Authors: Fangyuan Zheng, Huanxin Xiang, Liangjie Zhong, Xueling Wang, Xiaohan Zhang, Qingwen Su, Chunmei Tang, Ling Meng, Lei Du, Feng Jiao, Yoshitaka Aoki, Baoyin Yuan, Ning Wang, Siyu Ye
Published: 2025-05-30
Source: Full article
AbstractProtonic solid oxide cell (P‐SOC) is a novel type of solid oxide cell for hydrogen production and power generation. P‐SOCs have garnered significant attention due to their advantages, such as the elimination of precious metals and high conversion efficiency. However, the commercialization of P‐SOCs is currently hindered by suboptimal electrochemical performance, particularly at the air electrode side, where challenges in catalytic activity and ionic/electronic conductivity persist. Recently, strategies for designing advanced triple‐conducting oxides, exsolution, and optimizing the air electrode–electrolyte interfaces have been proposed to improve the electrochemical reactive area, kinetics, and durability of air electrodes. Thereinto, machine learning (ML) techniques have emerged as powerful tools, playing a crucial role in the above topics. Despite these advancements, a comprehensive review synthesizing these innovative strategies and ML‐guided advances and perspectives has been lacking in literature. This review comprehensively makes a summary of these methods and discusses their effects on cell performance. Importantly, the ML‐guided perspectives and challenges in accelerating the optimization of these strategies and P‐SOCs are proposed here. This paper not only offers valuable insights for understanding and optimizing performances at the air electrode side but also provides a roadmap for the rational design of superior air electrodes of P‐SOCs.