Introduction

The merger of photoredox catalysis with transition metal catalysts, named as “metallaphotoredox”, has become a popular strategy to construct C–C bonds in the past decade1. This dual catalytic platform has led to many new discoveries, which were unfeasible or difficult to achieve with a single catalytic system2,3,4,5,6. Such cooperative regime, especially with nickel, provides a novel catalytic mode in asymmetric cross-couplings under exceptionally mild conditions and avoids the use of air-sensitive organometallic reagents or stoichiometric metal reductants (Fig. 1a)7,8. Currently, dual Ni/photoredox-catalyzed asymmetric couplings typically involve redox-neutral methods, such as Suzuki–Miyaura reactions, but few reductive cross-couplings have been reported9,10,11,12,13,14,15. We speculated that the application of dual catalytic system in asymmetric reductive cross-couplings could enable the use of various electrophiles and increase the range of accessible reaction types16,17,18. In addition, more ligand selections are likely to achieve high enantioselectivity in the absence of stoichiometric metal reductants because of circumventing the possible coordination between the metal reductant and the ligand.

Fig. 1: TM-catalyzed asymmetric cross-couplings.
figure 1

a Catalytic modes on asymmetric cross-coupling. b The routes to access chiral benzylic boronic esters via coupling. c Ni/photoredox-catalyzed asymmetric process. (TM: transition metal; PC: photoredox catalyst).

However, the challenges also remain. For instance, rigorous coordinative control over several reduction cycles is needed to avoid homo-couplings or other side products. Furthermore, because of the high tendency to protodehalogenation of both halide components in the presence of reducing reagents (amines, Hantzsch ester (HEH)), the efficiency is limited. Based on these challenges, the application of this strategy upon the enantioselective reductive cross-couplings is still very less to date.

As a privileged structural class, chiral benzylic boronic esters are important precursors in organic synthesis due to their versatilities on the stereospecific conversion to many kinds of bonds19. In addition to the ease of handling and high air- and moisture-stability, many methods have therefore been devised to access enantioenriched benzylic boronic esters, such as hydroboration of alkenes20,21,22,23 or alkynes24, and others25,26,27,28. Meanwhile, asymmetric cross-couplings are also very popular (Fig. 1b). A palladium-catalyzed asymmetric coupling of bis(pinacolato)diboron (B2Pin2) with aryl halides, mainly bearing electron-donating groups, was reported by the Morken group29. A nickel- and a copper-catalyzed enantioselective borylation of secondary benzylic chlorides with B2Pin2, were published by the Fu30 and Ito31 group, respectively. However, in these cases, the enantiomeric excess (ee) values of the products were generally not ideal (<87%). Although several racemic couplings under reductive conditions have been developed32,33,34,35,36, to the best of our knowledge, asymmetric reductive cross-coupling with the dual catalytic system constitutes a potentially powerful yet unreported route to construct chiral benzylic boronic esters. Herein, we report a Ni/photoredox-catalyzed asymmetric coupling with broad substrate scopes and excellent enantioselectivities (basically above 90% ee) by using a simple, commercially available ligand (Fig. 1c).

Results

Reaction conditions development

We commenced our investigation of this asymmetric coupling with the reaction between 4-iodobenzotrifluroride (1a) and α-chloroborane (2a) with dual Ni/photoredox catalysts37. Systematic evaluation of all the reaction parameters indicated that the transformation proceeded smoothly under the conditions listed in entry 1, Table 1, to give the product 3a in 94% yield with 93% ee (Table 1, entry 1). Other Ni catalysts gave lower conversions (entries 2–3). The use of other ligands (L2, L3), especially the bis(oxazoline) ligand L4, which is particularly common in cross-couplings30, resulted in diminished outcomes (entry 4–6). Examination of other photocatalysts or solvents identified a rather significant decrease in yields and/or enantioselectivities. (entries 7–10, see supporting information). Although the exact reasons remain elusive, the mixed solvent system possibly increased the dielectric constant of the reaction medium30,38, thus improving the catalytic efficiency and suppressing the protodehalogenation product. A trace amount of product was detected in the absence of Et3N and a low yield was obtained with Li2CO3 (entries 11–12). Considering Et3N can act as a reductant in photoredox-catalyzed reactions, it is reasonable that the reaction still worked without HEH, albeit with lower efficiency (entry 13). The reaction can proceed for one or fewer catalytic cycles without the photocatalyst, which may suggest a mechanistic scenario more complex than previous proposition (entry 14). Control reactions performed in the absence of Ni catalyst or light resulted in no detectable product formation, confirming the essential role of each of these components in the dual catalytic process (entry 15).

