A Two-Coordinate Iron(II) Imido Complex with NHC Ligation: Synthesis, Characterization, and Its Diversified Reactivity of Nitrene Transfer and C–H Bond Activation

Abstract

Iron terminal imido species are typically implicated as reaction intermediates in iron-catalyzed transformations. While a large body of work has been devoted to mid- and high-valent iron imidos, to date the chemistry of iron(II) imidos has remained largely unexplored due to the difficulty in accessing them. Herein, we present a study on the two-coordinate iron(II) imido complex [(IPr)Fe(NAr^Trip)] (3; IPr = 1,3-bis(2′,6′-diisopropylphenyl)imidazol-2-ylidene; Ar^Trip = 2,6-bis(2′,4′,6′-triisopropylphenyl)phenyl) prepared from the reaction of an iron(0) complex with the bulky azide Ar^TripN₃. Spectroscopic investigations in combination with DFT calculations established a high-spin S = 2 ground spin state for 3, consistent with its long Fe–N multiple bond of 1.715(2) Å revealed by X-ray diffraction analysis. Complex 3 exhibits unusual activity of nitrene transfer and C–H bond activation in comparison to the reported iron imido complexes. Specifically, the reactions of 3 with CH₂=CHAr^CF3, an electron-deficient alkene, and CO, a strong π acid, readily afford nitrene transfer products, Ar^CF3CH=CHNHAr^Trip and Ar^TripNCO, respectively, yet no similar reaction occurs when 3 is treated with electron-rich alkenes and PMe₃. Moreover, 3 is inert toward the weak C(sp³)–H bonds in 1,4-cyclohexadiene, THF, and toluene, whereas it can cleave the stronger C(sp)–H bond in p-trifluoromethylphenylacetylene to form an iron(II) amido alkynyl complex. Interestingly, intramolecular C(sp³)–H bond functionalization was observed by adding (p-Tol)₂CN₂ to 3. The unique reactivity of 3 is attributed to its low-coordinate nature and the high negative charge population on the imido N atom, which render its iron–imido unit nucleophilic in nature.

Short abstract

The two-coordinate iron(II) imido complex (IPr)Fe(NAr^Trip) (3) exhibits a high-spin ground state (S = 2) and was found to be reactive toward electron-deficient alkene, diazo compounds, terminal alkyne, et al., in which diversified reactivities of nitrene transfer, C−H bond activation, and C−N bond formation have been observed. The reactivity pattern reflects the nucleophilic nature of the imido moiety of the high-spin iron(II) complex.

Introduction

Iron terminal imido species are often implicated as key intermediates in iron-catalyzed nitrogen-group-transfer reactions: e.g., alkene aziridation, sulfur imidation, and C–H bond amination.¹⁻⁶ A range of iron-mediated nitrogen fixation reactions have also proposed with iron imido species as intermediates en route to the final products, NH₃ and N(SiMe₃)₃.⁷⁻¹⁰ The importance of these chemical transformations has attracted great interest in the formation, structural and spectroscopic features, and reactivity of iron imido species. Traditional wisdom had viewed such complexes as highly reactive, nonisolable intermediates, and this knowledge had been largely restricted to theoretical and mass spectrometry studies on the bare iron imide [FeNH]⁺.¹¹ Recent studies showed that, by the use of appropriate ancillary ligands, iron terminal imido species can be effectively stabilized and are amenable to isolation. Since Lee’s report on the first isolable iron terminal imido complex, [Fe₄(μ₃-NBu^t)₄(NBu^t)Cl₃], in 2000,¹² more than 50 structurally well characterized iron terminal imido complexes have been reported.¹³⁻³⁹ In addition, modern spectroscopic methods have enabled probing iron imido species that merely persist at low temperatures,⁴⁰⁻⁴² further enriching our knowledge of this type of metal–ligand multiple-bond species.

So far, the known iron imido complexes span a series of oxidation states ranging from iron(V) to iron(II).^2,11−41 Among them, mid- to high-valent iron species, iron(V),^20,41 iron(IV),^{18,30,31,33,36,40} and iron(III) imidos^{13,17,22,35,37,38} and iminyls^27,35 have been subjected to intensive studies, and diversified reactivity, for example nitrene transfer to CO,^13,22 isocyanides,²² and phosphine,^22,37 alkene aziridation,⁴³ hydrogen atom abstraction (HAA),^{16,24,28,33,42,44} and C–H bond amination,^6,27,35,37 has been observed. In general, the nitrogen groups in high-valent iron imidos are found to be electrophilic in character.^2,4 In contrast, few explorations on iron(II) imido species have been published. To our knowledge, Peters’ complexes with trisphosphine ligation [[Buⁿ₄N][(PhB(CH₂PPh₂)₃)Fe(NAd)]¹⁵ and [(B(o-C₆H₄PPrⁱ₂)₃)Fe(NC₆H₄-p-OMe)]²⁶ (Chart 1) are the only known examples of iron(II) terminal imido complexes. Both iron(II) complexes exhibit a low-spin (S = 0) ground state. However, no reactivity study of them has been reported yet, which leaves a question as to whether iron(II) imido species are capable of performing C–H bond activation and nitrene transfer reactions that are typically found for mid- to high-valent iron imido complexes.

Chart 1. Iron(II) Imido Complexes.

