End-On Copper(I) Superoxo and Cu(II) Peroxo and Hydroperoxo Complexes Generated by Cryoreduction/Annealing and Characterized by EPR/ENDOR Spectroscopy

Abstract

In this report, we investigate the physical and chemical properties of monocopper Cu(I) superoxo and Cu(II) peroxo and hydroperoxo complexes. These are prepared by cryoreduction/annealing of the parent [LCu^I(O₂)]⁺ Cu(I) dioxygen adducts with the tripodal, N₄-coordinating, tetradentate ligands L = ^PVtmpa, ^DMMtmpa, TMG₃tren and are best described as [LCu^II(O₂^•−)]⁺ Cu(II) complexes that possess end-on (η¹-O₂^•−) superoxo coordination. Cryogenic γ-irradiation (77 K) of the EPR-silent parent complexes generates mobile electrons from the solvent that reduce the [LCu^II(O₂^•−)]⁺ within the frozen matrix, trapping the reduced form fixed in the structure of the parent complex. Cryoannealing, namely progressively raising the temperature of a frozen sample in stages and then cooling back to low temperature at each stage for examination, tracks the reduced product as it relaxes its structure and undergoes chemical transformations. We employ EPR and ENDOR (electron–nuclear double resonance) as powerful spectroscopic tools for examining the properties of the states that form. Surprisingly, the primary products of reduction of the Cu(II) superoxo species are metastable cuprous superoxo [LCu^I(O₂^•−)]⁺ complexes. During annealing to higher temperatures this state first undergoes internal electron transfer (IET) to form the end-on Cu(II) peroxo state, which is then protonated to form Cu(II)–OOH species. This is the first time these methods, which have been used to determine key details of metalloenzyme catalytic cycles and are a powerful tools for tracking PCET reactions, have been applied to copper coordination compounds.

Graphical Abstract

INTRODUCTION

Iron- and copper-containing metalloenzymes capable of the reductive transformation of molecular oxygen (O₂, dioxygen) to generate powerful oxidizing intermediates are ubiquitous in biological systems. During catalysis these intermediates are capable of efficient substrate oxidation (dehydrogenation) or O atom insertion reactions, which produce critically needed biomolecules and/or metabolic precursors. Heme (iron porphrinate)-containing enzymes are perhaps the most recognized, including the widespread cytochrome P450 (CYP) enzyme family of monooxygenases, which perform reactions such as the synthesis of hormones and cholesterol, as well as the transformation of many molecules and toxins.^1,2 Nonheme iron enzymes also activate dioxygen for oxidative processes. These include the diiron-containing soluble methane monooxygenase (sMMO) and related desaturases,³ as well as monoiron enzymes, such as catechol dioxygenases and α-ketoglutarate (αKG)-dependent monooxygenases.⁴

Copper proteins likewise exhibit a diversity of active-site structures and reactivities. Enzymes with dicopper active sites include monooxygenases/oxidases such as tyrosinases, NspF (converts aminophenol natural product to a N-nitroso derivative), and catechol oxidases.⁵ However, the reaction mechanisms of peptidyl glycine α-hydroxylating monooxygenase (PHM) and dopamine-β-monooxygenase (DβM) are most often described by a mechanism in which substrate prebinding leads to a [Cu^II(O₂^•−)]⁺ (cupric superoxide)-mediated hydrogen atom abstraction (HAA) and net O atom transfer, all occurring at a single-copper ion active site that possesses an unusual S_thioether Met ligand.^5,6,7 A recent theoretical-computational study⁷ suggests an alternative reaction mechanism involving a protein open-to-closed conformational change and substrate HAA chemistry carried out by a Cu₂–O₂-derived binuclear intermediate. Finally, lytic polysaccharide monooxygenases (LPMOs)^4e,8,9 and particulate methane monooxygenases (pMMO’s)¹⁰ use a mono-Cu center to oxygenate their respective substrates (both with C–H bond dissociation energies >100 kcal/mol).

Our interest here is in the chemistry of mono-Cu centers, in particular ligand–copper(I)–dioxygen complexes relevant to practical copper-ion-mediated substrate oxidations,¹¹ and modeling O₂ activation in copper biochemistry.^5,12 In this report, we study the reduction/protonation of several previously characterized cupric superoxo complexes, [Cu^II(O₂^•−)]⁺, formed by low-temperature oxygenation of LCu^I complexes. The three complexes studied employ the family of tripodal tetradentate ligands ^DMMtmpa, ^PVtmpa, and TMG₃tren, as shown in Figure 1.

Figure 1.

Parent ligand TMPA and cationic copper(II) superoxo complexes employing derivatized TMPA ligands, as formed from Cu(I) complexes on reaction with dioxygen, used in this study.

The Cu^I/O₂ chemistry of the two ligands that are derivatives of TMPA (Figure 1) has been extensively characterized. On the basis of physical-spectroscopic and structural studies, these ligand–Cu(I)–dioxygen adducts (designated below as [LCu^I(O₂)]⁺ species) are best described as Cu^II superoxo complexes possessing an end-on (η¹-O₂^•−)¹³ superoxo coordination (ν_O–O = 1119 cm⁻¹ for the TMPA parent ligand)¹⁹.¹³ The third complex, [(TMG₃tren)Cu^II(O₂^•−)]⁺, has been well established as an end-on cupric superoxo complex (Figure 1).¹⁴ These pentacoordinate [LCu^II(O₂^•−)]⁺ solution species possess S = 1 ground states, having ferromagnetically coupled copper(II) (i.e., d⁹) and O₂^•− lone electrons, and do not exhibit X-band EPR.¹⁵ The [(tmpa)Cu^II(X)]^+,2+ complexes (X = Cl⁻, N₃⁻, H₂O, MeCN) possess trigonal-bipyramidal (TBP) coordination geometries,¹⁶ as do [{(tmpa)Cu^II}₂(μ-1,2-O₂²⁻)]^{2+ 17} and [(TMG₃tren)Cu^II(O₂^•−)]⁺.^14b TBP geometries are maintained in solution, as shown by distinctive UV–vis d–d band patterns. This geometry yields a d-orbital manifold with the Cu(II) odd electron in a d_z²-based orbital, as signaled by a diagnostic “reverse-axial” EPR spectra seen for [LCu^II(X)]^+,2+ (g_⊥ > g_‖ ≅ 2.00; A_⊥ ≈ 200–300 MHz, A_∥ ≈ 300–400 MHz), as opposed to the familiar spectrum of a tetragonal Cu(II) with an odd electron in a d_x²–y²-based orbital (g_∥ > g_⊥ ≅ 2.00; A_∥ ≈ 500 MHz, A_⊥ ≈ 50 MHz).^16a,d,18

