Introduction

Difluorocarbene represents a versatile synthetic building block, readily accessible from commercially available and inexpensive halodifluoroalkyl reagents1,2,3,4,5,6,7,8,9,10,11, with wide applications in organic synthesis1,2,3,4,5, drug development6,7,8,9,10, and advanced functional materials11. As a singlet carbene, difluorocarbene is intrinsically electrophilic due to the existence of an empty p-orbital and exhibits conventional carbene reaction properties1, such as cycloaddition reactions with alkenes or alkynes for formation of gem-difluorocyclopropanes12,13,14,15,16,17, the Wittig reactions with carbonyl to generate gem-difluoroalkenes, etc (Fig. 1a, left)18,19,20. In addition, the difunctional reactions of difluorocarbene3, in which two chemical bonds are simultaneous formed at the carbene carbon center via reacting with nucleophile, followed by coupling with an electrophile, enable difluorocarbene to be a bipolar CF2 linker and production of the difluoroalkylated compounds (Fig. 1a, left)21,22,23,24,25,26,27,28,29. Compared to these conventional CF₂-containing compounds constructing, unconventional transformations of difluorocarbene involving deconstructive functionalization of C-F bonds, beyond its role as a difluoromethyl synthon, have garnered significant attention4,30. The reactions are initiated by the electron-deficient characteristics of difluorocarbene that cause the C-F bond scission, enabling difluorocarbene as a versatile C1 synthon for the assembly of valuable N-containing compounds31,32,33,34,35,36, heterocycles31,32, and aliphatic ethers36 via different intermediates such as isocyanides32, cyano anions33, formamides34, and others (Fig. 1a, right)31,35,36. However, despite its thought-provoking properties and fascinating application potential, the elusive reactivity of difluorocarbene poses great challenges to the control of the reaction pathways and limits its widespread use in synthetic chemistry, resulting in the restricted reaction types.

Fig. 1: Transformations involving difluorocarbene.
figure 1

a Free difluorocarbene-involved reactions; b Pd-catalyzed difluorocarbene transfer via [Pd] = CF₂; c Cu-catalyzed difluorocarbene transfer via [Cu] = CF₂; d This work: Cu-catalyzed multi-component reactions via [Cu] = CF₂ as carbonyl source.

Full size image

In contrast to free difluorocarbenes, metal difluorocarbenes ([M] = CF2) in which transition metals coordinate with carbene carbon to modulate the reactivity offer a promising approach to overcome the limitations mentioned above37,38,39,40. However, transition-metal catalyzed reactions involving metal–difluorocarbene complexes remain severely underdeveloped, although various [M] = CF2 have been synthesized and characterrized over the past 40 years37,38,39,40,41,42,43,44,45. Recently, Zhang’s group successfully synthesized, isolated, and characterized a [Pd⁰]=CF₂ complex for the first time and disclosed its application in the palladium-catalyzed coupling reaction of difluorocarbene with arylboronic acids (Fig. 1b)46,47,48. Following this breakthrough, several catalytic reactions involving palladium-difluorocarbene complexes have been reported, where metal difluorocarbenes exhibit varying reactivity controlled by the valence state of palladium49,50,51,52,53,54,55,56. Among them, the strongly nucleophilic [Pd⁰]=CF₂ can be protonated to produce PdII–CF₂H, which can subsequently couple with aryl boronic acids57 and terminal alkynes (Fig. 1b, left)49,50. While the electrophilic [PdII]=CF₂ undergoes hydrolysis with water to generate CO, serving as a CO surrogate in carbonylation reactions (Fig. 1b, right)51,52,53,54,55,56,57. Although [Cu]=CF₂ complexes were proposed by Burton decades ago and later suggested by Ichikawa in 2016, the development has significantly lagged behind that of [Pd]=CF₂ complexes58,59,60. Until 2023, the isolation and structural characterization of [CuI]=CF₂ complex was first achieved by Zhang’s group61.The study demonstrated the electrophilic nature of [CuI]=CF₂ complex, which allows it to be attacked by silyl enol ethers, thereby enabling the catalytic modular synthesis of fluorinated compounds (Fig. 1c)60,61,62,63. The intriguing discovery of difluorocarbene presents an exciting opportunity to expand the fluorine chemical space. However, catalytic transformations involving [CuI]=CF₂ are still in their infancy, and application of such electrophilic copper difluorocarbene complexes to organic synthesis remains a significant challenge.

