Biomimetic Reactivity of Oxygen-Derived Manganese and Iron Porphyrinoid Complexes

Abstract

Heme proteins utilize the heme cofactor, an iron porphyrin, to perform a diverse range of reactions including dioxygen binding and transport, electron transfer, and oxidation/oxygenations. These reactions share several, key metalloporphyrin intermediates, typically derived from dioxygen and its congeners such as hydrogen peroxide, and which are comprised of metal-dioxygen, metal-superoxo, metal-peroxo, and metal-oxo adducts. A wide variety of synthetic metalloporphyrinoid complexes have been synthesized to generate and stabilize these intermediates, and then employed to determine the spectroscopic features, structures, and reactivities of such species in controlled and well-defined environments. In this review, we summarize recent findings on the reactivity of these species with common porphyrinoid scaffolds employed for biomimetic studies. The proposed mechanisms of action are emphasized. The review is organized by structural type of metal-oxygen intermediate, and broken into subsections based on the metal (manganese, iron) and porphyrinoid ligand (porphyrin, corrole, corrolazine).

TABLE OF CONTENTS FIGURE

1.1. Introduction

Porphyrins and related macrocycles have been well studied because of their role as biological ligands. The heme cofactor, which consists of an iron-bound porphyrin, is at the active site of many metalloproteins and performs a diverse range of chemistry such as oxygen transport and storage, electron transfer, and oxidation/oxygenation reactions (using O₂/H₂O₂ as the oxidant and in many cases the O-atom source).^1-3

Examples of monooxygenation reactions catalyzed by heme enzymes include aromatic and alkane hydroxylations, olefin epoxidations, as well as N-, S- or O-dealkylations. One class of enzymes that performs many of these transformations, which use dioxygen (O₂) as the oxidant and O-atom source, is a large class of thiolate-ligated heme enzymes called Cytochrome P450s (CYPs), which are found in most mammalian tissues and organs as well as other organisms such as plants, bacteria, and yeast. Heme-dependent dioxygenases, which incorporate both oxygen atoms into the substrate, are a smaller class of enzymes that oxidize tryptophan or other indole derivatives to give N-formylkynurenine products. While there is a lot of information regarding the mechanism of CYP, less is known about the mechanism of dioxygenases.^1,4 Peroxidases, catalases, and choroperoxidases (CPOs) all utilize hydrogen peroxide as oxidant and substrate but differ in their functions and proximal ligand; peroxidase heme iron is bound by histidine, catalase heme iron is bound by tyrosinate, and CPO heme is bound by cysteinate. Peroxidases utilize hydrogen peroxide (H₂O₂) to oxidize such substrates as phenols and aromatic amines. Catalases perform the disproportionation of hydrogen peroxide to dioxygen and water. Heme chloroperoxidase is a halogenating enzyme that uses hydrogen peroxide to chlorinate aliphatic substrates; however, it is functionally tolerant of other substrates in that it is also capable of catalyzing peroxidase, catalase, and P450-type reactions. Finally, heme degradation is performed by heme oxygenase enzymes, in which the oxygenation of the porphyrin ligand itself occurs.

All of the enzymes discussed here utilize only a few common intermediates derived from dioxygen or hydrogen peroxide in their mechanistic cycles. These intermediates are metal-superoxo (O₂^•−), - peroxo (O₂²⁻), - hydroperoxo (OOH⁻), and - oxo (O²⁻) species (Scheme 1). Synthetic analogs target these intermediates and allow for the structural and spectroscopic features of these species to be analyzed. The development of synthetic analogs of these intermediates provides systems in which bond-making and bond-breaking events that occur at the metal can be studied in well-controlled environments. In this review, we focus on the biomimetic reactivity of synthetic oxygen-derived metalloporphyrinoid compounds.

Scheme 1.

Common Metal-Oxygen Intermediates Utilized by Metalloproteins

The environment surrounding the heme in the active site varies depending on the system. The iron center is coordinated axially on the proximal side by an N, O, or S donor derived from the amino acids histidine, tryptophan, tyrosine, or cysteine. The identity of the amino acids surrounding the distal side of the metalloporphyrin also play important roles, such as providing hydrogen bond donors or acceptors that can interact with the reactive intermediates during the catalytic cycle. These structural features combine to provide selectivity and control reactivity of the metalloenzyme, allowing for a wide variety of reaction outcomes (Scheme 2). For example, within the broad class of heme proteins, myoglobin and CYP utilize heme iron at their active sites, but have entirely different functions and O₂ reactivities due to differences in the chemical environment surrounding the heme sites. Myoglobin is a simple dioxygen- binding and transport protein. At the active site, the heme is ligated by an axial histidine nitrogen, and contains a nearby histidine on the distal side, which provides selectivity for O₂ over other toxic small molecules such as carbon monoxide (CO). In CYP, a monooxygenase, the heme iron is ligated by a cysteinate on the proximal side, and several residues and a hydrogen-bonded water network on the distal side control O₂ binding, cleavage, and subsequent substrate oxidation. Scheme 3 illustrates the “push/pull” effects of these interactions in CYP that influence this reactivity.¹

Scheme 2.

Structural Factors that Affect Reactivity in Heme Proteins

Scheme 3.

Push/Pull Effects of Axial Ligand and Distal Residues on O–O Bond Cleavage in Cytochrome P450

Because the heme cofactor utilizes the porphyrin framework, synthetic metalloporphyrin complexes have proved useful as model systems to study the factors that control the reactivity of heme proteins. Porphyrins are heterocyclic macrocycles composed of four pyrrole units joined by methine bridges. They are aromatic, 18-π electron systems, and this high degree of conjugation allows for intense absorption bands in the visible region. Physicochemical properties of the porphyrin ring, such as electronic effects, steric effects, and solubility, can be varied by substituting at the meso- positions of the macrocycle or the β- positions of the pyrrole unit. The porphyrinoid family encompasses the common structural types shown in Chart 1. This family includes ring-contracted analogs in which one meso position is missing from the framework (as in corroles and corrolazines). Ring contracted analogs such as subporphyrins, subphthalocyanines, and subporphyrazines, which are missing a pyrrole unit, are also members of the porphyrinoid family, but are only known as the boron-bound derivatives.^5,6 Ring-expanded porphyrins are another large category within the porphyrinoid family, in which one or more meso- positions or pyrrole units are added to the framework. There are some excellent reviews on the synthesis of ring-expanded porphyrinoid compounds and their reactivities including articles by Dolphin,⁷ Sessler,⁸ and most recently Osuka.⁹ In this review, we focus on porphyrins and the contracted corrole and corrolazine systems, which have been used in a range of biomimetic studies over the past several years.

Chart 1.

Core Structures of Common Members of the Porphyrinoid Family Used for Biomimetic Studies

In metalloporphyin models of heme proteins, the “active site” has been synthetically constructed in levels by building out from the porphyrin framework. The first level of synthetic design is to modify the porphyrin substituents on the periphery of the ring. The second level is to change the first coordination sphere of the central metal atom by adding axial ligands. The third level is to add steric encumbrance on functional groups orthogonal to the macrocycle plane, which may interact with oxygen-derived intermediates through non-covalent interactions such as hydrogen-bond donors or acceptors. An example of the latter comes from the synthesis of so-called “Pacman” porphyrins.¹⁰ The fourth level of synthetic modification is through the encapsulation of the metalloporphyrin in a larger scaffold such as MOFs, dendrimers, or short peptides.^11,12 Finally, the ability to encapsulate the metalloporphyin in artificial or engineered proteins has seen a recent surge in effort.^13-16

In many cases, the reactive intermediates of interest (e.g. metal-oxo, metal-O₂) in metalloporphyrin systems are transient species at room temperature and must be generated or trapped at low temperatures for characterization and study. Ring-contracted porphyrinoid complexes such as corrole and corrolazine (Chart 1) provide enhanced thermal stabilization for some of these intermediates and allow for their direct characterization or even isolation in some cases. Corroles contain a direct pyrrole-pyrrole linkage and thus differ by the deletion of a meso-carbon atom relative to porphyrins. Fully deprotonated, they are trianionic (3−) ligands, and the higher charge and smaller cavity size provide a stabilizing environment for metal ions in higher oxidation states (≥ 3). Corrolazines are also ring-contracted like corroles, but contain meso-nitrogen atoms in place of meso-carbons. The biomimetic reactivity of these three ligand systems (porphyrins, corroles, and corrolazines) will be covered here. The synthesis of corroles, corrolazines, and their metalated forms has been recently reviewed in the Handbook of Porphyrin Science (2011).¹⁷ Herein recent developments on the biomimetic reactivity of such systems are reviewed. The review is organized by structural type of metalloporphyrinoid species. Complexes are restricted to those able to participate in one or more of the biomimetic reaction types defined in the following section.

1.2. Biomimetic reaction types

Shown in Chart 2 are several biomimetic reactions performed by metalloporphyrin complexes. Oxygen atom transfer (OAT) is a two-electron process which results in the net transfer of an oxygen atom to a substrate and can occur by at least two different mechanisms, with two of the most common for metal-oxo species shown in Scheme 4. The first of these is a concerted mechanism, in which the oxygen atom is transferred in one step. Another common mechanism is a stepwise mechanism, in which the substrate is oxidized by one electron either by an outer-sphere or inner-sphere oxidation, followed by fast oxygen transfer. Typically this reaction is facile when a substrate is nucleophilic and the oxidant is an electrophilic oxo complex. The substrate for OAT is often an alkene, sulfide, or phosphine, resulting in an epoxide, sulfoxide, or phosphine oxide, respectively.

Chart 2.

Biomimetic Reactions Performed by Metalloporphyrinoid Complexes

Scheme 4.

Possible Mechanisms for Oxygen Atom Transfer from a Metal-Oxo Complex

Hydrogen-atom abstraction is the net transfer of an H-atom, which can occur by several different mechanisms. H-atom abstraction is usually proposed as the first step in the oxidation of inert C–H bonds by high-valent metal-oxo species in biological systems, followed by rebound of the newly formed carbon radical with a metal-hydroxo intermediate. Mechanistically, understanding H-atom abstraction can be complex because H• transfer can occur in a step-wise fashion (proton transfer followed by electron transfer or vice versa) or in a concerted fashion. The kinetics and thermodynamics are intertwined, and the thermodynamic parameters (pK_a, E°) associated with H-atom transfer are illustrated in the square scheme in Scheme 5. The square scheme is similar to a thermodynamic Hess cycle, in which the sum of the ΔG°s of the horizontal (from the E° of ET) and vertical (from the pK_a of PT) steps adds up to give the ΔG° of the diagonal step.

Scheme 5.

Square Scheme Depicting H-atom Transfer to a Metal-Oxo Complex

There are many examples of hydroxylation reactions by high-valent metal-oxo porphyrin species. These reactions typically are initiated by H-atom abstraction of a substrate by the metal-oxo species to give a high-valent metal hydroxide and a substrate radical, followed by oxygen rebound in which the substrate radical “combines” with the hydroxyl radical from the metal-hydroxide. Thus far, there are no examples in the biomimetic literature, in which this second step, involving oxygen rebound, has been examined directly via reaction of a high-valent metal-hydroxide complex with an organic radical to give an ROH product.

Oxygen (O₂) activation is another common biomimetic reaction. This umbrella term encompasses O₂ binding/transport (reversible O₂ activation), the two-electron/two-proton reduction of O₂ to hydrogen peroxide (H₂O₂), and the four-electron/four-proton reduction of O₂ to water. The binding and activation of H₂O₂ is related to this process, and shares common metal-oxygen intermediates. The extent of O₂ activation by a metal porphyrinoid complex is controlled by many different factors, including the metal identity, oxidation state, and ligand electronic properties, as well as any exogenous or ligand-appended hydrogen-bond donor groups in the secondary coordination sphere.

2. Metal-Oxo Porphyrinoid Complexes as Models for Biological Oxidations

2.1. Iron-Oxo Complexes

2.1.1. Iron-Oxo Porphyrins

Reactions carried out by iron-oxo porphyrins have received considerable attention because of their synthetic utility in the oxidation of organic substrates and their relevance to enzymatic heme-catalyzed processes such as those found in cytochrome P450, peroxidases, and catalases. The key reactive intermediates are postulated as high-valent iron-oxo porphyrin species defined in the biological context as Cpd I (Fe^IV(O)(porphyrin^•+)), formally a Fe^V(O) complex, and Cpd II ((Fe^IV(O)(porphyrin)). The terms “compound I” and “compound II” were historically applied to the reactive ferryl species found in horseradish peroxidases (HRP), but are now used to describe isoelectronic iron-oxo porphyrin species in other heme enzymes.¹⁸ Studies of synthetic high-valent iron-oxo porphyrin complexes have provided valuable insights into the reactivity of these species because of the relative ease of control of the ligand steric and electronic properties, and reaction environment, as compared to heme enzymes. In 1979, Groves et al. demonstrated alkene epoxidation and alkane hydroxylation using iodosylbenzene (PhIO) with an Fe^III(Cl)(TPP) (TPP = tetraphenylporphyrin) catalyst at room temperature.¹⁹ One of the proposed intermediates for this catalytic oxidation was an iron-oxo complex. An early analog of the proposed catalytic intermediate came from the oxidation of Fe^III(Cl)(TMP) (TMP = tetramesitylporphyrin) with 1.5 equiv of mCPBA at −78 °C to form Fe^IV(O)(TMP^•+), a porphyrin π-radical cation complex (Scheme 6).²⁰ The large downfield shifts in the ¹H NMR peaks of the mesityl groups and the UV-visible spectrum suggested a porphyrin π-radical cation. The small Mossbauer isomer shift of 0.05 mm/s (ΔE_Q = 1.49 mm/s) was in the range expected for an iron(IV) species. This complex is also reactive with olefins to form epoxides even at low temperatures. When the oxidation reaction was performed in the presence of H₂¹⁸O, almost complete (99%) ¹⁸O incorporation in the epoxide was observed.

Scheme 6.

Synthesis of Fe^IV(O)(TMP^•+), an Early Compound I Analog

After these initial reports, a number of studies on iron-oxo porphyrins were made in efforts to define possible mechanisms of substrate oxidation by heme enzymes. Excellent reviews on the reactivity of iron-oxo porphyrin complexes can be found in the literature, including those from Fujii (2000,²¹ 2002²², 2016²³) Meunier et al. (2004),²⁴ Groves (2006),²⁵ Newcomb (2006),²⁶ Nam (2007),²⁷ Costas (2011),²⁸ Ray and Meyer (2012),²⁹ and Rutkowska-Zbik (2014)³⁰. Dedicated chapters in The Porphyrin Handbook (2000),^18,31-34 Handbook of Porphyrin Science,³⁵ and other compilations^36-38 can also be found. This section will focus on recently published works in this area.

