Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction

The kinetic theory of mixtures with a multistep potential (MSP) is considered. A passage to the limit of the smooth continuous potential for the kinetic equation and the potential energy density balance equation is analyzed. When the “hard spheres + soft tail” form for the limiting potential is chos...

Full description

Saved in:
Bibliographic Details
Date:2010
Main Authors: Humenyuk, Y.A., Tokarchuk, M.V.
Format: Article
Language:English
Published: Відділення фізики і астрономії НАН України 2010
Subjects:
Online Access:https://nasplib.isofts.kiev.ua/handle/123456789/13437
Tags: Add Tag
No Tags, Be the first to tag this record!
Journal Title:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Cite this:Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction / Y.A. Humenyuk, M.V. Tokarchuk // Укр. фіз. журн. — 2010. — Т. 55, № 4. — С. 450-457. — Бібліогр.: 26 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-13437
record_format dspace
spelling Humenyuk, Y.A.
Tokarchuk, M.V.
2010-11-08T17:31:55Z
2010-11-08T17:31:55Z
2010
Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction / Y.A. Humenyuk, M.V. Tokarchuk // Укр. фіз. журн. — 2010. — Т. 55, № 4. — С. 450-457. — Бібліогр.: 26 назв. — англ.
2071-0194
PACS 05.20.Dd, 05.60.Cd, 51.10.+y
https://nasplib.isofts.kiev.ua/handle/123456789/13437
The kinetic theory of mixtures with a multistep potential (MSP) is considered. A passage to the limit of the smooth continuous potential for the kinetic equation and the potential energy density balance equation is analyzed. When the “hard spheres + soft tail” form for the limiting potential is chosen, the kinetic equation reduces to that of the kinetic variational theory (KVT), while the limiting balance equation for the kinetic energy density differs from its KVT counterpart.
Розглянуто кiнетичну теорiю сумiшей з багатосходинковим потенцiалом взаємодiї. Проаналiзовано граничний перехiд до плавного неперервного потенцiалу для кiнетичного рiвняння i рiвняння балансу для густини потенцiальної енергiї. Коли граничний потенцiал вибрано у формi “твердi кульки + плавний хвiст”, кiнетичне рiвняння зводиться до рiвняння кiнетичної варiацiйної теорiї (КВТ), однак граничне рiвняння балансу для густини кiнетичної енергiї вiдрiзняється вiд вiдповiдного рiвняння КВТ.
en
Відділення фізики і астрономії НАН України
Загальні питання теоретичної фізики
Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction
Гранична поведінка кінетичної теорії для систем із багатосходинковим потенціалом
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction
spellingShingle Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction
Humenyuk, Y.A.
Tokarchuk, M.V.
Загальні питання теоретичної фізики
title_short Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction
title_full Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction
title_fullStr Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction
title_full_unstemmed Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction
title_sort limiting behavior of the kinetic theory for systems with multistep interaction
author Humenyuk, Y.A.
Tokarchuk, M.V.
author_facet Humenyuk, Y.A.
Tokarchuk, M.V.
topic Загальні питання теоретичної фізики
topic_facet Загальні питання теоретичної фізики
publishDate 2010
language English
publisher Відділення фізики і астрономії НАН України
format Article
title_alt Гранична поведінка кінетичної теорії для систем із багатосходинковим потенціалом
description The kinetic theory of mixtures with a multistep potential (MSP) is considered. A passage to the limit of the smooth continuous potential for the kinetic equation and the potential energy density balance equation is analyzed. When the “hard spheres + soft tail” form for the limiting potential is chosen, the kinetic equation reduces to that of the kinetic variational theory (KVT), while the limiting balance equation for the kinetic energy density differs from its KVT counterpart. Розглянуто кiнетичну теорiю сумiшей з багатосходинковим потенцiалом взаємодiї. Проаналiзовано граничний перехiд до плавного неперервного потенцiалу для кiнетичного рiвняння i рiвняння балансу для густини потенцiальної енергiї. Коли граничний потенцiал вибрано у формi “твердi кульки + плавний хвiст”, кiнетичне рiвняння зводиться до рiвняння кiнетичної варiацiйної теорiї (КВТ), однак граничне рiвняння балансу для густини кiнетичної енергiї вiдрiзняється вiд вiдповiдного рiвняння КВТ.
issn 2071-0194
url https://nasplib.isofts.kiev.ua/handle/123456789/13437
citation_txt Limiting Behavior of the Kinetic Theory for Systems with Multistep Interaction / Y.A. Humenyuk, M.V. Tokarchuk // Укр. фіз. журн. — 2010. — Т. 55, № 4. — С. 450-457. — Бібліогр.: 26 назв. — англ.
