BH Vertices Research Articles

Directing-Group-Assisted Transition-Metal-Catalyzed Selective BH Functionalization of o-Carboranes

AbstractCarboranes are a type of molecular clusters consisting of carbon, hydrogen, and boron atoms. They possess unique characteristics, such as three-dimensional aromaticity, icosahedral geometry, and robustness. Functionalized carboranes have been utilized in various fields, including medicine, materials, and organometallic/coordination chemistry. In this context, selective functionalization of o-carboranes has received tremendous attention, specifically in the regio- and enantioselective modification of the ten chemically similar BH vertices within the carborane cage. In recent years, significant progress has been made in catalytic vertex-specific BH functionalization, as well as achieving enantioselective functionalization of the cage BH. This review provides an overview of the recent advancements in this research field.1 Introduction2 Carboxy-Assisted BH Functionalization2.1 Formation of B–C Bonds2.2 Formation of B–N Bonds2.3 Formation of B–O Bonds2.4 Formation of B–X Bonds2.5 Consecutive Formation of B–C and B–Y (Y = N, O) Bonds3 N-Based Directing-Group-Assisted B–H Functionalization3.1 Acylamino as a Directing Group3.2 Amide as a Directing Group3.3 Pyridyl as a Directing Group3.4 Imine as a Directing Group4 Phosphinyl-Assisted Cage B–H Functionalization5 Bidentate-Directing-Group-Assisted B–H Functionalization6 Other Directing-Group-Assisted B–H Functionalization7 Conclusions

Recent Advances in Transition Metal-Catalyzed Selective B-H Functionalization of o-Carboranes

Abstract Carboranes are a class of carbon-boron molecular clusters, possessing extraordinary characteristics including three-dimensional aromaticity conjugated by σ-bonds, icosahedral geometry and inherent robustness. They are finding growing applications as valuable building blocks in boron neutron capture therapy agents, pharmacophores, nanomaterials, optoelectronic, organometallic/coordination chemistry and more. Therefore, the effective and controlled functionalization of carboranes has attracted enormous research interests, particularly in regio- and enantio-selective cage BH derivatization among ten chemically similar BH vertices of o-carboranes. Only in the recent few years, significant progress has been made in transition metal catalyzed vertex-specific BH functionalization. This review summarizes recent advances in this research realm.

Synthesis, Structures and Chemistry of the Metallaboranes of Group 4–9 with M2B5 Core Having a Cross Cluster M–M Bond

In an attempt to expand the library of M2B5 bicapped trigonal-bipyramidal clusters with different transition metals, we explored the chemistry of [Cp*WCl4] with metal carbonyls that enabled us to isolate a series of mixed-metal tungstaboranes with an M2{B4M’} {M = W; M’ = Cr(CO)4, Mo(CO)4, W(CO)4} core. The reaction of in situ generated intermediate, obtained from the low temperature reaction of [Cp*WCl4] with an excess of [LiBH4·thf], followed by thermolysis with [M(CO)5·thf] (M = Cr, Mo and W) led to the isolation of the tungstaboranes [(Cp*W)2B4H8M(CO)4], 1–3 (1: M = Cr; 2: M = Mo; 3: M = W). In an attempt to replace one of the BH—vertices in M2B5 with other group metal carbonyls, we performed the reaction with [Fe2(CO)9] that led to the isolation of [(Cp*W)2B4H8Fe(CO)3], 4, where Fe(CO)3 replaces a {BH} core unit instead of the {BH} capped vertex. Further, the reaction of [Cp*MoCl4] and [Cr(CO)5·thf] yielded the mixed-metal molybdaborane cluster [(Cp*Mo)2B4H8Cr(CO)4], 5, thereby completing the series with the missing chromium analogue. With 56 cluster valence electrons (cve), all the compounds obey the cluster electron counting rules. Compounds 1–5 are analogues to the parent [(Cp*M)2B5H9] (M= Mo and W) that seem to have generated by the replacement of one {BH} vertex from [(Cp*W)2B5H9] or [(Cp*Mo)2B5H9] (in case of 5). All of the compounds have been characterized by various spectroscopic analyses and single crystal X-ray diffraction studies.