The complex [(tmpa)Cu^II(O₂^•−)]⁺ was first detected in UV–vis monitored stopped-flow studies under cryogenic conditions, as a very rapidly formed kinetic intermediate, with a rate constant for O₂ binding of k_on = 1.8 × 10⁴ M⁻¹ s⁻¹ at −90 °C, in EtCN as solvent.²⁰ Subsequent studies by flash- and-trap laser spectroscopy demonstrated that k_on = 1.3 × 10⁹ M⁻¹ s⁻¹ (RT) in the noncoordinating solvent tetrahydrofuran (THF), a value greater than those found for reactions of ferrous hemes (synthetic or biological) with O₂.²¹

At −80 °C, [(tmpa)Cu^II(O₂^•−)]⁺ is metastable and transforms rapidly to a thermodynamic product, the binuclear complex [{(tmpa)Cu^II}₂(μ-1,2-O₂²⁻)]²⁺ (Scheme 1).²² However, all of the superoxo complexes depicted in Figure 1 are stable in solution at T < −80 °C. Strong donor ligands and/or H-bonding substituents such as the pivalamido group in [(^PVtmpa)Cu^II(O₂^•−)]⁺ (Figure 1) (i) considerably enhance both the [LCu^II(O₂^•−)]⁺ complex thermal stability and reactivity with substrates—H-atom abstraction (HAA) from phenols or weak C–H bond substrates,^13,23 (ii) while they suppress diversion to binuclear peroxo dicopper(II) analogues.^19,24 In addition to the direct reactivity of peroxo and hydroperoxo Cu(II) complexes in H-atom abstraction and possibly in nucleophilic reactions, they may be important intermediates in Cu-based monooxygenase catalytic cycles (vide supra). For these, oxidative chemistries are initiated by the formation of an O₂ adduct, such as a cupric superoxo (Cu^II–O–O^•−) species with end-on (η¹) or side-on (η²) O₂^•− ligation²⁵ or a binuclear peroxo dicopper(II) complex, and proceed through peroxo, hydroperoxo, or even other highly oxidizing intermediates (see Figure S1). Thus, as part of the effort to understand Cu-catalyzed oxidations in general, it is imperative to elucidate the structures and reactivity of [LCu^II(O₂^•−)]⁺, [LCu^II(O₂²⁻)], and [LCu^II(⁻OOH)]⁺ intermediates.

Scheme 1.

Formation of the Binuclear Complex [{(tmpa)Cu^II}₂(μ-1,2-O₂²⁻)]²⁺

In this report, we extend earlier investiations of the physical and chemical properties of Cu^II superoxo and Cu^II hydroperoxo compounds^25b–d,26 and present a detailed examination of all three products of reduction–protonation of the [LCu^I(O₂)]⁺ adducts given in the title, for the complexes with L = ^PVtmpa, ^DMMtmpa, TMG₃tren (Figure 1), in MeTHF (2-methyltetrahydrofuran) as solvent. We employ EPR and ENDOR (electron–nuclear double resonance) as powerful spectroscopic tools to examine the properties of the states that form. ENDOR, a method that monitors the NMR spectra of nuclei associated with an EPR-active center, allows for the observation of nuclear hyperfine couplings that are unresolved in the EPR spectra of those centers.²⁷ From the g values of such a center and the hyperfine couplings to its nuclei, its metal-ion valence, ligand covalency parameters, and coordination geometries can be identified.

The investigations are initiated by cryoreduction of the EPR-silent parent complexes. In this procedure, γ-irradiation at 77 K, using a ⁶⁰Co source, ionizes solvent molecules, generating energetic mobile electrons. These “hot electrons” proceed to reduce the [LCu^II(O₂^•−)]⁺ centers within the frozen matrix, typically creating a reduced form trapped in a structure comparable to that of the parent complex for study by EPR/ENDOR methods.^27a,28 Reactions of the reduced species can then be examined by a procedure denoted cryoannealing: namely, progressively raising the temperature of a frozen sample in stages and then cooling back to low temperature at each stage for spectroscopic characterization. This process allows a reduced complex to be tracked as it relaxes its structure and possibly undergoes chemical transformations at higher temperatures. In the present case the final step in cryoannealing the MeTHF frozen matrix occurs upon warming to T > 116 K, at which temperature the solvent melts, with further changes possible upon warming of the fluid phase. This is the first time these methods, which have been used to determine key details of metalloenzyme catalytic cycles^27a–i and are powerful tools for tracking PCET reactions,^27b have been applied to copper coordination compounds.

MATERIALS AND METHODS

General Experimental Details.

All solvents and chemicals were purchased from commercial sources and used as received unless stated otherwise. Tetrahydrofuran (THF) and 2-methyltetrahydrofuran (MeTHF) were distilled over sodium metal with benzophenone under argon, deoxygenated with argon, and then stored over 4 Å sieves. Pentane was distilled 4× over calcium hydride under Ar and deoxygenated with Ar purging.

1-Hydroxy-2,2,6,6-tetramethyl-piperidine (TEMPO-H) and 1-Deuteroxy-2,2,6,6-tetramethylpiperidine (TEMPO-D) Synthesis.

TEMPO-H was synthesized by following a literature procedure.²⁹ TEMPO–D was synthesized by following the same procedure, but with D₂O instead of H₂O.

Ligand Syntheses.

^DMMtmpa, ^PVtmpa, and TMG₃tren were all prepared according to literature procedures.³⁰ Deuterated (d₆) ^DMMtmpa (^D-DMMtmpa, with the three ligand backbone methylene groups deuterated) was prepared by refluxing 200 mg of ^DMMtmpa in MeOD for 48 h. MeOD was removed by rotary evaporation, and the solid thus obtained was dissolved in 20 mL of hot diethyl ether and filtered through a cotton plug. The solution was then layered with 100 mL of n-hexanes and placed in a −30 °C freezer. After a solid product was obtained, the solvent was decanted and the precipitate was dried. An NMR spectrum of the product indicated 75% incorporation of deuterium into the ^DMMtmpa methylene positions.

Syntheses of LCu^I.

[(^PVtmpa)Cu^I(MeCN)]B(C₆F₅)₄, [(TMG₃tren)Cu^I(MeCN)]B(C₆F₅)₄, and both [(^DMMtmpa)Cu^ICO]-B(C₆F₅)₄ and [(^D-DMMtmpa)Cu^ICO]B(C₆F₅)₄ were prepared according to literature methods.^23,31

Syntheses of LCu^II(Cl)₂.

