The Stepwise Protonation and Electron-Transfer Reduction of a Primary Copper-Dioxygen Adduct

Abstract

The protonation-reduction of a dioxygen adduct with [LCu^I][B(C₆F₅)₄], cupric superoxo complex [LCu^II(O₂^•−)]⁺ (1), (L=TMG₃tren(1,1,1-tris[2-[N²-(1,1,3,3-tetramethylguanidino)]ethyl]amine)), has been investigated. Trifluoroacetic acid (HOAc_F) reversibly associates with the superoxo ligand in ([LCu^II(O₂^•−)]⁺) in a 1:1 adduct [LCu^II(O₂^•−)(HOAc_F)]⁺ (2), as characterized by UV-visible, resonance Raman (rR), nuclear magnetic resonance (NMR) and X-ray absorption (XAS) spectroscopies, along with density functional theory (DFT) calculations. Chemical studies reveal that for the binding of HOAc_F with 1 to give 2, K_eq = 1.2×10⁵ M⁻¹ (−130 °C) and ΔH° = − 6.9(7) kcal/mol, ΔS° = − 26(4) cal/mol•K). Vibrational (rR) data reveal a significant increase (29 cm⁻¹) in ν_O-O (= 1149 cm⁻¹) compared to that known for [LCu^II(O₂^•−)]⁺ (1). Along with results obtained from XAS and DFT calculations, hydrogen bonding of HOAc_F to a superoxo O-atom in 2 is established. NMR spectroscopy of 2 at −120 °C in 2-methyltetrahydrofuran are also consistent with 1/HOAc_F = 1:1 formulation 2 and that this complex possesses a triplet (S = 1) ground state electronic configuration, as previously determined for 1. The pre-equilibrium acid association to 1 is followed by outer-sphere electron-transfer reduction of 2 by decamethylferrocene (Me₁₀Fc) or octamethylferrocene (Me₈Fc), leading to the products H₂O₂, the corresponding ferrocenium salt and [LCu^II(OAc_F)]⁺. Second-order rate constants for electron transfer (k_et) were determined to be 1365 M⁻¹ s⁻¹ (Me₁₀Fc) and 225 M⁻¹ s⁻¹ (Me₈Fc) at −80 °C. The (bio)chemical relevance of the proton-triggered reduction of the metal-bound dioxygen-derived fragment is discussed.

Introduction

The reduction and protonation of molecular oxygen (dioxygen, O₂) to hydrogen peroxide or water is a critical component of 'dioxygen activation' chemistry effected by transition metal ions, i.e., for use of O₂ as an energy source in fuel cell applications (O₂ + 4e⁻ + 4H⁺ −−> 2H₂O)^{1 ,2} or in oxidative chemistries including oxygen-atom transfer from O₂ to substrates (e.g., as in monooxygenases such as cytochrome P450;³ R-H + O₂ + 2H⁺ + 2e⁻ ––> R-OH + H₂O), dehydrogenation (SH₂ + O₂ ––> S + H₂O₂) or oxidase reactions (e.g., 4AH + O₂ ––> 4A + 2H₂O).^3a,4 As relevant to these processes, considerable insight can be obtained from copper metalloenzyme studies⁵ which perform these biologically essential functions including the use of molecular oxygen for the functionalization/incorporation of oxygen–atom(s) into organic substrates,^5a,5b,6 the production of signaling agents,^5b,6b,7 or regulation of cellular energy flux through the utilization of dioxygen as a terminal electron acceptor.⁸ Nature has evolved to include a diverse repertoire of copper active site configurations containing one to four copper ions with some possessing another redox active site to facilitate dioxygen reduction chemistries in oxidases and oxygenases systems.

For this variety of copper enzymes involved in O₂-processing, partly reduced dioxygen species (e.g., superoxo O₂^•− or peroxo O₂²⁻) form at the protein active-sites during turnover and on the pathway to formation of hydrogen peroxide or water.^7,9 In a number of cases, for proteins and in synthetic inorganic systems, these exhibit diverse electronic, chemical, and physical-structural properties. The first step in dioxygen activation for cases involving O₂ addition to a single copper(I) ion involves formation of a cupric-superoxo (Cu^II-O₂^•−) adduct (A) (Figure 1) which may or may not be detected.¹⁰ This is followed by reduction-protonation by an H-atom donor, a substrate or proton-electron sources to give a cupric-hydroperoxo product (B). Further protonation could liberate hydrogen peroxide, such as occurs in the so-called T2 oxidases, galactose oxidase (GO)¹¹ or copper amine oxidase (CAO),¹² Figure 1.

Figure 1.

Copper/dioxygen intermediates and reactions catalyzed at mononuclear copper enzyme active sites. Shown in the red box are the reduced forms of the 'cofactors' which provide a proton & electron for step (A)–>(B) in T2 copper oxidases. In the orange box are given the source of an electron involved in turnover in the copper monooxygenases DβM/PHM or PMO's. As variously proposed, this electron serves either in (A)–>(B), (B)–>(C) or (C)–>(D).

However, in "noncoupled" binuclear copper monooxygenases, peptidylglycine-α-hydroxylating monooxygenase (PHM) and the homologous dopamine-β-monooxygenase (DβM), further O–O reductive cleavage from the Cu^II–OOH entity (B) occurs, likely via a copper-oxyl ("cupryl") intermediate (C) ultimately yielding hydroxylated substrate (R-OH) and water (Figure 1). For PHM and DβM, the (A)-->(B) step is postulated as that effecting H-atom abstraction from a R-H substrate,^6b,13 but other views exist.¹⁴

In the recently identified copper-dependent polysaccharide monooxygenases (PMOs), a single copper ion found at the enzyme active site apparently facilitates dioxygen binding and hydroxylation of cellulose by an as yet unidentified reaction mechanism.¹⁵ By analogy to the case of PHM, a proposal for a cupric-superoxo mediated C–H activation has been mentioned.^15b

Thus, the primary Cu^I/O₂ intermediate is the gateway complex and fundamental building block along the reductive transformation of other proposed partially reduced Cu/O₂ derived species (vide supra).^5e,16 However, the chemistry of this initial copper/dioxygen adduct, a Cu^IIO₂^•−) species, is considerably less understood when compared to binuclear copper/O₂ (Cu₂O₂) complexes such as dicopper-peroxo or bis–μ-oxo adducts. The inability thus far to detect and carry out in detail reactivity studies for biological Cu/O₂ 1:1 adducts has been part of the problem. However, there is more recently a body of work where the design and synthesis of small molecule 1:1 Cu/O₂ analogues has been achieved and where ongoing efforts are intended to provide fundamental chemical insights into their structure, electronics/bonding and reactivity. In particular, Kitajima/Fujisawa and Solomon,¹⁷ Tolman,¹⁸ Sundemeyer/Schindler,¹⁹ Itoh²⁰ and our own²¹ research groups have significantly contributed (also see Discussion, vide infra) and especially of late, ligand Cu^II–O₂^•− complexes capable of effecting intra-^20a or intermolecular^21d C–H bond oxidations have been described.

Such chemistry is the subject of the current report. Systematic exposure of Cu^II–O₂^•− complexes to exogenous reductant and acid reagents has received little attention even as such reactivity could lead to Cu^II-OOH formation and even further transformation to yield high-valent Cu-O intermediates via subsequent O–O cleavage chemistry.²² This approach may provide specific advantages when compared to C–H and O–H oxidation reactions, including but not limited to the determination of accurate pK_a (or basicity) values and redox potentials of Cu-bound oxygen intermediates. In addition, such studies could lead to mechanistic clarification of stepwise vs concerted (H⁺/e⁻) reductive pathways. Herein, we report the reactivity and physical and chemical properties of an acid associated cupric superoxo complex [(TMG₃tren)Cu^II(O₂^•−)(HOAc_F)]⁺[B(C₆F₅)₄]⁻ (2-B(C₆F₅)₄⁻ = [LCu^II(O₂^•−)(HOAc_F)]⁺ (2)) reacted with electron donors. Our findings are summarized/overviewed in Figure 2. [LCu^II(O₂^•−)]⁺ (1) is reversibly protonated by trifluoroacetic acid (HOAc_F) to give an adduct and only this species can undergo reduction. A cupric hydroperoxo product is implicated but it rapidly releases H₂O₂ producing a copper(II)-OAc_F complex product. The protonation equilibrium between 1 and 2 is established, the structure of the HOAc_F-adduct 2 is elucidated in detail and we also describe in detail on the kinetics of this complex's one-electron reduction.

Figure 2.

Reaction pathway delineated for the protonation-reduction of the cupric superoxo complex [LCu^II(O₂^•−)]⁺ (1) by exogenous ferrocene (Fc) reductants in the presence of trifluoroacetic acid (HOAc_F).

Experimental Section

2.1 General

All materials for ligand synthesis and metal complexes were obtained from commercial sources and used as received unless stated otherwise. Inhibitor free tetrahydrofuran (THF) and 2-methyltetrahydrofuran (MeTHF) were purchased from Sigma-Aldrich and distilled under argon from sodium/benzophenone prior to use. Tetrabutylammonium trifluoroacetate ([NBu₄]⁺[CF₃COO]⁻) and tetrabutylammonium BArF (= B(C₆F₅)₄⁻) ([NBu₄]⁺[B(C₆F₅)₄]⁻) were prepared according to literature procedures.^23,24 Pentane was distilled under argon over CaH₂. Acetonitrile was stored under N₂ and purified via passage through 2 × 60 cm columns of activated alumina. (Innovative Technologies Inc.) The synthesis of air sensitive materials was accomplished under an argon atmosphere using standard Schlenk line techniques and stored in a MBraun glove box with O₂ and H₂O levels at <1 ppm.

The preparation of the ligand TMG₃tren ( = L) and the corresponding copper(I) BArF salt was synthesized as previously reported.^21b d₁₈-TMG₃tren was synthesized by substituting d₆-tetramethyl urea, that was synthesized according to published procedures.²⁵ The¹H-NMR was recorded on a Bruker 300 MHz or a Bruker 400 MHz spectrophotometer.²H-NMR spectra were recorded on a the broad band coil on a Bruker 300 MHz instrument with heteroatom nuclei resonances for²H and¹⁹F occurring at 46 MHz and 282.4 MHz respectively. Unless otherwise specified, NMR spectral peak positions were referenced against TMS or residual solvent signals. Bench-top low temperature UV-visible experiments were carried out on a Cary bio-50 spectrophotometer equipped with an Unisoku USP-203A cryostat using a 1 cm modified Schlenk cuvette. Low temperature stopped-flow experiments were achieved using a Hi-Tech dual sequential mixing stopped flow setup (SF-61DX2 Hi-Tech Scientific) with either a 300–700 or 300–1000 nm range diode array detector (both from Hi-Tech Scientific). Temperature control for the stopped flow measurements was achieved by using an ethanol bath cooled by liquid nitrogen evaporation. Electrochemical experiments were performed using a BAS 100B electrochemical analyzer. EPR measurements were performed on a Bruker EMX X-band EPR spectrometer using 4 or 5 mm quartz EPR tubes. All reactions were performed in MeTHF unless specified otherwise.

2.2 Job’s plot

The total concentration of [LCu^II(O₂^•−)]⁺ (1) + HOAc_F was 3.0 mM at − 130 °C. The sample was then oxygenated by bubbling O₂ for ~ 30 seconds. After this, an appropriate amount of a 33 mM HOAc_F (50–200 µL) (total volume added of LCu(I) + HOAc_F = 250 µL) was added to the chilled solution and mixed by bubbling O₂ through the reaction mixture for a few seconds. The delta-absorbance at 550 nm, for the formation of [LCu^II(O₂^•−)(HOAc_F)]⁺, was recorded and plotted versus mole fraction (X) copper.

2.3 Binding constants and van’t Hoff analysis

Binding constants for the formation of [LCu^II(O₂^•−)(HOAc_F)]⁺, the protonated cupric-superoxide complex, were determined in 5 °C intervals from − 130 °C to − 95 °C by the titration of a stock solution of HOAc_F into a 0.17 mM solution of [LCu^II(O₂^•−)]⁺. The delta-absorbance at 448 nm corresponding to the conversion of [LCu^II(O₂^•−)]⁺ to [LCu^II(O₂^•−)(HOAc_F)]⁺ was equal to − 0.53 AU. The equilibrium constant (K) was determined by fitting the change of absorbance at 448 nm to the following equation: log(θ/(1−θ)) = log[HOAc_F] + log K where θ = ΔAbs₄₄₈(X) / ΔAbs₄₄₈ (Total). Titration data and a table with calculated equilibrium constants at each temperature employed is given in the Supporting Information.

Determination of the standard enthalpy and entropy change for the formation of [LCu^II(O₂^•−)(HOAc_F)]⁺ was accomplished by van’t Hoff analysis of the variable temperature titration data.

2.4 Resonance Raman (rR) measurements

A 7 mM copper complex solution was made by dissolving 83 mg of [LCu^I](B(C₆F₅)₄⁻) in MeTHF to make a total of 10 mL solution. A 500 µl aliquot of this copper(I) solution was added to the 5 mm NMR sample tube, capped with a septum, and chilled in a pentane/N_2(l) bath. Oxygenation of the copper samples was achieved by slowly bubbling an excess of dioxygen through the solution using a Hamilton gas–tight syringe equipped with a 3–way valve and needle outlet. After addition of dioxygen, 50 µL of a 132 mM solution of^XHOAc_F (X = 1 or 2) was added to the top of the sample tube. After 1–2 minutes, the solution was mixed via bubbling an additional 1 mL of dioxygen through the mixture; leading to the formation of a red–colored solution. The final sample concentration consisted of a 6.4 mM copper and 12 mM HOAc_F solution. Dioxygen,¹⁶O₂ (Airgas OX UHP-300) and¹⁸O₂ (Icon 6393), were added to an evacuated Schlenk flask fitted with a septum for the oxygenation reactions described above. Resonance Raman spectra were obtained using a Princeton Instruments ST–135 back–illuminated CCD detector on a Spex 1877 CP triple monochromator with 1200, 1800, and 2400 grooves/mm holographic spectrograph gratings. Laser excitation was provided by a Coherent I90C–K Kr⁺ ion laser and Innova Sabre 25/7 Ar⁺ ion laser with 20 mW of incident power at the sample in a ~ 135° backscattering configuration. Resonance Raman spectra were collected for 10 minutes on a spinning sample of frozen MeTHF solutions in NMR tubes cooled to 77 K in a liquid nitrogen finger Dewar (Wilmad). For the excitation profile, the area of the 1149 cm⁻¹ vibration was determined from a subtracted spectrum (spectrum of 2 prepared with¹⁶O₂ minus the solvent spectrum). The average area was determined from three independently prepared samples, normalized to the area of a solvent vibration, by fitting a Gaussian function to the two vibrations.

2.5 X–Ray absorption spectroscopy

A modified 10 mL Schlenk flask was charged with 5 mL of a 1 mM solution of [LCu^I](B(C₆F₅)₄⁻) and chilled to − 130 °C in a pentane/N_2(l) bath. After sample oxygenation, addition of 100 µL of a 500 mM solution of HOAc_F resulted in the copper solution to turn from green to a red hue. Liquid nitrogen chilled XAS sample holders were submerged into the reaction mixture, loaded, and flash frozen.

2.5.1 XAS Data Acquisition

The Cu K-edge X-ray absorption (XAS) spectra of [LCu^II(O₂^•−)]⁺ (1) in 2-MeTHF and of 2, formed after addition of TFA, to 1, were measured at the Stanford Synchrotron Radiation Lightsource (SSRL). XAS spectra of the frozen solutions in Delrin XAS cells with 37 µm Kapton windows were measured on the unfocused 20-pole, 2.0-T wiggler beam line 7-3. Storage ring parameters were 3 GeV and ~350 mA. A Rh-coated premonochromator mirror was used for harmonic rejection and vertical collimation. A Si(220) double-crystal monochromator was used for energy selection. The samples were maintained at a constant temperature of ~ 10 K during data collection using an Oxford Instruments CF 1208 continuous-flow liquid helium cryostat. A Canberra solid-state Ge 30-element array detector was used to collect Kα fluorescence. Internal energy calibration was performed by simultaneous measurement of the absorption of a Cu foil placed between two ionization chambers located after the sample. The first inflection point of the foil spectrum was assigned to 8980.3 eV. EXAFS data are reported to k = 12.85 Å⁻¹ in order to avoid interference from the Zn K-edge. Photoreduction was observed for both samples, as evidenced by a gradual decrease in the energy of the edge region and with slight changes in the EXAFS upon continuous scanning at the same spot. To minimize effects of photoreduction, data were collected on four physically separate spots on two sample cells for 1 and on three sample cells for 2. The data reported here include an average of 47 scans for 1 and 12 scans for 2.

2.5.2 XAS Data Analysis

The energy-calibrated averaged data were processed by fitting a second-order polynomial to the pre-edge region and subtracting this from the entire spectrum as a background. A three-region polynomial spline of orders 2, 3, and 3 was used to model the smoothly decaying post-edge region. The data were normalized by scaling the spline function to an edge jump of 1.0 at 9000 eV. This background subtraction and normalization was done using PySpline.²⁶ The least-squares fitting program OPT in EXAFSPAK²⁷ was used to fit the data. Initial ab initio theoretical phase and amplitude functions were generated in FEFF 7.0²⁸ using crystallographic parameters of 1 and DFT optimized structures of 1 and 2. During the fitting process, the bond distance (R) and the mean-square thermal and static deviations in R (σ2) were varied for all components. The threshold energy (E₀), the point at which the photoelectron wave vector k = 0 was allowed to vary for each fit but was constrained to the same value for all components in a given fit. Coordination numbers (N) were systematically varied to evaluate chemically viable structures but were fixed within a given fit to the EXAFS data and their Fourier Transform (FT).

The data for 1 show minor contamination from metallic Cu as seen from the FT peaks between 4–5 Å (Figure S12, blue line) and also evidenced during EXAFS fitting that required a peak at ~2.6 Å (Figures S13–14, Table S3, Fit 1B and Fit 1C). This ~2.6 Å peak can be fit with either a Cu-O/N or a Cu-Cu vector. Due to the absence of any light atom coordination around Cu at a distance of ~2.6 Å in the crystal structure of 1, the ~2.6 Å peak in the data is likely from metallic Cu contamination (Figure S14 and Table S3 Fit 1C). The EXAFS and FT reported here, thus, have a small amount (~8%) of metallic Cu signal subtracted followed by renormalization (Figure S12, green line). EXAFS fitting of 1 before and after subtraction of Cu contamination has no significant change to the first-shell parameters (Tables 1 and S3). The data for 2 were collected separately and show no evidence of Cu contamination.

