You can use the software accompanying DSC to compute the Tg irrespective of the cooling curve.

I need to add that heat-cool program at 10 K/min and 5 K/min helps you to improve the runs by checking the thermal history. However, this improves other parameters also such Tm etc in addition to Tg  but you still need to recognise your parameters. 

For perfect heating run, it would be proper to use 5 K/min, but also this depends on what your intention. If it is to gauge where your parameters fall, you can do it at 10 K/min. Cooling can be done as fast as you can, perhaps by placing the pan in ice or even liquid nitrogen before the second heating. 

Dear collegue,

Effectively, there are two thermodinamic transitions in you graph. The first one could be a solid to solid, or solid to solid+liquid transition. There is not a rule, it depends of the phase diagram of your system. The second one could be the melting...

I guest that your curve is the heating cycle, please include the cooling cycle, generally the curves are different and the information must be analyzed together.

With my best regards,

Pedro

There could well be two Tg events there.  This would be the case if the system showed liquid-liquid immiscibility (either stable or metastable), which may be reported in the phase diagram if it exists?  The two Tg events would then correspond to the two glass phases formed from the two immiscible liquids.

@Oliver_Alderman and @P_Gallardo Thank you very much for your reply.

Its actually an epoxy sample, I have mixed with another chemical to modify it before curing it with hardener. I don't have the phase diagram.

Is it possible that the system was unsuccessfully modified with addition of chemical and it forms separate phases?

The first drop was not observed in the control sample.

Yes the curve I have is a heating curve. I have never looked at cooling curve before. What information can I get from cooling curve.

You can use the software accompanying DSC to compute the Tg irrespective of the cooling curve.

I need to add that heat-cool program at 10 K/min and 5 K/min helps you to improve the runs by checking the thermal history. However, this improves other parameters also such Tm etc in addition to Tg  but you still need to recognise your parameters. 

For perfect heating run, it would be proper to use 5 K/min, but also this depends on what your intention. If it is to gauge where your parameters fall, you can do it at 10 K/min. Cooling can be done as fast as you can, perhaps by placing the pan in ice or even liquid nitrogen before the second heating. 

Dear Si Abdul, to decide whether there is one ore two Tg's you must do more: Run a heat-cool-heat program with a heating/cooling rate of 10 K/min. The first run gives the effects of the original sample (with its specific thermal history) the cooling run reproduces the glass transition(s) if they exist, the second heating run gives thermal events which are independent from the preparation and possible events like evaporation, reactions and other processes taking place during the 1. run. This procedure is always recommended for DSC-measurements. In addition it could be helpful to subtract an empty pan baseline. with such measurements it would be much easier to interpret your results. Good luck regards Gunther Höhne.

Dear Si Abdul, I fully agree  with Dr. Höhne.

In addition, I can suggest to run a Modulation Temperature DSC. It can be run with all DSC instruments, but from the output I recognize that you are working with a TA Instruments DSC. The TA software allows you writing directly the procedure for MDSC. Just follow the recommendation on scan rate, amplitude and frequency of the oscillation that you can find on the TA Instruments webpage or online. By means of MDSC you should separate the Tgs (in the Reversing signal) from the kinetically driven effects (in the Non Reversing).

Hi all,

Thank you for your reply. I did some extra reading and indeed modulated test would give me more information.

I have another question.

My first run produces endothermic relaxation peak near Tg while the second run removes it. Which Tg is more accurate or technical correct Tg to take as comparison to compare with my other samples?

It seems the Tg deviates after the second run.

Thank you in advance! :)

It looks like your sample crystallizes on cooling.  Therefore the Tg in the second heating is only for any residual glass.  This is also why you don't see crystallization exotherm in the 2nd heating.

Also, be careful, even if you had a glass for 2nd heating (which you don't) the thermal history and fictive temperature are likely different to the starting material in 1st heating run.  Therefore the morphology of the Tg event will be different.  See:

Moynihan CT, Easteal AJ, Debolt MA, & Tucker J (1976) Dependence of Fictive Temperature of Glass on Cooling Rate. J. Am. Ceram. Soc. 59(1-2):12-16.

Dear San Ty, concerning your question above: as I told you some month ago: The thermal glass transition (measured in a DSC) depends on the thermal history. As the thermal history of your sample is different for the 1st run (stored or annealed or otherwise treated during production) and the second run up (after cooling with 20 K/min) the measured Tg must be different as well. The rather large relaxation peak of the first run appears if the sample was cooled very slowly or the sample was stored for a longer time at temperatures somewhat below the Tg (what was not the case for the second run up!). Furthermore in the 1st run there is a larger Tg step, because the original sample is almost amorphous (there is a "cold crystallization" peak at 150 oC). During the cooling run there appears a crystallization peak af 190 oC. Consequently the sample is partial crystalline then and the glass transition step become much smaller during the second run up. To sum up: both Tg's are different but true because of the different thermal history.

If you will do a temperature modulated DSC measurement you get another Tg again, this is from the dynamic glass transition process which depends on the frequency but not on the thermal history. Please do not mix these different Tg's.

The behavior of your sample is as expected and well known for PET.