work_keys_str_mv AT humenyukya limitingbehaviorofthekinetictheoryforsystemswithmultistepinteraction
AT tokarchukmv limitingbehaviorofthekinetictheoryforsystemswithmultistepinteraction
AT humenyukya graničnapovedínkakínetičnoíteoríídlâsistemízbagatoshodinkovimpotencíalom
AT tokarchukmv graničnapovedínkakínetičnoíteoríídlâsistemízbagatoshodinkovimpotencíalom
first_indexed 2025-11-24T04:30:54Z
last_indexed 2025-11-24T04:30:54Z
_version_ 1850842154615701504
fulltext GENERAL PROBLEMS OF THEORETICAL PHYSICS 450 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 LIMITING BEHAVIOR OF THE KINETIC THEORY FOR SYSTEMS WITH MULTISTEP INTERACTION Y.A. HUMENYUK,1 M.V. TOKARCHUK1, 2 1Institute for Condensed Matter Physics, Nat. Acad. of Sci. of Ukraine (1, Svientsitskii Str., Lviv 79011, Ukraine; e-mail: josyp@ ph. icmp. lviv. ua ) 2National University “Lvivska Politekhnika” (12, Bandera Str., Lviv 79013, Ukraine; e-mail: mtok@ ph. icmp. lviv. ua ) PACS 05.20.Dd, 05.60.Cd, 51.10.+y c©2010 The kinetic theory of mixtures with a multistep potential (MSP) is considered. A passage to the limit of the smooth continuous potential for the kinetic equation and the potential energy den- sity balance equation is analyzed. When the “hard spheres + soft tail” form for the limiting potential is chosen, the kinetic equa- tion reduces to that of the kinetic variational theory (KVT), while the limiting balance equation for the kinetic energy density differs from its KVT counterpart. 1. Introduction Triple and higher-order collisions present a consider- able difficulty for the development of the kinetic the- ory for dense systems with realistic potentials, e.g., of the Lennard-Jones type and, specifically, for treating a long-range interaction in explicit way. The significant progress has been achieved, therefore, for model systems based on the Enskog kinetic theory for hard spheres [1–4] which include the long-range interaction in some approx- imate way. The addition of a Fokker–Planck-type term to the Enskog collision integral [5] results in that the hard spheres move as Brownian particles between hard- core collisions. However, the shear viscosity and thermal conductivity coefficients of the model differ appreciably [6] from the molecular dynamic data. Using the approach of maximizing the entropy with certain constraints, several kinetic mean-field theories (KMFT) for a system of particles interacting via a “hard- core + soft tail” potential were proposed [7–9]. They take the attraction into account through the mean-field collision integral and thermodynamic quantities. A mix- ture version was considered, and the theory in the Kac- tail limit was examined [10]. Though the proposed bal- ance equation for the kinetic energy density is of non- typical form, the theory is considered successful up till now [11]. There also appear the semiphenomenological ways to treat the interparticle attraction through, e.g., the average cross-section of the momentum transfer used for viscosity [12]. The kinetic theory for MSP [13–17] generalizes the theory for a square well [18–20]. The both are char- acteristic in that the long-range attraction is included into an appropriate collision integral explicitly and in irreversible fashion. The balance equation for the inter- action energy density is a necessary ingredient of these theories. It is worth noting that one can deduce the sim- plest version of KMFT, called KVT, by passing to the limit upon the potential, when the starting point is the MSP kinetic theory, the number of steps increases, and, simultaneously, separations between them decrease [14]. Thus, we get a way to verify KVT by the MSP theory. The question on the limiting behavior concerning the MSP kinetic equation alone was considered earlier [14], but for a one-component model and for such a closure re- lation [18] for the two-particle distribution function that the potential energy equation was neglected. It was re- vealed that the MSP kinetic equation transforms in the limit to the KVT-type one, while the question on the lim- iting behavior of the balance equations for kinetic and potential energy