[emim][EtSO₄] was acquired from Merck (≥95%). [emim][Br] was acquired from Across Organics (98%). [emim][Tf₂N] and [empyrr][EtSO₄] were kindly donated by Prof. Peter Licence’s ionic liquid research group of the School of Chemistry, University of Nottingham. Other chemicals used were AgNO₃ (Fischer Scientific, analytical reagent grade), ferrocene (Alfa-Aesar, 99.0%), acetonitrile (BDH VWR Prolab, 99.9%), HClO₄ (Fischer Chemicals, Fischer Scientific, United Kingdom, 60%), and tetrabutylammonium hexafluorophosphate [TBA][PF₆] (Sigma-Aldrich, 98%). CO₂ (99.5%), N₂ (99.9999%), and Ar (99.995%) were acquired from BOC, Ltd. (Nottingham, United Kingdom). The compounds were utilized in their original form. A model 760C potentiostat from CH Instruments (Austin, TX, United States) was used to conduct the electrochemical experiments.

We performed cyclic voltammetry (CV), linear sweep voltammetry (LSV), and chronoamperometry (CA) experiments in triplicate for data reproducibility check, using a three-necked glass cell and a conventional three-electrode configuration. The CV, LSV, and CA experimental data were found to be reproducible. The working electrodes were BDD, Au, Cu, and Ag disks of geometrical surface area 7.07 × 10⁻² cm², 3.14 × 10⁻² cm², 3.14 × 10⁻² cm², 3.04 × 10⁻² cm², and 2.51 × 10⁻² cm², respectively. The counter electrode used was a Pt flag (0.5 cm × 0.4 cm). For the Au electrode, ECSA was determined from the CV obtained in 0.1 M aqueous HClO₄ solution by the electronic charge of gold oxides (Ma et al., 2014) and was 9.5 × 10⁻² cm². The ECSA of the Ag and Cu electrodes were estimated to be 4.78 × 10⁻² cm² and 4.11 × 10⁻² cm², respectively, from the charge passed during the under-potential deposition of lead onto each electrode from 5 mM Pb (NO₃)₂ solution in 10 mM KCl/10 mM HNO₃ (Supplementary Figures S1, S2 in the supporting information) (Salehi-Khojin et al., 2013; Lu et al., 2012). The charge under the Pb_UPD stripping peak at a Pb²⁺ concentration of 5 mM or above corresponds to 600 μC cm⁻² Ag. The same integrated charge density under the Pb_UPD region was used to estimate the ECSA of the Cu electrode from a CV obtained at a Cu electrode (Supplementary Figure S2 in the supporting information) in 5 mM Pb (NO₃)₂ solution in 10 mM KCl plus 10 mM HNO₃ (Rosen et al., 2012). The currents in each measurement were normalized to the ECSA for calibration of catalytic activity of the electrodes for the CO₂ reduction in the room temperature ILs. In the case of the BDD electrode, the geometric surface area of the electrode was used to calibrate the currents for CO₂ reduction in the liquids. The reference electrode was a self-made Ag/Ag⁺ reference electrode, which was normalized versus Fc/Fc⁺, and the results are reported vs. Fc/Fc⁺.

