HI Shaunak,

Please see the attached document. It is by TA instruments on how to measure specific heat using DSC.

I hope this helps.

Regards,

Ali Amiri

The TA-Instrument paper is of course helpful to handle the Q1000 DSC, but it does not answer your question concerning the "3 run" procedure go get precise cp-values. In short words: The first run (with sample) gives the originally measurement; the second run (with empty pan of same mass) gives the baseline to be subtracted from the 1 run to eliminate the non-zero signal because of asymmetries of the apparatus and the third run is performed to calibrate the DSC for cp-measurements [for precise cp-measurements a cp-calibration is needed, which often gives a somewhat different calibration factor from that of say Indium peak area calibration]. In the Q1000 the calibration method is somewhat different (see in the paper you have got from Ali). If you perform the Q1000 calibration as recommended and find a very straight baseline at zero heatflow rate with the empty pan, it is possible to get rather precise cp-values ( about 5% error is possible)  from only one run with the Q1000. For better results you are in need of an empty pan run which you have to subtract from the sample run before you calculate the cp (by division of the corrected heat flow rate by mass and heating rate) which the software does for you. For details see the attached book.

Good luck Gunther Höhne

My colleague spent two years of his Master Thesis by trying to measure Cp using TA Q1000. We use the 3 run procedure with another calorimeter (Setaram mDSC IIIa) and we tried it here too, applying our own script for the evaluation. The results were disastrous, errors amounted up to 100%. As the main problems, I would highlight bad-defined position of the pans (the autosampler does not put them every time on the exactly same spot) and instability of the baseline. The slope correction, which has to be done during the Cp measurements amounts some few percent on mDSCIII, but tens or even more than  one hundred percent on TA Q1000. The slope correction moreover expects, that the baseline is linear, which is definitely not true on TA Q1000.

The only use of TA Q1000 for Cp I would reccomend is the "controlled extrapolation", where we fit the results to the mDSCIIIa in the overlapping range. V. Štejfa et al. / Fluid Phase Equilibria 402 (2015) 18–29, http://dx.doi.org/10.1016/j.fluid.2015.05.026.

Please, do not use TA Q1000 for Cp measurements. Or, if you do, be sure to confirm your calibration by measureing at least two reference compounds and evaluate the possible unceratinity.

We have had a lot of experiences both with the Perkin-Elmer DSC's and TA-Instruments Q1000 and we found an uncertainty of 15% for cp-measurements for the latter. But with precise calibration [i.e. the sapphire disk should be used inside the same type of pan with the same mass, or other cp-standards (always with pan!) should be used] it is possible to get cp-values with an uncertainty of 5 percent. The problem is indeed, that the thermal resistance between sample pan and the supporting area of the cylinder must be the same for all three measurements, i.e. very flat and clean areas and a non-deformed flat pan. If this is the case you may get reliable results for cp (but never better than 5 percent). [There is no doubt, that the power-compensated DSC of Perkin-Elmer is more suitable for cp-measurements, but this does not help if you don't have one at hands]

Thanks Ali Amiri for the document. Thanks Vojtech Stejfa for your advice on the Q1000. I actually used a Perkin Elmer before but it is not giving us good results. 

Thank you Dr. Hohne for the detailed advice. I have your book and will refer to it.

For the Q1000, I had calibrated for temperature with three transitions: 2 for cyclohexane and one for indium. For heat flow calibration I had used indium. I had measured the water values with this calibration and had gotten most results with about 6 % error with literature and a higher error for temperatures below -70C, about 12 %.

I had a question though. I subtracted a baseline run from my sample run to get Cp results for water. I did not do any empty pan runs here. The masses will be different as I will have to use different pans for empty and sample runs.

1. Is it necessary to do an empty pan run here to subtract it from sample run ?

2. Also when you do an empty pan run, you consider both the system baseline as well as the heat flow due to the pans ?

 For values below -70C, the error is about 12%, but about 6% for values above -70C.

I am going to do a heat capacity calibration now with sapphire and see if I get better results.

The errors you have found are reasonable, which tells me that the Q1000 you have is OK. To answer your questions I want to tell you more about the backgraund for the three run methot of cp-determination:

What you need for cp determination is i) the heat flow into the sample (and no other heat flow) and ii) a (temperature-dependent) calibration factor which corrects the measured heat flow rate to the true heat flow rate.

What you get in a heat flux type DSC is a differential heat flow i.e. the difference of the heat flows on sample and reference side. These heat flows are of course the sum of heat flow into the sample, the sample pan and the reference pan and possible assymmetries of the measuring head of the DSC. To get correct cp values you must eliminate al other influences then the heat flow into the pure sample. To eleminate the heatflow into the sample pan you put an identical crucible (same type, exact same mass) but empty on the reference side. The differentail signal then does not contain the pan heat flow anymore, but it contains still unavaoidable assymmertries caused by the apparatus. To eleminate them you perform a second measurement with two empty pans (same type, exact same mass as with the sample run) and same measuring cconditions. This gives the so-called baseline (or zeroline) of the DSC for the specific run procedure. Subtracting this from the sample run gives the heat flow into the sample and only the sample. From this heat flow rate function you may calculate the cp-function by dividing with sample mass and heating rate, in the case of well calibrated DSC. For precise cp measurements you should test the calibration as well. Unfortunately the peak area calibration (e.g. with Indium) yields an calibration factor which is a) true only for enthalpy determination (from other sunstances) in the same temperature region and b) is unfortunately somewhat differrent from the cp-calibrationwhich results in a temperature-dependent calibration function rather than a single factor. To get this precise function you have to determine the heat capacity of a standard (as saphire) with the same procedure writen above in the same temperature region with the same pantype at the same heating rate. The this way determined measured heat capacity has to be compared with the true one (given by the standard) and results in a correction function which has to be applied to your sample result.

The procedure is laborious but unavoidable for better values than around 10%.

I hope this will answer your questions an I could help you.

Good luck Gunther Höhne

Thank you Dr. Hohne! This was very helpful. I read the classical three step procedure in your book. It mentioned that we could bypass the temperature dependent factor in the final formula for Cp, hence we might not need to calculate it directly ?

This is a question of the precision you need and the error you can tolerate. For the Q1000 I feel it is necessary  to determine a temperature dependent calibration factor if you want cp-values with an error lower than 7%. I recommend this because of my experiences with the Q1000 [for Perkin Elmer DSC and the new DSC of TA Instruments it may be avoidable].