The dependence of the structure of (210) and (310) symmetrical [001] tilt boundaries in silicon, germanium and diamond on the Keating covalent force field (potential) has been investigated by computer modelling. We have found that the sensitivity of grain boundary structure to variations of the Keating potential depends on the local atomic arrangement at the grain boundary.

The reconstruction of the Si(100) surface has been studied using a potential consisting of two- and three-body interactions [1]. Minimum energy quenches of a slab of 320 Si atoms 10 layers thick with periodic boundary conditions in the xy plane found the p(2×l) reconstruction lower in energy than the c(2×2) reconstruction. The potential energy contours for the addition of an adatom to the (100) surface were determined for the unreconstructed surface and for the p(2×l) and c(2×2) reconstructions.

A new silicon (111) surface (7×7) reconstruction model is proposed. The model for the unit mesh consists of 13 benzene like rings formed by lowering triples of top layer atoms and 12 raised top layer atoms, with three of the rings combining to give depressed corners. The construction of the model is based on some considerations of the physical properties of the surface and on cluster calculations using the semi-empirical MNDOgeometrical optimization method. A cluster calculations with 9 silicon atoms gave the ring structure as an energy minimized result and also gives evidence that the two halves of the unit mesh should be at different height levels. A 13 silicon atom “flower” cluster calculation gave the height of the raised atoms to be .851A. The structure of the unit mesh resembles two types of (2×2) structure which can be formed from the rings and raised atoms (“flowers”), one of which is readily transformed into pi-bonded chain or molecular structures. The model gives a good qualititative account of the various hills and valleys seen in scanning tunneling microscopy, and since lateral displacement is small, it agrees with low energy channeling experiments. In addition good agreement with Cardillo s helium atom scattering experiment and Yang et al. s LEED experiment is found as the model has different heights for the two halves of the unit mesh and characteristics of a mixture of two types of structure.

We develop two and three-body classical interatomic potentials that model structural energies for silicon. These potentials provide a global fit to a database of firstprinciples calculations of the energy for bulk and surface silicon structures which spans a wide range of atomic coordinations and bonding geometries. This is accomplished using a new “separable” form for the 3-body potential that reduces the 3-body energy to a product of 2-body sums and leads to computations of the energy and atomic forces in n2 steps as opposed to n3 for a general 3-body potential. Simulated annealing is performed to find globally minimum energy states of Si-atom clusters with these potentials using a Langevin molecular dynamics approach.

The need for computational power in the modeling of physical systems is rapidly increasing. Realistic simulations of materials often require complex interactions and large numbers of particles. For most scientists, full-time access to supercomputers is not possible, and even this might not be sufficient to solve their problems. As most of the calculations involved are straightforward and repetitive in nature, a possible solution is to design and build a processor for a specific application with a low cost/performance ratio. This approach is to be contrasted with the use of a general purpose computer, which is designed to treat a large class of problems and includes many expensive features (e.g. software) that are not utilized in the simulations. The architecture of a special purpose computer can be tailored to the problem; e.g., parallel and pipelined operations can be incorporated to obtain efficient computational throughput, and memory organization and instruction sets can be optimized for this purpose. A few of such algorithm oriented processors have been built in the last decade and have been utilized for certain specific jobs; for example: molecular dynamics simulations of systems of Lennard-Jones particles and Monte Carlo calculations of Ising models. An overview of some existing algorithm oriented processors and expectations for the future will be presented.

Major developments in materials characterization instrumentation over the past decade have helped significantly to elucidate complex processes and phenomena connected with the microstructure of materials and interfacial interactions. Equally remarkable advances in theoretical models and computer technology also have been taking place during this period. These latter now permit, for example, in selected cases the computation of material structures and bonding and the prediction of some material properties. Two assessments of the state of the art of instrumental techniques and theoretical methods for the study of material structures and properties have recently been conducted. This paper will discuss aspects from these assessments of computational theoretical methods applied to materials. In addition, an approach will be presented which uses advanced instrumentation and complementary theoretical computational techniques in tandem in an effort to construct and verify hierarchies of models to translate engineering materials performance requirements into microscopic-level and atomic-level materials specifications (composition, structure, and bonding). Areas of practical interest include catalysis, tribology (contacting surfaces in relative motion), protective coatings, and metallurgical grain boundaries. A first attempt involving modeling of grain boundary adhesion in Ni3Al with and without boron additions will be discussed.

Extensive numerical simulations of random magnetic materials have been recently performed at AT&T Bell Laboratories with a fast specially designed computer. I will discuss certain issues concerning the use of specialized computers in research, and I will review some major results obtained in simulations of a three-dimensional spin glass and an antiferromagnet with random fields.

