Reliability prediction in electronic packages using molecular simulation

Abstract

Reliability of electronic packages is a great concern to packaging design engineers. During its design life, packages experience a wide range of temperature variations. The mismatch in coefficient of thermal expansion between the different layers in the packages can generate high interfacial stresses due to these thermal loading. If these stresses exceed the limiting value, delamination will occur. The present study is focused on the reliability of the epoxy molding compound (EMC) and cuprous oxide coated copper substrate. In order to verify whether the interfacial adhesion is dominant by the content of cuprous oxide on the cooper substrate, two models were built to simulate the thermal cycling test with a constant cuprous oxide and a changing content of cuprous oxide on the copper substrate. The thermal cycling test was conducted with a given thermal profile. The adhesion strength between EMC and cuprous oxide coated copper substrate at different thermal cycles was evaluated using the button shear test (BST). A simple molecular model of a bimaterial system, which consists of EMC and cuprous oxide coated copper substrate, was built to evaluate the interfacial energy of the Cu-EMC system. In order to dramatically reduce the computational time, the system was modeled with a limited number of atoms. A preset strain was applied to the model representing a forcing step as the EMC material was pulled away from the copper substrate. Equilibration was conducted to relax the whole system before the next strain step proceeded. The procedure was repeated using different strains. The interfacial energy at different thermal cycles was evaluated. The variation of the interfacial energy indicated the change in the adhesion strength between EMC and cuprous oxide coated copper during the thermal cycling test. The simulation results revealed that the cuprous oxide content in the copper substrate had a large effect on the adhesion between EMC and copper, which is consistent with the experimental observation.