Elucidation of the Electrocatalytic Nitrite Reduction Mechanism by Bio-Inspired Copper Complexes

Mononuclear copper complexes relevant to the active site of copper nitrite reductases (CuNiRs) are known to be catalytically active for the reduction of nitrite. Yet, their catalytic mechanism has thus far not been resolved. Here, we provide a complete description of the electrocatalytic nitrite reduction mechanism of a bio-inspired CuNiR catalyst Cu(tmpa) (tmpa = tris(2-pyridylmethyl)amine) in aqueous solution. Through a combination of electrochemical studies, reaction kinetics, and density functional theory (DFT) computations, we show that the protonation steps take place in a stepwise manner and are decoupled from electron transfer. The rate-determining step is a general acid-catalyzed protonation of a copper-ligated nitrous acid (HNO2) species. In view of the growing urge to convert nitrogen-containing compounds, this work provides principal reaction parameters for efficient electrochemical nitrite reduction. This contributes to the investigation and development of nitrite reduction catalysts, which is crucial to restore the biogeochemical nitrogen cycle.


General
All chemicals and solvents were purchased from commercial chemical suppliers without further purification. All electrolyte solutions for electrochemistry were prepared using Mili-Q grade water and high purity salts, NaNO2 (99.999% trace metals basis, Sigma Aldrich), NaH2PO4 (99.99% Suprapur, Merck), Na2HPO4 (≥99.999% TraceSELECT, Honeywell Fluka). D2O for the proton inventory experiments was obtained from Eurisotop (99.9 % D). The Ar gas used in electrochemical experiments was supplied by Linde. The pH of all electrolyte solutions was measured using a HI 4222 pH meter from Hanna Instruments. UV-vis measurements were conducted on a Agilent Varian Cary 50 spectrophotometer. EPR spectra were recorded on a Bruker EMXPlus X-band spectrometer. [Cu(tmpa)MeCN] 2+ (tmpa = tris(2pyridylmethyl)amine) was synthesized as previously reported. 1

Electrochemistry measurements
All electrochemical experiments were carried out in custom-build glass cells using a threeelectrode setup. In all measurements the working electrode was made from glassy carbon (A = 0.71 cm 2 , Metrohm), encapsulated in PEEK. The working electrode was polished on a Struers LaboPol-30 polishing machine in two steps. First the GC surface was cleaned with diamond polish (1.0 µm, DiaPro, Struers) on polishing cloth (Dur-type) for 1 minute. Next, the electrode was polished on a different polishing cloth from the same material with silica suspension (OP-S Non-dry, Struers) for 2 more minutes. After this, the electrode was sonicated in Mili-Q for 15 minutes. In all CV measurements the counter electrode was made of a Au wire, which was flame annealed and rinsed with Mili-Q at the start of the day. The reference electrode in all experiments was a double junction Ag/AgCl electrode (3M KCl, Metrohm) of which the equilibrium potential was regularly measured versus a reversible hydrogen electrode (RHE).
All glassware for electrochemistry measurements was regularly cleaned by overnight soaking in an aqueous solution of KMnO4 (1 g/l) and H2SO4 (0.5 M), after which it was further cleaned in a dilute solution of H2O2 and H2SO4 and boiled three times for 30 minutes in Mili-Q water.
Prior to every measurement, all glassware was boiled in Mili-Q for 30 minutes more. All electrochemical measurements were conducted on PGSTAT 12,204 or 128N potentiostats from Autolab, using NOVA software. Prior to recording a CV measurement in presence of catalyst, three CV scans were always recorded in absence of catalyst in a blank electrolyte solution to ensure the cleanliness of the working electrode. In addition, the electrode surface of the working electrode was polished every time after recording CVs in presence of catalyst.

