Recent advances in B–H functionalization of icosahedral carboranes and boranes by transition metal catalysis

Introduction

Clusters based on the frameworks 1,2-dicarba-closo-dodecaborane (o-C₂B₁₀H₁₂, 1), monocarba-closo-dodecaborate ([CB₁₁H₁₂]⁻, 2) and closo-dodecaborate ([B₁₂H₁₂]²⁻, 3) are boron cage compounds which exhibit icosahedral geometry and delocalization of electron density (Fig. 1) [1], [2], [3], [4]. Their three-dimensional electron distribution can be considered as an extension of the π aromaticity of classical organic arenes such as benzene [5]. This feature results in exceptional chemical and thermal stability as well as low toxicity. The unique steric and electronic properties of icosahedral boron clusters sets them apart from common organic or inorganic substituents. As a result, an large number of applications have emerged over the past years; among them are ligand design [6], [7], supramolecular [8], [9] and medicinal [10], [11], [12], [13] chemistry, as well as fluorescence/phosphorescence [14], [15] and materials science [16]. Furthermore, weakly coordinating halogenated derivatives of 2 and 3 are the anions of choice as counterions for cationic reactive intermediates and of highly active catalysts [17], [18].

Fig. 1:

Structures and atom numbering of icosahedral boron clusters.

The practical versatility of boron clusters has inspired chemists to develop various methods for derivatization. Carbon vertices can in many cases be functionalized by deprotonation and subsequent reaction with electrophiles. The traditional approach to bring about modification at B–H corners relies on electrophilic substitution or halogenation followed by cross coupling reactions [1], [3], [19]. However, vertex selectivity and the degree of substitution because of the high symmetry of the clusters are challenges of such methods. To address synthetic obstacles in the preparation of icosahedral carborane and borane derivatives, transition metal-mediated B–H activation is an attractive alternative with the potential to provide desired products regioselectively with a minimum number of steps. Early studies on dicarbadodecaboranes by Hawthorne demonstrated that transition metal complexes based on ruthenium, rhodium and iridium can induce formation of B–[M] (M=Ru, Rh, Ir) intermediates by oxidative addition to B–H bonds [20], [21]; notably, one example involved the use of a phosphine group attached to C1 to direct the metalation process. Work by Sneddon showed that polyhedral boranes can undergo B–C bond formation using substoichiometric amounts of transition metals, procedures that can formally be viewed as catalytic hydroboration reactions [22], [23], [24], [25], [26], [27]. The directing group-assisted, transition metal-mediated B–H activation of icosahedral boron clusters is outlined in Fig. 2. In a first step, the directing group coordinates to a transition metal fragment, thereby bringing it into proximity of a B–H moiety. Activation of the B–H bond by oxidative addition or cyclometalation–deprotonation produces an intermediate featuring a direct B–[TM] bond. Treatment of this intermediate with a suitable coupling partner then leads to boron vertex substitution while the transition metal fragment is released and can engage in further B–H activation steps.

Fig. 2:

Directing group-assisted, transition metal-mediated B–H functionalization of icosahedral boron clusters. DG, directing group; TM, transition metal; FG, functional group.

Among the transformations involving B–H functionalization of icosahedral boron clusters mediated by transition metals, the chemistry of 1 has been investigated most thoroughly. Several reviews on this subject are available, and three of them have appeared very recently [28], [29], [30], [31], [32]. The reaction pattern of 1 can be summarized as follows. The nucleophilicity of the boron vertices increases from B3/6 to B4/5/7/11 to B8/9/10/12 (Fig. 3). As a result, positions B3/6 are in many cases, but not exclusively, activated by electron-rich metal complexes. This process can occur with or without the presence of a directing group at C1. Positions B4/5/7/11 are usually selectively functionalized with the assistance of a directing group using electron-deficient metal complexes. Similarly, substitution of positions B8/9/10/12 requires elecron-deficient metal centers that induce B–[TM] bond formation by metalation–deprotonation, but without guidance by a directing group.

Fig. 3:

Polarization of o-carboranes and B–H activation strategies. DG, directing group; TM, transition metal.

Pioneering research in the field of selective catalytic B–H activation of 1 has been carried out by the groups of Xie and Cao. They have developed methodologies for cage arylation, alkenylation, fluorination, hydroxylation and cascade cyclizations [33], [34], [35], [36], [37], [38], [39], [40], [41]. These results underscore the significance of catalytic B–H activation as an attractive new tool to address synthetic challenges in the preparation of icosahedral carborane and borane derivatives. In this conference report, recent advances in the functionalization of boron clusters are summarized as they were presented by the groups of Prof. Yangjian Quan (The Chinese University of Hong Kong) and Prof. Zaozao Qiu (Shanghai Institute of Organic Chemistry) as well as our own group (Zhejiang University). Focus is placed on the actual transformations, while for proposed mechanisms the reader is referred to the original publications.