Table 1.

EXAFS Least-Squares Fitting Results for k = 2–12.85 Å⁻¹ for 1 (left) and 2 (right).

Species 1		Fit 1A		Species 2		Fit 2A
Coord No./Path		R (Å)a	σ2 (Å2)b	Coord No./Path		R (Å)a	σ2 (Å2)b
1	Cu-O/N	1.93	399	1	Cu-O/N	1.94	130
4	Cu-N/O	2.10	492	4	Cu-N/O	2.08	335
6	Cu-C	2.87	921	6	Cu-C	2.89	379
12	Cu-N-C	3.24	520	12	Cu-N-Cd	3.22	379
6	Cu-C	3.89	572	3	Cu-C	3.46	1005
6	Cu-N	4.05	843	6	Cu-N	3.90	1014
Ea	(eV)	−8.01		Ea	(eV)	−4.67
Fc		0.16		Fc		0.20

^aThe estimated standard deviations in R for each fit is ±0.02 Å.

^bThe σ² values are multiplied by 10⁵.

^cThe error, F, is given by Σ[(χ_obsd–χ_calcd)²k⁶]/Σ[(χ_obsd)²k⁶].

^dσ² for the multiple scattering path is linked to the corresponding single scattering path.

2.6 NMR spectroscopy (²H and¹⁹F)

²H–NMR experiments were performed at − 120 °C using a 2.0 mM d₁₈–TMG₃tren Cu(I) (B(C₆F₅)₄⁻) stock solution. Preparation of [LCu^II(O₂^•−)]⁺ of and [LCu^II(O₂^•−)(HOAc_F)]⁺ was performed at − 130 °C and frozen in liquid nitrogen prior to insertion into the spectrophotometer. For the formation of [LCu^II(O₂^•−)(HOAc_F)]⁺ the total concentration of HOAc_F (^XHOAc_F (X = 1 or 2)) was 12 mM. Using an ACD/NMR Processor Academic Edition (Advances Chemistry Development, Inc., Toronto, ON Canada, www.acdlabs.com), NMR data were processed with 5 Hz line broadening applied to the FID prior to Fourier transformation. Peak positions are referenced versus the methyl signal of natural abundance deuterium in the MeTHF solvent set to 1.25 ppm.

Fluorine NMR on [LCu^II(O₂^•−)(HOAc_F)]⁺ samples were performed as described above using a solvent mixture of 8:5:87 v/v (d₈–THF:α,α,α-trifluorotoluene:MeTHF) using the non–deuterated ligand supported copper complex at a concentration of 5 mM and 12 mM HOAc_F. The signal corresponding to α,α,α-trifluorotoluene was set to −63.9 ppm.

2.7 EPR spectroscopy

EPR measurements were taken at either 77 K or 20 K with a total copper concentration of either 0.9 mM or 2.3 mM.

2.8 Hydrogen peroxide quantification

The spectrophotometric quantification of hydrogen peroxide was achieved by recording the intensity of the 413 nm oxidized product of diammonium 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonate (AzBTS-(NH₄)₂) by horse radish peroxidase (HRP) which was adapted from published procedures. In a typical experiment 2500 µl of a 338 mM stock solution of [LCu^I](B(C₆F₅)₄) was chilled to − 130 °C and oxygenated for 15 seconds. At this time, two sample solutions of 200 µl containing ~ 5 equiv HOAc_F and 1 equiv Me₁₀Fc each were added and the reaction was allowed to proceed for 1 hour. Excess dioxygen was removed by performing 10 vacuum argon purge cycles at − 130 °C after which time the reaction mixture was warmed to room temperature and subjected to H₂O₂ analysis as described below. The yield of H₂O₂ is 70% in the presence of both exogenous HOAc_F and Me₁₀Fc. In the absence of exogenous reductant (i.e. only addition of HOAc_F) the yield of H₂O₂ is = 30%.

Hydrogen peroxide (H₂O₂) was detected using the procedure described below with the following stock solutions: 300 mM sodium phosphate butter pH 7.0 (solution A), 1 mg/mL AzBTS–(NH₄)₂ (solution B), 4 mg HRP (type II salt free (Sigma)) and 6.5 mg sodium azide in 50 mL water (solution C). Quantification of hydrogen peroxide was achieved by adding a 100 µl of a MeTHF solution containing H₂O₂ to a cuvette containing 1.3 mL water, 500 µl of solution A, 50 µl solution B and C. After mixing for 15 seconds, the samples were allowed to incubate for ~ 2 minutes until full formation of the 413 nm band was achieved.

2.09 Electronic Structure Calculations

DFT calculations were performed using the Gaussian 09 package (revision C.01).²⁹ All calculations used the spin-unrestricted hybrid density functional B3LYP³⁰ and a mixed basis set (6–311G(d) for Cu, O, coordinated N, and the acidic H atom in HOAc_F, and 6–31G(d) for all other atoms) on an ultrafine integration grid. Solvation effects were accounted for using the polarized continuum model (PCM) with THF as the solvent (using default parameters). Analytical frequencies were calculated on all geometry-optimized species to verify that the species were true local minima without negative frequencies. Reported vibrational frequencies are scaled by 0.966.³¹ All energies are reported as Gibbs free energies at 1 atm and 148 K (−125 °C). For time-dependent DFT calculations the first 25 roots were determined, and the calculated electronic absorption spectra were generated using SWizard³² with 3000 cm⁻¹ linewidths and Gaussian bandshapes. Wave-functions were visualized with Lumo³³ and orbital components were analyzed with QMForge.³⁴

2.10 Kinetic measurements on the reduction of 2

Kinetic measurements on the rate of Me₁₀Fc/Me₈Fc oxidation were carried out under pseudo-first-order conditions employing two experimental conditions: Condition 1. Acid dependent measurements: the concentration of [LCu^I](B(C₆F₅)₄) and reductant (Me₁₀Fc/Me₈Fc) were fixed at 0.679 mM and 6.79 mM respectively and were prepared and transferred anaerobically to the stopped-flow instrument. Dioxygen saturated MeTHF was used for the oxygenation premixing reaction and referred to as mix 1. This solution involved in mix 1, were allowed to react at for >10 seconds prior to the 2^nd mixing event in order to achieve full formation of [LCu^II(O₂^•−)]⁺. The HOAc_F acid solutions were made in the presence of air with freshly distilled MeTHF and were discarded every 45 min and used for the second mixing event referred to as mix 2. Condition 2. Reductant dependent measurements: A 0.679 mM stock solution of [LCu^I](B(C₆F₅)₄) was mixed with a sample containing a varied concentration of reductant which was purged with O₂ prior to loading into the instrument. These two solutions were used in the first mixing event (mix 1). The concentration of acid stock solution was fixed at 30 mM (15 mM after dilution) HOAc_F and prepared in the presence of air and discarded after every 45 min. This solution was used in the second mixing event (mix 2) in order to initiate the reaction.

Pseudo-first-order rate constants k_obs were determined by monitoring the decay of absorbance at 448 nm due to [LCu^II(O₂^•−)] (1), which obeyed single exponential decay at times >250 ms, by using the program kinetic studio supplied by Hi–TECH scientific. Normalized 448 nm time dependent absorption profiles were obtained using Origin 8. Values reported for k_obs are the average of 10 or greater shots.

Results and Discussion

Cupric Superoxo Complexes

As mentioned above in the Introduction, the isolation of discrete cupric superoxo (Cu^II–O₂^•−) complexes have been of considerable interest in efforts to provide chemical insight into dioxygen activation processing by copper (bio)chemical systems. In synthetic and biological systems, the lifetimes of Cu^II–O₂^•− adducts are dictated by multiple factors including the synthetic protocol and the ligand environment provided for the copper ion.^16a,16b,35 Another major problem is that follow up reaction of a second copper(I) precursor complex with the superoxo-copper(II) species is often faster than the initial Cu^I/O₂ binding event.³⁶ To date, there are three structurally reported Cu^II– O₂^•− complexes, as shown in Figure 3, with one example isolated in a biological setting (i.e., the enzyme PHM) and two additional structures observed using a synthetic ligand scaffold and carrying out the chemistry under cryogenic conditions.

Figure 3.

Crystallographically characterized Cu^II-superoxo complexes: A. Cu_M site in PHM displaying N_(His)2S_(Met) Cu– coordination (PDB Code:1SDW); B. Crystal structure of side-on cupric superoxo complex with tris-pyrazolylborate ligand displaying N₃ ligand Cu–coordination (CCDC Code: YINBAY). C. Crystal structure of TMG₃tren-Cu^II–superoxo complex [LCu^II(O₂^•−)]⁺ (1) displaying N₄ ligand Cu–coordination (CCDC Code: KEKPAS).

