Reactivity and Mechanism of Photo- and Electrocatalytic Hydrogen Evolution by a Diimine Copper(I) Complex

: The tetrahedral copper(I) diimine complex [Cu(pq) 2 ]BF 4 displays high photocatalytic activity for the H 2 evolution reaction with a turnover number of 3564, thus representing the ﬁrst type of a Cu(I) quinoxaline complex capable of catalyzing proton reduction. Electrochemical experiments indicate that molecular mechanisms prevail and DFT calculations provide in-depth insight into the catalytic pathway, suggesting that the coordinating nitrogens play crucial roles in proton exchange and hydrogen formation.

Distorted trigonal bipyramidal and square pyramidal Cu(II) complexes bearing polypyridine chelates, which are supposed to minimize the inherent lability of the d 10 Cu(I) ion, have been extensively explored as proton reduction electrocatalysts and photocatalysts. Their proposed catalytic pathways involve proton coupled electron transfer (PCET) processes leading to a Cu(II)-hydride that evolves H 2 via intramolecular coupling between the hydride and the proton of a pendant pyridine nitrogen [24][25][26][27]. Square planar copper corroles can evolve H 2 electrocatalytically via an heterolytic pathway, where a Cu hydride reacts with a solution proton [28], while a Cu(II) complex bearing a non-innocent thiosemicarbazone ligand exhibits metal-assisted ligand-centered electrocatalytic H 2 evolution [29]. Other Cu(II) molecular catalysts involving tetradendate triazenido [30], Schiff base [31,32], oxamato [33], oxamido [34], phenanthroline [35] and imidazole-based [36] ligands have been recently presented. Notably, many of those catalysts decompose under reductive conditions to afford H 2 evolving materials, thus part of their activity is attributed to heterogeneous deposits of nanoparticles [19].
Herein, we examined the photocatalytic H 2 production behavior of [Cu I (pq) 2 ](BF 4 ) ([1]BF 4 ), a homoleptic tetrahedral Cu(I) diimine complex bearing the bidentate 2-pyridin-2-yl-quinoxaline (pq) ligand (Figure 1a). A complex like [1] + but carrying two methyl groups on quinoxaline ring, has been previously reported to be an intermediate in CO 2 fixation catalysed by molecular [Cu(pq) 2 (H 2 O)](ClO 4 ) 2 [9]. Evaluation of the H 2 evolution mechanism by electrochemical and density functional theory (DFT) methods indicates that the quinoxaline nitrogen atoms are primarily protonated and subsequent metal centered PCET leads to the formation of a Cu(II)-hydride, that can evolve H 2 via a low energy transition state.

Synthesis
Complex [1] + was readily prepared from one equivalent of the Cu(I) starting [Cu(MeCN)4BF4] [37] and two equivalents of 2-pyridin-2-yl-quinoxaline (pq) [38] in acetonitrile. All operations were performed under a pure nitrogen atmosphere, using Schlenk and syringe techniques. [Cu(I)(MeCN)4BF4] [37] 1.0 mmol were dissolved in 5 mL acetonitrile and after a few minutes, 2-pyridin-2-yl-quinoxaline (pq) [38] 2.0 mmol dissolved in 10 mL acetonitrile was added and the mixture was stirred for 2 h. After this time, a deep purple solution was formed. Then the solvent was evaporated to dryness. The product was isolated in crystalline form with a yield of 63%. Calculated elemental analysis: C, 55 , Tables S1-S7), UV-visible (Figure 1b) 1 H-NMR ( Figure S1) and cyclic voltammetry. Although its crystal structure was reported before [39], we repeated it for accuracy reasons and for taking into consideration the different anion of the complex. Time dependent density functional theory (TD-DFT) calculations with the B3LYP functional adequately reproduce the key features of the absorption spectrum and, thus, allow a reliable band assignment (Figure 1b, Table S8). The absorption bands centered at 550 nm are attributed to metal to ligand charge transfer (MLCT) transitions of the Cu(I) d orbitals to the pq π* antibonding orbitals.

