PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Am Chem Soc. Author manuscript; available in PMC 2017 September 14.
Published in final edited form as:
PMCID: PMC5228556
NIHMSID: NIHMS836329

Activation of Dioxygen by Iron and Manganese Complexes: A Heme and Nonheme Perspective

Abstract

The rational design of well-defined, first-row transition metal complexes that can activate dioxygen has been a challenging goal for the synthetic inorganic chemist. The activation of O2 is important in part because of its central role in the functioning of metalloenzymes, which utilize O2 to perform a number of challenging reactions including the highly selective oxidation of various substrates. There is also great interest in utilizing O2, an abundant and environmentally benign oxidant, in synthetic catalytic oxidation systems. This Perspective brings together recent examples of biomimetic Fe and Mn complexes that can activate O2 in heme or nonheme-type ligand environments. The use of oxidants such as hypervalent iodine (e.g., ArIO), peracids (e.g., m-CPBA), peroxides (e.g., H2O2) or even superoxide is a popular choice for accessing well-characterized metal–superoxo, metal–peroxo, or metal–oxo species, but the instances of biomimetic Fe/Mn complexes that react with dioxygen to yield such observable metal–oxygen species are surprisingly few. This Perspective focuses on mononuclear Fe and Mn complexes that exhibit reactivity with O2 and lead to spectroscopically observable metal–oxygen species, and/or oxidize biologically relevant substrates. Analysis of these examples reveals that solvent, spin state, redox potential, external co-reductants, and ligand architecture can all play important roles in the O2 activation process.

1. INTRODUCTION

Dioxygen, arguably the most important molecule for sustaining aerobic life, plays a number of critical roles in biology, ranging from nutrient metabolism to the synthesis of various important biomolecules (e.g., aromatic amino acids, hormones, neuro-transmitters).1 Most of these processes rely on the oxidizing power of dioxygen, where it undergoes four-electron reduction to form water. This four-electron reductive activation of dioxygen, although thermodynamically favorable overall (0.815 V vs NHE in water at pH 7, 25 °C), is kinetically hindered, because the dioxygen molecule is found in a triplet spin ground state and has a high negative one-electron reduction potential (−0.33 V vs NHE in water at pH 7, 25 °C).2 Nature can overcome the spin state barrier by employing transition metal ions that also exist in open-shell spin ground states and can react directly with triplet dioxygen. These metal ions can facilitate the one-electron reduction of O2 by metal coordination, and also can serve as multielectron reductants to access thermodynamically more favorable two-electron, or even four-electron reduction pathways. A number of first-row transition metals ions (such as Mn, Fe, Cu) are employed by metalloenzymes for the purpose of activating O2, and the power and breadth of these enzymes have captured the imagination of biochemists, inorganic chemists, and other researchers for many years.

Iron-containing enzymes make up a large number of these O2-activating enzymes, partly because of the bioavailability of iron in Nature, and partly because iron can access multiple redox states. In addition, there are a number of open-shell spin states available to iron in its different common oxidation states, with high-spin (hs) iron(II) (S = 2) perhaps being the most important with regard to the binding and activation of O2. The Fe-containing enzymes can be classified into two types: heme enzymes, which contain a macrocyclic porphyrinoid ligand which houses the metal center, and nonheme iron enzymes, which have nonporphyrinoid ligand coordination (e.g., two-histidine-1-carboxylate binding motif) holding the iron ion in the protein. Both heme35 and nonheme1,6,7 Fe enzymes reductively activate O2 in their respective catalytic cycles, and often target the oxidation of organic substrates, including aliphatic C–H hydroxylation, aromatic hydroxylation, olefin epoxidation, and halogenation reactions. Although heme and nonheme enzymes have very different structural features, they share similar iron–oxygen intermediates (e.g., iron–superoxo, iron–(hydro)peroxo, high-valent iron–oxo) during their respective catalytic cycles. Spectroscopic evidence for some of these intermediates has been obtained,8 but trapping these intermediates, and studying their spectroscopic and reactivity properties inside the protein scaffold remains quite challenging because of their short lifetimes, general instability, and the inherent difficulties in studying large macromolecular complexes.

Synthetic biomimetic model complexes have been prepared and their reactivity has been studied over the years to aid in the understanding of these biological processes. The mechanisms and key intermediates involved are often more easily studied in these small-molecule analogue systems. Heme and nonheme iron complexes have received a great deal of attention because of their relevance to biology, and much effort has gone into their synthetic development and the study of their rich spectroscopic features as well as their reactivity. Mn-based complexes have gathered increasing attention as a close analogue of Fe-based systems, owing to the fact that both Fe and Mn share similar coordination geometries and multiple redox states, and also because of Mn sites in biology (e.g., photosystem II, superoxide dismutase, ribonucleotide reductase, Mn catalases).9,10 Although the direct reactivity of Fe and Mn complexes with dioxygen has led only to limited progress, some progress has been made by studying organic and inorganic oxidants (PhIO, m-CPBA, H2O2, NaOCl, ROOH) as surrogates for O2 and as tools for accessing proposed O2-derived intermediates. A number of review articles have been written describing this chemistry for both heme and nonheme model complexes.6,8,1019

In this Perspective, we present recent, significant findings on the activation of O2 by mononuclear iron and manganese complexes. As opposed to earlier reviews, we focus exclusively on reactivity with dioxygen as the oxidant, and we bring together both heme and nonheme systems for comparison. We also attempt to provide a brief description of the structural and electronic factors that help facilitate and control the binding and activation of dioxygen at the different metal sites. This Perspective separately considers heme and nonheme systems as a convenient way to categorize and discuss the progress that has been made in these areas over the past 10 years.

2. DIOXYGEN REACTIVITY OF IRON-PORPHYRIN COMPLEXES

Heme proteins are ubiquitous in nature and they perform a diverse range of biological functions including dioxygen transportation (e.g., hemoglobin) and storage (e.g., myoglobin), electron transfer (e.g., cytochrome c oxidase) and various organic transformations (mono and dioxygenases).20 Heme proteins employ a planar, tetradentate porphyrin macrocycle as the ligand for the iron center. The iron porphyrin cofactor typically is coordinated to the protein by one or two axial ligand(s) with N, S, or O donor atoms (e.g, histidine, cysteine, methionine, tyrosine), and these ligands play an important role in controlling the reactivity at the iron center.3,4 Cytochrome P450 (CYP) is one such enzyme and is the prototype of a dioxygen activating heme enzyme. The resting state of the CYP active site contains a ferric center with a cysteinate ligand in the axial position, which facilitates the dioxygen activation process at the iron center. The Cytochromes P450 carry out many chemical transformations, including the regioselective hydroxylation of challenging organic C–H substrates. The CYP enzymes utilize NADH as a source of reducing equivalents, and together with dioxygen give an FeIV(O) π-cation-radical species (Compound I (Cpd-I)) that is capable of activating inert C–H bonds.

An early synthetic analogue for Cpd-I was prepared in 1979 with [FeIII(TPP)Cl] (TPP = meso-tetraphenylporphyrin) as the precursor complex.21 However, an organic oxidant (PhIO, m-CPBA) was needed to generate this species, and most of the work on Cpd-I analogues that has followed since then still relies on the use of similar, high-energy oxidants.2123 Although efforts have been made to utilize O2 to generate and characterize Cpd-I and other related Fe/O2–derived intermediates during O2 activation, successful examples of such chemistry remain limited. Challenges in forming Cpd-I from O2 and synthetic porphyrins come from the difficulties in biasing synthetic systems toward favorable binding of O2, and in providing the appropriate stoichiometry and timing of electrons and protons that must be added to achieve the multielectron cleavage of the O–O bond from FeII- or FeIII-heme and O2. Heme proteins such as P450 have a considerable advantage in carrying out this operation through evolutionarily optimized e and H+ transfer chains that guide the injection of reducing equivalents from NADH and H+ from H2O into the active site heme iron. However, progress has been made in overcoming these obstacles, and some of these results are summarized in the following discussion.

From the 1960s, ferrous porphyrin complexes were known to bind and react with dioxygen to form the µ-oxo-bridged diiron(III) product.24,25 The mechanism for this autoxidation was not fully understood until the 1980s, when the reaction was shown to proceed through a peroxo-bridged diiron(III) and the ferryl FeIV(O) intermediates (Scheme 1).2629 Both the peroxo and oxo intermediates were characterized at low temperatures. Since that time, there have been relatively few studies describing the characterization of mononuclear, O2-derived intermediates. There are few examples where an iron–porphyrin complex was shown to mediate substrate hydroxylation/epoxidation reactions utilizing dioxygen.22,3032 Some of these systems required the use of an external reductant to observe the activity. Although a high-valent iron–oxo species was postulated as the active oxidant in most of these cases, there was very little spectroscopic evidence for the proposed oxidant, and the mechanism of the catalytic transformations were also poorly understood. The instability of these species makes their isolation and characterization quite challenging, and in the next section we highlight some of the recent examples in this area, where iron–oxygen species have been characterized. It is important to note that a large amount of work has been done on iron–porphyrin-based electrocatalysts for the reduction of oxygen;3338 however, these systems will not be covered in this Perspective.

Scheme 1
Dioxygen-Mediated Autoxidation Mechanism for Ferrous–Porphyrin Complexes

Most of the dioxygen activating heme and nonheme iron enzymes are proposed to proceed through an initially formed iron(III)–superoxide intermediate. In synthetic heme systems, formation of an FeIII(O2) complex was reported for a few cases.3941 A later example of a porphyrinoid FeIII(O2) species was reported in 2009 with [FeII(tmpIm)] as starting material (Scheme 2).42 Reaction of [FeII(tmpIm)] with O2 at −75 °C led to the formation of [FeIII(O2)(tmpIm)], which gave a characteristic UV–vis spectrum (426, 535, 589 nm). Stoichiometric addition of the one-electron reductant cobaltocene to the FeIII(O2) species with excess MeOH generated the FeIII(OOH) complex (427, 534, 564 and 610 nm). This result indicated that MeOH protonates an initially formed peroxo complex. Attempts to isolate the unprotonated peroxo species from dioxygen were unsuccessful. Interestingly, the same FeIII(OOH) species could be produced by reacting [FeII(tmpIm)] with KO2 and subsequent treatment with MeOH (Scheme 2). The formation of an FeIII(OOH) species was confirmed by resonance Raman (rR) spectroscopy (νO–O = 810; νFe–O = 570 cm−1) and isotope labeling experiments with 18O2 and MeOD. Zero-field Mössbauer and electron para-magnetic resonance (EPR) experiments revealed the presence of a low-spin (ls) Fe3+ center in the hydroperoxide complex. Introduction of a bulky xanthene substituent on a meso-carbon atom of the tmpIm macrocycle led to stabilization of the unprotonated peroxo species.43 The xanthene moiety was thought to provide steric shielding for the peroxo group. Reaction of the new xanthene appended [FeII(tmpIm)] complex with O2 at −30 °C generated the FeIII(O2) species, which was characterized by UV–vis (428, 550, 592 nm) and rR spectroscopy (νFe–O2 = 582 cm−1). Addition of cobaltocene to the FeIII(O2) species resulted in the formation of a ls-FeIII-(peroxo) complex. This complex was characterized by UV–vis (430, 568, 610 nm), EPR (g = 2.27, 2.16, 1.96), and rR spectroscopy (νFe–O = 585 cm−1, νO–O = 808 cm−1). Based on the available spectroscopic data, an end-on FeIII-(peroxo) species was proposed for the product from Cp2Co reduction.43 In 2016, the same group showed that a related imidazole-ligated, ferrous–porphyrin complex with a pendant anthrace-necarboxylic acid group reacts with O2 to give the FeIII(O2) complex, and also yields the one-electron-reduced hydroperoxo complex with the FeII starting porphyrin serving as reductant.44

Scheme 2
Formation of an FeIII(O2) Species Generated from the Reaction of Ferrous–Porphyrin with O2, and the One-Electron Reduction of the FeIII(O2) Complexa

The former work highlights the power of ligand design in improving the reactivity of these heme analogues. Tethering of the axial ligand to the synthetic porphyrin helps with facilitating O2 binding and increasing the stability of the O2 intermediates such that electrons and protons can be delivered. Further work on modifying axial ligation through ligand design may improve the O2 activation chemistry in synthetic systems. The addition of second-coordination-sphere groups orthogonal to the plane of the porphyrin, such as the pendant anthracene carboxylic acid described above, is another promising strategy for building discrete molecules with improved O2 activation properties. These strategies, however, necessarily involve multistep organic synthesis, which can be a drawback. Other promising strategies may be to imbed and sequester porphyrins in modifiable, three-dimensional frameworks that can be readily synthesized, such as metal–organic frameworks (MOFs). A study was recently reported that led to the binding of O2 to an iron porphyrin encapsulated in the MOF PCN-224FeII.45 The encapsulation led to the trapping of a novel Fe–O2 adduct that lacked a sixth axial ligand trans to the O2 binding site.

