Structure, Spectroscopy, and Reactivity of a Mononuclear Copper Hydroxide Complex in Three Molecular Oxidation States

Abstract

Structural, spectroscopic, and reactivity studies are presented for an electron transfer series of copper hydroxide complexes supported by a tridentate redox-active ligand. Single crystal X-ray crystallography shows that the mononuclear [CuOH]¹⁺ core is stabilized via intramolecular H-bonds between the H-donors of the ligand and the hydroxide anion when the ligand is in its trianionic form. This complex undergoes two reversible oxidation processes that produce two metastable “high-valent” CuOH species, which can be generated by addition of stoichiometric amounts of 1e⁻ oxidants. These CuOH species are characterized by an array of spectroscopic techniques including UV–vis absorption, electron paramagnetic resonance (EPR), and X-ray absorption spectroscopies (XAS), which together indicate that all redox couples are ligand-localized. The reactivity of the complexes in their higher oxidation states toward substrates with modest O─H bond dissociation energies (e.g., 4-substitued-2,6-di-tert-butylphenols) indicates that these complexes act as 2H⁺/2e⁻ oxidants, differing from the 1H⁺/1e⁻ reactivity of well-studied [CuOH]²⁺ systems.

Graphical Abstract

INTRODUCTION

High-valent copper complexes have been invoked as active intermediates in the hydroxylation of C─H bonds performed by Cu-dependent metalloenzymes.^1,2 For example, lytic polysaccharide monooxygenases (LPMOs) catalyze the oxidative degradation of cellulose using a mononuclear Cu center bound by two histidines, one acting as a bidentate ligand (Figure 1A).^3,4 It has been proposed that the active species responsible for the oxidation of the cellulose substrate is a [CuO]¹⁺ intermediate, which could be in equilibrium with a [CuOH]²⁺ analogue.⁵ In most of the members of this family of enzymes, a tyrosine residue is found in the vicinity of the active center. Recent reports have found that some LPMOs react with H₂O₂ (an alternative oxidant^6,7) to generate a metastable Cu-tyrosyl species,⁸ which could be the active 2e⁻ oxidant (Figure 1B).⁹ This species is similar to the active 2H⁺/2e⁻ oxidant found in galactose-oxidase in which the dehydrogenation of alcohols to aldehydes is observed.¹⁰

Figure 1.

First step of cellulose degradation catalyzed by Cu-dependent LPMOs (A) and putative 2e⁻ oxidants capable of performing C─H hydroxylation (B).

Inspired by our recent work on Cu complexes bearing bidentate redox-active ligands,^11,12 we describe the synthesis and characterization of a mononuclear copper hydroxide complex bound by a tridentate redox-active ligand scaffold that stabilizes the CuOH core via intramolecular H-bonding interactions (see Figure 2). This [LCuOH]²⁻ complex can be chemically oxidized using ferrocenium (Fc⁺) to generate two “high-valent” CuOH species, [LCuOH]¹⁻ and [LCuOH], that are characterized by multiple spectroscopies. The [LCuOH] species (generated by adding 2 equiv of Fc⁺ to the [LCuOH]²⁻ complex) is capable of performing two sequential H-atom abstractions from phenols. This reactivity differs from the well-established 1H⁺/1e⁻ oxidations promoted by mononuclear [CuOH]²⁺ cores.^13-15

Figure 2.

Synthesis of the mononuclear complex [^tBuCuOH]²⁻ and displacement ellipsoid plot (50% probability level) at 110(2) K (see SI for details).

RESULTS AND DISCUSSION

Synthesis and Characterization of the Copper Hydroxide Complex.

Inspired by the work of Heyduk¹⁶ and Borovik,¹⁷ we designed ^tBuLH₃, which contains a redox-active tridentate amine and two ureanyl H-bond donors (see Figure 2). ^tBuLH₃ was synthesized by addition of 2 equiv of tert-butyl isocyanate (^tBuNCO) to a THF solution of bis(2-aminophenyl)amine (see further details in the Supporting Information, SI).

In the glovebox, the addition of 3 equiv of KH to a solution of ^tBuLH₃ led to the formation of the deprotonated scaffold (^tBuL³⁻)K₃, which, upon treatment with 1 equiv of Cu(OAc)₂ and 2 equiv of NMe₄(OH)·5H₂O, formed the mononuclear copper hydroxide complex [Cu(^tBuL3−)(OH)](NMe₄)₂, [^tBuCuOH]²⁻ (see SI).

Complex [^tBuCuOH]²⁻ was characterized by single crystal X-ray diffraction analysis (Figure 2). The Cu ion exhibits a distorted square-planar geometry around the metal ion (τ₄ = 0.2. Note: for D_4h structures, a τ₄ value of 0.0 is expected¹⁸). Moreover, there is an intramolecular interaction between the hydroxide ligand and the ureanyl H-bond donors of the ligand scaffold (O…Hα’ distances ca. 1.9 Å). The average Cu─O distance is ca. 1.92 Å, which is comparable to the Cu─O distances found for the few mononuclear [CuOH]¹⁺ complexes reported to date (ca. 1.90 Å).^14,19-22

Electrochemistry.

