Synthesis, characterisation and properties of a Cr(III) diphenyl quaterpyridine complex towards CO₂ to CO reduction

Abstract

During the last decade, group VI metals have regained interest in synthesizing inorganic complexes that demonstrate catalytic properties for CO₂ reduction. However, these complexes often exhibit high overpotential and poor efficiency compared to those constructed with less-abundant metals. In this work, a polypyridinyl complex consisting of a Cr(III) centre coordinated to the tetradentate ligand 4′,4″-diphenyl-2,2′:6′,2″:6″,2‴-quaterpyridine and two monodentate chloride ligands ([Cr(dpqpy)Cl₂]⁺) was synthesized and characterized. Cyclic voltammetry performed under argon displayed three quasi-reversible one-electron systems in the cathodic part. The Zn(II) parent complex ([Zn(dpqpy)Cl₂]) was also synthesized and investigated to identify the nature of the reduction events. The three reduced species of [Cr(dpqpy)Cl₂]⁺ were generated by exhaustive electrolysis and characterized by ultraviolet-visible and electron paramagnetic resonance spectroscopies. Density functional theory calculations confirmed that the last two reductions are localized on the quaterpyridine ligand. Under a CO₂ atmosphere, with [Cr(dpqpy)Cl₂]⁺ dissolved in DMF in the presence of phenol, a catalytic current was observed during the last reduction at an onset potential of −1.72 V vs. Ag/AgNO₃ (10⁻² M). Exhaustive electrolysis of [Cr(dpqpy)Cl₂]⁺ performed at −1.85 V under CO₂ in DMF + 0.1 M TBAPF₆ in the presence of 1.25 M phenol led to the formation of CO with a faradaic efficiency of 60% after 2 h, with no H₂ being detected during the reaction time. Longer experiments led to a partial degradation of the complex.

---

Introduction

With a continuous increase in the consumption of fossil fuels as a result of anthropogenic activities, the emission of greenhouse gases such as carbon dioxide has led to a considerable change in the climate. It now becomes crucial to reduce its concentration in the atmosphere. One answer to the problem can be the transformation of the molecule after its capture. As CO₂ is inert and the most oxidized form of carbon, its conversion requires an electron source for reduction as well as selective catalysts.¹ Indeed, depending on the number of electrons and protons involved in the process, the CO₂ reduction reaction (CO₂RR) leads to several products, such as CO, HCOOH, H₂CO, CH₃OH, and CH₄, while the competitive reduction of protons occurs in the same potential range.² If the chemicals provided by CO₂ reduction are beneficial for the foreseeable future, this transformation remains a challenging process.³ Notably, it requires a large Gibbs free energy input of around 300 kJ mol⁻¹, which makes the reaction cumbersome. To circumvent this large energy barrier, a catalyst is necessary to open up routes for less energy-demanding reactions by stabilizing the intermediate transition states between the linear CO₂ molecules and the intended product.⁴ Among different molecular catalysts, polypyridyl complexes built with a noble metal centre such as Re(I), Ru(II), or Rh(III) have proven their efficiency in the 2e⁻ catalytic reduction of CO₂ since the 1980s.^5–7 Alternatives constructed with a first-row metal centre such as Mn(I), Ni(II), Co(III) or Fe(II) have also shown interesting behavior, and various examples have been reviewed in the literature.^8–11 Polypyridyl complexes of Cr are, however, still underexplored despite the large accessibility of this metal on Earth.^12,13 [Cr(bpy)(CO)₄] (bpy = 2,2′ bipyridine) was notably investigated by F. Hartl and co-workers, and showed moderate catalytic activity and quite high overpotential. The pentacoordinate species [Cr(bpy)(CO)₃]²⁻ formed upon two-electron reduction is hinted to be the active catalytic species for CO₂RR.¹⁴ In the mid-90s, Abruña and co-workers reported a Cr(III) complex designed with a bis-4′-vinylterpyridine ligand, which, upon electropolymerization onto a glassy carbon electrode, exhibited electrocatalytic activity under CO₂ towards the formation of formaldehyde (H₂CO) as a predominant reduction product. The formation of H₂CO was attributed to a cooperative effect between two or more metal centres in the electrodeposited film.¹⁵ More recently, Machan and co-workers proposed a series of Cr(III) complexes built with a tetradentate ligand that consists of a 2,2′-bipyridine with two phenolate moieties or a terpyridine core with a phenolate substituent.^16,17 These complexes are capable of selectively reducing CO₂ into CO, with phenol as a proton source, in acetonitrile. Following this approach, J. W. Wang et al. studied a quaterpyridinyl (qpy) complex of Cr(III) for CO₂ reduction. They reported an almost 100% faradaic efficiency for CO production in a DMF/phenol mixture with a turnover frequency of 86.6 s⁻¹; these values were obtained at −2 V vs. Fc⁺/Fc after 30 min of electrolysis.¹⁸ Quaterpyridine-based ligands have already been used to prepare other complexes with a first-row metal centre, such as Co(II),¹⁹ Cu(II)²⁰ or Fe(II),²¹ showing good activity towards the catalysis of CO₂ reduction. Indeed, as a good π-acceptor ligand resistant to oxidation, qpy leads to the formation of stable metal complexes in high oxidation state, and favours CO₂ reduction at relatively low potential values.²²

In such a context, we decided to synthesize a modified structure [Cr^III(dpqpy)Cl₂]⁺ (dpqpy = 4′,4″-diphenyl-2,2′:6′,2″:6″,2‴-quaterpyridine; Scheme 1) and investigate its electrochemical properties in the context of the CO2RR. The dpqpy ligand offers higher aromaticity than qpy, which could improve the electron storage capacity of the complex under reducing conditions. The synthesis of the ligand and the corresponding [Cr^III(dpqpy)(Cl)₂]Cl complex was carried out following the publications of Constable et al.^23,24 The related zinc complex [Zn^II(dpqpy)(Cl)₂] was also synthesized to study the electrochemical behaviour of the ligand with a redox-inactive metal centre. The study focuses on how the [Cr^III(dpqpy)(Cl)₂]⁺ complex works as an electrocatalyst in solution to convert CO₂ to CO. Spectroelectrochemistry and electron paramagnetic resonance (EPR) spectroscopy, both coupled with density functional theory (DFT) calculations, were also performed in order to characterise the different reduction states of this complex. Under a CO₂ atmosphere, the production of CO is confirmed and followed by GC-MS experiments under exhaustive electrolysis for 3 h to study the robustness of the complex.

