Hydrofluoromethylation of alkenes with fluoroiodomethane and beyond

A process for the direct hydrofluoromethylation of alkenes is reported for the first time. This straighforward silyl radical-mediated reaction utilises CH2FI as a non-ozone depleting reagent, traditionally used in electrophilic, nucleophilic and carbene-type chemistry, but not as a CH2F radical source. By circumventing the challenges associated with the high reduction potential of CH2FI being closer to CH3I than CF3I, and harnessing instead the favourable bond dissociation energy of the C–I bond, we demonstrate that feedstock electron-deficient alkenes are converted into products resulting from net hydrofluoromethylation with the intervention of (Me3Si)3SiH under blue LED activation. This deceptively simple yet powerful methodology was extended to a range of (halo)methyl radical precursors including ICH2I, ICH2Br, ICH2Cl, and CHBr2F, as well as CH3I itself; this latter reagent therefore enables direct hydromethylation. This versatile chemistry was applied to 18F-, 13C-, and D-labelled reagents as well as complex biologically relevant alkenes, providing facile access to more than fifty products for applications in medicinal chemistry and positron emission tomography.


Introduction
The introduction of uoroalkyl groups has garnered signicant interest in medicinal chemistry, enabling the modulation of biological and physicochemical properties of lead candidates for drug discovery. [1][2][3] Whilst the elds of radical tri-uoromethylation and diuoromethylation have been extensively explored, 4-10 the uoromethyl radical has received far less attention. [11][12][13] This is unexpected as the uoromethyl group features frequently in pharmaceutical drugs, more oen to improve metabolic stability by serving as a bioisosteric replacement of functional groups responsible for poor performance. 14,15 In recent years, several reagents for the generation of the CH 2 F radical have been developed. [16][17][18][19][20] Oen, efficient activation of these reagents requires harsh reaction conditions, such as elevated temperatures, strong oxidants, or strong reductants. Furthermore, many of these reagents are either expensive, highly toxic or non-commercial, requiring multistep syntheses for their preparation. As part of our growing interest in developing "minimalistic" procedures for the late-stage hydrouoroalkylation of alkene-containing biologically active molecules, [21][22][23] we sought to develop an operationally simple method for the direct hydrouoromethylation of alkenes, as an attractive strategy for the introduction of this motif to C(sp 3 )enriched backbones (Scheme 1).
In 2020, an indirect method for the hydrouoromethylation of alkenes was developed by Aggarwal and co-workers; 13 this elegant multi-step procedure starts with the conversion of alkenes into boronic esters, subsequent treatment at low temperature (À78 C) with in situ formed uoroiodomethyl lithium to generate uoroboronic esters, and a nal protodeboronation. Our aim was to develop a one-step method that avoids operational complexity and over-engineering, ideally using uoroiodomethane which is a non-ozone depleting, easy to handle and inexpensive commercial CH 2 F radical precursor. We noted that uoroiodomethane has found applications as an electrophilic or nucleophilic uoromethylation reagent as well as in cross-coupling reactions, 24-27 but has not been explored in the context of radical chemistry.