Metal complexes bearing ligands derived from the bis(2-aminophenyl)amine scaffold are able to reach multiple oxidation states by oxidation/reduction of the ligand scaffold. For example, Heyduk reported that Ta^V complexes bound by the reduced form of a similar ligand scaffold generated Ta^V-semiquinone and Ta^V-quinone species via ligand oxidation.¹⁶ Similarly, we envisioned that [^tBuCuOH]²⁻ ([^tBuCuOH]²⁻ = ^tBuCuOH^O.S.2) could be reduced to generate ^tBuCuOH^O.S.1 or oxidized to form “high-valent” ^tBuCuOH^O.S.3, ^tBuCuOH^O.S.4, and ^tBuCuOH^O.S.5 via ligand and/or metal oxidation (see Figure 3A in which ^tBuCuOH^O.S.3 and ^tBuCuOH^O.S.4 are depicted as [CuOH]⁺ complexes bound by the ligand scaffold oxidized to the semiquinone and quinone form, respectively). Cyclic voltammetry experiments revealed that ^tBuCuOH^O.S.2 was reversibly oxidized to ^tBuCuOH^O.S.3 and ^tBuCuOH^O.S.4 at −1.05 and −0.28 V, respectively (V vs Fc^0/+; see Figure 3B).

Figure 3.

(A) Possible oxidation states reached via reduction/oxidation of ^tBuCuOH^O.S.2. (B) Electrochemical and spectroscopic characterization of ^tBuCuOH in three oxidation states (i.e., ^tBuCuOH^O.S.2, ^tBuCuOH^O.S.3, and ^tBuCuOH^O.S.4).

The oxidation of ^tBuCuOH^O.S.2 was also accomplished using stoichiometric amounts of ferrocenium hexafluorophosphate (FcPF₆) (Figure 3B). Addition of 1 equiv of Fc⁺ to ^tBuCuOH^O.S.2 in DMF at −40 °C (880 nm with ε = 400 M⁻¹ cm⁻¹) led to the formation of a new species with intense UV–vis absorption features at 625 nm (ε = 3600 M⁻¹ cm⁻¹) and 950 nm (ε = 4000 M⁻¹ cm⁻¹), which we assigned as ^tBuCuOH^O.S.3 Addition of 1 equiv of Fc⁺ to ^tBuCuOH^O.S.3 produced strong UV–vis absorptions centered at 442 nm (ε = 6200 M⁻¹ cm⁻¹) and 965 nm (ε = 6200 M⁻¹ cm⁻¹) which we assign as ^tBuCuOH^O.S.4. Both isosbestic transformations required only 1 equiv of Fc⁺ (see titrations in Figure S6 and Figure S8 in the SI). The UV–vis absorption features of ^tBuCuOH^O.S.3 are similar to those exhibited by metal complexes bound to the semiquinone form of the tridentate ligand, suggesting that ^tBuCuOH^O.S.3 is a [CuOH]¹⁺ complex bound by the radical ligand ^tBuL^.2− (see Figure 3A).¹⁶ The UV–vis absorption spectrum of ^tBuCuOH^O.S.4 strongly resembles the spectra of some metal complexes containing the quinone form of similar triamine ligands, which implies that ^tBuCuOH^O.S.4 is a [CuOH]¹⁺ core bound by the doubly oxidized form of the ligand (^tBuL⁻).^16,23

After the formation of ^tBuCuOH^O.S.4 via addition of 2 equiv of Fc⁺ to ^tBuCuOH^O.S.2, 1 equiv of cobaltocene (CoCp₂) was added to generate ^tBuCuOH^O.S.3, which was further reduced to ^tBuCuOH^O.S.2 by adding a second equivalent of CoCp₂ (see Figure S10 in SI). The reversibility of these redox processes agrees with the cyclic voltammetry experiments described above in which the ^tBuCuOH^O.S.2 complex was oxidized to two “high-valent” species at low potentials (lower than Fc) and in which these species are sequentially reduced to ^tBuCuOH^O.S.2 at higher potentials than CoCp₂ (E_1/2(CoCp₂) = −1.29 V vs Fc^0/+ in DMF).

The ^tBuCuOH platform in the different oxidation states was also characterized by perpendicular-mode continuous wave X-band EPR (Figure 3B). We observed an axial S = ½ EPR signal for ^tBuCuOH^O.S.2 characteristic of square planar copper(II) complexes (g_⊥ = 2.05, g_∥ = 2.20, A_∥ = 542 MHz).²⁴ Upon addition of 1 equiv of Fc⁺, the EPR features of ^tBuCuOH^O.S.2 disappeared to produce an EPR silent species, which could be due to the formation of an S = 0 [CuOH]²⁺ core or the formation of a Cu-semiquinone-like species in which the unpaired electron of the [CuOH]⁺ center is ferromagnetically coupled to the semiquinone ligand (S = 1) or antiferromagnetically coupled (S = 0).²⁵ An addition of 1 equiv of Fc⁺ to the silent species ^tBuCuOH^O.S.3 produced an EPR signal similar to ^tBuCuOH^O.S.2 but distinct (g_⊥ = 2.05, g_∥ = 2.24, A_∥ = 508 MHz), which suggests that ^tBuCuOH^O.S.4 is a [CuOH]⁺ complex bound by the quinone form of the ligand. Magnetic susceptibilities were measured at room temperature using the Evans method for ^tBuCuOH^O.S.2 (μ_eff = 1.50 μ_B), ^tBuCuOH^O.S.3 (μ_eff = 0.00 μ_B), and ^tBuCuOH^O.S.4 (μ_eff = 1.81 μ_B).²⁶ This magnetic data supports the formulation of ^tBuCuOH^O.S.2 and ^tBuCuOH^O.S.4 as S = ½ complexes and indicates that the spin state of ^tBuCuOH^O.S.3 is S = 0.

