化工进展 ›› 2024, Vol. 43 ›› Issue (4): 1649-1654.DOI: 10.16085/j.issn.1000-6613.2024-0175

• 观点 •    

人工智能驱动新型界面分离材料的创制

何林1,2(), 贺常晴1,2, 隋红1,2   

  1. 1.天津大学化工学院,天津 300072
    2.精馏技术国家工程研究中心,天津 300072
  • 收稿日期:2024-02-07 修回日期:2024-02-15 出版日期:2024-04-15 发布日期:2024-05-13
  • 通讯作者: 何林
  • 作者简介:何林(1987—),男,博士,副教授,博士生导师,研究方向为资源与环境多介质体系界面分离与转化过程。E-mail:linhe@tju.edu.cn
  • 基金资助:
    国家自然科学基金(22078237)

Prospects for the creation of novel interfacial separation materials driven by artificial intelligence

HE Lin1,2(), HE Changqing1,2, SUI Hong1,2   

  1. 1.School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China
    2.National Engineering Research Center of Distillation Technology, Tianjin 300072, China
  • Received:2024-02-07 Revised:2024-02-15 Online:2024-04-15 Published:2024-05-13
  • Contact: HE Lin

摘要:

计算机技术与人工智能的发展给新材料的开发带来了新的研究范式(第四代),也为新型界面分离材料的创制带来新的发展契机。然而,受界面分离过程机理研究和材料、基团的理化数据的局限性,人工智能在新型界面分离材料的创制中的应用仍处于初期阶段。本文为解决特定体系数据缺失的问题,构建了百万级(106)的样板数据库,为大预测模型与生成模型提供了可靠、规范的示范数据基础;通过从大数据中挖掘分子结构与性质的隐藏关联,并借助分子模拟计算剖析其分子间相互作用机理,解析分子间相互作用机制和重组过程规律,进而实现高性能界面分离材料的构筑。最后,本文浅析了介尺度样板大数据、机器学习势理论、国产软件开发、有机合成、智能算法等多学科交叉在界面分离材料创制中的重要影响与发展机遇,以期推动虚拟实验室前沿发展。

关键词: 界面, 大数据, 人工智能, 分离材料, 分子间相互作用

Abstract:

The advancement of computer technology and artificial intelligence has presented a novel research paradigm, known as the "fourth generation". It has revolutionized the field of material development. Moreover, it has generated fresh prospects for the fabrication of interfacial separation materials. However, artificial intelligence is still in its infancy in this field. This is mainly ascribed to the constraints imposed by the interfacial separation process mechanism and the physical and chemical characteristics of materials and groups. This work developed a million-level (106) model database to address the issue of data scarcity in particular systems. The database furnished a substantial, dependable and standardized data foundation for generative models and large prediction models. Through extracting latent correlations between molecular properties and structures from large datasets and applying molecular simulation calculations to study the mechanism of inter-molecular interactions, the aim was to examine the governing laws of the recombination process. This enabled the development of interfacial separation materials with exceptional performance. Finally, this work concluded with a concise overview of the significant impact and prospects for growth that arose from interdisciplinary approaches, including meso-scale big data, machine learning potential theory, domestic software development, organic synthesis and intelligent algorithms. The intention was to facilitate the advancement of virtual laboratories to the forefront of innovation.

Key words: interface, big data, artificial intelligence, separation materials, inter-molecular interactions

中图分类号: 

京ICP备12046843号-2;京公网安备 11010102001994号
版权所有 © 《化工进展》编辑部
地址:北京市东城区青年湖南街13号 邮编:100011
电子信箱:hgjz@cip.com.cn
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn