Synthesis and Reactivity of Iron Complexes with a Biomimetic SCS Pincer Ligand

Abstract

Recent experimental evidence suggests that the FeMoco of nitrogenase undergoes structural rearrangement during N₂ reduction, which may result in the generation of coordinatively unsaturated iron sites with two sulfur donors and a carbon donor. In an effort to synthesize and study small-molecule model complexes with a one-carbon/two-sulfur coordination environment, we have designed two new SCS pincer ligands containing a central NHC donor accompanied by thioether- or thiolate-functionalized aryl groups. Metallation of the thioether ligand with Fe(OTf)₂ gives 6-coordinate complexes in which the SCS ligand binds meridionally. In contrast, metallation of the thiolate ligand with Fe(HMDS)₂ gives a four-coordinate pseudotetrahedral amide complex in which the ligand binds facially, illustrating the potential structural flexibility of these ligands. Reaction of the amide complex with a bulky monothiol gives a four-coordinate complex with a one-carbon/three-sulfur coordination environment that resembles the resting state of nitrogenase, and reaction with phenylhydrazine gives a product with a rare κ¹-bound phenylhydrazido group which undergoes N-N cleavage to give a phenylamido complex.

Graphical Abstract

Synopsis:

Two new pincer ligands with N-heterocyclic carbene anchors and S-containing arms form iron complexes which incorporate nitrogenase-relevant donors.

Introduction

Nitrogenases, which convert N₂ to bioavailable NH₃, play a vital role in the global nitrogen cycle.¹ In Mo-nitrogenase, this reaction occurs at the unusual iron-molybdenum cofactor (FeMoco), which in its resting state contains six sulfide-bridged iron atoms arranged around a central carbide atom.^2-5 In order to bind N₂ the FeMoco must undergo further reduction, which is potentially accompanied by changes in the coordination environment at the iron site(s).^6-8 In particular, crystallographic studies have revealed that in nitrogenases, the central “belt” sulfides can be replaced by exogenous ligands during substrate reduction.^9-12 The ability to lose a sulfide implies that prior to N₂ binding, the FeMoco may contain one or more coordinatively unsaturated iron atoms with two S donors and a C donor (Chart 1, top).

Chart 1.

The unusual coordination environment and reactivity of the Fe atoms in the FeMoco has stimulated the development of inorganic model complexes that capture features in common with the potential intermediates.¹³ While many earlier studies had abiological donor atoms and/or metals,^14-15 there has been significant recent progress towards iron complexes with S- and C-containing ligands that more closely mimic the coordination environment in nitrogenase.^16-21 For example, we reported iron complexes of a m-terphenyl-derived bis(thiolate) ligand (Chart 1) in which the sulfur donors are accompanied by a central arene ring that acts as a carbon donor through the π-system of the central arene.^22-23 Iron complexes with this ligand bind N₂ upon reduction, giving the first example of an Fe-N₂ complex containing only biomimetic S and C donors. However, the π-acceptor character of the central arene competes with π-backbonding to N₂, leading to weak N₂ binding. Furthermore, the SCS ligand dissociates in the presence of excess acid, causing decomposition and preventing catalytic N₂ reduction. We reasoned that incorporation of a C-based σ-donor in place of the arene ring might lead to more stable complexes that are more electron-rich and therefore suitable for substrate binding.

Although they are not as common as pincers containing N, C, and P donors, some SCS pincer ligands have been reported in the literature.^24-25 We sought an SCS pincer with both anionic S donors and a strongly-bound C anchor, which led us to the design of N-heterocyclic carbene (NHC) systems. Iron-NHC complexes are employed in a variety of catalytic reactions.^26-28 They have been used in iron-sulfur clusters, where they have been shown to support low oxidation states²⁹ and stabilize alkyl complexes.^30-32 Chelating bis(NHC) ligands can also support iron-thiolate and iron-selenolate complexes.^33-34 Other iron complexes of multidentate NHC ligands are able to coordinate N₂.^35-43 Peters has described a carbene complex of iron that can reduce N₂ to ammonia.⁴⁴

NHC ligands can be modified by changing the substituents attached to the aryl groups, allowing facile introduction of additional donor atoms.^45-49 Although several NHC ligands containing accessory S-donors have been reported, the majority have only one sulfur donor and their syntheses cannot be easily modified to incorporate a second sulfur donor.^50-51 One notable exception is a set of nickel, platinum, and palladium complexes with a bis(thiolate) NHC ligand reported by Sellmann and co-workers (Chart 1).^52-54 However, in this system the free ligand was not cleanly isolated or fully characterized due to its poor solubility in common organic solvents. We reasoned that the introduction of bulky organic substituents on the aryl groups of the ligand would facilitate isolation and purification of both the ligand and corresponding metal complexes.

