A Mononuclear Nonheme Iron(III)-Peroxo Complex Binding Redox-Inactive Metal Ions

Abstract

Redox-inactive metal ions that function as Lewis acids play pivotal roles in modulating reactivities of oxygen-containing metal complexes in a variety of biological and biomimetic reactions, including dioxygen activation/formation and functionalization of organic substrates. Mononuclear nonheme iron(III)-peroxo species are invoked as active oxygen intermediates in the catalytic cycles of dioxygen activation by nonheme iron enzymes and their biomimetic compounds. Here, we report mononuclear nonheme iron(III)-peroxo complexes binding redox-inactive metal ions, [(TMC)Fe^III(O₂)]⁺-M³⁺ (M³⁺ = Sc³⁺ and Y³⁺; TMC = 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane), which are characterized spectroscopically as a ‘side-on’ iron(III)-peroxo complex binding a redox-inactive metal ion, (TMC)Fe^III-(μ,η²:η²-O₂)-M³⁺ (2-M). While an iron(III)-peroxo complex, [(TMC)Fe^III(O₂)]⁺, does not react with electron donors (e.g., ferrocene), one-electron reduction of the iron(III)-peroxo complexes binding redox-inactive metal ions occurs readily upon addition of electron donors, resulting in the generation of an iron(IV)-oxo complex, [(TMC)Fe^IV(O)]²⁺ (4), via heterolytic O-O bond cleavage of the peroxide ligand. The rates of the conversion of 2-M to 4 are found to depend on the Lewis acidity of the redox-inactive metal ions and the oxidation potential of the electron donors. We have also determined the fundamental electron-transfer properties of 2-M, such as the reduction potential and the reorganization energy in electron-transfer reaction. Based on the results presented herein, we have proposed a mechanism for the reactions of 2-M and electron donors; the reduction of 2-M to the reduced species, (TMC)Fe^II-(O₂)-M³⁺ (2’-M), is the rate-determining step, followed by heterolytic O-O bond cleavage of the reduced species to form 4. The present results provide a biomimetic example demonstrating that redox-inactive metal ions bound to an iron(III)-peroxo intermediate play a significant role in activating the peroxide O-O bond to form a high-valent iron(IV)-oxo species.

Introduction

Metal-dioxygen adducts, such as metal-peroxo (M-O₂) species, are invoked as active oxygen intermediates in enzymatic reactions incorporating oxygen atoms into newly biosynthesized molecules as well as in the deleterious reactions of biological oxidative stress and enzymatic detoxification reactions of reactive oxygen species. Mononuclear nonheme iron(III)-peroxo species are often detected in the activation of dioxygen by nonheme iron enzymes,^1-7 and the structural and chemical properties of the intermediates have been intensively investigated in enzymatic and biomimetic reactions.⁸ Very recently, a mononuclear nonheme iron(III)-peroxo complex bearing a macrocyclic N-tetramethylated cyclam ligand, [(TMC)Fe^III(O₂)]⁺ (TMC = 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane), was synthesized and characterized with various spectroscopic techniques and X-ray crystallography.^9-11 It was also shown that the iron(III)-peroxo complex is converted to an iron(III)-hydroperoxo complex, [(TMC)Fe^III(OOH)]²⁺, upon protonation, and then the latter intermediate cleanly converts to an iron(IV)-oxo complex, [(TMC)Fe^IV(O)]²⁺, via O-O bond cleavage of the hydroperoxo ligand.^10,11 Thus, success of the generation and isolation of the thermally stable [(TMC)Fe^III(O₂)]⁺ complex provided us with an opportunity to use this intermediate in the investigation of chemical and physical properties of the biologically important iron(III)-O₂ species.

Redox-inactive metal ions that function as Lewis acids are known to play pivotal roles in a variety of oxidation reactions by oxygen-containing metal complexes, including dioxygen activation/formation and functionalization of organic substrates. One notable example is a redox-inactive Ca²⁺ ion in the oxidation of water to dioxygen in the oxygen-evolving complex (OEC, a heteronuclear Mn₄CaO₅ cubane complex) of Photosystem II (PS II),^12-14 although the role of the Ca²⁺ ion in making O-O bond has yet to be clarified.¹⁵ In biomimetic studies, the redox-inactive metal ions have shown remarkable effects on the reactivities of metal-oxo complexes in electron transfer, oxygen atom transfer, and C-H bond activation reactions.^16-26 It has been demonstrated that the binding of redox-inactive metal ions to the metal-oxo complexes changes their redox potentials, thereby increasing their oxidizing power that enhances reactivities of the metal-oxo species in oxidation reactions. The activation of dioxygen by metal complexes is also facilitated in the presence of redox-inactive metal ions (e.g., Ca²⁺).²⁷ Thus, the role(s) of redox-inactive metal ions is being unveiled recently in various oxidation/reduction reactions by oxygen-containing metal complexes.