Although iron complexes with the tetrakis(meso-phenyl) porphyrin general structure led to some metastable iron-oxo complexes that mimic cytochrome P450 compound I reactivity, their catalytic activities were plagued with porphyrin decomposition and μ-oxo dimer formation. These problems were overcome by introducing alkyl, halogen, or other substituents to either the meso-aryl or the β-pyrrole positions of the porphyrin ring. The result was increased catalytic activity and stability of the porphyrin ligand towards decomposition.^38,39 For example, Groves has reported high rates for C-H hydroxylation in aqueous solution for Fe^IV(O)(TMPyP^•+) (TMPyP = tetrakis(N-methyl-4-pyridinium)porphyrin) (Scheme 7).⁴⁰ Reaction of Fe^IV(O)(TMPyP^•+) with xanthene gave the hydroxylation product, 9-xanthydrol (90%), and performing the reaction in the presence of 47.5 % H₂¹⁸O led to 21% ¹⁸O incorporation. This incorporation is consistent with an oxygen-rebound scenario, where a high-valent Fe^IV(O)(porphyrin^•+) species abstracts an H-atom from xanthene to give a carbon radical, and an Fe^IV(OH) rebound intermediate. This species “rebounds” the •OH group to the carbon radical in the solvent cage to form the alcohol product and the resting ferric state. The rate constant for this reaction was found to be 3.6 × 10⁶ M⁻¹ s⁻¹, with a modest kinetic isotope effect of 2.1. A variety of C-H hydroxylation substrates (fluorene-4-carboxylic acid, 4-isopropyl and 4-ethylbenzoic acid) were tested as well, and the hydroxylation rates correlate well with the substrate BDEs (C-H). The BDE(O-H) for Fe^IV(O-H)(TMPyP) was estimated from a Brønsted–Evans–Polanyi relationship to be ~100 kcal/mol, much larger than the values estimated for Fe^IV(O-H)(TMP) (92 kcal/mol),⁴¹ and Fe^III(O-H)(TPFPP) (~86 kcal/mol)⁴². The high kinetic reactivity of this complex can be traced to the redox and basicity tuning of the iron-oxo porphyrin, expected from a porphyrin ligand with electronegative and positively charged substituents. It was further suggested that the increase in reactivity could be due to a spin-state crossing phenomena occurring in the course of the reaction.

Scheme 7.

Reactivity of Fe^IV(O)(TMPyP^•+), a Highly Reactive Compound I Model Complex

An alternative method of producing high valent iron-oxo species was reported by the groups of Zhang, Fung, and Newcomb.⁴¹ In this method, Fe^III(ClO₄)(TMP) was oxidized to the metastable Fe^IV(ClO₄)₂(TMP) which, when irradiated with 355 nm laser light, gave a new transient species which could be characterized only by UV-visible spectroscopy (Scheme 8). Preparative reaction of this transient iron-oxo species with cyclooctene gave cyclooctene epoxide in 28% yield. This species also reacts with alkanes and benzylic C-H substrates with rates up to 5 orders of magnitude larger than the cpd I analog Fe^IV(O)(TMP^•+). For this reason, this species was speculated to be an Fe^V(O)(TMP) complex on the basis of its reactivity. Although this speculation is reasonable from kinetic data, this oxidation state assignment should remain tentative until appropriate spectroscopic evidence such as EPR, Mössbauer, or X-ray absorption spectroscopy is presented. These laser flash photolysis (FLP) studies were further extended to other porphyrin analogs, TPP (tetraphenylporphyrin) and OEP (octaethylporphyrin), and the order of reactivity was found to be OEP>TPP>TMP (Scheme 8). These putative transient iron-oxo species, characterized by their unique UV-visible spectra, exhibited similar rate constants to the similarly generated Mn^V(O) porphyrins. Rate constants for benzylic C-H bond cleavage correlated with the C-H BDEs, and Hammett correlations with para-substituted styrenes gave slopes between −0.5 and −0.7, reflecting the electrophilic character of these reactions.⁴³

Scheme 8.

Synthesis of a Proposed Fe^V(O) Intermediate by Laser-Flash Photolysis, and Different Porphyrin Ligands Used in the Study

Zhang and coworkers recently reported a method of generating Fe^IV(O)(porphyrin^•+) or Fe^IV(O)(porphyrin) (porphyrin = TMP, TPP, TPFPP) complexes by simple visible light irradiation of Fe^III(OBrO₂)(porphyrin) complex.⁴⁴ The photolysis reaction is proposed to go through the heterolytic cleavage of the O-Br bond, forming a putative Fe^V(O) complex which then relaxes to an Fe^IV(O)(porphyrin^•+). This was observed in electron-rich porphyrins such as TMP and TPP. With Fe^III(TPFPP) (TPFPP = tetrakis(pentafluorophenyl)porphyrin), however, a final Fe^IV(O)(TPFPP) product was observed, which is proposed to result from the Fe^V(O)(TPFPP) comproportionating with residual Fe^III(TPFPP). The formation of cpd I and cpd II analogs was further confirmed by comparable rate constants obtained in the HAT and OAT reactions of iron-oxo porphyrins generated using visible light generation vs addition of mCPBA. DFT studies show that an Fe^V(O) ground state is favored in vacuo. However, the Fe^IV(O)(porphyrin^•+) configuration is favored in the solvent or protein environment.⁴⁵

Newcomb and coworkers have shown that single-turnover hydroxylation reactions of ethyl benzene and benzyl alcohol with Fe^IV(O)(TMP^•+), generated from the addition of m-CPBA to Fe^III(TMP) in CH₃CN, give large temperature-dependent KIEs, which are ascribed to H-atom tunneling.⁴⁶ The substrates were oxidized in the benzylic C-H position, to give 1-phenylethanol and benzaldehyde respectively, as determined by gas chromatography (GC).

It has been shown that axial ligands of heme enzymes also play an important role in their reactivity. In work from Fujii and coworkers, the factors that affect H-tunneling processes were explored by studying the C-H hydroxylations of xanthene and tetralin with Fe^IV(O)(TMP^•+)(L) with different axial ligands (L = NO₃⁻, Cl⁻, imidazole) (Scheme 9).⁴⁷ The Fe^IV(O)(TMP^•+)(L) complex was generated with excess O₃ (ozone) in CH₂Cl₂ at low temperatures (−20 to −95 °C). Non-linear Arrhenius plots were observed, providing strong evidence for the participation of H-tunneling in the rate-limiting H-transfer process. In addition, the authors showed that H-tunneling is controlled by temperature, the BDE of the C-H substrate, and the reactivity and Fe-O bond strength of the Fe^IV(O)(TMP^•+)(L) complex.

Scheme 9.

Reactivity of a Series of Fe^IV(O)(TMP^•+)(L) Complexes with Benzylic C-H Substrates Xanthene and Tetralin

The strong electron-donation from the cysteinyl thiolate ligand in cytochrome P450s is proposed to influence the driving force for C-H bond cleavage.⁴⁸ Efforts have been made to synthesize thiolate-ligated heme models with the goal of studying the role of thiolate ligation on reactivity, and these are summarized in a review.⁴⁹ The facile oxidation of thiolate ligands to disulfide remains an outstanding difficulty in the synthesis of thiolate-ligated heme models, and this is usually prevented by using a large excess of the thiolate ligand or by covalently attaching the thiolate group to the porphyrin scaffold.^38,49-51 A stable thiolate-ligated complex (called SR complex, Swan Resting) was prepared by Higuchi and co-workers by covalently attaching an alkane-thiolate group to the porphyrin, and bulky pivaloyl groups were added to protect the sulfur atom from oxidation, and amide NH groups for hydrogen-bonding with the sulfur atom (Scheme 10).⁵² However, although the SR complex is known to react with oxygen atom donors and peroxyacids, it does not form a stable dioxygen adduct due to the absence of a binding pocket. In addition to the instability of thiolate-ligated heme-dioxygen adducts, the increasingly complicated synthesis of these model complexes has limited the number of examples of these systems in the literature.

Scheme 10.

Swan Resting (SR) Complex, an Example of a Thiolate-Ligated Heme Model Complex, and its Derivatives

Substitution of a thiolate ligand (RS⁻) with a sulfonate (RSO₃⁻) ligand reduces the charge density at the metal and tunes the Fe^III/II redox couple to a more positive potential close to that found in P450_cam.⁵³ Woggon and van Eldik showed that the cpd I analog Fe^IV(O)(porphyrin^•+)(RSO₃⁻) can be generated with m-CPBA under saturation kinetics, which suggests the formation of an acyl-peroxo adduct intermediate, eventually converting to the cpd I analog (Scheme 11).⁵⁴ There has been speculation that an Fe^III(OOH)(porphyrin) complex (compound 0), the precursor to compound I, can act as a second oxidant for certain substrates in cytochrome P450. Thus, the generation of an Fe^III-acylperoxo complex with m-CPBA can be utilized as a model to study the possible involvement of an Fe^III-OOH species in heme enzyme oxidations.

Scheme 11.

Conversion of an RSO₃-ligated Fe^III(acylperoxo)(porphyrin) Intermediate into a Cpd I Analog Fe^IV(O)(RSO₃)(porpyrin^•+)

Nam and coworkers have shown that Fe^IV(O)(TMP^•+) is the active oxidant in olefin epoxidation with cyclohexene, aromatic ring oxidation with anthracene, C–H bond activation with dihydroanthracene, and alcohol oxidation with benzyl alcohol in CH₃CN/CH₂Cl₂ (1:1 v/v) at −40 °C. These results contrast with the poor oxidizing ability of the Fe^III(m-CPBA) adduct of this complex.⁵⁵ Similarly, van Eldik and coworkers reported kinetic evidence showing that oxygen atom transfer for Fe^III(m-CPBA)(TMP) was orders of magnitude slower than the Fe^IV(O)(TMP^•+) complex in epoxidation and sulfoxidation reactions.⁵⁶

Heme synthetic models with axial ligands covalently tethered to the porphyrin ligand can provide insights into effects of axial ligation on reactivity. However, the synthesis of these tethered systems are usually difficult and low-yielding. An alternative approach is by adding exogenous ligands to the cpd I and cpd II analogs. Large increases in cyclooctene epoxidation rates were observed upon addition of axial ligands to the Fe^IV(O)(porph^•+) (porph = TMP, TMTMP) (TMTMP = 2,7,12,17-tetramethyl-3,8,13,18-tetramesitylporphyrin) complex, and the epoxidation rates increase in the order of L = NO₃⁻ < Cl⁻ < 3-fluoro-4-nitrophenolate<imidazole.⁵⁷ Interestingly, the measured redox potentials for Fe^IV(O)(porph)(L)/Fe^IV(O)(porph^•+)(L) follow the opposite trend, i.e., a positive shift was observed with anionic axial ligands (more oxidizing) and a negative shift was observed with neutral ligands (less oxidizing).⁵⁸ The reactivity of Fe^IV(O)(L)(porphyrin^•+) correlates instead with the Fe^III/Fe^II redox potential of the reaction product, Fe^III(L)(porphyrin), suggesting that axial ligands affect the reactivity of Fe^IV(O)(porphyrin^•+) complexes by stabilizing the reaction product.⁵⁹ In a separate study, the reactivity of Fe^IV(O)(TMP^•+) in olefin epoxidation, aromatic hydroxylation, alcohol oxidation and alkane hydroxylation was measured in the presence of para-substituted pyridine N-oxide derivatives (para-Y-C₅H₅NO, Y = -OCH₃, -CH₃ -H, -Cl) as axial donors. It was found that the binding of these axial ligands lead to enhanced reactivities in all of the oxidation reactions.⁶⁰ The most electron-donating pyridine N-oxide (p-OCH₃-C₅H₅NO) gives the most reactive Fe^IV(O)(TMP^•+)(L) complex in both O-atom transfer and H-atom transfer reactions. This increased reactivity was attributed to a stronger Fe(O-H) bond formed in HAT reactions, and a weaker reactant Fe=O bond broken in OAT reactions.

DFT calculations on the effects of various anionic axial ligands on aromatic hydroxylation with Fe^IV(O)(porphyrin^•+) have shown that stronger axial ligands give rise to higher pKas of the corresponding Fe^IV(OH)(porphyrin) complex, while electronic effects are less affected. Moreover, the barrier for aromatic hydroxylation appear to be proportional to the strength of the O-H bond formed in the Fe^IV(OH) complex.⁶¹ A similar trend was observed in alkane hydroxylation,⁶² with the strength of the C-H bond of the substrate also correlating with the reaction barrier. For sulfoxidation⁶³ and alkene epoxidation^64,65 reactions however, the barrier height correlates with the ionization potential of the substrate, supporting an ET rate-determining step in the oxidation process. Calculations were also performed to determine the source of regioselectivity in the oxidation of ethylbenzene by Fe^IV(O)(porphyrin^•+) to form 1-phenylethanol (alkane hydroxylation product) and p-ethylphenol (aromatic hydroxylation product). The calculations indicate that an axial acetonitrile ligand favors aromatic hydroxylation, whereas a Cl⁻ axial ligand favors alkane hydroxylation.⁶⁶ The anionic ligand, such as Cl⁻, makes the oxidant less electrophilic due to orbital mixing which disfavors the aromatic hydroxylation mechanism and favors the alternative alkane hydroxylation mechanism.

A high-valent terminal iron-oxo species generated from a nitrido-bridged-diiron porphyrin complex was trapped and characterized, and was found to oxidize C-H bonds as strong as that of methane (BDE = 104 kcal/mol).⁶⁷ Addition of m-CPBA to [(TPP)Fe^III(N)Fe^IV(TPP)] in CH₂Cl₂ at −80 °C results in formation of a peroxo complex which undergoes heterolytic cleavage to form the diiron, terminal oxo species [(TPP)(m-CBA)Fe^IV(N)Fe^IV(O)(TPP^•+)]⁻, characterized by UV-vis, ESI-MS, EPR and Mössbauer spectroscopies (Scheme 12 and Figure 1). This species was found to have superior oxidizing properties towards alkanes when compared with the model complex Fe^IV(O)(TPP^•+), and KIEs of about 3 support a rate-limiting C-H cleavage. To avoid uncertainties as a result of the possible oxidation of organic solvent by strongly oxidizing species, heterogeneous oxidation of CH₄ was performed starting with the [(TPP)Fe^III(N)Fe^IV(TPP)] catalyst combined with silica and dried at 50 °C. The supported catalyst, m-CPBA and methane gas were added in an autoclave reactor heated to 60 °C. Methane was oxidized to HCOOH (formic acid) from GC and LC analyses. This report showed the enhancement of C-H bond oxidations induced by an N-bridging ligand on an iron-oxo complex.

