Methane Oxidation at Redox Stable Fuel Cell Electrode La0.75Sr0

The analysis of the reactor effluent was performed with an on-line HP ... Samples were prepared by milling with graphite in a 15:100 wt % ratio and ...
0 downloads 0 Views 155KB Size
J. Phys. Chem. B 2006, 110, 21771-21776

21771

Methane Oxidation at Redox Stable Fuel Cell Electrode La0.75Sr0.25Cr0.5Mn0.5O3-δ Shanwen Tao,* John T. S. Irvine,* and Steven M. Plint School of Chemistry, UniVersity of St. Andrews, Fife KY16 9ST, Scotland, United Kingdom ReceiVed: April 18, 2006; In Final Form: August 21, 2006

Because of its widespread availability, natural gas is the most important fuel for early application of stationary fuel cells, and furthermore, methane containing biogases are one of the most promising renewable energy alternatives; thus, it is very important to be able to efficiently utilize methane in fuel cells. Typically, external steam reforming is applied to allow methane utilization in high temperature fuel cells; however, direct oxidation will provide a much better solution. Recently, we reported good electrochemical performance for an oxide anode La0.75Sr0.25Cr0.5Mn0.5O3 (LSCM) in low moisture (3% H2O) H2 and CH4 fuels without significant coking in CH4. Here, we investigate the catalytic activity of this oxide with respect to its ability to utilize methane. This oxide is found to exhibit fairly low reforming activity for both H2O and CO2 reforming but is active for methane oxidation. LSCM is found to be a full oxidation catalyst rather than a partial oxidation catalyst as CO2 production dominates CO production even in CH4-rich CH4/O2 mixtures. X-ray adsorption spectroscopy was utilized to confirm that Mn was the redox active species, clearly demonstrating that this material has the oxidation catalytic behavior that might be expected from a Mn perovskite and that the Cr ion is only present to ensure stability under fuel atmospheres.

1. Introduction

2CH4 + O2 ) 2CO + 4H2

(4)

The solid oxide fuel cell (SOFC) is a highly promising system for power generation. It is an all-ceramic device operating at high temperatures.1 Both hydrogen and hydrocarbons are proposed as fuels for SOFCs, and unlike hydrocarbons, hydrogen itself is essentially an energy vector and not a primary energy source. Among the hydrocarbons, methane, which is the main component of natural gas, or indeed biogas, is of particular importance. A much higher efficiency of fuel cell conversion of methane would provide a dramatic reduction in CO2 production. When methane is used as the fuel, the cracking of CH4 (reaction 1) on the traditional Ni-YSZ anode is problematic due to the formation of carbon that may block the catalytic sites and the pores for gas diffusion resulting in the degradation of fuel cell performance.

CH4 + 2O2 ) CO2 + 2H2O

(5)

CH4 ) C + 2H2

(1)

One method to suppress carbon deposition is the gradual internal reforming of methane by steam or CO2 yielded at the SOFC anode side according to reactions 2 and 3.2

CH4 + H2O ) CO + 3H2

(2)

CH4 + CO2 ) 2CO + 2H2

(3)

Carbon deposition can be avoided if sufficient methane is directly oxidized by oxygen transported from the cathode side under the load.3 The partial oxidation of methane may yield CO and H2 by reaction 4, which could in turn be used as the fuel. Complete oxidation of methane under SOFC conditions to produce CO2 and H2O (reaction 5) would be ideal, although it is usually difficult to achieve. * Corresponding authors. (S.T.) Tel.: +44 1334 463680; fax: +44 1334 463808; e-mail: [email protected]. (J.T.S.I.) Tel.: +44 1334 463817; fax: +44 1334 463808; e-mail: [email protected].

The improved oxidation of hydrocarbons with new electrodes in high temperature fuel cells has been reviewed before.4-6 Among the potential anode materials, LaCrO3-based perovskites are promising due to their relatively good stability in a reducing atmosphere at high temperatures7,8 and good catalytic properties for methane reforming at high temperatures.2,9,10 The introduction of other transition elements onto the B-site in La1-xSrxCr1-yMyO3 (M ) V, Cr, Mn, Fe, Co, Ni, Cu, etc.) has the potential to improve the methane reforming catalytic properties, as reported by Sfeir et al., using 10 mol % transition element doped lanthanum chromite.9 The yielded CO from steam- or CO2 reforming may be used as the fuel, but it may not be completely oxidized into CO2 (reaction 6).

