2-Fluoroenones via an Umpolung Morita–Baylis–Hillman Reaction of Enones

Several methods have been reported for the formation of 2-fluoroenones. However, all these methods involve laborious multiple-step sequences with resulting low overall yields. In this paper, we report the first formal enone-α-H to F substitution, leading to 2-fluoroenones in a single step from ubiquitous enones in 63–90% yield. The reaction is applicable to a wide range of aromatic and alkenyl enones and is carried out at room temperature using HF-pyridine complex as the fluoride source. Mechanistic investigations support that the reaction takes place through a rare umpolung Morita–Baylis–Hillman-type mechanism.

T he proliferation of approved pharmaceuticals incorporating fluorine atoms 1−3 has led to the development of numerous synthetic methods for the synthesis of fluorinated organic compounds. Of special relevance for this work is the ability of vinyl fluorides to act as enolate bioisosteres. 4 2-Fluoroenones are interesting compounds with a large potential as building blocks for fluorinated medicines given their reactivity. 2-Fluoroenones undergo asymmetric hydrogenation to give chiral 2-fluoroketones, 5 asymmetric Diels− Alder cycloaddition reactions, 6,7 and participate readily in Meerwein arylations 8 as well as conjugate addition reactions with a large variety of nucleophiles. 9 Pannecoucke 10 and Hoveyda 11 demonstrated their conversion into biologically active 2-fluoro allylic amines.
Unfortunately, all reported syntheses of 2-fluoroenones require two-to five-step procedures from commercially available starting materials.
Even the shortest route of only two steps involves fluorination of enones with elementary fluorine followed by elimination and proceeds in a yield of 42−68%. 12 Alternatively, Horner−Wadsworth−Emmons (HWE) reactions of aldehydes may be done with 2-fluoro-3-ketophosphonates. 13 However, the preparation of the reagents for this HWE reaction require three to four steps. 13−15 Another option is to carry out a HWE reaction between aldehydes and commercially available 2fluorotriethylphosphonoacetate to give the corresponding esters followed by a two-step sequence to convert the ester to the ketone. 5,14,16 2-Fluoroenones may also be prepared by selenium-catalyzed α-oxygenation of vinyl fluorides. 6a,9 Addition of in situgenerated phenylselenyl fluoride (from PhSe-Br and AgF) to α-diazoketones and α-diazoesters was also reported. 17 1-Bromovinyl fluorides 18 or 1-chlorovinyl fluorides 11 may be coupled with alkoxyvinylzinc reagents using palladium catalysis. Again, these reactants are prepared in multistep sequences. Ketones may be fluorinated and then made to undergo an aldol condensation with aromatic aldehydes to give aromatic fluoroenones. 19 Yet another approach to 2fluoroenones involves the fluorination of 1,3-dicarbonyl compounds followed by a one-pot condensation−retro-Claisen reaction with an aldehyde. 20 In comparison with all these multistep procedures, an operationally simple single-step preparation of 2-fluoroenones from readily available enones using commercially available reagents would be highly desirable.
Recently, umpolung via discrete iodine(III)−enolonium species 21 has attracted much attention as a strategy to access new chemical reactions not possible through classical reactivity. In this context, we 22 and others 23,24 have reported extensively on the chemistry of iodine(III)−enolonium species, i.e., the electrophilic equivalent of classical lithium enolates. In 2020, we reported the first umpolung Morita− Baylis−Hillman reaction and showed that it could be used to prepare 1,2-diones and 2-tosylenones. 23d,25 It occurred to us that this novel concept could be useful for accessing 2fluoroenones from ubiquitous enones.
We therefore studied the reaction of enone 1 with various iodine(III) reagents (2), 26 fluoride sources (4), and amine bases (3) ( Table 1). The use of DIB (2 equiv) in conjunction with pyridine (1.5 equiv) with different inorganic fluoride salts failed to give any desired product 6 (Table 1, entries 1−3). Acetonitrile was used in order to ensure the solubility of DIB as well as partial solubility of these salts. Using more soluble organic salts such as TBAF or TBAT also failed (entries 4−6). However, using the inexpensive fluoride source triethylamine-HF complex (5 equiv) in combination with DIB (2 equiv) and using pyridine (1.5 equiv) as a MBH activator for the first time led to complete consumption of 1 and formation of a new product, later identified as 5, as observed by TLC. Adding triethylamine to both neutralize excess HF and cause elimination of the amine leaving group of 5 led to the formation of the desired product 6 in 36% isolated yield (entry 7). Knowing the propensity of DIB to oxidize triethylamine as a side reaction, 25a we increased the amount of DIB in the reaction, but this decreased the yield to 30% (entry 8). Carrying out the reaction in dichloromethane led to a decrease in yield to 15% (entry 9). Replacing DIB with 2 equiv of PhI(OPiv) 2 (entry 10) did not give any improvement. Exchanging DIB for PIFA, or Koser's reagent, or TolIF 2 led to no product formation at all (entries 11−13). However, using iodosyl benzene did lead to an improved yield of 55% (entry 14). A distinct issue is the low solubility of iodosyl benzene in most solvents. However, the protonated reagent generated by the action of the fluoride source Et 3 N·3HF does dissolve in acetonitrile. Using the more hindered and soluble 2iodosyl-1,3-dimethylbenzene (2 equiv) improved the yield of 6 further to 67% (entry 15).
