1. Department of Materials Science and Engineering, Northwestern University, Evanston, Illinois 60208, USA

Correspondence to: S. A. Barnett¹ Correspondence and requests for materials should be addressed to S.A.B. (e-mail: Email: s-barnett@nwu.edu).

Earlier studies of solid oxide fuel cells (SOFCs) have investigated the potential for both the direct use of hydrocarbons and internal reforming (conversion of the hydrocarbon to hydrogen at the SOFC anode). Internal reforming is generally achieved by using nickel-based anodes to catalyse the endothermic reforming reactions. In the case of methane, these are:

where H⁰₂₉₈ is the standard enthalpy of the reaction at 298 K. However, the challenges in maintaining appropriate gas composition and temperature gradients across large-area SOFC stacks make it difficult to achieve 100% internal reforming⁸,⁹. Instead, a partial external pre-reforming stage is typically used, followed by internal reforming within the stack¹⁰. The direct electrochemical oxidation of hydrocarbon fuels in SOFCs is, in principle, possible, but has resulted in carbon deposition³,⁵ atoperating temperatures above about 800 °C, and low power densities³,⁴ (10 mW cm^-2) at temperatures below 800 °C. But recently developed SOFCs produce high power densities by using hydrogen fuel at 600–800 °C (refs 6, 7, 11–15) and may thus have the potential to provide good power densities by directly using hydrocarbons in a temperature range where carbon deposition can be avoided.

The SOFCs we report here were fabricated on porous La_0.8Sr_0.2MnO₃ (LSM) cathodes. The LSM pellets were 2 cm indiameter and 1 mm thick, and were produced using standard ceramic processing techniques. All SOFC layers, starting with a 0.5-m-thick (Y₂O₃)_0.15(CeO₂)_0.85 (YDC) porous film, were deposited on the LSM pellet using d.c. reactive magnetron sputtering. The electrolyte, 8 mol% Y₂O₃-stabilized ZrO₃ (YSZ), was then deposited under conditions yielding a dense, 8-m-thick film. To complete the cell, another 0.5-m-thick YDC film was deposited, followed by a porous, 2-m-thick Ni-YSZ anode. Although sputtering was used in this study, similar cells can be made using other methods⁶,^{11, 12, 13}. The SOFCs and their performance on hydrogen fuel are described in detail elsewhere⁷. Anode reactions were studied using impedance spectroscopy with anodes that were sputter-deposited onto both sides of bulk YSZ single-crystal electrolytes.

Figure 1 shows measurements of current density and power density versus voltage, performed on a typical cell using air and methane. The results for dry and wet (with 3% H₂O) methane were nearly identical. Cell performance was stable in our preliminary 100-h life tests, except for wet methane at low voltages where the anode Ni gradually oxidized. The results in Fig. 1 are similar to those obtained for the same cells operated with humidified hydrogen fuel, except that the power densities are 20% lower⁷. The previous studies with hydrogen fuel⁷ showed that the current densities were limited primarily by cathode overpotential. Examination of the anodes after the cell tests (by visual observation, energy dispersive X-ray, and scanning electron microscopy) showed no evidence of carbon deposition after 100 h of operation.

Figure 1: Cell voltage and power density versus current density for an SOFC operated on air and wet methane.

The measurements were collected in atmospheric-pressure air, and the methane fuel was supplied at 50 cm³ STP min^-1.

High resolution image and legend (12K)

Successful cell operation on dry methane indicated that direct electrochemical oxidation,

was the primary anode reaction. The Gibbs free energy for reaction (3) is given by G ⁰. Methane reforming (reactions (1) and (2) may have also played a role in the cell operation, but only after H₂O and CO₂ were produced through reaction (3). Even after H₂O and CO₂ production, reforming rates were probably too low to compete with the primary anode reaction, because the small anode area (1 cm²), low temperature⁸ and flushing of products (H₂O and CO₂) from the anode compartment by the relatively high fuel flow rates suppress reactions (1) and (2).

More evidence of direct methane oxidation is provided by the different impedance spectra (Fig. 2) obtained from these Ni-YSZ/YDC anodes in humidified methane (trace a), humidified dilute H₂(trace b) and humidified H₂ (trace c). The 3%H₂ + 3%H₂O + 94%Ar mixture used to obtain trace b simulates a slightly reformed methane fuel. This mixture yielded an electrode spectrum with a shape different from that of the methane spectrum (a). Since the shape of the spectra are related to the mechanisms governing the fuel reaction, this result suggests that the primary anode reaction with methane was not oxidation of hydrogen produced by reforming. Trace c shows that the electrode resistance was much smaller for H₂ than for CH₄, explaining why the SOFC current densities were higher for hydrogen than methane.