Scheme 1 Synthesis procedure for the [Cr(dpqpy)(Cl)₂]⁺ complex.

Experimental section

Materials and compounds

All the starting materials and solvents for synthesis were obtained from commercial sources without further purification. CrCl₃·6H₂O (98%), ZnCl₂ (98%) and tetrabutylammonium hexafluorophosphate (TBAPF₆) were purchased from Sigma Aldrich. dpqpy was synthesized following published procedure.²⁴ A 3 mm diameter (0.071 cm² area) glassy carbon disk electrode was used in all cyclic voltammetry (CV) and rotating disk electrode (RDE) experiments. For bulk electrolysis, a porous carbon foam electrode (Duocel ERG aerospace, 80 dpi) was used.

Instrumentation

Ultraviolet-visible (UV-VIS) and infrared (IR) spectroscopies. UV-VIS spectra of complexes in solution were recorded on a Cary 50 Probe UV-VIS spectrophotometer (VARIAN) or on a Zeiss MC5601 spectrophotometer, using either a 1 mm quartz immersion probe (Hellma Inc.) or a quartz cuvette of optical path length l = 1 cm. IR spectra were recorded using a “Golden Gate” ATR system. This system features a brazed diamond ATR crystal and a signature “Golden Gate” sample loading bridge. The complex samples in solid form were loaded on this bridge, which contains the required mirrors for beam transmission.

Electrochemistry. Cyclic voltammetry scans were performed on a Model SP-300 potentiostat (BioLogic) in a standard three-electrode cell (glassy carbon working electrode; Ag/AgNO₃ 10⁻² M in DMF + 0.1 M TBAPF₆ as the reference electrode; platinum as the auxiliary electrode) (Fig. S1). Using this reference electrode, the potentials can be converted to Fc⁺/Fc by subtracting 87 mV or SHE by adding 542 mV.²⁵ Experiments under argon were performed in a glovebox (Jacomex).

Electrocatalysis. Experiments were carried out at room temperature in a CO₂-saturated environment in a sealed conventional three-electrode cell. The reference and counter electrodes were Ag/AgNO₃ 0.01 M and a Pt plate, respectively; carbon foam (Duocel ERG aerospace, 80 dpi) was used as the working electrode material. In a normal setting with 3 electrodes and 10 mL solution, the headspace volume was measured to be 140 mL (Fig. S1). During the experiments, at each time interval, a 100 μL sample was taken from the headspace gas using a gas-tight injection syringe. Gas products were analyzed on a PerkinElmer Clarus 590 gas chromatography (GC) instrument equipped with two columns, a Restek Hayesep Q (L = 2 m, diam. = 1 mm), and a Restek molecular sieve 5A (L = 2 m, diam. = 1 mm), using argon as the vector gas. In GC, thermal conductivity and flame ionisation detection (TCD, FID) were used for H₂ (TCD), CO (FID), and CH₄ (FID). The limits of quantification are estimated to be 80 ppm for H₂ and 30 ppm for CO and CH₄. Calibration was performed with a gas mixture standard H₂/CO/CH₄ 1% each in CO₂. At the end of the experiments, the solution was analysed by manual injection of 20 µL on a Metrohm Eco ionic chromatograph equipped with a Metrosep A Supp 19 column in which Na₂CO₃ 8 mM/NaHCO₃ 0.25 mM in water was used as a mobile phase for formate detection by conductivity. The uncertainties of the faradaic efficiency were estimated by comparison of the results of two similar experiments.

Mass spectroscopy. High-resolution mass spectrometry (HRMS) spectra were recorded on an LTQ Orbitrap XL Thermo Scientific instrument in electrospray ionisation (ESI) positive mode.

EPR spectroscopy. X- and Q-band EPR spectra were recorded with a Bruker EMX instrument, equipped with an ER-4192 ST Bruker cavity and ER-4121 VT for the X-band experiments at 100 K and an ER5106 QTW Bruker cavity and an Oxford Instruments ESR-900 continuous-flow helium cryostat for the Q-band experiments at 4.5 K.

DFT calculations

Full geometry optimizations were performed with the Gaussian 9 program,²⁶ by using the B3LYP exchange–correlation functional.^27,28 Dispersion was taken into account by using the Grimme D3 correction.²⁹ Two basis sets were investigated for the Cr, C, H, and N atoms: the double zeta 6-31g*,³⁰ as well as the triple zeta 6-311+G(2d,p).^31,32 Frequency calculations were systematically performed in order to ensure that the optimized structure corresponds to a minimum and not to a saddle point. The solvent (DMF) was considered by using a polarized continuum model.³³ The energies discussed in the text correspond to the sum of electronic and thermal free energies, expressed in kcal mol⁻¹, at 298 K.

Synthesis

[Cr(dpqpy)(Cl)₂]Cl. First, 0.13 mmol (0.027 g) of CrCl₃·6H₂O was mixed with 0.1 mmol (0.046 g) of 4′,4″-diphenyl-2,2′:6′,2″:6″,2‴-quaterpyridine in 15 mL of ethanol; this solution was refluxed for 30 h under N₂. After cooling, the solution was filtered, resulting in a greenish-grey solid product that was then rinsed with ethanol and dried using diethyl ether. Yield: 62% (38.8 mg). FT-IR: 1603s, 1545s, 1487w, 1461s, 1407s, 1242s, 1049w, 1024s, 886w, 830w, 787w, 769s, 731s, 689s cm⁻¹. HRMS: m/z [M]⁺ = 582.06754 (calculated for C₃₂H₂₂N₄Cl₂Cr: 582.06765). Elem. anal. calculated for C₃₂H₂₂N₄Cl₃Cr, 3H₂O: C: 56.95%; H: 4.18%; N: 8.30% found: C: 56.64%; H: 3.91%; N: 8.29%.