The reaction mechanism remained ambiguous at this point in time. At least two likely mechanisms could be postulated. In one (Scheme 1, path b), formation of "an F + -type reagent", such as ArI(OH)F, which could form in the reaction by the We therefore studied the reaction mixture by NMR and HRMS. 1 H NMR results were similar to those we observed in our previous work (see the Supporting Information (SI)). 25a Conclusively, direct injection HRMS of the reaction mixture, before addition of triethylamine, clearly showed the presence of 8 but no trace of difluoro intermediate 9 (see the SI). Thus, path a is likely the mechanism of the reaction.
Next, we examined the scope of the reaction (Scheme 2). Enones with E geometry and methyl or ethyl substituents led to 13 and 14 in 70 and 79% yield, respectively, with the double bond geometry shown (Scheme 2). The latter reaction was carried out on a 1 mmol scale of 10 (R 1 = 2-naphthyl, R 2 = Et). In contrast, enone 10 (R 1 = Ph, R 2 = Ph) did not afford any product when subjected to the standard reaction conditions. Reaction of 1-phenylprop-2-en-1-one led to 15 in 78% yield.
Intriguingly, enones with an additional potentially reactive conjugated double bond underwent highly selective reactions in good yields at the less substituted double bond. Thus, enones with both a terminal double bond and an E-phenylsubstituted double bond reacted selectively to give 16 and 17 in 74 and 78% yield, respectively. Similarly, the cyclohexenesubstituted 2-fluoroenone 18 could be prepared in 76% yield. In none of these three cases was the products of reaction at the more substituted double bond observed. This is further circumstantial support for a Morita−Baylis−Hillman mechanism.
Substituted 1-phenyl-2-enones with methyl in the para-, ortho-, or meta-positions of the phenyl group afforded vinyl fluorides 19, 20, and 21 in 85, 67, and 86% yield, respectively. This indicates that steric hindrance of an ortho-substituent could impair the successful reaction, while a meta-substituent would not affect the yield. Indeed, 1-mesitylprop-2-en-1-one was recovered unchanged when submitted to the reaction conditions. In contrast, two methyl groups in the para-and The para-tert-butyl-substituted 2-fluoroenone 23 was produced in 86% yield. Interestingly, phenyl-substituted 24 was formed in 90% yield. Despite the oxidizing and acidic conditions of the reaction, enones with electron-donating groups are quite compatible with the reaction conditions. Thus, 2-fluoroenones 25, 26, and 27 with methoxy groups in the para-, ortho-, and metapositions, respectively, were all isolated in similar yields of 76− 77%, thus indicating that methyl ethers of phenol are quite compatible with the reaction's conditions. Phenoxy-substituted 2-fluoroenone 28 was produced in 84% yield. Remarkably, two or even three methoxy-group-substituted benzenes were also tolerated despite being highly susceptible to both oxidation and acid hydrolysis. Thus, 2-fluoroenones 29, 30, and 31 could be prepared in 78, 76, and 80% yield, respectively. 2-Fluoroenone 32 with the adjacent oxygen atoms protected as a formaldehyde acetal could be synthesized in 84% yield.
Enones with a powerful electron-withdrawing nitro group might be less prone to oxidative conditions, and we therefore tested enones with nitro groups in the para-and meta-positions and were able to prepare 33 in 63% yield and 34 in 68% yield. Cyano-substituted fluoroenone 35 was successfully synthesized in 68% yield. Halogen atoms on the benzene ring were also compatible with 36, 37, 38, and 39 being formed in 79, 68, 70, and 76% yield, respectively.
Interestingly, attempts to prepare 40 from the corresponding enone failed. However, when the N-tosyl-protected indole enone was used instead, the deprotected product 40 was isolated in 84% yield. Unprotected pyrrole-substituted enone also did not give the desired product. Additionally, alkylsubstituted enones such as (E)-2-hexen-3-one did not work in the reaction.
Adding to the known synthetic use of 2-fluoroenones, 5−11 we demonstrated two synthetic transformations of 2fluoroenones (Scheme 3). Thus, α-fluoroenone 6 was converted to corresponding alkyl alcohol 41 (full reduction) and allyl alcohol 42 (partial reduction) under Pd−C/H 2 and NaBH 4 , respectively (Scheme 3). Esters of 2-fluoroallylic alcohols have been utilized in Claisen-type rearrangements 28 and Tsuji-Trost allylic functionalization reactions. 29 In conclusion, we have developed a single-step synthesis of 2-fluoroenones using HF-pyridine as a source of nucleophilic fluoride. The reaction likely takes place via an umpolung Morita−Baylis−Hillman mechanism. ■ ASSOCIATED CONTENT

Data Availability Statement
The data underlying this study are available in the published article and its online Supporting Information.

* sı Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.orglett.3c00313. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.