Enhanced Electron-Transfer Reactivity of Nonheme Manganese(IV)– Oxo Complexes by Binding Scandium Ions

Abstract

One and two scandium ions (Sc³⁺) are bound strongly to nonheme manganese(IV)–oxo complexes, [(N4Py)Mn^IV(O)]²⁺ (N4Py = N,N-bis(2-pyridylmethyl)-N-bis(2-pyridyl)methylamine) and [(Bn-TPEN)Mn^IV(O)]²⁺ (Bn-TPEN = N-benzyl-N,N′,N′-tris(2-pyridylmethyl)-1,2-diaminoethane), to form Mn^IV(O)–(Sc³⁺)₁ and Mn^IV(O)–(Sc³⁺)₂ complexes, respectively. The binding of Sc³⁺ ions to the Mn^IV(O) complexes was examined by spectroscopic methods as well as by DFT calculations. The one-electron reduction potentials of the Mn^IV(O) complexes were markedly shifted to a positive direction by binding of Sc³⁺ ions. Accordingly, rates of the electron transfer reactions of the Mn^IV(O) complexes were enhanced as much as 10⁷–fold by binding of two Sc³⁺ ions. The driving force dependence of electron transfer from various electron donors to the Mn^IV(O) and Mn^IV(O)–(Sc³⁺)₂ complexes was examined and analyzed in light of the Marcus theory of electron transfer to determine the reorganization energies of electron transfer. The smaller reorganization energies and much more positive reduction potentials of the Mn^IV(O)–(Sc³⁺)₂ complexes resulted in remarkable enhancement of the electron-transfer reactivity of the Mn^IV(O) complexes. Such a dramatic enhancement of the electron-transfer reactivity of the Mn^IV(O) complexes by binding of Sc³⁺ ions resulted in the change of mechanism in the sulfoxidation of thioanisoles by Mn^IV(O) complexes from a direct oxygen atom transfer pathway without metal ion binding to an electron-transfer pathway with binding of Sc³⁺ ions.

INTRODUCTION

High-valent metal–oxo complexes play pivotal roles as reactive intermediates in the reactions of heme (cytochromes P450 and peroxidases)^1–3 and nonheme metalloenzymes (taurine/α-ketoglutarate dioxygenase (TauD),⁴ soluble methane monooxygenase (sMMO),⁵ and oxygen-evolving center (OEC) in Photosystem II^6,7) as well as in their biomimetic oxidation catalysis.^8–13 A number of synthetic high-valent metal–oxo complexes have been synthesized and characterized by various spectroscopic techniques as well as by X-ray crystallography.^8–17 The oxidizing reactivity of the high-valent metal–oxo complexes has so far been finely controlled by the oxidation state of metals and the supporting and axial ligands.^8–28 Alternatively, binding of redox-inactive metal ions acting as Lewis acids to the metal–oxo moiety of high-valent metal–oxo complexes has also been reported to enhance the oxidizing power of the metal–oxo complexes.^29–32 For example, Lau and co-workers have shown that rates of oxidation of alkanes by MnO₄⁻ are accelerated dramatically by addition of Lewis acids.³³ We have reported the first X-ray crystal structure of a Sc³⁺ ion-bound iron(IV)–oxo complex³⁴ and found that the binding of Sc³⁺ ions to iron(IV)–oxo complexes resulted in the remarkable enhancement of the oxidizing reactivity of the iron–oxo complexes in various oxidation reactions.²⁹ Those findings are reminiscent of the indispensable role of Ca²⁺ in the manganese–oxo–calcium (Mn₄CaO₅) active site in OEC, which catalyzes four-electron oxidation of water to dioxygen, although the exact function of Ca²⁺ has yet to be clarified.^35,36 In this context, we have recently communicated preliminary results that binding of redox-inactive metal ions to a nonheme manganese(IV)–oxo complex affected the oxidizing ability significantly.³⁷ Goldberg and co-workers also reported the influence of a redox-inactive Zn²⁺ ion on a valence tautomerization of a manganese(V)–oxo corrolazine complex.³⁸ More recently, Agapie and co-workers reported that the redox potentials of tetranuclear heterometallic trimanganese dioxo clusters [Mn₃M(µ₄-O)(µ₂-O)] containing a redox-inactive metal ion, which were synthesized as an excellent structural model of the OEC active site,³⁹ were systematically controlled by the Lewis acidity of the redox-inactive metal ions.⁴⁰ The positive shift in the redox potentials of an iron(IV)–oxo complex was also observed in the presence of various redox-inactive metal ions, showing a correlation between the reactivity of the iron(IV)–oxo complex and the Lewis acidity of the redox-inactive metal ions.^29a However, the electron-transfer properties of metal ion-bound manganese-(IV)–oxo (Mn^IV(O)) complexes, which are the most fundamental factor in controlling the reactivity of metal–oxo species in oxidation reactions, have yet to be clarified.

