Palladium Catalyzed Regioselective B – C ( sp ) Coupling via Direct Cage B – H Activation : Synthesis of B ( 4 )-Alkynylated o-Carboranes

Icosahedral carboranes are a class of polyhedral boron hy­ dride clusters in which one or more of the BH vertices are re­ plac ed by CH units, which can be viewed as three­dimensional relatives of benzene. Their exceptional thermal and chemical stabilities as well as 3D structures make them useful building blocks for boron neutron capture therapy agents in medicine, versatile ligands in coordination/organometallic chemistry, and functional units in supramolecular design/optoelectronic materials.1 As a result, considerable attention has been direct­ ed towards the functionalization of carboranes. Classic routes to functionalized carboranes rely on the po­ larized cage C–H/B–H bonds: the weakly acidic C–H proton (pKa ~ 23) and basic B–H hydride.1 Generally, cage C–H bonds can be deprotonated by strong bases, followed by the reac­ tion with electrophiles to give carbon­substituted carboranes. Cage B–H bonds are preferentially subjected to electrophil­ ic substitution reactions, resulting in the formation of cage boron­substituted carborane derivatives. However, the latter suffers from poor regioselectivity due to the presence of diffe­ rent electronic environments of BH vertices. To tackle the regioselectivity problem, the group of Profes­ sor Zuowei Xie from the Chinese University of Hong Kong (P. R. of China) introduced a carboxyl group at the cage carbon to control the regioselectivity and facilitate cage B–H activa tion. Subsequently, transition­metal­catalyzed cage B(4)­alken yl­ ation,2a B(4,5)­dialkenylation,2b and B(4,5)­diarylation2c have been achieved (Scheme 1). Professor Xie said: “In view of the wide application of carboranyl acetylenes in molecular rods, nanomaterials and metal­organic frameworks,1 we have developed the first tran­ sition­metal­catalyzed regioselective cage B–C(sp) coupling via direct cage B(4)–H activation for the synthesis of B(4)­ alkynylated o­carboranes.” In the presence of 5 mol% Pd(OAc)2 and three equivalents of AgOAc, the reaction of carboranyl carboxylic acid with al­ kynyl bromide proceeds smoothly in DCE (DCE = 1,2­dichloro­ ethane) at 90 °C to give the desired B(4)­alkynylated o­carbo­ ranes in moderate to very good isolated yields. “Though this reaction is tolerant of many functional groups R1 at the cage C(2), it is compatible only with sterically bulky silyl groups Palladium Catalyzed Regioselective B–C(sp) Coupling via Direct Cage B–H Activation: Synthesis of B(4)-Alkynylated o-Carboranes


