Conversion Of Hydrocarbons In Sofc

4 Jun 2003

12:38

AR

AR189-MR33-13.tex

AR189-MR33-13.sgm

LaTeX2e(2002/01/18) P1: GJB 10.1146/annurev.matsci.33.022802.092713

Annu. Rev. Mater. Res. 2003. 33:321–31 doi: 10.1146/annurev.matsci.33.022802.092713 c 2003 by Annual Reviews. All rights reserved Copyright °

CONVERSION OF HYDROCARBONS IN SOLID OXIDE FUEL CELLS Mogens Mogensen and Kent Kammer Materials Research Department, Risø National Laboratory, DK-4000 Roskilde, Denmark; email: [email protected]; [email protected]

Key Words electrochemical oxidation, oxide anodes, cracking ■ Abstract Recently, a number of papers about direct oxidation of methane and hydrocarbon in solid oxide fuel cells (SOFC) at relatively low temperatures (about 700◦ C) have been published. Even though the conversion of almost dry CH4 at 1000◦ C on ceramic anodes was demonstrated more than 10 years ago, the reports about highcurrent densities for methane oxidation at such low temperatures are indeed surprising. Several papers indicate that a catalytic effect (due to the mixed ionic and electronic conductivity) of CeO2-x is partially responsible for this effect. However, this seems to contradict previous reports, and thus this issue deserves further analysis.

INTRODUCTION The requirement of pure hydrogen or hydrogen-rich fuels is a major obstacle for commercial application of fuel cells in general. The solid oxide fuel cell (SOFC), which is a high-temperature fuel cell with operation temperatures usually above 600◦ C, has an advantage of being tolerant to nonhydrogen fuels, e.g., CO is a SOFC fuel, in contrast to low-temperature fuel cells that are poisoned even by rather low levels of CO. Furthermore, the SOFC operating temperature is high enough to sustain the process of reforming hydrocarbons, which is the fuel of preference for most applications. This is generally regarded as another advantage of SOFC. However, steam reforming poses several problems in practice, as discussed briefly below. Therefore, several groups of researchers (see e.g., 1–12) have been working on the possibility of directly feeding the SOFC with hydrocarbons without any pretreatment of the fuel. Such direct conversion of methane and higher hydrocarbons is certainly a challenge. Methane is a fairly stable molecule that needs specific catalysts in order to react sufficiently fast even at the relative high temperatures of the SOFC, and both methane and the higher hydrocarbons have the problem of carbon precipitation (formation of solid carbon compounds) due to cracking. The higher hydrocarbons in particular are already prone to cracking at relatively low temperatures of about 500–600◦ C (13). The literature about the direct conversion of relatively dry hydrocarbons in SOFC is divergent, and several papers in peer 0084-6600/03/0801-0321$14.00

321

4 Jun 2003

12:38

322

AR

AR189-MR33-13.tex

MOGENSEN

¥

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

P1: GJB

KAMMER

review journals seem to be in direct conflict with many other papers and, to some extent, with generally accepted knowledge from the catalysis literature. We discuss this conflict, but before this literature is reviewed, some background information on steam reforming, hydrocarbon cracking, and carbon deposition in connection with SOFC is given. A minimum of this kind of knowledge is crucial for the discussion of the reports about the direct conversion of hydrocarbons in SOFC. Readers desiring a detailed description of SOFC, steam reforming, and hydrocarbon cracking are referred to review articles on these subjects (13–16).

