Metalloid impurities have a very low solubility in Tantalum, and therefore prefer to segregate at the grain boundaries (GBs). In order to analyze the energetics of the impurities on the Tantalum GB, the LMTO calculations were performed on a simple 8-atom supercell emulating a typical (capped trigonal prism) GB environment. The so-called “environment-sensitive embedding energies” were calculated for Hydrogen, Boron, Carbon, Nitrogen, Oxygen, Phosphorus, and Sulphur, as a function of the electron charge density due to the host atoms at the impurity site. The calculations showed that, at the electron density typical of a GB, Carbon has the lowest energy (followed by Nitrogen and Boron) and thus would compete with the other impurities for the site on the GB, tending to displace them from the GB. The above energies were then used in a modified Finnis-Sinclair embedded atom approach for calculating the cohesive energies and the equilibrium interplanar distances in the vicinity of a (111) Σ3tilt GB plane, both for the clean GB and that with an impurity. These distances were found to oscillate, returning to the value corresponding to the equilibrium spacing between (111) planes in bulk BCC Tantalum by the 10th-12th plane off the GB. Carbon, Nitrogen and Boron somewhat dampen the deformation wave (making the oscillations less than in the clean GB), while Oxygen, Phosphorus and Sulphur result in an increase of the oscillations. The cohesive energies follow the same trend, the GB with Carbon being the most stable. Thus, Carbon, Nitrogen and Boron may be thought of as being cohesion enhancers, while Oxygen, Phosphorus and Sulphur result in decohesion effects.

Recently we developed a new nonequilibrium molecular simulation method [1] that allows the direct study of interdiffusion in multicomponent mixtures. The method combines stochastic insertion and deletion moves characteristic of grand canonical (GC) simulations with molecular dynamics (MD) to control the chemical potential μi of a species i. Restricting the insertions and deletions to two separate control volumes (CV's) one can apply different μ's in distinct locations, and thus create chemical potential gradients. DCV-GCMD can be used to study transient phenomena such as the filling of micropores or used in steady-state mode to determine the diffusion coefficients in multicomponent fluid mixtures. We report on the effects of molecular interactions and demonstrate how in a sufficiently nonideal ternary mixture this can lead to up-hill or reverse diffusion. In addition we introduce a novel extension of DCV-GCMD that is specifically designed for the study of gradient-driven diffusion of molecules that are simply too large to be inserted and deleted.

Most perovskite structure oxides exhibit structural phase transitions from a hightemperature cubic phase to a distorted low-temperature phase which can be described by the freezing-in of one or more phonon modes of the cubic structure [1]. The first-order cubic-tetragonal ferroelectric transition in PbTiO3 at Tc = 763 K involves the freezing-in of a single F15 polar mode. In PbZrO3 , the structure of the antiferroelectric low-temperature orthorhombic phase is far more complicated, with forty atoms per unit cell and the freezing-in of R25 and Σ3 modes, perhaps accompanied by other modes as well [2][3].

The lattice vibrations influence the ordering tendencies in substitutional alloys through the configurational dependence of the vibrational free energy. In this paper we study the size of this effect and the trends with chemical affinity and size mismatch. By analyzing available experimental information and by studying simple model systems for which the phonon frequencies can be easily computed, we find that a) changes in transition temperatures of order-disorder transformations of 30% due to the lattice vibrations may not be uncommon; b) the transition temperatures are usually lowered when the vibrations are included in the modeling of the system, except for certain combinations of size mismatch and chemical affinity; and c) that the largest contribution to the configurational dependence of the vibrational free energy is due to the intermediate and high frequencies of the spectrum.

Simple pair potentials are constructed which give hexagonal- or squarelattice ground states in two dimensions, depending on the value of a single parameter controlling the width of the potential well. Molecular dynamics calculations for free clusters of a few hundred particles are used to examine a structural transition from the square lattice, at low temperatures, to the hexagonal lattice at high temperatures. Another structural transition (melting) is identified by the onset of diffusion. The melting temperature is found to be a minimum near the point where the barrier between the two structures, as a function of lattice strain, is also a minimum.

In order to get a better understanding of the energetics of ZrO2 (zirconia) ab initio calculations with the full potential linear muffin tin orbital method ( fp LMTO) have been performed on the tetragonal structure over a range of c/a and sublattice displacement. A new semi-empirical shell model is developed which makes use of Hartree-Fock calculations and includes compressible anions and quadrupolar distortions. The empirical model predicts energies for tetragonal distortion in agreement with the fp LMTO calculations. Furthermore, it enables us to understand why the seven-fold coordinated monoclinic phase is the low temperature equilibrium structure.

