Systematic Coarse-graining of Epoxy Resins with Machine Learning-Informed Energy Renormalization

Area of Science:

• Materials Science
• Computational Chemistry
• Polymer Science

Background:

• Predictive molecular modeling of thermoset polymers faces challenges in efficiently capturing the impact of chemical composition and degree of crosslinking (DC) on material properties.
• Accurate simulation of dynamical and mechanical properties requires computationally intensive methods, limiting large-scale investigations.

Purpose of the Study:

• To develop a computationally efficient coarse-graining (CG) approach for predictive molecular modeling of thermoset polymers.
• To accurately predict the effects of chemical composition and degree of crosslinking (DC) on polymer properties.

Main Methods:

• Established a new CG approach combining energy renormalization with Gaussian process surrogate models for molecular dynamics simulations.
• Developed a machine-learning informed functional calibration of DC-dependent CG force field parameters.
• Applied the framework to epoxy resins (Bisphenol A diglycidyl ether with specific curing agents) and compared predictions with all-atom simulations.

Main Results:

• Demonstrated excellent agreement between all-atom and CG predictions for density, Debye-Waller factor, Young's modulus, and yield stress across various degrees of crosslinking (DC).
• Successfully simplified functional forms of 14 non-bonded calibration parameters using surrogate models and uncertainty quantification.
• Validated the CG approach for versatile epoxy resin systems.

Conclusions:

• The established framework provides an efficient methodology for chemistry-specific, large-scale investigations of epoxy resin dynamics and mechanics.
• The ML-informed CG approach significantly enhances computational efficiency while maintaining high prediction accuracy for thermoset polymers.
• This method enables more accessible and scalable exploration of structure-property relationships in complex polymer systems.