Some Physical Properties of the Weeks-Chandler-Andersen Fluid David Michael Heyes, Hisashi Okumura To cite this version: David Michael Heyes, Hisashi Okumura. Some Physical Properties of the Weeks-Chandler-Andersen Fluid. Molecular Simulation, 2006, 32 (01), pp.45-49. 10.1080/08927020500529442. hal-00514977 HAL Id: hal-00514977 https://hal.science/hal-00514977 Submitted on 4 Sep 2010 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. F Some Physical Properties of the Weeks-Chandler-Andersen Fluid o r Journal: Molecular Simulation/Journal of Experimental Nanoscience P Manuscript ID: GMOS-2005-0102 e Journal: Molecular Simulation e Date Submitted by the Author: 18-Drec-2005 Complete List of Authors: Heyes, David; University of Surrey, Chemistry Okumura, Hisashi; Institute for Molecular Science Okazaki, R Department of Theoretical Studies e Keywords: WCA fluids, molecular dynamics, physical properties v i e w O n l y http://mc.manuscriptcentral.com/tandf/jenmol Page 1 of 16 1 Some Physical Properties of the Weeks-Chandler-Andersen Fluid 2 3 4 5 David M. Heyes(cid:3) 6 7 Division of Chemistry, 8 9 10 School of Biomedical and Molecular Sciences, 11 12 University of Surrey, 13 14 15 Guildford GU2 7XH, United Kingdom, 16 F 17 18 o 19 20 r 21 Hisashi Okumuray 22 23 P Department of Theoretical Studies, 24 25 e 26 Institute for Molecular Science, e 27 28 r Okazaki, Aichi 444-8585, Japan, 29 30 31 and Department of Structural Molecular Science, R 32 33 The Graduate Universityefor Advanced Studies, 34 35 v 36 Okazaki, Aichi 444-858i5, Japan. 37 38 z e 39 40 w 41 (Dated: December 18, 2005) 42 43 O 44 45 46 n 47 l 48 y 49 50 51 52 53 54 55 56 57 58 59 60 1 http://mc.manuscriptcentral.com/tandf/jenmol Page 2 of 16 1 Abstract 2 3 4 Molecular Dynamics simulations have been carried out of some properties of a Weeks-Chandler- 5 6 Andersen (WCA) system in its (cid:13)uid phase. Data for the potential energy components, mean square 7 8 9 force, in(cid:12)nite frequency elastic moduli and Poissons ratio are presented as a function of density and 10 11 temperature. The scaling behaviour of these quantities using reduced variables, such as an e(cid:11)ective 12 13 14 hard sphere diameter, (cid:27) , was investigated. The in(cid:12)nite frequency Poisson’s ratio, (cid:23) , was found HS 1 15 16 to increase with pacFking fraction and temperature towards the incompressible (cid:13)uid limit value of 17 18 o 19 1=3. Some success was achieved in scaling the mean square force, by the value of the force squared 20 r 21 evaluated at (cid:27) . 22 HS 23 P 24 25 e 26 e 27 PACS numbers: 28 r 29 30 31 R 32 33 e 34 35 v 36 i 37 38 e 39 40 w 41 42 43 O 44 45 46 n 47 l 48 y 49 50 51 52 53 54 55 56 (cid:3)Electronic address: [email protected] 57 yElectronic address: [email protected] 58 59 zSubmitted to Molecular Simulation; 10 text pages and 6 (cid:12)gures 60 2 http://mc.manuscriptcentral.com/tandf/jenmol Page 3 of 16 1 I. INTRODUCTION 2 3 4 5 The Weeks-Chandler-Anderson (WCA) potential is a popular pair potential, 6 7 8 4(cid:15)(((cid:27))12 ((cid:27))6)+(cid:15); r 21=6(cid:27) 8 9 (cid:30)(r) = > r (cid:0) r (cid:20) : (1) 10 >< 0 r > 21=6(cid:27) 11 12 > > : 13 This potential is the Lennard-Jones (LJ) potential, shifted upwards by (cid:15) and truncated at the 14 15 16 LJ potential minimum of 21=6(cid:27). Although the total potential is repulsive, it is composed of F 17 18 repulsive r(cid:0)12 (‘r’) and attractive r(cid:0)6 (‘a’) components. It was originally devised as a reference o 19 