Organoazides constitute efficient and sustainable precursors for the formation of reactive nitrene functionality, which engenders many useful transformations, including C-H bond amination, aziridination, formation of azoarenes and carbodiimides, and others [1][2][3][4][5][6][7]. Transition metals generally enable milder conditions for the formation of nitrenes and higher selectivity in their subsequent transfer. The majority of nitrene-transfer studies are focused on middle and late transition metals, which generally feature weaker metal-imido bonds and therefore exhibit more reactive nitrene functionalities [8][9][10][11][12][13][14][15][16][17][18][19][20][21]. In contrast, early transition metals (particularly chromium) typically exhibit less reactive nitrene functionalities due to the stability of multiple metal-nitrogen bonds [22][23][24][25][26][27][28]. However, recent studies demonstrated the ability of coordinatively unsaturated chromium(IV)-imido to transfer nitrene to olefins or isocyanides to form aziridines and carbodiimides, respectively [29][30][31]. C-C and C-H bond activation reactions featuring reactive chromium(IV) imido were also reported [32]. As part of our ongoing investigation of nitrene-and carbene-transfer chemistry mediated by 3d complexes in bulky alkoxide ligand environments [33,34], we have reported synthesis and reactions of chromium-imido complexes featuring two bulky monodentate alkoxides [OC t Bu2Ph] [31]. We have demonstrated that Cr IV (OC t Bu2Ph)2(NR') complexes, which were formed for bulky R' groups (R' = Mes, 2,6-Et2C6H3, Ad) enabled catalytic transfer of nitrene group [NR'] to isocyanides to form carbodiimides. In contrast, the reaction of less bulky para-substituted aryl azides N3R' (R' = 4-MeC6H4, 4-MeOC6H4, 4-CF3C6H4) formed Cr(VI) complexes Cr VI (OR)2(NR')2 which were catalytically inactive. Recently, we have reported a new chelating bis(alkoxide) ligand [OO] Ph [35], that created similar bis(alkoxide) ligand environment but warranted higher degree of stability of the resulting iron complex Fe[OO] Ph , and consequently led to a broader range of catalytic nitrene transfer compared with related Fe(OR)2 systems [36,37]. Herein we report our synthetic, structural, and catalytic studies on the chemistry of Cr II [OO] Ph with organoazides and diazoalkanes, and compare this reactivity with the previously reported Cr II (OC t Bu2Ph)2 system.
Treatment of the chromium(II) amide precursor with one equivalent of H2[OO] Ph in THF produces a green solid. Recrystallization of the product from toluene/THF mixture yields mononuclear pale green Cr[OO] Ph (THF)2 (1) related to the mononuclear Cr(OC t Bu2(3,5-Ph2Ph))2(THF)2 [38]. In contrast, recrystallization of the product from a mixture of two polar solvents (CH2Cl2 and THF) forms pale green-blue homodinuclear complex Cr2[OO] Ph 2 (2, Scheme 1) isolated in 79% yield. Homodinuclear complex 2 is structurally related to the previously reported alkoxide-bridged Cr2(OC t Bu2Ph)4 [31], Cr2(OC t Bu3)2(μ 2 -OC t Bu2H)2 [39], and siloxide-bridged Cr2(OSi t Bu3)4 [40]. We note that monodentate alkoxides [OC t Bu2Ph] and [OC t Bu2(3,5-Ph2Ph)] formed dinuclear or mononuclear complexes selectively [31,38], whereas the chelating ligand appears to support both types of geometries in the solid state, depending on the nature of the crystallization medium.
Scheme 1. Synthesis and reactivity of Cr(II) complex in the chelating bis(alkoxide) ligand environment.
The structures of complexes 1 and 2 are given in Figure 1. The structure of 1 is similar to the previously reported structure of Fe[OO] Ph (THF)2 [35]. However, the structure of Fe[OO] Ph (THF)2 exhibited very long distance between the metal and the central phenyl (>3 Å) indicating no interaction [35]. In contrast, the distance between the chromium center and the central phenyl in 1 is ion. Collectively, these computational data suggest a weak covalent interaction between Cr and the bridging phenyl moiety in 1.