2CO + O2 ) 2CO2

(6)

Unoxidized CO may cause local pollution as CO is very toxic. It would be much better if we could find a complete oxidation catalyst for methane that only generates CO2 without emitting toxic CO. Complete oxidation may also improve the fuel utilization and energy efficiency. Some perovskite oxides LaMO3 (M ) Fe, Mn) are actually known to be good methane oxidation catalysts.11-13 LaMnO3 also exhibits a good catalytic effect for CO oxidation.13,14 CO was completely converted to CO2 below 550 K for samples with more than 4% LaMnO3 loading on a zirconia support when the ratio of CO to O2 was 1:20. Manganese oxide is of particular interest because Mn promotes electrochemical reactions and as an electrocatalyst coated on a YSZ electrolyte reduces overpotentials.15 LaMnO3 itself is unstable in a reducing atmosphere at high temperatures; therefore, it cannot be directly used as an SOFC anode, but

10.1021/jp062376q CCC: $33.50 © 2006 American Chemical Society Published on Web 10/03/2006

21772 J. Phys. Chem. B, Vol. 110, No. 43, 2006 LaCrO3 is stable over a wide pO2 range at high temperature.7 The combination of LaCrO3 and LaMnO3 has given a good anode that seems to satisfy both redox stability and catalysis requirements. Recently, we reported that (La0.75Sr0.25)Cr0.5Mn0.5O3 (LSCM) is a redox stable, efficient SOFC anode using either hydrogen or methane as fuel.16-19 The polarization resistance is rather small using yttria stabilized zirconia (YSZ) as the electrolyte. The anode polarization is even smaller when perovskite-based electrolytes are used, probably due to a better anode/electrolyte interface.19 The main purpose of this study is to improve our understanding of how this perovskite anode functions, and it was found that LSCM is a complete methane oxidation catalyst under most conditions. Most reports on new anodes have focused on the reforming of methane, based on the idea that the generated H2 and CO may be further oxidized. Conversely, it was reported that direct utilization of hydrocarbons may be achieved over a Cu-CeO2YSZ anode.3,20 In later reports, it was found that carbon may enhance the anode performance due to better connectivity between anode particles, which improves local electronic transport. Carbon deposition has been observed when passing methane over a Cu-Νi cermet anode or butane over a CuCeO2-YSZ anode.21,22 Therefore, it is unclear whether full direct oxidation occurred over the Cu-CeO2-YSZ anode or simply a net oxidation process possibly involving cracking as an intermediate step. In this paper, it is clearly shown that LSCM is a good complete methane oxidation catalyst through catalytic tests. 2. Experimental Procedures Materials for the catalytic test were prepared by a combustion synthesis process.18 The catalyst was fired at 1100 °C for 4 h because this is the minimum required temperature to obtain pure La0.75Sr0.25Cr0.5Mn0.5O3-δ (LSCM) perovskite phase and is also pertinent to the minimum requirement to apply such electrodes to an SOFC device. The specific surface areas of the LSCM sample was 4.59 ( 0.14 m2/g measured by the BET method on an IGA Intelligent Gravimetric Analyzer using nitrogen as the test gas. Catalytic activity experiments were carried out on a fixedbed reactor using 120 mg of catalyst. The reactor was made of quartz with a sintered glass filter. The filter was P40 with a pore size of 16-40 µm. The catalyst was put on the filter with gas flow from the bottom. A standard gas mixture (Scotty II analyzed gases, MIX234) was used to calibrate the gas chromatography. The mole percentage of the standard mixture was 5.02% CO2, 4.99% CO, 4.00% H2, 3.99% CH4, 4.99% N2, and 5.00% O2 balanced by He. The flow rates of the reactive gases were controlled by a mass flow-meter. The flow rate for the helium carrier was fixed at 36 mL/h. Water was added into the system by a microsyringe pump with a 5.6 mm diameter. Some heating tapes were wrapped around the pipe for steam to avoid water condensation. The analysis of the reactor effluent was performed with an on-line HP chromatograph using a thermal conduction detector (TCD) and a column CarboPlot P7 fused with a silica PLOT 25 m × 0.53 mm. The data were analyzed by HP Chemstation software. The catalysts were directly exposed to reactive gases without pretreatment. The conversion and selectivity results under different conditions were obtained by analyzing gas chromatographic data. All the CH4, CO, CO2, H2, and O2 gases were directly analyzed from experimental data. For comparison, blank tests were carried out with different CH4/ O2 ratios at 700 and 900 °C, respectively. X-ray absorption spectrometry (EXAFS) spectra were collected using the synchrotron radiation source at Daresbury

