Previous session | Next session

Session S15 - Materials Theory and Simulation V.
FOCUS session, Wednesday afternoon, March 14
Room 211, Washington State Convention Center

[S15.001] Rigid Body Diffusion Monte Carlo studies of SF_6 in helium clusters

Mehul V Patel (Departments of Physics and Chemistry, U. C. Berkeley), K. Birgitta Whaley (Department of Chemistry, U. C. Berkeley)

The incorporation of rotational degrees of freedom into Quantum Monte Carlo calculations is necessary for studies of molecular clusters (e.g. water) and molecule doped quantum clusters (e.g. SF_6 in helium).

We describe our recent implementation of importance sampling in rigid body diffusion monte carlo and show that it correctly samples the angular degrees of freedom. We also demonstrate that importance sampling significantly improves the efficiency and accuracy of calculations of ground state properties.

Molecules doped in helium clusters exhibit very interesting rotational spectra that resemble the gas phase spectra with reduced rotational constants (increased moments of inertia). This phenomenon is explored by applying the above methodology to investigate the degree of adiabatic following of the helium density surrounding the SF_6 molecule in SF_6-He_N clusters.

Our results suggest incomplete adiabatic following of the helium density around the SF_6 molecule, even for the smallest cluster sizes.

[S15.002] Moving Atoms, Electrons and Spins with Molecular Dynamics

Roberto Car (Princeton Univ)

This abstract not available.

[S15.003] The SIESTA Program for Electronic Structure Simulations

Jose M. Soler (Dep. de F\'\isica de la Materia Condensada, Universidad Autónoma de Madrid, E-28049 Madrid, Spain)

We have implemented a method to perform standard density-functional calculations with CPU time and memory requirements that scale linearly with the system size, allowing calculations with thousands of atoms on a workstation. Exchange and correlation are treated within the local spin density or gradient-corrected approximations, and we use Troullier-Martins norm-conserving pseudopotentials in the Kleinman-Bylander form.

The basis set of atomic orbitals can be constructed from numerical solutions of the atomic pseudopotential, while they are constrained to be zero beyond a cutoff radius. Further improvements can be made by optimizing their radial shape for the chemical environment. Multiple-zeta and polarization orbitals can be included to achieve an arbitrarily rich basis set, and an accuracy comparable to that of plane-waves.

The basis orbitals are projected on a uniform real-space grid in order to calculate the electron density and the Hartree and exchange-correlation potentials and matrix elements. Other matrix elements, like those of the kinetic energy and nonlocal pseudopotentials are tabulated as two-center integrals.

Electron wave functions need not be explicitly orthogonalized. Instead, we use a modified energy functional, whose minimization produces orthogonal wave functions and the same energy and density as the Kohn-Sham functional. Additionally, confining the Wannier-like electron wave functions to a finite region allows the linear scaling of CPU time and memory, while the error introduced decreases rapidly with the confinement radius.

Forces and stresses are also calculated efficiently and accurately, thus allowing structural relaxation and molecular dynamics simulations.

I will briefly review some applications in physical, chemical, and biological systems.

note

[S15.004] First-principles simulations of shock front propagation in liquid deuterium

Francois Gygi, Giulia Galli (Lawrence Livermore National Laboratory, Livermore CA 94551)

We present large-scale first-principles molecular dynamics simulations of the formation and propagation of a shock front in liquid deuterium. Molecular deuterium was subjected to supersonic impacts at velocities ranging from 10 to 30 km/s. We used Density Functional Theory in the local density approximation, and simulation cells containing 1320 deuterium atoms. The formation of a shock front was observed and its velocity was measured and compared with the results of laser-driven shock experiments [1]. The pressure and density in the compressed fluid were also computed directly from statistical averages in appropriate regions of the simulation cell, and compared with previous first-principles calculations performed at equilibrium [2]. Details of the electronic structure at the shock front, and their influence on the properties of the compressed fluid will be discussed.

[1] J.W.Collins et al. Science 281, 1178 (1998).

[2] G.Galli, R.Q.Hood, A.U.Hazi and F.Gygi, Phys.Rev. B61, 909 (2000).