A parallel processing system with substantial computational power for the simulation of materials has been designed and constructed at Stony Brook. Although primarily designed to implement the molecular dynamics algorithm, the system is highly efficient in carrying out simulations that exhibit high degrees of local interaction. Our studies show that this system is capable of performance, for this application, which is comparable to the Cray-l at a small fraction of the cost.

We are constructing a highly parallel computer array well suited to the requirements of molecular dynamics simulations of the behavior of large systems of interacting particles. A tightly coupled message passing structure is used based on a specially designed processor with 12 MFLOPS peak performance. High speed synchronous communication over a full permutation network and hardware support for operations important in molecular dynamics codes are also provided. An analysis of this configuration indicates that efficient use can be made of 1000 processors working simultaneously.

The complete prediction of microstructural development in polycrystalline solids as a function of time and temperature is a major objective in materials science, but has not yet been possible primarily due to the complexity of the grain interactions. The evolution of the polycrystalline structure depends upon the precise specification of the coordinates of the grain boundary network, the crystallographic orientations of the grains, and the postulated microscopic mechanisms by which elements of the boundaries are assumed to move. Therefore, a general analytical solution to this multivariate problem has not yet been developed. Recently, we have been able to successfully incorporate these aspects of the grain interactions, and have developed a computer model which predicts the main features of the microstructure from first principles [1,2]., The polycrystal is mapped onto a discrete lattice by dividing the material into small area (2d) or volume (3d) elements, and placing the centers of these elements on lattice points. Interactions and dynamics are then defined for the individual elements which are analagous to those postulated in continuous systems. This discrete model preserves the topological features of real materials, and can be studied by computer simulation using Monte Carlo techniques. In this paper we report the application of the Monte Carlo method to the metallurgical phenomenon of grain growth with isothermal annealing. Extension of the model to treat primary recrystallization is presented elsewhere [3,4].

We have generated two-dimensional grain structures by considering constant growth of regions which nucleate under various conditions. We have considered continuous nucleation at a constant rate, nucleation site saturation, continuous nucleation with nucleation-exclusion zones surrounding the growing regions, and decaying nucleation rates. The microstructures resulting from these varying conditions differ markedly in their topological and geometric properties.

We have also developed a program which models grain boundary motion in structures resulting from nucleation and growth. Results from these models will be compared to observed microstructures in thin metallic and semiconductor films and should provide insight into the mechanisms of microstructural evolution and control.

The thermodynamicaL and structural behavior of a (110) face of a (12–6) Lennard-Jones fcc solid has been investigated by MoLecuLar Dynamics computer simulation on the solid-gas coexistence Line up to a temperature T= 0.94 TM (TM: melting point). We have found evidence for cooperative defect production on free surfaces which Leads to a structural transitiDn above T≈0.7 TM. This transition is studied using as an order parameter the excess energy for surface Layers due to missing bonds parallel to the surface with respect to the bulk. Furthermore we report the vaLues of the mean square displacement for surface and bulk atoms as a function of temperature. Despite their high values at the surface, surface Layers are not molten but only highly disordered above the transition temperature.

The ICECAP code is applied to charged and uncharged color centers in alkali halides and alkaline-earth oxides, to test the usefulness of complete-cation pseudopotentials for reproducing the cluster boundary conditions. The physical model includes consistency up to electrostatic octupole order between the Hartree-Fock cluster and the surrounding infinite shell-model lattice. The total energy of the system is determined variationally, including distortion and polarization of the cluster and lattice, and LCAO-MO gaussian-localized cluster wave functions. Electronic states with the lattice unrelaxed are also analysed, yielding color-center optical transition energies. Furthermore, consistency between quantum (cluster) and classical (shell-model) descriptions of the perfect lattice is tested.

A method for constructing Jahn-Teller coupled vibronic states is developed by projecting distorted electronic eigenstates onto undistorted states and, with the aid of irreducible tensor methods, the force constants can be identified with reduced matrix elements of the normal mode expansion of the vibrational Hamiltonian. A scheme is then realized for estimating the force constants as well as the total vibronic energy from the calculated energie of the distorted electronic states. An application of the method to the 1 E ⊗E vibronic state of the neutral silicon vacancy is presented as an illustration

A survey is given of various electron-phonon effects which have been vcalculated for the metals Nb, Mo, Ta, Pd, and Cu. These effects include the mass enhancement γ, superconducting Tc, electrical and thermal resistivity, Hall coefficient, magnetoresistance, and the successfully tested predictions of linewidths yQ of phonons. The calculations use local density approximations (LDA) energy bands, experimental phonons, and the rigid muffin tin (RMT) approximation. Mesh size noise is less than 1% and the Bloch-Boltzmann integral equation has been solved to unprecedented accuracy.

Computer simulations open new avenues in investigations of complex material systems and phenomena. The methodology, development and applications of molecular dynamics simulations are discussed and the wealth of microscopic information revealed via simulations of interphase-interfaces, liquid metals, metallic glasses and small aggregates are demonstrated.