3. MANGANESE–PORPHYRINOID COMPLEXES AND DIOXYGEN ACTIVATION

Manganese porphyrinoid complexes have been synthesized and examined for comparisons with the analogous iron complexes in heme enzymes. One of the early examples of dioxygen activation by a Mn–porphyrin complex was reported in 1975.46 A side-on MnIV–peroxo species was proposed to form from the reaction of [MnII(TPP)(Py)] complex (Py = pyridine) with dioxygen at −78 °C.47,48 A few other reports followed which described dioxygen reactivity of other Mn–porphyrin and Mn-phthalocyanine complexes.4953 Mn porphyrinoid complexes have also been used for substrate oxidations utilizing O2 as the oxidant.54,55 For example, one of the early examples of the use of O2 as oxidant was reported in 1987, where [Mn(TPP)(Cl)] was reacted with O2 in the presence of 1-methyl imidazole as cocatalyst and Zn as reducing agent for the epoxidation of olefins.54 In most of these earlier studies, the characterization of the proposed metal–oxygen intermediates was not extensive, and the nature of the active oxidant species in substrate oxidation reactions was poorly understood. There were very few reports in which well-characterized porphyrinoid Mn–oxygen intermediates were generated from dioxygen. Recently, our group reported the generation of a high-valent MnV(O) porphyrinoid complex from a MnIII–corrolazine and O2. The catalytic oxidation of certain organic C–H substrates was achieved with this system.5658

Corrolazine (Cz) is a ring-contracted member of the porphyrinoid family with a 3– charge when fully deprotonated, and was designed by our group to stabilize high-valent metal complexes. It has been used to generate stable high-valent metal–oxo complexes of iron, vanadium, chromium, rhenium, and manganese.14,59 In the case of Mn, the high-valent [(TBP8Cz)MnV(O)] (TBP8Cz = octakis(p-tert-butylphenyl)-corrolazinato3−) was isolated and characterized at room temperature. This complex was initially prepared from the reaction of a MnIII precursor [(TBP8Cz)MnIII] with the hypervalent iodine reagent PhIO. However, it was shown in 2012 that [(TBP8Cz)MnV(O)] could be synthesized from reaction of [(TBP8Cz)MnIII] with O2 as the oxidant in the presence of visible light (Scheme 3).58

Scheme 3
Photoinitiated Dioxygen Activation by a MnIII–Corrolazine Complex and the Mechanism of Formation of a MnV(O) Speciesa

Photoirradiation of [(TBP8Cz)MnIII] (λmax = 432, 687 nm) with a white light source under ambient conditions in cyclohexane led to the formation of dark green [(TBP8Cz)-MnV(O)] (λmax = 419, 639 nm) (Scheme 3). The production of [(TBP8Cz)MnV(O)] was confirmed by 1H NMR spectroscopy and laser desorption/ionization mass spectrometry (LDI-MS). Control experiments showed that the presence of both visible light and air was required for the conversion of MnIII to MnV(O) corrolazine. Isotope labeling experiments with 18O2 confirmed dioxygen as the source of the O-atom in the MnV(O) product. It was proposed that the light-driven, aerobic oxidation of MnIII to MnV(O) in cyclohexane involved the participation of solvent as a sacrificial reductant. This hypothesis was corroborated by the fact that the MnV(O) complex was not formed when the solvent was changed to benzonitrile.60 However, addition of toluene derivatives with relatively weak benzylic C–H bonds (e.g., hexamethylbenzene, HMB) resulted in the formation of [(TBP8Cz)MnV(O)] in PhCN (Scheme 3).60 GC-MS analysis of the reaction with HMB revealed the production of pentamethylbenzyl alcohol in good yield (87%), together with a small amount of pentamethylbenzaldehyde (8%). A primary kinetic isotope effect (KIE) for the substrates toluene (KIE = 5.4) and mesitylene (KIE = 5.3) was observed, indicating that the initial cleavage of the benzylic C–H bond was involved in the rate-determining step. It was shown that this strategy could be expanded to manganese porphyrin complexes, where the catalytic aerobic oxidation of an activated substrate (acridine) was performed with air, visible light, and MnIII(Porph)(X) (X = OH, OAc).61

Further mechanistic insights came from femtosecond transient absorption spectroscopy.60 The [(TBP8Cz)MnIII] complex has a singquintet (5S0) ground state, due to high-spin d4 electronic configuration. Femtosecond laser excitation of [(TBP8Cz)MnIII] converted the singquintet (5S0) ground state to a tripquintet (5T1) excited state at 530 nm (Figure 1). This state undergoes rapid intersystem crossing to a long-lived tripseptet (7T1) state with a new peak at 774 nm (Figure 1). The decay rate of the 7T1 state was sensitive to the presence of O2. It was suggested that the 7T1 state reacts directly with O2 to form a MnIV–superoxo intermediate, which abstracts a H-atom from the toluene derivative to form a MnIV(OOH) complex and benzyl radical. This latter step would be the rate-determining step based on KIEs. The subsequent breaking of the O–O bond could go through either a homolytic or heterolytic pathway, leading to the formation of [(TBP8Cz)-MnV(O)] and benzyl alcohol, as shown in Scheme 3.

Figure 1
Transient absorption spectral changes of the [(TBP8Cz)MnIII]* (5T1) (530 nm) and [(TBP8Cz)MnIII]* (7T1) (774 nm) states, generated from photoexcitation of (TBP8Cz)MnIII in benzonitrile. Reprinted with permission from ref 60. Copyright 2013 American Chemical ...

The [(TBP8Cz)MnV(O)] complex can oxidize substrates with relatively weak C–H bonds such as dihydroacridine (AcrH2, bond dissociation free energy (BDFE) = 69 kcal/mol) to give acridone (Acr═O), or O-atom acceptor substrates such as triphenylphosphine to give triphenylphosphine oxide.58,60 These substrates allow for catalytic turnover with MnIII as catalyst in the presence of air and light (turnover number (TON) for (OPPh3) = 535; TON(Acr═O) = 11). In a subsequent study, we hypothesized that strong acid would activate the photochemically generated MnV(O) complex toward stronger C–H bonds (e.g., toluene (BDFE = 87 kcal mol−1) and its derivatives), giving back the MnIII complex and completing a catalytic cycle.11,56,57 Photoirradiation (λirr > 400 nm) of a catalytic amount of [(TBP8Cz)MnIII] in the presence of excess HMB and a strong proton donor [H(OEt2)2]+[B-(C6F5)4] (H+[B(C6F5)4]) in benzene resulted in the catalytic formation of oxidized alcohol PMB–OH and aldehyde PMB-CHO products in 18 and 9 turnovers, respectively. Although the TONs were modest, catalytic reactivity was achieved, and importantly, the resting state of the catalyst was shown to be MnIII, with no evident buildup of the stable MnV(O) complex under acidic conditions. The nature of the catalytically active species under acidic conditions was examined by spectroscopic methods and single-crystal X-ray diffraction (XRD). The corrolazine ligand contains meso-nitrogen atoms that are possible sites for protonation, and both the monoprotonated [MnIII(TBP8CzH)]+ and diprotonated [MnIII(TBP8CzH2)]2+ complexes were isolated and characterized by UV–vis. The structures of these complexes, together with the neutral [MnIII(TBP8Cz)] were definitively proven by XRD.57 It was found that monoprotonated MnIII(TBP8CzH)+ was the catalytic resting state, as seen by UV–vis analysis of the catalytic reaction mixtures. The diprotonated complex was found to be catalytically unreactive. However, it was noted that a second equivalent of H+ was necessary for catalytic turnover. The MnIII complex was not protonated by this second equivalent, but the second H+ instead helped activate the transient Mn–oxo species for further reaction with substrate and subsequent closure of the catalytic cycle (Scheme 4).

Scheme 4
Proposed Mechanism for the Acid-Assisted Catalytic Oxidation of Hexamethyl Benzenea

In another study in 2016, the scope of the acid was also broadened to triflic acid, which led to a dramatic increase in catalytic turnover for the oxidations of HMB (TON = 563).56 In addition, the same system was shown to carry out the catalytic oxidation of the sulfur substrate thioanisole, giving the corresponding sulfoxide with high turnover (TON = 902). As in the case of HBArF, the relevant mono- and diprotonated species were crystallographically characterized, and once again it was found that the monoprotonated complex was the catalytic resting state. Interestingly, the triflate anion was coordinated to MnIII, and it was suggested that coordination of this counterion throughout the catalytic cycle (Scheme 4) may be responsible for the increase in catalytic activity.56

It remains to be seen if the photochemical activation seen for MnIII corrolazine can be extended to other metals, including iron. The analogous FeIII corrolazines are known, although the equivalent FeV(O) corrolazine (characterized as FeIV(O)-(Cz+•)) is much less thermally stable than the MnV(O) complex.62,63 Thus, the study of the photochemical/O2 activation pathway will be more challenging in the analysis of the desired Fe–oxo complex. The closely related MnIII and FeIII corroles are also interesting future candidates for testing with O2 and light.6467 These methods might lead to new corrole-derived high-valent metal–oxo species and/or catalytic oxidations.

The studies of the Mn porphyrinoid compounds described above are useful for comparison with the closely related iron systems. Examining the propensity to bind O2 and the requirements for stabilizing high-valent metal–oxo species in Mn yields information that may be relevant to the analogous Fe systems, and provides continued motivation to study the Mn systems. There are also, of course, the Mn-containing biological systems pointed out in the Introduction that interact with and process dioxygen species. In addition, employing Mn porphyrinoid compounds as catalysts for the selective oxidation of organic compounds with O2 as the oxidant remains an attractive, major goal in catalysis from an economic and environmental standpoint.

4. DIOXYGEN ACTIVATION BY NONHEME IRON COMPLEXES

Dioxygen activating nonheme iron enzymes are instrumental in a number of key biological processes such as synthesis of antibiotics and other biomolecules, DNA repair, and bio-degradation of aromatic compounds.1,68 Most of these enzymes employ a two histidine and one carboxylate binding motif for the iron center coordination, while solvent molecules occupy the rest of the coordination sites. This 2-his-1-carboxylate binding motif exerts a weak ligand field, which is reflected in the high spin state for the ferrous center. However, there are a number of enzymes (for example, cysteine dioxygenase) that have different coordinating ligands other than the 2-his-1-carboxylate triad. Although the majority of nonheme iron enzymes share similar ligand coordination, they can be classified into several categories (as shown in Figure 2), based on the identity of the cofactors and the functions they perform.1,69 For example, extradiol cleaving catechol dioxygenase performs a C–C bond cleavage reaction of a catechol substrate, Rieske dioxygenase acts on an aromatic substrate to carry out a cis-dihydroxylation reaction, and pterin-dependent aromatic amino acid hydroxylases hydroxylate aromatic amino acid side chains. Similar to heme enzymes, various Fe–oxygen intermediates (such as superoxo, peroxo, oxo) were proposed in the catalytic cycles for nonheme iron enzymes. Few of them were characterized spectroscopically.8 Model complexes provide important tools for detailed studies that are aimed at understanding the nature and reactivity of such iron–oxygen species in nonheme environments.

Figure 2
Nonheme iron enzymes and the various transformations that they catalyze. Adapted with permission from ref 1. Copyright 2008 Macmillan Publishers Ltd.

Biomimetic studies involving nonheme iron model complexes have received significant attention over the past 20 years with the advent of structural and spectroscopic data for a number of nonheme enzymes.69 A range of polydentate ligands with different combinations of both nitrogen and oxygen donors were designed to model different features of enzyme active sites and led to a large class of nonheme iron complexes that were used to study reactivity and generate a number of biologically relevant “Fe–oxygen” intermediates (e.g., hydro-peroxo–, peroxo–, and oxo–iron species). A number of review articles have summarized the work in this area.6,8,12,19 A major part of this nonheme iron chemistry, especially in the earlier efforts, relied on the use of external oxidants such as PhIO, m-CPBA and other peracids, H2O2 and tBOOH, instead of utilizing dioxygen as the oxidant and/or O-atom source. There were only a few examples of nonheme Fe complexes reacting with O2 for substrate oxidation reactions.6

The activation of O2 by synthetic nonheme iron complexes generally follows two possible pathways as shown in Scheme 5. The initial step involves coordination of O2 to an FeII center together with charge transfer to give a ferric–superoxo species. Addition of a proton/electron source can lead to further activation of the O2 adduct, producing a (hydro)peroxo– iron(III) complex (pathway A, Scheme 5). For those studies involving O–O cleavage, mechanistic details are unclear in many cases, and both homolytic and heterolytic pathways are proposed. The final metal-bound iron–oxygen species is usually an FeIV(O) species. Alternatively, a second iron(II) complex can take the place of the H+/e source, leading to the peroxo-bridged dinuclear species in pathway B. Homolytic cleavage of this intermediate gives the same terminal FeIV(O) species as in pathway A. The high-valent FeIV(O) intermediate acts as a potent oxidant in performing substrate oxidation reactions. In this section we describe some exciting, recent examples of nonheme iron mediated O2 activation with a focus on the mechanism and the key intermediates as implied in Scheme 5.

