A Balancing Act: Stability versus Reactivity of Mn(O) Complexes

CONSPECTUS

A large class of heme and nonheme metalloenzymes utilize O₂ or its derivatives (e.g. H₂O₂) to generate high-valent metal-oxo intermediates for performing challenging and selective oxidations. Due to their reactive nature, these intermediates are often short-lived and very difficult to characterize. Synthetic chemists have sought to prepare analogous metal-oxo complexes with ligands that impart enough stability to allow for their characterization and an examination of their inherent reactivity. The challenge in designing these molecules is to achieve a balance between their stability, which should allow for their in situ characterization or isolation, and their reactivity, in which they can still participate in interesting chemical transformations. This review focuses on our recent efforts to generate and stabilize high-valent manganese-oxo porphyrinoid complexes, and tune their reactivity in the oxidation of organic substrates.

Dioxygen can be used to generate a high-valent Mn^V(O) corrolazine (Mn^VO(TBP₈Cz)) by irradiation of Mn^III(TBP₈Cz) with visible light in the presence of a C–H substrate. Quantitative formation of the Mn^V(O) complex occurs with concomitant selective hydroxylation of the benzylic substrate hexamethylbenzene. Addition of a strong H⁺ donor converted this light/O₂/substrate reaction from a stoichiometric to a catalytic process with modest turnovers. The addition of H⁺ likely activates a transient Mn^V(O) complex to achieve turnover, whereas in the absence of H⁺, the Mn^V(O) complex was an unreactive, “dead-end” complex. Addition of anionic donors to the Mn^V(O) complex also leads to enhanced reactivity, with a large increase in the rate of 2-electron oxygen-atom-transfer (OAT) to thioether substrates. Spectroscopic characterization (Mn K-edge X-ray absorption and resonance Raman spectroscopies) revealed that the anionic donors (X⁻) bind to the Mn^V ion to form six-coordinate [Mn^V(O)(X)]⁻ complexes. An unusual “V-shaped” Hammett plot for the oxidation of para-substituted thioanisole derivatives suggested that six-coordinate [Mn^V(O)(X)]⁻ complexes can act as both electrophiles or nucleophiles, depending on the nature of the substrate. Oxidation of the Mn^V(O) corrolazine resulted in the in situ generation of an Mn^V(O) π-radical cation complex, [Mn^V(O)(TBP₈Cz^•+)]⁺, which exhibited more than a 100-fold rate increase in the oxidation of thioethers. The addition of Lewis acids (LA: Zn^II, B(C₆F₅)₃) to the closed-shell, diamagnetic Mn^V(O)(TBP₈Cz) stabilized a paramagnetic valence tautomer Mn^IV(O)(TBP₈Cz^•+):LA, which was characterized as a second π-radical cation complex by NMR, EPR, UV-vis, and high resolution CSI-MS. The Mn^IV(O)(TBP₈Cz^•+):LA complexes are able to abstract H• from phenols and exhibit a rate enhancement of up to ∼100-fold over the parent Mn^V(O) valence tautomer. In contrast, a large decrease in rate is observed for OAT for the Mn^IV(O)(TBP₈Cz^•+):LA complexes. The rate enhancement for HAT may derive from the higher redox potential for the π-radical cation complex, while the large rate decrease seen for OAT may come from a decrease in electrophilicity for an Mn^IV(O) versus Mn^V(O) complex.

Graphical Abstract

1. INTRODUCTION

High-valent metal-oxo species play important roles in the functioning of both heme and non-heme metalloenzymes. An iron(IV)-oxo porphyrin π-radical cation is the intermediate that carries out the oxidation reactions for heme enzymes including Cytochrome P450, peroxidase, and catalase.^1–3 This species, labeled Compound I (Cpd-I) in the heme literature, is capable of performing a range of selective and challenging oxidations, from the hydroxylation of strong C– H bonds, to the epoxidation of alkenes, to the sulfoxidation of thioether substrates.^2–4 Non-heme iron enzymes rely on a similar ferryl intermediate, although the overall oxidation state (Fe^IV(O)) is lower by one unit because there is no porphyrin ring available for storing an extra positive charge.⁵ Manganese porphyrins have been examined as surrogates for iron hemes in heme proteins including P450^6,7 and myoglobin.^8,9 Additionally, interest in Mn-oxo chemistry has seen a large resurgence because of the possible roles of Mn(O) species in water oxidation carried out by Photosystem II.^10,11

Synthetic chemists have put much effort into the synthesis and study of high-valent metal-oxo complexes with two main objectives: 1) to determine the spectroscopic signatures and fundamental reactivity patterns of these species so this information can be compared to the biological systems and used to define plausible enzymatic mechanistic scenarios and 2) to construct bioinspired oxidation catalysts that take advantage of the reactivity/selectivity properties of M(O) species. The obvious challenge to studying high-valent Fe(O) and Mn(O) complexes of biological relevance is the lack of stability of these species. The goal for the synthetic chemist has been to devise ligands which can provide enough stability to M(O) species to allow for their spectroscopic characterization and/or isolation, and also allow for rational modification of the ligand environment to assess structure/function relationships.

