摘要
固态锂电池因其具有高能量密度等优势,被认为是最具潜力的下一代电池技术之一。固态锂电池的诸多优势源于其中固态电解质材料的使用。机器学习方法的兴起为锂电池固态电解质研究提供了新的机遇。机器学习可以拓展传统理论模拟方法在固态电池研究中的应用边界,推动高精度、跨尺度模拟方法的发展;预测固态电解质离子导率、力学性质等关键指标,在原子层面理解固态电解质的构效关系,实现高性能固态电解质的高通量筛选;辅助固态电池实验研究,指导新型固态电解质合成,深入定量解析固态电池谱学表征。因此,机器学习方法的引入及其与理论模拟、实验之间的深度耦合将极大地推动固态电解质研究,促进固态锂电池的实用化进程。
Solid-sate lithium battery(SSB)is considered as one of the most promising next-generation batteries due to its high energy density.The emergence of machine-learning(ML)techniques affords a possibility for the study of solid-state electrolytes(SSEs).ML is able to promote a deep application of theoretical simulations in SSB and build a high-accuracy and multi-scale simulation paradigm.Besides,ML can establish a quantitative structure-function relation of SSEs and achieve a high-throughput screening of advanced SSEs.In addition,ML-assisted experiments can synthesize advanced SSEs with a high efficiency and deliver a comprehensive understanding of working mechanism in SSBs with various characterizations such as synchrotron imaging.Therefore,the introduction of ML and its combination with the existing theoretical and experimental methods can promote the study of SSEs and the practical applicationofSSBs definitely.
作者
陈翔
富忠恒
高宇辰
张强
CHEN Xiang;FU Zhong-Heng;GAO Yu-Chen;ZHANG Qiang(Beijing Key Laboratory of Green Chemical Reaction Engineering and Technology,Department of Chemical Engineering,Tsinghua University,Beijing 100084,China)
出处
《硅酸盐学报》
EI
CAS
CSCD
北大核心
2023年第2期488-498,共11页
Journal of The Chinese Ceramic Society
基金
国家重点研发计划(2021YFB2500300)
国家自然科学基金项目(21825501,22109086)
中国博士后科学基金(2021TQ0161,2021M691709)
中国科协青年人才托举工程(2021QNRC001)。
关键词
锂电池
固态电解质
机器学习
离子导率
分子动力学模拟
lithium batteries
solid-state electrolytes
machine learning
ionic conductivity
molecular dynamics simulations
