Diverse reactivity of a tricoordinate organoboron L2PhB: (L = oxazol-2-ylidene) towards alkali metal, group 9 metal, and coinage metal precursors

The reactivity of a tricoordinate organoboron L2PhB: (L = oxazol-2-ylidene) 1 towards metal precursors and its coordination chemistry were comprehensively studied.


Introduction
Transition metal complexes featuring boron-based ligands have attracted signicant attention in boron chemistry because of the fundamental importance of understanding basic metalboron bonding properties and their potential applications as catalysts in organic synthesis. 1 Depending on the coordination number of boron and the metal-boron bonding mode, they are classied as boron cluster, borane, boryl, borylene, and boride complexes. In addition to several conventional strategies for the preparation of complexes featuring unique metal-boron bonding interactions, a new general methodology, namely direct transmetalation between isolable nucleophilic boron species and metal precursors, has been developed in recent years ( Fig. 1). In 2007, Nozaki, Yamashita et al. reported the rst nucleophilic attack of boryl lithium I 2 on coinage metal chlorides supported by triphenylphosphine (Ph 3 P) or N-heterocyclic carbene (NHC) ligands which afforded the corresponding boryl complexes Ia possessing two-center two-electron M-B (M ¼ Cu, Ag, Au) bonds. 3 Since then, a large number of boryl complexes Ib have been synthesized by a similar approach, in which the metals varied extensively from rare earth metals and transition metals to main group metals. 4 Similarly, Bertrand et al. illustrated that CAAC-stabilized boryl anion II can also be used as a nucleophile to synthesize the gold-boryl complex IIa. 5 Braunschweig et al. revealed the radical reactivity of the boryl anion III towards heavier tetrel halides to produce IIIa possessing rare B-Sn and B-Pb bonds. 6 They also showed that salt elimination between anionic dimetalloborylene IV and NHCcoinage metal chlorides afforded the trimetallic boride complexes IVa that contain delocalized Mn-B-M (M ¼ Cu, Ag, Au) bonding, 7,8 whereas IVb was obtained when Ph 3 P-gold chloride was used. 8 In 2012, the same group isolated a novel planar tetra-metalated boron complex Va from the reaction of V with (NHC)AuCl, and proposed unique s and p metal-boron bonding interactions in Va. 9 Although a variety of M-B bonding modes have been achieved through reactions between metal precursors and nucleophilic boron species as mentioned above, such an approach is limited to the construction of M-B single bonds mainly in their covalent fashion, and cannot be applied in the formation of the B:/M bonding mode despite the signicance from the fundamental and application points of view. 1 In addition, conventional synthetic methodology for the complexes bearing B:/M bonds cannot be applicable to all metals due to the lack of the reducing strength of some metals towards boron-halide bonds, such as for coinage metals. Undoubtedly, it will be straightforward if direct coupling between metal precursors and ligands possessing a lone pair on the boron is available. In 2011, Bertrand and co-workers reported the rst isolation of tricoordinate nucleophilic organoboron VI that is isoelectronic with amines (Fig. 2). 10,11 They also developed an efficient synthetic route to relevant derivatives VII and VIII. 12 Despite the development of unique compounds VI-VIII, studies dealing with them are limited to their protonation reaction with Brønsted acids that afforded boronium species, and oxidation to generate radical cations. Recently, we reported the synthesis of L 2 PhB: (1) (L ¼ oxazol-2-ylidene) and elucidated that 1 can readily react with (thf)Cr(CO) 5 to afford the complex (L 2 PhB) Cr(CO) 5 , underlining its nucleophilic character (Fig. 2). 13 This result prompted us to investigate the further reactivity of 1 towards metal precursors. Herein, we report the reactivity of L 2 PhB: (1) towards alkali and late transition metal derivatives, and discuss the outcome of these reactions based on spectroscopic analysis, single-crystal X-ray diffraction and computational studies.