Here, we report the synthesis of two novel SCS pincer ligands containing a central NHC and two thioether or thiolate ligands (Chart 1), which may also find broader use in coordination chemistry. We demonstrate that in iron complexes, the ligand binds in a planar fashion with [S^MeCS^Me], whereas it binds facially with [SCS]²⁻, showing the geometric flexibility of this ligand type. In addition, the facial SCS coordination resembles the proposed coordination environment of the iron atoms in FeMoco after dissociation of a sulfide. Both the flexibility and preference for facial coordination are conducive to mimicking the environment of iron in the FeMoco. We also report our unsuccessful attempts to use these complexes as pre-catalysts for N₂ reduction. Finally, we describe tetrahedral iron complexes containing [SCS]²⁻ with nitrogenase-relevant thiolate, phenylhydrazido, and amido groups.

Results and Discussion

Synthesis of bis(thiolate) and bis(thioether) imidazolium salts.

We selected an NHC ligand with tert-butyl substituted aryl thiolate groups. The initial synthetic steps were adapted from procedures reported by Sellmann.⁵⁵ In the original report, 3,5-di-tert-butylaniline (1) was synthesized using a Schmidt reaction, but we wanted to avoid large-scale reactions with sodium azide. Therefore, we instead prepared 1 from commercially available 3,5-di-tert-butylbromobenzene via Buchwald–Hartwig amination with lithium bis(trimethylsilyl)amide followed by aqueous hydrolysis of the resulting silylated aniline.⁵⁶ Following literature procedures,⁵⁵ 1 was converted to 2-aminobenzothiazole 2 and ring-opened to give 2-aminothiophenol 3. Crude 3 was then treated with glyoxal in order to generate a Schiff base, which is known to rearrange to give 4.⁵⁵ Upon deprotonation with strong base, 4 rearranges to the bis(thiolate) Schiff base,⁵⁵ allowing the net functionalization of the thiolate groups with electrophiles. Using this strategy, deprotonation of 4 with KHMDS at −78 °C followed by addition of methyl iodide afforded S-protected Schiff base 5a in 90% yield. Our initial attempts to convert 5a to an imidazolium salt using paraformaldehyde and HCl following the general procedures used for preparation of common unsaturated NHC ligands⁵⁷ afforded only trace amounts of 6a. We hypothesized that acid-induced decomposition of 5a occurs more rapidly than reaction with the heterogeneous paraformaldehyde. Accordingly, sonication of the paraformaldehyde pellets with HCl in dioxane to produce a fine suspension prior to addition to 5a resulted in a significant increase in yield. Compound 6a was isolated in 35% yield by triturating the crude product with pentane, followed by counterion metathesis to the tosylate salt, which precipitated from dichloromethane solution upon addition of diethyl ether.

Since methyl protecting groups are often difficult to remove, we also prepared the analogous benzyl-protected imine 5b in 72% yield by treating 4 with KHMDS and benzyl bromide at −78 °C. Cyclization with paraformaldehyde using a modified version of the procedure used for preparation of 6a gave the imidazolium salt 6b in 36% yield. The benzyl protecting group was then removed by addition of excess (> 3 equiv) potassium naphthalenide at −78 °C in THF followed by quenching with methanol. The potassium bis(thiolate) imidazolium salt 7 was isolated in 64% yield after workup.