Scheme 12.

Proposed Mechanism for the Formation of [(TPP)(m-CBA)Fe^IV(N)Fe^IV(O)(TPP^•+)]⁻

Figure 1.

UV-vis spectral changes from the starting [(TPP)Fe^III(N)Fe^IV(TPP)] (1), formation of the mcpba adduct (2), and [(TPP)(m-CBA)Fe^IV(N)Fe^IV(O)(TPP^•+)]⁻ (3) in CH₂Cl₂ at −80 °C. Reprinted by permission from Macmillan Publishers Ltd: Nature Chemistry, Nature Chemistry 4, 1024–1029 (2012), copyright 2012.

Chloroperoxidase (CPO) and myeloperoxidase (MPO) are heme enzymes that catalyze the oxidation of Cl⁻ with H₂O₂, and are responsible for the biosynthesis of chlorinated compounds in microorganisms.^68,69 Studies have suggested that a cpd I intermediate oxidizes Cl⁻ to form a transient Fe^III(OCl) species. Studies from the Fujii group have shown different ways of accessing Fe^III(OCl) species from an Fe^III(TPFP) complex (Scheme 13). Addition of Cl⁻ ions to a solution of Fe^IV(O)(tpfp^•+) at −90 °C leads to reduction of the Fe complex, giving Fe^IV(O)(tpfp). The Cl⁻ ion is proposed to be oxidized to chlorine radical (Cl•).⁷⁰ A study from Fujii⁵⁸ showed that the redox couple for Fe^IV(O)(tpfp)(ClO₄⁻)^+/0 is 1.39 V vs SCE in dichloromethane at −60 °C. This finding indicates that the iron-oxo porphyrin complex should be capable of oxidizing chloride ion (1.11 V vs SCE in H₂O).⁷¹ Single turn-over chlorination reactions with organic substrates led to modest yields of chlorinated products. In an alternative route, protonation of Fe^IV(O)(TPFP^•+) with TFA results in the tautomerization of this complex to Fe^III(L’)(TPFP⁺⁺) (L’ = anion other than Cl⁻), and further addition of Cl⁻ ions lead to conversion to Fe^III(L’)(Cl)(TPFP-Cl), a meso-chloro-isoporphyrin.⁷² This complex can chlorinate electron-rich benzene and alkene derivatives catalytically through an electrophilic aromatic substitution (for benzene) or electrophilic addition mechanism (for alkenes). An Fe^III(OCl) species was independently formed by addition of 4 equiv of OCl⁻ to Fe^III(OH)(TPFP) in 1:1 CH₂Cl₂-CH₃CN at −60°C.⁷³ Based on spectroscopic data, the species was characterized as Fe^III(OCl)₂(TPFP)⁻, which eventually decomposed to Fe^IV(O)(TPFP) (Scheme 13). Imidazole derivatives of this complex were also synthesized. The O-atom transfer and chlorination reactivity of the Fe^III(OCl) complexes were tested with thioanisole, cyclohexene and 1,3,5-trimethoxybenzene substrates. The imidazole derivatives were found to have attenuated reactivity compared to the bis-hypochlorite analog. These biomimetic studies find their importance in comparison with CPO/MPO heme enzymes, where Fe^III(OCl) species are indicated, but spectroscopic evidence for such species is scarce.

Scheme 13.

Generation of Fe^III(OCl)₂(porph)⁻, Fe^IV(O)(porph^•+), and Fe^IV(O)(porph) Species and Their Interconversions

The last step in perchlorate (ClO₄⁻) degradation by microbes is the decomposition of chlorite (ClO₂⁻) to dioxygen (O₂) and chloride (Cl⁻) mediated by the heme enzyme chlorite dismutase. From stopped-flow and EPR experiments, the mechanism was proposed to go through a cpd I intermediate.^74,75 This process was mimicked outside of a protein with water-soluble porphyrins, where chlorite is decomposed to chloride and chlorate in a 1:2 stoichiometry (Scheme 14).^76,77 Based on experimental observations and kinetic studies, two competing cycles were proposed. The first Fe^IV(O)P^•+/Fe^III cycle produces Cl⁻ and O₂, as in the heme enzyme. However, the Fe^IV(O)P^•+ can also comproportionate with Fe^III to form Fe^IV(O)P. The second competing cycle involves an O-atom transfer from Fe^IV(O)P to ClO₂⁻ to form Fe^II (P) and ClO₃⁻.

Scheme 14.

Proposed Mechanisms for Chlorite Dismutase and Water-Soluble Porphyrins

Aromatic hydroxylation is another reaction studied extensively with iron porphyrin models. Its mechanism differs from alkane hydroxylation in that small KIEs and 1,2-hydride shifts (NIH shift) are observed, precluding the usual oxygen rebound mechanism proposed for alkane hydroxylations. Mechanistic information on aromatic hydroxylation of a series of para-substituted benzene derivatives with Fe^IV(O)(TPFPP^•+) was obtained by the groups of Jang and Nam.⁷⁸ A large negative Hammett slope (ρ) of −8.0 and a KIE value of ~0.8 were observed, consistent with the formation of a cationic σ adduct upon addition of an electrophilic iron-oxo species. Fujii and coworkers studied the ability of a series of Fe^IV(O)(porphyrin^•+) with electron-withdrawing porphyrins to hydroxylate aromatic substrates (benzene, anisole, naphthalene), and were able to show the involvement of an electron transfer process in these reactions.⁷⁹ Plots of (RT/F) ln(k₂) vs ΔE_1/2 (where k₂ is the second-order rate constant of aromatic hydroxylation and E_1/2 is the redox potential for the cpd I/II couple) were constructed, and revealed slopes that are in between 0.5 and 1.0, indicating that the electron transfer process is coupled to the subsequent O–C bond formation (instead of proton transfer, which is not found to be in the rate-limiting step). Direct electron transfer between the substrate and Fe^IV(O)(porphyrin^•+) can be ruled out because the reaction is highly endergonic (ΔG_ET >> 0). However, this uphill electron transfer step is followed by a downhill O-C bond formation, making the over-all process thermodynamically favorable. The proposed mechanism (Scheme 15) brings together and reasonably explains the previously observed results.

Scheme 15.

Proposed Mechanism of Aromatic Hydroxylation with Fe^IV(O)(porphyrin^•+) Complexes Showing an Electron Transfer Process Coupled to the Subsequent Bond Formation Step

N-oxygenation and N-dealkylation reactions are some of the many oxidations performed by compound I. The mechanism for N-dealkylation is proposed to go through either an H-atom transfer mechanism or a rate-limiting electron transfer step coupled with a proton transfer step (Scheme 16). DFT calculations showed that both N-oxygenation and N-dealkylation proceed from the low spin state of compound I, and the computed KIEs for the rate-determining HAT step fits well with experimental data.⁸⁰

Scheme 16.

Proposed Mechanisms for Oxidative N-demethylation by Compound I, Fe^IV(O)(porphyrin^•+)

Aside from exogenous substrates, the high-valent iron porphyrin complex is also capable of performing heme degradation reactions as observed in the enzyme heme oxygenase. One of the proposed reaction pathways for the first step of this reaction is the formation of a transient cpd-I (Fe^IV(O)(porphyrin^•+)), through the heterolytic cleavage of an Fe^III(OOH)(porphyrin) species. Karlin and co-workers have described a heme oxygenase model system in which an iron(III) tetraarylporphyrin complex is selectively oxidized to the final verdoheme-like product using a sacrificial oxidant, cerium(IV) ammonium nitrate (CAN), in the presence of water (Scheme 17).⁸¹ A combination of UV-vis, EPR, and ESI-MS observations provided evidence for the formation of 1) an isoporphyrin complex resulting from the attack of water/hydroxide on a cpd-I analog, 2) a porphyrin degradation product resembling benzoyl biliverdin, and 3) a verdoheme-like product as well as a benzoyl fragment. From ¹⁸O labeling experiments, water was shown to be the O-atom source in the final oxidized products. In addition, the intermediacy of a cpd-II-like complex (Fe^IV(O)(porphyrin)) was ruled out, as no further oxidation with this complex was detected, while direct observation of the conversion of cpd-I (generated at −80 °C) to the isoporphyrin complex was observed.

Scheme 17.

Proposed Mechanism for the Selective Oxidation of Fe^III(OH)(porphyrin) with CAN to a Verdoheme-like Complex via Formation of Intermediates Resembling Cpd-I, Isoporphyrin, and Benzoyl Biliverdin

Iron(IV)-oxo porphyrins, one-electron reduced analogs of iron(IV)-oxo porphyrin π-radical cation complexes, were first synthesized in 1980 by Balch and La Mar by the addition of nitrogen bases (N-methylimidazole, pyridine, piperidine) to Fe^III peroxo-bridged porphyrins at −80 °C (Scheme 18).^82,83 The Fe^IV(O) complex is EPR silent and shows paramagnetic proton resonances, while its magnetic susceptibility and UV-visible absorption spectrum resembled that of compound II horseradish peroxidase.

Scheme 18.

Synthesis of Fe^IV(O)(TPP), an Early Compound II Analog

Other than homolytic cleavage of the O-O bond of Fe^III peroxo-bridged porphyrins, iron(IV)-oxo porphyrins can be prepared from the starting Fe^III porphyrin by a) chemical oxidation with mCPBA or PhIO under certain conditions, b) electrochemical oxidation and c) addition of hydroperoxides (see ref. ³⁸ and related references therein). Iron(IV)-oxo porphyrins were initially considered as poor oxidants.⁸³ However, it has been shown that iron(IV)-oxo porphyrins react with substituted styrenes to give epoxide, although with poor selectivity towards cis vs trans epoxide formation.⁸⁴ Nam and coworkers reported the reactivity of Fe^IV(O)(porphyrin) complexes with benzylic C-H substrates and NADH analogues to give the corresponding oxidized products (Scheme 19).⁴² A linear correlation is observed between the rate constant log k₂’ and the C-H substrate BDEs, in addition to large KIEs of up to 21 for DHA and xanthene, providing evidence for an H-atom abstraction mechanism. Among the three porphyrins tested, there is a modest increase in HAT and hydride transfer reaction rates with the more electron-deficient porphyrins.

Scheme 19.

Reactivity of a Series of Iron(IV)-Oxo Porphyrins with H-atom and Hydride Donors

A DFT study on the comparative reactivity of Fe^IV(O)(porphyrin^•+) and Fe^IV(O)(porphyrin) with 10-methyl-9,10-dihydroacridine show that Fe^IV(O)(porphyrin^•+) reacts with a much lower barrier.⁸⁵ The main difference, however, is that Fe^IV(O)(porphyrin) reacts via hydrogen atom transfer, while Fe^IV(O)(porphyrin^•+) reacts via hydride transfer. This difference in mechanism was rationalized to be a result of differences in electron abstraction abilities of these two oxidants.

The ability to selectively produce cpd 0, I, and II with Fe^III(TMP) under different conditions was exploited by van Eldik and coworkers to compare directly their reactivity towards O-atom transfer, H-atom transfer, and hydride transfer reactions at low temperatures.⁸⁶ As expected, the Fe^IV(O)(TMP^•+) (cpd I) complex turned out to be the most reactive species. Interestingly, for hydride transfer reactions, the reactivity order was found to be cpd II > cpd 0 > cpd I. In the initial electron-transfer step, Cpd II is reduced to an iron (III)–oxo species, which would be the most basic intermediate and would therefore promote the subsequent proton abstraction step. A similar study from the same group on the comparison of reactivity of Fe^IV(O)(TMP^•+)(2-MeIm) and Fe^IV(O)(TMP)(2-MeIm) also confirmed the high reactivity of the cpd I analog towards epoxidation, H-atom abstraction, and heteroatom oxidation.⁸⁷ Compound I and II mimics were also stabilized in aqueous solution using the water soluble porphyrin Fe^III(TMPS), and their formation was found to be pH-dependent.⁸⁸ Combined temperature and pressure measurements have shown that, for selected substrates, oxidation with the Cpd II mimic is an enthalpy-controlled process whereas it is entropy-controlled for Cpd I.

Iron(IV)-oxo porphyrins were also shown to carry out oxidative N-dealkylation of N,N-dialkylanilines via a rate-limiting electron transfer coupled with proton transfer (ET-PT.⁸⁹ The large negative slopes obtained from the plots of log k_obs vs Hammett parameter σ (ρ = −3.3) and E°_ox of para-substituted N,N-dimethylanilines (slope = −5.0) suggest a rate-limiting electron transfer. In addition, the use of several mechanistic probes and a modest inter- and intramolecular KIE of 2.8 and 4.2 all pointed to an ET-PT mechanism. An Fe^IV(O) porphyrin complexed with a per-o-methylated ß-cyclodextrin dimer was found to oxidize ROCH₃ to ROCH₂OH, which further decomposes to ROH and HCHO, and the corresponding Fe^II(porphyrin) complex (Scheme 20).⁹⁰ An initial H-atom abstraction followed by an oxygen rebound mechanism similar to compound I has been proposed for this system. Work from the same group showed that an Fe^IV(O) porphyrin complexed to a per-o-methylated ß-cyclodextrin dimer was capable of a direct oxygen transfer reaction to the sulfide bond found intramolecularly on the linker.⁹¹

Scheme 20.