REFORMING OF HYDROCARBONS The classical materials used by most developers of SOFC are the Ni-YSZ-cermets as the fuel electrode (YSZ = yttria-stabilized zirconia), YSZ as the electrolyte, and LSM-YSZ composite as the air electrode (LSM = lanthanum strontium manganate). The excellent electrochemical performance of Ni-YSZ-cermet electrodes is the reason for this choice. It is generally agreed that direct feeding of dry hydrocarbons into the fuel cell must be avoided when using a Ni-containing anode because Ni is an efficient catalyst for hydrocarbon cracking, and such cracking of hydrocarbon will destroy the Ni-YSZ anode (10, 16). The carbon whiskers will separate the fine Ni-particle from the YSZ. The reformation of hydrocarbons is a well-established technology that most SOFC developers are using in order to avoid the problems of cracking. The reforming of hydrocarbons may, in the SOFC context, take place either as external reforming (in a separate unit) or as internal reforming inside the hot volume of the SOFC stack. The simplest is the direct internal reforming on the Ni-YSZ-cermet anode. Steam reforming seems to be most widely used. The reactions taking place in steam reforming are Cn Hm + nH2 O ← → nCO + (n + m)/2H2 , followed by the shift reaction, CO + H2 O ← → CO2 + H2 . On the SOFC anode, these reactions are followed by electrochemical oxidation − CO + O2− ← → CO2 + 2e ,

and − H2 + O2− ← → H2 O + 2e .

There are disadvantages to using internal reforming compared with using an ideal anode (fuel electrode), which can tolerate dry hydrocarbons. The disadvantages are these:

4 Jun 2003

12:38

AR

AR189-MR33-13.tex

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

CONVERSION OF HYDROCARBONS

P1: GJB

323

(a) Steam reforming requires considerable amounts of water in the fuel [usually a steam to carbon ratio (S:C) of about two], which results in a diluted fuel. This has a negative effect on the electromotive force (EMF) of the cell and thereby the electrochemical efficiency is reduced. (b) Steam reforming is an extremely endothermic process. In the case of external reforming, this requires a considerable transport of entropy, which involves an energy loss. Internal steam reforming directly on the anode may cause large thermal gradients that can damage the cell. (c) Extra costs are involved in heat exchangers and additional equipment for steam raising or recycling of the anode exhaust gas.

DIRECT CONVERSION OF HYDROCARBONS Explanation of Concepts Over the past few years, many SOFC researchers have been working on producing SOFC anodes that can tolerate dry hydrocarbons as fuel. However, the terminology used in several of the reports has not been clear; therefore, we define our understanding of these concepts as follows. Direct conversion of a hydrocarbon means conversion in the SOFC without pre-mixing the fuel gas with steam or CO2, and without processing the fuel before it enters the cell stack. Thus direct conversion is either direct electrochemical oxidation or electrochemical oxidation of cracking products. Furthermore, direct electrochemical oxidation of a hydrocarbon means a 100% Faradaic-coupled reaction or, in other words, a hydrocarbon molecule is oxidized electrochemically only if all important (fast) reaction steps are electrochemical steps. A direct electrochemical oxidation scheme may theoretically be written as − Cx Hy + (2x + y/2)O2– ← → xCO2 + y/2H2 O + (4x + y)e .

1.

However, Reaction 1 is highly unlikely to occur in one step even in case of the simplest hydrocarbon, CH4. It might instead proceed as written below (in the case of methane): CH4 + O2– CH3 OH + 2O2– HCOOH + O2–

− ← → CH3 OH + 2e ; − ← → HCOOH + H2 O + 4e ; − ← → CO2 + H2 O + 2e .

2. 3. 4.

Cracking followed by electrochemical oxidation of the cracking products follows a path as given by Cx Hy ← 5. → xC + y/2H2 ; 2– ← − 6. C + 2O → CO2 + 4e ; − O + 2e . 7. H2 + O2– ← H → 2

4 Jun 2003

12:38

324

AR

AR189-MR33-13.tex

MOGENSEN

¥

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

P1: GJB

KAMMER

This is a greatly simplified reaction scheme because the cracking process is usually highly complex and will often result in many types of products (see below). It has been argued that the precise mechanism of the hydrocarbon conversion process is not important for practical applications. As long as the chemical energy of the hydrocarbon is transformed into electrical energy, the reaction pathway does not matter. We disagree with this. It is important to distinguish between the two reaction pathways because the demands on the anode material differ significantly for these pathways. Whereas an electrode used for direct chemical oxidation need not be inert to reactions with carbon (as no carbon is formed), oxidation of cracking products requires that the electrode material must be inactive with respect to reactions with carbon. In order to be effective, an electrode for direct electrochemical oxidation of hydrocarbon must have a very high turnover rate of the hydrocarbon bond breaking. Cracking followed by electrochemical conversion of the cracking products requires an electrode that is both highly active for the cracking of methane and active for the oxidation of the cracking products. Thus the demands on the properties of the electrodes are fundamentally different. Also, in a situation where cracking of the hydrocarbons occurs, it is possible that carbon may block the tubing of the SOFC system. In other words, it is important to know the pathway by which the conversion of hydrocarbon is to proceed in order to determine which properties of the anode should be optimized. Both processes can in principle occur in parallel. It is also noteworthy that the oxidation might not be complete but proceed only partially. An advantage of this can be the simultaneous production of electricity, heat, and a useful chemical compound, i.e., syngas production by partial oxidation of hydrocarbons as (in the case of methane) − 8. CH4 + O2– ← → CO + H2 + 2e . Indeed, this concept has been pursued with the main product being the chemical compound. In such cases, the device is usually not referred to as a SOFC; instead it is called a ceramic electrochemical reactor [see e.g., (1, 17)] or a dense oxygen separation membrane [see e.g., (18)].

Carbon Deposition In order to appreciate the published results about direct conversion of hydrocarbons in SOFCs, it is necessary to have a brief discussion of our general knowledge about carbon deposition by pyrolysis of the hydrocarbon. The text below is based on work of Albright et al. (13) and Rostrup-Nielsen et al. (16). Possible pathways to the formation of carbon could be Cx Hy ← → xC + y/2H2 , or by disproportionation of carbon monoxide CO ← → C + CO2

9.

10.

from incomplete oxidation of the hydrocarbon. Carbon can be formed both in the

4 Jun 2003

12:38

AR

AR189-MR33-13.tex

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

P1: GJB

CONVERSION OF HYDROCARBONS

325

Lighter hydrocarbons Initial hydrocarbon

Gases

+ Heavier hydrocarbons Cyclized hydrocarbons Polymerization

High-molecular-weight hydrocarbons

Condensation Tars

Soot Surface carbon

Figure 1 Diagram of the formation of carbon by pyrolysis of hydrocarbons through free-radical reactions [after Albright et al. (13) with modification].

gas phase (gas-phase carbon) and on solid surfaces (surface carbon or catalytic carbon). In general, the carbons consist of high-boiling-point polycyclic aromatics. The formation of carbon is often a minor reaction of the overall pyrolysis reaction, but in the SOFC context, it is a problematic reaction because carbon may build up in the system. The reactions leading to the formation of carbon are complex. Figure 1 illustrates some possible pathways. In the present discussion, cokes formed on catalytic surfaces are most interesting. Many types of coke can be formed, depending on the material (surface) used for the (catalytic) cracking of the hydrocarbon. At least three major families of coke exist: filamentous, amorphous, and graphitic. Filamentous coke can be divided into different subtypes depending on the structure of the coke; for example, filaments with a rope-like appearance have been found. Normally, filamentous coke contains metal particles. In contrast to this, amorphous coke contains few or no metal particles. Graphitic coke is formed mainly at high temperatures (above 900◦ C) and has a higher density than that of other types of coke. The results of the two pathways may be illustrated by the difference in carbon growth on Ni and Pt-Fe catalysts. Nickel is good catalyst for cracking of both methane and higher hydrocarbons. It catalyzes the formation of carbon nanotubes. The nanotube grows with the nickel particle at its tip. A nickel catalyst is therefore normally not stable in methane. Another mechanism for growing coke is observed for an alloy of the Pt-Fe where the tube of coke is extruded from the catalyst particle, which stays in contact with the support during the growth of the coke. The stability of the nickel catalyst can be changed by so-called stabilization, for example, formation of sufficiently small Ni particles. For example, this can be done by forming a solid solution of nickel with magnesia (Ni1-xMgxO). When the