Controlled functionalization of o-carborane via transition metal catalyzed B-H activation.

Carboranes, a class of carbon-boron molecular clusters, are often viewed as three-dimensional analogues of benzene. They are finding increasing applications as useful functional building blocks in materials science, medicine, organometallic/coordination chemistry and more. Thus, functionalization of carboranes has received considerable attention. In comparison with the weakly acidic cage C-H bonds that can be readily functionalized, selective cage B-H functionalization among ten chemically similar BH vertices in o-carboranes is very challenging. Only in the recent few years, considerable progress has been made in transition metal catalyzed vertex-specific BH functionalization. This review summarizes recent advances in this research area.

Heterometallic boride clusters: synthesis and characterization of butterfly and square pyramidal boride clusters*

Abstract A number of heterometallic boride clusters have been synthesized and structurally characterized using various spectroscopic and crystallographic analyses. Thermolysis of [Ru3(CO)12] with [Cp*WH3(B4H8)] (1) yielded [{Cp*W(CO)2}2(μ 4-B){Ru(CO)3}2(μ-H)] (2), [{Cp*W(CO)2}2(μ 5-B){Ru(CO)3}2{Ru(CO)2}(μ-H)] (3), [{Cp*W(CO)2}(μ 5-B){Ru(CO)3}4] (4) and a ditungstaborane cluster [(Cp*W)2B4H8Ru(CO)3] (5) (Cp*=η 5-C5Me5). Compound 2 contains 62 cluster valence-electrons, in which the boron atom occupies the semi-interstitial position of a M4-butterfly core, composed of two tungsten and two ruthenium atoms. Compounds 3 and 4 can be described as hetero-metallic boride clusters that contain 74-cluster valence electrons (cve), in which the boron atom is at the basal position of the M5-square pyramidal geometry. Cluster 5 is analogous to known [(Cp*W)2B5H9] where one of the BH vertices has been replaced by isolobal {Ru(CO)3} fragment. Computational studies with density functional theory (DFT) methods at the B3LYP level have been used to analyze the bonding of the synthesized molecules. The optimized geometries and computed 11B NMR chemical shifts satisfactorily corroborate with the experimental data. All the compounds have been characterized by mass spectrometry, IR, 1H, 11B and 13C NMR spectroscopy, and the structural architectures were unequivocally established by crystallographic analyses of clusters 2–5.

Iridium-catalysed regioselective borylation of carboranes via direct B\u2013H activation

Carboranes are carbon–boron molecular clusters, which can be viewed as three-dimensional analogues to benzene. They are finding many applications in medicine, materials and organometallic chemistry. On the other hand, their exceptional thermal and chemical stabilities, as well as 3D structures, make them very difficult to be functionalized, in particular the regioselective functionalization of BH vertex among ten similar B–H bonds. Here we report a very efficient iridium-catalysed borylation of cage B(3,6)–H bonds of o-carboranes with excellent yields and regioselectivity using bis(pinacolato)diboron (B2pin2) as a reagent. Selective cage B(4)–H borylation has also been achieved by introducing a bulky TBDMS (tert-butyldimethylsilyl) group to one cage carbon vertex. The resultant 3,6-(Bpin)2-o-carboranes are useful synthons for the synthesis of a wide variety of B(3,6)-difunctionalized o-carboranes bearing cage B–X (X=O, N, C, I and Br) bonds.

