Introduction

Bimetallic active sites in both enzymes and heterogeneous catalysts are well known to be effective in mediating challenging and highly valuable reactions, which is largely attributed to synergistic effects between metals in close proximity1,2,3,4. However, achieving a fundamental understanding on the bifunctional role at the atomic and molecular level is very challenging due to the structural complexity of the catalytic sites in metallic surfaces and metalloenzyme. Over the last several decades, many efforts have been devoted to modelling these multimetallic activation processes for discovery of novel catalysts, primarily using dinuclear metal complexes in homogenous catalysis5,6,7,8,9,10,11,12,13,14. In this context, several types of bimetallic interaction models (e.g., co-activation of a single reactant by ligation with both metal nuclei, and synergistic activation of both reactants on distinct metal centers etc) have been proposed to account for the significant binuclear synergistic effects. Compared to the classical mono-metallic catalysis wherein all the element steps generally occur on one metal site (Fig. 1a, left), binuclear metallic catalysis offers a unique approach wherein two metal sites may independently promote distinct elementary steps in a parallel way during the initial stage of the catalytic cycle, resulting the formation of new binuclear catalytic intermediates, which thus may lead to unusual reactivity unattainable with a typical mononuclear catalyst (Fig. 1a, right). This would be highly desirable for the development of new reactions and further exploration of novel binuclear catalytic mechanisms15,16,17.

Fig. 1: Challenges and strategies in transition-metal-catalyzed hydrocarbofunctionalization of alkynes.
figure 1

a Synergistic binuclear catalytic model. b Current status of transition-metal-catalyzed hydrocarbofunctionalization of alkynes. c This work: Binuclear Ni-catalyzed dimerization-hydrocarbofunctionalization of alkynes.

Full size image

Alkynes are one of the most readily available feedstocks and versatile synthons18,19,20,21,22,23. Transition metal-catalyzed hydrocarbofunctionalization of alkynes using organometallic reagents has emerged as one of the most reliable strategies to access a wide range of highly substituted olefins (Fig. 1b)24,25,26,27,28. However, the reactions developed so far generally result in products with only a single equivalent of alkyne incorporated, and dimerization-hydrocarbofunctionalization of unsymmetrical internal alkynes with organometallic reagents remains elusive (mainly as side-reactions or giving a mixture of isomers)29,30,31,32,33,34,35, even though this may constitute an attractive strategy for quick access to synthetically versatile yet hardly prepared poly-substituted 1,3-dienes36,37,38,39,40,41,42,43. Controlling the degree of alkyne insertion as well as regio- and stereoselectivity constitutes nontrivial challenges, especially for an unsymmetrical internal alkyne via a general vinyl-metal intermediate or metallacyclopentadiene intermediate44,45,46. Thus, novel catalytic modes involving distinct intermediates from the classic mono-metallic species may need to be invoked for alkyne dimerization-hydrocarbofunctionalization processes, which eventually allow for efficient access to highly substituted 1,3-dienes.

Herein, we report dimerization and hydroarylation/hydroalkylation of unsymmetrical internal alkynes with organoborons using a dinuclear nickel complex as the catalyst, providing a series of pentasubstituted 1,3-dienes with excellent regio- and stereoselectivities (Fig. 1c). The versatile synthetic utility, broad functional group compatibility, and excellent chemo- and regioselectivity are demonstrated by substrate scope evaluation and product derivatization. In addition, the protocol can be applied to a one-pot synthesis of valuable yet complicated bridged bicyclic scaffolds that are otherwise difficult to prepare. Notably, control experiments using several mononuclear Ni catalysts only afford low yields of the corresponding hydroarylation products with a single equivalent of alkyne being incorporated, demonstrating the unique reactivity and distinct selectivity pattern of the current dinuclear Ni catalyst. Mechanistic studies suggest that two vinyl Ni moieties are generated intramolecularly on two Ni atoms, respectively, followed by a reductive elimination to form the final product. By analogy to “convergent strategies in total syntheses”47,48, it is tempting to tentatively name this catalytic model as “binuclear convergent catalysis” in homogeneous catalysis, since binuclear core enables parallel substrates activation via distinct elementary steps on two metals and cooperative intermediate transfer across dual metal centers to deliver the final product, whereas monometallic catalysis is generally confined to sequential reaction pathways mediated solely through a single active site.