Redox-inactive metal ions also bind to metal-peroxo complexes (e.g., Ni-O₂ and Cu-O₂), and the peroxide O-O bond of the metal-peroxo complexes is cleaved by binding a second metal ion to give heterobimetallic M(μ-O)₂M’ complexes.^28-31 In one case, a crystal structure of a nickel-peroxo complex binding potassium ion, [LNi(μ,η²:η²-O₂)K(solvent)], was successfully obtained.³² However, reactions of nonheme iron(III)-peroxo complexes with redox-inactive metal ions have never been investigated, although crystal structures of iron porphyrin-O₂-copper complexes and mechanisms of the O-O bond cleavage of the intermediates have been intensively investigated as a chemical model of cytochrome c oxidase (CcO),^33-35 an enzyme that catalyzes the reduction of dioxygen to water via the formation of a presumed iron(III) porphyrin-O₂-copper(II) intermediate.^36,37 However, the final step of the catalytic cycle of CcO, which is the O-O bond activation of the iron porphyrin-O₂-copper intermediate to form iron(IV)-oxo and Cu(II) species, remains elusive.

As our on-going efforts to understand the effects of redox-inactive metal ions on the chemical and physical properties of oxygen-containing metal complexes in enzymatic and biomimetic reactions, we have synthesized and characterized mononuclear nonheme iron(III)-peroxo complexes binding redox-inactive metal ions, [(TMC)Fe^III(O₂)]⁺-M³⁺ (M³⁺ = Sc³⁺ and Y³⁺). We have also shown that these [(TMC)Fe^III(O₂)]⁺-M³⁺ complexes are reduced by one-electron donors (e.g., ferrocene) and converted to the corresponding iron(IV)-oxo complex via O-O bond cleavage. The rates of O-O bond cleavage depend significantly on the Lewis acidity of the redox-inactive metal ions and the reduction potential of the electron donors. The overall mechanism of the O-O bond activation of the iron(III)-peroxo complexes by binding redox-inactive metal ions and specifically the role of the redox-inactive metal ions in facilitating the O-O cleavage are discussed as well.

Results and discussion

The iron(III)-peroxo complex, [(TMC)Fe^III(O₂)]⁺ (1), which was prepared by reacting [Fe^II(TMC)(CF₃SO₃)₂] with 5 equiv. H₂O₂ in the presence of 2 equiv. triethylamine in CF₃CH₂OH or CH₃CN at −40 °C,^9-11 was isolated as a solid and used for further reactions. Addition of 3 equiv. of scandium triflate, Sc(CF₃SO₃)₃, to the solution of 1 in acetone/CF₃CH₂OH (v/v = 3:1) at −40 °C immediately generated a purple intermediate (2-Sc) with an electronic absorption band at λ_max = 530 nm (Fig. 1a). Similarly, the reaction of 1 with 5 equiv. of yttrium triflate, Y(CF₃SO₃)₃, in acetone/CF₃CH₂OH (v/v = 3:1) at −40 °C afforded the generation of a purple intermediate (2-Y) with an electronic absorption band at λ_max= 560 nm (Fig. 1a). When the reactions were carried out in CH₃CN/CF₃CH₂OH (v/v = 1:1), addition of one equiv. of Sc³⁺ and Y³⁺ afforded the full conversion of 1 to 2-Sc and 2-Y, respectively (ESI, Fig. S1). The intermediates 2-Sc and 2-Y were highly stable and persisted for several days at −40 °C, which allowed us to characterize them with various spectroscopic techniques, such as electron paramagnetic resonance(EPR) spectroscopy, electrospray ionization mass spectrometry (ESI MS), resonance Raman (rRaman) spectroscopy, and X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS).

Fig. 1.

