In Situ Fabrication and Reactivation of Highly Selective and Stable Ag Catalysts for Electrochemical CO₂ Conversion

Abstract

In this work, the highly selective and stable electrocatalytic reduction of CO₂ to CO on nanostructured Ag electrocatalysts is presented. The Ag electrocatalysts are synthesized by the electroreduction of Ag₂CO₃ formed by in situ anodic-etching of Ag foil in a KHCO₃ electrolyte. After 3 min of this etching treatment, the Ag₂CO₃-derived nanostructured Ag electrocatalysts are capable of producing CO with up to 92% Faradaic efficiency at an overpotential as low as 290 mV, which surpasses all of the reported Ag catalysts at identical conditions to date. In addition, the anodic-etched Ag retained ∼90% catalytic selectivity in the electroreduction of CO₂ to CO for more than 100 h. The Ag₂CO₃-derived Ag is able to facilitate the activation of CO₂ via reduction of the activation energy barrier of the initial electron transfer and provide an increased number of active sites, resulting in the dramatically improved catalytic activity for the reduction of CO₂ to CO.

The electrocatalytic conversion of CO₂ into carbon-based fuels and valuable chemicals powered by renewable electricity is an attractive solution to both the utilization of captured CO₂ and the storage of renewable energy.¹⁻⁶ An essential step for achieving this goal is to find a highly efficient and selective electrocatalyst with long-term stability.^7,8 Many transition metal catalysts have been evaluated for the selective reduction of CO₂ in CO₂-saturated aqueous solutions.^2,9 Currently, the electrochemical conversion of CO₂ into CO provides one of the most promising routes to form a cost-competitive product because syngas (CO and H₂) can be employed in Fischer–Tropsch synthesis to produce value-added chemicals and synthetic fuels using already existing industrial technologies.^1,4,10 Therefore, significant attention has been focused on finding electrocatalysts that can selectively produce CO from CO₂.¹¹⁻¹⁸

While Au is the most active surface for reducing CO₂ selectively to CO among the identified metal catalysts, its potential for industrial applications is currently limited by its low abundance and high cost. In this context, Ag has great potential for large-scale applications due to its significantly lower cost than Au and high catalytic selectivity for the reduction of CO₂ to CO.^{9−11,19−24} However, high overpotentials (η) required for driving selective CO₂ reduction and rapid catalytic deactivation in favor of H₂ evolution on Ag catalysts significantly restrict its practical utilization.^20,25

To overcome the limitations of Ag electrocatalysts, many attempts have focused on the development of nanostructured surfaces, which offer mass-transport advantages and contain more low-coordinated sites (edge sites and corner sites) that are more active for CO₂ reduction in comparison with a planar metallic surface.^{7,10,14,22,26,27} It has been demonstrated that nanoporous Ag catalysts prepared by dealloying,¹⁰ Ag nanocoral catalysts synthesized via an oxidation–reduction process using chloride anions²⁸ and oxide-derived nanostructured Ag²² are capable of electrochemically reducing CO₂ to CO with dramatically enhanced catalytic selectivity at reduced overpotentials, which are ascribed to fast initial electron transfer for CO₂ activation on these nanostructured catalysts. In addition, the catalytic stability of the electroreduction of CO₂ to CO has been improved on nanostructured Ag catalysts,^10,22,28 owing to the enhanced tolerance to heavy metal impurities in the electrolyte.²⁷ While these previous attempts have significantly improved the performance of Ag electrocatalysts, it is still critical to develop a very simple, fast, scalable and low-cost method for preparing and maintaining higher-performance Ag electrocatalysts for practical applications of CO₂ electroreduction.

Herein, we demonstrate a simple technique to prepare highly active, stable, and selective Ag electrocatalysts in CO₂-saturated KHCO₃ electrolyte that is used for CO₂ reduction. By first anodic-etching Ag to form Ag₂CO₃ and then further reducing the Ag₂CO₃ to metallic Ag with a highly porous scaffold structure, a significantly reduced overpotential for high CO catalytic selectivity was achieved with remarkable catalytic stability, which outcompetes the reported Ag catalysts to date at identical conditions. Thus, the robust performance of the nanostructured catalysts formed via this very easy and low-cost synthesis method may offer a platform for practical applications of CO₂ electroreduction.

