DFT-ReaxFF Hybrid Reactive Dynamics Method with Application to the Reductive Decomposition Reaction of the TFSI and DOL Electrolyte at a Lithium–Metal Anode Surface

Abstract

The high energy density and suitable operating voltage make rechargeable lithium ion batteries (LIBs) promising candidates to replace such conventional energy storage devices as nonrechargeable batteries. However, the large-scale commercialization of LIBs is impeded significantly by the degradation of the electrolyte, which reacts with the highly reactive lithium metal anode. Future improvement of the battery performance requires a knowledge of the reaction mechanism that is responsible for the degradation and formation of the solid-electrolyte interphase (SEI). In this work, we develop a hybrid computational scheme, Hybridab initiomolecular dynamics combined with reactive force fields, denoted HAIR, to accelerate Quantum Mechanics-based reaction dynamics (QM-MD or AIMD, for ab initio RD) simulations. The HAIR scheme extends the time scale accessible to AIMD by a factor of 10 times through interspersing reactive force field (ReaxFF) simulations between the AIMD parts. This enables simulations of the initial chemical reactions of SEI formation, which may take 1 ns, far too long for AIMD. We apply the HAIR method to the bis(trifluoromethanesulfonyl)imide (TFSI) electrolyte in 1,3-dioxolane (DOL) solvent at the Li metal electrode, demonstrating that HAIR reproduces the initial reactions of the electrolyte (decomposition of TFSI) previously observed in AIMD simulation while also capturing solvent reactions (DOL) that initiate by ring-opening to form such stable products as CO, CH₂O, and C₂H₄, as observed experimentally. These results demonstrate that the HAIR scheme can significantly increase the time scale for reactive MD simulations while retaining the accuracy of AIMD simulations. This enables a full atomistic description of the formation and evolution of SEI.