Ag/Ag⁺ reference electrodes were constructed using a 7.5-cm-long glass tube, a 3.2 mm porous Vycor frit, and a heat-shrink Teflon tubing from Gamry Instruments, United States. The Vycor tip was placed inside the piece of a heat-shrink tube and held to the glass capillary by sliding the heat-shrink Teflon over the end of the glass tube. The Teflon was then gently warmed with a heat gun until the tubing tightened around the Vycor tip and the end of the reference electrode tube. The electrode was then filled with 10 mM AgTfO (Sigma-Aldrich ≥98.0% (Ag) in [emim][EtSO₄], and an Ag wire connected to a Teflon top was immersed. In the study that involved AcN as a main electrolyte, Ag/Ag⁺ reference electrodes were made by the same procedure, but the Ag wire was immersed in 10 mM AgNO₃ in either 0.1 M [TBA][PF₆]/AcN or 0.1 M [emim][EtSO₄]/AcN. Reference electrode stability is a common issue when using ionic liquids as electrolytes in electrochemistry (Zhao et al., 2008). Therefore, the Ag/Ag⁺ electrode (Ag/AgTfO in [emim][EtSO₄]) was calibrated against the potential of the IUPAC-recommended ferrocene/ferrocenium (Fc/Fc⁺), redox couple (Ma and Kenis, 2013; Niu et al., 2015a). Supplementary Figure S3 in the supporting information shows CVs obtained in 5 mM ferrocene solution in [emim][EtSO₄] at Pt and Au electrodes at 100 mV s⁻¹. The reference electrodes containing 10 mM AgNO₃ in 0.1 M [emim][EtSO₄]/AcN and 0.1 M [TBA][PF₆]/AcN were also calibrated versus ferrocene in 0.1 M [TBA][PF₆]/AcN and 0.1 M [emim][EtSO₄]/AcN, respectively (Supplementary Figure S4 in the supporting information). The formal potential, E⁰΄ (taken as the average of the Ferrocene (Fc) oxidation peak potential and cathodic ferrocenium (Fc⁺) reduction peak potential), was determined in N₂-saturated solution vs. the Ag/Ag⁺ reference electrodes, and all potentials are reported versus this.

The electrodes were polished before use using soft polishing pads and an aqueous suspension of 0.05 μm alumina from Buehler, Lake Bluff, Illinois (Buehler consumable, USA). The electrodes were cleaned, thoroughly rinsed with deionized water, subjected to 2 min of sonication in a very small amount of deionized water, and then thoroughly rinsed once more with deionized water before being dried in a stream of N₂. The cell containing roughly 5 mL of [emim][EtSO₄] was submerged with the working, counter, and reference electrodes. After the solution was bubbled for 30 min with either N₂ or Ar to eliminate dissolved oxygen or for 60 min with CO₂ (the ideal duration for CO₂ saturation in the IL), cyclic voltammograms or LSVs were captured. To study the effect of temperature on CO₂ERR in [emim][EtSO₄], we did the celebration of the Ag/Ag⁺ reference electrode at room temperature and also at 50°C, 80°C, and 100°C. Before recording an LSV for CO₂ERR at each temperature, CO₂ gas was purged in the IL for 60 min.

A 3-electrode system was adopted for the electrochemical analysis of the product formed by the reduction of CO₂ in [emim][EtSO₄], as described above. However, an additional 2 mm-diameter Pt electrode was used as the second working electrode to adsorb CO if any was formed from CO₂ reduction at the Ag working electrode. The experiment was performed using an Ag/Ag⁺ reference electrode and a Pt flag counter electrode. Chronoamperometry measurements (current-time transient) were carried out by holding the Ag working electrode for 2,400 s at a −2.33 V vs. Fc/Fc⁺ redox couple in CO₂-saturated [emim][EtSO₄] and also in [emim][EtSO₄]/AcN. A 2 mm-diameter freshly polished Pt disk electrode (activated in aqueous HClO₄) was held at −0.4 V vs. Fc/Fc⁺ in the same solution for adsorption of the CO forming during the reduction reaction. After the experiment, the Pt electrode was removed from the CO₂/IL system and rinsed with de-ionized water, then immersed in 0.1 M aqueous HClO₄ in a three-necked cell to voltammetrically strip off any CO adsorbed on the electrode surface.