We report herein not only the synthesis and characterization of nonheme Mn^IV(O) complexes binding one and two Sc³⁺ ions, Mn^IV(O)–(Sc³⁺)₁ and Mn^IV(O)–(Sc³⁺)₂, but also the detailed kinetic data on electron transfer from various electron donors to Mn^IV(O), Mn^IV(O)–(Sc³⁺)₁, and Mn^IV(O)–(Sc³⁺)₂ complexes. The large positive shifts in one-electron reduction potentials of Mn^IV(O) complexes by binding of one and two Sc³⁺ ions have also been determined for the first time by carrying out redox titration experiments using electron donors with known one-electron oxidation potentials. Rates of the electron-transfer reactions by Mn^IV(O) complexes were remarkably enhanced by binding of Sc³⁺ ions. The driving force dependence of the rate constants of electron-transfer reactions has been analyzed in light of the Marcus theory of electron transfer⁴¹ to determine the reorganization energies of electron transfer of Sc³⁺ ion-bound Mn^IV(O) complexes in comparison with those of Mn^IV(O) complexes without Sc³⁺ ions. We have also shown that binding of redox-inactive metal ions enhances the reactivity of Mn^IV(O) complexes in sulfoxidation of thioanisoles and changes the sulfoxidation mechanism from direct oxygen atom transfer to electron transfer.

RESULTS AND DISCUSSION

Binding of Sc³⁺ Ions to Nonheme Mn^IV(O) Complexes

[(Bn-TPEN)Mn^IV(O)]²⁺ (1) and its scandium-bound species were synthesized and characterized spectroscopically, as reported for binding of Sc³⁺ ions to [(N4Py)Mn^IV(O)]²⁺ (2) (see Chart 1).^37,41 Addition of up to 2 equiv of Sc(OTf)₃ to a solution of 1 in a solvent mixture of CF₃CH₂OH/CH₃CN (v/v = 19:1) resulted in a blue shift of the absorption band of 1 at λ_max = 1020 nm to that at λ_max = 740 nm with an isosbestic point at 900 nm (Figure 1a). Such a blue shift suggests that the ground state of 1 is more stabilized by binding of Sc³⁺ ions. Further addition of Sc(OTf)₃ resulted in a more blue-shift to give the absorption band at λ_max = 690 nm (Figure 1c). No further spectral change was observed by addition of more than 9 equiv of Sc(OTf)₃. Such stepwise spectral changes indicate the binding of one and two Sc³⁺ ions to 1 to produce Mn^IV(O)–(Sc³⁺)₁ and Mn^IV(O)–(Sc³⁺)₂ complexes, respectively. The X-band EPR spectra of [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₁ and [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₂ exhibit signals that are characteristic of S = 3/2 Mn^IV (Figure S1 in Supporting Information (SI)). This result is the same as that reported previously for the complexes of 2 with Sc³⁺ ions.³⁷ The magnetic moments of [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₁ and [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₂ were also determined to be 4.4 and 4.3 µ_B by the modified NMR technique of Evans,⁴² respectively, confirming the spin state of S = 3/2 for both complexes.^41a

Chart 1.

Figure 1.

(a) UV–vis spectral changes showing the conversion from [(Bn-TPEN)Mn^IV(O)]²⁺ (red line) to [(Bn-TPEN)Mn (O)]²⁺– (Sc³⁺)₁ (green line) upon addition of incremental amounts of Sc³⁺ (from 0.0 to 2.0 mM) in the titration experiment. (b) Spectroscopic titration monitored at 1020 nm (red rhombuses) due to the decay of [(Bn-TPEN)Mn^IV(O)]²⁺ and at 740 nm (green squares) due to the formation of [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₁. (c) UV–vis spectral changes showing the conversion from [(Bn-TPEN)Mn (O)]²⁺– (Sc³⁺)₁ (green line) to [(Bn-TPEN)Mn^IV (O)]²⁺–(Sc³⁺)₂ (blue line) observed upon addition of Sc³⁺ ions (from 0.0 to 12 mM). (d) Spectroscopic titration monitored at 690 nm (blue squares) due to [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₂.

The formation constant of the [(Bn-TPEN)Mn^IV(O)]²⁺– (Sc³⁺)1 complex was determined to be 4.0 × 10³ M⁻¹, by analyzing the spectral change in Figure 1b (see the linear plot to determine the formation constant in Figure S2a in the SI).⁴³ The formation constant of the [(Bn-TPEN)Mn^IV(O)]²⁺– (Sc³⁺)₂ complex was also determined to be 1.2 × 10³ M⁻¹ from the titration curve in Figure 1d (see also Figure S2b in SI for the linear plot to determine the formation constant), which is somewhat smaller than that of the [(N4Py)Mn^IV(O)]²⁺– (Sc³⁺)₂ complex (6.1 × 10³ M⁻¹).³⁷

The structural details of Sc³⁺ binding was investigated using Mn K-edge EXAFS on 1, [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₁ and [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₂ (see the SI, Experimental section for X-ray absorption spectroscopy, Figure S3 and Table S1). The EXAFS results show that on going from no Sc³⁺ binding in 1 to one and two Sc³⁺ binding, the Mn═O bond elongates from 1.69 to 1.74(2) Å, which is consistent with a weakening of the Mn═O bond. The EXAFS data also reveal a short Mn–Sc distance (3.45(10) Å), which clearly indicates that Sc³⁺ ions bind to the Mn^IV(O) moiety in both [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₁ and [(Bn-TPEN)Mn^IV(O)]²⁺– (Sc²⁺)₂.

Density functional theory (DFT)⁴⁴ calculations of the Mn^IV(O)–(Sc³⁺)₂ complex suggest that one Sc³⁺ ion binds directly to the oxo moiety of the Mn^IV(O) complex but the second Sc³⁺ ion is located at the secondary coordination sphere (see the DFT-optimized structures in Figure 2).^37,45 As shown previously, the Mn–O bond is elongated upon one Sc³⁺ binding (from 1.68 Å to 1.75 Å), but no further elongation is seen upon binding of the second Sc³⁺ ion (Table S2 in SI). This result is in a good agreement with that obtained from EXAFS experiments. This is also in line with what has been observed previously with [(N4Py)Mn^IV(O)]²⁺ (2).³⁷ The significant blue shift in the absorption band of the Mn^IV(O)–(Sc³⁺)₂ complex as compared with that of the Mn^IV(O)–(Sc³⁺)₁ complex indicates that the electronic state of the Mn^IV(O) moiety is somewhat perturbed by binding of the second Sc³⁺ ion, but the conservation of the core geometry in Mn^IV(O)–(Sc³⁺)₂, compared to Mn^IV(O)–(Sc³⁺)₁, suggests that the changes mostly affect the unoccupied virtual orbitals (such as σ*_xy). The blue-shift indicates that the virtual orbitals are shifted higher in energy relative to HOMO, hence stabilizing the occupied orbitals.⁴⁶ The Mulliken spin density distribution shows negligible spin on the Sc³⁺ or counteranion atoms in all cases, while showing around three radicals in total for the Mn(O)-5xN moiety (Table S3 in the SI), indicating that the Mn^IV(O) configuration is kept after Sc³⁺ binding. Indeed, natural orbital analysis showed that the valence orbital occupation was (δ, π*_xz, and π*_yz) corresponding to highspin Mn^IV(O). In addition, none of the nine valence Mn^IV(O) orbitals with Mn d-orbital elements in it is mixing with Sc³⁺ orbitals, indicating a pure ionic Mn^IV(O)–Sc³⁺ bond despite the short distance (1.94 Å). Hence, the Mn^IV–O bond elongation can be attributed to pure Coulomb interactions rather than to changes in the electronic structure.

Figure 2.

DFT-optimized structures of (a) [(Bn-TPEN)Mn^IV(O)]²⁺ (1), (b) [(Bn-TPEN)Mn^IV(O)–Sc(OTf)₃]²⁺, and (c) [(Bn-TPEN)-Mn^IV(O)–[Sc(OTf)₃]₂]²⁺, calculated at the B3LYP/LACVP level in the solvent phase. The Mn–O bond lengths of [Mn^IV(O)]²⁺, [Mn^IV(O)]²⁺–Sc³⁺, and [Mn^IV(O)]²⁺–(Sc³⁺)₂ were calculated to be 1.68, 1.75, and 1.75 Å, respectively (see the DFT Calculations Section in the the SI for computational details). H atoms in Bn-TPEN and F atoms in the OTf-counterions have been omitted for clarity (Mn, pink; N, blue; O, red; C, green; Sc, purple; S, yellow). (d) Schematic drawing of the Bn-TPEN ligand.

Positive Shifts in One-Electron Reduction Potentials of Nonheme Mn^IV(O) Complexes by Binding of Sc³⁺ Ions

Electron transfer from ferrocene derivatives to Mn^IV(O) complexes 1 and 2 occurred to completion in CF₃CH₂OH/ CH₃CN (v/v = 1:1) at 273 K (Figure 3a). The solvent mixture was used due to the solubility of electron donors and the stability of 1 and 2; both of which are quite stable at 273 K. The redox titration for formation of dibromoferrocenium ion (Br₂Fc⁺) in Figure 3b indicates only one-electron reduction of 2 occurred without further reduction by Br₂Fc (see Table 1 for E_ox values of electron donors). This result is in agreement with the result of the redox titration of 1 with dibromoferrocene (Br₂Fc) reported previously.^41b

Figure 3.

(a) Absorption spectral changes observed in electron transfer from dibromoferrocene (Br₂Fc; 5.0 × 10⁻³ M) to [(N4Py)-Mn^IV(O)]²⁺ (2; 5.0 × 10⁻⁴ M) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K. (b) Plot of concentration of Br₂Fc⁺ produced in electron transfer from Br₂Fc to 2 vs initial concentration of Br₂Fc, [Br₂Fc]₀.

Table 1.

One-Electron Oxidation Potentials (E_ox) of Electron Donors and Second-Order Rate Constants of Electron Transfer from Electron Donors to 1 and 2 in the Absence and Presence of Sc(OTf)₃ (10 mM) with Driving Force of Electron Transfer (–ΔG_et) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K

