Electrochemical reduction of CO₂ in a proton conducting solid oxide electrolyser

Abstract

Synthetic hydrocarbon fuels from CO₂/H₂O are proposed as alternatives to hydrogen as an energy carrier to enable a carbon neutral energy cycle, given their inherent advantages of high H/C ratio and convenience of storage and transportation. Here we demonstrate the successful electrochemical reduction of CO₂ into CO and CH₄ in a proton conducting solid oxide electrolyser based on BaCe_0.5Zr_0.3Y_0.16Zn_0.04O_3 − δ (BCZYZ) electrolyte and a composite iron/iron oxide cathode. The production of CH₄ and CO reaches 0.07 and 3.25 ml min⁻¹ cm⁻², respectively, with 1.5 A cm⁻² at 614 °C. The overall CO₂ conversion rate in the electrochemical reduction process is 65%.

---

Introduction

There is growing evidence for, and acceptance of, the role of the rising level of anthropogenic emissions of CO₂ and other greenhouse gases into the atmosphere in contributing to potentially dangerous climate change.¹ Significant efforts are being devoted to the development of novel, decarbonized forms of energy technologies, such as solid oxide fuel cells^2–5 and CO₂ capture,^6,7 to advance the world from the heavy reliance on finite fossil resources towards an environmentally sustainable energy system. In the meantime, effective energy carriers are needed for the utilization of the new energy source in an efficient way. Hydrogen, with high gravimetric energy density and zero carbon emission widely considered as the most promising candidate, unfortunately still fails to find its widespread use due to the practical engineering and economic limitations with respect to its generation and distribution.^8–12 Synthetic hydrocarbon fuels are proposed to be alternatives to hydrogen, or indeed alternative hydrogen carriers, given their inherent advantages in high H/C ratio and convenience of storage and transportation.⁹ In addition, they match the existing energy infrastructure well because of their similarity to current fossil fuels. Direct generation of hydrocarbon fuels or their precursors (synthesis gas) using CO₂/H₂O would therefore present a feasible way to achieve such effective energy carriers within a carbon neutral cycle.

Proton conducting solid oxide electrolysers (SOEs) as the reverse mode of proton conducting solid oxide fuel cells have attracted considerable attention because they are able to operate in high efficiency in conjunction with renewable electricity and to effectively exploit available heat streams (such as exhaust heat from nuclear plants).^{13, 14} They can efficiently produce pure hydrogen and oxygen through high-temperature steam electrolysis.^15–18 In this process, steam is oxidized in the anode and simultaneously split into oxygen and protons under external electrical potential. Protons diffuse across the proton conducting electrolyte to the cathode where the formation of pure hydrogen from protons takes place at the three-phase boundary. This feature makes it possible to utilize the proton before the formation of hydrogen at the cathode three-phase boundary to in situ electrochemically reduce CO₂ into hydrocarbons in the presence of appropriate metal catalysts under sufficient external loads. Some possible reactions of electrochemical reduction of CO₂ that could take place in the cathode are as follows¹⁹:

2H⁺ + 2e⁻ → H₂ (E^θ = 0 V) (1)

CO₂ + 4H⁺ + 4e⁻ → C + 2H₂O (E^θ=−0.44 V) (2)

CO₂ + 8H⁺ + 8e⁻ → CH₄ + 2H₂O (E^θ = −0.48 V) (3)

CO₂ + 6H⁺ + 6e⁻ → CH₃OH + H₂O (E^θ = −0.62 V) (4)

CO₂ + 4H⁺ + 4e⁻ → HCHO + H₂O (E^θ = −0.72 V) (5)

CO₂ + 2H⁺ + 2e⁻ → CO + H₂O (E^θ = −0.77 V) (6)

CO₂ + 2H⁺ + 2e⁻ → HCO₂H (E^θ = −0.85 V) (7)

Where CH₄ is a thermodynamically favoured hydrocarbon product.