densities remained without attention. We consider a mixture version of the MSP kinetic the- ory, when the closure relation used takes the potential energy density into consideration [19]. The passage to the limit is analyzed both on the kinetic level and in hydrodynamic equations. The limiting equation for the LIMITING BEHAVIOR OF THE KINETIC THEORY kinetic energy density is found to differ from that pro- posed in KVT and to include contributions to the stress tensor and the heat flux from the soft tail of a limiting potential. In Section 2, we present the equations of the kinetic level, the closure, and the corresponding hydro- dynamics [21]. In Section 3, the main ideas and results of KVT are outlined. The passage to the limit of a con- tinuous potential for both the kinetic and hydrodynamic equations is considered in Section 4. In the last section, conclusions are made. 2. Kinetic Description and Hydrodynamic Equations The contribution from a long-range interaction to the to- tal energy density cannot be neglected at intermediate and high number densities. This is the reason for the kinetic equation for the one-particle distribution func- tion to be complemented with an appropriate balance equation for the potential energy density [19, 20, 22, 23]. This circumstance changes qualitatively the scenario of the kinetic stage of evolution as compared with that of the hard-sphere system and must be included into the construction of kinetic theories for dense systems with a realistic smooth interaction [24, 25]. We consider an M -component system of classical par- ticles interacting by pair central forces. The multistep potential φMS ij mimics the realistic one consisting of a hard core and a set of repulsive and attractive walls of finite heights. The geometry of MSP is determined by the following parameters [15,16,21]: σ0 ij , σrl ij , σal ij are allo- cations of the hard-core, repulsive, and attractive walls, Kr ij and Ka ij are numbers of the repulsive and attrac- tive walls (Fig. 1). Parameters φrl ij and φal ij denote values of MSP between walls, while εrlij = φrl ij − φ r,l+1 ij , εalij = φal ij − φ a,l−1 ij characterize heights of the walls and are in- troduced in such a way that εrlij , εalij > 0. The plateaux are numbered in the direction of increasing rij : from the hard core for the repulsive part and from the first attractive wall for the attractive part of MSP. The one-particle distribution functions f i1 are gov- erned by the system of kinetic equations [21] [∂t + vi · ∇] f i1(r,vi, t) = M∑ j=1 { IE ij [f ij 2 ] + IMS ij [f ij2 ] } , (1) where ∂t ≡ ∂ ∂t , ∇ ≡ ∂ ∂r , and f ij2 are the two-particle dis- tribution functions, the closure relations for which will be given below. The collision integrals IE ij and IMS ij de- scribe pair processes at the hard core and the walls of Model multistep potential φMS ij , with Kr ij = 2, Ka ij = 5 finite height. Expression for the Enskog-type term, IE ij , is conventional [4, 21] and not given here. The contribution IMS ij can be presented in a compact form [21] with the aid of a type-of-step parameter q and a type of process p at the step: q = ‘r’ for the repulsive steps and q = ‘a’ for the attractive ones. There are three types of (ij)-processes at a step: descending p = ⊕, ascending p = , and reflection p = ⊗, therefore IMS ij [f ij2 ] = r,a∑ q Kq ij∑ l=1 ⊕, ,⊗∑ p Iqlpij [f ij2 ]. (2) We ascribe formal numerical values to the parameters as follows: q = ( r a ) = ( −1 +1 ) , p = ⊕ ⊗  = +1 −1 0  , (3) so that contributions to the IMS ij can be given as Iql⊗ij [f ij2 ] = (σqlij ) 2 ∫ dvjdσ̂ vjiσ θ(vjiσ) θ(v ql ij − vjiσ)× ×[f ij2 (r,v′i, r− qσ ql ij ,v ′ j) −q − −f ij2 (r,vi, r + qσqlij ,vj) −q] (4) for the reflection process at step {q, l} and Iqlpij [f ij2 ] = (σqlij ) 2 ∫ dvjdσ̂ vjiσ θ(vjiσ + p− 1 2 vqlij )× ×[f ij2 (r,vqlpi , r + qpσqlij ,v qlp j )−qp − −f ij2 (r,vi, r− qpσqlij ,vj) +qp] (5) for the descending and ascending processes. Here, nu- merical values of p and q are used to determine the po- sition of particle j for the function f ij2 , the argument ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 