We also attempted to isolate the corresponding bis(thiol) imidazolium salt by quenching the deprotection with an ethereal HCl solution instead of with methanol. The ¹H NMR spectrum of an aliquot of this reaction mixture in CD₂Cl₂ indicated complete consumption of the starting material and formation of a symmetric major species with tert-butyl resonances at δ 1.36 and 1.59 ppm that are assigned to the desired bis(thiol) imidazolium salt (Figure S11). Stirring the solution overnight resulted in degradation of this species to a lower-symmetry product with tert-butyl resonances at δ 0.80, 1.20, 1.42, and 1.45 ppm. Two doublets (³J_H-H = 13 Hz) also appear at δ 4.87 and 5.63 ppm, consistent with formation of an olefin-containing product. These results suggest rearrangement to a benzothiazolium-aminothiol species (Scheme 2) with unknown stereochemistry. This reaction occurred slowly even in the solid state, and was also observed when isolated 7 was treated with HCl in THF (Figure S12). Due to this instability, we were unable to cleanly isolate the fully protonated bis(thiol) imidazolium form of 7, though the bis(thiolate) imidazolium form was useful in synthesis, as described below. Analogous reactivity was observed by Despagnet-Ayoub et al. during attempts to synthesize an aniline-NHC-phenol ligand, which suggests that facile rearrangement may be a general property of N-aryl imidazolium salts containing protic functional groups.^58-59

Scheme 2.

Proposed rearrangement of Bis(thiol) Imidazolium Salt Under Acidic Conditions

Synthesis of Bis(Thioether) NHC Iron Complexes.

Treatment of 6a with KHMDS in THF resulted in formation of the free N-heterocyclic carbene S^MeCS^Me, as demonstrated by the disappearance of the imidazolium proton at δ 9.35 ppm in the starting material and appearance of a characteristic ¹³C resonance for the carbene carbon at δ 218 ppm (Figure S10). S^MeCS^Me was separated from the salt byproduct by extraction into diethyl ether, and isolated in 92% yield. This material is stable for months when stored in the solid state at −35 °C under N₂.

Slow addition of S^MeCS^Me to a solution of Fe(OTf)₂ in THF at −78 °C followed by warming to room temperature gave a colorless iron complex (8). Layering a concentrated THF solution of 8 with pentane and storing at −40 °C resulted in precipitation of a white solid. Despite multiple attempts, we were unable to obtain high-quality crystallographic data for 8, presumably due to its complex speciation in THF (see below). In a low-resolution crystal structure (R = 12.5%), the ligand is bound meridionally with a molecule of THF trans to the NHC, and the two trifluoromethanesulfonate ligands are bound in the remaining sites (Figure S52). The Mössbauer spectrum of material isolated in this fashion indicates a single iron-containing species with δ = 1.10 mm/s and ∣ΔE_Q∣ = 3.74 mm/s (Figure 1, top), which is typical for an octahedral high-spin iron(II) complex with low symmetry.

Figure 1.

(Top) Mössbauer spectrum of solid 8 isolated by precipitation from THF with pentane. (Middle) Mössbauer spectrum of 50 mM 8 in frozen THF. (Bottom) Mössbauer spectrum of 9 in the solid state. The data are shown as circles, the lines correspond to fits with the parameters in Table 1, and the grey line is the residual.

The ¹H NMR spectrum of 8 in THF-d₈ contains multiple paramagnetically shifted resonances between δ −1 and 106 ppm (Figure S15). Based on the relative integrations of the NMR signals, we estimate that two species containing the S^MeCS^Me ligand are present in solution in a ratio of ca. 3:1 (see Supplementary Information for a detailed discussion of the assignment of the ¹H NMR spectrum of 8). The ¹⁹F NMR spectrum of 8 contains a peak at δ −7.8 ppm (v_1/2 = 220 Hz) and a much broader peak at δ −33 ppm (v_1/2 = 2100 Hz) also in a ratio of ca. 3:1 (Figure S16), indicating the presence of triflate ions in two different coordination environments, presumably corresponding to the same two species observed in the ¹H NMR spectrum. The Mössbauer spectra of 8 in frozen THF (Figure 1, middle) and of a THF solution of 8 dried under vacuum (Figure S17) also indicate the presence of two species. One species has parameters nearly identical to those of the precipitated 8 described above (δ = 1.07 mm/s and ∣ΔE_Q∣ = 3.74 mm/s) and likely corresponds to the crystallographically determined structure. The second species has a broader spectrum with nearly the same isomer shift (δ = 1.11 mm/s) but a smaller quadrupole splitting (∣ΔE_Q∣ = 2.16 mm/s). The ratio of the two species in the Mössbauer spectra is nearly 1:1. The difference of this ratio from that observed by ¹H NMR spectroscopy may arise from shifts in speciation with concentration and/or freezing. The identity of the other species in the solution spectra of 8 is unknown, but could correspond to a different ligand binding mode (for example, fac vs. mer binding, or dissociation of one or both thioether groups).

