Hydrogen Economy

Overview

Materials Challenges for Solid-Oxide Fuel Cells

Solid-oxide fuel cells (SOFCs) can be used for efcient energy conversion with a variety of fuels. The fuel tolerance of SOFCs is due primarily to the high operating temperature, which accelerates reactions in the fuel. However, the high temperature accelerates unwanted reactions and creates materials challenges, so intermediate-temperature SOFCs are being developed. Reduction of the operating temperature requires solid electrolyte materials with higher conductivities, which can create new compatibility issues. In this paper, materials challenges in the development of SOFCs will be discussed. INTRODUCTION The need for clean, efcient energy conversion is a critical technological and

economic challenge. One approach to meeting this challenge is through electrochemical energy conversion with fuel cells. There are several different types of fuel cells, but the focus of this paper is on solid-oxide fuel cells (SOFCs).14 Although SOFCs were developed rst for stationary power plants, smaller units for auxiliary power units (APUs) in mobile applications4,5 and even small micro-fuel cells68 have been developed. Solidoxide fuel cells operate at a relatively high temperature, which accelerates electrode reactions and allows for use of a wide variety of fuels,911 including ammonia,12 methane,13 and jet fuel,14 as well as biofuels such as waste vegetable oil15 and bioethanol.16 This fuel exibility provides a transition to the hydrogen economy because SOFCs can be imple-

mented before a hydrogen infrastructure is established. At the same time, they can operate with hydrogen as the fuel and thus provide a market for hydrogen. The high operating temperature also improves energy conversion efciency, especially when the SOFC is combined with another energy conversion system,17 such as a micro-turbine.18 In addition to improved efciency, heat produced by the SOFC can be used in CO2 sequestration processes, which is particularly useful in coal gasication systems.19 While the high operating temperature can improve performance, it can also lead to increased materials degradation, so materials development is critical for the implementation of SOFCs.20 These challenges have, in part, led to attempts to reduce the operating temperature in intermediate-temperature SOFCs (ITSOFCs). While the reduced operating temperature of IT-SOFCs relaxes some of the materials requirements, there are still signicant challenges and some additional restrictions are created. FUEL CELL COMPONENTS Solid-oxide fuel cells can be used in tubular or planar configurations. The planar conguration (Figure 1) has better specic power density. The electrochemical cell consists of a solid electrolyte separating two electrodes. The solid electrolyte conducts ions (O2 for SOFCs) to allow the air and fuel to react, but does not conduct electrons, so that the current passes through the external circuit. The fuel is oxidized at the anode and oxygen is reduced at the cathode. Because of the high operating temperature, platinum is not needed (as in polymer-electrolyte membrane [PEM] fuel cells), but the difference in oxygen partial pressures requires that different materials be used for the anode and
JOM December 2007

Figure 1. The schematics of a solid oxide fuel cell: (a) fuel cell stack, (b) electrochemical cell. b

56

cathode. In addition to the electrochemical cell, there are materials challenges in other components. One example is the interconnect material, which carries current between the electrodes and the external circuit. Another challenge is the sealant, which is not shown in Figure 1, but is required to provide gas-tight separation between the air and fuel gas channels. MATERIAL CHALLENGES Solid Electrolytes A solid electrolyte must have good ionic conductivity and low electronic conductivity, and optimization of these transport properties requires control of the electrolyte composition and structure.21 The ranges of conductivities for electrolytes commonly used in SOFCs are shown in Figure 2.22 Yttria-stabilized zirconia (YSZ) is the most widely used electrolyte for SOFCs. Although yttria doping of zirconia is most common, scandium doping has been shown to result in a higher conductivity, which is attributed to a lower migration due to smaller size mismatch.23 A rhombohedral-to-cubic phase change occurs in scandium-doped zirconia (ScSZ), but this can be avoided by limiting the amount of the dopant to about 8 mol. percent and/or by adding other co-dopants, such as bismuth.24 The lower operating temperatures in IT-SOFCs require higher electrolyte conductivity, so other electrolytes have been developed. Ceria oxide forms the same crystal structure as zirconia (cubic uorite) and, as shown in Figure 2, has a higher conductivity than YSZ or ScSZ at low temperatures. As a result, ceria-based electrolytes are attractive for low-temperature and high-power applications.25 The primary challenge for these electrolytes is stability in low-oxygen partial pressures.23,25 Like zirconia, ceria is doped to increase the concentration of oxygen vacancies, and thus the ionic conductivity. Ceria is most commonly doped with gadolinium, which is referred to as gadolinium-doped ceria (GDC) or cerium gadolinium oxide (CGO). Samarium doping has resulted in a higher conductivity than GDC26 and other dopants, such as lanthanum and yttrium, have also been used. Another electrolyte for low-tempera2007 December JOM

