Monte Carlo simulation, in conjunction with the embedded atom method, has been used to model the composition and structure of a semicoherent (001) interphase boundary separating coexisting Cu-rich and Ag-rich phases in a binary Cu-Ag alloy. The results are compared with earlier simulations of the same boundary in a Cu-Ag-Au alloy, in which Au was found to segregate to the interface, and the boundary was found to be unstable with respect to break-up into {111} facets. The boundary in the binary system is also unstable to faceting, but displays both {100} as well as {111} facets. It is concluded that Au segregation in the ternary alloy plays an important role in stabilizing the {111} facets. The interplay between the misfit dislocations present at the interface, and the compositional features of the boundary are also discussed.

Diffuse TEM and x-ray data has been reported by several research groups that has been interpreted in terms of both tweed and (ω -phase instabilities. Computer simulations of 1,024 atom arrays were run at various temperatures using the Voter-Chen Embedded Atom Method (EAM) interactions. If the extra Ni atoms were placed at random on the Al sub-lattice of the B2 structure, at low temperatures both <110>* and <112>* diffuse streaking were observed in reciprocal lattice plots, indicative of tweed and ω - phase instabilities, respectively. The atomic displacements causing the <110> diffuse streaking could generally be classified as {110}<1-10> local shears, in agreement with accepted models of tweed. However, the displacements producing the <112>* streaking were best described as coordinated atom row motions in <111> directions, which does not agree with traditional ω -phase models based on (111) interplanar collapse. Both types of diffuse streaking decreased with increasing temperature, but the <110> type was stronger at all temperatures and persisted to higher temperatures. Furthermore, the atomic displacements producing the streaking were static at low temperatures ( persistence times greater than 1000 vibrational periods) but became dynamic at higher temperatures ( persistence times of the order of a few vibrational periods). Another interesting result was that by merely rearranging the extra Ni atoms into an ordered superlattice pattern, virtually no diffuse streaking of any type was observed at any temperature, strongly suggesting that compositional disorder ( or other kinds of point defects) are essential to induce tweed instabilities.

*The asterisk represents a quantity in reciprocal space.

The basic nature of kinetic processes is examined theoretically. When an isothermal development of order is studied theoretically, sometimes an ordered phase different from the final equilibrium one transiently predominates, while the free energy (Fsys) decreases monotonically although very slowly. There is another category of kinetic changes for which the convenient interpretation is that the Fsys surface itself changes in time. In this second category, a state can initially be a local minimum of Fsys and thus can be called a ‘metastable’ state, but after a certain time, as the Fsys surface itself changes, a kinetic path appears along which Fsys decreases and the system can change along it. To include both categories, the concept of the PSEUDOSTABLE phase is proposed. It is shown that even in the nucleation process, Fsys for the entire system decreases monotonically, although the local free energy may fluctuate.

We use molecular-dynamics simulations to investigate relaxation processes near the liquid-glass transition for a realistic model of the metal-metalloid system Ni80P20. We find that relaxation proceeds in two stages, excluding phonons: fast (or conformational) relaxation, related to local rearrangements of atoms, and slow relaxation, connected with atomic transport, i.e. diffusion. These processes are usually referred to as β and α, respectively. Our simulations show that diffusion exists even in the glass state, where it proceeds mostly by jumps, in contrast to the liquid phase where it is continuous. This relaxation mechanism is well described by a stretched exponential (Kohlrausch) law, in accord with recent modecoupling theories for supercooled liquids. The fast relaxation regime, on the other hand, does not appear to be well described by the theory.

We calculate the shear elastic constants of the alloy NiZr2 by molecular dynamics simulations in the crystalline and amorphous phases as well as upon introduction of antisite defects in the crystal at T=300K. For S (long range order parameter) equal to 0.5, the system is amorphous and C' is larger than the same quantity relative to the crystal whereas C44 and C66 are smaller.

The problem of relaxation to the quasistationary mode of homogenious nucleation in the supersaturated vapor is being studied analytically. A simple analytical function of the nucleation rate behaviour is obtained, which is in a good agreement with the results of the numerical analysis.

The time lag in transient nucleation was originally introduced to characterize the temporal evolution of the nucleation current in the simplest possible way. Previous attempts to calculate the time lag have relied mainly on either the approximate solution of the continuous Fokker-Planck equation or the numerical solution of the discrete rate equations. It is shown that the exact time lag can be calculated using a different method which requires knowledge only of the initial and the steady-state cluster populations.