Tao et al.

Figure 1. Methane conversion and gas formation rate over LSCM as a function of temperature with CH4/H2O ) 1:1. Flow rate: CH4/H2O/ He ) 12:12:36 mL/h.

Laboratories on station 9.3., which used a double-crystal harmonic rejecting Si220 monochromator. Standard ion chambers filled with Ar (20% absorbing Io) and (80% absorbing It) backfilled to 1000 mbar with He were used for the detection of incident and transmitted flux. Samples were prepared by milling with graphite in a 15:100 wt % ratio and pressing into pellets. La1-xSrxCr0.5Mn0.5O3-δ with x ) 0.2 and 0.3 were also synthesized for EXAFS tests. 3. Results and Discussion The catalytic capability of LSCM for steam reforming of methane was first tested at different temperatures with a fixed molar ratio of CH4/H2O 1:1. (Figure 1). The steam reforming of methane (reaction 2) started above 700 °C for LSCM. However, the catalytic effect of LSCM for this reaction is fairly insignificant. Only 2.2% methane was converted at a temperature as high as 900 °C. As mentioned previously, steam reforming is only one of the possibilities that can occur under SOFC anode conditions when methane is used as the fuel. This test on its own should not lead to the conclusion that LSCM is not a good anode for a direct methane fuel cell, and indeed, it has been demonstrated that LSCM is a good anode for SOFCs.16,19 CO2 reforming and/or CH4 oxidation may thus play an important role in the LSCM anode process. To investigate the CO2 reforming of methane over the LSCM anode as a function of temperature, the ratio of methane to CO2 was fixed to 1:1 (Figure 2a). No significant reactions occurred between the two gases over LSCM below 600 °C. Methane conversion increased with increasing temperature. The expected product hydrogen according to reaction 3 was not observed. The yielded H2 from reaction 3 may further react with CO2 to form CO and water as reaction 7

CO2 + H2 ) CO + H2O

(7)

Therefore, the general reaction would be

CH4 + 3CO2 ) 4CO + 2H2O

(8)

Thus, the final products of the CO2 reaction with methane would be steam and CO without hydrogen (reaction 8) if the gas shift reaction 7 was kinetically favored when a suitable catalyst such as LSCM was applied. If reforming was more important, then reaction 3 would have dominated. In general, the yielded steam from the gas shift reaction could further steam reform methane according to reaction 2, but for LSCM, the steam reforming of methane is not very high, so

Fuel Cell Electrode La0.75Sr0.25Cr0.5Mn0.5O3-δ

J. Phys. Chem. B, Vol. 110, No. 43, 2006 21773

Figure 2. (a) Methane conversion and gas formation rate over LSCM as a function of temperature when CH4/CO2 ) 1:1. Flow rate: CH4/ CO2/He ) 12:12:36 mL/h. (b) The methane conversion and gas formation rate over LSCM as a function of CO2 content at 900 °C. The CH4 and He flow rates were fixed at 12 and 36 mL/h, respectively.