Figure 2. The conductivity of SOFC electrolyte materials.22

ture use is lanthanum gallate, LaGaO3, doped with strontium on the lanthanum site and magnesium on the gallium site(La1-xSrx)(Ga1-yMgy)O3 (LSGM).27 The maximum conductivity has been found with 20% strontium and 1520% magnesium.27 Transition metals such as cobalt and iron can increase conductivity,28 but can also lead to undesired increases in electronic conductivity.29 Solid-oxide fuel cells most commonly use oxygen-ion conducting electrolyte, but proton-conducting oxides have also been used in high-temperature fuel cells.30 With a proton-conducting electrolyte, water is produced at the cathode (from protons reacting with oxygen) instead of the anode (see Figure 1), so the fuel is not diluted with water. In general, electrolyte materials are relatively stable in the operating conditions. However, in some cases electrolytes can react with other components. Anodes The most common anode material is a porous nickel-YSZ cermet, in which nickel provides the electronic conduction and YSZ provides the oxygen-ion conduction.3134 The cermet is typically prepared by reduction of the NiO in a mixture of NiO and YSZ. While a small amount of NiO dissolves in YSZ,35 nickel is stable with YSZ, so degradation of a Ni-YSZ anode is typically due to morphological changes rather than chemical reaction. For example, changes in the size and shape of the nickel particles

due to sintering can lead to stresses and cracking during redox cycling.3638 Although SOFCs are more tolerant of low fuel quality than PEM fuel cells, fuel impurities can degrade performance. Nickel catalyzes the formation of C-C bonds, so carbon deposition on nickel can occur when a hydrocarbon is used as the fuel. Small amounts of carbon can be benecial to anode performance,39 but large amounts block the electrode surface and degrade performance. Copper does not catalyze C-C bonds, so carbon deposition can be reduced by using Cu-YSZ cermets anodes. Another detrimental impurity is sulfur, which can adsorb to the electrode surface and degrade performance.40,41 Copper-containing electrodes are also more tolerant of sulfur impurities than Ni-YSZ,42 but copper is not as stable as nickel in fuel cell operating conditions. Oxide materials can also improve fuel tolerance. For example, ceria inltration has been shown to reduce degradation due to sulfur impurities.43 Ceria is used with, or in place of, zirconia, because it has good catalytic behavior and its electronic conductivity extends the electrochemical reaction area.44 Other electronic-conducting materials offer advantages for microstructural stability and carbon/sulfur tolerance.33,45 In addition, if the oxide material is a mixed conductor, the electrochemical reaction can occur over the entire electrode surface rather than only at the three-phase (Ni-YSZ-gas) boundary (Figure 3).
57

Figure 3. Cermet and mixed-conducting electrodes.

Oxides forming the perovskite structure, especially chromites and titanates, are the most commonly used oxide anode materials.46,47 The titanate most commonly used as an SOFC anode is SrTiO3, which is doped with reactive elements such as yttrium or lanthanum to increase conductivity. As with Ni-YSZ, ceria improves the sulfur tolerance of yttria-doped strontium titanate anodes.48 The most commonly used chromite is lanthanum chromite, typically doped with strontium or calcium, which is also a commonly used SOFC interconnect material. Mixed chromite-titanates are also used and are stable with YSZ.49 Another approach is to dope the electrolyte materials to increase electronic conduction. For example, YSZ can be doped with titania, which, due to the formation of Ti3+ at low oxygen partial pressures, increases conduction.50 Chemical reaction between perovskite anodes and YSZ is generally not a problem. However, if a perovskite electrolyte such as LSGM is used, interdiffusion rather than additional phase formation can occur.51 Such interdiffusion will change the compositions of the two materials, which can be detrimental. For example, the diffusion of gallium or magnesium from LSGM into strontium titanate leads to a decrease in conductivity.52 A ceria layer can be added to minimize this interdiffusion.53 In addition, while nickel is stable with YSZ, it can react with LSGM.27 Cathodes The oxygen partial pressure at the cathode restricts potential metallic cathodes to noble metals, which are typically expensive. Although silver is considered for use in SOFCs, most cathode materials are conducting oxides. Cathode degradation can occur due to chromium poisoning or due to reaction with the
58

electrolyte.10 The degree of chromium poisoning depends more on the interconnect material, so this will be discussed in the section on interconnects. The most common cathode material is strontium-doped lanthanum manganate (La1xSrx)MnO3 (LSM).54,55 The coefcient of thermal expansion (CTE) of LSM matches well with that of YSZ, and although its ionic conductivity is relatively low, it performs well as a cathode. As discussed previously, mixed conduction expands the area over which the electrochemical reaction can occur, so materials with higher ionic conductivity can perform better as cathodes. (La1xSrx)Co3 (LSC) has higher oxygen ion conductivity, but also higher CTE, than LSM. However, the addition of iron to LSC decreases the CTE, so (La1xSrx)(Co1yFey)O3 (LSCF) is an attractive cathode material. LSCF has faster oxygen-exchange kinetics and lower polarization resistance than LSM. La(Ni1xFex)O3 (LNF) is also a potential cathode material and has better stability with chromia. The area over which the electrochemical reaction occurs can also be expanded by mixing the electrode material with the electrolyte material to form a composite (e.g., LSM/YSZ). Reaction between the cathode and electrolyte most commonly occurs
60 50 Conductivity (S/cm) 40 30 20 10 0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 x in La1-x (Ca, Sr)x CrO3La1-x Srx CrO3- at 1,000C La1-x Cax CrO3- at 1,000C