Multicomponent reactions (MCRs) are regarded as one of the strategies that most closely approach the ‘ideal synthesis’, which are flexible, selective, convergent, and atom-efficient processes to construct complex molecules by a single step64,65,66,67,68,69. Our research group has been dedicated to the development of multicomponent reactions involving metal carbene, which proceed through the interception of active ylide/zwitterionic intermediates by various electrophiles, enabling the synthesis of a series of multifunctional molecules68,69,70,71,72,73,74,75. However, these reactions predominantly focus on donor-acceptor carbene intermediates. As our continuing interest in multicomponent reactions and considering the intriguing chemical properties of metal difluorocarbene, we propose that metal difluorocarbene could be attacked by nucleophilic reagents to generate a ylide intermediate, which is subsequently captured by an electrophile, thereby enabling a catalytic multicomponent reaction60,68,69. This approach is initiated by formation of a difluorocarbene metal complex where metal control the reactivity of difluorocarbene. Thus, it would overcome the limitations of direct difunctionalization of free difluorocarbene that can only couple with limited nucleophiles such as organometallic regents, halide ions and phosphines3.

Herein, we report a multicomponent reaction of arylamine, aldehyde and BrCF2CO2K as difluorocarbene precursor under the catalysis of copper, providing cost-efficient access to multifunctional amide (Fig. 1d). The reaction is proposed to involve copper difluorocarbene intermediate that is attacked by amine to form copper associated ammonium ylide. Subsequent interception of the active ylide intermediate occurs and is accompanied by carbonyl migration with fluorine elimination. In addition, besides aromatic aldehydes, alkyl aldehydes and drug modified arylamine can also be tolerated in this MCRs, which demonstrates the practical applicability of this method. Moreover, this process could serve as an effective and ideal alternative to the Ugi or Strecker reaction, addressing key limitations such as the reliance on highly toxic cyanide salts or the use of toxic and unstable isonitriles76.

Results

We initiated our research with the reaction of 4-bromoaniline (1a), BrCF2COOK (2a), and benzaldehyde (3a) under different reaction conditions (Table 1). The multicomponent defluorination product 4a was obtained in 57% yield, rather than CF₂-containing compounds, when a 2.5:3:1 mixture of 1a, 2a and 3a was treated with CuCl and racemic BINOL-derived phosphoric acid (PPA) in acetonitrile at 50 °C for 12 h under argon atmosphere (Table 1, entry 1). After evaluating a series of copper catalysts (Table S1, in Supplementary Information), Cu(CH3CN)4PF6 was selected as the optimal one, leading the 4a in 78% yield (Table 1, entry 2). Subsequently, various solvents were tested, and it was found that acetonitrile was the ideal solvent for this model reaction (Table S1, in Supplementary Information). When BrCF2COOK 2a was replaced by TMSCF2Br (2b) or BrCF2COOEt (2c), no product was detected (Table 1, entries 3, 4). The reaction could proceed smoothly with the lower yield when ClCF2COONa (2 d) was employed as difluorocarbene precursor (Table 1, entry 5). The addition of a Brønsted acid was found to influence the yield of the reaction (Table 1, entries 6–11, and Table S2, in Supplementary Information). Among the acids tested, p-toluenesulfonic acid (TsOH) proved to be the most effective, affording 4a with in 83% yield. (Table 1, entry 11). The optimal reaction temperature was determined to be 50 °C. When the temperature was lowered, even with extended reaction times, the starting materials remained unchanged (Table 1, entries 12, 13). Additionally, increasing the temperature did not lead to a significant increase in the yield of the reaction (Table 1, entry 14). When the reaction occurred in air atmosphere, the yield of multicomponent product 4a would slightly decrease (Table 1, entry 15). Attempts at the asymmetric synthesis of the target product 4a were systematically explored. However, no significant stereoselectivity was observed in the transformation (Table S3, S4, in Supplementary Information).

