Mesoionic Carbenes in Low- to High-Valent Vanadium Chemistry

We report the synthesis of vanadium(V) oxo complex 1 with a pincer-type dianionic mesoionic carbene (MIC) ligand L1 and the general formula [VOCl(L1)]. A comparison of the structural (SC-XRD), electronic (UV–vis), and electrochemical (cyclic voltammetry) properties of 1 with the benzimidazolinylidene congener 2 (general formula [VOCl(L2)]) shows that the MIC is a stronger donor also for early transition metals with low d-electron population. Since electrochemical studies revealed both complexes to be reversibly reduced, the stronger donor character of MICs was not only demonstrated for the vanadium(V) but also for the vanadium(IV) oxidation state by isolating the reduced vanadium(IV) complexes [Co(Cp*)2][1] and [Co(Cp*)2][2] ([Co(Cp*)2] = decamethylcobaltocenium). The electronic structures of the compounds were investigated by computational methods. Complex 1 was found to be a moderate precursor for salt metathesis reactions, showing selective reactivity toward phenolates or secondary amides, but not toward primary amides and phosphides, thiophenols, or aryls/alkyls donors. Deoxygenation with electron-rich phosphines failed to give the desired vanadium(III) complex. However, treatment of the deprotonated ligand precursor with vanadium(III) trichloride resulted in the clean formation of the corresponding MIC vanadium(III) complex 6, which undergoes a clean two-electron oxidation with organic azides yielding the corresponding imido complexes. The reaction with TMS-N3 did not afford a nitrido complex, but instead the imido complex 10. This study reveals that, contrary to popular belief, MICs are capable of supporting early transition-metal complexes in a variety of oxidation states, thus making them promising candidates for the activation of small molecules and redox catalysis.


■ INTRODUCTION
Almost two decades after the first report of an abnormal 5imidazolinylidene carbene complex, 1 mesoionic carbenes have been developed into a distinguished ligand class. 2,3 Among them, 1,2,3-triazole derived mesoionic carbenes, namely 1,2,3triazolinylidenes, 4 stand out by their modular synthesis via the copper-catalyzed [3 + 2] cycloaddition between azides and alkynes. 5−7 After their initial reporting by Albrecht et al., 8 they quickly became prominent synthetic targets for (electro-) catalysis, 9−23 supramolecular chemistry, 24−27 magnetism, 28 and photochemistry due to their versatile synthesis and comparatively straightforward handling. 29−36 Throughout these studies, a great effort has been made to decipher their electronic structure. Mesoionic carbenes are commonly believed to be strong σ-donor ligands paralleling heteroaryls; however, recent reports emphasize their π-accepting properties 37−39 as demonstrated by the isolation of a reduced triazolinylidene ligand. 40 Nevertheless, most studies targeted hitherto late transition metals or main group elements, while early transition-metal complexes with mesoionic carbenes have been rarely explored. 41−44 Arguably, this can be attributed to the relatively weak bond between N-heterocyclic carbenes and early transition metals. 45 However, this weak bond may be enhanced by harnessing anionic linkers. This strategy has allowed the isolation of a number of interesting metal complexes 46−49 of the early transition metals, 50−71 the lanthanides, 72−80 and the actinides. 81 −86 Among the early transition metals, vanadium chemistry has witnessed a remarkable activity over the past 50 years and has been applied in heterogeneous and homogeneous catalysis, 87 small molecule activation, 88−92 molecular magnetism, 93−96 and spin qubits. 97−100 Narrowing down the field to NHC vanadium complexes, since the first two reports of vanadium NHC complexes in 1994 by Roesky et al. 101 and 2003 by Abernethy et al., 102 the utility of these complexes has mostly been explored in polymerization catalysis. 103−107 Beyond this, only a few other applications of vanadium-NHC complexes have been examined, including small molecule activation 108,109 and the neutralization of chemical warfare agents. 110 Still, most vanadium NHC complexes refer to diamagnetic vanadium (V) complexes, while low-valent vanadium complexes have been rarely investigated. 101,106,111−114 Inspired by Bellemin-Laponnaz's ligand design, employing two anionic redox-active phenolate linkers, 50−54,115−117 we have recently reported the first mesoionic carbene complexes of groups IV, V, and VI based on a 1,2,3-triazolinylidene scaffold. 44 Our initial report focused on niobium as a group V representative, and we thus expand this chemistry herein toward vanadium. We report the ligand's σ-donor strength by comparing the structural, spectroscopic, and electrochemical properties of the triazolinylidene complex 1 with its benzimidazolinylidene (benzNHC) congener 2, proving the triazolinylidene ligand L 1 as the stronger donor. Furthermore, the salt metathesis reactivity of the new triazolinylidene complex 1 is presented, revealing moderate scope. While phenols and secondary amides give good and clean conversion, all other nucleophiles investigated gave no tractable reaction products. We furthermore report on the isolation of various vanadium complexes in the oxidation states +IV and +III, where the latter are potent precursors to vanadium(V) imido complexes.

