Light–Driven, Proton–Controlled, Catalytic Aerobic C–H Oxidation Mediated by a Mn(III) Porphyrinoid Complex

Abstract

The visible light-driven, catalytic aerobic oxidation of benzylic C-H bonds was mediated by an Mn^III corrolazine complex. To achieve catalytic turnovers, a strict selective requirement for the addition of protons was established. The resting state of the catalyst was unambiguously characterized by X–ray diffraction as [Mn^III(H₂O)(TBP₈Cz(H))]⁺, in which a single, remote site on the ligand is protonated. If two remote sites are protonated, however, reactivity with O₂ is shut down. Spectroscopic methods revealed that the related Mn^V(O) complex is also protonated at the same remote site at −60 °C, but undergoes valence tautomerization upon warming.

Graphical Abstract

The activation of molecular oxygen is a critical transformation in biology, which is mediated by both heme and nonheme metal centers.¹ Synthetic, biomimetic transition metal complexes aimed at activating O₂ have been targeted by employing both porphyrinoid and nonheme polydentate ligands, and some of these efforts have resulted in important mechanistic insights as well as catalysts for the oxygenation of a range of organic substrates.² The catalytic, aerobic oxidation of C–H bonds with first–row, biologically relevant metal complexes remains a particularly important, yet challenging goal.³ Although some Fe and Mn metalloporphyrins, and related metallomacrocycles (e.g. phthalocyanines), can catalyze the oxidation of C–H bonds with O₂, drawbacks remain in most cases. Such drawbacks include the need for stoichiometric co–reductants, and the involvement of poorly controlled radical–type pathways.^{2j, 3} The mechanisms of these systems, including the role of high–valent metal–oxo species, is often poorly understood.

Corroles are modified porphyrinoid ligands that stabilize high–valent metal–oxo species, and can exhibit catalytic activity different from their porphyrin analogs. A Cr^III corrole catalyzes the aerobic oxidation of PPh₃ via a Cr^V(O) intermediate.⁴ Under visible light irradiation, diiron(IV) μ–oxo biscorroles have been shown to catalyze the aerobic oxidation of phosphine and C–H substrates, presumably through a disproportionation pathway that involves μ–oxo–bridge cleavage and leads to formation of Fe^III/Fe^V(O) intermediates.⁵ Diiron(III)–μ–oxo–bridged porphyrins also mediate photoactivated aerobic oxidations through similar μ–oxo cleavage mechanisms.⁶ Light-driven activation of metalloporphyrinoid complexes for catalytic aerobic oxidation is especially attractive from an environmental and energy perspective.

In previous work, we showed that a stable Mn^V(O) complex prepared with a corrolazine ligand, which is a meso–N–substituted analog of a corrole, could be synthesized via photoirradiation of a Mn^III precursor, O₂ (or air), and C–H substrates.^7a This work was the first example of the synthesis of a Mn^V(O) complex from air, light, and a proton/electron source.^7a The addition of toluene derivatives to the Mn^III/hv/O₂ system led to selective C–H hydroxylation of benzylic C–H bonds.^7b However, the inherent stability of the Mn^V(O) complex toward the toluene derivatives limited this chemistry to a strictly stoichiometric process, with the Mn^V(O) complex as the final Mn product. Only in the case of PPh₃, or dihydroacridine, which has a very weak C–H bond (BDFE = 69 kcal mol⁻¹), was catalytic turnover obtained.^7a,8

Herein we report the visible light–driven, catalytic aerobic oxidation of benzylic C–H bonds via the proton–controlled activation of a Mn^III corrolazine complex. The selective addition of a strong proton donor leads to the formation of new mono– and diprotonated Mn^III complexes, which were characterized by single crystal X–ray diffraction (hereafter, XRD). The monoprotonated complex is capable of reacting with air, light and benzylic C–H bonds to give a high–valent Mn–oxo complex. Under the right conditions, the controlled addition of protons provides access to a catalytic cycle. Low–temperature methods were used to trap a novel, proton–activated form of the Mn^V(O) complex which may play a role in the catalytic cycle.

