欢迎访问《人工晶体学报》官方网站,今天是 分享到:

人工晶体学报 ›› 2024, Vol. 53 ›› Issue (9): 1475-1493.

• 特邀综述 •    下一篇

数据驱动的新型非线性光学材料理论设计研究进展

储冬冬, 杨志华, 潘世烈   

  1. 中国科学院新疆理化技术研究所,晶体材料研究中心,乌鲁木齐 830011
  • 收稿日期:2024-06-03 出版日期:2024-09-15 发布日期:2024-09-19
  • 通信作者: 杨志华,博士,研究员。E-mail:zhyang@ms.xjb.ac.cn; 潘世烈,博士,研究员。E-mail:slpan@ms.xjb.ac.cn
  • 作者简介:储冬冬(1999—),女,安徽省人,博士研究生。E-mail:chudongdong20@mails.ucas.ac.cn
  • 基金资助:
    国家重点研发计划(2021YFB3601502);国家自然科学基金(22193044,61835014,51972336)

Research Progress on Theoretical Design of Nonlinear Optical Materials via Data-Driven Approach

CHU Dongdong, YANG Zhihua, PAN Shilie   

  1. Research Center for Crystal Materials, Xinjiang Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Urumqi 830011, China
  • Received:2024-06-03 Online:2024-09-15 Published:2024-09-19

摘要: 非线性光学晶体是全固态激光器的核心器件,在信息技术、国防安全等方面具有广泛且重要的应用。随着高性能计算技术的发展,计算辅助实验的“自上而下”靶向设计逐渐成为非线性光学材料设计的重要组成部分。同时,基于高通量计算获取的大规模结构性能信息,进一步为数据挖掘、机器学习的算法训练提供了坚实的数据基础,促进了材料设计的第四范式的发展。本文从非线性光学晶体的计算材料设计入手,探讨了数据驱动的非线性光学理论设计的新范式,对本团队近年来在高通量筛选、晶体结构预测以及机器学习加速非线性光学材料设计等方向的研究进展进行了综述。

关键词: 晶体结构预测, 高通量筛选, 机器学习, 非线性光学材料, 材料设计

Abstract: Nonlinear optical crystals are the core devices of all-solid-state lasers, and have extensive and important applications in information technology and national security. With the development of high-performance computing, the “top-down” computer-aided design methods have gradually become an important part of nonlinear optical materials design. In addition, the large-scale structural and properties information obtained based on high-throughput computing provides a solid data foundation for data mining and machine learning algorithm training, accelerating the development of the fourth paradigm of material design. This paper starts with the computational design of nonlinear optical materials, and then, discusses the new paradigm of data-driven nonlinear optical materials theoretical design. Finally, the recent research progress of our team in high-throughput screening, crystal structure prediction, and machine learning accelerated nonlinear optical materials are reviewed.

Key words: crystal structure prediction, high-throughput screening, machine learning, nonlinear optical material, material design

中图分类号: