Dioxygen Activation by a Macrocyclic Copper Complex Leads to a Cu₂O₂ Core with Unexpected Structure and Reactivity

Abstract

We report the Cu^I/O₂ chemistry of complexes derived from the macrocylic ligands 14-TMC (1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane) and 12-TMC (1,4,7,10-tetramethyl-1,4,7,10-tetraazacyclododecane). While [(14-TMC)Cu^I]⁺ is unreactive towards dioxygen, the smaller analog [(12-TMC)Cu^I(CH₃CN)]⁺ reacts with O₂ to give a side-on bound peroxo-dicopper(II) species (^SP), confirmed by spectroscopic and computational methods. Intriguingly, 12-TMC as a N4 donor ligand generates ^SP species, thus in contrast with the previous observation that such species are generated by N2 and N3 ligands. In addition, the reactivity of this macrocyclic side-on peroxodicopper(II) differs from typical ^SP species, because it reacts only with acid to release H₂O₂, in contrast with the classic reactivity of Cu₂O₂ cores. Kinetics and computations are consistent with a protonation mechanism whereby the TMC acts as a hemilabile ligand and shuttles H⁺ to an isomerized peroxo core.

Over the last 3 decades, our research interests have included a focus on the study of transient species derived from the reaction between Cu^I complexes with dioxygen.^[1] Rational design of the ligand employed has allowed our group, among others,^[2] to extensively characterize the formation of mononuclear^[3] and dinuclear^[4] copper–dioxygen assemblies (Figure 1A). Many of our findings are based on the tripodal ligand TMPA (tris((2-pyridyl)methyl)amine) and its derivatives, that have allowed fine tuning of the electronic and steric properties of Cu/O₂ derived species.^[5] Particularly relevant are our recent findings in the stabilization of mononuclear end-on LCu^II(O₂^•−) species (^ES) making use of electron-rich TMPA derivatives^[6] or by H-bonding interactions,^[7] and our recent reports on the first examples of end-on peroxo-dicopper(II) intermediates (^EP) that undergo reversible O–O cleavage via formation of dicopper(III) bis(μ-oxo) species (O).^[8] These results show that even after many years of study, new Cu/O₂ reactivity patterns are still emerging.

Figure 1.

Controlled generation of different LCu–O₂ species by rational ligand design A) First row metal/O₂ species bearing macrocyclic 14-TMC and 12-TMC ligands. B) Crystallographically characterized ^EP species (C).^[5a,10]

One of the most extensively employed ligand types in bioinspired O₂-activation chemistry is the macrocycle 14-tetramethylcyclam (14-TMC), along with its contracted analogue 12-TMC. Nam and co-workers have extensively characterized M-(O₂^•−) and M-(O₂(H)) complexes of the 1st row of transition metals (Figure 1B); associated O₂ activation levels and substrate reactivity are dependent on both metal identity and TMC ring size.^[9] However, the reactivity of analogous TMC/Cu/O₂ systems has remained elusive. Schindler and co-workers^[10] recently reported the first example of end-on peroxo Cu₂O₂ species (^EP) bearing a macrocylic ligand tetb (rac-5,5,7,12,12,14-hexamethyl-1,4,8,11-tetraazacyclotetradecane), a ^EP complex with a distorted square-pyramidal (τ = 0.3) geometry and equatorial bound bent bridging O₂²⁻ ligand (Figure 1C). This geometry differs from the TMPA-based ^EP species, where the O₂²⁻ ligand is found to be coordinated in an axial position of a trigonal bipyramidal (τ = 0.9)^[11] geometry (Figure 1C), and suggests macrocyclic ancillary ligands can form Cu/O₂ species with unique structure. Herein, we report the first study on the Cu–O₂ chemistry of the 14-TMC and 12-TMC systems and the first example of a side-on (i.e., μ-η² :η²) peroxodicopper(II) complex (^SP) bearing the 12-TMC ligand.

Mixing equimolar amounts of the 14-TMC ligand with [Cu^I(CH₃CN)₄](BAr^F) (BAr^F: B(C₆F₅)₄⁻) in Et₂ O at room temperature under Ar allows the growth of colorless crystals, which were analyzed by single crystal X-ray diffraction (Figure 2A); the 14-TMC ligand shows an unusual type-V (TV) conformation (TV, where 1,8-Me groups are oriented up; 4,11-Me groups down)^[12] in the Cu^I complex TV-[(14-TMC)Cu^I]⁺,^[13] where the Cu ion adopts a flattened tetrahedral geometry. Surprisingly, this macrocyclic type-V [(14-TMC)Cu^I]⁺ complex is completely unreactive towards O₂ (see the Supporting Information for details).^[14]

Figure 2.

Displacement ellipsoid plots (50% probability level) of TV-[(14-TMC)Cu^I]⁺. A) (un-reactive towards O₂) and TI-[(12-TMC)Cu^I(CH₃CN)]⁺ B) (reactive with O₂). For clarity, the BAr^F anions and the H atoms are not depicted. See the Supporting Information for synthetic details and structural features.