Results

Reaction development

Bis pyridyl diimine (PDI) ligands49,50,51 based binuclear metal complexes have been used as catalysts in olefin polymerizations52 and studied in activation of small molecules53,54. Inspired by the striking potency of binuclear metallic catalysis5,6,7,8,9,10,11,12,13,14, our group has developed a macrocyclic bi-PDI based dicobalt complex-catalyzed stereodivergent semi-reduction of alkynes55 and a dinickel complex-catalyzed Z-selective 1,4-hydroarylation of 1,3-dienes56. As our continuous interest in developing new reactions using binuclear metal complexes as catalysts, we aimed to investigate the coupling of organoboronic acids with alkynes using binuclear Ni complexes as the catalysts. The study was initiated using unsymmetrical internal alkyne, 1-phenyl-1-propyne 1a and phenylboronic acid 2a as the model substrates.

After a systematic screening of the reaction parameters (for details, see Supplementary Tables 1–10), we found that a mono-ene 3a and a pentasubstituted 1,3-diene product 4a were obtained in 9% and 84% GC yields, respectively, in the presence of the dinickel complex A, NaBHEt3 and t-BuONa in a mixed solvent of MeOH and DMF at 80 °C (Table 1, entry 1). Lower efficiency was observed when the reaction was carried out at 60 °C (entry 2). Only trace amount of the target product 4a was detected without MeOH, suggesting that the alcohol is crucial for the reaction (entry 3). Of note, the length of methylene linkers (where n = 2–4) of the catalyst proved to be a critical factor for the reaction efficiency. Compared to dinickel complex A (n = 3), the reaction using either dinickel complexes B (n = 2) or C (n = 4) led to the formation of 1,3-diene product 4a in significantly lower yields (entries 4 and 5), suggesting that the Ni-Ni distances of the binuclear Ni complexes may influence the reaction outcomes, both in terms of activities and selectivities. Moreover, the catalytic performance of the dinuclear Ni complex was evaluated in comparison with a series of mono-nickel complexes bearing structurally related N-donor ligands, such as PDI (D-G), bipyridine (L1), bis-oxazoline (L2), and pyridine bis-oxazoline (L3). In these cases, though varying degrees of alkyne consumption was observed, no significant amount of 1,3-diene product 4a was detected (entries 6–12). These results further attested the unique catalytic activity and distinct selectivity of the dinuclear catalysts.

Table 1 Optimization of the reaction conditionsa
Full size table

Reaction scope

With the optimized reaction conditions in hand, we set out to explore the generality of this methodology, by applying various aryl, alkyl-disubstituted internal alkynes in the reaction with phenylboronic acid 2a using dinickel complex A as the catalyst. The effect of substituents on the aromatic ring on the reaction outcomes was shown in Fig. 2a. Both electron-rich (methyl 1b-1c, tert-butyl 1d, methoxy 1e and methylthio 1f) and electron-deficient (fluoro 1g-1h, chloro 1i, bromo 1j and ester 1k) alkynes were well compatible with the protocol, to deliver the corresponding conjugated diene products (4b-4k) in 60–79% yields. It should be noted that the reaction of ortho-substituted aryl alkyne with large steric hindrance also proceeded smoothly to give the desired product 4h in 49% yield. Moreover, multi-substituent alkynes (1l-1n) and 2-naphthyl alkyne 1o underwent the hydroarylation reaction smoothly, resulting in the formation of 1,3-diene products 4l-4o in 52–83% yields. Importantly, this method was found to be compatible with an array of heteroaromatic alkynes, including indole, benzofuran, thiophene, pyrazole and ferrocene, furnishing the corresponding products (4p−4t) in good yields. Subsequently, the reaction scope with respect to the alkyl moieties of the aryl-alkyl acetylenes was also evaluated. Alkyne substrates bearing different linear alkyl substituents, including Et- and nPr-, and cyclic substituent, cyclopropyl, were well compatible in the reaction, albeit with moderate yields of the corresponding products (4u-4w). Of particular interest is that some functional groups on the alkyl chain, such as ester, chloro, phthalimide and free hydroxyl were well tolerated in the reaction, furnishing penta-substituted diene products 4x-4aa in good yields. In addition, diphenylacetylene was also proved to be a suitable substrate for this transformation, giving the product 4ab in 77% yield, with an olefin bond configuration different from that of the mono-Ni catalyzed transformation33. Under the standard conditions, the reaction of dialkyl acetylenes only led to mono hydroarylation products (for failed examples, see Supplementary Information Page S35). The structure of 4a was unambiguously confirmed by X-ray analysis.