Metalloenzymes such as P450 are capable of generating high-valent M(O) species from dioxygen. In contrast, synthetic methods typically require high-energy oxidants (e.g. iodosylbenzene) to prepare M(O) species that can be spectroscopically characterized or isolated. An ongoing challenge is thus to utilize O₂ to prepare identifiable, yet reactive metal-oxo complexes that can oxidize organic substrates.

The contraction of the porphyrin core has led to the development of corroles and corrolazines, and these porphyrinoid compounds have enjoyed significant success in stabilizing high-valent metal-oxo and related (e.g. metal-imide, metal-nitride) species.^12,13 Our laboratory has focused on the synthesis and reactivity of the corrolazine (Cz) scaffold, which has provided access to high-valent Fe-oxo, Mn-oxo, Mn-imido complexes, as well as other (Co, Cu, V) high oxidation state complexes.^14,15 The advantages in stability of M(O) complexes brought about by the Cz scaffold have allowed us to test fundamental structure/function paradigms, including those that involve axial ligand effects and the influence of the redox-active ability of the porphyrinoid framework. We have also discovered a method for utilizing O₂ to prepare a high-valent Mn-oxo complex, and to perform catalytic oxidations under certain conditions. In this review we will describe our recent results on high-valent manganese-oxo corrolazines.

2. Mn^III(TBP₈Cz) and Mn^V(O)(TBP₈Cz)

Our group reported the first synthesis of corrolazine in 2001. The starting material is a tetraazaporphyrin (or porphyrazine), which can be prepared from available nitrile derivatives. The porphyrin is ring-contracted by treatment with PBr₃ to give a phosphorous corrolazine with a P^V(O), P^V(OH), or P^V(OR)₂ group in the internal cavity. A difficult step is removal of phosphorus to give metal-free Cz, which requires Na/NH₃(l) reduction at –78 °C. Details of the synthesis were reviewed previously.^14,15 The structural relationships between the different porphyrinoid members can be seen in Figure 1. The core size of the corrolazine (N_pyrrole-N_pyrrole(trans) = 3.54 Å) is the smallest of the porphyrinoid ligands depicted, and when fully deprotonated, the corrolazine is trianionic with an 18 π-electron aromatic core.^16,17

Figure 1.

Core structures of porphyrinoid ligands.

The manganese corrolazine, Mn^III(TBP₈Cz) (TBP₈Cz = octakis(p-tert-butylphenyl)corrolazinato³⁻) is easily prepared by addition of Mn(acac)₃ to metal-free TBP₈CzH_3. X-ray diffraction studies of Mn^III(TBP₈Cz) have revealed two different axial ligands bound to the Mn ion, MeOH or H₂O, depending on crystallization conditions.^18,19 Hydrogen bonds are evident between these axial donors and nearby meso-N atoms in the crystal lattice. The high-valent metal oxo complex, Mn^V(O)(TBP₈Cz), was first prepared from the oxidation of Mn^III(TBP₈Cz) by iodosylbenzene in CH₂Cl₂. We designed the corrolazine ligand to stabilize high-valent states, and the Mn^V(O) complex was not only stable in solution under ambient conditions, but was also amenable to purification by chromatography and isolation as a solid. As seen in Figure 2, the Mn^III and Mn^V(O) complexes have formed the basis of a wide range of derivatives obtained from the addition of axial ligands, protons, oxidants and Lewis acids. These species are easily distinguishable by their spectroscopic features, and exhibit novel and varied patterns of reactivity. This review discusses these findings.

Figure 2.

Transformations of manganese corrolazines. (Ar = p-tert-butylphenyl)

3. OXYGEN ACTIVATION

A long-standing goal in model complex chemistry and in the pursuit of oxidation catalysts has been to generate high-valent metal-oxo species from dioxygen. It was shown that diiron µ-oxo porphyrins derived from O₂ are susceptible to photocleavage to generate postulated Fe^IV(O)(porph) intermediates involved in catalytic oxidations.²⁰ Chromium(III) corroles react with O₂ to give Cr^V(O) complexes and catalytically oxidize PPh₃.²¹ There are still few Fe or Mn complexes that can activate O₂ in the absence of co-reductants to give M(O) species and oxidize organic substrates. During our work with Mn corrolazines, we observed that aerobic solutions of Mn^III(TBP₈Cz) would, on occasion, slowly convert to the bright green typical of Mn^V(O)(TBP₈Cz). This transformation was difficult to reproduce, but careful examination of reaction conditions led to the finding that aerobic solutions of Mn^III(TBP₈Cz) required irradiation by visible light (>400 nm) to produce Mn^V(O)(TBP₈Cz) (Scheme 1).²² This reaction was also dependent on the nature of the solvent. Confirmation that O₂ was the source of the terminal oxo ligand was obtained by isotope labeling with ¹⁸O₂. In contrast, addition of H₂¹⁸O did not give any labeled product, ruling out H₂O as an oxygen source. The participation of singlet O₂ was ruled out by the use of a ¹O₂ trap.²²

Scheme 1.

Production of Mn^V(O)(TBP₈Cz) from Mn^III(TBP₈Cz), O₂, and Visible Light. Adapted from refs 22, 24. Copyright 2013 American Chemical Society.

