Mononuclear, Nonheme, High-Spin {FeNO}^7/8 Complexes Supported by a Sterically Encumbered N₄S-Thioether Ligand

Abstract

The synthesis of a new nonheme iron NO binding complex, [Fe^II(CH₃CN)(N3Py^2PhSEtCN)](BF₄)₂ (1), is reported. Complex 1, which contains two sterically encumbering phenyl substituents, exhibits a hs Fe^II (S = 2) ground state in contrast to the S = 0 ground state for unsubstituted [Fe^II(CH₃CN)(N3PySEtCN)(BF₄)₂. Reaction of 1 with NO_(g) in CH₃CN yields an S = 3/2 {FeNO}⁷ complex 2 which slowly decays at 25 °C with loss of NO∙ to regenerate 1. One-electron reduction of 2 with Cr(C₆H₆)₂ at −40 °C yields the metastable, S = 1 {FeNO}⁸ complex 3. The nitrosyl moieties in thioether-ligated 2 and 3 are significantly less activated than in thiolate-ligated [Fe(NO)(N3PyS)]^+/0, a structurally analogous pair of hs {FeNO}^7/8 complexes. Calculations reveal that reduction of 2 is iron-centered, which may be a general property of hs {FeNO}^7/8 complexes.

Graphical Abstract

Incorporation of sterically encumbered phenyl rings favors a high-spin ground state in a mononuclear, thioether-ligated nonheme iron(II) complex, which in turn activates reactivity toward nitric oxide. The resulting {FeNO}⁷ complex can be reduced by one electron to give a metastable {FeNO}⁸ complex, and these complexes can be compared to their thiolate-ligated analogs, providing information on the influence of sulfur donation at the metal on the spectroscopic properties and extent of NO activation.

Nature utilizes iron-containing enzymes to carry out the binding, reduction, and coupling of nitric oxide (NO∙) to nitrous oxide (N₂O).^1–5 This reaction is central to the regulation of NO_x levels, and is also important environmentally as N₂O is a potent greenhouse gas. Mechanistic proposals for NO∙ reductase involve formation of mono- or dinuclear {FeNO}⁷ adducts that can either mediate N-N coupling directly,^6–7 or must be reduced to the {FeNO}⁸ level prior to N₂O formation.^8–9 In contrast to {FeNO}⁷ species, which have been extensively characterized in heme and nonheme proteins, {FeNO}⁸ and {Fe(HNO)}⁸ species have proven more difficult to study because of their instability. For nonheme iron enzymes, {FeNO}⁷ adducts typically exhibit S = 3/2 ground states,¹⁰ and it is anticipated that their reduced {FeNO}⁸ products would exhibit an S = 1 configuration. Although these {FeNO}⁸ species have been proposed as intermediates in nonheme NO∙ reductases, no such species have been observed in a catalytically relevant setting. Likewise, characterization of {Fe(HNO)}⁸ species has proven elusive, with a myoglobin-HNO complex being the only Fe(HNO) adduct characterized by multiple spectroscopic methods.^11–13 Cryoreduction of the nonheme taurine dioxygenase (TauD) {FeNO}⁷ adduct was employed as a strategy to access putative {FeNO}⁸ (S = 1) and {Fe(HNO)}⁸ (S = 2) species;¹⁴ however, these species were thermally unstable and/or not fully formed, which limited characterization to ⁵⁷Fe Mössbauer spectroscopy and density functional theory (DFT) calculations.

The synthesis and study of well-defined {FeNO}⁸ and {Fe(HNO)}⁸ model complexes could inform on how these species participate in biological reactions, and provide spectroscopic signatures for such species. Reports of {FeNO}⁸ complexes have increased steadily over the past decade,^15–29 but to date, only three mononuclear {FeNO}⁸ (S = 1) complexes have been described.^15–19 A dinuclear S = 1 {FeNO}⁸-{Fe(NO)₂}¹⁰ complex was reported, and calculations indicated that two unpaired spins were localized on the {FeNO}⁸ unit.²¹ Only a few examples of Fe(HNO)}⁸ species have been characterized in heme models,^{26, 30} and nonheme {Fe(HNO)}⁸ species remain entirely unknown.

