Collaborative Computational Projects

  
   Home
  CCP1
  CCP2
  CCP3
  CCP4
  CCP5
  Organisation
  Joining CCP5
  Meetings, workshops and visitors
  Funding from CCP5
  Program Library
  Infoweb
  Links
  CCP6
  CCP9
  CCP11
  CCP12
  CCP13
  CCP14
  CCPB
  CCPN
   Other Projects
SRRTNet
CECAM

 

 

ABSTRACTS

INDEX

Simulation: How to do things that experimentalists can't do and wish they could.

Ruth M. Lynden-Bell
Atomistic Simulation Group, School of Mathematics and Physics, The Queen's University of Belfast, UK

In numerical experiments or simulations one has complete control over the experimental variables in a way that nature does not always allow. For example in classical simulations one can change potential parameters at will and in a systematic way.

Rasaiah and I used this possibility some years ago to measure the variation of solvation entropy with ion charge q and found a hydrophobic minimum near q = 0 and structure breaking maxima at positive and negative values of q. In recent work Bergman and I have used distorted models of water in an attempt to try to understand what causes the hydrophobic effect. As the model is distorted we observe the network structure to change and disappear while the solvation entropy of a hard sphere increases. The hydrophobic minimum in the solvation entropy versus solute charge also decreases until it disappears. We conclude that the high free energy of solvation of hydrophobic solutes in water is primarily due to the lack of free volume in the network structure and it is not really necessary to invoke solvent ordering or ``iceberg'' formation near the solute.

Index of talks

Simulating Non-equilibrium Dynamics in Glassy Systems: Aging, Effective Temperature and Behaviour under Shear

Jean-Louis Barrat
Département de Physique des matériaux (bat 203), Université Claude Bernard Lyon 1 , 43, bld du 11 novembre 1918, 69622 Villeurbanne Cedex- France

Glassy systems are, by definition, out of equilibrium, since their relaxation time is longer than the experimental (or simulation) time scale. I will show how these systems can nevertheless be characterized using molecular dynamics simulations, through the aging properties of their correlation and response functions [1]. The behaviour of simple glass-forming liquids will be discussed in analogy with that of spin-glasses [2]. The concept of an effective temperature related to the fluctuation dissipation relation will be introduced [3]. The behaviour of a glassy system undergoing external forcing (shear) will also be discussed [4 , 5].

References

1
``Fluctuations, response and aging dynamics in a simple glass-forming liquid out of equilibrium'' W. Kob and J-L. Barrat, Eur. Phys. Journal B13 319 (2000)
preprint cond-mat/9910305
2
``Mode-Coupling Approximations, Glass Theory and Disordered Systems'' Jean-Philippe Bouchaud, Leticia Cugliandolo, Jorge Kurchan, Marc Mézard Physica A 226 , 243 (1996);
cond-mat/9511042
3
``Energy flow, partial equilibration and effective temperatures in systems with slow dynamics'' Leticia F. Cugliandolo, Jorge Kurchan, Luca Peliti, Phys. Rev. E 55 3898 (1997);
preprint cond-mat/9611044
4
``Fluctuation-dissipation relation in a sheared fluid'' J-L. Barrat, L. Berthier,
cond-mat/0003346
5
``Two-time scales, two-temperature scenario for nonlinear rheology'' L. Berthier, J-L. Barrat, J. Kurchan, to appear in Phys. Rev E
preprint cond-mat/9910305
Index of talks

Geometric Integrators in Molecular Simulation

Ben Leimkuhler
Department of Mathematics and Computer Science and Centre for Mathematical Modelling, University of Leicester, UK

Geometric integrators are timestepping methods that preserve invariants or symmetries associated to the flow-map of a dynamical system. The most important instances of geometric integrators include unitary integrators (e.g. quantum propagators), symplectic integrators for Hamiltonian systems, and time-reversible integrators . Geometric integrators have become quite popular in recent years, particularly as tools for long term physical or chemical simulation. While our theoretical understanding of their properties remains incomplete, there is growing evidence---both empirical and analytic---to justify the use of these schemes.

In applications like molecular dynamics the systems are characterized by a multiplicity of timescales, generally meaning that a qualitatively important dynamic is manifest on a time interval much greater than the period of the fastest local oscillatory mode. In most cases, the long time-scale phenomena are of the greatest interest, but it is difficult or impossible to model the slow and fast parts separately; although their local effect is weak, the fast components may – over a long period of time – make essential cumulative contributions to the evolution of any state. Stability restrictions due to the presence of the fast modes greatly limit the effectiveness of numerical integrators, and it is for this reason precisely that many interesting phenomena lie beyond the reach of simulation, even on the fastest computers available. The situation is made still more challenging by the refinement of models to include additional quantum mechanics, since this only has the effect of increasing the the range of time-scales.

In the first part of the talk, I will survey some work on the development of geometric integrators for various dynamics problems such as rigid body systems [1 - 3] Heisenberg ferromagnets [4] and constant temperature/pressure models [5] In the second part of the talk, I will focus on a new reversible averaging method for fast-slow dynamics [6]

References

1
``Symplectic methods for conservative multibody systems'' E. Barth and B. Leimkuhler, Fields Inst. Comm. 8 (1996)
2
``Split-Hamiltonian methods for rigid-body molecular dynamics'' A. Dullweber, B. Leimkuhler and R. McLachlan, J. Chem. Phys. 107 5840 (1997)
3
``A symplectic method for rigid-body molecular simulation'' A. Kol, B. Laird and B. Leimkuhler, J. Chem. Phys. 107 2580-2588, (1997)
4
``Geometric integrators for classical spin systems'' J. Frank, W. Huang and B. Leimkuhler, J. Comput. Phys. 133 160-172, (1997)
5
``The Nos\'{e}-Poincar\'{e} Method for constant temperature molecular dynamics'' S. Bond, S, B. Laird and B. Leimkuhler, J. Comput. Phys. 151 114-134, (1999)
6
``A reversible averaging integrator for multiple time-scale dynamics'' B. Leimkuhler and S. Reich, Technical Report No. 2000/11, Dept. of Mathematics and Computer Science, University of Leicester.
7
``Integration methods for molecular dynamics'' B. Leimkuhler, S. Reich and R. Skeel in Mathematical Approaches to Biomolecular Structure and Dynamics , IMA Volumes in Mathematics and its Applications Vol. 82, Ed. by J. P. Mesirov, K. Schulten and D. W. Summers, Springer, (1996)
Index of talks

Many-body interaction effects on the properties of molten ionic mixtures

P.A. Madden;
Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QZ, UK.

Many binary systems (and their mixtures) which might be expected to be ``ionic'', from electronegativity considerations, are found to exhibit pronounced ``covalent effects'' in their condensed phase structure and dynamical properties. An extreme example is AlCl3, which melts from an ionic crystal to form a non-conducting molecular liquid: many other systems (like ZnCl2) form networks, which strongly influence the transport properties. Such systems provide interesting challenges for simulation studies at several levels.

Firstly there is the problem of providing a good description of the interactions. Recent work [1] has suggested that they are accounted for by an ``extended'' ionic model, in which it is recognised that anions in the condensed phase have profoundly different properties from their free counterparts as a consequence of the strong confining potential exerted on the anionic electron density by surrounding ions. Depending on the precise shape of the potential (which, in thermal motion, will vary from one instant to the next), the anion may be more-or-less compressed, deformed and polarized and hence its interaction potential with the other ions in the system will also vary. The resulting many-body effects promote remarkably rich changes in the intermediate-range structure of the liquid from the simplest ``rigid''-ionic model [2,3]. Secondly, there is the issue of connecting what might be calculated in a simulation with the experiment which most directly attests to the associated nature of these fluids - Raman spectroscopy , which shows discrete, quasi-molecular vibrational bands in associated liquids. To directly calculate a Raman spectrum for comparison with experiment a model for the fluctuating polarizability of the sample in terms of the ionic coordinates is needed. Considerable progress has been made with this problem; again, the key idea is the effect of the confining potential on the anion and its polarizability. Raman spectra in good agreement with experiment may be calculated [4]. Lastly comes the problem of providing a microscopic interpretation for the dynamical events affecting the transport properties of these liquids, particularly the ionic conduction. These events typically involve complicated network rearrangements, similar to those invoked in describing the proton transport in water, and do not conform to the type of model developed for the collective dynamics of unassociated fluids.

References

1
M. Wilson and P.A. Madden Chem. Soc. Rev. 25, 339 (1996).
2
F. Hutchinson et al. J. Chem. Phys. 110, 5821 (1999).
3
M. Wilson and P.A. Madden Molec. Phys. 92, 197 (1997).
4
M. C. C. Ribeiro et al. J. Chem. Phys. 110, 4803 (1999).
Index of talks

From Atoms to Continua, using Smooth Particles

Wm. G. Hoover
Department of Applied Science, University of California at Davis/Livermore,
and Methods Development Group, Lawrence Livermore National Laboratory,
Livermore, California 94551-7808

In 1977 Lucy and Monaghan discovered a method for solving the macroscopic continuum equations with particles, rather than with finite elements. The ``smooth particles'' which they introduced can range in size from microscopic to mesoscopic to macroscopic. Every smooth particle makes its own individual contribution to the collective constitutive properties in its vicinity--density, stress, energy, heat flux, and so on. Local values of any such ``field variable'' F are computed as sums:

F(r) $\displaystyle \equiv$ $\displaystyle \sum_{i}^{}$w(r - ri)Fi/$\displaystyle \sum_{i}^{}$w(r - ri),

where the ``weight function'' w is normalized, $ \int$4$ \pi$r2wdr $ \equiv$ 1, has two continuous derivatives, and a finite range, usually two or three particle diameters. This definition guarantees that all the field variables likewise have two continuous spatial derivatives, $ \nabla$F,$ \nabla$$ \nabla$F, enough to provide an approximate solution of the continuum equations for the conservation of mass, momentum, and energy:

$\displaystyle \dot{\rho}$ = - $\displaystyle \rho$$\displaystyle \nabla$ . v;
$\displaystyle \rho$$\displaystyle \dot{v}$ = $\displaystyle \nabla$ . $\displaystyle \sigma$;
$\displaystyle \rho$$\displaystyle \dot{e}$ = $\displaystyle \nabla$v : $\displaystyle \sigma$ - $\displaystyle \nabla$ . Q,

where the field variables {$ \rho$, v,$ \sigma$, e, Q} are the density, velocity, stress tensor, energy per unit mass, and heat-flux vector, respectively. The continuum equations of motion, formulated for the smooth particles, take the form:

{$\displaystyle \dot{v}_{i}^{}$ = m$\displaystyle \sum_{j}^{}$[($\displaystyle \sigma$/$\displaystyle \rho^{2}_{}$)i + ($\displaystyle \sigma$/$\displaystyle \rho^{2}_{}$)j] . $\displaystyle \nabla_{i}^{}$w(rij)}.

Notice that whenever stress and density are slowly varying in space these continuum motion equations look exactly like a set of atomistic motion equations (with the {w} playing the role of pair potentials). This resemblance is useful in understanding the numerical convergence properties of the continuum solutions on the basis of equilibrium statistical mechanics and kinetic theory.

Particle methods are not only simple to implement. They are also completely free of the geometric stability problems that plague finite-element solutions of the continuum equations. Smooth particles can interpenetrate one another and can undergo fracture and failure with no numerical instability. I illustrate this feature with smooth-particle simulations of the equilibration of extreme pressure and density gradients.

Smooth-particle methods can provide solutions converging to those of the continuum equations. This occurs in the limiting case where the particle size is small relative to the significant length scales of the problem being solved. I illustrate this convergence with smooth-particle simulations of thermal convection, comparing the particle results to those obtained with conventional finite elements.

My own introduction to smooth-particle methods is described in the article ``Mécanique de Nonéquilibre à la Californienne", Physica A 240, 1 (1997). For additional discussion and further references have a look at my most recent book--Time Reversibility, Computer Simulation, and Chaos, (World Scientific Publishing, Singapore, 1999).

Index of talks

Macromolecular Product Engineering Using a Simulation Toolkit

Ugur Tüzün
Dept Chemical & Process Engineering, School of Engineering in the Environment, University of Surrey, Guildford, GU2 7XH, UK

Computer simulations of particulate systems became popular in the late seventies when the maximum available computer power could only cope with simulations comprising 102-103 discrete elements. Today, it is possible to carry out simulations routinely comprising 105-106 element systems using parallel processing supercomputers. Further continuing advances of the parallel processor power should soon see systems of the size of 109-1010 elements used in discrete element simulations with only a fractional increase in cost. The availability of such large data sets currently offers real scope for statistically accurate calculations to describe the long-range evolution of the bulk flow and stress fields during industrial handling and processing of materials in granular and powder states. A number of emerging techniques used to post-process simulation data will be described and some recent applications at Surrey will be presented.

The primary fascination for the researchers is the computational potential to generate at relatively low cost, detailed descriptions of the stress and strain or strain-rate (velocity) fields under both quasi-static and dynamic conditions [1 , 2]. Similar data sets can be generated in real experiments using only the most sophisticated techniques such as tomography [3], particle tracking and MRI; all of which require several orders of magnitude greater expense in comparison with computational simulations. A comparative evaluation will be provided of the relative limitations as well as of the complementary features of the studies using simulations and real experiments. More importantly, the need for post-processing of large data sets to arrive at bulk mechanical phenomena remains a common feature of both types of studies. Based on this, a compelling case will be presented for the establishment of bulk mechanical evolution rules as a necessary first step to general reliable macroscopic models of bulk material behaviour under industrial process conditions.

