Stochastic multiscale modeling of random interphases in nanoreinforced polymers
Résumé
This presentation is concerned with the multiscale analysis of nanoreinforced heterogeneous materials. Because of the so-called surface effect, the latter typically exhibit a strong reinforcement effect at low volume fractions. Indeed, both experimental and numerical investigations have demonstrated the existence of a perturbed area, named the interphase hereafter, at the boundary between the filler and the matrix phase. In this work, we address the construction and identification of a stochastic model for the interphase properties making use of an atomistic representation of the composite system. To this aim, we first perform a set of molecular dynamics (MD) simulations on a model polymer containing spherical nano-inhomogeneities. These simulations provide a complete description of the local morphology of the polymer chains near the fillers, hence allowing for the construction of a suitable random field model. This model is subsequently identified through a statistical inverse problems involving MD results. The impact of such randomness on macroscale properties is finally characterized using a stochastic homogenization procedure.