The Wade–Mingos rules in seven-vertex dimetallaborane chemistry: Hydrogen-rich Cp2M2B5H9 systems of the second and third row transition metals

The complete series of hydrogen-rich seven-vertex cyclopentadienyl dimetallaboranes Cp2M2B5H9 (Cp = η5-C5H5; M = Ir, Ru/Os, Re, Mo/W, Ta), including the experimentally known Mo and W derivatives, have been examined by density functional theory. The central M2B5 polyhedra depend on the skeletal electron count and, in most cases, can be related to the Wade–Mingos rules. Thus the six lowest-energy Cp2Ir2B5H9 structures with 18 Wadean skeletal electrons are all based on the expected nido polyhedron obtained by removal of a degree 5 vertex from a D2d bisdisphenoid. Similar geometries are also found for most of the lowest energy Cp2M2B5H9 structures (M = Ru, Os). The three lowest-energy Cp2Re2B5H9 structures all have central Re2B5 capped octahedra in accord with their 14 Wadean skeletal electrons. The low energy structures of the more hypoelectronic Cp2M2B5H9 (M = Mo, W, Ta) systems have bicapped trigonal bipyramid geometries in which the two MB2 faces of a central trigonal bipyramid are each capped by BH vertices. For Cp2M2B5H9 (M = Mo, W) the experimentally realized symmetrical D2h bicapped trigonal bipyramid structure of this type is obviously very favorable since it lies more than 23 kcal/mol in energy below the next lowest energy structure.

ChemInform Abstract: Transition Metal Mediated Functionalization of O‐Carboranes

AbstractReview: 10 refs.

ChemInform Abstract: Synthesis, Structure, and Reactivity of 13‐ and 14‐Vertex Carboranes

AbstractReview: 39 refs.

Transition metal mediated functionalization of o-carboranes

Carboranes are a class of boron hydride clusters in which one or more of the BH vertices are replaced by CH units. Unlike small boranes, carboranes are kinetically and thermodynamically very stable as well as relatively chemically inert, which are often called three-dimensional relatives of benzenes. They are finding many applications in medicine as boron neutron capture therapy (BNCT) agents, in nanomaterials/supramolecular design as building blocks, and as ligands for transition metals [1]. However, their unique structures make derivatization difficult, which limits their application scope. To this end, there is a need to develop new methodologies for the functionalization of carboranes. In view of the widespread use of transition metals in synthetic chemistry, we initiated a research program to develop transition metal-mediated/catalyzed synthetic methods for cage C–H/B–H functionalization of carboranes. This short article highlights the recent development and addresses the problems in this growing area. Earlier experimental results show that the M–Ccage (Ccage: hypervalent carbon) σ bond in transition metal-carboranyl complexes are generally inert toward various electrophiles, and hence significantly different from traditional M–C (C: tetravalent carbon) bonds [2]. This can be ascribed to steric effects resulting from the carboranyl moiety. To overcome this steric problem and to activate the non-traditional M–Ccage bonds, a series of transition metal-o-carboryne (ocarboryne = 1,2-dehydro-o-carborane) complexes have been prepared as the formation of metallacyclopropane opens up the coordination sphere and creates ring strain, facilitating the reactions of M–Ccage bonds with electrophiles (Chart 1). In fact, metal-o-carboryne complexes are important intermediates for the preparation of a large class of cage C–H functionalized carboranes. For instance, the nickel-oChart 1 o-Carborane and its derivatives.

Synthesis, structure, and reactivity of 13- and 14-vertex carboranes.