Scheme 5
Proposed Dioxygen Activation Pathways for Synthetic Nonheme Iron Complexes

4.1. Iron(III)–Superoxide

An FeIII(O2) species is often proposed as the first intermediate in the catalytic cycle of nonheme iron enzymes that activate dioxygen. However, characterizing the superoxo species in the enzymatic systems has proven to be a difficult task and no such intermediate has been observed for mononuclear nonheme Fe enzymes, although one example has been characterized in the diiron enzyme MIOX.70 With the exception of iron, a number of metal–superoxo species in biomimetic systems have been characterized with a range of first-row transition metal ions (e.g., Cu, Cr, Ni). In some cases, the stability of the metal– superoxo species has lent itself to characterization by X-ray crystallography.7175 However, stabilizing a superoxo moiety derived from O2 at a nonheme iron center has proven particularly difficult. The earliest report of a well characterized nonheme iron(III)–superoxide species involved a dinuclear iron complex and was reported in 2005.76 This species was generated from a reaction of the dinuclear complex [Fe2(µ-OH)2(6-Me3-TPA)2](OTf)2 with O2 at −80 °C, resulting in an “end-on” or η1 FeIII–O2 intermediate (λmax = 325 nm (10 300 M−1 cm−1), 500 nm (1400 M−1 cm−1), 620 nm (1200 M−1 cm−1)). The nature of the dioxygen adduct was characterized as a superoxo ligand by rR spectroscopy (νO–O = 1310 cm−1).

Despite much effort aimed at examining the reactivity between mononuclear iron complexes and O2, it was not until 2014 that a mononuclear FeIII (O2) complex was described (Scheme 6).77 A doubly deprotonated BDPP ligand (H2BDPP = 2,6-bis (((S)-2- (diphenylhydroxymethyl)-1-pyrrolidinyl)-methyl)pyridine) with a sterically encumbered bis(alkoxide) binding motif was used to promote O2 activation and subsequent stabilization of the FeIII center. The red, square-pyramidal FeII complex [FeII(BDDP)] was reacted with dioxygen in THF at −80 °C to give a bright yellow superoxide species [FeIII(O2)(BDDP)] (λmax = 330 nm, ε = 9400 M−1 cm−1). The formation of this O2 adduct was reversible, as sparging the solution with N2 at −80 °C regenerated the FeII precursor. The key identification of the superoxo moiety comes from rR spectroscopy on a frozen solution at 77 K (λex = 413.1 nm), which revealed a resonance-enhanced vibration at 1125 cm−1. The energy of this vibration falls in line with other mononuclear metal–superoxide complexes (νO–O = 1100–1200 cm−1).7174 Mössbauer spectroscopy of [FeIII(O2)-(BDDP)] revealed an isomer shift (δ = 0.58(3) mm/s) and hyperfine splitting pattern which was consistent with a hs-FeIII (S = 5/2) center. An analysis of the Mössbauer data suggested a ferromagnetic spin-coupling interaction between the FeIII and O2 ligand, which would result in an S = 3 ground state.78 The reactivity of [FeIII(O2)(BDDP)] toward H-atom abstraction was examined.77 Treatment of the FeIII(O2) complex with excess 9,10-dihydroanthracene (bond dissociation energy, BDEC–H = 78 kcal/mol) at −70 °C resulted in the formation of anthracene in good yield, with a 1:1 reaction stoichiometry. These results imply that the proposed FeIII(OOH) intermediate reacts further with DHA radical, but this intermediate was not identified. A kinetic analysis gave k2 = 0.005 M−1 s−1 and KIE = 7 for the oxidation of DHA, pointing to rate-limiting H-atom abstraction. These results showed that a nonheme FeIII(O2) species is capable of abstracting a H-atom from a C– H bond even at the low temperature of −70 °C.

Scheme 6
Formation and Reactivity of the [FeIII(O2)(BDDP)] (Top) and [FeIII(O2)(TpMe2)(LPh)] (Bottom) Complexesa

More recently in 2015, a mixed-ligand system containing the tridentate hydrotris(3,5-dimethylpyrazolyl)borate (TpMe2) and the bidentate imidazolyl-based borate ligands [B-(ImN-Me)2MePh] (LPh) led to the characterization of a mononuclear iron(III)–superoxide complex (Scheme 6).79 The hs-FeII complex [FeII(TpMe2)(LPh )] reacted with O2 at −60 °C to form brown [FeIII(O2)(TpMe2)(LPh)] (λmax = 350 nm). Resonance Raman spectroscopy revealed a Fermi doublet band at 1168 cm−1 that shifted to 1090 cm−1 upon 18O substitution, and a low-energy vibration at 592 cm−1 which was also sensitive to isotope substitution. These bands can be assigned to the νO–O and νFe–O modes respectively, of a bound superoxide species. In this case, Mössbauer spectra were not reported, but 1H NMR spectroscopy showed a sharp, diamagnetic spectrum, pointing to a ls-Fe3+ center (S = 1/2) antiferromagnetically coupled with the superoxide anion (S = 1/2). An X-ray crystal structure of an analogous Co (O2) complex with the TpMe2 and OiPr-substituted LOiPr ligands was obtained. The superoxide ligand in the CoIII complex was bound in an η1 fashion, and by analogy it was assumed that the iron complex exhibited the same binding mode.

The [FeIII(O2)(TpMe2)(LPh)] complex exhibited similar UV–vis and rR features compared to [FeIII(O2)(BDDP)], which also contained an η1-bound O2 ligand.77,79 However, the former complex was characterized as hs-FeIII, whereas the latter complex was ls-FeIII. This difference in spin state was proposed to be an important factor in the difference in reactivity between these two complexes. The ls-[FeIII(O2)-(TpMe2 )(LPh)] is not capable of abstracting a H-atom from even weak C–H bonds such as found in 1-benzyl-1,4-dihydronico-tinamide (BNAH, BDEC–H = 67.9 kcal/mol), in contrast to the [FeIII(O2)(BDDP)], which was able to cleave the C–H bond in DHA (BDEC–H = 78 kcal/mol). The low spin complex can only abstract H-atoms from weak X–H bonds (X = O, N; BDEX–H < 73 kcal/mol) such as phenyl hydrazine, and 2-hydroxy-2-azaadamantane (AZADOL). One advantage of this system was that H-atom abstraction from these X–H substrates allowed for the spectroscopic identification (UV–vis, rR, EPR) of the FeIII–OOH product, which could not be detected in the BDDP complex.

4.2. Formation of FeIV(O) Complexes: Intermediacy of the (Hydro)peroxo Intermediate

A high-valent FeIV(O) species has been proposed to be the key intermediate responsible for substrate oxidation in the catalytic cycle of many nonheme iron enzymes. The first spectroscopically characterized FeIV(O) in a nonheme iron enzyme was seen in studies on the α-keto acid dependent taurine dioxygenase in 2003.80,81 Subsequently, ferryl species were detected and characterized in prolyl-4-hydroxylase,82 halogenases SyrB283 and CytC3,84 and aromatic amino acid hydroxylases tyrosine hydroxylase85 and phenylalanine hydroxylase.86 A common similarity among them is that they exhibit high-spin (S = 2) ground states.

For bioinorganic model systems, the first terminal oxo–iron complex synthesized from O2 was reported in 2000.87 A tripodal urea-based ligand, tris[(N′-tert-butylureaylato)-N-ethyl)]aminato ((H3buea)3−), was used to stabilize this metal–oxo complex via intramolecular H-bonding interactions with the terminal oxo ligand. The oxidation state of the iron center in this case was +3, not +4, leading to the unusual stabilization of a lower-valent FeIII(O) complex. The proposed mechanism (pathway A, Scheme 7) for the reaction involves a peroxo-bridged diiron(III) complex [1,2-µ-O2-(FeIIIH2buea)2]2−. Subsequently, the peroxo bond undergoes O–O bond homolysis to form a putative FeIV(O) species. A H-atom abstraction reaction from an exogenous C–H bond followed by an intramolecular proton transfer yields the final FeIII(O) complex. It was later shown that this FeIII(O) species could be oxidized to an FeIV(O) complex by ferrocenium tetrafluoroborate at −60 °C.88 X-ray crystallographic characterization of [FeIV(O)(H3buea)] revealed a short Fe–O distance (1.680(1) Å), and parallel mode EPR spectroscopy (X-band, 10 K) revealed resonances at g = 8.19, g = 4.06 indicative of a high-spin (S = 2) manifold. This complex is one of the few examples of a synthetic hs-FeIV(O) complex, and the only one derived from O2.8993

Scheme 7
Formation of Nonheme Iron(III)–Peroxo and Iron(IV)–Oxo Complexes Derived from the Reaction of FeII Complexes and O2a

The first spectroscopic observation of an FeIV(O) intermediate, obtained directly from the reaction of O2 with an iron precursor complex, was reported in 2005.94 The iron-cyclam derivative [FeII(TMC)(OTf)2] (TMC = 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane) was inert to O2 in CH3CN, but was made reactive with O2 upon changing the solvent to a mixture of CH3CN/solv (1:1) (solv = EtOH, Bu2O, THF). The resulting pale green species exhibited a broad, relatively weak UV–vis feature at λmax = 825 nm (ε = 370 M−1 cm−1), which was similar to that previously reported for [FeIV(O)(TMC)(CH3CN)]2+ (prepared from PhIO as the oxidant).95 These data suggested the formation of the FeIV(O) species under the mixed-solvent conditions. The marked difference in reactivity for the [FeII(TMC)(OTf)2] complex with O2 in different solvents was attributed to the influence of the solvent on the FeIII/FeII redox potential. The E1/2 value for the FeIII/FeII couple in a mixed CH3CN/solv (1:1) combination is more positive when solv = CH3CN (0.01 V), acetone (0.08 V), and CH2Cl2 (0.02 V), as compared to a more negative potential when solv = butyl ether (−0.28 V) or THF (−0.14 V). The latter two solvent combinations allowed for the activation of O2 by the FeII complex, providing strong evidence that the solvent-tuned FeIII/FeII redox potential was a critical factor in promoting O2 activation. A plausible mechanism for the generation of [FeIV(O)(TMC)(CH3CN)]2+ is through a diiron peroxo-bridged intermediate (pathway B, Scheme 7), although no evidence for this mechanism was presented. In a later study, it was argued that the binding of alcohol or ether to the proposed µ-1,2-peroxodiiron(III) intermediate may facilitate the homolysis of the O–O bond, providing a rationale for the formation of FeIV(O) in the mixed solvent system.96 A similar observation was made for iron–porphyrin, where amines were proposed to coordinate in the axial position of the metal and facilitate O–O cleavage in a peroxide-bridged intermediate.2829 The formation of a dimeric peroxo-bridged structure from the reaction of monomeric FeII complex + O2 was demonstrated in a separate study utilizing hydrotris-(pyrazol-1-yl)borate-bound iron(II)–diketonate complex.97

The influence of an additional, covalently linked axial donor atom on the O2 reactivity of the FeII(TMC) complex was examined by replacing one of the methyl groups in the TMC ligand with a 2-pyridylmethyl arm, giving the pentadentate ligand TMC-py (1-(2’-pyridylmethyl)-4,8,11-trimethyl-1,4,8,11-tetraazacyclotetradecane).96 As seen for the parent six-coordinate [FeII(TMC)(OTf)2] complex, the new five-coordinate [FeII(TMC-py)]2+ complex was air-stable in CH3CN. However, addition of stoichiometric amounts of a BPh4 salt and the strong proton donor HClO4 resulted in the rapid formation of [FeIV(O)(TMC-py)]2+max = 834 nm, ε = 260 M−1 cm−1) (pathway D, Scheme 7). The identification of phenol (PhOH) and biphenyl byproducts indicated that BPh4-was serving as a reductant during the O2 reaction according to the following reaction: BPh4 – e → BPh4 → BPh3 + Ph. The requirement of a 1:1:1 ratio of FeII/H+/BPh4 for the maximal formation of FeIV(O) suggested the intermediacy of an iron(III)–hydroperoxo (FeIII–OOH) complex in which one electron from FeII and one electron from BPh4 reduce O2 to the peroxide level. However, the peroxide intermediate could not be trapped for spectroscopic characterization.

