Towards the rational design of polymers using molecular simulation: Predicting the effect of cure schedule on thermo-mechanical properties for a cycloaliphatic amine-cured epoxy resin

• Published in: Reactive & functional polymers 2014-01, Vol.74, p.1-15
• Main Authors: Hamerton, Ian, Tang, Winnie, Anguita, Jose V., Silva, S. Ravi P.
• Format: Article
• Language: English
• Subjects: Applied sciences; Bisphenol A; Chemical properties; Density; Empirical analysis; Epoxy resins; Ethers; Exact sciences and technology; Glass transition temperature; Molecular simulation; Polymer industry, paints, wood; Properties and testing; Property prediction; Simulation; Technology of polymers; Thermal analysis; Thermal degradation
• ISSN: 1381-5148
• DOI: 10.1016/j.reactfunctpolym.2013.10.005

We report prediction of selected physical properties (e.g. glass transition temperature, moduli and thermal degradation temperature) using molecular dynamics simulations for a difunctional epoxy monomer (the diglycidyl ether of bisphenol A) when cured with p-3,3′-dimethylcyclohexylamine to form a dielectric polymer suitable for microelectronic applications. Plots of density versus temperature show decreases in density within the same temperature range as experimental values for the thermal degradation and other thermal events determined using e.g. dynamic mechanical thermal analysis. Empirical characterisation data for a commercial example of the same polymer are presented to validate the network constructed. Extremely close agreement with empirical data is obtained: the simulated value for the glass transition temperature for the 60°C cured epoxy resin (simulated conversion α=0.70; experimentally determined α=0.67 using Raman spectroscopy) is ca. 70–85°C, in line with the experimental temperature range of 60–105°C (peak maximum 85°C). The simulation is also able to mimic the change in processing temperature: the simulated value for the glass transition temperature for the 130°C cured epoxy resin (simulated α=0.81; experimentally determined α=0.73 using Raman and α=0.85 using DSC) is ca. 105–130°C, in line with the experimental temperature range of 110–155°C (peak maximum 128°C). This offers the possibility of optimising the processing parameters in silico to achieve the best final properties, reducing labour- and material-intensive empirical testing.