4 Jun 2003

12:38

326

AR

AR189-MR33-13.tex

MOGENSEN

¥

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

P1: GJB

KAMMER

oxide is reduced, nickel with a small grain size may be formed, and the growth of the Ni particles is hindered by the MgO, which cannot be reduced under practical SOFC anode conditions. If the size of the Ni particles is below about 5 nm, the formation of filamentous carbon is strongly retarded. Another possibility is alloying of nickel with copper. It is well known that copper suppresses the formation of carbon because copper is of limited use as a catalyst for the C–C bond formation, which in turn suppresses the formation of carbon. It should also be noted that the effect of nickel stabilization often seems to be a prolongation of the normally observed incubation time, which elapses before carbon deposition starts. Other strategies are to avoid nickel (or other cracking catalysts) by using anode materials that are either resistant to carbon formation/deposition or able to oxidize the formed carbon. Fluorite oxides such as ceria or Ti-doped YSZ, and perovskites such as doped lanthanum chromites or doped strontium titanates, have been suggested as anode materials (2–4, 11, 12, 19). These materials are mixed conductors at elevated temperatures under the conditions of the SOFC anode compartment. The chromites have been evaluated in a stream of flowing methane with good results in regard to stability; as for coke, only a limited formation of coke of two to three monolayers is observed (11). The activity for oxidation of methane of the chromites can be enhanced by doping with MgO, for example (20).

Electrochemical Oxidation of Carbon The literature on electrochemical oxidation of carbon in aqueous solution is vast. However only very little is known about the electrochemical oxidation of carbon in an all solid-state reactor. In part this may be due to the fact that it is not easy to perform the experiments, e.g., the addition of carbon in a continuous way is difficult. Pioneering work by the Danish company Dinex Filter Technology A/S on the electrochemical oxidation of carbon in a reactor constructed by porous ceria and catalytic active electrodes has been reported (21). The work is carried out as an attempt to clean up exhaust gas from diesel engines. It is claimed that the oxidation of carbon is facilitated by the application of a potential across the reactor. However, the work is carried out under net oxidizing conditions and not in an environment that is similar to that of the anode compartment in a fuel cell. Furthermore, the oxidation is not pure electrochemically because the process is stated to be enhanced by the NEMCA effect (21) [for explanation of the NEMCA effect see (22)]. However, ceria does prevent formation of carbon and/or oxidizes it at low partial pressures of oxygen, as is found in the anode compartment of a SOFC under operating conditions. One problem with the use of ceria is its volume expansion during reduction from Ce4+ to Ce3+ (23). The reduced ceria expands and peels off the electrolyte.

Direct Electrochemical Oxidation As background information for this section, it should be mentioned that in order to have a commercial potential, the area-specific internal resistance of the SOFC must

4 Jun 2003

12:38

AR

AR189-MR33-13.tex

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

CONVERSION OF HYDROCARBONS

P1: GJB

327

not be higher than about 0.3–0.4 Äcm2. This means that the polarization resistance of the anode should not be higher than about 0.2 Äcm2 at relevant SOFC anode conditions. The direct electrochemical oxidation of methane in a SOFC was first reported in detail by Steele et al. (1). It was observed that methane could be converted into CO2 and H2O with good activity on oxide-based electrodes using Pt as a current collector. Furthermore, it was reported that the oxidation on some electrodes was only partial and that the selectivity depended on the composition of the electrode material. The reaction rate was low, however, giving rise to high polarization resistances. Ceria was also mentioned as a particularly good candidate for an anode material for direct electrochemical oxidation of CH4 (2, 3). However, later detailed investigations proved that reduced ceria is almost inactive to C–H bond breaking (7, 24, 25). These findings are illustrated in Figures 2 and 3. The measurements of Figure 2 are from the group at Imperial College (25), and we have added the scales with current densities, assuming that each CH4 turnover releases eight electrons. It is noted that the resulting limiting current densities are in the range of 1 µA/cm2 in the temperature interval of 700–800◦ C. This corresponds to a polarization resistance, Rp, in the range of 10 MÄcm2. Figure 3 shows the result of an electrochemical measurement at 1000◦ C (7). A limiting current density of around 0.1 A/cm2 was measured at overpotentials of 0.2 to 0.5 V. It was shown by gas analysis that this current density corresponded well to the limitation of

Figure 2 Turnover frequencies for reaction of dry methane with gadolinia-doped ceria and for steam reforming withn 5% methane/5.5% steam over gadolinia-doped ceria. [After Aguiar et al. (25) with our additions of the current densities based on the assumption of eight electrons per turnover.]

