Electrochemical Reduction of Carbon Dioxide: II. Design, Assembly, and Performance of Low Temperature Full Electrochemical Cells

Jingjie Wu; Frank G. Risalvato; Pranav P. Sharma; Perry J. Pellechia; Fu-Sheng Ke; Xiao-Dong Zhou

doi:10.1149/2.030309jes

Electrochemical reduction of CO₂ using renewable sources of electricity is an alternative approach to produce fuels, akin to photosynthesis, thus potentially reducing the dependence on foreign fuels as well as mitigating the concentration of CO₂ in the atmosphere. The design of the electrode and its reactor is a critical step for the feasible production of fuels from CO₂ on a large scale. The electrochemical reduction process, a.k.a. electrolysis, can take place over a wide range of temperatures, from room temperature to over 1000°C. In the case of high temperature electrolysis, solid oxide electrolysis cells (SOECs) are exemplary for the electrochemical conversion of CO₂ into CO or co-electrolysis of H₂O and CO₂ to form syngas (CO and H₂). In general, SOECs exhibit a low overpotential and superior selectivity to CO formation as well as high current density.^1–3 The overarching challenges facing prolonged SOEC operation include the electrode degradation due largely to the formation of coke and volatile nickel-carbonyls.^4–6

Low temperature electrolysis of CO₂ has been investigated extensively in electrochemical cells consisting of solid polymer electrolyte.^7–15 Earlier work was carried out by using proton exchange membrane fuel cells (PEMFCs) to directly convert CO₂ to other species in gas phases.^7,8 Various hydrocarbons, including CH₄, C₂H₄ and C₂H₆, were observed on a copper electrode deposited on Nafion 117 membrane; the overall faradaic efficiency, however was ∼1% due to the acidity of the Nafion membrane which promoted the competitive hydrogen evolution reaction over CO₂ reduction reaction.^12,16 Modifications in the physical structure of PEMFCs were adopted to improve the activity and selectivity of the electrodes.^9–14 For instance, a 3D cathode consisting of a Sn-Cu mesh screen was investigated with Nafion 117 membrane as a separator. This design allowed a co-current ﬂow of CO₂ and catholyte through the cathode, thus providing an aqueous environment for CO₂ reduction.⁹ As a consequence, the faradaic efficiency toward formate formation was substantially improved to ∼86% at a cell potential of 3 V. The high operation voltage resulted from great ohmic loss and cathodic overpotential.⁹ In addition, the lower solubility of CO₂ in aqueous electrolytes¹⁷ significantly affected cell performance at higher current densities because of mass transfer constraints. In another design, a glass fiber saturated with 0.5 M KHCO₃ was inserted between the Nafion membrane and the Ag gas diffusion electrode (GDE) to prevent excessive buildup of protons near the cathode, which resulted in a significant increase of selectivity toward CO₂ reduction to CO.¹² The glass fibers, however could become dry during electrolysis due to the evaporation of water, resulting in degradation of cell performance. One approach to circumvent this drying issue was to operate the cell with a flowing electrolyte stream in place of saturated glass fiber.¹³ With a continuous flowing electrolyte, formate production on Sn GDE was achieved at a faradaic efficiency of ∼70% by using 0.5 M KCl as the electrolyte at pH = 7, while the efficiency increased to ∼90% at pH = 4. The cell, however, was operated at high voltages to obtain formate.¹³ Recently, an ionic liquid was used as the catholyte with Ag GDE to convert CO₂ to CO, which showed substantial improvement over both faradaic efficiency (∼100%) and the overpotential (less negative than −0.2 V) toward CO formation due to the decrease of the free energy for formation of the intermediate CO₂⁻.¹⁸

In this paper, we report a design of a full electrochemical cell based on the PEMFC configuration by inserting a circulating liquid-phase electrolyte buffer layer between the Nafion membrane and Sn GDE. This design demonstrates the feasibility to use current PEMFC hardware for electrochemical CO₂ conversion. This bench-scale full electrochemical cell precedes a future pilot-scale demonstration plant to generate fuels from CO₂ reduction. There are three key advantages of this cell design: (1) Insertion of a liquid-phase buffer layer significantly shifts the selectivity toward formate formation (2) In comparison with stationary electrolytes, the continuous circulation of the electrolyte provides moderate convective mixing and maintains adequate pH near cathode surface necessary for CO₂ reduction (3) A circulating electrolyte provides a cost-effective method for the accumulation of liquid products.