(a) UV-vis spectral changes observed in the conversion of 1 (black line) to 2-Sc (blue line) and 2-Y (red line) upon addition of Sc³⁺ (3.0 mM) and Y³⁺ (3.0 mM) to the solution of 1 (1.0 mM), respectively, in acetone/CF₃CH₂OH (3:1) at −40 °C. UV-vis spectrum of [(TMC)Fe^III(OOH)]²⁺ (3, green line) is shown for comparison. (b) Resonance Raman spectra of 2-Sc-¹⁶O (8.0 mM; black line),2-Sc-¹⁸O (8.0 mM; red line), and 2-Y-¹⁶O (8.0 mM; blue line)obtained upon excitation at 515 nm in acetone-d₆ at −20 °C. 2-Sc-¹⁶O and 2-Sc-¹⁸O were generated upon addition of Sc³⁺ ion to the solution of isolated [(TMC)Fe^III(¹⁶O₂)]⁺ and [(TMC)Fe^III(¹⁸O₂)]⁺, respectively. 2-Y-¹⁶O was generated upon addition of Y³⁺ ion to the solution of an isolated [(TMC)Fe^III(¹⁶O₂)]⁺ complex. The peak marked with “asterisk” is ascribed to acetone-d₆ solvent.

The EPR spectra of the intermediates, 2-Sc and 2-Y, exhibit signals at g = 9.2, 5.0, and 3.8 for 2-Sc and g = 8.8, 5.2, and 3.4 for 2-Y (ESI, Fig. S2), which are indicative of high-spin (S = 5/2) Fe(III) species.³⁸ The ESI MS of 2-Sc exhibits a prominent ion peak at a mass-to-charge (m/z) ratio of 835.9966 (ESI, Fig. S3), whose mass and isotope distribution pattern correspond to [(TMC)Fe(O₂)Sc(CF₃SO₃)₃]⁺ (calculated m/z of 835.9989). The ESI MS of 2-Y exhibits a prominent ion peak at m/z of 879.9496 (ESI, Fig. S4), whose mass and isotope distribution pattern correspond to [(TMC)Fe(O₂)Y(CF₃SO₃)₃]⁺ (calculated m/z of 879.9489).

The rRaman spectrum of the ¹⁶O-labeled 2-Sc in acetone-d₆, obtained by 515-nm excitation at −20 °C, shows two peaks with isotope shifts upon ¹⁸O-substitution (Fig. 1b). The peak at 543 cm⁻¹ shifts to 522 cm⁻¹ with a ^16,18Δ value of 21 cm⁻¹ (^16,18Δ (calcd) = 21 cm⁻¹) and is assigned as the Fe–O stretching vibration.^39,40 The other isotopically sensitive peak at 807 cm⁻¹ shifts to 761 cm⁻¹ with a ^16,18Δ value of 46 cm⁻¹ (^16,18Δ (calcd) = 47 cm⁻¹) and is assigned as the O–O stretching vibration.^39,40 The Fe–O and O–O stretching vibrations for 2-Y were observed at 540 and 799 cm⁻¹, respectively (Fig.1b). Thus, the O-O stretching frequencies of 2-Sc and 2-Y indicate that the O₂ unit in 2-Sc and 2-Y retains the peroxo character, O₂²⁻. Further, the O–O stretching frequencies of 2-Sc (807 cm⁻¹) and 2-Y (799 cm⁻¹) are smaller than that of 1 (825 cm⁻¹),^10,11 whereas the Fe–O stretching frequencies of 2-Sc (543 cm⁻¹) and 2-Y (540 cm⁻¹) are higher than that of 1 (487 cm⁻¹)^10,11 (Fig.1b). Furthermore, these values are significantly smaller than the Fe–O (658 cm⁻¹) and O–O (868 cm⁻¹) stretches in a high-spin end-on Fe(III)-hydroperoxo complex bearing the TMC ligand, [(TMC)Fe^III(OOH)]²⁺ (3).^10,11

