Abstract
Atomistic modeling has become an extensively used method for studying thermosetting polymers, particularly in the analysis and development of high-performance composite materials. Despite extensive research on the topic, a widely accepted, standardized, flexible, and open-source approach for simulating the cross-linking process from precursor molecules has yet to be established. This study proposes, tests, and validates a Molecular Dynamics (MD) protocol to simulate the cross-linking process of epoxy resins. We developed an in-house code based on Python and LAMMPS, enabling the generation of epoxy resin structures with high degrees of cross-linking. In our work, the epoxy network is dynamically formed within the MD simulations, modeling the chemical bonding process with constraints based on the distance between the reactive sites. To validate our model against experimental data from the literature, we then computed the density, thermal conductivity, and elastic response. The results show that the produced structures align well with experimental evidence, validating our method and confirming its feasibility for further analyses and in silico experiments. Beyond the case study presented in this work, focusing on bisphenol A diglycidyl ether (DGEBA) epoxy resin and diethylenetriamine (DETA) as curing agents in a 5:2 ratio, our approach can be easily adapted to investigate different epoxy resins.