Photocatalytic H2 Production
The photocatalytic H2 production activity of [1] + was assessed in a multicomponent artificial photosynthesis system in N,N-dimethylformamide-water (DMF:H2O) mixture, using fluorescein (Fl) as a photosensitizer (PS) and triethanolamine (TEOA) as a sacrificial electron donor and the sample pH was 10

Synthesis
Complex [1] + was readily prepared from one equivalent of the Cu(I) starting [Cu(MeCN) 4 BF 4 ] [37] and two equivalents of 2-pyridin-2-yl-quinoxaline (pq) [38] in acetonitrile. All operations were performed under a pure nitrogen atmosphere, using Schlenk and syringe techniques. [Cu(I)(MeCN) 4 BF 4 ] [37] 1.0 mmol were dissolved in 5 mL acetonitrile and after a few minutes, 2-pyridin-2-yl-quinoxaline (pq) [38] 2.0 mmol dissolved in 10 mL acetonitrile was added and the mixture was stirred for 2 h. After this time, a deep purple solution was formed. Then the solvent was evaporated to dryness. The product was isolated in crystalline form with a yield of 63%. Calculated elemental analysis: C, 55 4 was also characterized by X-ray crystallography ( Figure 1a, Tables S1-S7), UV-visible (Figure 1b) 1 H-NMR ( Figure S1) and cyclic voltammetry. Although its crystal structure was reported before [39] , we repeated it for accuracy reasons and for taking into consideration the different anion of the complex. Time dependent density functional theory (TD-DFT) calculations with the B3LYP functional adequately reproduce the key features of the absorption spectrum and, thus, allow a reliable band assignment ( Figure 1b, Table S8). The absorption bands centered at 550 nm are attributed to metal to ligand charge transfer (MLCT) transitions of the Cu(I) d orbitals to the pq π* antibonding orbitals.

Photocatalytic H 2 Production
The photocatalytic H 2 production activity of   (Table S9). Control experiments in the absence of light or any component of the system led to zero catalytic activity under the same conditions. In addition, mercury poisoning test was employed to investigate if catalytically active colloidal particles were formed during the catalytic process. Addition of mercury to this catalytic system resulted in a 30% reduction of the catalytic activity after 24h of irradiation ( Table 1), suggesting that the complex acts as a molecular photocatalyst, even though hydrogen evolution is partly attributed to in situ formed heterogeneous species.  Control experiments in the absence of light or any component of the system led to zero catalytic activity under the same conditions. In addition, mercury poisoning test was employed to investigate if catalytically active colloidal particles were formed during the catalytic process. Addition of mercury to this catalytic system resulted in a 30% reduction of the catalytic activity after 24 h of irradiation ( Table 1), suggesting that the complex acts as a molecular photocatalyst, even though hydrogen evolution is partly attributed to in situ formed heterogeneous species. In photocatalytic H 2 evolution, the catalyst may accept electrons either from the excited PS or the reduced PS, resulting in oxidative and reductive quenching, respectively. The luminescence of the excited photosensitizer FL*, in a deaerated H 2 O solution of fluorescein, was plotted as a function of catalyst concentration and Stern-Volmer analysis yielded a quenching constant k sv = 1.3*10 5 M −1 , indicating that the excited fluorescein can provide the potential needed to drive reduction of the catalyst (Figure 2b). However, the reductive quenching possibly predominates the photocatalytic process, because the TEOA concentration is six orders of magnitude higher than the concentration of the catalyst.
Due to the non-innocent nature of the pq ligand and the proposed ligand centered proton reduction mechanism (vide infra), a zinc analogue of complex [1] + , [Zn(pq) 2 ] 2+ was synthesized and was employed as a photocatalyst in identical conditions to complex [1] + , in order to elucidate whether the process is entirely ligand centered or mediated by the participation of the metal center ( Figure S2). The zinc complex showed no catalytic activity. In addition, photocatalytic hydrogen production experiments were performed using pq as a catalyst where the catalytic performance did not exceed 3 TON proving that copper is an essential component in the proton reduction process.

