plot of the number density of the dusty wind as a function of _ r _ and _ z _ for the fiducial model .
the inner wind launch radius is set by the dust sublimation radius .
the notable thinness of the wind is a consequence of the model parameters : in general , the outer wind launch radius is given by : @xmath10 for user - specified values of the inner launch radius @xmath11 , the number density at the inner launch radius @xmath12 , and the column density along the base of the wind @xmath13 .
a simulated high - resolution image of the wind at 9.5 @xmath14 m heated by the central continuum is shown in the right panel of fig .
the hourglass shape is characteristic of mhd winds , and reminiscent of the outflows seen in young stars .
it illustrates clearly that the base of the wind and the inner streamlines dominate the emission at this wavelength ; the brightest 9.5 @xmath14 m emission regions would correspond to the largest extinction values in the optical .
eight to 13 @xmath14 m interferometric observations have led hnig et al .
@xcite to infer a similar structure and characteristic size scale ( @xmath15 pc ) for the dust distribution in the seyfert 2 galaxy ngc 424 .
_ left : _ a 2d cross - section of the upper right quadrant of the number density map of the fiducial dusty wind model with @xmath16 @xmath8 , @xmath17 , and @xmath18 .
the color indicates number density according to the scale bar on the right .
the number density drops sharply as a function of height , @xmath19 , and the wind is quite narrow .
dotted straight lines indicate the values of the optical depth @xmath20 at h@xmath21 , as labeled . _
right : _ a simulated high - resolution 9.5 @xmath14 m image of the same model generated by mc3d @xcite .
the bulk of the emission comes from the inner streamlines heated by the accretion disk continuum .
, title="fig:",width=321 ] as luminous quasars are typically farther away and therefore fainter and of smaller angular size compared to nearby seyferts , they are beyond the reach of the current generation of interferometers .
nonetheless , within the context of our model we can explore what might be expected by investigating the effects of changing luminosity , @xmath22 , and @xmath23 on the structure of the wind .
figure [ fig2 ] illustrates the shape of the fiducial wind model ( labeled as a ) compared to two other models .
model b has a black hole mass of @xmath24 , ten times that of the fiducial model .
because the value of @xmath22 is fixed at 0.1 , this model is 10 times more luminous than the fiducial model .
the resultant shape of the wind is nearly identical to that of the fiducial model ; the streamlines are slightly more vertical .
the third model , c , is also similar to the fiducial model , this time changing only the inner wind launch radius to approximately one third of the fiducial model s inner radius .
the change in the inner radius was made to approximate the difference between the launch radius of graphites ( with a higher sublimation temperature ) versus silicates .
this decrease in the inner launch radius causes the wind to be notably broader in our model ( see fig .
[ fig3 ] ) , as expected from equation [ eq : thickness ] .
because the bulk of the near - infrared emission is coming from the innermost streamlines , it therefore follows that a smaller @xmath25 for a given @xmath22 generates more near - ir emission because of the higher dust temperatures @xcite .
the strength of the 35 @xmath14 m bump the manifestation of a larger contribution to the near - to - mid - infrared sed from hot dust is seen empirically to increase with increasing quasar luminosity ( e.g. , @xcite ) .
within the dusty wind paradigm , this implies some mechanism for decreasing the wind launching radius as a function of luminosity which is challenging to arrange . a second way of reducing @xmath25 is to reduce @xmath22 while keeping the black hole mass constant .
this is also consistent with the expected movement of the dust sublimation boundary to smaller radii . the slight change in wind structure from altering @xmath23 but keeping @xmath22 constant compared with the large change from decreasing @xmath22 by a factor of 10 shows that the eddington ratio is the relevant driver given our assumed wind structure .
however , while the effects on the wind structure are noticeable when compared to the fiducial model , this physical difference did not lead to a significant change in the shape of the output sed @xcite . a cross section of the wind , as in figure [ fig1 ] , comparing the shape of several wind models .
shown here is the fiducial model ( labelled `` a '' ) , with @xmath26 , @xmath6 , and @xmath27 ; a similar model with @xmath28 ( `` b '' ) ; and a third model with the same parameters as the fiducial model , but with a smaller inner launch radius ( `` c '' ) .
increasing the black hole mass by a factor of 10 does not significantly change the shape of the wind , but reducing the inner launch radius from @xmath29 cm to @xmath30 cm while maintaining the same density at the inner launch radius causes the wind to widen dramatically , a consequence of our assumptions .
the movement to smaller radii in model `` b '' also results in a strong near - infrared bump to appear in the output sed @xcite .
, title="fig:",width=377 ] the column densities of 30 elements at each ionization state as well as those of several important molecules are calculated by cloudy . because they are part of a wind , these column densities are smooth functions of inclination angle , monotonically increasing toward the plane of the disk , where the polar angle @xmath31 .
therefore , they offer no obvious inclination angle to choose as the boundary for our wind .
therefore , a plot of column density versus height does not show any obvious features that define a ` height ' for the wind . also available from cloudy
is the cross section per proton for the dust in the wind as a function of photon energy .
when multiplied by the total column density of hydrogen ( i.e. , both neutral and ionized ) , this serves as a measure of the optical depth of the wind .
figure [ fig3 ] shows the optical depth as a function of polar angle @xmath32 for a few selected wavelengths .
this optical depth is also a smooth function of @xmath32 for a wide range of wavelengths , and similarly provides no obvious angle to mark the boundary of the wind .
we have therefore chosen the angle at which the dust reaches an optical depth of 5 for light at a wavelength of 6563 ( the wavelength of @xmath33 emission ) to be a reasonable definition of the opening angle of the dusty wind . at this value of @xmath20 ,
only 0.7% of the broad @xmath33 emission is transmitted through the wind , and so a quasar viewed from this inclination angle ( or closer to the disk ) would be classified as an optical narrow - line object .
the optical depth at three characteristic wavelengths ( 1 @xmath14 m , 6563 , and 1000 ) through the dusty wind plotted as a function of inclination angle .
shown here is the fiducial model , with @xmath34 , @xmath6 , and @xmath35 .
the grey dashed lines indicate @xmath20 values of 1 , 2 , and 5 as labeled .
the intersection of the curve corresponding to 6563 and the line corresponding to @xmath36 determines the opening angle ( marked with a vertical black line ) .
the red dashed curve just above the optical depth curve at 6563 is the @xmath37-band extinction ; @xmath38 has a value of 7.7 for @xmath39.,title="fig:",width=491 ] typical opening angles using this criterion are around 70@xmath40 , which would lead to type 2 to type 1 ratios of @xmath41% .
this is on the low end of the range of estimates ( e.g. , @xcite ) .
however , if accretion disks are randomly oriented with respect to their host galaxies as expected , the observed type 2 fractions are inflated with respect to the torus - covering fraction because of line - of - sight obscuration occurring from dust within the host galaxy .
the modeled range of opening angles is tight , and shows little dependence on model input parameters with the exception of the column density at the base of the wind , and the inner wind launch radius .
this can be seen in figure [ fig4 ] , where there is a clear trend relating the opening angle and the base column density with relatively little spread among the suite of models from k12 .
the opening angle , @xmath42 , as a function of base column density , @xmath43 . at small base column densities ,
the opening angle is implausibly large . for @xmath16 @xmath8 ,
the values drop to @xmath44 .
this is also the value of @xmath43 that generates appropriate near - to - mid - infrared power in the sed @xcite .
the black crosses denote the opening angle as defined in the text : the angle at which the wind reaches an optical depth of 5 to the broad h@xmath45 emission line .
the grey circles show the angle at which the same models reach an optical depth of 2 to broad h@xmath45 emission .
the models are those presented in k12 .
, title="fig:",width=491 ]
getting the dusty - wind model to the point where it can plausibly account for the detailed structure in near - to - mid - infrared quasar seds requires additional refinement .
in particular , we have not taken into account the expected dust - free component of the wind that is interior to the dust sublimation radius . the strong absorption within the wind as seen in the x - ray and uv spectra of bal quasars
will significantly modify the continuum impinging upon the dusty wind .
specifically , we expect the dust sublimation radius to be pushed to smaller values .
however , the specific effects of the broad - line region gas shielding the dusty wind from the continuum will depend on its column density and ionization state ; this is the subject of ongoing investigations .
more generally , the viability of the dusty mhd disk wind paradigm requires an investigation into whether relatively strong ( mg to g ) , large - scale poloidal magnetic fields persist within the central parsecs of quasar host galaxies .
certainly quasars with jets have magnetic fields ; radio synchrotron emission is observed .
galaxies also show organized magnetic fields on large scales , but they are much weaker , of order 10 @xmath14 g ( e.g. , @xcite ) . within the dense cores of molecular clouds in our galaxy , magnetic fields with strengths up to a few mg have been measured ( e.g. , @xcite ) . the physical scale and @xmath46 number density of such cores ( 0.0220 pc and @xmath47 @xmath48 ) span that expected in tori . in the near future , the available sensitivity and spatial resolution of _ alma _ and the jansky vla are promising for placing meaningful constraints on the strengths of organized fields within the central few parsecs of radio - quiet active galactic nuclei .
this work was supported by the natural science and engineering research council of canada and the ontario early researcher award program . | mass ejection in the form of winds or jets appears to be as fundamental to quasar activity as accretion , and can be directly observed in many objects with broadened and blue - shifted uv absorption features . a convincing argument for radiation pressure driving this ionized outflow
can be made within the dust sublimation radius . beyond , radiation pressure is even more important , as high energy photons from the central engine
can now push on dust grains .
this physics underlies the dusty - wind model for the putative obscuring torus .
specifically , the dusty wind in our model is first launched from the outer accretion disk as a magneto - centrifugal wind and then accelerated and shaped by radiation pressure from the central continuum .
such a wind can plausibly account for both the necessary obscuring medium to explain the ratio of broad - to - narrow - line objects and the mid - infrared emission commonly seen in quasar spectral energy distributions .
a convincing demonstration that large - scale , organized magnetic fields are present in radio - quiet active galactic nuclei is now required to bolster the case for this paradigm . |
recent transport experiments @xcite have established that the resistivity of a two - dimensional electron system with weak disorder and valence orbital landau level filling @xmath4 close to @xmath5 is anisotropic when the valence orbital landau level index @xmath6 .
apparently the ground state spontaneously breaks orientational symmetry , a property believed to be associated with the uni - directional charge - density - wave ground states predicted @xcite in this regime by hartree - fock mean - field theory @xcite .
the charge - density - wave ( cdw ) state consists of electron stripes of width @xmath7 with local landau level filling factor @xmath8 , separated by hole stripes of width @xmath9 with local filling factor @xmath10 .
here @xmath11 , the cdw period , is comparable to the landau level s cyclotron orbit diameter . because of symmetry properties shared with the smectic state of classical liquid crystals , emphasized by fradkin and kivelson @xcite , these states
have been referred to as quantum hall smectics , a practice we follow here .
the most important transport property of quantum hall smectics is that dissipation occurs over a wide range of filling factors surrounding @xmath12 and is not activated at low temperatures .
this behavior is _ not _ consistent with the properties of a cdw state , which would be pinned by the random disorder potential and have a large gap for mobile quasiparticle excitations , and suggests that , although hartree - fock theory hints at the energetic motivation for a ground state with broken orientational symmetry , the description which it provides of the ground state is flawed .
several recent theoretical papers @xcite have addressed the properties of quantum hall smectics and the energetic competition between cdw states , compressible composite - fermion states , and paired incompressible quantum hall states .
in one recent paper @xcite we have described a theory of quantum hall smectics which starts from the hartree - fock theory ground state , recognizes the electron stripes as coupled one - dimensional electron systems , and treats residual interaction and disorder terms neglected by hartree - fock theory using the convenient bosonization techniques of one - dimensional electron physics .
the most important conclusions of this work are the following : i ) the quantum hall smectic state is _ never _ the ground state but instead is always unstable , for @xmath4 close to @xmath5 likely to an anisotropic electron wigner crystal state ; ii ) for @xmath13 the interaction terms responsible for the wigner crystal can be neglected at temperatures available in a dilution fridge ; and iii ) weak disorder which scatters electrons from stripe to stripe , enabling hard - direction transport , leads to non - linear transport . in this article
we emphasize and expand on an experimentally important prediction of this work , namely that the strength of the transport non - linearity is sensitive to the nature of the electron - electron interaction .
in particular we predict that the transport non - linearity can be enhanced by placing a screening plane close to the two - dimensional electron system .
in section ii we explain our theory of transport in quantum hall smectics and discuss how the coefficients which govern the power - law behavior of the differential resistivity for weak disorder are related to correlations in the coupled one - dimensional electron stripes . in section iii we briefly review our theory of these correlations and explain why long - range electron - electron interactions weaken transport non - linearities . in section
iv we present numerical results for the non - linear transport coefficients for a model in which interactions in the two - dimensional electron layer are screened by a metallic layer co - planar with the electron system but separated from it by a distance @xmath3 .
we conclude in section v with a brief summary .
our transport theory @xcite is built on a semiclassical boltzmann - like approach ; microscopic physics enters only in the calculation of scattering rates . we choose a coordinate system where the @xmath14 ( horizontal ) direction runs along the stripes which are separated in the @xmath15 ( vertical ) direction .
we assume that the charge density wave itself is pinned and immobilized by both the edges of the sample and weak impurities which couple to the electrons within the stripes . in this case ,
collective sliding motion of the charge - density will be absent , and electrical transport will be dominated by single - electron scattering across and between electron stripes .
an important property of electronic states in the quantum hall regime , is the spatial separation of states which carry current in opposite directions . in the case of the electron stripes in quantum hall smectics ,
the fermi edge states which carry oppositely directed currents along the stripes ( left - going and right - going ) are located on opposite sides of the stripe .
translational invariance along the @xmath14 direction , allows us to use a landau gauge where this component of wavevector @xmath16 is a good quantum number in the absence of interactions and disorder .
the single - particle states at the stripe edges have velocity magnitude @xmath17 , the fermi velocity . in the landau gauge , @xmath14 direction momenta
are related to @xmath15 direction positions by @xmath18 where @xmath19 is the magnetic length , so that states on opposite sides of the same stripe differ in momenta by a @xmath20 and the adjacent sides of neighboring stripes differ in momenta by @xmath21 .
we now summarize the basic assumptions on which our semiclassical transport theory is based and quote the expressions implied by these assumptions .
for further details see ref .
we assume that in the steady state , each edge of each stripe is characterized by a local chemical potential .
translational invariance in the @xmath15 direction implies that the chemical potential drops across each stripe and between any two adjacent stripes are the same ,
mf @xmath22 is the chemical potential drop across an electron stripe , it follows that the potential drop between stripes is @xmath23 , where @xmath24 is the hard - direction electric field , the field which supports a steady state transport across the stripes .
the electric field in the @xmath14 direction produces a semiclassical drift in momentum space which drives the system from equilibrium .
we assume that disorder scattering across and between stripes , then attempts to reestablish equilibrium and that the drift and scattering processes are in balance in the steady state .
the scattering currents in the hard - direction are characterized by relaxation times @xmath25 and @xmath26 respectively .
the current along a stripe is analogous to the quantized current in a long narrow hall bar and is proportional to the chemical potential difference across that stripe .
combining these ingredients leads @xcite to the following expressions for the resistivities : @xmath27 where @xmath28 , @xmath29 , and @xmath30 .
for @xmath12 , this theory makes a parameter free prediction for the product @xmath31 which has been confirmed experimentally @xcite .
in fact , as emphasized @xcite by van oppen _ et al .
_ , this feature of our results has a greater validity than would be suggested by our assumption of largely intact electron stripes .
our main interest here however , is in expanding on our predictions @xcite for non - linearities in the easy and hard direction differential resistivities .
these predictions were made on the basis of a simple lowest order renormalization group scheme for handling the infrared divergence which appear when disorder terms which scatter electrons either across or between stripes are treated perturbatively .
this analysis leads to @xmath32 where @xmath33 and @xmath34 are golden - rule scattering rates at the characteristic microscopic energy scale @xmath35 , here @xmath36 is @xcite the scaling dimension of the operator which scatters an electron across a stripe , which we discuss in the next section , and @xmath37 is the scaling dimension of the operator which scatters an electron between neighboring stripes .
the values of @xmath36 and @xmath37 depend on correlations induced by electron - electron interaction between stripes , and are sensitive in particular to the range of the microscopic electron - electron interaction . at @xmath38 , the case on which we will concentrate , @xmath39 . given these expressions
, it follows @xcite that the non - linear differential resistivity in the hard direction @xmath40 with an exponent @xmath41 .
similar considerations apply for the easy direction current : @xmath42 with an exponent @xmath43 . in the next section
we show that both @xmath44 and @xmath45 increase when the distance to the screening plane is comparable to or smaller than the cdw period .
the cdw state of hartree - fock theory @xcite is a single - slater - determinant . in the valence landau level
, groups of landau - gauge single - particle states with adjacent @xmath46 ( adjacent @xmath15 ) are occupied to form stripes and separated by groups which are unoccupied .
small fluctuations in the positions and shapes of the stripes can be described in terms of particle - hole excitations near the stripe edges .
the residual electron - electron interaction terms which scatter into these low energy states are ignored in hartree - fock theory and fall into two classes : forward " scattering interactions which conserve the number of electrons on each edge of every stripe , and
backward " scattering processes which do not .
the latter processes involve large momentum transfer and are unimportant @xcite at accessible temperatures for @xmath4 near @xmath5 .
the quantum smectic model @xcite , briefly described in this section includes forward scattering only .
the interactions are bilinear in the 1d electron density contributions from a particular edge of a particular stripe : @xmath47 , with @xmath48 .
since the density of a single filled landau level is @xmath49 , the displacement of an edge is related to its associated charge density contribution by @xmath50 .
the quadratic hamiltonian which describes the _ classical _ energetics for small fluctuations has the following general form : @xmath51 where @xmath52 . here
the @xmath53 integral is over the interval @xmath54 and a high momentum cutoff @xmath55 is implicit for @xmath56 .
symmetry considerations constrain the form of the elastic kernel . in position space ,
the kernel must be real and symmetric so that @xmath57 .
this implies @xmath58 and @xmath59 .
parity invariance ( under @xmath60 ) , implies moreover @xmath61 .
thus , the elastic kernel is fully specified by one real function , @xmath62 , and one complex function , @xmath63 .
it will be important for our present interest that the hamiltonian must be invariant under : @xmath64 . for short - range interactions
this implies that at long wavelengths @xmath65 characteristic of classical smectic elasticity . as we will discuss , this conclusion must be modified in the case of long - range interactions .
a _ quantum _ theory of the quantum - hall smectic @xcite is obtained by imposing kac - moody commutation relations on the chiral densities : @xmath66 = \frac{i}{2\pi } \alpha \delta_{\alpha,\beta } \delta_{n , n ' } \partial_{x } \delta(x - x').\ ] ] together with eq .
( [ hamiltonian ] ) , this relationship fully specifies the quantum dynamics .
electron operators in the chiral edge modes are related to the 1d densities via the usual bosonic phase fields : @xmath67 with @xmath68 .
quantum properties of the smectic can be computed from the imaginary - time action , @xmath69 where in an obvious matrix notation , @xmath70 correlation functions follow from wick s theorem and the momentum space correlator @xmath71 with @xmath72 the effect of weak disorder on transport in quantum hall smectics depends sensitively on the elastic constants _ at _ @xmath73 . in this
limit the relevant excited states are simply slater determinants with straight stripe edges displaced from those of the hartree - fock theory ground state . by evaluating the expectation value of the microscopic hamiltonian in a state with arbitrary stripe edge locations
we find that @xmath74 where the constant @xmath75 is such that the condition @xmath76 , and the positions @xmath77 are the ground state stripe edge locations . here , @xmath78 is the interaction potential between two electrons located in guiding center states a distance @xmath79 apart : @xmath80 @xmath81 the two terms in eq .
( [ gamma ] ) are direct and exchange contributions . in eq .
( [ matrixelement ] ) , @xmath82 is the fourier transform of the effective 2d electron interaction which incorporates form - factors @xcite dependent on the landau level index . @xmath83 and the ground subband wavefunction of the host semiconductor heterojunction or quantum well .
the smectic states have relatively long periods proportional to the cyclotron orbit radii . explicit calculations @xcite show that @xmath84 for @xmath85 .
it follows that the exchange contribution to @xmath78 is small and that @xmath78 decreases with stripe separation in the relevant range . in this paper
we address the influence of a metallic screening plane which cuts off this interaction at large distances . for @xmath86 , @xmath87 , decreasing extremely slowly with @xmath79 . for separation
@xmath79 larger than the distance @xmath3 to the screening plane , @xmath88 , making the sum over @xmath89 in eq .
[ elastic ] convergent .
the scaling dimension , @xmath36 , of the operator @xmath90 which scatters an electron across the @xmath89-th stripe is readily evaluated from eq .
( [ actionkernel ] ) .
we find that @xmath91 here , @xmath92 is the weight function , @xmath93}{[d_{++}^{2}({\bf q } ) - |d_{+-}({\bf q})|^{2 } ] ^{1/2 } } .
\label{wght}\ ] ] for 1d non - interacting electrons @xmath94 , so that disorder is relevant and eventually leads to localization . as discussed below , @xmath95 for quantum hall smectics .
disorder is even more relevant than in the non - interacting electron case . nevertheless ,
since the samples in which the quantum hall smectic is observed are of extremely high quality , there should be a wide range of temperature over which its effects can be treated perturbatively . if @xmath96 both hard direction ( @xmath44 ) and easy direction ( @xmath45 ) non - linear transport exponents vanish .
we see from eq .
( [ wght ] ) that @xmath97 if the average value of @xmath98 is one . to understand the dependence of @xmath36 on screening
, we have to understand the dependence of @xmath99 on both wavevector and @xmath3 .
note that @xmath92 is smaller than one , increasing the relevance of disorder , when @xmath100 and @xmath101 are opposite in sign and similar in magnitude . for each @xmath53 in eq .
[ integral ] the weighting factors are like those which enter in the calculation of the scaling dimension of the operator which describes backscattering from disorder in an isolated one - dimensional electron system . in continuum
1d models , @xmath100 has a contribution , proportional to the fermi velocity , from the band energy and a contribution proportional to the interaction between electrons traveling in the same direction , while @xmath102 has only an interaction contribution . for a continuum model ,
the effective interactions between electrons traveling in different directions is the same as that between electrons traveling in the same directions .
when the interaction term is much larger than the band term @xmath100 and @xmath103 are opposite in sign and nearly equal in magnitude and @xmath92 is very small .
this is what happens , for example , for a 1d electron system in which long - range makes the coulomb interaction very strong at long wavelengths . when @xmath92 is small , the 1d electron system is very close @xcite to an electron wigner crystal , and disorder is very strongly relevant . on the other hand
when @xmath103 is much smaller than @xmath100 we have a situation analogous to that in a very weakly interacting fermion system , in which disorder is relevant but the resistivity is linear when disorder can be treated perturbatively .
with this in mind we turn to a discussion of the quantum smectic , limiting our attention to the case @xmath12 .
useful insight comes from examining the value of @xmath92 at the end points of the integration interval , @xmath104 and @xmath105 where both @xmath100 and @xmath103 are real . for @xmath106 , invariance under a uniform translation of
the smectic implies that @xmath107 , so that @xmath108 .
when all the electron stripes move together , the energetics is precisely like that of a single 1d system . in the quantum hall regime
, there is no band energy , only interaction contributions from electrons traveling in the same direction , which appear in @xmath100 , and interaction contributions from electrons traveling in opposite directions , which appear in @xmath103 .
the absence of a band contribution means that , for @xmath109 , @xmath92 vanishes independent of the interaction s strength or range .
when @xmath110 , on the other hand , one has @xmath111 .\ ] ] we see that @xmath112 vanishes because the interaction between the top of one stripe and the bottom of the same stripe is identical to its interaction with the bottom of the next stripe up .
the interactions between oppositely directed electrons effectively cancel out , and we obtain @xmath113 , just as we would for a non - interacting 1d system .
thus @xmath36 is determined by the average over @xmath53 of a weighting factor which interpolates between that of a 1d electron wigner crystal at @xmath114 and that of a non - interacting 1d electron system at @xmath110 .
the average value of @xmath92 is determined by the rate at which @xmath92 goes from its @xmath106 limit to its @xmath115 limit . to obtain insight into what controls this ,
we consider first the limit of short - range interactions .
since @xmath116 vanishes due to the cancellations of its direct and exchange contributions , the short - range limit is obtained by taking only @xmath117 . in this case
, it follows from eq .
( [ elastic ] ) that @xmath118 and that @xmath119 , and therefore that @xmath120 .
the numerator of eq .
( [ wght ] ) vanishes like @xmath121 for @xmath122 , and the denominator , which is proportional to the collective mode velocity , vanishes like @xmath123 .
note that the expression for @xmath124 is independent of @xmath125 .
the average of @xmath124 may be evaluated analytically in this case , and we obtain for short - range interactions @xmath126 . for the realistic case , analytic calculations are no longer possible , but the behavior of @xmath92 can be simplified , at least at small @xmath127 if the exchange contribution to @xmath128 is neglected in constructing @xmath129 . in this case
@xmath128 is the simply the 1d transform to coordinate space of the reciprocal space interaction @xmath130 , i.e. it is the coulomb interaction between lines of charged smeared by @xmath83-dependent cyclotron orbit form factors .
the components of @xmath129 , are then fourier transforms back to reciprocal space , but with additional ` umklapp ' terms because this transform is discrete .
we find that @xmath131 and @xmath132 .
\label{dnotnoexchange}\ ] ] for the case of a coulomb interaction , @xmath133 /q $ ] where @xmath3 is the distance to a screening plane described by an image charge model . for large or infinite @xmath3 ,
the @xmath134 terms dominate the sums in eqs .
( [ noexchange ] ) .
the above expression for @xmath135 applies when @xmath136 , and therefore is always valid for the @xmath134 terms in eqs .
( [ noexchange ] ) .
note that both @xmath137 and @xmath138 are proportional to @xmath3 for large @xmath3 and diverge for @xmath139 . on
the other hand @xmath140 remains finite for @xmath139 because the @xmath134 term is excluded from this sum .
we emphasize that the numerical calculations whose results are shown below include the exchange contributions neglected in deriving eqs .
( [ noexchange ] ) , and important at any value of @xmath11 when @xmath3 is not large .
we now examine the large @xmath3 , small @xmath127 behavior of the numerator and denominator of eq .