Carboranes are a class of polyhedral boron hydride clusters in which one or more of the BH vertices are replaced by CH units. Their chemistry has been dominated by 12-vertex carboranes for over half a century. In contrast, knowledge regarding supercarboranes (carboranes with more than 12 vertices) had been limited merely to possible cage geometries predicted by theoretical work before 2003. Only in recent years has significant progress been made in synthesizing supercarboranes. Such a breakthrough relied on the use of Carbon-Atoms-Adjacent (CAd) nido-carborane dianions or arachno-carborane tetraanions as starting materials. In this Account, we describe our work on constructing and elucidating the chemistry of supercarboranes. Earlier attempted insertions of the formal [BR](2+) unit into Carbon-Atoms-Apart (CAp) 12-vertex nido-[7,9-C2B10H12](2-) did not produce the desired 13-vertex carboranes. Such failure is often attributed to the extraordinary stability of the B12 icosahedron (the "icosahedral barrier"). However, the difference in reducing power between CAp and CAd 12-vertex nido-carborane dianions had been overlooked. Our results have shown that CAd nido-carborane dianions are weaker reducing agents than the CAp isomers, allowing a capitation to prevail over a redox reactivity. This finding provides an entry point to the synthesis of supercarboranes and a series of 13- and 14-vertex closo-carboranes have been prepared and structurally characterized. They share some chemical properties with those of 12-vertex carboranes; on the other hand, they have their own unique characteristics. For example, a 13- vertex closo-carborane can undergo single electron reduction to give a stable carborane radical anion with [2n + 3] framework electrons, which can accept one additional electron to form a 13-vertex CAd nido-carborane dianion. 13-Vertex closo-carborane can also react with various nucleophiles to afford the cage carbon and/or boron extrusion products closo-CB11(-), nido-CB10(-), closo-CB10(-), and closo-C2B10, depending on the nature of the nucleophiles. Studies of supercarboranes remain a relative young research area, particularly in comparison to the rich literature of icosahedral carboranes with 12-vertices. Other supercarboranes are expected to be prepared and structurally characterized as the field progresses, and the results detailed here will further these efforts.

Evidence for an Intermediate in the Methylation of CB11 H12 - with Methyl Triflate: Comparison of Electrophilic Substitution in Cage Boranes and in Arenes.

The trideuteriomethylation of BH vertices in CB11 H12 - and its derivatives with CD3 OTf (OTf=triflate, trifluoromethanesulfonate) yields a mixture of BCD3 and BCHD2 substitution products, thus demonstrating the intermediacy of a species with a long enough lifetime for hydrogen scrambling between the boron vertex and the methyl substituent. No such scrambling is observed if CD3 OTf is used to methylate toluene. According to density functional theory calculations, the intermediate in BH vertex methylation is a three-center bonded σ adduct of a methyl cation to the BH bond and the proton scrambling occurs via a transition structure containing a distorted square-pyramidal methane attached axially to a "naked" boron vertex. The subsequent proton or deuteron loss is presently not understood in detail. A general comparison of electrophilic substitution on closo-boranes and arenes is provided and similarities as well as differences are discussed. A recalculation of the optimized geometry of the CB11 Me12 . radical produced a second Jahn-Teller distorted minimum and resulted in a somewhat improved agreement between calculated and measured proton hyperfine coupling constants.

Hypoelectronic diruthenaboranes and diosmaboranes having eight to twelve vertices: capped isocloso and bicapped closo structures

The oblate deltahedral structures for the experimentally known 2n − 4 Wadean skeletal electron systems Cp2Re2Bn−2Hn−2 (Cp = η5-R5C5 ligand; R = H, CH3; n = 8 to 12) differ radically from the more nearly spherical 2n skeletal electron isocloso deltahedra such as the known Cp2Fe2C2Bn−4Hn−2 and 2n + 2 skeletal electron closo systems such as the extensive series of known Cp2Co2C2Bn−4Hn−2 derivatives. Density functional theory has now been used to investigate the intermediate “missing” systems, namely the 2n − 2 Wadean skeletal electron systems Cp2M2Bn−2Hn−2 (M = Ru, Os). The lowest energy such structures are predicted to have central deltahedra with n − 1 or n − 2 vertices capped by the remaining BH vertices. This generates larger deltahedra having one or two degree 3 vertices leading to tetrahedral cavities. In all of the lowest energy Cp2M2Bn−2Hn−2 structures the metal atoms are located at the highest degree vertices, most frequently degree 6 vertices. These metal vertices share a deltahedral edge with M–M distances ranging from 2.7 to 2.9 A and Wiberg bond indices of ∼0.3 to ∼0.4. The structural pattern changes drastically for the 12-vertex systems Cp2M2B10H10 (M = Ru, Os) where the lowest energy structures are “isoisocloso” deltahedra having two adjacent degree 6 vertices for the metal atoms as well as eight degree 5 and two degree 4 vertices. Higher energy Cp2M2B10H10 structures are based on a regular icosahedron with all degree 5 vertices having unusually short MM edges of ∼2.2 to ∼2.3 A with corresponding high Wiberg bond indices of ∼1.3 to ∼1.5 suggesting formal metal–metal triple bonds.

