High Throughput Screening of Metal-Oxide Systems for Facile OER Kinetics in Electrochemical Mining

Many consumer products and industrial applications require availability of critical elements. Energy technologies are particularly demanding in part because elements that were scarcely used in the fossil-fuel energy economy become vital for renewable energy generation and storage, thanks to discoveries of new material properties. One example is lithium, which has been made critical due to demand for lithium-ion batteries. The need for exponential growth in the deployment and adoption of renewable energy technologies puts immense strain on the supply chains of elements like cobalt, lithium, vanadium, etc. These elements are primarily sourced from mining, which has tremendous environmental impact. While the oxygen evolution reaction (OER) kinetically limits the making of chemicals and fuels and the extraction of precious elements via mining, little is understood about the reaction thermodynamics and kinetics for many metal-oxygen systems.<br/>In this work we have built a tool for high throughput screening of metal oxide systems for facile OER kinetics in electrochemical mining of metals. This tool relies on the concept of the bulk free energy diagram, that elucidates the inherent thermodynamic limitations in M-OER using the formation energies of MxOy bulk intermediates rather than surface binding energies. We have shown in our previous work that bulk free energy diagrams accurately capture the trends in energetics of OER/ORR on metal-oxide surfaces. The formation energies used in this work is both from Materials Project and are supplemented by in-house calculations and machine-learned predictions. Finally, we use a Graph Convolutional Neural Network to predict ORR/OER activity directly from crystal structure, and identify bulk descriptors based on crystal structure and electronic structure for predicting overpotentials.