( [ wght ] ) . for the numerator @xmath141 and for the denominator @xmath142 \ ; [ q^2 a^2 ] .
\label{longrangedenom}\ ] ] the first factor is square brackets in eq .
( [ longrangedenom ] ) comes from @xmath143 for which the @xmath134 term dominates .
the second factor is proportional to @xmath144 for which only odd @xmath145 terms survive , implying no dependence on @xmath3 and analytic dependence on @xmath127 .
these formulas apply for @xmath146 ; for @xmath139 , the small @xmath127 behavior is obtained by replacing @xmath147 by @xmath148 , _
i.e. _ , by replacing @xmath3 by @xmath149 .
we plot the square of the denominator , proportional to the square of the collective excitation velocity for several values of the screening length @xmath3 in fig .
[ [ v2 ] ] .
the velocity increases as @xmath3 increases as expected . for @xmath150 ,
the quadratic small @xmath127 behavior predicted in eq .
( [ longrangedenom ] ) , applies only for @xmath151 , and is not apparent in the plot .
instead we see the long range interaction behavior , in which the velocity is proportional to @xmath152 . as is apparent from eqs .
( [ noexchange ] ) , screening is irrelevant except very close to @xmath153 when @xmath3 is large .
[ [ v2 ] ] shows that once @xmath154 , the large @xmath3 no - screening limit is approached . in fig .
[ [ wghtn2 ] ] we show the weight functions for @xmath155 , @xmath156 and @xmath157 . at each value of @xmath3
, the denominator of the weight function at small @xmath127 is proportional to @xmath158 and the numerator proportional to @xmath159 .
the weight function is therefore proportional to @xmath123 with a coefficient which varies as @xmath160 .
a larger value of @xmath3 ( less screening ) , leads to a weight function which increases more rapidly with @xmath127 and a scaling dimension for the scattering vertex which is closer to one . in the limit of unscreened interactions , which applies down to small @xmath127 for @xmath150 , @xmath161 in fig . [ [ nltrans ] ] we show the dependence of the scaling dimension @xmath36 and the non - linear transport exponents @xmath44 and @xmath45 on the distance @xmath3 to the screening plan .
for @xmath162 the numerical result is very close to that from the analytic short - range interaction model described above which leads to @xmath163 . for @xmath162 ,
the scaling dimension approaches @xmath164 , a value we have been able to obtain only numerically .
these relatively modest changes in the scaling dimension translate into relatively large changes in the transport exponents , particularly in @xmath44 which characterizes the hard - axis non - linearity .
we predict that these non - linearities will be much stronger if a screening placed in close proximity to the electron layer .
recent experiments @xcite have established a consistent set of transport properties for high - mobility two - dimensional electron systems with high orbital index ( @xmath165 ) partially filled landau levels .
these properties differ qualitatively from those which occur in the low orbital index ( @xmath166 ) fractional quantum hall effect regime . at large @xmath83 ,
the dissipative resistivities are large , strongly anisotropic , and non - linear for @xmath167 within each landau level .
this anisotropic transport regime is bracketed by regions of reentrant integer quantum hall plateaus .
we have recently @xcite developed a theory which is able to account for most features of these experiments .
an important prediction of the theory is that the easy and hard direction resistivities should have non - integral power - law temperature dependences . in this article
we have briefly summarized the theory and elaborated on its predictions for the dependence of these exponents on the distance @xmath3 between the two - dimensional electron layer and a remote screening plane , predictions which are summarized in fig .
[ [ nltrans ] ] .
we find that @xmath36 approaches two different values , both smaller than one , for @xmath162 and @xmath139 , and interpolates smoothly between these limits at finite values of @xmath3 .
verification of our prediction that non - linear transport can be enhanced by introducing a screening plane and reducing @xmath3 , would help substantiate our theory of quantum hall smectics .
we would like to acknowledge insightful conversations with herb fertig , michel fogler , tomas jungwirth , jim eisenstein , eduardo fradkin , steve girvin , and philip phillips .
ahm acknowledges the hospitality of the itp at uc santa barbara , where this work was initiated .
m.p.a.f . is grateful to the nsf for generous support under grants dmr-97 - 04005 , dmr95 - 28578 and phy94 - 07194 .
a.h.m . is grateful for support under grant dmr-97 - 14055 .
tsui , h.l .
strmer , l.n .
pfeiffer , k.w .
baldwin , and k.w .
west , solid state commun . * 109 * , 389 ( 1999 ) ; w. pan , r.r .
strmer , d.c .
tsui , l.n .
pfeiffer , k.w .
baldwin , and k.w .
west , phys .
* 83 * , 820 ( 1999 ) .
this expectation is based on hartree - fock theory which predicts charge - density - wave ground states throughout the quantum hall regime , usually incorrectly .
see for example a.h .
macdonald and d.b .
murray , phys . rev .
b * 32 * , 2291 ( 1985 ) .
the hartree - fock predictions of broken symmetry states are more plausible in higher landau levels , however , because the associated variational energies compete favorably with variational estimates of fluid state energies . | recent transport experiments have established that two - dimensional electron systems with high - index partial landau level filling , @xmath0 , have ground states with broken orientational symmetry . in a mean - field theory ,
the broken symmetry state consists of electron stripes with local filling factor @xmath1 , separated by hole stripes with filling factor @xmath2 .
we have recently developed a theory of these states in which the electron stripes are treated as one - dimensional electron systems coupled by interactions and described by using a luttinger liquid model . among other things ,
this theory predicts non - linearities of opposite sign in easy and hard direction resistivities . in this article
we briefly review our theory , focusing on its predictions for the dependence of non - linear transport exponents on the separation @xmath3 between the two - dimensional electron system and a co - planar screening layer . |
m87 is usually classified as a fanaroff - riley class i ( fr - i ) radio galaxy having a one - sided relativistic jet @xcite implying , in the context of unification models @xcite that its unresolved nuclear region is a misaligned bl lac object .
the viewing angle is estimated to be @xmath5 giving a maximum doppler factor @xmath6^{-1}$ ] of @xmath7 where @xmath8 is the jet lorentz factor . for simplicity
, we shall take @xmath9 but consider the range @xmath10 for the m87 jet at pc scales .
although the jet emission is not strongly doppler boosted towards us it may even be de - boosted its proximity to us ( 16.3 mpc @xcite ) partially compensates for this , and makes tev gamma - ray @xcite , cosmic ray @xcite and possibly neutrino emission from this object potentially interesting .
the strong variability of the optical flux of m87 suggests that the jet emits synchrotron radiation at optical frequencies somewhere within an unresolved central region less than 5 pc in diameter @xcite .
the jet has well - defined relativistic features , and it is remarkable that despite its low power the jet extends beyond 30 arcsec from the core @xcite , possibly being a remnant from when m87 was much more active . given the large black hole mass @xmath11m@xmath12 , and the low jet luminosity @xmath13 erg s@xmath14 @xcite , the accretion disk must be currently in a low radiative state and provide little power to the jet . as a consequence ,
the heating of any torus is currently inefficient and would produce little attenuation of tev gamma - ray signals although this would change if m87 returned to an active state commensurate with its high black hole mass @xcite .
bl lac objects , along with flat - spectrum radio quasars are collectively referred to as blazars .
bl lacs may be high - frequency or low - frequency peaked ( hbls and lbls ) .
their broad - band spectra consist of two spectral components which appear as broad ` humps ' in the spectral energy distribution ( sed ) , and are due to emission from a jet oriented at small angle with respect to the line - of - sight .
the low - energy component , is generally believed to be synchrotron emission from relativistic electrons , and extends from the radio to uv or x - ray frequencies .
the origin of the high - energy component , starting at x - ray or @xmath15-ray energies and extending in some cases to tev - energies , is uncertain . for bl lacs which have rather weak thermal emission , the favoured `` leptonic model '' is the so - called synchrotron - self compton ( ssc ) model in which the same relativistic electrons responsible for the low energy synchrotron hump in the sed up - scatter synchrotron photons to high energies via the inverse compton effect .
`` hadronic models '' were proposed more than 10 years ago to explain the @xmath15-ray emission from blazars @xcite . recently mcke & protheroe @xcite have discussed in detail the various contributing emission processes in the synchrotron proton blazar ( spb ) model . in hadronic models
the relativistic jet consists of relativistic proton and electron components , which again move relativistically along the jet .
high - energy radiation is produced through photomeson production , and through proton and muon synchrotron radiation , and subsequent synchrotron - pair cascading in the highly magnetized environment . in the case of bl
lacs internal photon fields ( i.e. produced by synchrotron radiation from the co - accelerated electrons ) serve as the target for pion photoproduction .
these models can , in principle , be distinguished from leptonic models by the observation of high energy neutrinos generated in decay chains of mesons created in the photoproduction interactions ( for a recent review see @xcite ) . in hadronic models
, agn would contribute also to the pool of extragalactic ultra - high energy cosmic rays ( uhecr ) through the decay outside the host galaxy of neutrons produced by photoproduction interactions ( see @xcite in the context of accretion shocks , and @xcite in the context of jets ) or by escape of protons directly accelerated at termination shocks of jets in giant radio lobes of fanaroff - riley class ii ( fr - ii ) radio galaxies @xcite .
greisen @xcite , and zatsepin and kuzmin @xcite ( gzk ) showed that the nucleonic component of uhecr above @xmath16 ev will be severely attenuated in the cosmic microwave background radiation ( cmbr ) , primarily due to pion photoproduction interactions . at @xmath17
ev the mean free path is @xmath18 mpc and the energy - loss distance is @xmath19 mpc @xcite ( see stanev et al .
@xcite for recent calculations ) .
thus , if the uhecr are extragalactic , one would expect their spectrum to cut off at @xmath20 ev , the `` gzk cut - off '' .
uhecr may have been observed with energies well above @xmath16 ev ( see nagano and watson @xcite for a review including a discussion of models for the origin of the uhecr ) .
however , very recent data from the two largest aperture high energy cosmic ray detectors are contraditory : agasa @xcite observes no gzk cut - off while hires @xcite observes a cut - off consistent with the expected gzk cut - off .
a systematic over - estimation of energy of about 25% by agasa or under - estimation of energy of about 25% by hires could account the discrepancy @xcite , and the continuation of the uhecr spectrum to energies well above @xmath16 ev is now far from certain .
nevertheless , if the spectrum does extend well beyond @xmath16 ev , even though it is only an fr - i radio galaxy m87 is an attractive possibility for the origin of the uhecr @xcite because of its proximity .
mcke et al .
@xcite identified the critical parameters determining the properties of bl lacs in the context of the spb model and constructed the `` average '' synchrotron spectrum for each class , hbls and lbls , which served as the target photon distribution for their hadronic cascade .
an extensive collection of blazar seds published by ghisellini et al .
@xcite were used to construct the `` average '' sed of hbls and lbls by overlaying all available hbl and lbl seds .
they found that a broken power law gave a reasonable representation of the low - frequency hump in the sed of both hbls and lbls @xmath21 with @xmath22 , @xmath23 and @xmath24ev .
the break energy @xmath25 of lbls varied between @xmath26 0.1 ev to 1.3 ev , while the maximum synchrotron photon energy @xmath27 ranged over two orders of magnitude , from @xmath26 41 ev to 4 kev .
the populated energy range of hbls is more restricted : @xmath28 26 ev to 131 ev and @xmath29 4.1 kev to 41 kev .
the peak of the low - energy sed was @xmath30 for lbls and @xmath31 for hbls .
mcke et al .
@xcite defined the `` average lbl '' by @xmath321.3 ev , @xmath334.1 kev and @xmath34 , and the `` average hbl '' by @xmath32131 ev , @xmath3341 kev and @xmath35 .
the parametrizations for hbls and lbls are visualized in fig .
[ fig1 ] , with the dashed lines showing the `` average '' sed and the hatched regions showing the range of the seds for the two cases .
assuming that the unresolved nuclear emission in m87 , i.e. from the pc scale jet , has @xmath9 we have added to fig .
[ fig1 ] the radio @xcite , infrared @xcite and x - ray @xcite data on the nuclear region of m87 doppler boosted by @xmath36 to mimic how m87 would appear if its jet were closely aligned towards us such that it would be classed as a bl lac rather than a fr - i .
since the doppler factor of m87 is not well known , taking the range from 0.66 to 1.5 , the observed sed as viewed along the axis with @xmath36 would move along the error bars added to the data . from fig .
[ fig1 ] , the rather high radio - infrared luminosity would suggest that m87 could be a mis - aligned lbl , in which case the x - ray emission would be part of the high - energy hump in the sed due to cascading initiated by photons and electrons from pion decay or by synchrotron photons emitted by protons and muons . however , the x - ray emission is at a level closer to that expected from an hbl , in which case it would be mainly due to synchrotron emission by directly accelerated electrons .
one should bear in mind that there is considerable variation in the seds of hbls and lbls , and so one should not expect the sed of an individual bl lac to be identical to our `` average '' spectrum .
this is particularly true for the m87 nuclear region where the observations at radio , infrared and x - rays were not simultaneous and are known to vary .
in addition , wilson and yang @xcite have calculated from the chandra x - ray flux and spectrum observed from the inner part of the jet that there is intrinsic absorption by cold matter with an equivalent hydrogen column density of @xmath37@xmath38 @xmath39 in m87 , and so the unattenuated x - ray flux could be higher than plotted .
thus we shall be satisfied if the observed sed is within a factor @xmath40 of an `` average '' sed . in fig .
[ fig2 ] we show the seds of hbls modelled by mcke et al .
@xcite in the context of the spb model as they would appear if observed at @xmath9 as fr - i radio galaxies at a distance of 16 mpc appropriate to m87 .
results taken from mcke et al .
@xcite for four peak luminosities of the low energy hump of the sed are shown after shifting to @xmath9 . as we see , there is little variation in the high energy hump which peaks at @xmath41@xmath42 gev at a level just below the sensitivities of egret and the whipple high energy gamma ray telescope , but well above the sensitivity of glast and future northern hemisphere large - area atmospheric cherenkov telescopes ( act ) such as magic and veritas ( sensitivities for all these telescopes taken from ref .
@xcite are shown together with the upper limit from whipple @xcite ) .
one should bear in mind , however , that the uncertainty in the doppler factor of m87 will give rise to a much larger uncertainty in the bolometric luminosity .
these uncertainties are indicated in the theoretical sed and neutrino spectra by the slanted error - bars .
we have extended the work of mcke et al .
@xcite by calculating the neutrino output for these four cases and plot the expected flux of muon - neutrinos ( @xmath43+@xmath44 ) .
the highest luminosity modelled for the low - energy hump of the sed ( solid curve ) is in closest agreement with the m87 sed and predicts the highest neutrino flux .
we have added an upper - limit for the amanda - b10 neutrino telescope by taking the diffuse neutrino background limit @xcite and multiplying by @xmath45 sr ( neutrino telescopes detect upward - going neutrinos ) .
sensitivities calculated by albuquerque et al .
@xcite of amanda - ii and icecube to @xmath46 diffuse intensity , again multiplied by @xmath45 sr to convert to point source sensitivity have also been added . unlike the 100 gev gamma - ray emission which should easily be detected by future telescopes ,
the neutrino flux is well below detection levels of future neutrino telescopes . in fig .
[ fig3 ] we model m87 as a mis - aligned lbl in the context of the spb model . in this case
, we see that the gamma - ray emission cuts off at 10 gev1 tev depending on the luminosity of the low energy hump of the sed .
this is because for the higher target photon densities in lbl , pion photoproduction losses determine the maximum proton energy , and hence the maximum gamma - ray energy which results from the cascades initiated by proton synchrotron radiation and pion decay ( including pion and muon synchrotron radiation ) .
this also affects the maximum neutrino energy in the same way .
the lowest luminosity modelled for the low - energy hump of the sed ( chain curve ) is in closest agreement with the m87 sed and predicts the highest gamma - ray and neutrino energies .
the gamma ray flux is a factor @xmath4710 lower than for the hbl case , but is still detectable by future telescopes . as with the hbl case
, the neutrino flux is below the sensitivity of icecube .
however , it is not impossible that in the case of rapid flaring , m87 might just be detected as it is at a declination of @xmath48@xmath49 and is ideally located for observation by a giant neutrino telescope to be located at the south pole such as icecube because it is sufficiently below the horizon to eliminate cosmic ray events , while not being so far below the horizon that neutrino absorption by the earth s core becomes important .
energetic protons magnetically trapped in agn jets lose energy predominantly by bethe - heitler pair production and pion photoproduction on ambient radiation fields , or by adiabatic deceleration as the jet expands .
neutron production in pion photoproduction sources ( e.g. @xmath50 ) provides a mechanism for escape of cosmic rays from the jet .
neutrons decay typically after travelling @xmath51ev ) mpc , which for uhecr is well outside the host galaxy .
recently ahn et al .
@xcite and biermann et al .
@xcite have shown that in the assumption of a parker - spiral galactic wind magnetic structure out to @xmath52 mpc , the galactic wind of our galaxy poses no restriction to the entry to our galaxy of uhecr from the general direction of m87 , and may even lead to clustering in the arrival directions of cosmic rays from m87 observed from earth .
these preferred arrival directions and clustering may , however , be mainly due to the galactic wind topology and contain little information about the direction of cosmic ray sources @xcite .
the observed intensity of uhecr @xcite ( multipled by energy squared ) is plotted in fig .
[ fig4 ] . for the most promising lbl model ( chain curve in fig . [ fig3 ] )
we plot ( chain curve ) the flux of neutrons divided by @xmath53 sr ( to convert flux to average intensity ) that would be observed at earth from m87 if the neutrons did not decay ( which of course they do ) . since these neutrons would travel to earth in straight lines , and
since we take m87 to have @xmath54 , the maximum neutron energy that would arrive at earth would be approximately the maximum jet - frame energy of the accelerated protons , which would be @xmath55 ev for the model shown .
however , neutrons decay into protons whose directions are isotropised in the intergalactic magnetic field before and during propagation to earth .
hence we would have contributions to the protons arriving at earth from neutrons emitted at all angles with respect to the jet axis , and hence from the full range of doppler factors corresponding to the jet lorentz factor , i.e. approximately @xmath56 .
for example , if the jet lorentz factor were @xmath57 a neutron with jet - frame energy @xmath58 ev emitted along the jet axis ( @xmath59 ) would decay to a proton with galaxy - frame energy @xmath60 ev .
what we would observe therefore depends on the galaxy - frame angle - averaged neutron luminosity _ on emission _ , and would also have contributions from both jets ( i.e. twice that for one jet ) .
given that the chain curve in fig .
[ fig4 ] is for @xmath61 , it is related directly to the jet - frame bolometric luminosity @xmath62 . for electromagnetic radiation and
neutrinos the _ observed _ bolometric luminosity is @xmath63 and the angle - averaged _ emitted _ bolometric luminosity is @xmath64 .
hence , by analogy , for ultra - relativistic protons of energy @xmath65 from neutron decay , and remembering that both jets contribute , @xmath66 we have added to fig . [ fig4 ] ( solid curve ) the resulting cosmic ray flux divided by @xmath53 sr assuming @xmath57 and straight - line propagation , i.e. @xmath67 corresponding to number density @xmath68 at @xmath69@xmath70 ev this is a factor @xmath71 below the observed uhecr . of course , so far this neglects effects of diffusive propagation to earth and interactions with the cmbr .
we shall consider first the effects of diffusion .
for the case of a constant injection rate of cosmic rays , and an infinite homogeneous diffusing medium with diffusion coefficient @xmath72 the number density at distance @xmath73 is simply @xmath74 if we take @xmath75 then the observed intensity will be enhanced with respect to straight - line propagation by a factor @xmath76 for a simple model of intergalactic space with a low magnetic field ( @xmath77 g ) and a cell size @xmath78 kpc , taking @xmath79 would give an enhancement factor up to @xmath80 , such that the predicted uhecr from m87 could explain the observed intensity above @xmath81 ev .
while diffusion appears to increase the cosmic ray intensity from m87 , it also increases the travel time or effective distance travelled from m87 to earth , @xmath82 hence for an enhancement factor of @xmath83 the effective distance is @xmath84 mpc . at @xmath81
ev the mean energy - loss distance is approximately 1 gpc and at @xmath16 ev the mean energy - loss distance is approximately 160 mpc , and so in this case there would only be a significant energy loss above @xmath16 ev
. the situation could actually be quite different if we consider an intergalactic magnetic field ( igmf ) structure based on the observation of microgauss fields in clusters of galaxies @xcite , and of clusters occurring in networks of `` walls '' separated by `` voids '' .
it would then be reasonable to expect relatively high fields of @xmath85@xmath86 g in the walls , and much lower fields , @xmath87@xmath88 g , in the voids .
recently , stanev et al .
@xcite have considered propagation of uhecr in three different models of the igmf in the local supergalactic structure in which the regular field in the high - field region was @xmath89 g. we consider propagation in a simple wall / void model similar to that used by medina tanco @xcite to illustrate how uhecr would be deflected in complex igmf structures . as m87 is at the centre of the virgo cluster and is close to the super - galactic plane we assume it is embedded in the higher field region . to illustrate the effect of propagation from m87 in a wall / void type igmf we define the origin of coordinates to be at m87 with the mid - plane of the wall corresponding to the @xmath90@xmath91 plane , the wall occupying @xmath92 mpc and the void occupying @xmath93 mpc .
following medina tanco @xcite we adopt a regular magnetic field of @xmath85 g in the @xmath90-direction in the wall and @xmath94 g in the void , with a transition region sandwiched between the wall and the void in which the magnetic field drops exponentially from @xmath85 g to @xmath94 g. the magnetic field contains an irregular component having @xmath95 equal to 30% of the regular component , and has a kolmogorov spectrum of turbulence .
the irregular field is modelled as described in ref .
@xcite with wavenumbers corresponding to turbulence scales @xmath96 with @xmath97 mpc and @xmath98 .
particles are injected isotropically at the origin and energy losses due to bethe - heitler pair production and pion photoproduction are included . a typical simulation for @xmath99 ev is shown in fig . [ fig5 ] .
whereas medina tanco @xcite was most interested in the angular deflection of arriving cosmic rays relative to the direction of the cosmic ray source , we are more interested in the travel times of cosmic rays to earth from the distance @xmath100 mpc to m87 , and the total amount of time spent within unit volume at distance @xmath73 from m87 .
we divide the spherical shell corresponding to 15 mpc @xmath101 mpc into 2160 equal cells of volume @xmath102 with a grid in spherical coordinates @xmath103 using 72 equally spaced @xmath104-values and 30 equally spaced @xmath105-values .
particle orbits are advanced in time steps of size @xmath106 mpc/@xmath107 chosen to be much smaller than the smallest magnetic structure simulated . at each step
, the proton energy is reduced according to the average energy - loss rate ( see e.g. figure 1 of ref .
a check is made at the end of each time step to determine whether the proton is within the spherical shell , and if it is then @xmath108 is added to the total time spent in the cell corresponding to its spherical coordinates @xmath103 i.e. if on @xmath109 occasions a particle was found to be located inside the @xmath110th cell at the end of a time step , then it would have spent approximately a total time @xmath111 inside the @xmath110th cell since being injected at the origin . dividing by the number of simulations one obtains the average time spent in a particular cell per injected proton @xmath112 .
similarly , the time spent travelling since injection at m87 is added to the ages of the protons `` detected '' in that cell .
dividing by the total number of occasions a particle was `` detected '' in that cell gives the average age @xmath113 of the protons present in that cell .
results based on the average time per unit volume in @xmath114 simulations are shown for two initial energies @xmath81 ev and @xmath16 ev in fig .
the gray - scale shows the enhancement factor @xmath115 the strong peaks at @xmath116 and @xmath117 show that despite the presence of the turbulent field component , cross - field diffusion is not strong enough to give rise to significant fluxes far away from the regular field threading the source , at least for source distances as small as 16 mpc for this field topology , and we note that stanev et al .
@xcite arrived at a similar conclusion about cross - field diffusion .
the peaks in @xmath118 shown in fig . [ fig6 ]
are at a level @xmath119 over a disk of radius @xmath120 ( @xmath121 mpc ) outside of which @xmath122 , indicating that if the magnetic field topology is such that the field lines connect the vicinity of m87 ( within @xmath123 mpc ) to our galaxy ( within @xmath123 mpc ) then cosmic rays from m87 should clearly be observed .
the average ages are @xmath124 mpc / c ( @xmath125ev ) and @xmath126
mpc / c ( @xmath99ev ) indicating their final energies would be @xmath127ev and @xmath128ev respectively .
we find the unresolved nuclear core of m87 to be consistent with it being a mis - aligned bl lac object . in the context of the synchrotron proton blazar model
, m87 could be either an hbl or an lbl . in both cases , we predict gamma - ray emission at levels detectable by glast and magic , and possibly by veritas .
neutrino detection is not expected , except possibly during an extreme flare in the lbl case .
ultra - high energy neutrons produced in pion photoproduction interactions will escape from the host galaxy where they decay into protons . because cosmic rays are deflected in the igmf
, protons resulting from neutrons emitted in all directions can contribute to the observed uhecr .
hence , even though the electromagnetic radiation we observe from the m87 jet is not significantly doppler boosted in energy , the cosmic ray output will be . by this mechanism
, the unresolved nuclear core of the m87 jet could emit uhecr with energies up to at least @xmath129 ev .
we predict the uhecr output from m87 to be at a level such that if uhecrs travelled in straight lines they would give an average intensity at earth a factor @xmath13020 below that observed . for
a constant uhecr output and simple isotropic diffusion with a scattering mean free path @xmath131 less than the distance @xmath73 to m87 , the predicted intensity would be enhanced by a factor @xmath132 relative to straight line propagation .
this enhancement is accompanied by increased travel times , and so higher energy losses in the cmbr .
nevertheless , taking reasonable scattering mean free paths we find that for simple diffusion models protons from decay of ultra - high energy neutrons produced by pion photoproduction in the m87 jet could easily account for all the observed uhecr . in perhaps more realistic wall / void models of igmf structure ,
m87 would only be a source of the observed uhecr if the topology of the igmf between m87 and our galaxy is favourable .
note , however , that because of its very high black hole mass m87 was probably much more active at earlier times than at present . many objects exhibit a high state for @xmath1305% of the time , and since we estimate the average travel time of uhecr from m87 to earth to be a factor @xmath130510 times the light propagation time it is indeed possible that the uhecr observed now were emitted when m87 was in a high state and possibly a more powerful cosmic ray source .
similarly , if m87 was in the recent past a powerful fr - ii radio galaxy , it could have been possible to accelerate uhecr at shocks in the giant radio lobes where the jets terminate as in the model of rachen and biermann @xcite , and these could also contribute to the observed uhecr .
we thank todor stanev for very helpful comments . this work was supported by a discovery project grant from the australian research council and a grant from the university of adelaide to rjp .
ar thanks the bundesministerium fr bildung und forschung for financial support through desy grant verbundforschung 05ch1pca/6 .
p l. biermannn , e .- j .
ahn e .- j .