X-ray Absorption Spectroscopy.

Cu K-edge X-ray absorption spectra (XAS) were obtained for ^tBuCuOH^O.S.2, ^tBuCuOH^O.S.3, and ^tBuCuOH^O.S.4 (Figure 4). Each spectrum exhibits a weak pre-edge absorption feature assigned by convention to Cu 1s → Cu 3d excitations. The maxima of these 1s → 3d bands occur at 8979.5 eV (^tBuCuOH^O.S.2), 8979.4 eV (^tBuCuOH^O.S.3), and 8979.2 eV (^tBuCuOH^O.S.4). While these energies lie within the range conventionally assigned to Cu^II,^27,28 recent studies have challenged the notion that pre-edge band energies correspond to distinct Cu physical oxidation states.^29-33 These effectively isoenergetic bands suggest that each complex bears Cu in a common physical oxidation state, with slight variations plausibly attributed to changes in ligand field strength attending ligand redox changes. The assumption of very small changes in ligand field strength about Cu accord with the very similar spin-Hamiltonian parameters extracted from fitting the EPR spectra of ^tBuCuOH^O.S.2 and ^tBuCuOH^O.S.4. Transitions to higher energy could arise due to excitations to ligand-localized molecular orbitals as well as excitations to Cu 4p. Rising edge inflection points occur at 8991.2 eV (^tBuCuOH^O.S.2), 8991.4 eV (^tBuCuOH^O.S.3), and 8992.0 eV (^tBuCuOH^O.S.4), further indicating minimal participation of Cu in the redox events linking these species.

Figure 4.

Cu K-edge XAS data obtained for ^tBuCuOH^O.S.2–4. Pre-edge features ca. 8979 eV are magnified for clarity.

H-atom Abstraction Reactivity of the CuOH Complexes.

The reactivity of the three members of the electron transfer series toward organic substrates containing weak C─H bonds [dihydroanthracene, xanthene, and 1,4-cyclohexadiene (BDE’s = 75–80 kcal/mol)] was investigated.³⁴ Disappointingly, none of the CuOH species under investigation was able to oxidize these substrates. However, the CuOH complexes reacted with phenols featuring relatively weak O─H bonds.

The 4-substituted 2,6-di-tert-butylphenols (4-X-2,6-DTBP) are typically used as H-atom donors since they generate metastable phenoxyl radicals (see Figure 5B) that can be observed and quantified by UV–vis or EPR (like in the oxidation of 4-methoxy-2,6-di-tert-butylphenol).³⁵ The products derived from coupling of the phenoxyl radicals can be also be analyzed and quantified (e.g., 2,6-di-tert-butylphenol oxidation).^36,37 The reactivity of the CuOH cores was studied in DMF at room temperature and the results are summarized in Figure 5A. Overall, ^tBuCuOH^O.S.4 abstracted H atoms from 4-substituted 2,6-di-tert-butylphenols to produce ^tBuCuOH-(H⁺)^O.S.3 (protonated form of ^tBuCuOH^O.S.3), which also reacted with phenols and was further reduced to ^tBuCuOH-(H⁺)₂^O.S.2. Thus, ^tBuCuOH^O.S.4 can be considered a 2H⁺/2e⁻ oxidant. However, our experimental evidence suggests that the reduced complex ^tBuCuOH(H⁺)₂^O.S.2 is unstable and undergoes a disproportionation reaction to generate the cuprous complex ^tBuCuOH(H⁺)₃^O.S.1 (that ultimately decomposes to release Cu^I and ^tBuLH₃; see SI) and ^tBuCuOH(H⁺)^O.S.3, which can oxidize the phenolic substrates. Hence, the stoichiometry of the reaction dictates that 1 equiv of ^tBuCuOH^O.S.4 can lead to three H-atom transfer events (details given below).

Figure 5.

Summary of the reactivity of CuOH cores with 4-substituted-2,6-di-tert-butylphenols (A) and organic products derived from the H-atom abstraction (B).³⁸

Reactivity with 4-Substituted 2,6-di-tert-Butylphenols: Stoichiometry.

