Experiment-Informed Stochastic Molecular Dynamics Modeling of Hygrothermal Degradation in Epoxy Networks

Epoxy resins exhibit excellent mechanical and thermal properties but suffer from molecular degradation under hygrothermal environments, compromising long-term reliability. Here, we introduce an experiment-informed stochastic molecular dynamics (MD) framework for directly modeling hygrothermal degradation in cross-linked epoxy networks without reactive force fields. Starting from an atomistically cured epoxy structure, C-O bond cleavage is introduced stochastically, with the local cleavage probability enhanced by nearby water molecules to account for their catalytic effect. This approach naturally reproduces the sequential release of singly reacted epoxy units and bisphenol A from doubly reacted units. The simulated evolution of low-molecular-weight species and the accompanying decrease in glass transition temperature agree well with experimental observations. This framework provides a computationally efficient and physically transparent route for simulating long-time scale degradation in epoxy resins.