this will not be important. CO2 conversion of methane only reaches 7.8% at 900 °C. To investigate the methane conversion as a function of CO2 to methane ratio, the methane conversion at different CO2 reactant contents at 900 °C is shown in Figure 2b. All the reactant contents displayed in this paper are the relative content as compared to methane excluding the carrier gas helium. Methane conversion linearly increases with increasing CO2 reactant content. The more CO2, the easier the reaction is according to reaction 8. The reaction equilibrium moves to the right as more reactant CO2 is introduced into the system. In the anode environment of a working solid oxide fuel cell, oxygen is transported from the cathode to the anode under polarization. The reactions between methane and oxygen are therefore very important, although passing gas mixtures over a catalytic bed does not exactly replicate the anodic condition, but it is perhaps the best approximation that can be achieved. Figure 3a shows the methane conversion as a function of temperature when pCH4/pO2 is fixed at 4:1. No reaction occurred at a temperature lower than 500 °C. The methane conversion increased to 3.1% at 550 °C. The oxygen conversion reached 96% at 600 °C. Above 650 °C, no oxygen was detected in the outlet gases from the reactor, indicating that all the oxygen reacted with methane. The reaction products may be CO and/ or CO2 according to reactions 4 and 5. Hydrogen and CO were not detected below 800 °C, indicating that complete oxidation according to reaction 5 occurred under those conditions. The appearance of small amounts of CO but without H2 above 800 °C indicating the partial oxidation of methane is as follows:

2CH4 + 3O2 ) 2CO + 4H2O

(9)

The measured methane conversion is close to 12.5%, the

Figure 3. (a) Conversion and gas formation rate over LSCM as a function of temperature when CH4 to O2 ratio is 4:1. Flow rate: CH4/ O2/He ) 12:3:36 mL/h. (b) CO and CO2 selectivity as a function of temperature. CH4/O2 ) 4:1.

equilibrium conversion when all oxygen was consumed according to reaction 5. To be an efficient SOFC anode, complete oxidation without the toxic CO would be beneficial. The CO and CO2 selectivities are calculated through the relative contents of the yielded CO and CO2.10 As shown in Figure 3b, the CO2 selectivity is 100% at 800 °C and below, indicating complete oxidation when the ratio of methane to oxygen is 4:1. When the temperature increases to 850 °C, the selectivity for CO is still only 1.4%. The higher the temperature, the more CO yielded, but it is still relatively small as compared to CO2. LSCM seems therefore to be a complete methane oxidation catalyst at 800 °C and below with some partial oxidation at higher temperatures. In a fuel cell, the oxygen partial pressure changes across the anode. The pO2 would be rather low at the fuel inlet but gradually increases toward the fuel outlet with the consumption of fuel along the anode. Therefore, it is necessary to study the methane conversion as a function of the CH4/O2 ratio. Figure 4a shows the degree of methane conversion with the increasing content of oxygen at 900 °C. As would be expected, methane conversion increases with increasing oxygen content. All the oxygen reacted with methane at an oxygen relative content less than 60%. All methane was oxidized when pCH4/pO2 ) 3:7 or lower. In this case, some oxygen was left unreacted; therefore, the oxygen conversion dropped. To completely oxidize methane at LSCM, the final effective methane-to-oxygen ratio should be larger than 3:7 at 900 °C. The CO and CO2 selectivities are shown in Figure 4b. The less oxygen introduced into the system, the more CO was observed, indicating partial oxidation. According to reaction 4, the yielded products would be CO and H2. However, hydrogen is not observed in the process even when the relative oxygen content is only 10% indicating LSCM is a

21774 J. Phys. Chem. B, Vol. 110, No. 43, 2006

Tao et al.

Figure 4. (a) Conversion and gas formation rate over LSCM as a function of oxygen content at 900 °C. The CH4 and He flow rate were fixed at 12 and 36 mL/h, respectively. (b) CO and CO2 selectivity as a function of oxygen content at 900 °C.

good catalyst for hydrogen oxidation, which is also confirmed by electrochemical tests.16,19 The formed hydrogen is simultaneously oxidized to water as follows:

2H2 + O2 ) 2H2O

(10)

