Green Ammonia Generation with a Polypyridine Cobalt Complex and Visible Light

Abstract

The use of a first-row transition metal complex and visible light was exploited to promote the photosynthesis of ammonia from nitrite ions. This is achieved in neutral aqueous solution by employing a cobalt polypyridine complex (CoL) within a light-driven catalytic scheme involving [Ru­(bpy)₃]²⁺ (where bpy = 2,2′-bipyridine) as the sensitizer and ascorbate as the electron donor. Under visible light irradiation, the resulting photosystem produces NH₃ with a quantum yield up to 3% and a selectivity approaching 100%, achieving a maximum turnover number (TON) up to 2150. Transient absorption spectroscopy and density functional theory provided a robust mechanistic picture, showing that binding of the NO₂ ^– substrate occurs at the one-electron reduced metal complex, while also ascertaining the energetic feasibility of subsequent elementary reaction steps. All in all, this work uncovers the great potential of first-row transition metal polypyridine complexes toward the synthesis of chemicals of industrial relevance.

Ammonia (NH₃) is one of the most important industrial chemicals worldwide, both in terms of production volume and breadth of applications. − Industrial synthesis is, however, dominated by the Haber-Bosch process that still relies on fossil fuels. Decarbonizing NH₃ synthesis has thus become a critical challenge. − Recently, the nitrite reduction reaction (NO₂ ^–RR) has emerged as a viable alternative for the production of NH₃. − Cobalt complexes (Scheme a) have been employed as molecular catalysts to promote the NO₂ ^–RR from neutral aqueous solutions mainly resulting in the formation of NH₃. − However, the investigation of the activity has been solely limited to electrochemical conditions with the catalysts either used as homogeneous species in solution − or immobilized onto solid-state supports. − Notably, to the best of our knowledge, no transposition of this reactivity under light-driven conditions has been documented with only one reported molecular system employed for light-driven generation of NH₃ based on a rare-earth element.

1. (a) Structures of Co Complexes Reported for Electrocatalytic NO₂ ^–RR and (b) Structure of CoL and Mechanism of Light-Driven NO₂ ^–RR to NH₃ Presented in This Work .

^a PS = Photosensitizer; ED = Electron Donor.

In recent years, we reported the use of metal complexes bearing the redox-active, hexadentate DBPy-PyA ligand (DBPy-PyA = (1-([2,2′-bipyridin]-6-yl)-N-([2,2′-bipyridin]-6-ylmethyl)-N-(pyridin-2-ylmethyl)­methanamine)) and related derivatives as catalysts for the hydrogen evolution reaction (HER) − and CO₂ reduction. − Their great performance mainly stems from the redox-active nature of the ligand and its ability to dissociate under turnover conditions, creating internal proton relays for efficient proton transfer (PT) to the catalytic site, which is pivotal for achieving efficient catalysis. Herein, we show that the cobalt polypyridine complex (CoL) (Scheme b) can be a competent catalyst for selective reduction of NO₂ ^– to NH₃ in neutral aqueous solution. Its catalytic activity can also be transposed under light-driven conditions using [Ru­(bpy)₃]²⁺ (where bpy = 2,2′-bipyridine) as the photosensitizer (PS) and ascorbate as the electron donor (ED), providing the first example of NH₃ photosynthesis from aqueous NO₂ ^– solution using a first-row transition metal complex.

CoL was obtained following a reported protocol, and the spectroscopic properties perfectly match the reported data. Cyclic voltammetry (CV) was performed in a phosphate buffer solution (0.25 M) at pH 7 (Figure a and SI for further details). Under these conditions, a hexacoordinate species, following removal of the monodentate ligand (i.e., a solvent molecule), is expected upon dissolution of CoL. , DFT calculations indeed predict endergonic binding of a seventh ligand at the Co­(II) center. During the cathodic scan, a reduction process can be observed at −1.30 V vs SCE, which can be assigned, for comparison with the CV in acetonitrile, to a formal Co­(II)/Co­(I) electron transfer (ET) step. A current increase is then observed upon further cathodic scan without any apparent observation of clear reduction waves. This can be explained because, at more negative potentials than −1.40 V vs SCE, the expected reduction process, involving a formal Co­(I)/Co(0) step, is presumably coupled with PT to result in HER catalysis. This is indeed anticipated considering the reactivity of CoL in acetonitrile with weak organic acids. During the return scan, a sharp anodic peak is detected at −1.15 V vs SCE. Similar features were found in electrochemical studies in aqueous solutions for related cobalt complexes and attributed to stripping upon reoxidation of the neutral Co(0) complex adsorbed onto the electrode during the cathodic sweep. − This feature disappears at more acidic pH (Figure S6), as Co­(I) protonation inhibits formation (and adsorption) of Co(0). ,

1.

(a) CV of 1 mM CoL in 0.25 M phosphate buffer at pH 7 with 0–0.2 M NaNO₂ at a scan rate of 0.1 V/s; (b) FEs of NH₃ and H₂ formation from CPE of 1 mM CoL in 0.25 M phosphate buffer at pH 7 with 0.05 M NaNO₂ and control experiments (applied potential of −1.4 V vs SCE).

