Ruthenium(II) pincer complexes featuring an anionic CNC bis(1,2,3-triazol-5-ylidene)carbazolide ligand coordinated in a meridional fashion

bis

This structural motif of a neutral CNC-tridentate ligand that coordinates to the ruthenium metal central at three adjacent co-planar sites, has been exploited also recently in the design of a bis(imidazolylidene)-lutidine based hydroboration catalyst (C) [6] and a bis[(imidazolylidene)alkyl]amino pincer ruthenium complex as a direct hydrogenation catalyst (D) [7]. In contrast, the ruthenium complex E, (Fig. 1) featuring the rigid carbene-based CNC pincer ligand (bimca), (bimca = bis(imidazolylidene)carbazolide) reported by Kunz et al. is the only example, where the anionic CNC pincer ligand coordinates as a tripodal cyclopentadienyl analogue to a RuCp * precursor (Cp * = pentamethylcyclopentadienyl), and not in the usual meridional fashion [8]. The use of other Ru(II) precursors yielded octahedral complexes with coordination of two of the monoanionic pincer ligands [9]. Parallel to this, pincer ligands based on abnormal NHCs (aNHCs) or mesoionic carbenes (MICs) [10] have been much less explored. This class of carbenes include the 1,2,3-triazolylidenes (TRZs), which are stronger donors than NHCs and yield more electron-rich metal complexes upon coordination [11]. The only examples of pincer-TRZ ruthenium(II) complexes reported, feature an extension of the central pyridine (F) [12]/lutidine motif (G) [13] to include the TRZs as the co-planar wing-tip groups in neutral ligand scaffolds. We have recently reported the preparation of the monoanionic CNC-pincer ligand, [C TRZ NC TRZ ] − , featuring a rigid carbazolide flanked by two 1,2,3-triazol-5-ylidenes as the co-planar C-donors, which is readily accessible from its salt precursor [H 3 C TRZ NC TRZ ]PF 6 ·Cl (Scheme 1), after deprotonation with KHMDS [14]. The strongly donating character of the central amido-group and the strong sigma-donating 1,2,3-triazol-5-ylidene moieties, and the steric bulk of the ligand scaffold, allowed for the isolation of reactive transition metal complexes [14a,15] and the use of the CNC-pincer metal complexes as highly selective alkyne functionalization catalysts [14b]. It was therefore reasoned that similar steric and electronic direction could be achieved in the coordination of [C TRZ NC TRZ ] − to an appropriate ruthenium(II) precursor.  Herein, we report the synthesis of three octahedral Ru(II)-pincer complexes with the complexation of one CNC-ligand, the TRZ-analogue of the bimca ligand (see Scheme 1).

Solvents and reagents
All synthetic manipulations, unless otherwise stated, were performed under N 2 gas or Ar gas atmosphere using oven or flame dried glassware and standard Schlenk or vacuum line techniques. Air sensitive solids where stored and handled in a PureLab HE glove box. Preparation of NMR and crystallization samples that also require an inert atmosphere were done in the glove box. [RuCl(H)(AsPh 3 ) 3 (CO)] was prepared as previously reported [16]. All other reagents were obtained from commercial sources and were used without any further purification.
Unless otherwise stated, only anhydrous solvents were used during experimental procedures. Solvents were dried using a solvent purification system (MBraun SPS). Anhydrous THF and Et 2 O were obtained after distillation over sodium and benzophenone ketyl under a N₂ gas atmosphere. Anhydrous PhMe and hexane were obtained after distillation over sodium under a N₂ gas atmosphere. Anhydrous CH₂Cl₂ was obtained after distillation over calcium hydride under a N₂ gas atmosphere. Deuterated benzene was dried over sodium and distilled under an Ar gas atmosphere.
Chemical shift assignment in the 1 H NMR spectra is based on first-order analysis and when required were confirmed by two-dimensional (2D) ( 1 H-1 H) homonuclear chemical shift correlation (COSY) experiments. The 13 C shifts were obtained from proton-decoupled 13 C NMR spectra. Where necessary, the multiplicities of the 13 C signals were deduced from proton-decoupled DEPT-135 spectra. The resonances of the proton-bearing carbon atoms were correlated with specific proton resonances using 2D ( 13 C-1 H) heteronuclear single-quantum coherence (HSQC) and heteronuclear multiple bond correlations (HMBC) experiments. Standard Bruker pulse programs were used in the experiments.
Solution IR spectra (ν(CO), and ν(Ru-H)) were recorded on a Bruker ALPHA FT-IR or on a JASCO FT/IR-6200 spectrometer with a NaCl cell, using CH 2 Cl 2 as solvent, or as a KBr pellet. The range of absorption measured was from 4000-600 cm −1.
Electrospray mass spectra (ESI-MS) were recorded on a Micromass Quatro LC instrument or on a Bruker QTOF Mass spectrometer with positive electron spray as the ionization techniques; nitrogen was employed as drying and nebulizing gas at a flow of 4 L/min. The m/z values were measured in the range of 100-1500 with acetonitrile as solvent. Accurate mass measurements were performed by use of a Q-TOF premier mass spectrometer with electrospray source (Waters, Manchester, UK) operating at a resolution of ca. 16 000 (fwhm). Elemental analyses were carried out on a EuroEA3000 Eurovector Analyzer. Single crystal X-ray diffraction data were collected on a Agilent SuperNova diffractometerdifractometer equipped with an Atlas CCD detector using Cu Kα radiation (λ = 1.54184 Å). The single crystal was mounted on a MicroMount polymer tip (MiteGen) in a random orientation.