■ RESULTS AND DISCUSSION
Despite our previous finding that protonolysis between the triazolium salt [H 3 L 1 ][Cl] and Ti(O i Pr) 3 Cl did not lead to quantitative deprotonation of the triazolium salt, 44 we decided to adopt this strategy using VO(O i Pr) 3 as the vanadium source. To our delight, the reaction between the triazolium salt [H 3 L 1 ][Cl] and VO(O i Pr) 3 , followed by the subsequent washing of the crude solids with hexane, afforded the MIC vanadium-oxo complex 1 as a dark green powder in yields of 85% (Scheme 1). The 1 H NMR spectrum of 1 in benzene confirmed the desired transformation due to the absence of the OH and triazolium-5H protons which revealed a C 1 symmetric species in solution. Unfortunately, due to the high quadrupolar moment of the 51 V nucleus, we were not able to observe the characteristic 13 C NMR resonance of the triazolinylidene carbon atom. Nevertheless, the absence of the triazolium-5C resonance at 131.9 ppm in the 13 C NMR of 1 confirms the formation of a triazolinylidene complex of high-valent vanadium (V). Furthermore, a shift of the 51 V resonance to −533 ppm ( Figure S5) in the 51 V NMR indicates the presence of a strong donor ligand. To set this value into context, we also synthesized the benzimidazolinylidene complex 2, recently reported by LeRoux et al. 118,119 The proton NMR of 2 shows a C s symmetric species in solution where the absence of the OH and the benzimidazolium-2H protons indicates also the formation of an NHC complex of vanadium (V). Similar to 1, we could not observe the carbene carbon resonance in the 13 C NMR spectrum of 2. The 51 V NMR signal of 2 is (compared to complex 1) strongly shifted to lower fields resonating at −503 ppm ( Figure S10), indicating a lower donor strength of the benzimidazolinylidene compared to the triazolinylidene ligand. 120 Unambiguous proof for the formation of the NHC and MIC complexes was obtained by single-crystal X-ray diffraction (SC-XRD) analysis. X-ray quality crystals of 1 and 2 could be grown by slow diffusion of pentane into a concentrated solution of the corresponding complexes in toluene or benzene, respectively ( Figure 2). Complex 1 crystallized in the monoclinic space group P2 1 /n as a toluene solvate, while complex 2 crystallized without additional solvent molecules in the asymmetric unit in the orthorhombic space group Pbca. The most striking difference in the molecular conformations of the complexes is that in the case of 2, the benzannulated heterocycle is shifted out of plane compared to the C1−V1 bond axis by 16.7(1)°, while for the triazolinylidene complex 1, this pitch angle along the C1−V1 bond axis was found to be only 0.7(1)°. These structural parameters are well reproduced by density functional theory (DFT) calculations (Table S1) and are also discernible in the coordination environment around the vanadium center. While 1 adopts an almost perfect square pyramidal coordination environment in the solid state (τ 5 = 0.05), complex 2 is distorted with τ 5 = 0.20. The C1−V1 distances in 1 and 2 are 2.055(3) Å and 2.131(3) Å, suggesting a stronger metal carbene interaction in 1 compared to 2. This agrees with the pronounced high-field shift of the 51 V NMR resonances in 1 relative to 2. The V1−O10 distances were determined to be 1.583(3) Å and 1.585(2) Å in 1 and 2, showing only a minor influence of the NHC moiety toward the strength of the VO bonds. However, the influence of the NHC unit on the VO stretching frequencies in the IR is discernible with resonances at 986 cm −1 (calculated: 1026 cm −1 ) and 1000 cm −1 (calculated: 1035 cm −1 ) in 1 and 2. These values are indicative for a weaker VO multiple bond character in 1 in comparison to 2 due to arguably reduced π-donation from the terminal oxo ligand, and thus, corroborate stronger donor properties of the MIC ligand.  To further probe the donor properties of the triazolinversus the benzimidazolinylidene donor, we investigated the complexes by electrochemical methods. Cyclic voltammetry measured in dichloromethane revealed a reversible reduction corresponding to the V(V/IV) redox couple for both complexes ( Figure 1). In agreement with the results from 51 V NMR and IR spectroscopy, the reduction potential (the reductions are vanadium centered, vide infra) for 1 appears 0.19 V cathodically shifted compared to the reduction potential of 2 (Table 1). This suggests a higher electron density at the vanadium center in 1, which is in line with the higher σ-donor character of MIC relative to benzNHC ligands. Additionally, complex 1 showed one reversible (ligandcentered) oxidation, whereas for 2 it is at the edge of the solvent window and thus could not be evaluated ( Figures  S69−S72).