Figure 1a shows the cathodic sweep profiles obtained in CO₂-saturated [emim][EtSO₄] at Ag, Cu, Au, BDD, and Pt electrodes. The initial scan direction was from positive toward negative potentials. CVs were obtained at these catalysts (electrodes), and the initial half of the CV (LSV) is reported here. By comparing the CO₂ electroreduction activity of these electrodes with that of the blank LSV (without CO₂), it is evident that some reduction of CO₂ is observed at each of these electrodes. However, by comparing the CO₂ reduction LSVs obtained at the studied electrodes, it is observed that at both the Ag and Cu electrodes, the onset potential for the reduction of CO₂ was −1.8 V compared with the onset potentials at Au (−2.0 V), BDD (−2.2 V), and Pt (−2.3 V), respectively. Thus, according to the current study, both Ag and Cu are better electrocatalysts for the electrocatalytic reduction of CO₂ in [emim][EtSO₄] medium, while at the Pt electrode, the electrocatalytic reduction of CO₂ occurs at a very high overpotential (0.5 V) compared to the Ag and Cu electrocatalysts used for CO₂ERR followed by BDD, at which the CO₂ reduction occurs at a 0.4 V overpotential and the overpotential observed at the Au electrode compared to Pt and BDD is 0.2 V. Although CO₂ERR occurs at the same onset potentials at both the Ag and Cu catalysts, a wide reductive wave with a higher current density is observed at the Ag than at the Cu catalyst. In addition, Ag is found to have known product selectivity that reduces CO₂ mainly to CO (Salehi-Khojin et al., 2013; Niu et al., 2015b; Sampson et al., 2014; Huang et al., 2009). Therefore, Ag was mainly used as a catalyst for further investigation of the CO₂ electroreduction in [emim][EtSO₄] and other ILs.

Figure 1b shows cyclic voltammograms of CO₂-saturated (green CV) and CO₂-free, that is, Ar-saturated [emim][EtSO₄] (black CV) recorded at an Ag electrode at 100 mV s⁻¹ at room temperature. In the Ar-saturated CV, only the background charging current flowed between −1.5 V and −2.5 V. After −2.5 V, a sharp increase in the reduction current started, which is attributed to the cation, [emim]⁺, reduction of the IL. In the CO₂-saturated liquid, a single irreversible broad reduction wave was observed at an onset of −1.8 V vs. Fc/Fc⁺, which peaked at approximately −2.4 V. The increase in the reduction current between −1.8 V and −2.5 V is attributed to an irreversible one-electron electroreduction of CO₂ to CO₂ ^·− in the IL (Rosen et al., 2012; Sánchez et al., 2019; Ma and Kenis, 2013). It is also evident from the figure that during the reverse sweep, a small reductive wave between −2.4 V and −2.3 V is seen, which is also due to the electroreduction of CO₂ on Ag catalysts.

APILs are thermally stable and can also be used for high-temperature studies. Room temperature is frequently used to study the CO₂ERR and practical electrolyzers; however, higher temperatures might be employed. Although it is well understood that the rates of reactions increase with an increase in temperature, fundamental knowledge of how temperature affects CO₂ERR is currently lacking. To learn more about how temperature affects this reaction, we conducted temperature-dependent investigations on the CO₂ERR using an Ag electrode. Figure 1c shows LSV profiles for CO₂ERR in [emim][EtSO₄] obtained at 298 K, 323 K, 353 K, and 373 K using polycrystalline an Ag disk electrode. Initially, CVs were recorded in the solution in a potential range from −1.8 V to −2.5 V vs. Fc/Fc⁺, and only the cathodic sweeps are shown in the figure. It is clear from the figure that at high temperatures, Ag shows better catalytic activity for CO₂ reduction. The onset potential and the reduction peak potential shift toward the more positive potential. The overpotential for CO₂ reduction in the ionic liquid decreased by ∼100 mV, and the onset reduction peak potential shifted from −2.37 V to −2.21 V when the temperature increased from 298 K to 373 K.