Amzel and coworkers were able to structurally characterize a dioxygen bound form of the PHM enzyme upon dioxygen exposure of reduced enzyme crystals grown in presence of an inhibitory substrate analogue.⁷ Figure 3A is the so-called enzyme Cu_M site from the crystal structure displaying an end-on coordination mode for dioxygen to the copper metal center with ∠Cu–O–O angle of 110° and a O–O bond distance of 1.23 Å; these values are consistent with this cupric superoxide electronic configuration. There are proposals for this intermediate to directly initiate H-atom (H•) abstraction from the substrate or facilitate oxidation of the Cu_H site with the participation of a proton yielding a cupric hydroperoxo complex. We also note that Marletta and coworkers very recently reported the copper-oxygen intermediates observed at the active-site of a copper dependent polysaccharide monooxygenase (PMO; see Figure 1); two different species were observed, one proposed to be a cupric superoxo (Cu^II–O₂^•−) moiety, the other a cupric hydroperoxo Cu^II-(H)OOH species.^15d

The two synthetically derived cupric superoxo complexes shown in Figure 3B and 3C exhibit different electronic and bonding geometries. The first example was reported by Kitajima, Fujisawa, and coworkers using a sterically hindered 3(R),5(R’)-disubstituted tris(pyrazolyl)borate ligand (Tp^R,R'; R = tBu, R’ = iPr = Tp^tBu,iPr), Figure 3B.^17a This Cu^II–O₂^•− complex binds superoxide anion in a side–on fashion and exhibits diamagnetic properties thus with an S = 0 ground state which is derived from the strong bonding interaction between the Cu_xy and the superoxo π*_σ orbitals. Here, the O–O bond length observed is 1.22 Å with an O–O stretching frequency of 1043 cm⁻¹ (Δ¹⁸O₂ = − 59 cm⁻¹) (Note: the latter was determined from a more sterically encumbered ligand Tp^Adm,iPr in order to remove a binuclear, side-on peroxo contaminant.^17b) It should be noted that isoelectronic complexes, best described as side–on bound Cu(III)–peroxides (Cu^III–O₂²⁻), with ν_O–O ~ 970 cm⁻¹ have been observed by Tolman and coworkers upon dioxygen exposure to strongly donating bidentate diketiminate anionic ligand supported copper(I) species.³⁷

Sundermeyer, Schindler and coworkers reported an end–on bound superoxo copper(II) compound structure using the TMG₃tren supporting ligand (Figure 3C), [LCu^II(O₂^•−)]⁺ (1), and this is the starting dioxygen adduct used for the chemical-spectroscopic studies described in this report.^19b Complex 1 shows geometric properties (∠Cu–O–O = 123°; O–O = 1.28 Å) similar to those found in the PHM Cu^II–O₂^•− structure (vide supra). Resonance Raman spectroscopic data reveals an O–O vibration at 1120 cm⁻¹ shifting to a Fermi doublet centered at 1057 cm⁻¹ when 1 is formed using¹⁸O₂.^21c Experimental and computational evidence demonstrate this copper(II) complex with end-on bound superoxo species has a ground state triplet electronic configuration, indicating separate spins on the Cu(II) ion and superoxide anion reside in orthogonal orbitals.^19c,21c [LCu^II(O₂^•−)]⁺ (1) displays slow and limited reactivity with added phenol or catechol substrates and over the course of reaction an internal ligand hydroxylation reaction has been observed.^21b

Formation and Physical Properties of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2)

As described above, the present study was initiated to elucidate aspects of the chemical reactivity of well-defined cupric-superoxo complexes, but for the first time employing proton and electron sources. Initially, we examined the reaction of [LCu^II(O₂^•−)]⁺ (1) with outer-sphere reductants. For this and other studies with copper-dioxygen complex O₂-reduction chemistries, ferrocene derivatives have been extremely useful since a wide variety are known over a range of redox potentials and the reactions are spectroscopically convenient to monitor. However, [LCu^II(O₂^•−)]⁺ (1) is completely unreactive to the strong reductant decamethylferrocene (Me₁₀Fc; E° = 0.02 V vs SCE) and there is also no reaction even with the very powerful cobaltocene reductant (E° = − 0.795 V vs SCE).³⁸ This implies that the reduction potential of 1 must be < − 0.8 V vs SCE. Thus, we turned to protonation chemistry. Initial surveys of the reactivity of [LCu^II(O₂^•−)]⁺ (1) employing acids of differing strengths (pK_a’s), we observed the formation of a new spectroscopic intermediate when utilizing trifluoroacetic acid (TFA = CF₃CO₂H = HOAc_F). As will be shown here, the complex formed is [LCu^II(O₂^•−)(HOAc_F)]⁺ (2), assigned as a (1)/HOAc_F 1:1 complex. This report will focus on the physical characterization of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) and its chemical reactivity towards electron transfer by outer-sphere reductants.

As is already well established for a variety of solvents employed, low temperature dioxygen exposure to a solution of [LCu^I] in MeTHF leads to the formation of the end–on cupric superoxide complex [LCu^II(O₂^•−)]⁺ (1) (Scheme 2, Figure 3C). This bright green species has a characteristic absorption spectrum as shown in Figure 4 (green spectrum) with λ_max = 448 (3400), 676, 795 nm [λ_max (ε_max: M⁻¹cm⁻¹)]. The addition of an excess of trifluoroacetic acid (HOAc_F) leads to the clean conversion to a new pale red metastable intermediate [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) possessing new absorption bands at 330 (5200), 382 (2600), 549 (370), 670 (170), 985 (430) nm (Figure 4 (black)). At − 130 °C, 2 is stable for hours with only minimal decomposition observed as followed by either UV-vis or EPR spectroscopies.³⁹ The addition of an excess of triethylamine (as a base) leads to restoration of the starting 1 complex indicating a reversible association of HOAc_F.³⁹

Figure 4.

Conversion of [LCu^II(O₂^•−)]⁺ (1) (green) to [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (black) upon the addition of HOAc_F at − 130 °C. Insert: Job’s plot for the formation of 2 which indicates a 1:1 ([LCu^II(O₂^•−)]⁺ : HOAc_F) binding stoichiometry for the formation of 2.

The stoichiometry of HOAc_F association with [LCu^II(O₂^•−)]⁺ (1) in the formulation of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) was obtained by Job’s analysis. A maximum absorbance change at 550 nm for the formation of 2 occurs at a molar ratio of 0.5 of 1 and HOAc_F as shown in the insert of Figure 4. This behavior is indicative that 2 consists of a 1:1 complex of 1 and HOAc_F. Spectrophotometric titration of [LCu^II(O₂^•−)]⁺ (1) with HOAc_F lead to equilibrium constants ranging from K = 1.19×10⁵ to 792 M⁻¹ over the temperature range − 130 °C to − 95 °C. Van’t Hoff analysis of the variable temperature equilibrium constants determined for the formation of 2 is shown in the Supporting Information. The results are that for the association of HOAc_F to 1, ΔH° = − 6.9 ± 0.7 kcal/mol and ΔS° = − 26 ± 4 cal/mol•K. The large and negative value of ΔS°, in conjunction with the 1/HOAc_F 1:1 binding stoichiometry indicated by Job’s analysis, suggests liberation of superoxide (O₂^•−) from the cupric center does not occur upon the association of HOAc_F to [LCu^II(O₂^•−)]⁺ (1).

Resonance Raman Spectroscopy of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2)

Laser excitation at 368.2 nm reveals a single dioxygen isotope sensitive vibration at 1149 cm⁻¹, which shifts to 1084 cm⁻¹ (Δ = 65 cm⁻¹) when¹⁸O₂ is employed (Figure 5A). The energy and isotope shift is consistent with the O–O stretch of previously characterized end-on, mononuclear copper(II) superoxide species (Table S1), including 1 (ν_O–O 1120 cm⁻¹; Δ(¹⁸O₂) = 63 cm⁻¹). The shift in ν_O–O to higher energy in complex 2 relative to 1 is inconsistent with an isomerization to a side-on superoxide species since the additional Cu–O bond results in a lower O–O frequency (~1040 cm⁻¹).^17b From the excitation profile, the ν_O–O is resonance enhanced in the low energy shoulder (382 nm, 2600 M⁻¹ cm⁻¹) of the intense 330 nm absorption band and is not observed in any of the other lower energy bands (Figure 5B). The significant distortion of the O–O bond in the excited state along with the energy and intensity of the 382 nm transition indicates that it is a superoxide to copper charge transfer transition. An isotope sensitive vibration corresponding to a ν_Cu–O is not observed in the rR spectra of 2.

Figure 5.

A. Resonance Raman spectra of protio and deutero [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) in MeTHF collected at 77 K with 363.8 nm excitation. B. Absorption spectrum of 2 and the relative intensity and standard deviation of the area of the 1149 cm⁻¹ vibration.

Preparation of 2 with deuterated acid (²HOAc_F) produces a rR spectrum that is indistinguishable from 2 prepared with protio acid (Figure 5A). This suggests that the proton has not been transferred to superoxide ligand of the cupric superoxo moiety; a large deuterium isotope shift on the O–O vibration of the hydroperoxyl radical is observed in the gas phase.⁴⁰

A comparison of the spectra of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) to the parent compound [LCu^II(O₂^•−)]⁺ (1), indicates that the addition of HOAc_F leads to a 3,860 cm⁻¹ increase in the energy of the CT transition (448 to 382 nm) and an increase in the O–O stretching frequency of 29 cm⁻¹.

A similar increase in superoxide to copper CT transition (422 to 410 nm) and the O–O stretch (11 cm⁻¹) was observed between tmpa and PV-tmpa, attributed to an intramolecular hydrogen bond between the 6’-amide substituent on one of the pyridyl ligand arms of PV-tmpa and the distal oxygen of superoxide (see Figure 6).^21d,41 Based on these observation, a hydrogen bonding interaction to the superoxide fragment would be consistent with the increase in ν(O– O) between 1 and 2. The change in these spectral features could also be attributed to differential protonation of the ligand. Both possibilities are evaluated computationally, see below.