References

1
``Microstructural Simulation and Imaging of Granular Flows in two- and three-dimensional Hoppers'' P. A. Langston, M. S. Nikitidis, U. Tüzün, D. M. Heyes and N.  M. Spyrou Powder Technology 94 , 59-72 (1997)
2
``The Physical Effects of Structures Formation in Granular Materials'' M. A. Koenders, N. Gaspar and U. Tüzün Mechanics of Materials in press. (2000)
3
``Measurement of Size Segregation by Self-Diffusion in Slow Shearing Binary Mixture Flows using Dual Photon Gamma-ray Tomography'' M. S. Nikitidis, U. Tüzün, and N. M. Spyrou Chemical Engineering Science 53 , 2335-2351 (1998)
Index of talks

A Link between the Two-Body and Three-Body Interaction Energies of Fluids from Molecular Simulation

Gianluca Marcelli and Richard J. Sadus
Centre for Molecular Simulation and School of Information Technology, Swinburne University of Technology, PO Box 218, Hawthorn, Victoria 3122, Australia

Advances in molecular simulation algorithms [1] coupled with rapid growth in the calculation speed of modern high performance computers, provide unprecedented opportunities to investigate natural phenomena from a molecular perspective. Molecular simulation has been applied to the design of beneficial pharmaceutical products, [2] the optimisation of important industrial processes, [2] and to the resolution of fundamental scientific questions. [3,4] Despite the use of high performance computing, molecular simulation is confined largely to the calculation of two-body interactions using ``effective'' intermolecular potentials because the inclusion of three- or more-body interactions remain computationally prohibitive. Generally, interactions between pairs [5] of molecules make the overwhelming contribution to the overall intermolecular interaction. However, it is also documented [6-13] that three-body interactions can make a significant contribution to intermolecular interactions in liquids.

The use of ``effective'' intermolecular potentials is a source of considerable inaccuracy and uncertainty in molecular simulations. For example, recent calculations [12] have shown that three-body interactions contribute significantly to the phase behaviour of fluids, whereas this effect had been hidden previously by the use of effective intermolecular potentials. The agreement between experiment and theory for the phase envelope was improved considerably by explicitly accounting for three-body interactions.

In this work, we discuss recently reported [14] molecular simulation data which indicate there is a simple empirical relationship between two body and three-body interaction energies for noble gas atoms. The significance of this relationship is that three-body interactions can be estimated accurately from two-body interactions without incurring the computational penalty of three-body calculations. The relationship has the potential of improving both the accuracy and predictive value of molecular simulation.

References

1
R. J. Sadus, Molecular Simulation of Fluids: Theory, Algorithms and Object-Orientation (Elsevier, Amsterdam, 1999).
2
K. E. Gubbins, and N. Quirke, (Eds), Molecular Simulation and Industrial Applications: Methods, Examples and Prospects (Gordon and Breach, Amsterdam, 1996).
3
P. S. Chialvo and P. T. Cummings, J. Chem. Phys. 101 4466 (1994).
4
J. I. Siepmann, S. Karaborni, and B. Smit, Nature 365 330 (1993).
5
A. J. Stone, The Theory of Intermolecular Forces (Clarendon Press, Oxford, 1996).
6
M. J. Elrod, and R. J. Saykally, Chem. Rev. 94 1975 (1994).
7
R. J. Sadus, and J. M. Prausnitz, J. Chem. Phys. 104 4784 (1996).
8
J. A. Anta, E. Lomba, and M. Lombardero,Phys. Rev. E. 55 2707 (1997).
9
R. J. Sadus, Fluid Phase Equilib. 144 351 (1998).
10
R. J. Sadus, Fluid Phase Equilib. 150-151 63 (1998).
11
R. J. Sadus, Ind Eng. Chem. Res. 37 2977 (1998).
12
G. Marcelli and R. J. Sadus, J. Chem. Phys. 111 1533 (1999).
13
M. A. van der Hoef and P. A. Madden,J. Chem. Phys. 111 1520 (1999).
14
G. Marcelli and R. J. Sadus,J. Chem. Phys. (2000), in press.
Index of talks

Simulations of Capillary Condensation in Porous Glasses

Lev D. Gelb 1 and Keith E. Gubbins 2

  1. Department of Chemistry, Florida State University, Tallahassee, FL 32306-4390, USA
  2. Department of Chemical Engineering, North Carolina State University, Raleigh, NC 27695-7905, USA

We investigate the effects of pore network structure and topology on capillary condensation phenomena using large-scale molecular simulation. Theoretical descriptions of capillary phenomena are based largely on classical thermodynamics applied to idealized pore systems, and are often not applicable to nanometer-scale porous matrices or complex pore topologies. Simulations can provide a molecular-scale understanding of gas adsorption in realistic models of porous materials, which can be used to interpret experimental results [1].

Materials such as xerogels and controlled-pore glasses have average pore sizes as small as a few nanometers, but in order to contain a statistically meaningful sample of the pore network the simulation cell must be many times larger than the average pore size. To accommodate this, we are using parallelized Grand Canonical Monte Carlo simulation techniques [2] to simulation cells of 20-30 nm edge length, which might contain several hundred thousand adsorbed gas molecules at condensation.

We have prepared models of controlled-pore glasses using molecular dynamics simulation mimicking the experimental preparations of these materials, leading to reasonable pore geometries and network topologies [3-5]. We report the properties of these models and compare them with properties of the real materials, and address issues of necessary system size and reproducibility of results. We also consider capillary phenomena in model materials with more regular pores and pore networks.

Xenon adsorption isotherms are simulated in a series of model porous glasses at several temperatures with the aim of obtaining the complete capillary phase diagram and correlating this with the known properties of the models. The phase diagrams of these systems are strongly shifted from that of the bulk fluid, and vary with both the mean pore sizes and porosities of the glass models.

References

1
L. D. Gelb et al, Rep. Prog. Phys. 62, 1573, (1999).
2
G. S. Heffelfinger and M. E. Lewitt, J. Comput. Chem. 17,250, (1996).
3
L. D. Gelb and K. E. Gubbins, Langmuir 14, 2097, (1998).
4
L. D. Gelb and K. E. Gubbins, Langmuir 15, 305 (1999).
5
L. D. Gelb and K. E. Gubbins, Mol. Phys. 96, 1795, (1999).
Index of talks

Diffusion of oxygen and nitrogen in graphite slit-pores using Dual Control Volume Grand Canonical Molecular Dynamics (DCV GCMD).

Dr Karl P. Travis
Department of Chemical and Forensic Sciences, University of Bradford, Bradford BD7 1DP.
k.travis@bradford.ac.uk

Air is commercially separated into its major components via pressure swing adsorption using molecular sieving carbon as adsorbent [1]. Oxygen selectivities in the range 3 - 30 have been reported despite the fact that the kinetic diameters of oxygen and nitrogen differ by less than 0.03 nm. The separation mechanism is based on the different diffusion rates of the two species through the porous adsorbent. This important industrial application provides the motivation for the present study. Dual Control Volume Grand Canonical Molecular Dynamics (DCV GCMD) [2 - 4] has been used to study the diffusion of oxygen and nitrogen through graphite slit pores under an imposed chemical potential gradient. The effect of pore width on the absolute diffusion rates and their ratio has been studied at two different temperatures with the aim of gaining greater insight into the separation at the molecular level. A complicated picture emerges in which the various contributions to the effective transport diffusivities show different pore width dependency.

References

1
H. Juntgen, K. Knoblauch and K. Harder, Fuel , 69 , 817 (1981).
2
G. Heffelfinger, and F. van Swol, J. Chem. Phys. , 100 , 7548 (1994).
3
J. M. D. MacElroy, J. Chem. Phys. , 101 , 5274 (1994).
4
R. F. Cracknell, D. Nicholson and N. Quirke, Phys. Rev. Lett , 74 , 2463 (1995).
Index of talks

Computer Simulation of the Sorption of Krypton in Silicalite

Mark Calmiano, C. Richard A. Catlow and Robert G. Bell
Davy Faraday Research Laboratory, The Royal Institution Of GB, 21 Albemarle Street, London W1X 4BS, UK

A computational study of the sorption of krypton in silicalite has been undertaken using a range of simulation techniques embracing Monte Carlo docking, energy minimisation, canonical Monte Carlo, grand canonical Monte Carlo and molecular dynamics methods. Previous experimental and computational studies have predicted sorption sites for krypton in silicalite as well as adsorption energies and adsorption isotherms. Our computational study has identified sorption sites for krypton in silicalite to be in the straight, sinusoidal and channel intersections in agreement with previous work. Sites in the straight and sinusoidal channels are shown to be of a lower energy than those in the channel intersections. Calculated adsorption isotherms show type I, single step behaviour again in good agreement with previous experimental and computational work and subsequent mass distribution plots have revealed occupation of the straight, sinusoidal and intersection positions at pressures of 0.15 kPa and temperatures of 78K. Molecular dynamics yield mean square distribution plots under NVT conditions at 400K for loadings of 2, 7, 14 and 18 Kr/uc which show decreasing diffusion on increased loading. Diffusion coefficients have been calculated for the entire unit cell as well as in the individual x, y and z directions; an activation energy of 2.5 kJ mol-1 for krypton diffusion in silicalite is calculated.

Index of talks

Simulation of Interfaces on Polar and Non-polar Systems

José Alejandre 1
Simulación Molecular, Instituto Mexicano del Petróleo, Eje Central Lázaro Cárdenas 152, Apdo. Postal 14-805, 07730 México D.F., Mexico

  1. Permanent address: Departamento de Química, Universidad Autónoma Metropolitana-Iztapalapa, Apartado Postal 55-534, 09340 México D.F., Mexico

In this work we present results on the liquid-vapor and liquid-liquid interfaces. It is established that canonical Molecular Dynamics and Monte Carlo simulations of interfaces produce the same results as those methods where the interface is not present if the same potential model is used [1] When the potential is spherically truncated the contribution from the discontinuity to the components of the pressure tensor has to be included in the canonical simulations to obtain a constant normal pressure along the interface. Wrong surface tensions are obtained if this term is ignored. In canonical simulations of interfaces the coexisting and critical properties depend on the truncation of the potential because the difficulty of including the long range corrections. We present an Ewald method to calculate the short range interactions to solve that problem. By combining this method with that of electrostatic interactions allow us to simulate interfaces of polar molecular systems. In the liquid-liquid interface the surface tension is obtained as a function of temperature and a maximum is found [2], this is in contrast with the liquid-vapor equilibrium where the surface tension decreases monotonically. The maximum is explained in terms of the surface free energy and of the repulsive interaction between the liquid phases. We also present results of interfacial properties for molecular systems of interest in the oil industry. In particular we simulate binary mixtures of nitrogen/hydrocarbons [3] and aqueous solutions of monoethanolamine. Force fields of nitrogen and monoethanolamine [4] were obtained to simulate those mixtures.

References

1
``Computer simulations of liquid-vapor interface in Lennard-Jones fluids: Some questions and answers.'' A. Trokhymchuk and J. Alejandre J.Chem. Phys 111 , 8510 (1999).
2
``Interfacial tension behavior of binary and ternary mixture of partially miscible Lennard-jones Fluids: A molecular dynamics simulations'' E. Díaz-Herrera, J. Alejandre, G. Ramírez-Santigo, and and F. Forstmann J.Chem. Phys 110 , 8084 (1999).
3
``Thermodynamic and transport properties of nitrogen-butane mixtures'' J. L. Rivera, J. Alejandre, S. K. Nath, and J. J. dePablo Molec. Phys , 98 , 43 (2000)
4
``Force field of monoethanolamine'' J. Alejandre, J. L. Rivera, M. A. Mora, and V. de la Garza J. Phys. Chem. B , 104 , 1332 (2000).
Index of talks

Simulating inelastic neutron scattering from proteins by using a simplified dynamical model

Gerald R. Kneller
Centre de Biophysique Moléculaire, CNRS Orléans, France

The first part of the talk focuses on a simplified atomic description of protein dynamics close to an equilibrium configuration. It is shown that harmonic force fields, in which all non-bonded interactions are replaced by an unspecific pair-wise additive harmonic term with a distance-dependent force constant, can describe the vibrational dynamics over the whole frequency range [1]. A simulation study of the influence of molecular flexibility on DNA radiosensitivity demonstrates that the force field may also be used for DNA molecules [2].