In this paper, we for the first time demonstrate the successful in situ electrochemical reduction of CO₂ to generate hydrocarbon in a proton conducting solid oxide electrolyser as shown in Fig. 1. The cell employs a 60 μm thick BaCe_0.5Zr_0.3Y_0.16Zn_0.04O_3 − δ (BCZYZ)²⁰ membrane as the proton conducting electrolyte and nickel as the anode electrocatalyst. Although a nickel metal anode needs a high concentration of hydrogen to maintain the reducing atmosphere, other ceramic anodes such as the p-type perovskite oxide LSCM²¹ could be adopted for steam oxidization without flowing hydrogen in the cathode. This cell utilises a composite cathode containing iron/iron oxide to support a thick porous iron cathode catalyst straddling an appropriate temperature gradient to facilitate the catalysis of hydrocarbon formation and reduce the decomposition of hydrocarbon products. From the point of thermodynamics, a high concentration of reactant is favorable for positive reaction, so in this work 100% CO₂ was supplied into the cathode to facilitate the electrochemical reduction.

Fig. 1 Schematic of testing conditions.

Experimental

BaCe_0.5Zr_0.3Y_0.16Zn_0.04O_3 − δ (BCZYZ) fine ceramic powders were synthesized by a modified glycine-assisted combustion method as reported by Xie et al.²² Ba(NO₃)₂, Ce(NO₃)₃·6H₂O, ZrO(NO₃)₂·xH₂O, Y(NO₃)₃·6H₂O and ZnO at the stoichiometric ratio for BCZYZ were added into hot distilled water (80 °C) containing an appropriate amount of nitric acid in a beaker under stirring until a transparent and homogeneous solution was obtained. Glycine was then added into the solution under stirring as a complexation/polymerization agent with the glycine : metal mole ratio set at 1.2 : 1. This solution was stirred further on a hotplate at 350 °C until it changed to a brown foam and ignited in a flame. The brown ash was subsequently calcined at 1050 °C for 2 h at the heating rate of 10 °C min⁻¹. The rapid temperature rate is to constrain the overgrowth of BCZYZ powders. XRD (Stoe, CuK_α1, Transmitting mode, 5° min⁻¹) was performed to confirm the phase formation of BCZYZ powders.

The anode powders were prepared by mechanically blending fine BCZYZ powder with green NiO (350 mesh) with the weight ratio of 40 : 60 in a mortar. Then green bi-layer pellets with the configuration of BCZYZ/BCZYZ-NiO (30 mm in diameter and 2 mm in thickness) were fabricated by the dry-pressing method.²² These freshly made pellets were then sintered at 1400 °C for 5 h on an alumina plate with a temperature ramp of 2 °C min⁻¹ in air. The cathode slurry was fabricated by milling Fe₂O₃ and BCZYZ (65 : 35 in weight ratio) in α-terpineol with cellulose additive. The cathode layer was then applied onto the BCZYZ electrolyte surface by the brush-printing method followed by heat treatment at 1000 °C for 2 h in air to assemble a single solid oxide electrolyser. Silver paste was printed onto both electrode surfaces followed by heat treatment at 550 °C (3 °C min⁻¹) for 30 min in air to make current collectors. Silver electrical wire (0.1 mm in diameter) was then connected to both current collectors using silver paste for connection followed by firing at 550 °C (3 °C min⁻¹) for 30 min in air.

The solid oxide electrolyser cell was then sealed onto a testing jig as shown in Fig. 1 using ceramic paste as sealing cement. Porous iron was directly placed onto the cathode to make a composite cathode catalyst straddling an appropriate temperature gradient to facilitate the catalysis of hydrocarbon formation and reduce the decomposition of hydrocarbon products. In order to check the gas leakage of the electrolyser, the open circuit voltage (OCV) was tested before performing electrochemical testing. As shown in Fig. 1, 5% H₂/Ar was supplied into the cathode compartment to convert metal oxide to metal while the anode was exposed to air. After voltage testing, pure argon was supplied to the cathode to prevent iron metal being oxidized. Then 5% H₂/Ar was input to the anode side to reduce the NiO based anode substrate. Then 3% H₂O/H₂ and 100% CO₂ were respectively supplied into the anode and cathode compartments while external loading was also applied to perform the electrochemical test. AC impedance spectroscopy was recorded using a Schlumberger Solartron 1255 Frequency Response Analyzer coupled with a 1287 Electrochemical Interface controlled by Zplot electrochemical impedance software over the frequency range 1 MHz to 100 mHz under different external loads. Current–voltage (I–V) testing and short-term performance under an external load were carried out with a Schlumberger Solartron 1255 controlled by Corrware software with applied current of 1.5 A cm⁻² at 614 °C. Gas flow rate was controlled with Mass flow meters and the output gas of the cathode was connected with a U-type cooling tube in dry ice to eliminate any steam in mixed gases before being supplied into online mass spectroscopy (MS) because a high steam content could destroy the MS analysis system.

