Abstract
Developing low-cost acidic oxygen evolution reaction (OER) catalysts is essential to scale proton exchange membrane water electrolysis (PEMWE). Ruthenium-based high-entropy oxides (Ru-HEOs) hold promise, yet the vast compositional space renders trial-and-error approaches inefficient. Here, we propose a data-driven, high-throughput strategy that accelerates the discovery of stable acidic OER Ru-HEO catalysts. First, literature mining with a large language model (LLM) identifies non-noble metal combinations with potential synergy with Ru. We then integrate automated powder synthesis and catalyst array fabrication to build a 60-member library. Screening using dual activity-stability metrics pinpoints the optimal composition, (RuNiFeMoCr)(3)O(4). Subsequent in situ characterization and theoretical calculations show that the high-entropy environment effectively suppresses Ru dissolution and, by optimizing local coordination and electronic structure, synergistically enhances reaction kinetics and structural stability. Last, in PEMWE device tests, an electrolyzer based on this catalyst operates stably for over 150 hour at 1 A cm(−2), validating the data-driven strategy for accelerated screening and proof-of-concept demonstration of Ir-free anode candidates.