An Open Implementation of MSEVB for Reactive Force Field Simulation with Simple Energy-Based Coupling

Abstract

Chemical reactivity modeling in large systems over extended time scales is computationally and theoretically challenging; ab initio methods are too expensive, while classical force fields cannot handle bond breaking/formation. Existing solutions like reactive force fields require extensive expertise to develop, and machine-learned approaches often need prohibitively large, high-quality data sets for training. These barriers have hindered the modeling of reactive chemical processes. Multistate empirical valence bond (MSEVB) theory offers an alternative by enabling conventional unreactive force fields to be mixed in a controlled manner to selectively describe chemical reactions. However, existing MSEVB implementations are often field-specific, have license-restricted distribution, or rely on complex parametrization schemes. This work presents a new open-source, general MSEVB implementation, designed to make reactive modeling more accessible to nonexperts. Specifically, the implementation here focused on simple energy-based coupling, thereby eliminating the need for complex geometry-based coupling approaches, significantly lowering the barrier to entry for reactive force field development. This generalized approach is demonstrated to reproduce proton transfer behavior in Y-doped and undoped BaZrO3, bulk water and liquid imidazole, as well as other forms of reactivity, such as electron migration in LiCoO2. We demonstrate that only two fitted parameters are needed to accurately model symmetric transfer across the three chemically diverse proton-conducting systems studied.