during processing, which is typically performed at temperatures higher than those used during fuel cell operation. LSM reacts with YSZ to form La2Zr2O7 and/or SrZrO3, depending on the level of strontium doping.56,57 The formation of La2Zr2O7 is most often attributed to the preferential diffusion of manganese into YSZ (relative to lanthanum) leading to an increase in the lanthanum in the LSM,58 so the onset of La2Zr2O7 formation can be delayed by starting with excess manganese (A-site decient) LSM. The reaction of LSC and YSZ is similar to that between LSM and YSZ, but is generally more severe.59 (La1xSrx)FeO3 (LSF) is less reactive than LSC, and the reactivity of LSCF is intermediate to those of LSC and LSF.59 LNF does not contain strontium, and thus forms only La2Zr2O7. Although the low oxygen pressure stability of ceria is inferior to that of YSZ, its stability with the cathode is superior as it does not react with any of the common cathode materials.59 Thus, ceria-based interlayers are often used to prevent reaction between YSZ and the cathode. In such cases, the interaction of ceria with YSZ is stronger than that with the cathode because ceria and YSZ form the same crystal structure, so interdiffusion can occur. In the case of CGO and YSZ, the intermediate composition has a lower conductivity than either phase, so the resistance of the interface can increase.60 Most cathode materials are oxides with the perovskite structure, so, as with perovskite anodes, interdiffusion rather than additional phase formation can occur when LSGM is used as the electrolyte. Manganese and cobalt diffuse most readily in LSGM,61 and cobalt is the major diffusing species even for

La1-x Cax CrO3- at 700C Figure 4. The conductivity of doped LaCrO3.69

0.35

0.40

JOM December 2007

cathodes with relatively low cobalt concentrations.62 Similar interdiffusion occurs for LSC63 and LSCF,64 both of which form the perovskite structure. As mentioned previously, the addition of cobalt to LSGM increases the conductivity, so this is not necessarily detrimental (i.e., if the electronic conductivity is not signicantly increased).22 Like nickel-containing anodes, nickelcontaining cathodes such as LNF react with LSGM,65 so interlayers such as ceria are often used to limit this reaction. Interconnects The primary function of the interconnect material is to transport the electrical current from the electrodes to the external circuit, which requires low electrical resistance, but there are other requirements.66,67 The interconnect material is exposed to both the anode and cathode, and thus must be stable in both oxidizing and reducing environments. As with all SOFC components, the CTE must match those of other components to minimize thermal stresses during thermal cycling. Ceramic and metallic interconnects have both been used. In general the advantages of metals are lower cost and ease of fabrication, while ceramics offer stability. Ceramic interconnects are semiconducting oxides that are stable in oxidizing environments. The most commonly used ceramic interconnect is LaCrO3, which is doped, typically with strontium or calcium, to increase the conductivity.68,69 Figure 4 shows that calcium more effectively increases conductivity and, as expected for a semiconductor, conductivity decreases with decreasing temperature. The strontium and calcium additions also increase the CTE to better match that of YSZ. Lanthanum chromite is a p-type conductor, so the conductivity decreases with decreasing oxygen partial pressure and the conductivity in the fuel is lower than that in air. In addition, a decrease in oxygen partial pressure leads to expansion of lanthanum chromite as Cr4+ (high pO2) is reduced to Cr3+ (low pO2), which has a larger radius.70 Although the most common dopants are strontium and calcium on the lanthanum site, smaller ions such as the alkaline earth magnesium and transition metals (e.g., copper, nickel) can be doped on the chromium site. Ceria has also been added
2007 December JOM

Figure 5. The coefcients of thermal expansion for YSZ and chromium-forming alloys.7582

to calcium-doped lanthanum chromium to increase the conductivity.71 Since the optimal material for the anode and cathode may be different, the use of a bi-layer consisting of p-type semiconductor at the cathode and an n-type semiconductor at the anode has been investigated.72 Although ceramic interconnects are generally stable with YSZ, their high cost and decreasing conductivity with decreasing temperature has led to the predominant use of metallic interconnects for IT-SOFCs. Except for precious metals which, with the possible exception of silver, are prohibitively expensive, metals are inherently unstable in air, so oxidation resistance is a major challenge for metallic interconnect materials.7375 In addition to forming on the metal surface, the oxidation scale forms at the interface between the interconnect and the electrodes and is thus part of the electrical circuit. This places an additional requirement that the electrical conductivity of the scale should be high to minimize the ohmic resistance and the associated overpotential. Due to this requirement, alumina- or silica-forming alloys are not used as interconnects because, although they have good oxidation resistance, the scale has low electronic conductivity, which increases the cell resistance. Thus, metallic interconnect alloys are usually chromia-formers since, although chromia grows faster than alumina or silica, it has higher electrical conductivity. The CTEs of interconnect alloys should match those of other materials used in