A comparison of the k³ weighted Fe K-edge EXAFS for 1 and its Sc(CF₃SO₃)₃ and Y(CF₃SO₃)₃ adducts, 2-Sc and 2-Y, respectively, along with their non-phase shift corrected Fourier transforms (k = 1 – 15 Å⁻¹) is shown in Fig. 2a. The data show that the first shell of 2-Sc and 2-Y are diminished in intensity relative to 1, indicating a perturbation in the first shell distances. Additionally, both 2-Sc and 2-Y have significant intensity between R’ 3 – 4 Å, which is absent in 1, indicating heavy atom coordination. FEFF⁴¹ fits to the data for 2-Sc and 2-Y are presented in Fig. 2b and 2c and ESI, Table S1. The best fit to the data of 2-Sc and 2-Y indicates a six-coordinate first shell in both complexes with two short Fe-O distances at 1.98 and 1.97 Å, respectively, and four longer Fe-N distances at 2.17 and 2.19 Å, respectively. The second shell is fit with single and multiple scattering components from the TMC ligand. The data for 2-Sc are best fit when a Fe-Sc and its corresponding multiple scattering contribution, Fe-O-Sc, are included. The Fe-Sc distance optimizes to 3.8 Å. In 2-Y, inclusion of Fe-Y and Fe-O-Y components are necessary to obtain a good fit and the Fe-Y distance optimizes to be 4.0 Å. Thus, the EXAFS analyses show that both 2-Sc and 2-Y are side-on bound Fe(III)-O₂ species complexed with Sc³⁺ and Y³⁺, respectively. A comparison with the EXAFS data of 1 shows that the first shell distances of 2-Sc and 2-Y are perturbed by metal ion binding, with the Fe-O bonds longer (by 0.06 and 0.05 Å, respectively) and the Fe-N bond distances slightly shorter (by 0.04 and 0.02 Å, respectively).

Fig. 2.

(a) A comparison of the non phase-shift corrected Fourier transforms and the corresponding EXAFS data (inset) for 1 (black line), 2-Sc (green line), and 2-Y (red line). (b) FEFF best-fit (black line) to 2-Sc (green line). (c) FEFF best-fit (black line) to 2-Y (red line).

To further shed light on the structural basis of Sc³⁺ and Y³⁺ bound to 1 and understand the structural differences, spin unrestricted density functional theory calculations were performed on the [(TMC)Fe(O₂)]⁺, [(TMC)Fe(O₂)-Sc(CF₃SO₃)₃]⁺, and [(TMC)Fe(O₂)-Y(CF₃SO₃)₃]⁺. All three converged to side-on bound Fe(III)-O₂ systems (see Fig. 3 for metrical parameters and ESI, Table S2 for spin populations and Mulliken⁴² charges). As can be seen from the Mayer⁴³ bond orders, upon binding to the innocent metal ion (e.g., Sc³⁺/Y³⁺), both the O-O and Fe-O bonds in 1 become weaker. This is because the Fe-O₂ orbital overlap is significantly diminished, as seen in the decrease in O₂ Mulliken populations in the d_xy and d_xz orbitals. This decrease restores electron density into the O₂ π* orbitals and weakens the O-O bond, as also seen in the increase in the calculated O-O distance from 1.42 in 1 to 1.49 Å in both 2-Sc and 2-Y and the decrease in the O–O stretching frequency from 825 cm⁻¹ in 1 to 807 cm⁻¹ in 2-Sc and 799 cm⁻¹ in 2-Y in rRaman experiments (Fig. 1b). However, it is of interest to note that the experimental Fe-O stretching frequencies for 2-Sc and 2-Y are higher than that of 1 (vide supra), even though EXAFS analysis shows that the M-O bonds become longer upon binding of Sc³⁺/Y³⁺. To understand this, numerical frequency calculations were performed on 1, 2-Sc, and 2-Y and the values of 430 cm⁻¹, 494 cm⁻¹, and 503 cm⁻¹, respectively, reproduce the experimental trend. In 2-Sc and 2-Y, the Fe-O mode is kinematically coupled with the Sc/Y-O mode, which results in the observed higher frequency, even though the Fe-O bond is weaker.

Fig. 3.

DFT-optimized structures of 1 (a), 2-Sc (b), and 2-Y (c). The distances obtained from EXAFS (red) are shown for comparison to the DFT distances (black). The average Fe-N distances from DFT are compared to experimental data. Atom colours: Fe, orange; N, blue; O, red; S, yellow; C, gray; F, green; Sc, purple; Y, cyan.