We have used constant pressure molecular dynamics calculations to explore the behavior at various temperatures of two molecular crystals: benzene and a brominated phenyl compound. We observed a melting transition by heating the crystals from a low temperature. In the case of benzene, we performed one heating run of about 1 ns and obtained agreement with the experimental melting point to within some 8%. We have also simulated the melting of a more complex molecular crystal that contains bromine and phenyl groups. We performed four heating runs, with different rates of heating. For total simulation times of about 100, 220, 770, and 1 I50ps, the heating runs predicted melting temperatures that differed from the experimental melting temperature by 53%, 33%, 25%, and 9% respectively.

Using the molecular dynamics (MD) method, we have studied the effect of pressure on the distribution of rings and their relationship to intermediate range correlations manifested as the first sharp diffraction peak (FSDP) for Si02 glass. A systematic analysis of the modifications observed in the FSDP for densities ranging from 2.0 to 3.2 g/cm3 and temperatures from 0 to 1500 K is reported. The decrease in the height of the FSDP with increasing density is found to be proportional to the decrease in the number of 6-fold rings. For the density and temperature ranges studied in SiO2 glass, the full width at half maximum (FWHM) of the FSDP remains unchanged.

We apply our recently developed first-principles theory of atomic short-range order in disordered multicomponent alloys to CuAuZn2 and AgAuZn2, two alloys which display partially-ordered B2 phases as well as fully-ordered Heusler structures. The calculated Warren-Cowley pair correlation functions for both alloys peak at the special points 〈½ ½ ½〉 and 〈100〉, with the peak at 〈100〉 dominant. This is indicative of a tendency to B2 order at high temperatures, and a subsequent tendency to Heusler-type order at lower temperatures. An analysis in terms of effective interactions shows that the Au Zn interaction drives the 〈100〉 ordering tendency, while the Ag Au and Cu Au interactions contribute more to the 〈 ½ ½ ½〉ordering tendency - this effect is larger in CuAuZn2 and should lead to a slightly higher transition temperature for the B2 to Heusler transition in CuAuZn2 as compared to AgAuZn2.

In this paper we report a molecular dynamics simulation at constant pressure and constant temperature of the structural phase transition occurring in epitaxial FeSi2 from the fluorite phase (metallic and pseudomorphic) to orthorhombic one (semiconductor and bulk stable). The evolution of the electronic density of states is carefully monitored during the transformation and we can show that the Jahn-Teller coupling between the density of states at the Fermi level and the lattice deformation drives the metal-semiconductor transition.

In a theoretical search for new hypotetical sp3-bonded carbon structure containing five-fold rings as a possible result of fullerene transformation under pressure, we have found a triclinic form of diamond with 16 atoms per unit cell which we called tcl-16. We have calculated the ground state structure, the cohesive energy, the bulk modulus and the electronic density of states by means of tight binding molecular dynamics (TBMD). Finally we have compared the phonon spectra at F to existing Raman data for a non-cubic- phase of diamond.

Semi empirical methods have shown considerable promise in aiding in the calculation of many properties of materials [1,2]. Materials used in engineering applications have defects that occur for various reasons including processing [3]. In this work we present the first application of the BFS (Bozzolo, Ferrante and Smith) method for alloys [1] to describe some aspects of microstructure due to processing for the Cu-Au system (Cu-Au, CuAu3 , and Cu3Au). We use finite temperature Monte Carlo calculations, in order to show the influence of ’heat treatment‘ in the low-temperature phase of the alloy. Although relatively simple, it has enough features that could be used as a first test of the reliability of the technique. The main questions to be answered in this work relate to the existence of low temperature ordered structures for specific concentrations, for example, the ability to distinguish between rather similar phases for equiatomic alloys (CuAu I and CuAu II, the latter characterized by an antiphase boundary separating two identical phases).

The clustering in alkali silica glasses (M2O) x (SiO2)1-x, where M is either Na or Rb, and x ranges between 0 and 0.4, was studied using molecular dynamics simulations. Computations were performed using a semi-empirical potential, including two- and three-body terms as well as dynamic partial charge transfer. Characterization of the structures was based on pair correlation functions, neutron static structure factor and ring statistics. Results have shown a much stronger tendency to cluster in case of Na than of Rb. The irregular arrangement of Na is evidenced by a decay pattern in the Na-Na pair correlation function, which can be associated with a fractal dimension. The clustering tendency can be attributed to the differences in the way the network structure is able to accommodate the introduction of modifying cations. As opposed to Rb, the smaller Na cation can fit on sites only slightly larger than is required for Si. This would distort the structure but hardly alter its topology. Rb on the other hand, requires larger interstices, which results in a reconstruction of the network and a significantly different intermediate range order. Na can achieve a similar influence on the network structure if they group in pairs.

