Introduction

Organofluorine chemistry has emerged as a highly significant field over the past few decades, enabling the on-demand synthesis of a wide range of unnatural fluorinated molecules, which has garnered considerable research interest, particularly in pharmaceuticals, agrochemicals, and specialized materials1,2,3,4,5,6,7,8,9,10,11,12. Standing out as one of the most significant fluorinated moieties due to its unique biological properties, the difluoromethyl group (CF₂H) has been recognized as an excellent bioisostere for hydroxyl, thiol, or amide groups, and as a lipophilic hydrogen-bond donor which could significantly enhance binding affinity, membrane permeability, and bioavailability while incorporated into parent molecules2,13, thus attracting increasing interest in drug design and screening. Indeed, several CF₂H-containing drugs and bioactive molecules have been developed14,15,16, including FDA-approved drugs and others in clinical trials (Fig. 1a). Given the inherent chiral nature of biological systems, drug chirality is widely recognized as a crucial aspect of medicinal research. Under these contexts, the synthesis of drug-like molecules with optically active CF₂H groups has been shown to be a very effective way to improve lead optimization in drug development.

Fig. 1: Selected bioactive molecules and recent synthetic methods.
figure 1

a Selected biologically active molecules containing 3 °C-CF2H stereocenter; b CF2H-containing synthons for asymmetric construction of difluoromethylated stereocenters; c Our strategy to enantioenriched difluoromethylated amines via development of CF2H-containing synthons. d This work.

Besides a few notable examples exist for the asymmetric construction of difluoromethylated stereocenters via fluorination17 or hydrogenation18,19,20,21, the most common approaches for obtaining enantioenriched difluoromethyl compounds have focused on direct asymmetric difluoromethylation22,23 and the functionalization of CF₂H-containing starting materials. Interestingly, nucleophilic addition with Me₃SiCF₂H, relying on chiral auxiliaries covalently linked to imines, has typically resulted in only moderate diastereoselectivities, even at low temperatures (−78 °C)24. In contrast, similar additions of carbanion nucleophiles to CF₂H-imines have shown significantly higher diastereoselectivity25,26,27,28. Consequently, various CF₂H synthons have been designed and applied in the enantioselective synthesis of chiral difluoromethylated compounds, employing both chiral auxiliary and asymmetric catalytic strategies. The most widely used electrophilic CF₂H synthons include difluoromethylated imines25,26,27,28, aldehydes/ketones29,30,31,32,33, and olefins34,35,36,37,38, where different nucleophiles attack C = X double bonds enantioselectively to form chiral stereogenic centers featuring a difluoromethyl group. Additionally, the only nucleophilic CF₂H synthon designed so far is a heterocyclic α-imino ketone, utilized as a bidentate directing group for asymmetric 1,4-additions39. However, these CF₂H synthons generally exhibit limited reactivity and enantioselectivity, especially compared to their trifluoromethylated analogs, and the reaction types are restricted. Therefore, a highly efficient and general strategy for the diversity-oriented synthesis of difluoromethylated compounds remains underdeveloped and is urgently needed.

Owing to their unique reactivity, structural diversity, and broad functional group tolerance, highly reactive radicals have facilitated a range of mild and efficient transformations, enabling the rapid assembly of functionalized, complex molecules, particularly through multi-step cascade reactions. Over the past few decades, catalytic enantioselective fluoroalkylation has seen remarkable progress, largely due to the relative stability of fluoroalkyl radicals conferred by the distinctive effects of fluorine. However, while asymmetric radical trifluoroalkylation has advanced rapidly40,41,42,43,44,45,46,47, transition-metal-catalyzed enantioselective radical difluoroalkylation has lagged behind, likely because of the weaker electron-withdrawing nature and smaller steric bulk of the difluoromethyl group. The only reported radical CF₂H synthon, described by Shen, enabled a nickel-catalyzed asymmetric arylation, yielding difluoromethylated 1,1-diarylalkanes with typically moderate enantiomeric excesses (ees), which were noticeably lower than those of their trifluoromethylated counterparts (Fig. 1b)48. There are various strategies to promote enantioselective control in the asymmetric transformation of radicals. As suggested, enhancing the stability of the radicals could offer sufficient recognition time within the chiral pocket of a metal complex, thereby improving the enantioselectivity (if oxidative addition is the enantio-determining step in the reaction). In addition, the application of directing groups is quite common in asymmetric transformations of radicals, where coordination between the metal center and the directing group often facilitates enantioselective control. Therefore, we aimed to design a CF₂H synthon that contains a directing group and can generate a stable radical in situ, to achieve the construction of a CF2H-substituted stereocenter via a radical pathway.