SOFCs. Figure 57582 shows that the CTEs of austenitic stainless steels and nickelbased alloys are considerably higher than that of YSZ. The CTEs of ferritic stainless steels and chromia-based alloys, however, are close to that of YSZ. The most promising chromia-based alloy is an oxide-dispersion-strengthened alloy (DucrolloyCr-5Fe-1Y2O3), which is relatively expensive and difcult to fabricate, so ferritic stainless steels are the most widely used metallic interconnects. In general, oxidation behavior depends on the atmosphere. Oxygen partial pressure affects oxide stability and growth kinetics, but in the case of SOFCs chromia is stable and grows at about the same rate in both the cathode and anode atmospheres.83 Water vapor, which is typically present in air and in the fuel, can affect oxidation behavior. Although the effect on the growth kinetics is equivocal, the presence of water vapor during oxidation does generally improve scale adherence. Anomalous behavior has been observed in a dual atmosphere, in which the alloy is exposed to air on one side and fuel on the other, as would be the case in an SOFC.84 The oxidation rate in methane, which can be used as a fuel in SOFCs, is similar to that in a hydrogen fuel.85 Another possible problem in carbon-containing environments is carbon dusting, but ferritic stainless steels, which are better suited for use in SOFCs, are more resistant to metal dusting than austenitic stainless steels. Although a protective chromia scale
59

0 Log Partial Pressure (Pa) 1 2 3 4 5 700

Gindorf et al. (Ref. 88)

Ebbinghaus (Ref. 89) CrO2(OH)2 (10 PA H2O, 100 ppm H2O)

Figure 6. The vapor pressures of Cr6+-containing species.8789 1,200

800

900 1.000 Temperature (C)

1,100

forms on ferritic stainless steels, during long exposure, chromium depletion can lead to breakaway oxidation.86 As the operating temperature is decreased in IT-SOFCs, the oxidation rate decreases, so this depletion and other detrimental effects of scale growth decrease. However, chromium poisoning, another degradation mechanism associated with the interconnect material, persists even at the lower temperatures. Chromium poisoning occurs when the chromia scale is oxidized to form a Cr6+containing gas species, such as CrO3 or CrO2(OH)2, which diffuses to the cathode and is reduced back to Cr2O3. Although reducing the temperature would generally be expected to decrease phenomena that are dependent on vaporization, the vapor pressure of CrO2(OH)2 has a relatively low temperature dependence (Figure 6).8789 The deposition of chromia occurs at the three phase boundary where electrons are available from the cathode and oxygen ions can be accommodated in the electrolyte (Figure 7). This is the same location at which reduction of oxygen occurs during fuel cell operation, so chromia blocks these active sites and degrades performance. Mixed conduction in either the cathode or electrolyte, however, can expand the area over which these reduction reactions occur so that the amount of degradation caused by chromia is decreased.90 To minimize chromium poisoning, alloys for SOFCs have been designed to minimize the volatilization of chromium oxides. One of the approaches is to add elements, such as manganese or titanium, that lead to the formation of oxides other than chromia on the outer surface of the scale.73,9193 For example, manganese additions lead to the formation of a manganese-chromium spinel on the scale surface that decreases
60

the amount of chromium poisoning.94 The interconnect is in contact with the cathode, so manganese can diffuse from LSM into the scale.9597 In addition to decreasing volatilization, the increase in manganese content increases the conductivity of the spinel phase and thus decreases the electrical resistance of the interface.98 Although volatilization is decreased by manganese additions, it is not eliminated,99 so alloys that form scales with little100 or no chromium101 have been developed. The reduction or elimination of the chromium leads to an increase in the oxidation rate, but if the conductivity of the scale is higher, a thicker scale can be tolerated. The protection can be further improved by applying a ceramic coating to the surface. The coating serves three functions which are illustrated schematically in Figure 8. The coating is part of the electrical circuit and thus must have low electrical resistance, while at the same time should have low oxygen permeability to reduce the growth rate of the chromia scale beneath the coating. In addition, the chromium permeability should be low to limit the outward diffusion of chromium and thus reduce chromium volatilization. An obvious choice for the ceramic coating is the same material that is used for ceramic interconnects: lanthanum chromite.102 Although lanthanum chromite contains chromia, the activity, and thus the equilibrium vapor pressure, is much lower than that of the chromia scale formed on the alloy. Volatilization is signicantly decreased by a lanthanum chromite coating, but chromium is still transported to the surface, which can lead to poisoning after long operating times.103 Solid-oxide fuel cell electrode materials are another logical source of potential ceramic coatings.102 Cathode