Experimental

Configuration of a full electrochemical cell

The buffer layer adopted in this research resides between the Nafion membrane and the cathode. A gas diffusion layer (GDL) and a half catalyst-coated membrane (CCM) were hot pressed together to serve as the anode. The Sigracet GDL 10BC was purchased from SGL group. The half CCM was made by spraying Pt/C catalysts (60%, Johnson Matthey) ink onto one side of a Nafion 212 membrane (DuPont). The Pt loading was 0.3 mg cm⁻². The Sn based cathode was made by spraying Sn (Alfa Aesar, APS 100 nm) catalyst ink onto the GDL. The Sn catalyst ink was prepared by mixing Sn powders, Nafion ionomer, deionized water, and isopropanol with an optimized volume ratio, followed by ultrasonicating the mixture for one hour. The cathode had a geometric area of 4 cm² and a loading of Sn ∼2 mg cm⁻². The thickness of buffer layer was approximately 2.4 mm. Figure 1 illustrates the schematic of the full electrochemical cell. All the reactant gases were supplied at ambient pressure (p° = 1.013 bar) and room temperature (T = 298 K). The cathode side of the cell (where the electrochemical reduction of CO₂ occurs) was supplied with gas-phase CO₂ (99.998%) at a flow rate of 45 mL min⁻¹. The anode side of the cell was supplied with either H₂ (99.99%) at 150 mL min⁻¹ or 1 M KOH (pH = 13.8, Sigma-Aldrich, reagent grade) solution at 10 mL min⁻¹. All the gases and solutions flowed along serpentine channels in graphite over which gold-plated metal hardware was chosen as the current collector. The aqueous electrolyte of 0.1 M KHCO₃ (pH = 7, Sigma-Aldrich, ACS reagent grade) was used as the buffer layer and circulated at a flow rate of 5 mL min⁻¹. The total volume of the electrolyte solution was 10 mL. The liquid products were accumulated in the electrolyte solution during electrolysis. The formate concentration was 2–100 μmol mL⁻¹ at the end of 0.5 h electrolysis in the potential range, −0.8 to −2.8 V.

**Figure 1.** A schematic of the full electrochemical cell featuring a buffer layer of circulating liquid-phase electrolyte.

The anode reaction is either hydrogen oxidation or oxygen evolution depending on the choice of the reactants. The reactions are shown in Eqs 1 and 2, respectively.

The electrochemical reduction of CO₂ into formic acid occurs at the cathode side and is shown below. The formic acid dissociates to formate (HCOO⁻) in a neutral electrolyte.

Electrochemical measurements

Potentiostatic electrolysis was conducted at room temperature (25 ± 1°C) and ambient pressure by means of a potentiostat (Princeton Applied Research PARSTAT 2273 potentiostat/galvanostats with a power booster up to 20 V/20 A and Bio-Logic VMP3 Potentiostats/Galvanostats) recording the current. Current density is expressed as the total current divided by the geometric surface area of the electrodes. The faradaic efficiency is determined by means of the charges consumed to produce a specific product divided by the total charges passed during the electrolysis for 0.5 h. The individual electrode potential during potentiostatic electrolysis was measured as referred to the Ag/AgCl reference electrode which was placed in the middle of the buffer layer. The cell ohmic resistance was measured by electrochemical impedance spectroscopy at open circuit voltage, with amplitude of 5 mV and frequency range of 100 kHz to 0.1 Hz. The total cell ohmic resistance was about 7.3 Ω corresponding to an area specific resistance (ASR) of 29.2 Ω cm².

