Modeling cure shrinkage and viscoelasticity to enhance the numerical methods for predicting delamination in semiconductor packages

Interfacial adhesion between the Epoxy Molding Compound (EMC) and the copper-based leadframe is one of the major concerns in the qualification of plastic packages. Since the conventional shear testing methods used in industry do not consider the residual stresses in the shear samples, they are only used as a qualitative testing method for the EMC qualifications. However, since these tests are based on the maximum force leading to interface delamination, they may cause erroneous results because of neglecting process-induced stresses, which may alter the required force needed to break the samples at the interface. Even classical fracture mechanics, based on mechanical load leading to crack propagation, may not fully characterize the interfacial fracture toughness, because the residual stresses available in the sample impede or facilitate the crack progress, depending on the state of the stresses at the crack tip. The aim of this work is to propose an effective selection criterion for finding the most suitable epoxy molding compound in terms of the intrinsic interfacial adhesion. The effect of residual stresses on the interfacial fracture toughness was investigated by performing an empirical approach to calculate the amount of the cure shrinkage by warpage measurement of the bi-material beams. The effective cure was implemented in the Finite Element Analysis (FEA) of the experimental fracture test to estimate the real interface adhesion. It was observed, that the proper molding compound candidate to fulfill the adhesion requirements was not the one which showed the maximum fracture force and material selection may be done wrongly, if the process-induced stresses are not considered in the FEA of the fracture tests.