Understanding Extrudate Swell: From Theoretical Rheology to Practical Processing

This thesis focusses on the measurement and prediction of extrudate (or die) swell of molten polymers. The overall aim of this work is to understand extrudate swell for complex, industrially relevant systems. This is performed by first understanding the causes of swelling for well defined, monodisperse polymers at a molecular level. The systems are then gradually built up in complexity from bidisperse to very polydisperse and/or branched samples. At each stage predictions for extrudate swell are obtained using the \textit{flowSolve} fluid dynamics package combined with a molecular constitutive equation and are compared to extrusion experiments using a novel Multi-Pass-Rheometer setup. The effects of both molecular weight and temperature can be ignored when shear rates are scaled by Rouse Weissenberg number as extrudate swell is a chain stretch controlled phenomenon. For monodisperse systems theoretical predictions using the Rolie-Poly constitutive equation match experimental results up to a $W_R$=7 above which simulations over-predict swelling ratios. This is justified in this work using reduction of monomeric friction at high deformation rates. Extrudate swell of polydisperse polystyrenes is successfully predicted up to high Weissenberg numbers using the Rolie-Double-Poly equation when combined with monomeric friction reduction. A slight under-prediction is seen at low Weissenberg number where the chain stretch times of long polymer chains are increased by dilution with shorter chains. Qualitatively correct but quantitatively poor predictions are obtained for highly polydisperse polyethylenes where the low shear extrudate swell is under-predicted. Branched polymers behave differently experimentally to linear samples, exhibiting extrudate swell below the Rouse time of the polymer backbone. A small amount of branching increases swelling ratios versus the linear case but moderate increases in branching above this point have little effect on the experimentally observed swelling ratios. Significantly branched polyethylenes swell more than this, especially at high shear rates. There is a similar trend in simulated results using the XPP model but only a partial agreement between simulated and experimental extrudate swell is observed