Molecular Dynamics Simulations Of Liquid Research Articles

Generalized coarse-grained model based on point multipole and Gay-Berne potentials

This paper presents a general coarse-grained molecular mechanics model based on electric point multipole expansion and Gay-Berne [J. Chem. Phys. 74, 3316 (1981)] potential. Coarse graining of van der Waals potential is achieved by treating molecules as soft uniaxial ellipsoids interacting via a generalized anisotropic Gay-Berne function. The charge distribution is represented by point multipole expansion, including point charge, dipole, and quadrupole moments placed at the center of mass. The Gay-Berne and point multipole potentials are combined in the local reference frame defined by the inertial frame of the all-atom counterpart. The coarse-grained model has been applied to rigid-body molecular dynamics simulations of molecular liquids including benzene and methanol. The computational efficiency is improved by several orders of magnitude, while the results are in reasonable agreement with all-atom models and experimental data. We also discuss the implications of using point multipole for polar molecules capable of hydrogen bonding and the applicability of this model to a broad range of molecular systems including highly charged biopolymers.

Computer simulation of liquid and amorphous selenium

We report molecular dynamics simulations of liquid and amorphous Se at pressures from 0 to 6 GPa for different quench rates. The atomic volume depends strongly on the quench rate, in contrast to the small variation of the atomic enthalpy. The glass transition temperature Tg, extrapolated to a quench rate of order K/s, agrees within 20% with experiment. Applying a pressure of 1 GPa, Tg is raised by 37 K.

Molecular Dynamics Simulation of Liquid H2O, MeOH, EtOH, Si(OMe)4, and Si(OEt)4, as a Function of Temperature and Pressure

We use molecular dynamics simulations to model pure liquids, including water (and heavy water), the simplest alcohols (MeOH and EtOH), and the simplest silica alkoxides (Si(OMe)4 and Si(OEt)4), at high pressures and temperatures and under ambient conditions. The same methodology is employed throughout to derive potentials for different species, and the same potentials are used for different thermodynamic conditions. The studies were carried out using two different MD codes, DISCOVER and DL_POLY, with slightly different implementations and force fields, to guarantee that the results would not be sensitive to details of the simulations. The results obtained with both codes for density, enthalpy of vaporization, and radial distribution functions compare very well with experiment, whereas the self-diffusion coefficients are slightly too high. The same general methodologies and sets of potentials should therefore be valid in describing complex liquids, such as silica-based sol−gel solutions, containing water, alcohol, and silica alkoxide, without having to redefine the force field on changing the composition or thermodynamic conditions.

Parallel aspects of quantum molecular dynamics simulations of liquids

A very straightforward parallel scheme to perform Born–Oppenheimer quantum molecular dynamics simulations of liquids is presented. It is analogous to Linear Scaling methods, used in electronic structure calculations. The instantaneous liquid configurations of molecules moving in the simulation cell with periodic boundaries and minimum image convention are divided into smaller spherical clusters around each molecule. Forces, acting on the nuclei of the molecule in the middle of each cluster, are calculated independently at any suitable level of quantum mechanical sophistication. From low-end semi-empirical schemes to high-end methods containing electron correlation corrections. In this communication selected results from simulations of liquid water simulated at 300K and normal density using ab initio Hartree–Fock level with three different basis sets are presented. Good results are obtained for the near-structure and other characteristic liquid properties of water. A simple way to implement the method by combining a conventional classical MD simulation code with any quantum chemistry package capable of performing force calculations is presented. Suggestions of suitable parallel schemes based on standard message passing techniques are given. Performance, scaling and load balancing properties are discussed and possible future directions are suggested.

Dynamic properties of simple liquids: Dependence on the softness of the potential core

The dependence of both individual and collective dynamic properties of simple liquids on the steepness of the potential cores is analyzed. Molecular dynamics simulations of liquids in identical conditions but assuming two repulsive interaction potentials with very different softness have been carried out. Two liquids at very different densities (with volume packing fractions η=0.5 and η=0.3) have been considered for each potential. The study of the dynamic collective properties includes the dynamic structure factor and the longitudinal and transverse current correlations. It has been corroborated that longitudinal modes associated with density fluctuations propagate up to higher wave numbers in liquids with softer potential cores. However, the propagating transversal modes in dense liquids are weakly influenced by the hardness of the potential core. Other properties such as the sound velocity in liquids at high density or the dynamic structure factor at the kinetic regime are rather insensitive to the details of the repulsive interactions. The velocity correlation functions for liquids with soft repulsive potential walls are markedly oscillatory. These oscillations are associated with the coupling of the atomic velocity to collective density fluctuations.