From kinetic data in cyclohexane we speculated that the mechanism of O₂ activation involved autoxidation of the solvent.²² A proposed mechanism involved a photochemically activated Mn^III complex reacting with O₂ to form an Mn^IV-superoxo species, which then abstracts an H atom from solvent to generate solvent radicals that can then propagate in a solvent-assisted autoxidation. The Mn^V(O) complex could arise from homolytic cleavage of Mn^IV(OOH), or other pathways.

The solvent dependence of the formation of the Mn^V(O) complex suggested to us that running the reaction in an inert solvent might allow us to control the oxidation of exogenous C–H substrates. We tested the inert solvent benzonitrile (benzene: BDFE (C-H gas phase) = 104.7 kcal mol⁻¹)²³ in the light-driven aerobic oxidation of Mn^III(TBP₈Cz), and no reaction was observed.²⁴ However, the addition of a series of toluene derivatives [Ph(CH₃)_n; n = 1 – 6] to the PhCN reaction as proton/electron sources led to production of Mn^V(O)(TBP₈Cz). Analysis of the reaction mixture, with hexamethylbenzene (HMB) (BDFE (C–H) = 83.2 kcal mol⁻¹)²³ as substrate, revealed the major product was pentamethylbenzyl alcohol (87% yield) with a minor amount of pentamethylbenzaldehyde (8% yield) (Scheme 1). The rate of this reaction increased with an increase in the number of methyl groups on the substrate, where toluene showed the slowest rate (4.0 × 10⁻⁷ M⁻¹ s⁻¹) and hexamethylbenzene showed the fastest rate (8.0 × 10⁻⁵ M⁻¹ s⁻¹). A kinetic isotope effect (KIE) of 5.4 for toluene and 5.3 for mesitylene was determined. From these findings we proposed that HAT from the toluene derivatives is the rate-determining step.²⁴

Femtosecond laser flash photolysis was employed to characterize the photochemical transformation. Upon femtosecond laser excitation (λ_exc = 393 nm) of Mn^III(TBP₈Cz), transient absorption difference spectra revealed short-lived excited states with features at 530 and 774 nm. Monitoring these bands under O₂ versus N₂ atmosphere showed that the 774 nm species decayed more rapidly in the presence of O₂. From the transient absorption data, a mechanism was proposed involving an excited state of Mn^III(TBP₈Cz) reacting with O₂.²⁴

The proposed mechanism is shown in Scheme 2. Photoexcitation generates a short-lived ⁵T₁ excited state (530 nm), which undergoes rapid intersystem crossing to the longer-lived ⁷T₁ excited state.²⁵ The tripseptet state is identified by the 774 nm peak and reacts with O₂ to give a putative superoxo complex, Mn^IV(OO^•−)(TBP₈Cz), which can either abstract a hydrogen atom from substrate to give Mn^IV(OOH)(TBP₈Cz) and benzyl radical, or undergo back electron-transfer (ET). Once formed, the hydroperoxo complex can produce Mn^V(O)(TBP₈Cz) and the benzyl alcohol derivative.

Scheme 2.

Mechanism for the Photochemical Oxidation of Mn^III(TBP₈Cz). Adapted from ref 24. Copyright 2013 American Chemical Society.

The experiments in inert PhCN showed that the oxidation of the toluene derivatives appeared to be a promising method for the aerobic oxidation of certain C-H substrates, but the final Mn^V(O)(TBP₈Cz) complex was stable, limiting this chemistry to a stoichiometric process. Catalytic turnover was seen only with weaker C–H bond substrates (e.g. dihydroacridine, BDFE = 69 kcal/mol) or with O-atom acceptors such as PPh₃.^22,24,26 We hypothesized that the addition of strong H⁺ donors might activate Mn^V(O)(TBP₈Cz) toward oxidation of the toluene derivatives, regenerating the starting Mn^III(TBP₈Cz) and yielding catalytic turnover.

Addition of the strong proton donor [H(OEt₂)₂]⁺[B(C₆F₅)₄]⁻ (H⁺[B(C₆F₅)₄]⁻) to the oxidation of HMB with air, light (hv > 400 nm) and Mn^III(TBP₈Cz) resulted in catalytic activity (scheme 3).¹⁹ Mn^III(TBP₈Cz) converted to a new species with UV-vis bands at 446 and 728 nm, which slowly bleached over the course of the reaction. Control reactions showed that light, O₂, and Mn^III(TBP₈Cz) were all required for production of oxidized products.¹⁹

Scheme 3.

Catalytic Aerobic Oxidation. Adapted from ref 19. Copyright 2015 American Chemical Society.