The electrochemical response of CoL was further assessed in the presence of nitrite to infer its catalytic ability toward the NO₂ ^–RR. Under these conditions, the comparison of the absorption spectra of CoL in 0.25 M phosphate buffer at pH 7 in the presence of NaNO₂ (Figure S19) indicates that nitrite does not coordinate to Co­(II). Nevertheless, the addition of NO₂ ^– triggers a current enhancement at the Co­(II)/Co­(I) reduction (Figure a), consistent with nitrite binding at the one-electron reduced species and ensuing catalysis. The current measured slightly increases with the concentration of the substrate leading to an i _c/i _p ratio up to 17.6 with 0.2 M NaNO₂. At fixed NO₂ ^– content, the current increases with decreasing pH and increasing phosphate buffer concentration (Figures S9 and S10), supporting the key role of proton delivery in the catalytic process. ,, Control tests (Figure S11) finally confirm that CoL is necessary to trigger any current enhancement.

Controlled-potential electrolysis (CPE) was performed to determine and quantify the reduction products. The results (Figure b) show that NH₃ is formed as virtually the sole reduction product with a faradaic efficiency (FE) of 83%. No NH₂OH is detected, while H₂ is generated with an FE of only 0.4%. Importantly, when CPE was conducted in the absence of CoL, the FE for NH₃ formation drastically decreased to only 16%, whereas the FE for H₂ increased to 36% (Figure b). Furthermore, an FE of 28% for NH₃ and a parallel FE of 23% for H₂ were measured during electrolysis of a catalyst-free solution using the electrode previously employed in the CPE with CoL (unpolished electrode). All of these controls confirm that selective NH₃ production requires the presence of CoL and does not originate from heterogeneous species formed on the electrode. Notably, the large FE for NH₃ production is in good agreement with the constant current response (Figure S12), supporting the stability of CoL under operation.

We then examined the possibility of extending this reactivity under light-driven conditions. To this purpose, we combined CoL with [Ru­(bpy)₃]²⁺ as the sensitizer and ascorbate as the electron donor. We first started by evaluating the effect of the concentration of CoL under 460 nm irradiation of 0.25 M phosphate buffer solutions at pH 7 containing 0.5 mM [Ru­(bpy)₃]²⁺, 0.1 M ascorbate, and 0.05 M NaNO₂. Figure a collects the results in terms of amount of NH₃ produced after 23 h of irradiation within a concentration range spanning from 2 μM up to 100 μM CoL. The activity in terms of product formation peaks at intermediate concentrations (Figure a), while the turnover number (TON) maximizes at 2 μM CoL, achieving a remarkable value of 2150 (Figure S15). We next selected a concentration of 25 μM CoL for further assays. Figure b depicts the kinetics of NH₃ formation as a function of the irradiation time. Production of NH₃ is appreciably linear in time within the first 10 h, and a quantum yield of 3% can be estimated (see SI). NH₃ photogeneration then starts slowing down at longer irradiation times and finally reaches a plateau producing up to 43 μmol of NH₃ after 72 h. NH₃ is produced with nearly 100% selectivity, as no NH₂OH is detected via ¹H NMR and only 0.1 μmol of H₂ is measured by GC after 23 h.

2.

(a) Amount of NH₃ generated after 23 h of irradiation (LED @460 nm) of 0.25 M phosphate buffer solution (5 mL) at pH 7 containing 0.5 mM [Ru­(bpy)₃]­Cl₂·6H₂O, 0.1 M sodium ascorbate, 0.05 M NaNO₂, and 2–100 μM CoL; (b) kinetics of NH₃ formation at [CoL] = 25 μM and comparison of the ¹H NMR signal associated with NH₄ ⁺ measured with 0.05 M Na¹⁴NO₂ or Na¹⁵NO₂ (inset).

Notably, addition of fresh [Ru­(bpy)₃]²⁺ after 23 h of irradiation fully restores the activity, leading to a cumulative NH₃ production of 72 μmol after 46 h of irradiation (Figure S16). These findings confirm catalyst stability and identify degradation of the sensitizer as the main limiting factor. Comparison of the absorption spectra before and after photocatalysis shows a marked attenuation of the MLCT transition of the chromophore (Figure S17), thus corroborating this interpretation.

NH₃ photogeneration after 23 h of irradiation is maximized at pH 7 (Table S1) and, at this pH, is only weakly dependent on both NO₂ ^– and phosphate buffer concentrations (Tables S2 and S3). The similar performance recorded at different nitrite loadings is consistent with the modest NO₂ ^– conversion recorded at the plateau (17%), whose value points out how substrate consumption does not represent a limiting factor herein. Control experiments (Table S4) confirm that CoL, [Ru­(bpy)₃]²⁺, and ascorbate are all simultaneously required for the effective generation of NH₃. Furthermore, experiments conducted with Na¹⁵NO₂ show the formation of isotopically labeled ¹⁵NH₃ (Figure b, inset), thus demonstrating that it unequivocally originates from the reduction of nitrite. As a final remark, the NH₃ photoproduct can also be quantitatively separated in the form of NH₄Cl salt, devoid of impurities from the photolyzed solution (Figure S18), upon vacuum transfer of the photolyzed mixture into a 3 M HCl aqueous solution, highlighting how the homogeneous approach used herein can also lead to clean and quantitative product capture.