Given the rarity of {FeNO}⁸/{Fe(HNO)}⁸ complexes, we were motivated to expand this class of compounds with a modular ligand (N3PyX) with tunable electronic and steric properties.^{18, 31–32} The N3Py^2PhSEtCN ligand was prepared to determine how thioether versus thiolate donors impact the properties of {FeNO}⁸. We hypothesized that incorporation of sterically demanding phenyl groups would enforce high spin states at Fe, as seen for other N3PyX systems.³³ The steric groups could be expected to improve the stability of {FeNO}⁸ or {Fe(HNO)}⁸ species by discouraging intermolecular reactivity (e.g., disproportionation). The nitric oxide reactivity of a new hs Fe^II complex, [Fe^II(CH₃CN)(N3Py^2PhSEtCN)](BF₄)₂ (1) is described, giving [Fe(NO)(N3Py^2PhSEtCN)](BF₄)₂ (2) ({FeNO}⁷). One-electron reduction of 2 at −40 °C gives metastable [Fe(NO)(N3Py^2PhSEtCN)]BF₄ (3) ({FeNO}⁸), which is characterized by UV-vis, ⁵⁷Fe Mössbauer, and resonance Raman spectroscopies and DFT calculations.

The synthesis of the N3Py^2PhSEtCN ligand is outlined in Scheme 1. Alkylation of N3Py^2Ph afforded the desired ligand in 55% yield. Metallation with Fe(BF₄)₂∙6H₂O in CH₃CN led to pale yellow, X-ray quality crystals of [Fe^II(CH₃CN)(N3Py^2PhSEtCN)](BF₄)₂ (1) (86%) from CH₃CN/Et₂O. Complex 1 is air-stable both in the solid state and solution (CH₃CN, CH₃OH).

Scheme 1.

The structure of 1, determined by single crystal X-ray diffraction (Figure 1), reveals a distorted octahedral geometry and Fe-N_py distances (2.1633(16)-2.2096(15) Å) that are consistent with a high-spin (hs) Fe^II center.³⁴ The thioether donor is oriented for a weak bonding interaction (d(Fe-S) = 2.6313(5) Å), and a CH₃CN ligand occupies the putative binding site for NO^●. The space-filling model for 1 (Figure S5) indicates that this binding site is congested due to steric contributions from the thioether and phenyl groups. A Mössbauer spectrum of 1-⁵⁷Fe (Figure 1) gives parameters (δ = 1.09, |ΔE_Q| = 2.93 mm s⁻¹) that are also in line with the hs Fe^II assignment. Incorporation of the phenyl substituents leads to a change in ground spin state from low-spin for [Fe^II(CH₃CN)(N3PySEtCN)](BF₄)₂³⁵ to high-spin for 1, likely due to sterically induced elongation of the Fe-N_py distances.

Figure 1.

Left: Displacement ellipsoid plot (50% probability level) of the dication of 1. H atoms have been omitted for clarity. Right: ⁵⁷Fe Mössbauer spectrum (80 K in frozen CH₃CN) of 1-⁵⁷Fe. Experimental data (black dots); best fit (red line).

Pale-yellow solutions of 1 in CH₃CN have two weak absorption bands at 378 (368) and 450 (43) nm (M⁻¹ cm⁻¹) as seen by UV-vis spectroscopy. Stirring of 1 under NO_(g) at 23 °C leads to a rapid color change to orange brown, forming a new species (2) (Scheme 1) with λ_max (ε) = 520 (400), 680 (200) nm (M⁻¹ cm⁻¹) (Figure 2). The EPR spectrum of this reaction gives an essentially axial signal with g = 4.07, 3.91, and 1.99, consistent with an S = 3/2 species (Figure 2, inset). The Mössbauer spectrum contains a broad, asymmetric doublet with δ = 0.67 and |ΔE_Q| = 2.20 mm s⁻¹ (90% of fit), which has an isomer shift typical of {FeNO}⁷ (Figure 3, top).^{7–8, 14, 36} Taken together, the data are consistent with the assignment of 2 as an {FeNO}⁷ (S = 3/2) complex. This assignment was confirmed by resonance Raman (RR) spectroscopy (vide infra). Complex 2 exhibits moderate stability in solution at 23 °C, but decays to give 1 with t_1/2 ≈ 1.5 h. The stability of 2 is greatly improved at −40 °C. Excess NO_(g) can be removed by sparging with argon at −40 °C, and only a 5% decay of 2 is seen after 2 h.