		[(Bn-TPEN)MnIV(O)]2+a		[(N4Py)MnIV(O)]2+
electron donor	Eox, V vs SCEa	kobs, M−1 s−1	−ΔGet, eV	kobs, M−1 s−1	−ΔGet, eV
dimethylferrocene	0.26	(1.0 ± 0.1) × 105	0.52
ferrocene	0.37	(6.1 ± 0.3) × 103	0.41	(8.2 ± 0.1) × 103	0.43
bromoferrocene	0.54	(9.3 ± 0.1) × 102	0.24	(8.4 ± 0.3) × 102	0.26
acetylferrocene	0.62	(9.3 ± 0.2) × 10	0.16	(8.2 ± 0.2) × 10	0.18
dibromoferrocene	0.71	(3.1 ± 0.1) × 10	0.07	(4.3 ± 0.1) × 10	0.09
		[(Bn-TPEN)MnIV (O)]2+ – (Sc3+)2		[(N4Py)MnIV(O)]2+ – (Sc3+)2
electron donor	Eox, V vs SCEb	kobs, M−1 s−1	−ΔGet, eV	kobs, M−1 s−1	−ΔGet, eV
[FeII (Me2-phen)3]2+	0.94	(1.7 ± 0.2) × 105	0.42	(2.3 ± 0.1) × 105	0.48
[FeII (Ph2-phen)3]2+	1.02	(2.1 ± 0.1) × 104	0.34	(2.4 ± 0.4) × 104	0.40
[FeII (bpy)3]2+	1.06	(3.6 ± 0.1) × 103	0.30	(6.9 ± 0.1) × 103	0.36
[FeII (5-Cl-phen)3]2+	1.20	(4.1 ± 0.8) × 102	0.16	(4.1 ± 0.2) × 102	0.22
[RuII (bpy)3]2+	1.24	(7.3 ± 0.1) × 10	0.12	(6.4 ± 0.3) × 10	0.18

^aTaken from ref 41b.

^bTaken from ref 30b.

When ferrocene derivatives were replaced by a weaker reductant such as [Ru^II(bpy)₃]²⁺ (bpy =2.2′-bipyridine; E_ox = 1.24 V vs SCE), no electron transfer was observed. This is consistent with the lower one-electron reduction potential of 1 (E_red = 0.78 V vs SCE), as reported previously.^41b The E_red value of 2 was also determined to be 0.80 V vs SCE by the redox titration using Br₂Fc (E_ox = 0.71 V vs SCE) (Figure 3b).

When the Mn^IV(O)–(Sc³⁺)₁ complex of 1 was employed with 2 equiv of Sc(OTf)₃ (2.0 mM), electron transfer from [Ru^II(bpy)₃]²⁺ to the Mn^IV(O)–(Sc³⁺)₁ complex became energetically feasible to occur as shown in Figure 4a, where the absorption band (λ_max = 650 nm) due to [Ru^III(bpy)₃]³⁺ increased. In this case, however, the electron transfer from [Ru^II(bpy)₃]²⁺ to the Mn^IV(O)–(Sc³⁺)₁ complex was not complete with 2 equiv of [Ru^II(bpy)₃]²⁺. Figure 4b shows the titration curve of the electron-transfer reaction, suggesting that electron transfer from [Ru^II(bpy)₃]²⁺ to the Mn^IV(O)–(Sc³⁺)₁ complex is in equilibrium with back electron transfer from the Mn^III(O)–(Sc³⁺)₁ complex to [Ru^III(bpy)₃]³⁺ (eq 1, where K_et is the electron-transfer equilibrium constant).

$${[{\text{Ru}}^{\text{II}}{(\text{bpy})}_3]}^{2+}+{[{\text{Mn}}^{\text{IV}}(\mathrm{O})]}^{2+}-{\text{Sc}}^{3+}\stackrel{K_{\text{et}}}{\rightleftharpoons }{[{\text{Ru}}^{\text{III}}{(\text{bpy})}_3]}^{3+}+{[{\text{Mn}}^{\text{III}}(\mathrm{O})]}^{+}-{\text{Sc}}^{3+}$$ (1)

Figure 4.

(a) Absorption spectral changes observed in electron transfer from [Ru^II(bpy)]²⁺ (5.0 × 10⁻³ M) to [(Bn-TPEN)-Mn^IV(O)]²⁺–Sc³⁺ (5.0 × 10⁻⁴ M) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K. (b) Plot of concentration of [Ru^III(bpy)₃]³⁺ produced in electron transfer from [Ru^II(bpy)₃]²⁺ to [(Bn-TPEN)Mn^IV(O)]²⁺– Sc³⁺ vs initial concentration of [Ru^II(bpy)₃]²⁺, [[Ru^II(bpy)₃]²⁺]₀.

From the redox titration curve in Figure 4b, the K_et value was determined to be 5.7 (see a linear plot to determine the K_et value in Figure S4d in the SI).⁴³ The E_red value of 1.28 V vs SCE for [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₁ was determined from the K_et value and the E_ox value of [Ru^II(bpy)₃]²⁺ (1.24 V vs SCE) using the Nernst equation (eq 2). This E_red value is significantly more positive than the E_red value of 1 (0.78 V vs SCE), which was reported previously.^41b

$$E_{\text{red}}=E_{\text{ox}}+(RT/F)\text{ln}{K}_{\text{et}}$$ (2)