, , p. p. kronberg , g. medina - tanco , t. stanev , in physics and astrophysics of ultra - high - energy cosmic rays , edited by m. lemoine and g. sigl , lecture notes in physics , vol .
576 , p. 181 .
`` glast : exploring nature s highest energy processes with the gamma ray large area space telescope '' nasa document np-2000 - 9 - 107-gsfc , february 2001 , p29 . | the unresolved nuclear region of m87 emits strong non - thermal emission from radio to x - rays . assuming this emission to originate in the pc scale jet aligned at @xmath0 to the line of sight
, we interpret this emission in the context of the synchrotron proton blazar ( spb ) model .
we find the observed nuclear jet emission to be consistent with m87 being a mis - aligned bl lac object and predict gamma - ray emission extending up to at least 100 gev at a level easily detectable by glast and magic , and possibly by veritas depending on whether it is high - frequency or low - frequency peaked . predicted neutrino emission is below the sensitivity of existing and planned neutrino telescopes .
ultra - high energy neutrons produced in pion photoproduction interactions decay into protons after escaping from the host galaxy . because energetic protons are deflected by the intergalactic magnetic field , the protons from the decay of neutrons emitted in all directions , including along the jet axis where the doppler factor and hence emitted neutron energies are higher , can contribute to the observed ultra - high energy cosmic rays .
we consider the propagation of these cosmic ray protons to earth and conclude that m87 could account for the observed flux if the extragalactic magnetic field topology were favourable .
astroparticle phys .
, submitted 11 october 2002 revised 19 november 2002 tev gamma rays and cosmic rays from the nucleus of m87 , + a mis - aligned bl lac object + r.j .
protheroe@xmath1 , a .- c .
donea@xmath2 , a. reimer@xmath3 + @xmath2department of physics and mathematical physics , the university of adelaide , adelaide , sa 5005 , australia + @xmath3ruhr - universitt bochum , institut fr theoretische physik , lehrstuhl iv : weltraum- und astrophysik , d-44780 bochum , germany + pacs : + 98.54.cm active and peculiar galaxies ( including bl lacertae objects , blazars , seyfert galaxies , markarian galaxies , and active galactic nuclei ) + 98.54.gr radio galaxies + 98.58.fd jets , outflows and bipolar flows + 98.70.rz gamma - ray sources + 98.70.sa cosmic rays ( including sources , origin , acceleration , and interactions ) + 13.85.tp cosmic - ray interactions + @xmath4email : [email protected] |
lithium is readily destroyed within a main sequence star and may survive only in a very thin surface layer provided that mixing between this layer and the interior does not occur .
surviving surface li is greatly diluted as a star evolves to become a red giant because an extensive convective envelope develops .
standard stellar evolutionary models predict li dilution by a factor of about 60 for low - mass solar - metallicity stars @xcite .
thus , a star of approximately solar metallicity with an initial ( say , meteoritic ) li abundance of @xmath1(li ) @xmath5 3.2 is predicted to have a li abundance of 1.5 on the red giant branch ( rgb ) . in general ,
observations of li abundances in k giants confirm these predictions in the sense that the predicted li abundance is the upper limit to observed li abundances except for extremely rare cases exhibiting a higher li abundance .
for many rgb stars , li is less than the predicted @xcite , and this is likely due to a combination of enhanced destruction of li by mixing within a giant s main sequence progenitor and continued destruction of li in a giant .
this comfortable picture of destruction and dilution was demonstrated to be incomplete by the serendipitous discovery of an unusually large amount of li in a k giant @xcite . since
then a few more li - rich k giants of solar metallicity have been discovered including examples with li abundances at or even exceeding the expected maximum li abundance of main sequence stars . the first systematic search for li - rich k giants @xcite demonstrated the rarity of li - rich giants , just about 1@xmath6 , in the solar neighborhood . in this _ letter _ , we present results from a large survey of k giants designed to find additional li - rich examples , locate their appearance in the hr diagram and so provide essential material for testing hypotheses about li - enrichment .
for example , li production may be associated with the bump in the luminosity function for stars on the rgb as suggested from observations by @xcite .
depending on mass and composition the bump occures on rgb between luminosities , log ( l / l@xmath7 ) = 1.5 - 2.2 , and this has been predicted @xcite as a region where evolution of a giant is slowed down creating a bump in the luminosity function of stars along the rgb .
internally , the bump is associated with the h - burning shell crossing over of h - discontinuity .
additionally or alternatively , the he - core flash may be the trigger for li production and the creation of li - rich red clump ( horizontal branch ) stars .
discrimination between bump and clump is possible because the two groups appear in somewhat different places in the hr diagram .
our 2000 giants provide a reference sample comfortably spanning the luminosity of the clump and the bump .
also , the survey designed to test the suggestion of planet engulfment scenario by including giants from all the way from the base of rgb to well above the clump / bump luminosities . with brown et al.s survey , a total of 2644 giants covering a wide range of luminosities and effective temperatures represent a potent database in which to search for clues to the origin of the li enrichment in k giants .
this _ letter _ is restricted to discussing the location of the li - rich stars in the hertzsprung - russell diagram .
full details of the survey will be provided in a subsequent publication . 0.1 cm
k giants with accurate astrometry ( parallaxes and proper motions ) were chosen from the hipparcos catalogue @xcite to span the luminosity bump on the rgb .
stars were selected according to their ( b - v ) color such that stars were confined to the mass range of 0.8 - 3.0@xmath8 .
filters were applied to select nearby ( d @xmath9 150 pc ) bright ( m@xmath10 @xmath9 8) stars within the declination range of @xmath1160 @xmath12 @xmath13 @xmath9 80 deg .
these criteria resulted in a total of 2000 k giants .
all of which were observed at low spectral resolution ( r=@xmath14/@xmath15 @xmath9 3500 ) at one of three telescopes : the 2-m himalayan chandra telescope ( hct ) with the faint object spectrograph camera ( hfosc ) at the indian astronomical observatory ( iao ) , hanle ; the 1-m zeiss telescope fitted with the universal astro - grating spectrograph ( uags ) , and the 2.34-m vainu bappu telescope ( vbt ) fitted with the optometrics medium resolution spectrograph ( omrs ) at the vainu bappu observatory ( vbo ) , kavalur .
sample low resolution spectra showing li and ca features is shown in fig . 1 .
the k giants with a moderate to strong li doublet at 6707.7 were identified and pursued at high resolution ( r=60,000 ) spectroscopy with the 2.34-m vbt equipped with the echelle spectrograph @xcite and/or with the 2.7-m harlan j. smith telescope with robert g. tull coud echelle spectrometer @xcite at the mcdonald observatory .
the raw data were processed under standard tasks in the @xmath16 package .
the parameters luminosity and effective temperature are obtained in identical fashion for the stars in our and brown et al.s survey .
luminosities are obtained using parallaxes from the hipparcos catalogue and the magnitudes ( m@xmath10 ) taken from the simbad data base .
effective temperatures ( @xmath17 ) are derived using the ( b - v ) color and the calibration given in @xcite .
since sample stars are nearby , the effects of interstellar extinction are not taken into account .
however , we note that by neglecting reddening we would have underestimated @xmath17 by about 50 - 100k considering canonical reddening of about e(b - v ) @xmath18 0.03 - 0.05 for giants of distances of 100 to 200 pc . neglecting reddening will not affect the @xmath17 of li - rich k giants because these temperatures were derived using iron lines in the high resolution spectra .
the temperatures derived from high resolution spectra are in good agreement ( within @xmath19100k ) with the values derived for the same stars using the ( b - v ) color and the @xcite calibration .
lithium abundances of the stars observed at low resolution were estimated from the strength of the lii 6707.7 doublet relative to that of the cai line at 6717.6 .
an empirical relation between the line - depth ratio of li and ca lines and the li abundance was derived from known li - rich k giants with li abundances from high - resolution spectra .
fifteen k giants in our survey were estimated to be li - rich on the basis of the low resolution spectra .
accurate abundances for these stars were then obtained from high resolution spectroscopic analysis ( table 1 ) .
previously known li - rich giants of which several were reobserved and reanalysed are listed in table 2 .
our survey confirms the earlier survey @xcite that the li - rich k giants are rare .
tables 1 and 2 give the atmospheric parameters ( @xmath17 , @xmath20 , and [ fe / h ] ) derived from iron lines in the high - resolution spectra .
a majority of the entries for table 2 are taken from the papers reporting the li abundances .
the lte and non - lte li abundances are tabulated where the latter are based on the recipe given in @xcite . in the tables
, we also present @xmath21c/@xmath22c values for the li - rich k giants .
the values of @xmath21c/@xmath22c measurements are based on molecular lines in the 8003 region using a similar procedure to that described in @xcite . for a few giants carbon ratios could nt be determined owing to their large rotation and as a result broad spectral features .
giants from our and brown et al.s surveys are shown on the hr diagram in fig . 2 along with evolutionary tracks @xcite computed for solar metallicity , [ fe / h ] = 0.0 , and stellar masses ranging from 0.8@xmath8 to 3@xmath8 .
all stars observed only at low resolution are assumed to have a solar metallicity .
for li - rich k giants actual metallicities measured from the high resolution spectra are adopted and most are close to solar values .
errors derived from the quoted uncertainties in the parallaxes and in the derivation of temperatures for the li - rich k giants are marked in fig .
all but a very few stars in the two samples have evolved beyond the base of the rgb ( the red dashed line in fig . 2 ) and , therefore , their surface li abundance has been greatly diluted . in panel
a of fig . 2 , there is the expected concentration of stars with luminosities of about @xmath23 .
this population in the hr diagram in the main are red clump ( horizontal branch ) he - core burning stars that have evolved from the tip of the rgb : the thick black line in fig .
2 shows the predicted locations of clump stars .
extension of the @xmath24 concentration to cooler temperatures may include stars at the bump luminosity of the first ascent rgb stars ( see red portions of the evolutionary tracks ) .
dominance of the hr diagram by clump stars is expected because the lifetime of a clump star is typically one to two orders of magnitude longer than for a bump star .
this argument is not strictly applicable to the li - rich stars because their frequency is also dependent on the probabilities that clump and bump stars produce lithium .
except for five luminous examples , the li - rich stars are either at the luminosity of the clump or within the range of expected values for the bump , a range spanning the clump s luminosity .
the five luminous exceptions must owe their li richness to different phenomena than those accounting for the majority of the li - rich stars .
for example , hd 39853 @xcite in table 2 is a li - rich k giant but with log l / l@xmath7=2.7 is probably a post - clump early agb candidate where , as discussed by @xcite , li - production by the cameron - fowler mechanism has occurred by a variant of its operation in bump stars ( see below ) . in contrast , the high luminosity warm ( @xmath25k ) li - rich star hd 21018 is , as suggested by @xcite , caught prior to dilution of lithium by the convective envelope , i.e. , it is a li - normal star exhibiting a li abundance close to its in initial abundance .
we focus here on the li - rich giants at bump and clump luminosities .
since the li - rich k giants are concentrated within a narrow luminosity range , one may suppose a significant event must be associated directly or indirectly with this range .
one suspects that the significant event is either the bump on the first ascent of the rgb and/or associated with the red clump formed of he - core burning giants that have experienced the he - core flash at the tip of the rgb .
the observed luminosity range is crossed by first ascent rgb stars in which the h - burning shell passes into the region that previously was the maximum extent reached by the convective envelope .
when the outward moving h - burning shell , as a result of increasing he - core mass , first encounters this region , a star reacts to the increased availability of h by lowering the surface luminosity . after a while
, the giant continues its ascent of the rgb .
this creates a kink on the evolutionary tracks for stars of masses up to 2m@xmath7 @xcite and a bump in the luminosity function along the rgb . as shown in fig . 2 , the kink on the evolutionary tracks @xcite is visible only up to 2m@xmath7 for solar metallicity stars .
internally , the kink is associated with an inversion in the run of mean molecular weight with distance from the stellar center .
this inversion is linked to destruction of @xmath3he by the reaction @xmath3he(@xmath3he,2@xmath26)@xmath27he which lowers the mean molecular weight , and the homogenization of the composition within the convective enevelope .
@xcite show that this inversion leads to ` compulsory ' mixing and changes to the surface abundances of c , n , and o isotopic abundances , i.e. , the @xmath21c/@xmath22c ratio is lowered relative to its value before the bump .
@xcite recognize too that mixing occurs as a result of the molecular weight inversion but include the effects of rotationally - induced mixing to drive the mixing .
this mixing referred to as @xmath28-mixing by @xcite or thermohaline mixing by @xcite is observationally confirmed by measurements of the @xmath21c/@xmath22c ratio in giants along the rgb showing a decrease in the ratio at and above the luminosity of the bump .
as @xmath26-captures on @xmath21c create @xmath22c , the reservoir of primordial and main - sequence synthesized @xmath3he is depleted .
it is this reservoir that is a potential source of @xmath4li from the cameron - fowler @xcite mechanism ( @xmath3he(@xmath27he,@xmath29)@xmath4be(@xmath30)@xmath4li ) but in order for the @xmath4li to enrich the stellar atmosphere it and its progenitor @xmath4be must be swept quickly to temperatures too cool for proton captures to occur .
@xcite calculations show that more than about 80% of the @xmath3he is destroyed in stars of masses less than about 1.5@xmath31 .
this destruction seems unlikely to produce lithium because the mixing is too slow for the @xmath4be and @xmath4li to avoid destruction by protons @xcite .
however , the initial subsurface @xmath3he reservoir is such that only a minor fraction of the @xmath3he need be converted with moderate efficiency to provide a li - rich giant .
it is anticipated that the lithium produced as the star crosses the bump s luminosity will be destroyed as the star with its convective envelope evolves to the tip of the rgb . the evidence from fig .
2 is that few of the li - rich stars are aligned along the run of bump stars in the hr diagram . although the cooler li - rich stars are likely to be bump stars , many li - rich stars are too warm to be so identified .
thus , we suggest that li - rich stars can not be identified exclusively with the bump .
rather the colocation of the li - rich stars in fig . 2 with
the high concentration of observed ( li - normal ) giants and the theoretical location of the clump suggests that they are clump stars . since li , if produced at the bump on the rgb is almost certainly destroyed by the time the star has evolved up the rgb to experience the he - core flash , the speculation is that the cameron - fowler mechanism operates at the he - core flash in at least some stars , i.e. , in stars of @xmath32 .
since @xcite predict survival of @xmath3he in stars with @xmath33 , there seems in principal the possibility that the he - core flash may be the key to synthesis of lithium in stars in a narrow mass range centered at about 2@xmath31 with the range s the upper limit set by the maximum mass for a he - core flash and the lower limit set by survival of sufficient @xmath3he following @xmath28 or thermohaline mixing on the rgb .
evidently , the concentration of li - rich stars at the clump implies that the synthesized li is swiftly destroyed as a clump star evolves along the early reaches of the agb .
if all stars evolving through the he - core flash synthesize copious amounts of li , they survive as a li - rich giant for about one per cent of their horizontal branch lifetime , i.e. , about 2 myr .
( the four luminous li - rich stars in fig .
2 have a mass in excess of those experiencing the he - core flash .
@xcite suggest that these tap their @xmath3he to produce li as the he - burning shell evolves outward . )
the principal novel result here is that the majority of li - rich k giants have a luminosity and effective temperature combination suggesting that li production occurs at the he - core flash in those stars .
although this speculation about the he - core flash has yet to be supported by calculations , the nuclear physics of li production is surely that described by the cameron - fowler mechanism , i.e. , conversion of @xmath3he to @xmath4li by @xmath34-capture with @xmath4be as a radioactive intermediary .
an earlier suggestion that li production , also by the cameron - fowler mechanism , occurs at the luminosity bump of the rib @xcite is required to account for cooler li - rich stars .
our discovery of that li - rich giants are concentrated within a narrow luminosity range does not support a view that li - rich giants result from the swallowing by the giant of a gaseous planet @xcite or some other external origin .
it remains to confirm and interpret suggestions that li - rich k giants may be unusual with respect to other k giants in exhibiting rapid rotation and/or an infrared excess ( see , for example , @xcite ; @xcite ) .
there are certainly li - rich giants with normal @xmath21c/@xmath22c ratios ( e.g. , hd 108471 in table 2 ) , and/or rapidly rotating surfaces ( e.g. , hd 217352 with @xmath35 km s@xmath36 , @xcite ) , and/or infrared excesses ( e.g. , pds 365 , @xcite ) .
perhaps , an area for observational scrutiny is a full determination of the c , n , and o elemental and isotopic abundances in order to search for discriminant between the candidate bump and clump li - rich stars themselves and between these li - rich stars and bump and clump stars exhibiting a normal li abundance .
finally , a radial velocity study should be undertaken to see if the collection of li - rich stars have the normal degree of binarity .
we are thankful to the trainees at crest and vbo observatory staff for their help during observations , and g. pandey for his observant comments .
dll thanks the robert a. welch foundation of houston , texas for support through grant f-634 .
this research has made use of the simbad database and the nasa ads service .
, j. b. 1967 , the observatory , 87 , 238 , a. , arribas , s. , & martnez - roger , c. 1999 , , 140 , 261 , s. c. , fekel , f. c. , henry , g. w. , & uitenbroek , h. 2000 , , 542 , 978 , d. , de castro , e. , fernandez - figueroa , m. j. , cornide , m. , & garcia lopez , r. j. 1998 , , 337 , 739 , s. v. & savanov , i. s. 1994 , astronomy letters , 20 , 639 , g. , girardi , l. , marigo , p. , & nasi , e. 2008 , , 484 , 815 , j. a. , sneden , c. , lambert , d. l. , & dutchover , e. j. , jr .
1989 , , 71 , 293 , a. g. w. & fowler , w. a. 1971 , , 164 , 111 , j. k. , smith , v. v. , cunha , k. , majewski , s. r. , & rood , r. t. 2010 , , 723 , l103 , c. & balachandran , s. c. 2000 , , 359 , 563 , c. & lagarde , n. 2010 , , 522 , a10 , n. a. , de la reza , r. , da silva , l. , & lambert , d. l. 2002 , , 123 , 2703 , p. p. , dearborn , d. s. p. , & lattanzio , j. c. 2008 , , 677 , 581 , r. g. & dantona , f. 1989 , , 215 , 66 , i. , jr . 1968 , , 220 , 143 , i. jr . 1967 , , 147 , 650 , i. jr .
1967 , , 147 , 624 , g. , parthasarathy , m. , de laverny , p. , & thvenin , f. 1999 , , 342 , 831 , y. b. & reddy , b. e. 2009 , , 703 , l46 , j. c. , dearborn , d. s. p. , & eggleton , p.
p. 2008 , in american institute of physics conference series , vol .
1001 , evolution and nucleosynthesis in agb stars , ed .
r. guandalini , s. palmerini , & m. busso , 279286 , k. , asplund , m. , & barklem , p. s. 2009 , , 503 , 541 , r. e. & heiter , u. 2007 , , 133 , 2464 , s. v. 1999 , , 352 , 495 , t. v. , bienaym , o. , gorbaneva , t. i. , et al .
2006 , , 456 , 1109 , m. a. c. , lindegren , l. , kovalevsky , j. , et al .
1997 , , 323 , l49 , n. k. , sriram , s. , jayakumar , k. , & gabriel , f. 2005 , japa , 26 , 331 , b. e. & lambert , d. l. 2005 , , 129 , 2831 , b. e. , lambert , d. l. , hrivnak , b. j. , & bakker , e. j. 2002 , , 123 , 1993 , k. , washuettl , a. , granzer , t. , scheck , m. , & weber , m. 2000 , , 142 , 275 , r. g. , macqueen , p. j. , sneden , c. , & lambert , d. l. 1995 , , 107 , 251 , f. , 2007 , hipparcos , the new reduction of the raw data , ( astrophys . space sci .
350 , springer , dordrecht ) , g. & sneden , c. 1982 , , 255 , 577 | in this _ letter _ , we report on a low - resolution spectroscopic survey for li - rich k giants among 2000 low mass ( @xmath0 ) giants spanning the luminosity range from below to above the luminosity of the clump .
fifteen new li - rich giants including four super li - rich k giants ( @xmath1(li ) @xmath2 ) were discovered .
a significant finding is that there is a concentration of li - rich k giants at the luminosity of the clump or red horizontal branch .
this new finding is partly a consequence of the fact that our low - resolution survey is the first large survey to include giants well below and above the rgb bump and clump locations in the hr diagram .
origin of the lithium enrichment may be plausibly attributed to the conversion of @xmath3he via @xmath4be to @xmath4li by the cameron - fowler mechanism but the location for onset of the conversion is uncertain .
two possible opportunities to effect this conversion are discussed : the bump in the first ascent of the red giant branch ( rgb ) and the he - core flash at the tip of the rgb .
the finite luminosity spread of the li - rich giants serves to reject the idea that li enhancement is , in general , a consequence of a giant swallowing a large planet . |
regge theory provides a useful tool for the phenomenological description of high energy hadron collisions @xcite .
the quantitative predictions of regge calculus are essentially dependent on the assumed coupling of participating hadrons to pomeron . in this paper high energy
elastic @xmath0 and @xmath1 scattering data including the recent lhc ones are analyzed in terms of the simple regge exchange approach in the framework of additive quark model ( aqm ) @xcite . in aqm baryon
is treated as a system of three spatially separated compact objects
constituent quarks .
each constituent quark is colored , has internal quark - gluon structure and finite radius that is much less than the radius of proton , @xmath2 .
this picture is in good agreement both with @xmath3 symmetry of strong interaction and the quark - gluon structure of proton @xcite .
the three constituent quarks are assumed in aqm to be the incident particles in @xmath0 or @xmath1 scattering .
elastic amplitudes for the large energy @xmath4 and the small momentum transfer @xmath5 are dominated by the pomeron exchange .
we neglect the small difference in @xmath0 and @xmath1 scattering coming from the exchange of negative signature reggeons , odderon ( see e.g. @xcite and references therein ) , @xmath6-reggeon etc , since their contribution is suppressed by @xmath7 . the single @xmath5-channel exchange results into @xmath7-channel amplitude of the constituent quarks scattering @xmath8 where @xmath9 is the pomeron trajectory specified by the values of intercept , @xmath10 , and slope , @xmath11 . the ( positive ) signature factor
, @xmath12 determines the complex structure of the amplitude .
the factor @xmath13 has the meaning of the pomeron coupling to the beam and target particles , the functions @xmath14 being the vertices of constituent quark - pomeron interaction ( filled circles in fig . [ 1p2p ] ) .
as the same set of the pomeron parameters @xmath15 , @xmath11 and @xmath16 describes proton and antiproton scattering , both @xmath0 and @xmath1 data have been commonly used to fix their values . and @xmath1 elastic amplitude .
we neglect their difference . ]
the one - pomeron exchange between two protons includes the sum over all possible exchanges between the quark pairs @xcite .
each term in the sum has a form ( [ mqq ] ) with the functions @xmath14 attributed to the individual constituent quarks . due to factorization of longitudinal and transverse degrees of freedom the longitudinal momenta
are integrated over separately in high energy limit .
after this the transverse part of the quark distribution is actually relevant only . writing the proton wavefunction in the transverse momentum space as @xmath17 , where @xmath18 are the quark transverse momenta , @xmath19 the proton - pomeron vertex , @xmath20 , @xmath21 , is given by the overlap function @xmath22 the function @xmath20 plays a role of the proton formfactor for strong interaction in aqm ( see section 2 ) . in what follows
we assume the pomeron trajectory in the simplest form @xmath23 the value @xmath24 defines the radius of quark - quark interaction while @xmath25 has the meaning of typical energy scale in regge theory .
putting together all 9 equal quark - quark contributions ( one of them is shown in fig . [ 1p2p ] a ) we get the first order elastic @xmath0 ( or @xmath1 , we do not distinguish between them here ) amplitude @xmath26 actually the formula ( [ m1 ] ) with a single pomeron gives amplitude in the impulse approximation @xcite .
similarly to light nuclear scattering the multipomeron exchange should be added ( see glauber theory @xcite ) giving rise to the terms @xmath27 , @xmath28 etc , so that the total amplitude @xmath29 if @xmath30 the multiple interactions become negligible leaving in the sum the first term only . the optical theorem , that relates the total elastic cross section and imaginary part of the amplitude , in the normalization adopted here reads @xmath31 the differential cross section is evaluated in this normalization as @xmath32.\ ] ] interference of the contributions generated by the various number of pomerons leads to the occurrence of local minima in the differential elastic cross section .
experimentally the minimum at @xmath33 gev@xmath34 for the energy @xmath35 tev is well observed at lhc @xcite .
the minima at another @xmath5 are also possible .
basically there is interplay between the minima in real and imaginary parts of the amplitude in the expression ( [ ds / dt ] ) so that the minimum in the imaginary part could be filled by the large real part at the same @xmath5 .
the present paper aims to give theoretical description of the experimental @xmath0 and @xmath1 elastic scattering data in the energy interval 546 gev @xmath36 7 tev in the framework of aqm .
the analysis of the differences in @xmath0 and @xmath1 amplitudes needs a more subtle treatment than that involving only pomeron exchanges and goes beyond our approach .
in aqm there are a total 9 orders of interactions .
the first order is the sum of all interactions between single @xmath37 pairs pair interaction in aqm may include the contributions of several gribovs pomerons @xcite .
we assume that high energy @xmath37 scattering is described by single effective pomeron exchange between each @xmath37 pair , the parameters of this effective pomeron could be different from those of gribov s bare pomerons . ]
it contains 9 terms .
similar to glauber theory @xcite one has to rule out the multiple interactions between the same quark pair .
aqm permits the pomeron to connect any two quark lines only once .
it crucially decreases the combinatorics leaving the diagrams with no more than 9 effective pomerons .
examples of various order diagrams are shown in fig [ 1p2p],[3p ] . elastic scattering .
the straight lines stand for quarks , the waved lines denote pomerons , @xmath38 is the momentum transferred , @xmath21 .
diagram ( a ) is the one of the single pomeron diagrams , diagrams ( b ) and ( c ) represent double pomeron exchange with two pomeron coupled to the different quark ( b ) and to the same quarks ( c ) , @xmath39 . ] let @xmath40 be the transverse momentum carried by the @xmath41-th effective pomeron , @xmath42 and @xmath43 denote the momenta transferred to the quark @xmath44 from the target proton or quark @xmath45 from the beam proton during the scattering process . if no pomerons are attached to the quark @xmath46 , that is it does not interact , then @xmath47 .
if only one pomeron carrying momentum @xmath40 is attached to the quark @xmath46 , then @xmath48 .
if two pomerons with the momenta @xmath40 and @xmath49 are coupled to it , then @xmath50 . in other words , @xmath51 @xmath52 with these notations the @xmath53 order amplitude is equal to @xmath54 where the sum in the last factor is taken over all distinct ways to connect the pairs of beam and target quarks by @xmath53 effective pomerons , each pair being connected no more than once .
the permutations of identical pomerons in the integrals is compensated by @xmath55 in front the sum .
there are two types of diagrams in the second order , @xmath56 where the first term comes from the diagrams with both pomerons coupled to the different quark lines whereas in the second one they are attached to the same line ( fig .
