Talk as part of a colloquium in Data-Driven Materials discovery.

Summary: Atomistic dynamics provide a direct route to understanding how microscopic processes govern the properties and performance of energy materials. In this talk, I will present studies that combine advanced atomistic simulations with machine learning and high-throughput approaches. I will first discuss hydrogen transport in layered materials, identifying the microscopic nature of the mobile hydrogen species, and quantum tunneling in Stone–Wales graphene. I will then show how machine-learning potentials enable long-timescale simulations of all-solid-state Li batteries, providing insights into Li dendrite growth at solid-state interfaces. Finally, I will show how high-throughput screening and data-driven analysis can identify design principles for NASICON-derived intermediate-temperature proton conductors. These studies demonstrate how physics based simulations, machine learning, and high-throughput approaches can bridge atomistic mechanisms and predictive materials discovery.