Coarse grained model of diffusion in entangled bidisperse polymer melts.

Chain diffusion is studied in mixtures of bidisperse linear polymers of same chemical identity by means of simulations. The two subpopulations are moderately to highly entangled, with the shorter chain length N(S), fulfilling N(S)N(e)> or =5. To this end, a coarse grained model calibrated to reproduce both the structure and dynamics of chains in monodisperse entangled melts is used [A. Rakshit and R. C. Picu, J. Chem. Phys. 125, 164907 (2006)]. Its performance in reproducing chain dynamics in a polydisperse melt is tested by extensively comparing the results with those obtained from an equivalent fine scale representation of the same system (a bead-spring model). The coarse grained model is used further to investigate the scaling of the diffusion coefficient with the length of the two types of chains and its dependence on the respective fractions. The model reproduces many features observed experimentally. For example, the diffusion coefficient of one of the chain types decreases with increasing the length of the other type chains. It is shown that, in this model, this effect is not linked to constraint release. When the matrix chains become sufficiently long, their length does not influence the diffusion coefficient of the short chains anymore. The diffusion coefficient of the short chains scales with their weight fraction in a manner consistent with experimental observations. In mixtures, the dynamics of the short chains is slower and that of the long chains is marginally faster than in their respective monodisperse melts.