4 Jun 2003

12:38

328

AR

AR189-MR33-13.tex

MOGENSEN

¥

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

P1: GJB

KAMMER

Figure 3 Current density versus overpotential and electrode potential against a Pt/air reference for a Ce0.6Cd0.4O1.8 (CG) electrode on an 8 YSZ electrolyte. pCH4 = 9 kPa, pH2O = 3 kPa, bal. N2 1000◦ C, Au current collector [taken from Marina & Mogensen (7)].

hydrogen diffusion to the anode from a gas with low-hydrogen concentration. The hydrogen originated from the cracking of CH4 in the test set-up. In spite of the facts mentioned above, fast direct electrochemical oxidation of hydrocarbons in SOFCs has been claimed in the recent literature by two groups: one from Northwestern University (8, 26) and the other from the University of Pennsylvania (9, 10, 27–34). In the case of the work at Northwestern, an Rp below 1 Äcm2 at 650◦ C has been demonstrated using relatively dry methane as fuel. A thin (2 µm) Ni/YSZ anode is used, but a large improvement is achieved when adding a layer of YDC ((Y2O3)0.15(CeO2)0.85) between the YSZ electrolyte and the Ni/YSZ anode. The electrode shows no sign of carbon formation after more than 100 h of operation. The natural question here is, how can the Ni/YSZ anode be useful for the oxidation of methane when it normally is destroyed within hours? It has been pointed out by the authors that carbon formation is less pronounced at low temperatures because no carbon is formed at the anodes operating at open circuit voltage (OCV) below

4 Jun 2003

12:38

AR

AR189-MR33-13.tex

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

CONVERSION OF HYDROCARBONS

P1: GJB

329

700◦ C, whereas carbon formation is observed at temperatures above 700◦ C. This is, however, in contrast to the fact that the rate of the formation of carbon is highest at temperatures between 500 and 650◦ C (13, 35). No other investigations of the Ni/YSZ anode for methane oxidation at intermediate temperature can confirm the results by Barnett et al. (8). The YDC interphase layer is also shown to decrease the rate of formation of carbon. Another point that must be raised is the low OCV obtained by the authors (8), which is significantly lower than the potential predicted from thermodynamics. This shows that equilibrium is not obtained at the anode. Therefore, the potential is most likely a mixed potential or one with H2/H2O ratio reflecting the amount (or rate) of methane being cracked. This indicates that the reaction at the anode is proceeding either by internal reforming (due to oxygen leakage through the electrolyte) or by cracking of methane followed by electrochemical oxidation of the cracking products. In the work at the University of Pennsylvania (10), both methane and higher hydrocarbons have been tested as fuel in a SOFC with good results. The authors use a copper-ceria composite electrode. No carbon formation is observed on the copper-ceria-based electrodes after prolonged time in methane. Copper is claimed to act as a current collector, and ceria is thought to act as an electrocatalytic material in contrast to the findings shown in Figures 2 and 3. The ability of copper to form C–C bonds is limited, which probably is the reason why no carbon is formed on the electrodes. Copper cannot break the C–H bonds in methane. Ceria aids in the combustion of carbon. Nonetheless, the OCV obtained by the authors is much lower than the voltage predicted by thermodynamics, indicating that the conversion of hydrocarbon is proceeding by a pathway other than by direct oxidation, i.e., by cracking followed by electrochemical oxidation of the cracking products. In spite of this, the results are very interesting. Gorte et al. has also used Ni-Cu-ceria electrodes. An electrode containing Cu and Ni in the ratio 80%/20% (w/w) shows an Rp as low as 1 Äcm2 at 800◦ C (33) after extended time in steam-humidified methane. The electrodes were observed to improve during time owing to carbon deposition, which improved the electronic conductivity of the electrodes. Copper suppresses carbon formation; ceria combusts carbon and also suppresses carbon formation. The results are in part confirmed by (36) but with much higher Rp values, about 25 Äcm2 at 750◦ C. Also it is has been shown that the Rp values are higher with the highest content of copper. However, the authors (36) did not use ceria, which might explain the difference in the performance. Other authors have investigated the oxidation of methane on different types of materials with poor results, as seen from a technological point of view (37). An example of these results is the use of lanthanum-chromites with different dopants, i.e., Ru. These chromites showed very high Rp values. The chromites seem to be poor electrocatalysts and are probably not adequate for use as anodes for the oxidation of methane in spite of their high stability toward coking. Also, doped titanates have been investigated as electrodes for the oxidation of methane, with a high Rp of 170 Äcm2 at 930◦ C (38). In summary, interesting results have been obtained using humidified hydrocarbons as fuel with ceria-based electrodes. This conversion of hydrocarbons is probably not due to direct oxidation of methane but rather due to cracking