In this section, we have shown the synthesis of mononuclear nonheme iron(III)-peroxo complexes binding redox-inactive metal ions. Based on the spectroscopic characterization of the iron(III)-peroxo complexes binding Sc³⁺ and Y³⁺ ions presented above, we were able to conclude unambiguously that 2-Sc and 2-Y are mononuclear iron(III)-peroxo complexes binding a peroxo ligand in a side-on η² fashion between two metal ions, [(TMC)Fe^III-(μ,η²:η²-O₂)-M^III(CF₃SO₃)₃]⁺ (see Fig. 3).

As mentioned above, the [(TMC)Fe^III(O₂)]⁺–M³⁺ (2-M) species, 2-Sc and 2-Y, are stable for several days in CH₃CN/CF₃CH₂OH (1:1) at −40 °C. Interestingly, addition of one equiv. of ferrocene (Fc) to the solution of 2-Sc and 2-Y resulted in the immediate disappearance of the absorption peaks corresponding to 2-Sc at 535 nm and 2-Y at 570 nm with the concomitant appearance of the absorption peaks corresponding to [(TMC)Fe^IV(O)]²⁺ (4)⁴⁴ at 820 nm and ferrocenium cation (Fc⁺)²⁰ at 620 nm (Fig. 4a for the reaction of 2-Sc and ESI, Fig.S5 for the reaction of 2-Y). Well-defined isosbestic points were observed at 713 and 759 nm in the titration reactions of 2-Sc and 2-Y, respectively (Fig. 4a and ESI, Fig. S5). These results demonstrate unambiguously that one-electron transfer from Fc to 2-Sc and 2-Y resulted in the formation of 4 via the O-O bond cleavage of the peroxide ligand (vide infra). It should be noted that the iron(III)-peroxo complex, [(TMC)Fe^III(O₂)]⁺ (1), did not react with Fc, indicating that the electron transfer from Fc to 2-M was facilitated by the binding of redox-inactive metal ions to 1. It should also be noted that the iron(IV)-oxo complex, [(TMC)Fe^IV(O)]²⁺ (4), which was produced in the reaction of 2-M and Fc, was not reduced by Fc under the reaction conditions,⁴⁵ suggesting that the reactivity of 2-M is much greater than that of 4 with Fc in the electron-transfer (ET) reaction (vide infra).

Fig. 4.

(a) UV-vis spectral changes showing the disappearance of 2-Sc at 535 nm and the formation of [(TMC)Fe^IV(O)]²⁺ (4) at 820 nm and Fc⁺ ion at 620 nm by addition of Fc (0.0 – 1.0 mM) to a solution of 2-Sc (0.50 mM) in increments of 0.2 equiv. in CH₃CN/CF₃CH₂OH (1:1) at −40 °C. Inset shows the spectroscopic titration at 820 nm for the formation of 4 as a function of the equiv. of Fc added to the solution of 2-Sc (0.50 mM) in increments of 0.2 equiv. (b) Spectral changes observed in the electron transfer from Fc (1.0 mM) to 2-Sc (0.10 mM) in CH₃CN/CF₃CH₂OH (1:1) at −40 °C. Inset shows time course monitored at 535 nm for the decay of 2-Sc.

Then, the rates of the electron transfer from Fc to 2-Sc and 2-Y were determined under the pseudo-first-order reaction conditions (e.g., with >10 equiv. of Fc to 2-M). First-order rate constants, determined by the pseudo-first-order fitting of the kinetic data for the decay of 2-Sc and 2-Y (Fig. 4b and ESI, Fig. S6), increased linearly with the increase of the Fc concentration, and the second-order rate constants (k_et), such as 1.6 × 10⁴ M⁻¹ s⁻¹ for 2-Sc and 5.8 × 10 M⁻¹ s⁻¹ for 2-Y at −40 °C, were determined from the slopes of linear plots of k_obs vs. [Fc] (ESI, Figs. S6a and S7b). This result clearly indicates that the reaction rates depend on the Lewis acidity of the metal ions in 2-M; the stronger the Lewis acidity of the binding metal ion is, the faster the ET reaction between 2-M and Fc is. Similarly, the k_et values of electron transfer from other Fc derivatives, such as octamethylferrocene (Me₈Fc), dimethylferrocene (Me₂Fc), bromoferrocene (BrFc), acetylferrocene (AcFc), and dibromoferrocene (Br₂Fc), to 2-M were determined and listed in Table S3 (also see Figs. S6 and S7 for the linear plots for the determination of k_et). In all of the reactions, the k_et values of 2-Sc were greater than those of 2-Y (i.e., k_et for 2-Sc > k_et for 2-Y), demonstrating again that 2-M binding a metal ion with stronger Lewis acidity is more reactive in the ET reaction. In addition, the reaction rates were different depending on the oxidation potential of the electron donors; the conversion of 2-M to 4 is faster with electron donors having lower oxidation potential (i.e., Me₈Fc > Me₂Fc > Fc > BrFc > AcFc > Br₂Fc) (see ESI, Table S3 and ESI, Figs. S6 and S7). These results imply that the electron transfer from the electron donor to 2-M determines the reaction rate (vide infra).