The reaction between ^tBuCuOH^O.S.4 and 50 equiv of 2,6-DTBP was carried out at 298 K and monitored by UV–vis absorption spectroscopy (see Figure 6A). The reaction proceeded in a two-step fashion. In the first step, UV–vis absorption features of ^tBuCuOH^O.S.4 decayed rapidly with concomitant formation of a UV–vis absorption band centered near 625 nm. This resembles the spectrum of ^tBuCuOH^O.S.3, which was assigned as the Cu product derived from the 1H⁺/1e⁻ reduction/protonation of ^tBuCuOH^O.S.4, ^tBuCuOH(H⁺) (green spectrum in Figure 6). ^tBuCuOH-(H⁺)^O.S.3 could be independently generated by adding 1 equiv of pyridinium triflate (pyr·H⁺) to a solution of ^tBuCuOH^O.S.3(^tBuCuOH^O.S.3+ H⁺ → ^tBuCuOH(H⁺)^O.S.3; see also Figure 12).³⁸ In the second step of the reaction, ^tBuCuOH-(H⁺)^O.S.3 reacted with the phenol to produce the diprotonated ^tBuCuOH(H⁺)₂^O.S.2 complex, which is unstable and decays in the presence of phenol to Cu^I products (see orange spectra in Figure 6A). Addition of 2 equiv of pyr·H⁺ to ^tBuCuOH^O.S.2 triggered the formation of 0.5 equiv of ^tBuCuOH(H⁺)^O.S.3 suggesting that ^tBuCuOH^O.S.2 undergoes disproportionation upon protonation (see Figures S78-S80). Both reaction steps exhibited the growth of very absorptive features near 425 nm, which correspond to the formation of 3,3′,5,5′-tetra-tert-butyldiphenoquinone (DPQ, λ_max 425 nm, ε = 7.2 × 10⁴ M⁻¹cm⁻¹ in DMF), the C─C coupling product derived from the 2H⁺/2e⁻ oxidation of 2,6-DTBP.³⁷ This product is formed via H-atom abstraction of 2,6-DTBP to produce a fleeting phenoxyl radical (2,6-DTBP•), which undergoes dimerization to the diphenol H₂DPQ (Figure 5B).³⁹ The 2H⁺/2e⁻ oxidation of the diphenol product leads to the final diphenoquinone DPQ. Under the employed reaction conditions, a UV–vis absorption band appeared corresponding to DPQ, but its high molar absorptivity led to spectral saturation (see Figure 6A).

Figure 6.

Analysis of the stoichiometry of the reaction between ^tBuCuOH^O.S.4 and 2,6-DTBP.

Figure 12.

Bordwell analysis of the reduction/protonation of the ^tBuCuOH cores (A) and analysis of the protonation ^tBuCuOH^O.S.3 (B) Note: the BDE, pK_a and E_1/2 values in italics were estimated using the Bordwell equation and were not obtained experimentally (see text below).

In independent experiments under the same conditions (room temperature, [Cu] = 0.25 mM, 50 equiv of 2,6-DTBP), DPQ evolution over time was quantified by UV–vis absorption via dilution of the crude reaction mixture. Meanwhile, H₂DPQ was quantified by ¹H NMR (Figure 6C and Figure 6D; see SI for details). After the short reaction times, H₂DPQ was the main product in solution (0.06 mM) while DPQ was formed in only small quantities (<0.02 mM). Over time, the concentration of H₂DPQ decreased slightly (~ 0.04 mM) but DPQ increased substantially (up to 0.11 mM), suggesting that 2,6-DTBP is first oxidized to H₂DPQ (via 2,6-DTBP• dimerization), DPQ is formed via subsequent oxidation of H₂DPQ. Overall, 1 equiv of ^tBuCuOH^O.S.4 (0.25 mM) produced 0.2 equiv of H₂DPQ (0.05 mM) and 0.5 equiv of DPQ (0.11 mM), indicating that ^tBuCuOH^O.S.4 is capable of performing three H-atom transfer reactions (yields ~ 70%). The data suggest that the 2H⁺/2e⁻ reduction product ^tBuCuOH(H⁺)₂^O.S.2 undergoes a rapid disproportionation to generate a Cu^I species and ^tBuCuOH(H⁺)^O.S.3, which can react again with the phenol to produce H₂DPQ and DPQ. In support of this hypothesis, the reaction between ^tBuCuOH^O.S.2, pyr·H⁺ and 2,6-DTBP results in the formation of H₂DPQ and DPQ with yields up to 50% (see Figure S25 in SI).

The reaction between ^tBuCuOH^O.S.4 and 50 equiv of 2,6-DTBP was monitored by EPR spectroscopy (Figure 6E). Addition of the substrate caused the disappearance of the ^tBuCuOH^O.S.4 features, suggesting the formation of the EPR silent ^tBuCuOH(H⁺)^O.S.4 (see also Figure 12). After full decay of ^tBuCuOH(H⁺)^O.S.3 (t ~ 10000 s), no S = ½ signal corresponding to the fully reduced ^tBuCuOH(H⁺)₂^O.S.2 complex emerged, which is in agreement with the proposed disproportionation of ^tBuCuOH(H⁺)₂^O.S.2 to EPR silent species.