20 r 21 (cid:13)uid in a perturbation treat ment for the LJ (cid:13)uid [1, 2]. The Lennard-Jones pair potential is 22 23 decomposed into two parts, onPe entirely repulsive (the WCA potential) and the other entirely 24 25 e attractive. The WCA (cid:13)uid has been examined in various dimensions [3, 4]. The phase diagram 26 e 27 of the WCA system is relatively simple, in that unlike the LJ potential it does not have a 28 r 29 30 critical point and liquid-vapour co-existence region. The melting line of the WCA (cid:13)uid has 31 R 32 been determined [5], and its thermomechanical properties calculated by Molecular Dynamics 33 e 34 35 simulations [6]. v 36 i 37 38 e 39 In this study we consider various properties of the WCA (cid:13)uid, including the potential energy, 40 w 41 42 the mean square force, in(cid:12)nite frequency elastic moduli and Poisson’s ratio. This study follows 43 O 44 on from our recent molecular dynamics simulation and theory of the generalised soft sphere po- 45 46 n 47 tential, (cid:30)(r) = (cid:15)((cid:27)=r)n where n is an adjustable parameter. This r(cid:0)n poltential, like the WCA, 48 y 49 is purely repulsive but exhibits a more convenient scaling of the static and dynamic properties 50 51 [7]. For the soft sphere potential, there is a direct mapping between temperature and density, 52 53 54 so that the complete equation of state can be computed from density dependent data carried 55 56 out at a single temperature. In the WCA case, one cannot perform such a temperature-density 57 58 59 interchange, although one would expect that at high temperature (kBT >> (cid:15)) the behav- 60 3 http://mc.manuscriptcentral.com/tandf/jenmol Page 4 of 16 1 ioroftheWCA(cid:13)uidshouldapproachthatofthesoftspherepotentialwiththeexponentn = 12. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 F 17 18 o 19 20 r 21 22 23 P 24 25 e 26 e 27 28 r 29 30 31 R 32 33 e 34 35 v 36 i 37 38 e 39 40 w 41 42 43 O 44 45 46 n 47 l 48 y 49 50 51 52 53 54 55 56 57 58 59 60 4 http://mc.manuscriptcentral.com/tandf/jenmol Page 5 of 16 1 II. RESULTS AND DISCUSSION 2 3 4 5 Molecular Dynamics simulations were carried out with N = 500 particles in volume V for 6 7 typically 70000 time steps of 0:005=pT (cid:27)(m=(cid:15))1=2 where T is the ‘reduced’ temperature or 8 9 10 k T=(cid:15), where k is Boltzmann’s constant. All quantities quoted below and on the (cid:12)gures are B B 11 12 in WCA reduced units, i:e:; (cid:15) (cid:27) m = 1 (m is the mass of the particle). The density 13 (cid:17) (cid:17) 14 15 of the (cid:13)uid is usually written in terms of a packing or volume fraction, (cid:16) = (cid:25)N(cid:27)3=6V. The 16 F 17 simulations were carried out at reduced temperatures in the range 0:25 5:0, the same set as 18 (cid:0) o 19 used in ref.[5]andatdensities uptoandabove the maximum (cid:13)uiddensity ateach temperature. 20 r 21 22 23 P 24 One of the main themes of this work is an attempt to collapse the computed quantities onto a 25 e 26 single master curve in each case. Ineorder to achieve this we need to introduce a prescription 27 28 r 29 for a state dependent characteristic length scale, which has had an extensive and rich history. 30 31 Although real molecules and the WCA parRticle are not hard spheres, the exercise of mapping 32 33 e 34 their properties onto those of an ‘equivalent’ hard sphere (cid:13)uid has proved such a longstanding 35 v 36 procedure in liquid state theory (e:g:; see ref. [8] Chapi. 