The aforementioned status quo prompted us to explore the chemistry of iron(II) imido species. Herein, we present a study on a two-coordinate iron(II) imido complex with N-heterocyclic carbene (NHC) ligation, [(IPr)Fe(NAr^Trip)] (Chart 1), which was prepared from the reaction of its iron(0) precursor with the bulky organic azide Ar^TripN₃. Spectroscopic characterizations and DFT calculations collectively established a high-spin (S = 2) ground state for this iron(II) imido complex, which is unprecedented for iron(II) imidos. Reactivity studies revealed that the iron(II) imido complex can react with electron-deficient alkenes, e.g. 3,5-bis(trifluoromethyl)phenylethylene, to produce enamines, but it is inert toward electron-rich alkenes. Furthermore, the reactions of [(IPr)Fe(NAr^Trip)] with CO and PMe₃ were found to give distinct outcomes; nitrene transfer products, an isocyanide and iron(0) species, were formed in the former reaction, whereas the ligand exchange product, the low-spin iron(II) imido complex [(Me₃P)₃Fe(NAr^Trip)], was obtained for the latter. [(IPr)Fe(NAr^Trip)] exhibits limited hydrogen atom abstraction activity toward 1,4-cyclohexadiene, THF, and toluene, but it is capable of activating a much stronger C(sp)–H bond to afford iron(II) amido complex when it is treated with a terminal alkyne. Interestingly, intramolecular C(sp³)–H bond activation was observed by adding an diazo compound to the iron(II) imido complex. These observations signify the unique reactivity of the iron(II) imido complex in comparison to those of mid- and high-valent iron imidos and also early-transition-metal imidos.^2,4,45,46 The high negative charge population on the imido nitrogen atom and the low coordination number of the iron center are considered the key factors that endow the iron imido complex with such unique reactivity.

Results and Discussion

Synthesis and Structure of [(IPr)Fe(NAr^Trip)]

Previously, we found that the reactions of the three-coordinate cobalt(0) complexes (NHC)Co(vtms)₂ (vtms = vinyltrimethylsilane) with the bulky organic azide Ar^MesN₃ (Ar^Mes = 2,6-bis(2′,4′,6′-trimethylphenyl)phenyl) could furnish two-coordinate cobalt(II) imido complexes in the form of (NHC)Co(NAr^Mes).^47,48 Aiming to synthesize analogous iron(II) imido complexes, we examined the reaction of the three-coordinate iron(0) complex [(IPr)Fe(vtms)₂] (1; IPr = 1,3-bis(2′,6′-diisopropylphenyl)imidazol-2-ylidene) with Ar^MesN₃; however, this afforded an iron(II) amide complex bearing a cyclometalated IPr ligand, [(IPr′)Fe(NHAr^Mes)] (2) (Scheme 1). Complex 2 was isolated in high yield (75%), and its molecular structure has been confirmed by a single-crystal X-ray diffraction study (Figure S2).

Scheme 1. Reactions of [(IPr)Fe(vtms)₂] with Ar^MesN₃ and Ar^TripN₃.

The attainment of 2 rather than the desired iron(II) imido species (IPr)Fe(NAr^Mes) hints that the iron(II) imido species might be highly reactive to initiate intramolecular C–H bond activation, which points to the necessity of further increasing steric hindrance in order to protect the FeN core. Toward this goal, the reaction of 1 with the sterically more demanding aryl azide Ar^TripN₃ (Ar^Trip = 2,6-bis(2′,4′,6′-triisopropylphenyl)phenyl)⁴⁹ was examined, which to our delight led to the preparation of the targeted iron(II) imido complex. Specifically, treatment of 1 with 1 equiv of Ar^TripN₃ in n-hexane results in the precipitation of a brown solid. After workup and recrystallization, an iron(II) imido complex, [(IPr)Fe(NAr^Trip)] (3), was isolated in 75% yield as a brown crystalline solid (Scheme 1). Complex 3 is paramagnetic, and its measured solution magnetic susceptibility (Evans method, μ_eff = 5.5(1) μ_B in C₆D₆ at room temperature) is close to those of the reported low-coordinate high-spin iron(II) complexes.^50,51 The ¹H NMR spectrum of 3 (measured in C₆D₆) features 13 broad signals in the range +125 to −180 ppm (Figure S21), suggesting C_2v molecular symmetry. Complex 3 is air-, moisture-, and heat-sensitive. Under a dinitrogen atmosphere at room temperature, complex 3 in C₆D₆ solution showed complete decomposition in 2 days, as evidenced by the disappearance of its characteristic ¹H NMR signals. ¹H NMR analysis indicated the formation of new species, but the attempts to isolate them were unsuccessful.