The second part of the talk deals with a more realistic description of protein dynamics in which a harmonic model with friction is used to describe the residue dynamics. The parameters for the model are extracted from Molecular Dynamics simulations. Using C-phycocyanin with a hydration shell as an example, it is demonstrated that internal protein dynamics can be decomposed into (a) translational diffusion of the residues in an effective harmonic potential (Ornstein-Uhlenbeck process), (b) vibrational motions of residues in a local harmonic potential, (c) rotational rigid-body motions of the side chain of the side-chains, and (d) side-chain deformations, making only a minor contribution. The above processes are nearly uncorrelated, which is reflected in a corresponding factorization of the intermediate scattering function for neutron scattering [3 - 4]

References

1
``A simplified force field for describing vibrational protein dynamics over the whole frequency range'', K. Hinsen and G. R. Kneller, J. Chem. Phys. 111(24), 10766-10769 (1999)
2
``The influence of molecular flexibility on DNA radiosensitivity: A simulation study'', D. Viduna, K. Hinsen, and G. R. Kneller , Phys. Rev. E 61(6), (June 2000)
3
``Inelastic neutron scattering from damped collective vibrations of macromolecules'', G. R. Kneller, Chem. Phys., special issue ``Condensed Phase Structure and Dynamics : A combined neutron scattering and molecular modelling approach'', in press.
4
``Harmonicity in slow protein dynamics'', K. Hinsen, A.-J. Petrescu, S. Dellerue, M. Bellissent-Funel, and G. R. Kneller, Chem. Phys., special issue ``Condensed Phase Structure and Dynamics : A combined neutron scattering and molecular modelling approach'', in press.
Index of talks

Developing force-fields for simulations of polymer-oxide

Tiffany R. Walsh and Adrian P. Sutton
Dept. of Materials, Parks Rd., Oxford, OX1 3PH
tiffany.walsh@materials.ox.ac.uk

Polymer-inorganic substrate interfaces form a class of technologically important systems, serving a wide variety of industrial demands. Common substrates used in such applications are oxides and metals, and often a polymer deposited onto these substrates will chemisorb at the surface. What makes the simulation of chemisorbed polymers demanding is the specificity of the interactions we seek to model. We will outline our plan to develop a force-field which will be used in simulations at polymer-oxide interfaces. We have investigated what types of bond may form across the polymer substrate interface, using ab initio calculations, since these are short-ranged interactions. However, longer-ranged interactions are also important. Experimentalists have long observed in such interfaces a region in the polymer film between the chains at the substrate and chains in the bulk [1], denoted here as the ``interfacial layer'', which typically exhibits a distinct gradient in chain packing compared with the bulk. Earlier modelling studies have noted that chains may adsorb in highly strained conformations [2]. Kinetics calculations [3] revealed that the chains relaxed over very long timescales. Therefore, to probe the structure the interfacial layer, we must consider not only large system sizes (containing at least tens of polymer chains), but also long time scales. To do this effectively we need a force-field which is suitably coarse grained to capture the variety of interactions which are important in such systems, from long-ranged packing effects to the short-ranged effects such as chain flexibility and bond strength and geometry across the interface. We aim to link up the various length scales in this system, and strip out as much atomistic detail from the system, using mesogenic units where possible. We have generated data from ab initio calculations in order to parametrise such a coarse-grained force-field. Results of our ab initio calculations from our work on the polyimide--silica interface will be presented.

References

1
M. Grunze, G. Hähner, Ch. Wöll and W. Schrepp, Surf. Interf. Anal., 20 , 393 (1993)
2
J. Scott Shaffer, A. K. Chakraborty, M. Tirrell, H. Ted Davis and J. L. Martins, J. Chem. Phys., 95 , 8616 (1991)
3
P. M. Adriani and A. K. Chakraborty, J. Chem. Phys., 98 , 4263 (1993)
Index of talks

From ion conductors to intercalation compounds: The modelling of new materials

M. Saiful Islam
Department of Chemistry, University of Surrey, Guildford GU2 5XH, UK
Email: m.islam@surrey.ac.uk

Computational techniques are now well established tools for probing structural and transport properties of crystalline materials. This presentation will highlight recent developments by focusing on new complex oxides, which are generating considerable interest owing to both the potential applications (fuel cells, lithium batteries) and the fundamental fascination in transport phenomena. Contemporary work will be illustrated by accounts of studies on oxygen ion and proton conduction in perovskite-structured oxides and on lithium intercalation in manganate spinels. We have used atomistic simulation, molecular dynamics (MD) and first-principles techniques to obtain detailed information on their defect and dopant properties, and on the mechanisms of ion migration.

References

1
M. S. Islam J. Mater. Chem. 10 , 1027 (2000)
2
C. A. J. Fisher and M. S. Islam Solid State Ionics 118 , 355 (1999)
Index of talks

Modelling Benzene Diffusion in a Zeolite with Constraint Molecular Dynamics

T. R. Forester and W. Smith
Computational Science and Engineering Department, CCLRC Daresbury Laboratory, Daresbury, Warrington WA4 4AD, United Kingdom

The diffusion of benzene on silicalite-1 (as a model for ZSM5) has been studied by molecular dynamics. On account of the slow diffusion rate ( $ \sim$ 2x10-14 m2/s) ordinary MD approaches are inadequate in this system and the constraint dynamics method has been employed to enable an effective extension of time scale. Constraint procedures are used to determine the mean diffusion path which is combined with the constraint force to determine the free energy profile, thus revealing the adsorption sites and permitting the calculation of the free energy of activation for the benzene to ``hop'' from one site to another. Transition state theory supplies the means to calculate the rate constants, from which the diffusion coefficients can be determined by Monte Carlo simulation.

Index of talks

Application of histogram techniques to phase diagram calculations

A. L. Ferreira * and M. A. Barroso
Departmento de Física, Universidade de Aveiro, 3810-193 Aveiro, Portugal

A generalization of the multiple histogram method [1] to simultaneous temperature and volume extrapolations is presented [2]. From a series of canonical ensemble Monte-Carlo simulations, the method allows the determination of free energy differences between thermodynamic states. The computation of the absolute free energy for a suitable chosen reference thermodynamic state together with the proposed method can be used to obtain absolute free energies [3] for a range of temperatures and volumes. From the volume dependence of the Helmholtz free energy for each temperature, and for each coexistent phase, the application of the double tangent construction gives us the phase coexistence properties. The method can be applied both to the study of solid-fluid coexistence and to the study of liquid-vapor coexistence. We present test-bed results for the Lennard-Jones system. Applications of the method to other interesting physical systems have been considered (see for example [4]).

References

1
A. M. Ferrenberg and R. H. Swendsen, Phys. Rev. Lett. 61 , 2635 (1988); A. M. Ferrenberg and R. H. Swendsen, Phys. Rev. Lett. 63 , 1195 (1989); R. H. Swendsen, Physica A 194 , 53-62 (1993)
2
A L Ferreira and M A Barroso Phys. Rev. E 61 , 1195 (2000)
3
D. Frenkel and B. Smit Understanding Molecular Simulation, Academic Press, San Diego, (1996).
4
A L C Ferreira, J M Pacheco and J P Prates-Ramalho, ``Phase diagram of C60 from ab initio intermolecular potential'', to appear in J. Chem. Phys.
* This work was partially supported by the projects PRAXIS/2/2.1/299/94 and PRAXIS/2/2.2/FIS/302/94.

Index of talks

Temperatures: Old, New and Middle Aged.

Jack Powles 1, Gerald Rickayzen 1 and David Heyes 2

  1. The Physics Laboratory, University of Kent, Canterbury, England.
  2. The Chemistry Department, University of Surrey, Guildford, England.

We simulators have supposed for fifty years that we only need one temperature for Monte Carlo, TMC and one temperature for Molecular Dynamics, Tk the kinetic temperature and that both are, as near as dammit, the macroscopic thermodynamic temperature. Now Hans Rugh [1 , 2] has proposed a new way of getting temperature, Tn which is partially kinetic and partially configurational, but is unusually rigorous and moreover is microcanonical. This can lead to various expressions for temperature which are all equivalent, to order 1/N, for large systems in thermodynamic equilibrium. Of particular interest is Tconfig, which is a function only of the particle configurations [3] which is in fact a particular hypervirial temperature [4]. This new approach to temperature promises to be a useful development both in simulation [3 , 5, etc.] and in theories [6, etc.] and a key question is whether any or all of these temperatures are useful for systems NOT in thermodynamic equilibrium. These matters will be discussed and illustrated with particular reference to the computer simulation of liquids [7].

References

1
``Dynamical Approach to Temperature.'' H. H. Rugh Phys. Rev. Lett. 78 , 772-4 (1997).
2
``A geometrical, dynamical approach to thermodynamics.'' H. H. Rugh J. Phys. A 31 , 7761-70 (1998).
3
``Configurational temperature: verification of Monte Carlo simulations.'' B. D. Butler, O. Ayton, O. G. Jepps and D. J. Evans J. Chem. Phys. 109 , 6519-22 (1998).
4
``Classical and quantum mechanical hypervirial theorems.'' J. O. Hirschfelder J. Chem. Phys. 33 , 1462-6 (1960).
5
``Definition of temperature in equilibrium and nonequilibrium systems.'' G. P. Morris & L. Rondoni Phys. Rev. E 59 , 1A, R5-8 (1999).
6
``On the configurational temperature of simple fluids.'' A. Baranyai J. Chem. Phys. 112 , 3964-6 (2000).
7
J. G. Powles, G. Rickayzen & D. W. Heyes Molec. Phys. (2000). in preparation.
Index of talks

Simulation study of rod-like molecules with terminal dipoles and flexible tails

J.S. van Duijneveldt 1, A. Gil-Villegas 2, G. Jackson 3 and M.P. Allen 4

  1. School of Chemistry, Cantock's Close, Bristol BS8 1TS, England
  2. Universidad de Guanajuato, Inst Fis, Leon, Guanajuato 37150, Mexico
  3. Department of Chemical Engineering, Imperial College, Prince Consort Road, London SW7 2BY, UK
  4. H.H. Wills Physics Laboratory, Bristol University, Bristol, BS8 1TL, UK

A primitive model for small mesogenic molecules is proposed, consisting of three elements:

  1. a rigid rod-like core, modelled as a hard spherocylinder of length/diameter ratio L/D = 5;
  2. a flexible end group, consisting of 5 segments of length D, which is ``ideal'' in the sense that it has no volume;
  3. a terminal dipole, located in the end cap opposite to the flexible tail.

This model is studied using Monte Carlo computer simulation, and the dipolar interactions are evaluated using the reaction field method. The hard spherocylinder model displays four phases: isotropic, nematic, smectic-A and crystal. Previously, it was found that the addition of the terminal dipole to hard spherocylinders without tails greatly enhances the range of stability of the nematic phase, at the expense of the smectic-A phase [1]. Conversely, adding the flexible tail to hard spherocylinders without dipoles is found to suppress the nematic phase, whereas the smectic-A and crystal phase are little affected. Combining the effects of the terminal dipole and the flexible tail, all four phases survive. Because of the dipoles, the particles prefer to adopt a staggered antiparallel arrangement. In the smectic-A and crystal phases, this gives rise to interdigitation of the smectic layers. In the crystal phase a tendency towards columnar ordering is observed. The results are compared with experimental observations.

References

1
McGrother et al., J. Phys.: Condens. Matter 8 9649 (1996).
Index of talks

Grain Boundary Diffusion and the Reactive Element Effect

J. H. Harding and D. J. Harris
Dept. Physics and Astronomy, University College London, London, U.K.

Passive oxide films are a common method of protecting metal surfaces. A thin, dense film of oxide prevents oxygen, water and the metal from coming into contact. Some metals and alloys readily grow such films. Other metals grow thick, porous films but can be induced to grow passive films by the addition of a so-called ``reactive element'' [1]. The discovery of suitable elements has been largely a matter of chance and the mechanism by which they work is still controversial. A common theory is that they prevent ion transport through the grain boundaries of the growing film and thus induce the formation of thin, dense films rather than thick porous ones.

Despite its obvious importance at low and medium temperatures, there have been few simulations of grain boundary diffusion in oxides (one of the few examples is [2]). The structure of grain boundaries is complex, with many possible pathways. Also, the obvious strategy of a direct molecular dynamics simulations encounters the difficulty that the migration energies, although lower than the bulk, are still so high that unrealistic temperatures must be used (see, for example [3]). Mishin [4] has shown that a combination of an encounter model and kinetic Monte Carlo can be used to obtain grain boundary diffusion rates.

We present calculations of diffusion rates for grain boundary diffusion for NiO, Al2O3 and Cr2O3 and consider the effect of a variety of reactive elements. It is clear that the relative ion size is important in determining the effectiveness of the reactive element, but it is not simply a matter of making the difference between host and impurity ion sizes as large as possible.

References

1
J. Strawbridge and R. A. Rapp J. Electrochem Soc 141, 1905 (1994)
2
D. M. Duffy and P. W. Tasker Philos. Mag. A 54, 259 (1986)
3
T. Karakasidis and M Meyer Phys. Rev. B 55, 13853 (1997)
4
Y. Mishin Philos. Mag. A 72, 1589 (1995)
Index of talks

A dissipative particle dynamics method for modelling the packing and flow of granular materials

J. A. Elliott and A. H. Windle
Department of Materials Science & Metallurgy, University of Cambridge, Pembroke Street, Cambridge, CB2 3QZ, UK.