Introduction
The development of efficient synthetic methodologies to incorporate alkyne motifs has received broad interest, as they are not only important building blocks in natural products, pharmaceuticals and materials 1 but also essential functional groups in cross-coupling, metathesis and cycloaddition reactions. 2 Meanwhile, carboranyl acetylenes have proved to be useful basic units in molecular rods, 3 nonlinear optical materials, 4 supramolecular design, 5 nanovehicles 6 and metalorganic frameworks. 7 As there is a lack of direct and efficient methodologies for the synthesis of B-alkynylated carboranes, the alkyne moieties in the aforementioned materials are generally connected to cage carbon atoms, 3-8 which limits the application scope of the carborane derivatives.
Though cage boron alkynylated carboranes can be prepared by two-step reactions, such as the selective iodination of an ocarborane, followed by Pd(0)-catalyzed cross-coupling with alkynyl Grignard reagents, 9 the installation of iodo groups to specic positions on the carboranes is necessary (Scheme 1). However, the selective iodination of cage B(4,5,7,11)-H is rather challenging, if not impossible. 8 Thus, we aim to develop new methodologies for the selective and direct alkynylation of carboranes via cage B-H activation.
Directing groups are essential in transition metal catalyzed C-H activation due to their ability to chelate the metal catalyst, position it for selective C-H cleavage, and reduce activation energy by stabilizing the metallacycle intermediates. 10 Nevertheless, strategies using directing groups suffer from limitations when the directing groups are not present in the target molecules. To overcome this problem, the use of traceless directing groups is obviously an ideal method. Recently, the use of -COOH as a weak coordinating yet efficient directing group for transition metal catalyzed phenyl C-H activation has been documented, and has been found to be easily removed by decarboxylation aer the reaction. 10h Subsequently, carboxylic acid directed phenyl C-H olenation, 11 arylation, 12 alkylation, 13 acylation, 14 carboxylation, 15 amination, 16 hydroxylation 17 and halogenation 18 have been successfully developed. However, to the best of our knowledge, the direct alkynylation of C-H bonds guided by -COOH is still elusive, although nitrogen-based directing-group-guided transition-metal catalyzed phenyl C-H alkynylation has been recently documented using alkynyl halides, 19 hypervalent iodine-alkyne reagents 20 and terminal alkynes 21 as the alkynylating reagents. Meanwhile, oxidative coupling of two C-H bonds for the formation of a C-C bond has received growing interest due to its benets which include atom-economy, step-economy and less waste. 22 Compared with the achievements of phenyl C-H bond oxidative coupling, the regioselective and direct oxidative coupling of an organic C-H bond with a cage B-H bond in o-carboranes is very rare. 23 Very recently, our group has developed a transition metal catalyzed -COOH guided cage B-H alkenylation 23,24 and arylation 25 of o-carboranes, in which the carboxyl group is removed in a one-pot fashion. Inspired by these results and other cage B-H activation reactions, [26][27][28][29] we have extended our research to investigate direct cage B-H alkynylation by alkynyl halides through a Pd(II)-Pd(IV)-Pd(II) catalytic cycle and by terminal alkynes via a Pd(II)-Pd(0)-Pd(II) catalytic cycle. These new ndings are reported in this article (Scheme 1).
of AgOAc in toluene at 90 C for 6 h did not give any of the desired product (entry 1, Table 1). Replacement of toluene with 1,2-dichloroethane (DCE) afforded the desired coupling product 4-( i Pr 3 SiC^C)-2-CH 3 -o-C 2 B 10 H 10 in 40% GC yield (entry 2, Table  1). Increasing the amount of AgOAc to 3 equiv. resulted in 90% GC yield of 3a (entry 4, Table 1). Higher or lower reaction temperatures led to decreased yields of 3a (entries 5 and 6, Table 1). Lowering the catalyst loading to 5 mol% did not change the reaction efficiency (entry 7, Table 1). In view of the yields of 3a, entry 7 in Table 1 was chosen as the optimal reaction conditions. A variety of carborane monocarboxylic acids (1) were examined under the chosen optimal reaction conditions, and the results are compiled in Table 2. All alkyl, alkenyl and aryl substituents on cage C(2), regardless of electronic properties, afforded the coupling products 3 in high isolated yields (entries 1-10 and 13, Table 2). For the heteroatom containing substrate 1j, the product 3j was isolated in 78% yield (entry 10, Table 2), whereas that bearing a thiophenyl group (1l) afforded the product 3l in 54% yield (entry 12, Table 2) probably due to the interaction of Pd with the S atom. Meanwhile, substrate 1k with a naphthyl substituent on cage C(2) gave 3k in only 40% isolated yield (entry 11, Table 2). For R 1 ¼ H, an inseparable mixture was produced (entry 14, Table 2). When R 1 ¼ Me 3 Si, the desilylation species 3n was isolated in 41% yield aer work up (entry 15, Table 2).
In contrast to R 1 at cage C(2), the scope of R 2 is highly limited in such a coupling reaction. t BuMe 2 SiC^CBr worked well to give 3p in 70% isolated yield (entry 16, Table 2). However, less hindered Me 3 SiC^CBr was not reactive, probably due to its propensity to coordinate with a Pd center via the p bond (entry 17, Table 2). Such a phenomenon was also observed in phenyl C-H alkynylations using R 3 SiC^CBr as reagents. 30 It was noted that other alkynyl bromides such as PhC^CBr and t BuC^CBr were not compatible with this reaction.

Alkynylation using terminal alkynes
As the previous method has a limited substrate scope, we wanted to develop a more atom-and step-economic method for cage B-H alkynylation using terminal alkynes as reagents. We commenced our studies by screening for a suitable base for the oxidative coupling of cage B-H in 1-COOH-2-CH 3 -o-C 2 B 10 H 10 (1a) with i Pr 3 SiC^CH under the aforementioned optimal reaction conditions. No reaction was observed in the absence of a base (entry 1, Table 3). The addition of 2 equiv. of K 2 HPO 4 afforded the target product 3a in 30% GC yield with i Pr 3 SiC^C-C^C i Pr 3 Si as the side product (entry 2, Table 3). To inhibit the formation of a homocoupling side product, i Pr 3 SiC^CH was added slowly via a syringe pump, leading to a signicantly increased yield of 3a to 56% GC yield (entry 3, Table 3). The yield was further improved to 75% if 2 equiv. of the terminal alkyne was used (entry 4, Table 3). Replacement of 1,2-dichloroethane (DCE) with toluene resulted in a slightly higher yield of 3a (entry 5, Table 3). Decreasing the reaction temperature to 80 C afforded 3a in 86% GC yield (entry 6, Table 3). In view of the yields of 3a, entry 6 in Table 3 was chosen as the optimal reaction conditions. This reaction has a much broader substrate scope (R 2 ¼ silyl, phenyl and carboranyl). The results are compiled in Table 4. For R 1 ¼ alkyl groups, the isolated yields of 3 are comparable to those observed in Table 2. However, if R 1 ¼ aryl unit such as 1g, the isolated yield of 3g is 30% (entry 4, Table 4), which is signicantly lower than that of 70% shown in entry 7, Table 2. On the other hand, compounds 1n (R 1 ¼ H) and 1o (R 1 ¼ Me 3 Si) give 3n in 35% and 74% yields, respectively (entries 5 and 6, Table 4). These yields are much higher than those found in the previous reaction (entries 14 and 15, Table 2). The reasons for this phenomenon are not clear at this stage.