Interestingly, ^tBuCuOH^O.S.3 did not react with 2,6-DTBP. Addition of 1 equiv of pyr·H⁺ to ^tBuCuOH^O.S.3 generated ^tBuCuOH(H⁺)^O.S.3, which oxidized 2,6-DTBP (Figure 7). The UV–vis absorption spectra obtained for this reaction (^tBuCuOH(H⁺)^O.S.3 + 2,6-DTBP) were very similar to those observed in the second step of the reaction between ^tBuCuOH^O.S.4 and 2,6-DTBP, suggesting that ^tBuCuOH(H⁺)^O.S.3 is an intermediate in the reaction between ^tBuCuOH^O.S.4 and 2,6-DTBP (Figure 7A). The ^tBuCuOH^O.S.2 complex also reacted with 2,6-DTBP but we found that this transformation entailed the protonation of ^tBuCuOH^O.S.2 without oxidation of the phenol substrate (see Figure S24 and Figure S26 in the SI). We hypothesize that this is due to the high basicity of ^tBuCuOH^O.S.2 (see further discussion below).

Figure 7.

Analysis of the stoichiometry of the reaction between ^tBuCuOH(H⁺)^O.S.3 and 2,6-DTBP.

The organic products formed in the reaction between ^tBuCuOH(H⁺)^O.S.3 and 2,6-DTBP were quantified (see SI) and the yields were consistent with ^tBuCuOH(H⁺)^O.S.3 undergoing two 1H⁺/1e⁻ events (via reduction of ^tBuCuOH(H⁺)^O.S.3 to ^tBuCuOH(H⁺)₂^O.S.2 followed by disproportionation and subsequent phenol oxidation).

The reaction between ^tBuCuOH(H⁺)^O.S.3 and 2,6-DTBP was also monitored using EPR spectroscopy (Figure 7E). Addition of one equiv of pyr·H⁺ to ^tBuCuOH^O.S.3 produced an EPR silent species, which reacted with the phenol over time to generate EPR silent solutions (Figure 7E). The fact that ^tBuCuOH(H⁺)^O.S.3 is capable of oxidizing the phenol (as seen by UV–vis absorption) and that it also led to EPR silent mixtures suggests that in both cases, the generation of ^tBuCuOH(H⁺)₂^O.S.2 is followed by fast disproportionation of this putative EPR active complex. Further corroboration was obtained by analyzing EPR samples of the reaction between ^tBuCuOH^O.S.2, pyr·H⁺ and 2,6-DTBP, which were also silent (see Figure S26 in SI).

The reaction of ^tBuCuOH^O.S.4 with 4-methoxy-2,6-di-tert-butylphenol (4-MeO-2,6-DTBP) was analyzed (Figure 8). Similar UV–vis absorption changes were found to those observed during the reaction between ^tBuCuOH^O.S.4 and 2,6-DTBP, in which the ^tBuCuOH^O.S.4 was sequentially reduced to ^tBuCuOH(H⁺)^O.S.3 and the products derived from ^tBuCuOH(H⁺)₂^O.S.2 disproportionation (Figure 8A). UV–vis absorption spectra revealed the formation of phenoxyl radical derived from H-atom abstraction (4-MeO-2,6-DTBP·, λ_max = 408 nm, ε = 3.0 × 10³ M⁻¹cm⁻¹ in DMF). Similarly, the reaction between ^tBuCuOH(H⁺)^O.S.3 and 4-MeO-2,6-DTBP produced phenoxyl radical, but in lower amounts (the intensity of the UV–vis features for 4-MeO-2,6-DTBP• produced in this reaction were ~50% lower than in the reaction between ^tBuCuOH^O.S.4 and 4-MeO-2,6-DTBP; see Figure 9A). In both cases, we analyzed the organic products derived from the oxidation of with 4-MeO-2,6-DTBP and we observed that 2,6-di-tert-butyl-1,4-benzoquinone (BQ in Figure 5) was produced in excellent yields.

Figure 8.

Analysis of the stoichiometry of reaction between ^tBuCuOH^O.S.4 and 4-MeO-2,6-DTBP.

Figure 9.

Analysis of the stoichiometry of reaction between ^tBuCuOH(H⁺)^O.S.3 and 4-MeO-2,6-DTBP.

The equivalents of phenoxyl radical formed in the reaction between ^tBuCuOH^O.S.4(or ^tBuCuOH(H⁺)^O.S.3) and 4-MeO-2,6-DTBP were quantified using EPR (see SI for details). As expected, the oxidation of 4-MeO-2,6-DTBP by ^tBuCuOH^O.S.4 produced higher quantities of the phenoxyl radical than the reaction between ^tBuCuOH(H⁺)^O.S.3 and 4-MeO-2,6-DTBP (0.25 mM vs 0.11 mM; see Figure 8D and Figure 9D). In both reactions, reduction of ^tBuCuOH(H⁺)^O.S.3 produced the phenoxyl radical but the S = ½ signal from the putative ^tBuCuOH(H⁺)₂^O.S.2 was not observed (Figure 8E and Figure 9E). As with the reactions with 2,6-DTBP described above, it appears that the putative ^tBuCuOH(H⁺)₂^O.S.2 species undergoes disproportionation to generate EPR-silent complexes. Full accumulation of the phenoxyl radical in these reactions was precluded due to its instability at room temperature (solutions of pure 4-MeO-2,6-DTBP radical in DMF decay at room temperature to produce BQ see the SI). The reaction between ^tBuCuOH^O.S.4 and 4-MeO-2,6-DTBP was carried out at −40 °C and was followed by UV–vis absorption and EPR (see Figures S42-S44 in the SI) to give the ^tBuCuOH(H⁺)^O.S.3 species, which did not react further with the phenol. This enabled quantification of the phenoxyl radical produced in the 1H⁺/1e⁻ reduction of ^tBuCuOH^O.S.4 (yields up to 85%).

