Initial Oxidation Kinetics of Al(111): A Monte Carlo Study
Doktorsavhandling, 2003

Almost all metals around us have a thin oxide layer on top of the bulk metal underneath. Often this oxide film protects the metal against further oxidation, or corrosion. Examples are the oxide film that makes stainless steel "stain-less" or the ones on titanium and aluminium alloys, which are very corrosion resistant. Thin metal oxide films have also many other technological applications, based on their mechanical, optical, electrical or biocompatible properties. In spite of their technological importance there are still many unanswered questions regarding the details of how the oxide films form and grow on metal (and semiconductor) surfaces. This concerns both the very initial nucleation of the oxide film and its subsequent growth. In this thesis the former regime is addressed. The situation of interest here is when a fresh metal, i.e. a clean metal surface with only metal atoms and no prior oxide or adsorbates on the surface, is exposed to oxygen molecules. A common sequence of events for many metals is that oxygen molecules initially adsorb and dissociate on the surface, and form a two dimensional layer of chemisorbed oxygen atoms. As the surface coverage increases, nucleation of oxide eventually occurs. It is this process that has been studied in this thesis by means of Monte Carlo (MC) simulations. The specific model system for this work is the Al(111) surface exposed to O2 molecules. This system constitutes an extremely interesting model system both because of its technical importance and because it exhibits a complex and rich behavior in the initial chemisorption and oxide nucleation regime. Furthermore there is a richness of experimental data available, both regarding the structure, i.e. the atomic arrangements of oxygen atoms on the surface, and about the kinetics of oxygen uptake and oxide nucleation. Finally there are a number of open, challenging questions about the whole process. The adsorption dynamics is peculiar for the O2/Al(111) system; upon dissociation the oxygen atoms can either take nearest neighbor sites on the surface ("normal dissociation"), or undergo so called "abstraction", meaning that one of the atoms is ejected into vacuum or to a distant (of order 10 nm) position on the surface. The branching ratio between these two channels depend strongly on the incident energy. This complex behavior opens up for very special kinetics and affects the distribution of oxygen atoms on the surface and the relative abundance of chemisorbed monomers, dimers, trimers etc. It also affects the oxide nucleation. The latter occurs when the chemisorbed O coverage exceeds 0.15 monolayers. In the present work a Monte Carlo simulation platform, based on the lattice-gas approach, has been established and applied to simulate the kinetics of the O2/Al(111) system. The adsorption kinetics in the low coverage regime and the initial oxide nucleation kinetics have been simulated, incorporating the available experimental and first principle calculations results. Elementary processes that are not yet known in sufficient detail, but suggested to play a role, have been explored in the simulations to see if they can explain the experimental data, and / or suggest new experiments. In the early part of the work the experimental adsorption kinetics and resulting distribution of adatoms in the thermal regime could be described quite well. In the later part incorporation of the results for higher incident energies were treated. In the last paper incorporation of short and long range O-O adatom interactions made it possible to reproduce the experimental data, including both the initial chemisorption phase and the transition to oxide formation on the surface.

oxide nucleation and growth

dissociative adsorption

metal oxidation

Monte Carlo simulations




surface oxidation


Dogan Ergun Oner

Chalmers, Teknisk fysik





Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 1954

Applied physics report - [Department of Applied Physics, Chalmers University of Technology and University of Göteborg]: 03:7