Product analysis and quantification

Liquid products were quantified by using 1D ¹H nuclear magnetic resonance (NMR) spectroscopy. The NMR data was acquired by using a Varian Mercury/VX 400 MHz spectrometer. More details can be found in our previous work.¹⁹ Gas products were quantified by using gas chromatography (GC, Inficon Micro 3000 GC). The residual reactant and product gases from the outlet of the cathodic compartment were vented directly into the gas-sampling loop of GC. The GC has two channels, one equipped with a packed 10 m MolSieve 5A column and the other with a packed 8 m Plot Q column. Each channel was connected with a thermal conductivity detector. Helium (Airgas, 99.999%) was used as the carrier gas to quantify CO concentration while argon (Airgas, 99.999%) was used to quantify hydrogen concentration. A standard gas mixture consisting of 1000 ppm CO and 1000 ppm H₂ in CO₂ was employed to calibrate the GC and to acquire the gas peak area conversion factor. The moles of CO produced during the electrolysis reaction were calculated from the GC peak area as follows:

where α is the conversion factor based on calibration of the GC with a standard sample, p° is 1.013 bar, T is 273 K, and R is the universal gas constant (8.314 J mol⁻¹ K⁻¹). The flow rate of CO₂ was 41.2 sccm.

Results and Discussion

Role of the liquid-phase buffer layer

Figure 2a shows the trend in faradaic efficiency toward hydrogen evolution in a conventional PEMFC without a buffer layer, in which Pt/C is replaced by Sn at the cathode side. The faradaic efficiency toward hydrogen production was ∼100% when helium was supplied to the cathode. When CO₂ was used instead of helium, hydrogen was the only product observed until a cell potential of −1.8 V. As cell potentials were increased to more cathodic values, the faradaic efficiency toward hydrogen evolution was found to decrease slightly by a few percentages. We used a chiller to condense volatile products in the cathode outlet tube. NMR analysis indicated the presence of formate in the liquid products; however, CO was not observed in GC trace analysis. In the absence of a buffer layer, hydrogen evolution reaction (HER) clearly dominated over the CO₂ reduction reaction. Similar results were also observed on Ag GDE in a PEMFC type electrolysis cell.¹² Moreover, the I-E curve indicated high activity toward HER as shown in Figure 2b. A high current density, ∼1 to 2 A cm⁻² was observed at moderate cell potentials, −1 to −2 V, which was consistent with the low total ASR of about 1.0 Ω cm² between this potential range (Figure 2b).

This dominance of HER over CO₂ electro-reduction can be understood by its favorable thermodynamics and facile kinetics.^20–25 HER in an aqueous medium occurs via two successive steps; the initial discharge of a hydronium ion (H₃O⁺) to give adsorbed hydrogen atom H_ads (Volmer step) followed by combination of two adsorbed H_ads atoms (Tafel step) or adsorbed H_ads with H₃O⁺ in the solution (Heyrovsky step) yielding H₂.^20,22 This competition between HER and CO₂ reduction suggests the importance of controlling the H⁺ concentration near the cathode surface to suppress HER, and thus facilitate CO₂ reduction. We used a buffer layer of circulating aqueous electrolyte to mediate this H⁺ concentration.

Full cell with a buffer layer and hydrogen supplied to anode

When a buffer layer of 0.1 M KHCO₃ was used, the cathode selectivity toward formate formation was significantly improved. Figure 3 shows NMR spectra of ¹H in the formate group as a function of the applied cell potentials. Formate was the only liquid product, which was formed at a cell potential of −0.8 V. This corresponded to an overpotential of −0.19 V with respect to the apparent standard cell potential of −0.61 V when taking into consideration the anodic (H₂ oxidation, Eq. 1) and the cathodic reactions (CO₂ reduction, Eq. 3, apparent standard potential, E°^* = −0.61 V vs. SHE for pH = 7). This low overpotential is mainly attributed to the optimized microstructure of Sn GDEs, which possess a large number of triple-phase boundaries where electrons, protons and CO₂ gases meet and react. Reduction of CO₂ at a lower overpotential is of substantial importance to achieve greater cell performance. Similar overpotential of ∼−0.2 V was recently reported in a cell mediated by ionic liquids, which provided more favorable thermodynamics for CO₂⁻ intermediate formation.^18,26