Recent progress in the catalytic functionalization of o-carboranes

Regioselective alkynylation of carborane carboxylic acids using bromoalkynes or terminal alkynes was reported by Xie et al. [42]. Starting materials 4 could be coupled with bromo-triisopropylsilylacetylenes to afford products 5 in moderate to high yields (Scheme 1). The Pd/Ag(I) catalyst system effected selective B4 substitution with concomitant removal of the acid moiety. Variation of the group at the C2 vertex showed that alkyl, benzyl and aryl substituents were well tolerated, while in the case of C2–H the desired product could not be isolated in pure form.

Scheme 1:

Palladium-catalyzed B4-alkynylation of o-carborane carboxylic acids.

Changing the catalyst system to Pd/Ag(I)/K₂HPO₄ enabled the direct coupling of o-carborane carboxylic acids with terminal alkynes. The versatility of the methodology was demonstrated by synthesizing a series of products 6 with H, alkyl and benzyl groups at the cage C2 position and silylated as well as arylated alkyne moieties (Scheme 2a). In addition, compound 5-Me was quantitatively converted to the desilylated intermediate using [Bu₄N]F (Scheme 2b). This intermediate was shown to undergo subsequent reactions typical of alkynes, such as palladium-catalyzed cross coupling, leading to 7.

Scheme 2:

Palladium-catalyzed B4-alkynylation of o-carborane carboxylic acids using terminal alkynes (a) and deprotection–cross coupling of the silylated alkynylcarborane (b).

Direct B–N bond formation was accomplished using o-carborane carboxylic acids 4 and N-benzoyloxyamines [43]. Under Pd catalysis, a range of carboranes with alkyl, benzyl and alkenyl groups at the C2 position was morpholinylated in moderate to good yields (Scheme 3). The substitution occurred regioselectively at the B4 position, and under the reaction conditions decarboxylation was observed to give products 8.

Scheme 3:

Palladium-catalyzed B4-amination of o-carborane carboxylic acids.

The substrate scope in terms of the amine coupling partner was investigated with 4-Me and a series of cyclic and acyclic N-benzoyloxyamines to give 9. Disubstituted aliphatic amino substituents were successfully introduced at the B4 position, including piperazine derivatives (Scheme 4). In the latter case, the heterocycle was protected either as an amide or as a carbamate.

Scheme 4:

Palladium-catalyzed B4-amination of 2-methyl o-carborane carboxylic acid.

Furthermore, B–N bond formation was also reported starting from o-carborane carboxylic acids 4 in combination with sulfonyl azides. [Ru(p-cymene)Cl₂]₂ as the catalyst effected regioselective sulfonamidation at the B4 position (Scheme 5). High yields were observed using C2-alkylated and benzylated carboranes as well as aromatic and aliphatic sulfonyl azides. As in the case of the Pd-catalyzed amination, concomitant decarboxylation took place to afford products 10.

Scheme 5:

Ruthenium-catalyzed B4-sulfonamidation of o-carborane carboxylic acids.

A combined B–H and C(sp²)–H bond activation was accomplished using o-carborane carboxylic acids and thiophenes [44]. The dehydrogenative cross coupling involved [IrCp*Cl₂]₂, Ag(I), Cu(II) and Li₂CO₃ as the catalyst system and afforded decarboxylated B4-(2-thienyl) derivatives with high regioselectivity with respect to cage and heterocycle substitution. Starting from 4-Me, investigation of the thiophene substrate scope showed that coupling to monosubstituted and fused aromatic systems occurred in moderate to high yields affording products 11 (Scheme 6a). Variation of the group at the carborane C2 position and reaction with 2-methylthiophene gave alkylated and benzylated compounds 12 (Scheme 6b). Even though in the majority of the reactions an excess of thiophene was used, the method allowed for a facile synthesis of carborane–heterocycle hybrids with potential applications in the areas of luminescence and photoelectronic materials.

Scheme 6:

Iridium-catalyzed dehydrogenative B4-thienylation of o-carborane carboxylic acids.

