Abstract The complex element combinations and synergistic effects of high-entropy alloys (HEAs) present significant challenges for efficient exploration and optimal design. Here, we propose a collaborative framework that integrates large language models (LLMs) with a high-throughput platform, aiming to reveal the correlation between HEA element systems and oxygen reduction reaction (ORR) activity. By domain-specific fine-tuning of the LLM, we developed ChatHEA as an assistant for the enumeration of HEA combinations, which enabled rapid synthesis via a high-throughput platform and facilitated the construction of a standardized ORR activity dataset through batch performance evaluation. Subsequently, ChatHEA performs multidimensional analysis and pattern recognition on the dataset to uncover the intrinsic relationships between HEA element systems and ORR activity, followed by detailed validation of the superior combinations. Density functional theory (DFT) and advanced pH-dependent microkinetic modeling further elucidate and confirm the facilitating effect of elemental synergism on reaction activity. This collaborative framework combining LLMs, high-throughput experimentation, and advanced modeling offers a new pathway for the efficient development of catalytic materials and mechanistic understanding.
Shan et al. (Fri,) studied this question.