Systematic coarse-graining of epoxy resins with machine learning-informed energy renormalization

This paper presents a coarse-graining (CG) approach for modeling epoxy resins, combining energy renormalization with Gaussian process surrogate models. The approach enables machine-learning informed calibration of degree-of-crosslinking (DC)-dependent CG force field parameters. The authors demonstrate excellent agreement between all-atom and CG predictions for density, Debye-Waller factor, Young’s modulus, and yield stress across various DCs for two epoxy resin systems. A surrogate model-enabled simplification of the functional forms of parameters is also introduced, improving efficiency for large-scale investigations of epoxy resin dynamics and mechanics.

Andrea Giuntoli, Nitin K. Hansoge, Anton van Beek, Zhaoxu Meng, Wei Chen, Sinan Keten