Introduction

The precise fluorine editing of organic molecules has emerged as a powerful tool in modern drug discovery due to the beneficial effect of fluorine atom(s) that can significantly improve the metabolic stability, lipophilicity, and binding affinity of bioactive compounds1,2,3,4. Consequently, impressive achievements have been made in the fluoroalkylation reactions over the past decades5,6,7,8,9. However, most developed methods focus on the transformations of fluorinated carbanions, carbocations, and carbon-centered radicals5,6,7,8,9. Compared to these three active intermediates, difluorocarbene, the smallest fluorocarbon unit, has the advantage of forming two chemical bonds10,11,12, providing a new dimension to expand the chemical space and create new fluorine structures for medicinal chemistry. Ideally, coupling difluorocarbene with two simple and readily available feedstocks would enable more efficient access to organofluorine compounds without the tedious synthesis of fluoroalkylating reagents (Fig. 1a). Nevertheless, this straightforward synthetic route is regulated by the high reactivity of difluorocarbene. As a result, only limited reaction types of difluorocarbene transfer reactions have been reported so far13,14,15,16. To overcome this limitation, the complexation of difluorocarbene with metal would be an attractive strategy, as the reactivity of difluorocarbene can be modulated by metal (Fig. 1b). However, due to the lack of catalytic activity in those isolated metal difluorocarbene complexes, the metal-catalyzed difluorocarbene transfer reaction remains a substantial challenge17, in sharp contrast to the classic metal catalyzed carbene transfer reactions that have been proven to be a powerful transformation in organic synthesis18,19,20. This problem is further underscored by the lack of understanding of metal difluorocarbene chemistry, though investigating metal difluorocarbene complexes has been around for over 40 years21.

Fig. 1: Metal difluorocarbene-involved catalytic coupling.
Fig. 1: Metal difluorocarbene-involved catalytic coupling.The alternative text for this image may have been generated using AI.
Full size image

a Coupling free difluorocarbene. b Metal difluorocarbene-involved catalytic coupling. c Previous nucleophilic or electrophilic addition of metal difluorocarbene. d Catalytic difluorocarbene transfer via 1,1-migration. e Our report on the copper difluorocarbene-involved catalytic gem-difluoropropargylation. f Unique properties of the CF2 group and representative CF2-containing bioactive molecules.

We recently isolated palladium(0)22 and copper(I)23 difluorocarbene complexes ([Pd0]=CF2 and [CuI]=CF2) and found they possess opposite reactivities ([Pd0]=CF2, nucleophilic; [CuI]=CF2, electrophilic), though Pd0 and CuI have the same d electron count. These findings have been applied in catalytic organic synthesis24,25,26,27. However, the initial step of these catalytic difluorocarbene transfer reactions requires the formation of the low valent metal difluorocarbene ([M] = CF2, M = Pd0, CuI) intermediates, followed by attacking the carbene carbon center with an electrophile or a nucleophile to generate a difluoroalkyl metal species ([M]-CF2R) (Fig. 1c)22,23. We envision that the formation of the M-CF2R species by migratory insertion of difluorocarbene into the C-M bond would open a new dimension to harness metal difluorocarbene chemistry for catalytic synthesis of organofluorine compounds, as the C-M bond can be easily constructed by transmetalation or oxidative addition28, which would provide a more general pathway for catalytic difluorocarbene transfer reactions (Fig. 1d). To realize this hypothesis, one critical factor is the rapid formation of a metal difluorocarbene complex C-[M] = CF2, followed by a facile migratory insertion pathway without the influence of coupling [M]-C with an electrophile or a nucleophile. Since copper is low-cost, earth-abundant, and easy to form a [CuI]-C species via transmetalation between [CuI] and a nucleophile29, we assume that using copper as a catalyst under suitable conditions may address the above crucial issue and provide a cost-efficient route for modular construction of fluorinated structures (Fig. 1e), thus expanding copper difluorocarbene chemistry and opening an avenue to efficient, precise synthesis of organofluorine compounds.

Here, we disclose a copper-catalyzed gem-difluoropropargylation reaction via 1,1-migration of copper difluorocarbene (Fig. 1e). This reaction uses inexpensive and industrial feedstock potassium bromodifluoroacetate (BrCF2CO2K) as the difluorocarbene precursor30,31, allowing various widely available potassium propiolates, terminal alkynes, and allyl/propargyl electrophiles to couple difluorocarbene, providing a facile route to accessing synthetically valuable gem-difluoropropargylated compounds. The distinct feature of this approach is its synthetic simplicity, eliminating the tedious synthesis of fluoroalkylating reagents or moisture-sensitive organometallic reagents. The diverse transformations of the resulting products, as well as the applications of the current protocol in the rapid synthesis of key intermediates for bioactive molecules, demonstrate the synthetic utility of this catalytic difluorocarbene transfer reaction, showing promise in modern drug discovery. Mechanistic studies reveal that the fast migratory insertion of difluorocarbene into the C-Cu bond of alkynylcopper(I) species is the key step for the catalytic difluorocarbene transfer.

