Computational Studies of Perovskite-Structured Ion Conductors.

This thesis describes the application of advanced computational techniques to the detailed study of the defect chemistry and ion transport properties of mixed metal oxides. In particular, we investigate the perovskite-type oxides based on the general formula LaBO3, (where B = Cr, Mn, Fe, Co) which have important applications in solid oxide fuel cells, and as effective heterogeneous catalysts. For the four compounds considered, a common set of interatomic potentials was derived that correctly reproduces their observed cubic structures. The simulations consider intrinsic disorder and find minimal deviation from ideal stoichiometry. We examine the energies of solution for a range of alkaline-earth and alkali metal dopants on the cation sites, and also investigate dopant-vacancy clusters. Both static lattice and molecular dynamics methods are used to study oxygen ion diffusion. The results support models in which diffusion is mediated by oxygen vacancies, with the calculated diffusion coefficients and migration energies in good agreement with available experimental values. We also explore protons in these perovskite oxides and find that the dissolution of water is an exothermic process. The mechanism and energetics of proton migration are investigated by ab initio quantum mechanical calculations. Finally, lattice simulations are performed on another important family of metal oxides, bismuth molybdates, which find use as selective oxidation catalysts. We successfully reproduce the complex structure of Bi2Mo2O9 and investigate key redox processes, oxygen migration mechanisms and surface structures. The reduction reaction involving loss of oxygen is predicted to be the most favourable process, which is consistent with the observed catalytic behaviour