In 2009, a similar approach of using external H+ (HClO4) and e (BNAH) sources was successfully employed to trap an FeIII–OOH species during O2 activation by [FeII(N4Py)]2+ (N4Py = N,N-bis(2-pyridylmethyl)-N-bis(2-pyridyl)-methylamine) and [FeII(Bn-TPEN)]2+ (Bn-TPEN = N-benzyl-N,N’,N’-tris(2-pyridylmethyl)-1,2-diaminoethane) complexes (pathway E, Scheme 7).98 In these cases, the dioxygen reactivity was achieved in CH3OH, instead of the previously used aprotic CH3CN. Both the N4Py and Bn-TPEN complexes exhibited ls-FeII (S = 0) centers in CH3CN, but gave hs-FeII complexes with CH3OH as solvent. The change in spin state was accompanied by a change in FeIII/FeII redox potentials. The hs-FeII complexes exhibited significantly lower redox potentials than their low-spin counterparts. As the lower FeIII/ FeII redox potentials favors oxidation of FeII, hs-FeII complexes are well-suited to activate O2. These results demonstrate the importance of solvent effects on the spin state of an iron complex, which in turn, can influence the O2 reactivity of such complexes.98

When the [FeII(TMC)(OTf)2] complex was exposed to similar H+/e sources with the addition of HClO4 (H+) and BNAH (e) in pure CH3CN, the activation of O2 was facilitated, but proceeded directly to the FeIV(O) complex without the observation of an FeIII(OOH) intermediate (pathway D, Scheme 7).98 The instability of the FeIII(OOH) intermediate in this case was suggested as the likely reason for the inability to trap this species. The same strategy of combining external H+/e (HClO4/BPh4) sources with O2 was employed in 2010 to give another example of an FeIIIOOH complex generated from a hs-FeII complex (pathway E, Scheme 7).99 A TPEN-based ligand (L52aH) was used in this work that incorporated a H-bond donor (pivalamido substituent) to stabilize and trap the FeIIIOOH complex.

The Brönsted acids (H+) used in the former FeII-mediated O2 activation processes can be replaced with Lewis acids (LAs) such as Sc(OTf)3. In 2013, The O2-derived formation of [FeIV(O)(TMC)(CH3CN)]2+ using NaBPh4 and Sc(OTf)3 in CH3CN was described.100 It was proposed that the Sc3+ binds to a putative iron(III)–peroxo intermediate to form a Fe3+– (µ–η22–O2)-Sc3+ core, which promotes O–O bond cleavage and formation of the FeIV(O) species. The plausibility of a species like Fe3+–(µ–η22-O2)–Sc3+ was independently demonstrated by preparing this species from the reaction of FeIIIOOH with Sc(OTf)3. However, a later report has suggested that this reaction occurs via a more complex Sc3+-promoted autocatalytic radical chain pathway, rather than via direct O2 activation.101 The addition of Brönsted and Lewis acids to facilitate the O2 activation process is reminiscent of the “push–pull” mechanism, well-developed for heme enzymes such as the peroxidases.4 In the biological systems, protons are carefully shuttled to an iron-bound dioxygen species by nearby protein residues to facilitate cleavage of the O–O bond and release of the distal O-atom as water. In the synthetic systems, simple bimolecular interactions with appropriate H+ or LA metal ions in sufficient concentrations appear to have a similar effect, although the structural aspects and exact timing of formation of the M–OO–H+ (or LA) intermediates are generally not well understood. These interactions are worthy of future study, as are the development of porphyrinoid ligands with tethered groups for shuttling protons or creating hydrogen bonds with M/O2-derived intermediates.

Most of the examples described above relied on the use of external proton and electron sources to reduce O2 and generate FeIIIOOH or FeIV(O) complexes.96,98,99 These separate proton and electron sources can also be replaced with a single H-atom donor to form a nonheme FeIV(O) complex (pathway C, Scheme 7).102 It was noted before that the [FeII(TMC)-(OTf)2]2+ is air-stable in CH3CN. However, the FeII–TMC complex in the presence of an olefin such as cyclohexene, cycloheptene, or cyclooctene led to the rapid formation of an FeIV(O) complex in high yield (>90%). A linear correlation between the rate of FeIV(O) formation and the allylic C–H bond strengths of the alkenes, along with a large KIE, indicated that C–H bond breaking was the rate-determining step. It was proposed that a putative FeIII(O2) complex abstracts an H-atom from the olefinic substrate to form the FeIIIOOH intermediate and alkenyl radical. The alkenyl radical could undergo a rebound reaction with FeIIIOOH to give the FeIV(O) species. Recently in 2015, formation of iron(III)–peroxo and iron (IV)–oxo complexes were shown with the same ligand platform TPEN.103 However, reductive activation of dioxygen in this case was achieved electrochemically.

More recently in 2016, an unusual example of the two-electron reduction of O2 at a single iron center to form an iron–peroxo intermediate was described for an organometallic precursor.104 In this case both of the reducing equivalents were provided by the FeII center, leading to an iron(IV)–peroxo complex. The X-ray crystallographic structures of the heterodinuclear [NiIIFeII] complexes [NiIILFeII(RCN)(η5-C5Me5)]+ (L = N,N′-diethyl-3,7-diazanonane-1,9-dithiolato, R = Et, Me) (Scheme 8) revealed that the NiII and FeII ions are bridged by two thiolato units of the ligand L. The spin state of the FeII ion was characterized as low-spin (S = 0) by various spectroscopic techniques, including Mössbauer, ESR, and 1H NMR spectroscopy. This complex is one of the few examples of a ls-FeII complex that is capable of O2 activation. The strong electron-donating nature of the η5-C5Me5 (Cp*) ligand may facilitate O2 reduction at ls-FeII in this system.104

Scheme 8
O2 Activation by [NiIIFeII] Complexes (Crystal Structure, Bottom Left) To Form an FeIV(O22−) Species (Crystal Structure, Bottom Right) and Subsequent 2e Reduction To Generate H2Oa

Both the FeII(CH3CN) and FeII(EtCN) complexes reacted with O2 at low temperature (−40 and −80 °C for MeCN and EtCN, respectively) to form a brown species with charge-transfer bands at 410 nm (ε = 3000 m−1 cm−1) and 520 nm (ε = 1500 m−1 cm−1), which was crystallized in CH3CN/Et2O. The crystal structure (Scheme 8) revealed that the O2 molecule was coordinated to the Fe center in an η2 (side-on) manner. The O–O bond distance (1.381(3) Å) was consistent with the O2 being reduced to the peroxide level, as seen in other side-on metal–peroxo complexes.105108 However, this distance is slightly shorter than the O–O distance found in another crystallographically characterized complex, [FeIII(TMC)-(OO)]+ (O–O = 1.463(6) Å).109 Isotope labeling experiments with 18O2 confirmed O2 as the sole source of oxygen in the complex, although addition of H2O2 also led to facile exchange of the side-on-bound peroxide ligand. Mössbauer parameters (δ = 0.42 mm s−1, ΔEQ = 0.33 mm s−1) supported the assignment of a +4 oxidation state for the Fe center, and variable-temperature magnetic susceptibility measurements revealed a low-spin (S = 0) diamagnetic ground state for the complex. The high-valent FeIV complexes, obtained in synthetic biomimetic systems, are normally intermediate-spin S = 1 complexes.110 Thus, this peroxo complex is a rare example of a ls-FeIV (S = 0), likely due to the strong-field Cp* donor. Addition of the esource BH4 in combination with the H+ donor EtOH to [NiIILFeIV2-O2)(η5-C5Me5)]+ at −40 °C led to the reduction of peroxide to H2O. A very modest TON = 1.3 was achieved. This work provided an example where 4e reduction of O2 was carried out at a nonheme iron center and both oxygen atoms in O2 were reduced to H2O.

4.3. Nonheme Iron-Mediated Substrate Oxidations Utilizing Dioxygen

There are a number of reports over the years where O2 has been used by nonheme iron complexes to oxidize organic substrates. Herein we focus on some of the select examples of these systems in recent years and categorize them in different subclasses based on the type of substrates that gets oxidized.

4.3.1. Aromatic C–C Bond-Cleaving Reactions

Cleavage of the C–C bonds, particularly in aromatic substrates, is an important class of reactions performed by a number of nonheme iron enzymes, including the intra- and extradiol-cleaving catechol dioxygenases and 2-aminophenol dioxygenase. These enzymes play a crucial role in the biodegradation of the aromatic compounds in bacterial systems. Diol-cleaving dioxygenases utilize a mononuclear iron center and convert catechol substrates to ring-opened products, whereas aminophenol dioxygenase acts on 2-aminophenol substrate to form the ring-opened 2-aminomuconic acid semialdehyde, which then loses water to form an aromatic ring (Scheme 9).111,112

Scheme 9
Reactions Catalyzed by Intra- and Extradiol Cleaving and 2-Aminophenol Dioxygenasesa

A number of synthetic model complexes have been prepared that exhibit intra- and extradiol cleavage activity, and a comprehensive review published in 2004 provided an account of these complexes.6 Extradiol/intradiol dioxygenase activity by nonheme iron complexes was shown with bis(1-alkylimidazol-2-yl)propionate (L) ligands (Scheme 10).111 Each ligand provides two imidazolyl N-atoms and a carboxylate O-atom for metal coordination, mimicking the 2-his-1-carboxylate binding motif observed in the extradiol cleaving dioxygenase. The [FeII(L)(Hdtbc)] complexes (Hdtbc = monodeprotonated 3,5-di-tert-butylcatechol), prepared in situ, reacted rapidly with O2 to give respective [FeIII(L)(dtbc)] complexes. The UV–vis spectrum for the reaction mixture after O2 reactivity (324, 490, 800 nm) matched well with that of independently synthesized [FeIII(L)(dtbc)] complexes. The initially formed [FeIII(L)-(dtbc)] underwent a subsequent slow O2 reaction to give oxidized organic product(s) (Scheme 10). The nature of the product(s) after the completion of the reaction was dependent on the solvent. Performing the reaction in CH3CN led to exclusive formation of the auto-oxidation product 3,5-di-tert-butylbenzoquinone, whereas both the auto-oxidation product (major) and intradiol cleavage products (minor) were obtained in CH3OH. Interestingly, non-coordinating solvents such as CH2Cl2 led to almost equimolar formation of both the intra-and extradiol cleavage products. This result was consistent with the hypothesis that extradiol cleavage product formation is favored when the metal center has a vacant site for dioxygen binding.111,113,114 Although the complexes were not selective for either type of cleavage pathway, addition of the proton donor, [Et3NH]BF4 increased the selectivity toward extradiol cleavage products. In a similar study, the carboxylate arm of the ligand L was replaced with a phenolate unit to mimic the active site coordination environment of intradiol cleaving dioxygenase.115 However, no greater selectivity for intra- or extradiol cleavage pathway was observed when the FeIII complex of the new ligand was reacted with O2.

Scheme 10
Ligands containg N,N,O Donor Atoms and Dioxygen Reaction Products for the FeII–Catecholate Complexesa

In 2008, dioxygenase reactivity of a number of FeIII-catecholate complexes with tetradenatate N4 ligands were investigated (Figure 3a).116 The oxygenation reaction of the FeIII complexes [L′FeIII(Hdtbc)]2+ with a monodeprotonated 3,5-di-tert-butylcatechol unit gave predominantly extradiol cleavage products. The fully deprotonated complex [L′FeIII(dtbc)]+ led to a higher yield for the intradiol cleavage products. It was proposed that an internal proton transfer from the −OH group to one of the cis-pyridyl rings in the monodeprotonated complex resulted in a vacant coordination site (as the protonated pyridine is a very weak donor) at the FeIII center for O2 binding (Figure 3a). As mentioned before, the availability of the vacant O2 binding site was postulated to be a critical component for the extradiol cleavage pathway.

Figure 3
(a) Tetradentate N4 ligands and dioxygenase reactivity for the corresponding mono- and doubly deprotonated catechol complexes. (b) Dioxygenase reactivity for FeIII–catecholate complexes with tridentate N3 ligands. (c) Reaction of iron(II)–aminophenolate ...