The Arrhenius equation ( ln ⁡ ⁡ K = l n A − E a R T ) can be used to determine the apparent activation energy for the CO₂ERR on Ag based on the temperature-dependent experiments. In the Arrhenius equation, k is the reaction rate constant, A is the pre-exponential factor, Ea is the activation energy, R is the gas constant (8.3143 J. K⁻¹. mol⁻¹), and T is the absolute temperature. Figure 1d shows the Arrhenius equation ( ln ⁡ j = l n A − E a R T plot of the natural logarithm of the current density (ln j in mA cm⁻²) versus 1/T (K⁻¹) at a constant potential of −2.2 V. It can be seen from the figure that the current density increases with an increase in the temperature; that is, for a lower 1/T value, a smaller value of current density is observed, and for a higher 1/T, value the magnitude of the current density increased. The higher intercept value from the graph indicates a higher ln j value, as from the Arrhenius equation, ln j is directly proportional to ln A. A small value of the activation energy (E_a), 13.04 J mol⁻¹ determined from the slope of the plot, shows that the CO₂ERR occurred at a fast rate at the Ag catalysts in the [emim][EtSO₄]. Lv et al. (2020) reported the activation energies for the [TETAH][Lys]-ethanol-water and [TETAH][Lys]- water solutions capturing CO₂ as 61.42 kJ/mol and 62.10 kJ/mol, respectively. Wu et al. (2016) reported the calculated activation energy of 41.89 kJ mol⁻¹ for CO₂ capture in 1-butyl-3-methylimidazolium glycinate aqueous solutions. Vos and Koper (2022) studied the effect of temperature on the cation-promoted electrochemical CO₂ reduction on gold and reported 60 kJ mol⁻¹ apparent activation energy with constant CO₂ concentration. Thus, compared to the literature-reported activation energy values for CO₂ reduction/capture in CO₂-saturated ILs, the activation energy calculated in this study is smaller, showing that an Ag catalyst surface, or the cation [emim]+, or the synergistic approach of both, activate the CO₂ERR to take place at lower overpotentials. The [emim]+ ---CO^∙−, on the one hand, reduces the activation energy for the reaction. On the other hand, the formation of surface intermediates like formate (HCOO) and carboxyl (COOH) species on the surface of thermally activated Ag can result in the final product at a fast rate (Álvarez et al., 2017). The data fit statistically well, as the regression coefficient, R², obtained from the Arrhenius plot is 0.9899, which is near the ideal fit R² = 1. The analysis confirms that the rate of the CO₂ERR increased as the temperature increased. At lower temperatures, ILs are more viscous. The viscosity of the IL decreases with increasing temperature, causing faster mass transport of CO₂ toward the electrode surface (Zare et al., 2019).

The diffusion coefficient (D) and concentration (C) of CO₂ in [emim][EtSO₄] were experimentally obtained by a potential-step chronoamperometry. A chronoamperometric transient was recorded using an Au microelectrode (nominal radius 12.5 µm) and is shown in Figure 2 (red line). The potential was stepped from −1.8 V vs. Ag/Ag⁺ (−2.05 V vs. Fc/Fc⁺) to −2.3 V (−2.55 V vs. Fc/Fc⁺) for 10 s. The black dotted line in Figure 2 shows the theoretical fitted data using the Shoup–Szabo equation shown as Equation 1, which allows determining both D and C of CO₂ simultaneously (Bard, 2010; Spendelow et al., 2004). i = − 4   n   F   D   C   r f τ τ , (1)

where n is the number of electrons involved in the redox process, which is taken as 1 for the reduction of CO₂ to CO₂ ^·−, and r is the radius of the Au microelectrode, which was determined using steady-state current. i_ss is used for oxidation of ferrocene at the microelectrode and is determined using Equation 2 (Supplementary Figure S5 in the supporting information) (Bard, 2010). i s s = 4   n   F   C   D   r , (2)

where i_ss was 7.97 × 10⁻⁹ A, and n is the number of electrons involved in the redox process, which is 1 for the Fc/Fc⁺ redox couple. The concentration, c, of the ferrocene solution was 2 × 10⁻⁶ mol cm⁻³. The diffusion coefficient of the Fc/Fc⁺ redox couple was 6.7 × 10⁻⁶ cm² s⁻¹ (Bard, 2010; Bond et al., 2010). The actual radius of the Au microelectrode calculated through this experiment was 15.6 µm.

The function f ( τ ) and the parameter, τ in the Shoup–Szabo equation are given by Equations 3, 4, respectively. f τ = 0.7854 + 0.8863 τ − 1 / 2 + 0.2146 ⁡ exp − 0.7823 τ − 1 / 2 . (3) τ = 4   D t / r 2 (4)

The D and C values obtained from theoretical fitted data were 4.78 × 10⁻¹⁰ m² s⁻¹ (or 4.78 × 10⁻⁶ cm² s⁻¹) and 1.83 × 10⁻⁵ mol cm⁻³ (0.0183 mol L⁻¹), respectively.

