One of the most significant benefits from fast scanning is the resultant increase in sensitivity. The faster the scan rate, the bigger the size of the peak or step that is measured, which is evident from the melting of indium and also the increase in the size of the Tg measurement in Figure 1. The increased sensitivity has many benefits, allowing small sample masses to be used, small transitions to be easily observed, and potentially increased accuracy for measurement of specific heat of materials. The reason for the increased sensitivity is that energy flows more quickly at the higher scan rates. The amount of energy involved remains the same but the time during which it flows is reduced as the scan rate is increased, so the y-axis response of the DSC records the energy flow increases with scan rate.

Figure 1. Increasing scan rate increases the size of the glass transition of PMMA. The increased sensitivity makes it possible to measure transitions that are difficult to see at low scan rates.

Sometimes there is a requirement to observe the changes that occur in a sample when it is heated slowly, for example when measuring cure or recrystallisation processes. On other occasions, these changes are undesirable and prevent measurement of sample properties at higher temperatures. Yet at the slow scan rates normally employed changes can easily occur, and with many polymer systems changes may occur without the analyst realising it. This can be seen very clearly with the melting of thermoplastic polymers such as polypropylene. The crystalline structure of polypropylene depends upon how it is cooled and the time given for crystals to develop. Upon fast cooling (typical of manufacturing processes) there is not enough time for large crystallites to form and a range of small less stable crystals will result. If then heated slowly, not only do the small less perfect crystals melt at lower temperatures, but larger more stable crystals develop from the molten material; the well-known process of annealing. However, this competing energetic process may lead to little if any net energy flow; in other words little if anything is observed in a DSC trace. The final melting profile is then that of the annealed material, which may be significantly different from that of the original un-annealed material placed in the pan. In this case, we have not measured the properties of the material put into the analyser, but those of the sample after it has changed during the slow heating rate. The actual material properties may be obtained by heating more quickly, though the scan rate required will vary from sample to sample depending upon the rate needed to prevent the sample changing before it is measured. An example is shown from the melt of polypropylene in Figure 2. The melting point and melting profile obtained at slow rates differ significantly from that obtained at 150◦C/min since the material is annealed during the slow scan.

Figure 2. In the example above, polypropylene heated at 150◦C/min melts at a lower temperature and with a broader profile than a sample heated at 10◦C/min. Heating faster prevents any annealing processes which will displace the melting profile to higher temperatures.

Many materials undergo changes in their crystalline structure when heated, including polymers, pharmaceuticals and foods. In some materials, particularly pharmaceutical materials, different crystal structures may exist (polymorphism), and on heating some materials can undergo changes from one form to another. Different changes may occur at different heating rates, but at faster rates such changes are less likely to happen due to the lack of time for crystal changes to occur. Thus, at high rates the original material is more likely to be characterised, just as for polypropylene referred to above.

If a material has been slowly cooled from the melt then the likelihood of changes occurring on slow heating is reduced. However, many materials are fast cooled, particularly in industrial processes, so that changes in structure under slow heating are likely to occur with many materials. This is particularly true of thermoplastic polymers where the crystal structure is likely to change during slow heating, the polymer having been annealed. Therefore, to prevent annealing and to measure the true properties of the material put into the DSC a fast heating rate is desirable, the actual scan rate being dependent on the kinetics of the processes involved.

Sometimes two events can occur over the same temperature range, often a large event obscuring a smaller event, or sometimes events are just poorly resolved. If the events are subject to different kinetics then altering the scan rate will allow separation on a temperature scale. This has been found to be effective in MTDSC where slow scan rates are employed but is also true where faster heating rates are used. One example is the loss of moisture from a damp material placed in a vented pan. In the case of a polyamide, Figure 3a, which normally contains some moisture, the glass transition of the amorphous material is completely obscured by the loss of the moisture. At faster heating rates not only is the Tg more obvious but the moisture loss, which is a slow process, is also displaced to much higher temperatures allowing the Tg to be clearly measured, Figure 3b. This effect is also of advantage with materials which are prone to decomposition. When fast scan rates are used the decomposition process, which is normally slow to begin with, is displaced to higher temperatures allowing measurement of transitions that would previously be obscured by the decomposition process. This can permit clear melting profiles and measurement of heats of fusion materials from materials which would otherwise have decomposed.

If adopting this method, great care must be taken to prevent contamination of the analyser. Frequent checks and cleaning may be required, and for this reason small samples should be used together with high purge gas rates to remove volatiles.

Figure 3 (a) Slow scan (10◦C/min) of a polyamide sample. The Tg of this material is obscured by the moisture loss which shows as the broad endotherm. At faster rates the Tg can be observed; see part b. (b) Moisture loss from a polyamide is displaced to higher temperatures as the scan rate is increased to 500◦C/min revealing the Tg of the polyamide, the step beginning around 50◦C. Significantly, this is the true Tg of the material plasticised by the water content the loss of which is seen as the broad endotherm above Tg.

Speed of analysis is not the main reason for using HyperDSC, but there is always a benefit from increased speed of analysis, particularly when the improvement is very marked. In a situation where many samples need to be run for screening or other purposes the fast analysis times using HyperDSC are a significant benefit. Where 20–30 min are required at slow rates, 2–3 min (or less) may be all that is required at high rates. The curves presented in this chapter may look similar to those taken at slow rates yet many will have taken less than a minute to produce, compared to the tens of minutes for a typical scan, and this gives a real advantage to the technique. In industry, there is always a push to reduce times and improve throughput and this is achieved by fast scanning DSC. There may be questions concerning the effect of increased scan speed on the data measured and some of these concerns will be addressed in subsequent sections, particularly with regard to resolution, and onset temperatures of melting transitions and glass transitions. Note though that with regard to resolution one of the key aspects is the use of helium (or other high conductivity gas) as a sample purge, since it has a higher thermal conductivity and gives much faster heat transfer to the sample than a system running under nitrogen or air.