Insight into the catalytically active species was obtained by examining the Mn^III complex in the presence of H⁺. Addition of 1 equiv of H⁺[B(C₆F₅)₄]⁻ to Mn^III(TBP₈Cz) resulted in a spectrum (446, 730 nm) that was nearly identical to that observed for the catalytic reaction. However, addition of a second equiv of H⁺ gave a new species with a spectral signature of 470, 763 nm. The distinct spectra for the three Mn^III species are shown in Figure 3. We were successful in crystallizing and characterizing all three of these complexes by X-ray diffraction (XRD). The neutral Mn^III complex is five-coordinate with an axial water molecule. Upon protonation with 1 H⁺, a remote site on the corrolazine ring is protonated to give [Mn^III(H₂O)(TBP₈Cz(H))][B(C₆F₅)₄]. Addition of 2 H⁺ results in the diprotonated [Mn^III(H₂O)(TBP₈Cz(H)₂)][B(C₆F₅)₄]₂ complex, in which two of the meso-nitrogen atoms are protonated (Figure 3). Dissolution of these crystalline complexes showed UV-vis spectra that were identical to in situ preparations, providing definitive characterization for the different protonation states in Figure 3.¹⁹

Figure 3.

(top) Displacement ellipsoid plots (50% probability level) for Mn^III(TBP₈Cz)(H₂O) (left), [Mn^III(H₂O)(TBP₈Cz(H))]⁺ (center) and [Mn^III(H₂O)(TBP₈Cz(H)₂)]²⁺ (right) at 110(2) K. (bottom left) UV–vis spectral changes for Mn^III(TBP₈Cz) upon addition of 1 equiv (blue line) and 2 equiv (red line) of H⁺[B(C₆F₅)₄]⁻ in CH₂Cl₂. (bottom right) Addition of H⁺[B(C₆F₅)₄]⁻ to Mn^III(TBP₈Cz). Adapted from ref 19. Copyright 2015 American Chemical Society.

The reaction of crystalline, monoprotonated [Mn^III(H₂O)(TBP₈Cz(H))][B(C₆F₅)₄] with light, O₂ and HMB under catalytic conditions led to PMBOH and PBMCHO, but only in sub-stoichiometric amounts. The formation of the valence tautomer [Mn^IV(O)(TBP₈Cz^•+)(H)]⁺ (418, 784 nm) (vide infra) was also noted by UV-vis. In contrast, the same reaction with crystalline diprotonated [Mn^III(H₂O)(TBP₈Cz(H)₂)][B(C₆F₅)₄]₂ did give catalytic turnover. It was shown by UV-vis that the diprotonated starting material rapidly converted to the monoprotonated complex in benzene during catalysis. Independent experiments showed that exogenous H₂O in C₆H₆ was a sufficient base to deprotonate the second meso-NH⁺ proton of the diprotonated Mn^III complex. The resulting monoprotonated complex that forms under catalytic conditions slowly decomposed over 5 h.¹⁹

The catalytic cycle in Scheme 4 was postulated for the catalytic aerobic oxidation of HMB. With 2 equiv of H⁺ in C₆H₆ under aerobic conditions, the monoprotonated Mn^III complex is the resting state of the catalyst. The monoprotonated complex exhibits similar photochemistry to the neutral complex, resulting in the generation of a tripseptet (⁷T₁) excited state which reacts with O₂ to form the proposed Mn^IV(O₂^•−) complex. The superoxo species then abstracts H• from HMB, leading to Mn^IV(OOH) and HMB radical, which can then recombine via O-O cleavage to give PMBOH and the Mn^V(O) complex. We proposed that the Mn^V(O) complex was activated by the excess H⁺ present under catalytic conditions to give further oxidation products and regenerate the Mn^III resting state.¹⁹

Scheme 4.

Proposed Catalytic Cycle for the Oxidation of HMB. Reproduced from ref 19. Copyright 2015 American Chemical Society.

To gain more information about the activated form of the Mn^V(O) complex in the presence of H⁺, the isolated Mn^V(O) complex was reacted with H⁺[B(C₆F₅)₄]⁻ and monitored by UV-vis. Interestingly, the reaction with acid leads to two distinct species depending upon the temperature. As shown in Figure 4, addition of H⁺ at 23 °C gives the valence tautomer Mn^IV(O)(TBP₈Cz^•+):H, in which electron transfer has occurred from the Cz ring to the metal ion and suggests that protonation occurs at the terminal oxo position. In contrast, the same reaction at −60 °C resulted in a new spectrum with peaks at 436 and 650 nm, which was reversible upon addition of proton sponge. NMR analysis, including deuterium exchange and two-dimensional NOESY and COSY spectra, identified the site of protonation as one of the meso-N atoms (Figure 4), similar to what was seen for the Mn^III complex (Figure 3). These experiments showed that there are two possible sites of protonation on the Mn(O) complex. The presence of two equiv of H⁺ under catalytic conditions may activate the Mn(O) complex through a combination of the potential proton binding sites shown in Figure 4.¹⁹

Figure 4.

Scheme showing the addition of H⁺ to Mn^V(O)(TBP₈Cz) (left). UV–vis spectra of Mn^V(O)(TBP₈Cz) + H⁺[B(C₆F₅)₄]⁻ (blue line) to give [Mn^IV(O)(TBP₈Cz^•+)(H)]⁺ (red line) in CH₂Cl₂ at 23 °C (top right). UV-vis spectral changes (0 – 1 min) for the addition of H⁺[B(C₆F₅)₄]⁻ to Mn^V(O)(TBP₈Cz) (blue line) to give [Mn^V(O)(TBP₈Cz(H))]⁺ (red line) in CH₂Cl₂ at –60 °C (bottom right). Adapted from ref 19. Copyright 2015 American Chemical Society.