Electrocatalytic Proton Reduction
The cyclic voltammogram (CV) of [1] + recorded in DMF with 0.1 M tetrabutylammonium hexafluorophosphate (n-Bu 4 NPF 6 ) as supporting electrolyte ( Figure 3a) features an irreversible redox process at −0.18 V vs. ferrocenium/ferrocene (Fc + /Fc), which may be assigned to Cu II /Cu I . Two irreversible reduction events at −1.70 V and −1.97 V are attributed mostly to the pq ligand with some contribution from the Cu metal center. The free ligand reduction potential in DMF is −2.0 V vs. Fc + /Fc ( Figure S3). The assignment of these reduction waves to the ligand-centered reduction processes is further supported by the calculated spin density of 0.18 on Cu at the reduced intermediate [1] 0 , as well as by the fact that the homoleptic Cu(I) complex with 2,2'-biquinoline (bq) has a similar redox process at the same potential window, which is attributed to bq/bq •− reduction [40]. The computationally derived oxidation and reduction potentials of [1] + are −0.25 V, −1.77 V and −2.18 V, respectively, which supports a reliable peak assignment.
Catalysts 2020, 10, x FOR PEER REVIEW 4 of 14 In photocatalytic H2 evolution, the catalyst may accept electrons either from the excited PS or the reduced PS, resulting in oxidative and reductive quenching, respectively. The luminescence of the excited photosensitizer FL*, in a deaerated H2O solution of fluorescein, was plotted as a function of catalyst concentration and Stern-Volmer analysis yielded a quenching constant ksv = 1.3*10 5 M −1 , indicating that the excited fluorescein can provide the potential needed to drive reduction of the catalyst (Figure 2b). However, the reductive quenching possibly predominates the photocatalytic process, because the TEOA concentration is six orders of magnitude higher than the concentration of the catalyst.
Due to the non-innocent nature of the pq ligand and the proposed ligand centered proton reduction mechanism (vide infra), a zinc analogue of complex [1] + , [Zn(pq)2] 2+ was synthesized and was employed as a photocatalyst in identical conditions to complex [1] + , in order to elucidate whether the process is entirely ligand centered or mediated by the participation of the metal center ( Figure S2). The zinc complex showed no catalytic activity. In addition, photocatalytic hydrogen production experiments were performed using pq as a catalyst where the catalytic performance did not exceed 3 TON proving that copper is an essential component in the proton reduction process.

Electrocatalytic Proton Reduction
The cyclic voltammogram (CV) of [1] + recorded in DMF with 0.1 M tetrabutylammonium hexafluorophosphate (n-Bu4NPF6) as supporting electrolyte (Figure 3a) features an irreversible redox process at −0.18 V vs ferrocenium/ferrocene (Fc + /Fc), which may be assigned to Cu II /Cu I . Two irreversible reduction events at −1.70 V and −1.97 V are attributed mostly to the pq ligand with some contribution from the Cu metal center. The free ligand reduction potential in DMF is −2.0 V vs Fc + /Fc ( Figure S3). The assignment of these reduction waves to the ligand-centered reduction processes is further supported by the calculated spin density of 0.18 on Cu at the reduced intermediate [1] 0 , as well as by the fact that the homoleptic Cu(I) complex with 2,2'-biquinoline (bq) has a similar redox process at the same potential window, which is attributed to bq/bq •− reduction. [40] The computationally derived oxidation and reduction potentials of [1] + are −0.25 V, −1.77 V and −2.18 V, respectively, which supports a reliable peak assignment.  