This paper presents a new computational method for simulating polymer network formation. There are four separate procedures involved in the methodology for this multiscale simulation: (i) mapping of the polymerizing monomers onto a coarse-grained model, (ii) cross-linking the monomers by applying Monte Carlo simulation to the coarse-grained model, (iii) reverse mapping of the coarse-grained model to a fully atomistic representation, and (iv) simulation of the atomistic model through standard molecular dynamics technique. Molecular dynamics simulations and experimental studies are carried out to check the algorithm on the basis of the determination of the physical properties of the cycloaliphatic epoxy resin which is prepared from 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate as resin monomers and 4-methylhexahydrophthalic anhydride as curing agents. Depending on the effective conversion and temperature, we determine the density, the glass transition temperature, and the thermal expansion coefficient of the cross-linked epoxy system. An increase in the degree of cross-linking is found to increase the glass transition temperature. Good agreement between computer simulation and experimental results is achieved for highly cross-linked networks, thereby showing that the simulation model is basically valid.

In this paper, an atomistic model for PI/SiO2 hybrid nanocomposites was designed for the investigation of physical properties of this material on the base of molecular dynamics simulations. The thermal properties of a reference pure PI matrix in the temperature range of 300–650 K were first investigated. The results for the CTE and the glass transition temperature showed good agreement with the experimental data. Then, the thermal expansion of the model of PI/silica composite material with different SiO2 fractions was investigated at normal conditions. The CTE of the composite model decreases with the increase in the SiO2 content in agreement with experimental studies. The results show a threshold for the SiO2 loading beyond which the material model exhibits ultra‐low thermal expansion.