Combining reactions 4 and 10, the overall reaction would be reaction 9. Methane and oxygen may react at high temperature without any catalyst. To clarify the contribution of those reactions to the CH4 and O2 conversion, blank tests were carried out under the same conditions. As shown in Figure 5, it was observed that less than 30% O2 was converted when the O2 content was less than 40% (Figure 5a). CO and CO2 were the sole products at O2 content 10 and 70%, respectively, and CO/CO2 mixtures were observed when the O2 content was 20-60% (Figure 5b). Under the same conditions, all of the oxygen was consumed when the reaction was performed over the LSCM catalyst (Figure 4a) and the major product was CO2 (Figure 4b). Thus, over LSCM, much more oxygen was converted, and full oxidation to CO2 was the dominant path. In the presence of the LSCM catalyst at 900 °C, the CO2 selectivity was ∼98% when the relative oxygen content was between 30 and 50%. However, large amounts of hydrogen and CO were observed when the ratio of methane to oxygen was 40:60 or 2:3 (Figure 4a). These data points were reproduced each time when the experiments were repeated. At this ratio, the product ratio was much closer to that in the absence of LSCM catalyst and seems to reflect direct combustion conditions becoming more important. This phenomenon was repeatedly observed at a CH4/O2 ratio of 40:60 for the blank tests as well (Figure 5a), indicating that the large amounts of H2 and CO in

Figure 5. (a) Conversion and gas formation rate as a function of oxygen content at 900 °C over a blank reactor. The CH4 and He flow rates were fixed at 12 and 36 mL/ h, respectively. (b) CO and CO2 selectivity as a function of oxygen content at 900 °C over a blank reactor.

Figure 4a may come from the combustion of CH4 and O2 at that particular ratio. When the relative oxygen content increased to 70%, complete oxidation of methane was realized again with both CO2 selectivity and methane conversion rising to 100% (i.e., reaction 5 occurred under these conditions). The relative percentage of oxygen for the partial oxidation reaction is 33%, whereas that for full oxidation is 67%; thus, it is not surprising that total oxidation is dominant at 70% oxygen and above. What is surprising is that at lower temperature and lower pO2, complete oxidation dominates. Under both types of conditions, the oxygen partial pressure is much reduced, and we suggest that the condition of the catalyst is also affected, increasing the oxygen vacancy concentration and enhancing oxidation properties. It is only in a fairly limited range where the pO2 is relatively high and the O2 content is less than what was required to provide the complete oxidation stoichiometry where significant partial oxidation was observed. A solid-oxide fuel cell was required to operate at a lower temperature when a metal interconnector was applied because stainless steel may corrode above 800 °C. The methane conversion as a function of oxygen content was also investigated at 700 °C. As shown in Figure 6, the methane conversion increased with increasing oxygen content, although it was slightly lower than that at 900 °C. Unlike at 900 °C, the oxygen conversion was 100% only when the relative oxygen content was less than 20%, and it dropped slowly at higher oxygen contents. According to the blank tests, no reaction occurred at 700 °C when the oxygen content was less than 30% (Figure 7a). The reaction over LSCM was much faster. For blank tests, when the O2 content reached 40%, only 0.33% CH4 reacted with O2 forming CO (Figure 7a). Under the same conditions,

Fuel Cell Electrode La0.75Sr0.25Cr0.5Mn0.5O3-δ

Figure 6. (a) Conversion and gas formation rate over LSCM as a function of oxygen content at 700 °C. The CH4 and He flow rates were fixed at 12 and 36 mL/h, respectively. (b) CO and CO2 selectivity as a function of oxygen content at 700 °C.

