文献类型：期刊文献

英文题名：Machine learning-assisted screening of efficient homogeneous catalysts for the aerobic reaction of aromatic hydrocarbons

作者：Li, Yudong[1];Wang, Xiashi[1];Wang, Shuangfu[1];Zheng, Weizhong[1];Sun, Weizhen[1];Zhao, Ling[1]

机构：[1]East China Univ Sci & Technol, Sch Chem Engn, State Key Lab Chem Engn & Low Carbon Technol, Shanghai 200237, Peoples R China

年份：2026

卷号：546

外文期刊名：CHEMICAL ENGINEERING JOURNAL

收录：;EI(收录号：20263221251590);Scopus(收录号：2-s2.0-105046463862);WOS:【SCI-EXPANDED(收录号:WOS:001845634600001)】；

基金：The financial support by the National Key Research and Develop-ment Project (2022YFC2104504) is gratefully acknowledged.

语种：英文

外文关键词：Machine learning; Catalyst screening; Aerobic reaction; Kinetics modeling

摘要：The ligand modification of the catalyst is a promising strategy to enhance the yields of the aerobic reaction of aromatic hydrocarbons. The approach of machine learning incorporating high-throughput density functional theory (DFT) calculations successfully achieves the efficient screening and design of catalysts. In this work, the impact of ligands on catalyst activity was investigated. The results reveal how ligands influence the catalyst's properties and structure, which subsequently affects its reaction activity. A unique descriptor, phi, which is the ratio of nonpolar and surface area, was identified using ML for predicting yields and efficiently capturing catalytic activity trends for a variety of ligands (R-2 = 0.999, error < 1%). The optimal catalyst, Cat II (cobalt bis(2-ethylhexanoate)), was selected by ML, which exhibits the highest yield (22.81 mol%). The kinetic model achieves great predictive ability for the variation in compound concentration under different conditions. The strategy incorporates ML, experimentation, and reaction kinetics, offering significant insights into catalyst screening, modifications, aerobic reaction, and modeling of aromatic hydrocarbons.