Fig. 2: The scope of the dimerization-hydroarylation of internal alkynes.
figure 2

a Scope of alkynes. b Scope of arylboronic acids. Reaction conditions A: alkyne 1 (0.40 mmol), arylboronic acid 2 (0.36 mmol), dinickel complex A (5.0 mol%), NaBHEt3 (20 mol%) and t-BuONa (30 mol%) in MeOH (1.0 mL)/DMF (2.0 mL) at rt for 1.0 h and then 80 °C for 12 h. Yields shown are of isolated products. For 4h and 4w, the reactions were conducted at 100 °C. PMP p-methoxyphenyl, NPhth N-Phthalimido, TMS trimethylsilyl.

Full size image

Encouraged by these results, we further examined the scope of arylboronic acids (Fig. 2b). The reactions of a series of para- and meta-substituted arylboronic acids, such as methyl, tert-butyl, trimethylsilyl, fluoro, chloro and ester, all afforded the desired 1,3-diene products 4ac-4ai in satisfactory yields with excellent regio- and stereoselectivity. In addition to monosubstituted arylboronic acids, polysubstituted arylboronic acids were also compatible with the procedure to give the corresponding products (4aj, 66% and 4ak, 33%). It is worth mentioning that boronic acids derived from several heterocycles, such as furan, benzofuran, dibenzo[b,d]thiophene, and carbazole, were also competent coupling partners, affording the corresponding pentasubstituted diene products 4al-4ao in moderate yields (42–62%).

To further expand the use of this methodology, we proceeded to examine the reaction of trialkylborane with alkyne using dinickel complex A as the catalyst. Under modified conditions, the reaction of 1-phenyl−1-propyne 1a and triethylborane 5a led to the desired ethyl-substituted 1,3-diene product 6a in 70% isolated yield (for details, see Supplementary Tables 11–14). Next, the substrate scope with regards to alkynes was examined (Fig. 3). Various aryl alkyl alkynes were amenable to this transformation, giving rise to the corresponding pentasubstituted diene products 6b-6m in good yields with high regio- and stereoselectivities. Thereafter, the scope of this reaction with other organoboranes was further explored. Commercially available tri-n-butylborane reacted well to furnish 6n in 54% yield. Additionally, the reactions using unpurified trialkylboranes (prepared from 1-hexene and 1-octene with BH3·SMe2) also gave 6o and 6p smoothly.

Fig. 3: The scope of the dimerization-hydroalkylation of unsymmetrical internal alkynes.
figure 3

Reaction conditions B: alkyne 1 (0.40 mmol), trialkylborane 5 (0.40 mmol), dinickel complex A (5.0 mol%), NaBHEt3 (30 mol%) and t-BuOLi (50 mol%) in MeOH (10 equiv)/DMF (0.30 mL) at rt for 1.0 h and then 80 °C for 12 h. Yields shown are of isolated products. For 6n-6p, B2Pin2 (10 mol%) was used instead of NaBHEt3.

Full size image

This strategy was subsequently applied into the functionalization of some alkynes bearing biologically relevant skeletons (Fig. 4a). The reactions of a range of internal alkynes derived from (-)-borneol, carbofuran phenol, canagliflozin, dapagliflozin, hymecromone, L-tyrosine or estrone with phenylboronic acid 2a proceeded smoothly to give the corresponding pentasubstituted diene products 4ap-4av in good yields with excellent regio- and stereoselectivity, respectively. All these results further underscored the good functional group compatibility of the reaction, attesting the robustness and practicality of the titled reaction.

Fig. 4: Late-stage functionalization and synthetic applications.
figure 4

a Transformations of complex alkyne substrates. b Transformations of products. c One-pot synthesis of bridged bicyclic compounds. Boc, t-butoxycarbonyl.