Proton inventory experiments
Prior to all proton inventory experiments, all glassware was placed in an oven at 160 °C overnight. In addition, [Cu(tmpa)(CH3CN)]OTf2, NaNO2, Na2HPO4, and NaH2PO4 were dried in a vacuum oven for one hour before preparing the electrolyte solutions. A 50 mM phosphate buffer solution was prepared in both D2O and H2O, and the pH of both solutions was measured using a H2O-calibrated pH meter. Using Equation 1, 2 the measured pH of the D2O solution (7.1) was converted, obtaining a pHD2O of 7.0. The electrodes were prepared as described above, except that the working electrode was sonicated in D2O after polishing, and the counter electrode and reference electrode were rinsed with D2O prior to the measurement. As experiments are very sensitive for the D2O fraction, the surface of the working electrode was saturated by the correct D2O fraction by measuring 10 CV cycles in a blank solution in absence of catalyst prior to recording the catalytic activity. In addition, prior to every experiment, the working electrode was soaked in the catalyst solution for a few minutes before recording the CV scans.

Rotating Disk Electrode measurements
All rotating disk electrode (RDE) experiments were carried out with an Autolab PGSTAT 12 potentiostat and a MSR rotator from Pine Instruments. All RDE measurements were carried out using a custommade three-electrode electrochemical cell with a volume of more than 40 ml, in which the counter electrode compartment is separated from the bulk solution by a glass frit.

Homogeneity of Cu(tmpa) during electrochemistry
The homogeneity of the catalyst during nitrite reduction catalysis was investigated in CV experiments ( Figure S2). In

Proton inventory experiments
Proton inventory experiments were carried out by recording CVs of Cu(tmpa) in solutions with varying fractions of D2O in the absence of substrate (See Figure S9) and in the presence of nitrite (See Figure   S10a). The basis of a proton inventory experiment is that each proton in the rate-determining step has its own fractionation factor (φ), which describes the extent of protonation at the transition state and is equal to kD/kH, the inverse of a kinetic isotope effect. The overall value of kD/kH is a multiplication of the individual φ values of each proton involved. By varying the D2O fraction (n) the φ values can be deconvoluted using Equation 2. A plot of kn/kH against n will only lead to a linear relationship if one proton is involved in the rate determining step, and a quadratic one if two protons are involved that have a different fractionation factor. 4 In order to determine the fractionation factor, a plot of (icat/ip) 2 n / (icat/ip) 2 H was prepared, in which n represents the fraction of D2O and H represents the measurements in pure H2O (See Figure S10b).
The value of ip in this plot was determined as the average current of all anodic peaks in Figure S9, and icat was determined for every D2O fraction as the average catalytic current recorded at -0.5 V vs.
Ag/AgCl in four separate measurements ( Figure S10a). The

DFT calculations
DFT calculations as well as visualization and structural analysis of the calculated structures were performed using the Amsterdam Density Functional (ADF) engines of the AMS2022 program package developed by SCM. 5,6 All DFT calculations were performed using the B3LYP exchange correlation functional 7, 8 including D4 dispersion corrections. 9 A triple zeta basis set with a polarization function (TZP) 10 was used and solvent effects in water were accounted for using the COSMO implicit solvent model. 11 The geometry that formed the starting point of the DFT calculations was obtained from the previously reported crystal structure of [Cu(tmpa)(NO2-κN)] + . 12 Thermodynamic corrections to the electronic energies were calculated using the harmonic-oscillator model at P = 1 atm and T = 298 K. Table S1 shows for all optimized geometries in this study the overall charge of the structure, its spin multiplicity and the calculated Gibbs Free Energy in kcal/mol.  Figure

Energy profile of the RDS
In order to identify a transition state structure belonging to the rate-determining, second protonation step, a linear transit calculation was carried out. In this calculation, the distance between a proton on a H2PO4molecule and the oxygen atom of HNO2, bound to the copper centre, was changed from 1.97 to 0.98 Å. This calculation resulted in identification of a transition state structure (See Figure S24), which showed only one single mode with a negative frequency. The reactant and product states connected to this transition state where found by an intrinsic reaction coordinate (IRC) calculation, resulting in the structures present in Figure S25 and Figure S26. Analysis of the product state showed that two other local minima could be accessed ( Figure S27 and Figure S28) upon changing the spin multiplicity of the system to the triplet state.