Figure 6.

Ligand supported cupric superoxo complexes; A. tmpa; B. PV-tmpa (amide (N–H) interacts with distal O-atom).

Protonated Superoxo Complex EPR and NMR Spectroscopy

Similar to the parent superoxo complex [LCu^II(O₂^•−)]⁺ (1), the HOAc_F adduct [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) is EPR silent (X–band perpendicular mode). Further hour-long incubation of this reaction mixture at − 130 °C leads to only slight decomposition, revealing a very weak non-stoichiometric typical copper(II) complex EPR signal.³⁹ A significant EPR signal corresponding to a full stoichiometric equivalent of the beginning sample of 2 is only observed after the sample is warmed to room temperature (RT).³⁹ The lack of either a Cu(II) or free superoxide EPR signal for [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) indicates it has either a singlet or triplet ground state configuration. Based on NMR spectroscopy (vide infra), 2 remains as possesses a triplet ground state.

We also probed possible structural rearrangements of the ligand environment upon the association of HOAc_F to [LCu^II(O₂^•−)]⁺ (1) using low temperature NMR spectroscopy employing the deuterium enriched supporting d₁₈–TMG₃tren ligand. This approach was applied to the study of 1 in a previous report by Roth and coworkers. In dimethylformamide (DMF) solvent at − 60 °C, they observed four paramagnetically broadened methyl signals between 5 −22 ppm; the notable paramagnetism thus suggested the triplet S = 1 electronic configuration ground state.^19c Here, the unpaired electron on the d⁹ copper(II) ion and that on the superoxide radical anion, reside in orthogonal orbitals, unlike that observed in the side-on bound cupric superoxide complex studied by Kitajima/Fujisawa and Solomon where S = 0 (vide supra). This NMR splitting pattern observed by Roth suggests that in solution the three tetramethylguanidine (TMG) groups (as ligand arms) on the TMG₃tren ligand are symmetrically bound to the copper ion and the associated four methyl groups on each arm of the TMG group are chemically distinct. Our data essentially reproduce that observed by Roth and coworkers; we observe a very similar splitting pattern for the d₁₈–TMG₃tren ligand supported [LCu^II(O₂^•−)]⁺ (1) complex in MeTHF at − 120 °C as shown in Figure 7 (green). The four chemically distinct methyl signals resonate at 28.8, 22.6 (shoulder), 19.8 and 7.8 ppm.

Figure 7.

²H-NMR spectroscopy of [LCu^II(O₂^•−)]⁺ (1) (green), [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (red), and [LCu^II(O₂^•−)(HOAc_F)]⁺ (2-²H) (Black) collected at − 120 °C in MeTHF. L is d₁₈–TMG₃-tren.

For the trifluoroacetic acid protonated complex [LCu^II(O₂^•−)(HOAc_F)]⁺ (2), a new chemically distinctive paramagnetically shifted²H-NMR spectrum is obtained (Figure 7, red). The observed paramagnetism is inconsistent with an isomerization to a side-on superoxide species in 2 since this species would be diamagnetic due to a stronger bonding interaction between the superoxide and the copper ion. This acid adduct possesses downfield shifted methyl signals at 32.5, 22.4, and 9.8 ppm that integrate in a ~ 2:1:1 ratio. When deuterated acid (²HOAc_F) is used in the formation of [LCu^II(O₂^•−)(²HOAc_F)]⁺ (2−²H) the three peaks do not shift and remain in a ~ 2:1:1 integration ratio, however a new broad 2-H signal at ~ 10.5 ppm now overlaps with the sharper 9.8 ppm resonance. We attribute this ~10.5 signal to residual excess²HOAc_F which may be involved in dynamic exchange with 2−²H.⁴² No other deuterium signals are observed for 2−²H at resonances from + 600 to − 35 ppm and from 0 to − 300 ppm. The observed decrease in the number of chemically distinct methyl signals raises the possibility that structural rearrangement of the ligand in 2 results in a gain of symmetry upon HOAc_F association with 1; two of the four methyl groups on the N=C-(NMe₂)₂ moieties of each ligand arm may be chemically equivalent on the time scale of the NMR spectroscopic experiment. An alternative conclusion is that two of the four -CD₃ resonances happen to have coinciding chemical shifts. The very similar paramagnetic shifts displayed by the ligand in the²H-NMR experiments for [LCu^II-(O₂^•−)(HOAc_F)]⁺ (2) and [LCu^II(O₂^•−)]⁺ (1) suggests a common triplet electronic ground state. This is consistent with the lack of an observed X-Band EPR signal for both complexes (vide infra).

The question of proximity of the trifluoroacetic acid molecule with respect to the paramagnetic Cu^II-O₂^•− core was also probed via low temperature ¹⁹F-NMR spectroscopy. The addition of HOAc_F to [LCu^II(O₂^•−)]⁺ (1) at − 120 °C leads to a new thermally sensitive broad and chemically distinct fluorine signal occurring at − 83.0 ppm attributed to the HOAc_F molecule associated with [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (see Figure 8). Integration of the − 83.0 ppm signal in comparison to the para-fluorine of the B(C₆F₅)₄⁻ counter anion as an internal standard indicates this adduct 2 is generated in ~ 80% yield. This stoichiometry is consistent with a 1:1 formulation of 1/HOAc_F of 2 as indicated by Job’s plot analysis and provides evidence against formulations involving multiple equivalents of HOAc_F. Also observed in this spectrum is residual excess HOAc_F resonating at − 78 ppm, which chemical shift appears to be unaffected by the presence of 2. It is notable that the − 83 ppm ¹⁹F signal is broadened compared to the non-associated excess HOAc_F and also compared to the fluorine atom resonances in B(C₆F₅)₄⁻. This behavior further supports the premise that the HOAc_F molecule associated with 2 is in close proximity to the paramagnetic triplet Cu^II-O₂^•− core. Additional experiments of interest include the finding that the addition of [NBu₄]⁺[OAc_F]⁻ to [LCu^II(O₂^•−)]⁺ (1) does not lead to formation of the − 83 ppm signal, clearly demonstrating that this signal is distinct from free ⁻OAc_F anion under these condition.⁴³

Figure 8.

¹⁹F-NMR spectrum of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) collected at − 120 °C. The paramagnetically broadened signal resonating at − 83.0 ppm is attributed to HOAc_F in complex 2. Also observed in this spectrum is residual (non-assocated) HOAc_F at − 78.0 ppm and three fluorine signals associated with the BArF (B(C₆F₅)₄⁻) anion resonating at − 133.8, − 164.0 and − 167.5 ppm. All chemical shifts are referenced vs CF₃C₆H₅(*) set to − 63.9 ppm.

In summary, the low-temperature²H-and¹⁹F-NMR spectroscopic experiments on [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) support the formulation for 2 and that it is a paramagnetic 1:1 adduct of HOAc_F and [LCu^II(O₂^•−)]⁺ (1). The experiments with the deuterium enriched ligand supported complex suggest a small structural rearrangement of the TMG₃tren ligand arms upon association of HOAc_F. Fluorine NMR spectroscopy suggests a close proximity of HOAc_F to the Cu^II-O₂^•− core. While the rRaman spectroscopic data implicate an acid with superoxide ligand interaction, the combined data do not address whether or not the HOAc_F molecule has undergone dissociation via proton transfer to the ligand in 2. Thus, we turned to XAS and DFT calculations to provide further insights (i) into the structural rearrangements occurring upon HOAc_F association with the cupric superoxide complex 1 and (ii) the electronic structural properties of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2).

XAS. Cu XANES

The Cu K-edge XAS spectra for [LCu^II(O₂^•−)]⁺ (1) (green line) and [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (red line) are shown in Figure 9. The second derivative of the data are shown in the inset that reveal a 1s → 3d pre-edge transition at ~ 8979 eV for both samples. The energy of this transition is characteristic of Cu(II). Conversely, Cu(III) species exhibit a pre-edge shift of ~ 1.0 − 2.0 eV to higher energy relative to analogous Cu(II) complexes.⁴⁴ Figure 9 also includes previously studied models, ([Cu^II(HB(3,5-Ph₂pz)₃)]₂(O₂)) and [Cu^III₂(TACN^Bn3)₂(O₂)]²⁺ as references.⁴⁵ The data show that 2 has no observable feature between 8980 and 8981 eV that would be attributed to the presence of Cu(III).

Figure 9.

Cu K-edge XAS spectra of 1 (green line), 2 (red line), and model ([Cu^II(HB(3,5-Ph₂pz)₃)]₂(O₂)) (---/○) and [Cu^III₂(TACN^Bn3)₂(O₂)]²⁺ (…/×) complexes. (Inset) Smoothed second derivative of the pre-edge region with simulated spectra generated with the following ratios of the Cu(II) and Cu(III) models: 80:20, 60:40, 40:60, and 20:80 (…).

