Unusual Sample Properties as an Origin of Artifacts

Sometimes, a DSC curve exhibits unexpected and unusual thermal events that cannot readily be explained. On the one hand, they might turn out to be artifacts, that is, thermal events that do not directly relate to changes in the physical properties of the sample. Such artifacts usually have their origin in the ambient operating conditions of the instrument. Occasionally, however, artifacts are caused by the sample itself. Since artifacts often overlap with real thermal effects, or can be wrongly interpreted, they must be identified and if possible eliminated. The following example describes an artifact caused by the unusual physical properties of a cured polymer and how the artifact was removed by suitable sample preparation.

Figure 1 shows several cooling curves (100 °C to −60 °C) of a polymer that had been cured at 120 °C in a 40-µL standard crucible. First, a glass transition occurs at an onset temperature of about 53 °C. Afterward, the DSC curve exhibits an irregular slope with several non-reproducible artifacts superimposed on it. This makes it difficult to evaluate the glass transition and any thermal effects that follow.

The Origin of the Artifacts

In addition to the artifacts described above, it was also observed that the sample crucible moved slightly sideways from its original position on the DSC sensor during cooling. The reference crucible, however, remained in exactly the same position. We were able to prove with certainty that the artifacts originated from the sample itself. First, empty aluminum crucibles were measured under identical experimental conditions. Then a sample of the original uncured material was cooled directly from 30 °C to −60 °C, that is, without the heating and curing step. In both cases, the signal artifacts and movement of the sample crucible were not observed. These experiments proved that the curing of the polymer and the cooling run afterward were responsible for the crucible movement and the artifacts.

This can be explained as follows. The curing process at 120 °C transforms the originally powdered uncured polymer into a cross-linked glassy material that adheres strongly to the bottom of the aluminum crucible. During the cooling measurement that follows, the polymer undergoes a glass transition at about 53 °C. This creates internal stresses in the region of contact between the sample and the aluminum crucible due to differences in the thermal expansion coefficients of the two materials. Under these conditions, the sample crucible undergoes elastic deformation, which leads to a change in the heat transfer between the crucible and sensor. This effect contributes to the unusual shape of the DSC curve. The relaxation of stress leads to movement of the crucible and thus to artifacts on the DSC signal.

The photograph of the sample in Figure 2 shows that induced stresses really do occur and relax during cooling. In this case, stress relaxation occurs through the formation of cracks in the sample material when the stresses exceed the stiffness of the polymer.

How to Avoid Sample Artifacts

The following measures prevent the movement of the crucible during a DSC measurement and hence the occurrence of artifacts such as those described above:

1. Use small samples in order to reduce the contact surface between the sample and crucible. Figure 3 shows that when the sample mass is reduced from 6 mg to 1 mg, crucible movement and artifacts no longer occur. This is due to the reduction in the surface stress between the aluminum crucible and sample.
   
2. Avoid direct contact between the sample and crucible, for example by wrapping the sample in a piece of aluminum foil. Surface stresses between the aluminum foil and sample will occur, but they are not transferred to the aluminum crucible. Figure 3 shows that artifacts are no longer observed on the DSC curve if this method is used.

Conclusions

Crucible movement on the DSC sensor and the sudden spikes on the DSC curve can sometimes be caused by the sample itself. The curing reactions of polymers or the crystallization of polymers and oils may lead to the sample adhering strongly to the bottom of the crucible. This generates internal stresses during cooling and especially below the glass transition of the material due to differences in the thermal expansion coefficients of the sample and crucible material. Stress relaxation can lead to the crucible moving slightly on the sensor, which produces artifacts on the DSC curve. The example above shows that this potential problem can be overcome with appropriate sample preparation and in particular by using smaller sample masses.

_{Unusual Sample Properties as an Origin of Artifacts | Thermal Analysis Application No. UC 236 | Application published in METTLER TOLEDO Thermal Analysis UserCom 23}