An experimental system has been found recently, a coagulated CaCO3 suspension system, which shows very variable yield behaviour depending upon how it is tested and, specifically, at what rate it is sheared. At Peclet numbers Pe > 1 it behaves as a si
mple Herschel Bulkley liquid, whereas at Pe < 1 highly non-monotonic flow curves are seen. In controlled stress testing it shows hysteresis and shear banding and in the usual type of stress scan, used to measure flow curves in controlled stress mode routinely, it can show very erratic and irreproducible behaviour. All of these features will be attributed here to a dependence of the solid phase, or, yield stress, on the prevailing rate of shear at the yield point. Stress growth curves obtained from step strain-rate testing showed that this rate-dependence was a consequence of Peclet number dependent strain softening. At very low Pe, yield was cooperative and the yield strain was order-one, whereas as Pe approached unity, the yield strain reduced to that needed to break interparticle bonds, causing the yield stress to be greatly reduced. It is suspected that rate-dependent yield could well be the rule rather than the exception for cohesive suspensions more generally. If so, then the Herschel-Bulkley equation can usefully be generalized to read (in simple shear). The proposition that rate-dependent yield might be general for cohesive suspensions is amenable to critical experimental testing by a range of means and along lines suggested.
The yielding of concentrated cohesive suspensions can be deformation-rate dependent. One consquence of this is that a single suspension can present in one several different ways, depending upon how it is tested, or more generally, how it is caused to
flow. We have seen variously Herschel-Bulkley flow, highly non-monotonic flow curves and highly erratic or chaotic yield, all in one suspension. In controlled-rate testing one sees a systematic effect of deformation rate. In controlled stress testing, matters are more subtle. Whereas step-stress creep testing will elicit reproducible behaviour, any attempt to determine a flow curve by, e.g. stepping up or sweeping stress at an inappropriate rate can lead to highly irreproducible behaviour.
Measuring yielding in cohesive suspensions is often hampered by slip at measurement surfaces. This paper presents creep data for strongly-flocculated suspensions obtained using vane-in-cup tools with differing cup-to-vane diameter ratios. The three s
uspensions were titania and alumina aggregated at their isoelectric points and polymer-flocculated alumina. The aim was to find the diameter ratio where slip or premature yielding at the cup wall had no effect on the transient behaviour. The large diameter ratio results showed readily understandable material behaviour comprising linear viscoelasticity at low stresses, strain-softening close to yielding, time-dependent yield across a range of stresses and then viscous flow. Tests in small ratio geometries however showed more complex responses. Effects attributed to the cup wall included delayed softening, slip, multiple yielding and stick-slip events, and unsteady flow. The conclusion was that cups have to be relatively large to eliminate wall artefacts. A diameter ratio of three was sufficient in practice, although the minimum ratio must be material dependent.
The behaviour in simple shear of two concentrated and strongly cohesive mineral suspensions showing highly non-monotonic flow curves is described. Two rheometric test modes were employed, controlled stress and controlled shear-rate. In controlled str
ess mode the materials showed runaway flow above a yield stress, which, for one of the suspensions, varied substantially in value and seemingly at random from one run to the next, such that the up flow-curve appeared to be quite irreproducible. The down-curve was not though, as neither was the curve obtained in controlled rate mode, which turned out to be triple-valued in the region where runaway flow was seen in controlled rising stress. For this first suspension, the total stress could be decomposed into three parts to a good approximation: a viscous component proportional to a plastic viscosity, a constant isostatic contribution, and a third shear-rate dependent contribution associated with the particulate network which decreased with increasing shear-rate raised to the -7/10th power. In the case of the second suspension, the stress could be decomposed along similar lines, although the strain-rate softening of the solid-phase stress was found to be logarithmic and the irreducible isostatic stress was small. The flow curves are discussed in the light of recent simulations and they conform to a very simple but general rule for non-monotonic behaviour in cohesive suspensions and emulsions, namely that it is caused by strain-rate softening of the solid phase stress.
