Machine‐Learning Assisted Screening Proton Conducting Co/Fe based Oxide for the Air Electrode of Protonic Solid Oxide Cell

Abstract Proton‐conducting solid oxide cells (P‐SOCs) as energy conversion devices for power generation and hydrogen production have attracted increasing attention recently. The lack of efficient proton‐conducting air electrodes is a huge obstacle to developing high‐performance P‐SOCs. The currently widely used air electrode material is Co/Fe based perovskite oxide, however, there is still no systematic research on studying and comparing the roles of diversiform elements at the B site for Co/Fe based perovskite oxide. Here, a machine learning (ML) model with eXtreme Gradient Boosting (XGBoost) algorithm is built to quickly and accurately predict the proton absorption amount of Co/Fe based perovskite oxides with 27 elements dopant at B site. Hereafter, La(Co_0.9Ni_0.1)O₃ (LCN91) is screened by a combination of the ML model and the density functional theory calculation. Finally, LCN91 is applied to the air electrode of P‐SOC, and the cell exhibits excellent electrochemical performances in fuel cell and electrolysis modes. The current study not only provides a useful model for screening air electrodes of P‐SOC, but also extends the application of ML in exploring the key materials for P‐SOCs and other fuel cells/electrolyzers.