(218c) Developing Ab Initio Methodology for Advancing Catalytic Reactions

Ab initio investigations have advanced the molecular-scale understanding of catalytic reactions. Current methods allow for predictions of catalytic reaction behavior, although reliability of calculations is a barrier to widespread adoption. The growing string method automates molecular-scale reaction path discovery and is a rapid, unbiased approach. It is enhanced for electrochemical systems which experience electrical potential bias. Although calculations without potential bias treatment are capable to make meaningful predictions for electrochemical systems, a more reliable approach considers the charge of the electrocatalyst surface and solvent effects on adsorbates. This improved growing string method is demonstrated on copper and silver for elementary reactions of electrochemical carbon dioxide reduction. In a second example of ab initio method strengthening, a study of platinum nanoparticles supported by titanium-dioxide catalyzing the water-gas shift reaction is presented. This study incorporates uncertainty quantification and the Bayesian inference from experiments to complement calculations. A microkinetic model connects the atomic-scale calculations with the reactor-scale kinetic experiments. Through the Bayesian inference process, the uncertainties of the ab initio calculations are refined, and a conclusion about the catalytically active site is achieved.