The diffusion coefficient of CO₂ in [emim][EtSO₄] (η = 108 cP) is comparable to the diffusion coefficient of O₂ in [bmim][BF₄] (η = 92 cP), which was 1.79 × 10⁻¹⁰ m²s⁻¹ measured at 298 K using a 10-µm-diameter Au electrode (Martindale and Compton, 2012). However, the D value for CO₂ in [emim][EtSO₄] was higher by two orders of magnitude than that observed in [bmim][Ac] (Wang et al., 2020), which was 2.65 × 10⁻¹² m²s⁻¹. This could be due to the low viscosity of [emim][EtSO₄], in which CO₂ molecules can diffuse faster than [bmim][Ac] (η = 140 cP). However, viscosity could not be the only reason for the high diffusion of CO₂ in [emim][EtSO₄]. The arrangement of atoms or alkyl groups around the central atom in the cation and anion may also play a role. For example, [bmim][BF₄] has viscosity η = 112 cP, which is very close to that of [emim][EtSO₄], 108 cP (Samjeské et al., 2009), and the D value for CO₂ in the latter case is higher by an order of magnitude than the former (7.3 × 10⁻¹¹ m²s⁻¹) (Wang et al., 2020).

The effect of [emim][EtSO₄] as a supporting electrolyte and/or co-catalyst was investigated during CO₂ERR in the conventional organic solvent, AcN, and the results were compared with the CO₂ERR carried out in AcN using the conventional supporting electrolyte, [TBA][PF₆]. Figure 3 shows cyclic voltammograms measured in AcN containing (a) 0.1 M [TBA][PF₆] and (b) 0.1 M [emim][EtSO₄] at a 2 mm-diameter Ag electrode. Each of the solutions was purged with Ar for 30 min to remove dissolved oxygen, as oxygen can actively reduce and hinder the main electrochemical process, and CVs were recorded as shown by black lines in the Figures. Then, CO₂ was bubbled for 60 min into the solutions, and CVs were measured at an Ag catalyst, as shown by the red and green lines in Figures 3a, b, respectively. In the blank (Ar saturated) [TBA][PF₆]/AcN, only capacitive current flowed up to −2.48 V, after which a large reductive current started flowing that corresponds to the decomposition of AcN. In CO₂-saturated [TBA][PF₆]/AcN, the current response was different from that in blank solutions. A reduction current started flowing at an onset potential of −2.1 V until −2.48 V. This cathodic reduction wave, which was also observed during the reverse sweep, is attributed to the electrocatalytic CO₂ reduction at the Ag catalyst in the solution.

Similarly, if we compare the CV obtained in the CO₂-saturated [emim][EtSO₄]/AcN solution (Figure 3b) with the blank CV (Ar-saturated), a sharp increase in the current density can be seen at an onset potential of −1.9 V. The reduction current kept increasing and can also be observed during the reverse sweep, which is attributed to the CO₂ electroreduction. This shows the Ag electrode’s ability to reduce CO₂ in AcN using both [TBA][PF₆] as well as [emim][EtSO₄] as supporting electrolytes. However, it is clear from Figure 3c that the reduction of CO₂ occurs at ∼200 mV lower potential when using [emim][EtSO₄] as a supporting electrolyte in AcN than when [TBA][PF₆] is used as a supporting electrolyte. The high current density and early onset potential for CO₂ electroreduction show that [emim][EtSO₄] plays a co-catalytic role. This could be attributed to the fact that [emim]⁺ may stabilize the intermediate radical anion, CO₂ ^·− during the intermediate step of the reduction of CO₂, that is, the electron uptake process (CO₂ + e− → CO₂ ^·−) and further help in its reduction (Niu et al., 2015b; Sampson et al., 2014).