Table 1 Condition optimizationa
Full size table

With the optimized reaction conditions in hand, we explored the scope of the multi-component reaction by investigating aromatic amines 2 (Fig. 2). The reaction with substrates containing para-substituents on the aromatic ring of aniline proceeded smoothly, and whether halogen (4a-4c) and trifluoromethyl (4 d) substituents with electron withdrawing properties or alkyl (4 f, 4 h and 4i), methoxy (4 g), and phenyl (4j) substrates with electron withdrawing properties gave the corresponding products in middle to good yields. The structure of 4e, derived from benzaldehyde, aniline, and BrCF2COOK, was unambiguously confirmed by X-ray crystallographic analysis. In addition, meta-substituents on the aromatic ring of aniline were well tolerated in such transformation, affording the multi-component products (4k-4n). It is worth noting that the ortho-substituted products (4o and 4p) were not detected, likely due to steric hindrance from the functional groups. Both disubstituted aromatic amines and β-naphthylamine were suitable for this reaction system (4q-4t). When halogen substituted aromatic aldehydes are used as substrates, aniline substituted with chlorine, methoxy, and trifluoromethoxy are also suitable substrates, yielding the corresponding products in good yields (4u-4w). Furthermore, aliphatic amines such as benzylamine (BnNH2) and 2,2,2-trifluoroethylamine (CF3CH2NH2) were also evaluated as substrates in our reactions, and the reactions underwent well to react with BrCF2CO2K and imine as the reaction partner. The desired products were obtained in 36% and 32% yield, respectively (4x and 4 y), albeit with competitive formation of aniline-derived products 4a. Excitingly, drug Prezista as HIV-1 protease inhibitor could also be applied to this transformation, and the target compound was obtained in 55% yield (4z).

Fig. 2: Substrate scope of amines.
figure 2

Unless otherwise noted, all reactions were conducted with 0.2 mmol of 3 in acetonitrile (3 mL) at 50 °C for 12 h, 1: 2a: 3 = 2.5: 3: 1 under argon atmosphere. Isolated yield.

Full size image

The scope of aldehydes was then investigated (Fig. 3). When aniline was used as the substrate, both monosubstituted 4-trifluoromethylbenzaldehyde and disubstituted 3,4-dichlorobenzaldehyde were well tolerated in this reaction system, to afford the target compounds (4aa and 4ab). When 4-bromoaniline participated in the reaction, aromatic aldehyde derivatives with electron-donating or electron-withdrawing functional groups on the para-position of aromatic ring, including halo, trifluoromethyl, nitro, methyl, methoxy, phenyl, and thiomethyl were harmonious with such catalytic system, yielding corresponding products (4ac-4aj). Once ortho- and meta-substituted aromatic aldehydes were applied to the reaction, the yields of the corresponding products (4ak-4aq) were favorable without significant decrease, which indicated that the steric hindrance of aromatic aldehydes did not interfere with the reaction yield. Among them, the structure of 4ap was unambiguously confirmed by X-ray crystallographic analysis. Heterocyclic aromatic aldehydes, polycyclic aromatic aldehyde and fused heterocyclic aromatic aldehydes were well tolerated in such reaction system, giving the multi-component products (4ar-4aw) in middle to good yields. To our delight, non-aromatic aldehydes such as cyclohexanecarbaldehyde, tetrahydropyran-4-carbaldehyde, and glyoxylic esters are also suitable substrates, affording desired products (4ax-4az). Moreover, aldehydes 3ab and 3ac derived from Gemfibrozil and Ciprofibrate were also suitable for this reaction system, to give the desired products (4 A and 4B), which demonstrated the utility of this reaction.