To reveal the site of reduction for the two vanadium complexes, we reduced the complexes with decamethylcobaltocene (Scheme 2). While a THF solution turned greyish upon reduction of 1, the desired product precipitated as a bright green powder for 2. Evans method in CD 2 Cl 2 revealed a magnetic moment of 1. 74 [2] could be grown by the slow evaporation of dichloromethane out of a hexane/ dichloromethane mixture ( Figure 2).
While the general structural factors (e.g., coordination environment) resemble the same trends as the parent vanadium (V) Table S3). This is in line with the larger ionic radius of a vanadium(IV) compared to a vanadium(V) ion, which suggests the reduction is vanadium centered. The center of the redox processes was further corroborated by EPR spectroscopy (Figure 3 [2], respectively, are comparable with other vanadium(IV) complexes, for example, a four-coordinate vanadium alkylidene. 121 The electronic structure was corroborated by scalar relativistic DFT calculations at the ZORA-PBE-D3BJ/def2-TZVPP//ZORA-PBE-D3BJ/def2-SVP level of theory, 122−132 which indicate a vanadium centered SOMO (quasi-restricted orbital QRO, Figure 3, left) with only small orbital overlap with the supporting ligand.
Further evidence for the stronger donor character of the mesoionic carbene ligand L 1 compared to the benzimidazolinylidene L 2 can be extracted from UV−vis spectroscopy. When changing from the MIC complex 1 to the benzNHC complex 2, the charge-transfer (CT) band located at 387 nm for 1 shifts  (see Figure 4). Overall, the results from NMR, electrochemistry, and UV−vis absorption spectroscopy support the notion of MICs being stronger donors than benzNHCs in vanadium complexes.
To further explore the chemical potential of the triazolinylidene complex 1, we turned our interest toward salt metathesis replacing the remaining chloride ligand (Scheme 3). Mixing the parent lithium salts with 1 in −40°C cold diethyl ether, followed by recrystallization from hexane, gave the pure mesitolate (3), 2,6-diisopropylphenolate (4), and 4,4′-ditolylamide (5) complexes in good to moderate yields. To our surprise, the reaction with primary amides (LiNHMes), thiophenolates (KSMes), and phosphanides (KPHMes) failed to give well-defined products under the above-described conditions. Similarly, anionic alkyl or aryl donors, unrelated to their source (lithium or Grignard reagents), resulted in complicated reaction mixtures from which no defined reaction products could be isolated. The NMR spectroscopic and structural analyses ( Figure 5) of the complexes 3−5 resemble the expected characteristics. For further information we refer to Figures S11−S25 and Tables S2 and S3 in the Supporting Information. Notably, we found that the coligand had a strong influence on the UV−vis spectroscopic features of the complexes. While the halide complex 1 was deep green (λ max = 582 nm; ε = 1000 L mol −1 cm −1 ), the phenolate complexes 3 and 4 are a deep purple color (λ max = 534 nm; ε = 4400 L mol −1 cm −1 for complex 4), and the amide complex 5 is dark teal (λ max = 660 nm; ε = 7200 L mol −1 cm −1 , compare also Figure S68).