Based on the summary and proposed conception above, we have incorporated isoindolinone—a key structural motif widely used in various biologically active compounds49,50,51,52,53,54,55,56 into alkyl halides to create a CF₂H synthon for asymmetric radical transformations. Our design is based on the following considerations: (1) the nitrogen atom can stabilize the in situ-formed difluoromethylated α-radical, enhancing chiral recognition during the radical oxidation step; (2) the carbonyl oxygen serves as a coordination site for the metal center, facilitating control over the preferential geometry of the lower-energy enantiomeric transition state; and (3) the isoindolinone ring acts as a sterically hindered group to discriminate the CF₂H moiety. Moreover, this CF₂H synthon offers several advantages, including simple preparation, bench stability, low cost, and convenient removal.

Herein, we present an efficient strategy for the general and asymmetric construction of carbon stereocenters containing a difluoromethyl group via a nickel-catalyzed enantioconvergent Negishi cross-coupling with arylzinc reagents. This method features mild reaction conditions, a broad substrate scope, good catalytic efficiency, and excellent enantioselectivity. The resulting difluoromethylated isoindolinone derivatives could be further converted into chiral difluoromethylated amines, allowing for the straightforward synthesis of chiral analogs bearing CF₂H-substituted stereogenic centers found in various known bioactive molecules. This strategy presents an innovative and practical method for the efficient synthesis of CF₂H-containing chiral drug-like molecules, as both optically pure difluoromethylated amines and isoindolinones are significant structural motifs in various biologically active compounds, thereby holding significant potential for drug research and development.

Results and discussion

Optimization of reaction conditions

Our initial study commenced with 2-(1-chloro-2,2-difluoroethyl) isoindolin-1-one (1) as the CF2H-containing coupling partner and arylzinc reagent bearing CO2Me at the para position (2) as the nucleophile to optimize conditions for this nickel-catalyzed stereoconvergent Negishi arylation (Table 1). Given their critical role in asymmetric reactions, several chiral bis-oxazoline ligands were initially evaluated under catalytic conditions, using NiBr2•DME (10 mol%) and NaBr (2.0 equiv) in THF at −20 °C (Table 1); for more details see Table S1 in the SI). To our delight, bis-oxazoline (Box) ligand L7 could deliver the desired product 3a with moderate yield (35%) and acceptable ee value (59%), whereas the 2,2-bis(2-oxazoline) (Bi-ox) ligand L1 was less effective in asymmetric induction. When sterically bulkier ligands L8 and L9 were employed, the reaction proceeded smoothly to furnish the CF2H-containing amide with higher yields and ee (46% yield, 76% ee and 48% yield, 88% ee), respectively. The use of L10, which carries a bulkier substituent, led to a significant loss in enantioselective control, resulting in a decrease in the ee value to 35%. Subsequently, different kinds of solvents have been carefully tested in this reaction (Table 1, entries 1–4;). Notably, the ee value increased to 90% while maintaining a comparable yield when 2-Me-THF was used in place of THF. Based on this outcome of solvent screening, 2-Me-THF was selected as a co-solvent with THF and a promising result was achieved. (entry 3; 64% yield, 91% ee). After the careful evaluation of additives and nickel catalysts, 0.5 equivalent of NaI was proved to be the best choice, providing the desired arylation product in 66% yield with slightly higher enantioselectivity (93% ee, entry 8) while NiCl2•DME or NiBr2 decreased the reactivity of this coupling process obviously. (Table 1, entries 9 and 10). Finally, extending the reaction time to 48 h, the α-CF2H amide could be obtained in acceptable yield without diminishing the enantioselectivity (Table 1, entries 11; 88% yield, 93% ee).