Dynamic properties of Lennard-Jones fluids and liquid metals.

The dependence of the dynamic properties of liquid metals and Lennard-Jones fluids on the characteristics of the interaction potentials is analyzed. Molecular-dynamics simulations of liquids in analogous conditions but assuming that their particles interact either through a Lennard-Jones or a liquid-metal potential were carried out. The Lennard-Jones potentials were chosen so that both the effective size of the particles and the depth of the potential well were very close to those of the liquid-metal potentials. In order to investigate the extent to which the dynamic properties of liquids depend on the short-range attractive interactions as well as on the softness of the potential cores, molecular-dynamics simulations of the same systems but assuming purely repulsive interactions with the same potential cores were also performed. The study includes both single-particle dynamic properties, such as the velocity autocorrelation functions, and collective dynamic properties, such as the intermediate scattering functions, the dynamic structure factors, the longitudinal and transverse current correlations, and the transport coefficients.

A Parallel Quantum Mechanical MD Simulation of Liquids

A parallel implementation to perform quantum molecular dynamics simulations of liquids on Born-Oppenheimer surfaces is evaluated. The MD method allows the use of atomic energy gradient forces at an arbitrary level of quantum chemical methodology and electronic state. The computational scheme is very simple to implement, although still expensive to use. The method can easily be developed further for studies which are beyond the capabilities of current classical simulations, e.g., simple chemical reactions or other chemical processes involving excited electronic states or radicals. In a preliminary study, the approach is applied to simulate liquid water (with all internal degrees of freedom included) at the semi-empirical AM 1 molecular orbital (MO), and at the ab initio SCF-MO Hartree-Fock level. Performance, scaling and load balancing properties of the parallel scheme are discussed. Timings of an implementation on an IBM SP-2 are presented. To test the approach, the simulation results are compared with those obtained from corresponding classical simulations of water, from simulations using the Car-Parrinello method and to experimental radial distribution functions.

Molecular dynamics simulations of liquids and glasses in the system NaAlSiO4-SiO2; physical properties and transport mechanisms

Molecular dynamics simulations of liquids and glasses in the system NaAlSiO₄-SiO₂; physical properties and transport mechanisms

First-principles molecular dynamics simulation of liquid

The liquid Mg - Bi system exhibits strong compound formation at the `octet' composition . We present results of first-principles molecular dynamics simulations of this alloy system at different compositions: the pure Mg and Bi liquid components, the stoichiometric liquid, and a Mg-rich composition . For the pure liquids, our results are in excellent agreement with experimental diffraction data. For , a significant modification of the characteristics of the local ordering is found w.r.t. the crystalline -phase: the ordering in the liquid is much more ionic. This structural modification is consistent with the structure of the superionic -phase, that was reported recently by Barnes et al 1994 J. Phys.: Condens. Matter 6 L467. Our simulations cannot reproduce the `reverse' metal - nonmetal transition observed upon melting, the computed conductivity being much larger than found in experiments. Instead, for the Mg-rich alloy, the calculated conductivity approaches closely to the experimental value.

Structural, electronic and magnetic properties of liquid, amorphous and quasicrystalline metals

Recently, the atomistic simulation of the properties of liquid, glassy, and quasicrystalline metals and alloys has made considerable progress on various levels. Ab-initio density-functional molecular dynamics techniques permit simulations based on the full set of quantum mechanical many-body forces. This allows materials where covalent contributions to the chemical bonding properties are important to be investigated, e.g. liquid transition metals, and the liquid-metal to amorphous semiconductor transition to be studied, e.g. in Si and Ge. Tight-binding techniques enable the calculation of interatomic pair- and many-body forces, and hence the molecular dynamics simulation of liquid and amorphous transition-metal and transition metal-metalloid alloys. Tight-binding techniques can also be used to study the electronic and magnetic structures of these materials, including non-collinear spin-structures. Pseudopotential perturbation theory allows the determination of effective pair forces in s, p-bonded materials. These have found new applications, e.g. in the investigation of quasicrystalline materials. Classical molecular dynamics techniques have now been developed to a point where systems with a size sufficient for the investigation of medium-range order effects can be treated. New results from all these areas are discussed.

