At the catalytic interface, catalysts are represented by an ensemble of cluster isomers or surfaces, each of which may contribute to the system’s stability, activity, selectivity, and resistance to sintering or poisons by reaction byproducts. I use ab-initio methods in conjunction with statistical-mechanical arguments to predict physico-chemical properties and guide development of premiering catalysts. Specifically, I utilize the work-horse of computation, density functional theory (DFT) calculations, to characterize systems of interest. Statistical-mechanical arguments such as Boltzmann-populations allow us to understand the role of relevant cluster isomers, surfaces, and reagents within the high temperature regime of real catalysis. My theoretical work applies this ensemble perspective of heterogeneous catalysis to diverse systems, from supported subnano-clusters to surfaces such as defective anatase for photocatalysis and Pt-Ni for fuel cells.

Supported metal clusters often display non-monotonic behavior in which a cluster of a specific n size may be especially active. My theoretical work at UCLA explores the high activity and tunability of supported clusters by investigating the system as an ensemble. Three of my four first author papers represent fundamental, surface science work into tuning supported clusters. In particular, the well-known Pt-Pd resistance to sintering i.e. agglomeration of the clusters to bulk inertness was explained successfully through this ensemble consideration. 1:1 ratios of Pt:Pd featured more accessible isomers than their pure and mixed counterparts, resulting in an entropic contribution to chemical stability. In collaboration with experiment, we identified the highly active Pt7 for ethylene dehydrogenation, each negatively charged cluster able to adsorb and activate a maximum of 3 ethylenes. Pt7’s high activity over that of a similar cluster size, Pt8, resulted from Pt7’s fluxionality, the ensemble is able to access isomer geometries with more exposed Pt sites. Selective de(hydrogenation) forms the basis for fossil fuel refinery through cracking of hydrocarbons3 and, due to its endothermic nature, can act as a self-cooling mechanism for jet engines. It also represents a tractable process for testing the tunability of cluster size and dopant effects. Successive de(hydrogenation) often results in deactivation as catalyst sites are blocked by carbon (coking). We therefore tempered that high activity and predilection towards coke formation by doping Pt7 with the electropositive boron, which sustained activity during successive reaction cycles by adsorbing and activating ~1 ethylene.

The ensemble perspective may also be extended to surfaces by accounting for non-equivalent defect sites, the local minima of reaction reagents and their subsequent products, or the distribution of facets present at the interface. The anatase surface remains ubiquitous in the field of catalysis for its unique photoactivity and reactivity, specifically, for CO2 reduction and water-splitting. CO2 reduction is an intermediate step towards the formation of organic products such as methanol and water-splitting for hydrogen evolution remains key to renewable energy. Both theory and experiment have cited surface defects such as oxygen vacancies to be a major contributing factor in anatase’s catalytic activity. I examined 9 non-equivalent oxygen vacancy sites under varying levels of theory and characterized a new surface oxygen vacancy minimum, whose electrons localized at unique Ti sites as compared to previous studies. This has important ramifications on the catalysis of reaction intermediates due to their interaction with surface oxygen vacancies. For example, the co-adsorption of CO2 and H2O at an oxygen vacancy results in spontaneous splitting of water (global minimum) and the formation of other organic species such as formic acid (local minima). Tilocca, et. al. estimated the barrier to be circa 0.1 eV and this barrier is eliminated in the presence of CO2 and an oxygen vacancy. In addition to a comprehensive study of defective anatase, I also characterized the interface of Pt-Ni nanowires in depth by considering varying lattice constants, facets, and Pt-skins on a sub-alloy of Pt-Ni. The Pt-Ni nanowires expressed high catalytic activity and durability for the oxygen reduction reaction (ORR). ORR remains the limiting factor in fuel cells due to cost (requiring high Pt-loadings) and kinetics (occurring at a rate of six orders of magnitude slower than the hydrogen oxidation reaction). By considering the interface in such complexity, I converged upon the same facet distribution and lattice constant as experiment’s high performer of (100) ~ (111) > (110) at a compressed lattice constant of ~ 3.7 �. This trend in stability observed by theory explained in part the durability of this high performer.

Thus, the system-specific investigations for catalysis presented in this dissertation show that I have successfully applied ab-initio methods in conjunction with statistical-mechanical arguments to understand the role of heterogeneity (e.g., cluster isomers, defect sites, facets) in determining catalytic properties. Our in silico predictions of stability and activity have complemented and informed or even prompted the development of novel catalysts from supported clusters to extended surfaces.