Reactivity with 4-Substituted 2,6-di-tert-Butylphenols: Kinetics.

The kinetics of the reaction between 4-substituted 2,6-di-tert-butyl phenols (X = MeO, Me, ^tBu, H, NO₂) and ^tBuCuOH^O.S.4 were analyzed (Figure 10). Under pseudo-first order conditions ([^tBuCuOH^O.S.4] ≪ [phenol]), the reaction rates (k_obs) increased linearly upon increasing the concentration of phenol, from which the second-order rate constants were calculated (Figure 10; see SI for details on the exponential fitting of the decay of ^tBuCuOH^O.S.4). ^tBuCuOH^O.S.4 reacted faster with phenols containing weaker O─H bonds (e.g., k(4-MeO-2,6-DTBP) = 4.6 M⁻¹ s⁻¹ vs k(4-NO₂-2,6-DTBP) = 0.013 M⁻¹ s⁻¹).

Figure 10.

Kinetic analysis of the reaction between ^tBuCuOH^O.S.4 and 4-X-2,6-DTBP.

A Marcus plot for the reactivity of ^tBuCuOH^O.S.4 against 4-substituted 2,6-di-tert-butyl phenols was obtained (Figure 11B) by plotting (RT/F) ln k vs E⁰_(PhOH/PhO.) where k is the second-order rate constant and E⁰_(PhOH/PhOH.) is the redox potential of 4-X-2,6-DTBP.^36,37 These plots are commonly utilized to analyze the reaction pathways by which phenols are oxidized (Figure 11A).^36,40-42 Slopes with values ca. −0.5 are indicative of single electron transfer (ET) during the rate-determining step.⁴¹ Meanwhile, slopes near −1.0 are found in reactions where proton transfer (PT) is rate limiting. Values close to 0.0 are indicative of hydrogen atom transfer or concerted proton electron transfer (HAT or CPET). The slope observed for ^tBuCuOH^O.S.4 (−0.12) is consistent with a concerted H-atom transfer and is similar to that found in the oxidation of 4-X-2,6-DTBP by a “pure” hydrogen atom abstractor such as cumylperoxyl radical (slope = −0.05).³⁶ Similar slope values have been reported for the H-atom transfer from phenols to formally manganese(IV)-hydroxo,⁴² copper(II)-superoxo,³⁶ nickel(III),⁴³ and copper(III) complexes.³⁷

Figure 11.

Mechanistic analysis of the reaction between ^tBuCuOH^O.S.4 and 4-X-2,6-DTBP to generate ^tBuCuOH(H⁺)^O.S.3 and phenoxyl radical.

A Bell-Evans-Polanyi plot was obtained by plotting ΔG^‡ vs BDE_O─H, where ΔG^‡ is the Gibbs energy of activation calculated from the k values, and BDE_O─H is strength of the O─H bonds for the phenols.^44,45 This data suggests that the reaction between ^tBuCuOH^O.S.4 and 4-substituted 2,6-di-tert-butyl phenols occurs via hydrogen atom transfer (slope = 0.48; note: the theoretical value for ideal HAT is 0.5;^34,46 see Figure 11C). The negative slope found in the Hammett plot is also in agreement with the proposed H-atom abstraction mechanism (Figure 11D).⁴⁷ The primary kinetic isotope effect measured (KIE ca. 2.1 for 2,6-DTBP vs 2,6-DTBP-OD; see Figure 11E) is consistent with O─H cleavage during the rate-determining step of the reaction. The modest KIE value is not unusual and has also been observed in the H-atom abstraction reactivity of phenols with metal-peroxo,⁴⁸ metal-oxo,^40,49 and metal-hydroxo systems.⁴²

Bordwell Analysis – BDE of the Reduction/Protonation of ^tBuCuOH^O.S.4.

The Bordwell equation⁵⁰ can be used to calculate the driving force associated with the protonation/reduction of oxidants (Figure 12A). Popularized by Mayer and co-workers, this analysis permits the determination of the thermodynamic parameters associated with the reactivity of oxidants; oxidants with higher BDE values are capable of oxidizing stronger C─H and O─H bonds, and usually with faster rates.^46,51 The BDE associated with the protonation/reduction of ^tBuCuOH^O.S.4 to ^tBuCuOH(H⁺)^O.S.3 was calculated using the E_1/2 for the reduction of ^tBuCuOH^O.S.4 (see Figure 4 above) and pK_a of the protonation of ^tBuCuOH^O.S.3 to ^tBuCuOH(H⁺)^O.S.3 (pK_a = 10.9), obtained by spectrometric titration via deprotonation of ^tBuCuOH(H⁺)^O.S.3 with triethylamine (Figure 12; see also SI). The BDE value for the reductive protonation of ^tBuCuOH^O.S.4 (83 kcal/mol in DMF) is very similar to the BDE values of the phenols that can be oxidized by ^tBuCuOH^O.S.4 (BDE_0─H = 78–85 kcal/mol in DMSO).^52,53

Bordwell Analysis – BDE of the Reduction/Protonation of ^tBuCuOH^O.S.3.