materials such as LSM, LSF, LSC, LSCF, and LNF do not contain chromium and have been used as interconnect coatings or contact materials. Among these, LNF is less reactive with chromia and has better tolerance to poisoning than other cathode materials. Another approach to interconnect coatings follows from the benecial effect of alloying ferritic stainless steels with manganese by depositing coatings of spinel oxides.82 The simple spinel Co3O4 has been used, but has a relatively high CTE,104 which leads to spallation,105 so mixed spinel oxides are preferred. One of the most promising coating materials is (Mn,Co)3O4, which has been shown to provide a good barrier to chromium diffusion.106 The coatings can be prepared by different methods, such as sputtering, plasma spraying, and electroplating following by oxidation or slurry coating, and the effectiveness depends on the deposition method.107 Because of challenges in the processing of ceramic materials, the coating quality is often more important than the particular coating used.102 Sealants Seals for SOFCs can be either rigid or compressive.108,109 Rigid seals do not require application of a load, but the CTE of the seal must closely match those of other SOFC components. On the other hand, compressive seals can accommodate larger differences in CTE, but a load must be applied continuously during operation. The most commonly used rigid seals are glasses or glassceramics.110,111 The CTEs of most glasses are less than that of YSZ, so additives such as barium and calcium are added because they form orthosilicate phases with high CTEs.112114 Although bariumcontaining glasses are commonly used, calcium-containing glasses are less reactive with other SOFC components.115 In addition to CTE, the viscosity must be low enough for the glass to ow and wet the surface, but high enough to provide adequate sealing. Boron oxide is added to decrease the viscosity and improve wetting116,117 but can limit the life, so the amount of boron added should be limited.118 In glass-ceramics, it is important to control the crystallization rate,111 so nucleating agents such as titanium, zirconium, nickel, and chromium are often added.112,119 LongJOM December 2007

which creates materials challenges for SOFC components. While some materials requirements are relaxed with the reduced operating temperatures in intermediate-temperature SOFCs, challenges remain in achieving long lifetimes and durability for expanded implementation of SOFCs. ACKNOWLEDGEMENT The financial assistance of the National Science Foundation (NSF-DMR0552896) is gratefully acknowledged.
References

Figure 7. A schematic of cathode poisoning.

term performance depends on stabilizing crystallization for long times, which is accomplished by additions such as yttrium120 or tantalum.121 Oxide additions can be melted into the glass or be present as second phases and their impact on the seal performance depends on how they are added. For example, the addition of NiO122 or MgO123 as separate phases can be used to tailor the CTE of the resulting composite. Other second phases, such as YSZ,124,125 monoclinic zirconia,126 or boron nitride (BN)127 can increase the strength of the seal. The sealant is in contact with the interconnect, so potential reaction with the stainless steel must be considered in selecting additions.128,129 Metallic brazes can provide rigid seals and are more accommodating of stress than ceramic seals. The two primary challenges with metallic seals are cost (i.e., noble metals are needed) and wetting of the ceramic components. Braze materials are based on gold or silver with additions such as titanium,130 aluminum,131 copper,132 or combinations of these,133 to improve wettability with

the ceramic. The other approach to SOFC sealants is to use compressible seals, such as metal gaskets, and silver has also been used as a gasket material.134 One of the problems with silver is that poor performance has been observed in a dual atmosphere, where hydrogen from the anode and oxygen from the cathode diffuse into the silver, react within the metal to form water, and cause seal failure.135 Compressive seals based on mica-based materials have also been used. The effectiveness of these seals is improved by inltration with silver,136 glass,137 silica,138 or talc139 to block the gaps between the mica akes. CONCLUSIONS The ability SOFCs to operate with a wide variety of fuelsfrom hydrogen to methane to biofuelsprovides an opportunity for implementation of fuel cells before a hydrogen infrastructure is established. The fuel tolerance that provides this capability is due primarily to the high operating temperature,

Figure 8. A schematic of transport in interconnect coating.

1. M.A.J. Cropper, S. Geiger, and D.M. Jollie, J. Power Sources, 131 (1-2) (2004), pp. 5761. 2. P. Singh and N.Q. Minh, Int. J. Appl. Ceram. Tech., 1 (1) (2004), pp. 515. 3. M.C. Williams, Fuel Cells (Weinheim), 7 (1) (2007), pp. 7885. 4. S.C. Singhal, Solid State Ionics, 152-153 (December 2003), pp. 405410. 5. P. Lamp et al., Fuel Cells, 3 (3) (2003), pp. 146152. 6. J.L. Hertz and H.L. Tuller, J. Electrochem. Soc., 154 (4) (2007), pp. B413B418. 7. G.J. La O et al., Int. J. Energy Research, 31 (2007), pp. 548575. 8. B.E. Buergler et al., J. Power Sources, 171 (2) (2007) pp. 310320. 9. M.C. Williams, J.P. Strakey, and W.A. Surdoval, Int. J. Appl. Ceram. Tech., 2 (4) (2005), pp. 295300. 10. H. Tu and U. Stimming, J. Power Sources, 127 (1-2) (2004), pp. 284293. 11. K. Sasaki et al., J. Electroceramics, 13 (1-3) (2004), pp. 669675. 12. N.J.J. Dekker and G. Rietveld, J. Fuel Cell Sci. Technol., 3 (4) (2006), pp. 499502. 13. M. Kawano et al., J. Electrochem. Soc., 154 (5) (2007), pp. B460B465. 14. Z.F. Zhou et al., J. Power Sources, 133 (2) (2004), pp. 181187. 15. Z.F. Zhou et al., J. Power Sources, 171 (2) (2007) pp. 856860. 16. W. Jamsak et al., ECS Trans., 7 (1) (2007), pp. 14751482. 17. D.L. Cresswell and I.S. Metcalfe, Solid State Ionics, 177 (19-25) (2006), pp. 19051910. 18. N. Autissier et al., J. Fuel Cell Sci. Tech., 4 (1) (2007), pp. 123129. 19. G. S owi ski, Int. J. Hydrogen Energy, 31 (2006), pp. 10911102. 20. B.C.H. Steele, J. Mater. Sci., 36 (5) (2001), pp. 10531068. 21. S. Hui et al., J. Power Sources, 171 (2) (2007) pp. 493502. 22. J.W. Fergus, J. Power Sources, 162 (1-2) (2006), pp. 3040. 23. V.V. Kharton, F.M.B. Marques, and A. Atkinson, Solid State Ionics, 174 (1-4) (2004), pp. 135149. 24. S. Sarat, N. Sammes, and A. Smirnova, J. Power Sources, 160 (2) (2006), pp. 892896. 25. B. Dalslet et al., J. Solid State Electrochem., 10 (8) (2006), pp. 547561. 26. S. Kuharuangrong, J. Power Sources, 171 (2) (2007) pp. 506510. 27. A.V. Joshi et al., J. Electroceram., 13 (1-3) (2004), pp. 619625. 28. J.W. Stevenson et al., J. Electrochem. Soc., 147 (9) (2000), pp. 32133218. 29. B.A. Khorkounov, H. Nfe, and F. Aldinger, J. Solid State Electrochem., 10 (7) (2006), pp. 479487. 30. T. Schober, Solid State Ionics, 162-163 (September

