已核实 DOI 相关产品: 10

Synthesis of highly active binary alloy acidic oxygen evolution catalysts assisted by machine learning

Journal 2026 · IF 12.5 · 发表于 2026 年 7 月 8 日
作者: Xinyu Wang,Chaowen Tan,Kechen Zhou,Xuewei Lv,Jie Dang
阅读全文 在 PubMed 查看

摘要

Constructed a closed-loop system integrating data, algorithms and experiments. • The catalyst exhibits superior bifunctional catalytic activity. • The FeRu alloy maintains stable operation for 500 h in PEMWE at 1 A cm −2 . • Mechanistic insights reveal the intrinsic stability origin of FeRu alloy catalyst. Proton exchange membrane water electrolysis (PEMWE) is promising for green hydrogen production, but its oxygen evolution reaction (OER) suffers from high overpotential and catalyst corrosion. While transition metal alloys exhibit excellent catalytic activity, traditional experiments and density functional theory (DFT) calculations are inefficient for exploring the vast binary alloy space. Machine learning enables rapid prediction and optimization by mining DFT data to map material properties with catalytic performance. Herein, this study integrates machine learning methods, including ABR, GBR, and other ensemble models, with DFT calculations to rapidly screen of acidic OER catalysts based on binary transition metal systems. Based on model predictions, we utilize electrodeposition and Joule heating techniques to synthesize FeRu alloy in a short period of time and perform experimental validation. The as-synthesized FeRu alloy exhibits exceptional bifunctional activity, requiring ultralow overpotentials of 171 mV (OER) and 47 mV (hydrogen evolution reaction) to achieve 10 mA cm −2 in acidic media. This catalyst reaches a current density of 1 A cm −2 at only 1.65 V in PEMWE. DFT calculations confirm that alloying optimizes electronic structure and reduces the rate-determining step (RDS) energy barrier, providing theoretical support for its excellent activity and stability. This work establishes a closed-loop framework that integrates data mining with experimental validation to shorten catalyst research and development cycles. It provides an efficient approach for screening and optimizing high-performance acidic PEMWE catalysts and facilitates the industrialization of green hydrogen production via PEMWE. The workflow integrates DFT, feature engineering, and model training to establish quantitative structure-activity-stability relationships. This data-driven framework drastically reduces trial-and-error costs, enabling the identification of FeRu alloys with exceptional acidic water electrolysis performance, and provides a scalable strategy for the rational design of advanced, low-cost electrocatalysts for industrial application. Download: Download high-res image (115KB) Download: Download full-size image

本研究中使用的阿拉丁产品

10 款产品在本文中被引用