[ 1p2p ] b , c ) .
the third order sum is @xmath57 @xmath58 here the first term arises when all three pomerons are attached to the same line , in the last term they connect three different quarks , diagrams ( a ) and ( d ) in the fig . [ 3p ] .
the second and the third terms correspond to the diagrams ( b ) and ( c ) .
the rest terms are provided by various permutations of the pomeron lines in these diagrams , in particular , by flipping a diagram as a whole .
the numerical coefficients encounter the number of connections resulting into equal expressions after variables changing in the integrals ( [ mn ] ) . in the highest order , containing 9 effective pomerons , @xmath59
@xmath60 each quark from the beam proton interacts with the quark from the target one .
the other orders , @xmath61 , have similar but more lengthly structure due to large combinatorics to redistribute @xmath62 momenta among @xmath63 and @xmath64 groups and @xmath65 scattering can be found in ( @xcite ) . ] .
the function @xmath66 is determined by the proton wavefunction in terms of the constituent quarks ( [ fp ] ) .
we model the transverse part of the wavefunction in a simple form of two gaussian packets , @xmath67,\ ] ] normalized to unity ( [ norm ] ) .
one packet parametrization , @xmath68 , is insufficient to reproduce experimental data as imposing too strong mutual dependence between the total cross section , the minimum position and the value of the slope at @xmath69 .
all parameters used in the calculation naturally fall into two different kinds : the parameters of the pomeron and those specifying the structure of colliding particles .
the former type , @xmath70 , @xmath71 , @xmath16 , refers to the high energy scattering theory while the latter , @xmath72 and @xmath73 , details the matter distribution inside the proton in the low energy limit(similar to density distribution in atomic nuclear ) . with the chosen values of @xmath72 and @xmath73 ( see below )
the wavefunction ( [ gausspack ] ) results into the density shown in fig .
[ rho ] ) and parameters ( [ set]).,title="fig : " ]
-1.5 cm it looks rather naturally having mean squared radius @xmath74 fm , that is close to the electromagnetic radius .
some discrepancy can be explained by the difference in the distributions of charged and strong interacting matter inside the proton .
here we present the numerical results obtained by summing over all 9 orders of interaction of the additive quark model for the energies in the interval @xmath75 gev @xmath36 7 tev .
they are compared with the experimental data taken from @xcite .
the following set of parameters have been used @xmath76 in choosing these values we have not distinguished @xmath0 and @xmath1 data . -1.cm scattering at @xmath35 tev and @xmath1 scattering at @xmath77 gev .
the experimental points have been taken from @xcite.,title="fig : " ] scattering at @xmath35 tev and @xmath1 scattering at @xmath77 gev .
the experimental points have been taken from @xcite.,title="fig : " ] -1.5 cm -1.cm scattering at @xmath78 gev and elastic @xmath1 and @xmath0 scattering at @xmath79 gev .
the experimental points have been taken from @xcite ( left panel ) , @xcite for @xmath1 and @xcite for @xmath0 ( right panel).,title="fig : " ] scattering at @xmath78 gev and elastic @xmath1 and @xmath0 scattering at @xmath79 gev .
the experimental points have been taken from @xcite ( left panel ) , @xcite for @xmath1 and @xcite for @xmath0 ( right panel).,title="fig : " ] -1.5
cm the calculated values of the total cross section @xmath80 , of @xmath81 , of the slope of the elastic cross section @xmath82 ( @xmath83 , and of the ratio @xmath84 are presented in the table 1 along with the experimental data .
the theoretical slope is calculated within the interval @xmath85 gev@xmath34 . here
we again neglect the difference between @xmath0 and @xmath1 collisions in the presented data . [ cols="^,^,^,^,^ " , ] -0.2 cm table 1 . the comparison of the calculated values of total cross sections @xmath80 , of @xmath81 , slope parameter @xmath82 and ratio @xmath86 with the available experimental data .
0.2 cm the values of the total cross sections and the slopes are in reasonable agreement with the experimental data , the value @xmath87 is also well reproduced at @xmath78 gev .
the aqm assumption @xmath88 continues to hold satisfactorily even at the lhc energy @xmath35 tev for @xmath89 , @xmath90 .
probably it would be better fulfilled if the pomeron slope @xmath71 would be lesser .
the value obtained for the total cross section for @xmath1 scattering at @xmath79 gev are smaller than the experimental data
. the reason might be in non - pomeron contributions , for example in @xmath91 reggeon , whose effect could be significant for the low energies but disappears when the energy grows . the experimental position of the local minimum of the differential @xmath0 cross section at @xmath92 tev is also reproduced although the theory predicts the more deep minimum . as the transferred momentum increases , @xmath93 gev@xmath34 , the interaction becomes sensitive to the internal structure of constituent quarks that have a finite size .
we have used a simple parametrization for their radius @xmath94 , but there is a possibility for additional constants or more complicated functions .
-1.cm scattering at @xmath35 tev ( left panel ) and elastic @xmath1 scattering at @xmath78 gev ( right panel ) . the dashed line ( @xmath95 ) presents the sum of first four orders of aqm calculated with the parameters @xmath96 , @xmath97 gev@xmath98 , @xmath99 gev@xmath98 , @xmath100 gev@xmath98 , @xmath101 gev@xmath98 , @xmath102 .
the solid line ( @xmath103 ) presents the sum of all aqm orders for the parameters ( [ set ] ) .
it exhibits the second local minimum for @xmath0 scattering at @xmath35 tev .
the experimental points have been taken from @xcite , title="fig : " ] scattering at @xmath35 tev ( left panel ) and elastic @xmath1 scattering at @xmath78 gev ( right panel ) .
the dashed line ( @xmath95 ) presents the sum of first four orders of aqm calculated with the parameters @xmath96 , @xmath97 gev@xmath98 , @xmath99 gev@xmath98 , @xmath100 gev@xmath98 , @xmath101 gev@xmath98 , @xmath102 . the solid line ( @xmath103 )
presents the sum of all aqm orders for the parameters ( [ set ] ) .
it exhibits the second local minimum for @xmath0 scattering at @xmath35 tev .
the experimental points have been taken from @xcite , title="fig : " ] -1.5 cm . [ 4 - 9 ] the experimental data for @xmath1 scattering show no clean dips at the smallest energies @xmath77 gev , 546 gev and 62 gev .
the reason , probably , is in substantial contribution that could be made to @xmath1 amplitude by by the negative signature reggeons ( see conclusion ) .
it is important to note that the complete 9 orders aqm calculation yields a second local minimum of the differential elastic @xmath0 cross section at @xmath35 tev for @xmath104 gev@xmath34 unobservable experimentally .
it may indicate to an invalidity of our description for @xmath105 gev@xmath34 , where the internal structure of the constituent quarks becomes important .
from another point of view there is an uncontrolled contribution of multipomeron diagrams , ( e.g. enhanced pomeron diagrams @xcite ) rapidly growing with the number of pomerons .
their numerical value is determined by unknown vertices of @xmath106 pomeron transitions , which increases the uncertainty of the next orders . in this context , it is worth stressing that the sum of only first four orders of aqm results ( with a little modified parameters ) into the theoretical curve that better fits the lhc @xmath0 data at the more broad @xmath5 interval and does not exhibit an extra local minimum as is shown in fig . [ 4 - 9 ] left .
an example of aqm predictions for @xmath1 scattering in the region @xmath93
gev@xmath34 for the energy @xmath78 gev is presented in fig .
[ 4 - 9 ] right .
the theory gives here the large dip , which is in fact absent in the experimental data .
it points to the essential effect of the contributions , that are not accounted for in aqm , such as negative signature reggeons etc .
nevertheless , the first four orders of aqm better describe the cross section behavior ( give much smaller dip than the complete sum ) even in this case .
both theoretical curves , @xmath95 and @xmath103 , have no extra local minima up to @xmath107 gev@xmath34 for this energy .
we show that the simple aqm model gives reasonable description of the high energy @xmath0 and @xmath1 elastic scattering at the not large momentum transferred , @xmath108 gev@xmath34 , that is at the distances where the internal structure of the constituent quarks probably do not show up .
our model contains 6 parameters , but only 3 of them , ( @xmath70 , @xmath71 , @xmath16 ) , are employed to describe high energy scattering while 3 others ( @xmath72 , @xmath73 ) determine the matter distribution in the proton .
note that the first group of parameters refers to the pomeron and has been chosen over together taken @xmath0 and @xmath1 data .
the interesting fact is that the matter distribution can not be parameterized by one gaussian packet ( see also @xcite ) . as mentioned in the introduction we neglect the contribution of the negative signature reggons that leads to the difference in @xmath0 and @xmath1 scattering especially around the dip .
some difference between them is experimentally observed @xcite at @xmath109 gev near the dip ( @xmath110 gev@xmath34 ) .
the energy behavior of this difference depends on the value of the odderon intercept , @xmath111 . the same odderon effects are responsible for the difference in @xmath112 and @xmath113 yields in the central region of the @xmath0 collision .
however , it was shown @xcite that the observed ratio @xmath114 is compatible with the value @xmath115 . in this case
the difference in @xmath116 for @xmath0 and @xmath1 scattering should very fast decrease with the initial energy growth and seems to be very small at lhc . as a consequence
the dip in the differential cross section clearly observed at lhc energy for @xmath0 scattering should manifest itself in @xmath1 scattering as well .
it can be considered as our prediction . | high energy @xmath0 and @xmath1 elastic scattering is treated in the framework of additive quark model .
the reasonable agreement with experimental data is achieved with the natural parameters for the strong matter distribution inside proton .
* high energy elastic pp scattering in additive quark model * yu.m . shabelski and a.g .
shuvaev + petersburg nuclear physics institute , kurchatov national research centre + gatchina , st . petersburg 188300 , russia + 0.9 truecm e - mail : [email protected] + e - mail : [email protected] pacs .
25.75.dw particle and resonance production |
understanding of the origin and evolution of galaxies , stars , planets , our earth and of life itself are fundamental objectives of science in general and astronomy in particular .
although impressive advances have been made in the last twenty years , our knowledge of how the universe has come to look as it does today is far from complete .
a full insight of the processes involved is only possible with observations in the long wavelength infrared waveband of the electromagnetic spectrum .
it is in this range that astronomical objects emit most of their radiation as they form and evolve in regions where obscuration by dust prevents observations in the visible and near infrared . over the past quarter of a century successive space infrared observatories ( iras , iso , spitzer and akari ) have revolutionized our understanding of the evolution of stars and galaxies .
mid to far infrared observations have led to stunning discoveries such as the ultra luminous infrared galaxies ( ulirgs ) , the basic processes of star formation from `` class 0 '' protostars through to the clearing of the gaseous proto - planetary disks and the presence of dust excesses around main sequence stars .
the herschel space observatory launched in 2009 continues this work in the far infrared and sub - mm and jwst , due for launch around 2014 , will provide a major boost in observing capability in the 228@xmath1 m range .
previous infrared missions have been hampered by the requirement to cool the telescope and instruments to @xmath25 k using liquid cryogens .
this has limited the size of the apertures to @xmath21 m and our view of the infrared universe has been one of poor spatial resolution and limited sensitivity .
mission addresses the first of these by employing a 3.5 m mirror to dramatically increase the available spatial resolution but , because it is only cooled to @xmath080k ( pilbratt et al .
2010 ) , only offers a modest increase in sensitivity in the 55210@xmath1 m range compared to previous facilities .
jwst will provide a major increase in both spatial resolution and sensitivity but only up to 28@xmath1 m .
the jaxa led mission spica is an observatory that will provide imaging and spectroscopic capabilities in the 5 to 210@xmath1 m wavelength range with a @xmath03.2 m telescope ( current design ) , but now cooled to a temperature of @xmath06 k. in combination with a new generation of highly sensitive detectors , the low telescope temperature will allow us to achieve sky - limited sensitivity over the full 5 to 210@xmath1 m band for the first time .
this unique capability means that spica will be between one and two orders of magnitude more sensitive than herschel in the far infrared band .
spica will cover the full 5 to 210@xmath1 m wavelength range , including the missing 28 - 55@xmath1 m octave which is out of the herschel and jwst domains , with unprecedented sensitivity and spatial resolution .
furthermore , spica will be the only observatory of its era to bridge the wavelength gap between jwst and alma , and carry out unique science not achievable at ( sub)mm wavelengths with alma . in the mid infrared spica
will be able to carry out medium and high - resolution ( @xmath5 ) spectroscopy , one order of magnitude higher than in jwst , and will include spatial high multiplexing imaging and medium spectral resolution capabilities .
in addition , the characteristics of the spica monolithic telescope will provide unique and optimal conditions for mid infrared coronagraphy in imaging and , uniquely , spectroscopic mode . at present , the scientific instruments foreseen for spica are : + @xmath6 a mid infrared camera and spectrophotometer , miracle ( 5 - 38@xmath1 m , 6@xmath7@xmath86@xmath7 fov ) .
+ @xmath6 a mid infrared medium resolution spectrometer , mirmes ( 10.32 - 36.04@xmath1 m with a spectral resolution of @xmath9900 - 1500 ) .
+ @xmath6 a mid infrared high resolution spectrometer , mirhes ( 4 - 18@xmath1 m with a spectral resolution of @xmath920,000 - 30,000 ) .
+ @xmath6 a mid infrared coronagraph , sci ( 5(3.5)27@xmath1 m , @xmath920 - 200 , contrast 10@xmath10 , inner working angle @xmath03.3@xmath3/d@xmath11 ) .
+ @xmath6 safari , a far infrared imaging fourier - transform spectrometer designed to provide continuous coverage in photometry and spectroscopy from 34 to 210@xmath1 m , with a field of view of 2@xmath7@xmath82@xmath7 and spectral resolution modes of @xmath12@xmath02000 ( at 100@xmath1 m ) , @xmath12 a few hundred and 20@xmath1350 .
the spectral sensitivity is required to be @xmath03.5@xmath810@xmath14wm@xmath15 at 48@xmath1 m ( 5@xmath16 , 1 hour ) .
tes bolometer detectors will be used .
safari is to be built by a consortium of european institutes lead by sron ( with canadian and japanese participation ) .
spica offers a sensitivity up to two orders of magnitude better than herschel covering the mid - to - far - ir ( mostly unreachable from the ground ) with imaging , spectroscopic and coronagraphic instruments ( see figure 1 ) .
thanks to this tremendous increase in sensitivity , spica will make photometric images in a few seconds that would take hours for herschel and will take a full 5 - 210@xmath1 m infrared spectrum of an object in one hour that would take several thousand hours for herschel .
we illustrate this major increase in sensitivity in figure 2 which shows the area covered by a full spatial and spectral survey in which we can detect spectroscopically all galaxies down to a luminosity of @xmath010@xmath17l@xmath18 at z=1 and @xmath010@xmath19@xmath20 at z=2 in 900 hours . in theory , in about twice this time
, the herschel - pacs spectrometer would just be able to detect a single object over its full waveband to the same sensitivity .
we can immediately see that this major increase in sensitivity , combined with a wide field of view and coverage of the full 5210@xmath1 m waveband , will revolutionize our ability to spectroscopically explore the nature of the thousands of objects that herschel , jwst , and spica will discover in photometric surveys .
+ with its powerful scientific capabilities , spica will provide unique and ground - breaking answers to many key questions , and it is with this goal that the spica science objectives are being defined : * ( i ) the formation and evolution of planetary systems * : by accessing key spectral diagnostic lines , spica will provide a robust and multidisciplinary approach to determine the conditions for planetary systems formation .
this will include the detection for the first time of the most relevant species and mineral components in the gas and dust of hundreds of transitional protoplanetary disks at the time when planets form .
spica will also be able to trace the warm gas in the inner ( @xmath230au ) disk regions and by resolving the gas keplerian rotation ( _ e.g. , _ by resolving the mid - ir h@xmath21 lines ) , will allow us to observe the evolution of disk structure due to planet formation .
spica will study debris disks and make the first unbiased survey of the presence of zodiacal clouds in thousands of exoplanetary systems around all stellar types .
it will allow us to detect both the dust continuum emission and the brightest grain / ice bands as well as the brightest lines from any gas residual present in the disk .
spica will have the unique capability to observe water ice in all environments , and thus fully explore its impact on planetary formation and evolution and in the emergence of habitable planets . in the closest debris disks , spica will spatially resolve the distribution of water ice and determine the position of the `` snow line '' around vega - like stars , which separates the inner disk region of terrestrial planet formation from that of the outer planets .
spica will also drastically enhance our knowledge of the solar system by making the first detailed characterization of hundreds of kuiper belt objects , and of different families of inner , hotter centaurs , comets and asteroids .
spica will provide the means to quantify their composition and determine unambiguously their size distribution : critical observational evidence for the models of solar system formation .
no other planned or present facility will be able to carry out these observations .
spica will provide direct imaging and low resolution mid infrared spectroscopy of outer young giant exoplanets ( _ e.g. , _ at @xmath010au of a star at @xmath010pc ) , which will allow us for the first time to study the physics and composition of their atmospheres in a wavelength range particularly rich in spectral signatures ( _ e.g. , _
h@xmath21o , ch@xmath22 , o@xmath23 , silicate clouds , nh@xmath23 , co@xmath21 ) and to compare it with solar system planet atmospheres .
in addition , mid infrared transit photometry and high resolution spectroscopy of `` hot jupiters '' will be routine with spica .
finally , by mapping very extended but faint regions , spica will also provide an unprecedentedly sensitive window into key aspects of the dust life - cycle both in the milky way and in nearby galaxies , from its formation in evolved stars , its evolution in the ism , its processing in supernova - generated shock waves and massive stars , to its final incorporation into star forming cores and protoplanetary disks . *
( ii ) the formation and evolution of galaxies * : spica observations will provide a unique insight into the basic questions about how galaxies form and evolve such as : what drives the evolution of the massive , dusty distant galaxy population , and what feedback / interplay exists between the physical processes of mass accretion and star formation ?
how and when do the normal , quiescent galaxies such as our own form , and how do they relate to ( ultra ) luminous infrared galaxies ( ulirgs ) ? how do galaxy evolution , star formation rate and agn activity vary with environment and cosmological epoch ?
substantial progress in this area can only be made by making the transition from large - area photometric to large - area spectroscopic surveys in the mid to far infrared , which will be possible with spica .
this is because the mid and far infrared region plays host to a unique suite of diagnostic lines to trace the accretion and star formation , and to probe the physical and chemical conditions in different regimes from agn to star - forming regions .
while the herschel - pacs spectrometer detects the brightest far infrared objects at z@xmath01 , spica will be able to carry out blind spectroscopic surveys out to z@xmath03 .
this will lead to the first statistically unbiased determination of the co - evolution of star formation and mass accretion with cosmic time .
spectroscopic surveys will provide direct and unbiased information on the evolution of the large scale structure in the universe from z@xmath03 and the unprecedented possibility to investigate the impact of environment on galaxy formation and evolution as a function of redshift .
the high sensitivity of spica will enable photometric surveys beyond z@xmath04 that will resolve more than 90% of the cosmic infrared background ( in comparison with 50% that herschel will achieve ) .
spica will also observe milky way type galaxies in the far infrared out to z@xmath01 , where the cosmic star formation rate peaks .
for the first time , we will be able to piece together the story of the evolution of our own galaxy and answer the question of whether we are in a
special place in the cosmos .
+ by observing at far - ir wavelengths , herschel is presently providing spectacular results in many fields of astronomy .
a new mission reaching better sensitivity in the mid- and far - ir domain must therefore follow to continue this work and gain a deeper understanding of the physics of the objects discovered by herschel : _
i.e. , _ to spectroscopically measure star formation rates in the obscured extra - galactic sources at different cosmic epochs and explore the physics and chemistry of planet formation .
to obtain this increase in sensitivity we need a cold ( @xmath26k ) telescope .
spica promises this in 2018 - 2020 .
spica will be an international mission and will be open to the worldwide astronomical community .
japan is in charge of the whole integration of the system .
the assessment study on the european contribution to the spica project has been carried out under the framework of the esa cosmic vision ( _ spica assessment study report for esa cosmic vision 2015 - 2025 plan _ , see 2010arxiv1001.0709s for details ) .
us and korean participation are also being discussed .
+ + _ acknowledgments _ : we warmly thank kate isaak , bruce swinyard , ana heras and the spica / safari science teams for fruitful discussions over the last years .
j.r.g . was supported by a _ ramn y cajal _ contract and by the aya2009 - 07304 and csd2009 - 00038 grants from the spanish micinn .
99 pilbratt , g. l. et al .
2010 , _ a&a _ , 518 , l1 .
springel , v. , frenk , c.s .
& white , s.d.m .
2006 , _ nature _ , 440 , 1137 .
spica study team collaboration , 2010 .
_ spica assessment study report for esa cosmic vision 2015 - 2025 plan _ ,
swinyard , b. , nakagawa , t. , et al .
2009 , _ experimental astronomy _ , 23 , 1 , 193 - 219 . | we present an overview of spica , the _ space infrared telescope for cosmology and astrophysics _ , a world - class space observatory optimized for mid- and far - ir astronomy ( from 5 to @xmath0210@xmath1 m ) with a cryogenically cooled @xmath03.2 m telescope ( @xmath26k ) .
its high spatial resolution and unprecedented sensitivity in both photometry and spectroscopy modes will enable us to address a number of key problems in astronomy .
spica s large , cold aperture will provide a two order of magnitude sensitivity advantage over current far ir facilities ( @xmath3@xmath430@xmath1 m wavelength ) . in the present design
, spica will carry mid - ir camera , spectrometers and coronagraph ( by jaxa institutes ) and a far - ir imager fts - spectrometer , safari ( @xmath034 - 210@xmath1 m , provided by an european / canadian consortium lead by sron ) .
complementary instruments such as a far - ir / submm spectrometer ( proposed by nasa ) are also being discussed .
spica will be the only space observatory of its era to bridge the far ir wavelength gap between jwst and alma , and carry out unique science not achievable at visible or submm wavelengths . in this contribution
we summarize some of the scientific advances that will be made possible by the large increase in sensitivity compared to previous infrared space missions . |
the nature of current accelerated expansion of the universe is one of the mysterious and interesting topics in cosmology @xcite . in order to interpret this expansion ,
some various candidates have been proposed by many authors , such as cosmological constant idea @xcite , dark energy models @xcite , ( exotic substance with large negative pressure @xmath3 ( @xmath4 is the dark energy density ) ) and modified gravities like lovelock gravity @xcite , brane world cosmology @xcite , scalar - tensor theories @xcite and also the so - called @xmath0 gravity theories @xcite .
on the other hand , amongst the nonlinear modifications of einstein gravity , @xmath0 models seem to provide an interpretation to dark energy , the hierarchy problem @xcite , the four cosmological phases @xcite , the power law early - time inflation @xcite , late - time cosmic accelerated expansion @xcite , singularity problem arising in the strong gravity regime @xcite , rotation curves of spiral galaxies @xcite and detection of gravitational waves @xcite .
in addition , when one considers @xmath0 theory as a modification of general relativity , it is quite natural to ask about black hole existence in this theory .
furthermore , it is notable that viable modifications in gravity should pass all sorts of tests from the large scale structure of the galaxy and cluster dynamics to the solar system tests .
hence , many different models of @xmath0 theory with various motivations have been proposed ( see @xcite and references therein for more details ) . moreover , regarding einstein gravity , one can find a large number of exact spherically symmetric solutions .
one of important works in the modified theories of gravity is obtaining the static spherically symmetric solutions , a requirement usually referred to as spatial isotropy and time independence .
therefore , the most widely explored exact solutions in @xmath0 gravity are the spherically symmetric solutions .
spherically symmetric solutions of @xmath0 gravity have been studied before @xcite .
furthermore , some of interesting black objects with various geometry have been investigated in @xmath0 gravity , for e.g. , static and rotating black hole / string solutions , magnetic string and so on @xcite .
also , the black hole / brane solutions with a nonlinear maxwell source in @xmath0 gravity has been investigated in @xcite . charged spherically symmetric black hole solutions
have been obtained in refs . @xcite and creation of the electric charge and cosmological constant from pure gravity , simultaneously ,
have been extracted in @xcite .
recently , sebestiani and zerbini @xcite have considered a class of interesting static spherically symmetric spacetime in the form @xmath5 with a model of @xmath0 gravity . in this work
, we would like to consider this metric and obtain the exact spherically symmetric black hole solutions for general value of @xmath6 and discuss about the behavior of these solutions .
we will consider @xmath7 theory , namely pure gravity ( @xmath8 ) , with a constant curvature scalar .
one of our purposes is finding the four - dimensional charged black hole solutions with cosmological constant in pure gravity similar to solutions of einstein-@xmath9-power maxwell invariant gravity @xcite .
in addition , we will investigate the behavior of these solutions and also discuss about dolgov - kawasaki ( dk ) stability @xcite .
the outline of our paper is as follows . in section [ fieldf(r ) ] , we obtain a static spherically symmetric solutions of a class of @xmath0 model and some special values of free parameter , @xmath6 , will be discussed in appendix . in section [ lifshitz ] , we compare obtained solutions with asymptotically lifshitz black holes .
we terminate our paper by a conclusion .
considering the field equation of @xmath10-dimensional pure @xmath0 gravity with the following form @xcite @xmath11 where @xmath12 is the ricci tensor , @xmath13 and @xmath0 is an arbitrary function of ricci scalar @xmath14 . here , we want to obtain the static solutions of eq .