Transformation of 3a
To demonstrate the applications of the resultant compounds 3 as building blocks, further transformation of 3a was carried out. The i Pr 3 Si group in 3a was readily removed by treatment with TBAF (TBAF ¼ tetra-n-butylammonium uoride) to afford quantitatively the terminal alkyne 4a (Scheme 2). Like other terminal alkynes, compound 4a can undergo various transformations to give different kinds of carborane-incorporated functional molecules. Sonogashira coupling of 4a with iodobenzene or 2-bromothiophene generated 3r or 5a in 92% and   (6a) in 84% isolated yield. A click reaction of 4a with phenyl azide afforded carboranefunctionalized 1,2,3-triazole (7a) in 95% isolated yield. All new compounds 3 and 4a-7a were fully characterized by 1 H, 13 C, and 11 B NMR spectroscopy as well as high-resolution mass spectrometry (HRMS). 31 Molecular structures of 4a and 6a were further conrmed by single-crystal X-ray analyses and are shown in Fig. 1. Experimental details are included in the ESI. †

Reaction mechanism
To gain some insight into the reaction mechanism, the following control experiments were carried out. No reaction was observed if 1a was treated with 1 equiv. of i Pr 3 SiC^CBr in the presence of 20 mol% Pd(dba) 2 (dba ¼ dibenzylideneacetone) in DCE at 90 C for 6 h in the absence of AgOAc. On the other hand, under the same reaction conditions, replacement of Pd(dba) 2 with Pd(OAc) 2 gave the alkynylation product 3a in 30% GC yield (Scheme 3a). Similarly, in the presence of 20 mol% Pd(OAc) 2 , the reaction of 1a with 2 equiv. of i Pr 3 SiC^CH afforded 3a in 16% GC yield without AgOAc as the oxidant. While, no 3a was observed when 20 mol% Pd(dba) 2 was used instead of Pd(OAc) 2 (Scheme 3b). These results suggest that both cross-coupling reactions are initiated by Pd(II) not Pd(0).
Decarboxylation of carboranyl carboxylic acids (1b and 3b-COOH) was also examined (Scheme 3c). Compound 1b was stable aer heating at 90 C for 12 h in DCE, whereas 3b-COOH underwent complete decarboxylation within one hour under the same reaction conditions. Notably, it only took ten minutes to  convert 3b-COOH to 3b in the presence of 1 equiv. of AgOAc.
These results clearly indicate that the introduction of an alkynyl group at the cage B(4) site can induce the decarboxylation, and the addition of a silver salt can accelerate such decarboxylation, which is crucial for controlling the mono-selectivity. On the basis of the aforementioned experimental data, two plausible reaction mechanisms are proposed in Scheme 4. For the Pd(II)-Pd(IV)-Pd(II) catalytic cycle: an exchange reaction of 1 with Pd(OAc) 2 , followed by regioselective electrophilic attack at Scheme 2 Transformations of 3a.   Reductive elimination affords the cage B(4)-alkynylated intermediate F and Pd(0). Decarboxylation of F results in the formation of the nal product 3, meanwhile Pd(0) is oxidized by AgOAc to regenerate Pd(OAc) 2 . It is noted that AgOAc acts as a bromide captor in the Pd(II)-Pd(IV)-Pd(II) catalytic cycle, but as an oxidant to regenerate Pd(II) from Pd(0) in the Pd(II)-Pd(0)-Pd(II) catalytic cycle. However, in both cross-coupling reactions, AgOAc plays a crucial role in promoting decarboxylation and thereby controlling the mono-selectivity.

Conclusion
We have developed two catalytic systems for regioselective and efficient alkynylation of cage B(4)-H bonds in o-carboranes using alkynyl bromides or terminal alkynes as alkynylating agents, where -COOH acts as a traceless directing group. A series of new cage B(4)-alkynylated o-carborane derivatives has been prepared for the rst time, which could nd many applications in the synthesis of carborane-based materials. 3-7 This opens up a new window for the functionalization of carboranes by direct oxidative coupling of the cage B-H and organic C-H bonds. This work also offers a useful reference for selective C-H alkynylation using carboxylic acid as a traceless directing group in other aromatic systems.
On the basis of control experiments and literature work, two catalytic cycles are proposed for the above two reactions: a Pd(II)-Pd(IV)-Pd(II) cycle for using alkynyl bromides as coupling agents and a Pd(II)-Pd(0)-Pd(II) cycle for employing terminal alkynes as coupling partners. The latter has a broader substrate scope than the former. This work also gives some hints for the development of new catalytic systems for the functionalization of carboranes.