Chain Motion in Nonentangled Dynamically Asymmetric Polymer Blends: Comparison between Atomistic Simulations of PEO/PMMA and a Generic Bead−Spring Model

Abstract

The polymer blend of polyethyleneoxide (PEO) and polymethylmethacrylate (PMMA) constitutes a miscible blend of high dynamical asymmetry; that is, the fully miscible components exhibit a large difference in their glass-transition temperatures, which are 200 K apart. To get a deeper understanding of the unusual PEO dynamics in this system, we have performed a fully atomistic MD simulation. Here we present all information and results obtained on the chain self-motion. We present the mean square displacements and the associated non-Gaussian parameters as a function of temperature. The associated self-correlation function is compared thoroughly with experiments. We display a Rouse analysis and find strongly modified mode friction coefficients but restoring forces that are identical to the pure melt. Thereby, the Rouse correlators are strongly stretched, and the mode number, p, dependence of the relaxation times deviates strongly from the p−2 Rouse behavior. We have also carried out simulations of a simple bead−spring blend, which exhibits the same qualitative dynamic features of the PEO/PMMA system. This suggests that such features are not specific of the PEO/PMMA system, but they are generic in real polymer blends with strong dynamic asymmetry. A further important issue was the test of different models that have been invoked to explain the anomalous PEO dynamics. We compare with a generalized Langevin equation (GLE) approach and with a random Rouse model dealing with a random distribution of friction coefficients. In all aspects, the GLE model agrees qualitatively very well with the results of the fully atomistic simulations. The random Rouse model may be considered to be a phenomenological instantaneous approximation valid for the case where the density fluctuations of the slow PMMA components are relaxing much slower than the relevant PEO dynamics.