Figure 4a illustrates the trend in faradaic efficiency for different products as a function of applied cell potential. These results were found reproducible in four independent operations within acceptable range. The faradaic efficiency toward formate formation started with 65% at −0.8 V and increased at more negative cell potentials. The maximum faradaic efficiency was observed ∼90% at −1.7 V. When 0.5 M Na₂SO₄ was used as buffer layer, formate was also formed at −0.8 V with a faradaic efficiency of ∼47%. The utilization of a circulating buffer layer and optimized microstructure of electrode were attributed to achieving a high faradaic efficiency at a low overpotential toward formate formation on Sn GDE by mediating H⁺ concentration near the cathode surface. The presence of H⁺ is essential to form formate, through the protonation of the carbon in the intermediate CO₂⁻, followed by a second electron transfer.²⁷ The excessive H⁺ concentration, however results in preferable hydrogen evolution over CO₂ reduction as observed in the cell without a buffer layer. The faradaic efficiency toward CO formation exhibited a similar trend as formate. The maximum faradaic efficiency of CO was ∼13% at −1.2 V. On the other hand, the faradaic efficiency for hydrogen formation drastically dropped to ∼20% and further decreased below 10% as potentials became more negative.

The energy efficiency is a key factor in determining the economic feasibility of converting CO₂ to fuels by utilizing electricity. The energy efficiency for formate was calculated using Eq. 5,^18,24 as shown in Figure 4a

where the apparent standard cell potential, E_st is −0.61 V, as predicted by Nernst equation. We observed a maximum energy efficiency of 50% for formate at −0.8 V with 0.1 M KHCO₃ as the buffer layer. The energy efficiency toward formate formation decreased at higher cell potentials due to increasing ohmic loss in the membrane and buffer layer.

The plots shown in Figure 4b are partial current densities of formate, CO and H₂ along with the total current density averaged over 30 min of electrolysis as a function of the cell potential. The total current density of the cell with a buffer layer was found to be much lower than that without a buffer layer due to the larger ASR (29.2 Ω cm²) with a buffer layer compared to about 0.8 Ω cm² without a buffer layer. The large ASR with a buffer layer results from the electrolyte resistance of the aqueous buffer solution. The partial current density of formate was ∼−1 mA cm⁻² at −0.8 V which increased at higher cell potentials and reached ∼−9.0 mA cm⁻², corresponding to a production rate of 3.0 μmol min⁻¹ cm⁻² at −1.7 V. The production rate of formate, corresponding to its partial current density, is limited by two factors: sluggish kinetics of CO₂ reduction and a high ohmic loss, which need to be improved in the future design of the cells and electrode materials.

Full cell with a buffer layer and water supplied to anode

Artificial photosynthesis in a photo-electrochemical cell involves the utilization of solar light for water oxidation at the anode side and CO₂ reduction at the cathode side. In this work, we demonstrated the feasibility of simultaneous water oxidation and CO₂ reduction in our cell with a 0.1 M KHCO₃ buffer layer. Specifically, an aqueous 1 M KOH solution and CO₂ were supplied to the anode and cathode, respectively, and the cell was operated in an electrolysis mode. Figures 5a and 5b depict the faradaic efficiency and partial current density as a function of cell potential. We observed the onset of formate production at −1.2 V with a faradaic efficiency of 70% and partial current density of ∼−1 mA cm⁻². A cell potential higher than when the anode was fed with hydrogen gas was needed to overcome the anodic overpotential required for water oxidation. Nevertheless, with respect to the apparent standard cell potential of −1.02 V for simultaneous water oxidation (Eq. 2) and CO₂ reduction (Eq. 3, E°^* = −0.61 V vs. SHE for pH = 7), this overpotential of −0.18 V toward formate production was found to be nearly the same with H₂ supplied to anode, as discussed previously. We observed the highest faradaic efficiency of 85% toward formate formation at −2.0 V. The measured partial current density and corresponding production rate for formate at this potential were −6 mA cm⁻² and 2.0 μmol min⁻¹ cm⁻², respectively. The maximum energy efficiency ∼60% for formate production was obtained at −1.2 V as shown in Figure 5a. The faradaic efficiency toward H₂ production reached a minimum of 8% at −2.2 V, at which the highest faradaic efficiency of 10% toward CO formation was observed.