As mentioned, ^tBuCuOH^O.S.3 was not able to oxidize 4-susbtituted-2,6-DTBP substrates (BDE_O─H = 78–85 kcal/mol in DMSO). ^tBuCuOH^O.S.3 reacted with 1,2-diphenylhydrazine PhNHNHPh (BDE_N─H = 71.7 kcal/mol in DMSO³⁴) to generate the corresponding PhN═NPh product (see SI). On the basis of this reactivity and using the Bordwell thermodynamic cycle, we can estimate that the pK_a for the protonation of ^tBuCuOH^O.S.2 is between 16 and 20 (see SI for details). In fact, ^tBuCuOH^O.S.2 underwent protonation and subsequent disproportionation with the addition of pyr·H⁺ (pK_a = 3.4 in DMSO⁵⁴), but also upon addition of 4-NO₂-2,6-DTBP (pKa = 7.3 in DMSO⁵⁵), 2,6-DTBP (pKa = 17.3 in DMSO⁵⁵), or 4-MeO-2,6-DTBP (pK_a = 18.2 in DMSO⁵⁵), suggesting that the pK_a for the protonation of ^tBuCuOH^O.S.2 is at least 19 in DMF (see SI).⁵² Likewise, the pK_a obtained for the protonation of ^tBuCuOH^O.S.3 (pK_a ~ 11) was also in agreement with the fact that ^tBuCuOH^O.S.3 was not protonated by 4-MeO-2,6-DTBP or 2,6-DTBP (pK_a > 17) but it reacted with 4-NO₂-2,6-DTBP (pK_a = 7.3) to form ^tBuCuOH(H⁺)^O.S.3 (see SI). On the basis of the reactivity of ^tBuCuOH^O.S.3 (capable of oxidizing PhNHNHPh but not 2,6-DTBP) and the pK_a of its protonation, we estimate the E⁰ for the reduction of ^tBuCuOH(H⁺)^O.S.3 to be between −0.5 and −0.8 V (see calculations in the SI). To validate this estimation experimentally, several 1e⁻ reductants were added to ^tBuCuOH(H⁺)^O.S.3 and the reactions were monitored by UV–vis spectroscopy. ^tBuCuOH(H⁺)^O.S.3 was reduced by CoCp₂ (E_1/2 = −1.29 V vs Fc^0/+ in DMF) and by Me₁₀Fc (E_1/2 = −0.50 V vs Fc^0/+ in DMF) but not by Me₈Fc (E_1/2 = −0.41 V vs Fc^0/+ in DMF) or Me₂Fc (E_1/2 = −0.07 V vs Fc^0/+ in DMF).

Bordwell Analysis – BDE of the Reduction/Protonation of ^tBuCuOH(H⁺)^O.S.3.

In the reaction between ^tBuCuOH^O.S.4 and the phenols, we observed the accumulation of ^tBuCuOH(H⁺)^O.S.3 which also oxidized phenols but with slower rates.⁵⁶ These findings suggest that the thermodynamic driving force (BDE) of the reaction between ^tBuCuOH(H⁺)^O.S.3 and H-atom donors should be similar, but probably slightly lower to the one calculated for ^tBuCuOH^O.S.4 + H• (i.e., lower reaction rates, lower BDE).^{46 tBu}CuOH(H⁺)^O.S.3 could also be reversibly protonated to generate ^tBuCuOH(H⁺)₂^O.S.3 (Figure 12B). Titration of ^tBuCuOH(H⁺)₂^O.S.3 with proton sponge (1,8-bis(dimethylamino)naphthalene) allowed us to determine the pK_a for the protonation of ^tBuCuOH(H⁺)^O.S.3 (pK_a = 8.9).⁵⁷ On the basis of the reactivity of ^tBuCuOH(H⁺)^O.S.3 (capable of oxidizing 2,6-DTBP and 4-MeO-2,6-DTBP), the BDE for its protonation/reduction could be estimated at 78–83 kcal mol⁻¹. This approximation was used to calculate the E_1/2 for the reduction of ^tBuCuOH(H⁺)₂^O.S.3, which we estimate to be between −0.1 and −0.4 V. To validate this calculation experimentally, several 1e⁻ reductants were added to ^tBuCuOH(H⁺)₂^O.S.3 and the reactions were monitored by UV–vis absorption spectroscopy, which revealed that reductants with E⁰ more negative than −0.4 V (CoCp₂, Me₁₀Fc and Me₈Fc) could reduce ^tBuCuOH(H⁺)₂^O.S.3 but Me₂Fc could not (E_1/2 = −0.07 V vs Fc^0/+ in DMF).

Bordwell Analysis – Comparison of the BDEs.

