arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.05141cond-mat.mtrl-sci
收藏

预测性机器学习分子动力学研究SEI在高浓度LiTFSI和LiPF6电解质中的形成用于锂金属电池

Solvation Restructuring Accelerates Early SEI Nucleation in Lithium Metal Batteries

University of Illinois Chicago(伊利诺伊大学芝加哥分校)

Syed Mustafa Shah, Mohammed Lemaalem, Anh T. Ngo

AI总结:

本文提出基于深度势的机器学习分子动力学框架,用于研究高浓度LiTFSI和LiPF6电解质中SEI的形成机制,揭示其对锂金属电池循环稳定性和倍率性能的影响。

摘要:

高能量密度的锂金属电池需要能够实现快速离子传输并形成稳定固电解质界面(SEI)的电解质,以维持高倍率循环,这一过程仍难以通过实验捕捉。在这里,我们开发了一种基于深度势的机器学习分子动力学(MLMD)框架,该框架在大量的从头计算数据集上进行训练,并通过实验传输性质进行验证,以量子精度解析锂金属界面的早期SEI成核。我们发现,在锂金属界面处,3.5 M LiTFSI/DMC诱导自发的热激活还原反应,产生快速生长的厚阴离子衍生SEI,富含O/F含物种。相比之下,1.5-2.5 M LiTFSI/DMC和1 M LiPF6/EMC/DMC/EC形成更薄的LiF主导界面,生长动力学较慢。我们的建模结果与实验观测一致,其中3.5 M LiTFSI增强了循环稳定性和倍率能力,而较低浓度导致保护能力较弱。我们的MLMD框架高效捕捉了锂金属电池中电解质传输和早期SEI形成机制。

英文摘要:

The development of high-energy-density lithium metal batteries is limited by electrolyte instability and the poorly understood onset of solid-electrolyte interphase (SEI) formation. Here, we use AIMD-trained Deep Potential molecular dynamics to link electrolyte solvation structure to early SEI nucleation in \ce{LiTFSI}/DMC electrolytes. Increasing \ce{LiTFSI} concentration drives the electrolyte from solvent-separated ion pairs toward contact ion pairs and aggregates, with \SI{3.5}{M} showing the strongest anion coordination. This anion-rich environment promotes earlier interfacial Li--F/Li--O bond formation, faster consumption of intact \ce{TFSI-} and solvent, and a denser, LiF/Li$_2$O-rich nascent interphase, in contrast to the more organic-laden, phosphorus/fluorine-based interphase formed by a \SI{1}{M} \ce{LiPF6} reference electrolyte. These results show that bulk solvation architecture biases both the timing and chemistry of early SEI formation and suggest solvation control as a design handle for Li-metal electrolytes.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