Similarly, the E_red values of 1 and 2 in the presence of various concentrations of Sc(OTf)₃ were determined from the K_et values obtained by the redox titrations with [Ru^II(bpy)₃]²⁺ (E_ox = 1.24 V vs SCE) and [Ru^II(5-Cl-phen)₃]²⁺ (5-Cl-phen = 5-chloro-1,10-phenanthroline; E_ox = 1.36 V vs SCE; see Figures S4 and S5 in the SI). Figure 5 shows plots of E_red of 1 and 2 in the presence of various concentrations of Sc(OTf)₃. The E_red values of 1 and 2 were shifted from 0.78 and 0.80 V vs SCE in the absence of Sc(OTf)₃ to 1.28 and 1.31 V vs SCE in the presence of 1 equiv (for N4Py) or 2 equiv (for Bn-TPEN) of Sc(OTf)₃, which correspond to the E_red values of the [Mn^IV(O)]²⁺–Sc³⁺ complexes of 1 and 2, respectively. The E_red values increase with increasing concentration of Sc(OTf)₃ to reach constant values with more than 6 equiv of Sc(OTf)₃. This is consistent with the UV–vis spectral change to produce the [Mn^IV(O)]²⁺–(Sc³⁺)₂ complex in Figure 1b. The saturated E_red values of 1.36 and 1.42 V vs SCE in the presence of large excess Sc(OTf)₃ correspond to those of the [Mn^IV(O)]²⁺– (Sc³⁺)₂ complexes of 1 and 2, respectively. These E_red values are summarized in Table 2. There is a large difference in the E_red values between [Mn^IV(O)]²⁺ and [Mn^IV(O)]²⁺–Sc³⁺, indicating that the binding of Sc³⁺ to [Mn^III(O)]⁺ is much stronger than that to [Mn^IV(O)]²⁺ as expected from the increased basicity of the oxo moiety in [Mn^III(O)]⁺. The smaller difference in E_red between [Mn^IV(O)²⁺]–(Sc³⁺)₁ and [Mn^IV(O)]²⁺–(Sc³⁺)₂ may support the previous conclusion based on XANES/EXAFS analysis together with DFT calculations: second Sc³⁺ ion is located at the secondary coordination sphere rather than direct binding to the oxo moiety.37

Figure 5.

Dependence of E_red of [(N4Py)Mn^IV(O)]²⁺ (blue circles) and [(Bn-TPEN)Mn^IV(O)]²⁺ (red circles) on [Sc³⁺] in CF₃CH₂OH/ CH₃CN (v/v = 1:1) at 273 K. The E_red values were determined from the equilibrium constants (K_et) between one electron donors and [(N4Py)Mn^IV(O)]²⁺.

Table 2.

One-Electron Reduction Potentials (E_red) and Reorganization Energies (λ) of Electron-Transfer Reduction of [Mn^IV(O)]²⁺, [Mn^IV(O)]²⁺–(Sc³⁺)₁, and [Mn^IV(O)]²⁺– (Sc³⁺)₂

complex	Ered, V vs SCE	λ, eV
[(Bn-TPEN)MnIV(O)]2+	0.78 ± 0.01a	2.24 ± 0.03a
[(Bn-TPEN)MnIV(O)]2+–(Sc3+)1	1.28 ± 0.02
[(Bn-TPEN)MnIV(O)]2+–(Sc3+)2	1.36 ± 0.02	2.12 ± 0.02
[(N4Py)MnIV(O)]2+	0.80 ± 0.01	2.27 ± 0.03
[(N4Py)MnIV(O)]2+–(Sc3+)1	1.31 ± 0.02
[(N4Py)MnIV(O)]2+–(Sc3+)2	1.42 ± 0.01	2.21 ± 0.02

^aTaken from ref 41b.

Comparison of Electron-Transfer Reactivity of Nonheme Mn^IV(O) Complexes with and without Sc³⁺ Ions

A remarkable positive shift of E_red values of [Mn^IV(O)]²⁺ complexes by binding one or two Sc³⁺ ions predicts the much enhanced electron-transfer reactivity. Rate constants of electron transfer from various electron donors to 1 and 2 in the absence and presence of Sc³⁺ were determined by monitoring the absorbance changes in reaction solutions due to the oxidized electron donors. Rates of electron transfer in the absence of Sc³⁺ obeyed pseudo-first-order kinetics in the presence of large excess of electron donors, and the observed pseudo-first-order rate constants (k_obs) increased linearly with no intercept with concentrations of electron donors (Figure S6 in the SI). Rate constants of electron transfer from one-electron reductants to 1 and 2 in the presence of Sc³⁺ were determined under second-order reaction conditions (Figures S7 and S8 in the SI). The k_et values increased with increasing concentration of Sc³⁺ in the conversion of [Mn^IV(O)]²⁺ to [Mn^IV(O)]²⁺– (Sc³⁺)₂ (Figure 6). The k_et values thus determined are listed in Table 1 along with the one-electron oxidation potentials (E_ox) of donors and the driving force of electron transfer (–ΔG_et).

Figure 6.

Dependence of k_et on [Sc³⁺] for electron transfer from [Ru^II(bpy)₃]²⁺ to [(Bn-TPEN)Mn^IV(O)]²⁺ in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K.

Figure 7 shows comparison of dependence of k_et on E_ox of electron donors for electron transfer from electron donors to [(Bn-TPEN)Mn^IV(O)]²⁺ (1) with that of electron transfer to [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₂ in the absence and presence of Sc(OTf)₃ (5.0 mM) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K. The k_et values of [Mn^IV(O)]²⁺–(Sc³⁺)₂ are much larger than those of [Mn^IV(O)]²⁺.

Figure 7.