The high reduction potential of CH 2 FI (E red ¼ À2.19 V vs. saturated calomel electrode (SCE) in MeCN), 28 much closer to MeI (E red ¼ À2.39 V vs. SCE in MeCN) 28 than CF 3 I (E red ¼ À1.22 V vs. SCE in MeCN), 29 encouraged the implementation of an activation pathway exploiting instead the favourable bond dissociation energy (BDE) of C-I (BDE (FH 2 C-I) ¼ 233 kJ mol À1 ) versus C-F (BDE (IH 2 C-F) ¼ 460 kJ mol À1 ). 30 Since the pioneering work of Chatgilialoglu, 31 tris(trimethylsilyl)silane (TTMSS) has found ample applications as a powerful tool for mild radical generation via the activation of alkyl halides. [32][33][34][35][36] In addition, TTMSS has valuably complemented Giese-type reactions, a commonly exploited platform for late-stage functionalisation, by providing a suitable alternative to traditional toxic tin-based reagents. 37 Consequently, we envisioned that the supersilyl radical (TMS) 3 Sic would be well suited to release cCH 2 F from CH 2 FI. Subsequent Giese-type addition of cCH 2 F to the electron-decient alkene would generate a carbon-centered radical intermediate. Hydrogen-atom transfer (HAT) between this electrophilic species and hydridic (TMS) 3 SiH would afford the desired hydrouoromethylated product, and (TMS) 3 Sic entering chain propagation. Initiation for this process would be triggered by photolytic C-I cleavage of FH 2 C-I. 38 This method offers the prospect of being applicable to a range of other halocontaining alkyl radicals, provided that competitive hydrogen atom abstraction with (TMS) 3 SiH does not occur prior to Giese addition. Herein, we report the realisation of this strategy with a wide range of haloiodomethanes for the direct hydrohalomethylation of electron-decient alkenes including biologically relevant molecules. The method was extended to 18 F-hydrouoromethylation and hydromethylation with iodomethane along with ve of its D and 13 C isotopomers.

Results and discussion
Preliminary experiments were conducted with N-phenyl acrylamide (1a) ( Table 1). 39 Various combinations of silanes and solvents revealed that the desired hydrouoromethylated product (2a) was obtained in 71% with (TMS) 3 SiH in MeCN at room temperature under blue light irradiation for 16 h (entry 1). 40 The addition of fac-Ir(ppy) 3 (0.5 mol%) did not lead to signicant improvement (entry 2). The simpler protocol was therefore retained for further investigations. Control experiments indicate that the reaction was not effective in absence of light (entry 3), and unsuccessful in absence of silane or in presence of the radical scavenger TEMPO (entries 4 and 5). No deuterium incorporation was observed in the product when the reaction was performed in CD 3 CN. 40 These data corroborate our proposed radical chain propagation mechanism, initiated by blue-light homolysis of the CH 2 F-I bond. 38 Giese addition of the uoromethyl radical to an electron-decient alkene furnishes an electrophilic carbon-centered radical intermediate, capable of undergoing HAT with (TMS) 3 SiH. The resulting silyl radical enables chain propagation by abstracting iodine from CH 2 FI to afford (TMS) 3 SiI along with cCH 2 F. 40 With the optimised reaction conditions in hand, we sought to explore the scope of this hydrouoromethylation protocol (Scheme 2A). Various functional groups, such as methoxy, nitrile, halide, ketone, ether, amide, ester, aniline, and sulfone were tolerated. The addition of fac-Ir(ppy) 3 (0.5 mol%) led to higher yields for selected substrates. 40 N-Aryl acrylamides bearing electron-withdrawing and electron-donating groups afforded the desired products in moderate to excellent yields (2a-d). The hydrouoromethylation of N-heteroaryl acrylamides, such as pyridyl and benzothiazyl was also successful (2e, 2f). Alkenes substituted with sulfones and esters were competent substrates generating 2g and 2h in moderate yield. As deuteration can improve metabolic stability, 41 we investigated the hydrouoromethylation of a deuterated alkene (1i) that was successfully converted into [D 3 ]2i. The gem-disubstituted alkene 1j provided 2j in 64% yield. Pleasingly, the internal alkene 1k was reactive under our reaction conditions and afforded uoromethylcyclobutane 2k in moderate yield. This result is signicant as 1,2-disubstituted uoroalkyl cyclobutanes currently require multiple steps for their preparation. 42 A non-Scheme 1 Hydro(per)fluoromethylation of alkenes. This work: direct silyl radical-mediated hydrofluoromethylation of electron-deficient alkenes and extension to numerous hydro(halo)methylation reactions. With fac-Ir(ppy) 3 (0.5 mol%) 75 3 No light Traces 4 No silane 0 5 With cyclic trisubstituted alkene afforded the product in 57% yield (2l). Styrene derivatives such as 1m and 1n afforded the desired products in synthetically useful yields (2m, 2n). Our protocol is amenable to scale-up as demonstrated by the 10 mmol scale hydrouoromethylation of N-benzylmaleimide affording 2o in 88% yield. The synthesis of uorinated pyrrolidine 2p, amine 2q, alcohol 2r and carboxylic acid 2s was performed in two steps, offering a pathway to diversify the range of products within reach from CH 2 FI. The late-stage hydro-uoromethylation of complex biologically active molecules was considered next. The anti-cancer drug ibrutinib as well as estrone, tyrosine and ethacrynic acid derivatives afforded the desired hydrouoromethylated products in good yields (2t-w).