4 Jun 2003

12:38

330

AR

AR189-MR33-13.tex

MOGENSEN

¥

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

P1: GJB

KAMMER

followed by the oxidation of the cracking compounds. The use of ceria seems to be a means for obtaining robust and highly active electrodes for the oxidation of hydrocarbons.

CONCLUDING REMARKS In a strict sense, direct electrochemical oxidation of hydrocarbon on mixed conductors such as reduced and/or doped ceria does not occur with a reaction rate that is of any technological interest. The high reaction rates/current densities reported in several recent papers are instead based on a cracking of the hydrocarbons on metals in the experimental set up or on homogeneous cracking in the gas phase, followed by electrochemical oxidation of hydrogen and, in the case of ceria, probably electrochemical oxidization of carbon. A distinction should be made between CH4, which is a fairly stable molecule that does not crack easily unless in the presence of a metal cracking catalyst such as nickel, and the higher hydrocarbons, which crack in the gas phase at temperatures of about 600◦ C. The fact that ceria is very resistant to carbon precipitation is probably also of great importance for maintaining the reactivity of the anode. Of further importance is the fact that carbon formation does not take place on copper, which is not an efficient methane cracker even at 900◦ C. Even though the literature about direct conversion of hydrocarbon in SOFCs seems promising, great care should be taken in the assessment of the possibility of using untreated hydrocarbons as fuels in SOFC systems. The fact that dry hydrocarbons are prone to cracking under SOFC conditions means that there is a need for much more work to be done, both with respect to clarifying the conversion mechanisms and with tests relevant to real SOFC technology. ACKNOWLEDGMENTS This review was done as part of the Danish solid oxide fuel cell program, DKSOFCb, contract No. 1713/01-0001. The authors thank the Danish Energy Agency for financial support. The Annual Review of Materials Research is online at http://matsci.annualreviews.org

LITERATURE CITED 1. Steele BCH, Kelly I, Middleton H, Rudkin R. 1988. Solid State Ionics 28–30:1547–52 2. Mogensen M, Bentzen JJ. 1989. In Proc. 1st Int. Symp. SOFC, Electrochem. Soc. Proceedings, ed. SC Singhal, 89–11:99– 110. Pennington, NJ: Electrochem. Soc.

3. Steele BCH, Middleton PH, Rudkin RA. 1990. Solid State Ionics 40–41: 388 4. Norby T, Osborg PA, Dyrlie O, Hildrum R, Seiersten M, Glenne R. 1994. In Proc. First Eur. Solid Oxide Fuel Cell Forum, ed.

4 Jun 2003

12:38

AR

AR189-MR33-13.tex

AR189-MR33-13.sgm

LaTeX2e(2002/01/18)

CONVERSION OF HYDROCARBONS

5. 6.

7. 8. 9. 10. 11.

12. 13.

14.

15. 16. 17. 18.

19. 20. 21.

22.