Crystal structure determination
A single crystal of 1 was selected and mounted on a SuperNova, Dual, Cu at zero, Atlas diffractometer. The crystal was kept at 293(2) K during data collection. Using OLEX2 [17], the structure was solved with the XS [18] structure solution program using Direct Methods and refined with the SHELXL [18]

Synthesis of complex 3, [RuH(CO) 2 (C TRZ NC TRZ )]
The red residue obtained from the reaction to prepare 2, was dissolved in diethyl ether at room temperature. In the absence of light, CO (g) was bubbled through the red solution for 10 min. The vessel was sealed, and the reaction stirred overnight, at which point a slight color change occurred from red to orange-brown. The reaction was left to settle, and the solution filtered off from the precipitate. The solvent was evaporated in vacuo, and the residue was washed with hexanes (4 x 20 mL). The solid was dried under reduced pressure, to yield 3 (68.0 mg, 6.5 x 10 −5 mol, overall yield 17%) as a yellow brown solid. 1 H NMR δ H (C 6 D 6 , 300 MHz) 8 2,141.8,140.6,139.6,137.7,137.1,136.3,136.2,135.5,135.3,130.1,129.8,129.3,129.1,127.5,117.9,117.5,112.9,34.5 (C(CH 3
Unlike the other two examples of Ru-bisTRZ-pyridine/lutidine pincer complexes reported previously, transmetallation from a Ag-precursor was not required for successful complexation [10,11]. The disappearance of the acidic triazolium CH and carbazole NH proton resonances in the 1 H NMR spectrum of 1 confirmed coordination of the ligand to Ru(II), as well as the carbene carbon resonance at 165.2 ppm in the 13 C NMR spectrum. This value corresponds to the TRZ-C carbene resonances for other reported Ru-TRZ complexes, ranging from 161 to 185 ppm, although occurring on the high field end of the range due to the overall neutral charge of the complex compared to mono-and dicationic Ru-TRZ complexes previously reported [10,11,20]. Similarly, the signals due to the carbonyl carbons also display relatively upfield chemical shifts, at 198.7 ppm and 196.0 ppm, compared to, for example, cationic complex F (Fig. 1) where the carbonyl carbon atom resonates at 208. 5 ppm [11], and the carbonyl stretching frequencies are observed at 2028 cm −1 and 1957 cm −1 .