[(^DMMtmpa)Cu^IICl]Cl and [(^PVtmpa)-Cu^IICl]Cl were synthesized according to a literature procedure.³² Solutions (1 mM) for EPR samples of [(^DMMtmpa)Cu^IICl]Cl were prepared in butyronitrile, while 1 mM EPR samples of [(^PVtmpa)-Cu^IICl]Cl were prepared in ethanol.

EPR/ENDOR Sample Preparation.

All stock solutions for EPR/ENDOR spectroscopy samples were prepared in a Vacuum Atmospheres glovebox. Samples (0.25 mM) were prepared in volumetric flasks by dissolving the appropriate amount of LCu^I solid in MeTHF, and 2.5 equiv trifluoroacetic acid or deuterated trifluoroacetic were added to the stock solutions as appropriate. Stock solutions were then placed in X- or Q-band EPR tubes that were then capped with septa and sealed with Parafilm. All samples were cooled using a cold bath, and the temperature was monitored using an Omega Engineering thermocouple. Cupric superoxide samples in MeTHF were prepared at −135 °C using a 4/1 MeTHF/THF bath cooled with liquid nitrogen, while samples of the peroxide dicopper(II) complex [{(L)Cu^II)₂(μ-1,2-O₂²⁻)](B(C₆F₅)₄)₂ were prepared at −80 °C with an acetone bath cooled with liquid nitrogen. Oxygenated samples using ¹⁶O₂ were prepared by passing O₂ through a Drierite column and needle into the solution. ¹⁷O₂ samples were prepared with 37% ¹⁷O₂ purchased from Cambridge Isotope Laboratories by using a breakflask attached to an evacuated Schlenk flask; the ¹⁷O₂ was then added to samples using a three-way syringe and the solution purged (to remove excess dioxygen) by bubbling with argon for 15 s. Product solutions were then immediately frozen in liquid nitrogen.

Synthesis of [(^DMMtmpa)Cu^II(⁻OOH)]B(C₆F₅)₄ and [(^DMMtmpa)Cu^II(⁻OOD)]B(C₆F₅)₄.

A stock solution (0.25 mM) of [(^DMMtmpa)Cu^ICO]B(C₆F₅)₄ in MeTHF was prepared in the glovebox, and 10 equiv of TEMPO-H or TEMPO-D were added. The solution was then transferred to the EPR tube, which was capped with a septum and sealed with Parafilm. The samples were cooled to −135 °C using a 4/1 MeTHF/THF bath cooled with liquid nitrogen; the temperature was monitored using an Omega Engineering thermocouple. Samples were oxygenated by passing O₂ through a Drierite column and syringe needle into the solution, and the solution was purged (to remove excess dioxygen) by bubbling with argon for 15 s. The cupric superoxide sample thus formed was allowed to react with the TEMPO-H or TEMPO-D already present for 1 h. Product solutions were then immediately frozen in liquid nitrogen.

To increase the relative population of monocopper [LCu^II(O₂^•−)]⁺ species (relative to binuclear peroxo complexes that could form; vide supra), all EPR and ENDOR spectroscopy experiments were conducted at low concentrations (e.g., 0.25 mM [(^DMMtmpa)Cu^I]⁺). Under these conditions, the concentration of the peroxo-bridged binuclear complex was less 0.04 mM.¹⁹

EPR and ENDOR Spectroscopy.

X-band continuous wave (CW) EPR spectra were recorded on a Bruker ESP 300 spectrometer equipped with an Oxford Instruments ESR 900 continuous He flow cryostat. All CW Q-band EPR and ENDOR spectra were recorded at 2 K in dispersion mode, under “rapid passage” conditions, which give an absorption-display line shape.^{10a 2}H Mims Pulsed ENDOR measurements were collected at 2 K on a spectrometer described previously,³³ with a SpinCore PulseBlaster ESR_PRO 400 MHz digital word generator and an Agilent Technologies Acquiris DP235 500 MS/s digitizer using SpecMan4EPR software.³⁴ The Mims pulse sequence (π/2–τ–π/2–TRF–π/2–τ–echo) was employed.³⁵ Simulations of EPR spectra were performed using the program Easyspin with the pepper function for powdered samples.³⁶

RESULTS

In this report we first describe the experiments that characterize the primary product trapped upon cryoreduction of the suite of Cu^IO₂-derived complexes. We then focus on the behavior of cryoreduced [(^DMMtmpa)Cu^I(O₂)]⁺, characterizing the conversion of its primary cryoreduction product to relaxed states upon successive stages of cryoannealing.

Primary Product of Cryoreduction.

1n the case of the end-on LCu^I(O₂) complexes being examined, given their actual electronic structure, [LCu^II(O₂^•−)]⁺, one must consider multiple conceivable cryoreduction products (Scheme 2), where species shown in red are EPR-active. First and foremost, the common result in previous cryoreduction studies of Fe^II(O₂) complexes is localization of the added electron on the O₂ moiety.^27j,37 In the present case such a result would yield a trapped copper(II) peroxo complex. However, in addition, there is precedent^27j,37 for localization of the added electron on the metal ion, which in this case would generate O₂^•− coordinated to copper(I). Moreover, as it is extraordinarily difficult to completely dry a solvent, either one of these products could pick up a proton during reduction to directly yield the protonated form of the O₂-derived moiety. This possibility is enhanced by adding trifluoroacetic acid (TFA) to the solution, with the prospect that it would H-bond to the [Cu^I(O₂)]⁺ reactant, thus offering an easily abstracted proton to the immediate product of reduction. Measurements now described identify the actual cryoreduction product.

Scheme 2.

Conceivable Products of [LCu^II(O₂^•−)]⁺ Cryoreduction

Radiolytic reduction of each of the diamagnetic [LCu^I(O₂)]⁺ complexes (≡[LCu^II(O₂^•−)]⁺) in MeTHF at 77 K results in the generation of a new paramagnetic species (A) that shows the same axial EPR spectra in Q-band and X-band EPR for all of the complexes studied (Figure 2 for L = ^DMMtmpa; see Figure S2 for L = ^PVtmpa, TMG₃tren). The simulation shown gives g_∥ = 2.52 ≫ g_⊥ ≈ 1.95; the strong broad signal at g ≈ 2 is due to trapped radiolytically generated free radicals. Multiple characteristics of this spectrum immediately indicate that it is not associated with a copper(II) center. First, although the “symmetry” of the g tensor, namely g_∥ > g_⊥, is reminiscent of a tetragonally coordinated copper(II) center with a d_x²–y² odd electron, the large value of g_∥ argues against this assignment. Even more telling is the absence of the large ^63,65Cu hyperfine splittings of the g_∥ feature of the X-band spectrum invariably seen for such a Cu(II) center (such splittings commonly are washed out in Q-band spectra). Furthermore, rapid electron-spin relaxation causes the loss of the X-band EPR spectrum for species A at temperatures above 50 K, whereas within our experience such copper(II) complexes never exhibit this characteristic.