Chemical bonding in oblatonido ditantalaboranes and related compounds

The recently discovered ditantalaboranes Cp2Ta2B n H n+6 (n = 4, 5) are isoelectronic with the previously discovered dimetallaboranes Cp2M2B n H n+4 of the group 6 metals Cr, Mo, and W where Cp = η5-cyclopentadienyl or substituted cyclopentadienyl. Their oblatonido polyhedral structures can be derived from the oblate (flattened) deltahedra of the oblatocloso dirhenaboranes Cp2Re2B n+1H n+1 by removal of an equatorial BH vertex with adjustment of the skeletal electron count by changing the metal atoms and adding hydrogen atoms. In these oblatocloso dirhenaborane deltahedra, the approximately antipodal rhenium atoms are close enough together to form a formal Re=Re double bond with lengths in the range 2.69–2.82 A. Similarly, short M=M distances are maintained in the related oblatonido derivatives Cp2Ta2B n H n+6 (n = 4, 5) and Cp2M2B n H n+4 (M=Cr, Mo, W). However, the synthesis of Cp2Ta2B n H n+6 (n = 4, 5) from CpTaCl4 + LiBH4/BH3 also gives a less-reduced product Cp2Ta2Cl2B5H11 with a longer Ta–Ta distance of ~3.2 A. This may be regarded as a formal single bond bridged by one of the hydrogen atoms. Vertices of degree 5 (excluding terminal atoms/groups but not edge-bridging hydrogens) are sites of highest stability/lowest chemical reactivity not only in metal-free boranes but also in the dimetallaboranes discussed in this paper. For example, all four boron vertices in Cp2Ta2B4H10 have the favorable degree or 5. The recently discovered ditantalaboranes Cp2Ta2B n H n+6 (n = 4, 5) are isoelectronic with the previously discovered dimetallaboranes Cp2M2B n H n+4 (M=Cr, Mo, W). They have oblatonido structures derived by removal of one BH vertex from the flattened oblatocloso polyhedra of the dirhenaboranes Cp2Re2B n H n . All of these structures have internal M=M formal double bonds inside the deltahedron.

Cobalt-Mediated Selective B−H Activation and Formation of a Co−B Bond in the Reaction of the 16-Electron CpCo Half-Sandwich Complex Containing ano-Carborane-1,2-dithiolate Ligand with Ethyl Diazoacetate