451 Y.A. HUMENYUK, M.V. TOKARCHUK of the θ function in Eq. (5), and to fix the right or left limiting values for f ij2 in Eqs. (4), (5). Otherwise, q and p are symbols being used for designation. Notations introduced for velocities of particles i and j mean: vi, vj – just before a process; v′i, v′j – just after a pair collision at the hard core or a reflection process at a step; vqlpi , vqlpj – after processes p = ⊕, at the step {q, l}; σ̂ is a unit vector from the center of particle j to the center of particle i, σ0 ij = σ0 ij σ̂, σqlij = σqlij σ̂, vjiσ = (vj −vi)· σ̂. The corresponding rules for the pair processes read v′i = vi + 2Mjivji · σ̂σ̂, v′j = vj − 2Mijvji · σ̂σ̂, (6) vqlpi = vi +Mji[vjiσ − √ v2 jiσ + p(vqlij )2] σ̂, vqlpj = vj −Mij [vjiσ − √ v2 jiσ + p(vqlij )2] σ̂, (7) where Mji = mj/(mi + mj), v ql ij = (2εqlij/µij) 1/2 is the step height in velocity units, µij = mimj/(mi + mj) is the reduced mass. The quantity f ij2 is discontinuous in configurations, for which MSP is, and f ij2 (.)± denotes the right (+) or left (−) limiting value. Averages with f i1 and f ij2 will be designated as 〈ψ1〉1,ivi df= ∫ dvi f i1 ψ1, 〈ψ2〉2,ijvixj df= ∫ dvidxj f ij 2 ψ2, where the subscripts indicate variables of integration. The balance equation for the potential energy density ep(r, t) df= M∑ i,j=1 〈1 2 φMS ij (rij)〉2,ijvixj ∣∣ ri→r (8) must be considered on the kinetic level of description together with the kinetic equations for f i1. Its heuristic derivation based on ideas of the numbers of direct and inverse collisions gives [21] ∂te p +∇· [Vep + qp] = sp, (9) where V(r, t) is the hydrodynamic velocity defined below Eq. (17) below and qp(r, t) = M∑ i,j=1 〈1 2 ciφMS ij (rij)〉2,ijvixj ∣∣ ri→r , (10) sp(r, t) =− M∑ i,j=1 r,a∑ q Kq ij∑ l=1 ⊕, ∑ p 1 2 p εqlij(σ ql ij ) 2 ∫ dvidvjdσ̂ × × vjiσ θ(vjiσ + p− 1 2 vqlij )f ij 2 (r,vi, r− qpσqlij ,vj) qp (11) are the flux in the local reference system and the source which concern to the potential energy of interaction; and ci ≡ vi −V is the thermal velocity. So far, Eq. (1) has been being considered as the first equation of the BBGKY hierarchy for MSP. The closure relation for the collision integrals IE ij , IMS ij and the source sp is chosen as in Ref. [19], when correlations in the velocity space are neglected, i.e., f ij2 is replaced with f̄ ij2 (xi, xj , t) ≡ fi(xi, t)fj(xj , t) gij2 (ri, rj , t), (12) and gij2 is a functional of number densities nk(r, t) and the inverse potential quasitemperature βp(r, t) gij2 (ri, rj , t) = gij2 (ri, rj |{n}, βp), (13) so that gij2 has the same cluster expansion (n-vertex, f -bond) as in equilibrium. However, in the nonequilib- rium case, nk(r, t) replaces each nk, and βp ij(ri, rj , t) = 1 2 [βp(ri, t)+βp(rj , t)] replaces 1/kBT at each bond. The quantity βp(r, t) is the Lagrange multiplier conjugated to the potential energy density [19, 23–25] and is treated in the theory using the balance equation for ep(r, t). The functional gij2 is discontinuous at each point of disconti- nuity of MSP and obeys the relation gij2 (r, r± σqlij , t) −qp = epβ p ijε ql ijgij2 (r, r± σqlij , t) qp. (14) Other closures are proposed for the case of a one- component system with the square-well potential using the equilibrium pair distribution function of the uniform system [18] or the pair distribution function of the hard sphere system in the state of nonuniform equilibrium [9]. The kinetic equation introduced can be used in the high-density region only, where the contributions from successive processes at two and more neighbouring steps may be neglected (the pair collision approximation for each step) [14, 15]. This approximation induces the re- striction [15, 17] Δσ σ0 m � 1 4 √ 2π n (σ0 m)3 gm 2 (σ0)+ , (15) where Δσ is the smallest separation between walls, σ0 m = min {σ0 ij}, n is the total number density, and gm 2 (σ0)+ = min {gij2 (σ0 ij) +} is the smallest contact value of gij2 ’s. On the hydrodynamic level, the system is described by balance equations for the densities of conserved quanti- ties, namely the mass ρ, momentum p, and energy e densities. Such equations for ρ, p, and the kinetic en- ergy