Figure 2.

(top) Q-band absorption (black) CW EPR spectrum for the 0.25 mM [(^DMMtmpa)Cu^I(O₂)]⁺ adduct (≡[(^DMMtmpa)-Cu^II(O₂ ^•−)]⁺) in MeTHF exposed to γ-irradiation at 77 K with a dose of 3 Mrad and a simulation (red). Instrument settings: modulation amplitude 1 G, microwave power 0.7 mW, microwave frequency 35.09 GHz, T = 2 K. (bottom) X-band derivative CW EPR spectra for 0.25 mM [(^DMMtmpa)Cu^I(¹⁶O₂)]⁺ and [(^DMMtmpa)-Cu^I(¹⁷O₂)]⁺ (37% ¹⁷O enrichment) adducts in MeTHF γ-irradiated at 77 K with a dose of 3 Mrad. The doublet of sharp peaks marked by asterisks is due to hydrogen atoms generated by γ-irradiation of the quartz tube, and the strong signal at g = 2 belong to radicals in the solvent matrix created by γ-irradiation. Instrument settings: modulation amplitude 8 G; microwave power 2 mW, microwave frequency 9.373, measurement temperature 10 K.

A more detailed examination confirms this conclusion. The coordination sphere of a metal center that undergoes radiolytic cryoreduction while it is trapped in a frozen matrix at 77 K cannot rearrange, and thus the cryoreduced product is expected to retain the geometry of the parent complex.²⁸ If species A trapped upon cryoreduction of [LCu^II(O₂^•−)]⁺ at 77 K were the end-on peroxo [LCu^II(O₂²⁻)] or hydroperoxo [LCu^II(⁻O₂H)]⁺ species that retains the trigonal-bipyramidal structure of the precursor, it would show a copper(II) EPR spectrum characteristic of this symmetry, with g_⊥ > g_∥^16a,b,18a accompanied by a well-resolved ^63,65Cu hyperfine structure at both g_∥ and g_⊥, as seen for other [LCu^II(X)]ⁿ⁺ compounds (e.g., X = Cl⁻; Figure S3) and in particular for [LCu^II(⁻O₂H)]⁺ (see below). However, this is not true for A. In addition, details of the EPR spectra of [LCu^II(X)]ⁿ⁺ complexes depend on the nature of the ligands L, whereas that of A does not vary with L (Figure S2).

Finally, ¹⁴N and ¹H ENDOR spectra of A, formed from the [(^DMMtmpa)Cu^II(O₂^•−)]⁺ complex (Figure 3), confirm that it does not have a copper(II) center. First, if A were a copper(II) species, the presence of nitrogens from ^DMMtmpa in its coordination sphere would generate well-resolved ¹⁴N ENDOR signals, such as those exhibited by the chemically prepared [(^DMMtmpa)Cu^II(⁻OOH)]⁺ complex, but species A does not show such signals (Figure 3; see further discussion in the Supporting Information). The absence of such signals is not merely the result of “detectability” issues. As further discussed in the Supporting Information, if A were a Cu(II) complex with a d_x²–y² odd electron, the ligand pyridyl ¹⁴N signals would be at a frequency (greater hyperfine coupling) even higher than those for the amine ¹⁴N of [(^DMMtmpa)-Cu^II(⁻OOH)]⁺ and their ENDOR signal would be far more intense. Thus, such a complex would give ¹⁴N signals that are far easier to see than the quite strong signals from [(^DMMtmpa)Cu^II(⁻OOH)]⁺, in contrast to the experimental absence of any ¹⁴N signals. Second, the ¹H ENDOR spectra of A formed by the cryoreduced L = ^DMMtmpa, ^PVtmpa, and TMG₃tren Cu^I(O₂) adducts (Figure S4) are all very similar and do not show the relatively strongly coupled nonexchangeable ¹H signals from the ligand methylene protons seen in LCu^II(X) centers, including the [(^DMMtmpa)Cu^II(⁻OOH)]⁺ complex (discussed below and shown in the Supporting Information). When they are taken together, these spectroscopic data indicate that the species A generated by cryoreduction of the [LCu^II(O₂^•−)]⁺ parent complexes are not copper(II) centers formed by electron addition to the superoxo moiety. The cryoreductive conversion of [LCu^II(O₂^•−)]⁺ to A must instead involve electron addition to the Cu(II) ion to generate end-on copper(I)-superoxo complexes, [LCu^I(O₂^•−)]⁺. As we now show in detail, the properties of A confirm this conclusion.

Figure 3.

35 GHz, 2 K CW ¹⁴N ENDOR spectra, (black) Spectrum for [(^DMMtmpa)Cu^II(⁻OOH)]⁺ in MeTHF taken at g_⊥ = 2.2, with responses from two types of ¹⁴N. As discussed with additional detail in the Supporting Information, the more intense pattern centered at ~18 MHz (“goalpost” centered at the ¹⁴N Larmor frequency (∇)) is assignable to the amine ¹⁴N; the low-frequency, and therefore less intense, pattern (which extends to still lower frequencies as indicated) is assigned to the three pyridyl nitrogens. (red) Spectrum of A, the product of cryoreduction of [(^DMMtmpa)Cu^II(O₂ ^•−)]⁺ in MeTHF, taken at g_⊥ = 1.95. Instrument settings: microwave power 25 db, modulation amplitude 2 G, RF sweep rate, 0.5 MHz/s, 35 scans.

This conclusion is supported by the observation of ¹⁷O hyperfine broadening of the A EPR spectrum even upon use of O₂ with a low level of ¹⁷O enrichment (37%) to prepare the parent adduct [(^DMMtmpa)Cu^I(^16,17O₂^•−)]⁺ (Figure 2, bottom). This assignment is further supported by the g tensor for species A, which is typical for the ²Π_3/2 state of O₂^•−. The influence of bonding interactions in a nonlinear Cu(I)–(O–O)^•− linkage can be described as creating a local C_2ν crystal field that splits the π-orbital degeneracy of the O₂^•− ion, placing the unpaired electron in the π*(2p_y) antibonding orbital normal to the Cu–O–O plane (Figure 4). As described by eq 1, this gives rise to a near-axial tensor with g_∥ ≫ g_⊥ ≳ g_e, with the g_∥ direction being along the internuclear O₂^•− axis (Figure 4).³⁸ In these equations, λ is the spin–orbit coupling constant, Δ is the π_y*–π_x* splitting, and l represents a correction to the angular momentum about the internuclear axis.³⁸

$$\begin{gathered} g_{\parallel }=g_{\text{e}}+2l{\left[\frac{{\lambda }^2}{{\lambda }^2+{\text{Δ}}^2}\right]}^{1/2};g_{\perp }\cong g_{\text{e}}{\left[\frac{{\text{Δ}}^2}{{\lambda }^2+{\text{Δ}}^2}\right]}^{1/2} \\ =g_{\text{e}}+2l{\left[\frac{1}{1+{\left(\frac{\text{Δ}}{\lambda }\right)}^2}\right]}^{1/2};\cong g_{\text{e}}{\left[\frac{{\left(\frac{\text{Δ}}{\lambda }\right)}^2}{1+{\left(\frac{\text{Δ}}{\lambda }\right)}^2}\right]}^{1/2} \end{gathered}$$ (1)

Figure 4.