( [ fe ] ) with positive , negative and zero curvature horizons . for this purpose , we assume that the metric has the following form @xcite @xmath15 where @xmath6 is an arbitrary constant and @xmath16 to find the function @xmath17 , we use the components of eq .
( [ fe ] ) with well - known @xmath18 model .
this special model and some of its properties have been investigated in refs .
hendigrg , hendiprd . using eq .
( [ fe ] ) with metric ( [ met1 ] ) , one can obtain a general solution in the following form @xmath19 where @xmath20 , @xmath21 and @xmath22 are integration constants and @xmath23 since we know that the cosmological constant may arises from various models of pure @xmath0 gravity ( see @xcite and references therein ) , one may set @xmath24 to obtain a consistent solutions for @xmath25 . in order to satisfy all components of the field equation ( [ fe ] ) , we should set the parameters of the @xmath0 model to satisfy the following equations @xmath26 @xmath27 where @xmath28 solving eqs .
( [ sol1 ] ) and ( [ sol2 ] ) for @xmath29 and @xmath30 with arbitrary @xmath31 lead to the following solutions @xmath32 @xmath33 after some cumbersome manipulation we obtain @xmath34 \frac{\lambda ^{2}}{9 } \label{riem2},\end{aligned}\ ] ] which confirm that there is a singularity at @xmath35 .
the position of the black hole event horizon is determined as a largest root of @xmath36 .
( [ ric1 ] ) and ( [ riem2 ] ) show that the curvature scalars are asymptotically finite and related to the cosmological constant .
but we can not redefine a unique effective cosmological constant for whole curvature scalars with an arbitrary @xmath6 .
therefore , we could not state these solutions are asymptotically ( a)ds with an effective cosmological constant .
nevertheless , the asymptotic behavior of the solutions are near to ( a)ds solutions . here , we give a brief discussion about the stability of the mentioned @xmath0 model .
it was showed that there is no stable ground state for @xmath0 models if @xmath37 and @xmath38 @xcite .
for the obtained solutions , we find that @xmath39 .
furthermore , it has been shown that @xmath40 is related to the the effective mass of the dynamical field of the ricci scalar @xcite .
therefore , the positive effective mass , a requirement usually referred to as the dk stability criterion , leads to the stable dynamical field @xcite .
in addition , one can discuss other instability criteria of the asymptotically lifshitz spacetimes considered in @xcite . in ref .
@xcite , the authors have employed the initial value problem to investigate the stability of such spacetimes .
we will investigate other instability criteria for the future works and in this paper , we consider dk stability . in order to check the dk stability criterion
, we should calculate the second derivative of the @xmath0 function with respect to the ricci scalar for this specific model @xmath41 the positivity of @xmath42 depends on the model parameters . in order to study the stability , we plot @xmath42 in fig .
this figure shows that one may obtain stable solution for special values of @xmath31 for negative cosmological constant . in other words
, we can set free parameters to obtain positive @xmath42 .
it is notable to mention that since eq .
( [ fe ] ) is a trivial relation for obtained solutions , we are not allowed to use the conformal transformation to discuss about the dynamic of the solutions .
in recent years , it has been discussed about duality of gravity and nonrelativistic scale invariant theories .
it may be considered a scale invariant fixed point that do not exhibit galilean symmetry with the metric of the corresponding gravity duals @xcite .
these metrics exhibit an anisotropic scale - invariant field theories in which time and space scale differently @xmath43 which is characterized by the so - called dynamical exponent @xmath44 .
one may obtain the standard scaling behavior of conformal invariant systems for @xmath45 .
in other words , for @xmath46 , it corresponds to relativistic invariance ( the scaling is isotropic ) . also for arbitrary values of @xmath44
, one can say that the system has lifshitz scaling . in some literature
a special solution was found , which corresponds to the asymptotically lifshitz black hole with various dynamical exponents @xcite .
now , let us compare the presented solutions with asymptotically lifshitz black holes . in order to accomplish this goal , we consider the following asymptotically lifshitz ansatz @xmath47 it is easy to show that eq .
( [ met1 ] ) is asymptotically lifshitz spacetime provided one apply the following transformation@xmath48 in other word , asymptotically lifshitz metric ( [ alifmet ] ) ( with the mentioned @xmath0 model ) leads to the following metric function @xmath49 where @xmath50 it is notable that the generic solution eqs .
( [ alifmet ] ) and ( [ h(r ) ] ) reduces to the solution of obtained in ref .
@xcite , for @xmath51 and @xmath52 .
in this paper , we have considered a new class of static spherically symmetric spacetime with a special model of @xmath0 gravity .
we have obtained some interesting solutions for different values of free parameter , @xmath6 , and studied the geometrical properties of the solutions .
we have investigated the analogy between obtained solutions with @xmath10-dimensional solutions of einstein-@xmath53 gravity in the presence of a nonlinear source , namely power maxwell invariant ( where @xmath54 denotes its nonlinearity parameter ) .
in other words , one can find that the integral constant @xmath21 may interpreted as @xmath55 provided @xmath56 .
in addition , for @xmath57 , one may interpreted @xmath22 as @xmath55 .
furthermore , we have calculated kretschmann scalar and found that there is a curvature singularity at @xmath35 . we have investigated the behavior of the solutions for large value of @xmath58 ( @xmath59 ) and showed that the asymptotic behavior of the solutions is neither flat nor ( a)ds . also , we have examined the dk stability criterion and showed that one may obtain a stable theory provided the parameters of the @xmath0 model are chosen suitably .
finally , we should remark that obtained solutions are interesting because of the following three main reasons : ( i ) both the geometry of the horizon(s ) and asymptotic behavior of the solutions depend on the values of free parameter @xmath6 , ( ii ) starting from pure @xmath0 gravity and extract the solutions of einstein-@xmath1-power maxwell invariant theory , ( iii ) obtained solutions are , exactly , the same as asymptotically lifshitz black holes , provided the transformation between them are chosen suitably .
now , we would like to discuss about the geometry of the spacetime with various values of metric parameters .
inserting @xmath63 in eq .
( [ gi ] ) , one can obtain @xmath64 with the following ricci and kretschmann scalars @xmath65 where confirm that the spacetime is asymptotically near to ( a)ds which its curvature scalars are the half of their counterpart in ( a)ds spacetime . one may consider @xmath66 in eq .
( [ gi ] ) to obtain@xmath67 where the ricci and kretschmann scalars are @xmath68 this solution is near to asymptotically ( a)ds spacetime , but as one may find , obtained scalars are different delicately .
now , we investigate the solutions with nonzero @xmath21 and @xmath22 . in order to have a real solution with nonzero @xmath21 or @xmath22
, we should consider @xmath71 or @xmath72 ( in other words : @xmath73 ) .
in general , considering the metric ( 2 ) with @xmath74 , one can show that the curvature scalars diverge at @xmath35 with arbitrary @xmath75 .
in addition , curvature scalars diverge at spatial infinity for @xmath76 . to avoid non - physical singularity ,
we restrict our discussions for @xmath77 and hence we only consider @xmath78 branch ( for a discussion about the essential singularity at infinity , we refer the reader to ref .
@xcite ) .
now , we discuss about some specific values of allowed @xmath6 . at first , we consider the trivial case ( @xmath60 ) and then we investigate the boundary values of @xmath6 and after that we analyze some interesting values of it which lead to integer number for the numerator of eq .
( [ gamma2 ] ) . with trivial rn solution as follows @xmath80 where @xmath81 where we redefined @xmath82 and @xmath83 .
this solution indicates that one may obtain cosmological constant as well as electric charge of maxwell field from pure gravity which has studied before @xcite . for this value of @xmath6 , we encounter with vanishing @xmath85 and @xmath86 , and so the solution is@xmath87 where the second term is dominant for large @xmath58 . calculating
the ricci and kretschmann scalars show that they diverge at the origin and they are finite for @xmath88 .
it is easy to show that the presented solutions may be interpreted as black hole solutions with two horizons , extreme black hole and naked singularity provided the parameters of the solutions are chosen suitably ( see fig .
[ fig3 ] for more details ) . considering @xmath89
, it is easy to show that the mentioned @xmath17 is the same as that in the einstein - power maxwell invariant gravity @xcite with vanishing mass when the nonlinearity parameter is chosen @xmath90 . one may consider another interesting value of @xmath6 to obtain@xmath92where eq .
( [ gi ] ) simplified to@xmath93despite of the first constant term in eq .
( [ gii ] ) , this solution is near to asymptotically ( a)ds charged solution of nonlinear maxwell gravity @xcite with @xmath94 ( for @xmath89 , they are exactly the same ) .
thus , one may conclude that @xmath21 and @xmath22 are related to the mass and charge of the spacetime , respectively .
in other words , we can extract the electric charge and cosmological constant form pure gravity , simultaneously .
one may calculate the ricci and kretschmann scalars to achieve @xmath95equations ( [ rq1])-([rr2q1 ] ) show that there is an essential singularity at the origin and this solution is , asymptotically , near to ( a)ds spacetime . for @xmath96 , and therefore @xmath97 and @xmath98 , eq . ( [ gi ] )
reduces to @xmath99 let us first consider @xmath100 . in this case eq . ( [ giii ] )
may be interpreted as approximately asymptotically ( a)ds charged solution with vanishing mass and therefore @xmath21 is related to electric charge .
in addition , one may consider @xmath101 with nonvanishing @xmath22 to obtain approximately asymptotically ( a)ds solution that for @xmath89 , this solution is the same one obtained in ref .
@xcite when the nonlinearity parameter is equal to @xmath102 .
we can obtain the ricci and kretschmann scalars in the following manner @xmath103 therefore as we expected , there is a curvature singularity at @xmath35 , and asymptotic behavior of the spacetime is near to ( a)ds .
the authors thank an unknown referee for good comments and advices which permitted to improve this paper .
s. h. hendi wishes to thank shiraz university research council .
this work has been supported financially by research institute for astronomy & astrophysics of maragha ( riaam ) , iran . | we consider a class of spherically symmetric spacetime to obtain some interesting solutions in @xmath0 gravity without matter field ( pure gravity ) .
we investigate the geometry of the solutions and find that there is an essential singularity at the origin .
in addition , we show that there is an analogy between obtained solutions with the black holes of einstein-@xmath1-power - maxwell - invariant theory .
furthermore , we find that these solutions are equivalent to the asymptotically lifshitz black holes . also , we calculate @xmath2 to examine the dolgov - kawasaki stability criterion . |
photons at optical frequencies are commonly used in quantum key distribution systems@xcite , and play a fundamental role in proposed devices for quantum information processing@xcite . while the majority of quantum cryptography systems currently in use are based on coherent light sources , most of the proposals for more advanced quantum information devices are based on non - classical states of light , mainly single photon states .
these type of devices include quantum repeaters @xcite , linear optics quantum computing devices@xcite , and quantum networks based on cavity quantum electrodynamics@xcite .
considerable research has been done to make deterministic single photon sources , which can generate indistinguishable single photons with very high efficiency .
one way to generate single photons is by parametric down conversion or attenuation of coherent laser beams such that the probability of having multi - photon states is considerably diminished .
however , this causes increased probability of the vacuum state , so most of the time there are no photons in the light source . to make a good single photon source one has to ensure suppression of both the multi - photon states as well as the zero - photon state
there are several proposals for making on demand single photon sources , like above - band or resonant excitation of solid state single emitters @xcite or adiabatic state transfer in single atoms @xcite . in the recent years
, there has been increased interest in generating non - classical states of light via strongly coupled cavity quantum electrodynamics ( cqed ) systems@xcite .
non - classical states of light that closely resemble the single photon states can be generated in this kind of systems by operating in the photon blockade regime@xcite with a single emitter strongly coupled to an optical resonator .
the photon blockade was first observed in atomic physics @xcite and recently has been demonstrated in solid state cqed using quantum dots in photonic crystal cavities@xcite .
these experiments demonstrated photon anti - bunching caused by photon blockade , a signature of enhanced probability of the single photon states at the output .
there are a few advantages for using photon blockade for single photon generation .
first , the produced single photons are free of jitter .
second , the collection efficiency of the photons is very high due to the presence of a cavity .
however , for on demand single photon generation , the source must be operated in pulsed regime and one photon should be generated for each pulse .
this requirement was not satisfied in any experiments performed so far on photon blockade . in this paper
, we analyze the optimal conditions that must be satisfied by the cqed system for on demand generation of single photon states via photon blockade .
the system considered here is a quantum dot ( qd ) strongly coupled to a photonic crystal cavity , although our conclusion is valid for any cavity qed system .
we analyze two different driving configurations that lead to on demand single photon generation . in section [ sec : pulse_shaping ] we consider the system coherently driven by laser pulses . in section [ sec : stark_shift_operation ] we consider a system driven by a continuous wave laser where the deterministic single photon generation is done by electrical control of the qd resonance frequency .
) such that most of the field coupled into the cavity is emitted at the output port .
( b ) schematic representation of the energy eigenstates for a strongly coupled cavity - emitter system.,width=336 ] the studied optical system consists of a single emitter ( qd ) strongly coupled to a cavity , as shown in fig.[fig_photon_blockade ] . when coherently driven by a laser field , the strongly coupled system is well described by jaynes - cummings hamiltonian : @xmath0\ ] ] where , @xmath1 and @xmath2 are the qd and the cavity resonance frequency respectively and @xmath3 is the driving laser frequency , @xmath4 is the coupling strength of the qd to the cavity mode .
@xmath5 is the rabi frequency of the laser and is given by @xmath6 , where @xmath7 is the dipole moment of the transition being driven by the laser and @xmath8 is the temporal electric field of the laser , inside the cavity ; @xmath9 is the annihilation operator for the cavity mode .
if the excited and ground state of the qd are denoted by @xmath10 and @xmath11 then @xmath12 , @xmath13 . in a frame rotating in laser frequency , under rotating wave approximation ( rwa ) the hamiltonian can be written as : @xmath14\ ] ] where @xmath15 and @xmath16 are the detuning of the qd and the cavity with the laser . to fully describe a real system
, loss must be included in the hamiltonian .
the two main loss mechanisms are the cavity field decay rate @xmath17 ( @xmath18 is the quality factor of the resonator ) and qd spontaneous emission rate @xmath19 . when the coupling strength @xmath4 is greater than the loss rates @xmath20 and @xmath19 , the system is in the strong coupling regime @xcite . in this regime ,
energy eigenstates are grouped in two - level manifolds with eigen - energies given by @xmath21 ( for @xmath22 ) , where @xmath23 is the number of energy quanta in the cavity - qd system .
the eigenstates can be written as : latexmath:[\[\begin{aligned } the splitting between the energy eigenstates in each manifold has a non - linear dependence on @xmath23 .
this anharmonicity in the splitting of the energy eigenstates gives rise to nonlinear optics phenomena at single photon level .
one of these phenomena is photon blockade , where the presence of one photon in the cavity blocks the coupling of subsequent photons .
for example , the system could be driven by a coherent light source ( see fig.[fig_photon_blockade ] ) with frequency resonant with one of the polaritons ( say @xmath25 ) .
once a photon is coupled , the system is excited into the state @xmath25 , so the coupling of another photon with energy @xmath26 would require the system to transition to energy state @xmath27 .
however , the system does not have an eigenstate at this energy , the closest being at @xmath28 .
for this reason , the probability of coupling the second photon is reduced .
a typical experimental configuration for observing photon blockade is depicted in fig.[fig_photon_blockade](a ) where light can be coupled into the cavity via the input port @xmath29 and the output can be collected at the port @xmath30 .
the resonator is represented by two mirrors with reflection coefficients @xmath31 and @xmath32 .
for on demand generation of non - classical states of light , it is desirable that the state is delivered at the output port at times that can be chosen deterministically .
this means that the interaction between the input beam and the cavity / qd system is controlled in time domain .
one way to achieve this is by driving the system with light pulses with controlled shape , as illustrated schematically in fig.[fig_shaping](a ) . in this section
, we investigate how the non - classical state of light observed at the output depends on the properties of the laser pulse at the input . in principle , the input laser pulse can have any specific waveform . to limit the number of free parameters
, we analyze the configuration where the system is driven with gaussian pulses .
in this case , the free parameters are the length , intensity and the center frequency of the pulse . to simulate the system we use quantum trajectory method @xcite .
the evolution of the density matrix @xmath33 of the system , including all the losses is given by ( in the normalized form with @xmath34 ) : @xmath35+\frac{\kappa}{2}(2a\rho a^\dag - a^\dag a \rho-\rho a^\dag a)+\frac{\gamma}{2}(2\sigma\rho \sigma^\dag-\sigma^\dag \sigma \rho-\rho \sigma^\dag \sigma)\ ] ] where @xmath36 is the system hamiltonian without any loss as given in eqn .
[ eqn : h2 ] . for the monte carlo simulation using quantum trajectory method
@xcite the schr@xmath37dinger equation is : @xmath38 where @xmath39 is given by : @xmath40 where , @xmath41 is the system hamiltonian without loss ( eqn .
[ eqn : h2 ] ) and @xmath42 is the collapse operator corresponding to the @xmath43 dissipation channel . in the experiment considered here ,
there are mainly two decay channels : the spontaneous emission of the qd @xmath44 and the cavity decay @xmath45 .
the dephasing rate of the qd and any non - radiative decay is neglected . as the non - classical state is collected from one of the output modes of the cavity ( @xmath30 )
, only the collapse operator corresponding to the cavity decay is monitored .
the driving term @xmath5 in the hamiltonian described in eqn .
[ eqn : h2 ] is assumed to be of the form @xmath46 , where @xmath47 is the amplitude and the time dependence is described by @xmath48 . for the simulation
we assume a gaussian shape for @xmath48 : @xmath49 where @xmath50 is the time when the pulse reaches its maximum value and @xmath51 is the pulse - width .
the non - classical state of light emitted at the output is analyzed using an ideal single photon detector .
for each quantum trajectory , a laser pulse is coupled to the cavity and the number of clicks detected at the output is monitored . ideally ,
for a deterministic single photon source , a single click should be registered by the detector every time the device is operated .
however , the output field is not in a pure single photon state , and in a fock state basis , it can be expressed as : @xmath52 where the @xmath53 is the coefficient of the fock state @xmath54 .
here we write the output state as a pure state considering that the dephasing rate of the system is negligible .
the normalized value ( @xmath55 ) of the coefficients @xmath56 can be estimated from the number of detected photons at the output when running a large number of trajectories .
for example , @xmath57 is well estimated by the relative number of trajectories for which @xmath23 counts were detected at the output .
if the desired output state should be as close as possible to a single photon state , then the simulation parameters should be optimized such that @xmath58 is maximized .
the experimental configuration considered here is as shown in figure [ fig_shaping](a ) .
the cavity has two mirrors , with decay rates @xmath59 and @xmath60 such that @xmath61 .
effectively , the total decay rate of the cavity is @xmath62 .
the driving laser is incident on the mirror with higher reflectivity and the output field is mainly collected from the lossier mirror .
this configuration allows for efficient collection of the non - classical field at the cavity output . to illustrate the behavior of the system operating in photon blockade under pulsed driving , we first analyze a system with parameters @xmath63 , @xmath64 , @xmath65 .
the value @xmath63 corresponds to a cavity with a quality factor of @xmath66 .
this is about 5 times larger than the state of the art values of @xmath18 observed in gaas cavities with coupled inas quantum dots operating around 930 nm , but still withing the theoretical limit for this material @xcite .
regarding the coupling rate @xmath4 , the typical values measured so far are around @xmath67@xcite .
however , with further improvements in the material system and the fabrication techniques it is expected that higher values for @xmath18 and @xmath4 , as considered here , will be achievable . for this simulation ,
the cavity and the quantum dot are assumed to be on resonance ( @xmath68 ) .
the center frequency of the driving field is set on resonance with the transition to the first order manifold ( @xmath69 ) and the pulse width is set to @xmath70 .
figure [ fig_shaping](b ) shows the zero - photon , single photon as well as multi - photon population ( @xmath71 ) as a function of the amplitude of a driving pulse .
the values for the coefficients @xmath55 are inferred from the number of detection events detected over multiple ( 3000 ) quantum trajectories . as shown in fig.[fig_shaping](b ) , the probability of obtaining a single photon at the output has a strong oscillatory dependence on the intensity on the incoming pulse .
when operating in blockade regime , the successful blocking of the second photon depends on how well the first photon is coupled to the qd - cavity system .
one could consider the ground state and the state @xmath72 to constitute an effective two - level system . under pulsed excitation ,
the population in @xmath72 rabi oscillates with the pulse area(fig . [ fig_photon_blockade](b ) ) . for optimum operation
, one would choose the pulse area such that the system transitions completely from @xmath73 to @xmath72 so one photon is coupled into the system and then is released at the output port . at the operation point with maximum single photon state probability , the output field has @xmath74 single photon , @xmath75 vacuum state and @xmath76 multi - photon state character ( i.e. @xmath77 , @xmath78 , @xmath79 ) .
the oscillations of the single photon population in the output field can also be observed when the maximum field intensity is kept constant and the pulse length is changed .
this is shown in fig.[fig_shaping](c ) , where the pulse intensity was chosen as the intensity that gave the maximum single photon output in fig.[fig_shaping](c ) ( @xmath80 ) .
a similar but less prominent effect can be observed in a system with @xmath81 , @xmath82 , parameters that are very close to those already achieved experimentally @xcite . for this set of parameters
the maximum achievable single photon probability is @xmath83 , with @xmath84 vacuum state and @xmath85 multi photon state , as shown in fig.[fig_shaping](d ) .
one of the main advantages of generating non - classical states of light using photon blockade instead of above band excitation and spectral filtering , is the high degree of indistinguishability of the output field .
the two main effects that affect the single photon indistiguishability is the quantum dot jitter and dephasing@xcite .
while jitter is completely avoided using this method , the output field still suffers from the dephasing of the quantum dot while the optical pulse passes through the cavity . to maximize the indistinguishability , it is thus desirable to operate with shorter laser pulses .
in this section we present an alternative method for on demand generation of non - classical light that is based on ultra - fast control of the optical dipole .
the principle is depicted in fig.[electrical_control](a ) , where a continuous wave field is incident on the input port and the frequency of the dipole ( @xmath86 ) is controlled at time - scales comparable to the cavity decay rate .
one method to achieve this kind of ultra - fast control in solid - state systems is by using an electric - field to shift the frequency of a quantum dot via quantum confined stark effect@xcite . in this configuration ,
the cavity frequency is kept constant and the electric field controlls the cavity - quantum dot detuning . with the quantum dot and the cavity on resonance , the input laser
is tuned to the frequency of one of the first order eigenstates ( say @xmath72 ) .
shifting the resonance of the quantum dot causes an energy shift of both first order eigenstates . to generate single photon states
, one should start with the quantum dot detuned from the cavity such that no light is transmitted to the output .
then the quantum dot is brought into resonance with the cavity such that @xmath72 becomes resonant with the input laser beam .
the qd is kept resonant with the cavity so only one photon is coupled and transmitted through the system , and then it is detuned back as shown in fig.[electrical_control](b ) .
the underlying hamiltonian for this system is given in rwa by eq.[eq : hamiltrwa ] : @xmath87 \label{eq : hamiltrwa}\ ] ] in this case , the coherent driving term @xmath88 is kept constant , @xmath89 is the time dependent detuning of the dipole with the cavity , and @xmath90 is the fixed detuning between the cavity and the driving laser . the quantum statistics of the output field as a function of @xmath88 is shown in fig .
[ electrical_control](c - d ) for different parameters of the strongly coupled system .
the quantum dot frequency is detuned by up to @xmath4 , as shown in in fig . [ electrical_control](b ) .
similar to the optical time domain pulse shaping , oscillations are observed in the magnitude of the single photon state for @xmath63 , @xmath65 ( fig.[electrical_control](c ) ) . for optimal operation parameters , @xmath91 , @xmath92 , @xmath93 .
the results for @xmath81 , @xmath82 are shown in fig .
[ electrical_control](d ) , and indicate an @xmath94 single photon , @xmath95 vacuum state , and @xmath96 multi photon fock states at the point of maxium single photon state probability .
in conclusion , we have shown that the non - classical states of light that can be generated using the strong optical nonlinearity of a solid - state cavity qed system ( photon blockade regime ) have a strong dependence on the properties of the input driving field .
the parameters of the driving field were optimized such that the main component of the output field is the single photon state .
we demonstrate that non - classical states that are @xmath97 single photon can be obtained using solid state systems with quality factor of @xmath66 and coupling rate @xmath98 which are realistically achievable .
these types of non - classical states represent a new tool in the developing toolbox of quantum techonlogies , that could be used effectively to improve the performance of quantum cryptography systems and linear optics quantum computation . | the generation of non - classical states of light via photon blockade with time - modulated input is analyzed .
we show that improved single photon statistics can be obtained by adequately choosing the parameters of the driving laser pulses .
an alternative method , where the system is driven via a continuous wave laser and the frequency of the dipole is controlled ( e.g. electrically ) at very fast timescales is presented . |
one of the central issues in the physics of the high - temperature superconductors is the nature of the ground state at low doping . in particular ,
the possible co - existence of ordering tendencies in the spin and charge sectors at low temperature is currently being actively investigated .
@xcite experimentally at low temperature la@xmath2sr@xmath3cuo@xmath4 ( lsco ) co - doped with nd ( lnsco ) exhibits static lattice deformation @xcite as well as incommensurate ( ic ) magnetic order at doping @xmath5 .
the lattice deformation indicates the presence of static charge order ( stripes ) .
recently similar behavior has also been found in la@xmath2ba@xmath3cuo@xmath4 ( lbco ) , @xcite @xmath5 .
however for @xmath6 and in nd - free lsco , charge order has not been detected , while ic magnetism persists down to @xmath7 .
@xcite thus while ic magnetic order is generically observed in the underdoped regime , static charge order seems to be confined to the neighborhood of @xmath8 . ) .
the orthorhombic @xmath9 and @xmath10 directions are shown . ] theoretically the ic magnetism in the cuprates is often modeled as originating from static charge stripes , and since these are not universally present , one is forced to introduce additional concepts , such as fluctuating " stripes , electronic liquid crystals , etc .