The reaction of the 16-electron half-sandwich complex CpCo(S(2)C(2)B(10)H(10)) (1; Cp = cyclopentadienyl) with ethyl diazoacetate (EDA) at ambient temperature leads to compounds CpCo(S(2)C(2)B(10)H(10))(CHCO(2)Et) (2), CpCo(S(2)C(2)B(10)H(8))(CHCO(2)Et)(CH(2)CO(2)Et)[CH(CO(2)Et)(CH(2)CO(2)Et)] (3), CpCo(S(2)C(2)B(10)H(9))(CH(2)CO(2)Et)(CHCO(2)Et)(2) (4), CpCo(S(2)C(2)B(10)H(9))(CHCO(2)Et)(CH(2)CO(2)Et) (5), and CpCo(S(2)C(2)B(10)H(9))(CHCO(2)Et)(2)(CH(2)CO(2)Et) (6). In 2, the EDA molecule has been inserted into one Co-S bond in 1 with the loss of N(2) to form an 18-electron compound containing a three-membered metallacyclic ring. In 3, two B-H bonds of the carborane cage have been activated and the unusual B4-H bond activation leads to the formation of a stable Co-B bond. Two EDA molecules are inserted into the Co-B3 bond to generate an unexpected six-membered heterocyclic ring Co-B-B-C-C-O. In 4, a stable Co-B bond is present as well but in the position B3/B6, and two EDA molecules are inserted into one Co-S bond to produce a five-membered heterocyclic ring Co-C-C-C-O. In 5, one EDA is inserted into the Co-B bond with the formation of a C-B bond in the position B3/B6. One more EDA is inserted into the Co-S bond in 5 to generate 6. Upon heating, 6 loses the BH vertex close to the two carbon atoms to lead to CpCo(S(2)C(2)B(9)H(9))(CHCO(2)Et)(CH(2)CO(2)Et)(2) (7) containing a nido-C(2)B(9) unit. All of the new compounds 2-7 were characterized by NMR spectroscopy ((1)H, (11)B, and (13)C), mass spectrometry, IR spectroscopy, and elemental analysis, and their solid-state structures were further characterized by X-ray structural analysis.

Reactivity Studies on 2,3,4,5‐Tetraethyl‐1,6‐diiodo‐2,3,4,5‐tetracarba‐nido‐hexaborane(6): Synthesis and Structures of New C4B2nido‐Carborane Derivatives

AbstractThe reactivity of the title compound 2,3,4,5‐tetraethyl‐1,6‐diiodo‐2,3,4,5‐tetracarba‐nido‐hexaborane(6) (1) towards a variety of nucleophiles is reported. When 1 is treated with RC2Li (R = Ph, tBu, SiMe3, p‐tolyl) and Ph2PLi, the corresponding new C4B2 nido‐carborane derivatives 2a−d and 3, respectively, are obtained by selective substitution at the basal B−I group, whereas the apical B−I bond remains inert. The reaction of 1 with K[(η5‐C5H5)Fe(CO)2] affords the novel species 4, which contains a CpFe(CO)2 fragment σ‐bonded to the basal boron atom of the C4B2 cluster. Attempts to isolate 4 by chromatography on silica gel led to cleavage of the Fe−B bond and formation of compound 5, containing a basal BH vertex, which can be separately prepared by the reaction of 1 and LiBEt3H. In the reactions of 1 with PhLi and Me3SnLi, respectively, 6a and 7a are formed as the predominant B6‐substituted carborane products. In addition, trace amounts of 1,6‐disubstituted species 6b and 7b are also detected. The substitution at the apical B−I group in 2a with an alkynyl group is effected by a Pd0‐catalyzed Negishi‐type cross‐coupling reaction to give compound 8. The reaction of 1 and monolithio‐o‐carborane affords the C4B2−C2B10 carborane 10, while the reaction of 2a and Co2(CO)8 furnishes the double cluster 9 (C4B2−C2Co2). In both 9 and 10 two different types of clusters are directly connected by a B−C bond. Compound 1 reacts with ClZnC6H4C6H4ZnCl in the presence of the catalyst Pd(PPh3)4 to give 11, in which two C4B2 clusters are linked by a C6H4C6H4 unit. The constitutions of the products follow from spectroscopic data, and X‐ray diffraction analyses for 2a, 2d, 4, 8, 9, and 10. (© Wiley‐VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004)

