Search

Stress relaxation in immiscible blends is studied for a well defined shear history, i.e. after prolonged steady state shearing. Model systems are used that consist of quasi-Newtonian liquid polymers. Hence the relaxation is dominated by changes in the morphology of the interface. Both shear stress and the first normal stress are considered. The measurements cover the entire concentration range. For dilute blends the interfacial contribution to the stress relaxation compares well with model predictions. Deviations occur when the matrix phase is slightly elastic. In that case the similarity between the relaxation of shear and normal stresses is also lost. The latter is attributed to a wider drop size distribution. Increasing the concentration of the disperse phase results in a complex evolution of the characteristic relaxation times. The normal stresses relax systematically slower than the shear stresses and the concentration curve includes two maxima. Even for equiviscous components the concentration curves are not symmetrical. It is concluded that even a slight degree of elasticity in the matrix phase drastically affects the morphology and the interfacial relaxation of such blends.

The relation between morphology and the steady state rheological behavior has been investigated for blends of immiscible polymers, i.e., polydimethylsiloxane and polyisobutylene, and this over the full composition range. Cryogenic scanning electron microscopy has been used to obtain information about the morphology during shearing. By means of relaxation experiments the stresses have been decomposed in interfacial and component contributions. From the interfacial stress components the orientation of the inclusions could be estimated. These stresses do not provide a clear indication of the location of the phase inversion as they hardly change with concentration in the midconcentration range, they even seem to develop a local minimum around the 50/50 concentration. In this region fibrils can be observed that are nearly oriented in the flow direction. The values of the interfacial stress components are consistent with such a structure. The component contributions to the viscosity can also be explained on the same basis although the normal stresses are underestimated.

We revisit the experiments of start-up shear of an immiscible dilute polymer blend presented in [J. Non-Newtonian Fluid Mech. 99 (2001) 167] and calculate the transient stresses via numerical simulations. As part of this investigation, we validate a stress-relaxation experimental technique [J. Non-Newtonian Fluid Mech. 99 (2001) 167; J. Non-Newtonian Fluid Mech. 93 (2000) 153] for the measurement of the interfacial contribution. The simulations assume Newtonian behavior, thus, not including bulk viscoelastic effects, and are performed using an adaptive boundary-integral method [Phys. Fluids 10 (8) (1998) 1781; J. Comput. Phys. 168 (2001) 445] to describe highly deformed drops and the effect of breakup events on the stresses. When viscoelasticity is negligible, we find agreement between simulations and experiments for the total stress and interface contribution at both subcritical and supercritical capillary numbers, and demonstrate that no existing analytical models are capable of describing correctly the stress evolution. Interestingly, we find that the breakup time is independent of the capillary number at high capillary number, and depends on the viscosity ratio only. When viscoelasticity is important, the comparison between the simulation results and the experimental data allows us to quantify the effect of viscoelasticity on the stresses by isolating correctly the Newtonian contribution.

The transient rheological behaviour of immiscible polymer blends is studied during start-up flows. The stresses can be divided in a contribution from the component polymers and a contribution caused by the presence of an interface. By means of relaxation measurements, the component stresses after inception of flow can be determined. In this manner, the different contributions to the stress can be compared separately with model predictions. For the contribution of the components, one normally uses mixing rules that do not take into account the morphology of the blend. As expected, a slight morphology effect can be detected. The stress contributions from the interface are calculated combining the Doi–Ohta approach with different models that describe the droplet deformation. For flows at small capillary numbers, the Maffettone–Minale model leads to accurate stress predictions. With larger capillary numbers, the normal stress profiles become more complex. The initial part is rather well described by the model of Almusallam et al. [J. Reol. 44 (2000) 1055]. It is demonstrated that the morphological changes in the blend can be deduced in quite some detail from the measured stress profiles.

The Maffettone–Minale analysis describes the shape of ellipsoidal droplets in immiscible two-phase polymer model blends during flow. It is used here to calculate the elastic interfacial contribution to the shear stress (σ12,int), and first normal stress difference (N1,int). The ratio N1,int/σ12,int can be linked to the orientation angle of the inclusions. Under steady state flow the Maffettone–Minale model gives an analytical expression between the orientation angle and the capillary number. In this manner the capillary number can be deduced from the orientation angle. From this the droplet size is calculated. Good agreement is found between the results and droplet sizes determined independently from dynamic measurements using the Palierne model. The stresses calculated from the model, with given values for the droplet size, compare quite well with the stresses measured on a model system for relatively small capillary numbers.