A foundational framework for the mesoscale modeling of dynamic elastomers and gels

TL;DR

This paper introduces a simplified mesoscale modeling framework for dynamic elastomers and gels that accurately predicts mechanical behavior while significantly reducing computational costs, enabling larger and longer simulations.

Contribution

It presents a reduced-order mesoscale approach that captures key molecular and mechanical features of polymer networks with 90% less computational effort.

Findings

Mesoscale model reasonably reproduces molecular predictions for short chains.

Model achieves 90% reduction in computational cost.

Enables simulation of larger deformations and longer timescales.

Abstract

Discrete mesoscale network models, in which explicitly modeled polymer chains are replaced by implicit pairwise potentials, are capable of predicting the macroscale mechanical response of polymeric materials such as elastomers and gels, while offering greater insight into microstructural phenomena than constitutive theory or macroscale experiments alone. However, whether such mesoscale models accurately represent the molecular structures of polymer networks requires investigation during their development, particularly in the case of dynamic polymers that restructure in time. We here introduce and compare the topological and mechanical predictions of an idealized, reduced-order mesoscale approach in which only tethered dynamic bonding sites and crosslinks in a polymer's backbone are explicitly modeled, to those of molecular theory and a Kremer-Grest, coarse-grained molecular dynamics approach. We find that for short chain networks at intermediate polymer packing fractions, undergoing relatively slow loading rates, the mesoscale approach reasonably reproduces the chain conformations, bond kinetic rates, and ensemble stress responses predicted by molecular theory and the bead-spring model. Further, it does so with a 90% reduction in computational cost. These savings grant the mesoscale model access to larger spatiotemporal domains than conventional molecular dynamics, enabling simulation of large deformations as well as durations approaching experimental timescales (e.g., those utilized in DMA). While the model investigated is for monodisperse polymer networks in theta-solvent, without entanglement, charge interactions, long-range dynamic bond interactions, or other confounding physical effects, this work highlights the utility of these models and lays a foundational groundwork for the incorporation of such phenomena moving forward.