Previously it was shown that photoirradiation (λ > 400 nm) of Mn^III(TBP₈Cz)(TBP₈Cz = octakis(p-tert-butylphenyl corrolazinato³⁻) in the presence of excess O₂ and hexamethylbenzene (HMB) in benzonitrile (PhCN) led to the selective oxidation of HMB to pentamethylbenzyl alcohol (PMB–OH) (87%) with clean formation of Mn^V(O)(TBP₈Cz).^7b This Mn^V(O) complex oxidizes relatively weak C–H bonds such as those found in dihydroanthracene (BDFE = 76 kcal mol⁻¹), but is essentially unreactive toward stronger C–H bonds such as those found in toluene (BDFE = 87 kcal mol⁻¹) or its derivatives.⁸ The selective hydroxylation of HMB was therefore limited to strictly stoichiometric reactivity.^7b We speculated that addition of a strong proton donor might activate the Mn^V(O) complex toward oxidation of the stronger C–H bonds found in HMB, providing entry into a catalytic oxidation pathway.

Addition of the strong proton donor [H(OEt₂)₂]⁺[B(C₆F₅)₄]⁻ (H⁺[B(C₆F₅)₄]⁻) to Mn^III(TBP₈Cz) in benzene (C₆H₆) caused an immediate color change of the brown Mn^III complex to a brown–red solution. The addition of excess HMB (1000 equiv) under ambient conditions to this solution, followed by photoirradiation with visible light (λ > 400 nm), led to the slow bleaching of the solution over 5.5 h. Monitoring of this reaction by UV–vis confirmed the slow decay of the protonated Mn^III complex, which exhibited distinct Soret and Q–bands at 446 and 728 nm (Figure S3). Analysis of the reaction mixture by GC–FID revealed the production of PMB–OH and the corresponding aldehyde PMB–CHO (Scheme 1) with 18 turnovers for PMB–OH and 9 turnovers for PMB–CHO. Both oxidation products increase steadily over time, and catalytic activity appears only limited by catalyst stability. In the absence of light, O₂, or Mn^III complex, no oxidized products were detected. As anticipated, the addition of H⁺ to the Mn catalyst leads to the catalytic oxidation of HMB with only air and light as additives.

Scheme 1.

The independent reaction of Mn^III(TBP₈Cz) with H⁺[B(C₆F₅)₄]⁻ was investigated to determine the effect of H⁺ on the Mn^III complex. These independent experiments were initially run in CH₂Cl₂, as opposed to C₆H₆, because of the available UV-vis data on Mn corrolazine complexes in CH₂Cl₂.^9,10 The starting Mn^III complex (435, 685 nm) was converted to a new brown–red species upon addition of one equiv of H⁺[B(C₆F₅)₄]⁻ (446, 730 nm) (Figure 1). This new spectrum (446, 730 nm) was almost identical to that seen for the resting state of the catalyst in C₆H₆. Introduction of a second equiv of H⁺ caused a further change to a second species with further red–shifted Soret and Q–bands at 470 and 763 nm. Spectral titrations (Figure S4) showed that no further changes occur upon addition of up to 5 equiv of H⁺. Similar spectra were obtained in dry C₆H₆ (Figure S7, S10).

Figure 1.

UV–vis spectral changes for Mn^III(TBP₈Cz) (12 µM, black line) upon addition of 1 equiv (blue line) and 2 equiv (red line) of H⁺[B(C₆F₅)₄]⁻ in CH₂Cl₂ (2 mL) at 25 °C.

