Rational Design for Efficient Bifunctional Oxygen Electrocatalysts by Artificial Intelligence

High-performance bifunctional electrocatalysts that simultaneously and efficiently catalyze oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) have become the bottleneck and main challenge for rechargeable zinc-air batteries. The components need to be comprehensively designed to enhance the formation possibility for the respective activities of ORR and OER. Nevertheless, the elements and types of chemical bonds generate an almost infinite space of potential candidates. In this work, we proposed a rational design strategy for efficient bifunctional oxygen electrocatalysts by a data-driven method. According to the inferred 3,E from E10 and E1/2 machine learning models among all the bond combinations, a bond combination of C-N, C-C, Fe-N, Ru-O, and C-P was considered with superior possibility to have bifunctional activity owing to its highest occurrence frequency. The experimental results further confirmed that this component and bonding method indeed have ORR/OER bifunctional activity, which has not been reported yet. This strategy brings novel and efficient insights for bifunctional electrocatalyst design from a huge potential exploration space.