Closed-Loop Framework for Discovering Stable and Low-Cost Bifunctional Metal Oxide Catalysts for Efficient...

Electrocatalytic water splitting, comprising the oxygen evolution reaction (OER) and hydrogen evolution reaction (HER), provides a sustainable route for hydrogen production. While low-cost metal oxides (MOs) are appealing as alternatives to noble metal electrocatalysts, their application in acidic media remains challenging. However, the dynamic nature of some MO surface structures under electrochemical conditions offers an opportunity for rational catalyst design to achieve bifunctionality in acidic OER and HER. Here, we present a closed-loop framework that integrates potential catalyst exploration (front-end), synthesis and electrochemical tests (mid-end), and advanced characterizations (back-end). This framework combines crucial steps in electrocatalyst exploration, including data mining, surface state analysis, microkinetic modeling, and proof-of-concept experiments to identify stable and cost-effective MO catalysts for acidic water splitting. Using this approach, RbSbWO₆ is identified as a promising bifunctional catalyst for the first time, with experimental validation demonstrating its exceptional stability and performance under acidic OER and HER. Notably, RbSbWO₆ outperforms many other reported non-noble stoichiometric MO catalysts that have not undergone major modifications for acidic water splitting. These findings, derived from our Digital Catalysis Platform (DigCat), establish RbSbWO₆ as a highly effective non-noble stoichiometric bifunctional MO catalyst and underscore the power of our closed-loop workflow for accelerating catalyst discovery. This framework begins with the DigCat platform, concludes with experimental validation, and feeds into the platform, demonstrating its potential for designing electrocatalysts in other systems such as metal nitrides or carbides. This study demonstrates the importance and high efficiency of data-driven approaches as a new scientific discovery paradigm.