A common calorimetry question I am asked is how a process can be red flagged for scale-up early on based on a small volume experiment. I just came across an example that illustrates the importance of:
- High Instantaneous Heat Release
- Delayed Initiation
- Heat Accumulation
The reaction was conducted on a 50-100 mL scale in an EasyMax equipped with heat flow calorimetry. It was a two-step process (scheme 1) involving the double deprotonation of a benzyl derivative using a strong alkyl lithium base, followed by oxidation into the corresponding benzyl alcohol using oxygen (one atmosphere). The presence of two equivalents of base is needed due to the presence of a more labile proton on the molecule.
Although the deprotonation step using a strong lithium base is, as expected, strongly exothermic (we found approximately 400 kJ/mol), its safety profile is perfectly managed using an adequate dosing rate. This was easily verified by interrupting dosing and seeing the heat signal dropping to zero. The reaction is fast and instantaneous, which allows for moderating the exotherm over a time range that matches the cooling capacity of small and large scale vessels. Just slowly increase the dosing rate at the very beginning to avoid the peak spike we observed (>15 W) – which was due to limited mixing. Interestingly, the initial high level of energy release (> 150 W/l) resulted in just a couple of ˚C increase in reaction temperature – which demonstrates the efficiency, and importance, of good temperature control in this particular instrument.
Further observations and results from analyzing the heat signal:
- The first deprotonation step was more exothermic than the second one: 6 W versus 2-3 W
- Gradual dilution when dosing caused the instantaneous heat release to decrease: from 6 W to 5.5 W during the first deprotonation to > 2 W down to < 2 W during the second deprotonation phase
- The instantaneous T adiabatic was approximately 50 ˚C. It is significant but would not cause the process temperature to exceed the boiling point of the solvents (THF 66 ˚C, and hexane 68 ˚C)
The second step consisted in exposing the reaction mixture to one atmosphere oxygen. As illustrated on figure 2., upon exposure to O2, the reaction mixture started releasing a small 3 to 3.5 W constant energy over approximately 15 minutes followed by a sudden and hazardous heat increase up to 80 W, or 800 W/l. The violence, the delay in onset, and lack of control of heat release makes this reaction especially hazardous to scale even though its specific enthalpy (400 kJ/mol), and instantaneous T adiabatic (60 ˚C) was similar to the previous deprotonation step. The relative complexity of the heat profile can probably be explained by a multi-step reaction mechanism. If this process is to be scaled, an effort toward better and easier heat management should be conducted, e.g. using additives as previously described1.
As also previously described by D.J. am Ende and colleagues at Pfizer, the unsafe nature of a chemical process is not as much related to specific enthalpy as it is to a delayed onset of significant reaction heat, resulting in high heat accumulation, and lack of control.
More generally, the availability of process heat information early-on during a development project from small scale experiments (< 100 mL) ensures “right-first-time” efficiency towards the development of safe, reliable and cost-efficient processes in the pharmaceutical, fine chemical, and specialty chemical industries.
1 Terrence J. Connolly,* Eric C. secure web browser . Hansen, and Michael F. MacEwan; Organic Process Research & Development 2010, 14, 466–469
2 D. H. Brown Ripin, G. A. Weisenburger, D. J. am Ende, D. R. Bill, P. J. Clifford, C. N. Meltz, J. E. Phillips; Pfizer Global Research; Org. Process Res. Dev. 2007, 11, 762-765