A higher selectivity for intradiol cleavage products was obtained when a series of five-coordinate FeIII complexes with isoindoline-based ligands were reacted with O2 (Figure 3b).117 Fully deprotonated 3,5-di-tert-butyl catechol was employed as the substrate and was shown to coordinate with the iron center in a bidentate manner. Here the meridional geometry imposed by the N3 donors of isoindoline was proposed to play a crucial role toward the observed intradiol selectivity for the reaction.117

The C–C bond cleavage reactivity of synthetic nonheme iron complexes was studied for a 2-aminophenol substrate as well.112,118121 Dioxygen reactivity was studied for the nonheme FeII complex [(6-Me3-TPA)FeII(4-tBu-HAP)]- (ClO4), which was prepared using tetradentate 6-Me3-TPA (6-Me3-TPA = tris(6-methyl-2-pyridylmethyl)amine) and bidentate 4-tBu-HAP (4-tBu-HAP = monoanionic 2-amino-4-tert-butylphenolate) ligand (Figure 3c).119 Reaction of [(6-Me3-TPA)FeII(4-tBu-HAP)](ClO4) with O2 in CH3CN immediately formed a metastable FeIII complex (UV–vis and EPR spectra matched well with the independently prepared FeIII complex) with absorption bands at 366, 600, and 934 nm. The 1e-oxidized FeIII complex, formed immediately after the reaction of FeII + O2, slowly converted into the [(6-Me3-TPA)FeIII(4-tert-butyl-2-picolinate)]+ complex (λmax = 660 nm), which was supported by EPR and ESI-MS experiments. 1H NMR and GC-MS analysis of the organic products also revealed the formation of 4-tert-butyl-2-picolinate, an extradiol cleavage product for 4-tBu-HAP. This is in contrast with the dioxygen reactivity of an analogous [(6-Me3-TPA)FeII(dtbc)]+ (dtbc =3,5-di-tert-butylcatecholate) and other FeII–catecholate complexes with tetradentate ligands, where C–C bond cleavage was observed to follow an intradiol pathway.6,116

4.3.2. Synthetic Models for α-Keto/Hydroxy Acid-Dependent Enzymes

α-Keto gluatarate-dependent enzymes are the largest subclass of nonheme iron enzymes that perform a wide range of organic transformations, including hydroxylation, desaturation, and ring closure.122 As the name suggests, α-keto glutarate cofactor is required for enzyme activity, and the binding of this cosubstrate promotes dioxygen activation by the iron center. The α-keto glutarate undergoes a decarboxylation reaction to give succinate, incorporating one O-atom from O2. A high-valent FeIV(O) intermediate was proposed to be the key intermediate in these enzymes and was spectroscopically characterized in a number of cases.8084,123

Despite being the largest member of the nonheme iron enzyme family, examples of synthetic functional model complexes of these systems that utilize O2 are limited.124,125 In 1999, dioxygen reactivity of an FeII complex, coordinated with hydrotris(3,5-diphenylpyrazol-1-yl)borate (TpPh2) and benzoylformate (BF), was reported.125,126 The iron(II) complex [FeII(TpPh2)(BF)] was shown to react with O2 at room temperature to form an arene hydroxylated [FeIII(TpPh2*)(OBz)] complex (Scheme 11). Although an FeIV(O) species was postulated to be the active oxidant, it was not detected by spectroscopic methods. In a subsequent study, it was shown that the FeIV(O) complex could be intercepted in the presence of external substrates such as thioanisole and cyclohexene (Scheme 11).127 When the oxygenation reaction of [FeII(TpPh2)(BF)] was performed in the presence of thioanisole, no ligand hydroxylation reaction was observed. Instead, decarboxylation of benzoylformate (BF) to benzoate (OBz) was observed along with the formation of methyl phenyl sulfoxide (70%). The C–H bond substrates such as DHA and cyclohexene were also used to intercept the putative FeIV(O) species in the oxygenation reaction.

Scheme 11
Interception of a Putative FeIV(O) Intermediate, Generated from the Reaction of [FeII(TpPh2)(BF)] and O2a

Dioxygen reactivity of FeII-α-hydroxy acid complexes received some attention lately128132 and were prepared as functional models for the enzyme CloR.133 The synthetic FeII complex [FeII(TpPh2)(benzilate)] reacted with O2 in benzene to form an FeIII-phenolate complex (λmax = 600 nm), along with the quantitative formation of benzophenone (generated from the decarboxylation of benzilate) (Scheme 12).132 The proposed mechanism for this reaction involves initial formation of an FeIII(O2) species that abstracts a H-atom from the hydroxyl group to generate an FeIIIOOH complex. A subsequent O–O bond cleavage reaction would form an FeIV(O)–OH complex, which was proposed to be the active oxidant for the ligand hydroxylation reaction. Although none of the proposed intermediates were characterized spectroscopically, the FeIV(O) oxidant was intercepted with a number of external substrates such as fluorene, cyclohexene and thioanisole.131 A subsequent study on the interception reactions with sulfides and cyclohexene revealed a nucleophilic reactivity profile for the oxidant, which was confirmed by a Hammett analysis (Scheme 12).130 The possibility of an FeIIOOH or an FeIV(O)–OH species was implicated in the absence of direct spectroscopic evidence.

Scheme 12
Reaction of [FeII(TpPh2)(benzilate)] with O2 and the Interception of Various Active Oxidant Speciesa

Interestingly, the presence of a LA (for example Sc(OTf)3) in the reaction of [FeII(TpPh2)(benzilate)] with O2 switched the nature of the oxidant from being nucleophilic to electrophilic (Scheme 12).129 Interception of the active oxidant with thioanisole substrate gave sulfoxide product only, whereas both the sulfoxide and sulfone were obtained previously in absence of Sc(OTf)3. Hammett analysis with para-substituted ArSMe substrates revealed a negative ρ value (−0.929), which was suggestive of an electrophilic oxidant. A negative ρ value was obtained for various para-substituted styrene substrates as well. An electrophilic FeIV(O)–OH species was implicated as the active oxidant here. The LA was proposed to facilitate the heterolytic cleavage of the O–O bond in the intermediate FeII(OOH) species, leading to the formation of the FeIV(O)– OH complex. It was shown that a protic acid (for e.g. pyridinium perchlorate) could also be used, instead of a LA, to generate the electrophilic oxidant.128 Interestingly, addition of a chloride source into the reaction made the oxidant much more electrophilic (based on Hammett analysis) and led to C–H halogenation along with C–H hydroxylation (Scheme 12). The intermediacy of an iron(IV)–oxo–chloride complex was hypothesized to explain the observed hydroxylation/halogenation reactivity.128

4.3.3. S-Oxygenation Reactions: Dioxygen Reactivity of Iron–Thiolate Complexes

Dioxygen activation by thiolate-ligated iron complexes is of particular biological relevance because of a range of nonheme iron enzymes that activate O2 and employ one or more cysteinate ligands. One example is cysteine dioxygenase (CDO), a nonheme iron enzyme that carries out the dioxygenation of cysteine to cysteine sulfinic acid. The CDO enzyme contains a mononuclear iron active site with three histidine ligands in a facial triad, which is different from the usual two histidine-1-carboxylate binding motif found in most of the other mononuclear nonheme iron enzymes. This enzyme is part of a larger group of related enzymes that can be classified as thiol dioxygenases (cysteamine dioxygenase, cysteine dioxygenase, 3-mercaptopropionate dioxygenase).134 These enzymes are related in that they utilize O2 to convert sulfur substrates to the dioxygenated sulfinic acid products. A number of mechanisms were proposed which involves formation of various iron–oxygen intermediates prior to the S-oxygenation reaction.135139 However, there is currently no direct experimental evidence for any of these O2-derived intermediates. Simplified synthetic model complexes can sometimes provide better access to the trapping and characterization of analogous Fe/O2 intermediates, with greater flexibility in tuning electronic and steric properties. In recent years, efforts have been undertaken to study the dioxygen reactivity of various mononuclear, thiolate-ligated iron(II) compounds with biomimetic ligand environments.

Controlling the oxygenation of sulfur coordinated to iron(II) is challenging because of the potentially facile oxidation of both Fe and S centers, and the possible range of products that could form. Until 2010, previous reports on FeII(SAr) + O2 chemistry described the formation of oxo-bridged dinuclear iron complexes or disulfide products.134 An S-oxygenation reaction (Scheme 13) occurring from FeII(SAr) + O2 was described by our group in 2010.140 A bis-imino pyridine (BIP) ligand (described as LN3S) providing three neutral N donors and a tethered thiolate donor was used in the former study. Reaction of [FeII(LN3S)(OTf)] with O2 in CH2Cl2 (or in MeCN, THF) resulted in an immediate color change, and MS analysis of the reaction mixture was consistent with sulfonate formation (i.e., triple oxygenation at S). Quantitative reversed-phase HPLC confirmed production of the triply S-oxygenated ligand. Isotope labeling (18O2, H218O) proved that oxygen gas was the sole source of O-atoms in the sulfonato complex. The lack of an EPR signal for the final product indicated that the oxygenated complex was in the FeII state. The necessity of a redox-active Fe center to mediate S-oxygenation was demonstrated by the synthesis of the Zn-analogue [ZnII(LN3S)(OTf)], which showed no reactivity toward O2. This work presented the first selective S-oxygenation reaction derived from the reaction of FeIISAr + O2.

Scheme 13
Examples of S-Oxygenation Reactions with Iron(II)–Thiolate Complexes and O2a

Subsequently in 2011, we described the dioxygen reactivity of two new bis-imino pyridine-based FeIISAr complexes, [(iPrBIP)-FeII(SPh)(Cl)] and [(iPrBIP)FeII(SPh)(OTf)] [iPrBIP = 2,6-(ArN═CMe)2C5H3N, Ar = 2,6- iPr2C6H3] (Scheme 13).141 However, unlike the previous example of [FeII(LN3S)(OTf)], the thiolate ligand was added from an exogenous source and not tethered to the BIP ligand. The complexes were prepared from the reaction of [(iPrBIP)FeII(X)2] (X = Cl, OTf) with NaSPh. The presence of hs-FeII(S = 2) in both complexes was confirmed by X-ray crystallography and 1H NMR spectroscopy (in CD2Cl2).

The binding of the exogenous thiolate ligand to [(iPrBIP)-FeII(X)2] facilitated dioxygen activation by the resulting FeIISAr complexes. The [(iPrBIP)FeII(SPh)(Cl)] complex with a PhSligand in a pseudoaxial position reacted with excess O2 in CH2Cl2 to form a green species (λmax = 690 nm, ε ≈ 1500 M−1 cm−1) (Scheme 13). The UV–vis features, mass-spectrometric data and the labeling experiments with 18O2 were suggestive of an iron–oxo product. The sulfur component was oxidized to disulfide (PhS-SPh) (85% yield). In comparison, the complex [(iPrBIP)FeII(SPh)(OTf)] which has the thiolate group in the pseudoequatorial position, reacted with O2 to form an S-oxygenated FeII(SO3Ar) complex (Scheme 13). The ability of both the [(iPrBIP)FeII(SPh)(Cl)] (E1/2 = −0.173 V vs Fc+/Fc) and [(iPrBIP)FeII(SPh)(OTf)] (E1/2 = −0.372 V) to activate O2 was attributed in part to their relatively low FeIII/FeII redox potentials. The nonthiolate-ligated [(iPrBIP)FeII(Cl)2] (E1/2 = 0.025 V) and [(iPrBIP)FeII(OTf)2] (E1/2 = 0.613 V) have significantly more positive E1/2 values, and do not show any reactivity toward O2. The difference in reactivity for [(iPrBIP)-FeII(SPh)(Cl)] versus [(iPrBIP)FeII(SPh)(OTf)] can be attributed to structural features. The former complex has a potential O2 binding site trans to the SPh group, whereas the latter complex has an open site cis to the SPh ligand. The cis orientation for [(iPrBIP)FeII(SPh)(OTf)] allows for close approach of a putative iron-bound superoxide toward the sulfur donor, leading to intramolecular S-oxygenation. This same process is not favored for the trans position of the O2 binding site, and thus [(iPrBIP)FeII(SPh)(Cl)] does not undergo S-oxygenation, but instead only gives disulfide.

These initial CDO model complexes led to triply oxygenated sulfur products.140,141 Our first evidence for the formation of an iron(II)–sulfinate complex from FeII–SR/O2 was obtained with a new tetradentate ligand, N3PySH, which allows for facial N3 coordination to the iron center and a cis orientation of the thiolate donor to the open site on the metal, as seen in the CDO active site.142 The resulting [FeII(N3PyS)(CH3CN)]-(BF4) complex reacted with O2 in CH3OH to form a doubly oxygenated sulfinate complex which was crystallized in the presence of SCN to give the neutral product [FeII(N3PySO2)-(SCN)] (Scheme 14). The crystal structure revealed that the sulfinate group was bound through the S-atom to the iron center. The IR spectrum of the sulfinate complex revealed peaks at 1129 and 1012 cm−1, which were assigned to the asymmetric and symmetric S–O stretching modes, respectively.