Proposed Mechanism for Intramolecular O-demethylation Mediated by an Fe^IV(O) Complex Encapsulated in a Per-O-methylated Cyclodextrin Dimer

As shown previously, there have been several instances of iron(IV)-oxo porphyrin complexes acting as competent oxidants for various substrates. A DFT electronic structure analysis of alkane hydroxylation with Fe^IV(O)(porph) and its hydrosulfide-ligated analog was performed, with methane as a model substrate for alkanes. DFT calculations on the ground triplet and excited quintet spin state surfaces revealed that H-atom abstraction is the rate-determining step, and a two-state reactivity (TSR) mechanism is plausible for C-H activation with Fe^IV(O)(porph), but not for the [Fe^IV(O)(porph)(SH)]⁻ analog, where C-H activation can only occur on the triplet surface.⁹²

It was shown by pH-jump stopped-flow UV-vis spectroscopy that iron(IV)-oxo porphyrins can be protonated twice to form the iron(III)-aquo porphyrin^•+ valence tautomer (Scheme 21).⁹³ The two-proton pK_a values of sulfonated, water-soluble Fe^IV(O) porphyrins were identified and it was found that the most basic Fe^IV(O) complex had the largest C-H bond cleavage rates. For Fe^IV(O)(TMPS^•+), a KIE of 1.80 was observed with xanthene-d₂, and a solvent KIE of 1.65 was seen with D₂O, giving a combined substrate-solvent isotope effect of 2.2. A mechanism which involves both the substrate proton and the solvent proton was proposed which accounts for all of the observed data (Scheme 21).

Scheme 21.

A) Proposed Mechanism for Solvent-Proton Assisted C-H Bond Cleavage by a Sulfonated, Water-soluble Iron(IV)-Oxo Porphyrin π-Radical Cation Complex in the Presence of an Exogenous Proton Source B) Valence Tautomer Equilibrium Observed upon Protonation of Iron(IV)-Oxo Porphyrins C) Water-soluble Porphyrins Used in the Study

2.1.2. Iron-Oxo Corroles and Corrolazines

The ring contracted porphyrinoid compounds corroles and corrolazines are designed to stabilize high-valent transition metals. Although a number of high-valent complexes have been prepared with these porphyrinoid ligands, the iron-oxo corroles and corrolazines are rare and less studied than the iron-oxo porphyrins.⁹⁴ The synthesis of well-characterized, high-valent, mononuclear iron-oxo corroles is challenging in part because of the propensity of iron corroles to form μ-oxo dimers upon oxidation of Fe(III) corrole.^95-97 In addition, the inherent non-innocence in iron(IV) corroles obscures their electronic structure assignment, with both Fe^IV(corrole) (S = 1) and Fe^III(corrole^•+) (S = 1) states as limiting possibilities.^98,99

The use of laser flash photolysis (LFP) – induced ligand cleavage of Mn^IV(OX) (X = ClO₂, NO₂) corroles to generate Mn^V(O) corroles was extended by Newcomb and Zhang to the analogous Fe^IV(OX) (X = ClO₂, NO₂ or Fe^IV) corroles^100-102 to generate a proposed Fe^V(O) corrole (Scheme 22). This species was formulated as Fe^V(O), as opposed to Fe^IV(O)(corrole^•+), because of its increased reactivity compared to the analogous Fe^IV(O)(porphyrin^•+). However, the inability to characterize this highly reactive complex by methods other than UV-vis spectroscopy has precluded definitive oxidation state assignment. The proposed species shows an intense Soret band that decays after 20 ms at 22 °C, which is similarly observed from the reaction of Fe^III corrole with excess mCPBA, and displays high oxidative reactivity towards alkenes (cyclohexene, cyclooctene). A KIE of 3.6 for the oxidation reaction of ethylbenzene was also observed, suggesting that C-H bond cleavage is the rate-determining step.

Scheme 22.

Formation of a Proposed Fe^V(O) Corrole

The isoelectronic Fe^IV(O)(corrole^•+) has been proposed to be energetically accessible based on calculations,¹⁰³ which also found the Fe^V(O)(corrole) complex to be equienergetic. Neither of these species have been experimentally observed. However, more recent DFT calculations from Valentine, Nam, and de Visser¹⁰⁴ find that the Fe^IV(O)(corrole^•+) valence tautomer is lower in energy than the Fe^V(O)(corrole). The H-atom abstraction reactivity of this complex towards a broad range of C-H bond substrates was explored by DFT calculations, and it was shown to react via a single, dominant low spin surface. These results are in stark contrast with the Fe^IV(O)(porphyrin^•+) analog, which is computationally predicted to react via a two-state reactivity surface.

Iron(IV)-oxo corroles, which are compound II analogs, are proposed as an intermediate in the catalytic decomposition of peroxynitrite (ONOO⁻) with water-soluble sulfonated Fe^III corroles.¹⁰⁵ An Fe^III(corrole)-mediated O-O bond cleavage of peroxynitrite forms nitrogen dioxide (N₂O) and a hydroxo or oxo-iron(IV) corrole, which was found to decay via multiple pathways of similar rates of about 10⁶ M⁻¹ s⁻¹ (Scheme 23). However, this highly reactive species has only been observed with stopped-flow UV-vis measurements. Fe^III corroles are also shown to catalyze the oxidation of hydrocarbons with t-butylhydroperoxide,¹⁰⁶ and xanthene-modified and hangman Fe^IV(Cl) corroles are shown to catalyze the disproportionation of hydrogen peroxide via the catalase reaction.^107,108 In both cases, the formation of an Fe^IV(O) corrole intermediate is invoked, but no direct evidence was provided.

Scheme 23.

Proposed Mechanism for the Catalytic Decomposition of Peroxynitrite with a Water-soluble Fe^III Corrole, Showing an [Fe^IV(O)(corrole)]⁻ Intermediate

In 2009, Goldberg and coworkers reported the synthesis and the first direct spectroscopic observation of a compound I analog in a corrole-like scaffold, an Fe^IV(O)(TBP₈Cz^•+) (TBP₈Cz = octakis(tert-butylphenyl)corrolazine) complex.¹⁰⁹ The Fe^III precursor complex was synthesized by reaction of metal-free corrolazine with Fe(acac)₃ in refluxing pyridine,^110,111 to give Fe^III(TBP₈Cz). Addition of an O-atom transfer reagent (mCPBA, PFIB, or PhIO) to the Fe^III precursor complex in 1:1 CH₂Cl₂/CH₃OH at −78 °C resulted in the formation of Fe^IV (O)(TBP₈Cz^•+) (Scheme 24). The Fe^IV(O)(TBP₈Cz^•+) complex (Figure 2) has been invoked as the key oxidizing species in the catalytic oxidation of olefinic and benzylic C-H substrates with PFIB as oxidant and Fe^III(TBP₈Cz) as catalyst.¹⁰⁹

Scheme 24.

Synthesis of Fe^III(TBP₈Cz) and Fe^IV(O)(TBP₈Cz^•+)

Figure 2.

Left: Time-resolved UV-vis spectrum upon addition of PFIB to Fe^III(TBP₈Cz) in CH₂Cl₂/CH₃OH at −78 °C. Right: X-band EPR spectrum of Fe^IV(O)(TBP₈Cz^•+) at 15 K. Reprinted with permission from J. Am. Chem. Soc., 2009, 131 (23), pp 8040–8048. Copyright 2009 American Chemical Society.

EPR¹¹⁰ and X-ray absorption spectroscopy (XAS) measurements¹¹² of this iron-oxo corrolazine led to an assignment of an S = 1 Fe^IV metal center antiferromagnetically coupled to an S = ½ corrolazine π-radical cation, giving a total spin ground state of S = 1/2. Best fits to the EXAFS data showed a short Fe–O bond distance of 1.64 Å, in excellent agreement with the Fe^IV(O) assignment. The Fe^IV(O)(TBP₈Cz^•+) complex is stable at −78 °C for hours, but reverts to Fe^III(TBP₈Cz) at room temperature. This iron-oxo corrolazine complex is capable of O-atom transfer, H-atom abstraction, and electron transfer (Scheme 1).

Addition of PPh₃ to this complex at −78 °C converts the species back to the starting Fe^III complex. The ability of Fe^IV(O)(TBP₈Cz^•+) to cleave activated C-H bonds was studied.¹¹² A KIE of 5.7 was obtained for xanthene, and a linear correlation between the rate constants (log k) and BDE(C-H) of the C-H substrates show that these reactions occur via a rate-determining H-atom transfer. However, the putative Fe^IV(OH) intermediate was not observed. A temperature dependence on the product distribution was observed for the reaction between Fe^IV(O)(TBP₈Cz^•+) and xanthene, where the oxygen-rebound product (xanthydrol) is favored at lower temperatures (−25 °C) over the radical dimerization (9,9’bixanthene) product. Thus, lowering the reaction temperature increases the efficiency of oxygen rebound.

2.2. Manganese-Oxo Complexes

2.2.1. Manganese-Oxo Porphyrins

Iron-oxo intermediates are found in the catalytic cycles of heme enzymes, and Mn has been used as an analog of Fe in models of these systems in part due to the enhanced stability of Mn-oxo species over the corresponding Fe-oxo species. While Mn(porphyrins) have yet to be identified as natural cofactors in metalloenzymes, synthetic manganese porphyrins have been used as catalysts in the biomimetic oxidation of organic substrates. These reactions usually proceed through a high-valent metal-oxo intermediate in mechanisms similar to those proposed for heme enzymes. Manganese-oxo porphyrins have been identified in the Mn^IV-oxo and Mn^V-oxo states. An early example of a well-characterized manganese(IV)-oxo porphyrin complex was reported in 1995, in which the Mn^IV(O) complex was synthesized from oxidation of the Mn(III)porphyrin with potassium peroxycarbonate, and structurally characterized by XAS.¹¹³ Typically, Mn^IV(O) porphyrin species are less reactive than their Mn^V(O) counterparts.³¹ Spectroscopic characterization of a Mn^V(O) porphyrin species remained elusive until relatively recently.^114,115 The complex (Mn^V(O)(TMPyP) (TMPyP = meso-tetrakis(4-methylpyridinium)porphyrinato²⁻)) was generated in aqueous solution at 25 °C and characterized by UV-vis and ¹H NMR spectroscopy. High-valent Mn-oxo porphyrins, including the previous example, have been reviewed in The Porphyrin Handbook (2000-2003)^18,31 and Chemical Reviews^116,117 and won’t be reiterated here.

In 2002, Nam and coworkers presented UV-vis and EPR evidence for the formation of a Mn^V(O) porphyrin complex from the reaction of Mn^III(por) and H₂O₂ in aqueous solution (Scheme 26).¹¹⁸ The reaction between Mn^III(TF₄TMAP) (TF₄TMAP = meso-tetrakis(2,3,5,6-tetrafluoro-N,N,N-trimethyl-4-aniliniumyl)porphyrinato²⁻) and H₂O₂ was highly pH dependent, with O–O bond homolysis proposed to occur at low pH to generate an Mn^IV(O)(TF₄TMAP) and O–O bond heterolysis proposed at high pH to generate an Mn^V(O)(TF₄TMAP). These two high-valent species could be differentiated by their UV-vis and NMR spectra. When alkyl or acyl hydroperoxide was used as oxidant, the electron-donating or -withdrawing properties of the oxidant controlled the mechanism of O–O cleavage, with the former cleaving homolytically to generate the Mn^IV(O) species and the latter cleaving heterolytically to give the Mn^V(O) species. The high-valent Mn-oxo species also had differing reactivity. The Mn^V(O)(TF₄TMAP) was a very efficient epoxidizing agent, whereas the lower-valent Mn^IV(O)(TF₄TMAP) was unable to oxidize alkenes. This result was consistent with earlier findings that a Mn^IV(O)(TM-2-PyP) (TM-2-PyP = meso-tetrakis(N-methyl-2-pyridyl)porphyrinato²⁻) complex was unable to oxidize olefins.¹¹⁵ Electrochemical evidence for Mn^V(O)(TF₄TMAP) generated from H₂O₂ and Mn^III has been obtained.¹¹⁹

Scheme 26.

Synthesis of High-Valent Manganese-Oxo Species Depends on the Identity of the Oxidant

A dimanganese complex of a dimeric tetraaryl porphyrin was reported by Naruta and coworkers to generate a dinuclear Mn^V(O) complex (Scheme 27).¹²⁰ When two equivalents of mCPBA were added to the Mn^III₂ porphyrin dimer [Mn₂(DTMP)(OH)] (TMP = meso-(tetramesityl) porphyrinato²⁻) in CH₂Cl₂/CH₃CN (1:1 v/v), a new species assigned as an (Mn^V(O))₂ species was generated. This new species was stable in the presence of excess tetrabutylammonium hydroxide; however, when excess triflic acid was added to the complex in the presence of 10% water, fast reduction to a dinuclear manganese(III) species occurred with simultaneous evolution of O₂. A 92% yield of O₂ relative to the starting dimeric complex was obtained by mass spectrometry and when 18-O labeled water/hydroxide was used, ¹⁸O₂ was observed. Based on the available data, the authors proposed that the (Mn^V(O))₂ species underwent O–O bond formation to evolve dioxygen, either by nucleophilic attack of solvent water on an Mn^V(O)(H₂O) species or by coupling the two Mn^V(O) units.

Scheme 27.

Oxygen Evolution from a Dimanganese-oxo Porphyrin Complex

Evidence for the formation of a Mn^V(O) porphyrin species has come from laser flash photolysis (LFP) studies.¹²¹ It is important to note that in the following photochemical studies, the high-valent Mn-oxo species were characterized mainly by transient UV-vis spectroscopy, and further characterization (e.g. Raman, XAS) is needed to confirm their identity. Irradiation of a Mn^III(TPFPP) (TPFPP = meso-tetrakis(pentafluorophenyl)porphyrinato²⁻) in acetonitrile in the presence of the perchlorate anion resulted in the formation of a transient UV-vis spectrum for a new species assigned as Mn^V(O)(TPFPP) based on its reactivity. In the presence of alkene substrates, this species decayed back to Mn^III(TPFPP) at a rate faster than the self-decay process. The new species was able to oxidize stilbene to stilbene oxide with high stereoselectivity in a >95:5, cis:trans ratio. Furthermore, when ethylbenzene and ethylbenzene-d₁₀ were used as substrates, a kinetic isotope effect of 2.3 was obtained, implicating C–H cleavage in the rate determining step. The LFP generation method was subsequently applied to a wider series of manganese porphyrin complexes (Mn^III(TPFPP), Mn^III(TPP), and Mn^III(TMPyP)) in which the electron-donating properties were varied (Scheme 28).¹²² When LFP was initiated in the presence of NO₃⁻ or ClO₃⁻, a species assigned as an Mn^IV(O)(por) complex was obtained, which results from the homolytic cleavage of the N–O or Cl–O bonds. LFP in the presence of ClO₄⁻ resulted in the formation of a species assigned as an Mn^V(O)(por) complex, which results from heterolytic Cl–O cleavage. The presence of water further stabilized the higher-valent species. The reactivity of the postulated Mn^V(O)(por) species were studied with various substrates. Smooth two-electron chemistry was observed, with no intermediacy of an Mn^IV(O)(por) species. The reaction rates with a given substrate depended on the electronics of the porphyrin ligand. The most electron-withdrawing porphyrin was TPFPP, and the Mn^V(O)(TPFPP) species was the most reactive, followed by the Mn^V(O)(TMPyP) and Mn^V(O)(TPP) species. The Mn^IV(O)(por) species were found to be less reactive than the Mn^V(O)(por) species. Furthermore, the effect of the electronics of the ligand in reactions with substrates was found to be reversed relative to the Mn^V(O)(por). Thus, the rates of reaction were found to be as follows: Mn^IV(O)(TPP) > Mn^IV(O)(TMPyP) > Mn^IV(O)(TPFPP). This surprising trend was rationalized by the rapid disproportionation equilibrium of Mn^IV(O)(por) to form Mn^V(O)(por) and Mn^III(por)(X), with the Mn^V(O)(por) performing the actual oxidation of the substrate. The disproportionation equilibrium controls the amount of Mn^V(O)(por) in solution and therefore the “apparent” rate of the reaction. A related photochemical method of generating high-valent Mn-oxo porphyrins was reported for a dinuclear Mn^III complex. Upon visible light irradiation, bis-porphyrin dimanganese(III)-μ-oxo complexes could be photo-disproportionated to Mn^IV(O) porphyrin and Mn^II porphyrin.¹²³

Scheme 28.

