Autonomous, Cloud-Based Machine-Learning-Guided Optical Evaluation of Durability of Oxygen Evolution Reaction (OER) Electrocatalysts

The durability of OER catalysts in the (Ni-Co-Sb)O_x composition system¹ is investigated via optical transmission spectroscopy, as a function of electrolyte pH, electrochemical potential, film composition, and film thickness. Time-dependent changes in optical spectra of these previously investigated OER catalysts are a convolution of electrochromic and corrosion processes, which depend upon the investigated parameters. For a given electrolyte and sub-composition space, a machine learning algorithm guides the selection of composition and the sequence of electrochemical experiments to efficiently optimize our ability to predict stability windows for OER conditions. This work represents an initial demonstration of integration of our labs automated experimental electrochemical characterization instruments utilizing the Hierarchical Experimental Laboratory Automation and Orchestration with asynchronous programming (HELAO-async),² and deployment of cloud-based materials data storage and analysis.³
1. “Surveying Metal Antimonate Photoanodes for Solar Fuel Generation” Lan Zhou, Yu Wang, Kevin Kan, Daphne M. Lucana, Dan Guevarra, Yungchieh Lai, John M. Gregoire ACS Sustainable Chem. Eng. 2022, 10, 48, 15898–15908.
2. “Orchestrating nimble experiments across interconnected labs” Dan Guevarra, Kevin Kan, Yungchieh Lai, Ryan J. R. Jones, Lan Zhou, Phillip Donnelly, Matthias Richter, Helge S. Stein, and John M. Gregoire Digital Discovery, 2023, 2, 1806-1812.
3. “Event-driven data management with cloud computing for extensible materials acceleration platforms” Michael J. Statt, Brian A. Rohr, Dan Guevarra, Santosh K. Suram, and John M. Gregoire Digital Discovery, 2024, 3, 238-242.