The UV–vis data in Figure 1 clearly show that there are two distinct species accessible via protonation of the starting Mn^III corrolazine. The three Mn^III complexes identified in Figure 1 were successfully crystallized and unambiguously characterized by XRD. The crystal structures of the neutral, mono– and diprotonated Mn^III derivatives are presented in Figure 2. The structure of the neutral Mn^III complex contains a 5–coordinate Mn^III ion with a water ligand located in the axial position, similar to the structure of a MeOH–bound Mn^III(MeOH)(TBP₈Cz).^9d The axial H₂O ligands of two adjacent metallocorrolazines are H–bond donors to the meso–N atoms of the neighboring Cz ligand via a set of 1+1 O—H…N hydrogen bonds (i.e., only one hydrogen bond per axial H₂O ligand), resulting in H–bonded dimers of Mn^III(H₂O) complexes in the solid state. Treatment of the starting Mn^III complex with 1 equiv of H⁺[B(C₆F₅)₄]⁻ in CH₂Cl₂, followed by removal of solvent and re-crystallization from toluene/heptane, led to dark brown crystals of [Mn^III(H₂O)(TBP₈Cz(H))][B(C₆F₅)₄]. The crystal structure reveals that monoprotonation most likely occurs at one of the meso–N atoms (N1(H)) adjacent to the direct pyrrole–pyrrole bond of the Cz ligand. The axial H₂O ligand seen in the neutral Mn^III starting complex and H–bonded dimers are retained. Dissolution of crystalline [Mn^III(H₂O)(TBP₈Cz(H))][B(C₆F₅)₄] yields the same UV–vis spectrum as seen for the in situ protonation of the Mn^III complex via addition of 1 equiv of H⁺[B(C₆F₅)₄]⁻ (λ_max = 446, 730 nm) (Figure S6). If two equiv of H⁺[B(C₆F₅)₄]⁻ are added to the neutral Mn^III complex, a new complex in the form of dark brown crystals can be isolated. XRD analysis shows that this new species is doubly protonated with the molecular formula [Mn^III(H₂O)(TBP₈Cz(H)₂)][B(C₆F₅)₄]₂ (Figure 2). Both H–atoms were unambiguously located on the opposite meso–N atoms (N1 and N5). Unlike the neutral and monoprotonated structures, this complex exhibits no H–bonding to nearest neighbors and crystallizes in isolated molecular units. The UV–vis spectrum of crystalline [Mn^III(H₂O)(TBP₈Cz(H)₂)][B(C₆F₅)₄]₂ redissolved in CH₂Cl₂ matches that seen for the in situ diprotonation of the Mn^III complex following treatment with 2 equiv of H⁺ (λ = 470, 763 nm) (Figure S7). The metrical parameters for each of the structures show little variation upon mono– and diprotonation, with the exception of a slight shortening (Δd(Mn–O) = 0.028(3) Å) of the Mn^III–OH₂ distance.

Figure 2.

(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) Addition of one and two equiv of H⁺[B(C₆F₅)₄]⁻ to Mn^III(TBP₈Cz).

The preparation of crystalline mono– and diprotonated Mn^III complexes provided pure material for testing as possible catalysts in the observed light–driven, aerobic C–H oxidation. Monoprotonated [Mn^III(H₂O)(TBP₈Cz(H))][B(C₆F₅)₄] was dissolved in CH₂Cl₂ or C₆H₆ with excess HMB (1000 equiv) under aerobic conditions. Photoirradiation of this solution for 30 min at 23 °C followed by analysis by GC–FID showed both alcohol and aldehyde products were formed, but only in stoichiometric amounts (PMB–OH: TON = 0.4; PMB–CHO: TON = 0.7), i.e. no catalysis was observed. Monitoring this reaction by UV–vis showed isosbestic conversion of the spectrum for [Mn^III(H₂O)(TBP₈Cz(H))]⁺ to a new spectrum with features at 418 and 784 nm. This new spectrum (Figure S12) matches that seen previously for the Mn^IV–oxo π–radical cation complex [Mn^IV(O)(TBP₈Cz^•+)(LA)]⁺ (LA = Lewis acid = Zn^II, B(C₆F₅)₃), a valence tautomer of Mn^V(O)(TBP₈Cz) stabilized by Lewis acids.¹⁰ In support of this analysis, reaction of crystalline [Mn^III(H₂O)(TBP₈Cz)(H)][B(C₆F₅)₄] with the O–atom donor PhIO leads to isosbestic conversion to [Mn^IV(O)(TBP₈Cz^•+)(H)]⁺ (Figure S13), instead of Mn^V(O)(TBP₈Cz), which is the major product from [Mn^III(H₂O)(TBP₈Cz)] + PhIO.^9d These results indicate that generation of the valence tautomer [Mn^IV(O)(TBP₈Cz^•+)(H)]⁺ occurs if 1 equiv of H⁺ is employed.

