Multiscale modeling of cross-linked epoxy nanocomposites to characterize the effect of particle size on thermal conductivity

Abstract

A sequential multiscale model to characterize the size effects of nanoparticles on the effective thermal conductivity of SiC/epoxy nanocomposites is developed through non-equilibrium molecular dynamics (NEMD) simulations and continuum micromechanics. Even at the fixed volume fraction condition of a spherical nanoparticle, a significant particle size effect on the thermal conductivity of SiC/epoxy nanocomposites has been demonstrated using NEMD simulations. The main contributions of the particle size dependency are Kapitza thermal resistance at the interface between the particle and matrix, and the formation of highly densified polymer sheathing (adsorption layer) near the particle. To account for these two effects in a continuum regime, both the Kapitza interface and the effective interphase are defined in a micromechanics model, and a four-phase multiscale bridging method is suggested. The thermal conductivity of the effective interphase is implicitly obtained from the four-phase micromechanics model. The accuracy and the relative concentration effect of the particle, Kapitza interface, and the effective interphase are discussed via finite element analysis (FEA). By defining the conductivity of the effective interphase as a function of the particle radius, the proposed bridging model accurately reproduced the particle size dependency observed from NEMD simulations. Using the proposed multiscale model, a parametric study is performed to examine the effect of the Kapitza thermal interface and the effective interphase on the overall thermal conductivity of nanocomposites.