Cu EXAFS

The k³-weighted EXAFS data and their FT of [LCu^II(O₂^•−)]⁺ (1) (green line) and [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (red line) are shown in Figure 10. The EXAFS fit parameters of 1 and 2 are given in Table 1 (EXAFS fits are given in Figures S15–16). The first-shell EXAFS data of 1 were best fit with one Cu−O/N at 1.93 Å and four Cu−N/O at 2.10 Å (Fit 1A). These distances well reproduce the crystallographic distances (Cu−O = 1.93 Å and average of the four Cu−N = 2.10 Å) of this complex 1.^19b The major change observed in the data in converting 1 to 2 via HOAc_F association is an increase in the EXAFS and FT intensities. A good fit to the data was again achieved with a split first-shell with one Cu−O/N at 1.94 Å and four Cu−N/O at 2.08 Å (Fit 2A) with smaller σ² values than those observed for 1 or with one Cu−O/N at 1.93 Å and five Cu−N/O at 2.08 Å (Fit 2B, Table S4) with similar σ² values as for 1 (see Table 1). Thus, the increased intensity in 2 is accounted for by either a coordination number of five with smaller σ² values or a coordination of six with larger σ² values. However, a six coordinate Cu(II) complex with six Cu–N/O distances less than 2.1 Å is inconsistent with Cu(II) complexes undergoing a Jahn-Teller distortion, which would result in at least two Cu-N/O distances ≥ 2.1 Å. Therefore, we disfavor this interpretation and conclude that 2 must be five-coordinate.

Figure 10.

Cu K-edge EXAFS data (left) and non-phase-shift-corrected FT (right) of 1 (green line) and 2 (red line). Phase shift in the first shell is ~0.4 Å.

Small changes in the first-shell distances were observed for different spline functions used for data reduction giving an error of ± 0.05 Å for both species. Other first-shell fits with a combination of two to four short and one to three long Cu−O/N bonds were also attempted for both 1 and 2. These fits were not justified due to higher error function, small σ² values (< 0.00050) or because the split in distance between the two first-shell paths were less than the allowed resolution of the data (0.14 Å). The FT peaks in the R = 2.0 − 2.8 Å range for both 1 and 2 were fit using single scattering (SS) and multiple scattering (MS) contributions from the alkyl backbone of the ethylamine. The peaks in the R = 2.8 − 4.0 Å range were fit with SS contributions from the guanidino amino and methyl groups.

All together, the rR, NMR and XANES/EXAFS comparisons of [LCu^II(O₂^•−)]⁺ (1) and [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) indicate that the coordination geometries, i.e., the number of ligands and their arrangement, are essentially unchanged upon addition of HOAc_F to 1. As indicated, large differences in electronic structure have been revealed based the significant shifts in the O₂^•−→Cu charge transfer band energy and O–O stretching frequency (vide supra).

DFT models of 2

Density functional theory calculations were performed on [LCu^II(O₂^•−)]⁺ (1) and several possible structural isomers of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) to identify a plausible site of protonation. From the experimental data, these structural models were restricted to a 1:1 Cu:HOAc_F stoichiometry as indicated by the Job's plot experimental results (see Figure 4; vide supra), and with a triplet (S = 1) ground state. Four of the candidate models of 2 (Figure 11) differed by the site of protonation: NMe₂ (model 2A), a coordinating imine (2B, with ⁻OAc_F occupying the vacant coordination site), the distal oxygen atom of the superoxo (2C), or the tripodal tertiary amine (2E). Finally, we also considered formation of a 1:1 Cu:HOAc_F adduct without proton transfer (2D).⁴⁶ The calculated ν_O–O frequencies, relevant bond lengths, and free energies (relative to 1 + HOAc_F) for these models are given in Table 2, and TD-DFT calculated UV-vis absorption spectra of 1 and the three lowest-energy models of 2 are shown in Figure 12.

Figure 11.

(Left) Isomers of complex 2 analyzed by DFT.

Table 2.

Selected Geometric, Vibrational, and Free Energy Data for DFT-Calculated Species 1 and 2A–2E.

	ΔGrela	νO-O (cm−1) (Δ ν18O) [Δ ν2H]b	rO-O (Å)	rCu-O (Å)	∠Cu-O-O (°)	rCu-Ligand (Å)c
1	0	1172 (−73)	1.298	1.952	120.7	2.154; 2.112/2.142/2.163
2A	+8.2	1186 (−68) [−0.2]	1.307	1.962	118.0	2.099; 2.005/2.007/2.635
2B	−10.4	1177 (−73) [0]	1.292	1.974	120.3	2.270; 1.993/2.230; 2.054
2C	+20.4	1060 (−61) [−11.3]	1.340	2.030	121.0	2.060; 2.056/2.082/2.145
2D	−7.1	1189 (−71) [−0.3]	1.309	1.958	119.6	2.114; 2.109/2.134/2.146
2E	+21.0	1183 (−66) [0]	1.287	1.938	98.3	3.549; 2.032/2.067/2.124

^aFree energy in kcal/mol at 148 K (− 125 °C) relative to 1 + HOAc_F. The outer-sphere ⁻OAc_F anion was included in computational models 2A, 2C, and 2E for the determination of ΔG_rel.

^bCalculated shift in the ¹⁶O–¹⁶O stretching frequency in the presence of ²HOAc_F.

^cThe Cu–N_amine distance is listed first, followed by the Cu–N_imine distances. For 2B, the Cu–O_Ac distance is listed last.

Figure 12.

TD-DFT calculated UV-vis absorption spectra for 1, 2A, 2B, and 2D. For each species, the calculated energy and oscillator strength of the O₂→ Cu LMCT is indicated with a vertical bar.

Structure 2C with a protonated superoxo can be immediately excluded as a candidate model since the calculated wavefunction has significant Cu³⁺ character (inconsistent with Cu²⁺ in the XAS pre-edge) and a much lower ν_O–O frequency than [LCu^II(O₂^•−)]⁺ (1), inconsistent with the + 29 cm⁻¹ increase observed by rRaman spectroscopy on [LCu^II(O₂^•−)(HOAc_F)]⁺ (2). Furthermore, the presence of an O–H bond requires a sizable ¹H/²H isotope effect on ν_O–O (calcd ν_O–O(H) – ν_O–O(D) = − 11.3 cm⁻¹), but no isotope-sensitive features were observed in the reaction of 1 with HOAc_F/²HOAc_F (see Figure 5). Although the protonated imine structure 2B is exergonic with respect to 1 + HOAc_F, it does not reproduce the large change in ν_O–O frequency (calcd + 3 cm⁻¹) and LMCT energy (calcd + 400 cm⁻¹) that is observed for 2 (obsd + 29 and ~+ 4000 cm⁻¹). Therefore, we also rule out structure 2B. Models 2A and 2D that include the acidic H-atom hydrogen-bonded to the superoxo each reproduce the shift in ν_O–O to higher frequency (calcd + 14 and + 17 cm⁻¹, respectively) and LMCT to higher energy (both calcd + 3300 cm⁻¹). Additionally, there is a minimal calculated ¹H/²H isotope effect on ν_O–O (< 0.3 cm⁻¹) for both 2A and 2D, consistent with identical rRaman spectra obtained with ¹HOAc_F and ²HOAc_F. However, model 2A has a pseudo-four coordinate geometry with a very long axial Cu⋯N_imine distance > 2.6 Å, which is inconsistent with the coordination number ≥ 5 that is required to fit the EXAFS data. Therefore, model 2D is the only models that reproduce the combined spectroscopic data, including the 5–coordinate EXAFS fit, the significant shift of both ν_O–O and LMCT to higher energy, and lack of an ¹H/²H isotope effect on ν_{O– O}. Model 2D is also consistent with the ²H and ¹⁹F NMR spectroscopic data for 2. Assuming rapid precession around the Cu-O bond in solution (which also must be present in 1), 2D would be expected to exhibit threefold symmetry in its ²H-NMR spectrum, and the proximity of HOAc_F to the paramagnetic core would explain the broad resonance in the ¹⁹F-NMR spectrum. Finally, model 2D is calculated to be both exothermic and exergonic from 1 and HOAc_F at − 125 °C (ΔH° = − 11.9 kcal/mol; ΔG_−125°C = − 7.1 kcal/mol), similar in magnitude to the experimentally determined values of ΔH° = − 6.9 kcal/mol and ΔG_−125°C = − 3.8 kcal/mol.

Effect of HOAc_F on the ν_O-O Frequency and LMCT Energy

Consistent with the experimental data, structure 2D predicts a shift in ν_O–O to a higher frequency than in [LCu^II(O₂^•−)]⁺ (1). However, the DFT model also predicts a longer O–O bond in 2D than in 1 (1.309 vs. 1.298 Å, respectively), which would be expected to result in a lower energy ν_O–O due to a weaker bond. To resolve this discrepancy, we investigated the effect of the HOAc_F dipole on the geometry and electronic structure of the Cu-O₂^•− fragment. First, using the molecular coordinates of the Cu-O₂^•− fragment in 2D with HOAc_F removed, ν_O–O was calculated to be 1129 cm⁻¹ (i.e., 43 cm⁻¹ lower than 1), as would be expected from its longer O-O bond. Next, we examined a modified model of 2D where the atoms of HOAc_F were replaced by point charges in order to mimic the electrostatic effect of the HOAc_F dipole without introducing covalent ffects of the hydrogen bond.⁴⁷ In the presence of the HOAc_F dipole, ν_O–O shifted to 1180 cm⁻¹ (i.e., 51 cm⁻¹ higher than without the dipole), which is close to that observed for the full model 2D (1189 cm⁻¹). Thus, the shift to a higher frequency ν_O–O in 2 results from an electrostatic interaction between HOAc_F and the copper superoxide.