[S15.005] A Molecular Dynamics Study of High Pressure Nitrogen

William Mattson, Daniel Sanchez-Portal, Richard Martin (Department of Physics, University of Illinois Urbana-Champaign)

The energies of several different phases of Nitrogen have been calculated, across a wide range of pressures, and compared with the literature. These calculations were done with the Density Functional Theory in both the Local Density Approximation and the Generalized Gradient Approximation. Molecular dynamics simulations have been performed, with the same approximations, under high pressure at both low and high temperature, where the initial structure is predicted to be unstable both by the previous energetics calculations and experiments. Simulation cells as large as 256 atoms have been used.

[S15.006] Transitions to Nonmolecular Structures and Decomposition of CO2 at High Pressures

Oliver Tschauner, Maddury Somayazulu, Ho-kwang Mao, Russell J. Hemley (Geophysical Laboratory and Center for High Pressure Research, Carnegie Institution of Washington, 5251 Broad Branch Road, N.W., Washington, D.C. 20015)

We studied the structure and stability of solid CO^2 at pressures up to 110 GPa and temperatures up to 3000 K focussing on potential nonmolecular phases. Such phases display remarkable optical, electronic and elastic properties [1]. We used in situ high-P Raman spectroscopy and x-ray diffraction as structural probes. The high temperature experiments were performed by CO^2-laser heating. At pressures above 42 GPa and temperatures below 2000 K we generated a novel, monoclinic phase of CO^2 which appears, like CO^2-V [1], to be nonmolecular. There is, however, some indication that this phase is a 2-d network structure rather than a 3-d one like CO^2-V. We further show that CO^2 in the solid state breaks down to oxygen and diamond along a negative P-T reaction boundary crossing 2000 K around 60 GPa. The phase relations in the C-O system at high pressures appear to be similar to the C-S [2] rather than the Si-O system, although expanded to much larger scales both in P and T. Supported by NSF DMR-9972750

[S15.007] Model Potential Calculations of the Thermal Properties of Argon

E. Roger Cowley (Rutgers University)

Interatomic potentials in the inert gas crystals are still usually represented by phenomenological expressions. Simple forms, such as the Lennard-Jones potential have been used for many years. More recently, sophisticated models have been based on a wide range of experimental properties and on Hartree-Fock calculations for pairs of atoms. A stringent test of such models is the calculation of solid state properties. We test several model potentials for Argon, including both nearest neighbor and all neighbor Lennard-Jones potentials, and a potential of the Tang Toennis form. The thermal properties of the crystal are calculated using the Improved Self-Consistent formalism, which we believe to be accurate over almost whole temperature range of the solid.

[S15.008] Monte Carlo calculation of interfacial free energy of hard-sphere fluid against structured walls

Atsushi Mori, Brian Laird (Department of Chemistry, University of Kansas)

Particle insertion method for evaluating the chemical potential is modified to apply to the calculation of the interfacial free energies of the hard-sphere fluid against fcc (001), (011), and (111) walls, respectively. For hard-body systems a free energy difference, i.e., the chemical potential or the interfacial free energy, is calculated as a logrithm of probability that a hard object is successfully inserted into the system. However, because the direct insertion of a hard wall into a dense fluid is less probable, we calculate the free energy difference by forming a structure on a structurless wall gradually. The interfacial free energy against a flat wall is also evaluated by gradual forming of a wall in a fluid, which is tested by comparing with the literature. These wall interfacial free energies are compared with the direct evaluation of the crystal/fluid interfacial free energies which has been recently published.

[S15.009] Direct calculationof the crystal/melt interfacial free energy for a system of hard-spheres

Brian Laird (Department of Chemistry, University of Kansas), Ruslan Davidchack (Department of Mathematics and Computer Science, University of Leicester)

We present a direct calculation by molecular-dynamics computer simulation of the crystal/melt interfacial free energy, \gamma, for a system of hard spheres of diameter \sigma. The calculation is performed by thermodynamic integration along a reversible path defined by cleaving, using specially constructed movable hard-sphere walls, separate bulk crystal and fluid systems, which are then merged to form an interface. We find the interfacial free energy to be slightly anisotropic with \gamma = 0.62\pm 0.01, 0.64\pm 0.01 and 0.58\pm 0.01 k_BT/\sigma^2 for the (100), (110) and (111) fcc crystal/fluid interfaces, respectively. These values are consistent with earlier density functional calculations and recent experiments measuring the crystal nucleation rates from colloidal fluids of silica spheres that have been interpreted [Marr and Gast, Langmuir 10, 1348 (1994)] to give an estimate of \gamma for the hard-sphere system of 0.55 \pm 0.02 k_BT/\sigma^2, slightly lower than the directly determined value reported here.

Part S of program listing