Reaction optimization
We initiated our study with the investigation of the reaction between o-carborane amide 1a (0.20 mmol) as a model substrate and methyl acrylate 2a (1.5 equiv) in the presence of transition metal catalyst (2.0 mol %) and Cu(TFA)2.H2O (1.0 equiv) (Table 1). When the reaction mixture was stirred in 1,2-dichloroethane (DCE) at 90 oC for 12 h, mediated by either [(p-cymene)RuCl2] or [Cp*RhCl2]2 catalysts, the desired transformation did not take place (entries 1 and 2). Interestingly, the B(9)-alkylated product 3a was obtained exclusively in 31% yield with the use of [Cp*IrCl2]2 catalyst (entry 3). After an examination of solvents such as toluene, PhCF3, MeCN, dioxane, and trifluoroethanol (TFE) (entries 4–8), the best result was obtained with TFE, affording the desired product 3a in 50% yield (entry 8). Encouraged by these results, a screening of additives, including NaOAc, AgOAc, AgTFA, and AgF, was conducted (entries 9–12), whereupon it was discovered that AgF was the optimal additive, which furnished 3a in 85% yield (entry 12). When 1a was treated with 2a without additive, 4a was obtained in trace, indicating that AgF was essential for the B(9)-alkylation reaction (entry 13). Gratifyingly, the use of hexafluoroisopropanol (HFIP) as solvent considerably improved the reaction, affording the B(9)-alkylated product 3a in 91% isolated yield (entry 14). Surprisingly, no alkenylated product is formed in any conditions.
Table. 1. Reaction Optimizationa. aReaction conditions: 1a (0.20 mmol, 1.0 equiv), 2a (1.5 equiv), cat. (2.0 mol %), and additive (1.0 equiv) were dissolved in solvent (2.0 mL). The resulting solution was stirred at 90 oC for 12 h. bNMR yields using CH2Br2 as an internal standard. cIsolated yield.
Substrate scope
With the optimized conditions in hand, we investigated how structural differences of the template could affect the efficiency in this remote B(9) − H alkylation reaction (Fig. 2, top). In the reaction with 1b in which the N-substituent of amide was replaced with a Ns (4-nitrophenylsulfonyl) group, the yield of corresponding product 3b was greatly reduced to 39%. When the ether linker was modified to an alkyl linker such as one seen in compound 1c, the reaction afforded the desired B(9)-alkylated product 3c alongside the undesired B(8)-alkylated compound as an inseparable mixture [B(9):B(8) = 7.3:1], indicating deteroriation of regioselectiviy. The two regioisomers were separated after methanolysis, and their exact structures were undoubtedly identified by X-ray crystallography (see the Supplementary Information for details). In the case of carborane amide 1d, which has a relatively less flexible linker due to the presence of a pyrrolidine ring, the yield of corresponding product 3d decreased to 50%. However, it is noteworthy that the reaction proceeded with high regioselectivity. On the one hand, secondary amide 1e, which is sterically encumbered due to germinal dimethyl groups at the α-position, gave the desired product 3e in 52% yield with poor regioselectivity [B(9):B(8) = 2.5:1] under optimized reaction conditions. The absence of the cyano group in 1a was ineffective, demonstrating that the nitrile template is essential for B(9) − H metalation and alkylation process (Fig. 2, bottom). Other template moieties containing carboxylic acid, pyridine, and another nitrile as a coordinating group were also evaluated in this regioselective B − H alkylation reaction but the reaction did not proceed, establishing the fact that the fine tuning of the template structure is also critical in this transformation.39,58,59 As a result, it was determined by us that the carborane amide 1a was the optimal template to employ in the regioselective remote B(9) − H alkylation reaction.