A method for di- and monoborylation of o-carboranes was developed by Qiu and coworkers [45]. In their study, the combination of carborane precursors 13 without directing groups and bis(pinacolato)boron effected selective introduction of pinacolatoboryl substituents to give products 14 with [Ir(cod)Cl]₂]/2-methylpyridine (cod=cyclooctadiene) as the catalyst system (Scheme 7). Various starting materials bearing groups such as Me, Bn, Ph and I at remote positions gave B3/6-diborylated compounds in high yields. Starting from B4-Ph, B4-(1,2-diphenyl)ethenyl or B4/7-I₂ substitution resulted primarily in monoborylation. Blocking the original B3 position with Ph caused monoborylation as well.

Scheme 7:

Iridium-catalyzed B3/6-diborylation and B6-borylation of o-carboranes using bis(pinacolato)diboron.

A switch in regioselectivity was observed when one of the carbon vertices was prefunctionalized with a dimethyl-t-butylsilyl group. o-Carboranes 15 afforded B4-monoborylated products 16 in high yields (Scheme 8a). As an important part of their study, the authors demonstrated that the pinacolatoboryl groups could successfully be transformed into other functionalities (Scheme 8b). Thus, carbon, halogen, acyloxy and nitrogen substituents were introduced under cross coupling or oxidative conditions furnishing 17 in yields of 73–90 %.

Scheme 8:

Iridium-catalyzed B4-borylation of silylated o-carboranes using bis(pinacolato)diboron (a) and derivatization of B3/6-diborylated o-carborane (b).

Catalytic functionalization of anionic carboranes and boranes

In contrast to the advances in the chemistry of 1, well-controlled B–H activation of anionic icosahedral boron clusters by transition metals is much less established. Recently, our group started a research program aimed at establishing methodologies for the selective derivatization of 2 and 3. Until then, procedures for the formation of B–X bonds (X=halogen, C, N, O) primarily relied on electrophilic substitution reactions, often under rather harsh conditions, or halogenation combined with cross coupling. On the other hand, little had been reported on metal-mediated functionalization.

With respect to B–H activation of monocarba-closo-dodecaborates, Weller and coworkers observed that agostic Rh(III) phosphine complexes of 2 undergo cage ethenylation and ethylation in the presence of ethene [46], [47]. Moreover, they were able to isolate a dinuclear Rh(I)/Rh(III) species containing two carborane cages that featured one direct B–Rh bond. However, the reactions relied on the stoichiometric use of rhodium, and challenges in terms of regioselectivity and degree of substitution were reported. Lavallo’s group investigated the divergent reactivity of two closely related carboranylphosphines 18 and 19 towards Ir(I) [48], [49]. Treatment of the o-carborane derivative with [Ir(cod)Cl]₂ (cod=cyclooctadiene) resulted in the formation of Ir(III) complex 20 by oxidative addition across the B3–H moiety (Scheme 9a). Conversely, under similar conditions the anionic monocarba-closo-dodecaborate congener did not undergo the same reaction but afforded Ir(I) complex 21 with two agostic-like cage–Ir interactions. The results involving these two carboranylphosphine precursors indicated the impact of subtle changes in structure upon the energy barrier associated with oxidative addition.

Scheme 9:

Divergent reactivity of o-carborane phosphine and monocarba-closo-dodecaborate phosphine towards iridium(I) (a) and formation of a monocarba-closo-dodecaborate complex with a direct B2–Ir bond using iridium(III) (b); cod=cyclooctadiene, Cp*=pentamethylcyclopentadienyl.

We speculated that a directing group allowing for formation of a five-membered iridacycle combined with a formal cyclometalation–deprotonation process could provide controlled access to an intermediate with a direct B–Ir bond. Reaction of tosyl amide 22 with a stoichiometric amount of IrCp*(OAc)₂ (Cp*=pentamethylcyclopentadienyl) indeed afforded complex 23 in acetonitrile at room temperature (Scheme 9b) [50]. This compound was isolated in 80 % yield and characterized by spectroscopic methods, X-ray crystallography and DFT calculations. The calculated transition state for the B–H activation step suggested H atom transfer from the B2 position to the carbonyl O atom of the acetate ligand. Monochlorination of the B2 vertex of 23 using N-chlorosuccinimide with concomitant removal of the IrCp* fragment demonstrated the possibility of selective metal-mediated cage functionalization. In parallel to our studies, Lavallo obtained a bis(carboranylphosphine) Pd(II) compound featuring a B–Pd bond as a by-product from a bis(carboranylphosphine) Pd(0) complex upon oxidation with chlorobenzene [51].