The use of crystalline diprotonated [Mn^III(H₂O)(TBP₈Cz(H)₂)][B(C₆F₅)₄]₂ as starting material, in contrast, leads to catalytic turnover in the light–driven oxidation of HMB, giving PMB–OH (TON = 16) and PMB–CHO (TON = 11) after 5 h in C₆H₆ at 23 °C. However, the initial spectrum for the diprotonated complex (470, 763 nm) rapidly (≤ 2 min) converts to a spectrum characteristic of the monoprotonated Mn^III species (446, 728 nm) under catalytic conditions, and no evidence for the formation of either the Mn^V(O) or Mn^IV(O)(π–radical cation)(Figure S14). The 446, 728 nm spectrum slowly bleaches over the course of the catalytic reaction. The spectrum for the diprotonated Mn^III complex can be maintained if a third equiv of H⁺ [B(C₆F₅)₄]⁻ is added prior to photoirradiation, but under these conditions no oxidation products of HMB were obtained. Taken together, these observations indicate that the presence of 2 equiv of H⁺ are required for catalytic turnover, and it is the monoprotonated Mn^III complex that is the catalytically active resting state.

Further insight regarding the role of 2 equiv of H⁺ came from examining the diprotonated Mn^III complex in “wet” C₆H₆ under aerobic conditions, versus dry C₆H₆ under inert atmosphere. It was found that residual H₂O, present under aerobic conditions, acted as a weak base and led to the deprotonation of [Mn^III(H₂O)(TBP₈Cz(H)₂)]²⁺ to give the monoprotonated complex, while under strictly anaerobic conditions (doubly-distilled C₆H₆, inert atmosphere), the diprotonated complex was observed (Figures S9–S11). Thus the monoprotonated Mn^III complex is formed under the aerobic conditions of the catalytic reaction.

The data presented herein, together with previous femtosecond transient absorption measurements,^7b lead to the proposed catalytic cycle shown in Scheme 2. The monoprotonated Mn^III complex is the resting state of the catalyst, which is activated by photoexcitation to a tripquintet state (⁵T₁) that rapidly converts to the O₂–reactive tripseptet (⁷T₁) state. Entry of O₂ gives a putative Mn^IV(superoxo) complex, which can either undergo rapid back electron–transfer to the ground state and release of O₂, or can abstract an H–atom from the HMB substrate. The resulting Mn^IV(OOH) complex can then hydroxylate the HMB radical via O–O cleavage and rebound of OH^•, to yield PMB–OH and the Mn^V(O) complex. The Mn^V(O) complex is activated in the presence of acid under the catalytic conditions, leading to oxidation of HMB (or PMB–OH) and completing the catalytic cycle.

Scheme 2.

Proposed Catalytic Cycle for the Oxidation of HMB

Attempts were made to gain insights into the final step of the catalytic cycle by reacting isolated Mn^V(O)(TBP₈Cz) with H⁺[B(C₆F₅)₄]⁻ in CH₂Cl₂ or benzene. This reaction, when run with one equiv of H⁺ at 23 °C, afforded the valence tautomer [Mn^IV(O)(TBP₈Cz^•+)(H)]⁺ (Scheme 3, Figure S15). This species was unreactive toward either excess HMB or PMBOH (1000 equiv) for up to 2 h at 23 °C, and photoirradiation of this solution caused significant decomposition (~40% bleaching). The addition of 2 equiv of H⁺[B(C₆F₅)₄]⁻ to isolated Mn^V(O)(TBP₈Cz) in the presence of light and excess HMB caused rapid (<10 s) formation of Mn^III, although organic oxidation products could not be identified. The exact nature of the catalytically active intermediate cannot be assigned, but it is clear that 2 equiv of H⁺ rapidly destabilizes the isolated Mn^V(O) complex, and these observations suggests that H⁺ may assist in activating the Mn^V(O) complex to induce turnover under the appropriate catalytic conditions.