Table 1 Optimization of reaction conditions.

In addition, perhaps most notable, the comparison reactions with other common reductive systems were tested to identify the difference and powerful efficiency of this Ni/photoredox-catalyzed coupling (Fig. 2). The results showed that whether the metal reductants Zn/Mn or the organic reagents B2Pin2 and tetrakis(dimethylamino)ethylene (TDAE) gave only traces of the product or no product, thus further strengthening the high superiority of the dual catalytic method on this C–Cl bond transformation.

Fig. 2: The reaction with other reductants.
figure 2

Comparison reaction with other reductive conditions.

Substrate scope

Having developed the enantioselective catalytic reaction of this cross-coupling, we sought to investigate the generality and utility of this dual Ni/photoredox-catalyzed method (Some products were found to be prone to decompose in column, thus isolated yields after oxidation were given. See Table 2). We firstly tested various aryl iodides with either electron-withdrawing or electron-donating groups (the reaction with p-CF3-C6H4Br gave the product 3a in 21% yield with 91% ee under the standard conditions). The excellent enantioselectivities and satisfactory yields were generally given whether the substituents on the para-, meta-, or ortho- positions, in an effort to showcase the mild nature of this transformation. The electron-poor substrates obviously facilitated the couplings, as demonstrated with compounds 3a3f. The reactions with strong electron-rich aryl iodides (3i) led to a decreased but still acceptable results, thus serving to highlight the complementary and discrepant nature of this method compared to the established method with palladium where the two-electron transmetalation regime was involved29. The chemoselectivity with different electrophiles was studied under the current conditions. The exclusive selection on aryl iodide was observed in the presence of aryl chloride or triflate group (3x, 3y, 3z) while a mixture was obtained with 1-bromo-4-iodobenzene as a substrate, possibly due to the existing reactivity of aryl bromides. It is worth mentioning that, this method can afford high enantioselectivities not only for the bulky or ortho-substituted aryl substrates as previous reports29,30,31, but also for the less sterically hindered components (such as 3g, 3h), thus expanding its application in organic synthesis.

Table 2 Scope of Ni/photoredox-catalyzed coupling of aryl iodides with α-chloroboranesa.

The scope of α-chloroborane was then explored in the dual catalytic system. Overall, a wide range of the chloroboronic esters could be smoothly transferred to the corresponding products with very high ee values. For the complex molecules (3nn-3pp), this Ni/photoredox-catalyzed coupling was also suitable, thus demonstrating the potential of this methodology in the late-stage functionalization. The reaction also exhibited comparable efficiency on a larger scale, as the products (3a, 3y) were isolated in good yields and undiminished ee on a 1 mmol scale. In addition, Table 2 also highlighted the robustness on functional group compatibility of this approach. The trifluoromethyl (3a, 3v), trifluoromethoxy (3b, 3p), ester (3c, 3hh, 3ii), amide (3f), nitrile (3e, 3n, 3dd), aryl halides (3d, 3w, 3x), triflate (3z), phenol (3m), ether (3i, 3ee), alkyl chloride (3ff, 3gg), alkenyl (3mm) groups were all compatible under the optimized conditions.

Competing and comparison reactions

To clarify the features on this dual catalytic system, competing reactions were conducted. Although both electron-rich and electron-poor aryl iodides can participate in this transformation, the substrate with an electron-withdrawing group favored when both are present, as evidenced by isolation of product 3a in Fig. 3a. Steric effect was also investigated. When 1d and 1u were used simultaneously, the less sterically hindered component gave the main product 3d.

Fig. 3: Competing and comparison reactions.
figure 3

a The reaction with aryl halides with different properties. b Test on the α-bromo- or α-iodoborane.

α-Bromo- and α-iodoboranes, which are often used in Ni-catalyzed cross-couplings34,39, were also tested (Fig. 3b). Surprisingly, when α-bromoborane was employed, the target product was isolated in only 30% yield with a similar enantioselectivity. A trace amount of product was observed in GC when the more active α-iodoborane was used. The protodehalogenation product was the main product in these two reactions. These results show the importance of concerted rates of reductive processes in this dual catalytic system and the difference from the single catalytic regime.

