Nickel-Catalyzed Decarbonylative Synthesis of Fluoroalkyl Thioethers

Abstract

This report describes the development of a nickel-catalyzed decarbonylative reaction for the synthesis of fluoroalkyl thioethers (R_FSR) from the corresponding thioesters. Readily available, inexpensive, and stable fluoroalkyl carboxylic acids (R_FCO₂H) serve as the fluoroalkyl (R_F) source in this transformation. Stoichiometric organometallic studies reveal that R_F–S bond-forming reductive elimination is a challenging step in the catalytic cycle. This led to the identification of diphenylphosphinoferrocene as the optimal ligand for this transformation. Ultimately, this method was applied to the construction of diverse fluoroalkyl thioethers (R_FSR), with R = both aryl and alkyl.

Graphical Abstract

Fluoroalkyl thioethers (R_FSR) have emerged as increasingly common motifs in bioactive molecules due to their unique physiochemical properties.¹ As shown in Figure 1A, thioethers bearing diverse fluoroalkyl substituents (for example, CF₂H, CFH₂, and CH₂CF₃) appear in lead structures relevant to both medicinal and agricultural chemistry.^2,3 The most common synthetic routes to R_FSR involve either the electrophilic fluoroalkylation of thiols (Figure 1B, i)^3–5 or the coupling of aryl/alkyl electrophiles with [M]–SR_F nucleophiles (Figure 1B, ii).^3,6–8 Both approaches have significant limitations with respect to the breadth of R_F substituents that can be introduced, since very few of the necessary R_F-containing electrophiles/nucleophiles are commercially available.^6,7 Furthermore, many of these methods require other toxic, unstable, or expensive reagents.^3,4,6,7 Overall, more general synthetic approaches to fluoroalkyl thioethers are of high interest, and the use of readily available fluoroalkyl carboxylic acids as R_F precursors would be particularly enabling in this context.

Figure 1.

(A) Representative examples of bioactive molecules containing fluoroalkyl thioethers (R_FSR). (B) Existing synthetic approaches to R_FSR (i, ii) and our approach (iii).

This report describes the development of a Ni-catalyzed reaction for constructing fluoroalkyl thioethers from the corresponding thioesters (Figure 1B, iii). Our approach leverages fluoroalkyl carboxylic acids as inexpensive, stable, and commercially-available R_F precursors.^9–12 As such, it enables the construction of a variety of different fluoroalkyl thioethers from a single thiol starting material.

Recent studies on Ni-catalyzed coupling reactions of carboxylic acid derivatives^13,14 led us to propose this decarbonylative route to fluoroalkyl thioethers. Recent reports from our group^13c and others^15–17 have demonstrated that Ni(0) phosphine complexes catalyze decarbonylative C–S coupling reactions of (hetero)aryl thioesters to afford (hetero)aryl thioether products (for example, see Figure 2A). We hypothesized that an analogous pathway, using fluoroalkyl thioesters as starting materials, could offer a route to R_FSR products. The proposed catalytic cycle (Figure 2B), involves initial oxidative addition of the fluoroalkyl thioester at a Ni(0) catalyst to form the acyl Ni(II)-intermediate, I (step i). Carbonyl de-insertion then generates the Ni(II)(fluoroalkyl)(thiolate) intermediate II (step ii). Finally, II undergoes C–S bond-forming reductive elimination (step iii) to yield the target fluoroalkyl thioether product and regenerate the Ni(0) catalyst.

Figure 2.

(A) Example of precedent for decarbonylative thioetherification. (B) Proposed catalytic cycle. (C) Initial catalysis studies. (D) Stoichiometric studies with PⁿBu₃ as ligand.

We initiated these investigations by targeting the conversion of difluoromethyl thioester 1a to thioether 2a (Figure 2C). We focused on catalysts based on a combination of Ni(cod)₂ and monodentate phosphine ligands (PR₃), which were previously employed for the transformation in Figure 2A.^13c,15–17 However, only traces (<1%) of product 2a were detected using PPh₃, P(o-Tol)₃, PCy_3, or PBu₃ (Figure 2C). In all of these systems, the majority of the mass balance was the unreacted starting material 1a.

We next conducted stoichiometric studies to identify the problematic step(s) in this sequence. The treatment of a toluene solution of Ni(cod)₂/PⁿBu₃ with 1 equiv of 1a resulted in the formation of (PⁿBu₃)₂Ni(SPh)(CF₂H) (II-PⁿBu₃) within 1 h at ambient temperature (Figure 2D). Complex II-PⁿBu₃ was characterized in situ via ¹⁹F and ³¹P NMR spectroscopy, which show resonances indicative of a trans configuration, with three-bond coupling between the CF₂H and PⁿBu₃ ligands (J_PF = 26.5 Hz). The formation of II-PⁿBu₃ implicates the feasibility of two key steps of the catalytic cycle: oxidative addition (step i) and carbonyl de-insertion (step ii). However, when in situ-generated II-PⁿBu₃ was heated at 130 °C for 2 h, none of the thioether product 2a was formed (step iii). Instead, the resonances associated with II-PⁿBu₃ slowly decayed, without the observation of identifiable organic products. This suggests that F₂HC–S bond-forming reductive elimination is challenging in this system and that alternative ligands are required to enable this step.