A method is presented for modelling the geometrical packing and rudimentary flow behaviour of mixtures of both spherical and non-spherical particles, which are models for filler particles used in the manufacture of polymer composites. The technique is based on the calculation of the dissipative dynamics [1 - 2] of an ensemble of fused soft spheres at constant temperature and pressure, using an adapted version of the DL_POLY molecular dynamics package developed by CCP5 [3] After validation of the method by comparison with analytical equations of state for monodisperse and binary mixtures of hard spheres, the random packing of various fused soft sphere particles, including fibres and cubes, is studied. The effects of mixing together particles of different shape and size is then examined, with the aim of developing an understanding of how to minimise the amount of void space in composites containing angular particles. Preliminary attempts will be made to characterise the dilatancy of such mixtures by applying crude shear forces to the simulations.

References

1
P. J. Hoogerbrugge and J. M. V. A. Koelman Europhys. Lett. 19 , 155-160. (1992)
2
J. M. V. A. Koelman and P. J. Hoogerbrugge Europhys. Lett. 21 , 363-368. (1993)
3
T. R. Forester and W. Smith DL_POLY molecular dynamics code CCP5 of the EPSRC. (1995)
Index of talks

Ion Transfer Processes Across Liquid | Liquid Interfaces

M. Natália D. S. Cordeiro, Pedro A. Fernandes, and José A. N. F. Gomes.
CEQUP/ Departamento de Química, Universidade do Porto, Portugal

Ion transfer processes across the interface between two immiscible liquids play a major role in areas like liquid chromatography, phase-transfer catalysis or drug-delivery problems, to name but a few. In spite of their great importance, our knowledge about the mechanism of the ion transfer is still scarce and most of what is known has been gathered by molecular simulations [1-4]. In fact, simulations are crucial due to the new and complementary of information they provide to experimental or other theoretical results. In this work, a systematic MD study is presented to address some fundamental questions regarding the mechanism of ion transfer across the interface between water and an organic phase. In particular, issues such as the potential of mean force for the ion transfer, the influence of the ion's size and charge, the ion's solvation shell exchange, and the nature of the organic solvent are focused. The transfer of several ions (namely Na+, K+, Rb+, Sr2+, N(CH3)4+ and I-) across two liquid interfaces (water | 2-heptanone and water | i-octane) are analysed. This study shows that the ion-transfer free-energy increases with the ion charge and decreases with its size. The transfer is a non-activated process and involves dragging of the ion's hydration shell when the ion is driven to the organic phase. The results obtained are in good agreement with the available experimental data.

References

1
K. Schweighofer and I. Benjamin, J. Phys. Chem. B , 99, 9974 (1995).
2
M. Lauterbach, E. Engler, N. Muzet, L.Troxler and G. Wipff, J. Phys. Chem. B , 102, 245 (1998).
3
Pedro A. Fernandes, M. N. D. S. Cordeiro and José A. N. F. Gomes, J. Phys. Chem. B , 103, 1176 (1999).
4
Pedro A. Fernandes, M. N. D. S. Cordeiro and José A. N. F. Gomes, J. Phys. Chem. B,, in press.
Index of talks

Colloidal clusters in shear flow near a wall

R. B. Jones
Department of Physics, Queen Mary and Westfield College, Mile End Road, London E1 4NS, United Kingdom

The motion of colloidal particles near surfaces or interfaces is important for understanding single particle adsorption and desorption processes as well as the time scales and morphology of multi-particle deposition [1-2]. Such phenomena are of technological importance in detergency or catalysis at solid boundaries. Externally imposed shear flows constitute an external probe field complementary to external electric or gravitational probe fields. We have recently developed a Stokesian dynamics code for simulating the motion of clusters of spherical colloidal particles near a wall [3]. Our algorithm computes the many-body mobility matrix, which incorporates many-body hydrodynamic interactions, to arbitrary accuracy while including two-body and particle-wall lubrication forces by an extension of the Brady-Bossis method [4]. We have used this to study the motion and deposition of clusters near a wall subject to external forces [5-6]. Recently , we have extended the algorithm to include an externally imposed shear flow. We discuss the form of our algorithm and certain instabilities that shear fields can introduce. Some examples of simulations with small clusters will be presented to illustrate both the effect of pure hydrodynamic forces as well as two-body potential interactions on the dynamics of sheared clusters near a wall.

References

1
G. Bossis, A. Meunier and J. D. Sherwood, Phys. Fluids A3 (1991) 1853.
2
A. T. Clark, M. Lal and R. B. Jones,``Dynamics of small clusters of particles bound to an interacting wall'', in Structure and Dynamics of Materials in the Mesoscopic Domain: Proceedings of the Fourth Royal Society-Unilever Indo-UK Forum in Materials Science and Engineering , edited by M. Lal, R. A. Mashelkar, B. D. Kulkarni and V. M. Naik (Imperial College Press and The Royal Society, London, 1999).
3
B. Cichocki, R. B. Jones, Ramzi Kutteh and E. Wajnryb, J. Chem. Phys. 112 (2000) 2548.
4
L. Durlofsky, J. F. Brady and G. Bossis, J. Fluid Mech. 180 (1987) 21
5
Robert B. Jones and Ramzi Kutteh, Phys. Chem. Chem. Phys. 1 (1999) 2131.
6
R. B. Jones and Ramzi Kutteh, ``Effect of hydrodynamic interactions on the irreversible deposition of colloidal particles: Deposition algorithm and simulations'', in Press, J. Chem. Phys .
Index of talks

EQUILBRIUM - An improved multiple-time-stepping scheme for molecular dynamics.

Dr. Sebastian Reich
Department of Mathematics and Statistics, University of Surrey, Guildford GU2 7XH, UK.

Simulations of the dynamics of biomolecules have been greatly accelerated by the use of multiple time-stepping methods. Indeed, numerical experiments have shown that time steps of 4 fs are possible for the slow forces but have also shown that a long time step of 5 fs results in a dramatic energy drift. To overcome this instability the slow part of the potential energy is evaluated at its equilibrium value with respect to the fast bonded interactions and one uses the gradient of this modified potential for the slow part of the forces [1 , 2]. The new method Equilibrium requires a SHAKE-type recursion per long time step and has been shown to be stable with a time step of 6-7 fs [2].

References

1
``Multiple time-scales in classical and quantum-classical molecular dynamics.'' S. Reich J. Comput. Phys. 151 , 49-73, (1999).
2
``Longer time steps for molecular dynamics.'' J. Izaguirre, S. Reich, and R. D. Skeel J. Chem. Phys. 110 , 9853-9864, (1999).
Index of talks

Constraint Methods for Stokesian Dynamics Simulations of Colloidal Particles with Hydrodynamic Interactions

Ramzi Kutteh
Department of Physics, Queen Mary and Westfield College, University of London, Mile End Road, London E1 4NS, UK
r.kutteh@qmw.ac.uk

Constraint methods [1 - 5] are widely used in Molecular Dynamics (MD) simulations. The presence of hydrodynamic interactions in Stokesian Dynamics (SD) simulations renders the inclusion of constraints into these simulations a more complicated affair than in the MD case. Algorithms for incorporating constraints into SD simulations, with the hydrodynamic mobility matrix computed by any desired scheme [6] are described and numerical results are presented. These constraint algorithms are useful, for example, for simulations of rigid/partially rigid colloidal clusters, possibly with adsorption effects, and for simulations of some polymer chain models.

References

1
M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids (Oxford University Press, 1992).
2
R. Kutteh and T. P. Straatsma, in Reviews in Computational Chemistry, edited by K. B. Lipkowitz and D. B. Boyd (Wiley, New York, 1998), Vol. 12.
3
R. Kutteh, J. Chem. Phys. 111 , 1394 (1999).
4
R. Kutteh, Comp. Phys. Comm. 119 , 159 (1999).
5
R. Kutteh and R. B. Jones, Phys. Rev. E 61 , 3186 (2000).
6
B. Cichocki, R. B. Jones, R. Kutteh, and E. Wajnryb, J. Chem. Phys. 112 , 2548 (2000).
Index of talks

Modelling the formation of corrosion inhibitor films and wax deposition on metal oxide surfaces

M. A. San-Miguel and P. M. Rodger
Department of Chemistry, University of Warwick, Coventry CV4 7AL UK
M.A.San-Miguel-Barrera@warwick.ac.uk, smiguel@cica.es

Efficiency of the transport processes in the oil and gas pipelines constitutes an important aim for the industry, and there are various phenomena that reduce it such as corrosion, scale and wax deposition, and clathrate hydrate formation. Most of the works concerning these processes have been studied separately and more extensively by experimental methods [1 - 2]. Our major interests are on the formation of corrosion inhibitor films and its undesirable effect inducing wax deposition [3]. In this work, by using molecular dynamics techniques we have firstly focused on wax crystal growth on different iron oxide surfaces [4] and the formation of corrosion inhibitor monolayers. Furthermore, wax deposition on these films has been considered in absence and in presence of a representative solvent as heptane.

References

1
S. Ramachandran, B-L, Tsai, M. Blanco, H. Chen, Y. Tang, and W. A. Goddard III Langmuir 12 , 6419 (1996)
2
S. Ramachandran, B-L, Tsai, M. Blanco, H. Chen, Y. Tang, and W. A. Goddard III J. Phys. Chem. A 101 , 83 (1997)
3
M. A. San-Miguel and P. M. Rodger J. Molec. Struct. (Theochem) 000 , 000 (2000)
4
M. A. San-Miguel and P. M. Rodger Molec. Simul. 000 , 000 (2000)
Index of talks

Partial Derivatives of Molecular Dynamics Calculations

J. Stefanovi{\'{c\/}} and C. C. Pantelides
Centre for Process Systems Engineering, Imperial College of Science, Technology and Medicine, London SW7 2BY, United Kingdom

Molecular dynamics can be viewed as a deterministic mathematical mapping between, on one hand, the force field parameters that describe the potential energy interactions and input macroscopic conditions, and, on the other, the calculated corresponding macroscopic properties of the bulk molecular system.

The differentiability of such a mapping in conventional molecular dynamics calculations is affected by the discontinuities introduced through the use of the minimum image convention and other simplifications commonly employed in the calculation of interparticle potential and forces (e.g. the cut-off potential, the cut-and-shift potential etc.). An alternative framework that employs a modified force function which is almost everywhere continuous and differentiable, and exhibits a natural periodicity is proposed. These characteristics make it possible to apply standard methods for the computation of the partial derivatives of the molecular dynamics mapping based on the integration of either the adjoint equations, or the sensitivity equations of the classical Newtonian equations of motion [1]. We present procedures for these computations in the standard microcanonical (N, V, E) ensemble, and compare the computational efficiency of the two approaches.

As an illustration, we apply these techniques to a system of flexible hydrocarbon molecules described by the NERD potential [2], computing the partial derivatives of the calculated temperature and pressure with respect to density, energy and all potential function parameters. These derivatives are computed within the same degree of accuracy as the calculated quantities themselves, and are, therefore, numerically consistent with them. Such accurate partial derivative information can be valuable in a wide variety of applications, such as the use of sophisticated techniques for the estimation of values of the potential parameters.

References

1
J. Stefanovi{\'{c\/}} On the Mathematics of Molecular Dynamics . PhD thesis, University of London, 2000.
2
S. K. Nath, F. A. Escobedo, and J. J. de Pablo. On the simulation of the vapor-liquid equilibria for alkanes. J. Chem. Phys. 108, 9905–9911 (1998).
Index of talks

Molecular Dynamics simulations of the growth of paraffin crystals with adsorbed inhibitors

Dr. D. M. Duffy and Dr. P. M. Rodger
Dept. of Chemistry, Warwick University, Coventry, CV4 7AL UK

The crystallisation and deposition of high molecular weight alkanes present in oil cause significant problems in industry. Certain comb shaped polymers are known to have an effect on both the crystallisation temperature [1 - 2] and the rate of paraffin deposition [3]. A comprehensive understanding of the mechanism for these effects should aid in the development of more effective deposition inhibitors.

We have modelled the interaction between an octamer unit of one such polymer (poly-octadecylacrylate) and the faces of an octacosane crystal. The polymer unit was found to interact strongly with certain crystal surfaces, with excellent crystallographic matching [4]. Crystal growth was simulated by the addition of alkane crystal layers to surfaces with adsorbed inhibitors. The inhibitor was found to act as a source of defects in the growing crystal, causing substantial disruption to the crystal structure. Such disruption would slow subsequent crystal growth and contribute to the inhibitor mechanism.

References

1
R. Kern and R. Dassonville J. Cryst. Growth 116 , 191 (1992)
2
X. Ding G. Qi and S. Yang Polymer 40 , 4139 (1999)
3
A. J. Hennessy, A. Neville and K. J. Roberts J. Cryst. Growth 830 , 198-199 (1999)
4
D. M. Duffy and P. M. Rodger, to be published
Index of talks

Generalization of Nosé dynamics

A. C. Bra{\'{n\/}}\kern.05emka and K. W. Wojciechowski
Institute of Molecular Physics, Polish Academy of Sciences, Smoluchowskiego 17, 60-179 Pozna{\'{n\/}}\kern.05em, Poland

The isothermal dynamics based on the Nosé and Nosé-Hoover methods are investigated. Their properties and criteria for selecting different isothermal dynamics determined by various scaling functions of the thermostat s-variable involved in the generalized Nosé Hamiltonian [1 , 2], are tested with molecular dynamics simulations and examined analytically. It is shown that time scaling is related to the scaling of the momenta [2 , 3]. The general form of the generalized Nosé-Hoover ($ \cal {G}$NH) equations of motion is discussed. The $ \cal {G}$NH equations with h = s$\scriptstyle \alpha$, u = s$\scriptstyle \vartheta$, and v $ \sim$ lns are studied in detail. With such a choice of the functions the extended Nosé-Hoover ($ \cal {E}$NH) equations are expected to produce more chaotic phase-space dynamics than the NH equations. For a system away from equilibrium the $ \cal {E}$NH thermostat is not able to provide dynamics consistent with the target temperature and thus, the $ \cal {G}$NH approach reduces to the original Nosé-Hoover thermostat. A simple modification [4] of the $ \cal {E}$NH equations is proposed which makes the $ \cal {E}$NH thermostat applicable also to nonequilibrium states.

