Molecular dynamics and model potential energy surfaces for two-dimensional film, nickel, iron and alloy systems

Computer simulation and density functional theory (DFT) in the local density approximation (LDA) are two important tools in scientific investigation. They play different but complementary roles in studying many physical phenomena concerned with finite temperature effects, and/or large number of particles of which can not be understood by simple analytic models. The former, based on classical mechanics with some model interparticle interaction potential, is now becoming the third dimension of scientific methodology, and is useful for dealing with systems of low symmetry. The latter, a theory constructed from first principles with suitable approximations, is widely used to study electronic structure, energy levels and total energy, and is very powerful in the investigation of complex systems such as transition metals and their alloys. It is a goal today to marry these two approaches in one way or another. On one hand, in the simulation scheme a suitable potential model is very important in order to give a description of physical systems with quantum dynamical features. On the other hand, it is forbidden by the practical computer memory size and speed to solve the Shrodinger equation for a large system of low symmetry.;This thesis contributes in developing a method of connecting the two methods discussed above through three steps: First, with a simple screened Coulomb potential in MD we investigate a two dimensional system through its thermodynamical properties, structural information, phase diagrams, and phase transitions. Gibbsian dynamics (which simulates the isotherm-isobaric ensembles) is built into our calculations. Second, we study the binding energy curves of Pt, Fe and Ni clusters with up to four atoms along certain symmetrical geometries and extract the effective pair interaction between particles, using discrete-variational (DV) X...