We next performed spectroscopic studies to attain mechanistic insights. Figure a depicts the laser flash photolysis (LFP) experiments conducted on a 0.25 M phosphate buffer solution at pH 7 with [Ru­(bpy)₃]²⁺, ascorbate, and CoL.

3.

(a) Transient absorption spectra between 1 and 50 μs measured by LFP (excitation at 532 nm) of an N₂-purged 0.25 M phosphate buffer solution at pH 7 containing 70 μM [Ru­(bpy)₃]­Cl₂·6H₂O, 0.1 M sodium ascorbate, and 100 μM CoL and (b) kinetic traces at 700 nm with 0–50 mM NaNO₂.

The transient spectrum measured at 1 μs (Figure a, top) features an absorption band at 500 nm which corresponds to the reduced [Ru­(bpy)₃]⁺ species , obtained by reductive quenching of the excited sensitizer with the ascorbate donor. , This absorption progressively decreases in the following 10 μs (Figure a, top), leading to the concomitant formation of a broad transient absorption characteristic of the reduced CoL, ,,− which is identified as a triplet ground state via the Evans method (Figure S24), in agreement with the calculations (Tables S5–S7). , This residual absorption finally decays to the baseline within 50 μs (Figure a, bottom) mainly due to recombination between Co­(I) and the ascorbate radical. , Interestingly, the transient absorption signal at 700 nm, associated with the reduced CoL, undergoes a faster decay in the presence of NaNO₂ (Figure b). This evidence can be explained by assuming binding of NO₂ ^– to the formal Co­(I) generated upon ET from [Ru­(bpy)₃]⁺. UV–vis spectroelectrochemistry (SEC) in acetonitrile confirms the abatement of the absorption characteristic of reduced CoL in the presence of 0.05 M TBANO₂ (Figure S22), supporting the proposed mechanistic hypothesis. Interestingly, the rate of this reaction increases with NO₂ ^– concentrations for values below 0.01 M, finally reaching a plateau above 0.05 M. Fitting of the observed rate k as a function of [NO₂ ^–] was attempted using an empirical saturation model (Figure S23), which allows one to extract an apparent maximum velocity of 3.5·10⁵ s^–1 for nitrite coordination at the reduced CoL under conditions relevant to the photocatalytic assays. The apparent saturation attained at larger NO₂ ^– concentrations corroborates the experimental observation of a light-driven catalytic activity independent of substrate loading in the range of 0.05–0.5 M (Table S3).

DFT-D3//COSMO-RS calculations (see SI for computational details) were finally conducted to propose a catalytic mechanism. A schematic representation, with relevant structures and free energy changes of key chemical steps, is presented in Scheme , while details of structural and energetic parameters are provided in Scheme S1 and Figure S26. As confirmed by LFP, nitrite binding occurs at Co­(I) (+12.4 kcal·mol^–1). The complex then evolves through several intermediates, whose formation, implying PT and dehydration steps, is generally thermodynamically favorable (see SI). All these processes ultimately lead to a Co-amino species from which NH₃ release can occur. This step is exergonic by −8.0 kcal·mol^–1, supporting efficient product release. In contrast, computations predict an endergonic NH₂OH release (+8.3 kcal·mol^–1), mainly ascribed to favorable H-bonding interactions with a dangling pyridine, corroborating the selective formation of NH₃, as experimentally observed.

2. Proposed Catalytic Mechanism for NO₂ ^–RR to NH₃ by CoL with Structures of Relevant Intermediates and Free Energy Changes (in kcal·mol^–1) Associated with Key Chemical Steps .

^a See SI for details.

In conclusion, we reported for the first time the integration of molecular catalysis by a first-row transition metal complex and visible light to promote selective conversion of NO₂ ^– into NH₃. This is achieved in aqueous solution at neutral pH by combining the cobalt complex CoL with [Ru­(bpy)₃]²⁺ as the photosensitizer and ascorbate as the electron donor. This work underscores the great potential of polypyridine complexes for advancing sustainable synthetic methods relevant to high-value industrial products.

Glossary

Abbreviations

NO₂ ^–RR

nitrite reduction reaction

HER

hydrogen evolution reaction

PT

proton transfer

ET

electron transfer

PS

photosensitizer

ED

electron donor

CV

cyclic voltammetry

CPE

controlled potential electrolysis

LFP

laser flash photolysis

GC

gas chromatography

SEC

spectroelectrochemistry

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.6c01477.

• Experimental and computational methodologies, additional electrochemical measurements, additional photochemical studies, additional mechanistic data, and detailed catalytic mechanism (PDF)

• Computational data: coordinates of all stationary points and their respective energies (ZIP)

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.