2007 December JOM

61

2003), pp. 277281. 31. S.P. Jiang and S.H. Chan, J. Mater. Sci., 39 (14) (2004), pp. 44054439. 32. A. McEvoy, High Temperature Solid Oxide Fuel Cells, ed. S.C. Singhal and K. Kendall (Oxford, U.K.: Elsevier Ltd., 2003), pp. 149171. 33. A. Atkinson et al., Nature Materials, 3 (1) (2004), pp. 1727. 34. C. Sun and U. Stimming, J. Power Sources, 171 (2) (2007), pp. 247260. 35. P.M. Delaforce et al., J. Am. Ceram. Soc., 90 (3) (2007), pp. 918924. 36. D. Sarantaridis and A. Atkinson, Fuel Cells (Weinheim), 7 (3) (2007), pp. 246258. 37. T. Klemens et al., J. Electrochem. Soc., 152 (11) (2005), pp. A2186A2192. 38. N.M. Tikekar, T.J. Armstrong, and A.V. Virkar, J. Electrochem. Soc., 153 (4) (2006), pp. A654A653. 39. S. McIntosh, M.J. Vohs, and R.J. Gorte, J. Electrochem. Soc., 150 (4) (2003), pp. A470A476. 40. M. Gong et al., J. Power Sources, 168 (2) (2007), pp. 289298. 41. Z. Cheng and M. Liu, Solid State Ionics, 178 (13-14) (2007), pp. 925935. 42. Y.M. Choi et al., J. Alloys Comp., 427 (1-2) (2007), pp. 2529. 43. H. Kurokawa et al., Electrochem. Solid-State Lett., 10 (9) (2007), pp. B135B138. 44. T. Horita et al., Solid State Ionics, 177 (19-25) (2006), pp. 19411948. 45. S. Tao and J.T.S. Irvine, The Chemical Record, 4 (2) (2004), pp. 8395. 46. J.W. Fergus, Solid State Ionics, 177 (17-18) (2006), pp. 15291541. 47. B.A. Boukamp, Nature Materials, 2 (5) (2003), pp. 294296. 48. H. Kurokawa et al., J. Power Sources, 164 (2) (2007), pp. 510518. 49. V. Vashook et al., Solid State Ionics, 175 (1-4) (2004), pp. 151155. 50. F. Tietz, I.A. Raj, and D. Stver, Brit. Ceram. Trans., 103 (5) (2004), pp. 202210. 51. Y. Du and N.M. Sammes, Electrochem. Soc. Proc., 2005-07 (SOFC IX) (2005), pp. 11271136. 52. S. Hui and A. Petric, J. Eur. Ceram. Soc., 22 (9-10) (2002), pp. 16731681. 53. J.-H. Lee et al., J. Mater. Sci., 42 (6) (2007), pp. 18661871. 54. H. Yokokawa and T. Horita, High Temperature Solid Oxide Fuel Cells, ed. S.C. Singhal and K. Kendall (Oxford, U.K.: Elsevier Ltd., 2003), pp. 119147. 55. L.J. Gauckler et al., Chimia, 58 (12) (2004), pp. 837850. 56. G. Stochniol, E. Syskakis, and A. Naoumidis, J. Am. Ceram. Soc., 78 (4) (1995), pp. 929932. 57. K. Kleveland et al., J. Am. Ceram. Soc., 82 (3) (1999), pp. 729734. 58. A. Mitterdorfer and L.J. Gauckler, Solid State Ionics, 111 (3-4) (1998), pp. 185218. 59. J.M. Ralph, C. Rossignol, and R. Kumar, J. Electrochem. Soc., 150 (11) (2003), pp. A1518A1522. 60. A. Tsoga, A. Naoumidis, and D. Stver, Solid State Ionics, 135 (1-4) (2000), pp. 403409. 61. K. Huang et al., J. Electrochem. Soc., 143 (11) (1996), pp. 36303636. 62. W. Chen et al., Mater. Res. Bull., 38 (8) (2003), pp. 13191328. 63. Z. Bi et al., J. Power Sources, 161 (1) (2006), pp. 3439. 64. N. Sakai et al., J. Electrochem. Soc., 153 (3) (2006), pp. A621A625. 65. C.N. Munnings et al., J. Fuel Cell Sci. Tech., 2 (1) (2005), pp. 3437. 66. W.Z. Zhu and S.C. Deevi, Mater. Sci. Eng., A348 (12) (2002), pp. 227243. 67. N.Q. Minh and T. Takahashi, Science and Technology of Ceramic Fuel Cells (Amsterdam: Elsevier, 1995), pp. 165198. 68. N. Sakai et al., Int. J. Appl. Ceram. Technol., 1 (1)