A single-crystal X-ray diffraction study established the molecular structure of 3 as a two-coordinate NHC-iron-imido complex with linear C(carbene)–Fe–N(imido)–C(aryl) alignment (Figure 1). Steric repulsion has rendered a staggered conformation between the core planes of the NHC ligand and the central aryl group of the imido moiety. The dihedral angle (45°) is comparable to that of the nickel complex [(IPr^#)Ni(NAr^Mes)] (41°, IPr^# = 1,3-bis(2′,6′-bis(diphenylmethyl)-4′-methylphenyl)imidazol-2-ylidene)⁵² and apparently different from the right angle (90°) observed for [(IPr)Co(NAr^Mes)].⁴⁷ Complex 3 is the first example of a two-coordinate iron(II) complex with NHC ligation.^53,54 Its Fe–C(carbene) distance (2.042(2) Å) is found to be close to those of the two-coordinate iron(I) complexes [(cyIDep)₂Fe][BAr^F₄] (1.996(7) Å)⁵⁵ and [(IPr)Fe(N(SiMe₃)(Dipp))] (2.014(2) Å).⁵⁶ The Fe–N(imido) bond distance of 1.715(2) Å determined for 3 is longer than those of the low-spin iron(III), iron(IV), and iron(V) imido complexes^2,4,20,33 and comparable to those of the high-spin iron(III) complexes [(^ArDP^Mes)Fe^III(NR)] (R = Mes, 1.708(2) Å)³⁷ and the intermediate-spin iron(III) complexes [(nacnac^terphenyl)Fe^III(NAd)] (1.700(5) Å)²³ and [(^•–PDI^Dipp)Fe^III(NAr)] (Ar = Dipp, 1.705(2) Å; Ar = Mes, 1.717(2) Å).¹⁷ In comparison to the low-spin iron(II) imido complexes [Buⁿ₄N][(PhB(CH₂PPh₂)₃)Fe(NAd)] (1.651(3) Å)¹⁵ and [(B(o-C₆H₄PPrⁱ₂)₃)Fe(NC₆H₄-p-OMe)] (1.67 Å),²⁶ the Fe–N(imido) bond in 3 is considerably lengthened, which hints at a different electronic structure.

Figure 1.

Molecular structure of 3, showing 30% probability ellipsoids and the partial atom-numbering scheme. Hydrogen atoms are omitted for clarity. Selected distances (Å) and angles (deg): Fe1–N1 1.7151(16), Fe1–C1 2.0417(18), N1–C15 1.346(2); N1–Fe1–C1 180.0, C15–N1–Fe1 180.0.

Spectroscopic Characterization and Electronic Structure of [(IPr)Fe(NAr^Trip)]

To shed light on the electronic structure of 3, magnetometry and ⁵⁷Fe Mössbauer measurements coupled with DFT calculations have been undertaken.

Magnetic susceptibility measurements (Figure 2) at variable temperature with a superconducting quantum inference device (SQUID) show that the effective magnetic moment of 3 at room temperature (μ_eff ≈ 6.2 μ_B) is higher than the spin-only value (4.9 μ_B) expected for an S = 2 state, suggesting an unquenched orbital momentum contribution. Interestingly, with decreasing temperatures μ_eff first increases, and then monotonically plummets; consequently, it reaches a maximum around 6.4 μ_B at 80 K. Similar situations were encountered for two-coordinate S = 2 iron(II) alkyl, amido, and aryloxide complexes.⁵⁷⁻⁵⁹ Our exploring simulations revealed that the overshooting behavior of μ_eff found for 3 originates mainly from large g-anisotropy. The variable-temperature and variable-field magnetization studies revealed that all isofield curves are nearly superimposable, which reflects an exceedingly large magnetic anisotropy in 3. The simulation of the combined SQUID data yielded g_|| = 2.08(6), and g_⊥ = 3.05(6), but the D and E/D values only can be narrowed down to D < −10 cm^–1 and 0 < E/D < 0.11, because the computed relevant error surface is rather flat (Figures S45 and S46).

Figure 2.

Temperature dependence of the effective magnetic moment, μ_eff, of a solid sample of 3 recorded in a 1.0 T magnetic field and variable-field and variable-temperature dependence of the magnetization of 3 (inset). The solid lines represent a reasonable fit using the following parameters: D = −42 cm^–1, E/D = 0.07, g_|| = 2.08, and g_⊥ = 3.05.

Zero-field ⁵⁷Fe Mössbauer measurements at various temperatures ranging from 80 to 293 K exhibited a broad and asymmetric quadrupole doublet (Figure S17), probably due to intermediate spin relaxation.⁶⁰ As depicted in Figure 3, magnetic Mössbauer spectra of complex 3 in the solid state, recorded at 1.7 K with 4.0 and 7.0 T applied perpendicular to the γ-rays, display well-resolved magnetic splitting. The relative intensity of the six lines indicates that complex 3 features extremely strong easy-axis magnetization at low temperatures. Assuming that the principal axis system of the electric field gradient tensor is largely collinear with that of D, a satisfactory fit gave δ = 0.56(2) mm/s, ΔE_Q= – 2.08(6) mm/s, and A_∥ = +65.8(6) T. Precise determination of the D and E/D values by the applied Mössbauer measurements is hampered by fast relaxation of 3, due to the existence of the unquenched orbital angular momentum. Therefore, during the simulations, D, E/D, g_||, and g_⊥ were fixed to those determined by the above magnetometry investigations. The simulations do not allow accurate determination of the A_⊥ and η values either, because nearly all ⁵⁷Fe nuclear spins are polarized along the z axis of the ZFS tensor. According to the observed A_∥ value, complex 3 possesses an internal magnetic field of ∼132 T. The measured Mössbauer parameters are comparable to those found for analogous two-coordinate high-pin ferrous complexes.⁶¹⁻⁶⁴

Figure 3.

Variable-field magnetic Mössbauer spectra of 3 in solid state. The solid sample of 3 was immobilized in eicosane (C₂₀H₄₂). Fields were applied perpendicular to the γ-rays. Solid lines represent the best fit using the following parameters: δ = 0.56 mm/s, ΔE_Q= – 2.08 mm/s, η = 0 (fixed), A_∥ = +65.8 T, D = −42 cm^–1 (fixed), E/D = 0.07 (fixed), g_|| = 2.08 (fixed), and g_⊥ = 3.05 (fixed). The spectra show the sample contains 25% diamagnetic impurity. However, the much sharper features of 3 can be readily identified and the existence of the impurity does not affect our further analysis.