Full size image

Furthermore, a series of synthetic elaboration of the products 4a and 6a were conducted to showcase the synthetic versatility of the methodology (Fig. 4b). For examples, 4a was quantitatively converted to substituted indene derivative 7 by methanesulfonic acid (MeSO3H)-mediated intramolecular Friedel-Crafts cyclization, while Diels-Alder reaction of 4a with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) and 4-phenyl-1,2,4-triazoline-3,5-dione delivered bicyclic products 8 and 9 in 80% and 92% yields, respectively. In addition, epoxidation of 6a with varying amounts of 3-chloroperbenzoic acid (m-CPBA) gave mono-epoxide 10 or di-epoxide 11, respectively, in moderate yields. It is noteworthy that sequential hydroarylation and double intramolecular Friedel-Crafts cyclization of two equivalents of alkyne 1 can be successfully achieved in one-pot via di-Ni catalysis followed by a treatment with trifluoromethanesulfonic acid (TfOH), leading to a series of biologically interesting dibenzobicyclo-[3.2.1]octadienes 12 in remarkable efficiency (Fig. 4c).

Experimental and theoretical mechanistic studies

To gain some insight into the reaction mechanism, a series of experiments were conducted (Fig. 5). The use of 2,6-di-tert-butyl-4-methylphenol (BHT, 1.0 equiv), a radical scavenger, in the reaction of alkyne 1a and phenylboronic acid 2a under the standard conditions showed almost no adverse effect on the formation of the desired product 4a, suggesting that the reaction was most likely not a radical-based process (Fig. 5a). Next, deuterium-labeling experiment was carried out to probe the origin of the hydrogen source for the hydroarylation. As shown in Fig. 5b, submission of 1a and arylboroxine 13 to the reaction using CH3OD resulted in 4a-D with 90% deuterium incorporation at the olefinic carbon position, indicating that methanol proton acts as the hydrogen source in the reaction. Moreover, kinetic isotope effect (KIE) was measured on the reaction of 1a and 2a in MeOH and MeOD, respectively, to deliver a kH/kD value of 1.3 (Fig. 5c), suggesting that proton transfer is unlikely to be the rate-determining step. Additionally, a competitive reaction of two electronically reversed alkynes, 1e and 1k, with 2a was run under the standard conditions (Fig. 5d). In this case, the homo-coupling products 4e and 4k were isolated in 17% and 45% yields, respectively, along with the formation of a 34% yield of cross-coupling product 4aw. These results suggested that alkyne with electron-deficient substituent on the phenyl ring is more favorable in the competition reaction. Then, kinetic orders of the reaction were studied under the standard reaction conditions (Fig. 5e). The rate data show a linear dependence on the concentration of the dinuclear nickel catalyst A and alkyne 1a, but a zero-order dependence is observed for arylboronic acid 2a. These results suggested the involvement of one catalyst molecule and one alkyne molecule in the rate-determining step, and the binuclear Ni structure remains largely intact during the catalytic process. Furthermore, Hammett studies were performed to investigate the electronic effect of substituents appended on the alkynes and arylboronic acids, respectively (Fig. 5f). A linear correlation between log(kX/kH) and the Hammett constant (σpara) with a slightly positive slope of 0.157, indicating that the effect of electronic effects of alkyne on the reaction rate is not significant. Moreover, a ρ value of −1.455 was obtained for a series of para-substituted arylboronic acids, suggesting that the presence of an electron-rich group on arylboronic acid is favorable for the reaction.

Fig. 5: Experimental mechanistic studies.
figure 5

a Radical inhibition experiment. b Deuterium-labeling experiment. c KIE experiment. d Competition experiment. e Kinetic order studies. f Hammett studies.