Synthesis and characterization of dicarbonylhydridoridebis(1,2,3-triazol-5-ylidene)carbazolide ruthenium(II) complex
The bis(triazolium) precursor [H 3 C TRZ NC TRZ ]PF 6 ·Cl was triply deprotonated by adding an excess of potassium hexamethyldisilazide (KHMDS). The in situ generated free bis(triazolylidene) reacted with the [RuCl(H)(AsPh 3 ) 3 (CO)] to yield complex 2, with a triphenyl arsine co-ligand (see Scheme 1). Purification of this complex proved non-trivial, and all our attempts to isolate complex 2 in the pure form were unsuccessful -the complex was always accompanied by a residual amount of AsPh s . Nevertheless, we were able to characterize the complex by NMR techniques (see Figs. S3-S4, Supplementary data). Complex [RuH(CO) 2 (CNC)], 3, was generated by bubbling carbon monoxide through an ethereal solution of 2. After workup, 3 was obtained in low yield (overall yield from precursor triazolium salt 17%, Scheme 1).
The 1 H NMR spectrum of 3 displays the signal due to the hydrido ligand at −3.83 ppm (see Fig. 3). This value is downfield shifted compared to the hydrido resonances reported for D and G, ranging from −7.15 to −5.10 ppm [6,11], and also compared to the hydrido chemical shift of 2 (−8.82 ppm). Correspondingly, the 13 C NMR spectrum of 3 shows that both the carbonyl and carbene carbon resonances are observed at 201. 3, 196.4 and 170.6 ppm, respectively, therefore downfield shifted compared to the same signals in 1. In addition, the strongly σ-donating hydrido ligand induces lower carbonyl stretching frequencies observed for 3 (ν CO = 2008, 1944 cm −1 ) in the FT-IR spectrum, than for 1. The Ru-H band observed in the IR spectrum vibrates at a high energy of 2040 cm −1 , although such a high wavenumber ν Ru-H is not unprecedented in the literature [21]. In the case of complex 2, carbene and carbonyl carbon resonances were observed at 205.9 and 178.8 ppm, respectively.

Transfer hydrogenation of ketones
The structural similarity of 1 to the transfer hydrogenation catalysts A and B (Fig. 1), prompted us to test the catalytic activity of 1 in the transfer hydrogenation of acetophenone to 1-phenylethanol, using isopropanol as sacrificial H-donor. The reaction conditions were optimized using different bases (potassium tert-butoxide or cesium carbonate), different substrate:base:catalyst ratios and by varying the temperature and time of the reaction (See Table S1, SI). The optimized conditions of 1 mol% catalyst loading, 10 mol% Cs 2 CO 3 as base at 95 °C for 14 h, yielded quantitative reduction of acetophenone (entry 6, Table S1). In the case of the reaction with 4-bromoacetophenone, the substrate was reduced to its related alcohol in 20% yield under the same reaction conditions (entry 8, Table S1). The possibility of the low activity of 1 as catalytic precursor due to steric factors is discarded [14b], and is rather ascribed to the presence of two carbonyls and a chloride ligand to complete the octahedral coordination sphere of the catalyst precursor as well as the absence of ligand cooperativity, unlike the transfer hydrogenation catalyst G described by van der Vlugt, Elsevier et al. [13], where the presence of more labile phosphine and reactive hydrido ligands bonded to the Ru(CNC)moiety combined with the pincer ligand cooperativity yield a more active catalyst.

Conclusions
We synthesized and characterized a new Ru(II) CNC-pincer complex, with a bis(1,2,3-triazol-5-ylidene)carbazolide ligand with two carbonyls and a chloride as ancillary ligands. The synthetic procedure to this complex circumvents the use of a silver complex intermediate as transmetallation agent. The rigidity of the central carbazole moiety, together with the presence of bulky mesityl substituents at the triazolylidene rings, enforce a mer-geometry of the CNC pincer ligand. This finding contrasts with the recent example reported by Kunz and co-workers, in which a related CNC di-NHC ligand with a carbazole linker displays a CNC-fac conformation [8]. In our case, the steric bulk produced by the mesityl fragments may disfavor the coordinating in the tripodal (fac) form. Our complex constitutes the first example of a CNC-pincer complex of ruthenium where the central N-donor coordinates as a monoanionic X-type ligand flanked by 1,2,3-triazol-5-ylidene moieties. Access to the hydrido-analogue of the complex was attained by the in situ reaction of the ligand precursor with KHMDS to yield the triply deprotonated ligand, which could then be reacted with an appropriate ruthenium(II) carbonyl hydrido precursor and then with carbon monoxide to afford [RuH(CO) 2 (CNC)]. Preliminary studies of the dicarbonyl chloride Ru(II) pincer complex 1 as a catalyst for the transfer hydrogenation of ketones did not yield results as satisfactory as expected. We are currently exploring the use of these complexes as catalysts in other model organic transformations, including amination reactions.