The enhanced oxidative power of ^tBuCuOH^O.S.4 compared to ^tBuCuOH^O.S.3 can be rationalized in terms of the higher reduction potential of ^tBuCuOH^O.S.4 (ΔE_1/2(^tBuCuOH^O.S.4/^tBuCuOH^O.S.3) = 0.77 V; 18 kcal/mol), whereas the change in the pK_a for the reduced products does not contribute as significantly to the overall BDE (ΔpK_a (^tBuCuOH^O.S.2/ ^tBuCuOH^O.S.3 = 5–9; 9–12 kcal/mol). Protonation of ^tBuCuOH^O.S.3 to generate ^tBuCuOH(H⁺)^O.S.3 led to small changes in the basicity of the complex (ΔE_1/2(^tBuCuOH^O.S.3/ ^tBuCuOH(H⁺)^O.S.3) = 2; 3 kcal/mol) but a substantial increase of the reduction potential of the protonated counterparts (ΔE_1/2(^tBuCuOH(H⁺)₂^O.S.3/^tBuCuOH(H⁺)^O.S.3) ~ 0.4 V; 9 kcal/mol), which enables ^tBuCuOH(H⁺)^O.S.4 to oxidize phenolic substrates like ^tBuCuOH^O.S.4. Overall, the versatility of the ^tBuCuOH platform, which is able to reversibly reach three physical oxidation states and three different protonation states, is fundamental to achieving the unprecedented 2H⁺/2e⁻ oxidation reactivity described in this work.

Comparison of ^tBuCuOH^O.S.4 with Other “High-Valent” Copper Complexes.

Previous reports extensively studied the reactivity of mononuclear [CuOH]²⁺ complexes toward C─H and O─H bonds and it was found that these species act as strong 1H⁺/1e⁻ oxidants (BDE of 90 kcal/mol), abstracting hydrogen atoms from the substrates via an HAT mechanism (Figure 13A).^13,58 Recent studies have also described that analogous “high-valent” [CuX]²⁺ complexes also behave as a 1H⁺/1e⁻ oxidants.^37,59 Zhang and co-workers have recently reported that a [CuF]²⁺ species carried out the fluorination of C─H bonds via H-atom abstraction (performed by one equiv of [CuF]²⁺) followed by radical rebound (performed by a second equiv of [CuF]²⁺), suggesting that this “high-valent” complex acted as 1e⁻ oxidant.⁵ This differs from the 2e⁻ reactivity of ^tBuCuOH^O.S.4, capable of performing two consecutive H-atom abstractions.

Figure 13.

2e⁻ oxidations performed by the putative Cu-oxyl intermediate of LPMOs compared with the 1e⁻ reactivity of [CuOH]²⁺ complexes (A).^13,58 2e⁻ reactivity found for the [CuOH]¹⁺ cores reported in this article (B).

We hypothesize that the superior oxidative power of an LCu^IIIOH complex⁵⁸ (depicted in Figure 13A) compared to ^tBuCuOH^O.S.4 (90 kcal/mol in THF vs 83 kcal/mol in DMF) results from the stronger basicity character of the reduced complex LCu^IIOH (pK_a in THF similar to the strong base DBU) and from the slightly higher redox potential of LCu^IIIOH (see Figure 13A). The inclusion of the H-bonding donors on the ligand scaffold, which we believe allow for stabilizing the CuOH cores in different molecular oxidation states, could also be preventing a richer oxidative reactivity toward external substrates. Karlin and co-workers have recently reported that fine-tuning of these H-bonding interactions can stabilize mononuclear Cu/O₂ entities but also enhance their oxidative reactivity.^35,60

CONCLUSIONS

Recent studies have proposed that the oxidation of C─H bonds in LPMO occurs via formation of a Cu-oxyl complex that can perform the 2e⁻ oxidation of strong C─H bonds via H-atom abstraction followed by OH rebound (Figure 13).^5,7 In this report, we describe that the [CuOH]¹⁺ cores bearing a redox-active ligand react with phenolic O─H bonds via two consecutive 1H⁺/1e⁻ events. This differs from the reactivity of galactose oxidase model systems in which Cu complexes bound by redox-active ligands (L•Cu^II) typically perform the oxidation of alcohols (RCH₂OH) to aldehydes (RCHO) via alkoxide coordination to the Cu center (L•Cu^IIOCH₂R) followed by an intramolecular 1H⁺/2e⁻ oxidation (L•Cu^IIOCH₂R → LHCu¹ + RCHO).^25,61 Interestingly, the H• abstraction reactivity of the CuOH complexes reported is not triggered by the oxidation of the CuOH core, which maintains its physical oxidation state, but rather by oxidation and/or protonation of the ligand scaffold, a strategy that might also be utilized by some Cu-dependent monooxygenase enzymes.^5,62 Ongoing studies are focused on modifying our copper complexes by tuning the electronics of the redox-active backbone, altering the stereoelectronics of the H-bond donors, and utilizing various anions, which we believe will boost the oxidative power of these reactive M─O(H) cores and allow for monooxygenase-like 2e⁻ functionalization of C─H substrates (H• abstraction followed by •OH rebound).