We were unable to fully purify 8 due to its poor solubility in common non-coordinating organic solvents and its similar solubility to free Fe(OTf)₂ and protonated ligand, which are consistently present as minor impurities in the isolated material. We therefore explored the behavior of 8 in a more strongly coordinating solvent. Upon dissolving 8 in CH₃CN, an immediate color change from white to intense pink occurred. The ¹H NMR spectrum of the resulting species in CD₃CN indicates formation of a diamagnetic acetonitrile adduct (9) and the ¹⁹F NMR contains a single sharp resonance at δ −80 ppm. The Mössbauer spectrum of 9 (Figure 1) contains a single quadrupole doublet with δ = 0.38 mm/s, which is significantly lower than that of 8 (see above) and is characteristic of a low-spin electronic configuration at iron(II). The CH₃CN binding is reversible; upon dissolving pink crystals of 9 in THF, a colorless solution is obtained and the ¹H NMR spectrum of the resulting material in THF-d₈ is identical to that of 8. Crystals of 9 suitable for XRD were grown by layering a concentrated CH₃CN solution with diethyl ether at −40 °C. In this structure (Figure 2), the iron is six-coordinate and S^MeCS^Me is bound meridionally. The three remaining open coordination sites are occupied by CH3CN molecules and the triflate counterions are outer-sphere.

Figure 2.

(Top): Thermal-ellipsoid plot of the XRD structure of 9 with thermal ellipsoids shown at the 50% probability level. Hydrogen atoms, the outer-sphere triflate counterions, and outer-sphere solvent molecules (CH₃CN and diethyl ether) have been omitted. Selected bond distances (Å): Fe1-S12: 2.2629(8); Fe1-S13: 2.2731(9); Fe1-C11: 1.884(2). (Bottom): View of 9 along the Fe-C bond with CH₃CN omitted, to show the conformation of S^MeCS^Me.

We attempted catalytic N₂ reduction by treating 8 (generated via co-evaporation of 9 with THF) with 6 equiv of KC₈ and 12 equiv of acid ([H(OEt₂)₂][BAr^F₄], BHT, diphenylammonium triflate, lutidine hydrochloride, H₂O, or HCl) at −78 °C in THF. These reactions generated little or no NH₃ and/or N₂H₄ (Scheme S1, Supporting Information). Due to the inability of 8 to catalyze N₂ reduction and the unclear nature of its solution speciation, we turned our attention to iron complexes of the bis(thiolate) NHC SCS²⁻.

Synthesis of Bis(Thiolate) NHC Iron Complexes.

Treatment of 7 with KHMDS in THF resulted in clean formation of the bis(thiolate)-carbene ligand SCS²⁻, as indicated by the loss of the imidazolium signal at δ 9.35 ppm in the ¹H NMR spectrum and appearance of a resonance at δ 214 ppm in the ¹³C NMR spectrum (Figures S13-S14). Addition of this in situ-generated NHC to FeCl₂, FeBr₂, or Fe(OTf)₂ in THF at −80 °C gave bright red solutions. However, upon warming to room temperature, these reaction mixtures turned black and a dark precipitate formed. Although we have not been able to structurally characterize the product(s), we suspect that a mixture of oligomeric species is formed. Dimerization is common for iron complexes with thiolate-containing ligands and, in particular, has been observed for ligands derived from the o-aminothiophenol 3.^60-62 The formation of polymeric species has also been reported for some iron complexes with a benzimidazole OCO ligand related to SCS²⁻.⁶³

In contrast, addition of 7 to Fe(HMDS)₂ in THF followed by addition of 18-crown-6 formed a bright yellow species (10) that is stable in solution at room temperature in the absence of air. Crystals of 10 suitable for X-ray diffraction were obtained from a concentrated diethyl ether solution stored at −40 °C overnight. XRD analysis reveals a complex in which iron is coordinated to SCS²⁻ and a bis(trimethylsilyl)amide ligand with a distorted tetrahedral geometry (τ₄ = 0.79) at iron (Figure 3).⁶⁴ The ligand is bound facially, which results in the iron lying 1.18 Å below the plane defined by the donor atoms of the SCS ligand. This contrasts with 8, 9, and the nickel dimer reported by Sellmann and co-workers⁵², in which the metal ion sits in the SCS plane and the flanking aryl groups twist to accommodate a planar binding (see Figure 2). Structurally analogous bis(phenolate) NHC ligands are also known to bind in both fac and mer geometries in molybdenum complexes, illustrating the conformational flexibility of this type of X(NHC)X ligand scaffold.⁶⁵

Figure 3.

