Modelling multi-phase oxide formation in aqueous environments

Surface oxidation of austenitic alloys, including stainless steels (e.g. 304 or 316) is well understood and the mechanisms of oxide formation widely accepted. Or so we think. Around 20 years ago, our group (and others elsewhere) started characterizing these oxides with high-resolution techniques such as scanning transmission electron microscopy (STEM) and atom-probe tomography (APT). As the spatial resolution and analytical accuracy of these techniques improved, it became clear that these oxides were not as simple as traditionally accepted. For instance, they are rarely stoichiometric. While the inner oxide layer is generally described as chromite (FeCr₂O₄), as confirmed by diffraction, a careful quantification revealed that there are obvious cation deficiencies, particularly trivalent ones. Another observation, with more important repercussions for safety, is the fact that voids and nanocavities can be ubiquitous. This project aims to provide a mechanistic explanation to all these phenomena by using modelling. The main objectives are:

1. Demonstrate that cation-deficient oxides can form (with different number of vacancies instead of 3+ or 2+ cations) as a way of compensating for the missing charge. Cr₂O₃ and FeCr₂O₄ would be studied.
2. Rank which oxides (including those with non-stoichiometric compositions) are likely to form first depending on local composition/cation availability
3. Demonstrate that defects/lattice strain promote matrix diffusion and compare how Fe, Cr and Ni behave (oxidation is the driving force). Our observations suggest that lattice strain can increase diffusion rates by several orders of magnitude. This changes the local availability of, e.g. Cr, influencing which oxides can form and changing passivity behaviour

This project will use a variety of first principles methods to investigate these questions.  Methods based on density functional theory (DFT) will be used to determine the chemical bonding and relative stability of different phases and defect structures, as well as investigate the effect of defects and strain on diffusion rates.  Structure searching techniques, using both density functional theory and machine-learned potentials, will be used to expand the structure space investigated.