Reactivity of 1 towards alkali metal salts
First, we attempted the reaction between 1 and lithium tri-uoromethanesulfonate (LiOTf). Treatment of 1 with an equivalent of LiOTf in toluene for 48 h at ambient temperature afforded a yellow precipitate. Aer removal of the solvent under vacuum, the residue was washed with n-hexane to afford a yellow powder of 2 in 90% yield (Scheme 1). The 11 B NMR spectrum of 2 in C 6 D 6 displays a singlet at 1.74 ppm, which is shied only 0.07 ppm downeld compared with that of 1, implying that the boron center is inert towards lithium salt.
Single crystals were obtained from a saturated toluene solution of 2 and the solid state structure was conrmed by an X-ray diffraction study (Fig. 3). 14 Compound 2 contains two L 2 PhB: units and two LiOTf molecules. Two oxygen atoms from each oxazol-2-ylidene coordinate to a lithium atom forming a BC 2 O 2 Li six-membered ring. The two lithium atoms are linked    by two OTf units in which two oxygen atoms from each OTf unit coordinate to different lithium atoms to form a LiO 2 S 2 -O 2 Li eight-membered ring, with the CF 3 groups on the S atoms in a trans orientation. The geometry around the boron centers remains trigonal planar (sum of the angles ¼ 360.0 ) and structural parameters in each of the L 2 PhB: moieties are    nearly identical to those in compound 1. To gain insight into the electronic properties, we performed quantum chemical calculations for 2. Electron delocalization over the two C-B-C units was conrmed in the HOMO (À5.355 eV) and the HOMOÀ1 (À5.357 eV) that are nearly degenerate (Fig. 3). Interestingly, these MOs are lower in energy than the HOMO (À4.80 eV) of compound 1, presumably caused by the coordination of oxygen atoms to the Lewis acidic Li atoms, which reduces the electron density in the L 2 PhB: units. Thus, the nucleophilicity of 1 might be tunable depending on the property of the units coordinating to these O-atoms. The formation of 2 rather than B:/Li coordination can be rationalized by the hard acidic and oxophilic properties of lithium in contrast to the nature of the boron center as a so base. Although compound 2 is thermally stable, it decomposes upon exposure to air. Attempts to react 1 with NaOTf and KOTf, containing metals slightly soer than Li, afforded neither B:/M (M ¼ Na, K) coordination nor the corresponding analogues to 2.

Reactivity of 1 towards group 9 metals
Reactivity of 1 towards late transition metals was investigated employing rhodium and iridium complexes. To a THF solution of 1, 0.5 equivalent of [RhCl(COD)] 2 was added at ambient temperature. Aer removal of the solvent in vacuo, the residue was washed with cold n-hexane to afford a yellow solid of 3 in 45% yield (Scheme 2). The 11 B NMR spectrum of 3 displays a doublet at À21.0 ppm which is due to boron-hydrogen coupling ( 1 J BH ¼ 34.0 Hz). In the 13 C NMR spectrum, a new peak for CH 2 appeared at 39.7 ppm as a doublet ( 1 J RhC ¼ 31.8 Hz), indicating the formation of a rhodium-carbon bond. The solid structure of 3 was determined by a single crystal X-ray diffraction study, which revealed the unique zwitterionic property of 3 involving a boronium cation and an anionic rhodium(I) center (Fig. 4, le). 14 Note that only a few anionic Rh(I) complexes have been described and structurally characterized thus far, and to the best of our knowledge, no rhodium complexes including a boronium fragment have been reported before. 15 The anionic Rh center coordinated by COD is covalently bonded to a chlorine atom and the CH 2 carbon [Rh-Cl: 2.408(3)Å. Rh-CH 2 : 2.100(10)Å]. The distance between the Rh and the B atoms is greater than 3.6791(5)Å, indicating no interaction between them. The geometric parameters around the boron center are almost identical to those previously reported in the boronium [L 2 PhBH]OTf. 13 Reaction of 1 with 0.5 equivalent of [IrCl(COD)] 2 led to the formation of a new complex 4 in 80% yield (Scheme 2). In the 1 H NMR spectrum of 4, a characteristic broad peak was observed at À7.54 ppm, indicating a strong interaction of the H atom with the iridium center. The 11 B NMR spectrum of 4 showed a broad peak at À22.7 ppm, corresponding to the tetracoordinate boron. Single crystals of 4 were obtained from a saturated toluene solution at room temperature. In the solid structure of 4, one of the oxazol-2-ylidenes coordinates to the Ir center in an h 1 -fashion (Fig. 4, right). 14  A proposed reaction pathway for the formation of 3 and 4 is drawn in Scheme 3. As no metal complexes containing B:/M bonding were observed in these reactions even when reactions were conducted at low temperature, the reactions may be