Dependence of O–Cl Bond Cleavage on Starting Manganese Porphyrin Axial Ligand

The effects of oxo-hydroxo tautomerism (Scheme 29) on the reactivity of Mn^V(O) porphyrins were studied in two isomeric porphyrin complexes.¹²⁴ It was found that Mn^V(O)(X)(4-TMPyP) was more reactive with Br⁻ ions to form hypobromite (BrO⁻) than the less electron-rich Mn^V(O)(X)(2-TMPyP). This result is due to the higher basicity of the Mn^V(O)(4-TMPyP) complex, resulting in the presence of more Mn^V(O)(H₂O) (which is the active oxidizing species) rather than Mn^V(O)(OH) in this complex. This finding is consistent with the idea that an anionic axial donor would make the Mn^V(O) less electrophilic. Density functional theory (DFT) calculations of the possible spin states (singlet, triplet, and quintet) of the three possible Mn^V-oxo isomers (dioxo, oxo-hydroxo, and oxo-aquo complexes) reveal a singlet ground state for all three complexes. However, for the oxo-hydroxo and oxo-aquo complexes, the triplet and quintet states are relatively close in energy to the singlet ground state. This small energy difference could allow for a possible spin-crossing event during oxygen transfer, leading to enhanced rates of reaction with substrates.

Scheme 29.

pH-Dependent Oxo-Hydroxo Tautomerism in High-Valent Manganese-Oxo Porphyrins

The importance of spin states on Mn-oxo reactivity has been studied extensively by DFT. A DFT study on C–H hydroxylation by Mn^V(O)(X)(TPP) (X = O²⁻, OH⁻, and H₂O) suggested that a rebound mechanism is possible in the triplet and quintet states, but not along the singlet pathway.¹²⁵ The higher selectivity for the rebound mechanism in the triplet and quintet states was attributed to an increased oxyl radical character found in these higher spin-states, which caused the reaction to occur in two distinct one-electron steps, with a spin cross-over from singlet to triplet in the initial step.¹²⁶ The extent of oxyl radical character influenced the bond order of the Mn–O interaction. The bond order was found to increase with protonation of the axial ligand X (X = O²⁻, OH⁻, and H₂O) according to the following series Mn^V(O)₂<Mn^V(O)(OH)<Mn^V(O)(H₂O).¹²⁷ However, the multiplicity of the ground states and the extent of oxyl character were highly dependent on the functional used.¹²⁶ Furthermore, the effect of the axial ligand in the calculations was found to decrease the singlet-triplet and singlet-quintet gaps along this series, and this gap is the dominant factor in controlling the calculated reactivity, with the Mn^V(O)(H₂O) species as the most reactive, as found experimentally.^124,128 The protonated bis-hydroxy complex Mn^V(OH)(OH)(por) was also found to be a viable, potent C–H hydroxylation oxidant by DFT.^129,130 Recently, an ab initio benchmarking study on the spin-state energetics and oxyl character of Mn-oxo porphyrins was reported, in which the authors suggest that the best electronic description of these species can be achieved with pure functionals.¹³¹

N-donor axial ligands are common additives to C–H bond activation reactions catalyzed by Mn-porphyrins, for which Mn^V(O)-porphyrin intermediates are often proposed. Addition of axial N-donors is in part motivated by the prominent role played by the axial ligands (e.g. histidine, tyrosine, cysteine) in heme enzymes. A DFT study on the effect of Mn–O bond length versus axial N–donor strength found that the Mn–O bond lengthens with decreasing Mn–N_axial bond distance, indicating that the stronger the axial ligand, the weaker the Mn–O bond, which in turn leads to predicted enhanced reactivity.¹³² Furthermore, the ability of the axial N-donor to interact with the meso-phenyl rings of an Mn(O)(por) through non-covalent interactions, such as hydrogen-bonding, was also an indicator of reactivity.¹³³ DFT calculations showed that the ortho-C–F bonds in the phenyl rings of Mn^V(O)(TPFPP) were able to hydrogen-bond with the C–H groups of axially ligated pyridine derivatives, providing a stronger Mn–N_axial interaction, and a lower barrier for C–H activation.

An unusual substrate for Mn^V(O) porphyrin is the Br⁻ anion. It was shown that Mn^V(O)(TDMImP) (TDMImP = meso-tetrakis(dimethylimidazolium)porphyrinato²⁻) was capable of oxidizing the bromide ion to form hypobromite (OBr⁻), and this reaction was reversible.¹³⁴ The equilibrium for the reaction Mn^V(O)(TDMImP) + Br⁻ ⇌ Mn^III(TDMImP) + OBr⁻ could be controlled by changing the pH. The formation of the Mn^V(O)(TDMImP) and Br⁻ was favored under basic conditions, whereas the formation of Mn^III(TDMImP) and the O–Br bond was favored under acidic conditions.

Nam, Solomon, and coworkers synthesized a high-valent Mn porphyrin complex by reaction of Mn(III) porphyrins with one of various oxidants including mCPBA, iodosylarenes, and hydrogen peroxide. This product was formulated as an Mn^V(O) complex based on UV-vis, EPR, ¹H and ¹⁹F NMR, resonance Raman, and XAS.¹³⁵ The XAS data led to identification of an Mn–O distance of 1.68 Å for the proposed Mn^V(O)(TDCPP) (TDCPP = meso-tetrakis(2,6-dichlorophenyl)porphyrinato²⁻) and indicated that the central Mn was likely 6-coordinate with an axial ⁻OH ligand. However, the Mn–O distance of 1.68 Å was significantly longer than other Mn^V–O distances, including that reported for a related Mn^V(O) corrolazine.¹³⁶ Reactions of the high-valent Mn complex with various substrates were analyzed kinetically. A linear Hammett plot was obtained for reactions with para-substituted thioanisoles of varying electron-donating or electron-withdrawing substituents. The negative slope for this plot indicated that the Mn complex was electrophilic in nature. When the reaction with cyclooctene and cyclooctane was examined, no oxidized organic products were observed. The authors postulated that, in light of the EXAFS results showing that an axial ligand was present, an anionic ⁻OH ligand could decrease the electrophilicity of the proposed Mn^V(O)(TDCPP) complex and stabilize it. However, the structure of this complex was not assigned correctly and it was subsequently reformulated as a trans-dioxo Mn^V porphyrin complex, based on reinterpretation of the spectroscopic data and independent generation of another example of a trans-dioxo Mn^V porphyrin complex.¹³⁷

The formal bonding picture of Mn^V(O) porphyrins, known since the 1960s, predicted that mono-oxo complexes should be stable species, but trans-dioxo Mn^V complexes would be inaccessible. However, an example of a trans-dioxo Mn^V was reported by Groves and Spiro in 2007. Trans-dioxo Mn^V-oxo porphyrins were generated by exploiting pH-dependent oxo-hydroxo tautomerism.¹³⁸ Under highly basic conditions, a series of trans-dioxo Mn^V(O) porphyrin complexes were generated in organic solvent. These species were half- or fully-labelled isotopically by using ¹⁸O-sources, allowing for the key identification of the symmetric and asymmetric Mn–O stretches by resonance Raman and IR spectroscopy. A correlation could be obtained between the force constant for Mn–O stretches and Mn–O bond lengths for a series of comparable manganese-oxo complexes. This correlation, known as Badger’s rule, allowed for a bond length prediction of ~1.7 Å for the Mn=O double bonds.

The isolated trans-dioxo complexes were unreactive towards olefins, but addition of one equivalent of trifluoroacetic acid caused instantaneous reaction with cyclooctene at −70 °C. The authors suggested that the acid protonated one of the oxo ligands and lowered the net negative charge on the [Mn^V(O)₂]⁻ complex, thereby enhancing its electrophilicity. While unreactive with olefins, trans-dioxo Mn^V-oxo porphyrins were found to perform hydride abstraction from dihydronicotinamide adenine dinucleotide (NADH) analogues, such as 10-methyl-9,10-dihydroacridine (AcrH₂).¹³⁹ The hydride abstraction rates were sensitive to the electron-donating properties of the AcrH₂ derivative, as well as the electronics of the porphyrin ligand, with the most electron-deficient porphyrin having the fastest hydride transfer rates. Large deuterium kinetic isotope effects were also observed for the AcrH₂ substrates. The kinetic data indicated that the hydride transfer reaction was occurring through a mechanism involving initial rate-determining proton-coupled electron transfer followed by fast electron transfer.

Manganese-porphyrins were found to photocatalytically oxygenate AcrH₂, exploiting the facile reaction between Mn^V(O)(por) and this hydride donor.¹⁴⁰ A mechanism was invoked in which the resting state of the catalyst, the Mn^III(por), is irradiated and generates a photoexcited [Mn^III]* which reacts with O₂ to generate a Mn^IV-superoxo species (Scheme 30). The superoxo species cleaves the C–H bond of the AcrH₂ to generate a Mn^IV-hydroperoxo species and the acridinyl radical AcrH•. The O–O bond of the Mn^IV-OOH species goes through oxygen rebound homolysis to generate the Mn^V(O) species and a hydroxylated AcrH–OH molecule. The Mn^III(por) catalyst is regenerated by the further oxidation of the AcrH–OH molecule to the ketone, acridone, or by the oxidation of another AcrH₂ molecule.

Scheme 30.

Proposed Mechanism of Photocatalytic Oxygenation of AcrH₂ by Manganese Porphyrins

Early work on high-valent Mn–oxo porphyrins showed that Mn^V(O) and Mn^IV(O) porphyrins had distinct reactivity, with the lower valent Mn^IV(O) complexes being less reactive in oxidations such as epoxidation of alkenes.³¹ A comparison of the differing reactivity in OAT and HAT reactions for trans-dioxo Mn^V and Mn^IV(O) generated in aqueous solutions was described by Fukuzumi and Nam. The trans-dioxo complex Mn^V(O)₂(tf₄tmap) (tf₄map = meso-tetrakis(2,3,5,6-tetrafluoro-N,N,N-trimethyl-4-aniliniumyl)porphyrinato²⁻), and the oxo-hydroxo complex, Mn^IV(O)(OH)(tf₄tmap), were generated in H₂O by oxidation with H₂O₂ and tert-butyl hydroperoxide, respectively, and characterized by UV-vis and EPR spectroscopies. The two high-valent Mn(O) complexes were reacted with various C–H substrates and O-atom acceptor reagents and the kinetics of these reactions were analyzed (Scheme 31).¹⁴¹ It was found that only the Mn^V(O)₂ complex was capable of OAT to thioanisole substrates, and this reaction gave a Mn^III product and the corresponding methyl phenyl sulfoxide. A linear Hammett plot for para-substituted thioanisole derivatives showed a negative slope which indicated a mechanism where the nucleophilic sulfide substrate attacks the electrophilic Mn^V(O)₂. The Mn^IV(O)(OH) species was able to perform OAT to triarylphosphines, a highly nucleophilic substrate, to give a Mn^II product and triphenylphosphine oxide. When C–H substrates were added to Mn^V(O)₂(tf₄map), the product of the H-atom abstraction reaction was Mn^IV(O)(OH)(tf₄tmap). When C–H substrates were added to the independently generated Mn^IV(O)(OH)(tf₄tmap), slow formation of Mn^III(tf₄tmap) was observed. No substrate hydroxylation was observed for either Mn-oxo species. A mechanism was suggested from DFT calculations by showing that the energy barrier for the oxygen rebound step is higher than that for the escape of the carbon radical from the cage. These DFT calculations along the reaction coordinate showed that oxygen rebound was disfavored, especially for a dioxo-manganese complex Mn^V(O)₂ relative to the oxo-hydroxo complex Mn^V(O)(OH), which suggests that the strongly bound oxo (O²⁻) ligand may play a role in controlling the reaction barrier associated with oxygen rebound.¹⁴² The H-atom abstraction rate constants for both Mn^V(O)₂ and Mn^IV(O)(OH) complexes were found to be linearly dependent on the strength of the C–H bond, with the stronger C–H bonds giving slower reaction rates, and large kinetic isotope effects (k_H/k_D = 8-24) for the deuterated substrates were also observed. These data support C–H cleavage as the rate-determining step and implicate a hydrogen atom transfer mechanism for both high-valent Mn-oxo species. The authors conclude that the reaction between H-atom donor and Mn^IV(O)(OH)(tf₄tmap) was direct, and, under these conditions, does not proceed via a disproportionation pathway as proposed earlier by Newcomb.¹²² In contrast, when the hydride donor AcrH₂ was added to a different Mn^IV(O) complex, Mn^IV(O)(TMP), a disproportionation mechanism was invoked in which the Mn^IV(O) is in equilibrium with Mn^III and Mn^V(O) (Scheme 32). The Mn^V(O) species was proposed to be the active oxidant in the reaction.¹⁴³ The reaction of Mn^IV(O)(TMP) with ferrocene derivatives, which are one-electron reductants, was proposed to go through a direct electron-transfer mechanism, and did not involve disproportionation. DFT calculations provided some insight into the different reaction pathways observed for Mn^V(O) vs. Mn^IV(O) with AcrH₂.¹⁴⁴ In the case of Mn^IV(O), an H-atom abstraction is more thermodynamically favored than overall hydride transfer, whereas with Mn^V(O), hydride abstraction is more favored. The Mn^V(O) was described as the more potent oxidant with that substrate. The different reaction pathways of these Mn^IV(O) porphyrins were highly controlled by the nature of the substrate.

