Shear banding in time-dependent flows of polymers and wormlike micelles

TL;DR

This paper develops a universal linear stability criterion for shear banding in time-dependent flows of complex fluids, based on rheological response shapes, applicable to polymers, micelles, and other disordered soft materials.

Contribution

It introduces a fluid-universal shear banding criterion dependent only on rheological response shape, applicable across various complex fluids and flow protocols.

Findings

Transient banding can occur even in fluids with monotonic constitutive curves.

Numerical support from rolie-poly and Giesekus models confirms predictions.

The rolie-poly model captures experimental phenomena better than the Giesekus model.

Abstract

We study theoretically the formation of shear bands in time-dependent flows of polymeric and wormlike micellar surfactant fluids, focussing on the protocols of step shear stress, step shear strain (or in practice a rapid strain ramp), and shear startup, which are commonly studied experimentally. For each protocol we perform a linear stability analysis to provide a fluid-universal criterion for the onset of shear banding, following our recent Letter [Phys. Rev. Lett. 110 (2013) 086001]. In each case this criterion depends only on the shape of the experimentally measured rheological response function for that protocol, independent of the constitutive properties of the material in question. (Therefore our criteria in fact concern all complex fluids and not just the polymeric ones of interest here. A separate manuscript [in preparation] will explore them in a broad class of disordered soft glassy materials including foams, dense emulsions, dense colloids, and microgel bead suspensions.) An important prediction is that pronounced banding can arise transiently in each of these protocols, even in fluids for which the underlying constitutive curve of stress as a function of strain-rate is monotonic and a steadily flowing state is accordingly unbanded. For each protocol we provide numerical results in the rolie-poly and Giesekus models that support our predictions. We comment on the ability of the rolie-poly model to capture the observed experimental phenomenology, and on the failure of the Giesekus model.