Differential Dihydrofunctionalization: A Dual Catalytic Three-Component Coupling of Alkynes, Alkenyl Bromides, and Pinacolborane

Abstract

A new method for differential dihydrofunctionalization of terminal alkynes enables the synthesis of allylic boronate esters through reductive three-component coupling of terminal alkynes, alkenyl bromides, and pinacolborane. The transformation is promoted by cooperative action of a copper/palladium catalyst system and results in hydrofunctionalization of both π-bonds of an alkyne. The synthesis of allylic boronate esters can be accomplished in the presence of a wide range of functional groups, including, esters, nitriles, alkyl halides, sulfonyl esters, acetals, protected terminal alkynes, aryl halides, and silyl ethers. Mechanistic experiments reveal the importance of subtle ligand effects on the performance of the palladium co-catalyst.

Graphical Abstract

A new method for differential dihydrofunctionalization of terminal alkynes enables the synthesis of allylic boronate esters through reductive three-component coupling of terminal alkynes, alkenyl bromides, and pinacolborane. The transformation is promoted by cooperative action of a copper/palladium catalyst system and results in hydrofunctionalization of both π-bonds of an alkyne.

Alkynes are extensively used in organic synthesis as readily available and versatile intermediates. They participate in a wide range of transformations, the most common of which are C-H functionalization of terminal alkynes, mono-addition to one of the π-bonds, and double addition to both π-bonds. Significantly less common are reactions that allow differential transformations of the two π-bonds present in alkynes.

Intrigued by the paucity^[1] and the synthetic potential of such transformations,^[2],[3] we became interested in the differential dihydrofunctionalization of alkynes. Our efforts towards this goal have focused on copper hydride chemistry,^[4] which emerged as a powerful tool for the hydrofunctionalization of alkynes.^[5] In addition, the Buchwald^[6] and Mankad^[7] groups have demonstrated that copper-catalyzed hydrofunctionalization of alkynes can be coupled with a subsequent reduction of the remaining π-bond.^[8]

Our approach to differential dihydrofunctionalization of alkynes is outlined in Scheme 1a. Inspired by a report from Sadighi,^[9] we targeted the formation of the heterobimetallic intermediate 2 through the hydrocupration of alkenyl boronate ester 1. Together with well-established copper-catalyzed hydroboration of alkynes,^[10] this transformation provides access to the bimetallic intermediate 2 directly from terminal alkynes (Scheme 1a). Further electrophilic functionalization of this intermediate completes the differential dihydrofunctionalization reaction.

Scheme 1.

Differential dihydrofunctionalization.

The key feature of the strategy outlined in Scheme 1a is that it allows systematic variation of the final functionalization step. In our initial report on differential dihydrofunctionalization of alkynes we described the synthesis of benzylic boronate esters through arylation of the heterobimetallic intermediate (Scheme 1b).^[11] Soon after, Engle and Liu^[12] used electrophilic amination of the heterobimetallic intermediate to develop an elegant and highly effective method for the synthesis of α-amino boranes (Scheme 1c).^[13] Herein, we describe further exploration of the dihydrofunctionalization of alkynes and its application to the synthesis of other classes of boronate esters.

Pursuing the further development of differential dihydrofunctionalization, we decided to focus on the synthesis of allylic boronate esters (Scheme 1d). These compounds have been an attractive target for reaction development^[14] as extremely versatile synthetic intermediates in which the interplay of the alkene and the boronate ester creates unique reactivity.^[15] We recognized that differential dihydrofunctionalization offers a highly convergent approach to this important class of compounds through the coupling of readily available terminal alkynes, alkenyl bromides, and HBpin.

The main challenge in developing this reaction was identifying a method for alkenylation of the heterobimetallic intermediate that can be effective in the context of the overall transformation. We were hoping to accomplish alkenylation of the heterobimetallic complex using a palladium co-catalyst according to the mechanism outlined in Scheme 2.

Scheme 2.

Proposed mechanism.

To test the feasibility of the proposed palladium catalytic cycle we explored catalytic alkenylation of a preformed heterobimetallic complex. An XPhos-supported^[16] palladium catalyst proved effective, providing the product of the desired cross-coupling reaction in 58% yield (Scheme 3a). Even more encouraging were preliminary experiments exploring the catalytic differential hydroboration/hydroalkenylation reaction. Using an XPhos-supported palladium catalyst and the reaction conditions we had previously developed for the synthesis of benzylic boronate esters,^[11] we obtained 32% yield of allylic boronate ester 5 (Scheme 3b).

Scheme 3.

Preliminary results. [a] Yield was determined by GC using 1,3,5-trimethoxybenzene as an internal standard. IPr = 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene. R = p-(CH₃O)C₆H₄(CH₂)₃ and R₁ = Ph(CH₂)₃.