Results

Mechanistic investigation

To test our hypothesis, we chose terminal alkynes as the nucleophiles, as the resulting gem-difluoropropargyl structure is a synthetically versatile synthon for diverse transformations. Notably, it has been widely used in copper-free click chemistry32,33 due to the unique properties of the difluoromethylene (CF2) group, which can lower the lowest unoccupied molecular orbital (LUMO) of the alkynes32,34. Furthermore, the CF2 group is a bioisostere of the oxygen atom and the carbonyl group (Fig. 1f)2,35. Incorporating the CF2 group at the metabolically labile position can enhance the metabolic stability of bioactive molecules1,2,3,4. It has been one of the valuable strategies for discovering new bioactive molecules by tactically site-selective difluoromethylenation (Fig. 1f)1,2,3,4,35,36. However, efficient methods for such a gem-difluoropropargyl structure are limited. The developed methods either rely on the deoxyfluorination of alkynyl ketones with sulfur fluorides37 or coupling gem-difluoropropargyl bromides with organometallic reagents38,39, aldehydes40, or imines41. However, the requirement of multiple steps to prepare the substrates, such as alkynyl ketones and organometallic reagents, as well as the poor functional group compatibility of sulfur fluorides, and the use of strong base n-butyllithium to prepare gem-difluoropropargyl bromides42 regulate the widespread applications of these methods. Yet, the current catalytic modular synthesis harnessing copper difluorocarbene chemistry would overcome these limitations and provide straightforward access to the gem-difluoropropargyl structure.

Initially, to ascertain the feasibility of the migratory insertion of difluorocarbene into the C-Cu bond, we prepared the 1,10-phenanthroline-supported alkynyl complex A143. Unexpectedly, treatment of A1 with inexpensive and widely available difluorocarbene precursor BrCF2CO2K 1 in CH3CN at 50 °C afforded a trifluoroalkene 3 instead of gem-difluoropropargyl copper complex C1 (Fig. 2a). A similar result was also observed in DMF. Although low yields of 3 were obtained due to the decomposition of A1, these results demonstrate the feasibility of the migratory insertion pathway. Once C1 was formed via the copper difluorocarbene complex B1, it underwent another difluorocarbene insertion to generate a tetrafluoroalkylcopper E1 (ref 23). Finally, the β-fluoride (β-F) elimination of E1 produced 3. This possible pathway indicates that the difluorocarbene elongation in the alkynyl copper complex C1 is favorable, and the tetrafluoroalkyl copper E1 is prone to β-F elimination due to its instability. Complex A1 could also be used as a nucleophile to react with 1 and allyl chloride 2a in CH3CN, providing the three-component coupling product 4 in 38% yield along with a side product 5 (13%) generated between A1 and 2a (Fig. 2b). Replacing CH3CN with DMF led to a lower yield of 4. No 4 was observed using DMSO. These results suggest that the formation of C1 via an alkynylcopper difluorocarbene complex B1 through 1,1-migration is reasonable, which should be faster than the cross-coupling of A1 with 2a in a suitable reaction media, such as CH3CN and DMF, thereby facilitating the formation of gem-difluoropropargyl structure in the catalytic reaction. Given the difficulty in obtaining C1 through the current difluorocarbene pathway, we prepared gem-difluoropropargyl cadmium species F1 and F2 by reaction of gem-difluoropropargyl bromide 6 with cadmium in DMF44. These two organocadmium reagents were assigned according to the literature45. Transmetalation of the mixture of F1 and F2 with CuI at −40 °C afforded the gem-difluoropropargyl copper C2 and bis(gem-difluoropropargyl)copper species C3 in 41% yield and 8% yield, respectively. Since it is hard to isolate these two species, they were directly used to react with allyl chloride 2a, providing 7 in 95% yield, thus demonstrating the feasibility of coupling gem-difluoroproparyl copper with an electrophile (Fig. 2c). To investigate the possibility of nucleophilic addition of alkynyl species to the carbene carbon center, we prepared copper difluorocarbene complex G23. However, no desired product 4 was obtained when G was treated with 2a and alkynyl nucleophiles, including alkynyl lithium/zinc reagents (8a, 8b) and potassium propiolate 9a (Fig. 2d). Thus, the pathway beginning with the formation of [CuI]=CF2, followed by a reaction with an alkynyl nucleophile, is less likely.