Our work on O₂ chemistry showed for the first time that an Mn^V(O) complex could be synthesized from an Mn^III precursor and air, visible light, and an H-atom donor. We also determined that Mn^III(TBP₈Cz) could function as a pre-catalyst for the light-driven, aerobic oxidation of the toluene derivative HMB in the presence of a strong proton donor.

4. INFLUENCE OF AXIAL LIGANDS ON REACTIVITY

In an early study on axial ligands, we showed that the addition of anionic donors (F⁻, CN⁻) to Mn^V(O)(TBP₈Cz) resulted in large rate enhancements (760 – 16,000-fold) for HAT reactions with C-H substrates.^15,27 It was hypothesized that coordination of these donors led to an increase in driving force for HAT and resulted in greatly enhanced reactivity. This work has been reviewed elsewhere.¹⁵ Kinetic analysis suggested a pre-equilibrium binding of the donor prior to the rate-determining step. A binding constant for F⁻ of K = (163 ± 7) M⁻¹ was obtained.²⁷ Mass spectral evidence was provided for [Mn^V(O)(TBP₈Cz)(X)]⁻ (X = F⁻, CN⁻), but the instability of these species did not allow for characterization by XRD.

The six-coordinate [Mn^V(O)(TBP₈Cz)(CN)]⁻ complex was characterized by X-ray absorption spectroscopy (XAS).²⁸ The decrease in the pre-edge peak is consistent with converting the Mn^V(O) complex from 5- to 6-coordinate (Figure 5). The extended X-ray absorption fine structure (EXAFS) for [Mn^V(O)(TBP₈Cz)(CN)]⁻ indicated that a sixth C/N donor at 2.21 Å was present, in accord with a coordinated cyanide ligand. Interestingly, the coordination of CN⁻ to Mn^V(O)(TBP₈Cz) did not cause a significant lengthening of the Mn–O bond.

Figure 5.

(top) Formation of six-coordinate [Mn^V(O)(TBP₈Cz)(X)]⁻ complexes. (bottom left) Normalized Mn K-edge XAS data for Mn^V(O)(TBP₈Cz) with none (red), 10 (dark blue) and 100 (light blue) equiv CN⁻. (bottom right) Resonance Raman spectra of [Mn^V(O)(TBP₈Cz)(X)]⁻ where X = no ligand (black), F⁻ (blue), N₃⁻ (red), or OCN⁻ (purple). Adapted from refs 28, 29. Copyright 2014 John Wiley & Sons, Inc. and Copyright 2014 American Chemical Society.

Resonance Raman (RR) spectra show a strong Mn–O vibrational mode at 981 cm⁻¹ for Mn^V(O)(TBP₈Cz) that downshifts 5 cm⁻¹ to 976 cm⁻¹ when the anionic donors F⁻, N₃⁻, or OCN⁻ were present (Figure 5).^29,30 All attempts to measure the RR spectra of the CN ⁻ adduct failed, due the rapid decay of the complex at higher concentrations. The assignment of the 976 cm ⁻¹ peak was confirmed by the predicted downfield shift of 40 cm ⁻¹ for the O¹⁸-labeled six-coordinate complexes with F⁻ and N₃⁻.²⁹ These data are consistent with coordination of the anionic donors trans to the oxo group.

Having previously looked at the influence of axial donors on HAT reactivity,²⁷ we sought to understand if similar effects on O-atom transfer (OAT) would be observed. Sulfoxidation was examined with the 5- and 6-coordinate Mn^V(O) complexes (Scheme 5). Smooth conversion of [Mn^V(O)(TBP₈Cz)(X)]⁻ to [Mn^III(TBP₈Cz)(X)]⁻ (X = none, CN⁻, F⁻, OCN⁻, N₃⁻, SCN⁻, NO₃⁻) was observed after addition of RSR (R = Me, n-Bu). Second-order rate constants revealed large rate enhancements for several of the anionic donors, with the largest (24,000-fold) seen for CN⁻. ²⁹ The trend in rate enhancement that was found, X = none ≈ SCN⁻ ≈ NO₃⁻ < OCN⁻ < N₃⁻ < F⁻ ≪ CN⁻, was supported by DFT calculations.²⁸ These results suggested that the stronger the binding interaction of X⁻ to Mn^V(O), the larger the rate enhancement. An Eyring analysis was conducted on [Mn^V(O)(TBP₈Cz)(CN)]⁻ plus dibutyl sulfide (DBS), and the activation parameters were ΔH^‡ = 14 ± 0.4, ΔS^‡ = −10 ± 0.8, ΔG^‡ = 17 ± 0.5 (298 K).²⁸ A comparison of the activation parameters for Mn^V(O)(TBP₈Cz) (ΔH‡ = 16 ± 1, ΔS‡ = −20 ± 1, ΔG‡ = 22 ± 2)³¹ showed that addition of CN⁻ lowered the overall barrier (ΔG^‡) through both enthalpic and entropic contributions.