The 12-TMC cuprous complex was synthesized similarly and the X-ray structural determination disclosed an N₅Cu(I) center in which the 12-TMC ligand adopts a type-I conformation (with all Me groups up) in TI-[(12-TMC)Cu^I(CH₃CN)](BAr^F) with a tightly coordinated acetonitrile molecule (d(Cu-N_CH3CN) = 1.92 Å) (Figure 2B). Unlike for TV-[(14-TMC)Cu^I]⁺, TI-[(12-TMC)Cu^I(CH₃-CN)]⁺ reacts with O₂ at −90 to −135 °C, forming a new green species in acetone or MeTHF (2-methyltetrahydrofuran). UV/Vis spectroscopy (Figure 3A) reveals an intense high-energy spectral feature at 364 nm (ε = 20 mm⁻¹ cm⁻¹) and two weak transitions at 450 and 635 nm, which are characteristic of a complex with a side-on peroxodicopper(II) coordination (^SP).^[2a] [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ is the first example of a ^SP complex bearing a tetradentate ligand; N₄ ligands typically provide for an end-on peroxo (^EP) coordination.

Figure 3.

A) Spectroscopic characterization (UV/Vis) of green [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ (λ_max = 364 nm) generated by the oxygenation of TI-[(12-TMC)Cu^I(CH₃CN)]⁺. Inset: rR spectra (λ = 380 nm) of [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ generated from ¹⁶O₂ and ¹⁸O₂ (*), and ¹⁶O₂–¹⁸O₂ subtraction. B) DFT optimized structure of ^SP [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ and TD-DFT calculated spectroscopic features (rR and UV/Vis).

Additional proof for the side-on [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ formulation is provided by resonance Raman (rR) spectroscopy (Figure 3A, inset). Laser excitation (λ = 380 nm) of frozen samples of [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ generated in both MeTHF and acetone led to observation of an isotope-sensitive feature at ν_Cu–O = 544 cm⁻¹ (Δ(¹⁸O₂) = −29 cm⁻¹) plus a ν_Cu–Cu = 264 cm⁻¹ isotope-insensitive feature that are typical of ^SP complexes.^[15] Higher energy features are observed at 1085 cm⁻¹ and around 1340 cm⁻¹, and are consistent with assignments as an overtone (2ν_Cu–O) and combination band (2ν_Cu–O+ν_Cu–Cu), respectively. Since the antisymmetric ν_Cu–O mode is not Raman active in a planar core, the unusually high intensity of ν_Cu–O in [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ suggests the Cu₂O₂ core adopts a butterfly-distorted geometry (see below). This is consistent with the perpendicular π*_v →Cu LMCT being observed (450 nm), which is a feature that has no intensity in planar Cu₂(O₂²⁻) cores.^[15b]

DFT calculations of the [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ intermediate led to the ^SP species as the lowest energy isomer, with both Cu centers presenting a distorted octahedral geometry around a butterflied Cu₂O₂ core (Figure 3 B), as suggested by the experimental UV/Vis and rR spectroscopies.^[16] The calculated Cu–O, O–O and Cu–Cu distances are consistent with other well-characterized ^SP complexes.^[17] The optimized Cu–N_av distance (2.32 Å) was found to be slightly longer than the ones found for other ^SP-type species (2.0–2.2 Å), due to the elongation of two of the Cu–N distances per Cu ion (2.39–2.56 Å). These weak Cu–N bonds have also been observed by Tolman and co-workers in the Cu/O₂ species bearing tridentate macrocyclic systems, inducing the formation of O-type species over ^SP.^[18] In our case, we propose that these weak, pseudoaxial, Cu–N bonds are a key factor to explain the unusual reactivity of the macrocyclic ^SP core (vide infra). DFT and TD-DFT calculations also reproduce the experimental rR data (calcd ν_Cu–O = 514 cm⁻¹ [exptl: 544 cm⁻¹]; calcd ν_Cu–Cu = 268 cm⁻¹ (exptl: 264 cm⁻¹); and absorption spectrum (calcd: 351, 450 and 622 nm; exptl: 364, 450 and 635 nm, see Figure S9).

The study of the reactivity of the different Cu₂O₂ centers (^EP, ^SP, O) towards external substrates has led to some general rules which depend on the core formulation.^{[19] E}P complexes are considered basic/nucleophilic species, while ^SP and O-type complexes are considered non-basic/electrophilic.^[19] However, this ^SP [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ species is found to be unreactive towards external substrates usually oxidized by electrophilic species (Figure 4A): phenols (2,4-tBu₂-PhOH/2,6-tBu₂-4-MeO-PhOH, for H^• abstraction), sodium phenolates (for ortho-hydroxylation) and PPh₃ (for O-atom transfer) are each unreactive when added in excess (up to 50 equiv, see SI). The 12-TMC-based ^SP species is also found to be stable when it was exposed to substrates that can be oxidized by nucleophilic M(O₂²⁻) species such as 4-substituted benzaldehydes (oxidation to benzoic acids) or cyclohexanecarboxaldehyde (decarboxylation to cyclohexene).^[9a]

Figure 4.