For the electrochemical synthesis of Ag electrocatalysts used in this work, a polycrystalline Ag foil electrode was immersed in a CO₂-saturated 0.1 M KHCO₃ electrolyte in a two-compartment cell using a three-electrode configuration (a Pt counter electrode and a Ag/AgCl reference electrode). The two-compartment cell was separated by a Nafion-115 proton exchange membrane to minimize impurity deposition during electrolysis. An anodic potential of 2.6 V vs the reversible hydrogen electrode (RHE) was applied on the Ag foil electrodes for 3 min to synthesize the Ag₂CO₃ layer, with an estimated thickness of ∼4.9 μm (Table S1). The Ag₂CO₃ electrodes were then directly utilized for electrocatalytic CO₂ reduction in CO₂-saturated 0.1 M KHCO₃ electrolyte and were electrochemically reduced to metallic Ag in the initial period (<2 min) of CO₂ reduction electrolysis (Figure S5d).

In the initial anodic-etching, the Ag electrodes formed short microporous polyhedral rod-like morphologies with smooth surfaces, as presented in Figure 1a, which is consistent with the reported morphologies of Ag₂CO₃.²⁹Figure 1b shows scanning electron microscope (SEM) images of the same electrodes after the electrolysis (∼0.5 h) of CO₂ reduction, revealing that a nanoporous structure was formed via the electroreduction of Ag₂CO₃. In addition, transmission electron microscopy (TEM) confirmed that the size of the nanostructured materials after the CO₂ reduction (Figure 1c), and the selected-area electron diffraction (SAED) pattern in the inset of Figure 1c exhibits the typical nature of crystals. To identify the phase of the prepared materials, X-ray diffraction (XRD) measurements were conducted. The XRD diffractograms in Figure 1d indicate that Ag₂CO₃ was formed by anodic-etching of Ag foil in KHCO₃ solution. After CO₂ reduction electrolysis, only Ag diffraction peaks were observed (Figure 1d) without any remaining Ag₂CO₃, indicating a full transformation from Ag₂CO₃ to metallic Ag.

Figure 1.

SEM images of anodic-etched Ag (AE-Ag) (4.9 μm) before (a) and after (b) CO₂ reduction electrolysis. (c) TEM image of AE-Ag after CO₂ reduction electrolysis (the inset is the SAED pattern). (d) XRD patterns and (e) XPS spectra of untreated polycrystalline Ag (blue line) and AE-Ag before (dark yellow line) and after (red line) CO₂ reduction electrolysis, respectively.

To verify the surface composition of our samples, X-ray photoelectron spectroscopy (XPS) measurements were performed. As shown in Figure 1e, the Ag 3d_5/2 peak at 368.2 eV was observed for polycrystalline Ag. For the anodic-etched Ag (AE-Ag) before CO₂ reduction electrolysis, the Ag 3d_5/2 peak shifted by about 0.3 eV to the binding energy of 367.9 eV compared to the metal Ag, which is consistent with the value of Ag⁺ in the synthesized Ag₂CO₃ according to previous work.²⁹⁻³¹ In addition, the binding energy of 288.7 eV in C_1s spectra (Figure S3) represents the carbon associated with the carbonate ((CO₃)^2–),³² which further confirms the formation of Ag₂CO₃. After electrolysis, the Ag 3d_5/2 peak shifted back to 368.2 eV, corresponding to metallic Ag⁰.²⁹⁻³¹ Furthermore, a surface valence band XPS spectrum of AE-Ag after electrolysis is in line with that of metallic Ag foil (Figure S4). All of the above results indicate that the reduction of Ag₂CO₃ to metallic Ag was complete, implying that only metallic Ag was present on AE-Ag after CO₂ reduction electrolysis.