Figure 4 illustrates the electronic structure computed by B3LYP calculations. The computations predicted an electron configuration of (d_x²–y²)²(d_xy)¹(d_z²)¹(d_xz/yz)² for the ground state of 3. The linear coordination environment of the iron center in 3 leads to formation of two π bonds between Fe 3d_xz/yz and N 2p_x/y atomic orbitals and one σ bond between Fe 3d_z² and N p_z. As borne out by the shape of the orbital labeled as Fe 3d_z², significant 3d_z²–4s mixing considerably reduces the antibonding interaction with the two ligand and hence stabilizes the σ* orbital relative to the two π* orbitals, a bonding feature found for the low-spin complex [(PhB(o-C₆H₄PPrⁱ₂)₃)Fe(NAd)]⁺ as well.⁴ Furthermore, the remaining Fe 3d_x²–y² and 3d_xy orbitals are essentially nonbonding in nature. The orbital occupation pattern of the ground state therefore defines a formal bond order of 1.5 for the Fe–N(imido) interaction. In line with this reasoning, 3 features a longer Fe–N bond distance relative to those measured for the low-spin Fe(II)-imido complexes.^15,26 The Fe–N π interaction in 3 is somewhat ionic, as suggested by the uneven contribution of the Fe and N atomic orbitals to the two Fe–N π* orbitals. As a consequence, the imido ligand of 3 donates less electron density to the Fe center and its N atom thus accumulates sizable negative charge. This notion is consistent with the computed charge population of N (−0.57), which is much higher than that of Fe (−0.18). Our further theoretical results showed that the computed negative charge population on the N atom of 3 is much higher than those of the related iron(III) imido/iminyl complexes [(nacnac^terphenyl)Fe(NAd)]²³ (−0.45), [(^ArDP^Mes)Fe(NMes)]³⁷ (−0.48), and [(^ArDP^Mes)Fe(NC₆H₄-p-Bu^t)Cl]²⁷ (−0.38) (Table S3), which hints at stronger basicity/nucleophilicity of the imido moiety in 3. On the basis of the metal–ligand interaction discussed above for 3, one can envision it has a rather low lying excited state with an electron configuration of (d_x²–y²)¹(d_xy)²(d_z²)¹(d_xz/yz). A CASSCF(12,13) calculation reveals that this excited state is only 190 cm^–1 above the ground state. Both states are hence nearly degenerate, which explains the unusual magnetic properties observed for 3.⁶¹⁻⁶⁴

Figure 4.

Frontier molecular orbital diagram of [(IPr)Fe(NAr^Trip)] (3) derived from a B3LYP calculation.

Reactivity of [(IPr)Fe(NAr^Trip)]

Complex 3 represents the first example of isolable iron(II) imido species featuring a high-spin (S = 2) ground state. As mentioned earlier, structurally well characterized iron(II) terminal imido complexes are scarce, and their chemistry has remained largely unexplored.^15,26 The status quo thus prompted us to further investigate the reactivity of 3.

Reactions with Alkenes

In iron-catalyzed alkene aziridation reactions, iron terminal imido complexes were proposed as key intermediates.^44,65−68 However, isolable iron imido species capable of reacting with alkenes are exceedingly rare. The only known example is Betley’s iron(III) iminyl complex [(^ArDP^Mes)Fe^III(^•–NAd)Cl], which can react with alkenes to afford aziridines and allyl amines.⁴³ As elaborated below, complex 3 is reactive toward electron-deficient alkenes to give distinct net nitrene transfer products, enamines; more importantly, the transformation displays a disparate reactivity pattern in comparison to that observed for the iron(III) iminyl complexes.