[Zn(dpqpy)(Cl)₂]. First, 7.5 × 10⁻² mmol (0.010 g) of ZnCl₂ was mixed with 5.2 × 10⁻² mmol (0.024 g) of 4′,4″-diphenyl-2,2′:6′,2″:6″,2‴-quaterpyridine in 20 mL of ethanol; the solution was refluxed for 2 h under Ar. After cooling, the solution was filtered and evaporated, resulting in a white solid product (12.0 mg). FT-IR: 1602vs, 1555s, 1486w, 1459s, 1402s, 1247s, 1162w, 1015s, 891w, 795w, 764s, 733w, 682w cm⁻¹. HRMS: m/z [M − 2Cl]²⁺ = 263.0561 (calculated for C₃₂H₂₂N₄Zn: 263.05625).

Results and discussion

Synthesis and photophysical characterisation of [Cr(dpqpy)(Cl)₂]⁺

The chromium [4′,4″-diphenyl-2,2′:6′,2″:6″,2‴-quaterpyridine] bis-chloro complex, with chloride as the counter anion ([Cr(dpqpy)(Cl)₂]Cl), was synthesized using hydrated chromium(III) chloride salt and diphenyl quaterpyridine (dpqpy) as starting materials, as depicted in Scheme 1, following an adapted procedure from the literature.²³ Given the inert nature of Cr³⁺, the complexation procedure was found to be kinetically slow. After a reflux reaction in ethanol for 30 h under nitrogen, a grey solid powder was collected by filtration. The synthesis was confirmed by HRMS, showing a mass peak and isotopic pattern in accordance with the expected formula (Fig. S2 and S3).

The complex is sparingly soluble in CH₃CN, THF or CH₂Cl₂ and more soluble in DMF, providing a faint brown-gold solution. The UV-VIS spectrum displayed a strong absorption band in the UV region corresponding to the ligand-centred allowed π → π* transition (λ_max. = 323 nm, ε = 4 × 10⁴ M⁻¹ cm⁻¹). Additionally, a weak absorption in the visible is detected with a maximum located around 540 nm (ε = 250 M⁻¹ cm⁻¹). This band is associated with the forbidden metal-centred transitions (Fig. 1).

Fig. 1 Electronic absorption spectrum of [Cr(dpqpy)(Cl)₂]⁺ in DMF; inset: zoomed in the visible region. ε: molar extinction coefficient.

The Zn(II) analogue complex was also synthesized in order to further elucidate the electrochemical properties of the Cr complex (see further in the text). The complexation occurs in 2 h, starting with ZnCl₂ and dpqpy in ethanol. The solution remains colourless in the course of the reaction as Zn(II) has a d¹⁰ configuration.

Electrochemistry under argon

All the compounds were characterized by cyclic voltammetry (CV) in an argon-saturated DMF solution with 0.1 M TBAPF₆ as a supporting electrolyte. While the CV of the free ligand (dpqpy) shows two chemically irreversible reduction waves in the cathodic part at cathodic peak potentials of E_pc = −2.36 V and −2.69 V vs. Ag/AgNO₃ (Fig. S4), [Zn(dpqpy)(Cl)₂] exhibits two quasi-reversible monoelectronic systems at E_1/2 (V) = −1.56 V (with a peak separation of ΔE_p = 75 mV) and −1.82 V (ΔE_p = 130 mV) (Fig. S5). As Zn(II) is redox inert, the 2 reduction events are centred on the ligand. [Cr(dpqpy)(Cl)₂]Cl shows in the same potential range a series of three quasi-reversible systems at E_1/2 (V) = −1.03 V (ΔE_p = 70 mV), −1.48 V (ΔE_p = 110 mV) and −1.79 V (ΔE_p = 70 mV) (Fig. 2). The two last ones appear at a closer potential than the reduction systems of [Zn(dpqpy)(Cl)₂] and can be attributed to the stepwise reduction of the ligand, whereas the first reduction event at E_1/2 (V) = −1.03 V does not appear in [Zn(dpqpy)(Cl)₂] and may be attributed to the reduction of the Cr(III) center into a Cr(II) one. EPR and DFT calculations were performed to confirm the attribution (see further in the text).

Fig. 2 Cyclic voltammetry scan of 1 mM [Cr(dpqpy)(Cl)₂]Cl in DMF/0.1 M TBAPF₆ under argon atmosphere on a GCE at a scan rate of 0.1 V s⁻¹: above in the anodic direction, below in the cathodic direction showing three different reduction waves recorded from −0.5 V to −1.43 V, −0.5 V to −1.78 V and −0.5 V to −2.18 V. Electrode surface 0.071 cm².

In the anodic part, [Cr(dpqpy)(Cl)₂]Cl exhibits a chemically irreversible system at an anodic peak potential of E_pa = 0.58 V, attributed to the oxidation of the chloride counter anion in solution.³⁴ Indeed, a related Cr(III) polypyridine complex such as [Cr(ttpy)₂](ClO₄)₃ (ttpy = 4′-(4-methylphenyl)-2,2′:6,2″-terpyridine) does not exhibit any oxidation wave in this potential range,³⁵ whereas tetraethylammonium chloride [N(Et)₄]⁺,Cl⁻ salt in DMF shows an irreversible oxidation wave at 0.58 V (Fig. S6). Rotating disk electrode (RDE) experiments show that this oxidation requires the same number of electrons as the first reduction step of the complex (Fig. S7). It indicates that the two chloro ligands are not exchanged with solvent molecules. Exhaustive bulk electrolysis performed at E_pa = 0.6 V consumed one electron per complex. The CV recorded after this exhaustive electrolysis is similar to the initial CV of the complex in the cathodic part, showing the three one-electron reduction events at similar potential values (Fig. S8). This behaviour is in accordance with the oxidation of the counter anion.