**Figure 5.** Simultaneous water oxidation and CO₂ reduction in the full electrochemical cell with a circulating aqueous-electrolyte buffer layer. (a) faradaic efficiencies toward the production of formate, CO and H₂, and energy efficiency of formate, (b) total and partial current densities of products as a function of applied cell potential. The anode and cathode were supplied with 1 M KOH (10 mL min⁻¹) and CO₂ (45 mL min⁻¹) at p° = 1.013 bar and T = 298.15 K, respectively. Electrolyte (0.1 M KHCO₃) was circulated at 5 mL min⁻¹.

In order to obtain individual electrode polarization curves, the Ag/AgCl reference electrode was placed in the buffer layer. The individual electrode potential was further compensated by IR drop, as shown in Figure 6. The high overpotentials required for both oxygen evolution and CO₂ reduction, suggests the choice of electrocatalysts holds the key to achieve higher efficiency and current density. We used IrO₂ instead of Pt as catalyst at anode because of its higher activity toward oxygen evolution.^28,29 Although the faradaic efficiency for formate only increased slightly from −1 to −3 V, the total current density was found to increase by about 50% (data not shown here). Further work needs to be done to optimize electrodes for both oxygen evolution and CO₂ reduction, and consequently scale-up this approach to meet the requirements of commercial processes with a typical production rate on the order of 100 μmoles per minute per cm².

**Figure 6.** Individual anode and cathode potentials in the full electrochemical cell with a buffer layer of 0.1 M KHCO₃ when 1 M KOH (10 mL min⁻¹) was supplied to the anode and CO₂ (45 mL min⁻¹) to cathode at p° = 1.013 bar and T = 298.15 K.

Conclusions

A full electrochemical cell with a PEMFC structure modified to include a buffer layer of circulating aqueous electrolyte is designed and assembled for the conversion of CO₂ to formate using a Sn GDE. In the presence of a buffer layer, the CO₂ reduction reaction dominates over HER at the cathode, while HER prevails in the absence of the buffer layer. Using this cell with a buffer layer of 0.1 M KHCO₃, at room temperature and ambient pressure, formate production is accomplished with high faradaic (∼90%) and energy efficiencies (∼60%). Moreover, the formate starts to form at the overpotential of ∼−0.2 V, irrespective of the fuel (H₂ or 1.0 M KOH) supplied to anode with the corresponding partial current density of ∼−1 mA cm⁻². The highest partial current density for formate observed is ∼−10 mA cm⁻², which can be increased by reducing ohmic resistance of the buffer layer and optimizing microstructures of the GDEs. The maximum formate concentration is ∼100 μmol mL⁻¹ at the end of 0.5 h electrolysis. This cell design employs current PEMFC configuration and thus can be easily scaled-up for the potential mass production of formate from CO₂, if the partial current density of formate can be increased to ∼100 mA cm⁻² at a moderate cell potential of −2.0 V.

Acknowledgments

Part of this work is supported by the Solid State and Materials Chemistry Program of the Division of Materials Research at the National Science Foundation. FGR is supported by the South Carolina Honors College Science Undergraduate Research Funding Program.

Electrochemical Reduction of Carbon Dioxide: II. Design, Assembly, and Performance of Low Temperature Full Electrochemical Cells

Article metrics

Permissions

Author affiliations

Author notes

Dates

Abstract

Experimental