Further applications

One advantage of this method is that the resulting molecules contain a versatile boron group, which is convenient for further functionalization to construct many other chiral bonds. We demonstrated this in a series of examples in Fig. 4. For example, the boranes can be oxidized to chiral alcohol (4a) or transformed to the potassium trifluoroborate salt (4b) in good yields. The chiral 1,1-diarylalkanes, which are widely useful but not easily accessible, can be conveniently obtained from these products with a Pd-catalyzed coupling reaction (4c). Meanwhile, some other Csp3–Csp2 bonds were readily formed to introduce alkenyl or heterocycles according to the reported methods (4d, 4e)40,41.

Fig. 4: Further application of the products.
figure 4

Derivatization of the chiral benzylic boronic esters.

Mechanistic studies

To give some insights into the mechanism of the coupling, some experiments were performed (Fig. 5). The substrate 2d with an alkene chain was investigated. The cyclization product 5a was isolated in 52% yield and direct cross-coupling 3qq was only detected in 5% GC yield, indicating the production of the radical intermediate. The same conclusion is also deduced from the reaction with radical clock substrate 2e (Fig. 5a). Although it was reported that an alkyl halide can be reduced to a radical species via a single-electron reduction by photocatalysts42, herein, the luminescence quenching experiments indicated that HEH and Et3N are more likely to quench the excited state of 4CzIPN (Fig. 6). The reactions with a radical probe also proved this point (Fig. 5b). In the absence of Ni/L1, the α-chloroborane can not react with 1,1-diphenylethylene 6 to give any radical-trapped product 7. However, in the stoichiometric studies with Ni(cod)2/L1, product 7 can be obtained in 30% yield with no enantioselectivity. Hence, the free alkyl radical should be generated from α-chloroborane by the reduction of Ni species rather than 4CzIPN43,44,45,46.

Fig. 5: Mechanistic studies.
figure 5

a Cyclization reactions. b Reactions with radical probe.

Fig. 6: Stern–Volmer quenching experiments.
figure 6

a With substrate 2a. b With HEH. c With Et3N.

Based on these experiments, a proposed mechanism is shown in Fig. 7. Light excitation of photocatalyst 4CzIPN generates the photoexcited [4CzIPN]*, which is reduced by HEH or Et3N to the reduced ground-state photocatalyst [4CzIPN]·-. Then, combination with the nickel cycle, it is responsible to produce a boron-stabilized radical I. Meanwhile, the oxidative addition of aryl iodide to the ligated Ni(0) II forms a Ni(II) species III. Then oxidation by radical I accesses the high-valent Ni(III) intermediate IV, and subsequent reductive elimination affords the product and releases Ni(I) species V. Reduction of V would regenerate the Ni(0) and 4CzIPN catalyst, closing the dual catalytic cycle. However, at this moment, we can not exculde another catalytic cycle in which the alkyl radical is trapped by Ni(0), then the alkyl Ni(I) species is oxidated by aryl hailde47,48.

Fig. 7: Proposed mechanism.
figure 7

A proposed catalytic cycle for the dual catalytic coupling.

Discussion

In summary, herein a dual Ni/photoredox-catalyzed asymmetric reductive cross-coupling of aryl iodides with α-chloroboranes was developed. The mild conditions enable the process to give excellent enantioselectivities and accommodate a broad palette of architectures with many functional groups. Mechanistic studies suggest the alkyl radical is probably generated via reduction by nickel. The resulting chiral benzylic boronic esters, which have important versatilities in organic synthesis demonstrated the high potential of this transformation.

Methods

General procedure for the dual catalytic coupling

In a N2-filled glovebox, an oven-dried 10-mL Schlenk tube containing a Teflon stir bar was charged with 4CzIPN (0.002 mmol), NiBr2·DME (0.02 mmol), L1 (0.024 mmol), HEH (0.4 mmol), TEA (1.0 mmol). Then the tube was sealed with a septum and taken out. DME/DMA (v/v = 5/1, 4 mL) were added via syringe under N2 atmosphere. After stirring for 30 min at room temperature, aryl halide (0.32 mmol) and α-chloroborane (0.2 mmol) were added. The reaction mixture was stirred and irradiated under blue light (λ = 450–455 nm) for 8–10 h, while the temperature was controlled at 18–25 °C. Upon completed, the mixture was diluted with EtOAc and quenched with water. The aqueous solution was extracted with EtOAc three times. The combined organic layers were dried, concentrated and purified by flash column chromatography using PE/EtOAc as the eluent to afford the coupling product. The ee value was determined by high-performance liquid chromatography analysis using the corresponding alcohol after oxidation of the product.