Fig. 3: Substrate scope of aldehydes.
figure 3

Unless otherwise noted, all reactions were conducted with 0.2 mmol of 3 in acetonitrile (3 mL) at 50 °C for 12 h, 1: 2a: 3 = 2.5: 3: 1 under argon atmosphere. Isolated yield.

Full size image

Moreover, to highlight the synthetic value of such approach, when 4-methoxybenzaldehyde was scaled up to 6 mmol, the desired amide 4ah was smoothly obtained in 67% yield, and the amides could be easily further functionalized (Fig. 4). First, compound 4ah was converted into reduced product 5 with 92% yield under the action of NaBH4. Then, compound 4ah could smoothly generate cyclization products with 90% yield in the presence of DBU. In addition, when amide 4 g was used as starting material, the reaction yielded the product 7 with 55% yield, which was deprotected from the p-methoxyphenyl group (PMP) with the assistance of ceric ammonium nitrate (CAN). Of note, when aldehydes were replaced by N, N-dibenzyl-1-methoxymethanamine 8, corresponding multi-component products 9 was obtained with 58% isolated yield.

Fig. 4: Synthetic application.
figure 4

a Gram-scale reactions and transformations of the product 4ah; b Remove the p-methoxyphenyl group in 4 g; c Multi-component reaction with dibenzyl-1-methoxymethanamine. DBU = 1,8-Diazabicyclo[5.4.0]undec-7-ene, CAN ceric ammonium nitrate.

Full size image

In order to better understanding the pathway for this transformation, several validation experiments were conducted. Firstly, 100 μL oxygen-18 water was added to template reaction, the yield of multi-component products was not affected, and no 18O labeled products were observed by HRMS. The standard template reaction results in roughly the same yield with or without water (100 μL). These results indicated that water is not involved in the reaction and the water molecules generated in situ were not the oxygen source of the product in multi-component reactions (Fig. 5a). Secondly, when benzaldehyde was omitted from the multi-component reactions, p-tert-butylaniline 1i was reacted with BrCF2COOK under the standard conditions, leading to formylation products 10a formed via N-H insertion followed by defluorination and direct amidation product 11a with nearly 1:1 ratio in yield (Fig. 5b). Notably, isonitrile was not detected by LC-MS at any stage of the reaction. According to the literature research, isonitriles can only be obtained in the presence of organic or inorganic bases. However, our optimal reaction conditions were not conducive to the generation of isonitriles30. Therefore, we speculated that there was no involvement of isocyanide intermediates in such transformation. Then, compounds 10a and 11a were added to the model reaction of 4-bromoaniline 1a, BrCF2COOK 2a, and benzaldehyde 3a, only the target product 4a was obtained, with no detectable formation of cross-multicomponent products bearing 4-tert-butyl substitution (Fig. 5c). This observation suggests that neither the formylation product 10a nor the amidation product 11a serve as intermediates in this transformation, and exclude a stepwise mechanism for this multicomponent reaction.

Fig. 5: Control experiments.
figure 5

a Control reaction with/ without water and 18O-isotope labeling; b Control reaction of aniline with BrCF2CO2K; c Cross-multicomponent reaction with 10a and 11a; d Control reaction without copper; e Competition experiment of BrCF2COOK, 4-bromoaniline 1a and silyl enol ether 13a.