Subsequently, [1] + was investigated with respect to its activity for electrocatalytic H2 production. As displayed in Figure 3b, upon addition of trifluoroacetic acid (TFA, pKa DMF = 6.0) [41] in a DMF solution of [1] + , a new reduction peak appears at −1.10 V and a significant catalytic current was observed over the potential of −1.66 V, attributed to catalytic H2 evolution. The overpotential of the catalyst was calculated [42] to be 720mV, based on the reduction potential of TFA in a 5mM DMF solution was calculated to be −0.94 V [43]. Several pieces of evidence suggest that the catalytic system is homogeneous. The linear dependence of catalytic current on the square of the scan rate (Figure 4a,b) indicates that the copper complex functions in a diffusion-controlled regime under the Subsequently, [1] + was investigated with respect to its activity for electrocatalytic H 2 production. As displayed in Figure 3b, upon addition of trifluoroacetic acid (TFA, pKa DMF = 6.0) [41] in a DMF solution of [1] + , a new reduction peak appears at −1.10 V and a significant catalytic current was observed over the potential of −1.66 V, attributed to catalytic H 2 evolution. The overpotential of the catalyst was calculated [42] to be 720 mV, based on the reduction potential of TFA in a 5 mM DMF solution was calculated to be −0.94 V [43]. Several pieces of evidence suggest that the catalytic system is homogeneous. The linear dependence of catalytic current on the square of the scan rate (Figure 4a,b) Catalysts 2020, 10, 1302 5 of 13 indicates that the copper complex functions in a diffusion-controlled regime under the experimental conditions, and that it is a molecular catalyst. Also, rinse test experiment (Figure 4c) indicates that absorbed materials on the electrode surface are not catalytically active. Furthermore, the current increases linearly with an increase in the acid concentration (Figure 4d). The electrocatalytic activity of [1] + was also assessed in aqueous solution in the presence of acetic acid as a source of protons (Figures 5a,b). A very useful method to investigate the possibility of a concerted electron and proton transfer process (proton coupled electron transfer, PCET) instead of two separate consecutive electron transfer (ET) and proton transfer (PT) processes is the construction of the Pourbaix diagram, in which the reduction potential versus the pH of the solution is plotted. [44] The Pourbaix diagram (Figure 5c) shows that the reduction potential of the first reduction peak shifts linearly with the pH value in the range of 2.5 to 6.0, with a slope of 66.1 mV, which indicates that electron reduction is coupled with proton uptake. Similar ligand centered PCET processes have been reported for polypyridine copper proton reduction catalysts [24][25][26][27]. The electrocatalytic activity of [1] + was also assessed in aqueous solution in the presence of acetic acid as a source of protons (Figure 5a,b). A very useful method to investigate the possibility of a concerted electron and proton transfer process (proton coupled electron transfer, PCET) instead of two separate consecutive electron transfer (ET) and proton transfer (PT) processes is the construction of the Pourbaix diagram, in which the reduction potential versus the pH of the solution is plotted [44]. The Pourbaix diagram (Figure 5c) shows that the reduction potential of the first reduction peak shifts linearly with the pH value in the range of 2.5 to 6.0, with a slope of 66.1 mV, which indicates that electron reduction is coupled with proton uptake. Similar ligand centered PCET processes have been reported for polypyridine copper proton reduction catalysts [24][25][26][27].