Examining the reactions of 3 with different alkenes revealed the higher activity of 3 toward electron-deficient terminal alkenes over electron-rich alkenes. For instance, treatment of 3 with 3 equiv of 3,5-bis(trifluoromethyl)phenylethylene (Ar^CF3CH=CH₂) at room temperature in Et₂O can lead to the full consumption of 3 in hours. In contrast, 3 is inert toward 4-methoxyphenylethylene, 1-octene, and 4-methylcyclohexene. Our further workup and characterization study revealed that the reaction of 3 with 3,5-bis(trifluoromethyl)phenylethylene gives a mixture of the iron(0) complex [(IPr)Fe(CH₂CHAr^CF3)₂] (4), the enamine trans-Ar^CF3CH=CHNHAr^Trip (5), and the iron(II) enamino complex [(aIPr)Fe(NAr^TripCHCHAr^CF3)(CHMeAr^CF3)] (6; aIPr = 1,3-bis(2′,6′-diisopropylphenyl)imidazol-4-ylidene) (Scheme 2), along with unidentified fluoro-containing species. Complex 4 has been characterized by NMR and elemental analysis. Its ¹H NMR spectrum displays a resonance pattern similar to that of 1. The yield of 4 in this reaction is low (ca. 6% based on ¹⁹F NMR analysis). On the other hand, the ligand replacement reaction of 1 with 2 equiv of Ar^CF3CH=CH₂ proved a more reliable synthetic route for 4. The enamine 5 was isolated in 62% yield when chromatographic separation was performed on the quenched reaction mixture. The high isolated yield of 5 partially comes from protonation of the enamindo complex 6. Compound 5 has been characterized by NMR and mass spectroscopy. Its ¹H NMR spectrum exhibits a diagnostic vicinal, trans coupling constant of 13.8 Hz for its olefinic hydrogens, and DEPT ¹³C NMR spectra also support its identity. ¹⁹F NMR analysis suggested a moderate yield (ca. 50%) for 6. Because of its high solubility in common organic solvent, only a small amount of red crystals of 6 could be isolated. Single-crystal X-ray diffraction established its structure as an iron(II) enamino complex (Figure 5). The enamindo ligand in 6 has Fe–N and C65–C66 distances of 2.020(3) and 1.344(5) Å, respectively, which are typical of Fe–N single bonds and C=C double bonds. The Fe–C(75) and C(75)–C(76) distances of 2.104(3) and 1.532(5) Å, respectively, support the identity of a 1-phenyl-1-ethyl ligand. Probably driven by the relief of steric congestion, the reaction also induced the conversion of IPr into its abnormal form aIPr. Similar normal to abnormal rearrangement has been observed in three-coordinate iron(II) amido NHC complexes.⁶⁹ In addition to 4–6, ¹⁹F NMR analysis indicated that a species which shows ¹⁹F NMR signals at −62.18 and −62.54 ppm were also formed in the reaction of 3 with Ar^CF3CH=CH₂. However, attempts to isolate the species were unsuccessful. In addition to the reaction with 3,5-bis(trifluoromethyl)phenylethylene, complex 3 could also react with ethylene, styrene, and vinyltrimethylsilane, which are more electron rich than 3,5-bis(trimethyl)phenylethylene. The latter reactions, however, proceeded more sluggishly. GC-MS analyses on these reaction mixtures indicated the formation of new organic species having molecular masses anticipated for the corresponding enamines. However, even when the alkenes were used in a large excess (more than 100 equiv), the yields of the desired enamines in 24 h were low (<10%), and further isolation attempts were unsuccessful.

Scheme 2. Reaction of 3 with 3,5-Bis(trifluoromethyl)phenylethylene.

Figure 5.

Molecular structure of [(aIPr)Fe(NAr^TripCHCHAr^CF3)(CHMeAr^CF3)] (6), showing 30% probability ellipsoids and the partial atom-numbering scheme. Hydrogen atoms are omitted for clarity. Selected distances (Å): Fe1–N1 2.020(3), Fe1–C1 2.111(3), Fe1–C75 2.104(3), N1–C65 1.367(4), C65–C66 1.344(5), C75–C76 1.532(5).

The production of enamine in the reaction of 3 with 3,5-bis(trifluoromethyl)phenylethylene is different from the formation of aziridines in the reactions of Betley’s iron(III) iminyl complexes [(^ArDP^Mes)Fe^III(^•–NAd)Cl] with substituted styrenes.⁴³ For the latter, irrespective of electron-donating and -withdrawing groups on the para position of styrenes, all reactions showed rates accelerated over that with unsubstituted styrene. According to this finding, the aziridation reaction was proposed to proceed via the sequential steps of the attack of iminyl radical on styrenes followed by radical rebound. The different reaction outcome reflects the influence of the electronic structure of the nitrogen group, [NAr^Trip]^2– in 3 versus [NAd]^•– in [(^ArDP^Mes)Fe^III(^•–NAd)Cl], on its reactivity toward alkene. On the other hand, the distinct coordination numbers, two-coordinate in 3 over four-coordinate in [(^ArDP^Mes)Fe^III(^•–NAd)Cl], should also be among the contributing factors, as they might induce the reaction to operate in an “inner-sphere” or “outer-sphere” pathway.^70,71

Although the reaction of 3 with 3,5-bis(trifluoromethyl)phenylethylene bears resemblance to the production of imine and enamine, respectively, in the reactions of [(IPr)Co(NAr^Mes)]⁴⁷ and [(IPr^#)Ni(NAr^Mes)]⁵² with ethylene, these divalent late 3d metal imido species seems to show differentiated affinity toward ethylene. Only low conversion was detected in the reaction of the iron(II) imido complex 3 with ethylene, whereas the reactions of cobalt(II) and nickel(II) imido complexes with ethylene took place quickly and gave the products in high conversions. This difference is likely associated with the different character of the nucleophilicity/electrophilicity of these divalent late 3d metal imido species. Our DFT calculation studies on [(IPr)Fe(NAr^Trip)] (3), [(IPr)Co(NAr^Mes)], and [(IPr^#)Ni(NAr^Mes)] revealed a stepwise decrease in the negative charge population on their imido N atom (−0.57, −0.51, and −0.46, respectively; see Table S3), as a consequence of the gradually increased effective nuclear charge of the metal center. Therefore, we hypothesized that the [NAr] group in the iron(II) imido species is more nucleophilic or less electrophilic than those in its cobalt(II) and nickel(II) imido congeners.