By further probing the versatility of complex 1, we examined oxo-exchange reactions, with emphasis on generating vanadium imido complexes. To reach this goal, we applied  Inorganic Chemistry pubs.acs.org/IC Article isocyanates as the imido source, liberating carbon dioxide to provide the driving force for this process. Although this strategy was successful for a plethora of vanadium imido complexes, 134,135 in the present case, even at elevated temperatures (60, 80, or 120°C), no conversion could be observed. In one experiment (120°C) with 3,5-bis-(trifluoromethyl)phenyl isocyanate, we observed its cyclotrimerization to yield the corresponding isocyanuric amide. It has been reported that electropositive metals, that is, Lewis acids, catalyze this process. 136−140 However, NHCs and NHOs are also known to be potent catalysts for this transformation. 141,142 As the presence of free carbenes at elevated temperatures (e.g., due to minimal thermal decomposition) cannot be fully ruled out, the catalytically active species remains unclear. Since the direct generation of imido complexes from the oxo complexes had failed, we sought other strategies. Another versatile access to metal imido complexes is the direct reduction of organic azides by low-valent metal complexes. Thus, we initially aimed to generate a low-valent vanadium(III) complex by deoxygenating the parent vanadium(V) complex 1 with triethylphosphine. Although this strategy met with success for the deoxygenation of molybdenum(VI) benzimidazolinylidene complexes, 143,144 no useful reaction products could be isolated in the present case. Similarly, switching to other phosphines such as triphenylphosphine or trimethylphosphine turned out to be unproductive. We consequently turned our focus to installing the triazolinylidene ligand L 1 directly on vanadium(III). 113 Although our attempts to isolate the fully deprotonated and free triazolylidene Li 2 [L 1 ] have failed so far, deprotonation of [H 3 L 1 ][Cl] with LiHMDS (HMDS = hexamethyldisilazide) at room temperature, followed by the immediate addition of the deprotonated triazolinylidene to VCl 3 (THF) 3 resulted in the formation of a brown suspension. After extraction with toluene and washing of the crude solids with hexane, we isolated the desired vanadium(III) complex 6 as an orange powder in 46% yield (Scheme 4). The complex is remarkably sensitive toward air and moisture and decomposes in the glovebox at room temperature within several days even in the solid state. However, it turned out to be stable for a couple of weeks at −40°C. The formation of the desired vanadium(III) complex was initially evident by the strong paramagnetic nature of its 1 H NMR, revealing an effective magnetic moment of 2.71 μ B , which is in line with the presence of a d 2 electron configuration and is comparable to previously reported vanadium(III) complexes. 109 Despite numerous attempts and due to the high sensitivity of the complex, only low-quality crystals of the complex could be obtained from concentrated diethyl ether solutions at −40°C. In any case, the connectivity of the molecule was unambiguously determined, confirming the vanadium(III) oxidation state ( Figure 6, left). In addition to the triazolinylidene and the halide ligands, complex 6 was found to further hold two tetrahydrofuran donors, creating an octahedral coordination environment around the vanadium center. As expected for an S = 1 spin system, the EPR spectrum at room temperature did not reveal any observable signals ( Figure S90). 121,145 Computational investigation revealed both SOMOs to be vanadium centered (Figure 7) with only small orbital overlap with the supporting ligand. Accordingly, and in agreement with the experiment, these calculations indicate an adiabatic singlet−triplet gap of ΔE = −47 kJ mol −1 in favor of the triplet state. Note that the prediction of two SOMOs is consistent with two irreversible waves of 6 observed in the cyclic voltammogram at 0.09 and 1.03 V vs Fc/Fc + in acetonitrile ( Figure S73).
Having the vanadium(III) complex 6 in hand, we turned back our interest to the generation of vanadium(V) imido complexes following an azide reduction strategy. As envisioned, complex 6 reacts smoothly with organic azides such as phenyl, 4-fluorophenyl, or 4-methylphenyl azide to   Figures S32, S38, and S43). To unambiguously prove the structural identity, X-ray quality crystals of complex 8 were grown from concentrated diethyl ether solution (Figure 6, middle). The complex crystallizes in the monoclinic space group P2 1 /n with one molecule in the asymmetric unit. The vanadium center is penta-coordinate by the ligand, the imido nitrogen atom, and a chlorido ligand in a slightly distorted square pyramidal coordination environment (τ 5 = 0.10). The V1−N40 distance was found to be 1.644(2) Å and is comparable to previously reported vanadium(V) imido complexes. The metal carbene distance V1−C1 was found to be 2.048(2) Å, which compares well to complex 1. The structural parameters of the ligand are similar to complex 1 and can be found in the Supporting Information in Tables S2 and  S3.