Plots of log k_et vs the one-electron oxidation potentials of electron donors (E_ox) for electron transfer from various electron donors [(1) dimethylferrocene, (2) ferrocene, (3) bromoferrocene, (4) acetylferrocene, (5) Br₂Fc, (6) [Fe^II(Me₂-phen)₃]²⁺, (7) [Fe^II(Ph₂-phen)₃]²⁺, (8) [Fe^II(bpy)₃]²⁺, (9) [Fe^II(5-Cl-phen)₃]²⁺, and (10) [Ru^II(bpy)₃]²⁺] to [(Bn-TPEN)Mn^IV(O)]²⁺ (red line) and [(Mn^IV(O)]²⁺–(Sc³⁺)₂ (blue line) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K.

The driving force dependences of the rate constants of electron transfer from electron donors to [Mn^IV(O)]²⁺ and [Mn^IV(O)]²⁺–(Sc³⁺)₂ for 1 are shown in Figure 8, where two different plots in Figure 7 are largely unified. The driving force dependence of log k_et is well fitted by the solid line in Figure 8 in light of the Marcus theory of adiabatic outer-sphere electron transfer (eq 3)⁴⁷

$$k_{\text{et}}=Z\text{exp}[-(\lambda /4){(1+\mathrm{\Delta }{G}_{\text{et}}/\lambda )}^2/k_{\mathrm{B}}T]$$ (3)

where Z is the collision frequency taken as 1 × 10¹¹ M⁻¹ s⁻¹, λ is the reorganization energy of electron transfer, k_B is the Boltzmann constant, and T is the absolute temperature. The closer look of the fitting afforded slightly different λ values, such as one for [(Bn-TPEN)Mn^IV(O)]²⁺ (2.24 ± 0.03 eV) and the other for [(Bn-TPEN)Mn^IV(O)]²⁺–(Sc³⁺)₂ (2.12 ± 0.02 eV). Similarly, the other λ values were determined with the E_red values, as listed in Table 2.

Figure 8.

Driving force (–ΔG_et) dependence of rate constants (log k_et) of electron transfer from one-electron donors [(1) dimethylferrocene, (2) ferrocene, (3) bromoferrocene, (4) acetylferrocene, (5) Br₂Fc, (6) [Fe^II(Me₂-phen)₃]²⁺, (7) [Fe^II(Ph₂-phen)₃]²⁺, (8) [Fe^II(bpy)₃]²⁺, (9) [Fe^II(5-Cl-phen)₃]²⁺, and (10) [Ru^II(bpy)₃]²⁺] to [(Bn-TPEN)Mn^IV(O)]²⁺ (red circles) and [(Bn-TPEN)Mn^IV(O)]²⁺– (Sc³⁺)₂ (blue circles) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K. The blue and red lines are the Marcus lines calculated with λ values of 2.12 and 2.24 eV, respectively.

The λ values of the electron-transfer reduction of [Mn^IV(O)]²⁺ are similar to those determined for the electrontransfer reduction of iron(IV)–oxo complexes.^23b This indicates that the one-electron reduction of high-valent metal–oxo complexes generally requires a large reorganization energy probably due to the significant elongation of metal–oxo bond upon one-electron reduction. The smaller λ value of [Mn^IV(O)]²⁺–(Sc³⁺)₂, as compared with [Mn^IV(O)]²⁺, may be explained by the elongation of the Mn^IV–O bond prior to electron transfer because of binding of two Sc³⁺ ions, which results in smaller change in the Mn–O bond distance upon the electron transfer. It should be noted that [(N4Py)Mn^IV(O)]²⁺– (Sc³⁺)₂ has the most positive E_red value as compared with those of iron(IV)–oxo complexes.²³ Thus, [(N4Py)Mn^IV(O)]²⁺– (Sc³⁺)₂ is regarded as the strongest oxidant among high-valent metal–oxo complexes reported so far.

Change in Mechanism from Direct Oxygen Atom Transfer to Electron Transfer in Sulfoxidation of Thioanisoles with Nonheme Mn^IV(O) Complexes by Binding of Sc³⁺ Ions

We have suggested previously that the mechanism of sulfoxidation of thioanisoles with [(N4Py)-Mn^IV(O)]²⁺ is changed from direct oxygen atom transfer (DOT) to electron transfer (ET) when Sc³⁺ ions are bound to [(N4Py)Mn^IV(O)]²⁺.³⁷ This proposal is now verified by the electron-transfer driving force dependence of log k_obs for sulfoxidation of thioanisoles with [(N4Py)Mn^IV(O)]²⁺ vs [(N4Py)Mn^IV(O)]²⁺–(Sc³⁺)₂ (Figure 9). The k_obs values of sulfoxidation of thioanisoles with [(N4Py)Mn^IV(O)]²⁺ are much larger than the predicted values of the solid line drawn based on eq 3 for outer-sphere electron-transfer reactions. This implies that much stronger interaction between thioanisoles and [(N4Py)Mn^IV(O)]²⁺ is required in the sulfoxidation reaction as compared with outer-sphere electron transfer from thioanisoles to [(N4Py)Mn^IV(O)]²⁺. Thus, the sulfoxidation reaction with [(N4Py)Mn^IV(O)]²⁺ proceeds via direct oxygen atom transfer. In contrast, the rate constants of sulfoxidation of thioanisoles with [(N4Py)Mn^IV(O)]²⁺– (Sc³⁺)₂ fit well in the Marcus line for electron transfer from one-electron reductants to [(N4Py)Mn^IV(O)]²⁺–(Sc³⁺)₂. Such an agreement clearly indicates that the sulfoxidation of thioanisoles by [(N4Py)Mn^IV(O)]²⁺–(Sc³⁺)₂ proceeds via an outer-sphere electron-transfer pathway (Figure 9).