We have also investigated the structures of 1 and 2 in solution. Both compounds give rise to paramagnetic NMR spectra, as anticipated. The UV-vis spectra of 1 and 2 (collected in THF) are very similar, exhibiting a peak around 710 nm, and a shoulder around 510 nm (Figures S39 and S40, SI).
The solution magnetic moment of 1 (3.0(3) μB, C6D6) was found to be significantly lower than the expected spin-only value of 4.9 μB. Recalculated for the structure of Cr2([OO] Ph )2, the observed magnetic moment is 2.6(2) μB (per Cr), that is close to the observed value of 2.5(3) μB reported for Cr2(OC t Bu2Ph)2 [31]. These data suggest that 1 undergoes dimerization in solution to form 2.
Following the synthesis of the chromium(II) precursor, its reactivity with organoazides was investigated. Complex 2 was utilized as a precursor in group-transfer reactions. Similar to Cr2(OC t Bu2Ph)4 [31], 2 was found to exhibit profoundly distinct reactivity with non-bulky (4-tolyl, 4-methoxyphenyl, and 3,5-dimethylphenyl) vs. bulky (mesityl, adamantyl) organoazides. Treating a green solution of 2 with two equivalents of para-tolylazide in THF leads to a dark-brown solution, from which diamagnetic 3 is isolated in 69% yield by recrystallization from THF/ether. Similarly, treating 2 with two equivalents of 4-MeOC6H4N3 or 3,5-Me2C6H3N3 forms brown solutions containing diamagnetic products 4 and 5 in 54% and 59%, respectively. We were not able to obtain X-ray quality crystals of 4 and 5, however, their NMR and UV-Vis data (see SI) suggests close similarity to 3. The overall connectivity of 3 was confirmed by X-ray structure determination (Figure 3). The structure reveals distorted tetrahedral geometry, consistent with Cr(VI) bis(imido)
formulation. No chromium-phenyl interaction is observed in this case, with the chromium-phenyl distance exceeding 3.5 Å. As the structure was of relatively low quality, its metric parameters will not be further discussed. In contrast, the reaction of the chromium(II) precursor with two equivalents of bulkier mesityl azide (per Cr) produces a paramagnetic brown solution, containing approximately one equivalent of unreacted azide. Similarly, treating chromium(II) precursor with two equivalents of adamantyl azide also produces paramagnetic product, along with unreacted azide. We were not able to obtain X-ray quality crystals of the products. The formation of trigonal Cr(IV) mono(imido) complexes is proposed, based on similar formation of brown paramagnetic Cr IV (OC t Bu2Ph)2(NMes) and
Cr IV (OC t Bu2Ph)2(NAd) upon reaction of [Cr(OC t Bu2Ph)2] with excess azide. We note that selective formation of Cr(IV) mono(imido) complexes upon reaction with bulkier organoazides was also observed for the chromium(II) system ligated by two monodentate siloxides, [Cr(OSi t Bu3)2] [40].
In addition to the reactions of Cr(II) precursor with organoazides, we have also studied its reactivity with azobenzene, with the goal to produce Cr(VI) bis(imido) via azobenzene cleavage.
Cleavage of the N=N double bond in azoarenes at a single metal center is rare [41][42][43][44]. Heating the mixture of complex 2 with azobenzene to 60 °C in C6D6 for 24 h did not lead to the formation of
Cr VI [OO] Ph (NPh)2 as indicated by 1 H NMR spectroscopy.