On the basis of the above discussion, we proposed the route shown in Scheme 3 to explain the production of enamine. The iron(II) imido complex 3 interacts with the alkene to give a [2π + 2π]-addition product, the four-membered metallacycle A (Scheme 3), as usually found for early-transition-metal imido complexes.⁷² The intermediate A then undergoes β-H elimination to form the iron(II) enamino hydride species B, which eventually initiates reductive elimination to form 4 and 5 (Scheme 3).⁵² As such, 6 is generated by inserting a 3,5-bis(trifluoromethyl)phenylethylene into the Fe–H bond of enamino iron(II) intermediate B, accompanied by the normal to abnormal NHC rearrangement (Scheme 3). Intriguingly, the C=C double bond of the enamino ligand in 6 exhibits a trans configuration like that of 5. This observation hints that the steric repulsion between the 3,5-bis(trifluoromethyl)phenyl group and IPr in the four-membered metallacycle A might govern the selectivity of the β-H elimination step. As the nitrogen atoms of amido ligands in low-coordinate iron(II) complexes usually show trigonal-planar geometry,^33,53,73 steric repulsion between Ar^Trip and IPr in intermediate A could be alleviated.

Scheme 3. Proposed Route for the Reaction of 3 with 3,5-Bis(trifluoromethyl)phenylethylene.

Reactions with CO and PMe₃

Nitrene transfer to CO and phosphine to form isocyanates and iminophosphines has been observed for both iron(III) and iron(IV) imido species.^2,4,22,37 To check whether or not iron(II) imidos can perform nitrene transfer reactions with these substrates, the reactions of 3 with CO and PMe₃ were examined. It turned out that a facile transformation was observed when 3 was treated with CO, but the reaction of 3 with PMe₃ does not yield Me₃PNAr^Trip.

The interaction of 3 with an excess amount of CO (1 atm) proceeded quickly and gave the nitrene transfer reaction products, the isocyanate Ar^TripNCO (7) and iron(0) species [(IPr)Fe(CO)₄] (8), in high yields (Scheme 4). The reaction of 3 with PMe₃, however, did not produce a nitrene transfer product but afforded ligand-replacement products, IPr and the low-spin iron(II) imido complex [(PMe₃)₃Fe(NAr^Trip)] (9). Complex 9 displays a distorted-tetrahedral coordination geometry that is similar to those of [(PMe₃)₃Ru(NDipp)]⁷⁴ and [(PMe₃)₃Co(NAr^Mes)]⁷⁵ (Figure 6). Its short Fe–N distance (1.675(2) Å) and nearly linear Fe–N–C alignment (171.8(3)°) resemble those of [(B(o-C₆H₄PPrⁱ₂)₃)Fe(NC₆H₄-p-OMe)].²⁶ To our surprise, 9 resists further reductive elimination to form Me₃PNAr^Trip even when a mixture of 9 with 20 equiv of PMe₃ in benzene is heated at 60 °C for 36 h.

Scheme 4. Reactions of 3 with CO and PMe₃.

Figure 6.

Molecular structure of [Fe(NAr^Trip)(PMe₃)₃] (9), showing 30% probability ellipsoids and the partial atom-numbering scheme. Hydrogen atoms are omitted for clarity. Selected distances (Å) and angles (deg): Fe1–P1 2.1726(10), Fe1–P2 2.1685(9), Fe1–P3 2.1342(10), Fe1–N1 1.675(2), N1–C1 1.380(3); N1–Fe1–P1 125.97(10), N1–Fe1–P2 122.33(9), N1–Fe1–P3 110.85(11), P2–Fe1–P1 96.64(4), P3–Fe1–P1 96.57(4), P3–Fe1–P2 98.77(4), C1–N1–Fe1 171.8(3).

Nitrene-transfer reactions of metal–ligand multiple bond species with phosphines involves orbital interactions of the frontier molecular orbitals of metal–ligand multiple-bond species with the lone pair and σ*(C–P) orbital of phosphines. This notion is well-demonstrated in the nitrogen atom transfer reaction of iron(V) nitrido with PPh₃ by Smith and Harvey.⁷⁶ Alkyl phosphines as σ-donating ligands possess rather limited π-accepting ability (high lying σ*(C–P) orbital), whereas CO has pronounced σ-donating and also good π-accepting ability. On the basis of this consideration, again we attributed the reaction product not being Me₃PNAr^Trip to the essentially nucleophilic nature of the imido moiety in 3, in striking contrast to mid- and high-valent iron imido/iminyl species. Interestingly, the low-spin iron(II) imido complex 9 was also found to be inert toward 3,5-bis(trifluoromethyl)phenylethylene. This finding is consistent with the notion that the nucleophilicity of the imido moiety in 3 plays an important role in functionalizing C=C bonds, because the calculated negative charge population on the imido nitrogen of 9 (−0.46) is lower than that of its counterpart in 3 (−0.57) (Table S3).

C–H Bond Activation Reactions

Facile hydrogen atom abstraction (HAA) reactions^77,78 have been observed for iron(V),⁴² iron(IV),^{33,39,40,79−81} and iron(III) imido^37,50,82 species. In this regard, we have also tested the activity of 3 toward cleaving C–H bonds in a range of hydrocarbons with different C–H bond strengths and found that 3 is able to activate a C(sp)–H bond but is inert toward the substrates with weaker C(sp³)–H bonds.

