ABSTRACT
Epoxy-based polymers are widely used in the semiconductor industry as thermal and/or electrical interfaces and as encapsulating material. In the automotive industry, epoxy-based molding compounds (EMCs) are often used to protect not only the single IC packages but also the entire electronic control units (ECUs) (or the power modules). The stress caused by the mismatch of the coefficient of thermal expansion (CTE) between EMC and adjacent materials is one of the major causes for premature failure. In the temperature range of interest, the mold material used in the specific applications exhibits a highly nonlinear behavior. Especially around the glass transition temperature, a dominant nonlinear viscous characteristic of the mold material can be observed. During operation of the aforementioned applications, the epoxy-based polymers are subjected to elevated temperatures around the glass transition temperature, where the material exhibits significant volumetric viscosity in addition to shear viscosity. Traditionally, the nonlinear viscoelastic constitutive models do not consider the volumetric viscoelastic behavior examined in epoxy-based polymers at elevated temperatures. In this contribution, we present the theoretical modeling of these materials. On the theoretical side, we propose a new finite viscoelastic constitutive model, which accounts for the nonlinear volumetric and isochoric viscoelasticity observed in this class of materials. To this end, the polymer network is additively decomposed into a strong elastic network and a superimposed secondary network responsible for rate-dependent response. A nonlinear equation is proposed for the evolution of the viscous deformations in the sense of Dal (2011, Ph.D. thesis, TU Dresden). Regarding an accelerated product development process using commercial software packages, accurate and predictive material modeling is vital. The proposed model is proven to be a very powerful tool for accelerating the development process of electronic control units or power modules by better prediction and evaluation of the limits.