The one-electron reduction potentials (E_red) and reorganization energies of electron transfer (λ) of 2-Sc and 2-Y were then estimated to understand the fundamental ET properties of the iron(III)-peroxo complexes binding redox-inactive metal ions. Since the reduction of 2-Sc and 2-Y by the electron donors resulted in a rapid O-O bond cleavage to yield [(TMC)Fe^IV(O)]²⁺ (4), the E_red values of 2-Sc and 2-Y could not be determined by a conventional cyclic voltammetry method. We therefore estimated the E_red values by analyzing the driving force dependence of k_et for the irreversible ET reduction of 2-Sc and 2-Y using the Marcus theory of electron transfer.⁴⁶ First, the k_et values were converted to the activation free energies (ΔG_et^‡) using eq 1, where Z is the frequency factor taken as 1 × 10¹¹ M⁻¹ s⁻¹. The calculated ΔG_et^‡ values are listed in Supporting Information, Table S3. According to the Marcus theory of electron transfer, ΔG_et^‡ is given as a function of ΔG_et (free energy change of electron transfer) and λ, as shown in eq 2. Since ΔG_et is given as the difference between the one-electron oxidation potential of an electron donor (E_ox) and the one-electron reduction potential of [(TMC)Fe^III(O₂)]⁺–Mⁿ⁺ (E_red) (i.e., ΔG_et = e(E_ox – E_red)), eq 3 is rewritten as a linear correlation between (ΔG_et^‡)^1/2 and E_ox (eq 4),⁴⁶ and plots of (ΔG_et^‡)^1/2 vs. E_ox for electron transfer from Fc and Fc derivatives to 2-Sc and 2-Y are shown in Fig. 5.

$$k_{\mathrm{et}}=\mathrm{Zexp}(-\mathrm{\Delta }{G}_{\mathrm{et}}^{‡}/k_{\mathrm{B}}T)$$ (1)

$$\mathrm{\Delta }{G}_{\mathrm{et}}^{‡}=(\lambda /4){(1+\mathrm{\Delta }{G}_{\mathrm{et}}/\lambda )}^2$$ (2)

$${(\mathrm{\Delta }{G}_{\mathrm{et}}^{‡})}^{1/2}={(\lambda /4)}^{1/2}(1+e(E_{\mathrm{ox}}-E_{\mathrm{red}})/\lambda )$$ (3)

$${(\mathrm{\Delta }{G}_{\mathrm{et}}^{‡})}^{1/2}={(1/4\lambda )}^{1/2}e{E}_{\mathrm{ox}}+{(1/4\lambda )}^{1/2}(\lambda ­e{E}_{\mathrm{red}})$$ (4)

Fig. 5.

Plots of (ΔG_et^‡)^1/2 vs E_ox values of Fc derivatives [Br₂Fc (1), AcFc (2), BrFc (3), Fc (4), Me₂Fc (5), and Me₈Fc (6)] in the electron transfer from Fc derivatives to 2-Sc (blue circles) and 2-Y (red circles) in CH₃CN/CF₃CH₂OH (1:1) at −40 °C.