Scheme 5.

OAT to Thioether Substrates. Adapted from ref 29. Copyright 2014 American Chemical Society.

Further mechanistic information was obtained by examining the OAT reactivity of the six-coordinate [Mn^V(O)(TBP₈Cz)(CN)]⁻ and [Mn^V(O)(TBP₈Cz)(F)]⁻ complexes with a series of para-substituted thioanisole derivatives (X-p-C₆H₄SCH₃; X = OMe, CH₃, H, Br, C(O)OMe, CN, NO₂). As with the alkyl sulfides, smooth conversion to Mn^III products were seen by UV-vis, and good yield of sulfoxide obtained. Labeling of the terminal oxo ligand with ¹⁸O led to 71% isotopic incorporation in the sulfoxide product.²⁸ A Hammett analysis of the thioanisole derivatives exhibited a negative slope for the electron-donating substituents, with ρ = − 1.39 for CN ⁻ and ρ = − 2.29 for F⁻. However, when electron-withdrawing substituents were tested, a positive linear slope was observed with ρ = 1.22 for CN ⁻ and ρ = 1.78 for F⁻, resulting in unusual “V-shaped” Hammett plots (Figure 6).²⁹ A V-shaped Hammett plot strongly suggested a fundamental change in the mechanism of OAT, leading to two proposed pathways (Figure 6). Pathway A depicts the typical electrophilic mechanism, in which the high-valent Mn^V(O) complex is the electrophile. This pathway accounts for the part of the Hammett plot with a negative slope and is the anticipated mechanism for OAT. The change in mechanism can be rationalized by pathway B in Figure 6. In this pathway, a quinoid-type resonance form of the thioanisole derivative is invoked, which is stabilized by strong electron-withdrawing substituents. The oxo ligand functions as a weak nucleophile in this case, and is attracted to the partial positive charge on the sulfur center. This “umpolung” reactivity provides the mechanistic switch. A quinoid-type X-ray structure was observed for a phenylthiolate-nickel(II) complex, providing confidence in the proposed mechanism in pathway B.³²

Figure 6.

(left) Hammett plots for [Mn^V(O)(TBP₈Cz)(X)]− (X = CN⁻ or F⁻) and para-X-substituted thioanisole derivatives. (right) Mechanistic pathways for electron-donating (Pathway A) and electron-withdrawing (Pathway B) para-substituted thioanisole derivatives. Adapted from ref 29. Copyright 2014 American Chemical Society.

The V-shaped Hammett plot in Figure 6 can be contrasted to the reactivity of the one-electron oxidized, cationic Mn^V(O)(TBP₈Cz^•+) complex (vide infra). Reaction of this complex with the same thioanisole derivatives resulted in a linear Hammett plot for all of the thioanisole derivatives, with ρ = −1.40. These data indicated that only an electrophilic mechanism (pathway A) is operative for the cationic Mn-oxo complex, and suggest that the positive charge on the metal complex removes any nucleophilic character. ²⁹ Together these results show that the six-coordinate [Mn^V(O)(TBP₈Cz)(X)] complexes are more reactive than the five-coordinate Mn^V(O)(TBP₈Cz) in OAT reactions, and that the anionic, axially ligated Mn-oxo complexes can exhibit both electrophilic and nucleophilic character.²⁹

5. INFLUENCE OF ONE-ELECTRON OXIDATION ON THE OAT REACTIVITY

Addition of the one-electron oxidants [(BrC6H₄)₃N^•+](SbCl₆⁻) or cerium(IV) ammonium nitrate (CAN) to Mn^V(O)(TBP₈Cz) led to the generation of the one-electron oxidized π-radical-cation complex [Mn^V(O)(TBP₈Cz^•+)]⁺. This complex could not be isolated in the solid state, but was characterized spectroscopically in situ. Oxidation of Mn^V(O)(TBP₈Cz) was accompanied by an immediate color change from bright green to orange-brown, and a new spectrum with peaks at 410 and 780 nm appeared (Figure 7). The broadened Soret band, and low-intensity, long-wavelength peak in the near-IR are characteristic of porphyrin π-radical cations, in which an electron has been removed from the aromatic π system. Electron paramagnetic resonance (EPR) spectroscopy revealed a sharp singlet at g = 2.001, consistent with the Cz ring as the locus of oxidation and the metal remaining in an Mn^V oxidation state (Figure 7, inset). Spectral titrations together with previous spectroelectrochemical assignments¹⁸ supported the one-electron oxidation of the Cz ring.³¹

Figure 7.

Scheme and UV-vis spectra for the oxidation of (TBP₈Cz)Mn^V(O) (10 µM) with increasing amounts of [(BrC₆H₄)₃N^•+](SbCl₆⁻) (0–1.2 equiv) at 25 °C in CH₂Cl₂. Inset: X-band EPR spectrum of [Mn^V(O)(TBP₈Cz^•+)]⁺ (10 µM) at 15 K in CH₂Cl₂. Adapted from ref 31. Copyright 2014 American Chemical Society.

