基于机器学习预测尖晶石氧化物的析氧催化性能

电解水制氢作为绿色氢能的核心技术之一而备受关注。开发高效的析氧反应（OER）电催化剂是电解水制氢技术的关键。尖晶石型催化剂具有低过电位和长期稳定性,被广泛用于碱性OER阴极材料研究。目前关于尖晶石型OER催化剂的研究通常需要通过实验逐个测试大量的材料,并进行活性评估。这种方法费时费力,且无法涵盖大量的材料组合。本文引入一种基于机器学习的尖晶石型催化剂OER性能预测方法,通过支持向量机（SVM）算法成功拟合了催化剂化学组成（AxByCzO4、载体）、合成方法、焙烧温度、焙烧时间、焙烧气氛、升温速率、形貌、电解质、催化剂载量和工作电极与过电位之间的关系。该模型平均均方误差为182.7,平均绝对误差为20.6,催化剂性能预测效果较好,为高效开发尖晶石型OER催化剂材料提供有效的方法。

Prediction of Oxygen Evolution Activity for Spinel Oxide via Machine Learning

Hydrogen production through water electrolysis has become a prominent technology for green hydrogen energy. The development of efficient electrocatalysts for the oxygen evolution reaction (OER) is crucial for this process. Spinel oxide, known for their low overpotential and long-term stability, have been extensively investigated as cathode materials for alkaline OER. However, it remains a great challenge to search for high performance spinel oxide using trial and error approaches in a reasonable timescale from a large number of possible candidates. To address these challenges, this study proposes a machine learning-based approach for predicting the OER performance of spinel-type catalysts. The support vector machine (SVM) algorithm is utilized to establish a relationship between the catalyst"s chemical composition (AxByCzO4), synthesis method, calcination temperature, calcination time, calcination atmosphere, heating rate, morphology, electrolyte, catalyst loading, working electrode, and overpotential. The model achieved an average mean squared error of 182.7 and an average absolute error of 20.6, indicating its effectiveness in predicting catalyst performance. This approach provides an efficient method for the development of spinel-type OER catalyst materials.