References

1
J. Jellinek and R. S. Berry Phys. Rev. A 38, 3069 (1988)
2
J. Jellinek J. Chem. Phys. 92, 3163 (1988)
3
S. Nosé Prog. Theor. Phys. Suppl. 103, 1 (1991)
4
A. C. Bra{\'{n\/}}\kern.05emka Phys. Rev. E, 61, 4769 (2000)
Index of talks

Solvation Shell Exchange Along the Ion Transfer Across Immiscible Liquids

M. Natália D. S. Cordeiro, Pedro A. Fernandes, and José A. N. F. Gomes.
CEQUP/Chemistry Department, Faculty of Sciences, University of Porto, Rua do Campo Alegre, 687, 4169-007 Porto, Portugal

Solvation Shell of ion. Figure 1

A detailed description of the ion transfer mechanism across liquid | liquid interfaces is essential to many scientific areas. Extraction processes, drug delivery to cells, building of selective electrodes or kinetics of mass transport across liquid phases are some of the processes that are intimately connected with the mechanism of ion transfer [1-2].

The key step for understanding such mechanism is the exchange of the ion's solvation shell, as the ion moves from one solvent to the other.

This work describes the results of molecular dynamics studies for the transfer of several ions across the water | 2-heptanone and water | i-octane [3-4]. An analysis of the ionic association as a function of the ion-interface distances is performed by looking at the corresponding radial distribution functions and coordination numbers. It is seen that the ions drag significant amounts of water during the transfer. Inside the organic phase, the ions still keep part of their hydration shells and form a mixed water/organic solvent shell. The influence of the ion's size and charge as well as the hydrophobicity of the organic solvent in the exchange of the ion's solvation shell are discussed.

References

1
``Chemical Reactions and Solvation at Liquid Interfaces: A Microscopic Perspective'' I. Benjamin Chem. Rev. 96 , 1449 (1996)
2
``Molecular Structure of Aqueous Interfaces'' A. Pohorille and M. Wilson J. Molec. Struct. (THEOCHEM) 284 , 271 (1993)
3
``Molecular Dynamics Study of the Transfer of Iodide Ion Across Two Liquid Interfaces'' P. A. Fernandes, M. N. D. S. Cordeiro and J. A. N. F. Gomes J. Phys. Chem. B 103 , 8930 (1999)
4
``Influence of Ion Size and Charge in the Ion Transfer process Across a Liquid | liquid Interface'' P. A. Fernandes, M. N. D. S. Cordeiro and J. A. N. F. Gomes J. Phys. Chem. B , 104 , 2278 (2000).
Index of talks

Vapor-Liquid Equilibrium of systems of hard spherocylinders plus a square well

B. Martínez-Haya 1, A. Cuetos 1, L. F. Rull 2 and S. Lago 1

  1. Universidad Pablo de Olavide, Facultad Ciencias Experimentales, Dpto Ciencias Ambientales, Ctra.Utrera Km 1, 41013 Seville (Spain)
  2. Universidad de Sevilla, Dpto.Física Atómica y Molecular, Área de Física Teórica Facultad de Física, Aptdo. 1065, 41080 Seville (Spain)

Gibbs ensemble Monte Carlo simulations have been performed for systems of square-well spherocylinders of different length-to-breadth ratio. Simulations are used as a test to check the validity of a recent perturbation theory proposed for this kind of systems [1]. In addition, the results are contrasted to similar simulations performed for a Kihara fluid of elongated molecules [2] with the aim of evaluating the possible use of the spherocylinder plus square-well interaction as a reference potential for a perturbative treatment of more realistic fluids. The system was found particularly difficult to simulate and special care was taken using the experience of a recent work for a related system [3].

References

1
D. C. Williamson and Y. Guevara J. Phys. Chem. 103 , 7522 (1999)
2
C. Vega, S. Lago, E. De Miguel and L. F. Rull J. Phys. Chem. 96 , 7431 (1992)
3
E.Ávalos,R. Espíndola,F. del Río, L. F. Rull, G. Jackson and S. Lago (to be published).
Index of talks

Proton migration and dopant substitution in the CaZrO3 proton conductor

R. A. Davies 1, M. Saiful Islam 1, A. V. Chadwick 2, G. E. Rush 2 and J. D. Gale 3

  1. Department of Chemistry, University of Surrey, Guildford, GU2 7XH, UK.
  2. School of Physical Sciences, University of Kent, Canterbury, CT2 7NR, UK.
  3. Department of Chemistry, Imperial College of Science, Technology and Medicine, South Kensington, SW7 2AY, UK.

High temperature proton-conducting oxides such as In-doped CaZrO3 [1 , 2] are of considerable importance due to their wide variety of electrochemical applications in fuel cells and as hydrogen sensors. Key questions such as activation energies, bonding changes, transition state structures and reaction pathways can only be realistically tackled by quantum mechanical methods which can effectively handle the changes in local bonding. Using a high quality planewave pseudopotential approach [3 , 4] embodied within CASTEP, we have identified stable proton configurations within a hypothetical Ca4Zr4O12H supercell. Low activation barriers ( $ \sim$ 0.15 eV) have been predicted for proton hopping between adjacent octahedra, while larger activation barriers (> 0.5 eV) have been predicted for proton hopping events within the same octahedra. Although the addition of dopants is essential for proton uptake, certain dopants will trap protons, causing an increase in the overall activation energy for proton migration. Trapping energies have been obtained for proton-nearest neighbour dopant clusters, enabling an initial qualitative prediction of proton hopping energetics in doped materials.

References

1
``Dopant and proton incorporation in perovskite-type zirconates'' R. A. Davies, M. S. Islam, J. D. Gale Solid State Ionics 126 , 323 (1999)
2
``Cation dopant sites in the CaZrO3 proton conductor: a combined EXAFS and computer simulation study'' R. A. Davies, M. S. Islam, A. V. Chadwick, G. E. Rush Solid State Ionics 130 , 115 (2000)
3
``Proton diffusion and Defect Association in the Orthorhombic Perovskite CaZrO3: a First Principles Simulation Study'' M. S. Islam, R. A. Davies, J. D. Gale Chem. Mater. (in preparation).
4
``Ionic transport in ABO3 perovskite oxides: a computer modelling tour'' M. S. Islam J. Mater. Chem. 10 , 1027 (2000)
Index of talks

Dual-control-volume grand canonical molecular dynamics simulation of transport diffusion of binary mixtures in carbon nanotubes

T. Düren 1, F. J. Keil 1, N. A. Seaton 2

  1. Department of Chemical Engineering, Technical University of Hamburg Harburg, Eissendorfer Str. 38, D-21073 Hamburg, Germany
  2. School of Chemical Engineering, University of Edinburgh, King's Buildings, Mayfield Road, Edinburgh EH9 3JL, United Kingdom

Carbon nanotubes are cylindrical structures which consist of coaxially rolled graphite sheets. The inner diameters range from 1 to 10 nm [1]. With the inner diameters in the nanometer range carbon nanotubes can be potentially used as molecular sieves, porous adsorbents and membranes. In order to further develop these applications it is necessary to have a clear understanding at the molecular level of the behaviour of molecules confined in nanotubes. The transport of fluids through nanopores is mainly diffusive. As diffusion is often the rate controlling step it is our aim to understand how transport diffusion in carbon nanotubes is influenced.

The driving force for transport diffusion is a gradient in chemical potential. Dual control volume grand canonical molecular simulations DCV-GCMD) mimic the real situation by establishing a gradient in chemical potential over the pore. The concentration in the control volumes at each end of the pore is kept constant by periodically performing a number of grand canonical Monte Carlo (GCMC) steps. These steps are followed by Molecular Dynamics (MD) steps throughout the pore that describe the actual movement of the particles inside the pore. DCV-GCMD simulations therefore allow the direct simulation of transport diffusion and the study of transport diffusion on a molecular level [2 - 4]. Using this simulation method, we studied the influence of a range of adjustable parameters in transport diffusion like pressure, temperature or mixture composition of different binary mixtures in order to optimise fluid separations.

References

1
K. Tanabe, T. Yamabe and K. Fukui (Eds), The Science and Technology of Carbon Nanotubes (Elsevier, Amsterdam, 1999)
2
G. Heffelfinger and F. van Swol J. Chem. Phys. 100 , 7548 (1994)
3
J. M. D. MacElroy J. Chem. Phys. 101 , 5274 (1994)
4
R. F. Cracknell, D. Nicholson and N. Quirke Phys. Rev. Lett 74 , 2463 (1995)
Index of talks

Local reorientation in polymer melts

Roland Faller and Florian Mueller-Plathe
Max-Planck-Institut fuer Polymerforschung, Ackermannweg 10, D-55128 Mainz, Germany

Recent NMR experiments [1] suggest in entangled polymer melts a quite high degree of local order. This phenomenon is investigated in simple bead-spring polymer melts with and without intrinsic stiffness. The static correlations increase with increasing local stiffness [2]. Additionally the reorientation depends on stiffness as well as on entanglements. It slows down with chain length and changes the dynamics with increasing stiffness [3]. A new type of dynamics is found if the local stiffness increases a (low) level. Reptation persists but is much stronger pronounced. Chains move strictly along their contour if entanglement and persistence length approach each other [4]. The entanglement length decreases with stiffness.

References

1
R. Graf, A. Heuer, and H. W. Spiess Phys. Rev. Lett 80 , 5738 (1998)
2
R. Faller, A. Kolb, and F. Mueller-Plathe Phys. Chem. Chem. Phys. 1 , 2071 (1999)
3
R. Faller, F. Mueller-Plathe, and A. Heuer submitted to Macromolecules
4
R. Faller submitted to Phys. Rev. Lett.
Index of talks

Discrete Element Simulation of Fine Particle Granular Materials with Chemical Discernment

Torsten Gröger 1, Ugur Tüzün 1, David M. Heyes 2

  1. Dept Chemical & Process Engineering, School of Engineering in the Environment, University of Surrey, Guildford, GU2 7XH, UK
  2. Department of Chemistry, School of Physics and Chemistry, University of Surrey, Guildford, GU2 5XH UK.

The design of engineering systems for reliable processing of powder materials requires can be assisted by the modelling of material behaviour. Often, the approach to model a powder or bulk material by means of continuum mechanical (macroscopic) methods fails. To improve existing or develop new macroscopic models, the micro- and mesoscopic behaviour must be known. For that reason the Discrete Element Method which is very similar to Molecular Dynamics) was developed more than two decades ago [1]. Promoted by the fast development of computer technology, this method has become a popular tool for theoretical investigations on powders and bulk materials [2].

Though, in the last years three-dimensional samples with irregular shaped particles and cohesive forces of various types have been investigated there are many problems still not well understood. Cohesive forces acting between the particles has a major influence on the macroscopic behaviour of fine powders. Various microscopic force models (e.g. for Van-der-Waals-forces [3 , 4] and liquid bridges [5]) will be tested, compared and assessed. The developed software will allow the handling of regular and irregular shaped particles assembled from component spheres. To link the microscopic forces via the mesoscopic (particle) scale to a macroscopic scale, direct and indirect shear experiments will be simulated. Amongst other things it is anticipated that we will gain insights into the experimentally observed dependency of flow behaviour on the stress and consolidation history of powder materials [6]. For the second phase we plan to investigate dynamical agglomeration processes. The growth of agglomerates with time, and their dependency on environmental conditions (e.g. humidity) and the properties (e.g. hardness) of these agglomerates, will be of special interest.