Previous attempts to simulate by Molecular Dynamics the spontaneous nucleation and growth of a crystalline Stillinger-Weber ‘silicon’ from the liquid have been essentially impossible because of constraints on system size and time scales. We have overcome these limitations by studying the related problem of the disintegration of crystalline ‘embryos’ into the liquid phase at temperatures slightly above the melting point. Molecular Dynamics simulations using the Stillinger-Weber potential were performed by embedding crystallites of 400 atoms in a liquid consisting of approximately 3600 atoms. During each simulation, the time-evolution of the size and shape of the embryo was followed until it became indistinguishable from the liquid. These simulations provide intriguing new information on the atomic processes involved in dissolution and on the macroscopic kinetics of small clusters. Comparisons of results at different temperatures, system sizes and initial configurations are shown and the implications of these cluster dynamics for crystal growth in supercooled liquids, homogeneous nucleation, and transient nucleation are discussed.

A theory is developed for the formation of precipitates in very dilute hypereutectic two phase systems during solidification. It is based on a mechanism for the growth of microdefects at the melt -interface. The model presents a solution of the diffusion problem for the solute atoms in the melt and provides an analysis of the factors which determine the particle size and their mean distance. The possibility of microdefect nucleation in front of the melt -interface, which can lead to very large precipitates, is discussed. The model is suitable to describe the formation of microdefects in semiconductors where impurity concentrations are generally low but may exceed the solubility limit in the melt. The results of the theory are compared with experimental investigations of the silicon carbide formation in carbon rich silicon crystals grown from the melt.

We study the dynamics of an interface driven far from equilibrium in three dimensions. The relationship of the phenomena to self-organized critical phenomena is discussed. Numerical results are obtained for three models which simulate the growth of an interface: the Kardar-Parisi-Zhang equation, a discrete version of that model, and a solid-on-solid model with asymmetric rates of evaporation and condensation. We show that the three models belong to the same dynamical universality class by estimating the dynamical scaling exponents and the scaling functions. We confirm the results by a careful study of the crossover effects. In particular, we propose a crossover scaling ansatz and verify it numerically. Furthermore, the discrete models exhibit a kinetic roughening transition. We study this phenomenon by monitoring the surface step energy which shows a drastic jump at a finite temperature for a given driving force. At the same temperature, a finite size scaling analysis on the bond energy fluctuation shows a diverging peak [1].

The phase behavior of silicon is studied using the Modified Embedded Atom Method (MEAM) proposed by Baskes, Nelson and Wright. We find this model to quantitatively reproduce aspects of both the solid and liquid phases with an accuracy comparable to the widely-used Stillinger-Weber (SW) potential, thus providing an opportunity to examine the consistency of results obtained previously using the SW model. Although the models are very different, they both produce solid-liquid interfaces on both silicon (100) and (111) which have very similar morphologies. We find that the MEAM predicts the melting point of silicon to be 1445K, or 14% lower than the experimental value. The model also predicts the heat of melting to be 34.9 kJ/mol, 45% lower than the experimental value of 50.6 kJ/mol, and a liquid density which is 5.4% larger than that of the solid at the melting point, which is the density ratio found by experiment. The liquid density is found to be too low with respect to experiment. We also suggest a correction which might be applied to the MEAM model to improve its description of the liquid phase.

A mathematical model is presented to describe the fundamental phenomena occurring in the vicinity of the interface between a polymerizing medium and a non-solvent. The model combines the principles of polymerization kinetics, mass transfer and polymer solution thermodynamics and is applied to the polymerization of methyl methacrylate with water. Numerical simulations indicate the development of large interface velocities and the formation of a dense polymer skin at very low monomer conversions in agreement with experimental observations.

Energy barriers for ‘magic’ strains, large pure shear strains which transform a lattice into itself, are computed for aluminum and iridium, using the linearized-augmented-plane-wave method, and for silicon, using the plane wave pseudopotential approach. The barrier energies are found to be of the order of the thermal energy available at the melting temperature. Volume changes associated with the barriers are consistent with the fact that the metals decrease in density while silicon increases in density upon melting.

The effects of scanning speed and supercooling were studied during zone-meltingrecrystallization (ZMR) of silicon-on-insulator (SOI) wafers. Using finite difference methods, a numerical simulation of the ZMR process was developed which captures all of the optical and thermal property changes during phase transformation. The effects of supercooling and scanning speed on the temperature profiles, the total width of the melt-zone and the width of ‘slush’ region were investigated. The melt-zone width increases for increasiongf thdee gfrreeeezsi nogf supercooling and decreases for increasing strip heater velocities. The combined effects on the melt-zone width were shown for various scanning speeds and degrees of supercooling. Supercooling also had a significant effect on the size of the freezing ‘slush’ region which was shown to decrease for increasing degrees of supercooling.