CV measurements recorded in reduction at different scan rates show a correlation between the square root of scan rate and the peak current intensity i_p (Fig. S9). The linear trend of the Randles–Ševčík plot is common to all the three redox processes and indicates that these three redox waves are diffusion-controlled processes, proving that the complex does not stick to the surface of the GC electrode. Using the Randles–Ševčík equation at 20 °C (eqn (1)), the diffusion coefficient of [Cr(dpqpy)(Cl)₂]⁺ is estimated to be 2.82 × 10⁻⁶ cm² s⁻¹

I_p = 2.69 × 10⁵ × A × D^1/2 × C × v^1/2 (1)

In this equation, A is the surface of the electrode in cm², C the concentration in mol cm⁻³ and v the scan rate in V s⁻¹.

To investigate the stability of the different reduction states of the complex in solution, exhaustive bulk electrolysis was performed at each reduction potential of [Cr(dpqpy)(Cl)₂]⁺. Fig. 3 shows the CV and RDE experiments obtained after electrolysis. RDE shows that the redox processes have a similar current magnitude prior to and after the first two electrolysis experiments performed at −1.07 and −1.55 V. After the third electrolysis performed at −1.83 V, the signal exhibits a lower intensity compared to the others due to a lower solubility of the fully reduced species. The CV scans recorded after each reduction are depicted in Fig. 3B. After the first electrolysis experiment at −1.07 V, the CV scan of the solution presents three redox waves at a similar position to that of the initial solution. It proves the full reversibility of the first electron transfer process. However, after the reduction performed at −1.55 V, the CV scan shows the emergence of two new waves, each marked by an asterisk, indicating that the complex undergoes a slow chemical transformation under such a reductive condition.

Fig. 3 (A) Cathodic (solid lines) and anodic (dash–dot lines) cyclic voltammetry scans prior to and after electrolysis of 1 mM [Cr(dpqpy)(Cl)₂]⁺ complex under argon on a rotating disk electrode (rate of rotation: 200 rpm, scan rate: 10 mV s⁻¹, electrode area: 0.071 cm²); (B) cathodic CV scans after electrolysis of 1 mM [Cr(dpqpy)(Cl)₂]⁺ complex under argon on a glassy carbon electrode in DMF/0.1 M TBAPF₆ at a scan rate of 0.1 V s⁻¹; electrode area 0.071 cm².

To characterize the nature of the transformation, CV experiments were performed on a wider potential window, including the anodic part (Fig. S10). Prior to electrolysis, the CV scan shows the oxidation peak at 0.58 V, attributed to the oxidation of the chloride counterion in solution. After the first electrolysis process, at −1.07 V, the current magnitude corresponding to the chloride oxidation process was not modified; however, upon electrolysis at the second reduction peak (−1.55 V), the increase in current density suggests an accumulation of released chloride ions in solution (Fig. S10). This should account for the discoordination of the Cl⁻ linked to the metal centre. This oxidation process that emerges after the second reductive electrolysis was deduced to be a 2-electron oxidation process per complex by RDE voltammetry (Fig. S10B). It is then attributed to one Cl⁻ counter ion and one coordinated Cl⁻ ion released in solution after electrolysis at −1.55 V. The bis-reduced complex is then assumed to undergo a transformation, associated with the release of one chloride ion, resulting in the formation of [Cr^III(dpqpy²⁻)(Cl)(DMF)]. This new species can account for the appearance of a new reduction peak (≈−1.95 V) in the cathodic CV scan (Fig. 3B).

Electronic absorption spectra were recorded after the three successive exhaustive electrolysis steps of [Cr(dpqpy)(Cl)₂]⁺ performed in DMF, in order to characterize the different redox states of the complex. Prior to electrolysis, one strong absorption band in the UV region and a weak absorption band at 540 nm were observed, as discussed previously (Fig. 4).

Fig. 4 Ultraviolet-visible-near infrared absorption spectra of 1 mM [Cr(qpy)(Cl)₂]⁺ in DMF/0.1 M TBAPF₆: (a) initial solution, (b) after exhaustive electrolysis at −1.03 V, (c) after exhaustive electrolysis at −1.49 V, and (d) after exhaustive electrolysis at −1.79 V. ε: molar extinction coefficient.

The mono-reduced [Cr(dpqpy)(Cl)₂] species exhibits in the visible range two absorption bands at 528 nm (ε = 4200 M⁻¹ cm⁻¹) and 628 nm (ε = 3500 M⁻¹ cm⁻¹), plus two other weak shoulder peaks around 713 nm (ε = 2100 M⁻¹ cm⁻¹) and 773 nm (ε = 1700 M⁻¹ cm⁻¹). Further electrolysis at −1.55 V led to the formation of a ligand-reduced species whose absorption bands extend to the whole visible part of the spectrum, reaching a maximum at 601 nm (ε = 8000 M⁻¹ cm⁻¹). In the course of electrolysis at −1.55 V, the solution colour changed from faint golden-brown to deep purple. The final tris-reduced species is less soluble and might have decomposed. The solution exhibits an absorption maximum at around 568 nm, as shown in Fig. 4.

EPR characterisation

Electron paramagnetic resonance (EPR) spectroscopy was performed to confirm the assignment of the added electron upon successive reduction of [Cr^III(dpqpy)(Cl)₂]⁺. Fig. 5 presents EPR spectra recorded at the X-band frequency (ν ≈ 9.65 GHz) at 20 K prior to and after successive bulk reductive electrolysis steps.

Fig. 5 Electron paramagnetic resonance spectra of 1 mM [Cr^III(dpqpy)(Cl)₂]⁺ in DMF/0.1 M TBAPF₆ prior to (A) and after electrolysis at −1.07 V (B) at −1.55 V (C) at −1.83 V (D).

The EPR spectrum of [Cr^III(dpqpy)(Cl)₂]⁺ (Fig. 5A) covers a large spectral window, with resonances at g = 5.7, 5, 2.6, 1.7 and 1.4. It is typical of a high-spin (S = 3/2) Cr^III complex.³⁶ By taking into consideration a one-electron reduction, the putative spin states are S = 2, 1 or 0. The spectrum depicted in Fig. 5B shows that the species generated upon one-electron reduction is X-band EPR silent. A ground spin state (S = 0) is very unlikely, given the geometry; thus, we propose that the one-electron reduced species is an integer spin system, but the zero-field splitting parameters are too large to allow for its detection at the X-band frequency.