Scheme 31.

Reactivity Comparison of Mn^V(O) and Mn^IV(O) Porphyrins

Scheme 32.

Substrate Dependence on Reactivity of a Mn^IV(O) Porphyrin

The reactivity of Mn^V(O) porphyrins with substrates is complicated by the possible presence of axial ligands (H₂O, OH⁻, or O²⁻) that are present in buffered aqueous solution. A “naked” 5-coordinate Mn^V(O)(TPFPP) complex was generated from Mn^III(TPFPP) and iodosylbenzene in the gas phase in methanol by electrospray ionization and the reactivity was monitored by Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometry.^145,146 The Mn^V(O)(TPFPP) was capable of OAT to olefins to give the reduced Mn^III(TPFPP) complex. The rate of OAT depended on the ionization energy of the olefin, with the fastest reactions occurring for olefins with lower ionization energies. Furthermore, selectivity of the gaseous Mn^V(O)(TPFPP) complex for Z isomers over E isomers of the alkene was also observed. In negative ion mode, the gas-phase trans-dioxo Mn^V(O)₂(TPFPP) complex could be generated, and was found to be completely unreactive towards olefins and sulfides, as found previously in aqueous solution.¹³⁸ The Mn^V(O)(TPFPP) complex was also found to oxidize p-substituted thioanisoles and perform N-dealkylation of N,N-dialkylamines.¹⁴⁷ The sulfoxidation reaction rates were found to depend on the electronics of the substrate. A negative linear correlation between the Hammett σ-constant of the p-substituted thioanisole and the log of the rate constants could be obtained. The reaction rates between Mn^V(O)(TPFPP) and p-substituted dimethylanilines were also linearly correlated with the Hammett σ⁺-parameters of the substrates, and the slope of the line had a negative value. In the case of this substrate, an initial electron-transfer mechanism from the dimethylaniline substrate to the Mn^V(O) species was proposed. The reaction rates slowed with increasing ionization energy of the dimethylaniline, providing further support for this mechanism.

High-valent Mn-oxo porphyrins are typically generated via O-atom transfer from an appropriate activated O-atom transfer reagent (mCPBA, PhIO) to the Mn^III(por). However, Fukuzumi and coworkers showed that water can be used as the O-atom source in the presence of a one-electron oxidant, [Ru(by)₃]³⁺.¹⁴⁸ Oxidation of a series of Mn^III-porphyrins in acetonitrile with added water generated either Mn^IV(O) or Mn^V(O) porphyrins, depending on the axial ligands present in the starting Mn^III(por). If Cl⁻ was present, the Mn^IV(O) species was formed, whereas if H₂O was the axial ligand, the Mn^V(O) species was formed. The Mn^III-porphyrin/oxidant/water system could also serve as a catalyst for epoxidation and alkane hydroxylation reactions, in which the Mn^V(O)(por) species is implicated as the active oxidant (Scheme 33).

Scheme 33.

Water as the O-atom Source in Manganese Porphyrin-Catalyzed Oxidations

Another unique set of oxidants, the mild, single-oxygen atom donors tetra-n-butylammonium periodate (n-Bu₄NIO₄) and tetra-n-butylammonium hydrogen monosulfate (n-Bu₄NHSO₅), were found to generate high-valent Mn-oxo porphyrins.^149,150 The stability and formation of the high-valent Mn(O) species in CH₂Cl₂ were found to be highly dependent on the sterics and electronics of an added alcohol co-solvent (Scheme 34). It was found that the rate of formation of Mn^V(O) species was faster for small linear alcohols than for the larger ones. This observation indicated that the hydrogen bonding between the alcohols and the periodate ion coordinated to Mn facilitated the heterolytic cleavage of the I–O bond. Furthermore, the authors found that the stronger the acidity of the alcohol, the better the hydrogen-bonding interaction. The stability of the Mn^V(O) species was enhanced generally by the presence of alcohols, but especially in the presence of bulkier alcohols. These results pointed to hydrogen bonding between the Mn=O unit and the H–O of the alcohols.

Scheme 34.

Synthesis of an Mn^V(O) Porphyrin Complex with Periodate as Oxygen Atom Donor in the Presence of Alcohol Cosolvents

Manganese(V)-oxo porphyrins have been found to oxygenate chloride ions, leading to a variety of halogenation chemistry with these molecules. A water soluble manganese porphyrin Mn^III(TDMImP) was found to catalytically generate chlorine dioxide (ClO₂) from chlorite ion (ClO₂⁻).¹⁵¹ The proposed mechanism (Scheme 35) involved oxygen transfer from ClO⁻ or ClO₂⁻ to Mn^III(TDMImP) to give Mn^V(O)₂(TDMImP), which can oxidize a molecule of ClO₂⁻ to give Mn^IV(O)(TDMImP) and one molecule of ClO₂. The Mn^IV(O)(TDMImP) intermediate then reacts with ClO₂⁻ to give another ClO₂ molecule. To provide support for this mechanism, chlorite ion added to authentic Mn^IV(O)(TDMImP) underwent facile reaction under the catalytic conditions to produce ClO₂ and the Mn^III(TDMImP) catalyst. The rate-determining step was determined to be the initial oxygen atom transfer to Mn^III(TDMImP) from the chlorite ion to give the high-valent Mn^V(O)₂ species.¹⁵² Halogenation of unactivated C–H bonds was observed with a Mn-porphyrin catalyst, and the presence of high-valent Mn^V(O) species has been observed in the UV-vis spectra during the reactions, although the direct C–H cleavage step has not been directly observed.¹⁵³

Scheme 35.

Catalytic Chlorite Oxidation by a Mn(III) Porphyrin

2.2.2. Manganese-Oxo Corroles and Corrolazines

Manganese-oxo corroles have been studied as meta-stable analogues of their reactive porphyrin counterparts. Relatively stable Mn^V(O) corroles can be synthesized from the Mn^III precursor and an appropriate O-atom donor. These species are low-spin diamagnetic S = 0 complexes and the lifetime of the Mn^V(O) depends on the corrole ligand and solvent. Manganese-oxo corroles have been found to participate in the catalytic oxidation of substrates, and the reaction pathways depend highly on the electronic nature of the ligand. Some recent reviews have covered the reactivity of these complexes, including a chapter in the Handbook of Porphyrin Science in 2011,¹⁷ which covers high-valent transition metal corroles and corrolazines, and a Coordination Chemistry Review¹⁵⁴ in 2013, which exclusively covers recent advances in manganese corrole chemistry. Since those reviews were published, several reports on the reactivity of Mn-oxo corroles have further advanced the field.

A survey of the reactivity of a series of Mn^V(O) corroles with a series of alkene substrates revealed a strong dependence of the rate constant as well as the mechanism of OAT on the solvent.^155,156 The product of decomposition, which typically changes with ligand identity, was also found to be solvent dependent. The more electron-rich Mn^V(O)-corroles decomposed to Mn^IV(X) species in toluene and dichloromethane, but in dimethylformamide and dimethylacetamide, the product of decomposition was an Mn^III corrole. In toluene and dichloromethane, the rate of decomposition was faster for the electron-deficient corroles, but in dimethylformamide and dimethylacetamide the trend was reversed. Addition of alkene to the self-decay reaction of Mn^V(O) was found to greatly accelerate them, with the most electron-rich and least-substituted alkenes giving the fastest rates. The OAT reaction rates between alkene and Mn^V(O)corroles in either CH₂Cl₂ or C₆H₅CH₃ depended on the electron-donating properties of the corrole, with the most electron-deficient corroles giving the fastest rates. This trend in OAT rates was reversed in dimethylformamide and dimethylacetamide. The authors attributed this reversed trend to a change in mechanism from direct OAT in toluene and dichloromethane to a disproportionation mechanism in dimethylformamide and dimethylacetamide, in which the Mn^V(O) disproportionates to an Mn^IV(X) and an Mn^VI(O) species, which could be the actual oxidizing agent.

The effect of peripheral substituents on the corrole ring on the reactivity of manganese-oxo corroles has been studied. The effect of β-bromination on the OAT reactivity of a series of A₃ and A₂B - meso-phenyl substituted Mn^V(O) corroles was determined by Chang, Liu, and coworkers.¹⁵⁷ It was found that the brominated Mn^V(O) corroles performed OAT at rates much higher than their non-brominated analogues. When the meso positions contained ortho-ortho′-dibromophenyl groups, the steric protection of these substituents provided enhanced stability for the Mn^V(O) species. DFT calculations provided insight into the mechanism of various β-substituted (with Br, H, or CH₃) high-valent Mn-oxo corrole complexes reacting with dimethylsulfide, an O-atom acceptor. Liu and coworkers studied different oxidation levels of the corroles by DFT including [Mn^IV(O)(tpfc)]⁻ (tpfc = 5,10,15-tris(pentafluorophenyl)corrolato³⁻), Mn^V(O)(tpfc), and the one electron-oxidized [Mn^V(O)(tpfc^•+)]⁺ complexes.¹⁵⁸ The calculated barriers suggested that the OAT reactivity followed the order: [Mn^V(O)(tpfc^•+)]⁺ > Mn^V(O)(tpfc) > [Mn^IV(O)(tpfc)]⁻, indicating that the overall charge on the complex plays an important role in the mechanism. Furthermore, the more electron-deficient corrole ligand gave a more reactive complex, consistent with earlier DFT work¹⁵⁹ as well as experimental studies. Spin states were also important for determining the energies of the barriers for the reaction between Mn^V(O)(tpfc) and DMS. The barrier for the triplet state, best described as high-spin Mn^V(O), was found to be lower in energy than the singlet state, implicating the possibility of a spin crossover event along the reaction pathway. This result is consistent with earlier calculations from Cao and coworkers that found that while the ground state of the starting Mn^V(O) corrole is the singlet, the OAT triplet pathway has a lower barrier and therefore might play a significant role in the reaction.¹⁶⁰

High-valent manganese-oxo and -imido corroles were shown to participate in the two-electron reduction of O₂ to H₂O₂, serving as precursors to the catalytically active manganese(III) corrole (Scheme 36).¹⁶¹ The addition of trifluoroacetic acid to a high-valent Mn^V-imido complex, Mn^V(NAr)(tpfc) (NAr = 2,6,-dichlorophenylimido²⁻), resulted in hydrolysis by residual water and production of a complex assigned as a protonated Mn-oxo corrole, [Mn^V(OH)(tpfc)]⁺. However, direct evidence for the structure of this complex was not obtained. This complex can be reduced rapidly to Mn^III(tpfc) by the reductant octamethylferrocene. The reduced Mn^III(tpfc) reacts with O₂ to give a putative Mn^IV-superoxo species, which can obtain a proton and electron from TFA and octamethylferrocene, respectively, giving an Mn^IV-hydroperoxo species. Rather than undergoing O–O bond cleavage, it was suggested that the Mn^IV-hydroperoxo species is protonated and hydrogen peroxide is released. Reduction of the resulting Mn^IV(tpfc) allows for the catalytic cycle to begin again. Others have observed the formation of Mn^V(O) corroles from Mn^III(corroles) during the metallation of free-base corroles with Mn^II sources in the presence of O₂, although the mechanism was not elucidated.¹⁶²

Scheme 36.

Proton-Coupled Reduction of Mn^V-imido Corrole to Give a Mn^III Corrole

The addition of trifluoroacetic acid to Mn^V(O)(tpfc) by Fukuzumi, Abu-Omar, and coworkers generated an electronic isomer of this complex, characterized as an Mn^IV(OH)(tpfc^•+) species, in which the oxo ligand was protonated.¹⁶³ The comparative reactivity of these two valence tautomers was studied (Scheme 37). There was little or no difference for the two complexes in H-atom transfer rates with a substituted phenol as H-atom donor. However, OAT from Mn^V(O)(tpfc) to thioanisole was extremely facile, whereas the Mn^IV(OH)(tpfc^•+) species was unreactive with thioanisole in the same time frame. In contrast, the electron-transfer reactivity exhibited the opposite trend, with the Mn^IV(OH)(tpfc^•+) complex having the faster rates with the reductant ferrocene.

Scheme 37.

Comparison of Reactivity for Two High-Valent Manganese-Oxo Corrole Valence Tautomers

Another method of entry into high-valent Mn-oxo corroles was found by Zhang and coworkers, in which visible light irradiation of Mn^IV(BrO₃⁻)(tpc) (tpc = triphenylcorrolato³⁻) and Mn^IV(BrO₃⁻)(tpfc) complexes resulted in Br–O bond homolysis to give the Mn^V(O)-corroles (Scheme 38).¹⁶⁴ This kind of light-induced O–X bond homolysis was used previously by Newcomb and coworkers.^121,122 The reactivity of the photo-generated species was then studied with various organic substrates. The products of the reaction of photo-generated Mn^V(O)(tpfc) with organic substrates were highly solvent-dependent, with Mn^III(tpfc) as the product in CH₃CN and Mn^IV (tpfc) as the product in CH₂Cl₂. In contrast, the reaction of the more electron-rich Mn^V(O)(tpc) with organic substrates gave Mn^IV(tpc) as the only product regardless of solvent. Despite the complexity of products, the kinetics of the reactions of Mn^V(O) species with organic substrates all followed a simple second-order rate law. The reaction of Mn^V(O)(tpfc) with para-substituted thioanisoles resulted in a linear Hammett plot with a negative slope, which indicated a mechanism in which nucleophilic thioanisole reacted with an electrophilic Mn-oxo complex. When the rates of Mn^V(O)(tpfc) and Mn^V(O)(tpc) are compared, the Mn^V(O)(tpc) was generally the faster oxidant. Because the tpc ligand is more electron-rich than tpfc, and the reactivity of electrophilic metal-oxo species is generally enhanced with a more electron-poor ligand, this result was unexpected. The reactivity was therefore rationalized by the disproportionation of Mn^V(O) to Mn^IV(X) and Mn^VI(O), with Mn^VI(O) proposed as the actual oxidant (Scheme 39). This idea also explains the ability of Mn^V(O) to promote the facile 2-electron oxidations of alkene and sulfide substrates to give the Mn^IV(X) species, which is only one oxidation state below Mn^V(O). However, direct experimental evidence of the putative Mn^VI(O) oxidant remains elusive.

