(426f) An Electronic Descriptor Based Machine Learned Model for Adsorption Energies on Oxide Materials
AIChE Annual Meeting
2022
2022 Annual Meeting
Catalysis and Reaction Engineering Division
Data Science & Machine Learning Approaches to Catalysis I: Interpretable and Theory-Guided Machine Learning For Catalysis Design and Understanding
Wednesday, November 16, 2022 - 9:48am to 10:06am
Constructing economically viable fuel cells requires discovering novel electrode materials that efficiently catalyze the OER and ORR reactions. The key to this is understanding and modeling the adsorption of O* and OH* on the relevant catalyst surfaces. Our previous work[1] showed that the bulk integrated COHP provides an excellent descriptor for both O* and OH* adsorption in rutile oxide materials. In this work, we further investigate these quantities in oxide materials seeking to model the adsorption behavior of O* and OH* across multiple materials and oxidation states utilizing machine learning models. To achieve this, we generate a consistent set of adsorption energies for pure oxides composed of the complete transition metal series (3d, 4d, and 5d) across three oxidation states (+3, +4, and +5) at the PBE+U DFT level. With this data and utilizing a modified graph-based approach, we construct a model for predicting the adsorption of O* and OH*. We find that using the integrated COHP of metal-oxygen bonds in the corresponding pure bulk materials improves accuracy over the base methods. Furthermore, this analysis is robust to the complicated spin effects seen in 3d metals, the low spin 4d and 5d metals, and the differing electronic structures of the 3+,4+, and 5+ oxidation states. Further, this work suggests that electronic features extracted from bulk materials may provide a path forward toward building more generalized adsorption models for oxides.
This research is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division, Catalysis Science Program to the SUNCAT Center for Interface Science and Catalysis.
[1] Comer, B., Li, J., Abild-Pedersen, F., Bajdich, M., Winther, K., (2022). Unraveling Electronic Trends in O* and OH* Surface Adsorption in the MO2 Transition-metal Oxide Series. The Journal of Physical Chemistry C. (Accepted)