(left) Schematic representation of the O₂^•− π* orbitals split by an energy Δ through the influence of progressively stronger bonding in a nonlinear Cu–O–O complex, (right) Plot of the superoxide g values as a function of Δ /λ, with λ being the spin–orbit coupling constant (eq 1).

As seen in Figure 4, these equations fully describe the g tensor for species A when Δ /λ ≈ 4.3. It is well-known that the dependence of the EPR spectra of superoxide ions (O₂^•−) on their surroundings can be described as reflecting variations in the crystal-field parameter that splits the π* orbitals (Δ), which are degenerate in an idealized axially symmetric environment.^38,39 In particular, superoxide ions in alkali halide hosts show a strongly anisotropic EPR signal, with g_‖ ≈ 2.46 and g_⊥ ≈ 1.95 for the KCl host, values remarkably similar to these for species A.^38,40 In short, the question posed by Scheme 2 has been answered by EPR and ENDOR measurements: cryoreduction of the [LCu^I(O₂)]⁺ = [LCu^II(O₂^•−)]⁺ complexes generates an end-on superoxide radical coordinated to a diamagnetic copper(I) ion.

The implicit assumption in the above discussion is that the superoxo ligand in [LCu^I(O₂^•−)]⁺ (A) has not abstracted a proton (i.e., from solvent or other source) upon cryoreduction to generate the hydroperoxo species, [LCu^I(^•O₂H)]. This is confirmed by a comparison of the g tensor and ¹H ENDOR of A with the properties of the well-known ^•O₂H radical.⁴¹ The π* orbitals of the nonlinear ^•O–O–H are split by such a large value of Δ /λ that g_‖ < 2.1, in sharp contrast with g_‖ ≈ 2.46 for A, and the ^•O₂H proton exhibits hyperfine couplings, A ≈ 40 MHz, not present for A (Figure S4). Indeed, the isolation of the parent freeze-quenched Cu^II(O₂^•−)⁺ moiety from interactions with its environment in the 77 K frozen solution is illustrated by cryoreduction of [(TMG₃tren)Cu^I(O₂)]⁺ with added trifluoroacetic acid. A previous study indicated that TFA forms an H-bond with the superoxo ligand of the [(TMG₃tren)Cu^II(O₂^•−)]⁺ complex.⁴² Nonetheless, we find here that adding TFA to [(TMG₃tren)Cu^II(O₂^•−)]⁺ in MeTHF neither alters the EPR signal of the cryoreduced sample (Figure S5) nor introduces an exchangeable ¹H ENDOR response from an interacting proton (Figure S6), further demonstrating that this state has not abstracted a proton to generate the hydroperoxo species [LCu^I(^•O₂H)]. Thus, A, the primary product of cryoreduction of the dioxygen adduct [(TMG₃tren)Cu^I(O₂)]⁺, indeed is “simply” the end-on superoxo complex, [(TMG₃tren)Cu^I(O₂^•−)]⁺ (A; Scheme 3).

Scheme 3.

Primary Product of [(TMG₃tren)Cu^I(O₂)]⁺ Cryoreduction A

Cryoannealing.

Overview.

For clarity, we first summarize here the multistep cryoannealing process and then proceed to describe the individual stages; additional data are presented in the Supporting Information. We progressively annealed A = [(^DMMtmpa)Cu^I(O₂^•−)]⁺ at multiple temperatures up to 260 K, at each stage recooling the sample to collect EPR spectra (Figure 5 and Figure S7). Cryoannealing the frozen solid containing the complex A at T = 90 K for 5 min allowed its relaxation and transformation to a new species with a dramatically different EPR spectrum, denoted B (Figure 5 and Figure S8), which is stable at this temperature. In turn B transforms to a third species, C, at T ≈ 116/117 K, where MeTHF becomes fluid. Species C has an EPR spectrum very similar to that of B but has a definitively different ¹H ENDOR spectrum (Figures 6 and 7). Complex C is stable in fluid solution to T ≈ 135 K, but attempts to warm it further were compromised by its instability along with autoxidation of unreduced [(^DMMtmpa)Cu^I(O₂)]⁺ (Figure S7). We now describe the characterization of complexes B and C.

Figure 5.

X-band CW EPR spectra: (black) radiolytically cryoreduced 0.25 mM [(^DMMtmpa)Cu^I(O₂)]⁺ adduct, A, in MeTHF; (red) after annealing at 90 K for 5 min to generate B; (blue) after further annealing at 117 K for 2 min, creating C; (pink) 0.25 mM [(^DMMtmpa)Cu^II(⁻OOH)]⁺ adduct in MeTHF prepared chemically as described in Materials and Methods. The 117 K annealed spectrum and the [(^DMMtmpa)Cu^II(⁻OOH)]⁺ spectrum are both well-simulated with the parameters g = [2.21, 2.20, 1.96], A_⊥ = 220 MHz, and A_‖ = 330 MHz. Instrument settings: microwave frequency 9.37 GHz; modulation amplitude = 10 G; microwave power = 2 mW; T = 10 K (black) and 20 K for others.

Figure 6.

(top) 2 K Q-band, CW ¹H ENDOR spectra (g = 2.23) for [(^DMMtmpa)Cu^II(⁻OOH)]+ (red) and [(^DMMtmpa)Cu^II(⁻OOD)]⁺ prepared chemically (black). (bottom) 2 K Mims pulsed ²H ENDOR spectrum for [(^DMMtmpa)Cu^II(⁻OOD)]⁺, with the frequency axis stretched to correspond to the ¹H axis. CW instrument settings: modulation amplitude 2 G, microwave power 25 db; RF scan rate 1 MHz/s (¹H CW) and 0.1 MHz/s (Mims), 30 scans, T = 2 K, microwave frequency 35.2 GHz. Mims instrument settings: π = 90 ns, τ = 500 ns, repetition time 100 ms, t_RF = 60 μs, RF tail 10 μs, 125 scans, microwave frequency 34.648 GHz.