Accordingly, the two-electron reduction of [Cr^III(dpqpy)(Cl)₂]⁺ results in the growth of a strong signal at g = 1.98, which is reminiscent of an (S = 1/2) spin state (Fig. 5C). The g value is close to 2, indicative of a significant radical character on the complex. Finally, the three-electron reduction is accompanied by a quenching of the S = 1/2 resonance, concomitant with the appearance of minor resonances attributed to unidentified decomposition products (Fig. 5D).

DFT calculations

To gain insights into the electronic structures of the electron-transferred series, we performed density functional theory (DFT) calculations, using the B3LYP hybrid functional. Full computational details, including Cartesian coordinates, energetic analyses, and selected Mulliken spin population data (B3LYP, D3-B3LYP, D3-B3LYP/PCM and D3-B3LYP/PCM with a triple zeta basis set), are provided in the SI (Fig. S11–S29 and Tables S1–S6). Table 1 summarizes the most stable electronic structures of the complex in its different reduction states. The trends in the energy ranking for all spin states considered, as well as their formulation, were similar when using the B3LYP, D3-B3LYP, and D3-B3LYP/PCM functionals (with a double zeta 6-31g* basis set), as well as D3-B3LYP/PCM with a split valence triple zeta basis set (6-311+G(2d,p)) (Table S2). For clarity, only results obtained using the D3-B3LYP/PCM/6-31g* approach are presented here, unless otherwise stated. As the three-electron reduced species was found to decompose during electrolysis, it was not included in the calculations.

Table 1 Electronic structures and Mulliken spin populations of the Cr atom and the central N atoms in the dpqpy ligand for chromium complexes in their ground spin state^a

Complex	Spin state	Mull. spin pop. at Cr/N	Formulation	Metal ion geometry
a From a D3-B3LYP/PCM/6-31g* calculation. The data for all the oxidation states and spin states, including the free energy differences between each spin state for a given complex, are detailed in SI (BS = broken symmetry).
[Cr(dpqpy)(Cl)2]^+	Quartet	3.07/−0.05	Cr^III/qpy	Octahedral
[Cr(dpqpy)(Cl)2]^0	BS triplet	2.97/−0.21	Cr^III/qpy˙^−	Octahedral
[Cr(dpqpy)(Cl)2]^−	BS doublet	2.89/−0.25	Cr^III/qpy^2˙^2−	Octahedral
[Cr(dpqpy)(Cl)]^0, DMF	BS quartet	3.67/−0.21	Cr^III/qpy^2˙^2− ↔ Cr^II/qpy˙^−	Square pyramidal

---

Geometry optimization of the quartet [Cr(dpqpy)(Cl)₂]⁺ yielded an octahedral metal ion (Fig. 6a). A comparison of structural parameters within the coordination sphere of the related complex based on unsubstituted quaterpyridine revealed a very good agreement between experiment and theory.¹⁸ The two distinct Cr–N and the Cr–Cl bond distances were indeed reproduced with deviations that do not exceed 0.01, 0.03, and 0.06 Å, respectively. Furthermore, frequency and TD-DFT calculations accurately predict the experimental IR and absorption spectra, validating our methodology (Fig. S22 and S27). The Mulliken spin population analysis revealed a value of 3.07 on the Cr atom, confirming that the spin density is predominantly localized on a Cr(III) centre. This value will be taken as a reference for assigning the metal's formal oxidation state in the electron-transferred series.

Fig. 6 Spin density plot from a B3LYP/6-31g* calculation. Isosurface shown with 0.005410 contour value: (a) [Cr(dpqpy)(Cl)₂]⁺ (quartet); (b) [Cr(dpqpy)(Cl)₂]⁰ (broken symmetry triplet); (c) [Cr(dpqpy)(Cl)₂]⁰ (quintet); (d) [Cr(dpqpy)(Cl)₂]²⁻ (broken symmetry doublet); (e) [Cr(dpqpy)(Cl)₂]²⁻ (sextet).

For the octahedral one-electron reduced species [Cr(dpqpy)(Cl)₂], both triplet (S = 1) and quintet (S = 2) spin states were considered (Table S2), and calculation predicts the first to be the ground spin state. The calculated Mulliken spin populations on the metal centre are 2.97 and 3.08 for the triplet and quintet, respectively. The second value is very close to that computed for [Cr(dpqpy)(Cl)₂]⁺, indicative of a Cr(III) metal ion, while the first is slightly lower, consistent with the lower spin state. Thus, both states mostly correspond to octahedral Cr(III)-ligand radical systems, denoted as [Cr(dpqpy˙⁻)(Cl)₂]. Notably, the ligand radical's spin density is primarily localized on the central bipyridine subunit in a symmetric manner, with negligible delocalization onto the phenyl rings. The broken-symmetry (3,1) nature of the triplet state was further confirmed by a negative Mulliken spin population on the ligand and a positive value on the Cr atom, whereas the quintet exhibits positive values only (Fig. 6b and c). This conclusion matches that reached by Wieghardt et al. on one-electron reduced octahedral Cr(III) bis(terpyridine) complexes³⁷ but contrasts with that recently reported for a related one-electron reduced quaterpyridine Cr(III) complex (metal-centred reduction), lacking axial ligands. While no electrochemical evidence for ligand dissociation upon one-electron reduction was found in the present series, we nonetheless reoptimized a structure in which the metal centre is formally uncoordinated in the axial positions ([Cr(dpqpy)]²⁺). Under these conditions, the calculations converged towards a quintet square-planar Cr(II) species (Table S2). This highlights the strong sensitivity of the electronic structure to the coordination number of the metal centre in this family of complexes and possibly to the theoretical approach used. Interestingly, TD-DFT calculations on both triplet [Cr(dpqpy)(Cl)₂] and quintet [Cr(dpqpy)]²⁺ revealed that only the first leads to multiple absorption bands in the 400–800 nm range in accordance with the experimental absorption spectrum recorded in DMF (Fig. S27–S29). The most intense transitions are predicted at 647 and 495 nm for [Cr(dpqpy)(Cl)₂], and are attributed to intra-ligand charge transfers (Fig. S23).

