Spectroscopic characterization and reactivity studies of a mononuclear nonheme Mn(III)-hydroperoxo complex.

We report the first example of a mononuclear nonheme manganese(III)-hydroperoxo complex derived from protonation of an isolated manganese(III)-peroxo complex bearing an N-tetramethylated cyclam (TMC) ligand, [Mn(III)(TMC)(OOH)](2+). The Mn(III)-hydroperoxo intermediate is characterized with various spectroscopic methods as well as with density functional theory (DFT) calculations, showing the binding of a hydroperoxide ligand in an end-on fashion. The Mn(III)-hydroperoxo species is a competent oxidant in oxygen atom transfer (OAT) reactions, such as the oxidation of sulfides. The electrophilic character of the Mn(III)-hydroperoxo complex is demonstrated unambiguously in the sulfoxidation of para-substituted thioanisoles.

Manganese(III)–peroxo (MnIII–O2) and manganese(III)–hydroperoxo (MnIII–OOH) species have been invoked as key intermediates in the activation and formation of dioxygen (O2) and the detoxification of reactive oxygen species by enzymes and their biomimetic compounds.[1,2] For example, manganese superoxide dismutase (Mn-SOD) catalyzes the disproportionation of superoxide to dioxygen and hydrogen peroxide.[1] At higher superoxide concentrations, the activity of Mn-SOD diminishes owing to the formation of a product-inhibited complex; the structure of this intermediate has been proposed to be a side-on MnIII–peroxo or an end-on MnIII–hydroperoxo species.[3,4] In water oxidation by the oxygen evolving complex (OEC) in Photosystem II (PSII), the final short-lived state (S4) accompanies an O–O bond forming event to generate a putative MnIII–OO(H)–CaII species, although no structural or spectroscopic information on the key intermediate has been obtained to date.[2]

In biomimetic studies, the first crystal structure of a MnIII–peroxo porphyrin complex was reported by Valentine et al. in 1987.[5a] Since then, a number of MnIII–peroxo complexes bearing nonheme ligands have been characterized structurally and/or spectroscopically and their reactivities have been well demonstrated in nucleophilic reactions.[5−8] In addition to the MnIII–peroxo species, Kovacs et al. reported very recently the first example of structurally characterized MnIII–alkylperoxo species and their reactivities in the alkylperoxo O–O bond cleavage.[9] However, to the best of our knowledge, no clear spectroscopic and/or structural data for MnIII–hydroperoxo species have been obtained yet in heme and nonheme systems.

Herein we report the first example of a mononuclear nonheme MnIII–hydroperoxo species derived from protonation of an isolated MnIII–peroxo complex bearing an N-tetramethylated cyclam (TMC) ligand (Scheme 1). Spectroscopic data, including electronic absorption (UV–vis), resonance Raman (rRaman), electron paramagnetic resonance (EPR), and X-ray absorption spectroscopy/extended X-ray absorption fine structure (XAS/EXAFS), of the MnIII–hydroperoxo species are reported. The MnIII–hydroperoxo complex shows high reactivity with electrophilic character in sulfoxidation reactions (Scheme 1).

Scheme 1

Addition of 3 equiv of HClO4 to a solution of [MnIII(TMC)(O2)]+ (1; TMC = 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane) in CH3CN at −40 °C under an Ar atmosphere resulted in formation of a yellow intermediate (2) with electronic absorption bands at λmax/nm (εM/cm–1 M–1) = 324 (6400) and 384 (8000) immediately (Figure 1a). 2 was highly stable at −40 °C under an Ar atmosphere (t1/2 ∼1 day) but decayed at a fast rate at room temperature (rt). The intermediate 2 reverted back to 1 upon addition of 3 equiv of triethylamine (Supporting Information (SI), Figure S1), suggesting that 1 and 2 are interconvertible through reactions similar to the previously reported acid–base chemistry between nonheme iron(III)–peroxo and iron(III)–hydroperoxo species.[10] Attempts to isolate 2 as crystalline material were not successful; therefore, the intermediate 2 was characterized using various spectroscopic techniques, including rRaman, EPR, XAS/EXAFS, and DFT.

Figure 1

(a) UV–vis spectra of 1 (red line) and 2 (black line) in CH3CN (0.13 mM) at −40 °C. (b) Resonance Raman spectra of 2 (black line) in CH3CN and 18O-labeled 2 (red line) in CD3CN at −40 °C. # and * denote solvent peaks from CH3CN and CD3CN, respectively.