Figure 9.

Plots of log k_obs for sulfoxidation of para-X-substituted thioanisoles [X = (1) MeO, (2) Me, (3) H, (4) F, (5) Br, and (6) CN] with [(N4Py)Mn^IV(O)]²⁺ (0.50 mM) in CF₃CH₂OH/CH₃CN (v/v = 19:1) at 273 K vs the driving force of electron transfer [–ΔG_et = e(E_red – E_ox)] from the thioanisoles to [(N4Py)Mn^IV(O)]²⁺ in the absence of Sc³⁺ (red circles) and the presence of 5.0 mM Sc³⁺ (blue circles). The k_obs values for the sulfoxidation were taken from ref ³⁷. The black and green circles show the driving-force dependence of the rate constants (log k_et) for electron transfer from one-electron reductants [(7) [Fe^II(Me₂-phen)₃]²⁺, (8) [Fe^II(Ph₂-phen)₃]²⁺, (9) [Fe^II(bpy)₃]²⁺, (10) [Fe^II(5-Cl-phen)₃]²⁺, (11) [Ru^II(bpy)₃]²⁺, (12) ferrocene, (13) bromoferrocene, (14) acetylferrocene, and (15) Br₂Fc] to [(N4Py)-Mn^IV(O)]²⁺ in the absence of Sc³⁺ (black circles) and the presence of 5.0 mM Sc³⁺ (green circles) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K. The blue, green, and black lines are the best fit of the Marcus lines calculated for blue, green, and black circles, respectively.

The occurrence of electron transfer was confirmed in the case of p-methoxythioanisole in the presence of Sc(OTf)₃ (5.0 mM), where the driving force for electron transfer is positive (–ΔG_et = 0.20 eV; i.e., a thermodynamically favorable process). As shown in Figure 10, the transient absorption band at 580 nm due to p-methoxythioanisole radical cation appeared instantly, followed by a relatively slow decay (see ref ^30b for the reference spectrum of p-methoxythioanisole radical cation). The absorption band at 635 nm due to [(N4Py)Mn^IV(O)]²⁺– (Sc³⁺)₂ was not detected even with a stopped-flow spectrometer since electron transfer from p-methoxythioanisole to [(N4Py)Mn^IV(O)]²⁺–(Sc³⁺)₂³⁷ was extremely fast. These results clearly demonstrate that the electron-transfer pathway (ET/OT in Scheme 1) becomes dominant over the oxygen atom transfer pathway (DOT in Scheme 1) when the sulfoxidation by the Mn^IV(O) complex is carried out in the presence of Sc³⁺.

Figure 10.

UV–vis spectral changes in the reaction of [(N4Py)-Mn^IV(O)]²⁺ (6.25 × 10⁻⁵ M) with p-methoxythioanisole (1.25 × 10⁻³ M) in the presence of Sc(OTf)₃ (6.25 × 10⁻⁴ M) in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K. Inset shows the time course of absorbance at 580 nm due to p-methoxythioanisole radical cation.

Scheme 1.

CONCLUSION

The electron-transfer reactivity of nonheme Mn^IV(O) complexes was enhanced as much as 10⁷-fold by binding of Sc³⁺ ions. A large positive shift in the one-electron reduction potential was observed by binding of Sc³⁺ ions to [Mn^IV(O)]²⁺ (e.g., from 0.80 V of [(N4Py)Mn^IV(O)]²⁺ to 1.42 V of [(N4Py)Mn^IV(O)]²⁺–(Sc³⁺)₂). The reorganization energy of electron transfer becomes smaller by binding of Sc³⁺ ions because of the elongation of Mn^IV–O bond, which results in a smaller change in the Mn–O distance upon the electron transfer. Thus, nonheme Mn^IV(O) complexes become much stronger oxidants by binding of Sc³⁺ ions, resulting in the change of mechanism in the sulfoxidation of thioanisoles by Mn^IV(O) complexes, such as from a direct oxygen atom transfer pathway by [Mn^IV(O)]²⁺ to an electron transfer pathway by [Mn^IV(O)]²⁺–(Sc³⁺)_x. The remarkable enhancement of the electron-transfer reactivity of Mn^IV(O) complexes by binding of Sc³⁺ ions may pave a new way to improve the catalytic reactivity of metal–oxo complexes in oxidation reactions.