Scheme 14
Double Oxygenation of Sulfur, Derived from the Reaction of an FeII–Thiolate Complex and O2a

Another example of a CDO model complex was reported at about the same time and described a thiolate-ligated tris-(pyrazolyl)borate complex [TpMe,Ph FeII(CysOEt)], in which the sulfur donor comes from cysteine ethyl ester (Scheme 14).143 This complex reacted slowly (5–6 h) with O2 in CH2Cl2, and ESI-MS together with isotope labeling (18O2) experiments suggested formation of a doubly oxygenated complex. Mixed labeling experiments with 16O2 /18O2 (1:1) indicated that both O-atoms in the product came from the same O2 molecule. Although crystallographic evidence for the dioxygenated complex was lacking, evidence for sulfinate formation came from isolation of the organic product. In a subsequent report, a similar S-oxygenation reaction was observed with the related cysteamine complex [TpMe,PhFeII-(SCH2CH2NH2)].144 It should be mentioned that there is also related work on a well-defined chromium(III)–superoxo complex, derived from CrII and O2, that reacts with thioether substrates to give sulfoxide products. This system was discussed in the context of the proposed mechanism for CDO.145

5. NONHEME MANGANESE COMPLEXES AND DIOXYGEN

Nonheme manganese centers play important roles in a number of metalloenzymes, such as superoxide dismutase, the oxygen-evolving complex in photosystem II, and ribonucleotide reductase.9,10 In synthetic chemistry, Mn has been targeted for catalyzing the oxidation of organic substrates (e.g., epoxidation), as well as inorganic substrates such as water for renewable energy applications. Studies on the dioxygen reactivity of nonheme Mn complexes goes back to the 1970s. One of the first examples of the reaction of a MnII complex with O2 was reported in 1978, in which a peroxo-bridged dimeric MnIII complex was proposed.146,147 However, subsequent studies with similar MnII complexes indicated that the product was more likely an oxo-bridged Mn2III complex.148 In fact, most of the early reports on the activation of O2 by Mn complexes describe the formation of oxo-bridged multinuclear structures.148152 A few studies reported the formation of mixed oxo/peroxo-bridged complexes.150,151 However, the reaction mechanism and relevant “Mn–O2” intermediates were not defined in most cases. The past few years has seen some development in mononuclear nonheme Mn chemistry, including the characterization of “Mn–O2” species. These studies have led to new mechanistic information, including the identification of different factors that contribute to the reactivity of Mn with O2.

Dioxygen-derived Mn–superoxo species are extremely rare for the nonheme systems. A stable MnIII(O2) species was generated in 2011 using a calixarene ligand platform and was structurally characterized by XRD.153 The peroxo complexes were characterized for a number of nonheme Mn systems. An example of a mononuclear MnIII–peroxo complex, which was synthesized from the reaction of a MnII complex and dioxygen, was reported in 2008.154 A derivative of the tren ligand, H2bupa, which contains two substituted urea arms and one carboxyamidopyridyl donor, was employed in this study. The ligand substituents provide H-bonding groups that can help stabilize a peroxo-bound complex. The five-coordinate MnII complex [MnIIH2bupa] reacted with O2 in the presence of diphenyl hydrazine (DPH) at room temperature to form a green MnIII–peroxo complex with UV–vis maxima at λmax = 660 nm, 490 nm(sh) (Scheme 15). Characterization by FTIR and ESI-MS along with isotope labeling (18O2) studies supported the formation of a monomeric MnIII–peroxo species, [MnIIIH3bupa(O2)]. Parallel mode EPR spectroscopy revealed a spectrum consistent with a quintet (S = 2) Mn ion, which was quantified as 80% of the sample and indicated a +3 oxidation state. However, the protonation state of the peroxo ligand could not be conclusively determined, with either an η1-hydroperoxo or an η2-peroxo species as possibilities (Scheme 15). It was proposed that a MnIII–superoxo intermediate forms initially and then abstracts a H-atom from DPH to give a MnIII–hydroperoxo complex. The expected DPH product, azobenzene, forms nearly quantitatively in this reaction. The MnIII(OOH) complex can convert to an η2-peroxo species by intramolecular proton transfer from MnIII(OOH) to the deprotonated carboxamido arm.

Scheme 15
Catalytic Reduction of O2 to H2O via the Intermediacy of MnIII–Peroxo and MnIII–O(H) Complexesa

The peroxo complex [MnIIIH3bupa(O2)] slowly converted into a MnIIIO(H) species, which was characterized by UV–vis (λmax = 677 nm), ESI-MS, Evan’s method, and X-ray crystallography.155 The hydroxyl proton was shared between the oxo ligand and a N-atom of the carboxamido unit. Interestingly, addition of the H-atom donor DPH to [MnIIIO-(H)(H2bupa)] led to production of H2O and regeneration of the MnII complex (Scheme 15). Thus, the [MnIIH2bupa] complex can serve as a catalyst for the reduction of O2 to H2O, and performing the oxygenation reaction of [MnIIH3bupa] in the presence of excess DPH (20 equiv) produced azobenzene and H2O in excellent yields.

The first example of a structurally characterized MnIII– peroxo complex derived from MnII + O2 was not reported until 2013.13,156 A pentadentate ligand, HSMe2 N4(6-Me-DPEN), containing a –SH group was utilized to prepare the monomeric complex [MnII(SMe2N4(6-Me-DPEN))](BF4) (Scheme 16). The thiolate ligation to the Mn center promoted O2 activation, and the 6-methyl substituents on the pyridine rings provided steric shielding to stabilize the peroxo intermediate. The peroxo species {[MnIII(SMe2N4(6-Me-DPEN))]2(µ-O2)}2+ was generated with O2 at −40 °C in MeCN (λmax = 640 nm). This species had a short lifetime even at −40 °C and converted to a dinuclear MnIII oxo-bridged structure within minutes. Resonance Raman spectroscopy performed on the {[MnIII(SMe2N4(6-Me-DPEN))]2(µ-O2)}2+ showed resonance-enhanced vibrations at 819 cm−1O–O) and 611 cm−1Mn–O), which were assigned with the help of 18O2 isotope-labeling experiments. X-ray-quality crystals of {[MnIII(SMe2N4(6-Me-DPEN))]2(trans-µ-1,2-O2)(BPh4)2· 2CH3CH2CN were obtained at −80 °C from reaction in propionitrile, and the crystal structure revealed a trans-orientation of the two MnIII centers bridged by a µ-1,2-O2 ligand. The O–O bond distance (1.452(5) Å) in the crystal structure was consistent with peroxide formation. The Mn···NPy distances (2.492(3) and 2.410(3) Å) in the crystal structure are significantly longer than the sum of covalent radii of Mn- and N-atoms (2.105 Å). This elongation of the Mn···NPy bonds was attributed to the Jahn–Teller distortions in the MnIII (d4) center and also due to the steric interaction of 6-Me substituents on the pyridine rings with the gem-dimethyl groups. Interestingly, dioxygen reactivity of an analogous alkoxide ligated [FeII(OMe2N4(6-Me-DPEN))](PF6) complex did not yield an oxo- or peroxo-bridged compound, instead a dihydroxo-bridging [FeIII(OMe2N4(6-Me-DPEN))]2(µ-OH)2-(PF6)2 species was isolated and characterized.157 Formation of an active FeIV(O) species was proposed for the reaction, which could abstract a H-atom from the solvent CH3CN molecule. Isotope labeling experiments with CD3CN supported this hypothesis. Although successful use of iodosylbenzene in this reaction indicated the intermediacy of an FeIV(O), no direct evidence for such species could be obtained even at −80 °C.

Scheme 16
Low-Temperature Formation of a Peroxo-Bridged Dimanganese(III) Species from the Reaction of a MnII–Thiolate Complex + O2, and Its Subsequent Conversion to a µ-Oxo-Bridged Dimeric Complexa

A high-valent MnIV(O) complex in which the oxo ligand derives from O2 was synthesized by following the same strategy employed for a related nonheme FeIV(O) complex (pathway A, Scheme 7).87,88 The complex [MnIIIH3buea(O)]2− was synthesized by reacting H6buea and KH with Mn(OAc)2 in the presence of O2.158,159 Subsequent oxidation of the MnIII(O) complex to a MnIV(O) complex was achieved by treating [MnIIIH3buea(O)]2− with [Cp2Fe]BF4.

Recently in 2016, a nonheme mononuclear MnII complex was shown to perform stepwise oxidation of benzylic C–H bonds with O2 as the oxidant.160 In an attempt to synthesize MnII complexes with the dpeo ligand in air, the desired [MnIIBr2(dpeo)2] complex was isolated along with the oxidized complex [MnIIBr2(hidpe)2] (Scheme 17). The crystal structure of the [MnIIBr2(hidpe)2] complex revealed that the benzylic C–H bonds of the dpeo ligand were converted into a ketone group. An intermediate, two-electron-oxidized MnIII(alkoxide) complex [MnIII(hdpeo)2]+ was also isolated, when MnII(ClO4)2 was employed instead of MnBr2. The isolated [MnIII(hdpeo)2]+ complex was shown to react with O2 to form [MnII(hidpeo)2]2+, indicating that the alkoxide complex was an intermediate in the overall four-electron oxidation process. Isotope labeling experiments with 18O2 and H218O suggested that O2 was the source of oxygen in the product. A mechanism involving MnIII(O2) and MnIV(O) intermediates was proposed, but spectroscopic evidence for such intermediates was lacking. The O2-mediated ligand oxidation reaction was specific for Mn2+, as it was shown that Fe2+, Ni2+ or Zn2+ were incapable of performing similar C–H oxidation reactions.

Scheme 17
Stepwise Oxidation of Benzylic C–H Bonds Using O2 by a MnII Complexa

6. CONCLUSIONS AND PERSPECTIVE

Metalloenzymes that activate dioxygen have highly evolved metal active sites that provide both first- and second-coordination-sphere environments optimized for processing O2. These systems provide a roadmap for the synthetic chemist to prepare small-molecule transition metal complexes that are designed to perform similar O2 activation chemistry. However, it has been challenging to follow this roadmap, because of the subtleties in identifying the key components, and incorporating these structural elements in synthetic ligands for practical use in transition metal chemistry. The chemistry of biomimetic model complexes has relied for a long time on utilizing oxidants other than O2, such as H2O2 or O-atom transfer agents such as ArIO or organic peracids, to access and study proposed intermediates along the O2 activation pathway. These studies have led to useful information regarding metal–oxo, metal–peroxo and other intermediates, although the direct use of O2 remains a relatively rare occurrence.

In this Perspective, we have provided an overview of the few iron and manganese biomimetic systems that have been employed to carry out O2-mediated oxidation reactions, and that have provided new insights on the mechanism of O2 activation over the past 10 years. Advanced spectroscopic techniques (e.g., rR, Mössbauer, EPR) and low-temperature methods have allowed researchers to trap and spectroscopically characterize key metastable intermediates. Ligand development has also played a crucial role in the activation of O2 at both heme and nonheme metal centers, and has been an essential factor in the subsequent stabilization of the intermediates. It is evident from the nonheme metal studies that dioxygen activation is favored by a high-spin ground state in the starting metal complex. The requirement for high-spin starting complexes suggests that relatively weak-field ligands are more apt to promote O2 reactivity. The ligand also needs to induce an MIII/MII (M = Fe, Mn) redox potential that is in a region appropriate for O2 reduction. Significant steric encumbrance of the ligand is another key requirement, crudely mimicking the ability of proteins to sequester the metal active site, and preventing oxo-bridged dimerization and other unwanted bimolecular side-reactions. Second-coordination-sphere effects also need to be controlled, and H-bond donors incorporated into the ligand in the appropriate position can sometimes help to stabilize metal–oxygen species.

Despite O2 activation at Fe or Mn being a prominent target for more than 40 years, there are still relatively few complexes that activate O2 in a rationally designed and controlled manner. From this Perspective, it is evident that O2-activating iron complexes are higher in number than the corresponding manganese complexes, and there are clear opportunities to develop Mn complexes that can utilize O2. The generation and characterization of various metal–oxygen intermediates derived exclusively from O2 remains a challenge for both heme and nonheme systems. There are only three synthetic nonheme iron(III)–superoxide species that have been well characterized, and only two of these come from specific reaction with O2. The latter two species are end-on (η1)-bound Fe(O2) and display electrophilic reactivity toward substrates, while the only side-on (η2) superoxide, derived from KO2, revealed both electrophilic and nucleophilic reactivity.161 The differences in reactivity of these Fe(O2) species are not well understood, and more examples are needed of this fundamental O2 adduct to understand the origins of the different reactivities. Similarly, identifying the M/O2 intermediates involved in substrate oxidation, such as the S-oxygenation reactions described herein, is an important, unmet goal. The S-oxygenation reaction is a good example of a highly selective nonheme iron enzyme-mediated reaction that remains challenging to control in a synthetic system.

The use of benign and inexpensive O2 to perform specific and controlled oxidation reactions with the readily available biologically relevant metals Fe and Mn remains a significant challenge for the synthetic chemist. Metalloenzymes can efficiently catalyze O2-dependent oxidations, but their mechanisms of action remain poorly defined in many cases, and advances in synthetic model systems should provide future insights regarding plausible pathways for these transformations. Most of the dioxygen activating biomimetic systems also lack catalytic capability, or their catalytic efficiency is far from the enzymatic scale. Building catalytic reactivity into the synthetic systems with O2 as the oxidant not only would provide systems for comparison with metalloenzymes, but also could provide novel Earth-abundant transition metal catalysts for synthetic transformations.