Figure 2a presents a comparison of the electrocatalytic activity of CO₂ reduction for untreated polycrystalline Ag (in blue) and 4.9 μm AE-Ag (in red) at various applied potentials (iR-corrected potentials). Both the AE-Ag and untreated Ag experienced a gradually enhanced Faradaic efficiency (FE) for CO production at more negative potentials (Figure 2a), simultaneously accompanying with a decrease in the related FE for H₂ formation (Figure S6). Notably, the overpotential required for achieving >90% FE for CO production was shifted toward the positive potential by ∼800 mV on AE-Ag compared to that of untreated Ag. More importantly, a high FE of more than 92% for CO formation was achieved on AE-Ag at a potential of −0.4 V vs RHE, which corresponds to an overpotential (η_co) as low as 0.29 V relative to the CO₂/CO equilibrium potential of −0.11 V vs RHE, representing the highest catalytic selectivity for CO₂ reduction to CO among the reported Ag catalysts at the same η_co (Table S2). In contrast, no CO production was detected on untreated polycrystalline Ag at identical conditions (η_co = 0.29 V). A plot of the partial current density for CO production (j_co) as a function of potential in Figure 2b suggests that the onset potential for the reduction of CO₂ to CO on AE-Ag was −0.3 V vs RHE (η_co = 0.19 V), which is a positive shift of ∼250 mV in comparison with that (−0.55 V vs RHE) of untreated polycrystalline Ag. These results show that Ag₂CO₃-derived nanostructured Ag is a highly selective electrocatalyst for the electrocatalytic reduction of CO₂ to CO while inhibiting H₂ evolution at significantly reduced overpotentials.

Figure 2.

Comparison of the electrocatalytic activity of polycrystalline Ag and AE-Ag (4.9 μm). (a) FE for CO at various potentials in CO₂-saturated 0.1 M KHCO₃ (pH 6.8). (b) Current density for CO formation at various potentials. (c) Catalytic stability performance for AE-Ag. The inset shows the catalytic stability of untreated polycrystalline Ag. All of the potentials were iR-corrected.

To test the electrocatalytic stability of AE-Ag catalysts, a long-term CO₂ reduction measurement was performed on AE-Ag at a fixed potential of −0.55 V vs RHE (KHCO₃ electrolyte without any purification was used). As shown in Figure 2c, AE-Ag exhibited an initially high geometric current density (j_tot) at the early stage of electrolysis owing to the electroreduction of Ag₂CO₃ to Ag and subsequently a stable j_tot of ∼1 mA/cm² with a FE of ∼90% for CO production over ∼37 h. After ∼37 h, slight catalytic deactivation for CO formation was found, which may result from the deposition of impurities on the surface of the catalyst during the electrochemical reduction of CO₂.^9,27 To overcome this slight deactivation, the same Ag catalyst electrodes were then held at an anodic potential of 2.6 V vs RHE for 3 min in the same CO₂-saturated KHCO₃ electrolyte, and after returning back to −0.55 V vs RHE, a FE as high as ∼90% for CO formation was recovered and maintained for more than 60 h (no replacement of electrolyte during >100 h electrolysis). This remarkable stability significantly surpasses the currently reported durability for CO₂ reduction on single-element catalysts (Table S3) under similar conditions. The in situ reactivation of the catalysts may be attributed to the removal of impurities on the surface of the catalysts by anodic-etching the contaminated surface. In contrast, the polycrystalline Ag electrodes had a very low j_tot (∼0.07 mA/cm²) and a very low FE for CO, which decreased from 3.4 to 0% over the course of 2.5 h at −0.55 V vs RHE, which indicates fast catalytic deactivation. Thus, the Ag resulting from AE-Ag exhibited high catalytic selectivity and activity with long-term stability for the electrocatalytic reduction of CO₂ to CO.

The electrochemical active surface area (EASA) of nanoporous Ag catalysts reduced from AE-Ag and untreated polycrystalline Ag was measured by forming a monolayer oxide on Ag surface in 0.1 M KOH.²⁰ The charge used for oxidizing the monolayer of the Ag surface was calculated in Figure S7, which shows that the EASA of nanoporous Ag catalysts reduced from AE-Ag (3 min) is more than 10-fold larger than that of untreated Ag, resulting in the discrepancy of j_tot between AE-Ag (∼1 mA/cm²) and untreated Ag (∼0.07 mA/cm²), as shown in Figure 2c. Thus, the increased number of active sites (increased EASA) reflects the enhanced catalytic reaction rate. In addition, the normalized j_co of AE-Ag by EASA (∼0.08 mA/cm²) is ∼40 times higher in comparison with that (∼0.002 mA/cm²) of untreated Ag, indicating significantly improved intrinsic CO₂ reduction activity on nanoporous Ag catalysts reduced from Ag₂CO₃.

