This is indeed a common feature for most materials: you need to do a first heating to melt the material and therefore get a good thermal contact with the DSC crucible. This first cycle is very important, also for metal powders for example. Only the successive cycles are worth, there is no need to have the 1st cycle perfect, as it is not possible for the aformentioned reason of thermal contact. Heating the aluminium pan with the material inside (in inert atmosphere) before doing the DSC measurement will allow to get the 1st cycle as perfect as the following ones.

Alain

I agree with Prof. Alain Celzard, It is a common feature of most of the films & flakes (particularly polymers and electrodeposited  samples). This is due to the stress and strain  released from the sample during heating cycles. This can be removed as suggested by Prof. Alain.

Dear Prof.Celzard and Prof. Arumainathan, thank you for your responses. I know that the first heating is usually used to eliminate the thermal history of the materials. However, because i would like to identify the effect of the post-treatment on the thermal behavior (such as Tg, Xc and Tm) of the same polymer film, I therefore need to use the first heating curves.

Dear Jalali Dil, thank you for your comments. Indeed, I tested a PVA film prepared by hot water solution. Even though the thin film was well dried before the DSC test, it still had some water. It might be used to explain the samll endothermic peak around 150 degree and the wight loss of the material. In addition, i did not conduct a TGA test to measure the decomposition temperature of the PVA film, I chose a temperature range from room temperature to 250 degree according to the literature (similar PVA film preparation method). From the DSC curves, the melting just finished around 230 degree.

OK, so would you be able to prepare your PVA sample DIRECTLY in the aluminium span ? It would probably solve the problem ...

Good luck

Alain

I also investigated the first run DSC after processing my materials by cast extrusion to investigate the drawing effect. When you deal with thin films you have some special issues. I was asking my self the same question so I hope that the few tricks which work for me will work for you. First, I superposed the film to have a decent thickness (I took around 1.5 mm) and then I used a punch to cut into my superposed films. In the better case scenario, maybe your lab has a punch of the exact same size than you pan. The punch allows you to have a smaller and regular sample (maybe you'll need to cut again into it to do an even smaller one, but cut properly with a cutter) and the layer will be stick to each other because of the pressure on it.

Then, I always take the same mass of material (approximately 10 mg (+/-1mg) in a 50µL pans with holes) to compare. And I see that your compressed pan is all twisted and dented, you need to be more carefull with it. It has to be properly sealed. With properly sealed and especially not dented pans you'll also have less baseline problems.

Hope that it will be useful, goof luck

Violette

Dear Violette, thank you for your suggestions. My colleague suggested me to use a small sample to reduce the effect of  contact surface changing. I therefore used a single layer film 2-3 mg. Moreover, because i saw sometimes the lips opened after the first heating. I fold the DSC pans to fix the lips' position. I will try your way to conduct my test. Thank you!

Yes it sure do solve the interlayers slipping effect but I am pretty sure that taking a so little mass won't be enough if you want to see the Tg. About the lips open after the first heat, did you consider taking pans with holes? I think that folding the pans may lead to baseline deviation (the sample pan has to be exactly similar to the reference pan). Please keep me inform of the result of taking a 2-3 mg sample mass.

I agree with more of the previous answers: 

- solvent weight losses are evident, have you weighted your samples pan after scan?

- stress (entropic by segment orientation) and enthalpy relaxation may occur at the first run. The first one may give sample displacements and Tg is start occurring at lower temperature then (once segments orientation is gone) the true glass transition occurs. Enthalpy relaxation gives peak at the glass transition. You may have these two phenomena occurring simultaneously. Increasing the heating rate you can better differentiate the two events. Lower heating rates give you a more smooth curve but the two phenomena may superimpose and give a "false" thermogram.

u'd better to use the lid of sample holder

First reduce the size of the slab (there is a problem of dimensinal stability with one dimension much smaller). You can cut the same sample in several pieces and you shoul observe a more clean curve.

Anyway, after having compressed the lid, have you pinched. it?

Since that was the 2014 question, can we know the result of your experiment Sir, Kui Wang if you don't mind, that's a very interesting topic to learn?

Thank you, and kind regards.

You may check also my answer given to a related question ─ «Will bi-layer film affect the Differential Scanning Calorimetry results?»: https://www.researchgate.net/post/Will_bi-layer_film_affect_the_Differential_Scanning_Calorimetry_results