(2004), pp. 2330. 69. J.W. Fergus, Solid State Ionics, 171 (1-2) (2004), pp. 115. 70. R.D. Shannon, Acta Cryst., A 32 (5) (1976), pp. 751767. 71. X. Zhou et al., J. Power Sources, 164 (1) (2007), pp. 293299. 72. W. Huang and S. Gopalan, J. Power Sources, 154 (1) (2006), pp. 180183. 73. W.J. Quadakkers et al., Materials at High Temperatures, 20 (2) (2003), pp. 115127. 74. J.W. Fergus, Mater. Sci. Eng. A, 397 (1-2) (2005), pp. 271283. 75. Z. Yang et al., J. Electrochem. Soc., 150 (9) (2003), pp. A1188A1201. 76. T. Uehara et al., Solid Oxide Fuel Cells VIII Electrochem. Soc. Proc., PV 2003-07, ed. S.C. Singhal and M. Dokiya (Pennington, NJ: The Electrochemical Society, 2003), pp. 914922. 77. M. Ueda and H. Taimatsu, European Solid Oxide Fuel Cell Forum Proceedings, Vol. 2, ed. A.J. McEvoy (Lucerne, Switzerland: The European Fuel Cell Forum, 2000), pp. 837843. 78. J. Pirn-Abbelln et al., Solid Oxide Fuel Cells VII Electrochem. Soc. Proc., PV 2001-16, ed. H. Yokokawa and S.C. Singhal (Pennington, NJ: The Electrochemical Society, 2001), pp. 811819. 79. H. Hayashi et al., Solid State Ionics, 176 (2007), pp. 613619. 80. S. Linderoth et al., J. Mater. Sci., 31 (19) (1996), pp. 50775082. 81. W. Glatz et al., European Solid Oxide Fuel Cell Forum Proceedings, Vol. 2, ed. A.J. McEvoy (Lucerne, Switzerland: The European Fuel Cell Forum, 2000), pp. 855864. 82. M. Stanislowski et al., J. Power Sources, 164 (2) (2007), pp. 578589. 83. H. Kurokawa, K. Kawamura, and T. Maruyama, Solid State Ionics, 168 (1-2) (2004), pp. 1321. 84. Z. Yang et al., J. Electrochem. Soc., 151 (12) (2004), pp. B669B678. 85. T. Horita et al., J. Power Sources, 118 (1-2) (2003), pp. 3543. 86. P. Huczkowski et al., J. Fuel Cell Sci. Tech., 1 (2004), pp. 3034. 87. E.J. Opila et al., J. Phys. Chem. A, 111 (10) (2007), pp. 19711980. 88. C. Gindorf, L. Singheiser, and K. Hilpert, Steel Research, 72 (11/12) (2001), pp. 528533. 89. B.B. Ebbinghaus, Combustion Flame, 93 (1-2) (1993), pp. 119137. 90. J.W. Fergus, Int. J. Hydrogen Energy, 32 (16) (2007) pp. 36643671. 91. S. Geng and J. Zhu, J. Power Sources, 160 (2) (2006), pp. 10091016. 92. M. Stanislowski et al., J. Electrochem. Soc., 154 (4) (2007), pp. A295A306. 93. J.E. Hammer et al., Oxidation of Metals, 67 (1-2) (2007), pp. 138. 94. K. Fujita et al., J. Power Sources, 131 (1-2) (2004) 270-277. 95. Y.D. Zhen et al., J. Eur. Ceram. Soc., 26 (15) (2006), pp. 32533264. 96. Z. Yang, G. Xia, and J.W. Stevenson, Ceram. Eng. Sci. Proc., 26 (4) (2005), pp. 217224. 97. M.C. Tucker et al., J. Power Sources, 160 (1) (2006), pp. 130138. 98. W. Qu et al., J. Power Sources, 153 (1) (2006), pp. 114124. 99. N. Sakai et al., Solid State Ionics, 177 (19-25) (2006), pp. 19331939. 100. S. Geng et al., J. Power Sources, 171 (2) (2007) pp. 775781. 101. M.P. Brady et al., Oxidation of Metals, 65 (3-4) (2006), pp. 237261. 102. H. Kurokawa et al., Solid State Ionics, 178 (3-4) (2007), pp. 287296. 103. I. Belogolovsky et al., J. Electrochem. Soc., 154 (9) (2007), pp. B987B980.