Scheme 38.

Generation of Mn^V(O) Corroles by Br–O Bond Homolysis of Mn-Bromate Complexes

Scheme 39.

Disproportionation of Mn^V(O) Corrole to Give Mn^IV(X) and Mn^VI(O) Corrole Complexes

The effects of axial ligation on high-valent manganese-oxo corroles have been studied by ligands with tethered nitrogenous-bases.¹⁶⁵ The reactivity between the 6-coordinate acetamido-, pyridyl-, and imidazolyl- appended Mn^V(O) corroles and alkenes was found to be enhanced relative to the 5-coordinate Mn^V(O)corrole analogue. The reaction rates increased with increasing axial donor strength in the order: acetamido < pyridyl < imidazolyl. It was postulated that the coordination of the axial ligand destabilizes the Mn^V(O) species, which then increases the rate of OAT. These appended corroles were also capable of catalytic epoxidation, and the turnover frequency (TOF) also increased with the strength of the axial donor, following the same trend as the stoichiometric reactions.

The chemistry of high-valent manganese-oxo corrolazines was reviewed in 2011 in the Handbook of Porphyrin Science¹⁷ and a survey of recent work from Goldberg and coworkers on these complexes appeared in 2015 in an Accounts of Chemical Research article.¹⁶⁶ A summary of work since these previous reviews is given here.

Although a number of spectroscopic methods had been used to characterize a high-valent manganese-oxo corrolazine complex Mn^V(O)(TBP₈Cz) (TBP₈Cz = octakis(p-tert-butylphenyl)corrolazinato³⁻) since the initial report of this complex in 2001, it was not until 2015 that the structure was determined by single crystal x-ray diffraction (XRD).¹³⁶ The structure revealed a short Mn–O bond distance of 1.55 Å and the Mn ion was displaced 0.55 Å relative to the plane of the four pyrrole nitrogens. Importantly, the Mn ion was five-coordinate, in contrast to Mn(O) porphyrin complexes, which typically have a sixth ligand.

The effect of modifying the peripheral aryl substituents at the β-position of the corrolazine ring had subtle but measurable effects on the reactivity of the Mn^V(O) complexes.¹⁶⁷ The one-electron reduction potential of Mn^V(O)(MeOP₈Cz) (MeOP₈Cz = octakis(p-methoxyphenyl)corrolazinato³⁻) (E_1/2 = −0.57 V vs Fc⁺/Fc) was shifted negative compared to Mn^V(O)(TBP₈Cz) (E_1/2 = −0.53 V vs Fc⁺/Fc), indicating a slightly more electron rich metal center for the methoxy derivative. A Hammett study of the OAT reaction of Mn^V(O)(MeOP₈Cz) with various p-substituted triarylphosphines to give triarylphosphine oxides revealed a more negative slope than with the Mn^V(O)(TBP₈Cz) derivative, consistent with Mn^V(O)(MeOP₈Cz) as a less electrophilic oxidant. In H-atom abstraction, the Mn^V(O)(MeOP₈Cz) gave slightly faster rates than the Mn^V(O)(TBP₈Cz) complex, and this result was rationalized by an enhanced basicity for the putative [Mn^IV(O)(MeOP₈Cz)]⁻ species which compensates for the more negative redox potential of the Mn^V(O) complex.

The corrolazine scaffold provides additional opportunities for modification through attachment of electrophiles to the meso-nitrogen atoms. An Mn^III(TBP₈Cz) complex could be singly- or doubly- protonated on the meso-nitrogens using triflic acid (HOTf).¹⁶⁸ The mono- protonated [Mn^III(TBP₈CzH)]⁺ complex was the resting state of a catalyst capable of oxidizing the benzyl C–H bonds of toluene derivatives. One proposed intermediate in the catalytic cycle was a protonated high-valent Mn^V(OH)(TBP₈Cz) complex. Addition of HOTf to independently synthesized Mn^V(O)(TBP₈Cz) was found to generate the valence tautomer Mn^IV(OH)(TBP₈Cz^•+). This species oxidized substrates very slowly; however, when excess HOTf was added, it was capable of rapid C–H oxidation by a PCET mechanism consistent with this species being competent during catalysis. From these results, the proposed catalytic cycle is shown in Scheme 40, whereby the photoexcited protonated manganese corrolazine complex [Mn^III(OTf)(TBP₈CzH)]⁺ reacts with dioxygen to generate a putative Mn^IV-superoxo species, which abstracts an H• from a toluene derivative (hexamethylbenzene) to generate a putative Mn^IV-hydroperoxo species and the substrate radical. O–O bond cleavage of the Mn^IV(OOH) species occurs in concert with oxygen rebound to the substrate radical to give the high-valent manganese-oxo species, Mn^IV(OH)(TBP₈Cz^•+), and hydroxylated organic product. The Mn^IV(OH)(TBP₈Cz^•+), in the presence of a second H⁺ equivalent, reacts with another equivalent of hexamethylbenzene to give another equivalent of hydroxylated product along with regeneration of the resting state of the catalyst.

Scheme 40.

Mechanism for Photocatalytic Oxygenation of the Toluene Derivative Hexamethylbenzene by a Protonated Manganese Corrolazine Complex with Dioxygen

Theoretical calculations have been used as insightful tools into the nature of high-valent metal-oxo species. DFT calculations are popular for studying these types of complexes, but pitfalls in these methods remain, especially with respect to describing spin state energetics. Ghosh and coworkers used a higher level of theory to examine the low-lying excited states of Mn^V(O) corroles and corrolazines using ab initio methods. The calculations uniformly indicated that a low-spin (S = 0) d² ground state was favored for these compounds, and confirmed that the corrole and corrolazine macrocycles are unambiguously innocent in the ground states.¹⁶⁹ The next low-lying excited states exhibited a high-spin Mn^V configuration, with the Mn^IV(O) macrocycle-π-cation-radical states much higher in energy. Furthermore, the authors proposed that the meso-nitrogens of the corrolazine do not impart special stability to the ligand, but rather the steric bulk provided by the eight tert-butylphenyl β-substituents allow for the enhanced stability of the Mn^V(O) species relative to the corrole.

DFT calculations have provided some insight into the reactivity of high-valent Mn-oxo corrolazines. The oxidation of C–H bonds by the low-spin Mn^V(O)(TBP₈Cz) showed dramatic rate accelerations in the presence of anionic axial donors.¹⁷⁰ In a separate computational study, Shaik and coworkers proposed that the origin of the rate enhancement was due to a low-energy triplet transition state, as the most likely C–H bond cleavage pathway.¹⁷¹ Higher-level NEVPT2:CAS calculations have shown that the energetically favored ground state is the singlet state for the same six-coordinate complexes, and showed that the singlet-triplet gap is much larger than predicted by the former DFT calculations. The six-coordinate complexes [Mn^V(O)(X)(TBP₈Cz)]⁻ (X = CN⁻, F⁻) also exhibited dramatically enhanced OAT rates relative to the non-axially ligated Mn^V(O)(TBP₈Cz) in reactions with sulfide substrates, including para-substituted thioanisole derivatives.^172,173 Calculations on this OAT reaction predicted that the six-coordinate complex [Mn^V(O)(X)(TBP₈Cz)]⁻ reacted through the low-spin singlet pathway, avoiding spin-crossover during the reaction mechanism.¹⁷⁴ Experimental data showed a rare V-shaped Hammett plot for OAT, and this result could only be reproduced by calculations on the singlet pathway (Figure 3). When the Hammett plot was calculated for the reaction along the triplet pathway, a linear Hammett plot was obtained. This result provided an experimental test for distinguishing spin state pathways.

Figure 3.

Computational Hammett plot for the reaction of singlet and triplet [Mn(O)(H₈Cz)(CN)]⁻ with para-Z-substituted thioanisole derivatives. Data calculated at RIJCOSX-TPSSh-D3/def2-QZVPP/ZORA//RIJCOSX-B3LYP-D3/SDD/BS2 and includes zero-point, thermal, and solvent corrections. (a) Correlation for singlet spin barriers. (b) Correlation for triplet spin barriers. Reproduced with permission from reference ¹⁷⁴. Copyright 2016 American Chemical Society.

3. High-Valent Metal-Hydroxo Complexes

For discrete, isolated synthetic complexes of Mn or Fe, we are not aware of any high-valent (oxidation state = +4 or +5) metal-hydroxo porphyrinoid species. The analogous porphyrin and corrole metal(IV)-halogen (chloride, bromide, iodide, fluoride) complexes have been isolated, but have not exhibited biomimetic reactivity.^154,175,176

4. Metal-Oxygen Complexes

4.1. Iron-Oxygen Complexes

4.1.1. Iron-Oxygen Porphyrins

Many studies of heme-dioxygen chemistry have focused on functional mimics for myoglobin and hemoglobin. Several reviews exist that cover dioxygen binding and activation by synthetic iron porphyrin complexes, including spectroscopic and structural features of the dioxygen adducts as well as their reactivity with proton/electron sources. A comprehensive review was published in Chemical Reviews in 1994 by Momenteau and Reed, covering synthesis and characterization of heme dioxygen adducts and some reactivity.¹⁷⁷ Extensive auto-oxidation reactivity of these species was summarized in a Coordination Chemistry Reviews article by Shikama in 1988.¹⁷⁸ In The Porphyrin Handbook (2000), Watanabe described the reactivity of ferric porphyrin peroxo complexes, and Collman et al. summarized the literature of functional synthetic analogues of terminal oxidases.^33,179 In 2004, a Chemical Review article by Collman¹⁸⁰ summarized the literature on reversibly-formed synthetic heme-dioxygen adducts as functional analogs of the heme enzymes Cytochrome c oxidase, myoglobin, and hemoglobin. The same year, a Chemical Review article by Karlin¹⁸¹ summarized the literature on synthetic heme-copper dioxygen chemistry as models for the active site of Cytochrome c oxidase. It is important to note that this review contained information on well-defined mononuclear heme-dioxygen adducts as a starting point for heme-copper dioxygen chemistry. In 2013, a Coordination Chemistry Reviews article by Ohta and Naruta described their recent progress in understanding the vibrational properties of mononuclear heme-peroxo complexes with regards to their influence on the reactivity of these species.¹⁸² In this section, we focus on recent progress in the reactivity of well-defined dioxygen adducts of mononuclear iron porphyrins since the latest review (2012-present).

A report by Ivanović-Burmazović and coworkers showed that the formation of iron-O₂ porphyrin adducts in DMSO is reversible (Scheme 41).¹⁸³ The addition of excess potassium superoxide to Fe^II(tBuTPP) (tBuTPP = meso-tetrakis(p-tert-butylphenyl)porphyrinato²⁻) generated the peroxo species [Fe^III(O₂²⁻)(tBuTPP)]⁻. This process could be reversed by addition of a proton source, triflic acid (HOTf). The HOTf attacked excess free superoxide in the solution, which shifted the equilibrium back towards the Fe^II(tBuTPP) complex, according to Scheme 41. This result lies in contrast to other reports in which the presence of acid has been found to induce O–O cleavage to give the high-valent Fe-oxo species rather than disproportionation. By calculation, DMSO was found to have a strong influence on the O₂ binding mode and therefore the electronic structure of the Fe-O₂ adduct. However, potassium ion, present in solution from the superoxide source, was found to exert very little influence on these factors.

Scheme 41.

Effect of Acid Addition on a Peroxo Adduct of an Iron-Porphyrin

A study by van Eldik examined the influence of axial ligand and spin state on the O–O bond cleavage and reactivity of Fe-hydroperoxo porphyrin complexes.¹⁸⁴ The proposed five-coordinate, high-spin [Fe^III(TPFPP)(OOH)] was observed to give Fe^IV(O)(TPFPP), a Cpd-II-like species, suggesting homolytic cleavage of the O–O bond, while the proposed six-coordinate, low-spin [Fe^III(TPFPP)(OH)(OOH)]⁻ was observed to give an Fe^IV(O)(TPFPP^•+) species, suggesting heterolytic cleavage of the O–O bond. The authors attributed this difference in reactivity to the spin-states of the reactant and products for each reaction. The reactivity of the [Fe^III(TPFPP)(OH)(OOH)]⁻ complex with various substrates was tested. No reaction was observed upon addition of olefins and sulfides to this complex, and addition of triphenylphosphine resulted only in displacement of the hydroxide ligand rather than oxidation (Scheme 42).

Scheme 42.

Comparative Reactivity of a Ferric-Hydroperoxide Porphyrin and the O–O Bond Cleavage Product, a Cpd-I-like Species

DFT calculations suggest that a doubly-protonated Fe^III-(H₂O₂) porphyrin complex could be a competent oxygen-atom donor for sulfide substrates.¹⁸⁵ The reaction was found to proceed via proton-coupled heterolytic cleavage of the O–O bond in the Fe^III(H₂O₂)(Por) complex. The mechanism was initiated by nucleophilic attack of the distal oxygen by the sulfur lone pair of dimethyl sulfide, an electron-rich sulfur substrate. In tandem with the heterolytic cleavage, the Fe^III(OH)(Por) abstracts the proton from the protonated sulfoxide product to give dimethylsulfoxide and Fe^III(OH₂)(Por). An alternative mechanism was ruled out, in which the O–O bond of Fe^III(H₂O₂)(Por) is cleaved homolytically to generate an [Fe^III + •OH] and a hydroxyl radical, •OH. The •OH abstracts an H-atom from [Fe^III + •OH] to give Fe^IV(O)(porphyrin π-cation-radical) (Cpd-I), which can then perform sulfoxidation. However, the homolytic cleavage step was found to have a much higher activation barrier than the heterolytic/sulfur attack mechanism, indicating that the heterolytic mechanism will undergo sulfoxidation much faster than Cpd-I can be generated (Scheme 43). While an Fe^III(H₂O₂) species has not been observed in heme enzyme catalysis, the peroxide shunt is often invoked in mechanistic studies where hydrogen peroxide is used to generate the high-valent iron-oxo species.

Scheme 43.

