Speaker
Description
A highly intricate interplay of underlying processes governs certain materials properties and functions. This prevents a realistic description by physical models or atomistic simulations. AI can identify nonlinear correlations between materials’ parameters and the measured performance. Thus, AI might better capture the materials’ behavior compared to the theory of the past. However, the data is often inconsistent and the flexibility of AI usually comes together with a lack of interpretability. To address these issues, we combine systematic experiments and simulations with interpretable, data-efficient AI to identify key physical parameters that describe complex materials properties, the “materials genes”.[1] We discuss the concept and recent applications in heterogeneous catalysis.[2,3]
References
[1] L. Foppa, L.M. Ghiringhelli, F. Girgsdies, M. Hashagen, P. Kube, M. Hävecker, S. Carey, A. Tarasov, P. Kraus, F. Rosowski, R. Schlögl, A. Trunschke, and M. Scheffler, Materials genes of heterogeneous catalysis from clean experiments and artificial intelligence, MRS Bulletin 46, 2021; https://doi.org/10.1557/s43577-021-00165-6
[2] G. Bellini, G. Koch, F. Girgsdies, J. Dong, S. J. Carey, O. Timpe, G. Auffermann, M. Scheffler, R. Schlögl, L. Foppa, A. Trunschke. CO Oxidation Catalyzed by Perovskites: The Role of Crystallographic Distortions Highlighted by Systematic Experiments and Artificial Intelligence. Angew. Chem. Int. Ed. 2024 https://doi.org/10.1002/anie.202417812
[3] R. Miyazaki, K. S. Belthle, H. Tüysüz, L. Foppa, M. Scheffler, Materials Genes of CO2 Hydrogenation on Supported Cobalt Catalysts: An Artificial Intelligence Approach Integrating Theoretical and Experimental Data. J. Am. Chem. Soc. 146, 5433, 2024; https://doi.org/10.1021/jacs.3c12984