The rRaman spectrum of 2, obtained upon 407 nm excitation in CH3CN at −40 °C, exhibits an isotopically sensitive band at 792 cm–1 (Figure 1b). The peak shifted to 748 cm–1 when 2 was prepared with [MnIII(TMC)(18O2)]+ (1-18O) (SI, Experimental Section). The observed shift of 16,18Δ = 44 cm–1 is in good agreement with the calculated value of 45 cm–1 for a harmonic O–O oscillator. The isotopically sensitive bands disappeared upon warming the samples to rt, indicating that these peaks are not derived from solvents. It is noteworthy that the O–O vibrational frequency of 2 (792 cm–1) is lower than those for MnIII–O2 (∼900 cm–1) and MnIII–O2R (∼880 cm–1)[6a,7a,8c,11] species, but falls in the range of the spectroscopically characterized mononuclear FeIII–hydroperoxo species (780–840 cm–1).[12] Similar to 1,[7c]2 was EPR silent.[13] The spin state of 2 was determined to be S = 2 (μeff = 5.1 μB) by 1H NMR spectroscopic analysis using Evans’ method.[14]

XAS analysis, including XANES and EXAFS, clearly supports the formulation of 2 as a [MnIII(TMC)(OOH)]2+ dication species. The normalized Mn K-edge XAS data for 1 and 2 are compared to those of [MnII(TMC)]2+ in Figure 2. The data show a shift in the rising-edge position of 1 and 2 to higher energy by over 2 eV (2.4 eV for 1 and 2.2 eV for 2), clearly indicating that the oxidation states of both Mn ions in 1 and 2 are +3. The local structure of these two MnIII species was investigated using EXAFS. A comparison of the EXAFS and Fourier transform data for 1 and 2 is shown in Figure 3a. The comparison shows that, on going from 1 to 2, the intensity of the first shell peak drops significantly and is slightly shifted to a higher R′ value. Redistribution of the outer shell single and multiple scattering is also observed, but the change is smaller compared to that in the first shell.

Figure 2

Normalized Mn K-edge XAS spectra for 1 (blue line) and 2 (red line) compared to the starting [MnII(TMC)]2+ species (black line). The arrows point to the rising-edge at half maxima (0.5 normalized intensity).

Figure 3

(a) A comparison of the non phase-shift corrected Fourier transforms and the corresponding Mn K-edge EXAFS data (inset) for 1 (blue line) and 2 (red line). FEFF best fits to 1 and 2 together with original data (blue and red lines) are shown in (b) and (c) as a black line, respectively. Fits were performed over k = 2–12 Å–1 range.

FEFF fits to the non phase-shift corrected Fourier transform of 1 and the corresponding EXAFS data are shown in Figure 3b. Least-squares fit results show that the data are most consistent with a first shell composed of two Mn–O at 1.88 Å and four Mn–N components at 2.24 Å (SI, Table S2). The second and third shells are fit with single and multiple-scattering components from the TMC ligand. The number of independent parameters used for the fit was 11 (allowed independent parameters using Stern’s rule: 2δkδr/π + 2 = 14). Together, the edge and EXAFS data are consistent with a side-on bound MnIII–O22– description for 1. The least-squares EXAFS fit results for 2 are presented in Figure 3c. In 2, the first shell is fit with one Mn–O at 1.93 Å and four Mn–N at 2.17 Å. The outer shells are fit with single and multiple scattering contributions from the TMC ligand (SI, Table S2). These results reveal that 2 is an end-on bound MnIII–OOH species.

DFT calculations were also performed on the peroxo monocation 1 (SI, Tables S3–S5) and hydroperoxo dication 2 (SI, Tables S6–S8) species in the S = 1 and 2 spin states. Total energy comparisons show that both 1 and 2 energetically prefer the S = 2 state by about 20 kcal/mol (SI, Tables S3 and S6). The structures with the four methyl groups syn to the O2(H) group(s) are preferred, as they are both energetically favorable and in good agreement with the EXAFS data (see SI, Tables S5 and S8). Numerical frequency calculations yielded an O–O stretching frequency of 934 and 857 cm–1 for 1 and 2, respectively, which is in good agreement with the rRaman data indicating the weakening of the O–O bond upon protonation.