References

1
``A discrete numerical model for granular assemblies'' P. A. Cundall, O. D. L. Strack Geotechnique 29 1 , 47-65 (1979)
2
``Partikelmechanische Untersuchungen zur senkrechten Schlauchgurtförderung'' T. Gröger PhD Thesis Otto-von-Guericke-University of Magdeburg, Logisch GmbH, Germany, (1999)
3
``Surface energy and the contact of elastic solids'' K. L. Johnson, K. Kendall, A. D. Roberts Proc. R. Soc. Lond. A 324 , 301-313 (1971)
4
``Effect of Contact Deformations on the Adhesion of Particles'' B. V. Derjaguin, V. M. Muller, Yu. P. Toporov J. Coll. Interface. Science 53 , 314-326 (1975)
5
``A theoretical study of the liquid bridge forces between two rigid spherical bodies'' G. Lian, C. Thornton, M. J. Adams J. Colloid Interface Sci. 161 138-147 (1993)
6
``Testers for Measuring Flow Properties of Particulate Solids'' J. Schwedes Proc. Int. Symp. Reliable Flow of Particulate Solids III Porsgrunn, Norway, 3-40 (1999)
Index of talks

Intermediate-time Tracer-diffusion of Non-spherical Brownian Particles

F. de J. Guevara-Rodríguez 1 and M. Medina-Noyola 2

  1. Coordinación de Simulación Molecular, Instituto Mexicano del Petróleo, Eje Lázaro Cárdenas 152, 07730, México Distrito Federal, México.
  2. Instituto de Física ``Manuel Sandoval Vallarta'', Universidad Autónoma de San Luis Potosí, Alvaro Obregón 64, 78000 San Luis Potosí, S.L.P., México

The time-dependent tracer-diffusion properties of a non-spherical Brownian particle that interacts with a suspension of spherical particles are studied in terms of an idealized but non-trivial model system for which the predictions of the Generalized Langevin Equation approach to tracer diffusion can be calculated, and compared with the results of a computer simulation experiment. In the model, the non-spherical particle is represented by a linear array of NT (=2 or 3) spherical particles with nearest-neighbour separation $ \Delta$L. For this model, we calculate the rotational and the (transversal and longitudinal) translational mean squared displacements, both, directly from the computer simulation, and approximately using the Generalized Langevin Equation approach. The theory is found to reproduce qualitatively and quantitatively the main features of the results of the simulation experiment for these properties.

Index of talks

Modelling the Uptake of Magnesium into Calcite

N. H. de Leeuw 1 , J. H. Harding 2 and S. C. Parker 3

  1. Department of Chemistry, University of Reading, Whiteknights, Reading RG6 6AD
  2. Department of Physics and Astronomy, University College London Gower Street, London WC1E 6BT
  3. Department of Chemistry, University of Bath, Claverton Down, Bath BA2 7AY

Magnesium is often found in large quantities in calcite crystal (CaCO3), which is isomorphic with both magnesite (MgCO3) and dolomite, a mixed calcium magnesium carbonate. We employed atomistic energy minimisation techniques to model the absorption and segregation of magnesium ions to the low-index surfaces of calcite. Magnesium ions are calculated to absorb at the surfaces from solution, with the absorption energies being surface dependent, due to distinct relaxations of the different surfaces. The calculated absorption energies are large, due to the close coordination of the water molecules to the magnesium-substituted surfaces, which ensures that the magnesium ions prefer to remain at the surface rather then segregate into the bulk crystal. However, segregation energies for the magnesium ions in second and further layers of some surfaces are positive, indicating that once a calcium carbonate layer has overgrown the substituted surface layer segregation to the bulk is energetically possible.

As crystal growth occurs from steps and dislocations, we also used classical molecular dynamics simulations to model growth of MgCO3 units onto two experimentally found steps of the major calcite cleavage plane. We found that it is energetically more favourable to grow MgCO3 rather than CaCO3 molecules onto the step edges. At the fastest-growing step the process of adding MgCO3 at the edge is even exothermic. Hence, on thermodynamics grounds it is easy to incorporate magnesium into the calcite crystal.

Index of talks

Monte Carlo simulation of the adsorption of water/hydrocarbon mixtures on activated carbon

M. Jorge, C. Schumacher and N. A. Seaton
Department of Chemical Engineering, University of Edinburgh, King's Buildings, Mayfield Road, Edinburgh EH9 3JL, United Kingdom.

Our aim is to develop a method for the prediction of the adsorption of mixtures of water and light hydrocarbons on activated carbon. This problem is highly relevant to several practical applications, such as air purification. In such applications, the objective is to adsorb organic species, but water also adsorbs competitively in a way that is difficult to predict. Our approach is to use Monte Carlo simulation to identify the features of the activated carbon that most strongly affect the adsorption of these substances, and then to make quantitative predictions of the extent of adsorption and the composition of the adsorbed phase.

Since hydrocarbons are non-polar molecules, their adsorption will be affected mainly by the physical structure of the adsorbent. Hydrocarbon adsorption can therefore be described by a model that considers the structural heterogeneity of the carbon (such as the size distribution of a network of slit-shaped pores). On the other hand, the behaviour of water - a polar molecule - is strongly affected by any polar surface groups that appear as a consequence of the activation process. Thus, for the adsorption of water/hydrocarbon mixtures the surface polarity must be taken into account along with the pore size distribution.

To accomplish our goal, it is essential to fully understand the way in which polar surface sites affect the adsorption of both water and the non-polar hydrocarbons. We have carried out Grand Canonical Monte Carlo simulations of adsorption of pure water and water/hydrocarbon mixtures in model pores with different kinds of structural and chemical heterogeneities. These simulations allowed us to determine in which way those heterogeneities affect the adsorption. The results obtained form a basis for the development of a method for the prediction of adsorption of water/hydrocarbon mixtures in real activated carbon, using a limited set of experimental data as an input.

Index of talks

An molecular dynamics investigation of the use of surface waves in molecular desorption from graphite

Michelle Kerford and Roger Webb
Electronic Engineering Department, University of Surrey, Guildford GU2 7XH

A molecular dynamics investigation into the effects of energetic fullerene impacts on graphite with molecular overlayers has been performed. The impact of the fullerene upon the surface causes the propagation of an acoustic wave at certain energies. Such a wave can be used to eject molecules contained within an overlayer. The energy and size of the fullerene used can control the ejection process.

References to this work may be found on the web page:
http://www.ee.surrey.ac.uk/SCRIBA/simulations/index.html

Index of talks

Effect of short- and long-range forces on the orientational structure of fluid water, acetone, and carbon dioxide

Ji{\v{r\/}}\kern.05emí Kolafa and Ivo Nezbeda
E. Hála Laboratory of Thermodynamics, Institute of Chemical Process Fundamentals, Academy of Sciences, 16502 Praha 6 – Suchdol, Czech Republic
E-mail: jiri@icpf.cas.cz, ivonez@icpf.cas.cz

Effect of short- and long-range interactions on the structure of selected polar fluids has been studied in detail by computing the following properties:

  • full pair correlation function, visualized in 3D [1 , 2 , 3]
  • site-site correlation functions
  • two-dimensional site-site correlation functions [2 , 3]
  • dipole-dipole correlation function (2nd-order axis-axis function for CO2)
  • radial Kirkwood gl -factors (running integrals)
  • dielectric constant (not for CO2).

For water, two model potentials (TIP4P and ST2), and their short-range versions [4] have been considered at ambient, elevated, and supercritical conditions. The Ewald summation under different conditions has been used to investigate also their effect on results. Similar investigations have been performed for simple models of acetone [5] and CO2 [6].

An analysis of the results shows that although all site-site correlation functions for the short- and long-range systems are similar, the orientational ordering in systems of different range may be considerably different, this evidence being provided mainly by the dipole-dipole correlation function and the radial Kirkwood factor. For water [3], the orientational ordering is only short-range in long-range systems, whereas in short-range systems the hydrogen bonding gives rise to a damped long-range regular pattern of alignment. Nonetheless, the resulting dielectric constants for the short- and long-range systems coincide within the combined error bars. All findings are more pronounced at low temperatures but they are otherwise only marginally temperature and density dependent.

References

1
I. M. Svishchev, P. G. Kusalik, J. Chem. Phys. 99, 3049 (1993)
P. G. Kusalik, I. M. Svishchev, Science 265, 1219 (1994)
2
http://www.icpf.cas.cz/jiri/water
3
J. Kolafa, I. Nezbeda, Mol. Phys., in the press.
4
I. Nezbeda, J. Kolafa, Mol. Phys. 97, 1105 (1999)
5
P. Jedlovszky, G. Pálinkás, Mol. Phys. 84, 217 (1995)
6
J. G. Harris, K. H. Yung, J. Chem. Phys. 99, 12021 (1995)
Index of talks

Force field and molecular dynamics of molten AlCl3/LiI and AlCl3/LiSCN

Yi-Chia Lee 1, 2, Ji{\v{r\/}}\kern.05emí Kolafa 3, 1, and Larry A. Curtiss 2

  1. Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, IL 60208, USA
  2. Chemistry and Material Science Division, Argonne National Laboratory, Argonne, IL 60439, USA
  3. E. Hála Laboratory of Thermodynamics, Institute of Chemical Process Fundamentals, Academy of Sciences, 16502 Praha 6 – Suchdol, Czech Republic
E-mail: jiri@icpf.cas.cz, yi-chia@chem.nwu.edu

Adducts of aluminium chloride with different alkali metal salts are crystalline materials with low melting points (LiI/AlCl3) or glasses (LiSCN/AlCl3). Complexing of anions with AlCl3 delocalizes the electric charge and thus increases the ionic conductivity. These systems could therefore be useful as electrolytes in high energy density batteries.

The proposed force field is based on the exp-6 (Buckingham) potential with the Busing combining rules for the exp-term and the r-6 attraction for anions. The Li+-anion interactions are modified by an r-12 term. The anions are polarizable and the induced dipole is damped at close cation-anion separations by the shell-core model [1]. The model of the thiocyanate anion (•–S–•––C––•–N–•) uses four auxiliary charges (•) and polarizability tensors at S and N. The parameters for Al3 +, Cl-, and I- are taken from [1], the model of Li+ is based on the LiCl and LiI crystals and the properties of the LiCl pair calculated at the MP2/6-31G* level, and the model of SCN- is fitted to geometry and vibrational spectra (MP2/6-31G*) of SCN- [2] and different conformers of AlCl2SCN, AlCl3SCN-, and LiSCN [3]. The MD calculations are performed using MACSIMUS [4] implementing a novel predictor-corrector scheme for polarizable force fields [5]. Various structural characteristics (correlation functions, coordination numbers, oligomer analysis) as well as diffusion coefficients and conductivity are measured and analyzed.

References

1
Z. Akdeniz, G. Pastore, M. P. Tosi, Phys. Chem. Liq. 32, 191 (1996)
Z. Akdeniz, G. Pastore, M. P. Tosi, Nuovo Cimento 20, 595 (1998)
2
P. W. Schultz, G. E. Leroi, J. F. Harrison, Molec. Phys. 88, 217 (1996)
S. T. Howard, Molec. Phys. 85, 395 (1995)
3
T. Veszprémi, T. Pasinszki, M. Fehér, J. Am. Chem. Soc. 116, 6303 (1994)
4
http://www.icpf.cas.cz/jiri/macsimus .
5
J. Kolafa, Molec. Simul. 18, 193 (1996)
Index of talks

Particle Simulations of Colloidal Particle Flocculation and Gel Formation

J. F. M. Lodge 1 and D.M. Heyes 2

  1. Present address: University of Delaware, Department of Chemical Engineering, Colburn Laboratory, Newark, Delaware 19716, U.S.A.
  2. Department of Chemistry, School of Physics and Chemistry, University of Surrey, Guildford, GU2 5XH UK.

Brownian Dynamics, BD, simulations have been used to model the structural evolution and rheology of model attractive spherical colloidal particles as they self-assemble into long-range networks. The procedure used was to 'quench' the particles from a supercritical state point into the vapour-liquid or vapour-solid parts of their phase diagrams. The solids volume fractions $ \phi$ , were in the range 0.05 - 0.20. The interactions between the model colloidal particles were of a generalised Lennard-Jones (n : m) form: $ \phi$(r) = 4$ \epsilon$[($ \sigma$/r)n - ($ \sigma$/r)m] where r is the separation between the particle centres and $ \sigma$ and $ \epsilon$ set the distance and energy scales of the particles, respectively. Simulations were performed using 12 : 6, 24 : 12 and 36 : 18 potentials. Along this series the attractive part of the potential becomes shorter ranged. We explored the effects of quenching the systems into regions where the (original) single phase was unstable and consequently it proceeded to assemble into 'vapour-liquid' or (at lower temperatures) 'solid-liquid' mixtures. These systems developed a gel-like morphology during the simulations, with the aggregate morphology and rheology sensitive to the range of the attractive part of the potential and the position in the phase diagram of the quench (see refs. [1 - 4] for early publications from this programme of work). The long-range 12 : 6 potential induced compact structures with thick filaments, whereas the systems generated using the shorter-ranged 24 : 12 and 36 : 18 potentials persisted in a more diffuse network for the duration of the simulations and evolved more slowly with time. The rheology of these systems was characterised using the linear shear stress relaxation function, Cs(t), computed using a Green-Kubo fluctuation formula. The rheology of many of the systems displayed gel-like viscoelastic features, especially for the long-range attractive interaction potentials, which manifested a non-zero plateau in Cs(t), the so called equilibrium modulus, Geq, considered a useful indicator of a gel. The infinite frequency shear rigidity modulus, G$\scriptstyle \infty$, was extremely sensitive to the form of the potential. Despite being the most short lived, the 12 : 6 potential systems gave the most pronounced gel-like rheological features, which suggests that the traditional picture of a particle gel as being formed by thin filamentary networks might require reconsideration. The local and long-range structure, and rheology showed similarities with real gels. All properties were found to be quite sensitive to the range of the potential. In this poster the recent results and conclusions are presented.