Table 1 The optimization of reaction conditionsa

Enantioselective Construction of Difluoromethylated Tertiary Stereocenters by Nickel-catalyzed C-C Coupling Reaction

With the optimized conditions established, we proceeded to investigate the substrate scope of this enantioselective Negishi cross-coupling. As illustrated in Fig. 2, a diverse array of arylzinc reagents was initially examined in this reaction, coupling with 1a to yield products with excellent enantioselective control (Fig. 2). Notably, both electron-withdrawing groups such as ester (3a, 3b), trifluoromethyl (3c, 3d), fluoro (3e), chloride (3f) and electron-donating groups, including OMe (3g, 3h), OBz (3i), methyl (3k), phenyl (3m) were well tolerated under the standard conditions. Besides, fused ring derivatives such as phenyl (3n) and naphthyl (3o), were also smoothly difluoroalkylated to afford the desired products. Furthermore, Polysubstituted benzenes were successfully obtained under these optimized conditions(3p, 3q, 3r, 3s), yielding 59–75% with 90%-95% ee. Notably, the Rivastigmine intermediate(3j) was synthesized with a yield of 61% and an ee of 95%.

Fig. 2: Scope of substrate.
figure 2

Reaction conditions were as follows: 1 (0.2 mmol, 1.0 equiv), 2 (0.36 mmol, 1.8 equiv), NiBr2·DME (0.02 mmol, 10 mol%), L9 (0.026 mmol, 13 mol%), NaI (0.1 mmol, 0.5 equiv), 2-MeTHF (1.0 mL), Ar, −20 °C. Isolated yields. The ee values were determined by HPLC analysis on a chiral stationary phase.

Next, CF2H-substituted alkyl chlorides 1 were investigated in this Negishi cross-coupling reaction. As shown in Fig. 2, secondary α-CF2H chlorides with electron-withdrawing substituents were examed, such as fluorine (3t, 3u), chloride (3v,3w), bpin(3x) and CF3(3y), furnishing the coupling products with excellent ee values. Additionally, electron-donating groups such as OMe(3z), TMS (4a) and tBu (4b) were also tested, yielding corresponding products with good yields and enantioselectivity. Notably, this reaction is compatible with 3,4-dihydro-2H-isoquinolin-1-one (4cb4h), resulting in moderate yields (55–75%) and excellent ee values (97–98%).

Motivated by the favorable functional group compatibility demonstrated in our reactions, we set out to explore the potential application of this asymmertic difluoroalkylation in modifying biologically active molecules. Notably, the desired products were obtained with good yields and excellent enantiomeric excess (ee) values in all cases, including (L)-menthol (4i), (L)-borneol (4j), majantol (4k), (S)-ibuprofen (4l), canagliflozin (4m), leaf alcohol (4n), and gemfibrozil (4o), which demonstrates the applicational potential of incorporating CF2H-substituted stereogenic centers in commercially available drugs and natural products.

Synthetic applications

To further illustrate the synthetic utility of this transformation, we conducted a large-scale experiment under the optimized condition. The corresponding product 3b was obtained with a 68% yield accompanied by an excellent ee value (95%) (Fig. 3a). Additionally, given the modification potential of substituting a methyl group with a difluoromethyl group in drug molecules, we achieved a 0.5 mmol-scale synthesis of α-CF2H amide 3j via this asymmetric Negishi cross-coupling, achieving a 58% yield with 95% ee. This intermediate could be further converted into a Rivastigmine analogue in three steps, with an overall yield of 49% while maintaining enantiocontrol (Fig. 3b).