Results and discussion

The solid oxide electrolyser was tested under the conditions depicted in Fig. 1 and the microstructure of the cell after testing is shown in Fig. 2 The BCZYZ electrolyte membrane (60 μm in thickness) supported on a NiO-BCZYZ anode substrate is quite uniform and dense. The porous cathode layer in thickness of 40 μm adheres to the electrolyte very well. As presented by the OCV test in the experiment, the OCV reaches 0.999 V when using 5% H₂/Ar to reduce the cathode at ∼600 °C and exposing the NiO-based anode in air, which means the BCZYZ membrane is dense and the sealing is quite good as well because small gas leakage will lead to a significant OCV drop. At this point, the electrochemical cell is indeed a solid oxide fuel cell; and the OCV mainly comes from the electrochemical potential 2H₂ + O₂ → 2H₂O. The OCV quickly dropped to about −0.07 V when the anode was being reduced with 3% H₂O/H₂, relating to the difference in hydrogen concentration across the two electrodes. Then pure CO₂ was supplied into the cathode for electrochemical reduction while a constant current was simultaneously applied to perform the electrochemical process.

Fig. 2 Microstructure of the electrolyser Ni/BaCe_0.5Zr_0.3Y_0.16Zn_0.04O_3 - δ/Fe solid oxide electrolyser after testing.

Fig. 3 shows the measurement of voltage versus current at a set temperature of 614 °C. The absolute open circuit voltage is around −0.065 V coming from the pre-reduced electrodes in absence of significant oxygen flux. The I–V (current–voltage) curve appears non-linear at low voltages with clear changes in slope at around 0.5 and especially 1.0 V. At these potentials hydrogen pumping and perhaps electrode activation are the main processes. At higher potentials steam oxidation and carbon dioxide reduction processes become viable and current rapidly increases. The onset of H₂O oxidation to protons (2H₂O → 4H⁺ + O₂) is anticipated at around 1.2 V with CO₂ reduction occurring at similar potentials considering that there may exist some overpotentials for CO₂ reduction under operation temperature. The linear I–V at high voltage demonstrates a stable cell process, where the detected voltage reaches 2.6 V with a final applied current density of 1.5 A cm⁻².

Fig. 3 I–V curve of the electrolyser operated at 614 °C.

In situ AC impedance spectroscopy was carried out to investigate the cell process with loads ranging from 0 to 2 V as shown in Fig. 4. The R_s mainly arises from the BCZYZ electrolyte and is almost constant; though it slightly decreases with increasing voltage, which could be attributed to increased electronic conduction as well as some oxygen-ion conduction at higher potentials. The R_p from electrode polarization shown in Fig. 4 demonstrates sharp increases from 5.1 to 11.2 Ω cm² with potential rising from 0 to 0.5 V and then decreases to 6.3 Ω cm² at 1.0 V, 3.9 Ω cm² at 1.5 V and 0.4 Ω cm² at 2.0 V, respectively. R_p remains stable above 2.0 V in accordance with the linear I–V curve at high voltages. The rapid R_p increase at low voltage may come from the overpotential associated with activation of the redox processes. Accordingly, the favourable thermodynamics and kinetics at high voltage would result in desirable electrode reactions and a large drop in R_p. The frequency below 10 kHz mainly relates to the electrode process. The low frequency processes seem to be the rate limiting step. These may relate to adsorption and diffusion processes at the electrode. The summit frequency of the low frequency process shifts from 0.5 to 19.9 Hz with the increase of external load from 0.5 to 2.0 V, which demonstrates higher external loads are favorable to the improvement of the low frequency process.