Full size image

Next, the role of the copper in the catalytic process was further investigated. The control experiments of 4-bromoaniline 1a, BrCF2COOK 2a, and benzaldehyde 3a were performed under the standard conditions in the absence of copper or with alternative Lewis acidic metal catalysts, including Sc(OTf)3, Yb(OTf)3, Zn(OTf)2, Ni(OTf)2 and AgSbF6. Notably, no significant formation of the target product was observed in any of these cases, even at elevated temperatures (Table S5, in Supplementary Information). These findings demonstrate that the copper catalyst fulfills a critical and distinctive role beyond conventional Lewis acid catalysis. Previous studies have unequivocally demonstrated that free difluorocarbenes readily undergo cycloaddition with silyl enol ethers to form gem-difluorocyclopropanes, whereas copper difluorocarbene species exhibit complete incapacitation of cyclopropanation activity60,61. Capitalizing on this distinct reactivity profile, the competitive experiments of BrCF2COOK, 4-bromoaniline and silyl enol ether 13a with or without copper catalyst were carried out. The formylation product 10b was obtained in 37% yield with no detectable cyclopropanation product 12a under the catalysis of Cu(CH3CN)4PF6. In contrast, in the absence of copper catalyst, ¹⁹FNMR analysis revealed the formation of difluorocyclopropane 12a in 19% yield, alongside a significant reduction in formylation product yields (10b < 5%, Fig. 5e). In addition, the template reaction of BrCF2COOK, 4-bromoaniline and benzaldehyde was also conducted under the previously reported conditions for transformations involving a copper-difluorocarbene intermediate (copper salt, 2,9-diMe-1,10-phen, CH3CN, 50 °C.)60,62, yielding the desired multi-component product in 60% yield (Figure S5, in Supplementary Information). These results are consistent with prior reports, corroborating the generation of a copper-difluorocarbene intermediate in our reaction system, and supporting the potential involvement of a copper-difluorocarbene species in the present multi-component reaction. We performed density functional theory (DFT) calculations to illustrate the formation of the intermediate IV as shown in Fig. 6. The free difluorocarbene I preferentially coordinate with [CuI]⁺ to form the copper-difluorocarbene complex II, which is energetically favorable (−15.3 kcal/mol) compared to the direct interaction with ArNH₂, yielding a weakly bound van der Waals complex II’ (4.9 kcal/mol). Subsequent coordination of ArNH₂ to copper-difluorocarbene complex II generates intermediate III (−15.0 kcal/mol), which undergoes C-N bond formation via transition state TS1 (−4.2 kcal/mol) to afford the copper-associated ammonium ylide intermediate IV (−24.1 kcal/mol). The overall process for formation the intermediate IV via copper difluorocarbene isexergonic, releasing 24.1 kcal/mol, indicating a thermodynamically favorable pathway. In the absence of copper catalyst, formation of a weakly bound van der Waals intermediate II’ from free difluorocarbene with ArNH2 exhibits significantly slower kinetics (4.9 kcal/mol versus −15.3 kcal/mol for the copper-mediated pathway). Computational attempts to locate a transition state for formation of C-N bond from II’ were unsuccessful, excluding the possibility of intermediate IV forming directly from free difluorocarbene (Figs. S7, S8, in Supplementary Information). The DFT calculations highlight the crucial role of copper(I) in the transformation, which initiates with coordination of copper(I) to difluorocarbene, generating a key copper-difluorocarbene intermediate. The key intermediate subsequently undergoes C-N bond formation through nucleophilic attack by ArNH₂, ultimately affording intermediate IV.

Fig. 6: Proposed reaction mechanism and the DFT calculated free-energy profile from difluorocarbene to the product.
figure 6

Copper-difluorocarbene was demonstrated as the key intermediate rather than free difluorocarbene for formation of ylide intermediate IV, and subsequently ylide interception, defluorination and carbonyl migration process were energetic feasibility.

Full size image

On the basis of the above results and previous works30,68,69, the plausible subsequently reaction pathway involving ylide interception, defluorination and carbonyl migration of the multicomponent reaction is proposed and demonstrated its energetic feasibility through DFT calculations (Fig. 6). The nucleophilic ammonium ylide intermediate IV is captured by the activated imine via transient state TS2 (1.6 kcal/mol), in which p-toluenesulfonic acid and bromodifluoroacetate ion facilitate the proton transfer, leading to the product VII (−54.4 kcal/mol) with releasing the copper(I). The resulting product VII containing a fragment of CF2 adjacent to nitrogen atom is vulnerable, and the consecutive scission of Csp3-F bond occurs under the assistant of BrCF2COOK. According to the DFT, the monofluoroimine species VIII (−33.3 kcal/mol) is formed via nucleophilic substitution of bromodifluoroacetate ion with VII through transition state TS3 (−33.1 kcal/mol), followed by the second Csp3-F bond cleavage to generate intermediate IX (−21.3 kcal/mol). Intramolecular carbonyl migration of intermediate IX via TS5 (−20.6 kcal/mol) with a barrier of 0.7 kcal/mol and subsequent deprotonation eventually render α-aminoamide products.

