Development of electronic structure and kinetics methods for the rational design of electrocatalysts

Abstract

Computational modeling has untapped potential for novel material and chemical discovery. In this thesis, we explore ways to improve existing modeling methods, and how to apply these methods to design novel graphite-conjugated catalysts (GCCs). For improving electronic structure methods, we first present an extended study of bootstrap embedding theory (BET) and its ability to recover static correlation, as well as a proof on BET's ideal convergence properties. We then present a theoretical analysis using density functional theory (DFT) on a class of GCCs containing cationic nitrogen atoms, which are particularly active for catalyzing the oxygen reduction reaction (ORR). Using a mixture of high-throughput screening, statistical analysis, and computational exploration guided by chemical intuition, we design several novel GCCs, several of which DFT predicts would have enhanced activity above existing GCCs. Furthermore, our analysis reveals that known ORR scaling relations hold for GCCs, but hint at the possibility of breaking these relations with careful molecular engineering of the GCC active sites.