The FeCr sigma phase is a good example of a complex structure: it. has 30 atoms in the unit cell and 5 inequivalent lattice sites, and it belongs to the class of tetrahedrally close packed structures, also known as Frank-Kaspar structures. So far. such structures have riot been treated within a first-principles statistical thermodynamics framework. It will be shown that dtlme to advances in algorithms and hardware important features of the phase stability of complex phases can be computed. The factors which affect the stability of the sigma phase have been studied using carefully selected supercells for electronic total energy calculations. cluster variation calc:ulations in the tet.rahedron approximation were performed to evaluate the effect of partial disorder and of finite temperature. The preferred occupancy of the 5 lattice sites has been investigated and is compared with experimental determinations.

The semiempirical BFS method for alloys is generalized by replacing experimental input with firstprinciples results thus allowing for the study of complex systems. In order to examine trends and behavior of a system in the vicinity of a given point of the phase diagram a search procedure based on a sampling of selected configurations is employed. This new approach is applied to the study of the/3 phase of the Ni-Al system, which exists over a range of composition from 45–60 at. % Ni. This methodology results in a straightforward and economical way of reproducing and understanding the basic features of this system. At the stoichiometric composition, NiAl should exist in a perfectly ordered B2 structure. Ni-rich alloys are characterized by antisite point defects (with Ni atoms in the Al sites) with a decrease in lattice parameters. On the Al-rich side of stoichiometry there is a steep decrease in lattice parameter and density with increasing Al content. The presence of vacancies in Ni sites would explain such behavior. Recent X-ray diffraction experiments suggest a richer structure: the evidence, while strongly favoring the presence of vacancies in Ni sites, also suggests the possibility of some vacancies in Al sites in a 3:1 ratio. Moreover, local ordering of vacant sites may be preferred over a random distribution of individual point defects.

Mercury has perhaps the strangest behavior of any of the metals. Although the other metals in column IIB have an hcp ground state, mercury's ground state is the body centered tetragonal βHg phase. The most common phase of mercury is the rhombohedral αHg phase, which is stable from 79K to the melting point and meta-stable below 79K. Another rhombohedral phase, γ71Hg, is believed to exist at low temperatures. First-principles calculations are used to study the energetics of the various phases of mercury. Even when partial spin-orbit effects are included, the calculations indicate that the hexagonal close packed structure is the ground state. It is suggested that a better treatment of the spinorbit interaction might alter this result.

Accurate density functional calculations are performed to investigate the structure and vibrational dynamics of carbon monoxide adsorbed to the (100) surface of copper. The adsorbate and substrate are considered as a unified system, with atoms of each treated on an equal footing. Coupling between the two components is found to have a significant effect. In particular, frustrated translational motion mixes strongly with transverse phonons of the substrate to form a broad resonance. Direct computation of anharmonic coupling between the internal CO bond stretching mode and other adsorbate-weighted modes confirms the experimental conclusion that the transient CO-stretch response seen in recent pump-probe studies is an indirect probe of the transient dynamics of frustrated translations. In this light, the computed resonance between this mode and substrate phonons suggests a dephasing mechanism to account for the observed relaxation dynamics.

An atomistic Monte Carlo (MC) method has been used to predict equilibrium segregation of isovalent cations to (001) surfaces in (Fex.Mn-x)O and (NixCol-x)O. The surface is found to be enriched with solvent in both systems. Long-range electrostatic interactions and atomic motions that occur on small time scales make the MC approach very computationally demanding. The Free Energy Minimization (FEM) method is a more efficient alternative for performing such segregation simulations, but involves several approximations. Comparison of the surface segregation profiles determined using the MC and FEM simulation methods show that the two are essentially indistinguishable. The FEM results can be obtained about 1,000 times faster than the MC predictions. Therefore, the FEM method is a practical and accurate alternative to the more cumbersome MC approach.

The spatial redistribution of niobium atoms near the surface of pure nickel has been considered in the low niobium concentration limit. The calculation of free energy includes lattice relaxation around niobium atoms by using molecular dynamics (MD) incorporating atomistic potentials based on ab-initio quantum mechanical calculations and includes vibrational entropy phenomenologically within the local harmonic approximation.

An overview is provided of an evolving atomistic and kinetic model of semiconductor growth that unifies the main features of strain relaxation in low and high lattice misfit heteroepitaxy. The model reveals a kinetic pathway for dislocation formation during growth with little or no energy cost at low misfits, thus providing a way out of the longstanding dilemma of too high dislocation nucleation energies predicted by classical theories of the equilibrium behaviour of a fixed number of particles at low misfits. The essential kinetic processes underlying the model are identified on the basis of comparison of the predictions of kinetic Monte-Carlo simulations of growth with real-time or in-situ data obtained in such experiments as reflection high-energy electron diffraction (RHEED) and scanning probe microscopy (SPM). Relative significance of these atomistic kinetic processes is shown to naturally lead to strain relaxation via defect initiation at low misfits while maintaining smooth surface morphology or at high misfits change to 3-dimensional morphology while initially maintaining coherence. The potential role of steps in providing sources for defect formation is examined through molecular dynamics simulations of Ge overlayers on Si (001) stepped surfaces.