The Mn^V(O) π-radical-cation was examined for OAT reactivity. Addition of DMS to [Mn^V(O)(TBP₈Cz^•+)]⁺ generated from (BrC₆H₄)₃N^•+/Mn^V(O)(TBP₈Cz) showed isosbestic conversion to a new spectrum (440, 470, 722 nm) characteristic of [Mn^IV(TBP₈Cz)]⁺. The oxidation state of this complex was confirmed by EPR, revealing the spectrum of an S = 3/2 Mn^IV ion with well-resolved ⁵⁵Mn hyperfine coupling.³³ The OAT product DMSO was obtained in good yield (88%). For Mn^V(O)(TBP₈Cz) and [Mn^V(O)(TBP₈Cz^+•)]⁺, second order rate constants with DMS were (2.0 ± 0.2) × 10 ⁻³ M⁻¹ s⁻¹ and 0.25 ± 0.05 M⁻¹ s⁻¹, respectively, indicating a large rate enhancement (> 100-fold) for [Mn^V(O)(TBP₈Cz^•+)]⁺. An Eyring analysis gave ΔH^‡ = 16 ± 1 kcal mol ⁻¹ for the neutral Mn^V(O) complex, but a much smaller ΔH^‡ = 7 ± 0.8 kcal mol ⁻¹ for the one-electron oxidized complex. Both complexes gave negative ΔS^‡ values as expected for a bimolecular OAT mechanism, but ΔS^‡ was unusually large for [Mn^V(O)(TBP₈Cz^•+)]⁺ (−45 ± 3 cal K⁻¹ mol ⁻¹). The origin of this large negative entropic factor was not determined, but it attenuates the effect of the enthalpic change, giving a more modest rate enhancement than might be expected.³¹

DFT calculations on the potential energy profiles of sulfoxidation corroborated the trends seen from kinetics. A concerted bimolecular mechanism for both the Mn^V(O)(TBP₈Cz) and [Mn^V(O)(TBP₈Cz^•+)]⁺ was found by DFT, and a lower barrier of OAT was seen for the one-electron oxidized complex. An increase in electrophilicity for the [Mn^V(O)(TBP₈Cz^•+)]⁺ complex may lower the reaction barrier for OAT.³¹

5. LEWIS ACIDS AND VALENCE TAUTOMERISM

The addition of Lewis and Bronsted acids to Mn^V(O)(TBP₈Cz) provided a new method for tuning the reactivity and electronic structure of this species. Addition of Zn(OTf)₂, a redox-inactive Lewis acid, to the Mn^V(O) complex resulted in a UV-vis spectrum similar to that of the one-electron oxidized [Mn^V(O)(TBP₈Cz^•+)]⁺, a puzzling result.³⁴ The spectral change was fully reversible with the addition of a Zn(II) chelator. Evans method NMR measurement revealed that the new species was paramagnetic, with µ_eff = 4.11 µ_B, (S = 1 (2.83 µ_B) and S = 2 (4.90 µ_B)). The data suggested that Zn^II may be stabilizing an alternate electronic configuration, or “valence tautomer” of Mn^V(O)(TBP₈Cz). We postulated that Zn^II was binding to the oxo group and weakening the metal-oxo π-bonding, thereby favoring the Mn^IV(O)(TBP₈Cz^•+) valence tautomer.³⁴

The Mn^IV(O)(π-radical-cation):Zn^II complex exhibited electron-transfer reactivity significantly different from that seen for the Mn^V(O) starting material. The Mn^V(O)(TBP₈Cz) valence tautomer is reduced by decamethylferrocene (Cp*₂Fe, E°_red = −0.59 V vs. Fc⁺/Fc), but not by the weaker reductant ferrocene (Fc, E°_red = 0.00 V vs. Fc⁺/Fc). However, Mn^IV(O)(π-radical-cation):Zn^II is easily reduced by Fc and gives an Mn^IV(Cz⁰) product. Reduction of the Mn^V(O) complex by Cp*₂Fe leads to an Mn^III product. These data indicated that the Mn^IV(O)(TBP₈Cz^+•):Zn^II complex was a more powerful oxidant than the Mn^V(O) complex, but leads only to the one-electron reduced Mn^IV state. Reaction of Mn^IV(O)(TBP₈Cz^+•):Zn^II with hydrogen atom donors (e.g. substituted phenols) yielded similar results.³⁴

A nonmetallic Lewis acid, B(C₆F₅)₃, exhibited similar chemistry to Zn^II, with UV-vis and Evans method confirming the stabilization of the valence tautomer Mn^IV(O)(TBP₈Cz^•+):B(C₆F₅)₃).³⁵ Spectral titrations for both Lewis acids revealed similar association constants for one-to-one binding, with K_a = 2.0 × 10⁷ M⁻¹ for B(C₆F₅)₃ and K_a = 4.0 × 10⁶ M⁻¹ for Zn^II. Low-temperature, high-resolution electrospray mass spectrometry data on Mn^IV(O)(TBP₈Cz^•+):B(C₆F₅)₃ showed the formation of a 1:1 adduct. For both Zn^II and B(C₆F₅)₃, the data are most consistent with coordination of the Lewis acid at the terminal oxo ligand, however conclusive structural characterization for these species has not yet been obtained.