Full size image

To further unveil the mechanistic details of this multicomponent reaction, density functional theory (DFT) calculations were further performed at SMD-M06L-D3/def2-TZVP//M06L/def2-TZVP/def2-SVP level of theory (see detailed introduction in Supplementary Information). A formally reduced [(PDI)Ni0]2 ([Ni2]) is firstly generated via in situ reduction of the catalyst precursor A (Fig. 6)56. Based on our previous work, herein the formal [(PDI)Ni0]2 can be described more appropriately as [(PDI−)NiI]2, a dinuclear Ni(I) complex in ground state of triplet56, which facilitates the activation of alkyne 1a (1INT1A, −22.7 kcal/mol) rather than methanol (3INT1C, −13.3 kcal/mol, see detailed mechanism of Path C in Supplementary Fig. 15). Oxidative cyclization of the alkyne π bond across the two Ni centers forms a four-membered bimetallic intermediate 1INT1A with the C1−C2 bond length being elongated to 1.34 Å, which is close to a C=C double bond of alkenes (1.33 Å)45. This similar reactivity and coordination model was disclosed on the addition of diphenylacetylene to a neutral bis-pyridyldiimine dicobalt complex by Tomson group57. In addition, we failed to locate the side-on mode (µ-η2:η2) of alkyne with the C≡C triple bond perpendicular to the Ni-Ni direction58,59, which would spontaneously transfer to the end-on mode (µ-η1:η1) 1INT1A, possibly due to the steric linkage (-C3H6-) of PDI ligands. Therefore, the reaction may be initiated by the oxidative addition of the first alkyne triple bond across the dinuclear Ni core (1INT1A). The viability of synergistic oxidative cyclometallation of two alkynes on 3[Ni2] to form 3INT2B was also assessed as Path B58,59, however, it was found unfavorable due to the much higher energy barrier (∆G‡ = 32.2 kcal/mol, Supplementary Fig. 16). More pathways can be envisioned starting from 1INT1A to generate the intermediate 3INT4A or 3INT4B, depending on the order of protonation, transmetalation and alkyne insertion (Paths A1–A4 in Fig. 6). The protonation of 1INT1A with MeOH to give 1INT2A would be a pathway more favorable than the alkyne insertion of 1INT1A with the second alkyne (1a) on di-Ni core to 3INT2B in Path A1 (see detailed mechanism in Supplementary Fig. 14). After the formation of intermediate 1INT2A, the direct transmetalation to form 1INT3A needs to overcome an energy barrier of +21.9 kcal/mol, which is more accessible than the second alkyne (1a) insertion to form 3INT3B (Path A2, 41.8 kcal/mol via 3TS2B in Supplementary Fig. 14). However, the following alkyne insertion into the Nia-C2 bond of 1INT3A as depicted in Path A3 is associated with a rather high energy barrier (32.5 kcal/mol), suggesting that the butadienyl intermediate 3INT4B might not be accessible. Alternatively, the second alkyne (1a) can be selectively inserted into the Nib-Cph bond at the C4 site with a favorable energy barrier (∆G‡ = 28.7 vs 33.1 kcal/mol at C3 site, path A4), resulting in two different vinyl-Ni moieties within intermediate 3INT4A and finally generates the diene products through a bimetallic reductive elimination. Detailed free energy profiles were depicted in Fig. 7.

Fig. 6: Possible reaction pathways for the dimerization-hydroarylation of alkynes.
figure 6

The Positive values (kcal/mol) are free energy barriers, and the negative one (kcal/mol) is the free energy change of the reaction. NPA charges are shown in purple and bond length (Å) in black. IGMH analysis of 3INT4A with δginter = 0.004. For DFT calculation results on unfavored pathways, see Supplementary Figs. 14–16.

Full size image
Fig. 7: Calculated mechanism of dimerization-hydroarylation of alkynes.
figure 7

See unfavored pathways of singlet states in Supplementary Fig. 17; NPA charges are shown in purple; Spin population in green and bond length (Å) in black. IGMH analysis of 3TS3A with δginter = 0.004. MECP the minimum energy crossing points.