(Top) Thermal-ellipsoid plot of the XRD structure of 10 with thermal ellipsoids shown at the 50% probability level. One of two chemically identical but crystallographically inequivalent molecules in the asymmetric unit is shown. Hydrogen atoms, K(18-crown-6), and outer-sphere solvent molecules (diethyl ether) are omitted. Selected bond distances (A): Fe1-S12: 2.378(2) Fe1-S13: 2.345(2) Fe1-C11: 2.062(6) Fe1-N14: 1.984(5) (Bottom) View of 10 along the Fe-C bond to illustrate the conformation of SCS²⁻. Note that the Fe does not lie in the plane of the imidazolium ring. The bis(trimethylsilyl)amide ligand is omitted for clarity.

The ¹H NMR spectrum of 10 in C₆D₆ contains six broad, paramagnetically shifted resonances between +78 and −13 ppm, suggesting that the crystallographically observed C_s symmetry is retained in solution. The solution magnetic moment of μ_eff = 5.1 ± 0.1 is consistent with a high-spin (S = 2) configuration at iron(II). The solid-state Mössbauer spectrum of 10 (Figure S26) features a quadrupole doublet with δ = 0.67 mm/s and ∣ΔE_Q∣ = 1.82 mm/s which is also consistent with high-spin iron(II) in a tetrahedral geometry. The Mössbauer spectra of 10 in the solid state and in a frozen THF solution are nearly identical (Figure S27), implying that the coordination mode observed in the solid-state structure is retained in coordinating solvents.

In order to create a vacant coordination site through loss of silylamine, 10 was treated with 1 equiv of [H(OEt₂)₂][BAr^F₄] in THF at −78 °C, which resulted in immediate color change from yellow to dark red. The Mössbauer spectrum of a frozen aliquot of this solution indicates the presence of at least three iron-containing species (Figure S47). The same behavior was observed in an analogous experiment with the weaker acid [HNEt₃][BF₄] (Figure S48). Furthermore, the Mössbauer spectra of reaction mixtures generated under N₂ and Ar are identical (Figure S47), indicating that the complexity of the reaction mixture is not caused by partial N₂ binding. Upon warming these reaction mixtures to room temperature, a slow color change to black occurred and Mössbauer spectra indicate the presence of at least two iron species different from those present in aliquots that were frozen before warming. As discussed above, the same behavior was observed during our attempts to metallate the free carbene SCS²⁻ to iron(II) salts, and may arise from oligomerization of transiently-formed three-coordinate iron complexes.

Treatment of 10 with 5 equiv of [H(OEt₂)₂][BAr^F₄] at −80 °C resulted in immediate formation of a light orange solution and precipitation of a white solid. The Mössbauer spectrum of this crude reaction mixture (Figure S49) indicates a complex mixture of products, suggesting decomposition of the complex (perhaps via ligand protonation). Reduction with a mixture of stoichiometric KC₈ and [H(OEt₂)₂][BAr^F₄] also led to the formation of multiple species (Figure S50). Our attempts to use 10 as a pre-catalyst for N₂ reduction in the presence of excess (55 equiv) KC₈ and a variety of acids ([H(OEt₂)₂][BAr^F₄], diphenylammonium triflate, H₂O, and BHT) resulted in release of only one equivalent of NH₃, presumably formed via hydrolysis of the bis(trimethylsilyl)amide ligand in 10 (Scheme S2, Supporting Information).

Generation of SCS-Supported Iron Complexes of Nitrogenase-Relevant Ligands.

By protonating the amide in 10 with weaker acids, we were able to isolate complexes of their conjugate bases. Treatment of 10 with a bulky m-terphenyl monothiol (Scheme 5) gave tris(thiolate) iron(II) complex 11, which has an S₃C coordination environment. The crystallographic structure of this complex (Figure 4) features a distorted (τ₄ = 0.66) tetrahedral geometry in a coordination environment that resembles the individual iron sites in the resting state of nitrogenase. Compared to a similar complex with an m-terphenyl-derived bis(thiolate) ligand (Chart 1)²², 11 exhibits similar Fe-S bond lengths of ~2.3 Å. However, in the m-terphenyl analogue, the shortest Fe-C(arene) distance was 2.484(7) Å indicating little or no Fe-C interaction in the iron(II) state. In contrast, the Fe-C bond length in 11 is 2.035(2) Å which is similar to the 2.0 Å Fe-carbide bond lengths in the resting state of the nitrogenase FeMoco.⁹

Scheme 5.