Encouraged by the results involving tosyl amide 22, our goal was to establish conditions that would enable catalytic derivatization of monocarba-closo-dodecaborate amides. As a model reaction, the coupling with ethyl acrylate was chosen. A screening of various directing groups, transition metals and solvents showed that pyrrolidine amide 24 with [RhCp*Cl₂]₂/Ag(I)/Cu(II) as the catalyst system in dimethylacetamide lead to tetrasubstitution at positions B2–5 (Scheme 10) [52]. The product contained three reductively coupled ester moieties and one ethyl acrylyl substituent with a remaining double bond. Other activated alkenes such as phenyl acrylate, styrene and 4-fluorostyrene exhibited similar reactivity to give tetrasubstituted carboranes 25. In the case of 4-fluorostyrene, the disubstituted compound dominated when acetonitrile was used. Diphenylacetylene as the coupling partner gave the tetraalkenylated product. Ir(III) as the catalyst proved to be useful as well, leading to monoalkenylation in combination with diphenylacetylene and sulfonamidation at the B2/4 positions with tosyl azide as the coupling partner.

Scheme 10:

Rhodium- and iridium-catalyzed ortho-functionalization of amide-substituted monocarba-closo-dodecaborates.

Selective monochlorination was achieved using Rh(III) and N-chlorosuccinimide as the chlorination agent. Starting from amides 24 with B12–H, B12–Br or B12–I vertices afforded B2-substituted products 26 in 1,2-dichloroethane (Scheme 11). This reaction was remarkable because electrophilic halogenation of 2 causes substitution at the B12 vertex, followed by B7–11. In the absence of Rh(III), pyrrolidine amide 24 (B12=H) was indeed chlorinated at the B12 position, while with the catalyst system Rh(III)/Ag(I)/Cu(II) only the B2 isomer was detected.

Scheme 11:

Rhodium-catalyzed B2-chlorination of amide-substituted monocarba-closo-dodecaborates.

Until very recently, the B–H activation chemistry of dodecaborate 3 and its derivatives was unexplored with the exception of two reports. Hawthorne and coworkers described the perhydroxylation of 3 in fuming sulfuric acid in the presence of Pt(II) or Pd(II) [53]. B–[M] bond formation followed by hydroxylation was suggested for this reaction although the reaction mechanism was not investigated. In 2016, Schleid’s group presented structural evidence for a bismuth(III) complex with a direct B–Bi interaction [54]. This compound was obtained via formal metalation–deprotonation using bismuth(III) oxide. While constituting a landmark in dodecaborate chemistry, the potential of the complex to undergo further B vertex functionalization was not studied.

The facile monoaminatin of 3 lead us to explore directing group-substituted dodecaborates based on the [B₁₂H₁₁(NH₂)]²⁻ scaffold. Upon testing several functionalities, we identified the ureido group (–NHC(O)NR₂) as a suitable moiety for metal coordination and B–H activation. A screening of conditions showed that N,N-dimethyl derivative 27 could be selectively coupled with alkynes under rhodium catalysis. Using [RhCp*Cl₂]₂ in combination with Cu(II) afforded ortho-monoalkenylation with subsequent ring closure to give B1,2,3-trisubsituted products 28 (Scheme 12) [55]. Notably, the twofold B–H activation process occurred in acetonitrile at 25°C and did not require oxygen- or moisture-free conditions. Diarylalkynes with different electronic properties coupled efficiently; in particular, 3/4-alkyl, 4-halide and 4-methoxy substituents gave yields of 60–85 %. The mild conditions allowed for the use of carbonyl groups and thiophene rings as well. Unsymmetrical phenylpropyne and the dialkyl substrate hex-3-yne afforded the monoalkenylated products in moderate yields. The effect of varying the directing group was studied by using –NHC(O)NEt₂ instead of –NHC(O)NMe₂. In this case the coupling with diphenylacetylene occurred in 48 % yield, which underscored the substantial influence of subtle structural changes of the starting material on the outcome of the transformation.

Scheme 12:

Rhodium-catalyzed alkenylation–cyclization of dimethylureido-substituted dodecaborate.

The reactions summarized in this conference report demonstrate the rapid progress that has been made in terms of derivatization of boron cages. The B–H activation processes not only allow for product preparation within a minimum number of steps, but they also provide routes to compounds hitherto inaccessible by electrophilic substitution or halogenation–cross coupling. Control over regioselectivity and degree of substitution as well as novel substituents that can be introduced at the boron vertices are outstanding features of transition metal-catalyzed B–H functionalization. This field has undoubtedly gained momentum, and it should be mentioned that this fact is emphasized by very recent results not presented at the IMEBORON-16 conference. Among them are protocols to synthesize novel halogenated, arylated, acyloxylated and sulfonamidated o-carboranes as well as fused heterocycles of monocarba-closo-dodecaborates [56], [57], [58], [59], [60], [61], [62]. Such an extension of the synthetic palette will have a significant impact on tailor-made boron clusters and corresponding applications.