Full size image

The calculated pathway with the most favorable energy profile was found to consist of species mostly occurring in the triplet state, and spin populations of Nia and Nib in the relevant intermediates and transition states may vary throughout the reaction course, and some of intermediates were found to adopt singlet as their ground states (See unfavored pathway of singlet states in Supplementary Fig. 17). Oxidative addition of alkyne (1a) with two Ni centers would result in the formation of a bimetallic intermediate 1INT1A via a MECP with a barrierless process (see Supplementary Fig. 18 for further discussion), which could react with MeOH via a binuclear concerted protonation process to give 1INT2A. NPA charges of 1a and 1INT1A indicated that binuclear Ni2 core acts as an electron-donor to achieve two-electron transfer to the triple bond of 1a, and hence the accumulated electron density on C1 rather than C2 in 1INT1A would favor the capture the proton of MeOH by C1 (20.4 kcal/mol in 3TS1A vs 21.9 kcal/mol in 3TS1A’), which is consistent with the regioselective protonation of C1 site as revealed by the deuterium labelling experiment (Fig. 5b). Upon generation of 1INT2A, the methoxy group in 1INT2A readily associates with phenylboronic acid 2a, which undergoes transmetalation on Nib (21.9 kcal/mol in 3TS2A) under the triplet state, which is more stable than the corresponding singlet state species (40.3 kcal/mol in 1TS2A, see Supplementary Fig. 17). After release of MeOB(OH)2 in 1INT3A, the following reductive elimination of the aryl and alkenyl on Nia and Nib, respectively, has to overcome a relatively high energy barrier of 30.2 kcal/mol, which would prevent the formation of the hydroarylation product 3a with incorporation of only one alkyne (for details, see Supplementary Fig. 19). Alternatively, through the ligand exchange of the second alkyne 1a with MeOB(OH)2 ligand of 1INT3A, the second alkyne insertion of Nib-Cph bond proceeds via transition state 3TS3A to form intermediate 3INT4A featured by π-π interactions between styrenyl-Nia and the phenyl group from the second alkyne as well as the π-π interactions between PDI and the phenyl group from phenylboronic acid (IGMH analysis in Fig. 6). The activation barrier for the migratory insertion of the phenyl group to C4 site was 4.4 kcal/mol lower than that for C3 site (28.7 kcal/mol in 3TS3A vs 33.1 kcal/mol in 3TS3A’), which is consistent with the regioselectivity observed in the reaction. Closer inspections of 3TS3A revealed that the electronic interaction between aryl group and alkyne substrate is the key to account for this regioselectivity of the insertion step along with the π–π stabilization of two aryls, which was further supported by the energy decomposition analysis (EDA, in Supplementary Fig. 20) and IGMH of 3TS3A in Fig. 7. This is also in agreement with the Hammett studies that electron-rich arylboronic acid is favorable for the reaction. For the following reductive elimination step, we failed to locate the seemingly plausible concerted bimetallic transition state (Nia-C2-C3-Nib) from intermediate 3INT4A. Instead, calculations indicated that the reductive elimination would occur on a single Ni site (Nia) of 3INT5A, formed by migration of the diphenyl substituted alkenyl group from Nib to the low-coordinated Nia center in 3INT4A, a process facilitated by the dissociation of one N-Nia coordination bond in PDIa (Nia-N1 = 3.21 Å). The activated Nia of 3INT5A favors the acceptance of electrons by its empty d orbitals to facilitate reductive elimination of two alkenyl groups via a three-membered ring transition state 3TS4A (26.2 kcal/mol). Finally, upon the ligand exchange with substrate 1a would release the product 4a and regenerate the active 1INT1A for the next catalytic cycle. Over the catalytic cycle, the two Ni centers of the dinuclear nickel complex can mediate the individual steps of transformation of two alkynes, respectively, and ultimately achieve double alkynes incorporation and hydroarylation efficiently with high regioselectivity in a “bimetallic convergent catalysis” manner. DFT calculation suggested that the insertion of the second molecule of alkyne should be the rate-limiting step, which is consistent with the experimentally determined first-order kinetic dependence of 1a together with the Hammett studies. Alkyne with electron-deficient substituent on the phenyl ring is more favorable in the competition experiment (Fig. 5d), which might be due to its favorable coordination with low-valent di-Ni core during the generation of the intermediate 1INT1A.