@xcite the microscopic origin of such states is still not clear and is currently being debated .
an alternative explanation for the ic magnetism which follows naturally from the @xmath1 model , is the formation of a non - collinear spiral state , partially relieving the frustration due to the hole motion .
@xcite such a state would not co - exist with charge order , as can be shown in the context of effective landau theory .
@xcite however it is still possible that in the presence of anisotropic dzyaloshinski - moriya interactions the spin spiral could cause weak lattice modulation with period generically half that of the spiral .
while the spiral ground state has a tendency to be unstable toward phase separation , we have shown recently that in the extended @xmath11 model the spiral can be stabilized by the presence of the additional hoppings .
@xcite the spiral description was successfully applied to explain magnetic properties of lsco , such as the location of the elastic neutron scattering peaks and the change of the incommensurability direction by 45@xmath12 across the superconductor - insulator boundary ( @xmath13 ) .
@xcite the spiral was also proposed as a candidate for the ground state in the spin - glass ( insulating ) phase for @xmath14 .
@xcite in the present work we address transport properties within the spiral framework .
our main motivation comes from the recent experimental data in the spin - glass phase of lsco showing transport anisotropies as large as 50% to 70% , both in dc and ac measurements .
@xcite it has been argued that these data provide indirect support to the notion of fluctuating stripes or electronic liquid crystals , but no quantitative theory exists that takes these concepts into account .
@xcite recent infrared experiments that carefully identify phonon modes in lsco once again find no charge ordering tendencies in the spin - glass phase of this material .
@xcite therefore we take the point of view that the ground state has a spiral spin structure , and calculate the transport anisotropy in the variable - range hopping ( vrh ) regime for low temperature and frequency , where the anisotropy has the largest value .
we show that the spatial anisotropy of the hole wave - function in a spiral translates into anisotropy of the hopping transport .
our main result is that for microscopic parameters appropriate for lsco ( within the @xmath11 model ) , the magnitude of the anisotropy is large ( 40%-90% , depending on temperature ) , and close to the one found experimentally .
thus we demonstrate that the transport anisotropy data can be explained _ quantitatively _ within the spiral theory which does not involve any tendency of the holes to self - organize into charge stripes .
the anisotropy was also analyzed in ref .
on the basis of topological defect scattering in a spiral ; however the results are applicable to the quasi - metallic ( higher temperature ) regime where the anisotropy is very small , of the order of several percent .
we emphasize that in the present work we consider the low - temperature , strongly - localized regime where it is clear from experiment that the anisotropy is the largest ( 10 - 20 times larger than the value in the metallic regime ) .
finally we mention that the problem was also addressed within the spin - charge separation scenario which provides a description of the pseudogap phase @xcite and could lead to effective insulating " behavior ; however within this approach the transport anisotropy is linked to the magnetic correlation length anisotropy which can be taken from experiment but can not be explicitly calculated .
our starting point is the description of the spin - glass phase ( @xmath14 ) developed in ref . which is briefly summarized below .
first , a single hole resides near the points @xmath15 , with a quadratic dispersion around them @xmath16 , where @xmath17 is perpendicular to the face of the magnetic brillouin zone and @xmath18 is parallel to the face . within the effective @xmath11 model for lsco ,
the parameters are taken to be : @xmath19 , where @xmath20 . from now on we measure energies in units of @xmath21 ( i.e. set @xmath22 ) and lengths in units of the lattice spacing @xmath23 ( we set @xmath24 ) . by using the self - consistent born approximation one finds that these parameters lead to an almost isotropic dispersion @xmath25 , and a quasiparticle residue around the nodal points @xmath26 .
a hole trapped by the coulomb field of the sr ion generates a spiral distortion of the nel background , which can be parametrized as : @xmath27 , @xmath28 , where @xmath29 is an arbitrary unit vector perpendicular to the quantization axis of the states @xmath30 , @xmath31 .
the angle @xmath32 , measuring deviations from collinearity is given by : @xcite @xmath33,\ ] ] where @xmath34 is the spin stiffness , and @xmath35 is the localization length of the orbital part of the wave function .
the value of @xmath36 is extracted from experimental data , @xcite and is generally expected to increase with doping .
the unit vector @xmath37 , in the usual square - lattice coordinate system . at finite doping ( @xmath14 )
the interaction of the long - range dipole distortions from holes trapped at different sr ions leads to spiral magnetic order , @xcite characterized by average @xmath38 , as shown in fig .
[ fig1](a ) . in a perfect square lattice
the spiral can be directed along any diagonal @xmath39 .
however from the location of the elastic neutron scattering peaks @xcite the incommensurability is determined to be along the orthorhombic @xmath10 direction . in our picture
this means that the orthorhombic deformation pins the direction of the spiral , as shown in fig .
[ fig1 ] , and thus we set @xmath40 . in
what follows the exact nature of the pinning mechanism , to be discussed elsewhere , will not be important .
experimentally the magnetic correlation length is finite @xmath41 ( decreases with increasing doping ) .
@xcite on the theoretical side it was argued that @xmath42 is finite due to topological defects that lead to frustration of the long - range spiral order .
@xcite since the localization length is less than the magnetic scale , @xmath43 , we expect that the effective description of transport properties , developed below in terms of the one - hole wave function , is quantitatively valid also at finite ( small ) doping ( i.e. the topological frustration mechanism does not affect our considerations ) .
it should be noted that the above inequality while being explicit is also the strongest ( most restrictive ) condition we could give .
its refinement would have to come out of a detailed theory of ic magnetism in the spin - glass phase ( i.e. a self - consistent theory that takes into account both the effective disorder and the spiral formation , generated by the doped holes ) .
purely theoretical arguments aside , transport measurements in the doping range 0%-4% suggest that the system is in a strongly - localized regime at low temperatures , and we thus expect our calculations to be valid as long as doping is not too close to the insulator - metal boundary ( at 5.5% ) . the rest of the paper is organized as follows . in section
ii we analyze the properties of a localized hole in a spiral background , and in section iii we use the hole s wave function to calculate the in - plane transport anisotropies in the variable - range hopping regime .
section iv contains our conclusions .
the coupling @xmath44 between the spin of the magnetic background ( angle @xmath32 ) and the orbital wave - function of the hole @xmath45 generates the spiral and has the form : @xcite @xmath46 the effective schrdinger equation is then : @xmath47 here the coulomb potential of the @xmath48 ion which keeps the hole localized is @xmath49 , where @xmath50 is the unit charge and @xmath51 is the effective dielectric constant known to be quite large in the copper oxides , @xmath52 ( increases with doping ) .
@xcite ) in terms of the expansion ( [ solution ] ) , for @xmath53 only the first two ( largest ) components are shown .
the normalization of @xmath54 is arbitrary .
inset : the ratio @xmath55 . ]
the last term in ( [ sch ] ) determines the anisotropy of @xmath45 and is given by @xmath56 \ , \nonumber \\ & & g(z ) = 1 - ( 1 + 2z + 2 z^{2 } )
e^{-2z } , \
f(z ) = 2 z^{2 } e^{-2z}. \end{aligned}\ ] ] the coordinate system is chosen so that @xmath57 is parallel to the @xmath10 axis and @xmath58 is the polar angle ( fig . [ fig1 ] ) .
the value of @xmath59 ( and consequently @xmath60 ) can be found by a variational minimization of the total energy : @xcite @xmath61 we will use @xmath62 , consistent with experiment , @xcite rather than rely on ( [ lambda ] ) which contains the uncertainty related to the exact value of the dielectric constant . the dimensionless parameter @xmath63 , as defined in ( [ lambda ] ) , characterizes the coupling between the orbital motion of the hole and the deformation of the spin background . substituting the values of @xmath64 appropriate for lsco ( discussed in section i.a ) , we obtain @xmath65 , which will be used from now on .
let us look for solution of ( [ sch ] ) in terms of the series in angular harmonics @xmath66 then eq .
( [ sch ] ) leads to a set of coupled equations for @xmath67 .
we have solved these equations numerically and the results for the first two harmonics are presented in fig .
the next harmonics are found to be small , for example @xmath68 , and can be in fact safely neglected . at large distances all the wave - functions behave as @xmath69 , while at finite distances this behavior is substantially modified , as shown in fig .
below characteristic temperature @xmath70 , the low - temperature resistivity of lsco in the spin - glass phase can be described by the 2d version of the mott vrh formula @xmath71 .
@xcite here @xmath72 depends somewhat on doping and sample quality and generally decreases when doping increases ( and thus conduction becomes easier ) . as an estimate , for example at 4% doping the data of ref .
are well fit with @xmath73 ; @xcite analyzing the curves of ref .
we obtain @xmath74 , and from ref .
we have extracted @xmath75 . within the spiral framework the physical idea behind
the resistivity anisotropy is quite simple .
according to ( [ solution ] ) the hole wave function acquires an elliptic deformation induced by the spin spiral , as shown schematically in fig .
[ fig1](b ) . hence hopping in the @xmath10-direction is less probable than hopping in the @xmath9-direction . using ( [ solution ] ) , and essentially following the derivation of mott s result , @xcite we obtain for the resistivity anisotropy : @xmath76 here @xmath77 is the vrh length , @xcite and in order for the approach to be justified we must certainly have @xmath78 .
( [ dca ] ) reflects the difference in the wave - function overlap for the two electric field directions .
the factors @xmath79 arise from the fact that the conductivity varies with the square of the carrier jump distance projection in the electric field direction .
@xcite since the hole s dispersion is isotropic ( @xmath80 ) , the exponential part of the wave function is isotropic and consequently the resistivity anisotropy is expected to arise from the anisotropy of the exponential prefactors , whose overlap leads to ( [ dca ] ) .
both the vrh length and @xmath72 thus remain isotropic within this formulation .
it is clear from ( [ dca ] ) that as @xmath81 decreases ( i.e. @xmath82 increases ) the anisotropy grows , due to the increase of @xmath83 ( fig .
[ fig2](inset ) ) and consequently the more pronounced angular dependence of @xmath84 .
the results are summarized in fig .
[ fig3 ] , where also the evolution of the zero temperature anisotropy @xmath85 as a function of @xmath63 is shown for completeness in the inset .
at very low temperature ( @xmath86 , @xcite i.e. @xmath87 ) we would have to take into account a crossover to the coulomb gap regime , which however would practically not influence the curve in fig .
[ fig3 ] .
the data of ref . were taken at temperatures @xmath88 , meaning that the lowest ratio @xmath89 ( we take @xmath90 ) .
as the temperature increases beyond @xmath91 when @xmath92 , the approach based on vrh conduction ceases to be valid , as the conduction mechanism changes to impurity band conduction and eventually quasi - metallic behavior . in the most relevant low - temperature range below @xmath93 ( corresponding to largest anisotropy ) , both the calculated magnitude of @xmath94 and its temperature dependence are very close to the experimental results , @xcite although in these experiments the temperature is not low enough to penetrate the `` deep '' vrh regime @xmath95 where the anisotropy should increase even further .
it should be also noted that there exists quite a bit of uncertainty in the determination of @xmath96 and hence in the determination of the exact value of @xmath94 from fig .
[ fig3 ] . .
inset : maximum ( @xmath97 ) anisotropy versus @xmath63 . ] at finite frequency and temperature the calculation of the ac conductivity is a very complicated problem .
however in the `` quantum '' limit when the frequency @xmath98 , the vrh ac conduction is expected to be dominated by resonant absorption by singly occupied pairs ( without involvement of phonons ) , and is usually relevant in doped semiconductors .
@xcite the vrh length in the ac regime ( neglecting for a moment the angular dependence of the states ) is logarithmic : @xcite @xmath99 , which is the main difference from the dc case . from now on we denote @xmath100 .
the formula for @xmath101 follows from the fact that upon evaluation of the conductivity , the most effective pairs are the ones that satisfy : @xcite @xmath102 , where @xmath103 is the overlap integral , and a typical estimate of the prefactor is @xmath104 .
the functional dependence of the conductivity in 2d at low frequency @xmath105 , in the leading logarithmic approximation , is given by the mott - shklovskii - efros expression : @xmath106 $ ] , where the second term takes into account the coulomb interaction in the resonant pair .
@xcite the generalization of the mott - shklovskii - efros formula to the anisotropic case is straightforward , as it amounts to taking into account the non - exponential ( angle - dependent ) part of the wave - function , leading to the replacement ( with logarithmic accuracy ) : @xmath107 , where @xmath108 is defined in ( [ dca ] ) . taking also into account the expressions for @xmath109 ( [ lambda ] ) , we obtain for the resistivity anisotropy :
@xmath110 the expression ( [ aca ] ) is plotted in fig .
[ fig4 ] . due to the logarithmic dependence , in the theoretical limit of zero frequency the anisotropy would vanish ( albeit very slowly ) , i.e. @xmath111
. however one must keep in mind that the above expressions are not valid at arbitrary low frequencies , as we must have @xmath112 .
generally we find that the ( maximum ) ac anisotropy of 30 - 40% is somewhat smaller than the dc anisotropy . in the regime @xmath113 , and
@xmath114 , for @xmath115 . ] when attempting to compare our results to the experimental data of ref .
we realize that @xmath116 and @xmath81 are not sufficiently low , nor is the difference between them sufficiently high to justify the separation into quantum " and thermal " vrh conduction and consequently the mott - shklovskii - efros approach .
the temperature used is @xmath117 and at the lowest frequency of @xmath118 , estimating for the hole energy @xmath119 , we have @xmath120
. moreover , upon increasing the frequency to @xmath121 , a change of behavior from insulating to conducting takes place , a broad peak develops , @xcite and consequently it is not clear to us that the ac data are ever in a clean vrh regime . in spite of all this the magnitude of the calculated
anisotropy agrees reasonably well with experiment
. it should be experimentally possible to lower the temperature ( towards the mk range ) as well as @xmath116 in order to probe the vrh ac anisotropy more reliably and compare with our theory .
we would like to reiterate that in our theory there is no tendency of the holes to form charge stripe - like structures .
nevertheless the resistivity shows anisotropic behavior ( and so does the uniform magnetic susceptibility , to be discussed separately ) .
it is particularly important , in our view , that the calculations presented in this work produce quantitative results , and we would thus hope that the successes of the present formulation would stimulate further exploration of the spiral and similar scenarios , not involving any charge - ordering tendencies .
the present work is relevant to the spin - glass regime of la@xmath2sr@xmath3cuo@xmath4 , @xmath14 .
however we have also argued @xcite that the spin spiral structure could exist in the superconducting state @xmath122 .
the theory naturally explains why the incommensurate direction , determined from elastic neutron scattering , rotates by 45@xmath123 exactly at the insulator - superconductor transition point .
on the other hand it is usually argued that at @xmath124 both in charge - ordered lnsco @xcite and lbco , @xcite non - collinear spiral order is not consistent with experiment .
we would indeed not claim that the spiral ground state is stable at that particular doping since in fact the spiral becomes commensurate with the lattice and therefore significant changes in the ground state could occur due to the spin - lattice coupling . in conclusion ,
starting from a spiral ground state which unambiguously follows from the extended @xmath1 model , we have calculated the in - plane anisotropy of the low - temperature dc and low - frequency ac resistivity of la@xmath2sr@xmath3cuo@xmath4 in the spin - glass phase .
the theory has no fitting parameters and the calculations are performed for the variable - range hopping regime . within the spiral description
the transport anisotropy is due to the anisotropy of the hole wave - function since the hole hops in a non - collinear ( spiral ) spin background .
the ac anisotropy reasonably agrees with experiment in spite of the fact that the data are on the border of the vrh regime .
the experimental data for the dc resistivity are well within the vrh regime and here both the calculated magnitude and temperature dependence of the anisotropy agree with experiment very well .
s. wakimoto , r. j. birgeneau , m. a. kastner , y. s. lee , r. erwin , p. m. gehring , s. h. lee , m. fujita , k. yamada , y. endoh , k. hirota , and g. shirane , phys . rev .
b * 61 * , 3699 ( 2000 ) ; m. matsuda , m. fujita , k. yamada , r. j. birgeneau , m. a. kastner , h. hiraka , y. endoh , s. wakimoto , and g. shirane , phys . rev .
b * 62 * , 9148 ( 2000 ) . | we study the in - plane resistivity anisotropy in the spin - glass phase of the high-@xmath0 cuprates , on the basis of holes moving in a spiral spin background .
this picture follows from analysis of the extended @xmath1 model with coulomb impurities . in the variable - range hopping regime
the resistivity anisotropy is found to have a maximum value of around 90% , and it decreases with temperature , in excellent agreement with experiments in la@xmath2sr@xmath3cuo@xmath4 . in our approach
the transport anisotropy is due to the non - collinearity of the spiral spin state , rather than an intrinsic tendency of the charges to self - organize . |
flavor - changing neutral current ( fcnc ) decays are forbidden at tree level in the standard model ( sm ) .
they can proceed through loop processes an example is displayed in fig .
[ fig : feyndiag ] at a considerably suppressed rate .
the modes under study have additional helicity and gim suppression , and have sm branching fractions beyond the reach of current experiments .
the largest predicted branching fraction is about @xmath4 for the decay @xmath5 , the others being more than an order of magnitude smaller .
a primary interest in these modes is as probes of new physics ( np ) .
np can enter either at tree level , or through loop processes , and can increase the branching fraction by orders of magnitude .
these modes provide the best limits on a number of np scenarios . lepton - flavor violating ( lfv )
decays are strictly forbidden in the sm , so observation is a clear sign of np .
( the lfv in neutrino oscillations plays an insignificant role in the decays studied here . )
cdf searched for the decays @xmath0 , @xmath1 , @xmath3 , and @xmath2 .
( inclusion of change conjugate modes is assumed throughout . )
the searches share many features .
the search is performed relative to a normalization mode chosen for its similarity to the search mode .
this avoids the difficulty of determining absolute efficiencies , and largely cancels many systematics .
the selection criteria for the search mode are optimized in an unbiased manner , with the events in the search window blinded .
a monte carlo with a detailed detector simulation is used to simulate signal events .
background is estimated from sideband data , data before lepton identification , and monte carlo when appropriate .
the production , acceptance , and reconstruction efficiency of the search mode are determined relative to the normalization mode . with the selection optimized , the data is unblinded .
none of the investigated decays showed a significant signal after unblinding , so branching fraction limits are determined , for example , as : @xmath6 where the superscript @xmath7 indicates that the corresponding quantity is a credibility level limit , @xmath8 is the number of decays observed to final state @xmath9 , the ratio of @xmath10 s is the ratio of acceptance and efficiency between the search and normalization final states , @xmath11 is the branching fraction of @xmath12 to final state @xmath9 , and , because the search mode is for @xmath13 while the normalization mode is for @xmath14 , the ratio of fragmentation fractions @xmath15 is needed .
equation [ eq : limits ] is for @xmath16 , and the obvious substitutions yield the expressions for the other modes . in the following section we present some of the specific challenges of each analysis ,
then summarize the results and discuss limits on np that can be derived .
the @xmath0 analysis for @xmath17 of integrated lumninosity is published @xcite and will be omitted in this proceeding in the interest of devoting more space to the unpublished results .
the limits are displayed in table [ limits ] .
the @xmath18 and @xmath19 decays are normalized to the charmless @xmath20 decay @xcite .
the number of @xmath21 decays is extracted from data collected by a two - track , separated vertex trigger ( ttt ) from @xmath17 of integrated luminosity .
this trigger selects events with two oppositely charged tracks with transverse momenta above @xmath22 , impact parameters between @xmath23 and @xmath24 to the beam spot , and a transverse decay length greater than @xmath25 .
the decays @xmath26 reconstructed from the trigger track - pairs are indistinguishable in the invariant mass plot ( fig .
[ fig : mpipi ] ) , and are separated on a statistical basis as described in @xcite .
the @xmath27 and @xmath28 decays are extracted from this same data by reconstructing the trigger track - pairs as either @xmath29 or @xmath30 and then requiring the tracks be matched to hits in the muon detectors , or to an energy cluster in the electromagnetic calorimeter .
additional background suppression for electrons is provided by requiring an ionization energy loss , @xmath31 , in the outer tracker inconsistent with hadrons . for electrons ,
bremstrahlung in the detector can result in significant energy loss producing a tail in the reconstructed invariant mass distribution . to maintain good acceptance ,
a generous search window is used , @xmath32 around the nominal @xmath13 or @xmath14 mass for the @xmath27 modes , and @xmath33 for the @xmath28 modes .
the backgrounds are primarily combinatorial and from misidentified @xmath34 decays .
the unblinded mass plots ( fig .
[ fig : beeemumass ] ) reveal two events in the @xmath35 and @xmath30 mass windows , one event in the @xmath36 and @xmath30 mass windows .
the @xmath37 and @xmath38 bayesian credibility level limits determined for these modes is displayed in table [ limits ] .
the @xmath2 search is normalized to the cabibbo suppressed @xmath39 decay @xcite .
two body @xmath40 decays are reconstructed from the trigger tracks in events selected by the ttt from @xmath41 of integrated luminosity .
combinatorial background to the normalization and search channels is greatly reduced by requiring an associated pion from the decay chain @xmath42 .
the small mass difference between pions and muons leads to invariant mass distributions that overlap significantly as shown in fig .
[ fig : dmmmass ] .
we identify a track as a muon based on hits in either the central muon detector ( cmu ) or the muon extension ( cmx ) , as determined by projecting the track from the central outer tracker to the muon detectors .
a careful study of background sources revealed that a significant background exists from pairs of real muons coming from the sequential decay chain @xmath43 and @xmath44 .
this background is suppressed using a likelihood ratio based on decay length , decay length significance , and impact parameter of the candidate @xmath40 to the primary vertex . when the data is unblinded we observe three central - central events , zero central - extension events , and one extension - extension event , consistent with the background estimate .
the derived @xmath37 and @xmath38 bayesian credibility limits are displayed in table [ limits ] .
the limits from these analyses are displayed in table [ limits ] , along with the previous best published limit and the experiment that produced it . from these limits
we derive limits on np models .
the @xmath5 limit is used to restrict the allowed parameter space for supersymmetry models .
the @xmath18 limits have been used to derive mass limits on pati - salam leptoquarks @xcite ( fig .
[ fig : lqmasslimits ] ) .
the mass limits are in the range of @xmath45 for both decays . | decays that are highly suppressed in the standard model are excellent places to search for effects of new physics .
decays mediated by flavor - changing neutral currents are forbidden at tree level in the sm , and are often further suppressed by helicity and the glashow - iliopoulos - maiani mechanism .
exclusive final states with charged lepton pairs are a particular strength of the cdf experiment due to the large bottom and charm production cross section and the ability to efficiently separate signal from background . cdf has searched for and set the world s best limits on the rare flavor - changing neutral current decays @xmath0 , @xmath1 , and @xmath2 , and the lepton flavor violating decay @xmath3 . |
as the atmospheres of more exoplanets are being characterized , astronomers are finding that a non - negligible fraction of them are cloudy ( e.g. , @xcite ) .
cloudy atmospheres are challenging to decipher , because the interpretation of their chemical abundances is degenerate with the degree of cloudiness ( e.g. , @xcite ) .
it motivates the construction of dimensionless indices for quantifying the degree of cloudiness of the atmosphere of a transiting exoplanet .
if trends between these indices and the properties of the exoplanet or its star are found , then one may select a sub - sample of cloudfree objects for more indepth atmospheric characterization using future facilities such as the james webb space telescope .
essentially , any trends found allow us to perform triage on the cloudy objects .
@xcite has previously defined a cloudiness index based on near - infrared diagnostics : the strength of the water feature at about 1.4 @xmath2 m , as probed by wfc3 on the hubble space telescope , and the j band .
the study of @xcite found that exoplanets with equilibrium temperatures higher than 700 k and surface gravities greater than @xmath3 ( cgs units ) are more likely to be cloudfree .
@xcite found a correlation between the strength of the 1.4 @xmath2 m water feature and the difference in transit radii between the near- and mid - infrared wavelengths .
the approach of @xcite is based on using one - dimensional , self - consistent model atmospheres in chemical and radiative equilibrium , along with an assumed value for the metallicity , as a baseline for defining what cloudfree " means . in the current letter
, we take a different and complementary approach by constructing a cloudiness index based on the notion that , in a cloudfree atmosphere , the difference in transit radii between the line center and wing of the sodium or potassium lines should be straightforwardly calculable .
a cloudy atmosphere would have a value of this difference in transit radii that is _ less _ than the cloudfree value @xcite , thereby naturally allowing us to define a cloudiness index .
we apply our cloudiness index to the sample of exoplanets presented in @xcite and report a tentative trend between cloudiness at visible wavelengths and the strength of stellar irradiation .
before constructing our cloudiness index , we examine in detail if measurements of the spectral slope in the visible range of wavelengths may be used , in isolation , to quantify the degree of cloudiness of an atmosphere .
@xcite have previously published , in their appendix , approximate algebraic solutions to infer the mean molecular mass of an atmosphere based on a pair of transit measurements at different wavelengths .
we wish to point out that these formulae have an earlier origin and also elucidate the assumptions behind them . in work preceding @xcite , @xcite have shown that , in an isothermal atmosphere , the temperature may be inferred from @xmath4 where @xmath5 is the mean molecular mass , @xmath6 is the surface gravity , @xmath7 is the boltzmann constant , @xmath8 is the transit radius and @xmath9 is the cross section for absorption or scattering .
equation ( [ eq : isotemp ] ) may be trivially re - arranged to yield equation ( a5 ) of @xcite , @xmath10^{-1}.\ ] ] we now assume that the cross section is associated with scattering by _ both _ molecules and aerosols / condensates , and that it has a power - law functional form , @xmath11 where @xmath12 denotes the wavelength and @xmath13 is a dimensionless index .
it immediately follows that @xmath14 independent of @xmath15 and @xmath16 .
physically , it implies that measuring the spectral slope alone does not allow us to distinguish between scattering by molecules or aerosols , since the preceding expression has no dependence on either @xmath15 or @xmath16 .
equation ( [ eq : isotemp ] ) becomes @xcite @xmath17 where @xmath18 and @xmath19 correspond to the transit radii at the wavelengths @xmath20 and @xmath21 , respectively .
the approximate expression in the preceding equation , which is equation ( a6 ) of @xcite , derives from assuming that @xmath22 for rayleigh scattering .
it is important to note that only @xmath23 is an observable , which allows @xmath24 to be inferred .
it does not allow one to ascertain if rayleigh scattering is at work , i.e. , that @xmath22 .
this is an assumption .
if we are associating the rayleigh slope with aerosols / condensates , then we are assuming that the particles have a radius of @xmath25 . by further manipulating equation ( [ eq : scaleheight ] ) , we obtain equation ( a7 ) of @xcite
, @xmath26 the preceding expression carries an additional assumption : the stellar radius ( @xmath27 ) is assumed to be the same at both wavelengths . in summary , @xcite and @xcite
are using the same algebraic formula to essentially infer the isothermal pressure scale height by assuming that rayleigh scattering is at work ( @xmath22 ) .