As mentioned above, [emim][EtSO₄] has some electrocatalytic effect toward CO₂ERR when used alone as a non-aqueous medium or as a supporting electrolyte in AcN at an Ag catalyst. To verify whether cation or anion of the ILs plays a role during CO₂ERR, experiments were performed using different ionic liquids as supporting electrolytes in AcN. Figure 3d displays linear sweep voltammograms for the reduction of CO₂ in AcN containing 0.1 M of each of [TBA][PF₆], [empyrr][EtSO₄], [emim][Tf₂N], and [emim][EtSO₄] obtained at a scan rate of 100 mV s⁻¹on an Ag catalyst. Generally, it is clear from the figure that compared to [TBA][PF₆], the CO₂ ERR occurs at a lower onset potential when ILs are used as supporting electrolytes in AcN. This confirms further that ILs play a co-catalytic role during the reduction process. Furthermore, it can be seen from the figure that a low overpotential is observed for the CO₂ERR when [emim]⁺ cation-based ILs, that is, [emim][EtSO₄] and [emim][Tf₂N], were used as supporting electrolytes as in these ILs. The onset potential for the reduction is approximately the same (∼1.9 V), although the anions are different. However, while comparing the catalytic role of the different ILs toward CO₂ERR, it can be seen that the overpotential for the reduction of CO₂ in AcN solution containing 0.1 M [empyrr][EtSO₄] as a supporting electrolyte is higher by 80 mV than that containing [emim][EtSO₄] and [emim][Tf₂N]. This shows that the imidazolium-based ILs, especially [emim]⁺, play some co-catalytic role during the reduction of CO₂ at the Ag catalyst, also signifying the role of the cation of the ILs during CO₂ electrocatalysis [51].

It is suggested by Rosen et al. (2012) that after a one-electron reduction in imidazolium-based ILs in acidic media, an intermediate complex with [emim]⁺ cation is formed, which further results in CO formation. In order to confirm this in our IL medium, we performed product analysis by using the cyclic voltammetry technique. Figure 4 shows cyclic voltammograms recorded at the Pt electrode that was held at −0.4 V vs. Fc/Fc⁺ during CO₂ reduction at an Ag electrode in [emim][EtSO₄] at 100 mV s⁻¹. The figure shows that a strong oxidative wave started flowing during the first positive going scan (dotted line) at an onset potential of 0.846 V with a peak potential of 0.908 V vs. RHE. This sharp oxidative wave is attributed to the voltammetric oxidation of Pt-adsorbed CO. The adsorbed CO was completely stripped off the surface of Pt during the first scan, as no CO oxidation peak was observed in the second scan (solid line) in the same potential region, showing that the surface of Pt was CO-free during the second scan. It is well established that at potentials ≥0.85 V (here onset potential of CO oxidation), Pt surfaces oxidize to form PtOH_ads/PtO_ads. in aqueous acid solutions. The surface Pt hydroxide is reduced back to Pt on the reverse scan, as evident from the broad reductive wave in Figure 4, at an onset potential of approximately 1.0 V with a peak potential of 0.75 V. The oxygen-containing Pt surface acts as a catalyst to oxidize CO-covered Pt to CO₂ during oxidative sweep by the Langmuir–Hinshelwood mechanism (Martindale and Compton, 2012; Samjeské et al., 2009). Formation of CO during CO₂ERR in [emim][EtSO₄] at the Ag electrode and its adsorption on and then removal from the Pt surface was reproducible, as a similar behavior was observed when the experiment was performed again at 323 K under the same set of other experimental conditions. In addition, we have shown that the CO₂ reduction rate at the Ag electrode in [emim][EtSO₄] increased as the overpotential gradually decreased with the increase in temperature of the system, probably due to a decrease in the viscosity of the liquid with rising temperature. The formation of CO was also detected during the electroreduction of CO₂ at an Ag electrode in AcN containing 0.5 M [emim][EtSO₄] as a supporting electrolyte (inset of Figure 4).

Based on the proposed mechanism in the literature (Zhang et al., 2024), it is suggested that the CO₂ ⁻ intermediate formed by a one-electron reduction in the [emim]⁺⋅⋅⋅CO₂ medium at the Ag electrode is stabilized by forming an intermediates complex through interaction between CO₂ ^·− and [emim]⁺. This can facilitate the binding of CO₂ ^·− at the electrode interface to accept electrons for further reduction to CO. Additionally, during electrochemical reduction, the imidazolium cation can continue to exhibit catalytic activity because the structures of [emim]⁺ do not change. The schematics are shown in Figure 5. The optimization of CO₂ conversion at the Ag electrode in terms of maximizing CO production would be beneficial to produce liquid fuels when combined with H₂ via the Fischer–Tropsch process as well as by Sabatier’s process (Ma and Kenis, 2013).