At the recent ‘Scale Up of Chemical Processes Conference’ in Baveno, Italy, John Snoonian of Sunovion discussed the development and scale up of a manufacturing route to make a 5HT4 partial agonist containing a 1,2,4-oxadiazole ring. During route scouting studies on the synthesis of the benzopyrazole an interesting solvent effect was observed – see Scheme 2. As the ketone was the desired product this was manufactured via ortho-metalation of 1,2-difluorobenzene followed by reaction with isobutyryl chloride.
Scheme 2: Solvent effect in reaction between iPrMgCl and 2,3-difluorobenzonitrile.
There were other solvent issues later in the synthesis during the formation of the oxadiazole ring. The first occurred during the reaction of the pyrazole with cyanogen bromide (see Scheme 3).
Scheme 3: Reaction with cyanogen bromide (BrCN)
Under the original reaction conditions solid cyanogen bromide was added to the pyrazole in DMF using Cs2CO3 as base which was clearly not going to be suitable for scale up. Cyanogen bromide is not stable in DMF so it was purchased as a solution in acetonitrile so acetonitrile was used as reaction solvent as well, which avoided the formation of a guanidine impurity when BrCN in CH3CN was reacted with the pyrazole in DMF (the base was changed to Et3N as well).
The nitrile from the cyanogen bromide reaction was reacted with hydroxylamine to give the N-hydroxy-amidine intermediate which was cyclised with the acid shown to produce the oxadiazole (see Scheme 4).
Scheme 4: Oxadiazole formation
Coupling of the N-hydroxyamidine and the carboxylic acid with T3P as coupling reagent rapidly formed the intermediate which cyclised slowly on heating to 100°0C. Originally T3P was purchased as a solution in EtOAc, but this was changed to T3P in toluene for scale up because of the work up conditions. The reaction mixture is quenched in to water to destroy excess T3P. The aqueous phase is now pH 1-2 and contains the product. The pH is adjusted to 13 to return the product to the organic phase, but EtOAc and strongly basic aqueous phases are not a good option when reactions are scaled up because of the extra time required for phase separation and the likelihood of ester hydrolysis.