6). The requirement is to attribute an 37 38 e e(cid:11)ective hard sphere diameter, (cid:27) to the real molecule. The WCA data can be rescaled by 39 HS 40 w 41 de(cid:12)ning an e(cid:11)ective hard sphere diameter, (cid:27)HS, which leads to an e(cid:11)ective hard sphere packing 42 43 fraction through (cid:16) (cid:25)N(cid:27)3 =6V. A number of presciptions for (cid:27) exist in the literature 44 HS (cid:17) HS OHS 45 46 (e:g:; see ref. [6] and references therein) which involve, for example, inntegrating the Boltzmann 47 l 48 factor of the potential. Unfortunately for WCA (indeed for most commonly used potentials) y 49 50 these do no lead to simple analytic expressions for the usual potential forms. An alternative 51 52 53 scheme is to simply use the distance at which the WCA potential equals the thermal energy 54 55 k T (or a factor times the thermal energy), which leads to a simple analytic expression for (cid:27) . 56 B HS 57 58 This criterion for (cid:27)HS has also been used by Hess et al. [6]. In reduced units this requires us 59 60 5 http://mc.manuscriptcentral.com/tandf/jenmol Page 6 of 16 1 to solve the quadratic equation, T = 4X2 4X +1 for X, where (in reduced units) X = r(cid:0)6. 2 (cid:0) 3 4 Then substitution of X for r gives, 5 6 21=6 7 (cid:27) = ; (2) HS 8 (1+pT)1=6 9 10 on setting r (cid:27) . The formula in Eq. (2) indicates that the e(cid:11)ective hard sphere diameter 11 (cid:17) HS 12 13 decreases with increasing temperature. This prescription a speci(cid:12)c example of a more general 14 15 formula derived by Ben-Amotz and Herschback [9], 16 F 17 (cid:11) 18 0 19 o (cid:27)HS = (1+ T=T )1=6; (3) 0 20 r p 21 where (cid:11) and T are variables that could be density dependent [9, 10]. 22 0 0 23 P 24 25 e 26 In Fig. 1 the potential energy per particle divided by k T , u=k T and the repulsive (u =k T) e B B r B 27 28 r 29 and attractive components (ua=kBT) of the WCA potential energy are presented, plotted 30 31 against (cid:16)HS derived from Eq. (2). It can bRe seen that both components are important at all 32 33 e temperatures (even though the total potential energy is always positive). Figure 2 shows the 34 35 v 36 total potential energy per particle only. There is no obivious way of scaling < u > as at (cid:27) the HS 37 38 e Eq. (2)energy perparticle is always k T (this factwas used inderiving the formulain Eq. (2)). B 39 40 w 41 42 43 Figure 3 shows the mean square force per particle plotted against e(cid:11)ective hard sphere packing O 44 45 fraction, (cid:16) de(cid:12)ned with (cid:27) from the formula in Eq. (2). The force is normalised by the 46 HS HS n 47 l 48 value derived from the pair potential, evaluated at r = (cid:27)HS. It can be seen that the scaling is y 49 50 not good. However, it can be improved if the de(cid:12)nition of (cid:27) is modi(cid:12)ed empirically to, HS 51 52 22=9 53 (cid:27) = ; (4) 54 HS (1+pT)1=6 55 56 as can be seen in (cid:12)gure 4. The di(cid:11)erence between the de(cid:12)nitions of (cid:27) in Eqs. (2) and (4) is 57 HS 58 59 in the coe(cid:14)cient 21=6 being replaced by 22=9. These two de(cid:12)nitions still fall within the general 60 6 http://mc.manuscriptcentral.com/tandf/jenmol Page 7 of 16 1 de(cid:12)nition of (cid:27) of Ben-Amotz and Herschback [9] given in the formula of Eq. (3). Using 2 HS 3 4 Eq. (4), convergence of the data onto a single line is reasonable, at least for T > 1. 