@xcite then infer the temperature by assuming a value for the mean molecular mass is the mass of the hydrogen atom , then the mean molecular weight is given by @xmath28 . ] .
@xcite infer the mean molecular mass by assuming a value for the temperature . from measuring the spectral slope alone , one can not distinguish between rayleigh scattering by molecules ( e.g. , hydrogen , nitrogen ) or aerosols / condensates . [ tab : data ] [ cols="<,^,^,^,^,^,^,^,^",options="header " , ] +
we have demonstrated that measuring the spectral slope alone does not allow us to infer the mean molecular mass or the degree of cloudiness in an atmosphere .
next , we proceed to construct a cloudiness index that does not depend on the spectral slope .
our cloudiness index is meaningfully defined only if there is a detection of the sodium or potassium line .
we first wish to establish that , when the line is undetected , it is most probably not due to it being dominated by rayleigh scattering associated with hydrogen molecules .
rather , the absence of a sodium or potassium line is either due to a vanishingly low abundance of sodium or potassium or a very cloudy atmosphere . since there is no way to distinguish between these scenarios without additional information , we will focus on applying our cloudiness index only to objects with reported sodium / potassium line detections . for the sodium or potassium line center to be unaffected by rayleigh scattering associated with hydrogen molecules , the relative abundance of sodium or potassium ( to h@xmath29 ) , by number , has to greatly exceed a threshold value , @xmath30 the cross section for rayleigh scattering by molecules is @xcite , @xmath31 where @xmath32 is a reference number density
, @xmath33 is the real part of the index of refraction and @xmath34 is the king factor . for molecular hydrogen ,
we set @xmath35 and @xmath36 @xmath1 and take the refractive index to be @xcite @xmath37 + 1.\ ] ] for the sodium / potassium cross sections at line center ( @xmath38 ) , we take equation ( 14 ) of @xcite .
the condition in equation ( [ eq : condition ] ) becomes @xmath39 for both the sodium and potassium doublets .
given that the elemental abundances of sodium and potassium in the solar photosphere are @xmath40 and @xmath41 @xcite , respectively , this condition is expected to possess some generality . for atmospheres in which this condition holds , the sodium / potassium line center ( and the wavelengths near it , in the line wings )
is unaffected by rayleigh scattering .
other caveats include the assumption of an isothermal atmosphere in hydrostatic equilibrium and our neglect of non - local thermodynamic equilibrium ( non - lte ) effects @xcite . for cloudfree atmospheres
, we may directly infer the isothermal pressure scale height using equation ( [ eq : scaleheight ] ) . for cloudy atmospheres
, this approach is invalid .
equation ( [ eq : scaleheight ] ) is robust in the sense that it is based on measuring differences in the transit radii and wavelengths , but this also makes it blind to whether rayleigh scattering is associated with molecules or aerosols .
are taken from the analysis of the sodium lines .
the labels w6 " , w17 " , w31 " , w39 " , h1 " ,
h12 " and hd189 " refer to wasp-6b , wasp-17b , wasp-31b , wasp-39b , hat - p-1b , hat - p-12b and hd 189733b , respectively .
the spearman rank coefficient for @xmath42 versus @xmath43 , @xmath6 and @xmath44 are @xmath45 and @xmath46 , respectively .
the corresponding @xmath47-values are @xmath48 and @xmath49 , where the null hypothesis is that the quantities are uncorrelated .
a least - squares linear fit to @xmath50 yields coefficients of @xmath51 k@xmath52 and @xmath53 for the linear and constant components , respectively .
the thin dotted lines are 100 monte carlo realizations of the linear fit using these coefficients.,title="fig : " ] are taken from the analysis of the sodium lines
. the labels w6 " , w17 " , w31 " , w39 " ,
h1 " , h12 " and hd189 " refer to wasp-6b , wasp-17b , wasp-31b , wasp-39b , hat - p-1b , hat - p-12b and hd 189733b , respectively
. the spearman rank coefficient for @xmath42 versus @xmath43 , @xmath6 and @xmath44 are @xmath45 and @xmath46 , respectively .
the corresponding @xmath47-values are @xmath48 and @xmath49 , where the null hypothesis is that the quantities are uncorrelated .
a least - squares linear fit to @xmath50 yields coefficients of @xmath51 k@xmath52 and @xmath53 for the linear and constant components , respectively .
the thin dotted lines are 100 monte carlo realizations of the linear fit using these coefficients.,title="fig : " ] are taken from the analysis of the sodium lines . the labels w6 " ,
w17 " , w31 " , w39 " , h1 " , h12 " and hd189 " refer to wasp-6b , wasp-17b , wasp-31b , wasp-39b , hat - p-1b , hat - p-12b and hd 189733b , respectively . the spearman rank coefficient for @xmath42 versus @xmath43 , @xmath6 and @xmath44 are @xmath45 and @xmath46 , respectively .
the corresponding @xmath47-values are @xmath48 and @xmath49 , where the null hypothesis is that the quantities are uncorrelated .
a least - squares linear fit to @xmath50 yields coefficients of @xmath51 k@xmath52 and @xmath53 for the linear and constant components , respectively .
the thin dotted lines are 100 monte carlo realizations of the linear fit using these coefficients.,title="fig : " ] in a cloudfree atmosphere , the difference in transit radii , between the line center and wing of the sodium or potassium line , should take on a specific value .
we denote this quantity by @xmath54 . using equations ( 16 ) and ( 17 ) of @xcite
, we obtain @xmath55},\ ] ] based on the reasoning that the chord optical depths associated with the transits at line center and wing have the same value .
the line - center wavelength is denoted by @xmath56 . the line profile , in the line wings , is well approximated by a lorentz profile @xcite , @xmath57 where the einstein a - coefficient is @xmath58 and the speed of light is @xmath59 . for the sodium d@xmath60 and d@xmath29 lines , we have @xmath61 s@xmath52 and @xmath62 s@xmath52 , respectively , corresponding to @xmath63 @xmath2 m and 0.5891582 @xmath2 m @xcite . for the potassium d@xmath60 and d@xmath29 lines
, we have @xmath64 s@xmath52 and @xmath65 s@xmath52 , respectively , corresponding to @xmath66 @xmath2 m and 0.766701 @xmath2 m @xcite . for both the sodium and potassium doublets
, we estimate that @xmath67 with a gentle dependence on temperature .
it implies that the isothermal pressure scale height need not be known precisely to compute @xmath68 . in other words ,
the uncertainty on @xmath54 is linearly proportional to the uncertainty on @xmath69 .
a possible concern is that we have not taken pressure broadening of the line into account , which may render our calculations inaccurate . by using the expression for
the chord optical depth @xcite and the ideal gas law , one may obtain an expression for the pressure probed during a transit , @xmath70 the preceding expression is essentially the photospheric pressure with a correction term for transit geometry . here , @xmath71 is interpreted as being the mean opacity of the atmosphere . @xcite
have shown that the rosseland and planck mean opacities are @xmath72 @xmath73 g@xmath52 and @xmath74 @xmath73 g@xmath52 , respectively .
if we adopt typical numbers ( @xmath75 cm s@xmath76 , @xmath77 km and @xmath78 , where @xmath79 cm is the radius of jupiter ) , then we obtain @xmath80@xmath81 mbar .
we expect the sodium and potassium lines to probe pressures that are similar .
@xcite have previously shown that pressure broadening of the sodium lines is only important for @xmath82 bar .
therefore , we conclude that pressure broadening is not a concern . by measuring the actual difference in transit radii between the line center and wing ( @xmath83 ) , we may construct an index for the degree of cloudiness in the atmosphere , @xmath84 completely cloudfree atmospheres have @xmath85 .
atmospheres remain nearly cloudfree when @xmath86 .
cloudy atmospheres have @xmath87 .
we estimate the uncertainty associated with @xmath42 , denoted by @xmath88 , using @xmath89 where the uncertainties associated with the transit radii at line center ( @xmath90 ) and wing ( @xmath8 ) are denoted by @xmath91 and @xmath92 , respectively .
we do not include the uncertainties on the stellar radius . as a bonus
, one may directly infer the _ absolute _ abundance of sodium or potassium associated with the line center @xcite , @xmath93 where @xmath90 is the transit radius at line center , @xmath94 is the mass of the electron , @xmath95 is the elementary charge and @xmath96 is the oscillator strength . for the sodium d@xmath60 and d@xmath29 lines , we have @xmath97 and 0.641 , respectively @xcite . for the potassium d@xmath60 and d@xmath29 lines
, we have @xmath98 and 0.682 , respectively @xcite .
the preceding expression assumes a cloudfree atmosphere . in a cloudy atmosphere
, it would yield an upper limit for the abundance of sodium or potassium .
to illustrate the usefulness of the cloudiness index , we estimate its value for a sample of hot saturns / jupiters taken from @xcite .
table 1 lists the data gleaned from @xcite .
the stellar radii are taken from @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and @xcite .
we augment the hubble space telescope data of @xcite with ground - based data of the potassium line from @xcite for hat - p-1b .
for the transit radius in the sodium line wing , we choose to extract the data point immediately redward ( instead of blueward ) of line center to minimize the effects of rayleigh scattering by molecules .
for the potassium line wing , we extract the data point immediately blueward . to begin our analysis requires that we have knowledge of the isothermal pressure scale height .
but as we have already demonstrated , this can not be reliably inferred in a cloudy atmosphere from measuring the spectral slope . instead , we begin by computing the isothermal pressure scale height assuming that the temperature is the equilibrium temperature ( @xmath99 ) , @xmath100 which
is then used to compute @xmath54 and @xmath42 .
we assume @xmath101 , where @xmath102 is the mass of the hydrogen atom .
in other words , we are assuming that @xmath103 , following @xcite and @xcite .
we set the error or uncertainty associated with this assumption to be @xmath104 ( see [ subsect : herror ] ) .
table 2 shows our estimates for @xmath42 , @xmath105 and @xmath106 .
we subscript @xmath42 with either na " or k " , depending on whether it was derived using the sodium or potassium lines generally , there is consistency between them .
we have taken the average of the values of @xmath58 and @xmath96 .
following @xcite , we omit wasp-12b , wasp-19b and wasp-31b from the analysis for sodium .
we also exclude hd 209458b as there are no data points that precisely align with the sodium line centers .
however , we include wasp-6b and hat - p-12b . for the potassium line ,
we analyze the data of wasp-6b , wasp-31b , wasp-39b and hat - p-12b , following @xcite .
we also use the ground - based data of @xcite for hat - p-1b .
we consistently obtain @xmath107 @xmath1 .
only for the nearly cloudfree wasp-17b , wasp-31b and hat - p-1b are the values of @xmath105 and @xmath106 actual estimates for the absolute abundances of sodium and potassium . for the other objects , they are upper limits as their atmospheres are cloudy .
note that we can not estimate the mixing ratios ( relative abundances ) based on analyzing the sodium or potassium lines alone , because we do not have an independent estimate of the _ total _ pressure being sensed .
if the atmosphere is hydrogen - dominated ( which is our assumption ) , then this would be the pressure associated with molecular hydrogen . generally , the temperature being sensed is not the equilibrium temperature , implying that there is an error associated with assuming @xmath103 .
we can estimate what this error is by calculating the true pressure scale height ( @xmath69 ) for the nearly cloudfree objects using equation ( [ eq : scaleheight ] ) and comparing it to @xmath108 .
specifically , we may perform this analysis for wasp-17b , wasp-31b and hat - p-1b , since they have @xmath86 .
we note that the claim for hat - p-1b being cloudfree is consistent with the work of @xcite . by specializing to @xmath22 , equation ( [ eq : scaleheight ] )
tells us that @xmath109 we may directly estimate the value of @xmath110 by performing a linear fit to the spectral slopes , in the visible range of wavelengths , measured by @xcite for wasp-17b , wasp-31b and hat - p-1b .
we use the data points from the bluest available wavelength up to the data point just blueward of the peak of the sodium line . for wasp-17b , we have @xmath111 km .
our linear fit and use of equation ( [ eq : scaleheight2 ] ) yields @xmath112 km , which translates into an error of @xmath113 . for wasp-31b
, we obtain @xmath114 km , @xmath115 km and @xmath116 . for hat - p-1b
, we obtain @xmath117 km , @xmath118 km and @xmath119 .
if we take the average of these three values , we obtain @xmath120 .
this is the value we assume when using equation ( [ eq : cerror ] ) .
figure [ fig : trends ] plots @xmath42 versus the equilibrium temperature , surface gravity and mass ( @xmath44 ) of the exoplanets .
curiously , there seems to be a tentative trend of increasing @xmath42 with decreasing @xmath43 , which is a proxy for the incident stellar flux .
the uncertainties associated with @xmath42 are larger for cooler objects , because of the larger uncertainties on the transit radii .
by contrast , it has previously been shown that the measured geometric albedos of hot jupiters exhibit no trend with equilibrium temperature , surface gravity or stellar metallicity , with kepler-7b being the oddball of having an unusually high geometric albedo @xcite .
physically , if this trend is real , it implies that clouds consisting of sub - micron - sized particles become less prevalent as the atmosphere becomes more irradiated .
any trend of @xmath42 with @xmath6 or @xmath44 is less apparent . for 5 out of 7 objects
, we may directly compare our values of @xmath42 with the values of the near - infrared index of @xcite , which is denoted by ( h@xmath29o@xmath121j)/@xmath108 .
@xcite regards index values of ( h@xmath29o@xmath121j)/@xmath122 to correspond to cloudfree atmospheres , while index values between 1 and 2 correspond to partially cloudy to nearly cloudfree atmospheres .
cloudy atmospheres have ( h@xmath29o@xmath121j)/@xmath123 .
one would expect that atmospheres with @xmath86 would also have ( h@xmath29o@xmath121j)/@xmath124 . within the uncertainties associated with both indices ,
hat - p-1b fulfills the designation of being nearly cloudfree .
based on both indices , hat - p-12b is cloudy .
however , the comparison becomes less clear for wasp-17b , wasp-31b and hd 189733b .
it is conceivable that the visible and near - infrared wavelengths are probing different atmospheric layers and one layer being cloudfree does not preclude the other being cloudy .
future work with a larger sample , where both indices may be calculated , is needed to shed light on these discrepancies .
we have developed a dimensionless index to quantify the degree of cloudiness of an atmosphere , based on measuring the transit radii at the line center and wing of the sodium and/or potassium lines .
the value of this index has a lower limit of 1 , which corresponds to a completely cloudfree atmosphere .
larger values of the index correspond to cloudier atmospheres .
physically , our index measures the influence of clouds consisting of sub - micron - sized particles and is complementary to the near - infrared index of @xcite , which probes somewhat larger particles . we have computed the index for a small sample of 7 hot saturns / jupiters taken from @xcite .
we find a tentative trend of decreasing cloudiness with increasing equilibrium temperature ( figure [ fig : trends ] ) .
future work should measure the sodium and potassium lines at even higher resolutions for a larger sample of exoplanetary atmospheres , in order to confirm or refute this trend .
if the trend holds , it will allow us to screen for a sub - sample of cloudfree objects for further scrutiny by the james webb space telescope , via reconnaissance of a large sample of objects using ground - based measurements .
i acknowledge financial support from the swiss national science foundation , the planets nccr ( national center of competence in research ) framework and the swiss - based merac foundation .
i thank the referee for a constructive report . | we present a dimensionless index that quantifies the degree of cloudiness of the atmosphere of a transiting exoplanet .
our cloudiness index is based on measuring the transit radii associated with the line center and wing of the sodium or potassium line . in deriving this index
, we revisited the algebraic formulae for inferring the isothermal pressure scale height from transit measurements .
we demonstrate that the formulae of lecavelier et al . and
benneke & seager are identical : the former is inferring the temperature while assuming a value for the mean molecular mass and the latter is inferring the mean molecular mass while assuming a value for the temperature .
more importantly , these formulae can not be used to distinguish between cloudy and cloudfree atmospheres .
we derive values of our cloudiness index for a small sample of 7 hot saturns / jupiters taken from sing et al .
we show that wasp-17b , wasp-31b and hat - p-1b are nearly cloudfree at visible wavelengths .
we find the tentative trend that more irradiated atmospheres tend to have less clouds consisting of sub - micron - sized particles .
we also derive absolute sodium and/or potassium abundances @xmath0 @xmath1 for wasp-17b , wasp-31b and hat - p-1b ( and upper limits for the other objects ) .
higher - resolution measurements of both the sodium and potassium lines , for a larger sample of exoplanetary atmospheres , are needed to confirm or refute this trend . |
the first celestial maps in x - rays , in the early 1970s , show that globular clusters harbour more x - ray sources than one would expect from their mass . as a solution to this puzzle
it was suggested that these bright ( @xmath0erg / s ) x - ray sources , binaries in which a neutron star captures mass from a companion star , are formed in close stellar encounters .
a neutron star can be caught by a companion in a tidal capture , or it can take the place of a star in a pre - existing binary in an exchange encounter .
verbunt & hut ( 1987 ) showed that the probability of a cluster to harbour a bright x - ray source indeed scales with the number of stellar encounters occurring in it ; whereas a scaling with mass does not explain the observations . with the _ einstein _
satellite a dozen less luminous ( @xmath1erg / s ) x - ray sources were discovered in the early 1980 .
_ rosat _ enlarged this number to some 55 , and now thanks to _ chandra _ we know hundreds of dim x - ray sources in globular clusters .
the nature and origin of these dim sources is varied .
those containing neutron stars , i.e. the quiescent low - mass x - ray binaries in which a neutron star accretes mass from its companion at a low rate and the recycled or millisecond radio pulsars , have all formed in processes involving close stellar encounters .
the magnetically active binaries , on the other hand , are most likely primordial binaries , with stars that are kept in rapid rotation via tidal interaction .
cataclysmic variables are binaries in which a white dwarf accretes matter from a companion . in globular clusters
they may arise either via stellar encounters , or from primordial binaries through ordinary binary evolution this is expected to depend on the mass and density of the globular cluster .
in this paper we describe the classification and identification of the dim sources in section2 , and make some remarks on the theory of their formation in section3 . in section4
we will discuss a new , and in our view more accurate , way to compare the numbers of these sources with theoretical predictions .
work on the dim sources is progressing along various lines .
grindlay and coworkers study one cluster , 47tuc , in great detail ( grindlay et al
. 2001 , edmonds et al .
2003 , heinke et al .
webb and coworkers use xmm to obtain high - quality x - ray spectra ( e.g. webb et al . 2006 , servillat this meeting ) .
dim sources are also found in clusters in which individual sources are the main target , such as terzan1 and 5 , and m28 ( wijnands et al .
2002 , heinke et al .
2003 , becker et al .
lewin initiated a large program to observe clusters with very different central densities and core radii , and thereby to provide material for tests on the dependence on these properties of the numbers of dim sources .
further references to all this work may be found in the review by verbunt & lewin ( 2006 ) ; and in the remainder of this section .
the first classification of the dim sources may be made on the basis of the * x - ray properties only * ( fig.[xcol ] ) .
the brightest sources in the 0.5 - 2.5kev band , at @xmath2 erg / s , tend to be quiescent low - mass x - ray binaries .
to better use the chandra range , one may also select the brightest sources in the 0.5 - 6.0kev band , and select soft sources , with a high ratio of fluxes below and above e.g. 2kev : @xmath3 .
such sources also are mostly quiescent low - mass x - ray binaries . between @xmath4 and @xmath5
erg / s most sources are cataclysmic variables , especially when they have hard spectra @xmath6 .
the faintest sources include magnetically active binaries , often with soft x - ray spectra . for many faint sources
the number of counts is too low to decide on the hardness of the spectrum .
the second step in classification can be made when * identification * with a source * at other wavelengths * is made .
positional coincidence of an x - ray source with the accurate radio position of a millisecond pulsar provides a reliable identification and classification .
positional coincidence with optical sources is only significant if the highest possible astrometric accuracy is used to limit the number of possible counterparts ( e.g. bassa et al .
the position of these possible counterparts in a colour - magnitude diagram is then used to select the probable counterparts .
cataclysmic variables are bluer than the main sequence stars , and magnetically active binaries may lie above the main sequence .
systems on the main sequence can not be unambiguously classified : they may either be cataclysmic variables in which the optical flux is dominated by the donor star , or main - sequence binaries with unequal masses whose optical light is dominated by the brighter star . if a periodicity is found in the x - rays that corresponds to a period at another wavelength , e.g. the pulse period of a pulsar or the orbital period of a binary , identification and clasification are secured simultaneously ( e.g. ferraro et al .
a very useful discriminant in x - ray astronomy in general is the x - ray to optical flux ratio . in the case of globular clusters
we can use the known distance to determine the * optical to x - ray luminosity ratio * ( fig[lxmv ] ) . on the basis of in particular the extensive data on 47tuc ( edmonds et al .
2003 and references therein ) one finds that the lines of constant optical to x - ray luminosity ratio @xmath7 separates the quiescent low - mass x - ray binaries above it from the cataclysmic variables below .
the line @xmath8 roughly separates the cataclysmic variables from the magnetically active binaries .
this latter separatrix leads to a surprise when one compares it with the x - ray luminosities of nearby stars and of known magnetically active binaries , i.e. rs cvn systems , near the sun . for main - sequence stars in the solar neighbourhood
, the x - ray luminosity increases with the rotation speed , up to an upper bound given approximately by @xmath9 as illustrated in fig.[lxmv ] ( left ) . this bound is lower than the separatrix given by eq.[sepb ] , especially for brighter stars
this would imply that active binaries in globular clusters can have higher x - ray luminosities than similar binaries near the sun .
we suggest , however , that the classification must be reinvestigated , and that some of these objects are cataclysmic variables .
the absence of the blue colour expected for a cataclysmic variable ( see fig.[xcol ] ) then requires explanation
e.g. as a consequence of the non - simultaneous measurements at different colours combined with source variability . and @xmath10 erg / s
, it actually lies below the line given by eq.[lxmax ] . ]
binaries in a globular clusters change due to their internal evolution and/or due to external encounters . to describe the current cluster binary population one must track the events for each primordial binary and for each binary that is newly formed via tidal capture , throughout the cluster .
the first estimates of the formation of binaries with a neutron star necessarily made a number of drastic simplifications .
the sum of all encounters ( of a neutron star with a single star , or with a binary ) was replaced with an integral over the cluster volume of the encounter rate per unit volume .
four assumptions followed : the number density @xmath11 of each participant in the encounter scales with the total mass density @xmath12 , the relative velocity between the encounter participants scales with the velocity dispersion @xmath13 , the interaction cross section @xmath14 is dominated by gravitational focussing so that @xmath15 , and the encounter rate is dominated by the encounters in the dense cluster core . hence one writes the cluster encounter rate @xmath16 as @xmath17 where @xmath18 is the central density and @xmath19 the core radius . if one further eliminates the velocity dispersion through the virial theorem , @xmath20 , one has @xmath21 where @xmath22 is referred to as the collision number . with a life time @xmath23 the expected number of binaries of a given type is @xmath24 a major advantage of these simple estimates is the clear connection between ( the uncertainty in ) the input and ( the uncertainty of ) the output .
thus , if @xmath11 are the number densities of neutron stars and of binaries , respectively , eqs.[gama]-[gamc ] indicate that the uncertainty in the number @xmath25 of neutron star binaries scales directly with the uncertainties in @xmath26 and @xmath27 .
similarly , if we overestimate the life time @xmath23 of a binary by a factor 10 , the estimated number @xmath25 is overestimated by the same factor .
thanks to a concerted effort by various groups fairly detailed computations of the happenings in globular clusters are now undertaken .
this is a fortunate and necessary development , as many details can not be understood from the simple scalings above .
for example , the wide progenitors of cataclysmic variables are destroyed by close encounters before they evolve in a dense cluster core ( davies 1997 ) , but evolve undisturbed into cataclysmic variables in the outer cluster regions , from where they can sink to the dense core to form a significant part of the current population there ( ivanova et al .
a disadvantage of complex computations is that they tend to hide the uncertainties .
if the cross sections @xmath14 are described in paper i of a series , and the life times @xmath23 in paper iii , the large uncertainties in them tend to be less than obvious in paper v where the final computations are described .
the confidence expressed in summaries of the results of such computations is sometimes rather larger than warranted . an uncertainty in @xmath14 or @xmath23
has an equally large effect in complex computations as in simple estimates . as a further illustration
we discuss two other uncertainties .
the first relates to the question what happens when mass transfer from a giant to its binary companion is dynamically unstable .
it is usually assumed that a spiral - in follows , in which the companion enters the envelope of the giant and expells it through friction .
the outcome of this process is computed using conservation of energy , which implies a drastic shrinking of the orbit ( webbink 1984 ) .
however , the study of nearby binaries consisting of two white dwarfs shows that the mass ratios in them are close to unity ( e.g. maxted et al .
such binaries can only be explained if the consequences of dynamically unstable mass transfer are governed by conservation of angular momentum , rather than by the energy equation .
if the mass leaving the binary has roughly the same specific angular momentum as the binary , the orbital period changes relatively little during the unstable mass transfer and concomitant mass loss from the binary ( van der sluys et al .
the standard prescription of dynamically unstable mass transfer hitherto implemented in globular cluster computations must be replaced .
the second uncertainty relates to the conversion of the mass transfer rate in a cataclysmic variable to the x - ray luminosity . this conversion does not affect the evolution of the binary but it is important for comparison with observations , as most cataclysmic variables in globular clusters are discovered as x - ray sources . it is generally assumed that the x - ray luminosity scales directly with the mass transfer rate : @xmath28 .
alas , reality is more complicated , and indeed in most cases the x - ray luminosity goes down when the mass transfer rate goes up .
this is demonstrated unequivocally in dwarf novae whose x - ray luminosity drops precipitously during outbursts ( fig.[yzcnc ] ) , but there is evidence that it is true in the more stable nova - like variables as well ( verbunt et al .
but exceptions are also known : the dwarf nova ss cyg has higher x - ray flux during outburst than in quiescence ( ponman et al .
theoretical predictions of the numbers of cataclysmic variables in globular clusters that radiate detectable x - ray fluxes , are not believable when based on proportionality of x - ray flux and mass - transfer rate .
observational evidence for the numbers of binaries of various types , and of the dependence of these numbers on cluster properties , may be collected and used to constrain the theories on formation and evolution of various types of binaries in globular clusters ( e.g. pooley et al .