Based on the afore mentioned mechanistic studies, a plausible mechanism was depicted in Fig. 8. After reduction of di-Ni complex A with NaBHEt3, the resulting binuclear complex 3[Ni2] activates the first alkyne via oxidative cyclization to produce a bimetallacycle intermediate 1INT1A, allowing for a binuclear concerted protonation by MeOH to the C1 position of 1a. The formed intermediate 1INT2A then undergoes transmetalation with PhB(OH)2 on Nib and the second alkyne insertion to generate the di-vinyl di-Ni intermediate 3INT4A rather than the direct hydroarylation with one alkyne. Finally, a migration of the diphenyl substituted alkenyl group occurs to generate 3INT5A for the next reductive elimination, delivering the product 4a. Benefited from the binuclear core, the catalytic cycle can be regarded as a manner of “binuclear convergent catalysis”, wherein two alkynes are activated and transferred to the corresponding alkenyl-Ni on two different Ni atom: Nia is responsible for the activation and transfer of the first alkyne to alkenyl-Nia, and arylation of the second alkyne occurs on the Nib to generate the other alkenyl-Nib. Then the reductive elimination of these two “activated moieties” on binuclear Ni core generates the coupling product. The arylation of the second alkyne on Nib plays a key role in the current dimerization and hydrocarbofunctionalization of alkynes over that of mono-alkyne. The electronic effect of the alkyne insertion with aryl group from boronic acid and the π–π interaction of two aryls regulated the activity and regioselectivity.

Fig. 8: Proposed mechanism.
figure 8

Proposed catalytic cycle for binuclear Ni-catalyzed dimerization-hydrocarbofunctionalization of alkynes.

Full size image

Discussion

In summary, we have developed an efficient dimerization and hydrocarbofunctionalization of alkynes with organoborons using a bis-PDI dinickel complex as the catalyst, affording a wide spectrum of pentasubstituted 1,3-dienes with excellent regio- and stereoselectivities. This protocol exhibits high functional group tolerance and broad substrate scope, and the resulting products are amenable to a variety of transformations, especially for the efficient synthesis of bridged bicyclic compounds. Experimental and theoretical mechanistic studies suggested that each Ni site can facilitate distinct elementary steps of two alkynes to generate a key di-vinyl di-Ni intermediate, followed by a reductive elimination to release the final product, which is substantially different from classic mono-metallic catalysis. This work may stimulate the development of novel reactions via binuclear catalysis and provides an improved understanding of the binuclear catalytic mechanisms in industrial / biocatalytic systems.

Methods

Procedure for dimerization-hydroarylation of alkynes

Under argon atmosphere, arylboronic acid (0.36 mmol, 1.8 equiv), dinuclear nickel complex A (10.8 mg, 5.0 mol%) and tBuONa (5.7 mg, 30 mol%) were added to an oven-dried 10 mL Schlenk tube equipped with a magnetic stir bar. Then the Schlenk tube was moved to a nitrogen-filled glovebox, alkyne (0.40 mmol, 2.0 equiv), MeOH (1.0 mL), DMF (2.0 mL) and NaBHEt3 (40 μL, 20 mol%, 1.0 M in THF) were added sequentially. After that, the tube was sealed and moved out of the glovebox. The reaction mixture was stirred at room temperature for 1.0 h, then heated to 80 °C and stirred for 12 h. After cooling to room temperature, saturated brine was added, and the reaction mixture was extracted with ethyl acetate. The organic layer was concentrated in vacuum and the residue was purified by chromatography on silica gel to give the product.

Procedure for dimerization-hydroalkylation of alkynes

Under argon atmosphere, dinuclear nickel complex A (10.8 mg, 5.0 mol%) and tBuOLi (8.0 mg, 50 mol%) were added to an oven-dried 10 mL Schlenk tube equipped with a magnetic stir bar. Then the Schlenk tube was moved to a nitrogen-filled glovebox, alkyne (0.40 mmol, 2.0 equiv), trialkylborane (400 μL, 2.0 equiv, 1.0 M in THF), MeOH (81 μL, 10 equiv), DMF (0.30 mL) and NaBHEt3 (60 μL, 30 mol%, 1.0 M in THF) were added sequentially. After that, the tube was sealed and moved out of the glovebox. The reaction mixture was stirred at room temperature for 1.0 hour, then heated to 80 °C and stirred for 12 h. After cooling to room temperature, saturated brine was added, and the reaction mixture was extracted with ethyl acetate. The organic layer was concentrated in vacuum and the residue was purified by chromatography on silica gel to give the product.