2024 AIChE Annual Meeting

(569aw) Generalized Bulk Descriptors for Predicting O/OH/Chx Adsorption Energies on Metal Oxides across Varied Coordination Environments

Authors

Winther, K., SLAC National Accelerator Laboratory
Bajdich, M., SLAC STANFORD
In the search for active and stable catalysts, density functional theory (DFT) and machine learning (ML)-based models can accelerate the screening of materials, aiming to derive more accurate and useful outputs from simpler inputs. Recent studies have demonstrated that the O and OH adsorption energies on octahedral oxide surfaces can be predicted by an ML model incorporating the integrated crystal orbital Hamilton population (ICOHP) of bulk and non-DFT features. [ACS Catal. 2024, 14, 7, 5286–5296] To broaden its applicability, we have extended this approach to various crystalline oxides, including tetragonal and square planar structures. Each crystallinity possesses unique bond properties because the primary participating orbitals depend on their distinct crystal field splitting.

Our research interest lies not only in broadening the capabilities of an ML model to predict the adsorption energies of oxygen and carbon-containing intermediates across an expanded range of crystalline oxides, but also in identifying the link between coordination and M-O bonding strength. Establishing a connection between the coordination of metal atoms and M-O bonding strength is essential, as O and OH adsorption can be viewed as a local environmental change, transitioning from lower coordination (e.g., tetragonal, square planar) to higher coordination (e.g., square pyramidal, octahedral). Moreover, resistance to dissolution in aqueous conditions is also correlated with metal coordination; for example, the dissolution of IrO42- from the oxidized IrO2 surface is associated with a structural transformation from octahedral to tetragonal. We hope that this research will serve as a stepping stone to understanding the nature of the M-O bond and to examining the activity and stability of catalysts from more complex oxide systems.