Research

The ephemeral nature of active sites and the dynamic evolution of catalyst morphology are relatively well-known aspects of heterogeneous catalysis. Despite this, computational approaches have favored a static scheme in describing catalysts, often leading to inconsolable discrepancies with experimental measurements. We address this gap by leveraging advances in constructing deep neural network potentials and enhanced sampling methods to capture the spatial and temporal evolution of complex catalytic systems and their influence on the catalyzed reaction. We aim to predict dynamic changes in catalysts that could be favorable or detrimental to the chemical transformation.

The electronic state of the catalyst principally dictates the "enthalpic" component of the reaction free-energy. We develop chemisorption models rooted in physics to help identify electronic structure factors that influence the interaction between the adsorbates and the catalyst surface. Combining electronic structure methods and concepts from solid-state physics with deep learning accelerated molecular dynamics and enhanced sampling methods, our goal is to go beyond the pristine case and account for the influence of the dynamic changes to the catalyst morphology on its electronic structure.

Mineral-water interfaces play a crucial role in geochemical processes ranging from the deep carbon cycle to aiding the origins of life. We use a combination of different computational methods and thermodynamic models to understand the behavior of aqueous mixtures and mineral oxide-water interfaces, including the effects of confinement that are unique to geochemical reactions. We aim to address existential problems such as carbon capture and storage and explore the possibilities of becoming a multi-planetary species by probing geochemical processes in other celestial objects.