scieee Science in your language
[en] (orig)

Mesoscopic theory of critical fluctuations in isolated granular gases

Abstract

Fluctuating hydrodynamics is used to describe the total energy fluctuations of a freely evolving gas of inelastic hard spheres near the threshold of the clustering instability. They are shown to be governed only by vorticity fluctuations that also lead to a renormalization of the average total energy. The theory predicts a power-law divergent behavior of the scaled second moment of the fluctuations, and a scaling property of their probability distribution, both in agreement with simulations results. A more quantitative comparison between theory and simulation for the critical amplitudes and the form of the scaling function is also carried out.

Read accessible full text

Mesoscopic theory of critical fluctuations in isolated granular gases

Author: Domínguez Álvarez, Álvaro; García de Soria Lucena, María Isabel; Maynar Blanco, Pablo; Brey Abalo, José Javier
Publisher: American Physical Review
Year: 2006
DOI: 10.1103/PhysRevLett.96.158002
Source: https://idus.us.es/bitstreams/3a04ce9f-6d4b-49a9-be3f-904ae828e785/download
Mesoscopic Theo y o C i ical Fluc ua ions in Isola ed G anula Gases
J. Ja ie B ey,*A. Domı
´nguez, M. I. Ga cı
´a de So ia, and P. Mayna
Fı
´sica Teo
´ ica, Uni e sidad de Se illa, Apa ado de Co eos 1065, E-41080 Se illa, Spain
(Recei ed 24 No embe 2005; published 21 Ap il 2006)
Fluc ua ing hyd odynamics is used o desc ibe he o al ene gy luc ua ions o a eely e ol ing gas o
inelas ic ha d sphe es nea he h eshold o he clus e ing ins abili y. They a e shown o be go e ned only
by o ici y luc ua ions ha also lead o a eno maliza ion o he a e age o al ene gy. The heo y p edic s
a powe -law di e gen beha io o he scaled second momen o he luc ua ions, and a scaling p ope y o
hei p obabili y dis ibu ion, bo h in ag eemen wi h simula ions esul s. A mo e quan i a i e compa ison
be ween heo y and simula ion o he c i ical ampli udes and he o m o he scaling unc ion is also
ca ied ou .
DOI: 10.1103/PhysRe Le .96.158002 PACS numbe s: 45.70.n, 05.20.Dd, 51.10.+y
G anula gases a e assemblies o mac oscopic pa icles
e ol ing independen ly be ween inelas ic collisions [1].
The me hods o nonequilib ium s a is ical mechanics, ki-
ne ic heo y, and hyd odynamics ha e been success ully
ex ended o desc ibe he obse ed mac oscopic beha io
and also, al hough in a much mo e limi ed o m, he
luc ua ions a ound i [2]. The lack o ene gy conse a ion
makes hese sys ems beha e qui e di e en ly om mo-
lecula luids. A simple widely used model o hem con-
sis s o smoo h inelas ic ha d sphe es (IHS’s), wi h mo-
men um conse ing dynamics. Inelas ici y is cha ac e ized
by means o a cons an coe icien o no mal es i u ion .
Recen ly, molecula dynamics (MD) simula ion esul s
ha e been epo ed o he o al ene gy luc ua ions o a
wo-dimensional eely e ol ing IHS gas, nea he h esh-
old o he clus e ing ins abili y [3]. The dimensionless
second momen was ound o exhibi a powe -law di e -
gen beha io wi h he dis ance o he ins abili y. Also, he
scaled cooling a e was ound o end o ze o acco ding o a
powe law, al hough in a weak way. Besides, he dis ibu-
ion unc ion o he ene gy luc ua ions, when p ope ly