Fig. 2: Mechanistic studies of gem-difluoropropargylation via 1,1-migration of copper difluorocarbene.
Fig. 2: Mechanistic studies of gem-difluoropropargylation via 1,1-migration of copper difluorocarbene.The alternative text for this image may have been generated using AI.
Full size image

a Stoichiometric reaction of alkynylcopper complex A1 with difluorocarbene. b Stoichiometric reaction of A1 with difluorocarbene and allyl chloride 2a. c Preparation of alkynylcopper species and their reactions with 2a. The number in parentheses is the 19F NMR chemical shift. d Reaction of alkynyl nucleophiles with 2a and copper difluorocarbene G. e Possible pathway for the copper difluorocarbene-involved catalytic gem-difluoropropargylation.

Based on the above results, a copper-catalyzed difluorocarbene transfer reaction should be feasible for the catalytic modular synthesis of gem-difluoropropargylated compounds. In this copper-catalyzed process, the reaction is initiated by the formation of an alkynylcopper species A, which subsequently undergoes complexation with a difluorocarbene to generate an alkynylcopper difluorocarbene intermediate B. This key intermediate undergoes 1,1-difluorocarbene migratory insertion to produce the gem-difluoropropargyl copper species C. Finally, C reacts with an electrophile to produce the gem-difluoropropargylated compound and releases copper catalyst simultaneously (Fig. 2e).

Optimizations for catalytic reactions

Inspired by the above observations and the possible pathway illustrated in Fig. 1e, we explored a catalytic coupling reaction between terminal alkyne 8c and allyl chloride 2a to couple with difluorocarbene (Table 1). When 8c (1.0 equiv) was treated with 2a (1.5 equiv) and difluorocarbene precursor 1 (2.0 equiv) in the presence of CuCl (10 mol%) and 1,10-phenanthroline L1 (10 mol%) in CH3CN at 50 °C using K2CO3 as the base, 10% yield of the desired product 4 was obtained along with 5% yield of side product 5 (entry 1). The use of CH3CN as the solvent is due to its good solubility for BrCF2CO2K. A survey of the ligands showed that ligand L4 could suppress the generation of 5 and increase the yield of 4 to 34% (entries 2-4, Supplementary Table 2). This finding is likely attributed to the preferential complexation between the electron-rich alkynyl copper species, stabilized by ligand L4, and the electron-deficient difluorocarbene. This interaction is more favorable than the complexation of alkynyl copper species with the carbon-carbon double bond of the electrophile 2a. Consequently, the formation of the gem-difluoropropargyl copper complex occurs more rapidly via migratory insertion of difluorocarbene into the Cu-C bond than through the oxidative addition of the alkynyl copper to the electrophile. This kinetic preference effectively suppresses the formation of byproduct 5. Replacing K2CO3 with Na2CO3 slightly improved the reaction efficiency (entry 5, Supplementary Table 3). However, the undesired defluorination of 8c and the sensitivity of [CuI]=CF2 to base make it challenging to further increase the yield. To circumvent these limitations, we chose readily available potassium propiolate 9b as an alternative substrate. We envisioned that the relatively faster release of the alkynyl nucleophile through decarboxylation of 9b, without the need for a base, would benefit the reaction efficiency. Similar to terminal alkyne 8c, the ligand is critical for the reaction (entries 6-8, Supplementary Table 5), and L4 remained the optimal ligand, providing 7 in 73% yield at 80 °C (entry 6). Decreasing the reaction temperature to 70 °C increased the yield to 85% (entry 9). To further optimize the reaction conditions, we examined a series of reaction parameters, including copper catalysts, solvents, catalyst loading amounts, reactant ratios, and reaction times (entries 10, 11, Supplementary Tables 6–11). Finally, the optimized reaction conditions were identified by shortening the reaction time to 30 min with 7.5% mol CuCl/L4 as the catalyst (entry 12). Under these conditions, 1.5 equiv of 1 and 1.2 equiv of 2a could provide 7 in 80% isolated yield. Notably, this reaction proceeded smoothly, even shortening the reaction time to 10 min (entry 13). This distinct feature is in sharp contrast to conventional copper-catalyzed fluoroalkylation reactions, which typically require a long time, thereby underscoring the advancement of the current copper-catalyzed difluorocarbene transfer reaction. No product was observed without copper salt or ligand (entries 14, 15), demonstrating the essential role of Cu/L in promoting the reaction. It should be mentioned that we did not observe the formation of the tetrafluorohomoproargylated byproduct during the reaction process. This result suggests that the oxidative addition of the gem-difluoropropargyl copper complex with the allyl electrophile is faster than the difluorocarbene elongation process. This preference is likely attributed to the relatively stronger Cu-CF2 alkynyl bond, which results in a slower migratory insertion of the difluorocarbene compared to the oxidative addition of the gem-difluoropropargyl copper complex.