density ek only ρ(r, t)p(r, t) ek(r, t)  df= M∑ i=1 〈 mi mivi 1 2miv 2 i 〉1,i vi (16) 452 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 LIMITING BEHAVIOR OF THE KINETIC THEORY can be derived from Eq. (1). In the case of M species, they were obtained in Ref. [21]: ∂t ρp ek +∇· VρVp + P Vek + P·V + qk  =  0 0 sk  , (17) where V(r, t) df= p(r, t)/ρ(r, t) is the hydrodynamic ve- locity. The stress tensor P and the heat flux qk of the kinetic energy contain contributions of the kinetic type (k), from repulsion at the hard core (c) due to IE i , and from processes at steps (MS):[ P qk ] = [ Pk + Pc + PMS qk k + qc k + qMS k ] , where PMS = P⊗+P⊕+P , qMS k = q⊗k +q⊕k +q k . Here, we give only expressions for the contributions from the processes at steps [21]:( P qk )⊗ = M∑ i,j=1 r,a∑ q Kq ij∑ l=1 (−q 2 )(σqlij ) 3 ∫ dvidvjdσ̂ vjiσ × × θ(vjiσ) θ(vqlij − vjiσ) σ̂ 1∫ 0 dλ ( mi[c′i − ci] 1 2mi[c′2i − c2i ] ) × ×f ij2 (r− λqσqlij ,vi, r + [1− λ]qσqlij ,vj) −q, (18)( P qk )⊕+ = M∑ i,j=1 r,a∑ q Kq ij∑ l=1 ⊕, ∑ p qp 2 (σqlij ) 3 ∫ dvidvjdσ̂ × ×vjiσ θ(vjiσ + p− 1 2 vqlij ) σ̂ 1∫ 0 dλ ( mi[c qp il − ci] 1 2mi[(c qp il )2 − c2i ] ) × ×f ij2 (r + λqpσqlij ,vi, r− [1− λ]qpσqlij ,vj) qp. (19) The source sk on the right-hand side of Eq. (17) for ek is such that sk = −sp identically, see Eq. (11). As a result, the balance equation for the total energy density e = ek+ ep has no source, so that the local energy conservation law is satisfied. 3. Kinetic Variational Theory The main idea of KVT lies in the construction of a col- lision integral for the potential given as a sum of the hard-sphere repulsion and an arbitrary smooth tail φhs+t ij (r) = φhs ij (r) + φt ij(r), (20) when a form of the functional dependence of the N - particle distribution function is searched through maxi- mizing the entropy subjected to certain constraints [8,9]. As a result, the pair correlations in the velocity space are neglected, and the dependence of gij2 on {n} and βp is determined by the type of a constraint. In the approximation called KVT-III, the constraint consists in the requirement that the local potential energy density ep(r, t) is recovered correctly by the searched N - particle distribution. The entropy maximization princi- ple results in the function gij,hs+t 2 for φhs+t ij in form (13). In addition to the contribution IE ij from φhs ij , there is a term of mean-field type linear in φt ij [10]: IE+KVT i ≡ M∑ j=1 { IE ij [f i 1, f j 1 ] + IKVT ij [f i1, nj ] } , where IKVT ij is given by IKVT ij [f i1, nj ] = 1 mi ∫ r12>σ0 ij dr2 g ij,hs+t 2 (r1, r2|{n}, βp)× ×nj(r2, t)[∇1φ t ij(r12)]· ∂1f i 1(x1, t). (21) The approximation accepted for gij2 is used in both IKVT ij and IE ij . For the collision integral IE+KVT i , the authors of KVT gave a slightly unconventional balance equation for ek in both one- [7] and many-component [10] cases: ∂te k+∇· [ Vek+Pk+c·V+qk+c k ] +V· ( ∇·Pt ) = Δt ek , (22) where the contribution to the stress tensor from the IKVT i reads Pt(r1, t) df= −1 2 M∑ i,j=1 ∫ ds s ∂φt ij(s) ∂s × × 1∫ 0 dλnij,hs+t 2 (r1 − λs, r1 + [1− λ]s), (23) nij,hs+t 2 (r1, r2) ≡ ni(r1, t)nj(r2, t)g ij,hs+t 2 (r1, r2|{n}, βp). It is not hard to recover for a general case the remainder term Δt ek (given in Ref. [10] in the Kac-tail limit): Δt ek(r1, t) = M∑ i,j=1 [Vi −V]× × ∫ dr21 n ij,hs+t 2 (r1, r2|{n}, βp)φt′ ij(r12) r̂21, (24) where Vi ≡ 〈vi〉1,ivi /ni is the average velocity of species i. The following distinctions of Eq. (22) should be pointed out: a) the contribution with Pt is not of the standard ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 453 Y.A. HUMENYUK, M.V. TOKARCHUK form −∇· (Pt·V), as those with Pk or Pc; b) there is no contribution qt k from IKVT i to the heat flux of the kinetic energy; and c) due to the difference Vi −V, the source Δt ek is of the diffusion type and therefore vanishes in the one-component case. The balance equation for ep of KVT was introduced in Ref. [10] only, where many-component system was con- sidered, and simpler constraints which do not allow the appearance of βp were used. It was shown in the Kac- tail limit that the total energy density obeys a correct balance equation through second order in gradients. The indicated shortcomings require a more detailed analysis of this theory. 