Purification by Reaction: Another Example of Removing a Serious Impurity

KiloMentor has suggested in other blog entries that process chemists should think ahead of time about possible impurities that could contaminate their products that come particularly from impurities in starting material when these impurities are unlikely to be removed by the downstream reactions or process step separations. 

Such an example has appeared in a US patent application: US2005/0250961A1, authored by Muddasani Pulla Reddy is such an application that has since lapsed. The invention taught methods for performing the Friedel-Craft reaction between glutaric anhydride and fluorobenzene catalyzed by aluminum chloride, characterized in that even when standard commercial fluorobenzene containing 300-700 ppm of benzene is used as starting material, the isolated product would contain less than 0.05% of the nonfluorinated, 4-benzoylbutyric acid as an impurity in the desired 4-(4-fluorobenzoyl)butyric acid.  This 4-(4-fluorobenzoyl)butyric acid is an important intermediate for some syntheses of the medicine ezetimibe.

The full disclosure does not explain how or why the procedures taught decrease the amount of the defluorinated impurity. The document, in fact, does not provide a comparative example of 4-(4-fluorobenzoyl)butyric acid made according to the prior art literature that shows clearly, by contrast, the improvement taught. All that is said is that benzene is about 5 times more reactive than fluorobenzene to Friedel-Craft acylation.  The teaching does not describe how the benzene reacts, what product the benzene forms or at what stage in the process the impurity(ies) is(are) removed.

The requirement of a halogenated solvent such as methylene chloride or dichloroethane suggests to KiloMentor that the halogenated alkane solvent may react selectively with the benzene producing products that are more easily removed in one or more of the isolation steps. It appears that in each of the patent document’s specific examples one-half the fluorobenzene is mixed with aluminum chloride and halogenated solvent before adding glutaric anhydride. In this mixing period before glutaric acid is introduced, the benzene impurity could have reacted with chloroalkane solvent catalyzed by the aluminum chloride.