Table 1 Representative results for the optimization of the reaction conditionsa
Full size table

Substrates scope

With the viable reaction conditions in hand, we examined the scope of this copper difluorocarbene-involved catalytic gem-difluoroproparylation reaction (Fig. 3). Various potassium arylpropiolates were applied to this transformation (Fig. 3a), providing the corresponding gem-difluoropropargylated products efficiently (4, 7, 12-33). Generally, aromatic propiolates bearing an electron-donating substituent provided higher yields than electron-deficient substrates. The reaction exhibited high functional group tolerance. Base and nucleophile sensitive functional groups, such as ketone (13), ester (14), and nitrile (16), were compatible with the reaction; aryl fluoride (4), chloride (21), bromide (17, 19-21), and iodide (18) moieties underwent the current copper-catalyzed process smoothly. Additionally, the position of bromide in the aromatic ring did not affect the reaction efficiency. Para-, meta-, and ortho-aryl bromides efficiently delivered the corresponding gem-difluoropropargylated products (17, 19, 20). The high compatibility of the chlorobromoaryl moiety (21) offers a good opportunity for diversified transformations by sequential aryl bromide and chloride functionalization. Ferrocene- and thiophene-containing substrates were also employed in the reaction, yielding moderate to good yields (22-24). However, low yield was obtained with the pyridine-containing substrate (25). The reaction was not restricted to allyl chloride 2a, as substituted allyl chlorides, including linear, branched, and cyclic allyl chlorides (26-33), underwent smooth coupling. Even highly reactive allyl chlorides bearing vinyl chloride (31) or unsaturated ester (32) were still amenable to the reaction. In the case of linear allyl chloride (26), no branched product was observed. The observed regioselectivity is likely due to the steric effect that influences the reductive elimination of the Cu(III) species, generated between the gem-difluoropropargyl copper complex and the allyl electrophile. Specifically, steric interactions between the gem-difluoropropargyl group and the substituent on the allyl moiety favor reductive elimination at the less hindered position, thereby yielding a linear product.

Fig. 3: Copper difluorocarbene involved catalytic gem-difluoropropargylation.
Fig. 3: Copper difluorocarbene involved catalytic gem-difluoropropargylation.The alternative text for this image may have been generated using AI.
Full size image

a Substrate scope of potassium propiolates and allyl chlorides. b Substrate scope of propargyl sulfonates. a9 (0.5 mmol, 1.0 equiv), 1 (1.5 equiv), 2 (1.2 equiv), MeCN (5 mL). b11 (0.5 mmol, 1.0 equiv), 1 (1.5 equiv), 9 (1.2 equiv). All reported yields are isolated yields.

In addition to arylpropiolates, alkyl- and silyl-substituted propiolates were competent coupling partners (34-39). The aliphatic side chain bearing a benzyloxy (35), chloride (36), sulfamide (37), or cyclopropyl (38) group did not interfere with the reaction efficiency. This approach could also be extended to propargyl electrophiles (Fig. 3b). One problem with this type of substrate is the formation of a copper-allenylidene complex between the copper catalyst and the propargyl electrophile46. This competitive side reaction significantly influences the current copper-catalyzed difluorocarbene transfer process. After extensive efforts (Supplementary Table 12), we found that using propargyl sulfonates as the limiting substrates could suppress this undesired side reaction, producing various gem-difluoropropargylated allenes with high efficiency. Although the synthesis of allenes has been well established47,48,49,50, efficient methods for such fluoroalkylated allenes have yet to be reported. Given the synthetic versatility of allene and alkyne, the resulting gem-difluoroporpargylated allenes should be a valuable structure for diverse transformations. As depicted in Fig. 3b, arylpropiolates underwent smooth coupling with good functional group tolerance (40-44). Versatile synthetic handles, such as nitrile (41), thiophene (42), aryl bromide (43,44), and alkyl chloride (44) moieties, tolerate the reaction well. In contrast to the allyl electrophiles, arylpropiolate bearing an electron-withdrawing group provided a higher yield (41). The reaction is also compatible with complex molecule-containing substrates, as demonstrated by the efficient synthesis of gem-difluoropropargylated allenes 45 and 46. Furthermore, trisubstituted allene 47 could also be achieved with high efficiency. However, alkylpropiolates yielded only 20–30% of the products, and using benzyl bromide as the electrophile resulted in a 13% yield (for details, see the Supplementary Information). Notably, for all the coupling reactions described above, no [2 + 1] cycloaddition side products were generated between difluorocarbene and the unsaturated carbon-carbon bond15, thereby further advancing this copper catalytic system.