EXPERIMENTAL SECTION

Materials

Commercially available chemicals were used without further purification unless otherwise indicated. Solvents were dried according to published procedures and distilled under Ar prior to use. Scandium triflate, Sc(OTf)₃ (OTf = SO₃CF₃⁻), was purchased from Aldrich and used as received. Iodosylbenzene (PhIO) was prepared by a literature method.⁴⁹ N4Py and Bn-TPEN ligands and Mn^II(OTf)₂ 2CH₃CN were prepared by literature methods.^20b,50 The [(N4Py)Mn^II(CH₃CN)](CF₃SO₃)₂ and [(Bn-TPEN)-Mn^II(CH₃C(CF₃SO₃)₂ complexes were prepared in a drybox.^37,41 N4Py (0.54 mmol, 0.20 g) and Mn^II(OTf)₂ 2CH₃CN (0.82 mmol, 0.32 g) were dissolved in CH₃CN, and the reaction solution was stirred at ambient temperature overnight. The resulting solution was filtered and added to a large volume of Et₂O. The product was obtained as a white solid with 60% yield (0.24 g). Bn-TPEN (0.47 mmol, 0.20 g) and Mn^II(CF₃SO₃)₂ 2CH₃CN (0.57 mmol, 0.25 g) were dissolved in CH₃CN and stirred at ambient temperature overnight. The resulting solution was filtered and added to a large volume of Et₂O. The product was obtained as a white solid in 85% yield (0.38 g). One-electron reductants, such as [Fe^II(Me₂-phen)₃]²⁺, [Fe^II(Ph₂-phen)₃]²⁺, [Fe^II(bpy)₃]²⁺, [Fe^II(5-Cl-phen)₃]²⁺, and [Ru^II(bpy)₃] ⁺, were synthesized according to literature methods.⁵¹

DFT Calculations

Calculations were done with density functional theory (DFT)⁴⁴ using the Gaussian 09 package⁵² and the B3LYP functional.⁵³ Optimizing and single-point frequency calculations were done with the LACVP basis set⁵⁴ (except for S, which required 6– 311+G*), while a single-point evaluation was done using the LACV3P*+ basis set⁵⁴ in order to obtain more accurate Mulliken spin density distribution. All calculations (including the optimizations) were done in solvent (acetonitrile) using the CPCM scheme.⁵⁵

Redox Titrations

Electron transfer from dibromoferrocene to [Mn^IV(O)]²⁺ (5.0 × 10⁻⁴ M) was examined from the spectral change in the various concentration of dibromoferrocene (1.5 × 10⁻⁴ to 5.0 × 10⁻² M) at 273 K. Typically, a deaerated CH₃CN solution of dibromoferrocene (7.5 × 10⁻⁵ – 5.0 × 10⁻⁴ M) was added to a deaerated trifluoroethanol solution containing [Mn^IV(O)]²⁺ (5.0 × 10⁻⁴ M). The concentration of dibromoferrocenium ion (Br₂Fc⁺) was determined from the absorption band at λ = 700 nm due to Br₂Fc⁺ (ε = 4.0 × 10² M⁻¹ cm⁻¹). The ε value of Br₂Fc⁺ was determined by the electron-transfer oxidation of Br₂Fc with cerium(IV) ammonium nitrate (5.0 × 10⁻³ M) in CF₃CH₂OH/MeCN (v/v = 1:1) at 273 K. Likewise, electron transfer from [Ru^II(bpy)₃]²⁺ to [Mn^IV(O)]²⁺– (Sc³⁺)₁ (5.0 × 10⁻⁴ M) and [Ru^II(5-Cl-phen)₃]²⁺ to [Mn^IV(O)]²⁺– (Sc³⁺)₂ (5.0 × 10⁻⁴ M) were examined from the spectral change in the various concentration of [Ru^II(bpy)₃]²⁺ (1.5 × 10⁻⁴ to 5.0 × 10⁻² M) and [Ru^II(5-Cl-phen)₃]²⁺ at 273 K.

Kinetic Measurements

Kinetic measurements were performed in CF₃CH₂OH/CH₃CN (v/v = 1:1) at 273 K. Rates of electron transfer from ferrocene derivatives to [Mn^IV(O)]²⁺ (2.5 × 10⁻⁴ M) were investigated by the formation and decay of absorption bands due to ferrocenium ion and [Mn^IV(O)]²⁺, respectively. Kinetic measurements for rates of electron transfer from various electron donors to [Mn^IV(O)]²⁺–(Sc³⁺)₁ and [Mn^IV(O)]²⁺–(Sc³⁺)₂ were carried out under second-order conditions, where both concentrations of electron donors and [Mn^IV(O)]²⁺–(Sc³⁺)₁ or [Mn^IV(O)]²⁺–(Sc³⁺)₂ were 1.25 × 10⁻⁴ M, because electron transfer rates were too fast to follow under pseudo-order conditions even with use of a stopped-flow equipment.

Instrumentation

UV-vis spectra were recorded on a Hewlett-Packard 8453 diode array spectrophotometer equipped with a UNISOKU Scientific Instruments Cryostat USP-203A for low-temperature experiments or on a UNISOKU RSP-601 stopped-flow spectrometer equipped with a MOS-type highly sensitive photodiodearray. X-band EPR spectra were recorded at 5 K using X-band Bruker EMX-plus spectrometer equipped with a dual mode cavity (ER 4116DM). Low temperature was achieved and controlled with an Oxford Instruments ESR900 liquid He quartz cryostat with an Oxford Instruments ITC503 temperature and gas flow controller. The experimental parameters for EPR measurement were as follows: microwave frequency = 9.646 GHz, microwave power = 1.0 mW, modulation amplitude = 10 G, gain = 1.0 × 10⁴, modulation frequency = 100 kHz, time constant = 40.96 ms, and conversion time = 81.00 ms.