Literature reports have shown that 1,1’-bis(diphenylphosphino)ferrocene (dppf) is particularly effective for promoting challenging reductive elimination reactions.¹⁸ As such, we next conducted an analogous stoichiometric experiment with Ni(cod)₂/dppf. As shown in Figure 3A, the treatment of a toluene solution of Ni(cod)₂/dppf with 1 equiv of 1a resulted in 70% consumption of 1a within 1 h at 50 °C. This was accompanied by the formation of 2a (in 12% yield) along with broad signals in the ¹⁹F NMR spectrum. Based on previous reports,^18a these broad signals are indicative of fluxional (dppf)Ni^II intermediates. Subsequent heating at 130 °C for 1 h resulted in S–CF₂H bond-formation to generate 2a in 90% yield by ¹⁹F NMR spectroscopy (Figure 3A).¹⁹ Dppf was next examined as a ligand for the catalytic transformation of 1a to 2a. As shown in Figure 3B, the combination of 10 mol % Ni(cod)₂ and 12 mol % dppf afforded 2a in 58% yield over 20 h at 130 °C in toluene. Further optimization of the reaction solvent and time resulted in nearly quantitative yield over 4 h in THF (Figure 3B).²⁰

Figure 3.

(A) Stoichiometric and (B) catalytic studies with dppf.

The scope of this transformation was first explored with respect to the substitution on sulfur (Figure 4). The difluoromethyl thioester substrates 1a-w were prepared via the reaction of RSH with difluoroacetic anhydride. These were typically obtained in quantitative yield without the need for purification by column chromatography. Aryl thioesters bearing electron-donating and -neutral substituents (1b-f) afforded good yields of the difluoromethyl thioether products (Figure 4A). Substituents such as ethers, amines, and amides were compatible. Aryl thioesters bearing electron withdrawing groups resulted in lower yields (see products 2h-2l), with the exception of 4-fluorothiophenol derivative, 2g. In these systems, the major side products were diarylthioethers, which are likely formed via competing activation of the aryl–S bond of the product by the Ni(0) catalyst.²¹ This transformation showed modest sensitivity to sterics on the aryl ring, and substrates containing either one or two electron-donating ortho-substituents afforded 2m-2o in moderate to good yields.

Figure 4.

Scope of (A) aryl and (B) alkyl thioethers. ^a% conversion of 1 to 2 as determined by ¹⁹F NMR spectroscopy. ^bY-ield determined by ¹⁹F NMR spectroscopy with 4–fluorotoluene as internal standard. ^cCatalyst loading was increased to 15 mol% Ni(cod)₂, 18 mol% dppf.

Primary, secondary and tertiary alkyl thiols were also effective substrates for this transformation (for example 2p, 2s, and 2t in Figure 4B). Thiol-containing biologically active compounds such as captopril (2v) and thioglucose (2w) underwent conversion to the corresponding difluoromethyl thioethers in good yields. In these systems, unreacted starting material accounted for the remaining mass balance when the yields were modest. Importantly, the catalytic cycle does not require an exogenous base. This limits racemization of substrates like 2v during catalysis.

Finally, we used this approach to synthesize a series of different fluoroalkyl thioethers. As shown in Figure 5, the substrates for this transformation were synthesized from commercially available R_FCO₂H and thiols. Catalytic decarbonylation then provided the partially fluorinated thioether products 2x-2ab in good to excellent yields. Importantly, these products are challenging to access using most existing approaches (Figure 1B), due to the inaccessibility of the required fluoroalkylating reagents. One current limitation of this approach is that fluorinated derivatives (e.g., SCF₃, SCF₂CF₃) afford none of the desired fluoroalkyl thioether product.²² A stoichiometric study of the CF₃ system showed the formation of Ni–CF₃ intermediates; however, no thioether product was detected upon heating these species. This result suggests that the S–R_F reductive elimination step remains a challenge in these systems.²³

Figure 5.

Scope of fluoroalkyl groups derived from commercial R_FCO₂H. Isolated yields. See the SI for details. ^aCatalyst loading was increased to 20 mol% Ni(cod)₂ and ligand was Xantphos (24 mol %).

In summary, a nickel-catalyzed decarbonylative coupling reaction was developed to convert fluoroalkyl thioesters to the analogous thioethers. This method leverages readily available fluorocarboxylic acids as commercial and stable fluoroalkyl sources to install these functional groups, which are increasingly prevalent in biologically active molecules.