∼30% CH4 was converted to CO2 over LSCM (Figure 6a). LSCM is therefore a good complete methane oxidation catalyst. CO was the dominant product for the blank tests at 700 °C until O2 content reached 70% (Figure 7b). The CO2 selectivity was almost 100% over the entire oxygen content, indicating that LSCM was a complete methane oxidation catalyst at 700 °C. In the measurement, the flow rate of methane and carrier helium was fixed. The increased oxygen content was realized by enhancing the oxygen flow rate. The slight drop of CO2 selectivity at high oxygen content may be caused by the higher total gas flow rate in which case some CO passed across the catalyst before being oxidized to CO2. The relative higher total flow rate may be responsible for the formation of some CO at 70% O2 content at the blank test at 700 °C (Figure 7) because CO has not been completely oxidized to CO2 before passing through the hot zone in the reactor. It is instructive to compare the catalytic properties of LSCM with La0.75Sr0.25Cr0.5Fe0.5O3-δ (LSCrF), its Fe analogue measured under the same conditions.23 At 900 °C, an 11% conversion for methane steam reforming was observed when the steam-to-methane ratio was 1:1, about 5 times higher than for LSCM. At 900 °C, a conversion of 68% for methane oxidation was achieved over LSCrF when an equimolar mixture of CH4 and O2 was introduced into the reactor, whereas 100% conversion was achieved at LSCM. Above 850 °C, partial oxidation was completely dominant for most CH4/O2 ratios at La0.75Sr0.25Cr0.5Fe0.5O3-δ, whereas at La0.75Sr0.25Cr0.5Mn0.5O3-δ, full oxidation to CO2 was the dominant process at all temperatures. Complete oxidation of methane in these perovskite oxides is favored by Mn, whereas partial oxidation and reforming seem more favored at Fe.

J. Phys. Chem. B, Vol. 110, No. 43, 2006 21775

Figure 7. (a) Conversion and gas formation rate as a function of oxygen content at 700 °C over a blank reactor. The CH4 and He flow rate were fixed at 12 and 36 mL/h, respectively. (b) CO and CO2 selectivity as a function of oxygen content at 700 °C over a blank reactor.

To further understand the mechanism of the catalytic processes occurring at La1-xSrxCr0.5Mn0.5O3-δ, the X-ray adsorption spectra of samples with x ) 0.2, 0.25, and 0.3 were investigated, with the XANES spectra shown in Figure 8. There is no change in Cr XANES spectra between as-prepared and reduced samples; hence, there are no significant redox changes of Cr on reduction. Similarly, the EXAFS spectra of Cr are very similar in as-prepared and reduced samples, indicting that the local environment of the Cr ions does not change significantly on reduction. The situation for Mn is very different with the XANES spectra showing a strong shift on reduction for all compositions, confirming that the Mn species changes in oxidation state on reduction, and further, there are also very significant changes in the EXAFS spectra for Mn on reduction. These results clearly indicate that oxygen is lost in the coordination environment of Mn on reduction but not in that of Cr. In terms of local coordination, this can be understood if one looks at the two-dimensional representation of the perovskite lattice shown in Figure 9, where it is only the oxygen atoms between two Mn ions that are lost on reduction. As the amount of oxygen in reduction approaches 0.25 per perovskite unit,18 this corresponds to reduction in coordination of all the Mn ions in the lattice from 6-fold coordination to 5-fold coordination (i.e., the coordination of Mn changes from octahedral toward square pyramidal). As the Cr coordination environment is dominated by octahedral oxygen coordination, the extent of oxygen reduction implies that the core MnO5 units tend to share vacant oxygen sites between pairs of such units. On reduction, the oxygen vacancies formed between two manganese ions in LSCM may act as active centers for oxidation reactions. The redox chemistry of manganese may significantly benefit the

21776 J. Phys. Chem. B, Vol. 110, No. 43, 2006

Tao et al. insignificant; however, it is a good oxidation catalyst. According to the catalytic tests, the direct complete oxidation of methane plays a very important role, although there would also be some CO2 reforming and steam reforming. It is therefore concluded that steam reforming is not the only way of using methane in a high temperature fuel cell. With an anode containing LSCM, an external reformer before the fuel cell system is not necessary. With direct oxidation, and in particular complete oxidation without toxic CO, we can directly utilize methane as the fuel. In this paper, we reveal that complete methane oxidation can play a very important role in the SOFC anode process. Under most conditions, LSCM is a complete methane oxidation catalyst. Unlike LaMnO3, it is redox stable and therefore may be used as a SOFC anode. EXAFS experiments reveal that LSCM is basically a stabilized manganite because the oxidation state of chromium remains unchanged on reduction. Acknowledgment. We thank EPSRC, Rolls Royce, Carbon Trust, and the EU RealSOFC IP for support. We also thank EPSRC and CCLRC for provision of synchrotron time at Daresbury station 9.3 and Drs. Ian Harvey and Alistair Lennie for assistance and advice in running these experiments. We thank Dr. Paul Connor at St. Andrews for helpful discussions on the EXAFS analysis. One of the authors (S.T.) thanks EaStCHEM for a fellowship. References and Notes

Figure 8. Cr and Mn K XANES spectra for La1-xSrxMn0.5Cr0.5O3-δ.