Mechanism of the opening of the closo-NB11 clusters by bases

Abstract The icosahedral aza-closo-dodecaboranes RNB11H11(R = H, Me, Ph) are opened by neutral bases L or anionic bases X– to give the nido-species RNB11H11L or [RNB11H11X]–, respectively, with a nonplanar pentagonal aperture; the N atom and a BHB hydrogen bridge are situated at opposite sides of the aperture. The BL or the BX vertex is found in the aperture either adjacent to the hydrogen bridge (type 1; C1) or adjacent to nitrogen (type 2;C1) or off the aperture adjacent to nitrogen on a mirror plane (type 3;Cs). At any rate, the isomer of type 1 is the primary product, which may rearrange to yield an isomer of type 3 via an isomer of type 2. Working in deuteromethanol shows that the bridging H atom originates from the primarily attacked BH vertex. The process from RNB11H11 to its base adduct of type 3 includes the opening and the closure of skeletal BN bonds and the jumping of the extra-H atom from endo into bridging positions and vice versa, whereas the base does not alter its position. The application of the opening process to a series of aza-closo-dodecaboranes with non-hydrogen boron ligands confirms that only atoms of the ortho-belt of the starting material are involved in structural changes. The elementary steps from the closo -species to the three isomers are identified as a [3c, 1c] collocation and subsequent [3c, 2c] translocations in the picture of molecular orbitals localized over three or two centers or to one center (lone pair).

Synthesis and X-ray Structure of the Monocarbon Rhodacarborane Complex [RhBr(PPh3)(η5-7-NH2But-8-CH2CO2Et-7- CB10H9)

The electronically unsaturated complexes [RhX(PPh3)(η5-7-NH2But-7-CB10H10)] (X = Br, Cl) react with 1 mol equiv of the diazo compound N2C(H)CO2Et in CH2Cl2 at room temperature to give the species [RhX(PPh3)(η5-7-NH2But-8-CH2CO2Et-7-CB10H9)] resulting from insertion of the alkylidene fragment C(H)CO2Et into a BH vertex α to carbon in the CBBBB ring ligating the rhodium.

Synthesis of the complexes [N(PPH 3) 2][M(CO) 2(η 3-C 3H 5)-(η 6-7, 9-C 2B 10H 10Me 2)](M = Mo or W): crystal structure of [N(PPH 3) 2][WBr(CO) 3(η 6-7,9-C 2B 10H 10Me 2)

Treatment of the compounds [MBr(CO) 2(NCMe) 2(η 3-C 3H 5)] (M = Mo or W) with Na 2[7,9-C 2B 10H 10Me 2] in tetrahydrofuran, followed by addition of [N(PPh 3) 2]Cl, affords the salts [N(PPh 3) 2][M(CO) 2(η 3-C 3H 5)(η 6-7,9-C 2 B 10H 10Me 2)]. Reactions of [N(PPh 3) 2] 2[M(CO) 3(η 6-7,9-C 2B 10H 10Me 2)] with an excess of allyl bromide generate the complexes [N(PPh 3) 2][MBr(CO) 3 (η 6-7,9-C 2B 10H 10Me 2)] via an allyl-coupling process. The structure of the latter (M = W) has been determined by X-ray diffraction. The preparation of the molybdenum salt [N(PPh 3) 2] [MoBr(CO) 3(η 6-7,9-C 2B 10H 10Me 2)] by this route was accompanied by the formation of small amounts of the complex [N(PPh 3) 2][MoBr(CO) 3(η 5-7,9-C 2B 9H 9Me 2)] resulting from ejection of a BH vertex from the cage. Indeed, it was observed that in CH 2Cl 2 solutions, the former species was slowly converted into the latter at ambient temperatures. Protonation (aqueous HBr) of the salts [N(PPh 3) 2][M(CO) 2(η 3-C 3H 5)(η 6-7,9-C 2B 10H 10Me 2)] in the presence of an atmosphere of CO also yields the compounds [N(PPh 3) 2][MBr(CO) 3(η 6-7,9-C 2B 10H 10Me 2)], but if the molybdenum compound is protonated in the absence of CO the complex [N(PPh 3) 2][MoBr(CO) 3(η 5-7,9-C 2B 9H 9Me 2)] is formed directly. The NMR data for the new compounds are reported and discussed, as are the possible pathways of their formation.