The reaction of 3 with a terminal alkyne, p-trifluoromethylphenylacetylene, forms the iron(II) amido alkynyl complex [(IPr)Fe(NHAr^Trip)(CCC₆H₄-p-CF₃)] (10; Scheme 5). The structure of 10 was established by a single-crystal X-ray diffraction study (Figure 7). The presence of a terminal alkynyl ligand, a long Fe–N(amido) distance of 1.922(2) Å, and a bent Fe–N(amido)–C(aryl) angle of 139.24(19)° in 10 clearly indicated the occurrence of C(sp)–H bond activation in this reaction. On the other hand, no reaction occurred when 3 was treated with THF, toluene, or 1,4-cyclohexadiene (CHD) at room temperature despite the presence of weaker C(sp³)–H bonds in these molecules in comparison to the C(sp)–H bond in the alkyne. We note that the observed reactivity pattern for 3 is exclusively distinct from those for HAA reactions of Nam’s iron(V) imido species, wherein the reaction rate of C(sp³)–H bond activation decreases as the bond dissociation energy of the target C–H bond increases.⁴² As such, it is unlikely that the reaction of 3 with the terminal alkyne involves a direct HAA mechanism. As our calculation study on 3 indicated a high negative charge population on the imido N atom, we reasoned that the polarized N–Fe multiple bond might be amenable to invoke a [2π + 2σ]-addition pathway to interact with the H–C(sp) bond and form 10 at the end. Alternatively, a stepwise mechanism that involves the deprotonation of the alkyne by the nucleophilic imide to form iron(II) amido species and alkynyl anion, followed by subsequent binding of the anion to the iron center, could also explain the formation of 9. Notably, such reactivity is typical of early-transition-metal imido complexes^83,84 and is rarely known for late-transition-metal imidos.⁴⁷

Scheme 5. Reaction of 3 with p-CF₃C₆H₄CCH.

Figure 7.

Molecular structure of [(IPr)Fe(NHAr^Trip)(CCC₆H₄-p-CF₃)] (10), showing 30% probability ellipsoids and the partial atom-numbering scheme. For simplicity, all hydrogen atoms are omitted. Selected distances (Å) and angles (deg): Fe1–N1 1.922(2), Fe1–C1 2.154(3), Fe1–C2 2.006(3), N1–C37 1.376(3), C2–C29 1.212(4); N1–Fe1–C1 122.92(10), N1–Fe1–C2 135.67(11), C2–Fe1–C1 99.08(10), C37–N1–Fe1 139.24(19).

In contrast to the reluctance of 3 to perform intermolecular C(sp³)–H bond activation reactions, we found that the interaction of 3 with (p-Tol)₂CN₂ can promote an intramolecular C(sp³)–H bond activation reaction. Adding 1 equiv of (p-Tol)₂CN₂ in an n-hexane solution of 3 at room temperature led to the formation of a brown solution. After workup, the iron(II) amido complex [(IPr)Fe(NHAr^Trip-H)] (11) was isolated as orange crystals in 49% yield (Scheme 6). NMR analysis of the reaction mixture indicated the formation of the azine (p-Tol)₂CNNC(p-Tol)₂ as the byproduct in ca. 45% NMR yield relative to (p-Tol)₂CN₂, which might arise from iron-promoted decomposition of the diazo compound.⁸⁵ The molecular structure of 11 established by XRD clearly shows that the reaction has led to the conversion of the terminal imido ligand [NAr^Trip]^2– of 3 into a bidentate amido-alkyl chelate denoted as [NHAr^Trip-H]^2– in 11 (Figure 8), indicating the involvement of benzylic C–H bond activation on one of the flanking arenes of the imido ligand. The lack of coordination of the diazo ligand (p-Tol)₂CN₂ or its derivatives in 11 suggests that in the course of the intramolecular C(sp³)–H bond activation reaction the diazo compound seems to play the role of “catalyst” in mediating the C–H activation reaction. Accordingly, it can be proposed that the interaction of 3 with (p-Tol)₂CN₂ might result in the three-coordinate iron imido complex [(IPr)Fe(NAr^Trip)(η¹-NNC(Tol-p)₂)] (12), which might perform an intramolecular C–H bond activation reaction to form the cyclometalated complex C. Complex C might be attacked by another molecule of the diazo compound to form cyclometalated complex 11 along with the azine and N₂ (Scheme 6).

Scheme 6. C(sp³)–H Bond Activation Reaction of 3 Trigged by a Diazo Compound and a Proposed Reaction Route.

Figure 8.

Molecular structure of [(IPr)Fe(NHAr^Trip-H)] (11), showing 30% probability ellipsoids and the partial atom-numbering scheme. For simplicity, only one of the two crystallographically independent molecules in the unit cell is shown and all hydrogen atoms are omitted. Selected distances (Å) and angles (deg): Fe1–N1 1.925(3), Fe1–C1 2.112(3), Fe1–C62 2.092(3); N1–Fe1–C1 120.99(12), N1–Fe1–C62 108.55(12), C62–Fe1–C1 130.42(12), C29–N1–Fe1 124.3(2).

