Thermoset Modelling and Simulation Analysis, with a Focus towards Epoxies.

This work covers a number of aspects central to, and relating to molecular modelling and simulation. Section 1 introduces the diamine, 4,4-diaminodiphenylsulphone (DDS) and the epoxide monomers Bis[4-(glycidyloxy)phenyl]methane (BFDGE) and N,N-Diglycidyl-4-glycidyloxyaniline (TGAP) which are used in this research. The primary and side reactions are explained along with the description and origin of the known impurities. Molecular modelling concepts are explained, covering energy calculations, geometry optimisation and molecular dynamics (MD). The Polymer Consistent Forcefield (PCFF) is used throughout this work, Section 1 includes a comparison of common forecefields, ensembles, thermostats and barostats. Section 2 explains the experimental methodology used for the empirical measurement of glass transition temperatures (Tg) using Differential Scanning Calorimetry (DSC) and Dynamic Mechanical Thermal Analysis (DMTA). Also covered are the more technical and specific matters of this research such as how the impurities are modelled and how the molecular ratios are calculated. There are seven primary models built, based on four formulations. Epoxy A is an epoxy rich mixture of BFDGE, TGAP and DDS, Epoxy B is a stoichiometric mix of BFDGE and DDS, Epoxy C is also stoichiometric, but comprised from TGAP and DDS, Epoxy D is epoxy rich, but only uses TGAP and DDS. The Epoxy F, G and H formulations are the second generation models of Epoxy B, C and D respectively, the models differ from the originals slightly on a number of points. Molecular models were used to predict Tg for comparison with empirical values. Two programs were written as part of this study, the first will process a molecular model, comprised of epoxy and amine monomer, returning a cured resin model, coded to simulate the epoxy and amine bonding reaction. The second is used to processed density vs. temperature data and determine the Tg. The source codes for both programs are available as appendices in Section 6. Materials Studio by Accelrys was used extensively for molecular modelling throughout this work. Section 3 contains the results and discussion from the DSC, DMTA, molecular modelling and lesser experiments. Owing to inter-model similarity, only 4 formulations were subject to thermal analysis. DSC measurements for Tg yielded 215 C for Epoxy A, 161 C for Epoxy B and 229 C for Epoxy C, the DSC analysis of Epoxy D did not establish a value for Tg. The DMTA measurement of Tg gave good agreement for Epoxy A and Epoxy B, about 46 C higher than DSC for Epoxy C and an average of 280 C for Epoxy D. The closest simulations to the empirical values were Epoxy B (+3. 0%), Epoxy C (+4. 3%), Epoxy D (-5. 0%) and Epoxy A (-6. 0%). It was noticed the simulation of stoichiometric formulations predicted lower, and epoxy rich models predicted higher than empirical measurements. The final sections include the summary (Section 4), discussion on possible further work (Section 5) and the appendices (Section 6)