Abstract:The discovery of high-performing catalysts for the oxygen reduction and oxygen evolution reactions (ORR/OER) and related oxidative reactions is critical for the development of energy-efficient hydrogen fuel cells and electrolyzers. To achieve this goal, affordable and abundant alternatives to precious metal catalysts with sufficient stability and catalytic activity must be discovered. Ab-initio catalytic simulations combined with machine learning (ML) methods can play an important role in screening a larger search space of potential electrocatalysts materials of higher structural and compositional complexity, such as class of transition metal oxides (TMOs). In this talk, we will show how electronic and structural descriptors derived from bulk DFT calculations can improve ML models for the OER and ORR on transition metal oxides. In our previous work [1] we demonstrated that the electronic structure of the bulk TMOs, obtained as the crystal orbital Hamiltonian populations (COHP) of the metal-oxygen bond, is an accurate descriptor for O and OH adsorption. Here, we will discuss our extended model for O* and OH* adsorption across crystal structures and oxidation states, utilizing a new dataset of adsorption on binary (AxOy) oxide surfaces spanning the entire transition metal series. Building on our recent work, we extend the COHP-based model to a ML-based prediction of adsorption across multiple oxidation states to obtain a MAE < 0.2 eV for the O-OH and OH adsorption energy descriptors. These results can enable a more efficient screening of catalysts on the bulk level of DFT computation to significantly reduce the computational cost [2]. Furthermore, we will discuss how the COHP analysis can be directly applied to oxide surfaces to understand the impact of metal valency, structural reorganization, and magnetization on adsorption energetics [3]. Our prediction models of stability of bulk phases from COHP and oxidation state beyond binary alloys will also be highlited [4]. Lastly, we will discuss recent progress on collecting and accessing computational and underlying experimental data on catalysis-hub.org. This research was 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. M., Li, J., Abild-Pedersen, F., Bajdich, M., Winther, K. T. (2022). Unraveling Electronic Trends in O∗ and OH∗ Surface Adsorption in the MO2Transition-Metal Oxide Series. , JPCC 2022 126 (18)(2022) [2] Comer, B. M., Bothra, .N, Lunger, J., Abild-Pedersen, F., Bajdich, M., Winther, K. T., Generalized Prediction of O and OH Adsorption on Transition Metal Oxide Surfaces from Bulk Descriptors (under review) [3] Bothra, .N, Comer, B. M., Winther, K. T., Bajdich, M., Understanding the Effects of Surface Bonding in Metal-Oxides Beyond the Limit of the Active Site (in preparation) [4] Craig, M. J., Kleuker, F., Bajdich, M., & Garcia-Melchor, M. (2023). FEFOS: A method to derive oxide formation energies from oxidation states. Catalysis Science & Technology.

Changes in Strength over a 2-Year Period in Professional Rugby Union Players

Invariant Molecular Representations for Heterogeneous Catalysis

Adsorption Enthalpies for Catalysis Modeling through Machine-Learned Descriptors

Explainable Data-driven Modeling of Adsorption Energy in Heterogeneous Catalysis

Predicting binding motifs of complex adsorbates using machine learning with a physics-inspired graph representation

AdsorbML: A Leap in Efficiency for Adsorption Energy Calculations using Generalizable Machine Learning Potentials

Exploring Structure-Sensitive Relations for Small Species Adsorption Using Machine Learning

Machine-learning driven global optimization of surface adsorbate geometries

Electric Dipole Descriptor for Machine Learning Prediction of Catalyst Surface-Molecular Adsorbate Interactions.

Toward Next-Generation Heterogeneous Catalysts: Empowering Surface Reactivity Prediction with Machine Learning

Decoupling Analysis of O2 Adsorption on Metal-N-C Single-Atom Catalysts Via Data-Driven Descriptors.

Machine learning models predict calculation outcomes with the transferability necessary for computational catalysis

Prediction of O and OH Adsorption on Transition Metal Oxide Surfaces from Bulk Descriptors

Machine learning approach for screening alloy surfaces for stability in catalytic reaction conditions

Computational catalyst discovery: Active classification through myopic multiscale sampling

Machine-Learning Adsorption on Binary Alloy Surfaces for Catalyst Screening

Catalyst design using actively learned machine with non-ab initio input features towards CO2 reduction reactions

Adaptive Catalyst Discovery Using Multicriteria Bayesian Optimization with Representation Learning

Development of Activity-Stability Descriptors for Transition Metal Oxide Electrocatalysis

Fine-structure sensitive deep learning framework for predicting catalytic properties with high precision

Adsorb-Agent: Autonomous Identification of Stable Adsorption Configurations via Large Language Model Agent