The incorporation of difluoromethyl group (CF₂H), a lipophilic hydrogen bond donor, has become a valuable tool in pharmaceutical research and drug development, as the CF₂H can serve as an alternative and complementary bioisostere for alcohols, thiols, amines and hydroxamic acid [1-3]. In the meantime, alkenes are present in a wide range of natural products and bioactive compounds [4]. Thus, the hydrodifluoromethylation of alkenes, which allows installing CF₂H group in a wider scope, has attracted increasing research interest [5-16]. In recent years, an emerging powerful strategy is photoredox-catalyzed hydrodifluoromethylation of alkenes with different CF₂H radical precursors. In this context, Dolbier and co-workers have shown the application of CF₂HSO₂Cl in visible-light-induced hydrodifluoromethylation of activated alkenes in the presence of Ir-based photocatalyst [17]. Qing and co-workers developed several elegant photoredox-catalyzed hydrodifluoromethylation of activated and unactivated alkenes with CF₂Br₂ [18] or difluoromethyltriphenylphosphonium salts [19, 20]. An electron donor and acceptor complex strategy was also developed by Xiao and co-workers for the generation of CF₂H radical from difluoromethyltriphenylphosphonium salt [21]. Recently, Gouverneur and co-workers successfully developed an alternative photocatalyst-free protocol by using difluoroacetic acid and superstoichiometric amount of phenyliodine(Ⅲ) diacetate (Scheme 1A) [22]. Despite these progresses, these methods involve either photocatalysts or superstoichiometric amounts of oxidants and additives. In this regard, the development of a photocatalyst- and additive-free method would benefit the synthetic and biological applications of hydrodifluoromethylation.

Fluoromethylphosphonium salts are readily available and bench stable and have been used as fluoromethyl source for the synthesis of a variety of fluoro-containing compounds [23, 24]. In particular, the photoredox-catalyzed single-electron reduction of difluoromethylphosphoniums to produce CF₂H radical has been extensively investigated since the pioneering study of Qing [25-31]. Very recently, we found that the σ-hole effect could enable the photolysis of monofluoromethyltriphenylphosphonium iodide in the presence of TMEDA as the electron donor [32]. On the basis of these studies, we envisioned that the introduction of a stronger electron-withdrawing CF₂H group can enhance the acidity of α-hydrogen and deepen the σ-hole [33-41], which would facilitate the formation of an intramolecular charge-transfer complex (ICTC) [42-47]. The resultant ICTC can undergo single electron transfer (SET) from I⁻ to [PPh₃CF₂H]⁺ to generate CF₂H radical, thus without using additional Lewis base as the electron donor (Schemes 1B and C).

In our previous work, we found that the difluoromethyltriphenylphosphonium iodide salt can react with activated alkenes for the cascade reactions and difluoromethylation reactions [32]. As an ongoing research in weak interaction enabled photoreactions [48, 49], we set out to explore the possibility of developing a photocatalyst- and additive-free, visible-light-induced hydrodifluoromethylation of alkenes with difluoromethyl-triphenylphosphonium iodide salt. Initially, the model reaction of difluoromethyltriphenylphosphonium iodide salt 1 and unactivated alkene 2 was investigated to validate our hypothesis (Table 1). We were pleased to find that the reaction went smoothly with THF as the solvent, giving the desired product 3 in 78% yield, wherein THF serves as the hydrogen atom donor (entry 1). A solvent screening showed that the reaction in DMF, DMA, CH₃CN, acetone, toluene or DCM provided poorer results than in THF (entries 2–7). In addition, no reaction was observed without irradiation (entry 8).

With the optimized conditions in hand, the scope of the alkenes was briefly investigated (Scheme 2). Reactions of a series of α, β-unsaturated esters, amide, 4-vinyl-1, 1′-biphenyl, dimethyl(phenyl)(vinyl)silane and (vinylsulfonyl)benzene all reacted well with difluoromethyltriphenylphosphonium iodide salt 1 and gave the desired products 4–10 in 50%−75% yields. The use of unactivated alkenes also allowed moderate to good yields. The reaction of (allyloxy)benzene gave the desired product 11 in 57% yield. Both electron-withdrawing (12–15) and electron-donating (16 and 17) groups on the phenyl ring were tolerable. Alkyl substituents (18–20) worked as well. This was also true for the 4-allyl-1, 2-dimethoxybenzene (21). A substrate with a free hydroxyl group was also successful and provided the desired product 22 in 56% yield.

To explore the synthetic utility of this simple procedure, a convenient method for the late-stage structural modification of the natural products and drugs containing terminal double bond was developed. The alkenes derived from various scaffolds were all converted into the corresponding products 23–33 in moderate to good yields.