The influence of B(C₆F₅)₃ on HAT by Mn^IV(O)(TBP₈Cz’⁺):B(C₆F₅)₃ was greater than that of Zn^II, with an ∼100-fold increase in HAT rate with phenols (Scheme 6). A kinetic isotope effect of k_H/k_D = 3.2 ± 0.3 was measured for 2,4,6-TTBP, pointing to an HAT mechanism. The trend in reactivity is shown in Scheme 6, where the Mn^IV(O)(TBP₈Cz’⁺):Zn^II tautomer shows a moderate rate enhancement, and Mn^IV(O)(TBP₈Cz^•+):B(C₆F₅)₃ shows rate enhancements up to ∼100-fold relative to Mn^V(O). Thus the HAT reactivity of Mn^IV(O)(TBP₈Cz^•+):LA increases with the increasing strength of the Lewis acid.³⁵

Scheme 6.

Summary of HAT reactivity of Mn^IV(O)(TBP₈Cz^•+) with substituted phenols. Adapted from ref 35. Copyright 2014 American Chemical Society.

In earlier work we found that the low spin Mn^V(O) tautomer reacts rapidly with triphenylphosphine (PPh₃), an O-atom accepting reagent, and we decided that triarylphosphines were good test substrates to compare the two-electron O-atom transfer reactivity of Mn^V(O)(TBP₈Cz) with the Mn^IV(O)(π-radical cation) valence tautomer. Stopped-flow UV-vis spectroscopy was required for these fast reactions, and we obtained second-order rate constants ranging from 16 ± 1 to 1.43 ± 6 × 10⁴ M⁻¹ s⁻¹ for a series of substituted triarylphosphine derivatives reacting with Mn^V(O)(TBP₈Cz) (Figure 8).³⁶ Variation of the para substituents of the phosphine derivatives led to a linear Hammett correlation with a negative slope (ρ = − 0.91 ± 0.05), consistent with the expected mechanism where the Mn^V(O) complex is an electrophilic oxidant.

Figure 8.

Reaction of Mn^V(O)(TBP₈Cz) with a series of para-substituented phosphines. Adapted from ref 36. Copyright 2015 American Chemical Society.

There was also a strong dependence on the steric encumbrance of the PAr₃ derivatives, consistent with concerted nucleophilic attack of the PAr₃ substrate.

The reaction of PPh₃ with the paramagnetic Mn^IV(O)(TBP₈Cz^•+):LA species (LA = Zn^II, B(C₆F₅)₃, H⁺) led to a dramatic decrease in the rate of OAT, with a ratio of second-order rate constants = k_none/k_Mn(O)LA = 14,000 – 71,000. This rate inhibition strongly contrasts the increase in rate constants we observed for HAT with the π-radical cation valence tautomer. We suggested that perhaps it is the inherent electrophilicity of an Mn^IV(O) versus an Mn^V(O) unit that may be responsible for the inhibition, since the local oxidation state of +4 may lower the electrophilicity of the terminal oxo ligand compared to a local oxidation state of +5.³⁶ The lower electrophilicity could influence the two-electron phosphine oxidation reactions as proposed, and would be exerting a different influence than the increase in redox potential seen for the π-radical cation complex, which results in faster one-electron HAT. A summary of the reactivity for the two valence tautomers is shown in Scheme 7.

Scheme 7.

Reactivity of Mn(O)(Cz) Valence Tautomers. Adapted from ref 36. Copyright 2015 American Chemical Society.

6. Summary

The formation of a stable high-valent manganese(V)-oxo corrolazine has led to a wide variety of discoveries pertaining to the reactivity of metal-oxo complexes. It was shown that reacting manganese(III) corrolazine with light, air and a C–H substrate allowed for clean formation of the Mn^V(O) complex, and addition of acid to the reaction mixture allowed for catalytic turnover. The addition of anionic donors, protons, oxidants or Lewis acids to the Mn^V(O) corrolazine each affected the reactivity of the high-valent species in different ways. These new mechanistic insights contribute to our understanding of how the reactivity of high-valent metal-oxo intermediates is controlled in enzymatic systems. The knowledge gained from these studies also may suggest strategies for designing potential new synthetic oxidation catalysts.

Biographies

Heather Neu received her B.S. degree in Chemistry from the University of Wisconsin Superior in 2007, where she conducted independent undergraduate research through the McNair Scholars program under the direction of Professor Troy S. Bergstedt. In 2010, she earned her M.S. from the University of Minnesota Duluth, with Professor Victor N. Nemykin. She is currently a Ph.D. candidate at Johns Hopkins University working with Professor David P. Goldberg.

Regina Baglia received her B.S. degree in Biochemistry from Temple University in 2011, where she conducted undergraduate research through the Diamond Research Scholars program under the direction of Professor Michael J. Zdilla. She is currently a Ph.D. candidate at Johns Hopkins University working with Professor David P. Goldberg.

David Goldberg received his B.A. degree from Williams College (1989) and a Ph.D. degree from M.I.T (1995). After completing a postdoctoral fellowship at Northwestern University, he moved to Johns Hopkins University in 1998, where he is currently a Professor of Chemistry.