Fig. 4 AC impedance of solid oxide electrolyser with different potentials.

Fig. 5 shows the short-term performance of the electrochemical process with applied current of 1.5 A cm⁻² at a set temperature of 614 °C. The observed voltage is stable and time-independent, indicating a steady electrochemical process. Table 1 shows the gas contents of cathode input and output. Methane was directly produced at the rate of 0.07 ml min⁻¹ cm⁻² with content of 1.2% in a single step, implying the successful electrochemical reduction of CO₂ to hydrocarbon. CO with concentration of 61% was produced at the rate of 3.25 ml min⁻¹ cm⁻² as well; and the corresponding CO₂ conversion is around 65%, which confirms the effective reduction of CO₂ in the electrolyser. Hydrogen was produced at the rate of 0.43 ml min⁻¹ cm⁻² as well and the content was about 8%, which demonstrates that 88.3% of the proton flux is efficiently utilized to reduce CO₂ while 11.7% is released in the form of hydrogen. However, the methane yield is much lower than CO, indicating CO generation is more competitive than methane formation though the overpotential of CO formation is larger than that of methane formation. Kinetic limitations and methane cracking at such temperature may together reduce the degree of methane generation. The high concentration of CO₂ in the gas input may be advantageous for the high efficiency of proton utilization. The Faradaic yield of CO, H₂ and methane reach 29.5%, 4.0% and 2.4%, respectively. The Faradaic losses may well arise from parasitic losses due to transport of other species such as holes or oxide ions as would normally be expected for a proton conducting perovskite operating at temperatures of 610 °C. The OCV in fuel cell mode at this temperature was only about 90% of that expected for a pure protonic conductor, consistent with prior studies of protonic fuel cells.²² In addition, no serious carbon deposition such as carbon powder in the cathode, from the reduction of CO₂ or CO to carbon, was observed though we did observe some black coloration, consistent with electrolyte reduction.

Fig. 5 Cell voltage vs. time at applied 1.5 A cm⁻² current at 614 °C.

Table 1 Content of gases detected with online mass spectroscopy^a

Electrochemical reduction	CO2	CH4	CO	H2
a — denotes no data available.
Input	100%; 5 ml min^−1 cm^−2	—	—	—
Output	29%; 1.75 ml min^−1 cm^−2	1.2%; 0.07 ml min^−1 cm^−2	61%; 3.25 ml min^−1 cm^−2	8%; 0.43 ml min^−1 cm^−2
Conversion	65%	—	—	—
Faraday yield	—	2.4%	29.5%	4.0%
Fischer–Tropsch synthesis	CO2	CH4	CO	H2
Input	58%; 5.0 ml min^−1 cm^−2	—	—	42%; 3.62 ml min^−1 cm^−2
Output	51%; 3.15 ml min^−1 cm^−2	0.7%; 0.04 ml min^−1 cm^−2	29%; 1.85 ml min^−1 cm^−2	19%; 1.0 ml min^−1 cm^−2
Conversion	37%	—	—	—

---

As a control for the electrolysis study, a mixture of CO₂ and H₂ was supplied to the cell cathode without current under equivalent conditions. The amount of hydrogen input was calculated according to the generated proton in the above electrochemical process. As shown in Table 1, methane generation in this process only reaches 0.7% via Fischer–Tropsch synthesis (FTS),²³ which is lower than the value through the electrochemical reduction described above. In addition, the conversion of CO₂ is 37% through reverse water gas shift reaction²⁴ and up to 18% hydrogen is left as well in mixed gases. It means that for the same materials composition and morphology, methane generation and CO₂ reduction in the electrolyser mode is more effective than conventional Fischer–Tropsch synthesis and water gas shift reaction, respectively.

