3.1. DSC Thermal Curves Obtained from This Study
Some examples of the resulting DSC thermal curves obtained for both n-dotriacontane and n-hexatriacontane are given here in
Figure 3,
Figure 4 and
Figure 5. As was mentioned earlier, a major difficulty for these DSC studies is the weakly observed (low-energy) event for the monoclinic to orthorhombic form of n-dotriacontane. As a result, these analyses were performed at relatively low heating rates and small analyte sample mass in an effort to improve peak resolution in the resulting DSC thermal curves. The heating rate and sample mass for each study is given in the figure caption located above in each of the corresponding figures. Comparatively speaking, a large portion of DSC analyses performed throughout the world routinely employ sample masses ranging from 8 to 12 mg and heating rates of 10 or 20 °C/min.
Figure 3 and
Figure 4 were obtained from both heating and cooling experiments for the two n-paraffins of this study. One will observe that all three of the expected endothermic transitions are exhibited for each of the paraffins, but only two exothermic transitions are present in the cooling curves. This is not always the case for n-C
36H
74 but is the usual observation for the cooling curves obtained for the n-C
32H
66 specimen. The most probable reasons for the absence of this weak exothermic transition in n-dotriacontane are those given by Broadhurst [
1], as mentioned earlier in this study. One will also note that when both heating and cooling curves for these particular sample specimens are compared, there is not a very large difference in the onset temperatures for these observed transitions.
The use of the earlier mentioned approach to achieving increased resolution of the two solid–solid transitions (by decreasing the heating rate of the DSC experiment) also leads to two important consequences. First, the peak height of the DSC thermal curve is decreased when the heating rate is decreased. Thus, there is a decrease in the ability of the instrument to observe this weak transition. Second, the temperature calibration of the instrument can also be affected by change in heating rates. A decrease in heating rate, most often, will lead to a lower value obtained for the temperatures of the measured transition. On the other hand, an increase in heating rate leads to a larger value of the onset temperatures for the transition being assigned. Therefore, to reduce this effect, the DSC instrument should be temperature-calibrated using the same heating rate as that used in this experimental DSC analytical procedure.
The assignment of temperature for transitions observed by DSC, for this type of compound, is based on the extrapolated onset temperature (To) obtained from the thermal curve. It is not assigned from the observed DSC peak maximum in the experimentally obtained DSC thermal curve. The To is assigned from the thermal curve, by the thermal analysis software, as the temperature of intersection of the extrapolated pre-transition baseline and a tangent line drawn from the point of maximum slope of the leading edge of the measured DSC peak.
Figure 5 displays the cooling curve, which was given as a part of
Figure 4. This figure also shows the data-handling procedure for obtaining the extrapolated onset temperature for the transitions in a DSC cooling curve. In this case, the extrapolated tangent line is made from the high-temperature side of the exothermic (downward) peaks. In the ideal case (complete thermal reversibility), the extrapolated onset temperatures obtained (from both heating and cooling experiments) should be exactly the same when the heating and cooling rates used are the same. This is seldom the case due to several reasons. Supercooling is one of the possible reasons, while another is that the instrument is usually temperature-calibrated using the heating mode of operation. For these reasons, the cooling curve generally gives a slightly lower onset temperature. When large differences between the onset temperatures in the first heating and cooling runs for the same sample specimen are observed, this is usually an indication of lack of purity of the sample specimen. For this reason, it is recommended by the authors of this paper that the first heating runs be used only when assigning temperatures of transitions for n-paraffins by DSC.
3.2. DSC Results for the n-Paraffins of This Study
The experimental DSC results obtained in our laboratory for the two n-paraffins of interest represent the work of several groups of two students each. The values for the experimentally observed thermal transitions, in the DSC thermal curves, were averaged and are reported in
Table 1. The procedure for the experimental DSC work was described earlier in this paper. It was found that the best results for n-dotriacontane were obtained with heating rates as low as one or two degrees per minute in order to obtain better separation between the monoclinic to orthorhombic transition and the subsequent orthorhombic to hexagonal (rotational) transformation in the n-paraffin. One can see from the listed temperatures that the average temperature difference between these two thermal events, obtained by the participants in this study, is slightly more than 1 degree Celsius in n-C
36H
74 and is less than 1.5 degrees Celsius of separation in the DSC thermal curves for n-C
32H
66. As was reported earlier, sample masses for the analytical sample size were often as small as 1.120 mg when obtaining DSC data for the n-C
32H
66 sample. The desired DSC thermal curves for obtaining the temperature and calorimetric data would look very similar to that obtained for n-hexatriacontane shown in
Figure 5.
In the study of the n-dotriacontane specimen, some of the participating analysts did not report a value for the monoclinic to orthorhombic transition temperature (T
1) nor the enthalpy change for this endothermic event. This was primarily due to poor separation of this transition peak in the DSC thermal curves, which were obtained in their experimental efforts. Some examples of this type of overlap are shown in the heating curves of
Figure 3, as well as in the lower thermal curve of
Figure 4. However, in many of these cases, the onset temperature (T
o) for this transition can be approximated by the data-handling software using a tangent line drawn to the leading edge of the overlapped endothermic peak. When this was done here, the assigned T
o values were averaged, and a mean value of 63.71 °C was included in
Table 1 for the endothermic transition temperature T
1. This method of approximation of T
1 from the overlapped peaks would seem to always give a slightly lower value depending on the degree of overlap of the two endothermic peaks.