A) Reactivity towards external substrates shown by ^SP [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ (right) compared to generic reactivity of ^EP and ^SP/O cores. B) UV/Vis changes and kinetic analysis (inset) for the reaction of [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ (λ_max = 364 nm) with acid to generate 2 equivalents of [(12-TMC)Cu^II]²⁺ (λ_max = 590 nm) and H₂O₂.

On the other hand, when [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ was exposed to a strong acid (H⁺•DMF•CF₃SO₃⁻), decay of its UV/Vis features was observed (Figure 4B) with concomitant formation of new weak features centered at 590 nm, corresponding to the mononuclear [(12-TMC)Cu^II(L)]²⁺ complex (for Cu^II quantification and EPR characterization see Figure S7). H₂O₂ formation was confirmed by iodometric titration of the final reaction solution (95% based on [Cu₂O₂]₀, see Figure S6). To the best of our knowledge, this is one of the first examples of a ^SP species that can be protonated quantitatively to release H₂O₂, a reaction that is typical of ^EP complexes.^[20]

In order to obtain mechanistic insight, kinetic analysis on the reaction of [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ with (H⁺•DMF•CF₃SO₃⁻) was conducted under pseudo-first-order conditions (5–20 equiv of acid). The decrease of the ^SP UV/Vis features over time was fit using pseudo-first-order exponential decays to obtain the kinetic constants (k_obs) at different [H⁺] (Figure 4B, inset). The k_obs increased parabolically with increasing acid concentration (rate law: d[Cu₂O₂]/dt = −k_H+[Cu₂O₂][H⁺]²), suggesting an unusual involvement of 2 molecules of acid during the rate determining step (r.d.s.). The same kinetic behavior was observed with mild acids such as substituted acetic acids (RCO₂H, R: Cl₃C, Cl₂HC, CF₃, CH₃). The k_obs values were found to be independent on the H⁺ strength and an inverse kinetic isotope effect (KIE) was observed (k_CF3CO2H/k_CF3CO2D = 0.8, see Figure S5 in the Supporting Information).

Protonation of Cu₂O₂ species to generate H₂O₂ is characteristic of ^EP species; ^SP cores are usually stable under acidic conditions. However, the basicity of this [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ complex is not matched with a nucleophilic character. We propose that the unique reactivity of [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ is dictated by the ligand, since the 6-coordinate Cu ions have no coordination positions available in the Cu₂O₂ core for incoming substrates. The lack of a pK_a effect in the O₂²⁻ protonation indicates the r.d.s. involves protonation by a common intermediate, regardless of the identity of added acid. This suggests the TMC ligand is initially protonated by acid, and subsequently transfers H⁺ intramolecularly to the peroxide. DFT calculations support this idea, showing that two sequential TMC protonation events lead first to isomerization of the ^SP to a mono-protonated η₂ :η¹-Cu₂O₂ and subsequently a di-protonated ^EP species (Figure 5), both stabilized by H-bonding interactions. We propose that this nucleophilic ^EP core is now activated towards intramolecular H⁺ transfer from the protonated N atoms of TMC (see Figures S10–S13 in the Supporting Information).

Figure 5.

Proposed reaction mechanism for the protonation of [(12-TMC)₂Cu^II₂(O₂²⁻)]²⁺ based on experimental findings and DFT calculations.

The experimental observations (2nd order in [H⁺] with no pK_a effect) and DFT calculations point towards a mechanism where a first fast protonation to 12-TMC (uphill)^[21], enabled by the weak Cu–N bonds, is followed by a second rate-determining protonation. Rate-limiting proton transfer from 12-TMC to the peroxo moiety is also consistent with the inverse KIE since the O–H bond being formed is stronger than the N–H bond being broken. We propose that the macrocyclic ligand has a dual role upon protonation: the flexibility to accommodate different Cu₂O₂ isomers (control of the primary coordination sphere) and the ability to act as a proton shuttle (control of the secondary coordination sphere by H-bonding interactions). A similar duality is observed in the reactivity of selected metalloenzymes (e.g. Cu,Zn-SOD^[22]) and bio-inspired catalysts (e.g. Ni-catalyzed H₂ formation^[23]).

In conclusion, we have succeeded in generating and describing previously elusive LCu/O₂ chemistry of the macrocyclic ligands 14-TMC and 12-TMC. The dramatic effects on structure and chemistry previously known for differing TMC ring-size,^[9] along with the finding of a 12-TMC Cu^I/O₂ adduct with distinctive structure and behavior, are manifest by the new findings. Current studies are focused on understanding the Cu-dioxygen chemistry of other TMC isomers generated in the in situ reactions of 14-TMC/Cu^I,^[14a] along with exploring other macrocyclic systems beyond 12-TMC and 14-TMC likely to provide Cu/O₂ derived species with new structural, spectroscopic and reactivity features.