Many thanks for the interesting talk. One remark regarding "Tz" in ABAQUS (for Arrhenius model). According to the official documentation: " ... Tz is the absolute zero in the temperature scale being used ..." What I read as: "Tz = 0" in case of Kelvin temperature scale, and "Tz = -273.15" for Celcius temperature scale being used. Therefore, Tz in not an independent parameter of the Arrhenius model. Note that the WLF equation uses a temperature difference (T - T0), which is the same for both temperature scales. So we don't have to pay attention to "Tz" in this case.
Very nice presentation, many thanks for posting it. I have three comments. (1) I am confused why ABAQUS called their model Arrhenius. As you presented, this shift factor form is the standard Vogel-Fulcher-Tammann-Hesse (VFTH) form, which is indeed equivalent to WLF. The ANSYS and LS-Dyna form is the true Arrhenius and this is why they have one fewer parameter. (2) According to many authors (e.g., van Krevelen, Properties of Polymers), the shift factor in reality has two regions: the Arrhenius region at temperatures T > 1.2Tg, and the WLF region at temperatures T < 1.2Tg. One can come up with interpolation shift factors encompassing both regimes (I proposed one called TS2 a couple years ago, see Soft Matter 2020, 16 (3), 810-825). (3) WLF has a divergence issue as T --> T0 - C. This may not matter for most practical calculations but can be unpleasant if one writes an automated code that attempts to blindly go into the low-temperature region. Recently, there have been a lot of discussions in the community whether the WLF-predicted divergence is "real" and there are indications that it is not. My bottom line -- WLF is probably fine if one does not attempt to study viscoelastic melts or crosslinked rubbers where T > 1.2Tg. For those, it is better to use the "true" Arrhenius (with Tz = 0).
Many thanks for the interesting talk. One remark regarding "Tz" in ABAQUS (for Arrhenius model).
According to the official documentation: " ... Tz is the absolute zero in the temperature scale being used ..."
What I read as:
"Tz = 0" in case of Kelvin temperature scale, and
"Tz = -273.15" for Celcius temperature scale being used.
Therefore, Tz in not an independent parameter of the Arrhenius model.
Note that the WLF equation uses a temperature difference (T - T0), which is the same for both temperature scales. So we don't have to pay attention to "Tz" in this case.
Very nice presentation, many thanks for posting it. I have three comments.
(1) I am confused why ABAQUS called their model Arrhenius. As you presented, this shift factor form is the standard Vogel-Fulcher-Tammann-Hesse (VFTH) form, which is indeed equivalent to WLF. The ANSYS and LS-Dyna form is the true Arrhenius and this is why they have one fewer parameter.
(2) According to many authors (e.g., van Krevelen, Properties of Polymers), the shift factor in reality has two regions: the Arrhenius region at temperatures T > 1.2Tg, and the WLF region at temperatures T < 1.2Tg. One can come up with interpolation shift factors encompassing both regimes (I proposed one called TS2 a couple years ago, see Soft Matter 2020, 16 (3), 810-825).
(3) WLF has a divergence issue as T --> T0 - C. This may not matter for most practical calculations but can be unpleasant if one writes an automated code that attempts to blindly go into the low-temperature region. Recently, there have been a lot of discussions in the community whether the WLF-predicted divergence is "real" and there are indications that it is not.
My bottom line -- WLF is probably fine if one does not attempt to study viscoelastic melts or crosslinked rubbers where T > 1.2Tg. For those, it is better to use the "true" Arrhenius (with Tz = 0).