Supportive of the aforementioned mechanism, an intermediate, which can further convert into 11, has been detected from the NMR-scale reaction of 3 with (p-Tol)₂CN₂ (Figure S41). A preparative-scale reaction at −30 °C led to the isolation of [(IPr)Fe(NAr^Trip)(η¹-NNC(Tol-p)₂)] (12). While the low isolated yield and instability of 12 at room temperature do not allow full spectroscopic characterizations, a single-crystal X-ray diffraction study has unambiguously established its structure as a three-coordinate NHC iron imido complex bearing η¹-NNC(Tol-p)₂ ligation. As shown in Figure 9, the coordination of an η¹-NNC(Tol-p)₂ ligand has rendered a face-to-face contact of the flanking arene rings of the NHC ligand with respect to those of the imido ligands in 12. The Fe–N(imido) distance in 12 (1.707(2) Å) remains close to that in 3 (1.715(2) Å). However, its bent Fe–N(imido)–C(aryl) alignment (155.79(17)°) is distinct from the linear Fe–N(imido)–C(aryl) alignment in 3. Concurrent with the bent Fe–N(imido)–C(aryl) motif, a benzylic C–H bond on the Ar^Trip moiety is pointing to the imido N atom, as indicated by their short N···C(55) separation of 3.32 Å. The η¹-NNC(Tol-p)₂ ligand in 12 has a short Fe–N(diazo) distance of 1.727(2) Å. Its N–N and N–C distances (1.248(3) and 1.308(3) Å, respectively) and bent N–N–Fe and N–N–C angles (153.8(2) and 135.5(2)°, respectively) are comparable to those of the diazo radical ligands in [Li(THF)₄][(tpe)Fe(NNCPh₂)] (tpe = tris(5-mesitylpyrrolyl)ethane)⁸⁶ and [(((Dipp)NC(Bu^t))₂CH)Co(NNCPh₂)],⁸⁷ implying the radical anion nature of the η¹-NNC(Tol-p)₂ ligand. Thus, in addition to exersion of a steric effect, the coordination of the diazo ligand might also change the electronic nature of the iron imido moiety. These changes could be beneficial for further intramolecular C(sp³)–H bond activation reactions. It should be noted that at this stage whether the C(sp³)–H bond activation reaction from 12 to 11 undergoes a stepwise mechanism of HAA followed by benzylic radical rebound to iron center or a [2π + 2σ]-addition mechanism remained unclear. However, the long Fe···C(benzyl) separations in 12 should make it difficult to accommodate a four-membered transition state for the [2π + 2σ]-addition mechanism.

Figure 9.

Molecular structure of [(IPr)Fe(NAr^Trip)(η¹-N₂C(Tol-p)₂)] (12), showing 30% probability ellipsoids and the partial atom-numbering scheme. For simplicity, all hydrogen atoms are omitted. Selected distances (Å) and angles (deg): Fe1–N1 1.7066(19), Fe1–N2 1.727(2), Fe1–C1 2.061(2), N2–N3 1.248(3), N3–C64 1.308(3); N1–Fe1–C1 141.30(9), N2–Fe1–C1 96.56(9), N1–Fe1–N2 122.07(10), C28–N1–Fe1 155.79(17), N2–N3–C64 135.5(2), N3–N2–Fe1 153.8(2).

Conclusion

In the present work, we reported the preparation of the first two-coordinate iron(II) imido complex, [(IPr)Fe(NAr^Trip)], via the reaction of [(IPr)Fe(vtms)₂] with the bulky organic azide Ar^TripN₃. Single-crystal X-ray diffraction revealed a long distance of 1.715(2) Å for its Fe–N multiple bond. Variable-temperature and variable-field magnetometry and ⁵⁷Fe Mössbauer measurements coupled with DFT calculations collectively established an S = 2 ground spin state for the two-coordinate iron(II) imido complex, different from Peters’ low-spin (S = 0) iron(II) imidos with tris(phosphine) ligation.

The iron(II) imido complex [(IPr)Fe(NAr^Trip)] has been subjected to a detailed reactivity study with the aim of probing the hitherto unknown reactivity of iron(II) imido species. Our study revealed that [(IPr)Fe(NAr^Trip)] performs nitrene transfer reactions with electron-deficient terminal alkenes and CO, producing enamines and isocyanide, respectively. In contrast, the interaction of [(IPr)Fe(NAr^Trip)] with PMe₃ afforded the low-spin iron(II) imido complex [(Me₃P)₃Fe(NAr^Trip)] instead of iminophosphine. [(IPr)Fe(NAr^Trip)] can activate the C(sp)–H bond of a terminal alkyne to afford an alkynyl iron(II) amido complex, whereas it exhibits limited HAA activity toward weaker C–H bonds. When it is treated with a diazo compound, [(IPr)Fe(NAr^Trip)] can convert into an iron(II) complex bearing a cyclometalated arylamido ligand [NHAr^Trip-H], wherein an intramolecular HAA reaction might occur.

The outcomes of the aforementioned C–H bond activation reactions and nitrene-transfer reactions with alkenes are different from those of the mid- to high-valent iron imido/iminyl complexes. The difference can be attributed to the less covalent iron–imido interaction and the resulting higher negative charge population accumulating on the nitrogen atom of the iron(II) imido complex versus mid- and high-valent iron imidos/iminyls as revealed by DFT studies. This bonding feature renders the imido moiety pronounced basicity and nucleophilicity. Equally important is the low-coordinate nature of the metal center, as it might allow the iron center to interact with substrates via a four-membered transition state required for the proposed [2π + 2σ]-addition mechanism, which is hard to occur with iron imido complexes featuring higher coordination numbers. These results hint at the potential utility of iron(II) imido intermediates for the development of new iron-catalyzed organic transformations.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.9b01147.

• Detailed experimental procedures, details of the calculation study, characterization data, molecular structures of 1 and 2, ⁵⁷Fe Mössbauer spectra, NMR spectra, absorption spectra, and electronic configurations of other iron imido species (PDF)

Accession Codes

CCDC 1892821–1892828 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_request@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

The authors declare no competing financial interest.