Next, the reactivity of the Cr(II) bis(alkoxide) precursor with diazoalkanes was investigated, targeting formation of Cr(IV)/Cr(VI)-alkylidene complexes. Cr-alkylidene complexes are rare [26,45,46], and their reactivity is relatively unexplored. The related cobalt(II) bis(alkoxide) complex Co(OC t Bu2Ph)2(THF)2 formed high-valent cobalt-alkylidene upon reaction with N2CPh2 and N2C(Ph)(CO2Me) [47], whereas the corresponding iron complex Fe(OC t Bu2Ph)2(THF)2 reductively coupled diazoalkanes/diazoesters via terminal nitrogens to form bridging tetrazene species [48].
Different reactivity mode was observed for the chromium bis(alkoxide). Treatment of 2 with diphenyldiazomethane results in the formation of brown Cr[OO] Ph (NNCPh2)2 species (6), which was characterized by X-ray crystallography. The complex appears unstable in both solid state and solution. Our repeated attempts to collect NMR spectrum on the X-ray quality crystals produced spectra with varying but significant number of decomposition products (benzophenone azine, free ligand); peaks attributable to the complex were broad and not informative. Wolczanski and coworkers have reported similar reactivity of Cr2(OSi t Bu3)4 with N2CPh2 to yield Cr(OSi t Bu3)2(N2CPh2)2 complex [40]; their complex also demonstrated limited stability.
The structure of 6 and selected metrics are given in Figure 3 DFT calculations on 6 were performed to analyze the electronic structure of this species. Singlet (6S), triplet (6T), and quintet (6Q) states were optimized and the singlet was found to be lowest in free energy followed by the triplet (+15.1 kcal/mol) then the quintet (+33.9 kcal/mol). Only the optimized structure of 6S is consistent with the experimental x-ray structure (Table S3, SI) and has no unpaired with (NO)(BF4) yields the same product 7. Cyclization of terphenyl bis(carbinol)s via carbocation intermediates to give various substituted dihydroindenofluorenes had been previously reported [49,50]. We surmise that the reaction proceeds via nitrosonium abstraction of hydroxyl (to yield nitrous acid) followed by cyclization as described in Scheme 2.
Given the ability of the Cr II [OO] Ph precursor to undergo stoichiometric reactions with organoazides and isocyanides, we turned next to investigate its catalytic reactivity in formation of carbodiimides. A matrix of 4 representative aryl azides of varying steric bulk (mesityl, 2-methylphenyl, 3,5-dimethylphenyl, and 4-methylphenyl) and 4 representative isocyanides (xylyl, 4-methoxyphenyl, adamantyl, and cyclohexyl) was investigated (Scheme 3). All reactions were carried out in C6D6 at 60 °C for 24 h; the nature of the reaction products and their yields were determined by 1 H NMR spectroscopy. The outcome of the catalytic reaction appears to be a primary function of the organoazide nature. Catalytic formation of carbodiimides is observed for the bulkier mesityl azide. The reaction of mesityl azide with xylyl isocyanide gives nearly quantitative yield; good to moderate yields (72-50%) are observed for other isocyanides. For the ortho-monosubstituted 2-methylphenyl azide, the reaction with xylyl isocyanide produces carbodiimide in a good yield, whereas other isocyanides give low yields of the corresponding products. No carbodiimide formation is observed for 3,5-dimethylphenyl azide and 4-methylphenyl azide. These results are rationalized by the nature of the chromium-imido intermediate, as was described in our previous study [31]. Whereas bulkier azide forms a chromium(IV)-imido, which is capable of isocyanide coordination and subsequent N-C bind formation, tetrahedral chromium(VI) bis(imido) (formed with non-bulky aryl azides) does not coordinate an isocyanide and therefore does not catalyze N-C bond formation.
Scheme 3. Catalytic formation of carbodiimides mediated by 2.
CrCl2 and LiN(Si(CH3)3)2 were purchased from Strem chemicals and Sigma-Aldrich, respectively, and used as received. Previously reported methods were used to synthesize Cr(N(SiMe3)2)2(THF)2 and the ligand precursor H2[OO] Ph [35,51]. Deuterated solvents were purchased from Cambridge Isotope Laboratories and stored over 3 Å molecular sieves.