References

1
J. F. M. Lodge and D. M. Heyes, J. Chem. Soc., Faraday Trans. 93 , 437-448 (1997).
2
J. F. M. Lodge and D. M. Heyes, Molec. Sim. 18 , 155-177 (1996).
3
J. F. M. Lodge and D. M. Heyes, J. Chem. Phys. 109 , 7567-7577 (1998).
4
J. F. M. Lodge and D. M. Heyes, J. Rheol. 43 , 219-244 (1999).
Index of talks

Simulation study of equilibrium and dynamics of linear polar molecules sorbed in the ORTHO, PARA and MONO structures of silicalite.

K. Makrodimitris 1,2, G. K. Papadopoulos 1, N. Kanellopoulos 1, K. Philippopoulos 2 and D. N. Theodorou 3

  1. Institute of Physical Chemistry NCSR ``Demokritos'', Athens, HELLAS
  2. Department of Chemical Engineering National Technical University of Athens, Athens, HELLAS
  3. Department of Chemical Engineering University of Patras , PATRAS-HELLAS

We have used grand canonical Monte Carlo (GCMC) in order to study the adsorption isotherms of carbon dioxide and nitrogen sorbed in silicalite crystal. The calculation of the diffusivity values inside the crystal, was performed via equilibrium molecular dynamics (EMD) for various densities using the LEN algorithm [1].

The zeolite was modeled as a framework of fixed atoms at the crystallographic positions (known from X- rays diffraction data), in the three symmetries Pnma, P212121 and P21/n.1.1. For the representation of the sorbate molecules we employed and compared selected models from literature [2].

The GCMC results for CO2, modeled as a three LJ sites three partial charges model, and assuming Pnma (ORTHO) symmetry for the silicalite, are in good agreement with experimental measurements given elsewhere. The two LJ sites three partial charges N2 model, predicts successfully the isotherm step observed experimentally at 77 K (assuming ORTHO silicalite); the simulation results of N2 at 300 K, (assuming ORTHO and MONO silicalite) verifies the experimental isotherm.

Our EMD simulations for both sorbates, assuming Pnma symmetry for the crystal, agree with other microscopic techniques (experimental and simulation) [3 , 4], as well as with macroscopic experiments (supported membrane permeation methods).

References

1
D. Fincham Mol. Simul. 11 , 79 (1993)
2
C. S. Murthy, K. Singer, M. L. Klein, I. R. McDonald Molec. Phys. 41 , 1387 (1980)
3
Ed. J. Maginn, A. T. Bell, D. N. Theodorou J. Phys. Chem. 97 , 4173 (1993)
4
J. Kärger, D. M. Ruthven, Diffusion in Zeolites and Other Microporous Solids; Wiley-Interscience; New York, (1992)
Index of talks

Dissolution of Porous Media

Alistair S. McLeod 1 and S. Bordia 2

  1. School of Chemical Engineering, University of Edinburgh, Kings Buildings, Mayfield Road, Edinburgh, EH9 3LJ, Scotland
    amcleod@chemeng.ed.ac.uk
  2. Shell Netherland, PO Box 1414, 3000DN, Rotterdam, The Netherlands.

We analyse a deterministic cellular automaton for describing interface growth during the dissolution of a soluble porous solid by an invading solvent and compare our results to the more widely studied case of fluid imbibition into an insoluble medium. In an insoluble porous medium, imbibition experiments conducted in Hele-Shaw cells suggest that the displacement of a non-wetting fluid by a wetting fluid in a disordered medium leads to the formation of a self-affine interface separating the two fluids. This roughening of the fluid interface can be driven by two noise sources. Annealed noise, resulting from thermal fluctuations at the fluid interface, and quenched noise, arising from the disordered structure of the medium in which the fluid is flowing.

For the case of a soluble solid phase, for solid phase void fractions below the site percolation threshold, we observe the formation of a self-affine solvent front. The values of the roughness and growth exponents, $ \alpha$ $ \approx$ 1/2 and $ \beta$ $ \approx$ 1/3 in d = (1 + 1) and $ \alpha$ $ \approx$ 0.38 and $ \beta$ $ \approx$ 0.26 in d = (2 + 1), describing the interface dynamics are in agreement with the exponents predicted for the Kardar-Parisi-Zhang equation with additive annealed noise.

Index of talks

Zirconia Addition in Lime-Silicate Glasses: A Molecular Dynamics Study.

M. Montorsi 1*, M.C. Menziani 1, C. Leonelli 1, G.C. Pellacani 1 and A.N. Cormack 2

  1. Department of Chemistry, University of Modena and Reggio Emilia, via Campi 183 Modena ITALY.
  2. NYSCC, Alfred University Alfred NY (USA)
* montorsi.monia@unimo.it

In the recent past computer simulations were largely used, as a complementary tool in the glasses structure characterization, together with the traditional experimentally measures [1 , 2]. This work we analyze the structural modifications inducted by zirconia addition to a lime-silicate basic system. Particular attention has been directed to the setting of an appropriate computational procedure in order to test the effect of the equilibration, using different cooling cycle, and the effect of the spatial position of the atoms in the starting input structure [3]. Moreover a detailed characterization of the SRO environment around the Zr ion has been performed in order to investigate a) the role played by this ion in the glasses materials and b) the structural modifications directly responsible of the macroscopic properties of these systems [4]. The results obtained from the computer simulations has been compared with the experimentally measurement performed on the same glasses system.

References

1
T. F. Soules J. Chem Phys 71 , 4570-4578.
2
A. N. Cormack and Yuan Cao Mol. Eng. 6 , 183-227, (1996)
3
M. Montorsi, M. C. Menziani, C. Leonelli, and A. N. Cormack, Mol. Eng., in press.
4
C. Meneghini, A. F. Gualtieri and C. Siligardi J. of Appl. Cryst. 32 , 1090-1099, (1999)
Index of talks

Different treatments of solvation of small organic molecules in aqueous solution

Stuart Murdock and Prof. Ruth Lynden-Bell 1, Dr. Graham Sexton 2

  1. Physics Department, Queen's University of Belfast, BT7 1NN, Belfast, N.Ireland,U.K.
  2. Zeneca Agrochemicals, Jealott's Hill Research Station, Bracknell, Berkshire, RG42 6E7, UK
    Graham.Sexton@aguk.zeneca.com

Various equilibria (Tautomeric and Conformational) have been studied using both a continuum description of water, and explicitly molecular methods. A comparison of these two techniques and other methods incorporating characteristics of both are needed to understand why the continuum description of water fails in some instances. I wish to know in what circumstances I am able to use a continuum to describe water molecules. Under what conditions will the macroscopic continuum be able / not be able to describe the microscopic water molecules to a qualitative accuracy? Results will be presented showing where individual water molecules solvate the organic molecule under consideration and hence reveal places where the continuum may fail.

Index of talks

Rugh's theorem for simulators

G Rickayzen
The Physics Laboratory, The University, Canterbury, Kent, CT2 7NR, UK

Although Rugh has provided a formal definition of temperature, his proof [1 , 2] is couched in mathematical language which is unfamiliar to many scientists. In this poster the proof is cast in language which we believe is more familiar to simulators. We point out that the theorem leads to an infinite number of formulae for the temperature of which the kinetic temperature and the configurational temperature are but two examples.

References

1
H. H. Rugh Phys. Rev. Letters 78 , 772-4 (1997)
2
H. H. Rugh Journal of Physics A - Mathematical and General 31 , 7761-7770 (1998)
Index of talks

Computer Simulation of Compaction of Powders

F. X. Sanchez-Castillo 1, J. Anwar 1, D. M. Heyes 2

  1. Computational Pharmaceutical Science, Department of Pharmacy, King's College London, Franklin-Wilkins Building, 150 Stamford Street, London SE1 8WA, United Kingdom
  2. School of Physics and Chemistry, University of Surrey, Surrey, GU2 7XH, United Kingdom

In the manufacturing of pharmaceuticals tablets there are two significant processing problems: capping and lamination of the formed tablets. Both represent the breakdown in tablet integrity subsequent to compression. They can be caused by a number of factors related to the compaction process or to the physicochemical properties of the powder. Two of the main factors to which capping and lamination are attributed are the distribution of particles within the die and deformation behaviour of the particles. Many details of the compaction process cannot be measured experimentally because of both the short time scales involved and because the process taking place at the molecular level are inaccessibly. An alternative approach is to use computer simulation models. A three dimensional Molecular Dynamics simulation of the compaction process is presented here. Features of the computational model include built-in time dependent viscoelastic behaviour, granule packing, deformation and bonding, and temperature control. The simulated configurations are connected in time and therefore the simulation may be used to calculate time dependent properties. Some facets of the program are the structural analysis of the system in real space via the radial distribution function g(r), n(r), volume fraction ( $ \phi$ ) and porosity [1 , 2]; evaluation of plastic-elastic deformation of granules through force-displacement profiles, loss of kinetic energy as a function of time, the creep test [3] and measurement of granules size; and finally also the usual thermodynamics parameters (T,P,F,PE,KE). The model has the potential to make the compaction process transparent, hence enable the exploration of the many facets and parameters that characterize compaction. Insights gained could be prove to be invaluable for developing formulations and technology to yield capping and lamination-free tablets.

References

1
David M. Heyes ``The Liquid state- applications of molecular simulations'', Wiley Series in theoretical chemistry, John Wiley & Sons (1998).
2
J F M Lodge, D M Heyes ``Brownian dynamics computer simulations of quenched Lennard-Jones fluids: I morphology and local structural evolution'' Molecular Simulation 23 , 203-241 (1999).
3
M. Celik, M E Aulton ``The viscoelastic deformation of some tableting materials as assessed by indentation rheology'' Drug Dev. Ind. Pharm. 22(I) , 67-75 (1996).

Figures

Compaction of powders. Figure 1 Compaction of powders. Figure 2

Index of talks

Molecular dynamics of a dense fluid of polydisperse hard spheres

Richard Sear
Department of Physics, University of Surrey, Guildford, Surrey GU2 7XH, United Kingdom

Slow dynamics in a fluid are studied in one of the most basic systems possible: polydisperse hard spheres. As the dynamics slow they become more heterogeneous, the spread in the distances travelled by different particles in the same time increases [1]. However, the dynamics appears to be less heterogeneous than in hard-sphere-like colloids at the same volume fraction [2]. The particles which move least far in a characteristic relaxation time and, particularly, the particles which move farthest in the same time are clustered, not randomly distributed throughout the sample [1]. For different polydispersity widths, the relaxation time is the same function of the compressibility factor, suggesting that this determines the relaxation time for hard spheres.

References

1
C. Donati, S. C. Glotzer, P. H. Poole, W. Kob and S. J. Plimpton Phys. Rev. E 60 , 3107 (1999).
2
E. R. Weeks, J. C. Crocker, A. C. Levitt, A. Schofield and D. A. Weitz Science 287, 627 (2000).
Index of talks

Theoretical analysis of the multi-particle electrostatic colloidal interactions.

W. Richard Bowen 1, Adel. O. Sharif 2 and Zinat Tabatabaian 2

  1. Centre for Complex Fluids Processing,, Department of Chemical and Biological Processes Engineering, University of Wales Swansea, Swansea SA2 8PP, UK
  2. Department of Chemical and Process Engineering, School of Engineering, University of surrey, Guildford, Surrey GU2 5XH, UK

Email: z.tabatabaian@surrey.ac.uk

The electrostatic force between two identical macroions is due to: repulsive interaction between two like-charged macroion spheres, attractive interaction between macroion and small counterions layer and attractive interaction connected to density fluctuations of the small ions in the double layers surrounding the two particles. Relatively small changes in above individual contributions can, therefore, shift the force balance from overall repulsion to overall attraction. The repulsive term between two macroions and the attractive term between macroions and counterions double layers depends on the charges of macroions and small ions. Whereas the attractive term due to the density fluctuation of the small counterions in the double layer depends on the charges and sizes of the small ions and macroions. In view of the fact that less number of multivalent counterions can neutralise a macroion in compare with monovalent counterions. Therefore they occupy smaller area of the double layer and leave more space for the other small counterions to produce attraction due to their variations. Consequently attraction force which have been experienced in recent reports overweight for only specific conditions of the macroions potential and diameters; and the small ions charges, concentrations and diameters.

In this work the influence of monovalent and multivalent counterions on the space potential of the adjacent macroion in an electrolyte solution have been quantified and their effect on the electrostatic interactions between two like-charged spheres have been evaluated. The force between two charged spheres confined in a long charged tube are also compared with the results obtained for two isolated spheres in electrolyte solution of different type of valances. The results show the strong effect of the wall on the reduction of the repulsive force between the spheres.

Index of talks

Vibrational modes of single-wall carbon nanotubes (SWNT) by computer simulation

V. P. Sokhan, D. Nicholson, and N. Quirke
Department of Chemistry, Imperial College of Science, Technology and Medicine, London SW7 2AY

Since their discovery in 1993, SWNT have become an important topic in nanotechnology due to their unique optical and elastic properties [1]. With the development of experimental techniques for mass-production of nanotubes in bundles, or nanotube ropes, it becomes possible to investigate the fundamental properties of individual nanotubes. Synthesized nanotube ropes contain nanotubes of various diameters and helicity and reliable methods of their characterisation are in high demand. Raman scattering experiments show that vibrational spectra of nanotubes contain several characteristic diameter-dependent features with the prominent line in the low frequency part identified with the so-called ``radial breathing mode'' [2].