Figure 9. Formation of oxygen vacancies of LSCM in a reducing atmosphere.

formation of intermediates that may accelerate the oxidation process. From this point of view, LSCM is a stabilized manganite because its catalytic properties are close to La1-xSrxMnO3. The introduction of chromium makes the perovskite lattice stable, which makes it possible to use as an anode material for SOFCs. 4. Conclusion In conclusion, we have investigated the possible catalytic reaction mechanisms at the redox stable SOFC anode LSCM. The steam and CO2 reforming of methane on LSCM is fairly

(1) Minh N. Q. J. Am. Ceram. Soc. 1993, 76, 563. (2) Vernoux, P.; Guindet, J.; Kleitz, M. J. Electrochem. Soc. 1998, 145, 3487. (3) Park, S.; Vohs, J. M.; Gorte, R. J. Nature 2000, 404, 265. (4) Irvine, J. T. S.; Sauvet, A.-L. Fuel Cells 2001, 1, 205. (5) Atkinson, A.; Barnett, S.; Gorte, R. J.; Irvine, J. T. S.; McEvoy, A. J.; Mogensen, M.; Singhal, C.; Vohs, J. M. Nat. Mater. 2004, 3, 17. (6) Tao, S. W.; Irvine, J. T. S. Chem. Rec. 2004, 4, 83. (7) Nakamura, T.; Petzow, G.; Gauckler, L. J. Mater. Res. Bull. 1979, 14, 649. (8) Yokokawa, H.; Sakai, N.; Kawada, T.; Dokiya, M. Solid State Ionics 1992, 52, 43. (9) Sfeir, J.; Buffat, P. A.; Mo¨ckli, P.; Xanthopoulos, N.; Vasquez, R.; Mathieu, H. J.; Van herle, J.; Thampi, K. R. J. Catal. 2001, 202, 229. (10) Sauvet, A. -L.; Fouletier, J. J. Power Sources 2001, 101, 259. (11) Spinnicci, R.; Delmastro, A.; Ronchetti, S.; Tofanari, A. Mater. Chem. Phys. 2002, 78, 393. (12) Marchetti, L.; Forni, L. Appl. Catal., B 1998, 15, 179. (13) Spinicci, R.; Tofanari, A.; Delmastro, A.; Mazza, D.; Ronchetti, S. Mater. Chem. Phys. 2002, 76, 20. (14) Cimino, S.; Colonna, S.; Rossi, S. D.; Faticanti, M.; Lisi, L.; Pettiti, I.; Porta, P. J. Catal. 2002, 205, 309. (15) Thampi, K. R.; McEvoy, A. J.; Van herle, J. J. Electrochem. Soc. 1995, 142, 506. (16) Tao, S. W.; Irvine, J. T. S. Nat. Mater. 2003, 2, 320. (17) Boukamp, B. A. Nat. Mater. 2003, 2, 294. (18) Tao, S. W.; Irvine, J. T. S. J. Electrochem. Soc. 2004, 151, 252. (19) Tao, S. W.; Irvine, J. T. S.; Kilner, J. A. AdV. Mater. 2005, 17, 1734. (20) Gorte, R. J.; Park, S.; Vohs, J. M.; Wang, C. H. AdV. Mater. 2000, 12, 1465. (21) Kim, H.; Lu, C.; Worrell, W. L.; Vohs, J. M.; Gorte, R. J. J. Electrochem. Soc. 2002, 149, 247. (22) McIntosh, S.; Vohs, J. M.; Gorte, R. J. J. Electrochem. Soc. 2003, 150, A470. (23) Tao, S. W.; Irvine, J. T. S. Chem. Mater. 2004, 16, 4116.