In a subsequent screen, we evaluated the performance of various classes of phosphine ligands using commercially available representatives. With a range of monodentate triaryl and trialkyl phosphine ligands, as well as with a range of bidentate ligands, we observed very little (< 5% yield) of the desired allylic boronate ester. The alkyne was generally consumed in all reactions and was mostly converted into oligomeric materials identified by GPC analysis. The alkenyl bromide was often reduced to an alkene or participated in Suzuki-type cross-coupling with the intermediate alkenyl boronate ester. Although the same side reactions were observed with dialkylbiaryl phosphine ligands, the desired product was also formed in varying yields. We identified XPhos as the best performing dialkylbiaryl phosphine ligand under a wide range of conditions. Unfortunately, extensive optimization of the reaction parameters around the palladium-XPhos catalyst never yielded more than 40% of the desired product.

The key breakthrough that allowed the efficient synthesis of allylic boronate esters came from fine-tuning of the XPhos ligand through the modification of the alkyl substituents on phosphorus. Using PentXPhos ligand, which features cyclopentyl substituents, we identified reaction conditions that afforded desired product 5 in 80% yield (Table 1). The remarkable effect of PentXPhos ligand is demonstrated by results presented in Table 1 (entries 1-6). Closely related XPhos and i-PrXPhos provided only 24% and 25% yield of the desired product, respectively. Modifications of either aryl ring of the PentXPhos ligand backbone had a negative impact on the reaction. Dicyclopentyl SPhos,^[17] RockPhos^[18] and BrettPhos^[19] ligands all gave significantly lower yields of the desired boronate ester (entries 4, 5, and 6).

Table 1.

Reaction development.

Unless otherwise noted, reactions were performed on 0.05 mmol scale. Yields determined by GC using n-octyl ether as an internal standard. [a] Isolated yield of the oxidized product allylic alcohol 5 from a reaction performed at 0.5 mmol shown in parenthesis. dmdba = 3,5,3’,5’-dimethoxydibenzylidene-acetone. R = p-(CH₃O)C₆H₄(CH₂)₃, R¹ = Ph(CH₂)₃ and dipp = 2,6-diisopropylphenyl.

The contributions of other reaction variables are also summarized in Table 1. Commercially available Pd(dmdba)₂ was a superior palladium source, with Pd(OAc)₂ and Pd₂(dba)₃ providing lower yields (entries 7 and 8). The performance of the dmdba precursor could be attributed to the destabilization of the L_nPd⁰-η²-dba complex, in favor of active L_nPd⁰, commonly observed with more electron rich dba ligands.^[20] Copper catalysts supported by NHC ligands gave the best results, although IPrCuCl (entry 9) performed worse than TriCuCl. A catalyst prepared in situ from CuCl and (R)-DTBM-SEGPHOS (entry 10) showed exclusive formation of the alkyl diboronate. Changes in copper catalyst loading had little effect on reaction outcome (entry 11), while increasing palladium catalyst loading led to lower yields, isomerization of the product alkene, and the reduction of the alkenyl electrophile (entry 12). A bulky base was necessary for efficient turnover, with KOTMS performing better than closely related tert-butyl alkoxides or sodium siloxides (entries 13 and 14). The reaction was viable only in hydrocarbon solvents, with toluene and benzene giving the highest yields of the desired product (see Table S4).

Having developed a set of conditions for this differential dihydrofunctionalization, we investigated the scope of the reaction. Allylic boronate esters initially formed in the reaction were oxidized in situ using NaOH/H₂O₂ to furnish allylic alcohols. We found that under the standard conditions, shown in Scheme 4, allylic alcohols could be synthesized from unactivated alkynes in the presence of a wide range of functional groups. Alkynes containing, esters (10), nitriles (15), acetals (12, 21), tosylates (19), aryl and alkyl halides (8, 11, 13, and 22), amines (18, 25 and 27), protected and unprotected alcohols (12, 16, and 23), and nitrogen heterocycles (20 and 24) were well tolerated. Functionalization of aryl acetylenes (9, 13, and 17) also provided moderate yields. Selectivity for the functionalization of terminal alkynes was shown in the presence of a terminal alkene (26), an internal alkyne (29), and a silyl-protected alkyne (28).

Scheme 4.

Substrate scope. Yields of isolated products are reported. Reactions performed on 0.5 mmol scale. [a] Reaction was run for 4 h followed by oxidation. [b] Shown in parenthesis is the NMR yield of the allylic boronate ester from a separate run using 1,3,5-trimethoxybenzene as an internal standard. [c] Following the initial 2 h reaction time the mixture was filtered through a plug of silica with Et₂O and purified by silica gel column chromatography. [d] β-bromostyrene was purchased as a (5:1) mixture of (E:Z) isomers, which was retained in the reaction. [e] Reaction was run for 24 h followed by oxidation, 1.5 equiv of alkenyl bromide was used. [f] Reaction was performed on a 1.0 mmol scale. [g] Following the initial 2 h reaction time 5 equiv of benzaldehyde was introduced and allowed to stir for 24 h at 60 °C before purification by silica gel column chromatography. [h] (Z:E) ratio.

We also explored variations of the alkenyl electrophile. Different substitution patterns were tolerated, including 1,1-disubstituted (38), cyclic trisubstituted (37), acyclic trisubstituted (34) and Z-alkenyl bromides (39). Alkene substitution at the position cis to the bromide resulted in lower yields despite longer reaction times and increased electrophile loadings. Additionally, conjugated alkenyl electrophiles (32 and 33) gave moderate to good yields. Esters, chlorides, and silyl ethers (30, 35, and 36) as well as heteroatoms in the propargylic and vinyl positions (36 and 41)^[21] were also tolerated.