Figure 2.

UV-vis spectral titration of 2 (black line) to 3 (red line) using Cr(C₆H₆)₂ (1.0 equiv) in CH₃CN/toluene at −40 °C. Inset: EPR spectrum (20 K) of 2 in CH₃CN.

Figure 3.

⁵⁷Fe Mössbauer spectra (80 K, 50 mT applied field) for the {FeNO}^7/8 complexes. Experimental data (black dots); best fits (red lines). Top: 2-⁵⁷Fe (green line, 90% of fit). Bottom: 3-⁵⁷Fe (purple line, 81% of fit). See Figure S4 for complete fits.

Reaction of 2 with Cr(C₆H₆)₂, a one-electron reductant (E_1/2 = −1.15 V vs Fc⁺/Fc in CH₂Cl₂),³⁷ leads to formation of a dark red-purple species 3 with λ_max = 495 (2400), 580 (1800) nm (M⁻¹ cm⁻¹) (Figure 2). Spectral titrations indicate that complete formation of 3 requires one equivalent of Cr(C₆H₆)₂. When complex 3 is reacted with ferrocenium hexafluorophosphate, a one-electron oxidant, the intense UV-vis bands for 3 diminish with concomitant regeneration of 2. These results provide strong evidence for the assignment of 3 as an {FeNO}⁸ complex, and indicate that the structural integrity of the iron complex remains intact during reduction/oxidation cycles.

A Mössbauer spectrum for 3-⁵⁷Fe is shown in Figure 3. The majority species gives a sharp doublet with δ = 0.89 and |ΔE_Q| = 1.42 mm s⁻¹. This isomer shift is consistent with the few reported S = 1 {FeNO}⁸ complexes, which have shifts of 0.80 – 0.90 mm s⁻¹.¹⁶ DFT-optimized structures for ⁴[2] (S = 3/2) and ³[3] (S = 1), obtained at the B3LYP/def2-TZVP level of theory, give isomer shifts of 0.66 and 1.00 mm s⁻¹, respectively. The ~0.1 mm s⁻¹ error in δ for ³[3] is within the known errors associated with these calculations.³⁸ Calculations on alternative spin states, ⁵[3] (S = 2) and ¹[3] (S = 0), give structures that are 16.9 and 20.4 kcal mol⁻¹ higher in energy, thus providing strong support for the assignment of a triplet ground state for 3.

To assess the degree of nitrosyl activation in 2 and 3, RR spectroscopy was employed in combination with DFT calculations (Figure 4). The RR spectrum of 2 reveals a prominent ν(NO) mode at 1812 cm⁻¹ that shifts −36 cm⁻¹ (¹⁵N¹⁶O: 1776 cm⁻¹) upon ¹⁵N labeling, in good agreement with the predicted shift for a diatomic oscillator (−32 cm⁻¹). The intensity of the ν(NO) mode relative to solvent bands greatly decreases in the RR spectra of frozen samples at 110 K, but the ¹⁴NO and ¹⁵NO frequencies are unchanged (Figure S2) and correspond to some of the highest recorded N-O stretching frequencies in {FeNO}⁷ species.³⁹ It has been observed that there is a linear correlation between the Fe-N-O angle and ν(NO) for six-coordinate hs {FeNO}⁷ complexes,³⁹ and this correlation suggests that the angle for 2 is >160°. Indeed, the DFT-optimized structure for ⁴[2] gives ¹⁴N-O (¹⁵N-O) stretches at 1805 (1771) cm⁻¹ that are in excellent agreement with the rR data, and exhibits an Fe-N-O angle of 165.4°.

Figure 4.

RR spectra of 2 and 3 in CH₃CN (¹⁴N¹⁶O, black; ¹⁵N¹⁶O, red) and resulting isotope-edited difference spectra (blue); as indicated, different optimum excitation wavelengths and sample temperatures are used for these RR analyses of 2 and 3.