Further reduction yields the monoanion [Cr(dpqpy)(Cl)₂]⁻, which could adopt doublet, quartet, or sextet spin states, all representing half-integer spin systems. The Mulliken spin population on the Cr centre is 2.89, 2.96 and 3.09, respectively, again arguing for a mostly ligand-centred redox process, the last value comparing well with [Cr(dpqpy)(Cl)₂]⁺ (Table S5). The doublet state is a broken symmetry (3,2) spin state, characterized by an antiferromagnetic interaction between a triplet diradical ligand (dpqpy²˙²⁻) and a Cr(III) metal ion (S = 3/2) (Fig. 6d), and it is the ground state. This marked radical character is consistent with EPR spectroscopy, which reveals a low anisotropy of the g-tensor and a low spin state.

The sextet state (+7.9 kcal mol⁻¹ relative to the doublet, Fig. 6d) results from ferromagnetic coupling, while the quartet lies intermediate and is spin-contaminated due to its proximity to the sextet. Note that attempts to converge a Cr(III) complex featuring a closed-shell (dpqpy²⁻) ligand, especially using the fragment approach, failed, highlighting the propensity of this ligand framework to support diradicals. This hypothesis is further supported by calculations on both the free ligand (dpqpy²⁻) (Fig. S11) and its Zn(II) complex (Fig. S12), which point to a diradical ground spin state of the ligand, with geometrical specificities (tilt or umbrella shape of the bipyridine core). This is also consistent with prior studies on terpyridine-based chromium complexes by Wieghardt et al., who demonstrated that the terpyridine dianion is stabilized under its diradical state (relative to the closed-shell state) by magnetic coupling.³⁷

It is worth noting that electrochemistry suggests that two-electron reduction of [Cr^III(dpqpy)(Cl)₂]⁺ may be accompanied by the release of one chloride ion in solution. Thus, the electronic structure of the neutral species [Cr(dpqpy)(Cl)(DMF)] was also computed, again under the doublet, quartet and sextet spin states (Table S2). The doublet and quartet are close in energy (ΔE = 3.3 kcal mol⁻¹), the second being the most stable spin state, while the sextet is much higher in energy, 13.7 kcal mol⁻¹ above the quartet. The doublet state is characterized by an octahedral Cr(III) centre, harbouring a Mulliken spin population of 2.83, with negative spin populations on the coordination bipyridine nitrogens (−0.24 and −0.14). Thus, it is formulated as a Cr^III(qpy²˙²⁻) radical species, similar to the triplet [Cr(dpqpy)(Cl)₂]⁻. The sextet is also better described as Cr^III(qpy²˙²⁻), although it involves ferromagnetic interactions. In contrast, in the quartet state, the Mulliken spin population is 3.67 on the Cr atom, e.g., between high-spin Cr(II) and Cr(III), while the DMF ligand is released during the optimization, affording a square pyramidal metal centre (Table S5). This suggests that the energy gap between the Cr^III(qpy²˙²⁻) and Cr^II(qpy˙⁻) valence tautomers is reduced, due to the square pyramidal geometry and the less efficient metal charge compensation (due to the lack of one anionic ligand), both favouring a Cr(II) centre. This Cr(II) stabilization becomes even more pronounced when optimization is performed on a species lacking axial ligands ([Cr(dpqpy)]⁺): the broken-symmetry quartet state is now formally a Cr(II) centre (Mulliken spin population of 3.94) coordinated to a ligand-based radical, and it is the ground state, lying just 0.5 kcal mol⁻¹ above the sextet. Notably, TD-DFT calculations consistently predict a rich visible absorption spectrum, irrespective of the nature of the axial ligand, and even in its absence (Fig. S27–S29). The assignment of the most intense bands is provided in SI (Fig. S26).

Electrocatalytic reduction of CO₂

Further electrochemical analyses of [Cr(dpqpy)(Cl)₂]⁺ were carried out under a CO₂-saturated atmosphere to investigate its electrocatalytic properties. As shown in Fig. 7, the CV scan under CO₂-saturated conditions resulted in a current increase at an onset potential of −1.77 V. This is accompanied by the loss of the reversibility behaviour of the third reduction waves as a more probable consequence of an inner-sphere mechanism for the catalysis.

Fig. 7 Cathodic scan cyclic voltammograms of [Cr^III(dpqpy)(Cl)₂]⁺ under argon (a), CO₂ (b) and of a blank CO₂-saturated electrolyte solution (c) on GCE in DMF/0.1 M TBAPF₆ at a scan rate of 0.1 V s⁻¹ (electrode area 0.071 cm²).

To improve the current response of the complex under CO₂, a weak acid, specifically phenol (PhOH), acting as a source of protons, has been added in solution (Fig. 8).

Fig. 8 Cathodic cyclic voltammetry scans of a 1 mM [Cr(dpqpy)(Cl)₂]⁺ complex under argon (a) and under CO₂-saturated conditions without (b) and with varying concentrations of [PhOH] (c) (1 M, 1.25 M, 1.5 M and 1.75 M) on GCE in DMF/0.1 M TBAPF₆ at a scan rate of 0.1 V s⁻¹.

A significant onset potential shift of ∼50 mV was noted upon addition of 1 M of PhOH; the second reduction wave is also shifted to a less negative potential and suggests that PhOH interacts after the first reduction event of the complex. A further increase in the phenol concentration did not show a noticeable difference in the onset potential value; it mainly resulted in an increase in the catalytic peak current. The use of PhOH gave the best current response among those observed for other weak acids considered, such as water or methanol in DMF (Fig. S30 and S31, respectively).