Computational Study of the Catalytic Mechanism
DFT calculations were performed, in an effort to provide further insight on the electrocatalytic mechanism by exploring the possible steps of the H 2 formation. The computed structures of all the intermediates involved in the catalytic pathways are presented in the Supplementary Information. The proposed proton reduction mechanism derived from the calculated free energies of possible intermediates of the reaction pathway is shown in Figure 6a.

Computational Study of the Catalytic Mechanism
DFT calculations were performed, in an effort to provide further insight on the electrocatalytic mechanism by exploring the possible steps of the H2 formation. The computed structures of all the intermediates involved in the catalytic pathways are presented in the Supplementary Information. The proposed proton reduction mechanism derived from the calculated free energies of possible intermediates of the reaction pathway is shown in Figure 6a.
In the presence of TFA (pKa = 6.0 in DMF) the first reduction to generate the doublet species [2] + is a PCET process, in which ligand reduction is concomitant with a nitrogen atom protonation. Three different isomers can be considered in terms of whether the site of protonation is the coordinating pyridine nitrogen (N1), the coordinating quinoxaline nitrogen (N2) or the non-coordinating quinoxaline nitrogen (N3) ( Figure S5).
The [Cu-N3-H] + isomer was found to be the most stable one, while the free energies of [Cu-N2-H] + and [Cu-N1-H] + are 7.2 and 10.6 kcal mol −1 higher than that of [Cu-N3-H] + , respectively. Protonation of Cu I to form a Cu III hydride was also considered, but the energy of this step is found to be 28.5 kcal mol −1 higher than the most stable species. This redox peak assignment is also supported by the cyclic voltammetry data of the free ligand, where a new reduction peak at −1.1 V appears in the presence of TFA ( Figure S3).
A second PCET step to afford intermediate [3] + is feasible on the basis of thermodynamics with a calculated potential of −0.90 V. The most favourable protonation site when the N3 is already protonated is the N2 ligating atom, on the same quinoxaline ring, which takes a tetrahedral conformation, thus it maintains the Cu-N2 bond as well as the N2-H bond. Protonation of the non-coordinating quinoxaline N3 is accompanied by reduction of the quinoxaline moiety under applied potential, which increases the basicity of the coordinating N2. Upon the second PCET, N2 In the presence of TFA (pKa = 6.0 in DMF) the first reduction to generate the doublet species [2] + is a PCET process, in which ligand reduction is concomitant with a nitrogen atom protonation. Three different isomers can be considered in terms of whether the site of protonation is the coordinating pyridine nitrogen (N1), the coordinating quinoxaline nitrogen (N2) or the non-coordinating quinoxaline nitrogen (N3) ( Figure S5).
The [Cu-N3-H] + isomer was found to be the most stable one, while the free energies of [Cu-N2-H] + and [Cu-N1-H] + are 7.2 and 10.6 kcal mol −1 higher than that of [Cu-N3-H] + , respectively. Protonation of Cu I to form a Cu III hydride was also considered, but the energy of this step is found to be 28.5 kcal mol −1 higher than the most stable species.
This redox peak assignment is also supported by the cyclic voltammetry data of the free ligand, where a new reduction peak at −1.1 V appears in the presence of TFA ( Figure S3).
A second PCET step to afford intermediate [3] + is feasible on the basis of thermodynamics with a calculated potential of −0.90 V. The most favourable protonation site when the N3 is already protonated is the N2 ligating atom, on the same quinoxaline ring, which takes a tetrahedral conformation, thus it maintains the Cu-N2 bond as well as the N2-H bond. Protonation of the non-coordinating quinoxaline N3 is accompanied by reduction of the quinoxaline moiety under applied potential, which increases the basicity of the coordinating N2. Upon the second PCET, N2 protonation induces loss of the π-conjugation of the pyrazine ring, which is reflected in the calculated C-C and C-N bond lengths, and N2 now lies above the plane of the pyrazine C atoms without being dechelated from the metal ([3] + , Figure S6). Other possible protonation sites and the corresponding relative free energies are shown in Figure S6. Notably, the formation of a Cu hydride is 30.2 kcal mol −1 energetically higher. Two further PCET processes result in protonation of the second quinoxaline ligand and the 4H + /4e − reduced intermediate [5] + is generated. Since protonation of the ligating nitrogens induces weakening of the Cu-N bonds, the possibility of [5] + decomposition was examined, but was found unfavorable ( Figure S7). In the presence of acid, five oxidation peaks appear at the anodic scan of the cyclic voltammogram at −0.70 V, −0.62 V, −0.41 and −0.26 V, which can be assigned to the four ligand centered redox processes, and at −0.07 V which is attributed to Cu I /Cu II oxidation (Figure 3b). Cyclic voltammogram within a narrow potential region (0.3 to −1.2 V) to prevent the formation of Cu(0) species, show that the four anodic peaks are attributed to the first reduction peak (Figure 5b). The reduction potential of intermediate [5] + to generate Cu(0) species is calculated to be −2.58 V, while a metal centered PCET at a calculated potential of −1.92 V (experimental −1.66 V) in the presence of TFA affords a trigonal pyramidal Cu(II) hydride, [6] + , which is proposed to be the H 2 evolving species.
[6] + can evolve H 2 via a low energy (+6.8 kcal mol −1 ) transition state (TS, Figure 6b) with a single imaginary frequency at 1245.6, that corresponds to H 2 formation, as confirmed by IRC (integrated reaction coordinates) calculation. The energy coordinate diagram of the hydrogen evolution reaction was generated from the free energy change of each step in the catalytic cycle ( Figure 7).

