Organoboron compounds represent significant structural motifs in organic synthesis [1-6], developing novel and practical strategies for the synthesis of organoboron compounds is highly demanding in nowadays chemical society. Among various organoboron compounds, benzylboronic esters are unique alkylboron compounds. They are relatively reactive and can be a practical benzylation reagent in the presence of transition metal catalysts [7-14]. Up to date, tremendous efforts have been made for the synthesis of benzylboronic esters. In general, benzylboronic esters are often synthesized by the borylation of Grignard or lithium reagents [15]. However, this method shows poor functional group compatibility and the difficult preparation of benzylic Grignard and lithium reagents make this classic approach less practical. Recently, transition metal (including Pd, Cu, Ni, Fe, et al.) catalyzed borylation of benzyl (pseudo)halides has been reported for the synthesis of benzylboronic esters [14, 16-22]. Moreover, the cross-coupling of aryl halides [23, 24] or sulfonates [25] with 1, 1-diborylalkanes under palladium catalysis is also an effective method. Benzylic alcohols have also been utilized as electrophiles for the synthesis of primary benzylic boronic esters in the presence of palladium or copper catalyst [26, 27], in which secondary benzylic alcohols cannot be compatible in those catalytic system. Directly utilizing benzylic C–H as the electrophile under the transition metal catalyzed borylation condition is an ideal approach, while it currently still suffers from the chemoselectivity issues on benzylic or aromatic C–H borylation and mono- or di-borylation [28-30]. Styrenes could also be used for the synthesis of benzylboronic esters via unusual transition metal catalyzed Markovnikov selective hydroborations [31-37]. Through 1, 2-metelate rearrangement, benzylboronic esters could be synthesized from tosylhydrazones and HBpin or B₂pin₂ under a metal-free condition [38]. Tosylhydrazones are usually prepared from carbonyl compounds. The direct utilization of aromatic carbonyls for the synthesis of benzylboronic esters would be more step-economy [39]. Herein, we describe the first deoxygenative gem-hydroborylation of aromatic aldehydes and ketones under copper catalysis to access both primary and secondary benzylboronic esters (Scheme 1).

In 2017, we have demonstrated a deoxygenative gem-diborylation of aliphatic aldehydes and ketones [40]. During this study, when benzaldehyde (1a) was utilized as the substrate, a significant amount of gem-hydroborylation product benzylboronic ester (2a) was obtained instead of gem-diboron product. These initial results promoted us to further improve the preparative procedure by using MeOH as the [H] source for achieving the catalytic synthesis of benzylboronic esters from aromatic aldehydes. As a result, 2a was isolated with 71% yield under an optimal catalysis system: ICyCuCl (5 mol%), B₂pin₂ (2.2 equiv.), NaO^tBu (1.3 equiv.), with MeOH (1.0 equiv.) as the proton source in hexane at 100 ℃ for 5 h (Scheme 2).

With the optimized conditions in hand, a wide range of aromatic aldehydes were first examined for its generality in this transformation. As shown in Scheme 3, the reaction system is efficient for substrates containing various functional groups and afford the corresponding products in moderate to good yields. For instance, the reaction of substrates with electron-donating groups (‒OMe, ‒^tBu or other oxygen-containing groups) and substrates with electron-withdrawing groups (‒F and ‒Cl) all proceeded well, affording the corresponding products in moderate to good yields (2b–2e, 2g, 2i–2m). The sulfur-containing group ‒SMe was also tolerated under standard condition (2f). It was noteworthy that steric effect did not have a strong influence on the reactivity. For instance, when 2, 4, 6-trimethylbenzaldehyde was used as the substrate, the corresponding product 2h was isolated in 52% yield. Moreover, other aromatic aldehyde such as 2-naphthaldehyde and hetero-aromatic aldehyde such as 2-thenaldehyde proceeded smoothly to generate the corresponding products (2n, 2o).

