Development of Experimental Tools and Data Analysis Pipelines for Autonomous Electrochemistry

Many of the world’s most pressing environmental challenges such as energy conversion, energy storage, and remediating environmental degradation rely on the rapid discovery of new and extraordinary materials. As outlined in the 2016 Mission Innovation Materials Acceleration Platform report, the development of self-driving labs which can be driven by artificial intelligence will be crucial to finding these materials quickly. In this talk, I will present recent work from the Alliance for AI-Accelerated Materials Discovery and the Acceleration Consortium on building a self-driving lab for electrochemistry. Using the oxygen evolution reaction (OER) as an example, I will cover three key aspects of developing reliable autonomous workflows. First, I will discuss the design (and re-design) and validation of the electrochemical characterization platform for investigating the performance and degradation of OER catalysts. Next, I will explain how this platform, along with the development of a mixed acceleration AI model ensemble, has been used to identify new catalysts that advance the performance-durability pareto frontier. Finally, I will discuss the remediation of bias in spectroscopic data analysis, specifically electrochemical impedance spectroscopy, and how reliable physics-based tools can be applied to track material degradation at the level of transport processes.