Scheme 3.

Addition of H⁺ to Mn^V(O)(TBP₈Cz)

Further examination of the reaction of Mn^V(O)(TBP₈Cz) and H⁺[B(C₆F₅)₄]⁻ at low temperature (−60 °C) led to an isosbestic change together with the formation of a distinct, new species with λ_max = 436, 650 nm (Figure S16). This spectral change is reversible upon addition of one equiv of l,8–bis(dimethylamino)naphthalene (proton sponge), consistent with an acid–base transformation (Figure S17). The structure of the new Mn^V species (436, 650 nm) was characterized by low–temperature ID and 2D NMR spectroscopy. The ¹H NMR spectra (−50 to −60 °C) of the starting Mn^V(O) complex and the monoprotonated derivative are compared in Figure 3. Both spectra are diamagnetic, confirming that the low–spin (d²) Mn^V oxidation state is retained at −60 °C (Figure 3). The most notable difference between the two species is the appearance of a new peak at 13.45 ppm following addition of the H⁺[B(C₆F₅)₄]⁻. The chemical shift of this resonance is consistent with protonation of a meso–N atom, as seen for the Mn^III complexes, and is shifted downfield due to a ring current effect. This new peak disappeared upon exchange with D₂O, while the other peaks remained intact (Figure S21). The protonated complex also exhibits a new multiplet at 7.87 ppm, along with additional splitting in both the aromatic and ^tBu regions (Figures S20). These perturbations are consistent with a lowering of the molecular symmetry, as expected for protonation at a single meso-N position adjacent to the direct pyrrole–pyrrole bond of the Cz ligand. Definitive assignment of the 13.45 ppm peak comes from ¹H–¹H 2D nuclear Overhauser effect spectroscopy (NOESY) (Figure S22). Well–defined cross–peaks are observed between the new peak at 13.45 ppm and peaks at 8.04 and 7.70 ppm, indicating an interaction between the meso N–H and two nearby (<5 Å) aryl C–H resonances. From the ¹H–¹H 2D correlation (COSY) spectrum, we can clearly assign the aryl C–H resonances at 8.04 and 7.70 ppm as arising from protons on separate benzene rings (Figure S23). Taken together, the NMR data show that a meso–N atom closest to the direct pyrrole–pyrrole bond is protonated upon addition of H⁺[B(C₆F₅)₄]⁻ to the Mn^V(O) complex at −60 °C (Scheme 3). Thus the Mn^V(O) complex can be protonated at the meso-N position as seen for Mn^III, and this protonation site, together with possible protonation at the oxo position, may play a role in the proton–assisted, light–driven, catalytic aerobic oxidation of HMB.

Figure 3.

Comparison of ¹H NMR spectra (aryl region) for [Mn^V(O)(TBP₈Cz(H))]⁺ (red) at −60 °C and Mn^V(O)(TBP₈Cz) (blue) at −50 °C.

In summary we have shown that a Mn^III porphyrinoid complex can catalyze the visible light-driven, aerobic oxidation of benzylic C–H bonds. Strict control of the proton content is necessary to achieve catalytic turnover. The Mn^III resting state of the catalyst was definitively characterized by XRD. Low temperature spectroscopic methods have identified a new form of an Mn^V(O) corrolazine, in which a remote site is selectively protonated, and suggests a potential binding site for H⁺ along the catalytic pathway. Although we cannot definitively characterize the precise protonation state of the Mn^V complex under catalytic conditions, taken together, the data indicate that 2 equiv of acid are necessary for catalysis. This work furthers our knowledge regarding the criteria for developing biomimetic heme complexes for O₂ activation catalysts, and suggests a promising strategy in catalyst design could involve the incorporation of remote ligand sites for modification by non-covalent interactions.