(629b) Kinetics without Dynamics: Reactive Coarse-Graining with a Real Time Coordinate

Lithium-ion batteries (LIB) are found everywhere from portable electronics to electric cars. The primary components of lithium-ion batteries are the graphite anode, the separator, an organic electrolyte, such as ethylene carbonate (EC) and propylene carbonate (PC), and the transition metal-oxide cathode. Another crucial component of the LIB that has a major impact on its stability and life span is a passivation layer that forms on the anode called the Solid-Electrolyte Interphase (SEI). The SEI imparts stability to the battery and its breakdown is a major cause of catastrophic degradation of the battery. However, there is limited understanding of the molecular-scale events that lead to the catastrophic failure of batteries.

Molecular Dynamics (MD) simulations are widely used to study systems at the molecular scale; however, MD simulations are unsuitable to study the molecular processes leading to battery failure as this is a highly reactive system and classical MD suffers from the inability to account for reactions that take place within the simulated system. Methods that can model reactive systems such as Ab-Initio Molecular Dynamics (AIMD) are prohibitively expensive to reach timescales that are relevant to battery degradation. Therefore, there is a need for computationally inexpensive methods that can be used for modeling the dynamics of reactive systems. One approach to solve this methodological gap is hybrid Monte Carlo/ Molecular Dynamics Methods (MC/MD). This method combines MD and MC, integrating MD simulations for transport processes and MC simulations for chemical reactions, our approach offers a comprehensive framework for modeling reactive systems in a computationally tractable manner.

Here, we report the development of the hybrid Kinetic Monte Carlo/ Molecular Dynamics (KMC/MD) method along with the validation of the method by studying the combustion of alkanes, such as methane. Initially targeting Solid-Electrolyte Interphase (SEI) formation kinetics presents challenges due to diffusion limitation making it a confounding factor for interpreting SEI formation kinetics. The complexity and unique challenges posed by methane combustion render it an ideal non-trivial benchmark for evaluating the reactive MD algorithm's efficacy. By focusing on methane combustion, we aim to validate the KMC/MD algorithm's ability to accurately capture the dynamics of reactive systems characterized by rapid chemical transformations and the generation of diverse molecular species. This preliminary validation step will provide crucial insights into the algorithm's performance and pave the way for its application to more complex systems, including SEI formation in Li-ion batteries.

For the benchmark problem, the reaction network was synthesized using an automated reaction network generation framework to find the possible products for a set of reactants and the transition state (TS) connecting the reactants and products along with the activation barriers using Density Functional Theory (DFT). Once a comprehensive list of reactions and the species involved in these reactions have been generated, force field (FF) parameters are required to capture the transport processes in the MD step. The commonly used FFs use experimental data during the parameterization process. This reliance on experimental data results in limited molecular coverage and impedes the study of reactive systems. To address this problem, a general-use FF parameterized purely on DFT calculations was also developed in this work. The reactive MD simulation itself consists of MD runs of constant time intervals followed by KMC steps to execute reactions. This alternating cycle of MC and MD continues until the maximum simulation time is reached. The hybrid KMC/MD algorithm incqorporates acceleration schemes such as selective downscaling of fast reactions that are in a pseudo-steady state (PSS) to enable the simulation of long timescales.

The resultant simulation outcomes are then compared against ground truth kinetics obtained from microkinetic modeling. Our analysis demonstrates a notable concordance between the kinetics profiles derived from the hybrid MD/MC simulations and those obtained from microkinetic modeling (MKM). This agreement persists across varying temperatures and initial concentration ratios, underscoring the algorithm's efficacy in capturing the dynamics of methane combustion systems. Furthermore, this work sets the initial groundwork for exploring the degradation of more complex reaction chemistries, particularly of the SEI degradation in Li-ion batteries.