Figure 4.

Thermal-ellipsoid plot of the XRD structure of 11 with thermal ellipsoids shown at the 50% probability level. Hydrogen atoms and K(18-crown-6) are omitted. Selected bond distances (A): Fe1-S12: 2.3258(9), Fe-S13: 2.3396(9), Fe1-C11: 2.035(2), Fe1-S14: 2.2940(8)

In the crystal structure of 11, the monothiolate ligand is bound asymmetrically with one mesityl ring lying below the iron center. However, spectroscopic studies suggest that the solid-state geometry is not fully preserved in solution. First, in the ¹H NMR spectrum of 11 in THF-d₈ at 25 °C, the peaks for the monothiolate ligand are broadened beyond detection with the exception of a sharp singlet at δ −24 ppm that integrates to 1H. This singlet is assigned to the proton para to the sulfur of the monothiolate, which is the only proton that is not exchanged by rotation of the monothiolate about the S─C bond. Upon cooling to −40 °C, a new set of resonances grows in (Figure S31). Integration of these resonances indicates that at low temperature, the two mesityl groups are inequivalent, which is consistent with slow S─C bond rotation. Furthermore, the Mössbauer spectrum of a frozen THF solution of 11 exhibits significant broadening in comparison to the solid-state spectrum (Figure S34). Taken together, these studies suggest that multiple conformations of the monothiolate ligand are accessible in solution, and that they interconvert on the ms–s timescale in solution at room temperature.

Addition of 1 equiv of phenylhydrazine to 10 in benzene resulted in a slow color change from yellow to red-orange over the course of 2 h at room temperature, and ¹H NMR spectroscopy indicated clean formation of a new iron(II) species (12). The XRD structure of 12 (Figure 5) reveals a four-coordinate complex with a distorted tetrahedral geometry (τ₄ = 0.74). Importantly, the observation that the tetrahedral geometry of 10 and 11 is retained in 12 indicates that the low coordination number of these complexes is not due solely to steric congestion at the metal center. Furthermore, the phenylhydrazido ligand is bound κ¹ rather than the typical η² or bridging binding modes for deprotonated phenylhydrazine. To our knowledge, 12 is the first structurally characterized complex featuring a phenylhydrazido ligand with this binding mode.

Figure 5.

Thermal-ellipsoid plot of the XRD structure of 12 with thermal ellipsoids shown at the 50% probability level. Hydrogen atoms and K(18-crown-6)(THF)₂ are omitted. Selected bond distances (A): Fe-S1: 2.3542(7), Fe-S2: 2.3501(8), Fe-C11: 2.057(2), Fe-N1: 1.963(2), N1-N2:1.470(5)

Only two other iron(II) phenylhydrazido complexes have been reported in the literature. In cis-[Fe(dmpe)₂(η²-NH₂NPh)]⁺, the phenylhydrazido moiety is bound side-on and has an N─N bond length of 1.400(5) Å.⁶⁶ In the second example, the phenylhydrazido ligand binds μ-κN¹:κN² between two sulfide-bridged β-diketiminate-ligated iron centers and has an N─N bond length of 1.417(8) Å.⁶⁷ In 12, the N─N bond is elongated to 1.470(5) Å, which is the same as the bond length for free phenylhydrazine⁶⁸ (1.471 Å) suggesting that the N─N bond may be weakened relative to those of the other iron complexes.

Although it is stable in the solid state, 12 slowly decomposes in THF or benzene solutions over the course of several days at room temperature or several hours at 70 °C. Although we were unable to fully purify 12 due to this instability as well as its high moisture sensitivity, we were able to identify the major product of its degradation. After heating for 2.5 hours at 70 °C, ¹H NMR spectroscopy reveals the presence of a new species formed in 35% spectroscopic yield (Figure 6). The similarity of the ¹H NMR spectrum of this product to those of 11–12, particularly the sharp singlet at δ −55 ppm that integrates to 1H, suggests that the new product is an iron(II) complex with an anionic ligand containing a phenyl group. We therefore reasoned that the new species was likely a phenylamido complex.