The increase in ν_O–O frequency is rationalized by the antiparallel orientation of the HOAc_F and O₂ dipoles (Figure 13A); stretching the O–O bond results in an increased O–O dipole which is aligned opposite to the HOAc_F dipole, resulting in a net higher frequency ν_O–O. The weaker O–O bond in 2 results from an increase in the amount of filled π*_σ antibonding character upon binding HOAc_F. In 2, the occupied β-MOs (Figure S18) account for 84% of the total π*_σ character (the remaining 16% is mixed with the Cu d_z² in the β-LUMO), compared to 74% π*_σ in occupied β-MOs in 1 (26% π*_σ in the unoccupied Cu d_z²). Together, these results confirm that the O–O bond is weaker in 2, yet the observed higher ν_O–O frequency is due to the presence of the HOAc_F dipole that causes a shift to higher energy, dominating the opposing shift to lower energy due to the change in geometry.

Figure 13.

(A) Mulliken charges of core atoms in DFT models of 1 (left) and 2D (right). Note that the O–O and O–H dipoles are oriented in opposite directions. (B) Effect of HOAc_F on calculated LMCT energy. The donor and acceptor states for the LMCT have predominantly π*_σ and d_z² character, respectively. The corresponding β-spin orbitals of π*_σ and d_z² character are visualized for 1 (left) and 2D (right). Orbital contours are plotted with an isovalue of 0.05.

The blue shift in the O₂^•− → Cu CT transition in the UV-Vis absorption spectrum of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (see Figure 4 and Supporting Information) can also be explained by the presence of the HOAc_F dipole in proximity to the Cu-O₂ core. TD-DFT calculations on the "dipole only" form of 2D indicate that the key O₂ (π*_σ) → Cu (d_z²) charge transfer increases by ~ 2200 cm⁻¹ when the HOAc_F dipole is included in the model. This is perhaps not surprising, since the HOAc_F dipole will have a larger electrostatic stabilization of the doubly-occupied O₂^•− π*_σ (the LMCT donor MO) than the singly-occupied Cu d_z² (LMCT acceptor orbital) due to the proximity of the positive end of the dipole to the O₂^•− moiety, resulting in an increase of the donor/acceptor orbital splitting and a higher energy LMCT (Figure 13B).

In summary, these results indicate that the trifluoroacetate remains protonated and hydrogen bonds to the distal oxygen of the superoxo fragment in [LCu^II(O₂^•−)(HOAc_F)]⁺ (2). The spectroscopic features displayed by 2 are a result of the electrostatic interaction with the dipole of HOAc_F and not a change in orbital covalency imparted by the hydrogen bond.

Reactivity of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) toward Exogenous Substrates

A facile electron transfer is observed upon the addition of the exogenous electron donors Me₁₀Fc (E_ox = 0.020 V vs SCE) and Me₈Fc (E_ox = 0.120 V vs SCE) to complex 2 yielding one equivalent of oxidized ferrocenium ion, Me₁₀Fc⁺ and Me₈Fc⁺ respectively. This reactivity is in stark contrast to that of the parent [LCu^II(O₂^•−)]⁺ (1), in which no electron transfer was observed with Me₈Fc, Me₁₀Fc or even cobaltocene (vide supra). Reactivity studies performed at − 130 °C on 2 with an excess of exogenous substrates including thioanisole, 2,6-di-t-butyl-4-methoxyphenol, and 9,10-dihydro-10-methylacridine lead to little to no change in the optical spectrum. A lack of exogenous C–H, O–H, and oxo transfer reactions between 2 and substrates may have to do with its inherent reaction properties, or it could be that the associated HOAc_F molecule limits access and approach of the substrate to the Cu^II-O₂^•− core.

In the electron-transfer reactions of 2 with Me₁₀Fc and Me₈Fc, the overall chemistry is highlighted in Figure 2. While it requires excess HOAc_F to fully form [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (vide supra), the overall stoichiometry requires the addition of 2 equiv HOAc_F and one equiv of Me₈Fc/Me₁₀Fc to the [LCu^II(O₂^•−)]⁺ (1) complex to yield one equiv hydrogen peroxide (H₂O₂), one equiv oxidized ferrocene (Me₈Fc⁺/Me₁₀Fc⁺) and a [LCu^II(OAc_F)]⁺ product complex (Figure 2). Experiments where the amount of reductant was varied indicate that only one equiv is required.³⁹ When only one equiv acid is employed in the presence of excess reductant, the reaction only proceeds half-way.³⁹ We can conclude that when [LCu^II(OOH)]⁺ forms it undergoes protonation more efficiently than does occur ferrocenyl reduction of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) (see Figure 2). Thus, one-half the added one equiv acid produces [LCu^II(OOH)]⁺ while the other half protonates the latter to give H₂O₂ and [LCu^II(OAc_F)]⁺. This cuprictrifluoroacetate complex was identified by EPR spectroscopy, in comparison to an 'authentic' sample generated by addition of a stoichiometric amount of AgOAc_F to [LCu^II(Cl)]⁺.³⁹ Quantitation of the H₂O₂ formed (when a small excess of acid was employed) led to a 70% yield (see Experimental Section).

Kinetics of the ferrocenyl reduction of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2)

As described and demonstrated below, 2 is reduced by ferrocenyl derivatives in an overall second-order process, with a first-order dependence on both 2 and Me₈Fc (or Me₁₀Fc). The formation of complex 2 under the conditions employed (− 80 °C) is related to complex 1 via a rapid HOAc_F equilibrium-binding event. The disappearance of 1 (and thus 2) at − 80 °C in the presence of excess reductant (Me₈Fc and Me₁₀Fc) obeys a first-order decay model.³⁹ Thus, the overall kinetics are described by a rapid preequilibrium (K_HOAcF) followed by the irreversible and ratelimiting ferrocenyl electron transfer reduction event (k_et), eqs 1 & 2. Eq. 3 describes the overall kinetic process.

$$\mathbf{1}+\mathrm{H}\mathrm{O}\mathrm{A}{\mathrm{c}}_{\mathrm{F}}\rightleftharpoons \mathbf{2}(K_{\mathrm{H}\mathrm{O}\mathrm{A}\mathrm{c}\mathrm{F}})$$ (1)

$$\mathbf{2}+\mathrm{F}\mathrm{c}\to \text{Products}(k_{\mathrm{e}\mathrm{t}})\mathrm{r}.\mathrm{d}.\mathrm{s}.$$ (2)

$$\mathrm{d}[\text{Products}]/\mathrm{d}t=k_{\mathrm{e}\mathrm{t}}[\mathbf{2}][\mathrm{F}\mathrm{c}]=k_{\text{obs}}[\mathbf{1}]$$ (3)

$$k_{\text{obs}}=(k_{\mathrm{e}\mathrm{t}}{K}_{\mathrm{H}\mathrm{O}\mathrm{A}\mathrm{c}\mathrm{F}}[\mathrm{H}\mathrm{O}\mathrm{A}{\mathrm{c}}_{\mathrm{F}}][\mathrm{F}\mathrm{c}])/(1+K_{\mathrm{H}\mathrm{O}\mathrm{A}\mathrm{c}\mathrm{F}}[\mathrm{H}\mathrm{O}\mathrm{A}{\mathrm{c}}_{\mathrm{F}}]))$$ (4)

The determination of the kinetics of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) promoted ferrocenyl electron transfer was achieved by examining the loss of [LCu^II(O₂^•−)]⁺ (1) under two separate experimental conditions. The concentrations of HOAc_F and ferrocene reductant were systematically and independently varied. Each reaction condition displays different kinetic behavior with respect to the concentration of varied substrate. As stated above, 2 is rapidly formed at − 80 °C, being observed in the first spectrum (collected at t = 1.5 ms) after the introduction of acid to 1 with the initial amount of 1 and 2 observed being dependent on the concentration of HOAc_F employed.³⁹ Thus, the rate constant for association of HOAc_F to 1 (k₁) is very much larger than k_et and k_et is associated with the rate-determining step. However, as the adduct 2 is responsible for ferrocenyl oxidation and is formed in a non-linear fashion with respect to increasing HOAc_F concentration (K_HOAcF = 165 ± 90 M⁻¹ at − 80 °C), k_obs will display similar behavior with respect to HOAc_F until a maximum rate (k_obs) is observed, i.e., saturation behavior (Figure 14A).

Figure 14.

A. First order decomposition rate constants for the disappearance of [LCu^II(O₂^•−)]⁺ (1) observed at 448 nm with 1.69 mM (10 equiv) reductant Me₁₀Fc (red circles) and Me₈Fc (gold circles) at − 80 °C as a function of increasing concentrations of HOAc_F. Black and green lines are the linear fits for the Me₁₀Fc and Me₈Fc respective k_obs values according to Eq. 4. Extrapolation for the second order rate constant (k_et) of 1365 ± 75 M⁻¹s⁻¹ and 225 ± 8 M⁻¹s⁻¹ for the oxidation of Me₁₀Fc and Me₈Fc respectively by [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) are obtained. B. First-order rate constants for the disappearance of [LCu^II(O₂^•−)]⁺ (1) followed at 448 nm at a concentration of 15 mM HOAc_F as a function of increasing concentration of the reductant Me₈Fc (10 – 20 equiv vs 1). A slope and thus a second order rate constant of 158 ± 3 M⁻¹ s⁻¹ is obtained. See text.

