High-throughput computational search for high performance energy materials

The development of new materials for energy conversion and storage processes is significantly limited by the time it takes to synthesize new materials. Computational techniques can provide insights into a much wider range of materials in a short time-scale, but quantum chemical methods remain too slow to tackle the vast chemical material space. In this project, we are aiming therefore from detailed quantum chemical calculations and kinetic modeling to develop insights into descriptors that accurately depict catalytic activity and selectivity trends across materials. Such descriptors are planned to be learned by high-performance machine learning algorithms, so that they can be quickly estimated for a giant class of materials.

• S. Ringe* et al., The importance of a charge transfer descriptor for screening potential CO₂ reduction electrocatalysts, Nat. Commun. 2023, 14, 2598.