By the very definition of Hot Melt Extrusion, the polymer processing is to be done when the polymer is in the melt state.  This would naturally imply that you should be above the melting temperature of the polymer.  In the case of thermoplastics, the Tg is less than the Tm and hence working at Tg is not the right thing to do if you are trying to do hot melt extrusion.  Table 1.5 in the book A. V. Shenoy and D. R. Saini, Thermoplastic Melt Rheology and Processing, Marcel Dekker Inc., New York (1996) gives Information on Selected Homopolymers (LDPE, HDPE, PP, PS, PMMA, POM, PA, PET, PC, PVDF, PPO, PPS, PES, PAS, PEEK, PEI, PAr, PVC), and gives the Glass Transition Temperature Tg, Melting Temperature Tm, Highest Processing Temperature Tp and Degradation Temperature Td.  Since the attempt is to find the lowest working temperature with different polymers in the Haake Mini Lab II extruder, you can use the values of Tm and Tp given in the Table 1.5 to determine what value you need to choose between Tm and Tp but as close as possible to Tm.  It is important that, at the chosen temperature, the hot melt extrusion takes place without producing extrusion defects.  The section on Extrusion in Chapter 8 of the book may also help in this endeavor.

Dr. Shenoy answered quite comprehensivly. I would only like to add that the Tg of many commonly extruded polymers is far below Tm. Take PE as an example: the Tm of common PE grades is between 120 and 140°C, while the Tg is below 0°C!

Regards,

Josef

Thank you very much for your detailed explanations...

For melt extrusion crystalline polymer - melting point + 10-15 ° C; amorphous polymer - glass transition temperature x2 (2 Tg).

On the extrusion temperature based on the Williams-Landel-Ferry (WLF) viscosity model:

The dynamic viscosity of a polymer at a temperature (T), between the glass transition temperature (Tg) and the melting temperature (Tm), is typically well correlated by the Williams-Landel-Ferry (WLF) viscosity model: μ = μº·exp [-C1·ΔT / (C2 + ΔT)]. Here μº denotes the dynamic viscosity at the reference temperature, Tº ≥ Tg, while we can identify Tº with Tg. The temperature T, exceeds Tº by a ΔT difference: ΔT = T-Tº, while C1 and C2 are fitting parameters, which can be determined from viscosity measurements. The viscosities at Tg and Tm, respectively μº and μ*, are often identified with generic widely accepted values, regardless of the actual polymer considered. We may rewrite the WLF correlation as: Tm - Tg = - C2 ·ln (μ* / μº) / [ln (μ* / μº) + C1].

Let us admit that for some conveniently selected extrusion temperature, Tx, the viscosity should be μx, which can possibly be found acceptable for many polymers. We shall have: Tx - Tg = - C2 ·ln (μx / μº) / [ln (μx / μº) + C1]. The folowing dimensionless temperature ratio can now be obtained in terms of a single polymer-dependent parameter (C1): (Tx - Tg) / (Tm - Tg)  = {ln (μx / μº) / [ln (μx / μº) + C1] } / {ln (μ* / μº) / [ln (μ* / μº) + C1]}.

I think you must have atemperature over the melting point of polymers (Tm).