4. Passage to the Limit in the MSP Kinetic Theory Starting from MSP, an arbitrary potential φ̃ij(r) can be approximated better and better by increasing the num- ber of steps and by a simultaneous decrease in separa- tions between them [14]. This passage to the limit can be denoted as lim {εql ij }→0 {Δσql ij }→0 (. . .) ≡ lim φMS→φ̃ (. . .). (25) Though the applicability condition (15) of the MSP ki- netic theory is violated, we consider consequences of the passage, when the model potential (20) plays the role of the limiting potential φ̃ij(r). In this case, IE i remains the same, while the contribution IMS i undergoes a change. Transformations below resemble those applied to the derivation [26] of the Landau kinetic equation from the Boltzmann equation and use expansions into series in a small value of the momentum transferred in pair colli- sions. 4.1. The limit of the kinetic equation Like in Ref. [14], we search for the limiting form of IMS ij , Eqs. (2), (4), (5), when f i1(ri,v qp il , t) f j 1 (rj ,v qp jl , t) g ij 2 (ri, rj |{n}, βp) (26) is expanded into a series near vi and vj ; here, f̄ ij2 in- cludes closure (12), (13). For small step heights, {εqlij} → 0, the square root in the pair collision rule (7) for processes ⊕ and is expanded into a series and restricted to the linear order in εqlij :[ vqpil vqpjl ] = [ vi vj ] − + [ Mji Mij ] × 1 2 p 2εqlij µijvjiσ σ̂. (27) For the contributions Iql⊕ij or Iql ij , this means f̄ ij2 (r,vqpil , r + qpσqlij ,v qp jl ) −qp = [ 1 + p εqlij vjiσ σ̂ × × ( − ∂i mi + ∂j mj ) + . . . ] f̄ ij2 (r,vi, r + qpσqlij ,vj) −qp, (28) where, e.g., ∂i ≡ ∂/∂vi. The term Iql⊗ij describes the reflection at steps of a very small height. Due to the second θ function in Iql⊗ij , the projection vjiσ has an upper bound 0 ≤ vjiσ ≤ (2εqlij/µij) 1/2. Therefore, v′i,v ′ j are close to vi,vj , see Eq. (6): f̄ ij2 (r,v′i, r− qσ ql ij ,v ′ j) −q = [ 1 + 2µijvjiσσ̂ × × ( − ∂i mi + ∂j mj ) + . . . ] f̄ ij2 (r,vi, r− qσqlij ,vj) −q. (29) Inserting these expressions into the formulas for Iqlpij , it can be observed that the main contributions ∼ 1 from processes and ⊗ in the square brackets cancel with that from process ⊕. The first-order term ∼vjiσ in Iql⊗ij , Eq. (29), gives the function with a fixed upper bound being integrated over the interval with the size tending to zero. As a result, Iql⊗ij vanishes. The first-order terms ∼ εqlij in Eq. (28) gives a non-zero contribution. We now consider expres- sions from ⊕ and , e.g., for a repulsive step: Kr ij∑ l=1 εrlij(σ rl ij) 2 ∫ dσ̂σ̂ × × { θ(vjiσ) ( − ∂i mi + ∂j mj ) f̄ ij2 (r,vi, r− σrl ij ,vj) +− −θ(vjiσ − vrl ij) ( − ∂i mi + ∂j mj ) f̄ ij2 (r,vi, r+σrl ij ,vj) − } . (30) Some remarks concerning the passage to the limit are as follows: i) by increasing the number of steps and de- creasing the separations between them, the sum ∑Kq ij l=1 is transformed into an integral with respect to the contin- uous relative distance R: Kr ij∑ l=1 Δσr;l,l+1 ij εrlij Δσr;l,l+1 ij −→ ∫ φ̃′ij<0 dR ( − ∂φ̃ij ∂R ) , (31) 454 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 LIMITING BEHAVIOR OF THE KINETIC THEORY where the region with φ̃′ij(R) < 0 corresponds to re- pulsive steps and that with φ̃′ij(R) > 0 corresponds to attractive ones; ii) ∫ dσ̂ changes to the integral over the orientations R̂ of the relative distance vector R; iii) for the continuous part of φ̃ij , the right and left limiting values of f̄ ij2 are equal to each other: f̄ ij2 (.)+ = f̄ ij2 (.)