Non-deuterated solvents were purchased from Sigma-Aldrich chemicals and purified using MBraun solvent purification system. Characterization of compounds was carried out using 1 H and 13 C NMR spectroscopy, high-resolution mass spectrometry and elemental analysis. Selected chemicals were characterized by X-ray crystallography. NMR spectra of carbodiimides and complexes were recorded using Agilent 400 MHz spectrometer and Agilent DD2 600 MHz spectrometer, respectively, in C6D6 at room temperature at Lumigen Instrument center. Chemical shifts and coupling constants (J) are reported in parts per million and Hertz, respectively. Elemental analysis analyses were carried out by Midwest Microlab LLC under air-free conditions. Thermo Fisher Scientific LTQ Orbittrap XL mass spectrometer at the Lumigen Instrument Centre was used for high resolution mass spectra. IR spectra of powdered samples were recorded on Shimadzu IR Affinity-1
FT-IR spectrometer outfitted with a MIRacle10 attenuated total reflectance accessory with a monolithic diamond crystal stage and pressure clamp. Shimadzu UV-1800 spectrometer was used to collect UV-visible spectra. All air-sensitive reactions were carried out in a nitrogen-filled glovebox.
A solution of 4-azidotoluene (0.020 g, 0.15 mmol) in THF was added to a 5 mL solution of Cr2([OO] Ph )2 (0.050 g, 0.038 mmol). Upon addition, the solution color changed from light green to brown; the release of N2 was also observed. The reaction was stirred for 24 h. The solvent was removed under vacuum and the solid residue was recrystallized from ether-THF mixture at -35 °C to afford dark brown crystals in 69% yield (0.046 g, 0.054 mmol). General Procedure for Catalytic Studies. C6D6 solution containing 20 equivalents of an organoazide, 24 equivalents of an isocyanide and 2 equivalents of 1,3,5-trimethoxybenzene (internal standard) was mixed with 2.5 mol% catalyst in C6D6. The reaction was heated at 60 °C for 24 h. The products were identified and the yields were calculated using 1 H NMR spectroscopy.
Structures of complexes 1, 2, 3, and 6, and 6,6,12,12-tetraphenyl-6,12-dihydroindeno-[1,2-b]fluorene (7) were confirmed by X-ray structure determination; data collection and refinement details are given in Table 1. The crystals were mounted on a Bruker APEXII/Kappa three circle goniometer platform diffractometer equipped with an APEX-2 detector. The structures of 1, 3, 6, and One of the para-tolyl groups in the structure of 3 was found to be disordered; the disorder was successfully modeled by finding two alternative conformations. The structure of 1 was also of relatively poor quality, due to the very small size of the crystals. Despite data collection at 60 s per frame, little diffraction was observed beyond ~1 Å. The structure of 1 contained several toluene molecules per asymmetric unit, one of which was found to be disordered over two conformations.
The structure of 2 contained significant amount of highly disordered THF solvent, which was removed using SQUEEZE. a R1 = ∑||Fo -|Fc||/∑|Fo|. b wR2 = (∑ (w(Fo 2 -Fc 2 ) 2 )/∑(w(Fo 2 ) 2 )) 1/2 . c GOF = (∑w(Fo 2 -Fc 2 ) 2 /(n -p)) 1/2 where n is the number of data and p is the number of parameters refined.
DFT calculations were performed using Gaussian 09 [52]. All calculations were performed at the BP86-D3/def2-SVP level of theory [53][54][55][56][57], employing density fitting, ultrafine grids [58], and Becke-Johnson damping for the dispersion corrections. Optimized structures were confirmed to be minima by ensuring all harmonic frequencies were real. All wavefunctions were verified to be stable. Orbital analyses were performed in GaussView 6.0.16 [59], and Mayer bond orders were calculated using Multiwfn 3.5 [60].
We have demonstrated that chelating bis(alkoxide) ligand