文献类型：期刊文献

英文题名：Active-learning-based rational inverse design of ionic liquids for interfacial property optimization

作者：Zhao, Shuheng[1];Tian, Yifan[1];Ma, Zhihong[2];Zheng, Weizhong[2];Lu, Jingyi[1];Sun, Weizhen[2];Du, Wenli[1]

机构：[1]East China Univ Sci & Technol, Key Lab Smart Mfg Energy Chem Proc, Minist Educ, Shanghai, Peoples R China;[2]East China Univ Sci & Technol, Sch Chem Engn, Shanghai, Peoples R China

年份：2026

外文期刊名：AICHE JOURNAL

收录：;EI(收录号：20260920195274);WOS:【SCI-EXPANDED(收录号:WOS:001702696000001)】；

基金：Key Project of Science and Technology Innovation 2030, Grant/Award Number: 2023ZD0121001; Shanghai Committee of Science and Technology, Grant/Award Number: 23ZR1416000

语种：英文

外文关键词：active learning; interfacial performance prediction; ionic liquid; molecular motif control; rational inverse molecular design

摘要：Ionic liquids (ILs) are thermally stable, nonvolatile salts widely used in catalysis and separations, yet their rational design is challenged by the vast cation-anion combinatorial space and strongly coupled interfacial properties. In this work, we propose an inverse design framework that combines data augmentation, dual-surrogate optimization, and active learning to enable efficient exploration of IL chemical space. A DFT-guided augmentation strategy is employed to expand the limited training data, while a dual-surrogate scheme guides optimization by jointly predicting interfacial properties and structural validity. To maintain reliability in unexplored regions, an active-learning loop is introduced to continuously refine the validity classifier. Furthermore, substructure constraints derived from interpretability analysis are incorporated to allow motif-controlled generation. Computational results show that the proposed framework achieves markedly improved optimization efficiency while consistently maintaining high structural validity, providing a practical route for the rational design of ILs with tailored interfacial properties.