Fig. 3: Synthetic utility and mechanistic studies.
figure 3

a mmol-scale reactions. b synthesis of Rivastigmine analogues. c radical trapping experiment. d control experiment. e proposed mechanism.

Mechanism investigation

To gain a deeper understanding of the mechanism underlying this reaction, a series of control experiments were conducted. Firstly, the reaction was inhibited by the addition of radical scavenger 2,2,6,6-Tetramethylpiperidinooxy (TEMPO), high-resolution mass spectra (HRMS) indicated the presence of 5n, suggesting a radical pathway in the catalytic cycle (Fig. 3c). As illustrated in Fig. 3d, the addition of 0.02 equivalents of TEMPO to the reaction (green line) resulted in an induction period of approximately 40 hours, continuing beyond 48 hours. This observation suggests that a radical process might be involved in the catalytic cycle. Furthermore, the addition of Ni(cod)₂ significantly reduced the induction period induced by TEMPO (red line) while Ni(cod)₂ was unable to initiate the reaction (black line) in the absence of NiBr₂•DME. Based on these results, the Ni(I) species that is created when Ni(II) and Ni(0) react with each other might make it easier for radicals to form.

Based on the experimental results and previous literature57,58,59,60,61,62,63,64,65,66,67,68,69, we hypothesize a plausible reaction mechanism. Firstly, the alkyl radicals E may be generated by the reduction of the alkyl chloride 1 by Ni(I) species (D), and then E is captured by the aryl nickel species to form Ni (III) species C. Finally, the Ni (III) species C undergoes a reduction elimination process to produce product 3 and regenerates Ni(I) species D into the next catalytic cycle (Fig. 3e).

In conclusion, we have developed an efficient and versatile nickel-catalyzed asymmetric Negishi cross-coupling for difluoroalkylation, utilizing designed radical CF2H synthons. This approach allows the synthesis of a diverse array of chiral amines with difluoromethylated stereogenic centers. The method is characterized by straightforward operations, mild reaction conditions, excellent functional group tolerance, and high catalytic activity with exceptional enantioselectivity. It facilitates the late-stage modification of complex bioactive molecules and the synthesis of chiral analogues with CF2H-substituted stereogenic centers from known bioactive amines. This asymmetric radical transformation provides a practical and efficient solution for the synthesis of chiral difluoromethylated drug-like molecules, offering significant potential for drug discovery and development. Ongoing work in our laboratory focuses on the design, synthesis, and application of various fluorine-containing synthons for asymmetric radical fluoroalkylation to enable the rapid and efficient construction of biologically active chiral fluorinated compounds.

Methods

General procedure C for enantioselective construction of difluoromethylated tertiary stereocenters by Nickel-catalyzed C–C coupling reaction

NiBr2•DME (10 mol%, 0.02 mmol), L9 (13 mol%, 0.026 mmol) were firstly combined in a 25 mL oven-dried sealing tube. The vessel was evacuated and backfilled with Ar (repeated for 3 times), 2 mL 2-MeTHF was added via syringe and the complex was allowed to pre-stir at 25 °C for 30 min. Then CF2H-substituted secondary alkyl chloride 1 (1.0 equiv, 0.20 mmol) and NaI (0.5 equiv. 0.10 mmol) was added and the tube was cooled to −20 °C. And arylzinc reagent 2 (1.8 equiv, 0.36 mmol) was added dropwise. The tube was sealed with a Teflon lined cap and stired at −20 °C for 48 h. The reaction mixture was then diluted with EtOAc (~10 mL) and filtered through a pad of celite. The filtrate was added to brine (20 mL) and extracted with EtOAc (3 × 15 mL), the combined organic layer was dried over Na2SO4, filtrated and concentrated under vacuum. The residue was then purified by flash column chromatography to give the desired products.