When it comes to thermal hazard assessment, differential scanning calorimetry (DSC) takes a special place in the landscape of possible testing methods. Both for rapid screening of a wide variety of samples, and for the gathering of the necessary high-quality data for a detailed kinetic study of decomposition reactions, DSC is usually the instrument of choice. It should therefore come as no surprise that many scientists working both in industry and academia use this technique for safety analysis, but this widespread use may come with a downside as well. Fact is that a uniform testing method is often lacking, and even worse, due to lack of expertise in this field sometimes wrong and possibly dangerous conclusions are drawn.
In an excellent paper in Angewandte Chemie, researchers from both Imperial College London and GSK made an analysis of the reported DSC safety analysis in a number of chemical peer reviewed journals for the year 2019. A first conclusion is that when DSC data are reported, the exact experimental details needed for comparison with other available data in literature is often lacking. This is of concern, since differing experimental conditions may lead to very differing and even contradicting conclusions, as will be explained underneath. The two most important parameters that should always be mentioned are without any doubt the type of crucible and the heating rate used.
DSC is widely used for the determination of the thermal properties of materials, like the glass transition temperature of polymers. In this field of analysis, aluminium DSC crucibles are most often used, either open, with a pierced lid, or crimp capped. Although these latter are sometimes called “hermetic” pans, they are by no means pressure resistant. In reality, they will only withstand a pressure of a few bar at best, which is certainly insufficient for safety hazards testing. Many samples will decompose at higher temperatures with the generation of gaseous byproducts. When the thermal measurement is performed in an open test cell, this gas release will often lead to an endothermic signal, obscuring the underlying exothermic decomposition. Also when measuring liquid samples, solutions or wet products, evaporation will only lead to measuring an endothermic signal, making an assessment of the thermal stability of the compound itself impossible. Therefore, truly high-pressure crucibles should always be used, being available as either stainless steel or gold-plated alternatives, both as single use (with high pressure sealing) or reusable (with screw cap). As a more economical alternative, sealed glass capillaries can be used as DSC crucibles as well. Using these types of crucibles, evaporative heat losses can be excluded, leading to more reliable results. In order to make sure the crucibles were properly sealed, it is also good practice to weigh the crucible before and after the experiment, with any significant weight loss pointing towards a leak during the run. The importance of using the correct crucible for DSC measurements was demonstrated by 2 publications in the Journal of Organic Chemistry. In a first paper, Xie et al reported on the synthesis of a novel diazo transfer reagent, 2-azido-4,6-dimethoxy-1,3,5-triazine (ADT). Quite surprisingly, they reported that this azide functionalized triazine was perfectly stable at elevated temperatures, only showing 2 endothermic events in the DSC. The authors attributed the first one to the melting of the compound, and the second one to an endothermic decomposition. No experimental details were provided, but these measurements were most probably not conducted in appropriately sealed crucibles, and retesting of this compound by Green et al in a high-pressure crucible did reveal a very high exothermic decomposition energy of more than 1000 J/g. This fact, that depending on the DSC crucible being used a compound can be either classified as an “intrinsically safe” compound, versus a “highly energetic” (if not explosive) compound should be understood by anyone making safety assessments based on DSC data. Moreover, it should be a wakeup call to all authors and reviewers to make sure that whenever DSC data are reported for safety evaluation, the experimental procedure being used should be reported in sufficient detail to enable reproduction of the results.
A second vital experimental detail when it comes to DSC data is the heating rate being used. It is well known that the peak (or the onset) of an exothermic event in a DSC experiment will shift to higher temperatures when using a faster heating rate. This effect can be quite significant, as can be seen from the figure underneath. In this illustrative example, the overlay of the DSC run of the same sample with 5 different heating rates is shown, in this case 8, 4, 2, 1 and 0,5 °C/min. In this example, the peak at the highest heating rate is situated about 30 °C higher than the one at the lowest heating rate. This is another illustration of the influence of the experimental details on the observed thermal stability in a DSC experiment. If an author claims: “according to DSC, the onset of decomposition is situated at 150 °C”, this claim is of little significance without clarifying the heating rate used. Moreover, when an accurate determination of the thermal stability of a compound is needed, the slowest heating rate will yield the lowest onset temperature, obviously at the expense of a longer experimental time. Whereas in material science heating rates of 10 °C/min or higher are common, it is advisable to use slower heating rates for safety testing. In our experience, using a scan rate of 2,5 to 5 °C/min offers a good compromise between accuracy and experimental time.
All in all, we warmly encourage reporting experimental thermal stability data based on DSC experiments, but care is needed to use the correct type of crucible, and to report the experimental conditions in sufficient detail.
About the Author:
Dr Wim Dermaut graduated from the University of Antwerp where he obtained his Ph.D. in physical organic chemistry in 2002. He then joined Janssen Pharmaceutica where he worked in the process safety lab, focusing mainly on reaction calorimetry, adiabatic calorimetry and kinetic modelling. In 2011 he moved to the Chemical Process Development group of Agfa, where he is responsible for general process development projects, with a strong focus on process safety. His role also includes technology scouting, acting as a project leader in consortium projects on flow chemistry and ultrasound technology. Wim is also a guest professor at the University of Antwerp teaching process safety to masters curricula in chemistry and chemical engineering. In 2022, Wim will be delivering an ONLINE training course entitled ‘Chemical & Process Safety for Development Chemists‘ where he will impart much of his experience to anyone interested in learing more about process safety.
 Green et al, Angew. Chem. Int. Ed, 2020, 59, 15798-15802
 Sheng et al, Org. Process Res. Dev., 2019, 23, 2200-2209
 Xie et al, J. Org. Chem, 2018, 83, 10916-10921
 Green et al, J. OIrg. Chem, 2019, 84, 5893-5898