The thickness effect of AE-Ag on the catalytic performance was also evaluated. In this work, the average thickness of AE-Ag was tuned by systematically varying the anodic-etching time (Table S1). As shown in Figure 3a, the thickness of AE-Ag is linearly correlated with the anodic-etching time, and an gradually enhanced FE for CO formation was observed along with decreased FE for H₂ evolution at −0.55 V vs RHE with increasing thickness (≤3.5 μm) of AE-Ag. While a high FE of >90% for CO formation was achieved on both 3.5 and 4.9 μm AE-Ag (Figure 3a), the distinct thickness leads to a discrepancy of j_tot between the two catalysts (Figure 3b). The EASA was enhanced with increasing thickness of AE-Ag (Figure 3b), which led to the correspondingly increased j_tot and j_co (Figure S8). These results indicate that the thicker nanoporous Ag is able to provide more active sites for the reduction of CO₂ to CO.

Figure 3.

Electrocatalytic activity as a function of the thickness of AE-Ag at −0.55 V vs RHE. (a) FE for CO and H₂ on AE-Ag with different thickness in CO₂-saturated 0.1 M KHCO₃ (pH 6.8). The inset shows the thickness for different anodic-etching times. (b) Geometric current density and normalized EASA.

It has been reported that the surface facets of Ag could significantly influence the catalytic activity of CO₂ reduction.^20,33 To reveal the variation of Ag surface facets before and after anodic treatment, the adsorption/desorption of OH^– was performed on AE-Ag and untreated Ag in argon-purged 0.1 M KOH in the potential range from ∼−0.3 to ∼1 V vs RHE (double layer region) at room temperature.³⁴⁻³⁶Figure 4a exhibits the difference in peak potentials for the OH^– adsorption/desorption processes between AE-Ag (red) and polycrystalline Ag (blue), which correlate with the distinct surface facets of the two catalysts.³⁴⁻³⁶ In addition, cyclic voltammetry of oxide-derived Ag reported in our previous work²² was also conducted in argon-purged 0.1 M KOH, which shows different peak potentials compared to AE-Ag (Figure S11), reflecting different dominant facets of the Ag surface. While the specific facets of the Ag surface could not be identified, the obvious discrepancy in Ag surface facets may contribute to the difference of the catalytic performance in the reduction of CO₂. Therefore, in addition to the increased EASA in the anodic-etched nanostructured Ag compared to polycrystalline Ag, we also provide evidence of a different surface electronic structure that could also influence the catalytic activity and selectivity.

Figure 4.

(a) Cyclic voltammetry curves of untreated polycrystalline Ag (blue line) and AE-Ag (4.9 μm) (red line) in argon-purged 0.1 M KOH at room temperature with a sweep rate of 50 mV/s. (b) Tafel plots of the CO partial current density for polycrystalline Ag and AE-Ag (4.9 μm). (c) Bicarbonate concentration at constant potentials of AE-Ag. (d) Charge transfer resistance at various overpotentials. The inset shows the equivalent circuit for the metal/solution interface (R_ct, R_s, CPE and W_s are charge transfer resistance, solution resistance, constant phase element, and Warburg-short circuit terminus, respectively).

In order to gain insight into the electrokinetic mechanism of CO₂ reduction on AE-Ag and untreated polycrystalline Ag, Tafel analysis was performed. It has been demonstrated that a two-electron transfer is involved for CO₂ reduction to CO, and each electron transfer is followed with one proton donation step (or proton-coupled electron transfer steps based on computational studies³⁷⁻⁴⁰).^7,9,10,22,27 Of particular note, the initial electron transfer for CO₂ activation (stabilization of CO₂^•– or COOH^•) is the rate-determining step (RDS) in the whole process due to the much higher activation energy barrier for the first electron transfer compared to the following steps.^10,20 In our study, a Tafel plot of untreated Ag (overpotential versus log of the partial current density for CO production) in Figure 4b shows a Tafel slope of 133 mV/dec, which implies that the initial electron transfer for CO₂ activation is the RDS for the overall process (Scheme 1).^10,16 In contrast, a low Tafel slope of 60 mV/dec was obtained on AE-induced nanostructured Ag catalysts (4.9 μm) at relatively low overpotentials, indicating fast initial electron transfer to a CO₂ molecule for CO₂ activation (Scheme 1).^10,25 In addition, this low Tafel slope is consistent with a fast pre-equilibrium of the initial electron transfer prior to a RDS according to previous work.²⁵ Furthermore, a dramatically increased Tafel slope for nanostructured Ag was observed at relatively high overpotentials, implying that the electrocatalytic CO₂ reduction likely reaches a mass transport limitation.

