research overview
- Our research focuses on the development and application of computer simulation and integrated machine learning methods, focusing on diverse structures from atoms to micrometers, interpretable property predictions, and guidance in materials selection. Much if the work is at the intersection of AI methods, simulation, chemistry, physics, biology, engineering, as well as exploring applications of quantum computing. We work in teams with other experimental and theoretical groups, explain the function of nanomaterials and biomaterials at the atomic or electronic scale, develop community modeling resources and data science/machine learning tools. Specifically, we develop the Interface force field (IFF) and a surface model database for the simulation of inorganic and organic compounds, solid-electrolyte interfaces, and gases, as well as complex interfaces in one single platform. IFF covers metals, oxides, 2D materials, minerals, polymers, gases, and has compatibility with standard biomolecular force fields. Applications include catalysts for fuel cells and energy conversion, perovskites, battery materials, polymer composites, biominerals, drug delivery, protein function, and building materials. We develop, validate, and deploy methods to improve and expand chemistry-driven and physics-driven models based on fundamental chemical theory, experimentally verified knowledge, and insights into electronic structure via quantum methods. We develop chemistry and physics informed graph neural networks and feature-based machine learning methods, as well as integrated AI agents to solve materials design and property prediction problems orders of magnitude faster than with conventional methods. We have extensive experience working with multinational companies and start-ups who deploy our methods for materials design (Amazon, BASF, P&G, Corning, Sika AG, Goodyear).
