Automatic Generation of Chemical Mechanisms for Electrochemical Systems: Solid Electrolyte Interphase Formation in Lithium Batteries

Electrolytes in many lithium ion batteries decompose at the low potentials near the anode. The decomposition products form a layer termed the solid electrolyte interphase (SEI). The composition and growth of the SEI layer significantly affect both the capacity fade and safety of lithium ion batteries. However, SEI formation and growth kinetics are not well understood. In this work, we present an extension of the Reaction Mechanism Generator (RMG) software to automatically generate mechanisms for SEI formation. We extend RMG’s solvation correction framework to account for kinetic solvent effects and demonstrate the accuracy of our technique. We calculate thermochemical parameters for 252 species and rate coefficients for 69 reactions, most with associated solvation corrections. This and additional quantum chemistry data are used to extend RMG’s thermodynamic group additivity and solute parameter estimation schemes to handle lithiated species and add 14 new reaction families to RMG. RMG is additionally extended to simulate electrocatalytic systems. Lastly, we demonstrate RMG on the decomposition of acetonitrile and ethylene carbonate near a battery anode. While this framework does not yet resolve individual ions, as appropriate thermochemistry estimators are not available, and thus, cannot yet resolve more complex electrochemical pathways, RMG is able to generate reasonable pathways for SEI formation that match literature pathways and products. In particular, RMG identifies a new important reaction pathway that is not present in literature.多数锂离子电池中的电解质会在靠近负极的低电位环境下发生分解，其分解产物会形成一层被称为固体电解质界面相（solid electrolyte interphase，SEI）的薄膜。该界面相的组成与生长过程，会对锂离子电池的容量衰减与安全性能造成显著影响。然而，当前学界对于SEI的形成与生长动力学机制尚未形成充分的认知。在本研究中，我们对反应机理生成器（Reaction Mechanism Generator，RMG）软件进行了扩展，以自动生成SEI形成的反应机理。我们拓展了RMG的溶剂化校正框架，以纳入动力学溶剂效应，并验证了所提出方法的准确性。我们为252种物种计算了热化学参数，为69个反应计算了速率系数，其中绝大多数反应均配有对应的溶剂化校正项。结合该数据集与额外的量子化学数据，我们对RMG的热力学基团贡献法以及溶质参数估算方案进行了扩展，使其能够处理锂化物种，并为RMG新增了14个反应家族。此外，我们还对RMG进行了拓展，使其能够模拟电催化体系。最后，我们以电池负极附近的乙腈与碳酸乙烯酯分解反应为例，对扩展后的RMG进行了应用验证。尽管当前该框架尚未能够解析单个离子——这是由于缺乏适配的热化学估算方法——因此也无法处理更为复杂的电化学路径，但RMG仍能够生成与已有文献报道的路径及产物相符的SEI形成合理路径。尤为关键的是，RMG还识别出了一条此前未在文献中被报道的重要新型反应路径。