−. In those terms of the expressions obtained where f̄ ij2 depends on r − R, the integration variable should be changed to R̂′ = −R̂. Keeping in mind that ∂φ̃ij(R) ∂R R̂ = ∂φ̃ij(R) ∂R and taking the equality of the limiting values of f̄ ij2 into account, it can be deduced that IMS i transforms into M∑ j=1 ∫ dvjdR ∂φ̃ij(R) ∂R · ( − ∂i mi + ∂j mj ) f̄ ij2 (r,vi, r + R,vj). (32) The integration by parts shows that the term with ∂jf ij 2 vanishes as f̄ ij2 ∣∣ |vj |→+∞ → 0. We can integrate further with respect to vj due to the absence of velocity corre- lations, Eq. (26), with the result − 1 m i M∑ j=1 ∫ dR gij2 (r, r + R|{n}, βp)× ×nj(r + R, t) ∂φ̃ij(R) ∂R · ∂i f i1(r,vi). (33) This expression coincides with the collision integral (21) of the kinetic variational theory, when gij2 is allowed to depend on βp (the version KVT-III, [8]). It is important to stress this dependence, as the corresponding system of limiting equations for f i1 is unclosed, until an equation for βp(r, t) is proposed. It should be derived from the limiting equation for ep complementing the system of kinetic equations. The final expression (33) was obtained for the first time in Ref. [14] for a one-component case, but no analysis of the gij2 functional dependence was given. It is worth noting that Eq. (32) corresponds to the integral term of the first equation of the BBGKY hier- archy for a mixture with continuous potential φ̃ij . In other words, we have shown that the pair collision oper- ator of MSP takes, in the limit, the form of a differential operator of pair interaction for the appropriate smooth potential. 4.2. The limit of the equation for ep Equation (9) is not changed, but the quantities in it are. The quantities ep and qp contain φMS ij and, in accordance with Eqs. (8) and (10), go over into[ ẽp q̃p ] df= M∑ i,j=1 〈1 2 [ 1 ci ] φ̃ij(rij)〉2,ijvixj ∣∣ ri→r . (34) The source sp depends on parameters of φMS ij and de- scribes the processes at steps. The contributions from ⊕ and in Eq. (11) for sp can be written explicitly as M∑ i,j=1 r,a∑ q Kq ij∑ l=1 1 2 εqlij(σ ql ij ) 2 ∫ dvidvjdσ̂ vji ·σ̂ × × { θ(−vjiσ)f̄ ij2 (r,vi, r + qσqlij ,vj) q + +θ(vjiσ − vqlij )f̄ ij 2 (r,vi, r + qσqlij ,vj) −q } , (35) where the variable σ̂ in the term with p = ⊕ has been changed to σ̂′ = −σ̂. Relation (14) for the left and right limiting values of gij2 at the discontinuity points of φMS ij can be expanded into a series as {εqlij} → 0, gij2 (r, r + qσqlij , t) −q= [1 + βp ijε ql ij + . . .]gij2 (r, r + qσqlij , t) q, (36) and inserted into expression (35). But the latter already contains the factor εqlij . Therefore, only the main term in square brackets of Eq. (36) is retained. After changing to σ̂′ = −σ̂ in the term with q = −1, constructions (31) can be built up with the same rules of transformation. The two θ functions in Eq. (35) cover almost the whole region of vjiσ with the exception of the interval of size ∼(εqlij) 1/2 which tends to zero in the limit. Replacing the sum with the integral over the relative distance, we get s̃p(r, t) = M∑ i,j=1 〈1 2 vji · ∂φ̃ij(rij) ∂rji 〉2,ijvixj ∣∣ ri→r . (37) This limiting expression coincides with its counterpart for a system with smooth interaction which can be de- duced [2], by starting immediately from the second equa- tion of the BBGKY hierarchy and the definition of ẽp. 4.3. The limit of equations for one-particle densities At the passage to the limit, only the expressions for con- tributions PMS and qMS k are changed. The source sk changes in the same way as sp, Eq. (37). The contributions P⊗ and q⊗k tend rapidly to zero, as 1) due to the product θ(vjiσ) θ(v ql ij − vjiσ), the size of ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 455 Y.A. HUMENYUK, M.V. TOKARCHUK the integration interval over vijσ vanishes; 2) owing to θ(vqlij − vjiσ), the factors mi(c′i − ci) and 1 2mi[c′2i − c2i ] vanish, as the upper bounds of their absolute values are proportional to (εqlij) 1/2. The contributions ⊕ and to PMS and qMS k depend on the height of steps through the differences which, for small {εqlij}, are equal to Eq. (27): [ cqpil − ci (cqpil )2 − c2i ] = − [ 1 2 1 ] pMji (vqlij ) 2 cjiσ [ σ̂ ci ·σ̂ ] . Inserting these into Eq. (19) and going over to the con- tinuous variable R as we have done above for IMS ij and sp yield[ P̃ q̃k ]⊕+ = −1 2 M∑ i,j=1 ∫ dvi dvj [ 1 ci· ]∫ drji r̂jirji × ×φ̃′ij(rij) ∫ 1 0 dλf̄ ij2 (r− λrji,vi, r + [1− λ]rji,vj). (38) The expression for P̃⊕+ can be integrated over vi and vj and, in the case of the potential φhs+t ij , coincides with the expression [7] for the contribution Pt from the soft tail φt ij , Eq. (23). However, the limiting heat flux of the kinetic energy q̃⊕+ k does not have such a counterpart in KVT. The limiting equation for ek is of the same form as Eq. (17), in which PMS, qMS, and sk must be replaced with P̃⊕+ , q̃⊕+ , and s̃k = −s̃p, Eq. (37). It does not have the shortcomings inherent in Eq. (22). 