In sight of the swift reduction of organic azides, we next turned our interest toward the use of "inorganic" azides, which we envisioned to form vanadium nitrido complexes. These complexes are relevant intermediates in the context of dinitrogen activation and valorization. Indeed, reacting complex 6 with 1 equiv of TMS-N 3 resulted in the clean conversion to diamagnetic 10 at 60°C (Scheme 4). After workup, the 1 H NMR spectrum of the new complex shows a single resonance at −0.31 ppm integrating nine protons, which is indicative of the remainder of the TMS group on the nitrogen atom and thus the formation of a TMS-imido complex. Indeed, Mindiola and co-workers recently reported that the cleavage of a TMS group from a TMS-imido ligand to form the corresponding nitride complex is not trivial also for other group 5 metals, for example, tantalum (V). 146 The 51 V-NMR shows a single resonance at −473.5 ppm ( Figure S48) which is high-field shifted compared to the aryl-imido complexes 7−9. However, this high-field shift can be attributed to the stronger donor character of alkyl/TMS imido versus and aryl imido ligands. Unfortunately, due to the quadrupolar moment of vanadium, the detection of both the imido nitrogen atom as well as of the silicon atom using 29 Si NMR spectroscopy failed. Nevertheless, X-ray quality crystals of complex 10 could be grown from slow evaporation of a concentrated diethyl ether solution at room temperature ( Figure 6, right). The structural analysis confirmed the formation of the terminal imido instead of the desired nitrido complex. The structural properties of complex 10 resemble those of complex 8 and are given in the Supporting Information in Tables S2 and S3. In contrast to previous reports, 147 applying sodium azide as a nitrogen/azide source resulted in complicated (paramagnetic) mixtures, from which no defined products could be isolated.

■ CONCLUSIONS
We have extended the use of mesoionic carbenes with an early transition metal, vanadium. Using combined spectroscopic, electrochemical, and computational methods, we have shown that mesoionic carbenes are stronger donors than classical NHCs in early transition-metal chemistry as well. The highvalent oxo-vanadium(V) complexes are of moderate use for salt metathesis, reacting cleanly only with phenolates and secondary amides. Remarkably, the mesoionic carbene ligand supports vanadium in three oxidations states (III/IV/V). This is a rare report of a structurally characterized low-valent vanadium(III) complex supported by an NHC ligand [112][113][114]148 and the first of a mesoionic carbene stabilizing a low-valent early transition metal. Complex 6 is a powerful two-electron reductant and forms the corresponding highvalent vanadium(V) imido complexes with azides. In recent years, NHC-imido vanadium complexes have attracted a large interest as polymerization catalysts as well as in nitrene transfer reactions. 149 Following the concept of extreme π-loading effects 150 and given the numerous examples of superior activity of MIC based catalysts over their NHC congeners, 151 we believe that our findings will create further interest in the use of mesoionic carbenes in early transition-metal-mediated reactions. Additionally, the redox-active nature of the phenolate tethers will also be of large interest in other catalytic reactions as well as small molecule activation.

■ EXPERIMENTAL SECTION
General Remarks. If not otherwise mentioned, all transformations were carried out in an argon-filled glovebox under inert conditions. Solvents were dried by an MBraun SPS system and stored over activated molecular sieves (3 Å) for at least 1 day. C 6 D 6 was dried over sodium/benzophenone and CDCl 3 and CD 2 Cl 2 over calcium hydride, followed by vacuum transfer and three freeze− pump−thaw cycles. The proligand [H 3 L 1 ][Cl] was synthesized following a literature known procedure. 44 LiOMes, LiN(Tol) 2 , and LiNHMes were obtained by deprotonating the corresponding phenol or aniline in pentane using n-BuLi and filtering off the products. In a similar way, KSMes and KPHMes were obtained by deprotonating the corresponding thiophenol and primary phosphine using KHMDS in toluene. 152 4-Methylphenyl azide 153 and 4-fluorophenyl azide 154 were synthesized following previously reported methods using tertbutyl nitrite and trimethylsilyl azide in acetonitrile. Decamethylcobaltocene, triethylphosphine, trimethylsilyl azide, sodium azide, and VO(O i Pr) 3 were used as received by commercial suppliers. NMR spectra were collected at ambient temperature on a Bruker AV-300, Ascent 400, AV-500, or an Ascent 700 spectrometer. 