A Novel Cure Reaction Model Fitting Technique Based on DSC Scans

Abstract

A numerical methodology has been developed to fit cure kinetic reaction models based on DSC scans only, and not isothermally as traditionally done. A non-linear least squares Levenberg-Marquardt algorithm was used to fit the reaction rates with the autocatalytic Kamal-Sourour kinetic model. The technique avoids the use of isothermal DSC data which causes additional problems in the fitting procedure, such as oscillations in low temperature scans or lost information for fast reaction materials at high process temperature scans. The new methodology is based on a series expansion of the kinetic parameters as a function of temperature. The technique was first tested with the vulcanization process of silicone rubber which is dominated by a cross-linking single-event curing reaction. The method was also tested with phenolic formulations, which according to the hardener, present one or two reaction events. Additionally, the glass transition temperature, of these materials, during the curing process is mostly below the vitrification line, thus, diffusion effects can be neglected. The results show good agreement between the experimental DSC scans and the predicted reaction rates. The technique presented in this paper is intended for use in process simulation, and not to infer kinetic reaction mechanisms.