5. Conclusions We have considered a passage to the limit of the poten- tial of “hard spheres + arbitrary tail” for the kinetic the- ory for the MSP system. The limiting kinetic equation obtained coincides with that of the kinetic variational theory [7, 9], namely KVT-III [8]. It is found out that the contribution IMS i from the processes at steps trans- forms into the mean-field term of this theory. Beside the earlier consideration for a one-component fluid [13], the limiting form of the balance equation for the potential energy density is additionally obtained with the explicit expression for the corresponding source. Contrary to the result for the kinetic equation, the lim- iting equation for the kinetic energy density differs from its KVT counterpart for both one- and many-component cases. It includes additional terms with contributions to the stress tensor and the heat flux from a smooth tail. The reason for this disagreement will be analyzed sepa- rately. 1. S. Chapman and T.G. Cowling, The Mathematical The- ory of Non-Uniform Gases (Cambridge University Press, Cambridge, 1970). 2. J.H. Ferziger and H.G. Kaper, Mathematical Theory of Transport Processes in Gases (North-Holland, Amster- dam, 1972). 3. H. van Beijeren and M.H. Ernst, Physica (Utrecht) 68, 437 (1973). 4. H. van Beijeren and M.H. Ernst, Physica (Utrecht) 70, 225 (1973). 5. S.A. Rice and A.R. Allnatt, J. Chem. Phys. 34, 2144 (1961). 6. R. Vogelsang and C. Hoheisel, Phys. Rev. A 39, 6391 (1989). 7. J. Karkheck and G. Stell, J. Chem. Phys. 75, 1475 (1981). 8. G. Stell, J. Karkheck, and H. van Beijeren, J. Chem. Phys. 79, 3166 (1983). 9. J. Karkheck and G. Stell, Phys. Rev. A 25, 3302 (1982). 10. J. Karkheck, E. Martina, and G. Stell, Phys. Rev. A 25, 3328 (1982). 11. K.M. Dyer, B.M. Pettitt, and G. Stell, J. Chem. Phys. 126, 034502 (2007). 12. S. Morioka, Phys. Rev. E 72, 051203 (2005). 13. M.V. Tokarchuk and I.P. Omelyan, Preprint of Bo- golyubov Institute for Theor. Physics, ITF–87–152R (ITP, Kiev, 1987) (in Russian). 14. M.V. Tokarchuk and I.P. Omelyan, Preprint of Bo- golyubov Institute for Theor. Physics, ITF–89–49R (ITP, Kiev, 1989) (in Russian). 15. M.V. Tokarchuk and I.P. Omelyan, Ukr. Fiz. Zh. 35, 1255 (1990). 16. I.P. Omelyan and M.V. Tokarchuk, Physica A 234, 89 (1996). 17. M.V. Tokarchuk, I.P. Omelyan, and A.E. Kobryn, Phys. Rev. E 62, 8021 (2000). 18. H.T. Davis, S.A. Rice, and J.V. Sengers, J. Chem. Phys. 35, 2210 (1961). 19. J. Karkheck, H. van Beijeren, I. de Schepper, and G. Stell, Phys. Rev. A 32, 2517 (1985). 20. H. van Beijeren, J. Karkheck, and J.V. Sengers, Phys. Rev. A 37, 2247 (1988). 456 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 LIMITING BEHAVIOR OF THE KINETIC THEORY 21. M.V. Tokarchuk and Y.A. Humenyuk, Condens. Matter Phys. 10, 151 (2007). 22. J. Karkheck, G. Stell, and J. Xu, J. Chem. Phys. 89, 5829 (1988). 23. J.A. Leegwater, H. van Beijeren, and J.P.J. Michels, J. Phys.: Condens. Matter 1, 237 (1989). 24. D.N. Zubarev and V.G. Morozov, Teor. Mat. Fiz. 60, 270 (1984). 25. J.A. Leegwater, J. Chem. Phys. 95, 8346 (1991). 26. V.P. Silin, Introduction to the Kinetic Theory of Gases (Nauka, Moscow, 1971) (in Russian). Received 23.10.09 ГРАНИЧНА ПОВЕДIНКА КIНЕТИЧНОЇ ТЕОРIЇ ДЛЯ СИСТЕМ IЗ БАГАТОСХОДИНКОВИМ ПОТЕНЦIАЛОМ Й.А. Гуменюк, М.В. Токарчук Р е з ю м е Розглянуто кiнетичну теорiю сумiшей з багатосходинковим потенцiалом взаємодiї. Проаналiзовано граничний перехiд до плавного неперервного потенцiалу для кiнетичного рiвняння i рiвняння балансу для густини потенцiальної енергiї. Коли гра- ничний потенцiал вибрано у формi “твердi кульки + плавний хвiст”, кiнетичне рiвняння зводиться до рiвняння кiнетичної варiацiйної теорiї (КВТ), однак граничне рiвняння балансу для густини кiнетичної енергiї вiдрiзняється вiд вiдповiдного рiв- няння КВТ. ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 4 457