Figure 2: Comparison of electrode impedance spectra for Ni-YSZ/YDC anodes.

Spectra were measured at 600 °C in 97%CH₄ + 3% H₂O (trace a), 3%H₂ + 3%H₂O + 94%Ar (trace b) and 97%H₂ + 3%H₂O (trace c).

High resolution image and legend (7K)

The rapid direct electrochemical oxidation of methane at these temperatures was due to the anodes employed in these SOFCs¹⁶, which combined Ni-YSZ and YDC layers. Impedance measurements taken in humidified methane for Ni-YSZ/YDC and Ni-YSZ anodes (Fig. 3) indicate that the YDC layer caused the interfacial resistance to reduce by a factor of 6. This agrees with previous studies that indicate that ceria promotes hydrocarbon oxidation¹⁷. Ceria is beneficial for several reasons. First, it becomes a mixed conductor in the reducing fuel environment¹⁴, a condition which should expand the reaction zone beyond three-phase boundaries. Second, the ionic conductivity of ceria is higher than that of YSZ, which improves the transport of oxygen ions from the electrolyte tothe anode. Third, ceria is known to readily store and transfer oxygen, and adding zirconia enhances the storage capability¹⁸. The present anodes have a ceria layer and two ceria/zirconia interfaces where enhanced oxygen storage may increase methane oxidation rates.

Figure 3: Impedance spectroscopy results for Ni-YSZ and Ni-YSZ/YDC anodes in 97% CH₄ + 3%H₂O at 600 °C.

The interfacial resistance corresponds to the difference between the real-axis intercepts.

High resolution image and legend (7K)

An important result from the cell tests was the absence of carbon deposition. In general, carbon deposition can occur by methane pyrolysis,

or carbon monoxide disproportionation,

The role of methane pyrolysis was tested by flowing pure methane over SOFC anodes without operating the SOFC, so that no reaction products were present. No carbon deposition was observed at <700 °C: that is, there was no methane pyrolysis. The amount of carbon deposited increased with increasing temperature above 700 °C, but less carbon deposition was observed at a given temperature on Ni-YSZ/YDC anodes than on Ni-YSZ.

The present results are for nearly pure methane, as the small-area anodes and relatively high fuel flow rates resulted in small concentrations of reaction products. The situation that would be encountered in a methane SOFC stack is more complicated, as the products of reactions (1)–(3) would be present at substantial partial pressures, allowing carbon deposition by reaction (5). Although tests onSOFC stacks were beyond the scope of this study, we have performed a simple equilibrium calculation to determine the conditions where one would expect carbon-deposition-free stack operation in CO–CO₂ mixtures (Fig. 4). The results show that low CO/CO₂ ratios suppress carbon deposition through reaction (5). The optimal temperature range for SOFC stack operation on dry methane is 500–700 °C. At temperatures below this range, carbon deposition through reaction (5) will occur unless CO/CO₂ ratios are kept very low. Temperatures 700 °C would tend to suppress carbon deposition by reaction (5), but would allow carbon deposition by direct pyrolysis of methane (equation (4)). We also note that carbon deposition may be suppressed by oxygen ions reacting with carbon at the anode during SOFC operation⁵. Finally, we point out that some internal reforming would be necessary in a SOFC stack to produce a small amount of CO and H₂. These compounds would balance the CO₂ and H₂O produced by direct oxidation, preventing the too-low CO/CO₂ and H₂/H₂O ratios at which the nickel anode may oxidize.

Figure 4: Calculated CO/CO₂ ratios at equilibrium with graphitic carbon, based on reaction (5).