Conclusions
We have presented that the Cu(I) diimine complex [1] + can serve as a molecular photocatalyst for water reduction in combination with fluorescein as a photosensitizer and triethanolamine as a sacrificial electron donor achieving 3564 turnovers, after 24h of irradiation. Mechanistic studies

Conclusions
We have presented that the Cu(I) diimine complex [1] + can serve as a molecular photocatalyst for water reduction in combination with fluorescein as a photosensitizer and triethanolamine as a sacrificial electron donor achieving 3564 turnovers, after 24 h of irradiation. Mechanistic studies based on electrochemical experiments and DFT calculations support that protonation of the nitrogen atoms of the pq ligand via consecutive PCET processes allows more anodic reduction potentials for the Cu(II)-hydride formation and lowers the overpotential needed for proton reduction. Furthermore, the non-ligating quinoxaline nitrogen also increases the stability of the catalyst under reductive conditions and, thus, the present study provides important guidelines for future robust catalysts design. Detailed mechanistic studies using spectroelectrochemistry experiments are at the center of ongoing work, as well as the development of polydentate quinoxaline containing ligands, that are expected to further stabilize the Cu(I) species.

General Information
Reagents (starting materials, solvents) used in the experiments were purchased from Sigma Aldrich (Darmstadt, Germany) Merck (Darmstadt, Germany), Alfa Aesar (Heysham, Lancashire, UK) and Panreac Applichem (Barcelona, Spain). It was at least analytical grade and underwent further purification, where it was necessary to carry out the experiments. The solvents used to measure the absorption and emission spectra were of spectroscopic purity and originated from the above companies. Also, the water used in the experiments was from the Milli-Q instrument. In complex synthesis [39], the reagents that were used were purchased from the above companies.
Absorption spectra were measured on a Cary 300 (Varian Inc.-Agilent Technologies, Grenoble, France) or a U-2000 (Hitachi, Berkshire, United Kingdom ) spectrophotometers at 25 • C. Quartz alveoli were used with a stopper, a 1.00 cm optical path. Emission spectra were measured on a RF-5301PC spectrophotometer (Shimadzu, Hesse, Germany). Quartz alveoli were used with a 1.00 cm optical path, using the software OriginPro 9.0 (Arezzo, Italia). 1 H-NMR spectra were measured on a Unity Plus 300/54 spectrophotometer (Varian, Grenoble, France).
For the X-ray diffraction studies, suitable crystals covered with paratone-N oil were placed onto the tips of glass fibers or scooped up in cryo-loops at the end of a copper pin and transferred to a goniostat for data collection with a SuperNova A diffractometer (Oxford Diffraction, Frankfurt, Germany,) equipped with a CCD area detector and a graphite monochromator utilizing Mo-Kα radiation (λ = 0.71073 Å). Empirical absorption corrections were applied using CrysAlis RED software [45] (multi-scan based on symmetry-related measurements). Structures were solved using SIR92 [46] and refined on F 2 using full-matrix least squares with SHELXL97 [47]. Software packages used: CrysAlis CCD for data collection, CrysAlis RED for cell refinement and data reduction, WINGX for geometric calculations [48] and MERCURY [49] for molecular graphics. The program SQUEEZE [50], a part of the PLATON crystallographic software package, was used to remove contribution of highly disordered solvent molecules. All non-hydrogen atoms were refined anisotropically, whereas the hydrogen atoms were placed in calculated, ideal positions and refined as riding atoms with relative isotropic displacement parameters. Unit cell data and structure refinement details are listed in Table S2.
Photocatalysis experiments were performed in special led headlamps where the emitted radiation is in the range of the visible radiation spectrum (λ > 400 nm). The samples were irradiated for at least 72 h until the maximum performance of each system was found. The gaseous composition of the mixture and therefore the hydrogen detection was done by gas chromatography. The gas chromatograph was a 430-GC type (Bruker, Karlsruhe, Germany) equipped with a thermal conductivity detector (TCD) and operated under the following conditions: carrier gas: nitrogen, oven temperature 70 • C, detector temperature 150 • C, injector temperature 80 • C, molecular sieves column 5Å. The analysis of the chromatograms was done using the Galaxie software (Varian). The flask under the photocatalyst experiment was degassed by argon for 15 min and the system was sealed and illuminated for as long as the experiment lasts to find the maximum hydrogen output. The hydrogen produced is measured by integration of the chromatographic peaks and the quantification of the amount of hydrogen was based on calibration using Zn in aqueous HCl solution. Turnover number (TON) is used to quantify catalyst performance and represents the overall number of moles of H 2 produced upon irradiation of a photocatalytic system per mole of catalyst in the system.

Cyclic Voltammetry
All cyclic voltammetry (CV) measurements were recorded using an AFCBP1 (Pine Instrument Company, Lyon, France) and using the Aftermath Data Organizer software version 1.4.7881. Experiments were conducted in a three-electrode single-compartment cell with a glassy carbon working electrode, Ag/AgCl reference electrode and Pt counter-electrode at room temperature and using degassed solutions by purging with N 2 . All measurements were performed in N,N-dimethylformamide (DMF) in the presence of 0.1 M tetrabutylammonium hexafluoro-phosphate (n-Bu 4 NPF 6 ) as a supporting electrolyte and ferrocenium/ferrocene (Fc + /Fc) redox couple was used as an internal potential reference.

Theoretical Studies
Geometry optimizations were carried out using the B3LYP density functional [52,53] in combination with the 6-31 g(d,p) basis set [54,55] as implemented in the Gaussian16 software package [56]. The choice of this functional was made based on agreement between the calculated optimized geometry and the experimentally defined structure (Table S10) as well as on its successful implementation in previous studies on copper-based catalysts. Gibbs free energies were calculated as single-point corrections on the optimized structures employing the larger basis set 6-311+g(2d,p) and D3 dispersion corrections [57] to obtain a more accurate value of the absolute free energy of each species. The solvent effects were considered through single-point energy calculations using the SMD continuum solvation model [58]. Frequency calculations were performed on all the optimized structures at the same level of theory, to identify all the stationary points as minima (zero imaginary frequency) or transition states (only one imaginary frequency) and to obtain Gibbs free energy corrections at 298.15 K. Time-dependent DFT (TD-DFT) calculations were performed to simulate the absorption spectra of [1] + .