Acknowledgments

The NIH (D.P.G., GM101153 and GM119374) is acknowledged for financial support.

Footnotes

The authors declare no competing financial interest.

REFERENCES

1. Kovaleva EG, Lipscomb JD. Nat. Chem. Biol. 2008;4:186–193. [PMC free article] [PubMed]
2. Wood PM. Biochem. J. 1988;253:287–289. [PubMed]
3. Poulos TL. Chem. Rev. 2014;114:3919–3962. [PMC free article] [PubMed]
4. Sono M, Roach MP, Coulter ED, Dawson JH. Chem. Rev. 1996;96:2841–2888. [PubMed]
5. Meunier B, de Visser SP, Shaik S. Chem. Rev. 2004;104:3947–3980. [PubMed]
6. Costas M, Mehn MP, Jensen MP, Que L., Jr Chem. Rev. 2004;104:939–986. [PubMed]
7. Abu-Omar MM, Loaiza A, Hontzeas N. Chem. Rev. 2005;105:2227–2252. [PubMed]
8. Ray K, Pfaff FF, Wang B, Nam W. J. Am. Chem. Soc. 2014;136:13942–13958. [PubMed]
9. Law NA, Caudle MT, Pecoraro VL. In: Advances in Inorganic Chemistry. Sykes AG, editor. Vol. 46. New York: Academic Press; 1998. pp. 305–440.
10. Pecoraro VL, Baldwin MJ, Gelasco A. Chem. Rev. 1994;94:807–826.
11. Neu HM, Baglia RA, Goldberg DP. Acc. Chem. Res. 2015;48:2754–2764. [PMC free article] [PubMed]
12. Nam W. Acc. Chem. Res. 2015;48:2415–2423. [PubMed]
13. Kovacs JA. Acc. Chem. Res. 2015;48:2744–2753. [PMC free article] [PubMed]
14. Goldberg DP. Acc. Chem. Res. 2007;40:626–634. [PubMed]
15. Oloo WN, Que L., Jr Acc. Chem. Res. 2015;48:2612–2621. [PubMed]
16. Liu W, Groves JT. Acc. Chem. Res. 2015;48:1727–1735. [PubMed]
17. Cook SA, Borovik AS. Acc. Chem. Res. 2015;48:2407–2414. [PMC free article] [PubMed]
18. Ohta T, Liu JG, Naruta Y. Coord. Chem. Rev. 2013;257:407–413.
19. Kryatov SV, Rybak-Akimova EV, Schindler S. Chem. Rev. 2005;105:2175–2226. [PubMed]
20. Smith LJ, Kahraman A, Thornton JM. Proteins: Struct., Funct., Genet. 2010;78:2349–2368. [PubMed]
21. Groves JT, Nemo TE, Myers RS. J. Am. Chem. Soc. 1979;101:1032–1033.
22. Che CM, Lo VK, Zhou CY, Huang JS. Chem. Soc. Rev. 2011;40:1950–1975. [PubMed]
23. Groves JT. J. Inorg. Biochem. 2006;100:434–447. [PubMed]
24. Cohen IA, Caughey WS. Biochemistry. 1968;7:636–641. [PubMed]
25. Kao OHW, Wang JH. Biochemistry. 1965;4:342–347.
26. Balch AL, Chan YW, Cheng RJ, Lamar GN, Latosgrazynski L, Renner MW. J. Am. Chem. Soc. 1984;106:7779–7785.
27. Chin D-H, Lamar GN, Balch AL. J. Am. Chem. Soc. 1980;102:5945–5947.
28. Chin D-H, Lamar GN, Balch AL. J. Am. Chem. Soc. 1980;102:4344–4350.
29. Chin D-H, Balch AL, Lamar GN. J. Am. Chem. Soc. 1980;102:1446–1448.
30. Takeuchi M, Kodera M, Kano K, Yoshida Z-i. J. Mol. Catal. A: Chem. 1996;113:51–59.
31. Grinstaff MW, Hill MG, Labinger JA, Gray HB. Science. 1994;264:1311–1313. [PubMed]
32. Ellis PE, Lyons JE. Coord. Chem. Rev. 1990;105:181–193.
33. Sengupta K, Chatterjee S, Samanta S, Dey A. Proc. Natl. Acad. Sci. U. S. A. 2013;110:8431–8436. [PubMed]
34. Carver CT, Matson BD, Mayer JM. J. Am. Chem. Soc. 2012;134:5444–5447. [PubMed]
35. Costentin C, Dridi H, Savéant J-M. J. Am. Chem. Soc. 2015;137:13535–13544. [PubMed]
36. Collman JP, Devaraj NK, Decréau RA, Yang Y, Yan Y-L, Ebina W, Eberspacher TA, Chidsey CED. Science. 2007;315:1565–1568. [PMC free article] [PubMed]
37. Bettelheim A, Kuwana T. Anal. Chem. 1979;51:2257–2260.
38. Kobayashi N, Nevin WA. Appl. Organomet. Chem. 1996;10:579–590.
39. Kim E, Helton ME, Wasser IM, Karlin KD, Lu S, Huang HW, Möenne-Loccoz P, Incarvito CD, Rheingold AL, Honecker M, Kaderli S, Zuberbuhler AD. Proc. Natl. Acad. Sci. U. S. A. 2003;100:3623–3628. [PubMed]
40. Ghiladi RA, Kretzer RM, Guzei I, Rheingold AL, Neuhold YM, Hatwell KR, Zuberbuhler AD, Karlin KD. Inorg. Chem. 2001;40:5754–5767. [PubMed]
41. Mittra K, Chatterjee S, Samanta S, Sengupta K, Bhattacharjee H, Dey A. Chem. Commun. 2012;48:10535–10537. [PubMed]
42. Liu J-G, Ohta T, Yamaguchi S, Ogura T, Sakamoto S, Maeda Y, Naruta Y. Angew. Chem., Int. Ed. 2009;48:9262–9267. [PubMed]
43. Liu J-G, Shimizu Y, Ohta T, Naruta Y. J. Am. Chem. Soc. 2010;132:3672–3673. [PubMed]
44. Nagaraju P, Ohta T, Liu JG, Ogura T, Naruta Y. Chem. Commun. 2016;52:7213–7216. [PubMed]
45. Anderson JS, Gallagher AT, Mason JA, Harris TD. J. Am. Chem. Soc. 2014;136:16489–16492. [PubMed]
46. Weschler CJ, Hoffman BM, Basolo F. J. Am. Chem. Soc. 1975;97:5278–5280. [PubMed]
47. Hoffman BM, Szymanski T, Brown TG, Basolo F. J. Am. Chem. Soc. 1978;100:7253–7259.
48. Hoffman BM, Weschler CJ, Basolo F. J. Am. Chem. Soc. 1976;98:5473–5482. [PubMed]
49. Moxon NT, Fielding PE, Gregson AK. J. Chem. Soc., Chem. Commun. 1981;0:98–99.
50. Lever ABP, Wilshire JP, Quan SK. J. Am. Chem. Soc. 1979;101:3668–3669.
51. Uchida K, Naito S, Soma M, Onishi T, Tamaru K. J. Chem. Soc., Chem. Commun. 1978:217–218.
52. Dismukes GC, Sheats JE, Smegal JA. J. Am. Chem. Soc. 1987;109:7202–7203.
53. Watanabe T, Ama T, Nakamoto K. Inorg. Chem. 1983;22:2470–2472.
54. Battioni P, Bartoli JF, Leduc P, Fontecave M, Mansuy D. J. Chem. Soc., Chem. Commun. 1987:791–792.
55. Sakurai H, Mori Y, Shibuya M. Inorg. Chim. Acta. 1989;162:23–25.
56. Jung J, Neu HM, Leeladee P, Siegler MA, Ohkubo K, Goldberg DP, Fukuzumi S. Inorg. Chem. 2016;55:3218–3228. [PMC free article] [PubMed]
57. Neu HM, Jung J, Baglia RA, Siegler MA, Ohkubo K, Fukuzumi S, Goldberg DP. J. Am. Chem. Soc. 2015;137:4614–4617. [PMC free article] [PubMed]
58. Prokop KA, Goldberg DP. J. Am. Chem. Soc. 2012;134:8014–8017. [PubMed]
59. Zaragoza JPT, Siegler MA, Goldberg DP. Chem. Commun. 2016;52:167–170. [PMC free article] [PubMed]
60. Jung J, Ohkubo K, Prokop-Prigge KA, Neu HM, Goldberg DP, Fukuzumi S. Inorg. Chem. 2013;52:13594–13604. [PMC free article] [PubMed]
61. Jung J, Ohkubo K, Goldberg DP, Fukuzumi S. J. Phys. Chem. A. 2014;118:6223–6229. [PMC free article] [PubMed]
62. Cho K, Leeladee P, McGown AJ, DeBeer S, Goldberg DP. J. Am. Chem. Soc. 2012;134:7392–7399. [PubMed]
63. McGown AJ, Kerber WD, Fujii H, Goldberg DP. J. Am. Chem. Soc. 2009;131:8040–8048. [PubMed]
64. Liu H-Y, Mahmood MHR, Qiu S-X, Chang CK. Coord. Chem. Rev. 2013;257:1306–1333.
65. Schwalbe M, Dogutan DK, Stoian SA, Teets TS, Nocera DG. Inorg. Chem. 2011;50:1368–1377. [PubMed]
66. Simkhovich L, Goldberg I, Gross Z. Inorg. Chem. 2002;41:5433–5439. [PubMed]
67. Vogel E, Will S, Tilling AS, Neumann L, Lex J, Bill E, Trautwein AX, Wieghardt K. Angew. Chem., Int. Ed. Engl. 1994;33:731–735.
68. Solomon EI, Light KM, Liu LV, Srnec M, Wong SD. Acc. Chem. Res. 2013;46:2725–2739. [PMC free article] [PubMed]
69. Bruijnincx PCA, van Koten G, Klein Gebbink RJM. Chem. Soc. Rev. 2008;37:2716–2744. [PubMed]
70. Xing G, Diao Y, Hoffart LM, Barr EW, Prabhu KS, Arner RJ, Reddy CC, Krebs C, Bollinger JM., Jr Proc. Natl. Acad. Sci. U. S. A. 2006;103:6130–6135. [PubMed]
71. Schatz M, Raab V, Foxon SP, Brehm G, Schneider S, Reiher M, Holthausen MC, Sundermeyer J, Schindler S. Angew. Chem., Int. Ed. 2004;43:4360–4363. [PubMed]
72. Peterson RL, Himes RA, Kotani H, Suenobu T, Tian L, Siegler MA, Solomon EI, Fukuzumi S, Karlin KD. J. Am. Chem. Soc. 2011;133:1702–1705. [PMC free article] [PubMed]
73. Cho J, Woo J, Nam W. J. Am. Chem. Soc. 2010;132:5958–5959. [PubMed]
74. Cho J, Kang HY, Liu LV, Sarangi R, Solomon EI, Nam W. Chem. Sci. 2013;4:1502–1508. [PMC free article] [PubMed]
75. Qin K, Incarvito CD, Rheingold AL, Theopold KH. Angew. Chem., Int. Ed. 2002;41:2333–2335. [PubMed]
76. Shan X, Que L., Jr Proc. Natl. Acad. Sci. U. S. A. 2005;102:5340–5345. [PubMed]
77. Chiang CW, Kleespies ST, Stout HD, Meier KK, Li P-Y, Bominaar EL, Que L, Jr, Münck E, Lee WZ. J. Am. Chem. Soc. 2014;136:10846–10849. [PMC free article] [PubMed]
78. Stout HD, Kleespies ST, Chiang C-W, Lee W-Z, Que L, Münck E, Bominaar EL. Inorg. Chem. 2016;55:5215–5226. [PubMed]
79. Oddon F, Chiba Y, Nakazawa J, Ohta T, Ogura T, Hikichi S. Angew. Chem., Int. Ed. 2015;54:7336–7339. [PubMed]
80. Price JC, Barr EW, Tirupati B, Bollinger JM, Jr, Krebs C. Biochemistry. 2003;42:7497–7508. [PubMed]
81. Krebs C, Fujimori DG, Walsh CT, Bollinger JM., Jr Acc. Chem. Res. 2007;40:484–492. [PMC free article] [PubMed]
82. Hoffart LM, Barr EW, Guyer RB, Bollinger JM, Jr, Krebs C. Proc. Natl. Acad. Sci. U. S. A. 2006;103:14738–14743. [PubMed]
83. Matthews ML, Krest CM, Barr EW, Vaillancourt FH, Walsh CT, Green MT, Krebs C, Bollinger JM., Jr Biochemistry. 2009;48:4331–4343. [PMC free article] [PubMed]
84. Fujimori DG, Barr EW, Matthews ML, Koch GM, Yonce JR, Walsh CT, Bollinger JM, Jr, Krebs C, Riggs-Gelasco PJ. J. Am. Chem. Soc. 2007;129:13408–13409. [PubMed]