Figure 7.

35 GHz 2 K ¹H ENDOR spectra collected near g_⊥, in the field region where there is no overlap of the EPR signal of B with radiation-produced solvent radicals (Figure S8). Spectra: complex B (black), the radiolytically cryoreduced 0.25 mM [(^DMMtmpa)-Cu^I(O₂)]⁺ adduct in MeTHF after annealing at 90 K for 5 min; complex C (blue), generated by further annealing at 117 K for 2 min; [(^DMMtmpa)Cu^II(⁻OOH)]+ (red) prepared chemically. Instrument settings: T = 2 K, modulation amplitude 2 G, radio frequency bandwidth broadened to 60 kHz, 1 MHz/s RF scan speed, 30 scans.

Annealing at 90 K.

Complex A = [(^DMMtmpa)Cu^I(O₂^•−)]⁺, formed directly by cyroreduction of the [(^DMMtmpa)Cu^I(O₂)]⁺ trapped at 77 K, relaxes upon annealing for 5 min at 90 K to the new species B showing a reverse-axial EPR signal with g_⊥ ≈ 2.23 > g_‖, characteristic of a trigonal-bipyramidal copper(II) d_z² ground state (Figure 5 (X-band) and Figure S8 (Q-band)). As is typical for TBP Cu(II) complexes, this signal exhibits a resolved ^63,65Cu hyperfine structure at g_⊥; there is clearly ^63,65Cu hyperfine structure at g_‖ as well, but details of the g_‖ ≈ 2 region are obscured by a signal from free radicals generated during cryoreduction (Figure 5 and Figure S8). The EPR data thus show that the annealing/relaxation at 90 K of the superoxo radical primary cryoreduction product A has led to internal electron transfer (IET) in which the copper(I) of [(^DMMtmpa)Cu^I(O₂^•−)] (A) is oxidized to the EPR-active copper(II) state while the superoxo radical is reduced to the closed-shell peroxo state (Scheme 4). In the next subsection we show that comparisons of proton ENDOR spectra of this species with those of the state obtained upon annealing B to higher temperature establish that the peroxo moiety of species B remains unprotonated.

Scheme 4.

Conversion of A to B by IET during Cryoannealing

Annealing to T ≈ 116 K.

The solvent-matrix radical signal almost completely disappears upon annealing B to T ≥ 116 K, the melting point of MeTHF, revealing the complete EPR signal for a further-relaxed copper(II) species, C (Figure 5 and Figure S8). The EPR signal of C has g_⊥ ≈ 2.23 and g_‖ ≈ 2.06, with well-resolved ^63,65Cu hyperfine splittings associated with both g values, as expected for a TBP geometry (legend of Figure 5). The spectra of B and C are quite similar, but g_⊥ and the ^63,65Cu hyperfine structure of C differ subtly from those in the spectrum of B. Instead, the g values and ^63,65Cu hyperfine splittings of C are those of the mononuclear end-on hydroperoxo [(^DMMtmpa)Cu^II(⁻OOH)]⁺ complex state prepared by HAA from TEMPO-H by the [(^DMMtmpa)Cu^I/O₂]⁺ adduct (Figure 5 and Figures S7 and S8).¹⁹

Identity of C.

To test the resulting expectation that C is indeed the hydroperoxo state, we first characterized the ¹H ENDOR of the Cu(II)–OOH hydroperoxo proton by comparing the properties of the [(^DMMtmpa)Cu^II(⁻OOH/D)]⁺ isotopologue complexes, as separately prepared by the chemical reaction of [(^DMMtmpa)Cu^II(O₂^•−)]⁺ with TEMPO-H/D: their EPR spectra are identical, but their ENDOR spectra are not. The ¹H ENDOR response of [(^DMMtmpa)-Cu^II(⁻OOD)]⁺ shows a loss of intensity and subtle changes in shape in comparison to those of [(^DMMtmpa)Cu^II(⁻OOH)]⁺ in spectra taken across their EPR envelopes (Figure S9). In particular, at fields in the g_⊥ region (Figure 6 and Figure S9) the spectra of [(^DMMtmpa)Cu^II(⁻OOH)]⁺contain a doublet feature with a splitting of A ≈ 7 MHz, which upon closer inspection is a superposition of two signals. As a result of H/D substitution, those peaks in the ¹H spectra of [(^DMMtmpa)-Cu^II(⁻OOD)]⁺ with ^A ≈ 6–7 MHz lose intensity and are no longer composite, while in the ¹H spectra of [(^DMMtmpa)-Cu^II(⁻OOH)]⁺ there are broader “wings” as the field approaches g_⊥, and these are suppressed upon H/D substitution. The association of these differences in the ¹H ENDOR spectra with contributions from the hydroperoxo proton is confirmed by the appearance of ²H ENDOR signals for [(^DMMtmpa)Cu^II(⁻OOD)]⁺ that correspond to the loss of ¹H intensity in the ¹H ENDOR response of [(^DMMtmpa)-Cu^II(⁻OOH)]⁺ (Figure 6).

This exchangeable ¹H coupling of A ≈ 7 MHz for [(^DMMtmpa)Cu^II(⁻OOH)]⁺ is similar to that of the hydroperoxo proton of low-spin Fe(III)–OOH centers, A ≈ 10 MHz.^37b The nonexchangeable ¹H signals for [(^DMMtmpa)-Cu^II(⁻OOH/D)]⁺ (Figure 6) are attributable to protons of the tripodal ligand. For example, as seen in Figure S10, deuteration of the hydrogen atoms of the three backbone methylene groups causes the loss of ¹H intensity with a broad range range of couplings, with maximum values of A ≈ 10–12 MHz.

Having thus characterized the chemically formed [(^DMMtmpa)Cu^II(⁻OOH)]⁺, we then compared its properties to those of C. Not only do their EPR spectra agree (Figure 5) but also the ¹H ENDOR spectra of the cryoannealed species C correspond well, both showing the ¹H intensity from the exchangeable hydroperoxo proton (Figure 7). Together, these EPR/ENDOR measurements establish that annealing B at T ≥ 116 K leads to the formation of C = [(^DMMtmpa)-Cu^II(⁻OOH)]⁺. A plausible source of the hydroperoxo proton acquired during cryoannealing is a trace of water in the MeTHF that becomes available in the fluid solution.

Identity of B.