RR spectra of 3 obtained with 458 and 647 nm excitations (Figures 4 and S2, respectively) reveal a prominent ν(NO) band at 1656 cm⁻¹ that shifts down to 1623 cm⁻¹ with ¹⁵NO, consistent with an increase in nitrosyl π* orbital occupancy upon reduction of 2. These N-O stretching frequencies are equally well-matched by the DFT-predicted values for ³[3] (¹⁴NO: 1657 cm⁻¹; ¹⁵NO 1626 cm⁻¹). As in our prior study of the one-electron reduction of [Fe(NO)(N3PyS)]BF₄, the ν(NO) frequency of 3 is ~150 cm⁻¹ lower than in the hs {FeNO}⁷ complex 2. Neither 2 nor 3 provide resonance enhancement mechanisms for ν(Fe-NO) or δ(Fe-N-O) modes in the low frequency region of the RR spectra (data not shown).

Complexes 2 and 3 exhibit ν(NO) modes that are 75 cm⁻¹ and 68 cm⁻¹ higher in energy than those seen for thiolate-ligated [Fe(NO)(N3PyS)]⁺ and [Fe(NO)(N3PyS)]⁰, a pair of structurally analogous hs {FeNO}^7/8 complexes. The redox potential for 2 has not been determined directly due to its instability, but DFT calculations predict an E_1/2 value of ca. −500 mV vs Fc⁺/Fc in acetonitrile, which is ~700 mV less negative than that of [Fe(NO)(N3PyS)]⁺.³¹ Thus, the more electropositive Fe centers in complexes 2/3 can promote greater π-donation from NO, and since these interactions are primarily NO(π*) to Fe(d_xz/yz), they lead to stronger N-O bonds overall. It is tempting to suggest that steric congestion from the Ph substituents could lead to Fe-NO linearization and increased NO(π*)/Fe(d_xz/d_yz) overlap. However, DFT calculations on structures minus the Ph substituents exhibited only minor changes in the Fe-N-O angles, ν(NO) frequencies, or redox potentials, supporting the conclusion that the differences between 2/3 and [Fe(NO)(N3PyS)]^+/0 are due to changes in the sulfur ligation.

An unrestricted corresponding orbital (UCO) analysis of ⁴[2] and ³[3] was performed. The UCO diagram for ⁴[2] (Figure S9) contains five unpaired electrons with majority Fe(d) character in the α manifold and two unpaired electrons with majority NO(π*) character in the β manifold. The NO(π*) and Fe(d_xz/d_yz) orbitals exhibit significant overlap, indicative of magnetic coupling. This configuration is reminiscent of other hs {FeNO}⁷ complexes, which are described as hs Fe^III antiferromagnetically coupled to triplet nitroxyl.¹⁰ Reduction of ⁴[2] to ³[3] places an additional β electron in an Fe(d_xy) orbital that is nonbonding with respect to the nitrosyl ligand (orbital 175β, Figure S10); thus the reduction is iron-centered, as seen previously for [Fe(NO)(N3PyS)]⁺.¹⁹ In each case, the modest ~150 cm⁻¹ downshift in ν(NO) is explained by weakening the NO(π*)-to-Fe donation, rather than direct electron transfer into an NO(π*) orbital.

In conclusion, we have prepared two new sulfur-ligated {FeNO}⁷ and {FeNO}⁸ complexes. Installation of phenyl substituents in the second coordination sphere in N3PySEtCN generates a hs Fe^II ground spin state. This change in spin state from S = 0 (for the unsubstituted N3PySEtCN analogue) to S = 2 makes the Fe^II center active toward binding of NO^●. This ligand design enabled the first examination of thioether versus thiolate ligation in structurally analogous, hs {FeNO}^7/8 complexes. It is found that the nitrosyl ligands in 2 and 3 are appreciably less activated than in the thiolate-ligated analogues. Computations on ⁴[2] and ³[3] nicely corroborate the experimental findings, and show that reduction occurs at Fe (not NO), providing support for the hypothesis that reduction of hs {FeNO}⁷ species is invariably metal-centered. These findings show that strong, anionic donors, in synthetic or enzymatic ligand environments, should help facilitate NO activation.