104. A. Petric and H. Ling, J. Am. Ceram. Soc., 90 (5) (2007), pp. 15151520. 105. A.N. Hansson et al., Oxidation of Metals, 68 (1-2) (2007), pp. 2336. 106. Z. Yang et al., Int. J. Hydrogen Energy, (16) (2007) pp. 36483654. 107. Z. Yang et al., Surf. Coating Tech., 201 (7) (2006), pp. 44764483. 108. P.A. Lessing, J. Mater. Sci., 42 (10) (2007), pp. 34653476. 109. J.W. Fergus, J. Power Sources, 147 (1-2) (2005), pp. 4657. 110. R.N. Singh, Ceram. Eng. Sci. Proc., 25 (3) (2004), pp. 299307. 111. C. Lara, M.J. Pascual, and A. Duran, J. NonCrystalline Solids, 348 (15) (November 2004), pp. 149155. 112. S.-B. Sohn et al., J. Am. Ceram. Soc., 87 (2) (2004), pp. 254260. 113. R. Zheng et al., J. Power Sources, 128 (2) (2004), pp. 165172. 114. C. Lara et al., Solid State Ionics, 170 (3-4) (2004), pp. 201208. 115. F. Smeacetto et al., J. Eur. Ceram. Soc. 28 (1) 6168. 116. M.J. Pascual, A. Guillet, and A. Durn, J. Power Sources, 169 (1) (2007), pp. 4046. 117. C.A. Lewinsohn, M.A. Wilson, and D.M. Taylor, Proc. Electrochem. Soc., 2003-23Solid-State Ionic Devices III (2003), pp. 346371. 118. A. Flgel et al., J. Electrochem. Soc., 154 (6) (2007), pp. B601B608. 119. D. Bahadur et al., J. Electrochem. Soc., 151 (4) (2004), pp. A558A562. 120. Y.-S. Chou, J.W. Stevenson, and P. Singh, J. Electrochem. Soc., 154 (7) (2007), pp. B644B651. 121. M.D. Dolan and S.T. Misture, J. Electrochem. Soc., 154 (7) (2007), pp. B700B711. 122. Y.-S. Chou and J.W. Stevenson, J. Power Sources, 168 (2) (2007), pp. 426433. 123. K.A. Nielsen et al., Ceram. Eng. Sci. Proc., 25 (3) (2004), pp. 309314. 124. M. Brochu et al., J. Eur. Ceram. Soc., 26 (15) (2006), pp. 33073313. 125. M. Brochu et al., J. Am. Ceram. Soc., 89 (3) (2006), pp. 810816. 126. S.M. Gross et al., Fuel Cells Bull., 2006 (9) (2006), pp. 1215. 127. S.R. Choi, N.P. Bansal,and A. Garg, Mater. Sci. Eng. A, 460-461 (15) (July 2007), pp. 509515. 128. P. Batfalsky et al., J. Power Sources, 155 (2) (2006), pp. 11281137. 129. V.A.C. Haanappel et al., J. Mater. Sci., 40 (7) (2005), pp. 15831592. 130. M. Singh, T.P. Shpargel, and R Asthana, Int. J. Appl. Ceram. Technol., 4 (2) (2007), pp. 119133. 131. J.Y. Kim, J.S. Hardy, and S. Weil, Int. J. Hydrogen Energy (16) (2007) pp. 37543762. 132. J.Y. Kim, J.S. Hardy, and K.S. Weil, J. Am. Ceram. Soc., 88 (9) (2005), pp. 25212527. 133. M.C. Tucker et al., J. Power Sources, 160 (2) (2006), pp. 10491057. 134. J. Duquette and A. Petric, J. Power Sources, 137 (1) (2004), pp. 7175. 135. P. Singh et al., J. Mater. Eng. Perf., 13 (3) (2003), pp. 287294. 136. Y.-S. Chou et al., J. Electrochem. Soc., 153 (8) (2006), pp. A1591A1598. 137. Y.-S. Chou et al., J. Power Sources, 160 (2) (2006), pp. 115122. 138. S. Le et al., J. Power Sources, 168 (2) (2007), pp. 447452. 139. F. Wiener et al., J. Mater. Sci., 42 (8) (2007), pp. 26432651. Jeffrey W. Fergus is with the Materials Research and Education Center at Auburn University, 275 Wilmore Laboratories, Auburn, AL 36849; (334) 844-3405, fax (334) 844-3405; e-mail jwfergus@eng.auburn.edu.

62

JOM December 2007