The plot presents a strong and somewhat narrow exothermic peak obtained on cooling, which is distorted (tilted backwards, to higher temperature). Seemingly, the peak shape corresponds to thermogram traced as a function of the sample temperature; not versus the temperature of the reference (empty pan) as it is more usual. Heat released at the sample pan possibly was not dissipated soon enough for the sample temperature to decrease as scheduled, and its temperature might have increased for short time, resuming cooling only shortly after. That being the case; the peak would certainly look much more symmetrical if you may want to consider the following options: (i) reduce the quantity of sample; (ii) dilute the sample with α-Al2O3 powder or other thermally inert substance; (iii) select the temperature of the standard for the horizontal axis (rather than sample temperature); (iv) select time for the horizontal axis; (v) select pan material for higher thermal conductivity.

Carlos Araújo Queiroz

I've reduced the volume of sample to 10 µl and got more narrow peak. I seems like the higher volume of the sample allow to obtain more simmetrical peak.

What could be the problem?

I will check other your sugestions soon.

Could there be a problem in that I use an open aluminum crucible for a liquid sample?

Hi

May be the reference pan material is problem... Try for alumina crucible it will be stable in temperature varients.

Best

Arunprakash Vincent

Dear Elizaveta:

I would not call the peak for the smaller quantity of sample more symmetrical. In both cases the peak for the cooling thermogram is tilted tilted backwards (to higher temperature) and its higher temperature side is nearly straight. The lower temperature side has a sigmoidal shape for the larger quantity sample, but for the smaller one, is more close to the straight line shape of opposite side, except at its foot.

Comparing both assays, it can be noticed that the peak returns to base line at closer temperatures (peak offset) than at the peak temperature, and at its departure from base line (peak onset). I find the peak obtained for the smaller quantity narrower, nearly as well tilted, but actually more symmetrical in its overall shape appearance.

The effect you have characterized is indeed very characteristic for the crystallization behavior of a rather pure substance on cooling. The effect is simply the supercooling effect: The substance starts crystallization always at temperatures lower than the melting temperature because of missing nuclei. When the crystallization starts the temperature of the sample rizes emedeately to the equlilibrium melting temperature and when all substance is crystallized the temperature of the samlpl drops again. This is exactely what yousee in your measurement. [Remark: The purer a substance is, the larger is the supercooling value and the tilting angle, this can go up to several tens of degees.] To get more "normal" crystallisation peaks you may choose "program temperature" as x-axis in your plot. (Obviously you use a power-compensation DSC 4000).

Günther W. H. Höhne

Thanks a lot! It is literally what I was thinking about this curve, but I wasn't sure.