Scheme 1. Proposed Reaction Paths for CO₂ Reduction to CO on Untreated Ag and Ag₂CO₃-Derived Nanostructured Ag.

The grey, red, and white balls represent C, O, and H atoms, respectively. Larger arrows indicate the relatively fast reaction steps.

It has been demonstrated that the initial proton donation is derived from HCO₃^–.^20,25 Thus, to further uncover the reaction mechanism (first proton donation step) on nanostructured Ag, the effect of HCO₃^– concentration on the CO₂ reduction activity was investigated. A plot of log(j_co) versus log([HCO₃^–]) in Figure 4c exhibits a slope of ∼0.8, which corresponds to first-order dependence of the HCO₃^– concentration on the reaction rate, indicating that proton donation from HCO₃^– is a RDS for nanostructured Ag in the reduction of CO₂ to CO.^10,25 Thus, the RDS is switched from the first electron transfer for untreated Ag to the initial proton donation for nanostructured Ag (Scheme 1).

To better understand the charge transfer process at the electrode/electrolyte interface, electrochemical impedance spectroscopy (EIS) was performed at various potentials. The comparison of charge transfer resistance (R_ct) between untreated Ag and AE-Ag as a function of overpotential was extracted from EIS (Figure S9) based on the equivalent circuit (Figure 4d). As presented in Figure 4d, AE-Ag exhibited a much lower R_ct than that of polycrystalline Ag at identical conditions, suggesting a significantly accelerated charge transfer process on AE-Ag,¹⁴ which may reflect the reduced activation energy barrier of electron transfer on nanostructured Ag. This result is consistent with fast initial electron transfer on nanostructured Ag according to Tafel analysis. In addition, a clear mass transport limitation for nanoporous Ag was observed at relatively high overpotentials in Nyquist plots (Figure S10), which is also in line with Tafel analysis (the dramatic increase in the Tafel slope at relatively high overpotentials). These results indicate that, while a mass transport limitation may be reached on nanoporous Ag at high overpotentials, the dramatically improved initial electron transfer for CO₂ activation enhances the intrinsic CO₂ reduction activity, resulting in high catalytic selectivity and activity for the electrocatalytic reduction of CO₂ to CO. EIS has seldom been used in CO₂ reduction experiments; thereby, the consistency between the Tafel analysis and EIS shows the potential for this technique to give meaningful information relating to mechanistic charge transfer processes for electrochemical CO₂ reduction.

In summary, a simple and fast anodic-etching procedure was used to fabricate highly active, selective, and stable Ag electrocatalysts for the reduction of CO₂ to CO. A high FE of >92% for CO was achieved on AE-Ag at a potential of −0.4 V vs RHE (overpotential of 290 mV). Notably, the AE-Ag was capable of maintaining a high catalytic selectivity of ∼90% for CO production for >100 h, which remarkably outcompetes the currently reported durability of single-metal catalysts. The improved CO₂ reduction performance is attributed to the increased number of active sites for CO₂ reduction and the improved intrinsic CO₂ reduction activity by fast initial electron transfer. In this study, after prolonged CO₂ reduction, the procedure of anodic-etching can be performed subsequently on the same Ag electrocatalysts in the same KHCO₃ electrolyte that is used for CO₂ reduction to recover the robust catalytic performance. Thus, the Ag electrocatalysts, prepared by this fast, simple, and cost-effective approach, is capable of reducing CO₂ to CO with high catalytic selectivity and excellent stability, offering a very promising platform for industrial applications.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsenergylett.8b00472.

• Experimental details of Ag₂CO₃ fabrication, SEM, TEM, XRD, XPS, and CO₂ reduction measurement, thickness calculation of anodic-etched Ag, summarized tables of reported electrocatalysts, EASA measurement, calculation of j_CO and normalized j_CO, EIS, OH^– adsorption/desorption, iR correction, and error bars (PDF)

The authors declare no competing financial interest.