Computational defect engineering of d⁰ oxide & oxynitride photocatalysts

Sunlight driven water splitting holds great potential as a renewable source of H2 fuel or driving a separate reaction for the chemical transformation of CO2 in additional solar fuel products. In optimizing a semiconductor material for photo(electro)catalysis, research in recent years have demonstrated the significant and complex impact that defects have on the water splitting process. Much of defect engineering has been established thus far in oxide semiconductor photo(electro)catalysts while separately mixed-anion materials have also garnered significant research interest in the space of water splitting. This thesis addresses the two burgeoning fields in sunlight drive water splitting research and aims to highlight and explore the implications of defect engineering in oxynitride photocatalysts through quantum chemical modelling. The accuracy of quantum chemical modelling of transition metal containing photo(electro)catalysts (especially in the system sizes required for insight into defects) with density functional theory (DFT) can potentially be improved through the use of a Hubbard U correction, where higher levels of theory would be prohibitively expensive. Unfortunately, the choice of Hubbard U correction in modelling the defect chemistry and optical properties of such transition metal containing semiconductors (d0 and d10) is non-trivial. In particular, the effect of the Hubbard U correction on transition state characterization is poorly understood. As such, this thesis address these problems by benchmarking the effect of a two-site Hubbard U correction: Ud,p (applied to the transition metal d orbitals and anion p orbitals), on the optical properties and the oxygen vacancy diffusion barriers of widely studied photocatalysts materials.