85. Eser BE, Barr EW, Frantom PA, Saleh L, Bollinger JM, Jr, Krebs C, Fitzpatrick PF. J. Am. Chem. Soc. 2007;129:11334–11335. [PMC free article] [PubMed]
86. Panay AJ, Lee M, Krebs C, Bollinger JM, Jr, Fitzpatrick PF. Biochemistry. 2011;50:1928–1933. [PMC free article] [PubMed]
87. MacBeth CE, Golombek AP, Young VG, Jr, Yang C, Kuczera K, Hendrich MP, Borovik AS. Science. 2000;289:938–941. [PubMed]
88. Lacy DC, Gupta R, Stone KL, Greaves J, Ziller JW, Hendrich MP, Borovik AS. J. Am. Chem. Soc. 2010;132:12188–12190. [PMC free article] [PubMed]
89. Bigi JP, Harman WH, Lassalle-Kaiser B, Robles DM, Stich TA, Yano J, Britt RD, Chang CJ. J. Am. Chem. Soc. 2012;134:1536–1542. [PubMed]
90. Biswas AN, Puri M, Meier KK, Oloo WN, Rohde GT, Bominaar EL, Münck E, Que L., Jr J. Am. Chem. Soc. 2015;137:2428–2431. [PubMed]
91. England J, Martinho M, Farquhar ER, Frisch JR, Bominaar EL, Münck E, Que L., Jr Angew. Chem., Int. Ed. 2009;48:3622–3626. [PMC free article] [PubMed]
92. Puri M, Que L., Jr Acc. Chem. Res. 2015;48:2443–2452. [PMC free article] [PubMed]
93. Puri M, Biswas AN, Fan R, Guo Y, Que L., Jr J. Am. Chem. Soc. 2016;138:2484–2487. [PubMed]
94. Kim SO, Sastri CV, Seo MS, Kim J, Nam W. J. Am. Chem. Soc. 2005;127:4178–4179. [PubMed]
95. Rohde J-U, In JH, Lim MH, Brennessel WW, Bukowski MR, Stubna A, Münck E, Nam W, Que L., Jr Science. 2003;299:1037–1039. [PubMed]
96. Thibon A, England J, Martinho M, Young VG, Jr, Frisch JR, Guillot R, Girerd JJ, Münck E, Que L, Jr, Banse F. Angew. Chem., Int. Ed. 2008;47:7064–7067. [PMC free article] [PubMed]
97. Park H, Bittner MM, Baus JS, Lindeman SV, Fiedler AT. Inorg. Chem. 2012;51:10279–10289. [PMC free article] [PubMed]
98. Hong S, Lee YM, Shin W, Fukuzumi S, Nam W. J. Am. Chem. Soc. 2009;131:13910–13911. [PubMed]
99. Martinho M, Blain G, Banse F. Dalton Trans. 2010;39:1630–1634. [PubMed]
100. Li F, Van Heuvelen KM, Meier KK, Münck E, Que L., Jr J. Am. Chem. Soc. 2013;135:10198–10201. [PMC free article] [PubMed]
101. Nishida Y, Lee Y-M, Nam W, Fukuzumi S. J. Am. Chem. Soc. 2014;136:8042–8049. [PubMed]
102. Lee YM, Hong S, Morimoto Y, Shin W, Fukuzumi S, Nam W. J. Am. Chem. Soc. 2010;132:10668–10670. [PubMed]
103. Ségaud N, Anxolabéhère-Mallart E, Sénéchal-David K, Acosta-Rueda L, Robert M, Banse F. Chem. Sci. 2015;6:639–647.
104. Kishima T, Matsumoto T, Nakai H, Hayami S, Ohta T, Ogo S. Angew. Chem., Int. Ed. 2016;55:724–727. [PubMed]
105. Cho J, Sarangi R, Kang HY, Lee JY, Kubo M, Ogura T, Solomon EI, Nam W. J. Am. Chem. Soc. 2010;132:16977–16986. [PMC free article] [PubMed]
106. Cho J, Sarangi R, Annaraj J, Kim SY, Kubo M, Ogura T, Solomon EI, Nam W. Nat. Chem. 2009;1:568–572. [PMC free article] [PubMed]
107. Seo MS, Kim JY, Annaraj J, Kim Y, Lee Y-M, Kim S-J, Kim J, Nam W. Angew. Chem., Int. Ed. 2007;46:377–380. [PubMed]
108. Karlsson A, Parales JV, Parales RE, Gibson DT, Eklund H, Ramaswamy S. Science. 2003;299:1039–1042. [PubMed]
109. Cho J, Jeon S, Wilson SA, Liu LV, Kang EA, Braymer JJ, Lim MH, Hedman B, Hodgson KO, Valentine JS, Solomon EI, Nam W. Nature. 2011;478:502–505. [PMC free article] [PubMed]
110. McDonald AR, Que L., Jr Coord. Chem. Rev. 2013;257:414–428.
111. Bruijnincx PCA, Lutz M, Spek AL, Hagen WR, Weckhuysen BM, van Koten G, Gebbink RJM. J. Am. Chem. Soc. 2007;129:2275–2286. [PubMed]
112. Chakraborty B, Bhunya S, Paul A, Paine TK. Inorg. Chem. 2014;53:4899–4912. [PubMed]
113. Jo DH, Que L., Jr Angew. Chem., Int. Ed. 2000;39:4284–4287.
114. Bugg TDH, Lin G. Chem. Commun. 2001:941–952.
115. Bruijnincx PCA, Lutz M, Spek AL, Hagen WR, van Koten G, Gebbink RJM. Inorg. Chem. 2007;46:8391–8402. [PubMed]
116. Mayilmurugan R, Stoeckli-Evans H, Palaniandavar M. Inorg. Chem. 2008;47:6645–6658. [PubMed]
117. Váradi T, Pap JS, Giorgi M, Párkányi L, Csay T, Speier G, Kaizer J. Inorg. Chem. 2013;52:1559–1569. [PubMed]
118. Chatterjee S, Paine TK. Inorg. Chem. 2015;54:1720–1727. [PubMed]
119. Chakraborty B, Paine TK. Angew. Chem., Int. Ed. 2013;52:920–924. [PubMed]
120. Bittner MM, Lindeman SV, Popescu CV, Fiedler AT. Inorg. Chem. 2014;53:4047–4061. [PMC free article] [PubMed]
121. Bittner MM, Lindeman SV, Fiedler AT. J. Am. Chem. Soc. 2012;134:5460–5463. [PMC free article] [PubMed]
122. Hausinger RP. Crit. Rev. Biochem. Mol. Biol. 2004;39:21–68. [PubMed]
123. Wong SD, Srnec M, Matthews ML, Liu LV, Kwak Y, Park K, Bell CB, III, Alp EE, Zhao J, Yoda Y, Kitao S, Seto M, Krebs C, Bollinger JM, Jr, Solomon EI. Nature. 2013;499:320–323. [PMC free article] [PubMed]
124. Ha EH, Ho RYN, Kisiel JF, Valentine JS. Inorg. Chem. 1995;34:2265–2266.
125. Mehn MP, Fujisawa K, Hegg EL, Que L., Jr J. Am. Chem. Soc. 2003;125:7828–7842. [PubMed]
126. Hegg EL, Ho RYN, Que L., Jr J. Am. Chem. Soc. 1999;121:1972–1973.
127. Mukherjee A, Martinho M, Bominaar EL, Münck E, Que L., Jr Angew. Chem., Int. Ed. 2009;48:1780–1783. [PMC free article] [PubMed]
128. Chatterjee S, Paine TK. Angew. Chem., Int. Ed. 2016;55:7717–7722. [PubMed]
129. Chatterjee S, Paine TK. Angew. Chem., Int. Ed. 2015;54:9338–9342. [PubMed]
130. Paria S, Chatterjee S, Paine TK. Inorg. Chem. 2014;53:2810–2821. [PubMed]
131. Paria S, Que L, Jr, Paine TK. Angew. Chem., Int. Ed. 2011;50:11129–11132. [PubMed]
132. Paine TK, Paria S, Que L., Jr Chem. Commun. 2010;46:1830–1832. [PMC free article] [PubMed]
133. Pojer F, Kahlich R, Kammerer B, Li SM, Heide L. J. Biol. Chem. 2003;278:30661–30668. [PubMed]
134. McQuilken AC, Goldberg DP. Dalton Trans. 2012;41:10883–10899. [PMC free article] [PubMed]
135. Kumar D, Thiel W, de Visser SP. J. Am. Chem. Soc. 2011;133:3869–3882. [PubMed]
136. Ye S, Wu X, Wei L, Tang D, Sun P, Bartlam M, Rao Z. J. Biol. Chem. 2007;282:3391–3402. [PubMed]
137. Pierce BS, Gardner JD, Bailey LJ, Brunold TC, Fox BG. Biochemistry. 2007;46:8569–8578. [PubMed]
138. Simmons CR, Liu Q, Huang Q, Hao Q, Begley TP, Karplus PA, Stipanuk MH. J. Biol. Chem. 2006;281:18723–18733. [PubMed]
139. McCoy JG, Bailey LJ, Bitto E, Bingman CA, Aceti DJ, Fox BG, Phillips GN., Jr Proc. Natl. Acad. Sci. U. S. A. 2006;103:3084–3089. [PubMed]
140. Jiang Y, Widger LR, Kasper GD, Siegler MA, Goldberg DP. J. Am. Chem. Soc. 2010;132:12214–12215. [PMC free article] [PubMed]
141. Badiei YM, Siegler MA, Goldberg DP. J. Am. Chem. Soc. 2011;133:1274–1277. [PMC free article] [PubMed]
142. McQuilken AC, Jiang Y, Siegler MA, Goldberg DP. J. Am. Chem. Soc. 2012;134:8758–8761. [PMC free article] [PubMed]
143. Sallmann M, Siewert I, Fohlmeister L, Limberg C, Knispel C. Angew. Chem., Int. Ed. 2012;51:2234–2237. [PubMed]
144. Sallmann M, Braun B, Limberg C. Chem. Commun. 2015;51:6785–6787. [PubMed]
145. Cho J, Woo J, Nam W. J. Am. Chem. Soc. 2012;134:11112–11115. [PubMed]
146. Mabad B, Tuchagues JP, Hwang YT, Hendrickson DN. J. Am. Chem. Soc. 1985;107:2801–2802.
147. Coleman WM, Taylor LT. Inorg. Chim. Acta. 1978;30:L291–L293.
148. Kipke CA, Scott MJ, Gohdes JW, Armstrong WH. Inorg. Chem. 1990;29:2193–2194.
149. Kitajima N, Singh UP, Amagai H, Osawa M, Morooka Y. J. Am. Chem. Soc. 1991;113:7757–7758.
150. Bossek U, Weyhermüller T, Wieghardt K, Nuber B, Weiss J. J. Am. Chem. Soc. 1990;112:6387–6388.
151. Bhula R, Gainsford GJ, Weatherburn DC. J. Am. Chem. Soc. 1988;110:7550–7552.
152. Frederick FC, Taylor LT. Polyhedron. 1986;5:887–893.
153. Liu L-L, Li H-X, Wan L-M, Ren Z-G, Wang H-F, Lang J-P. Chem. Commun. 2011;47:11146–11148. [PubMed]
154. Shook RL, Gunderson WA, Greaves J, Ziller JW, Hendrich MP, Borovik AS. J. Am. Chem. Soc. 2008;130:8888–8889. [PMC free article] [PubMed]
155. Shook RL, Peterson SM, Greaves J, Moore C, Rheingold AL, Borovik AS. J. Am. Chem. Soc. 2011;133:5810–5817. [PMC free article] [PubMed]
156. Coggins MK, Sun X, Kwak Y, Solomon EI, Rybak-Akimova E, Kovacs JA. J. Am. Chem. Soc. 2013;135:5631–5640. [PMC free article] [PubMed]
157. Coggins MK, Toledo S, Kovacs JA. Inorg. Chem. 2013;52:13325–13331. [PMC free article] [PubMed]
158. Parsell TH, Behan RK, Green MT, Hendrich MP, Borovik AS. J. Am. Chem. Soc. 2006;128:8728–8729. [PubMed]
159. MacBeth CE, Gupta R, Mitchell-Koch KR, Young VG, Jr, Lushington GH, Thompson WH, Hendrich MP, Borovik AS. J. Am. Chem. Soc. 2004;126:2556–2567. [PubMed]
160. Deville C, Padamati SK, Sundberg J, McKee V, Browne WR, McKenzie C. J. Angew. Chem., Int. Ed. 2016;55:545–549. [PubMed]
161. Hong S, Sutherlin KD, Park J, Kwon E, Siegler MA, Solomon EI, Nam W. Nat. Commun. 2014;5:5440. [PMC free article] [PubMed]