Logo image
Comparison of Crosslinking Algorithms in Molecular Dynamics Simulation of Thermosetting Polymers
Journal article

Comparison of Crosslinking Algorithms in Molecular Dynamics Simulation of Thermosetting Polymers

Changwoon Jang, Timothy W Sirk, Jan W Andzelm and Cameron F Abrams
Macromolecular theory and simulations, v 24(3), pp 260-270
May 2015

Abstract

The generation of fully cross‐linked structure is a central task in molecular dynamics simulations of thermosetting polymers. Several algorithms for generating such structures exist, but it is so far not clear what impact the choice of algorithm has on thermal and material properties one typically wants to use simulations to predict. We generated cross‐linked systems comprised of stoichiometric amounts of diglycidyl ether of bisphenol A and poly(oxyproplylene) diamine using two methods: (1) a single‐step approach, in which cross‐link bonds are assigned based on a Monte‐Carlo algorithm that minimizes aggregate bond lengths, and (2) a multi‐step approach, which uses an incrementally increasing capture radius to identify bonding partners. The choice of cross‐linking method has only a minimal impact on thermal and mechanical properties. The minimum nitrogen‐to‐nitrogen contour‐length distributions are also insensitive to the method. However, significant differences were found in the molecular weight distribution of fragments formed by cutting each POP cross‐linker between the amines: the single‐step method results in fewer, larger fragments compared to the multi‐step method. This indicates that the networks formed by the two methods are qualitatively different, and underscores the need for further studies to characterize the influence of polymer network connectivity both in simulations and experiments. MD simulations with ad‐hoc cross‐linking algorithms have been used to generate molecular models of fully cross‐linked epoxy materials. Two such algorithms are compared, and it is shown that glassy‐state thermal and mechanical properties were not significantly influenced by the algorithm choice. This result notwithstanding, it is shown that the two algorithms result in very different network isomers, pointing toward a possible experimental method of structure validation in future work.

Metrics

13 Record Views
60 citations in Scopus

Details

InCites Highlights

Data related to this publication, from InCites Benchmarking & Analytics tool:

Collaboration types
Domestic collaboration
Web of Science research areas
Polymer Science
Logo image