DFT Calculations on the Heterolytic vs Homolytic Cleavage Mechanisms in Oxidation of Sulfide Substrates by an Fe(H₂O₂)-Porphyrin Complex

Spectroscopic detection of a ferrous-superoxo porphyrin could be achieved by cryo-generation (Scheme 44).¹⁸⁶ Irradiation of a ferric-superoxo complex with a γ-ray source at 77 K resulted in one electron reduction to the ferrous superoxo species, which was characterized by EPR and resonance Raman spectroscopies. DFT calculations also confirmed that the reduction of the Fe^III(O₂^•−) species occurred at the iron center rather than the superoxo-ligand. When the ferrous-superoxo species was warmed to 193 K, the ferric-hydroperoxide species was observed, as characterized previously by Naruta and coworkers,¹⁸² indicating that the Fe^II(O₂^•−) species was basic enough to scavenge a proton even in the absence of an exogenous acid source. This work gave insight into the thermodynamics of the transformation between ferric-superoxo and ferric-hydroperoxo, which differ by a hydrogen atom, and both of which are implicated in the catalytic cycles of heme monooxygenase enzymes.

Scheme 44.

Cryo-Reduction of a Ferric-Superoxo to the Ferrous-Superoxo Porphyrin Complex

The influence of the secondary coordination sphere on the reactivity of a ferric-superoxo porphyrin was investigated (Scheme 45).¹⁸⁷ An imidazole as an axial ligand and a carboxylic acid group as a proton donor were tethered to the porphyrin ligand. As a control, the secondary coordination sphere effects of a similar ligand containing an ethoxy carbonyl group instead of the carboxylic acid were compared. When the iron-superoxo species was generated by addition of O₂ to the ferrous iron in the ligand without the proton donor, addition of exogenous reductant and acid resulted in controlled formation of the ferric hydroperoxo species. However, when O₂ was added to the ferrous iron porphyrin containing the pendant proton donor, some ferric-hydroperoxide was observed, in addition to the ferric-superoxo species. The presence of ferric-hydroperoxide from the oxygenation reaction, in the absence of added proton and electron donors, implied that the reduced ferrous porphyrin itself could act as an H⁺/e⁻ donor, but only if the carboxylic acid group was present.

Scheme 45.

Comparison of the Reactivity of a Ferric Superoxo Porphyrin in the Presence and Absence of a Ligand-tethered Proton Donor

4.1.2. Iron-Oxygen Corroles and Corrolazines

To our knowledge, there are no examples of well defined iron-(hydro)peroxo or iron–superoxo corrole or corrolazine complexes.

4.2. Manganese-Oxygen Complexes

4.2.1. Manganese-Oxygen Porphyrins

In heme enzymes, the binding and activation of dioxygen or hydrogen peroxide at the metal center are typically the initial steps in the catalytic cycles. Cleavage of the O–O bond leads to the generation of high-valent metal-oxo species. There are very few examples of well-defined manganese-superoxo or –(hydro)peroxo porphyrins.¹⁸⁸ This may be due to the relative stability of the O–O bond-cleaved product, a high-valent manganese-oxo species, in these systems. However, those rare Mn–superoxo or – (hydro)peroxo species that have been observed can be generated by either O₂, H₂O₂, or ROOH binding and activation. Some reviews on Mn-superoxo/peroxo porphyrin species include a section of a chapter by Koga and Tabushi in Biomimetic Chemistry from the Advances in Chemistry (1980) series³⁷ and a section of a chapter by Groves and coworkers in the book Biomimetic Oxidations Catalyzed by Transition Metal Complexes (2000).¹⁸⁹ In this section, we focus on the reactivity of the Mn–O–O adducts, which is intimately tied to the extent of O–O bond activation.

The first example of a Mn(porphyrin)(O₂) adduct was reported in 1975. Addition of O₂ at −79 °C to Mn^II(TPP) resulted in the reversible formation of a five-coordinate dioxygen-bound Mn(O₂)(TPP) species.¹⁹⁰ Debate about the electronic structure and mode of dioxygen binding followed, with some suggesting an Mn^IV(O₂²⁻)(Por)(O₂²⁻ = peroxide²⁻) with the peroxide ligand binding to the Mn in a side-on fashion^191-194 and others an Mn^III(O₂^•−)(Por) (O₂^−• = superoxide¹⁻) with the superoxide ligand binding to Mn in an end-on fashion.^195,196 The one-electron reduced [Mn(O₂)(Por)]⁻ adduct generated from addition of potassium superoxide to Mn^II (generated from outer sphere reduction of Mn^III(Por) by another equivalent of potassium superoxide) was subjected to a similar debate regarding geometry and electronic structure.^197,198 Another way to generate the [Mn(O₂)(Por)]⁻ species was found when O₂ was added to Mn^III(por) in the presence of hydroxide.¹⁹⁹ Ultimately, a crystal structure of an [Mn(O₂)(TPP)]⁻ adduct showed that a peroxide ligand was bound to Mn^III(TPP) in the side-on mode, resolving the debate. This binding mode for the reduced complex also served as precedent for the oxidation state assignment of the neutral Mn^IV(O₂²⁻)(TPP) complex.²⁰⁰

The equilibrium constant for the binding of O₂ to Mn(II) porphyrins was found to depend on the electron-donating or electron-withdrawing nature of the porphyrin ligand.²⁰¹ The effect of neutral axial ligands (L) bound to Mn^II(por) complexes trans to the O₂ binding site was studied and a correlation between the proton affinities of the series of ligands and the equilibrium constants for O₂ binding was found.²⁰²

The acylation of an oxygenated manganese-porphyrin complex could be achieved at low temperature by addition of acylchloride to Mn(II) or Mn(III) porphyrins in the presence of O₂ or superoxide, respectively.²⁰³ The Mn-acyl-peroxo species was capable of the oxidation of cyclooctene to cyclooctene oxide. Addition of hydroxide ion to the acyl-peroxo species and warming of the reaction solution induced O–O bond cleavage to give the high-valent Mn-oxo species, which was also able to undergo epoxidation of cyclooctene.

Changing the electronics of the peroxyacid allowed for tuning of the O–O bond cleavage reactivity of an acylperoxo-Mn(TMP) adduct, but the relative rates of O–O cleavage to form the high valent Mn-oxo species were strongly dependent on the presence or absence of hydroxide ion.²⁰⁴ In the presence of hydroxide, the more electron-rich peroxyacids were more reactive than the electron-poor peroxyacids. In the absence of hydroxide ion, the opposite trend was observed. These differences in reactivity were attributed to a change in mechanism from homolytic to heterolytic O–O cleavage in the presence and absence of hydroxide, respectively.

A manganese(III)peroxycarbonate complex could be formed from the reaction of [Mn(O₂)(Por)]⁻ adducts with carbon dioxide.²⁰⁵ Homolytic or heterolytic O–O bond cleavage of the peroxycarbonate adducts was dependent on the solvent, temperature, and porphyrin ligand. Formation of [Mn(O₂CO₂)(Por)]⁻ adducts was achieved by addition of potassium superoxide to either Mn^II(TPP), Mn^II(TPFPP), or Mn^II(TP_ivP) (TP_ivP = meso-tetrakis(α-pivalamidophenyl)porphyrinato²⁻), followed by addition of carbon dioxide at − 70 °C. Depending on the solvent, O–O cleavage occurred to give one of two high-valent Mn(O) species. In THF, a ligating solvent, homolytic O–O cleavage to give the Mn^IV(O) was favored, whereas in toluene heterolysis to give the Mn^V(O) was preferred.

The reactivity of peroxomanganese(III) porphyrins with organic substrates was studied and compared with other metal-peroxo complexes.²⁰⁶ It was found that [Mn^III(O₂)(Por)]⁻ acted as a nucleophile of moderate strength. The peroxo complex [Mn^III(O₂)(Por)]⁻ was able to oxidize an extremely electron-deficient alkene, tetracyanoethylene, in addition to acyl halides and CO₂. However, [Mn^III(O₂)(Por)]⁻ was unable to oxidize other olefins, including some electron-deficient olefins and all electron-rich olefins tested, as well as triphenylphosphine and butyllithium. It was suggested that the relative stability of this complex was due to strong orbital overlap between the metal and O₂ ligand.

The mechanism of heterolytic O–O bond cleavage of a Mn^III(OOH)(TDMImP) was studied.²⁰⁷ The O–O bond cleavage step was found to be rate determining and followed a concerted “push-pull” mechanism, in which an axial ⁻OH was partially deprotonated and provided the “push”, while the terminal oxygen in the ⁻OOH ligand was partially protonated and acted as a “pull” to release an H₂O molecule into the solvent medium. This “push-pull” model has also been invoked to explain the heterolytic O–O bond cleavage to generate Cpd-I during the catalytic cycle of Cytochrome P450s and other heme enzymes.

4.2.2. Manganese-Oxygen Corroles

Well-defined manganese-peroxo corrole complexes are rare, while manganese-superoxo corrole complexes are unknown. The following examples constitute the only evidence for these species to date, and their reactivity is still not well understood.

In 2009, Åkermark, Sun, and coworkers showed evidence for the formation of O₂ by addition of hydroxide to an Mn^V(O) corrole complex (Scheme 46).²⁰⁸ Nucleophilic attack of hydroxide on Mn^V(O)(tnpc) (tnpc = 5,10,15-tris(4-nitrophenyl)corrolato³⁻) initially generated a putative Mn^III hydroperoxy complex, which was not observed. Formal H• abstraction by an unidentified oxidant from this complex gave an Mn^IV species, which was assigned as an Mn^IV-peroxo complex based on UV-vis and mass spectrometry. Further oxidation through an unknown mechanism results in the Mn^III corrole and release of dioxygen. Formation of O₂ was quantitated by online gas analysis, and addition of H₂¹⁸O to the hydroxide reaction resulted in mixed-labeled Mn-peroxo complex as well as mixed-label dioxygen (^16,18O₂), confirming that O–O bond formation occurs in this system.

Scheme 46.

O-O bond Formation from Nucleophilic Attack of Hydroxide on a High-Valent Metal-Oxo Corrole

Following this work, Nam and coworkers provided evidence for the reversible interconversion of an Mn^V(O) corrole and an Mn-peroxo corrole complex (Scheme 47).²⁰⁹ The addition of H₂O₂ to Mn^III (tfmpc) (tfmpc = 5,10,15-tris(trifluoromethylphenyl)corrolato³⁻) in the presence of hydroxide resulted in an intermediate assigned as an Mn^IV(O₂)⁻ species based on UV-vis, EPR, and mixed labeling mass spectrometry studies using ¹⁸O-labeled reagents. The addition of hydroxide to Mn^V(O)(tfmpc) also resulted in the formation of this new species, and could be reversed back to the starting Mn^V(O) complex with the addition of perchloric acid. The mechanism of interconversion is not understood with regard to electron-count, since Mn^V(O) and hydroxide should give an Mn^III(peroxo) complex and H₂O₂ addition to Mn^III (tfmpc) should give the same Mn^III species. The overall pathway is shown in Scheme 47.

Scheme 47.

Reversible Interconversion of a Manganese-oxo and Manganese-peroxo Corrole Complex in the Presence of Acid or Base

5. Conclusions and Final Remarks

In this review, we discussed the biomimetic reactivity of iron and manganese -oxo and -dioxygen intermediates in porphyrinoid systems. The reactivity of these species depends heavily on the identity of the porphyrinoid ligand (either porphyrin, corrole, or corrolazine), the metal identity and oxidation state, as well as the type of substrate. These results are analogous to natural systems that depend on tuning environment to control reactivity. Experimental and theoretical studies both gave crucial insights into the fundamental reaction mechanisms.

While biomimetic reactivity of metal-oxo and metal-dioxygen intermediates has been extensively studied, we still foresee interesting challenges ahead. For example, the direct observation of the combination of an organic radical and a high-valent metal-hydroxide compound has not yet been achieved. Understanding the factors that select for rebound chemistry over other reactions (such as dehydrogenation) could allow for insight into the design of metalloporphyrinoid catalysts that undergo selective hydroxylations with substrates that contain many possible sites of oxidation. The design principles from these studies could also give insight into C–H halogenation performed by heme chloroperoxidase. High-valent iron-oxo corrole complexes and their reactivity are notably absent from the literature. Only one corresponding high-valent iron-oxo corrolazine has been characterized. The field is open for ligand development to access high-valent mononuclear Fe(O) corroles and corrolazines. There is still much to learn about metal-oxygen species and what controls the extent of O₂ activation, the O–O bond formation/cleavage steps, and the mechanism of O–O bond formation/cleavage. Manganese or iron corroles could allow for the study of these species, but there are few examples in the literature showing that they can form stable metal-oxygen complexes. The syntheses of these species would allow for unknown reactivity to be revealed.

Achieving answers to these questions and opening up the field to new reactivity could give valuable insight into the natural systems and provide a foundational basis for the development of new metalloporphyrinoid-mediated reactions.

Scheme 25.

O-atom Transfer, H-atom Abstraction, and Electron Transfer Reactions of Fe^IV(O)(corrolazine^•+)

Author Biographies

Regina A. Baglia received her B.S. degree in biochemistry in 2011 from Temple University, in Philadelphia, PA, where she performed undergraduate research as a Diamond Research Scholar in the Department of Chemistry with Prof. Michael J. Zdilla. She received her PhD in chemistry at The Johns Hopkins University in 2016, where she worked with Prof. David P. Goldberg. While at Hopkins, she was awarded the George E. Owens Graduate Fellowship, an E²SHI fellowship, and the Harry and Cleio Greer fellowship.

Jan Paulo T. Zaragoza was born in Manila, Philippines in 1990. He received his B.S. in Biochemistry from the University of the Philippines—Manila in 2010. He is currently pursuing his doctoral studies under the supervision of Prof. David P. Goldberg at the Johns Hopkins University. His research involves the study of high-valent metal-oxo corroles and corrolazines as synthetic models for heme enzyme active sites.

David P. Goldberg received his B.A. degree from Williams College in 1989, and his Ph.D. degree in Inorganic Chemistry in the research group of Prof. Stephen J. Lippard. Following his Ph.D. in 1995, he was an NIH Postdoctoral Fellow in the group of Prof. Brian Hoffman at Northwestern University. He began his independent career at Johns Hopkins University in the department of Chemistry in 1998, where he is currently a Professor of Chemistry. His research interests include synthetic inorganic chemistry, models of heme and nonheme metalloenzymes, dioxygen and nitric oxide activation, and other fundamental transformations important to synthetic and biological oxidation/reduction catalysis.