Gas chromatography measurements were performed to analyse the gas produced during 4 h of exhaustive electrolysis of 1 mM [Cr(dpqpy)(Cl)₂]⁺ in the presence of different concentrations of phenol at a bias of −1.85 V. These analyses confirmed the formation of CO, with, in most cases, no H₂ detected (cf. Fig. S32 as an example of the chromatogram obtained). Tables S7–S10 gather the results obtained during the bulk electrolysis. A small amount of H₂ is only observed when using a concentration of PhOH of 1.75 M and after 2 h of electrolysis (Table 2). Fig. 9 shows the evolution of CO throughout the 4 h electrolysis measurements with different concentrations of PhOH in solution. In all cases, the production of CO increased during the first 2 h and then stabilized for the remaining time. Concomitantly, we noticed in all cases a decrease in the magnitude of the cathodic catalytic current after 2 h (Fig. S33). This suggests a degradation of the complex or a poisoning of the electrode surface, rather than the consumption of CO towards more elaborate products. Electrolysis in the presence of 1.25 M PhOH resulted in a higher amount of produced CO among all varied concentrations, yielding a 60% ± 10% faradaic efficiency after 2 h (Table 2). After the 4 h electrolysis, no formation of HCOO⁻ in solution was detected by ionic chromatography.

Fig. 9 Evolution of the amount of CO produced during electrolysis of 1 mM [Cr(dpqpy)(Cl)₂]⁺ complex at −1.85 V Ag/10⁻² M AgNO₃ under CO₂-saturated conditions with varying concentrations of PhOH (a: 1 M, b: 1.25 M, c: 1.5 M, d: 1.75 M) on a carbon foam electrode in DMF/0,1 M TBAPF₆.

Table 2 Faradaic efficiency (F.E.) for the production of CO and H₂ during the electrochemical CO₂ reduction with a 1 mM [Cr(dpqpy)(Cl)₂]⁺ complex as the catalyst in the presence of a varying concentration of phenol (PhOH) after a 2 h electrolysis at −1.85 V vs. Ag/AgNO₃ on a carbon foam electrode in CO₂-saturated DMF/0,1 M TBAPF₆. Uncertainties are estimated to ±10%

[PhOH]	Q (C)	F.E. CO (%)	F.E. H2 (%)
1 M	−12.8	18	—
1.25 M	−11.4	60	—
1.5 M	−10.3	55	—
1.75 M	−11.2	35	1

---

The foot of the wave analysis (FOWA),^38,39 following eqn (2)–(4), was performed on cyclic voltammograms of a 1 mM solution of [Cr(dpqpy)(Cl)₂]⁺ in DMF containing 0.1 M TBAPF₆, under CO₂-saturated conditions, both in the absence and presence of 1.25 M PhOH. This analysis was used to estimate the catalytic rate constant (k_obs), as well as the turnover frequency (TOF) and turnover number (TON) after 2 hours under both conditions, without and with the addition of PhOH.

TON = TOF × t (4)

In these equations, v is the scan rate, E° = −1.79 V, and i and i_p⁰ are the peak current under catalytic conditions and under argon, respectively.

The FOWA plots (Fig. S34) yield k_obs values of 0.62 s⁻¹ and 8.74 s⁻¹ in the absence and presence of PhOH (1.25 M), respectively, indicating a 14-fold increase in the reaction rate upon addition of the proton source. The TOF at E = −1.85 V is estimated to be 0.56 s⁻¹ without PhOH and 8 s⁻¹ with PhOH. In the presence of PhOH, the TOF is comparable to values reported for other Cr(III)-based catalysts;^18,40 however, it remains approximately one order of magnitude lower than that obtained for [Cr(qpy)(Cl)₂]⁺, indicating that diphenyl substitution has a detrimental effect on catalytic performance. After 2 h, the maximum TON is estimated to be 5.7 × 10⁴, significantly lower than the value of 1.56 × 10⁵ reported for [Cr(qpy)(Cl)₂]⁺, suggesting reduced catalyst stability.¹⁸

Conclusion

In the present work, the synthesis and reactivity of a diphenyl quaterpyridine Cr(III) complex that displayed substantial catalytic activity towards CO₂ reduction have been described. Catalysis is initated only after the complex has undergone three succesive one-electron reduction step. Its efficiency is improved in the presence of 1.25 M phenol as a proton source in DMF, reaching a maximum of 60% ± 10% faradaic efficiency for the formation of CO after 2 h. The complex undergoes degradation after 2 h of electrolysis, resulting in a cessation of the catalytic activity. It confirms, however, the interest of Cr(III)-based complexes in designing new molecular catalysts. Characterization of the complex under argon by electrochemistry and EPR, supported by DFT calculations, shows that the first reduction event of the complex is more probably centred on the metal, whereas the second and third reductions are localized on the ligand, highlighting the propensity of this ligand framework to support diradicals. This suggests a sufficient degree of electronic coupling between the metal centre and the redox-active ligand to enable efficient reactivity under CO₂.^41,42 Indeed, the electron storage ability of the ligand associated with an open coordination site on the metal is required for the two-electron, two-proton CO₂ reduction process. Further work needs to be carried out to deeply investigate the catalytic cycle and then to improve the long-term stability of such a complex upon catalysis. One way to improve stability is to anchor the complex onto the electrode surface, either by forming a polymer or by creating a self-assembled monolayer following specific modifications on the quaterpyridine ligand. Conversely, the use of less sterically hindered quaterpyridines appears to enhance catalytic performance, particularly in terms of faradaic efficiency and TOF.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6dt00748a.

Acknowledgements

The authors thank the ANR project “Solar-Driven Chemistry” (ANR-20-SODR-0001-01) and Labex Arcane (ANR-17-EURE-0003) for financial support, and the referees for their comments to improve the quality of the work. J. C. thanks Alain Deronzier for fruitful discussions. Part of the work was performed using the ICMG characterization platform and mass spectrum facilities at Université Grenoble Alpes.