U Bossel, pp. 217–26. Baden, Switz.: Eur. Fuel Cell Forum Putna ES, Stubenrauch J, Vohs JM, Gorte RJ. 1995. Langmuir 11:4832 Marina OA, Bagger C, Primdahl S, Mogensen M. 1999. Solid State Ionics 123:199–208 Marina OA, Mogensen M. 1999. Appl. Catal. A 189:117–26 Perry Murray E, Tsai T, Barnett SA. 1999. Nature 400:649 Park S, Vohs JM, Gorte RJ. 2000. Nature 404:265–67 Gorte RJ, Kim H, Vohs JM. 2002. J. Power Sources 106:10–15 Sfeir J, Buffat PA, Moeckli P, Xanthopoulos N, Vasquez R, et al. 2001. J. Catal. 202:229–44 Kaiser A, Feighery AJ, Fagg DP, Irvine JTS. 1998. Ionics 4:215–19 Albright LF, Crynes BL, Corcoran WH. 1983. Pyrolysis: Theory and Industrial Practice, pp. 47–87. New York: Academic Minh, NQ, Takahashi T. 1995. Science and Technology of Ceramic Fuel Cells. Amsterdam: Elsevier Rostrup-Nielsen JR. 1984. Catalytic Steam Reforming. Berlin: Springer-Verlag Rostrup-Nielsen JR, Sehested J, Nørskov JK. 2002. Adv. Catal. 47:65–139 Velle OJ, Andersen A, Jens K-J. 1990. Catal. Today 6:567–74 Hendriksen PV, Larsen PH, Mogensen M, Poulsen FW, Wiik K. 2000. Catal. Today 56:283–95 Marina OA, Canfield NL, Stevenson JW. 2002. Solid State Ionics 149:21–28 Saracco G, Scibilia G, Iannibello A, Baldi G. 1996. Appl. Catal. B Environ. 8:229–44 Christensen H, Dinesen J, Engell H, Larsen LC, Hansen KK, Skou EM. 2000. SAE Tech. Pap. 01–0478 Nicole J, Tsiplakides D, Pliangos C, Very-

23.

24.

25.

26.

27. 28. 29. 30. 31.

32. 33. 34. 35. 36. 37. 38.

P1: GJB

331

kios XE, Comninellis C, Vayenas CG. 2001. J. Catal. 204:23–34 Mogensen M, Lindegaard T, Hansen UR, Mogensen G. 1994. J. Electrochem. Soc. 141:2122–28 Ramirez-Cabrera E, Atkinson A, Chadwick D. 2000. In Proc. 4th Eur. SOFC Forum, ed. AJ McEvoy, pp. 49–58. Oberrohrdorf, Switz.: Eur. Fuel Cell Forum Aguiar P, Ramirez-Cabrera E, Atkinson A, Kershenbaum LS, Chadwick D. 2001. In Electrochem. Soc. Proc. SOFC VII, ed. H Yokokawa, S Singhal, pp. 703. Pennington, NJ: Electrochem. Soc. Liu J, Madsen BD, Ji Z, Barnet SA. 2002. Electrochem. Solid State Lett. 5(6):A122– 24 Park S, Gorte RJ, Vohs JM. 2000. Appl. Catal. A 200:55–61 Gorte R, Park S, Vohs JM, Wang C. 2000. Adv. Mater. 12:1465–69 Park S, Gorte RJ, Vohs JM. 2001. J. Electrochem. 148:A443–47 Kim H, Park S, Vohs JM, Gorte RJ. 2001. J. Electrochem. 148:A693–95 Wang C, Worrell WL, Park S, Vohs JM, Gorte RJ. 2001. J. Electrochem. 148:A864– 68 Kim H, Vohs JM, Gorte RJ. 2001. Chem. Commun. 22:2334–35 Kim H, Lu C, Worrell WL, Vohs JM, Gorte RJ. 2002. J. Electrochem. 149:A247–50 McIntosh S, Vohs JM, Gorte RJ. 2002. Electrochem. Acta 47:3815–21 Zhu T, Flytzani-Stephanopoulos IM. 2001. Appl. Catal. A 208:403 L¨u Z, Pei L, He T-M, Huang X-Q, Liu Z-G, et al. 2002. J. Alloys Comp. 334:299–303 Sauvet A-L, Fouletier J. 2001. Electrochem. Acta 47:987–95 Kelaidopoulou A, Siddle A, Dicks AL, Kaiser A, Irvine JTS. 2001. Fuel Cells 3– 4:219–25