Discussion

In summary, we have developed copper-catalyzed MCRs for synthesis of multifunctional amide derivatives from amine, aldehyde and BrCF2COOK without the need for any ligands. The mild reaction conditions, non-toxic nature, and use of readily available raw materials demonstrate that this reaction serves as an effective alternative strategy to the Strecker or Ugi reactions, enabling the synthesis of versatile and valuable products. Additionally, the high functional group tolerance, accommodating not only aromatic aldehydes but also alkyl aldehydes and even complex drug-like molecules, underscores the practical applicability of this method. Control experiments and DFT calculations systematically support that the copper difluorocarbene complex serves as the key intermediate in this transformation and acts as the carbonyl source for the formation of the amide group, and exclude the formation of isonitriles under the reaction conditions. The reaction is proposed to through the formation of copper difluorocarbene, nucleophilic attack by the amine to produce a copper-associated ammonium ylide, interception of the active ylide intermediate with imine, and subsequent defluorination via carbonyl migration. This sequence accounts for the overall high efficiency and distinctiveness of the reaction.

Methods

General

All 1H NMR (500 MHz, 600 MHz) and 13C NMR (125 MHz, 150 MHz) and 19F NMR (471 MHz) spectra were recorded on 500 or 600 MHz spectrometers in in CDCl3, DMSO-d6 and Methanol-d4. Chemical shifts were reported in ppm with the solvent signal as reference, and coupling constants (J) were given in Hertz. The peak information was described as: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, br = broad. High-resolution mass spectrometry (HRMS) was recorded on a commercial apparatus (ESI Source). Single crystal X-ray diffraction data were recorded on Bruker-AXS SMART APEX II single crystal X-ray diffractometer.

General procedure for synthesis of product 4

To an oven-dried 10 mL Schlenk tube equipped with a stir bar was added Cu(CH3CN)4PF6 (7.5 mg, 0.02 mmol, 10.0 mol%), TsOH (6.9 mg, 0.04 mmol, 20.0 mol%), aromatic amines 1 (0.5 mmol, 2.5 equiv), BrCF2COOK 2a (128 mg, 0.6 mmol, 3.0 equiv), and aldehydes 3 (0.2 mmol, 1.0 equiv), and suspended in CH3CN (3.0 mL) under dry argon atmosphere. The resulting mixture was stirred at 50 °C for 12 hours. The progress of the reaction was monitored by TLC. After the reaction was complete, the reaction was cooled to room temperature and concentrated under reduced pressure. The residue was purified by flash column chromatography (eluent: EA:PE = 1/20 ~ 1/5) to give the pure product 4.

General procedure for synthesis of product 9

To an oven-dried 10 mL Schlenk tube equipped with a stir bar was added Cu(CH3CN)4PF6 (7.5 mg, 0.02 mmol, 10.0 mol%), TsOH (6.9 mg, 0.04 mmol, 20.0 mol%), aromatic amine 1a (0.3 mmol, 1.5 equiv), BrCF2COOK 2a (128 mg, 0.6 mmol, 3.0 equiv), and N,N-dibenzyl-1-methoxymethanamine 8 (0.2 mmol, 1.0 equiv), and suspended in CH3CN (2.0 mL) under dry argon atmosphere. The resulting mixture was stirred at 50 °C for 12 hours. The progress of the reaction was monitored by TLC. After the reaction was complete, the reaction was cooled to room temperature and concentrated under reduced pressure. The residue was purified by flash column chromatography (eluent: EA:PE = 1/20 ~ 1/10) to give the pure product 9 (58% yield).