The tolerance of functional groups was investigated with a robustness screening. 40 These experimental data provide an overview of the many heteroarenes (e.g. pyridazine, 1,3,5triazine, indole, benzothiazole or oxazole) that are tolerated under the optimised reaction conditions. Whilst additives containing nucleophilic functional groups such as alcohols and anilines were tolerated, side reactivity arising from nucleophilic substitution was observed. 40 Competitive alkylation was suppressed when using 1.0 equivalent of CH 2 FI, albeit at the expense of reduced yield for the hydrouoromethylated product. Aliphatic amines were tolerated but yields did not exceed 30%. 40 The hydro-uoromethylation of alkenes not bearing electron-withdrawing groups was possible albeit signicantly less efficient. 40 With a protocol relying on the favourable C-I bond dissociation energy and considering the importance of homologation in medicinal chemistry, 43 we considered the generation of products from a series of homologated uoroiodoalkanes (Scheme 2B). 44,45 Hydrouoroalkylation of alkenes 1g, 1j and 1l provided effortlessly the homologous series of products 3d-i. Specically, the uoroethyl radical was efficiently generated applying similar silyl radical activation, and 3a was isolated in good yield. The introduction of the uoroethyl radical was successfully performed on linear terminal, gem-disubstituted, and trisubstituted alkenes (3d, 3f, 3h). The method was further extended to uoroiodopropane as shown with the synthesis of 3e, 3g, and 3i. Precursors featuring additional uorine atoms were less suitable with the diuoroethylated product 3b isolated in 30%, and no product observed when attempting to prepare the hydrotriuoroethylated product 3c. Increased uorine content enhances radical electrophilicity, thereby encouraging undesired H-atom abstraction from (TMS) 3 SiH. 40 Given the success of our protocol, we further investigated the applicability of our method for the generation of [ 18 F]CH 2 F radical from [ 18 F]CH 2 FI (Scheme 2C). [46][47][48] Compounds labelled with the radioisotope F-18 are important for applications in Positron Emission Tomography (PET). [49][50][51][52][53] The synthesis of [ 18 F] CH 2 FI in high molar activity (A m ) is well-established and has been automated. [54][55][56][57] To date, this labelled reagent is mainly employed for the electrophilic 18 F-uoromethylation of phenols. 58, 59 We now demonstrate that [ 18 F]CH 2 FI is well suited for [ 18 F]CH 2 F radical chemistry. Specically, Ibrutinib, an estrone, a tyrosine, and an ethacrynic acid derivative underwent 18 F-hydrouoromethylation in radiochemical yields up to 81% ([ 18 F]2t-w). This reaction was best performed for 20 minutes at ambient temperature in the presence of fac-Ir(ppy) 3 under bluelight irradiation. This method offers an alternative to nucleophilic 18 F-uorination with [ 18 F]uoride for precursors that are either unstable, require complex multiple steps synthesis, or lead predominantly to elimination products. Haloiodomethanes other than uoroiodomethane were also considered as they would allow for the one-step introduction of reactive halomethyl groups to alkenes (Scheme 2D). Controlled activation of reagents such as ICH 2 X (X ¼ Cl, Br, I) would enable their use for example as cCH 2 + synthon. To date, only few examples for the generation and use of halomethyl radicals have been reported. [60][61][62][63][64] When diiodomethane was employed under the standard reaction conditions, N-benzylmaleimide underwent hydroiodomethylation in 62% yield (4a). Similarly, hydrobromomethylation (from dibromomethane or bromoiodomethane), hydrochloromethylation (from chloroiodomethane), and hydrobromouoromethylation (from dibromouoromethane) provided the corresponding halomethyl alkanes in moderate yields (4b-4d). 23,65 Other alkenes afforded the hydrochloromethylated products in moderate yields (4e-4g).
