Assessment of the accuracy of cure kinetics models and fitting approaches utilised in analysis of microelectronics encapsulation materials

The accuracy of four widely used cure kinetics models in predicting the cure rate of a commercially available encapsulant material is assessed. Four n^th order phenomenological cure kinetics models and the single step autocatalytic model are outlined. These models are fitted to Differential Scanning Calorimetry data using two differing approaches. The Borchardt-Daniels approach is outlined and utilised in conjunction with a Levenberg Marquardt solver to determine model coefficients correlating the models to the experimental data. A particle swarm optimization approach to model fitting is also outlined and is used to develop an alternate set of model coefficients. The accuracy of each of the models combined with each of the fitting methods is defined using an error metric. Optimal model coefficients and related error metric data are presented. The results obtained indicate that the particle swarm optimization approach is able to fit the models more closely to the experimental data, resulting in lower error values than the Borchardt Daniels fitted data. The single step model is also shown to approximate the cure kinetics of the encapsulant material more closely than the n^th order models.