Figure 6.

¹H NMR spectra (400 MHz, THF-d₈, 25 °C) of 12 before (top) and after (middle) heating at 70 °C for 2.5 h, leading to partial formation of 13 (35% based on comparison to a nickelocene standard). The spectrum of independently synthesized 13 is shown in the bottom panel.

To confirm this hypothesis, the phenylamido complex 13 was independently synthesized by heating 10 with excess aniline at 65 °C in THF. Crystalline 13 was isolated by layering a concentrated THF solution with diethyl ether at −40 °C, and XRD analysis confirms that it is structurally analogous to 10–12, with a distorted tetrahedral (τ₄ = 0.73) geometry (Figure 7). The ¹H NMR spectrum of this material is identical to that of the major NMR-active species formed from thermolysis of 12 (Figure 6). Analysis of the thermolysis product reaction mixture by Mössbauer spectroscopy confirms the presence of 13 and indicates the presence of multiple additional species which we have been unable to identify (Figure S51).

Figure 7.

Thermal-ellipsoid plot of the XRD structure of 13 with thermal ellipsoids shown at the 50% probability level. Hydrogen atoms except H1 and K(18c6) are omitted. Selected bond distances (A): Fe-S1: 2.342(1), Fe-S2: 2.3649(8), Fe-C11: 2.069(3), Fe-N1: 1.974(3).

It is notable that this system can break N─N bonds. A number of previously studied iron systems are also capable of cleaving the N─N bonds in hydrazines. Some of the literature systems even contain sulfur-containing ligands, such as β-diketiminate-supported sulfide-bridged diiron systems,^{67, 69} as well as thiolate-supported,^70-71 phosphine-supported⁷² and pentamethylcyclopentadienyl-supported¹⁶ diiron complexes. In our current NHC/thiolate-supported system, the yield of the N─N cleaved product was too low for mechanistic studies that would enable this synthetic system to give insight into the enzymatic N─N cleavage process.

Conclusions

Here, we have reported a new pincer ligand with an SCS donor set that mimics nitrogenase intermediates. The isolation of these ligands is complicated by the acid-catalyzed rearrangement to a benzothiazole, which can be overcome with appropriate modifications. Importantly, the synthetic route allows incorporation of both thioethers and thiolates as S donors in a bulky SCS scaffold, which may be useful for a broad range of applications in addition to nitrogenase modeling.

The bis(thiolate) NHC ligand SCS²⁻ is particularly notable as a ligand that gives four-coordinate high-spin anionic iron(II) complexes with X ligands in the fourth position. SCS²⁻ can fold its core resulting in facial binding, which contrasts with the mer binding of the bis(thioether) NHC analog S^MeCS^Me. These examples illustrate the conformational flexibility of this scaffold, which contrasts with the more rigid scaffold in a set of iron complexes with an NHC-containing PCP pincer ligand recently reported by De Ruiter.⁷³ Though these iron complexes were not pre-catalysts for N₂ reduction due to their decomposition in the presence of acid, we were able to generate complexes of thiolate, hydrazido, and amido groups that take advantage of the biomimetic scaffold. Most notable is the unique κ¹-phenylhydrazido complex, which models a potential N₂ reduction intermediate and undergoes N─N cleavage.

Scheme 1.

Synthesis of Imidazolium Salts 6a and 6b

Scheme 3.

Synthesis of Bis(thioether) NHC Iron Complexes

Scheme 4.

Synthesis of 10

Table 1.

Fit parameters for Mössbauer spectra of 8 – 13 (mm/s)

	Solid State			Frozen THF Solution
	δ	∣ΔEQ∣	Γ	δ	∣ΔEQ∣	Γ
8	1.10a	3.74a	0.28a	1.07b	3.63b	0.37b
				1.11c	2.16c	0.56c
9	0.38	1.41	0.30
10	0.67	1.82	0.30	0.68	1.78	0.31
11	0.68	1.76	0.33	0.67	2.07	0.42
12	0.65	1.60	0.32
13	0.67	1.69	0.34

^aParameters for material precipitated from THF with pentane.

^bParameters for major component (53%, see Figure 1).

^cParameters for minor component (47%, see Figure 1).