The kinetic behavior just described applies to both Me₈Fc and Me₁₀Fc as reductants. The largest reaction rate observed in the two cases are not equal, displaying a maximum value of k_obs as 0.38 and 2.36 s⁻¹ respectively for Me₈Fc and Me₁₀Fc when k_obs is modeled using Eq. 4 and is shown in Figure 14A. These results indicate that the rate-determining step contains a redox component as the stronger reductant Me10Fc can more rapidly reduce [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) when compared to Me₈Fc. These saturation values of k_obs lead to an estimation of the second order electron transfer rate constant (k_et) of 225 ± 8 M⁻¹s⁻¹ and 1365 ± 75 M⁻¹ s⁻¹ for the oxidation of Me₈Fc and Me₁₀Fc respectively by 2. Equation 4 also allows for an estimate of the equilibrium constant for K_HOAcF of 154 ± 15 (from the Me₈Fc data) and 103 ± 13 M⁻¹ (from the Me₁₀Fc data) for the redox reactions which are in reasonable agreement with the value of 165 ± 90 M⁻¹ obtained from van’t Hoff analysis at this temperature (− 80 °C), see above. Along these lines, the much weaker reductant dimethylferrocene (Me₂Fc; E° = 0.435 vs SCE) does not lead to any observable oxidized ferrocenium products thus bracketing the reduction potential of 2 between 0.120 and 0.435 V vs SCE (not accounting for solvent or temperature effects). Note that this is very different from the reduction potential of 1 (<− 0.8 V vs SCE), which highlights the substantial effect of hydrogen-bonded HOAc_F on reduction potential, even without proton transfer.

Further, this affect is consistent with DFT calculations, which predict that association of HOAc_F leads to spontaneous transfer of the proton upon one-electron reduction, which increases the electron affinity by ~1.1 eV (see Supporting Information). Thus, the hydrogen-bonded HOAc_F provides an extra 25 kcal/mol driving force for substrate (ferrocenyl reductant) oxidation by 2 relative to 1 through a PCET process.

We were also able to investigate the effects of varying reductant concentrations on k_obs at a fixed concentration of HOAc_F (i.e. 15 mM). With this large amount of acid, the predominant copper species present in solution is [LCu^II(O₂^•−)(HOAc_F)]⁺ (2). Thus, the kinetic situation will more accurately represent that predicted by eqs 3 and 4. Figure 14B displays results of such experiments, demonstrating a linear correlation with respect to reductant concentration revealing a second order rate constant of 158 ± 3 M⁻¹ s⁻¹, comparing to the value of 225 ± 8 M⁻¹s⁻¹, see above. This behavior is consistent for a reaction mechanism involving rate determining electron transfer with a transition state comprised of a 1:1 stoichiometry of [LCu^II(O₂^•−)(HOAc_F)]⁺ and Fc.

Further Remarks and Conclusions

The kinetic investigation into the oxidation of ferrocenes by [LCu^II(O₂^•−)]⁺ (1) in the presence of HOAc_F supports a two–step reaction mechanism; where HOAc_F associates to 1 to form [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) and this complex is responsible for facilitating the electron transfer/oxidation of the substituted ferrocene. A question arises as to the identity of the electron accepting orbital and if this resides on either the Cu(II) metal ion, the super-oxide O₂^•− moiety, or a [Cu-O₂]⁺ molecular orbital. Our calculations indicate that the electron accepting orbital (LUMO of 2, see Figure S18) is the superoxide π*_v orbital. The associated proton on the HOAc_F moiety in 2 is poised for rapid proton transfer that is triggered upon one electron reduction.

It is important to mention the hydrogen bonding interactions in [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) are proposed to be to the distal oxygen atom of the cupric superoxo moiety. The interaction with this particular O–atom may be crucial in accessing the electron transfer reactivity. In related cupric hydroperoxide complexes (i.e., the one e⁻/H⁺ reduced form of [LCu^II(O₂^•−)]⁺) also possessing tripodal tetradentate chelates, Masuda and coworkers elegantly demonstrated that the positioning of ligand pendant hydrogen bonding amide N–H moieties in relation to the dioxygen fragment dictated the stability and physical properties of the [(ligand) Cu^II(OOH)]²⁺ complex (Figure 15A). For H-bonding interaction with the proximal “O”–atom, a stabilizing interaction is observed and in one case, the cupric hydroperoxide complex could be crystallographically characterized.⁴⁸ However, when H-bonding occurred to the distal “O”–atom an increased O–O bond stretching frequency was observed and the complex had markedly lower stability, decomposing ca. three times faster than for the case where H-bonding was absent.⁴⁹ In iron (bio)chemistry, interactions with the distal oxygen atom of ferric hydroperoxo complexes (Fe^III– OOH) can facilitate O–O bond cleavage, such as in cytochrome P450 monooxygenase^3b (Figure 15B) and non-heme⁵⁰ synthetic systems undergoing conversion to Fe^IV=O products. A further related system is where distal O-atom proton transfer occurs during the dioxygen binding event in the non-heme diiron dioxygentransport protein hemerythrin, Fe^II–OH–Fe^II → Fe^III–OOH‥(O)– Fe^III (Figure 15C).⁵¹

Figure 15.

Interactions with the proximal vs distal O–atom in metal hydroperoxide coordination complexes. See text for further discussion and explanation.

As mentioned, the reduction/protonation of [LCu^II(O₂^•−)]⁺ (1) to give the Cu^II-OOH product (Figure 2) can be effected by Me₈Fc, where the Me₈Fc/Me₈Fc⁺ couple has a reduction equal to + 0.120 V vs SCE (0.370 V vs NHE), but the even weaker reductant Me₂Fc (E° = + 0.435 vs SCE)⁵² does not effect reduction of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2), allowing us to bracket the reduction potential at + 0.120 < E_1/2 < + 0.435 V vs SCE. The present meaning of this reaction's standard reduction potential (E°) is unclear because of (i) the experimental conditions employed (MeTHF solvent at low temperature), and (ii) because the process is in some manner coupled to a proton transfer event; the chemistry certainly does not comprise a pure outer-sphere electron transfer event. We can never-the-less conclude that the reaction, [LCu^II(O₂^•−)(HOAc_F)]⁺ + e⁻ ⇌ [Cu^II–OOH]⁺, is not a highly "uphill" process and the reaction of cupric-superoxo intermediates by reductants such as the coordinated cysteine crosslinked tyrosine ArOH group (in GO) or the aminosemiquinone radical in CAO's (see Introduction) may well be expected to occur. Our work with the present cupric-superoxo complex 1 indicates that a H-bond interaction or full proton donor is also required.

An interesting corollary to these observations is that the reverse reaction, oxidation of a [Cu^II–OOH]⁺ moiety may also be possible within biological milieu. Such a situation could help explain an atmospheric dioxygen exchange pathway observed by Blackburn and coworkers in the enzyme PHM, see Figure 16.^13b The PHM active site was shown to separately display catalase activity, implying a peroxide to O₂ step can occur. This peroxide plus oxidizedenzyme chemistry in PHM diverges from the situation observed in heme and non-heme iron enzymes (e.g., cyt. P450 monooxygenase, soluble methane monooxygenase and naphthalene 1,2-dioxygenases) which undergo "peroxide shunt" biochemistry.^3b,53 There, no peroxide oxidation (to O₂) occurs, and only formation of O–O cleaved higher valent oxidizing intermediates form from metal(s)/H₂O₂ interactions.

Figure 16.

Proposed pathway for O₂ scrambling and incorporation of unlabeled oxygen incorporated in substrate hydroxylation observed by Blackburn and coworkers^13b when the fully oxidized form of PHM is exposed to labeled hydrogen peroxide.

In summary, we have described the synthesis and characterization of a mononuclear copper/O₂ adduct possessing an acidic moiety (as the form of a neutral HOAc_F molecule) which is in close proximity and directly interacting with the cupric superoxo (Cu^II– O₂^•−) fragment. The hydrogen bonding interactions in this complex result in little structural rearrangement and no change in spin state of the complex. The copper remains in a pseudo-trigonal bipyramidal geometry with the copper superoxo moiety in the triplet electronic ground state coordinated to the copper ion in end–on fashion. However, the H–bonding interaction to the superoxo moiety plays a vital role in altering the chemistry of the dioxygen fragment. As described above, the association of HOAc_F to the superoxo moiety in 2 leads to a longer and weaker O–O bond, but a polarization effect induced by the adjacent full trifluoroacetic acid molecule causes an observed increase of 29 cm⁻¹ in the O–O stretching frequency.

This pre-equilibrium acid association to [LCu^II(O₂^•−)]⁺ (1) leads to novel chemical reactivity, the oxidation of outer–sphere electron transfer reduction of [LCu^II(O₂^•−)(HOAc_F)]⁺ (2) to yield hydrogen peroxide (H₂O₂). This work begins to provide previously unavailable information and insights into the magnitude and nature of reduction potentials and relative pK_a’s involved in the interconversion of molecular oxygen and peroxide as it relates to (i) biological transformations facilitated by copper oxidases/monooxygenases and (ii) practical copper ion mediated oxidative processes possibly including dioxygen fuel cell technologies. To probe for detailed and deeper insights into the PCET superoxide reduction/protonation chemistry described here, further investigations needed are kinetics (i) employing reductants with a wider range of redox potentials, and (ii) with variations of the acid-proton source. Notably, the electron and proton which both end up on the copper(II) coordinated O₂^•− species derive from different sources, the former from an outer-sphere ferrocenyl reducing agent and the latter from the associated H-bonded HOAc_F molecule.