Comparison of the Performance of a Program for Molecular Dynamics Simulations of Liquids on Different Computers

Comparison of the Performance of a Program for Molecular Dynamics Simulations of Liquids on Different Computers

MD Simulations of Liquids withTd, OhMolecular Symmetry

Abstract A new method is proposed for the calculation of intermolecular interactions in Molecular Dynamics simulations of liquids with Td, Oh molecular symmetry. The new algorithm is based on the separation of the pair potential into a short-range and a long-range contribution described by a site-site and a spherical centre-centre potential model respectively using an additional cutoff distance. Test calculations for the Lennard-Jones fluids CCl4 and SF6 show significant savings in CPU time. We compare thermodynamic properties, pair correlation functions and a few dynamic autocorrelation functions obtained with the novel strategy with results of the commonly used algorithm for systems containing 864 molecules. Since no significant differences appear the new algorithm may be suggested as a useful contribution to the area of Molecular Dynamics simulation of liquids with these rather high molecular symmetries.

SHORT-RANGE ORDER OF LIQUID AND AMORPHOUS TRANSITION-METAL ALLOYS

We present molecular dynamics simulations of liquid and amorphous transition-metal alloys based on quantum-mechanically derived interatomic forces. Using the interatomic potentials calculated within the hybridized nearly-free-electron tight-binding-bond (NFE-TBB) theory we construct models for Ni,TMl-, (TM = Y, Zr, Nb, Ti, and V) glasses. We show that a clear trend from trigonal-prismatic to polytetrahedral local order and from strong to moderate chemical order exists in the series Ni-Y, Zr, Nb and with increasing Ni-content within a given system. These realistic structure models can be used also as a basis for selfconsistent supercell calculations of the electronic structure. Preliminary results are presented.

Molecular dynamics simulation of liquid and gaseous hydrogen selenide: Cross correlation functions

The effect of the phase change from compressed gas to liquid in hydrogen selenide is explored in terms of cross correlation functions. There are considerable differences in amplitude and time dependence of the cross correlation functions which are recorded at several state points from the liquid just below the boiling point to the compressed gas. Results are obtained for 108 and 500 molecules in order to investigate the effect of sample size on cross correlation functions in the laboratory frame and in the frame (1,2,3) defined by the molecular moments of inertia.

A new form for intermolecular potential energy functions

A new intermolecular potential energy function is presented for Monte Carlo and molecular dynamics simulations of liquids, solutions and other molecular assemblies. The potential energy function is expressed in terms of intermolecular overlap integrals over localized molecular orbitais of isolated molecules and Coulomb potentials between fractional point charges placed on the nuclei. The potential function is easy to generate and is applicable to a wide range of molecules. As examples the potential functions are generated for the water, ammonia and hydrogen fluoride dimers.

Molecular dynamics simulation of hydrocarbon molecules in condensed phases. II. Benzene

Molecular dynamics simulations of liquid (305 K) and solid (10 K) benzene have been carried out to investigate the vibrational properties of benzene in the condensed phase. The benzene molecule is modeled as a fully flexible system of 12 atoms with intramolecular interactions based on ab initio harmonic potentials and intermolecular interactions given by semiempirical atom–atom potentials. Vibrational frequencies are analyzed in terms of the frequency spectrum of the velocity autocorrelation function and the results are found to correlate well with optical absorption and scattering data. Single particle density fluctuations are calculated over a range of wave numbers and results are obtained that can be directly compared to the intensity distributions of inelastically scattered thermal neutrons. The different comparisons indicate that the existing potential energy function provides an essentially quantitative description of molecular vibrations of benzene.

Molecular dynamics simulation using the cray-1 vector processing computer

Algorithms are presented which enable molecular dynamics simulations of liquids to be performed most efficiently on the Cray-1 vector processing computer. For 256 atoms interacting via the Lennard-Jones potential the CPU time per step of the simulation is 22 ms.