The Tran Group investigates and designs functional materials where microscopic quantum dynamics dictate macroscopic utility in catalysis, clean energy, and information technology. We probe fundamental electron-nuclear dynamics to predict and control energy transfer mechanisms relevant to solar cells and computing architectures. To overcome traditional scaling bottlenecks, we develop computationally tractable, high-fidelity modeling frameworks by uniting the local nature of physical theories with physics-informed machine learning for data-driven materials discovery.
We train and deploy machine learning interatomic potentials (MLIPs), using active learning strategies to minimize the required correlated electronic structure data. Current targets include CO adsorption on metal surfaces and the electronic structure of the nitrogenase active site, where conventional methods are prohibitively expensive.
Charge transport in organic semiconductors is governed by the interplay between electronic coupling and nuclear dynamics. We apply tensor-network methods (TD-DMRG) within an embedding framework to simulate polaron formation and mobility, enabling predictions of device-relevant properties in materials like organic crystals and hybrid perovskites.