The initially formed boronate esters can also be isolated or used in other transformations. After careful column chromatography, the allylic boronate ester (28) was obtained in 61% yield. A one-pot transformation of allylic boronate esters can be accomplished by the addition of benzaldehyde to the crude reaction mixture. The transposed homoallylic alcohol (42) was isolated in 79% overall yield based on the starting terminal alkyne and (4:1) diastereoselectivity.^[22]

We also observed some limitations of the alkenyl electrophile scope. Little to no allylic boronate ester product was observed with tetrasubstituted alkenyl bromides, dienyl bromides, or alkenyl bromides with strongly electron withdrawing or donating substituents in the vinyl position. Other halides and pseudo halides, such as alkenyl chlorides, iodides, and triflates, were not competent coupling partners in this differential dihydrofunctionalization.

The key features of the mechanism shown in Scheme 2 are the formation of the heterobimetallic complex and subsequent palladium catalyzed cross-coupling. Our previous work with an IPr-supported copper catalyst suggests that the formation of the heterobimetallic intermediate proceeds through two copper hydride additions, first to the alkyne and then to the alkenyl boronate intermediate.^[11] We observed similar results with the triazole based NHC ligand (Tri). Heterobimetallic complex (44) is formed in the stoichiometric reaction of phenylacetylene, TriCuOt-Bu, and HBpin (Scheme 5a). Additionally, the reaction of alkenyl boronate 45 with an in situ generated copper hydride resulted in the formation of 44, in good yield. After careful recrystallization, X-ray quality crystals of the heterobimetallic complex (44) were obtained (Fig. 1).^[23]

Scheme 5.

Exploration of the reaction mechanism. [a] Yield was determined by GC using 1,3,5-trimethoxybenzene as an internal standard. [b] Reaction was run both with and without KOTMS and gave the same result. R¹ = Ph(CH₂)₃.

Figure 1.

X-ray crystal structure of the heterobimetallic complex 44 with thermal ellipsoids at 50% probability. Disorder omitted for clarity.

Next, we turned our attention to the part of the reaction that involved alkenyl bromide. We demonstrated the feasibility of the palladium catalyzed cross-coupling of alkenyl bromide 4 and our isolated heterobimetallic complex (44) using the PentXPhos-supported palladium catalyst (Scheme 5b).

An alternative mechanism for the cross-coupling event, based on a report by Morken, would involve an alkyl diboronate ester as the active coupling partner.^[14c] To test this hypothesis, alkenyl bromide 4 and alkyl diboronate 47 were treated with Pd(dmdba)₂, PentXPhos and KOTMS. Under these conditions we did not observe the desired allylic boronate ester and instead only the homocoupled product of the electrophile was formed (Scheme 5c).

During reaction development, PentXPhos was found to be a uniquely effective ligand, with other XPhos variants providing significantly lower yield of the desired product. To better understand this effect, we explored the cross coupling of heterobimetallic complex 44 and alkenyl bromide 4 using palladium catalysts supported by different XPhos-type ligands.

While t-BuXPhos and EtXPhos gave inferior results, XPhos, PentXPhos, and i-PrXPhos all performed well and provided the allylic boronate ester in ~ 80% yield (Scheme 6a). A more detailed analysis of the reaction shown in scheme 6a revealed that the reaction rate with PentXPhos is nearly double the rate of the reaction with XPhos or i-PrXPhos (Scheme 6b). The observed rate difference does not affect the overall outcome of the alkenylation of the preformed bimetallic intermediate (Scheme 6a). However, in the context of the dual catalytic cycle of differential dihydrofunctionalization, these rate differences become consequential. Increasing the concentration of palladium and XPhos or i-PrXPhos ligands to account for the lower rate of reaction observed with these catalysts does not improve their performance (see Table S14). These results are another example of the stringent requirements that dual catalytic cycles can impose on individual catalysts. Further investigation is necessary to fully elucidate the role of the phosphine ligands in the intricate interplay of the palladium and copper catalytic cycles of this reaction.

Scheme 6.

Effects of phosphine ligands. Reactions were performed on a 0.025 mmol scale with 1.2 equiv of alkenyl bromide 4. To determine yields, aliquots taken from the reaction mixture and quenched with dibromotetrachloroethane before being analyzed by GC using 1,3,5-trimethoxybenzene as an internal standard. R¹ = Ph(CH₂)₃.

In summary, we have developed a new method for the synthesis of allylic boronate esters through differential dihydrofunctionalization of terminal alkynes enabled by cooperative action of copper and palladium catalysts. The new reaction has a broad substrate scope and is compatible with a variety of functional groups and substituted alkenyl bromides. Mechanistic experiments support the proposed catalytic cycle that involves hydrocupration of the initially formed alkenyl boronate ester, followed by palladium catalyzed alkenylation of the heterobimetallic intermediate. We also provide an insight into the unique performance of PentXPhos relative to other closely related ligands.