High resolution image and legend (17K)

References

1. Ren, X. , Wilson, M. S. & Gottesfeld, S. High performance direct methanol polymer electrolyte fuel cells. J. Electrochem. Soc. 143, L12–L14 (1996). | ISI | ChemPort |
2. Wang, J. , Wasmus, S. & Savinell, R. F. Evaluation of ethanol, 1-propanol, and 2-propanol in a direct oxidation polymer-electrolyte fuel cell. A real-time mass spectrometry study. J. Electrochem. Soc. 142, 4218–4224 (1995). | ISI | ChemPort |
3. Sfeir, J. , Van herle, J. & McEvoy, A. J. in Proc. 3rd European Solid Oxide Fuel Cell Forum (ed. Stevens, P.) 267–276 (European Fuel Cell Forum, Switzerland, 1998).
4. Putna, E. S. , Stubenrauch, J. , Vohs, J. M. & Gorte, R. J. Ceria-based anodes for the direct oxidation of methane in solid oxide fuel cells. Langmuir 11, 4832–4837 (1995). | ISI | ChemPort |
5. Aida, T. , Abudala, A. , Ihara, M. , Komiyama, H. & Yamada, K. in Proc. 4th Int. Symp. on Solid Oxide Fuel Cells (eds Dokiya, M., Yamamoto, O., Tagawa, H. & Singhal, S. C.) 801–809 (Electrochemical Soc., Pennington, 1995).
6. Fung, K-Z. , Chen, J. , Tanner, C. & Virkar, A. V. in Proc. 4th Int. Symp. on Solid Oxide Fuel Cells (eds byDokiya, M., Yamamoto, O., Tagawa, H. & Singhal, S. C.) 1018–1027 (Electrochemical Soc., Pennington, 1995).
7. Tsai, T. & Barnett, S. A. Increased solid-oxide fuel cell power density using interfacial ceria layers. Solid State Ionics 98, 191–196 (1997). | Article | ISI | ChemPort |
8. Janssen, G. J. M. , DeJong, J. P. & Huijsmans, J. P. P. in Proc. 2nd European Solid Oxide Fuel Cell Forum (ed. Thorstense, B.) 163–172 (European SOFC Forum, Switzerland, 1996).
9. Hendriksen, P. V. in Proc. 5th Int. Symp. on Solid Oxide Fuel Cells (eds Stimming, U., Singhal, S. C., Tagawa, H. & Lehnert, W.) 1314–1325 (Electrochemical Soc., Pennington, 1997).
10. Ray, E. R. in 1992 Fuel Cell Seminar Abstr. (eds Fuel Cell Seminar Committee) 415–418 (Courtesy Associates, Washington, 1992).
11. Huebner, W. , Anderson, H. U. , Reed, D. M. , Sehlin, S. R. & Deng, X. in Proc. 4th Int. Symp. on Solid Oxide Fuel Cells (eds Dokiya, M., Yamamoto, O., Tagawa, H. & Singhal, S. C.) 696–705 (Electrochemical Soc., Pennington, 1995).
12. deSouza, S. , Visco, S. J. & DeJonghe, L. C. Thin-film solid oxide fuel cell with high performance at low-temperature. Solid State Ionics 98, 57–61 (1997). | Article | ISI | ChemPort |
13. Minh, N. Q. in Proc. 4th Int. Symp. on Solid Oxide Fuel Cells (eds Dokiya, M., Yamamoto, O., Tagawa, H. & Singhal, S. C.) 138–145 (Electrochemical Soc., Pennington, 1995).
14. Godickemeier, M. , Sasaki, K. & Gauckler, L. J. in Proc. 4th Int. Symp. on Solid Oxide Fuel Cells (edsDokiya, M., Yamamoto, O., Tagawa, H & Singhal, S. C.) 1072–1081 (Electrochemical Soc., Pennington, 1995).
15. Tsai, T. , Perry, E. & Barnett, S. Low-temperature solid-oxide fuel cells utilizing thin bilayer electrolytes. J. Electrochem. Soc. 144, L130–L132 (1997). | ISI | ChemPort |
16. Tsai, T. & Barnett, S. A. Effect of mixed-conducting interfacial layers on solid oxide fuel cell anode performance. J. Electrochem. Soc. 145, 1696–1701 (1998). | ISI | ChemPort |
17. Steele, B. C. H. , Middleton, P. H. & Rudkin, R. A. Material science aspects of SOFC technology with special reference to anode development. Solid State Ionics 40/41, 388–393 (1990). | Article |
18. Hori, C. E. et al. Thermal stability of oxygen storage properties in a mixed CeO₂-ZrO₂system. Appl. Catal. B 16, 105–117 (1998). | Article | ISI | ChemPort |

Acknowledgements

E.P.M. was supported by the Illinois Minority Graduate Incentive Program.