Since we have shown recently that a high-spin iron(III)–hydroperoxo complex, [FeIII(TMC)(OOH)]2+, is an active oxidant in both oxygen atom transfer (OAT) and hydrogen atom transfer (HAT) reactions,[15] we tested the manganese analog, 2, in those reactions. Unlike the reactivity of the [FeIII(TMC)(OOH)]2+ complex in HAT reactions, 2 did not react with cyclohexadiene in CH3CN at −40 °C, indicating that 2 is not capable of activating C–H bonds. However, 2 disappeared immediately upon addition of 4-methoxythioanisole (40 equiv) in CH3CN at −40 °C, suggesting its activity in OAT reaction. Therefore, the reactivity of 2 in sulfoxidation reactions was investigated using a stopped-flow spectrometer. The first-order rate constant (kobs) was determined to be 1.6 × 102 s–1 at −40 °C in the reaction of 4-methoxythioanisole (Figure 4a), and the first-order rate constants increased proportionally with an increase of the substrate concentration to give a second-order rate constant (k2) of 2.9 × 103 M–1 s–1 at −40 °C (SI, Figure S2a). The reactivity of 2 was further investigated with para-substituted thioanisoles, para-X–Ph–SCH3 (X = OCH3, CH3, F, H, Br), to investigate the electronic effect of para-substituents on the oxidation of thioanisoles by 2 (Figure 4b). A Hammett plot of the second-order rate constants vs σp+ gave a ρ value of −5.0 (Figure 4b, left panel; SI, Table S1), confirming the electrophilic character of the hydroperoxo group in 2.[16] In addition, we observed a good linear correlation when the rates were plotted against oxidation potentials (Eox) of thioanisoles (Figure 4b, right panel; SI, Table S1). The negative slope of −17 for 2 is significantly larger than that obtained in the sulfoxidation reaction by FeIII–hydroperoxo (slope = −2.5).[15c] Based on the previous reports in high-valent metal–oxo species,[16] we propose that the oxidation of sulfides by 2 proceeds via electron transfer, followed by oxygen atom transfer, although 2 is a metal–hydroperoxo species.

Figure 4

(a) UV–vis spectral changes, monitored by stopped-flow technique, showing the decay of [MnIII(TMC)(OOH)]2+ (2, 0.060 mM) upon mixing of 4-methoxythioanisole (40 equiv; 2.4 mM) in CH3CN at −40 °C. Inset shows the time course of the absorbance change of 2 (black line) and the corresponding pseudo-first-order fit (red line) at 384 nm. (b) Hammett plot of log k2 against σp+ of para-X–Ph–SCH3 (left panel) and plot of log k2 against Eox of para-X–Ph–SCH3 (right panel).

Organic product analysis of the reaction solution by HPLC revealed that methyl phenyl sulfoxide was produced in good yield (>60% based on the amount of 2 used). In addition, by analyzing the reaction solution with EPR and electrospray ionization mass spectrometry (ESI MS), we found that a MnII species was formed as a major product in the sulfoxidation reaction by 2 (SI, Figures S3 and S4). Although the exact mechanism of the formation of the MnII species is not clear, we propose that the Mn product formed in this reaction (e.g., [(TMC)MnIII(OH)]2+) is not stable and decomposes to the MnII species under reaction conditions, similar to that observed in the sulfoxidation reaction by Fe(III)–hydroperoxo species.[15c] DFT calculations have been performed for this reaction, modeled by a direct attack of 2 on a thioanisole substrate (SI, Tables S9–S11). The calculations show a heterolytic cleavage ([MnIIIO]+ + OH+) over a barrier of 16.5 kcal/mol, followed by S–OH bond formation and a simultaneous barrierless H+ transfer back to form [MnIIIOH]2+ (see SI, Figure S5). The heterolytic cleavage is shown by the Mulliken spin density distribution of 0.07 on the leaving OH+ group (see SI, Table S10). This is analogous to the earlier reported Fe case, where it has been shown that a singly occupied π* orbital (as in 2) renders a heterolytic cleavage.[15c] Detailed studies on the mechanism of the sulfoxidation by 2 and the reactivity comparison of 2 and [FeIII(TMC)(OOH)]2+ are underway in this laboratory.

In summary, we have reported the first clear spectroscopic and reactivity properties of a mononuclear nonheme MnIII–hydroperoxo species. The MnIII–hydroperoxo complex has been successfully generated by direct protonation of the parent MnIII–peroxo species. The successive interconversion of hydroperoxo to peroxo by the addition of a base further supports their conjugated acid–base nature. The most striking outcomes from this simple protonation reaction of MnIII–peroxo species are as follows: First, the electronic transition is significantly perturbed (showing a color change from green to yellow).[17] This optical change will be further investigated by using spectroscopic and computational methods to understand the general property of the mononuclear MnIII–hydroperoxo system in detail. Second, unlike the MnIII–peroxo in 1, the MnIII–hydroperoxo in 2 performs an electrophilic sulfoxidation reaction, which might be the consequence of electronic regulation by the proton as a Lewis acid. It has been reported recently that metal–oxidant adducts (M–OX where X = IAr, OH, OR, and halides) are capable of oxidizing substrates prior to the O–X bond cleavage to form their high-valent metal–oxo species.[18] With the MnIII–peroxo (1) and MnIII–hydroperoxo (2) complexes bearing a common supporting ligand, our future studies will focus on the structural and spectroscopic comparison as well as the reactivity comparison in substrate oxidation and electron transfer reactions, which may provide clues to understand the structures and functional roles of the proposed MnIII–peroxo and MnIII–hydroperoxo intermediates in the reactions of nonheme Mn-containing enzymes.[1−4]