From the slopes and intercepts in Fig. 5, the E_red values of 2-Sc and 2-Y were estimated to be 0.40 and 0.16 V vs. SCE, respectively, and the λ values of both 2-Sc and 2-Y were evaluated to be 1.29 eV. The λ values of both 2-Sc and 2-Y were evaluated to be virtually the same as 1.29 eV, which is significantly smaller than that of the iron(IV)-oxo complex (4, λ = 2.37 eV). The smaller λ values of 2-Sc and 2-Y, as compared with 4, may result from the difference in the bond order between the peroxo complexes (e.g., single bond) and the oxo complex (e.g., double bond), since the reduction of the oxo double bond results in the significant elongation of the Fe–O bond, which is accompanied by a larger reorganization energy as compared with those of the peroxo complexes (2-Sc and 2-Y). The same λ values of 2-Sc and 2-Y suggest that the electron-transfer reduction occurs at the metal center and that the difference in the Fe–O bond distance caused by the electron-transfer reduction remains virtually the same between 2-Sc and 2-Y. The higher E_red value of 2-Sc than that of 2-Y results from the stronger Lewis acidity of Sc³⁺ than that of Y³⁺,⁴⁷ which further indicates that 2-Sc is more electron deficient than 2-Y. As a result, 2-Sc is reduced faster than 2-Y by the electron donors, as discussed above. To the best of our knowledge, this is the first time to report the fundamental ET properties of iron(III)-peroxo complexes binding redox-inactive metal ions (see Fig. S8 for the k_et value of 4 and the comparison of the E_red, λ, and k_et values of 2-Sc, 2-Y, and 4 in Supporting Information).

In conclusion, we have reported for the first time the effects of redox-inactive metal ions on the chemical properties of a mononuclear nonheme iron(III)-peroxo complex in ET reactions. While the iron(III)-peroxo complex, [(TMC)Fe^III(O₂)]⁺ (1), cannot be reduced by electron donors (e.g., Fc derivatives), the iron(III)-peroxo complexes binding redox-inactive metal ions, [(TMC)Fe^III(O₂)]⁺–M³⁺ (2-M, M³⁺ = Sc³⁺ and Y³⁺), are reduced by the electron donors, resulting in the peroxide O-O bond cleavage of 2-M to form the corresponding iron(IV)-oxo complex, [(TMC)Fe^IV(O)]²⁺ (4) (Scheme 1; see ESI, Fig.S9 for the ESI MS spectrum of the M³⁺(O) product). We have also shown that the rates of the conversion of 2-M to 4 are dependent on the Lewis acidity of the redox-inactive metal ions and concentration and oxidation potential of the electron donors.⁴⁸ These results indicate that the reduction of 2-M by electron donors to form the one-electron reduced species (i.e., [(TMC)Fe^II(O₂)]–M³⁺ (2’-M)) is the rate-determining step (r.d.s.), followed by a rapid heterolytic O-O bond cleavage of the peroxide ligand of 2’-M that results in the formation of 4 (Scheme 1a). The fundamental ET properties of 2-M, such as the reduction potential and the reorganization energy in ET reactions, have also been estimated and compared with those of the corresponding iron(IV)-oxo complex (4). The faster electron transfer rates and the greater reactivities of 2-M are ascribed to the smaller reorganization energy of 2-M than that of 4 (ESI, Fig. S8). In addition, the linear plots of (ΔG_et^‡)^1/2 vs E_ox shown in Fig. 5 support the proposed mechanism that the one-electron reduction of 2-M to [(TMC)Fe^II(O₂)]–Mⁿ⁺ is the r.d.s. (Scheme 1a). Finally, the observation that one-electron reduction of the [(TMC)Fe^III-O₂]-M³⁺ species affords the formation of the corresponding iron(IV)-oxo species leads us to propose that the iron(III) porphyrin-O₂-Cu(II) intermediate at the active site of CcO is reduced by one-electron to iron(II) porphyrin-O₂-Cu(II), followed by the heterolytic O-O bond cleavage of the iron(II) porphyrin-O₂-Cu(II) moiety to give the S = 1 Fe(IV)-oxo and Cu(II) species (Scheme 1b).^33,35-37,49 In this regard, the Cu(II) ion in the iron(III) porphyrin-O₂-Cu(II) intermediate in CcO may function as Lewis acid that facilitates the one-electron reduction of iron(III) porphyrin-O₂-Cu(II) (Scheme 1b), although no experimental evidence supporting our assertion has been obtained so far. Detailed mechanistic studies, including density functional theory (DFT) calculations for the one-electron reduction and O-O bond cleavage steps of the nonheme iron(III)-peroxo complex binding redox-inactive metal ions, are underway in this laboratory to correlate these biomimetic observations to enzymatic reactions (e.g., CcO).^50,51

Scheme 1.