5 6 7 8 9 We now consider the mechanical properties of the WCA (cid:13)uid. The in(cid:12)nite frequency elastic 10 11 shear modulus, G for pair-wise additive interactions can be written as follows [11], 1 12 13 14 G = (cid:26)k T + 2(cid:25)(cid:26)2 1drg(r) d (r4(cid:30)0); (5) 15 1 B 15 Z dr 0 16 F 17 where g(r) is the radial distribution function and (cid:30)0 d(cid:30)=dr. (cid:26) = N=V is the number density 18 o (cid:17) 19 20 for N particles in volume rV. In a Molecular Dynamics, MD, or Monte Carlo, MC, simulation 21 22 G can be computed from Eq. (5) rewritten as a sum over pair interactions, 23 1 P 24 25 e1 N(cid:0)1 N d 2267 G1 = (cid:26)kBT +e15V < Xi=1 jX=i+1ri(cid:0)j2drij(ri4j(cid:30)0ij) > : (6) 28 r 29 where r and (cid:30) are the separation and p air potential between particles i and j respectively. ij ij 30 31 < > stands for a simulation time or ensRemble average. Similarly for the in(cid:12)nite frequency 32 (cid:1)(cid:1)(cid:1) 33 e 34 bulk or compressional modulus, K , [11], 1 35 v 36 i 2(cid:25)(cid:26)2 1 37 K = 5(cid:26)k T=3+ drg(r)r3(r(cid:30)00 2(cid:30)0); (7) 38 1 B 9 Z0 e(cid:0) 39 40 where, (cid:30)00 = (d=dr)(cid:30)0, which in terms of pair interactions iws, 41 42 N(cid:0)1 N 43 1 44 K1 = 5(cid:26)kBT=3+ 9V < rij(rij(cid:30)0i0j (cid:0)O2(cid:30)0ij) > : (8) 45 Xi=1 jX=i+1 46 n 47 At liquid-like densities the kinetic contributions to G and K are relaltively small compared 1 1 48 y 49 with the (cid:30)-dependent terms. 50 51 52 53 54 Figure 5 shows the packing fraction dependence of G along the isotherms considered. The 1 55 56 total G , and its repulsive and attractive components are also shown on the (cid:12)gure. The 57 1 58 59 repulsive component can be seen to dominate the total. Because of the complex nature of the 60 7 http://mc.manuscriptcentral.com/tandf/jenmol Page 8 of 16 1 formulas for this quantity, given in Eq. (5), one would not expect any simple scaling of the 2 3 4 data. The corresponding plot for K is qualitatively similar, although the magnitude of each 1 5 6 term is larger. The (in(cid:12)nite frequency) Poisson’s ratio, (cid:23) can be obtained from the computed 7 1 8 9 G1 and K1 using the formula [12], 10 11 3K 2G 12 (cid:23) = 1 (cid:0) 1 (9) 1 13 6K +2G 1 1 14 15 which is plotted as a function of equivalent hard sphere packing fraction in Fig. 9. The 16 F 17 18 (cid:23) increase with packing fraction and temperature towards the incompressible (cid:13)uid limit 1 o 19 20 r of (cid:23) = 1=3. Note that Poisson’s ratio, like the compressibility factor, Z is a dimensionless 21 22 23 quantity and therefore cannotPbe scaled by any function of (cid:27)HS. 24 25 e 26 e 27 In this report we have explored the density and temperature dependence of some of the less 28 r 29 30 well studied properties of the WCA (cid:13)uid. We have illustrated how these properties vary with 31 R 32 temperature and density. We have also made some attempts to scale the various data sets onto 33 e 34 master curves using an e(cid:11)ective hard sphere diameter treatment. These quantities we (cid:12)nd are, 35 v 36 i 37 however, rather di(cid:14)cult to collapse onto master curves, presumably because of the relatively 38 e 39 complicated functional form of the potential. 40 w 41 42 43 O 44 45 46 n Acknowledgements 47 l 48 y 49 50 DMH and HO thank The Royal Society (London) and the Japan Society for the Promotion 51 52 of Science for funding this collaboration. The authors thank Dr. A.C. Bran(cid:19)ka, Institute for 53 54 Molecular Physics, Polish Academy of Sciences, Poznan(cid:19), Poland, for useful discussions. 55 56 57 58 59 60 8 http://mc.manuscriptcentral.com/tandf/jenmol
Description: