| Abstract: | Almost all polymer processing operations involve moving and shaping the polymer as a melt and then cooling it, usually quite rapidly, to a solid state. In order to develop models for such processes we have begun systematic studies in non-isothermal rheology; here we interpret the results in the context of melt spinning. Theoretical predictions of stress vs distance from the spinneret were calculated from generalized (non-isothermal) viscoelastic theory and compared with Dees' melt spinning data on high density polyethylene. Despite certain experimental and theoretical difficulties, the agreement is good. Surprisingly, an additional theoretical curve, based on a simple Trouton viscosity, also gave a reasonable approximation over much of the distance, despite the transient nature of the flow. To understand this phenomenon further, a more tractable theoretical problem was studied in detail; the problem, of constant elongational flow (? = constant) in the presence of a constant rate of temperature change (dT/dt = constant). The results depend on two dimensionless groupings; the first is the usual product of a relaxation time and ?; the second involves the ratio of normalized dT/dt to ?. When the second group is large, a quasi-viscous state exists. The melt spinning data for the HDPE may be near this state. |
