Epoxy adhesive used in optical fiber/passive component: kinetics, voids and reliability

Abstract

The cure kinetics of a bisphenol epoxy system commonly used in optical fiber assembly and passive component has been analyzed by using a modulated dual scanning calorimeter (MDSC) under isothermal and dynamic conditions. The kinetic data are well represented by a two-parameter autocatalytical reaction rate model. The results suggest that curing of the bisphenol epoxy with imidazole as a curing agent can be fast cured at higher-temperatures. However, the size and numbers of voids increase as the isothermal cure temperature increases indicating that a high temperature snap cure may not provide adequate long term reliability and or fiber pull strength. Based on the thermogravimetric analysis (TGA) results we developed a step cure process for curing the epoxy. On a comparative basis we have found that the site and density of voids at epoxy to plate glass interfaces after a step cure is less than that of after isothermal cures. For the purpose of the verification, fiber assemblies were, prepared by using different curing profiles. Some of these assemblies were cross-sectioned for optical microscopy for estimation of void density before and after temperature cycles, for other fiber assemblies the pull strength before and after temperature cycling was determined. We observed a significant decrease in the void density for the step-cured samples as compared to the isothermal cured samples. The average destructive fiber pull test after temperature cycle testing for the step cured assemblies is higher than the as made isothermal cure assemblies. The results indicate that a proper curing profile can restrict void formation and increase adhesion strength. Finally, we found that a two-parameter Frechet cumulative distribution function (cdf) could be used to represent the statistical behavior of the void distributions in the epoxy used in the fiber assemblies.