Reactivity of 1 towards coinage metal compounds
Numerous coinage metals featuring boron-based ligand have been isolated and structurally characterized. 18 Although various types of bonding interaction between coinage metals and boron atoms have been described, 19 to the best of our knowledge, the B:/M (M ¼ Cu, Ag, Au) bonding mode is still unknown to date. We were interested in introducing 1 as the ligand onto coinage metals, not only for the fundamental curiosity in developing the novel B:/M bonding, but also considering their further application since a variety of coinage metal complexes exhibit catalytic activity in organic transformation. 20 First, we investigated the reaction between 1 and coinage metal chlorides MCl (M ¼ Cu, Ag, Au). Addition of two equivalents of MCl to a THF solution of 1 at room temperature immediately afforded a black precipitate, presumably metal (0), indicating the reduction of coinage metal chlorides by 1. In fact, quantitative formation of the resulting boronium 5 was observed (Scheme 4, top), and 5 was fully characterized by HRMS and NMR spectroscopy which are identical to previous data for 5 with tri-uoromethanesulfonate ( À OTf) as the counter anion. 13 We reasoned that the presence of a strong s-donating ligand on the M(I) would increase the electron density at the metal center, which will prevent the electron transfer from 1 to the metals. To examine this hypothesis, we chose N-heterocyclic carbene-gold chloride (IPr)AuCl as a precursor. A THF solution of 1 and (IPr) AuCl was stirred at ambient temperature for 10 min. As expected, no precipitation of black metal was detected, and a clear yellow solution was generated instead. Aer evaporation of the solvent in vacuo and washing of the residue with toluene, 6 was obtained as a white solid in 95% yield (Scheme 4, bottom). In the 11 B NMR spectrum of 6, an upeld shi from the precursor was observed at À15.1 ppm, which is in line with the formation of a new gold complex bearing the L 2 PhB: ligand. Recrystallization from a saturated THF solution of 6 at À25 C under argon afforded single crystals, and the solid-state structure of cationic gold(I) complex 6 was conrmed by a single-crystal X-ray diffraction study (Fig. 6). 14   Reactions with (IPr)CuCl and (IPr)AgCl were also examined. A CD 3 CN solution of 1 and (IPr)CuCl was prepared in a J. Young NMR tube. 1 H and 11 B NMR spectra were measured immediately, and showed unidentied complex mixtures. Meanwhile, a reaction mixture of 1 with an equivalent of (IPr)AgCl in CD 3 CN displayed a characteristic signal at À15.6 ppm in the 11 B NMR spectrum, presumably indicating the generation of [(L 2 PhB) Ag(IPr)]Cl. However, this intermediate was not thermally stable: thus, it decomposed completely aer 4 h at ambient temperature, and the formation of [(IPr) 2 Ag]Cl was conrmed concomitant with the reproduction of 1 (see the ESI †). The reactivity of 1 towards triphenylphosphine-ligated group 11 metal chlorides (Ph 3 P)MCl (M ¼ Cu, Ag, Au) was also investigated. A mixture of 1 and an equivalent of (Ph 3 P)CuCl was dissolved in CD 3 CN and the reaction was monitored by NMR spectroscopy (see the ESI †). In the 11 B NMR spectrum, a singlet appeared at À16.8 ppm which is in line with the formation of (L 2 PhB)CuCl 7 (Scheme 5, top). In fact, a quantitative generation of free Ph 3 P was conrmed by 31 P NMR. However, compound 7 gradually decomposed to 5 within 24 h at room temperature. Reaction of 1 with (Ph 3 P)AgCl afforded 5 directly, and no corresponding Ag complex intermediate was detected even when the reaction was conducted at low temperature (Scheme 5, middle). Interestingly, a contrasting result was obtained when (Ph 3 P)AuCl was employed. An acetonitrile solution of 1 and (Ph 3 P)AuCl was stirred at ambient temperature for 10 min. Aer removing the solvent in vacuo and washing the residue with n-hexane, the neutral gold chloride complex 8 was obtained as a white solid in 80% yield. Formation of 8 via the smooth replacement of triphenylphosphine by 1 on the gold atom demonstrates the strong nucleophilicity of the L 2 PhB: ligand 1. Note that a variety of gold complexes featuring strong s-donating neutral ligands are widely used as efficient catalysts in organic synthesis. 23 The 11 B NMR of 8 displays a sharp singlet at À18.6 ppm which is shied upeld compared to that of 1 (À1.1 ppm). The complex 8 is not only moisture sensitive but light sensitive, and quickly decomposes if exposed to air or light.