2003 , heinke et al . 2006 ,
pooley & hut 2006 ) . in the next section
we describe a new , and we hope more accurate , method of analysing source numbers : this method is based on direct application of poisson statistics .
this topic brings one of us , fv , to a brief intermezzo .
the first time that i had need of understanding of binomial and poisson statistics was in the early 1980s , when i bought a book by douglas heggie ( 1981 ) called _ megalithic science , ancient mathemathics and astronomy in northwest europe_. the first half of this book discusses the question whether the megalith builders used a standard measure of length ( the answer is no ) , and the second half studies the question whether megalithic structures had astronomical orientations . whereas douglas is still rather modest about this book
, it is still _ the best introduction to the issues in megalithic astronomy _ ( mccluskey 1998 , p.11 ) .
it , for the first time , explained the statistics of the study of astronomical orientations of megalithic structures . in my view
it is not an exaggeration to state that archaeo - astronomy may be divided in a pre - heggie era , in which the statistics is usually wrong , and a post - heggie era , in which at least some people , following douglas s prescription , do their statistics right . among the latter i may specifically mention ruggles ( 1999 ) . as an example of an application of binomial statistics
i paraphrase the discussion on p.198 of the book , of the four station stones in stonehenge .
these stones were placed , near 2500 b.c .
, in a rectangle with the short sides parallel to the main axis of stonehenge ( fig.[heggie ] ) .
this main axis points from the center of the rectangle through a gap in the bank - plus - ditch ( the _ henge _ ) surrounding it toward the location on the horizon of the most northern annual rising of the sun , at summer solstice .
the moon ranges in a 19yr cycle from about 5@xmath29 north of the sun to 5@xmath29 south of the sun , which leads to a range of its most northern and southern annual risings .
the extremes of these ranges , together with the two solstitia , define 6 positions on the eastern horizon . accepting a range of @xmath30 for each position
, we find a probability that an arbitrary direction towards the eastern horizon hits one of these positions @xmath31 .
the rectangle of the station stones defines three new directions towards the east : along the long side and along the diagnonals .
two of these hit one of the 6 positions : extremes of the southernmost annual setting of the moon in the 19yr cycle .
the probability of 1 chance hit in 3 trials with @xmath32 is 28% , the probability of 2 chance hits 4% .
most likely one hit is due to chance , and one hit intentional .
thanks to new research at stonehenge , done after the publication of _ megalithic science _ , we now know that the central structure , circle and horse - shoe form , were put up simultaneous with the station stones , which implies that the view along the diagonals was blocked .
the intentional hit is therefore the one along the long side .
since the long side of a rectangle is necessarily perpendicular to the short side , this implies that the location of stonehenge was selected for its latitude .
for the denser clusters it is found that the x - ray sources lie well within the half - mass radius ( e.g. ngc6440 , fig.1 of pooley et al .
2002b ) , whereas for ( apparently ) large clusters only the area within the half - mass radius is covered ( e.g. 47 tuc , grindlay et al .
thus , the analysis generally deals with the sources within the half - mass radius . if @xmath33 is the observed number of x - ray sources , and @xmath34 the number of background ( or foreground ) sources not related to the cluster , the number of cluster sources within the half - mass radius is @xmath35 . from a model
we may obtain an estimate of the expected number of cluster sources @xmath36 , and we also may estimate the expected number of background sources @xmath37 . in clusters with a small number of sources , @xmath38 , say , one can not apply chi - squared statistics .
one way of solving this is by adding such clusters together , to obtain sufficiently large numbers .
this is done by pooley & hut ( 2006 ) . in this process
, the information on the indicidual clusters is lost . to avoid this problem ,
we fit as follows .
the probability of observing @xmath25 when @xmath39 is expected according to a poisson distribution is @xmath40 an important aspect of the poison function for our application is its asymmetry : @xmath41 , i.e. the probability of observing 3 sources when 0.1 is expected is very much smaller than the probability of observing 0 sources when 3 are expected .
we now consider @xmath42 both @xmath43,@xmath34 are realizations of poisson distributions given by @xmath36 and @xmath37 respectively .
@xmath37 is determined for each cluster separately , e.g. from the observed number of sources well outside the half - mass radius . for each value of @xmath43
we take @xmath44 , and then compute the combined probability @xmath45 .
we then select the @xmath43,@xmath34 pair with the highest combined probability ( fig.[fitex ] ) .
the fitting procedure consists of varying @xmath46 and @xmath47 to maximize @xmath48_j \label{totprob}\ ] ] where @xmath49 indexes the clusters .
.collision number @xmath22 and mass @xmath50 , normalized on the values for ngc6121 , for clusters studied with chandra , together with the number @xmath33 of sources within the half - mass radius , the expected contribution of background sources @xmath37 , and the number of secure members @xmath51 .
most values @xmath33 and @xmath37 are from pooley et al .
( 2003 ) ; those for ngc288 , ngc6366 and ngc6809 are from estimated conversion of @xmath52 as given in kong et al .
( 2006 ) and bassa et al .
( 2007 ) to @xmath53 .
collision numbers and masses are derived from the central density ( in @xmath54pc@xmath55 ) , core radii , and absolute magnitudes given in the feb 2003 version of the harris ( 1996 ) compilation , and they are scaled on the values for m4 = ngc6121 .
thus we assume that the mass to light ratio is the same for all clusters .
[ numbers ] [ cols= " < , > , > , > , > , > " , ] table[numbers ] lists the numbers that we use in our fitting .
we note from the table that mass can not be used as a proxy for collision number .
the cluster with the highest collision number has a ratio @xmath56 ( normalized on m4 ) , the cluster with the highest mass has @xmath57 .
this suggests that mass and collision number can be well discriminated , and we confirm this below .
the derived values for @xmath33 and @xmath37 ( for sources with @xmath58 erg / s ) correlate with the derived values of @xmath22 and @xmath50 , since all are based on assumed values for the cluster distance @xmath59 and interstellar absorption @xmath60 . for the clusters in the table
the correlations thus introduced are small compared to the ranges of @xmath22 and @xmath50 , because the cluster distances are relatively well known , and we ignore them in what follows .
the models we fit all have the form given by eq.[modelfit ] .
we first fit models with @xmath61 , i.e. assuming that the number of cluster sources depends only on the collision number .
the best solution has @xmath62 and is shown in figure[besta ] .
the total probability as defined in eq.[totprob ] is @xmath63 . to see whether this is acceptable ,
we have used a random generator to produce 1000 realizations by drawing random numbers from poisson distributions with values @xmath36 as given by the best model and @xmath37 as listed in table[numbers ] , and for each realization computed the total probability @xmath64 .
the cumulative distribution of @xmath65 is also shown in figure[besta ] : the best solution is within the 95% range around the median value , and thus acceptable .
our next fit allows non - zero values for @xmath46 and @xmath47 , i.e. allows a linear dependence both on mass and on collision number .
the best solution now has @xmath66 and @xmath67 ; contours of 1,2,3 @xmath68 around these best values are shown in figure[fitb ] , where we assume that the distribution of @xmath69 is given by a @xmath70 statistic with one degree of freedom .
( @xmath71 is the total probability of the best solution . )
we see that the solution with @xmath61 is marginally acceptable , i.e. the evidence for the mass dependence is marginal . because we find an acceptable solution for @xmath22 as given by eq.[gamb ]
, there is no need for a different dependence on central density .
our final fit uses more information , _ viz .
_ the number of ( almost ) certain cluster members among the x - ray sources , as determined though optical identifications .
the importance of this may be seen from fig.[fitex ] for ngc288 . with an estimated background @xmath72 and , for a model depending on collision number only , @xmath73 , the most probable solution is that all 11 sources within the half - mass radius are background sources ( cf .
fig.[besta ] ) .
if we add the constraint that at least two sources included in @xmath33 are cluster members ( kong et al .
2006 ) this solution is no longer possible , and the post probable solution now has @xmath74 and @xmath75 . because the collision number is so small , any cluster source must be a primordial binary , as indeed argued by kong et al .
the best solution for all clusters combined now has @xmath76 and @xmath77 , and the solution with @xmath61 now lies well outside the 3-@xmath68 contour , i.e. the dependence on mass @xmath50 is significant ( fig.[fitb ] ) .
* mass @xmath50 is _ not _ a proxy for collision number @xmath22 * the number of dim sources scales both with collision number @xmath22 and with mass @xmath50 * scaling with mass only is not acceptable * correct treatment of the background is important , esp . for faint sources * to prove the mass - dependence optical identifications are essential | low - mass x - ray binaries , recycled pulsars , cataclysmic variables and magnetically active binaries are observed as x - ray sources in globular clusters .
we discuss the classification of these systems , and find that some presumed active binaries are brighter than expected .
we discuss a new statistical method to determine from observations how the formation of x - ray sources depends on the number of stellar encounters and/or on the cluster mass .
we show that cluster mass is not a proxy for the encounter number , and that optical identifications are essential in proving the presence of primordial binaries among the low - luminosity x - ray sources . |
in the monochalcogenides of anomalous rare earths ( re ) like sm or tm the equilibrium between two valence states ( sm@xmath5/sm@xmath6 or tm@xmath5/tm@xmath6 ) governs the release of a 5@xmath7 electron @xcite .
let us start with our new guided tour of the canonical example of sms .
the full line @xmath8 of the ( @xmath9 ) phase diagram ( see fig.[figure1 ] ) represents the first order isostructural transition between the black ( b ) insulating ( i ) sm@xmath5 phase and the so called gold ( g ) phase where the sm ions jumps to an intermediate valence state @xmath10 . in this gold phase ,
smoothly increasing the pressure must lead to recover a trivalent sm@xmath6 ; it is a kramers ion , long range magnetic order must appear above this threshold @xmath11 .
furthermore a sound experimental fact is that up to @xmath12 kbar in the low pressure gold phase a many body insulating phase is smoothly formed on cooling .
whatever the sample quality and the pressure conditions , @xmath13 is a fixed pressure above which the sample is clearly metallic i.e. its resistivity @xmath14 decreases on cooling after reaching a maximum @xmath15 at @xmath16 @xcite .
the recent combination of macroscopic measurements by transport and ac specific heat and of the microscopic probe of the hyperfine field on the sm nuclei by nuclear forward scattering ( nfs ) using esrf facility demonstrates that long range antiferromagnetic ( af ) order appears at @xmath17 . as the pressure variation of @xmath16 now identified as the nel temperature @xmath18 is huge
, @xmath4 is presumably a first order point associated with a lattice parameter discontinuity i.e. a tiny valence jump at @xmath19 towards the trivalent state .
by contrast to the sm case where the valence fluctuations occur between the sm@xmath5 state with a zero angular momentum ( @xmath20 ) and sm@xmath6 with a @xmath21 angular momentum , for the tm ions the mixing is between two finite angular momentum states , respectively @xmath22 ( tm@xmath5 ) and @xmath23 ( tm@xmath6 ) .
whatever the pressure i.e. the valence and even the electronic conduction , the ground state is magnetic .
however , the similarity with sms is a link between magnetic properties and 4@xmath2 localisation . for tmte as shown on fig .
[ figure2 ] , the @xmath24 sequence of ground states is af / i up to @xmath25 , ferromagnetic ( f ) and metallic ( m ) up to @xmath26 and finally the entrance to a af metallic state @xcite ; the rapid collapse of @xmath27 just below @xmath26 and the fast increase of @xmath18 just above @xmath26 suggests a first order transition . for tmse , already in the gold phase at @xmath28
, the switch at @xmath26 occurs between an af insulating phase @xmath29 and another af metallic state above @xmath26 @xcite .
tms itself appears already at @xmath28 as a quiet quasitrivalent metallic state , in principle the ideal kondo lattice @xcite .
as pointed out already , the cerium heavy fermion problem can be summarized by two major ( @xmath30 ) phase diagrams . the wellknown @xmath31 first order transition of the cerium metal and the up to date magnetic phase diagram @xmath18 or @xmath27 versus pressure where the ordering temperature may collapse at @xmath4 @xcite . for the cerium metal ,
the treatment of the magnetism can be crude as all characteristic temperatures @xmath31 are high by comparison to any hypothetical @xmath18 or any fine structure as the crystal field splitting @xmath32 of the trivalent @xmath33 phase . at @xmath28 , @xmath34 is already positive , finite and near 100 k and the critical end points @xmath35 k , @xmath36 kbar are large . for typical
cerium heavy fermion compounds @xmath4 is lower or comparable to @xmath37 and of course @xmath38 and @xmath39 will be negative .
however , the 4@xmath2 electron will feel the distance from @xmath37 and cause the crossing through a characteristic pressure @xmath26 where the local susceptibility loses the anisotropy given by its reaction to the crystal field ( that corresponds to the well known condition @xmath40 ( kondo energy ) greater than @xmath32 ) . above @xmath26 , the 4@xmath2 electron stays rigidly linked to the full degenerancy of the @xmath21 angular momentum . at @xmath26 ,
the sensitivity of the 4@xmath2 electron to the crystal field has vanished ; @xmath26 is characteristic of a valence crossover .
the relative location of @xmath4 and the hidden parameter @xmath26 is one of the key issue .
the idea that singular properties is underpinned by a quantum critical point i.e. pressure is one of the trends of strongly correlated electronic systems ( see references @xcite for the hidden order of uru@xmath0si@xmath0 or @xcite for the metamagnetic quantum criticality in sr@xmath41ru@xmath0o@xmath42 and @xcite for the conversion of a kondo gas to heavy electron kondo liquid ) .
recently special focus has been given on systems just on the paramagnetic verge of @xmath4 i.e. at @xmath43 @xcite .
the non fermi liquid ( nfl ) label refers to the difficulty to reach the standard fermi liquid regime where usual laws are found , even for non - interacting fermi gas ( specific heat @xmath44 linear in temperature @xmath45 , susceptibility in @xmath46 , scattering rate proportional to @xmath46 ) ; the nfl label guarantee `` vintage '' describes a very large temperature crossover before reaching the fermi liquid graal or a new state of matter . to clarify the complexity of this strongly correlated electronic system , an old fashioned approach via thermodynamic languages
will be carried out @xcite . in the same spirit as the clapeyron and ehrenfest relations on the @xmath24 , @xmath45 feedback for first or second order transitions , the grneisen parameter @xmath47 ( where @xmath48 , @xmath49 and @xmath50 are respectively the volume thermal expansion
, the molar volume and the isothermal compressibility ) is a unique experimental gauge of interfering hierarchies .
it will be reduced to a constant @xmath51 only if the free energy @xmath52 can be expressed by a unique parameter @xmath53 i.e. @xmath54 figure [ figure3 ] represents the temperature variation of the grneisen parameter of ceru@xmath0si@xmath0 at @xmath28 i.e. roughly 3 kbar above the critical pressure @xmath55 kbar .
the two singular points are : ( i ) the huge extrapolated value of @xmath56 and ( ii ) the slow entrance in a simple regime ( @xmath57 k ) where @xmath48 and @xmath58 reach their proportionality @xcite . the pressure dependence of @xmath59 is shown in the insert of fig.[figure3 ] . at low pressure close to @xmath4 ,
@xmath51 is weakly pressure dependent changing from 190 to 220 for a negative shift of @xmath60 kbar @xcite while @xmath59 drops to 80 at @xmath26 @xcite .
other values of @xmath61 can be listed crudely for other heavy fermion system . in a double logarithmic representation of the volume dependence of the @xmath62 coefficient of the well known @xmath63 fermi liquid resistivity assuming the usual kadowaki - woods rule @xmath64 , @xmath51 is directly linked to the slope of the curve @xcite .
the compound ceru@xmath0si@xmath0 is a quasi - ideal kondo lattice with a nice ordered tetragonal lattice , a robust ising character of the local susceptibility , the realization of a clean limit condition i.e. a mean free path higher than 1000 far larger than an hypothetical superconducting coherence length with temperature @xmath65 mk and good enough to obtain the complete determination of the fermi surface ( fs ) at least above @xmath4 ( see references in @xcite ) .
the application of a magnetic field @xmath66 along the easy @xmath67 axis leads to a switch through a drastic continuous crossover at @xmath68 from a low field paramagnetic phase ( pa ) dominated by af correlations at a finite wavevector @xmath69 to a highly polarized state ( pp ) dominated by the low wavevector @xmath70 excitation .
the microscopic vision of ceru@xmath0si@xmath0 was nicely given by inelastic neutron scattering experiments , ( i ) with at @xmath71 , the balance between the local response and the intersite coupling , ( ii ) with , under magnetic field , the continuous spreading of the af response with the same characteristic glumly energy @xmath72 mev up to @xmath73 @xcite and the observation just in the vicinity of @xmath73 of an inelastic ferromagnetic signal at far lower energy transfer @xmath74 mev than @xmath75 ( see fig.[figure4 ] ) @xcite .
a skilful mechanism characteristic of the kondo lattice ceru@xmath0si@xmath0 controles the `` ping - pong '' between the magnetic interactions .
quite remarkably , the field and temperature response is well reproduced by a simple form of the thermodynamic as the entropy .
@xmath76 with equal grneisen parameters @xmath77 . with this hypothesis
one can predict the field variation of @xmath78 knowing the field variation of the linear thermal expansion @xmath79 . a satisfactory agreement is found between this scaling derivation of @xmath80 and the bare determination either by specific heat or by magnetization@xcite .
physically , the adaptive formation of the many body kondo lattice bands is characterized by a pseudogap .
the ising spin character of the bare local magnetic ion plays a capital role in the sharpness of this electronic substructure and correlatively for the pseudo - metamagnetism @xcite .
the pseudogap ingredient is a key input in any theory which needs to respect the scaling behavior @xcite .
reminiscent of the @xmath81 collapse of the cerium metal , the associated spectacular lattice softening ( @xmath82% ) is again in agreement with the push pull between magnetic , electronic and lattice instabilities @xcite and the underpinned vicinity of a quantum critical end point .
after a lecture describing the `` beauty '' ceru@xmath0si@xmath0 to quantum criticality , the first attack is to produce the proofs that ceru@xmath0si@xmath0 is close to a magnetic instability .
looking back to the primitive results one decade ago on the ceru@xmath0si@xmath0 series either sweeping electronic density ( doping or pressure ) or magnetic field , a nfl increase of @xmath83 on cooling in a given temperature window justifies the proximity to @xmath4 @xcite . furthermore , this variation is not drastically different from that popularized in the canonical example of cecu@xmath84au@xmath85 @xcite after a rescaling of the temperature by some phenomenological parameter describing the proximity to @xmath4 and also differences in physical ingredients . as on approaching @xmath4 ,
no drastic pressure dependence of @xmath51 occurs .
collapsing to @xmath4 on the paramagnetic side does not lead to an `` earthquake '' even if of course @xmath53 will collapse when @xmath86 .
the deviation from any singular scaling law of the imaginary part of the dynamical susceptibility @xmath87 becomes also quite difficult to point out .
the yet unanswered dilemma is if a new state of matter occurs right at @xmath4 . the obsessive focus just at @xmath4 or on the paramagnetic ( pa ) side ( @xmath88 ) leads to neglect the disappearance of long range magnetism .
as many tunings were realized by doping i.e. at the cost of an increase of disorder , @xmath4 appears as a second order quantum phase transition .
comparison with the previous chalcogenide cases , recent nmr experiments on cein@xmath41 close to @xmath89 with evidence of af and pa magnetic phase separation @xcite as well as drastic fermi surface changes at @xmath4 in cerh@xmath0si@xmath0 @xcite point out that at the limit of a pure lattice at @xmath4 the transition may be first order . today one key question is to determine precisely the fermi surface on both sides of @xmath4 knowing the two limit conditions of a localized 4@xmath2 picture at low pressure ( @xmath90 ) and an itinerant 4@xmath2 picture at high pressure ( @xmath91 ) .
progress will imply a deep interdisciplinary collaboration between experimentalists and theoreticians eager to track carefully the @xmath24 and @xmath66 variations including the modification of the brillouin zone and the differentiation between spin up and spin down effective mass carriers . for unconventional superconductivity ,
the nature of the phase transition at @xmath4 is not critical as proved by the occurrence of superconductivity in cein@xmath41 and cerh@xmath0si@xmath0 .
there are examples like cepd@xmath0si@xmath0 , cein@xmath41 and cerh@xmath0si@xmath0 where the superconducting pocket is stocked to @xmath92 ; there are cases like cecu@xmath0si@xmath0 @xcite , cecu@xmath0ge@xmath0 @xcite and ceni@xmath0ge@xmath0 @xcite where two superconducting domains are centered on distinct pressures @xmath4 and @xmath26
. we will test soon if superconductivity appears on the verge of @xmath93 kbar in ceru@xmath0si@xmath0 ; near @xmath94 kbar , the ising character of the spin allows only longitudinal magnetic fluctuations which does not favor the observation of @xmath7 wave pairing @xcite while near @xmath26 this superconductivity can be promoted by the emergence of transverse magnetic modes as well as valence fluctuations @xcite .
it is time to concentrate on the complete bundle of curves on each side of @xmath4 in ceru@xmath0si@xmath0 as well as in the other cerium tetragonal lattices and yb homologues following our dresden colleagues @xcite .
y. haga et al . to be published .
a. barla et al . to be published .
j. flouquet et al .
physica b * 319 * ( 2002 ) 251 .
p. wachter , _ handbook of physics and chemistry of rare earths _ vol . 19 , p. 177
, edited by k. a. gscheidner , l. eyring , g. h. lander , and g. r. choppin ( north holland - amsterdam , 1994 ) . f. holtzberg and j. wittig , solid state commun . * 40 * ( 1981 ) 315 . f. lapierre et al .
, solid state commun .
* 40 * ( 1981 ) 317 .
h. konzykowski et al .
solid state commun .
* 40 * ( 1981 ) 517 .
m. ohashi et al . , rev .
high pressure sci .
technology * 7 * ( 1998 ) 611 .
et al . , phys .
80 * ( 1997 ) 173 . j. m. mignot et al . ,
j. magn .
mater * 211 * ( 2001 ) 226 .
m. ribault et al .
* 45 * ( 1980 ) 1295 .
f. holtzberg et al .
, j. appl .
* 57 * ( 1985 ) 3067 .
s. kambe et al . , j. phys .
* 69 * ( 2000 ) 41 .
f. bourdarot et al .
, to be published .
m. chiao et al . , cond - mat/0207657 .
s. nakatsujii et al . , cond - mat/0305427 .
a. schrder et al . ,
nature ( london ) * 407 * ( 2000 ) 351 .
a. benoit et al . , conference on valence fluctuations in solids , p. 283
, edited by l. m. falicov , w. hanke , m. b. maple ( santa barbara , 1981 ) . j. flouquet et al . , j. appl
* 53 * ( 1982 ) 2127 .
s. kambe et al . , j. phys . : condens .
matter * 9 * ( 1997 ) 4917 .
l. zhu et al . , to be published , cond - mat/0212335 .
a. lacerda , thesis , university of grenoble , 1990 , unpublished , and a. lacerda et al .
b * 40 * ( 1989 ) 8759 .
s. holtmeier , thesis , university of grenoble , 1993 , unpublished , and k. payer et al .
, physica b * 186 - 188 * ( 1993 ) 503 . j. flouquet , to be published in : progress low temperature physics s. raymond et al . , physica b * 259 - 261 * ( 1999 ) 48 .
m. sato et al . , j. phys .
* 70 * ( 2001 ) suppl .
a , 118 . s. raymond et al . to be published . c. paulsen et al . , j. low temp . phys . * 81 * ( 1990 ) 317 .
y. aoki et al . , j. magn . magn
. mater . * 177 - 181 * ( 1998 ) 271 .
f. ohkawa , phys .
b * 57 * ( 1998 ) 5891 .
r. a. fisher et al . , j. low temp
* 84 * ( 1991 ) 49 .
h. v. lhneysen et al . , phys .
* 72 * ( 1994 ) 3262 .
s. kawasaki et al . ,
private communication .
s. araki et al . , phys .
b * 64 * ( 2001 ) 224417 .
b. bellarbi et al . , phys . rev .
b * 30 * ( 1984 ) 779 .
d. jaccard et al . , phys .
a * 163 * ( 1992 ) 475 see d. braithwaite et al .
, j. phys . : condens .
matter * 12 * ( 2000 ) 1339 .
see p. monthoux and g. g. lonzarich , phys .
b * 66 * ( 2002 ) 224504 .
g. g. lonzarich , this conference .
y. onishi and k. miyake , j. phys .
* 69 * ( 2000 ) 3955 .
a. holmes et al . , cond - mat/0306054 .
see s. paschen et al .
, this conference . | the meccano of heavy fermion systems is shown on different cases going from anomalous monochalcogenides to cerium intermetallic compounds with special focus on the ideal case of the ceru@xmath0si@xmath0 series .
discussion is made in the frame of the interplay between valence , electronic structure ( fermi surface ) , and magnetism .
the nice tools given by the temperature , the pressure , and the magnetic field allow to explore different ground states as well as the slow downhill `` race '' before reaching a fermi liquid finish line at very low temperature .
experimentally , the grneisen parameter i.e. the ratio of the thermal expansion by the specific heat is a coloured magic number ; its temperature , pressure , and magnetic field dependence is a deep disclosure of competing hierarchies and the conversion of this adaptive matter to external responses . , , , , , , kondo lattice , valence instability , magnetism , superconductivity 71.10.hf , 71.27.+a , 74.70.tx , 75.30.mb heavy fermion physics are approached by pedestrian looking of basic questions on the electronic occupancy @xmath1 of the 4@xmath2 shell and the magnetic properties .
the monochalcogenide cases of sm and tm compounds are revisited with recent data obtained on sms @xcite . for the cerium intermetallic compounds ,
special attention is given close to the so called magnetic quantum criticality i.e. near the critical electronic density @xmath3 or pressure @xmath4 where drastic changes occur on the ground state .
pressure and magnetic field are elegant tool to modify the weight or even the sign of the magnetic interactions by comparison to the strong local fluctuations inherent to kondo lattices .
emphasis is given on the results on the ceru@xmath0si@xmath0 familly since the pure compound i.e. with vanishing disorder allows to scan through different temperature and field regimes @xcite . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.