AmericanChemicalSociety.com

Fluorine-containing organic molecules serve as critical pharmaceuticals, agrochemicals, performance materials, and radiotracers for imaging processes. Unfortunately, the availability of methods for the construction of C–F bonds severely limits the discovery and synthesis of these important targets. Particularly challenging in this context is the development of methods for the mild and selective incorporation of nucleophilic fluoride, the naturally occurring and least expensive form of fluorine, into organic targets. Use of anhydrous HF is plagued by difficulties with its safety and handling, metal fluorides suffer from sparing solubility in organic solvents, and reagents with enhanced solubility like tetraalkylammonium fluorides display high Brznsted basicity.

Instead, we hypothesized that commercial benzoyl fluoride could serve as bench-stable, organic-soluble latent fluoride source for catalytic C–F bond-forming reactions. In particular and in analogy to well-developed Lewis base-catalyzed acyl transfer, we proposed that a Lewis base catalyst could liberate HF from the combination of benzoyl fluoride and an alcohol additive. Because the uncatalyzed solvolysis of acyl fluorides is slow, we expected that (1) a catalytic procedure could overcome background reaction and (2) the protocol would be amenable to an enantioselective process through the intermediacy of chiral Lewis acid/amine co-catalysis. We first chose to evaluate this proposal in the context of epoxide fluorination, which provides access to synthetically versatile b-fluoro alcohols.  In the event, we determined that commercial (–)-tetramisole and (R,R)-(salen)Co serve as co-catalysts for desymmetrizations of five- through eight-membered epoxides, affording b-fluoro alcohols in up to 95% ee (Scheme 1).

We have also explored terminal epoxides as substrates for fluoride ring-opening kinetic resolutions and found that a broad range of substrates undergo reaction with extremely high regioselectivity and k_rel values (Scheme 2). In these reactions, 1,5-diazabicyclo[4.3.0]non-5-ene (DBN) was identified as the optimal amine co-catalyst. Notably, the co-catalytic fluorinations proceed at room temperature in standard glassware and do not require exclusion of moisture or air. A consequence of these mild conditions is that substrates containing traditionally fluoride-labile substituents such as silyl ethers are stable to the reaction conditions and undergo C–F bond formation with good reaction efficiency.

The fluorinations with (–)-tetramisole and (salen)Co(II) in Scheme 1 exhibit a pronounced matched/mis-matched effect on rate and enantioselectivity (Table 1). Although amine/Lewis acid co-catalysis has been described, a cooperative effect on enantioselectivity has little precedent. We are currently undertaking kinetic and non-linear effect studies in order to elucidate the role of the two catalysts in the catalytic fluorinations. We anticipate that these studies will lead to improvements to the catalytic methodology and to the development of new methodologies for the incorporation of fluorine into organic substrates. Along these lines, we have begun to undertake studies aimed at the development of an enantioselective fluoride ring-opening reaction of aziridines. The Doctoral New Investigator grant from the ACS PRF has been instrumental in jump-starting this program and we are extremely grateful for the support.