Single crystals of 8 were obtained from an acetonitrile/ toluene mixed solvent at room temperature, and X-ray diffraction analysis revealed that 8 features an almost linear geometry (B1-Au1-Cl1: 175.2(4) ), and the tetracoordinate boron is bound to the gold(I) center in an h 1 -fashion (Fig. 8). 14 The Au-B distance of 2.166 (14)Å is nearly identical to that of 6. These parameters are in good accordance with the calculated results for the model compounds (Ph 3 P) 2 (BH)AuCl and (NHC Me ) 2 (BH) AuCl. 24 We performed quantum chemical calculations to investigate the property of the Au-B bond in 8. The HOMO of 8 displays the coordination of electrons of the boron to the gold atom (Fig. 9). NBO analysis gave a Wiberg bond index (WBI) value of the Au-B bond (0.50), which is formed from the high-pcharacter hybrid orbital (s: 15.49%, p: 84.47%) of the boron atom and mainly the s-orbital of the gold atom (s: 84.66%, p: 1.72%, d: 13.60%). Natural Population Analysis (NPA) indicates an overall charge transfer of 0.43e from the L 2 PhB: fragment to the gold chloride. The calculated BDE of the donor-acceptor bond between 1 and the gold chloride in 8 is 78.4 kcal mol À1 , which is comparable to that (77.3 kcal mol À1 ) of (Ph 3 P) 2 (BH) AuCl. 24 Comparatively, much lower BDEs were estimated for the corresponding bonds in 1-CuCl (64.3 kcal mol À1 ) and 1-AgCl (55.8 kcal mol À1 ), which may account for the instability of compound 7. An analogous trend of the bond energies for the metal fragments (Au > Cu > Ag) was also found in their carbene, silylene and germylene complexes. 25 In the solid state IR spectrum of 8, a characteristic peak was observed at 586 cm À1 , which is assigned to a Au-B stretch based on computational results (mode 41: 577 cm À1 , see the ESI †).

Preliminary investigation of catalysis with complex 8
Next, we turned our attention to the potential ability of complex 8 as a catalyst. We chose the hydroamination reaction between aniline 9a and phenylacetylene 10a as a model reaction. A brief screening of solvents (Table 1, entries 1-5) revealed that benzene is suitable for the reaction; an imine derivative 11aa was obtained in 94% yield aer 12 h at room temperature (entry 5). During this reaction we observed a main signal in the 11 B NMR spectrum at À16.1 ppm, in addition to the peak at À16. 6 for B(C 6 F 5 ) 4 and a small peak at À19.3 ppm. Control reactions suggest that the main signal may correspond to either (L 2 PhB) Au(C^CPh) or s,p-acetylide complex [(L 2 PhB)Au] 2 (C^CPh)$ B(C 6 F 5 ) 4 , proposing the resting state of the catalyst (see the ESI †). Hence, this result supports the innocence of colloidal gold, demonstrating the essential role of 8 in the catalytic reaction. With AgOTf as an additive, little formation of 11aa (<5%) was observed (entry 6).