Although full conversion of starting material was observed for these reactions, purication via silica gel chromatography led to elimination, which is reected in the lower yield for these compounds upon isolation.
Competition experiments were performed to calibrate the reactivity of uoroiodomethane versus other alkyl iodides (Scheme 3). When equimolar amounts of iodomethane and uoroiodomethane were subjected to the standard reaction conditions, product resulting from uoromethyl radical addition was obtained in 74% yield (2n), along with 25% of the hydromethylated product 5a. When the reaction was carried out with equimolar amounts of iodoethane, products 2n and 6 were formed in close to 1 : 1 ratio. Additional competition experiments showed that the iso-propyl and tert-butyl adducts (7,8) were formed preferentially over the hydrouoromethylated product. The reactivity of these alkyl iodides therefore decreases in the following order: tBuI > iPrI > CH 2 FI $ EtI > MeI.
A notable outcome of this study was the observation that net methane addition across the double bond took place with iodomethane. Currently, protocols for the generation of the methyl radical from iodomethane (BDE CH3-I ¼ 239 kJ mol À1 , E red ¼ À2.39 V vs. SCE in MeCN) remain underdeveloped. 34,66 In recent years, the methyl radical has been generated from numerous precursors. [67][68][69][70][71] The formation of the methyl radical oen requires harsh reaction conditions, limiting the applicability of these protocols. Furthermore, the use of the methyl radical towards application to isotopic labelling is far from trivial. Iodomethane, on the other hand, can provide effortless access to a variety of useful isotopologues that would otherwise be beyond reach. The straightforwardness of our protocol prompted us to optimise the hydromethylation of alkenes using iodomethane as methyl radical precursor (Scheme 4). We noted signicant gas release when applying our reaction conditions, attributed to methane resulting from competitive HAT between the methyl radical and MeCN (BDE NCCH2-H ¼ 389 kJ mol; BDE CH3-H ¼ 439 kJ mol À1 ). 69 A screen of solvents, reactants stoichiometry and photocatalysts allowed for hydromethylation to occur in up to 93% yield (5a). 40 Under the optimised reaction conditions consisting of 4.0 equivalents of MeI, 3.0 equivalents of (TMS) 3 SiH and 1,2-diuorobenzene as solvent, in combination with photocatalyst MesAcrBF 4 (0.5 mol%), the hydromethylation of various alkenes took place in good to excellent yield (5b-f). Considering that bioactive compounds containing stable heavy isotopes are useful for example as mass spectroscopy standards, 41,72 the hydromethylation of an ethacrynic acid derivative was performed with CH 3 I, CH 2 DI, CHD 2 I, CD 3 I, 13 CH 3 I, and 13 CD 3 I. All six isotopologues (5h-5m) were obtained in moderate yield. [73][74][75]

Conclusions
In conclusion, the rst direct hydrouoromethylation of a broad range of electron-decient alkenes has been developed using uoroiodomethane. Mechanistically, the process harnesses known principles; so its core value is rooted in its immediate synthetic power. With the current global necessity "to do more with less", this minimalistic and mild chemical method stands out as it is operationally simple with the supersilyl radical precursor (TMS) 3 SiH being the only chemical required in addition to the reaction partners. The mild reaction conditions are compatible with complex biologically active molecules such as Ibrutinib. The methodology was successfully adapted for the 18 F-labelling of complex alkenes, and offers a new C-CH 2 18 F disconnection strategy for radiotracer development. The method was extended to additional uoroiodoalkanes enabling facile product homologation, as well as multiple (halo)methyl radicals including the methyl radical itself and ve of its D and 13 C isotopomers.

Data availability
The datasets supporting this article have been uploaded as part of the ESI. †

Conflicts of interest
There are no conicts to declare.