We report the results of classical molecular dynamics simulation of the phonon density of states of single-wall carbon nanotubes of various chiralities and as a function of temperature and nanotube diameter using the empirical many-body potential for the carbon-carbon interaction based on the bond-order concept [3]. The calculated frequencies of the radial breathing mode are in good agreement with results of empirical force constant model and with ab initio and tight-binding DFT calculations. The results obtained allow us to assign low-frequency resonance enhanced Raman modes observed in SWNT bundles.

References

1
R. Saito, G. Dresselhaus, and M. S. Dresselhaus. Physical Properties of Carbon Nanotubes (Imperial College Press, 1998).
2
A. M. Rao et al. Science 275, 187 (1997).
3
J. Tersoff Phys. Rev. Lett. 56 , 632 (1986); Ibid. 61 , 2879 (1988); Phys. Rev. B 37 6991 (1988); D. W. Brenner Phys. Rev. B 42 , 9458 (1990).
Index of talks

Differentiable Force Functions for Molecular Dynamics and Their Efficient Evaluation

J. Stefanovi{\'{c\/}} and C. C. Pantelides
Centre for Process Systems Engineering, Imperial College of Science, Technology and Medicine, London SW7 2BY, United Kingdom

The molecular dynamics technique can be viewed as a deterministic mathematical mapping between, on one side, the force field parameters that describe the potential energy interactions and the input macroscopic conditions, and, on the other, the calculated macroscopic properties of the bulk molecular system.

The differentiability of such a mapping in the conventional molecular dynamics calculations is affected by the discontinuities in particle positions introduced by the periodic boundary conditions and the discontinuities introduced by the minimum image convention and other methods commonly employed to approximate the calculation of interparticle potential and force.

We propose an alternative molecular dynamics framework [1] based on modified force functions which are almost everywhere continuous and differentiable, and exhibit a natural periodicity. These characteristics obviate the need for both the periodic boundary conditions and the minimum image convention, as well as for any corrections for long-range interactions. They also make it possible to apply standard methods of variational calculus for the computation of partial derivatives of the molecular dynamics mapping [2].

The fully continuous and differentiable framework for performing molecular dynamics calculations requires the evaluation of rather complex force functions and their spatial partial derivatives. We present an efficient interpolation scheme [3] for the evaluation of these quantities over a finite spatial domain.

The modified force function is approximated by a linear combination of Hermite cubic basis functions such that both the interpolant of the force and its spatial derivatives are continuous across the grid boundaries. In order to achieve better accuracy for a given grid size, a non-uniform rectilinear grid is constructed via iterative refinement procedure. The latter guarantees the accuracy of the force computed by interpolation within any specified tolerance $ \varepsilon$ > 0 .

For many potential functions of practical interest, it is possible for polynomial interpolants to be constructed for parts of the force functions which are independent of the potential parameters and system density (the so-called ``separable force functions''). In such cases, a single interpolation grid which is applicable for a wide range of potential parameters and system densities can be constructed a priori.

References

1
J. Stefanovi{\'{c\/}} and C. C. Pantelides. ``Molecular dynamics as a mathematical mapping. I. Differentiable force functions.'' Accepted for publication. J. Mol. Sim. , (2000).
2
J. Stefanovi{\'{c\/}} and C. C. Pantelides. ``Molecular dynamics as a mathematical mapping. II. Partial derivatives in the microcanonical ensemble.'' Accepted for publication. J. Mol. Sim. , (2000).
3
J. Stefanovi{\'{c\/}} and C. C. Pantelides. ``Molecular dynamics as a mathematical mapping. III. Efficient evaluation of the differentiable force functions and their derivatives.'' Accepted for publication. J. Mol. Sim. , (2000).
Index of talks

Molecular Simulations Applied to Carbon Molecular Sieve Membranes

Alexandre Vieira-Linhares 1 and Nigel A. Seaton 2
School of Chemical Engineering, University of Edinburgh Edinburgh EH9 3JL, UK

  1. alexl@chemeng.ed.ac.uk
  2. nigel@chemeng.ed.ac.uk

Molecular Simulations are widely used to predict both equilibrium and non-equilibrium properties of gas mixtures. A Grand Canonical Molecular Dynamics (GCMD) program is being prepared to calculate non-equilibrium transport properties for a binary mixture of gases onto slit-like graphitic pores. This research involves the mathematical modelling of adsorption and diffusion in carbon molecular sieve membranes (CMSM) under realistic conditions of temperature, pressure and bulk gas compositions. Adsorption and diffusion can be predicted using two standard methods: Grand Canonical Monte Carlo (GCMC) and Molecular Dynamics (MD), respectively. Several authors are using these two algorithms with satisfactory results for a wide range of cases. However, GCMC and MD by themselves have practical limitations - they fail to describe some important industrial applications (for example membrane separations). Systems such as membranes, catalysts and some adsorbents (e.g. [1 , 2]) cannot be described with conventional algorithms. GCMD is uniquely suited for these systems with flow and chemical potential. We are working on the roles that transport takes in separation of gas mixtures by CMSM. Here, we will present some GCMD preliminary results for methane and methane/hydrogen mixtures assuming graphitic slit-like pores.

References

1
J. M. D. MacElroy and M. J. Boyle, personal communication, to be published.
2
A. P. Thompson, D. M. Ford and G. S. Heffelfinger J. Chemical Physics 109 , 6406-6414 (1998)
Index of talks

Aggregation and Crystallisation of Particles with Short Ranged Interactions

Gerrit A. Vliegenthart 1 and Els H. A. de Hoog 2

  1. School of Chemistry, University of Bristol, BS8 1TS Bristol, United Kingdom
  2. Van 't Hoff laboratory, University of Utrecht, Padulaan 8, 3584 CH Utrecht, The Netherlands

In recent years there has been increasing interest for the role of the range of the (isotropic) attractive interactions on phase behaviour and structure of atomic fluids and colloidal suspensions (for references: see [1 , 2] and therein). It has now been established that the global features of the phase diagram i.e. the relative location of the fluid-solid phase boundary and the gas-liquid boundary with respect to one another is determined by the range of the attractive interactions relative to the range of the repulsive interactions. For short ranged attractions, the gas-liquid phase boundary lies metastable with respect to the fluid-solid boundary.

Globular proteins like lysozyme, are an important example of particles which are believed to interact through these short ranged attractive interactions. The prediction and understanding of successful crystallisation of proteins has been a widespread motivation to study the role of the interaction range on crystallisation/aggregation behaviour [3 - 7]. Protein crystallisation experiments are usually performed at low concentrations and the (empirically established) successful crystallisation conditions (ionic strength, temperature etc.) are bounded by values of the second osmotic virial coefficient B2 typically between -5 v0 and -40 v0 (v0 is the volume of a particle) [8 , 9]. Under conditions that B2 is much lower, the protein molecules rapidly aggregate instead.

To investigate the coupling of aggregation and crystallisation in the so-called crystallisation slot, we have performed both Brownian Dynamics simulations on particles interacting through short ranged interactions and experiments on colloid-polymer mixtures.

The simulations were carried out on particles interacting through a Lennard Jones 36-18 potential of which the phase diagram is well documented [10 , 1] and shows a metastable gas-liquid phase boundary.

Temperature quenches were done at low densities from the one phase region to various temperatures below the fluid-solid boundary. In the two phase region, three different types of aggregation are found.

At 20% undercooling ( B2 $ \approx$ - 20 v0), a few large aggregates are slowly formed in coexistence with single particles through a nucleation like process. These aggregates then slowly transform into more crystalline structures. At slightly lower temperatures ( B2 $ \approx$ - 40 v0), the initial clustering is much faster and compact solid structures are formed. Lowering the temperature even further ( B2 $ \ll$ - 100 v0), leads to a somewhat lower clustering rate while the shape of the clusters becomes more elongated.

Colloid-polymer mixtures are excellent model systems to study the role of interactions on phase behaviour and aggregation (see [2] and references therein). These mixtures allow for manipulating the range and strength of the attractive (depletion) interaction through the size ratio of polymer to colloid and concentration of polymer. The colloidal particles can be visualised on particle level in real time and real space using confocal microscopy [11]. The experimental work on aggregation in a colloid-polymer mixture of small size ratio (corresponding to short ranged attractive depletion interactions) [12] confirms the existence of the various aggregation scenarios found in the simulations in great detail.

References

1
G. A. Vliegenthart, J. F. M. Lodge and H. N. W.  Lekkerkerker, Physica A 263, 378 (1999)
2
W. C. K Poon, Curr. Opinion Coll. Interface Sci. 3, 593 (1998)
3
D. Rosenbaum and P. C. Zamora and C. F.  Zukoski, Phys. Rev. Lett. 76, 150 (1996)
4
M. Muschol and F. Rosenberger, J. Chem. Phys. 107, 1953 (1997)
5
R. Piazza, V. Peyre and V. Degiorgio, Phys. Rev. E 58, R2733 (1998)
6
W. C. K Poon, Phys. Rev. E 55, 3762 (1997)
7
P. R. Ten Wolde and D. Frenkel, Science 277, 1975 (1997)
8
G. A. Vliegenthart and H. N. W. Lekkerkerker, J. Chem. Phys. 112, 5364 (2000)
9
A. George and W. W. Wilson, Acta Cryst. D 50, 361 (1994)
10
M. Hasegawa and K. Ohno, J. Phys.: Condens. Matter 9, 3361 (1997)
11
A. van Blaaderen and P. Wilztius, Science 270, 1177 (1995)
12
E. de Hoog and H. N. W. Lekkerkerker, to be published
Index of talks

Computer simulation of switching in thin liquid crystal films

Richard Webster, D. J. Cleaver, C. M. Care
Modelling Group, Materials Research Institute, Sheffield Hallam University, Howard St., Sheffield S1 1WB

This work aims to simulate the switching that occurs on removal of an initial applied bulk field from a thin nematic liquid crystal film confined between two aligning substrates.

Simple liquid crystal displays exploit switching caused by competition between director alignment due to confining surfaces and alignment due to an applied field. On removal of the field, director relaxation can induce flow in the liquid, which can in turn affect the director orientation. These related processes of ``backflow'' and ``orientational kickback'' play a role in limiting switching speeds [1].

The focus will be on the initial surface-induced order throughout the film and its effect on the relaxation dynamics. This is often idealised or ignored in analytical studies [2 , 3].

Systems of interest have surface director orientation parallel to the surface and the applied field normal to the surface with the two surface easy axes both parallel and perpendicular.

Molecular interactions are modelled using the Gay-Berne [4] anisotropic intermolecular potential. The aligning surface model is based on the half-space integral of the particle interaction with one particle reduced to a sphere [5]. Simulations are performed using parallel replicated-data molecular dynamics [6].

References

1
S. Chandrasekhar Liquid Crystals Cambridge University Press, (1992).
2
``A calculation of orientational relaxation in nematic liquid crystals'' M. G. Clark and F. M. Leslie Proc. R. Soc. Lond. A. 361 , 463, (1978).
3
``General hydrodynamic equations for nematic liquid crystals'' T. Qian and P. Sheng Phys. Rev. E 58 (6) , 7475, (1998).
4
``Modification of the overlap potential to mimic a linear site-site potential'' J. G. Gay and B. J. Berne J. Chem. Phys. 74 (6) , 3316, (1981).
5
``Computer simulation studies of confined liquid crystal films'' G. Wall and D. J. Cleaver Phys. Rev. E 56, (4) , 4306, (1997).
6
``Replicated data and domain decomposition molecular dynamics techniques for simulation of anisotropic potentials'' M. R. Wilson et al J. Comput. Chem. 18, (4) , 478, (1997).
Index of talks

Brownian Dynamic Simulation of Polymer Chains with End-Attractions in Solution

C. Xiao and D.M. Heyes
Department of Chemistry, School of Physical Sciences, University of Surrey, Guildford GU2 5XH, United Kingdom.

Brownian dynamics simulations have been carried out for model polymer chains with end-attraction in solution over a range of concentration. The polymers are treated as beads linked by FENE springs and the repulsion between any two unlinked beads is modeled by a pair potential with a Gaussian analytic form, $ \beta$u1(r) = Aexp(- r2/$ \sigma^{2}$), where $ \beta$ = 1/kT, A and $ \sigma$ are characteristic energy and distance scales respectively. The basics of the analytic techniques are given in refs. [1 , 2].

For the beads at the end of a chain an additional attractive potential of a similar form, $ \beta$u2(r) = BAexp(- (r - 4$ \sigma$)2/$ \sigma^{2}$), is added. Three different systems were examined: system A1 with head-head attraction, system A2 with head-head and end-end atractions, and system A3 with head-end attraction. The dimensions of the chains, the site-site radial distribution function, as well as the dynamic properties such as time-correlation functions, infinity frequency elastic modulus, self-diffusion coefficient are studied as function of solution density and attraction types.

References

1
C. Xiao and D.M. Heyes J. Chem. Phys. 111 10694-10705 (1999)
2
C. Xiao and D.M. Heyes Phys. Rev. E 60 5757-5767 (1999)
Index of talks
Last modified 28 October 2002 		 
 
   
 
 
back to top