scaled, u ned ou o be independen o he pa ame e s
de ining he sys em. This was associa ed wi h a scaling
p ope y o he dis ibu ion. Qui e ema kably, he scaling
unc ion was e y well i ed by he same exp ession as
se e al equilib ium and nonequilib ium molecula sys ems
[4,5]. The main goal o his Le e is o p o ide an expla-
na ion o he abo e esul s on he basis o luc ua ing
hyd odynamics [6].
Conside an isola ed sys em o NIHS’s o mass mand
diame e . The o al (kine ic) ene gy ~
Eo he sys em can
be exp essed in he o m [7]
2~
E Zd d~
n ; ~
T ; m~
n ; ~
u2 ; ;(1)
whe e dis he dimension o he sys em, ~
n ;  he numbe
densi y ield, ~
T ;  he empe a u e ield, and ~
u ;  he
low ield. The ildes indica e ha all he quan i ies a e
unde s ood as luc ua ing a iables. In he ollowing, sys-
ems in he homogeneous cooling s a e (HCS) will be
conside ed. A a mac oscopic le el, his s a e is cha ac e -
ized by a cons an uni o m densi y nH, a anishing low
ield uH0, and a uni o m ime-dependen empe a u e
obeying he law [8] @ TH HTHTH , whe e H/
TH 1=2is he cooling a e. Mo eo e , we will es ic
ou sel es o he egion in which he ampli udes o he
luc ua ions o he ields a ound hei HCS alues emain
small on he a e age [see below Eq. (12)]. Then e aining
up o quad a ic o de in he de ia ions, Eq. (1) yields
~
E ~
E EH 
1
2Zd dnH~
T ; d~
n ; ~
T ; 
mnHj~
u ; j2:(2)
He e, EH dNTH =2,~
n ; ~
n ; nH,
~
u ; ~
u ; , and ~
T ; ~
T ; TH .I is
now con enien o in oduce dimensionless posi ion, l,
and ime, s, scales by l =l0and ds  H d =l0, e-
spec i ely, whe e H2TH =m1=2is he he mal e-
loci y and l0nHd11is p opo ional o he mean
ee pa h. Mo eo e , dimensionless ields a e de ined by
l;s~
n ; =nH,!l;s~
u ; = H , and
l;s~
T ; =TH . Then, Eq. (2) akes he o m
s0s
V1
V2X
kksks2
dj!ksj2;(3)
whe e s~
Es=EHs,VLdis he olume o he
sys em in he new uni s, and he Fou ie ans o ms o he
ields ha e been in oduced. I is assumed ha a e a ime
o he o de o he mean ee ime, he sys em eaches a
egime in which all i s ene gy is s o ed in he hyd ody-
namic modes. In his egime, he hyd odynamic ields a e
expec ed o be desc ibed a a mesoscopic le el by luc ua -
ing hyd odynamic equa ions. He e, hey will be assumed o
be linea Lange in equa ions ob ained by linea izing he
Na ie -S okes equa ions o a g anula gas a ound he
HCS. Mo eo e , i is pos ula ed ha he noise e ms a e
de ined by he same p ope ies as o molecula , elas ic
PRL 96, 158002 (2006) PHYSICAL REVIEW LETTERS week ending
21 APRIL 2006
0031-9007=06=96(15)=158002(4)$23.00 158002-1 ©2006 The Ame ican Physical Socie y
gases [9]. This is no expec ed o be ue, excep in he
nea ly elas ic limi , i.e., when is e y close o uni y.
Consequen ly, he heo y will be es ic ed in he ollowing
o his limi . Thus, he ans e sal low ield o o ici y
ield, !k?, obeys he ollowing equa ion in he scaled
a iables [6,9]:
@s=2k2!k?sk?s:(4)
In his exp ession, HTH l0= H and 
HTH =mnHl0 H ,Hbeing he shea iscosi y.
The noise e m k?sis Gaussian, wi h
hk?sk0?s0i  V2
Nss0k;k0k2l;(5)
lbeing he uni enso in he subspace pe pendicula o k,
and he angula b acke s deno ing a e age o e he noise
ealiza ions. A main ad an age o using he scaled a ia-
bles is ha he coe icien s in Eq. (4) and he s eng h o he
noise a e ime independen , con a y o wha happens in he
o iginal a iables. The equa ion shows ha !k?g ows in
ime o hose alues o ksuch ha ?k=2
k2>0. Al hough his does no imply by i sel ha he
HCS is linea ly uns able, due o he ime-dependen scaling
o he eloci y in oduced abo e, simula ion esul s and
nonlinea analy ical analysis o he Na ie -S okes equa-
ions ha e shown ha his g ow h is he o igin o he
clus e ing ins abili y [10,11]. The minimum alue o k
o a sys em o linea ex en L, measu ed in he lscale,
is kmin 2=L. Then, o gi en alues o he o he pa-
ame e s, he sys em becomes uns able i L>L
c, wi h
Lc22=1=2.Fo L<L
c, he HCS is s able and
he long ime solu ion o Eq. (4) is
!k?sZs
1
ds0ess0?kk?s0:(6)
F om his exp ession, i is easily ob ained
h!k?s!k0?s0i   V2k2
2N?kess0?kk;k0l;(7)
o ss01. This shows ha as Lapp oaches Lc om
below, he ampli udes o he luc ua ions o he ans e sal
componen s o he eloci y inc ease e y as due o con-
ibu ions om alues o kclose o kc. Fo he same eason
he decay o hese luc ua ions becomes e y slow. This is
no he case o he luc ua ions o he o he hyd odynamic
ields, whose Lange in equa ions a e decoupled om
Eq. (4) [6]. The e o e, i seems possible o conside a ange
o alues o
L LcL=Lcwhe e he luc ua ions o
!k;?, al hough s ill small, domina e o e he luc ua ions
o densi y and empe a u e. Bu , al hough his is ue o
componen s wi h k>0, some ca e is needed when analyz-
ing Eq. (3), since i in ol es 0s. The Lange in equa ion
o sis ob ained om he linea iza ion a ound he HCS
o he mac oscopic a e age equa ion o he o al ene gy,
@ E d
2Zd n ; Hn; TT ; :(8)
The esul is
@sss 3
2V0s:(9)
He e, he dependence o he cooling a e on he empe a-
u e has been aken in o accoun . Mo eo e , he noise e m
discussed in Re . [12], associa ed wi h he localized cha -
ac e o he ene gy dissipa ion, has been omi ed. Al hough
i can be expec ed o be negligible a om he ins abili y in
he quasielas ic limi , his may no be he case nea he
ins abili y. Equa ion (9) shows a coupling be ween he
luc ua ions o he olume a e aged empe a u e and hose
o he o al ene gy. Use o Eq. (3) in o Eq. (9) and neglec -
ing con ibu ions om he densi y and longi udinal eloc-
i y luc ua ions gi es
@ss
2s !s;
!s 6
V2dX
kj!k?sj2;
(10)
alid in he egion
L 1. The long ime limi o he
a e age alue o sis, he e o e,
his lim
s!1h!si  3d1
Nd X
k
k2
?k:(11)
Since we a e conside ing
L 1, he sum o e kin he
abo e exp ession is domina ed by he 2dmodes wi h he
la ges wa eleng h, o which ?kmin’
L. Using
his in o Eq. (11), i ollows ha he e is a eno maliza ion
by luc ua ions o he a e age o al ene gy o he HCS,
E h~
E i, gi en by
E EH 13d1
nHLd
c
L1:(12)
Consis ency o he heo y we a e de eloping equi es ha
nHLd
c
L11, a condi ion in ol ing he inelas ici y
and he dis ance o he ins abili y. Simila ly, he e is also
a eno maliza ion o he empe a u e o he HCS, T 
h~
T is , ha can be e alua ed di ec ly om he long ime
limi o he a e age o Eq. (9),
T TH 1h0is
VTH 12d1
nHLd
c
L1:
(13)
Al e na i ely, an e ec i e empe a u e Te  can be de-
ined as Te  2E =Nd. O cou se, he o m o he
eno malized law o he empe a u e depends on he de i-
ni ion used o he la e . In Re ., [3], wha was ac ually
measu ed was 
e e Te l0= HTe , wi h e de ined by
@ Te e Te Te . Then, i is ound
PRL 96, 158002 (2006) PHYSICAL REVIEW LETTERS week ending
21 APRIL 2006
158002-2

e "13d1
nHLd
c
L1#1=2
:(14)
This esul p edic s ha nea he clus e ing ins abili y
h eshold, 2
e 2=2A
L1wi h A3d
1=nHLd
c, ha is jus he beha io obse ed in Re . [3].
De ine
Es~
EsEs=Esshis 
EHs=Esand !s !sh!is . No e ha we a e
conside ing de ia ions om he eno malized a e age al-
ues, i.e., including he luc ua ions e ec s, and no om he
mac oscopic ba e alues. A s anda d calcula ion using
Eq. (7) and exploi ing he Gaussian cha ac e o he noise,
gi es ha a he ins abili y h eshold and o ss01i is
h!s!s0is 9d1
n2
HL2d
cd
L2ess0=sc;(15)
whe e sc2
L1is a di e gen ‘‘c i ical’’ elaxa ion
ime. Now Eq. (10) can be easily sol ed wi h he esul
h
Es
Es0is h!s!s0is ;(16)
alid o ss01. Thus below he ins abili y, he scaled
o al ene gy luc ua ions decay wi h he same a e as he
luc ua ions o he kine ic ene gy associa ed wi h he ans-
e sal modes o he eloci y. Fo ss0, Eq. (16) yields
2
Eh
E2is A2

L2;(17)
wi h A2
9d1=n2
HL2d
cd. The e o e, close o he in-
s abili y poin , he ela i e dispe sion o he o al ene gy
luc ua ions Ep esen s a di e gen beha io wi h a c i i-
cal exponen 1, and an ampli ude Adepending on nH
and ( h ough he alue o he c i ical leng h Lc). Again,
his is he same beha io as epo ed in Re . [3] om MD
simula ions.
To ca y ou a mo e de ailed check o he heo y p e-
sen ed he e, we ha e pe o med MD simula ions o wo-
dimensional sys ems wi h di e en alues o and nH(see
Table I). In all cases, he dependence on
L o bo h he
cooling a e and he dispe sion o he o al ene gy, i.e., he
exponen s in he powe laws (14) and (17), was in ag ee-
men wi h he heo e ical p edic ions. This was illus a ed
in Figs. 1 and 2 o Re . [3] and no mo e de ails will be
gi en he e. The compa ison be ween he p edic ed c i ical
ampli udes and he MD esul s gi en in Table I can be
conside ed as sa is ac o y, in he sense ha he heo y
co ec ly p edic s he o de o magni ude o he ampli udes,
especially aking in o accoun he smallness o he quan i-
ies being measu ed.
Nex , le us p oceed o in es iga e he o m o he
p obabili y dis ibu ion o he ene gy luc ua ions.
Pa icula iza ion o Eq. (4) o he modes wi h he smalles
possible alue o kin he limi
L 1gi es
@s
L!k?sk?s;(18)
whe e i is unde s ood ha jkjkmin. De ine a new ime
scale d 
Lds, and a new ans e sal eloci y ield by
!
k?!k?=Ld
c1=2
E. Equa ion (18) becomes
@1!
k?
k?;(19)
wi h
h
k?
k0?0i  d1=2
6d11=2k;k00l:(20)
Equa ion (19) implies ha he p obabili y dis ibu ion o
!
k?wi h jkjkmin nea he clus e ing ins abili y depends
only on he dimension do he sys em. In ac , since he
noise e m 
k?is Gaussian, i is i ial o w i e he long
ime o m o his dis ibu ion using Eq. (7) wi h ss0,
Ps !
k?22
!d1=2e!2
k?=22
!;(21)
wi h 2
!d1=2=12d11=2. In he ime scale , and
keeping only he dominan modes, Eq. (10) eads
L@y1
2y6
dX
jkjkminj!
?kj2;(22)
whe e y=Eand he sum is es ic ed o ec o s kwi h
jkjkmin. F om he compa ison o Eqs. (19) and (22) i is
seen ha , on he scale and in he h eshold o he
ins abili y, ydecays much as e han he dominan com-
ponen s o !
k?. Consequen ly, o la ge  he solu ion o
Eq. (22) is y6
dPjkjkmin j!
?kj2, whe e he p obabili y
dis ibu ion o he modes !
?kis gi en by Eq. (21). Since
he la e does no depend on he pa ame e s o he sys em
o he han he dimensionali y, he same p ope y ollows
o he p obabili y dis ibu ion o bo h yand he a iable
E
Edd11=26
dX
jkjkminj!
k?j2:(23)
TABLE I. Compa ison be ween he p edic ed and MD alues
o he c i ical ampli udes o he cooling a e Aand he o al
ene gy dispe sion A. All he alues o he ampli udes ha e been
mul iplied by 103.
nH2A
heo y
AMD
A heo y
AMD

0.02 0.9 0.88 1.11 0.62 0.6
0.02 0.8 1.62 3.63 1.15 1.5
0.1 0.98 1.06 0.50 0.75 0.47
0.1 0.95 2.4 2.38 1.7 1.45
0.2 0.98 1.97 2.34 1.4 1.34
0.2 0.95 4.59 7.78 3.24 3.6
PRL 96, 158002 (2006) PHYSICAL REVIEW LETTERS week ending
21 APRIL 2006
158002-3
This is equi alen o saying ha he p obabili y dis ibu ion
o
E e i ies he scaling ela ion
P
E 1
E

E
E;(24)
whe e is a scaling unc ion. This is jus he p ope y
assumed in Re . [3] and e i ied by MD simula ions. Since
he p obabili y dis ibu ion unc ion o !
k?is known, i is
possible o nume ically gene a e he p obabili y dis ibu-
ion unc ion P
E. The esul is shown in Fig. 1. Also
plo ed is he unc ion

