A Deeper Understanding of Material Properties Through Computational Model Development.

Abstract The pace of materials development can be significantly increased by computational models that accurately capture the characteristics of experimental materials. While advances in computational architecture have enabled researchers to calculate the properties of pure materials with high fidelity, it is often disorder of the perfect crystal that gives rise to technologically relevant properties. This thesis aims to contribute a robust modeling framework for disordered materials by developing improved criteria to access metal oxide point defects, interfaces, and nanoparticles using theories that are efficient and easily generalizable. Insights into technologically relevant materials are developed, including discovery and assignment of spectroscopic signals to new defect geometries in doped SrTiO3, identification of ring-like atomic motifs in amorphous Al2O3 responsible for experimentally observed electronic properties, and combination molecular dynamics and density functional theory approaches to produce more realistic amorphous TiO2 nanoparticles and interfaces.