Monte Carlo and molecular dynamics simulations of near-surface phenomena

Abstract

In this thesis, applications of the Monte Carlo and molecular dynamics simulation methods on different surface and near surface phenomena are presented. Monte Carlo simulation is used to investigate positron and electron slowing down in solid matter. The description of elastic scattering is based on accurate cross sections of effective crystalline atom potentials. Inelastic processes are described separately for each energy level by Gryzinski's excitation function. Various materials are studied and several electron and positron slowing down parameters and distributions are extracted. For example, backscattering and transmission energy and angular distributions, backscattering and transmission yields, and penetration depth and energy deposition distributions are calculated. The results are used to analyze and interprete a number of recent experiments utilizing keV electron and positron beams. Positrons lose less energy and scatter to smaller angles than electrons when slowing down in material. This is demonstrated by several extracted parameters. In particular, the differences in elastic scattering cross sections result in some drastic differences between positrons and electrons. Positrons have larger probability to penetrate through thin films and they penetrate deeper into material than electrons. A clear consequence of the different elastic scattering cross sections for positrons and electrons is that the backscattering probability from materials is for positrons about half of that for electrons. Differences between positrons and electrons increase as a function of atomic number of the target material and decrease as a function of incident energy. The implantation profile of particles is found to be a negative of the derivative of a Gaussian function in contradiction to earlier assumptions that the profile is an exponentially decreasing function. Analytic fits are presented for the stopping profiles and their Laplace transforms as well for the energy deposition and ionization profiles. Molecular dynamics simulation methods are used to study (i) damage production in aluminum (110) surfaces due to low-energy argon ion bombardment and (ii) the premelting effects of solid noble gas surfaces. Appropriately constructed pair potentials were assigned between the particles, and an electronic friction term proportional to the velocity was used for energetic ions. Of particular interest in (i) are the defect and implanted atom distributions, which are compared against recent experiments. In (ii), the simulations show the equilibrium existence of liquid-like layers on the densely packed surfaces well below the bulk melting temperature. The main results in (i) are the follows. The mean vacancy concentration depth depends only slightly on the incident angle. The total number of vacancies is almost independent of the incident ion dose for very oblique angles of incidence (0 > 45°). For small incident angles (near to normal) the number of vacancies per incident ion is small due to effective channeling of Ar+ ions, but the total number of vacancies increases considerably with increasing dose. Vacancy profile is found to have a clear peak in the topmost atomic layers and a broader tail deep in the material. The interstitial and Ar+ ion profiles are clearly deeper in the material than the vacancy profile. In (ii), a layer-by-layer premelting of Lennard-Jones (111) surfaces is observed. Also the (100) surfaces premelt, but the disordering mechanism for the loosely packed (110) surfaces is roughening. Furthermore, a general rule seems to be that melting proceeds along the directions of high packing densities.