EKexexa;xbs
E;a=2;
(25)
wi h K2:14,b0:938, and s0:374, ha i s ex-
emely well he MD esul s o EP
E[3]. I is impo -
an o ema k ha luc ua ions in a la ge numbe o
equilib ium and nonequilib ium sys ems exhibi ing sel -
o ganized c i icali y as well as con ined u bulen lows
p esen he same kind o beha io [4,5]. Al hough he
ag eemen be ween bo h plo ed cu es is no so bad o
posi i e alues o
E, s ong disc epancies a e obse ed
o nega i e alues. A majo sou ce o hem is easily
iden i ied om Eq. (23), ha o d2implies
E=E
2
p, while smalle alues a e ound in he MD simula-
ions. Since Eq. (23) is a consequence o Eq. (9), i seems
plausible ha in o de o elabo a e a mo e accu a e heo y
he in insic noise associa ed wi h he cooling a e mus be
aken in o accoun .
In summa y, we ha e de eloped a mesoscopic heo y o
he luc ua ions o he o al ene gy o an isola ed g anula
gas nea he h eshold o he clus e ing ins abili y. The
heo y desc ibes accu a ely he quali a i e beha io ob-
ained in MD simula ions, namely, he di e gen beha io
o he dimensionless second momen and he dec ease o
he appa en cooling a e. Also, i is consis en wi h he
obse ed scaling p ope y o he p obabili y dis ibu ion
unc ion o he luc ua ions. On he o he hand, i seems
clea ha a mo e e ined o mula ion is needed in o de o
ge a mo e sa is ac o y quan i a i e ag eemen , especially
o he dis ibu ion unc ion. Also, i should be in e es ing
o check whe he a simila beha io occu s in mo e eal-
is ic models o g anula gases in which depends on he
ela i e eloci y and, al hough p esen , he clus e ing in-
s abili y seems o be a ansien phenomenon [13].
This esea ch was suppo ed by he Minis e io de
Educacio
´n y Ciencia (Spain) h ough G an No.
FIS2005-01398 (pa ially inanced by FEDER unds).
*Elec onic add ess: [email p o ec ed]
[1] H. M. Jaege , S. R. Nagel, and R. P. Beh inge , Re . Mod.
Phys. 68, 1259 (1996).
[2] I. Goldhi sch, Annu. Re . Fluid Mech. 35, 267 (2003).
[3] J. J. B ey, M. I. Ga cı
´a de So ia, P. Mayna , and M. J. Ruiz-
Mon e o, Phys. Re . Le . 94, 098001 (2005).
[4] S. T. B amwell, P. C. W. Holdswo h, and J.-F. Pin on,
Na u e (London) 396, 552 (1998).
[5] S. T. B amwell, K. Ch is ensen, J.-Y. Fo in, P. C. W.
Holdswo h, H. J. Jensen, S. Lise, J. M. Lo
´pez,
M. Nicodemi, J.-F. Pin on, and M. Selli o, Phys. Re .
Le . 84, 3744 (2000).
[6] L. Landau and E. M. Li shi z, Fluid Mechanics
(Pe agamon P ess, New Yo k, 1959).
[7] R. B i o and M. H. E ns , Eu ophys. Le . 43, 497 (1998).
[8] P. K. Ha , J. Fluid Mech. 134, 401 (1983).
[9] T. P. C. an Noije, M. H. E ns , R. B i o, and J. A. G. O za,
Phys. Re . Le . 79, 411 (1997).
[10] I. Goldhi sch and G. Zane i, Phys. Re . Le . 70, 1619
(1993); S. McNama a and W. R. Young, Phys. Re . E 50,
R28 (1994).
[11] J. J. B ey, M. J. Ruiz-Mon e o, and D. Cube o, Phys.
Re . E 60, 3150 (1999).
[12] J. J. B ey, M. I. Ga cı
´a de So ia, P. Mayna , and M. J. Ruiz-
Mon e o, Phys. Re . E 70, 011302 (2004).
[13] N. B illian o , C. Saluen
˜a, T. Schwage , and T. Po
¨schel,
Phys. Re . Le . 93, 134301 (2004).
-6
-4
-2
0
-2 0 2 4
ln [σE P(δ E)]
δ E
~
~
/ σE
FIG. 1. P obabili y densi y unc ion o he ela i e o al ene gy
luc ua ions EP
E o a sys em o inelas ic ha d disks. The
b oken line is he heo e ical p edic ion de i ed in his Le e and
he solid line Eq. (7).
PRL 96, 158002 (2006) PHYSICAL REVIEW LETTERS week ending
21 APRIL 2006
158002-4