1 H and 13 C NMR chemical shifts (δ) are reported in ppm and were calibrated to residual solvent peaks. 51 V NMR chemical shifts have been calibrated to VOCl 3 in CDCl 3 as an external standard. It needs to be mentioned Inorganic Chemistry pubs.acs.org/IC Article at this point that 51 V NMR is extremely sensitive and minor impurities (0.5% < ) can still be observed, even though the remaining characterization data appear to be clean. This explains the minor impurities observed in the 51 V NMR spectra of the complexes 1, 4, 5, 7, 8, 9 and 10. Elemental analysis was performed using an Elementar vario microcube instrument. IR spectra were collected using a Bruker Alpha IR spectrometer. Cyclic voltammetry was recorded using a BioLogic potentiostat and a three-electrode array (working electrode: glassy carbon, counter electrode: platinum, reference electrode: silver). EPR spectra were recorded from 5 mM solutions at 300 K using a Bruker Magnettech 5000 EPR spectrometer (microwave frequency, 9.46 GHz; microwave power, 5 mW; modulation amplitude, 0.5 mT). CW spectra were processed using MATLAB and EasySpin software package (see Supporting Information for details). 155 Synthetic Procedures. General Procedure for the Synthesis of 1 and 2. The synthesis of the complexes was adapted from the literature. 54 If not otherwise stated, the corresponding azolium salt (1 equiv, 1 mmol) was mixed with [VO(O i Pr)] (1 or 1.2 equiv) in THF (30 mL) and stirred at room temperature for 2 days. The solvent was evaporated, and the resulting solids were suspended in hexane (20 mL) and stirred for 1 h at room temperature. The resulting suspension was filtered, and the solids were washed with minimal amounts of pentane (10 mL) and dried on the frit inside the glovebox to give the desired vanadium(V) complexes in yields of 78% and higher.
[ . The parent vanadium(V) complexes 1 or 2 (1 equiv.) were dissolved in THF and stirred for 10 min at room temperature. A solution of decamethylcobaltocene (1 equiv.) in THF was added, and the reaction mixtures were stirred for 5 h at room temperature.
[Co(Cp*) 2 ][VOCl(L 1 )] ([Co(Cp*) 2 ][1]). From complex 1 (1 equiv., 0.25 mmol, 148 mg) and Co(Cp*) 2 (1 equiv., 0.25 mmol, 83 mg). After 5 h the reaction was filtered, and the solvent was evaporated. The greenish-gray residue was dissolved in CH 2 Cl 2 and filtered again and concentrated to 1 mL. Hexane was added until the solution became turbid. One drop of dichloromethane was added to redissolve all solids, and the mixture was left to stand in an openly capped vial inside the glovebox for 2 days to let the dichloromethane evaporate. This formed large green blocks of [Co(Cp*) 2  ). From complex 2 (1 equiv., 0.25 mmol, 157 mg) and Co(Cp*) 2 (1 equiv., 0.25 mmol, 83 mg). After 5 h, the green suspension was filtered, and the green solids were washed with another 5 mL of THF and 5 mL of pentane. The green solids were then dissolved in 5 mL of dichloromethane, and the solution was concentrated to 1 mL. Hexane was added until the solution became turbid. One drop of dichloromethane was added to redissolve all solids, and the mixture was left to stand in an openly capped vial for 2 days inside the glovebox to let the dichloromethane evaporate. This formed large green blocks of [Co(Cp*) 2 ] [1]. Yield: 69% (0.172 mmol, 165 mg). Elemental analysis (%) calcd for C 55  General Procedure for the Salt Metathesis Reactions. In a 20 mL scintillation vial, vanadium complex 1 was dissolved in 5 mL Et 2 O and cooled to −40°C. In a separate vial, the corresponding lithium salt (phenolate or amide) was dissolved/suspended in 2 mL Et 2 O and cooled to −40°C as well. This solution was then added dropwise at −40°C to the solution of the vanadium complex and slowly warmed to room temperature overnight while stirring. The deeply colored solutions are filtered to remove any lithium chloride formed during the reaction, and the solvent was evaporated under high vacuum. The colored residues were then dissolved in hexane, filtered again, and concentrated to approximately 0.5 mL. Storing these solutions at −40°C resulted in the formation of the corresponding complexes in moderate to good yields overnight.
General Procedure for Imido Complexes. In a 10 mL J. Young Schlenk flask, vanadium(III) complex 6 (1.0 equiv., 0.10 mmol, 72 mg) was dissolved in 5 mL of benzene. The corresponding azide was added to the solution, and the mixture was heated to 60°C. After 24 h, the solvent was lyophilized. Residues were washed several times with hexane to afford clean product. X-ray quality crystals were grown from concentrated diethyl ether solution at ambient temperature.