To understand the reaction mechanism, the UV-vis spectrum of difluoromethyltriphenylphosphonium iodide salt was first measured. The tail of the spectrum extends to over 400 nm and overlaps with the emission spectrum of the blue LED employed (Fig. S4 in Supporting information), indicating that the salt could be activated by blue light irradiation. Next, we performed control experiments (Scheme 3). When the difluoromethyltriphenylphosphonium bromide salt 34 was employed as the substrate instead of difluoromethyltriphenylphosphonium iodide salt 1, only trace amounts of the desired product were observed, indicating the essential role of I⁻ anion for the formation of ICTC (Scheme 3A). In addition, when we subjected the 2, 2, 6, 6-tetramethylpiperidin-1-oxyl (TEMPO) to difluoromethyltriphenylphosphonium iodide salt 1 and irradiated the mixture, the corresponding trapped radical species 35 was obtained in 32% NMR yield (Scheme 3B). These results suggested that the CF₂H radical was generated by just irradiation of the difluoromethyltriphenylphosphonium iodide salt.

Furthermore, we carried out density functional theory (DFT) and time-dependent DFT (TDDFT) calculations (see Supporting information for computational details) to gain deeper insight into the reaction mechanism, using the model reaction in Table 1. We optimized several structures for difluoromethyltriphenylphosphonium iodide salt 1. As compared in Fig. 1A, the structure (IM1) of the salt with I⁻ staying at the opposite of CF₂H group is optimal in terms of both stability and absorption wavelength. The structure IM1a with hydrogen bonding is slightly more stable than IM1, but it has apparently shorter absorption wavelength. The computed UV-vis spectrum of IM1 and IM1a are in reasonable agreement with experimental UV-vis spectrum of the salt (Fig. S4 in Supporting information). The structure IM1b featuring halogen bonding has comparable absorption wavelength close to that of IM1, but it is 3.0 kcal/mol less stable than IM1. The structure IM1c involving interaction with a THF molecule is inferior in terms of both stability and photoactivity. These results suggest that the σ-hole effect of phosphorus cation in IM1 may contribute to its stability and photoactivity. In the following, we use IM1 to compute the reaction pathway (Fig. 1B).

To start the reaction, IM1 is vertically excited to its first excited state (¹IM1*) under the blue light irradiation. Examination of the HOMO and LUMO of IM1 indicates that the excitation results in SET from I⁻ to [PPh₃CF₂H]⁺. Then ¹IM1* relaxes to an equilibrium structure ¹IM1eq* in the first excited state. Compared to the P—C bond length of 1.89 Å in IM1, the bond in ¹IM1eq* is elongated to 1.94 Å. Thus, the bond tends to break, delivering CF₂H radical (i.e., ²). The bond cleavage may take place via two possible pathways. Along pathway (a), the P—C bond breaks in the first excited state via ¹TS*. Unfortunately, attempts to locate the transition state were unsuccessful. Proceeding along pathway (b), ¹IM1eq* undergoes inter-system crossing (ISC), reaching triplet ³IM1eq, followed by crossing ³TS1 in which the C—P bond is elongated to 2.35 Å. The resultant CF₂H radical then couples with the terminal carbon of alkene 2 via ²TS2a, giving alkyl radical ²IM3. Because a primary alkyl radical is less stable than a secondary alkyl radical, the CF₂H^• attacks at the internal carbon of 2 via ²TS2b is less favorable, which accounts for the regioselectivity of the reaction. Finally, the alkyl radical ²IM3 extracts a hydrogen atom from THF via ²TS3, affording the product 3. The resultant THF radical can couple with I radical generated from photoexcitation, giving the by-product 36. Alternative to THF, some substrates featuring benzylic or/and hydroxyl hydrogen (i.e., those of 19– 21, 22) could also serve as hydrogen source. The DFT calculations show that such substrates have somewhat lower barriers than THF to undergo HAT with alkyl radical like ²IM3, but deuterium labeling experiments indicate that THF solvent suppressed such substrates to undergo HAT with the alkyl radical, because THF is the solvent and has much higher concentration than the substrates (Fig. S5 in Supporting information for details). Overall, referring to ¹IM1*, the reaction is energetically downhill with a rate-determining barrier of 21.6 kcal/mol for C—C coupling and hydrogen atom abstraction. The results account for the occurrence of the reaction.

In summary, we have described a photocatalyst- and additive-free method for the hydrodifluoromethylation of alkenes with readily available difluoromethyltriphenylphosphonium iodide salt. Mechanistic studies show that the salt could undergo SET from I⁻ anion to [PPh₃CF₂H]⁺ cation under blue light irradiation, giving CF₂H radical. The radical then adds to the alkene to fulfill the hydrodifluoromethylations. The method adds a new choice for the generation of difluoromethyl radical.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

We acknowledge financial support from the National Natural Science Foundation of China (Nos. 22001248 and 22173103) and the Fundamental Research Funds for the Central Universities and the University of the Chinese Academy of Sciences.

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.04.071.