Conclusion

In this work, we demonstrated the successful electrochemical reduction of carbon dioxide into methane and carbon monoxide in a proton conducting solid oxide electrolyser. The high conversion of CO₂ (65%) demonstrated efficient reduction in an electrochemical cell with an applied current of 1.5 A cm⁻² at 614 °C. The higher content of carbon monoxide (3.25 ml min⁻¹ cm⁻²) than methane (0.07 ml min⁻¹ cm⁻²) also shows more competitive formation of carbon monoxide than methane in the overall process. Control experiments show that the electrochemical reduction of carbon dioxide is more efficient than reverse water gas shift reaction under the same geometry and compositions. Although the methane yield requires some development work, effective improvements could be implemented by lowering operation temperature, utilizing large currents and improving cell performance. The cycling of CO₂ would also open a new pathway for a carbon neutral renewable energy cycle.

Acknowledgements

The authors wish to thank EPSRC and Carbon Trust for Platform, Senior Fellowship and Carbon Vision awards and the China scholarship council for support.

References

1. K. M. K. Yu, I. Curcic, J. Gabriel and S. C. E. Tsang, ChemSusChem, 2008, 1, 893–899.
2. A. Atkinson, S. Barnett, R. J. Gorte, J. T. S. Irvine, A. J. Mcevoy, M. Mogensen, S. C. Singhal and J. Vohs, Nat. Mater., 2004, 3, 17–27.
3. S. W. Tao and J. T. S. Irvine, Adv. Mater., 2003, 2, 320–323.
4. J. C. Ruiz-Morales, J. Canales-Vazquez, C. Savaniu, D. Marrero-Lopez, W. Z. Zhou and J. T. S. Irvine, Nature, 2006, 439, 568–571.
5. S. W. Tao, J. T. S. Irvine and J. A. Kilner, Adv. Mater., 2005, 17, 1734–1737.
6. F. Bidrawn, G. Kim, G. Corre, J. T. S. Irvine, J. M. Vohs and R. J. Gorte, Electrochem. Solid-State Lett., 2008, 11, B167–B170.
7. C. M. White, B. R. Strazisar, E. J. Granite, J. S. Hoffman and H. W. Pennline, J. Air Waste Manag. Assoc., 2003, 53, 645–715.
8. C. S. Song, Catal. Today, 2006, 115, 2–32.
9. A. Midilli and I. Dincer, Int. J. Hydrogen Energy, 2008, 33, 4209–4222.
10. L. Schlapbach, Nature, 2009, 460, 809–811.
11. L. F. Wang and R. T. Yang, Energy Environ. Sci., 2008, 1, 268–279.
12. B. Sakintuna, F. Lamari-Darkrim and M. Hirscher, Int. J. Hydrogen Energy, 2007, 32, 1121–1140.
13. N. Q. Minh, J. Am. Ceram. Soc., 1993, 76, 563–588.
14. J. S. Herring, J. E. O'Brien, C. M. Stoots, G. L. Hawkes, J. J. Hartvigsen and M. Shahnam, Int. J. Hydrogen Energy, 2007, 32, 440–450.
15. K. D. Kreuer, Annu. Rev. Mater. Res., 2003, 33, 333–359.
16. H. Iwahara, Solid State Ionics, 1998, 28, 573–578.
17. R. Haugsrud and T. Norby, Nat. Mater., 2006, 5, 193–196.
18. P. Colomban, Ann. Chim. Sci. Mater., 1999, 24, 1–18.
19. A. J. Morris, G. J. Meyer and E. Fujita, Acc. Chem. Res., 2009, 12, 1983–1994.
20. S. W. Tao and J. T. S. Irvine, Adv. Mater., 2006, 18, 1581–1584.
21. S. W. Tao and J. T. S. Irvine, Nat. Mater., 2003, 2, 320–323.
22. K. Xie, R. Q. Yan, D. H. Dong, S. L. Wang, X. R. Chen, T. Jiang, B. Lin, M. Wei, X. Q. Liu and G. Y. Meng, J. Power Sources, 2008, 179, 576–583.
23. G. P. van der Laan and A. A. C. M. Beenackers, Catal. Rev. Sci. Eng., 1999, 41, 255–318.
24. H. D. A. L. Viana and J. T. S. Irvine, Solid State Ionics, 2007, 178, 717–722.

---

Footnote

† Electronic supplementary information (ESI) available: Testing Schematic, half-cell microstructure and mass spectroscopy. See DOI: 10.1039/c0jm02205e

---