With these measurements and conclusions regarding C, we can now complete the assignment of B. Despite the similarity between the EPR spectra of species B, generated by annealing A at 90 K, with that of C = [(^DMMtmpa)Cu^II(⁻OOH)]⁺, formed by annealing of B at T > 116 K, their ¹H ENDOR spectra taken near g_⊥ differ tellingly: the response from the exchangeable hydroperoxide proton that appears in the ¹H spectrum of chemically prepared [(^DMMtmpa)Cu^II(⁻OOH)]⁺ and of C is absent in the ¹H spectrum of B (Figure 7), indicating that B lacks a hydroperoxo proton. It is important to note that the absence of this proton and its ENDOR signal for B is supported by an examination of the spectrum of the chemically prepared [(^D-DMMtmpa)Cu^II(⁻OOH)]⁺, in which the methylene hydrogen atoms of the ^DMMtmpa ligand are deuterated (Figure S9). Comparing the signal of B with that of the latter shows that the residual, low-intensity features in the ¹H spectrum of B that extend to frequencies associated with the hydroperoxo proton of C are in fact instead attributable to signals from the ligand methylene protons, as described above (Figure S10).

This absence of intensity from a hydroperoxo ¹H in the spectrum of species B thus indicates that B is indeed the unprotonated mononuclear peroxo intermediate [(^DMMtmpa)-Cu^II(O₂²⁻)] (Scheme 4). The further conclusion that B exhibits end-on peroxo binding (Cu^II–O–O configuration), rather than side-on binding, is supported by its EPR spectrum with g_⊥ > g_‖ ≈ 2, which is characteristic of trigonal-bipyramidal copper(II) coordination enforced by the tetradentate ligand, and thus of the peroxo ligand bound axially to copper by only a single oxygen atom.

DISCUSSION

In this report, we have applied the cryoreduction/annealing protocol to frozen solutions of [LCu^I(O₂)]⁺ Cu(I) dioxygen adducts, utilizing a set of related tripodal tetradentate N₄ ligands to copper, ^DMMtmpa, ^PVtmpa, and TMG₃tren (Figure 1), whose O₂ adducts had been characterized as cupric superoxide ([LCu^II(O₂^•−)]⁺) coordination complexes. Scheme 5 summarizes the overall outcome of the cryoreduction/annealing of these species.

Scheme 5.

Overall Outcome of Cryoreduction/Cryoannealing of [LCu^II(O₂^•−)]⁺ Complexes

γ-irradiative cryoreduction of frozen solutions of the [LCu^II(O₂^•−)]⁺ species produces the reduced species (A) trapped in the geometry of the parent complex at the 77 K temperature of irradiation. Progressive annealing (warming) of the frozen solution to higher temperatures, first below and then above the solvent (MeTHF) freezing point, causes successive transformation of the initial product to the new species B and C through electron-transfer and/or protonation processes (Scheme 5). These intermediates have been characterized by EPR/ENDOR spectroscopy and correspond to those formed prior to O–O cleavage in P-450 monooxygenases and in the proposed mono-Cu monooxygenase cycles of Figure S1. Thus, their physical properties, reactivity, and an understanding of their chemical interrelationships are of central importance in chemical and biochemical copper-ion-mediated O₂ activation through reductive transformation to highly reactive oxidizing intermediates.

Species A and Superoxo Complexes.

Athough it was perhaps to be expected that the initial reduction of a [LCu^II(O₂^•−)]⁺ complex (Figure 1) would convert the superoxide radical anion to the peroxide dianion, creating the EPR-active Cu(II) peroxide, instead it is the copper(II) ion that is reduced, yielding A = [LCu^I(O₂^•−)], with a superoxide radical anion bound to a diamagnetic cuprous ion (Scheme 3). This assignment of A is based on multiple observations, as detailed in Results.

An early analysis of the EPR signal of the superoxide radical anion showed that the components of the g tensor vary with the splitting of the two mutually perpendicular π* superoxide anion orbitals (Figure 4). The properties of the [LCu^I(O₂^•−)]⁺ complexes can be usefully compared to those of the analogous dioxygen adducts of cobalt(II) macrocycles, which have long been recognized as exhibiting an end-on superoxide coordinated to the diamagnetic cobalt(III) ion,⁴³ [Co^III(O₂^•−)]²⁺. In those cases, g_‖ ≈ 2.09 is markedly less than that for A, and the description by eq 1 thus implies a far larger π-orbital splitting, Δ / λ > 15, with the precise value depending on assumptions regarding spin delocalization. In those cobalt complexes, the large π-splitting is induced by strong bonding in the bent Co–O–O linkage (angle of ~120°). The far larger value of g_‖ and thus smaller splitting for A implies that the nonlinear Cu^I–O–O^•− moiety exhibits a substantially weaker interaction between the Cu(I) and the superoxide.

While we are unaware of any examples of previously reported copper(I) superoxo complexes, analogous heme–iron(II)–superoxo species have been generated. Davydov, Hoffman, and co-workers^37a first established that cryoreduction of a monomeric oxy-ferrous hemoglobin Fe^II(O₂) center (Fe^III(O₂^•−)), an oxy-ferrous octaethylpoprhyrin, and an oxy-ferrous cyclen complex all produced a superoxo-ferrous state Fe^II(O₂^•−) state, where the macrocycle-iron(II) is low-spin and thus diamagnetic.^37a Upon annealing/warming to T > 150 K, all of these complexes transform by internal electron transfer (IET) to Fe^III–(hydro)peroxo species, driven at least in part by H-bonding or H⁺-donation to the superoxo moiety.

In synthetic systems, superoxide radical anion binding to heme iron(II) has also been observed. Ohta, Naruta, and co-workers⁴⁴ demonstrated that γ-ray irradiation could be used to cryoreduce a heme–Fe^III-superoxo, giving a porphyrinate–iron(II)–superoxide species. Ivanović-Burmazović and co-workers⁴⁵ showed that an iron(II) porphyrinate with a K⁺ crown ether appendage undergoes reversible O^2•− binding to generate an iron(II) superoxo complex shown to be in equilibrium with an iron(III) peroxo complex. Such interconversions suggest that an Fe^II(O₂^•−) moiety may function as a superoxide H atom abstractor and/or a peroxide anion nucleophilic oxidant.

Thus, on the basis of this iron(II)-superoxo complex literature, it would be of future interest to further probe the formation and reactivity of [LCu^I(O₂^•−)] (A) species, if they could be isolated in proteins or synthetic complexes. What are the chemical and redox properties of the superoxo binding in the parent [LCu^II(O₂^•−)]⁺ complexes and likewise for the initial product of cryoreduction, [LCu^I(O₂^•−)] (A)?

Species B and Peroxo Complexes.