References

1. R. Francke, B. Schille and M. Roemelt, Chem. Rev., 2018, 118, 4631–4701.
2. D. D. Zu, J. L. Liu and S. Z. Qiao, Adv. Mater., 2016, 28, 3423–3452.
3. A. M. Appel, J. E. Bercaw, A. B. Bocarsly, H. Dobbek, D. L. DuBois, M. Dupuis, J. G. Ferry, E. Fujita, R. Hille, P. J. A. Kenis, C. A. Kerfeld, R. H. Morris, C. H. F. Peden, A. R. Portis, S. W. Ragsdale, T. B. Rauchfuss, J. N. H. Reek, L. C. Seefeldt, R. K. Thauer and G. L. Waldrop, Chem. Rev., 2013, 113(8), 6621–6658.
4. B. Kumar, M. Llorente, J. Froehlich, T. Dang, A. Sathrum and C. P. Kubiak, Annu. Rev. Phys. Chem., 2012, 63, 541–569.
5. J. Hawecker, J. M. Lehn and R. Ziessel, Helv. Chim. Acta, 1986, 69, 1990–2012.
6. H. Ishida, K. Tanaka and T. Tanaka, Organometallics, 1987, 6, 181–186.
7. C. M. Bolinger, N. Story, B. P. Sullivan and T. J. Meyer, Inorg. Chem., 1988, 27, 4582–4587.
8. M. Stanbury, J.-D. Compain and S. Chardon-Noblat, Coord. Chem. Rev., 2018, 361, 120–137.
9. K. E. Dalle, J. Warnan, J. J. Leung, B. Reuillard, I. S. Karmel and E. Reisner, Chem. Rev., 2019, 119, 2752–2875.
10. J.-W. Wang, W.-J. Liu, D.-C. Zhong and T.-B. Lu, Coord. Chem. Rev., 2019, 378, 237–261.
11. E. A. Palade, R. Gobetto and C. Nervi, Inorg. Chim. Acta, 2024, 566, 122029.
12. K. A. Grice, Coord. Chem. Rev., 2017, 336, 78–95.
13. M. E. Moberg and C. W. Machan, Acc. Chem. Res., 2024, 57, 2326–2335.
14. J. Tory, B. Setterfield-Price, R. A. W. Dryfe and F. Hartl, ChemElectroChem, 2015, 2, 213–217.
15. J. A. Ramos Sende, C. R. Arana, L. Hernandez, K. T. Potts, M. Keshevarz-K and H. D. Abruña, Inorg. Chem., 1995, 34(12), 3339–3348.
16. S. L. Hooe, J. M. Dressel, D. A. Dickie and C. W. Machan, ACS Catal., 2020, 10(2), 1146–1151.
17. A. G. Reid, S. L. Hooe, J. J. Moreno, D. D. Dickie and C. W. Machan, Inorg. Chem., 2022, 61, 16963–16970.
18. J.-W. Wang, Z.-M. Luo, G. Yang, M. Gil-Sepulcre, S. Kupfer, O. Rudiger and G. Ouyang, Proc. Natl. Acad. Sci. U. S. A., 2024, 121(14), e2319288121.
19. K.-M. Lam, K.-Y. Wong, S.-M. Yang and C.-M. Che, J. Chem. Soc., Dalton Trans., 1995, 1103–1107.
20. Z. Guo, F. Yu, Y. Yang, C. F. Leung, S.-M. Ng, C.-C. Ko, C. Cometto, T.-C. Lau and M. Robert, ChemSusChem, 2017, 10, 4009–4013.
21. C. Cometto, L. Chen, P. K. Lo, Z. Guo, K.-C. Lau, E. Anxolabéhère-Mallart, C. Fave, T.-C. Lau and R. Robert, ACS Catal., 2018, 8, 3411–3417.
22. L. Chen, G. Chen, C.-F. Leung, C. Cometto, M. Robert and T.-C. Lau, Chem. Soc. Rev., 2020, 49, 7271–7283.
23. E. C. Constable, S. M. Elder and D. A. Tocher, Polyhedron, 1992, 11, 1337–1342.
24. E. C. Constable, M. J. Hannon, P. Harverson, M. Neuburger, D. R. Smith, V. F. Wanner, L. A. Whall and M. Zehnder, Polyhedron, 2000, 19, 23–34.
25. V. V. Pavlishchuk and A. W. Addison, Inorg. Chim. Acta, 2000, 298, 97–102.
26. M. J. Frisch, et al., Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford CT, 2009.
27. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789.
28. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
29. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104G.
30. A. Petersson and M. A. Al-Laham, J. Chem. Phys., 1991, 94, 6081–6090.
31. A. D. McLean and G. S. Chandler, J. Chem. Phys., 1980, 72, 5639–5648.
32. K. Raghavachari, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem. Phys., 1980, 72, 650–654.
33. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3093.
34. G. Cauquis, A. Deronzier, B. Sillion, B. Damin and J. Garapon, J. Electroanal. Chem., 1981, 117, 139–146.
35. R. Farran, L. Le-Quang, J.-M. Mouesca, V. Maurel, D. Jouvenot, F. Loiseau, A. Deronzier and J. Chauvin, Dalton Trans., 2019, 48, 6800–6811.
36. R. P. Bonomo, A. J. Di Bilio and F. Riggi, Chem. Phys., 1991, 151, 323–333.
37. C. C. Scarborough, K. M. Lancaster, S. DeBeer, T. Weyhermüller, S. Sproules and K. Wieghardt, Inorg. Chem., 2012, 51(6), 3718–3732.
38. C. Costentin, S. Drouet, G. Passard, M. Robert and J.-M. Savéant, J. Am. Chem. Soc., 2013, 135, 9023–9031.
39. C. Costentin and J.-M. Savéant, ChemElectroChem, 2014, 1, 1226–1236.
40. M. E. Moberg, A. G. Reid, D. A. Dickie and C. W. Machan, Dalton Trans., 2024, 53, 16849–16860.
41. C. R. Hess, T. Weyhermüller, E. Bill and K. Wieghardt, Inorg. Chem., 2010, 49, 5686–5700.
42. J. J. Moreno, S. L. Hooe and C. W. Machan, Inorg. Chem., 2021, 60, 3635–3650.

---