The evaluations of Tg shown in the graphs are wrong, because of the following reason: Tg appears in a DSC plot as a step-like endothermic change of the heat flow rate heating curve. What is seen in the plots is an exothermic flat peak between 70 and 130 °C. I assume, that this is some reaction heat from further hardening taking place during the heating run. As the peak shifts somewhat to higher temperatures in the second run (red curve) support this assumption. To detect Tg which is hidden by reaction heat, it is recommended to use temperature modulated DSC (TMDSC). The endothermic peak between 110 and 150 °C may be interpreted as evaporation of some remaining solvent. To test these ideas more measurements are needed. For details of Tg measurements see chapter 6.4 and Fig. 6.31 in our textbook "Differential Scanning Calorimetry" 2nd Ed. Springer 2003 and a lot of publications on epoxid hardening.

Regards Gunther Höhne

Dear colleague Simon Nübler,

Have a nice time and good day.

The glass transition temperature depends on many parameters; softening, hardening, localized deformation, applied stress, temperature, the strain rate, viscosity, inhomogeneous deformation region, bond strength  bond breakage, the mobility of the cations and anions, connectedness of the network, the average coordination number, the quality of connections, bond energies of heteropolar and homopolar bonds and heat capacity. try to check the mentioned parameters.  

Having a look to your attached data, the maximum temperature of your attached Fig. is 190 C. this is may be  low temperature so you may need for higher temperature may be 1000 C, it depends on the material (you don't report the focused material).

So you can try more high temperature.

 There is an attach file supported by references contains useful equation may help and be useful for you.

Yours

A. Eldenglawey 

Hello Simon,

I think there a two points to mention:

1) Your method of destination the Tg is not possible, because there is no Tg to see.

2) If you want to measure the Tg of a epoxy material it shoud be heated up to 300°C.

Then the Tg shoud be around 195-200°C measured with DSC. Better you use a DMA-Methode, it's much more sensitive than DSC

I think the attached file may help you.

I agree that you're probably not running your experiment to a high enough temperature.  But 1000 C is more than you need.  It should be sufficient to run it to 400 C to 500C. 

What you're likely seeing at the lower termperatures is simply water being expelled from the sample. 

See you my papers

Dear Simone,

Dear All,

I think just another important thing that we tend to underestimate while performing Tg analyses on materials is that, glass transition being a reversible thermal effect, it can also be seen in the cooling segment. So, if you carry on typical heating-cooling-heating cycle experiments, you may want to analyse also the cooling portion of the curve.

The identification of Tg in cooling can be sometimes more easy than on heating (especially the first one). Furthermore, it can be used to confirm the Tg presence and numerical evaluation.

Hope this helps. Samuele

you can also try out a search on http://www.mt.com/us/en/home/search/matchar-search.html

Dear Simon,

currently I'm curing a bis-phenol A epoxy with TETA. Curing run 0 - 270C at 10C/min.

1) Your DSC thermograms has a peculiar exothermic trend. "Normal" DSC baseline is slightly endothermic. Maybe your DSC cell needs recalibration? Perhaps rotate the baseline?

2) To understand complex Tg's on DSC, I normally plot the 1st der of heatflow vs temp. This gives great help for pinpointing a Tg, remember Tg is a step transition to another Cp baseline.

3) To make sure your sample is fully cured, open up the pan and visually see what state the sample is in. Also you could put a sample in an oven to see what happens isothermally.

The evaluations of Tg shown in the graphs are wrong, because of the following reason: Tg appears in a DSC plot as a step-like endothermic change of the heat flow rate heating curve. What is seen in the plots is an exothermic flat peak between 70 and 130 °C. I assume, that this is some reaction heat from further hardening taking place during the heating run. As the peak shifts somewhat to higher temperatures in the second run (red curve) support this assumption. To detect Tg which is hidden by reaction heat, it is recommended to use temperature modulated DSC (TMDSC). The endothermic peak between 110 and 150 °C may be interpreted as evaporation of some remaining solvent. To test these ideas more measurements are needed. For details of Tg measurements see chapter 6.4 and Fig. 6.31 in our textbook "Differential Scanning Calorimetry" 2nd Ed. Springer 2003 and a lot of publications on epoxid hardening.

Regards Gunther Höhne

From the curves you provided the midpoint is about 107°C, and 125°C for the two runs. It looks not bad at all. However, to have a better Tg, there is few step you need to do:

-Clean up you DSC, since I saw some noises on the curves at low and high temperature especially at about 180-200°C.

-Run the baseline to make sure there is no noise.

-Begin the run at lower temperature, because the step change on the curve is too close to 35°C degree, I suggest at least start at 0°C.

-Change the ramp rate to 20°C/min.

You should use procedure somewhat as below:

Under nitrogen atmosphere, approximately 10 mg of dry material was used in an aluminum pan. The samples were cooled from room temperature to at least 0°C at cooling rate of 20°C/ min, isothermal for 1 min, then heated to about 200°C at heating rate of 20°C/min (first measurement cycle), isothermal for 1 min, cooled to 0°C at cooling rate of 20°C/min, isothermal for 1 min, then reheated to 200°C at heating rate of 20°C/min (second measurement cycle).

Hope this will help.

TD