Cure modelling and monitoring of epoxy/amine resin systems

Abstract

Thermoanalytical techniques and dielectric analysis were used in this study to describe and characterise the cure processes occurring during the isothermal and dynamic cures of four epoxy/amine resin systems. The complexity of the cure reactions was illustrated by results from DSC and FTIR experiments and was attributed to the variety of chemical reactions between the epoxy and the amine groups. Several phenomenological and mechanistic cure kinetics models were constructed, based on the cure reaction mechanisms, in order to simulate the degree of conversion during the cure. A one-to-one relationship was established between the degree of cure and the glass transition temperature of the curing resin, which was finther used in the construction of chemoviscosity models and in a simulation of the viscosity advancement during the cure. A number of mathematical techniques were utilised to evaluate the parameters involved in all the models, varying from simple linear regression methods to complex non-linear least squared estimation procedures. An in-situ dielectric monitoring technique was used in combination with the above mentioned chemorheological models, to investigate the feasibility of a quantitative correlation between the changes in the dielectric signal, the cure advancement and the major physical transformations, namely gelation and vitrification. The imaginary impedance response of the curing resin, as measured by the dielectric technique, showed good agreement with the degree of conversion, depicting all the crucial characteristics of the curing mechanism, such as autocatalysis and diflusion. The endset of the cure reaction was also identified from the endset of the conductivity changes and correlated to the vitrification time. The analytical chemorheological models developed in this study to describe the cure processes for some epoxy/amine resin systems, along with the dielectric monitoring technique used, suggest that a real-time link between the above mentioned models and the cure monitoring technique can be achieved. This would greatly enhance the predictive capability of the technique and form the basis of a future feedback-loop control system.