Gram-scale reaction and applications

The reaction is readily scalable, as demonstrated by the gram-scale synthesis of 7 with high yield (Fig. 4a). The resulting gem-difluoropropargyl products can be elaborated through a myriad of transformations to create a diverse range of new organofluorine compounds. Selective oxidative cleavage of the carbon-carbon double bond of 7 with ozone, followed by reduction with NaBH4, afforded alcohol 48 efficiently. Cyclization of 38 with phenidone 49 via rhodium catalysis produced difluoroalkylated indole 50 with high efficiency (Fig. 4b)51. The gem-difluoropropargyl structure could also be used to construct the difluoroalkylated pyrrole 53 through deprotection of 39, followed by silver-catalyzed [3 + 2] reaction with ethyl isocyanoacetate 52 (Fig. 4c)52. Given the unique properties of the CF2 group and critical applications of indole and pyrrole in medicinal chemistry, the rapid access to these complex fluorinated molecules that otherwise require tedious steps to prepare through conventional methods provides a good opportunity to discover new interesting bioactive molecules. Remarkably, this copper-catalyzed difluorocarbene transfer reaction could be used as a key step to introduce the CF2 group at the metabolically labile allylic position of bioactive molecules. As shown in Fig. 4d, pheromone derivative 56, a probe used to study hydrophobic interaction in pheromone reception, was rapidly accessed from 54 via two steps, followed by a reported procedure53. Since the Z-difluoroalkylated alkenes have been found in a series of pheromone analogs53, this copper difluorocarbene-involved catalytic gem-difluoropropargylation should have applications in such compounds. Furthermore, using the (-)-corey lactone diol-derived terminal alkyne 8d as the substrate could directly afford 53 by harnessing the current copper difluorocarbene chemistry (Fig. 4e). Although a 42% yield of 57 was obtained, 46% of 8d could be recovered. Notably, compound 57 could be used as a potential key intermediate for synthesizing tafluprost 58, a PGF2α agonist for treating glaucoma36, thus underscoring the synthetic utility of this transformation. The transformations of gem-difluoropropargylated allene 40 were also performed. Hydrogenation of 40 efficiently produced difluoroalkylated compound 59. Treatment of 40 with MTBD (7-Methyl-1,5,7-triazabicyclo[4.4.0]decene-5) at room temperature selectively cleaved one of its C-F bonds, yielding a medicinally interesting enediyne 60 in good yield.

Fig. 4: Diverse transformations of the gem-difluoropropargylated products.
Fig. 4: Diverse transformations of the gem-difluoropropargylated products.The alternative text for this image may have been generated using AI.
Full size image

a Gram-scale synthesis of 7 and its transformation. b Synthesis of difluoroalkylated indole 46. c Synthesis of difluoroalkylated pyrrole 49. d Synthesis of the key intermediate 51 for pheromone derivative 52. e Synthesis of the key intermediate 53 for PGF2a analogue 54. f Transformations of compound 40.

In summary, a copper difluorocarbene-involved catalytic coupling reaction has been developed. The stoichiometric reactions reveal that a difluorocarbene migratory insertion into the C-Cu bond is a key step in the catalytic cycle. This approach enables the rapid, modular synthesis of valuable gem-difluoropropargyl structures using a wide range of readily available, simple components, including potassium propiolates, terminal alkynes, and allyl/propargyl electrophiles. This reaction opens an avenue for the efficient and precise synthesis of organofluorine compounds. Most importantly, this work should also prompt the development of new metal difluorocarbene-involved catalytic coupling reactions in methodology development.

Methods

General procedure of the gem-difluoropropargylation

To a 25 mL Schlenk tube, CuCl (7.5 mol%), L4 (7.5 mol%), potassium propiolate 9 (0.5 mmol, 1.0 equiv), and BrCF2CO2K (1.5 equiv) were added under Ar. Anhydrous MeCN (5 mL) was added, and the mixture was stirred for 2 min before adding the corresponding allyl chloride 2 (1.2 equiv). The tube was screw-capped and heated at 70 °C (oil bath). After stirring for 30 min, the reaction was cooled to room temperature and concentrated in vacuo. The residue was purified by silica gel column chromatography and reverse-phase column chromatography to provide the desired product.