Solvents are Chemicals Too: A Topical Tip from OPRD Barcelona
I recognise that we are all aware of this, but since we still see incidents of solvent-related impurities these two examples serve as reminders. We should all be vigilant with our solvent selection, because like all other reagents, they contain impurities that can be reactive and of course they have their own reactivity.
First, an ester exchange example. Philippe Vayron from Sanofi-Aventis described the formation of some isopropyl ester when using IPA for final stage crystallisation. The process required the use of IPA for control of physical form so the issue was how to alter the process. They established that just 0.1mol% of residual HCl was sufficient to catalyse the transesterification. The crude SR58611A had been isolated from ethyl acetate following a BOC-deprotection with HCl/EtOAc. It appeared that the crystals in the crude product trapped some residual HCl, so upon recrystallisation from IPA to both improve purity and establish the correct morphic form, the transesterification would take place.
The chosen solution was to add some ethanol at the end of the BOC deprotection, then strip back out to regenerate the mixture in ethyl acetate; this served to remove excess HCl from the system and provide the HCl salt without significant trapped HCl. The IPA recrystallisation then performed perfectly and was executed on 450kg scale.
The message: transesterification is always an issue to be contemplated as we design our processes, but careful examination of the underlying issues is invaluable, especially if there is no other solvent option owing to solid state (or regulatory) requirements.
Next came DMF and a gentle reminder of the side-reactions one sees due to the frequent presence of dimethylamine. In this case Christopher Nichols of GSK described a case where DMF was used in the discovery chemistry route in the formation of an oxadiazole from an acid chloride. The development team soon identified the dimethylamide as an impurity.
The chosen solution was to replace DMF/triethylamine by toluene/pyridine which both gave the required reaction and avoided problems associated with the formation of thick slurries seen in other solvents.
Finally on the subject of solvents, we should all remind ourselves that solvents (as well as their impurities) are not inert. Ian Moses from Pfizer reminded us all about the reactivity of DCM. In this case the development team were well aware of the potential issue and checked the literature, finding that piperidine reacts with DCM with a t1/2 of about 15h (and triethylamine with a t1/2 of about 580h). By analogy with their substrate it was clear DCM needed to be avoided. In this case they used 2-methyl THF.
The general message for me is that DCM will certainly react with competitive rates with the more nucleophilic amines such as piperidines and pyrrolidines and should be avoided. I have worked on a project wherein we saw quaterinisation of a tertiary amine by reaction with DCM, so its use should be carefully considered and appropriate stability studies undertaken to mimic plant durations.