As shown here, cryoannealing A = [(^DMMtmpa)Cu^I(O₂^•−)] complexes in the frozen solid leads to IET, giving the end-on B = [(^DMMtmpa)-Cu^II(O₂²⁻)] Cu^II peroxo complex (Scheme 5), with a dramatically altered EPR spectrum (see Results). While one O atom of the Cu^II–O–O⁻ center is ligated to copper(II), the other oxygen apparently is “naked”, as its properties were not changed by incorporation of TFA in the frozen solution and it does not show the ¹H ENDOR signal from an H-bond to the terminal O atom. Further verification that B is not a hydroperoxo species comes from the contrasting result that C indeed does possess an exchangeable proton, with spectroscopic properties (EPR and ¹H ENDOR) which match those of authentically generated Cu–OOH complex(es) (see Results).

Some time ago, the Karlin group characterized a peroxo dicopper(II) complex, [(XYLO⁻)Cu^II₂(O₂²⁻)]⁺ (XYLO⁻ is a binucleating ligand with phenolato bridging group), whose spectroscopic properties indicated a terminal end-on peroxo, Cu^II–O–O⁻ coordination.⁴⁶ The peroxo group is exceedingly basic; the conjugate acid of this complex has a pK_a value of 24 in MeTHF at −125 °C.⁴⁷

To our knowledge, other copper coordination complexes with a terminal end-on peroxo group, such as those we find for intermediate B, are unknown. However, two LPMO protein derivative X-ray structures are suggested to possess such entities.⁴⁸ In these, the active-site copper(II) exhibits tridentate nitrogen coordination: a histidine imidazole and an imidazole + amino −NH₂ group from the terminal histidine.^8,9,49 In both structures the Cu^II–O–O⁻ group is severely bent (~120°); the O–O moiety is assigned as a peroxo group on the basis of O–O distances calculated to be d_O–O = 1.44 or 1.46 Å. One does not know unambiguously if these peroxo groups might be protonated. Otherwise, side-on binding of a peroxo group to a monocopper(II) center has been demonstrated in both a protein⁵⁰ and a synthetic chemistry environment.⁵¹ In both of these cases overall highly favorable pentacoordination for Cu(II)^18b is achieved by peroxo side-on coordination, whereas for the [LCu^II(O₂²⁻)] studied here, with L as a tetradentate ligand, end-on peroxo binding achieves pentacoordination.

The relative reactivity toward substrates of an end-on bound copper or iron metal peroxo versus a metal hydroperoxo is of considerable interest. The expectation would be that the former possesses a nuclophilic oxidative behavior while the latter acts as an electrophile. While there are particular examples known for nucleophilic non-heme iron Fe–O–O⁻⁵² or Cu–O–O⁻ (vide supra)^46,47 species, there are also documented cases of non-heme iron⁵³ and copper hydroperoxo⁵⁴ complexes with nucleophilic or amphoteric⁵⁵ reactivity. A future comparative study of the reactivity of complexes such as B and C should thus be highly informative.

Examples of synthetic hemes^56,57 with end-on peroxo binding are known. This coordination is also seen as the primary product of cryoreduction of Fe^II(O₂) complexes in heme proteins,^27j,37 as confirmed via X-ray crystal structure determinations. For example, a ferric peroxo myoglobin (Mb) complex was generated by in situ synchrotron radiation of the oxy Fe^II(O₂) species at 100 K.⁵⁸ Another case comes from the cryoradiolytic reduction of a Mb Compound II.⁵⁹ In both, the end-on-bound peroxo group terminal O atom (with formal negative charge) is stabilized by protein H-bonding. Most recently, an end-on heme-peroxo moiety was generated from cryoreduction of an oxy-heme in an engineered protein studied by Lu and co-workers.⁶⁰

Sligar and co-workers⁶¹ recently reported on the chemistry of a steroidogenic P450, CYP17A1, which effects a biosynthetic aldehyde oxidation involving C–C bond cleavage (lyase activity); their spectroscopic interrogation points to a nucleophilic end-on peroxo heme-Fe^III(OO)²⁻ intermediate that adds to the substrate carbonyl group, leading to products. The possibility that P450 end-on bound ferric peroxo anion intermediates effect substrate transformation via nuclophillic attack had been discussed extensively,^2b,62 but it is now clear that this is not usually the case (i.e., in NO synthases or aromatase). Rather, the peroxo intermediates in general simply protonate to give heme-Fe^III(⁻OOH) intermediates that transform to the Compounds I which effect the substrate oxidative (net hydroxylation) reactions.^27j,37b

Species C and Hydroperoxo Complexes.

As has already been described, annealing the end-on peroxo complex [(^DMMtmpa)Cu^II(O₂²⁻)] (B) causes protonation and relaxation to the terminal hydroperoxo complex possessing TBP coordination, C = [(^DMMtmpa)Cu^II(⁻OOH)]⁺. Complex C is an analogue of heme-Fe^III(⁻OOH) hydroperoxo species, which are the reactive states in heme oxygenases and which form in P450s prior to further protonation and O–O heterolytic cleavage. Most recently, there has been an intense interest in the O–O cleavage, both homolytic and heterolytic, of synthetic ligand–Cu^II(⁻OOR) (R ≠ H) compounds, studied as a function of the ligand environment, presence of H-bonding groups, etc.⁶³ However, a deep understanding of O–O cleavage chemistries, which likely occur in synthetic hydroperoxo–copper(II) complexes⁶⁴ or in copper monooxygenases, remains to be achieved.

CONCLUSIONS

We have for the first time performed γ-irradiation cryoreduction of mononuclear copper(I) dioxygen complexes and have characterized reduced and/or protonated intermediates important in O₂-activation chemistry, analogues of species observed in the early intermediates observed for the cytochrome P450 enzyme and Cu monooxygenase (Figure S1) catalytic cycles. Surprisingly, the primary product of reduction of the Cu(II) superoxo species is a set of metastable cuprous superoxo [LCu^I(O₂^•−)]⁺ species. During annealing to higher temperatures this state first undergoes IET to form the Cu(II) peroxo state, which then protonates to form the Cu(II)–OOH species, according to Scheme 5.

Future studies suggested by these findings would include (a) probing for synthetic systems exhibiting a Cu^I superoxo/Cu^II peroxo (end-on) equilibrium, (b) the design of complexes that generate an analogue to complexes B, with an end-on peroxo group bound to Cu(II), (c) introduction of oxidizable organic substrates into studies such as those described here, and (d) cryoreduction of Cu(II) hydroperoxo complexes C, to study O–O reductive cleavage, a key step in all O₂-derived biological and abiological chemistries.