VISCOELASTIC PROPERTIES OF BINARY BLENDS OF LINEAR POLYSTYRENES - FURTHER EXAMINATION OF CONSTRAINT RELEASE MODELS

The viscoelastic behavior of probe polystyrene (PS) chains in PS/PS blends was compared with the predictions of the so-called CICR and CDCR models for a combined constraint release (CR) plus reptation process. In the CICR model, the CR rate constant was assumed to be independent of the probe chain configuration. On the other hand, in the CDCR model, the CR process was assumed to proceed through configuration-dependent Rouse dynamics. In the eigenfunction expansion procedure of this model for viscoelastic relaxation, only diagonal Rouse eigenfunctions were used to evaluate the relaxation rate due to CR. This approximation was referred to as the diagonal-dominance approximation (DDA). As the matrix molecular weight M(s) approached the probe molecular weight M(L), the M(s) dependence of the relaxation time of the probe chain became gradually weaker, and the terminal mode distribution became narrower. Such behavior was much better described by the CDCR model than by the CICR model. This indicated the configuration dependence to be essential for actual CR processes. However, even for the CDCR model, agreements with the data were poor at short time scales, as most clearly seen in the CR-dominant regime (M(L) >> M(s)). In that regime, the CDCR model with DDA was reduced to a pure CR model rigorously using Rouse dynamics (without DDA) and was found to overestimate fast viscoelastic modes. Thus, Rouse dynamics was valid as the actual CR dynamics only at long time scales. This in turn suggested that reasonably good agreements between the slow viscoelastic data and the present CDCR model in the entire range of M(s) may be related to DDA used in the model. For M(S) not much smaller than M(L), DDA underestimated fast viscoelastic modes and seemed to cancel to some extent the overestimation for those modes due to use of Rouse dynamics in the model.