The scope of the catalytic reaction was briey examined with various primary amine and terminal alkyne derivatives at ambient temperature ( Table 2). The addition reactions of aniline 9a, p-anisidine 9b or 3,5-dimethylaniline 9c to phenylacetylene 10a or 4-ethynylanisole 10b were effectively catalyzed to afford the corresponding imines in excellent yields: 11aa (98%), 11ab (93%), 11ba (90%), 11bb (94%) and 11ca (91%) (entries 1, 2, 4, 5, 7). Lower activity was observed when the sterically hindered amine substrates 9d, e were used (entries 8-10); i.e. from the reaction of 9e and 10b, only moderate formation of the corresponding imine 11eb (44%) was observed aer 72 h (entry 10). Although a longer reaction time was required, the catalyst 8 was also effective for the hydroamination with more challenging electron-poor alkynes, such as 1-bromo-4-ethynylbenzene 10c. Thus, imine derivatives 11ac, 11bc and 11ec were slowly formed from the reaction of 10c and the corresponding amines (9a, 9b and 9e) in low to good yields (entries 3, 6 and 11, respectively). Not surprisingly, complex 8 also catalyzed intramolecular hydroamination (Scheme 6a and b), in which N-heterocyclic products 13 and 15 were obtained quantitatively. We also tested O-H additions to alkynes with a catalytic amount of 8 (5 mol%). At room temperature, compound 16 was smoothly converted to 2-phenylbenzofuran 17 (96%) (Scheme 6c). When N-hydroxy benzotriazole 18 was treated with phenylacetylene, 19 was formed in good yield (83%) via an intermolecular O-H addition (Scheme 6d). Because selective C-H activation as well as the formation of C-C bonds is signicant in organic synthesis, we also investigated the catalytic transformation of compounds 20 and 22. In the reaction with 1-naphthalenyl propargyl ether 20, addition of a C(sp 2 )-H bond across the C-C triple bond took place which afforded a benzo[h]chromene derivative 21 in 93% yield aer 6 h (Scheme 6e). The cyclisation reaction of compound 22 also occurred effectively via activation of a C(sp 3 )-H bond, and b-ketoester 23 was obtained in good yield (86%) (Scheme 6f). Since it has been already demonstrated that with colloidal gold, higher temperature, irradiation of light, or longer reaction times are required to conduct the relevant reactions, 26 its involvement may be ruled out here. Note that some of the results shown here were comparable to or even better than those with (Ph 3 P)AuCl as a precatalyst under similar conditions. 27 It is also noteworthy that these results demonstrated the rst application of metal complexes featuring a neutral boron-based ligand in catalysis.

Conclusions
In conclusion, we have demonstrated the diverse reactivity of L 2 PhB: (1) towards metal precursors. Addition of hard acidic LiOTf to 1 induced coordination of the oxygen atoms in the oxazol-2-ylidenes to the Li atoms, and the unique compound 2 containing two L 2 PhB: moieties bridged via a cyclic Li(OTf) 2 Li unit was obtained. Reaction of 1 with [RhCl(COD)] 2 afforded the hitherto unknown zwitterionic complex 3 involving an anionic rhodium and a boronium cation whereas iridium borane complex 4 featuring a 3c-2e Ir-H-B bond was isolated from reaction with [IrCl(COD)] 2 . In the formation of complexes 3 and 4, deprotonation of the metal hydrides by the boron atom is proposed, which illustrates the strongly basic property of the boron center in 1. Direct complexation between 1 and gold chloride derivatives bearing carbene or phosphine ligands produced the rst isolable gold complexes 6 and 8 featuring B:/Au bonding. Depending on the ligand on the gold chloride precursors, either a cationic or a neutral gold(I) complex supported by the L 2 PhB: ligand can be obtained selectively via Au-Cl bond dissociation or ligand exchange, respectively. X-ray diffraction analysis and computational studies revealed the unique Au-B bonding formed by the coordination of lone-pair electrons on the boron atom to the gold center as well as the strong s-donating property of the L 2 PhB: ligand. Complex 8 showed good catalytic activity when employed as a precatalyst for both intra-and inter-molecular X-H (X ¼ N, O, C) addition to alkynes. Further developments of metal complexes bearing 1 and its application in catalysis are currently under investigation in our laboratory.