The assignment of the enthalpy change (∆H1) for this overlapped (monoclinic to orthorhombic) peak in n-dotriacontane is more difficult. For this case of overlapped peaks, the best approximation of this value is obtained by using the following subtraction (or “difference”) method:
First, the ΔH2 value (for the orthorhombic to hexagonal transition) may be assigned from the absolute value of the exothermic transition observed in the DSC cooling curve (hexagonal to orthorhombic transition).
Mathematically, ΔH2 of the endothermic heating value is equal to −ΔH2 of the exothermic transition. That is, (ΔH2)heating = −(ΔH2)cooling.
Now, the total energy observed for the overlapped endothermic peaks, ΔHTotal, in the DSC thermal curve is the sum of ΔH1 and ΔH2. That is, ΔHTotal = ΔH1 + ΔH2.
Therefore,
ΔH1 is obtained by the subtraction
When this is carried out, the mean of the assigned values of
∆H2 (
54.28 J/g) obtained from the cooling curves is subtracted from the average value obtained for the total change in enthalpy of the overlapped peaks in the heating thermal curves (
71.66 J/g). The total change in enthalpy for the overlapped peaks is assigned by use of the data-handling software of the DSC, using the total area of the combination peak formed by the two endothermic polymorphic transitions. When this subtraction procedure was completed here, an enthalpy of transition value of
17.38 J/g was obtained for the monoclinic to orthorhombic transition of the n-C
32H
66. One will note that cooling curves for n-C
32H
66, such as that shown in
Figure 5, which lead to the transition energy (
∆H2) value, are needed in this case. The value obtained for
∆H1, using the subtraction method described here, was therefore not obtained by a direct assignment and is probably less accurate than other thermodynamic values listed in
Table 1.
Table 2, given below, restates the results given in
Table 1 in different units. The temperatures of transition were converted to degrees Kelvin, and the heats of transition were changed to molar quantities of kilojoules/mole (kJ/mol). These values were then used to calculate the molar entropy change (∆S) using the following equation:
Use of Equation (2) gives the value for the change in entropy of the particular transition in units of Joules per mole per degree Kelvin (J mol
−1 K
−1). These units are often abbreviated as Entropy Units (eu), and they are employed as this manuscript continues. Unfortunately, there are no reference standards for physical properties of n-C
32H
66 or n-C
36H
74 available. The values given in the NIST WebBook SRD 69 [
19] are published data found in the literature by various laboratories. When the DSC results given in
Table 2, obtained by the student lab analysts, are compared to available sources of published data [
19,
20,
21,
22], one will see that reasonable agreement was observed for the average values for this study. In a few cases, the temperatures of transition were lower by as much as two degrees Kelvin. On the contrary, the student values matched the literature values for the transition temperature of the orthorhombic to hexagonal (rotational transition) by an exceptionally small 0.1 to 0.2 K margin in both of the n-paraffins studied. It will also be seen that the average value experimentally obtained for the ∆H value for the monoclinic to orthorhombic transition in the n-hexatriacontane (19.57 J/g or 9.92 kJ/mol) was exactly the same as that given by the data referenced by NIST [
19].
The comparison of the molar quantities of both the ∆H and ∆S values for the solid to liquid (fusion) transition, in each of the two paraffins, is interesting. The experimental values obtained are smaller than the NIST WebBook [
19] listing by very similar amounts (ca. 4–5%). The ∆S values in
Table 2 are 211.6 eu for n-dotriacontane and 243.4 eu for n-hexatriacontane. When compared to the NIST WebBook values of 222.93 and 254.5 eu, respectively, an absolute difference of 11.3 eu (or a relative difference of 5.06%) is obtained for the n-dotriacontane specimen of this study. When the same is done for the n-hexatriacontane value, an absolute difference of 11.1 eu or a 4.36% relative difference is the result. These differences in ∆S values are primarily due to the fact that the heat of fusion (∆H
m) values, obtained for each of the paraffin specimens, were lower than the values listed by NIST WebBook by the amounts of 5.78% (n-dotriacontane) and 4.84% (n-hexatriacontane).
Comparisons of results for n-paraffins, such as those given above, are affected by both the purity and thermal history of the specimens being compared. Without detailed information, it is difficult to say that one set of data is better than another. It must be remembered that, in this case, the NIST values are not those of NIST reference standards. They are values previously published in the literature by different authors in separate laboratories with different n-paraffin obtained from many different sources. This issue could be overcome, to some degree, by the establishment of a reference standard for physical properties of these long-chain normal paraffins.
It should also be mentioned that a specimen of n-C
32H
66, commonly employed as a reference for retention time (or elution volume) assignments in gas chromatographic analyses, was recently obtained from Separation Systems, Inc., Gulf Breeze, Fl. It is a purified sample of Sigma Chemical D-223107, Lot # 05718JM. This sample was analyzed chromatographically at the Separation Systems Laboratory and assigned a purity value of 99.7%. When this specimen of n-C
32H
66 was analyzed by DSC in our laboratory, a melting temperature of 67.98 °C and a heat of fusion of 156.01 J/g was obtained. When these values are compared with those listed for the n-C
32H
66 of our study in
Table 1, the heats of fusions are essentially identical, and the melting temperatures differ by only 0.25 °C. This strongly suggests that values reported in this study for the n-C
32H
66 specimen do not exhibit colligative property effects, such as melting point lowering [
23], due to impurities.