However, when extending the catalytic system to aromatic ketones, problem occurred and only trace amounts of products were detected, in which most of the starting ketones were reduced to their corresponding alcohols. The generation of alcohols indicated the presence of hydride species in this catalytic system. It has been demonstrated that proton could be used as hydride source in the presence of B₂pin₂ [41-49]. Furthermore, Clark et al. [50] has shown that the addition rate of a copper-boron species to ketone C=O group was slower than that of aldehydes. Therefore, the addition of hydride species to ketone carbonyls dominated the transformation of ketones. As a result, alcohols were produced as the major products. In order to solve this problem, a more reactive KO^tBu was applied to activate B₂pin₂ instead of NaO^tBu, for which we expected to increase the nucleophilicity of boryl group for the addition to ketone carbonyl groups. Meanwhile, a less acidic alcohol was utilized as the proton source instead of MeOH, for which we expected to slower down the generation of hydride species. To our delight, after a series of attempts, when the reaction was carried out in the presence of 1.0 equiv. KO^tBu as base and EtOH as the proton source, the aromatic ketones were successfully transformed to their corresponding secondary benzylboronic esters.

Subsequently, a series of aromatic ketones were investigated under the optimized conditions (Scheme 4). Aromatic ketones, such as acetophenone, propiophenone and butyrophenone, could be smoothly transformed to the corresponding benzylboronic esters in moderate yields (4a, 4b, 4d, 4e). Substrates bearing electron-rich group, such as p-methoxyacetophenone, have been evaluated and the desired product was obtained in moderate yield albeit prolonged reaction time was required (4c, 4i). A significant amount of 1, 1-diborylation products was detected by GC-MS if the reaction time was shortened to 1 h. This interesting phenomenon indicated that the ketones may firstly transformed to 1, 1-diborylakanes, followed by selective protodeboronation to afford the final products. Benzocycloanone can be smoothly converted into the desired products (4f, 4g). 2-Acetonaphthone was also a suitable substrate to produce the corresponding product in moderate yields (4h).

Mechanistically, the copper-catalyzed 1, 2-borylation of carbonyl was believed to initiate the reaction, in which α-OBpin benzylboronic ester was generated. Our previous report and the results from the monitor of the reaction course demonstrated above exhibited that the borylation of α-OBpin to give benzylic gem-diboron followed by protonation with alcohol is most likely the reaction mechanism (Scheme 5, Path A) [38, 40]. In this case, the α-OBpin containing tetracoordinated boron species with boryl as the migrating group was constructed. However, another pathway is also possible. As demonstrated above, there might be hydride species presented in the reaction system. Therefore, a tetracoordinated boron species with hydride as the migrating group would also generate the desired deoxygenative gem-hydroborylation product (Scheme 5, Path B).

To verify Path A, the benzylic gem-diboron 4a-1 was synthesized according to the literature [51]. Then it was subjected to react with EtOH in the presence of KO^tBu. As a result, the protodeborylation product 4a was obtained in 99 % yield (determined by ¹H NMR analysis). In this case, catalytic amount of KO^tBu was used as the promotor, because of the KO^tBu was almost consumed in those standard reaction systems. This result indicated that the benzylic gem-diboron was presumed to be an intermediate in this transformation (Scheme 6, Eq. (1)). Next, in order to verify Path B, α-OBpin benzylboronate 2a-1 was synthesized from benzaldehyde according to the literature [52]. Then it was subjected to react with MeOH, NaO^tBu and B₂pin₂ in hexane at 100 ℃ for 5 h. As a result, the desired 2a was obtained in 47% yield, indicating the generation of α-OBpin benzylboronate as an intermediate in this transformation (Scheme 6, Eq. (2)). Furthermore, HBpin was applied as the hydride source to test the hydride-migration hypothesis (Scheme 6, Eq. (3)). 2a-I was converted to the desired 2a smoothly in the presence of HBpin and NaO^tBu in 57% yield. This result indicated that the hydride-migration pathway is also possible for this transformation. Using CD₃OD and HBpin resulted in a mixture of deuterated-2a and 2a (See supporting information for details), also indicated the possibilities of both pathways for this transformation. At current stage, either pathway could not be ruled out.

In summary, we have developed a straightforward and effective method for the synthesis of benzylboronic esters through Cu-catalyzed deoxygenative gem-hydroborylation of readily available aromatic aldehydes and ketones. A series of substituted aryl aldehydes and ketones were successfully converted to their corresponding products. Alcoholic proton was utilized as the hydride source. At current stage, either boryl migration followed by protodeborylation or hydride migration is possible for the mechanistic consideration. Further detailed mechanism is currently underway in our laboratory.