To demonstrate the efficiency and scope of the iridium-catalyzed remote B(9) − H alkylation reaction, we applied this catalytic system to a variety of alkenes 2 with 1a. For example, alkyl acrylates bearing n-Bu, benzyl, and t-Bu were successfully employed, affording B(9)-alkylated o-carborane amide (4a-4c) in high to excellent yields, ranging from 82–99%. After evaluation of the aryl acrylates possessing methoxy and halogens, we found that electronic and steric effect of the aryl substituents were not obvious, which provided the corresponding product (4d, 4e, 4f, 4h, and 4i) in high yields, varing from 86–99% with the extended reaction time. In particular, 4-nitrophenyl acrylate underwent the alkylation reaction at 60 oC for 24 h, leading to the formation of 4g in 51% yield. Besides acrylate derivatives, it was seen that methyl vinyl ketone, acrylonitrile, and phenyl vinyl sulfone were also compatible in these reaction conditions. In this context, the ketone and nitrile products (4j and 4k) were obtained in 85% and 73% yields, respectively, under the modified conditions. Phenyl vinyl sulfone gave the corresponding product 4l in 91% yield. When the complex estrone-derived acrylate was employed as the substrate, the B(9)-alkylation reaction smoothly proceeded, affording the product 4m in 92% yield. It is worth mentioning that no alkenylated product was detected.
Furthermore, a variety of o-carboranes bearing nitrile template were investigated with methyl acrylate 2a. The reaction of C(2)-unsubstituted o-carborane amide 1f was accomplished with 2a (3.0 equiv), providing the product 5a in 82% yield. n-Butyl, benzyl, cyclohexyl, and t-butyl groups were all compatible under optimized reaction conditions without undue steric interference, affording the corresponding products 5b-5e in high yields ranging from 85–99%. The structure of 5b was unambiguously confirmed by X-ray crystallography (see the Supplementary Information for details). It has also been found that the reaction proceeds successfully even with a variety of aryl groups on the C(2)-position. Phenyl-substituted o-carborane amide was quantitatively converted to the alkylated product 5f. Besides, o-carborane amides possessing 3-methyl, 4-methoxy, 4-bromo, or 4-fluoro group on the aryl ring were subjected to the remote B − H alkylation reaction and provided the corresponding products 5g-5j in high yields, indicating the electronic effect does not largely affect this transformation. Notably, the reaction of 2-thiophenyl o-carborane amide afforded the product 5k in 90% yield with 2a (3.0 equiv).
Gram Scale Reaction and Synthetic Application
To demonstrate the utility of our regioselective remote B(9) − H alkylation reaction, a gram-scale reaction of o-carborane amide 1a was carried out, affording 3a in 98% yield with methyl acrylate and 4l in 91% yield with phenyl vinyl sulfone (Fig. 3). The amide template facilitating effective B(9) − H alkylation can be easily removed through methanolysis, providing the o-carboranes 6 and 7 in high yield. B(9)-Alkylated o-carborane 6 was hydrolyzed to form carboxylic acid 8 in 87% yield. Furthermore, reduction with DIBAL-H provided carboranyl propanol 9 in 88% yield. The carboxylation reaction of the sulfone 7 was smoothly proceeded, affording the B(9)-alkylated o-carborane acid 10 in 83% yield. Since carboxylic acid is versatile ortho-directing group, we applied the o-carborane 10 to the previously reported decarboxylative B(4)-H functionalization reaction.47,48 The Ir(III)-catalyzed B − H amidation reaction with phenyldioxazolone was performed successfully, leading to the B(9)-alkyl-B(4)-amido o-carborane 11 in 92% yield. On the other hand, o-carborane 10 was converted to B(9)-alkyl-B(4)-phenacyl o-carborane 12 in 81% yield.
A Proposed Mechanism
On the basis of the previous studies in the literature60–63, a plausible mechanism for the present remote B(9) − H alkylation reaction is proposed (Fig. 4). The complex I is generated by coordination of o-carborane amide 1 to a Ir(III) catalyst followed by template-assisted B(9) − H activation. Then, it would undergo consecutive coordination (II) and 1,2-insertion of alkene 2 to afford the iridacyclic intermediate III. Finally, B(9)-alkylated product (3, 4, and 5) are released through protodemetallation from the intermediate III, resulting in the regeneration of Ir(III) catalyst. It is noteworthy that alkenylated product, which could be produced through β-hydride elimination from the intermediate III, was not observed at all61.