Efficient and Flexible SOFC syst

Document Sample
Efficient and Flexible SOFC syst Powered By Docstoc
					(Registration number: 2001EF004)

                               Efficient and Flexible SOFC system

Research Coordinator
       Harumi Yokokawa                     National Institute of Advanced Industrial Science and Technology
                                                                           : JAPAN
Research Team Members
       Truls Norby                         University of Oslo              : NORWAY
       Koichi Eguchi                       University of Kyoto             : JAPAN
       Ellen Ivers-Tiffée                  University of Karlsruhe         : GERMANY

Duration    April, 2001 – March, 2004

    Scandia     stabilized   zirconia    (ScSZ),    iron   doped     calcium    titanate   (Ca(Ti0.9Fe0.1)O3)    and
(Zr0.5Ce0.5)0.65Y0.35O1.825 were selected as common anode components inside FLEXSYS team.              Tests on cell
performance revealed that the Ni/ScSZ cermet anode exhibited the best performance for the hydrocarbon
fuels among the presently tested anodes, namely, Ni/YSZ, Ni/ScSZ, Ni/(Ce,Gd)O2, Ni/(Zr0.5Ce0.5)0.65Y0.35
O1.825, Ni/Ca(Ti0.9Fe0.1)O3. The surface reaction rate and the proton solubility of ScSZ were essentially the
same as of YSZ despite high activity of Ni/ScSZ; the high ability of supplying oxygen seems crucial in
avoiding carbon deposition.       For Ca(Ti0.9Fe0.1)O3 having the proton conductivity, high isotope exchange
reaction rate and high hydrogen dissociation ability, the activity was not good but the stability against carbon
deposition was better than Ni/YSZ.       The FLEXSYS cell consisting of ScSZ electrolyte with Ni/ScSZ was
tested and demonstrated to show better performance than the conventional Ni/YSZ.

Keywords: Solid oxide fuel cell, Fuel flexibility, Energy conversion, Hydrocarbon fuels, Carbon deposition

1. Introduction
    Solid oxide fuel cells (SOFCs)[1,2] have many advantages against the polymer electrolyte fuel cells (PEFCs).
Particularly, higher energy-conversion efficiency can be achieved without using expensive precious metals.             In
addition, the long-term stability can be also achieved by selecting appropriate materials combinations. On the other
hand, SOFCs have a week point that it is technologically difficult to construct stacks because of the chemical and
mechanical instability during fabrication and thermal cycles[2]. Even so, well designed stacks have been constructed
and have survived for frequent thermal cycles or operations longer than 70,000 h. This makes it clear that SOFCs are
suitable for large power generators and also for many other applications in a rather small size. It should be noticed,
however, that in a moderate size, it is not easy to obtain high efficiency and fuel flexibility simultaneously because gas
circulation and heat transfer could not be made
                                                                                 VO•• + 0.5O2(g) = OO + 2h•
effectively in a small system.    Instead, the present                           VO•• + 0.5O2(g) + 2e- = OO
FLEXSYS team aims at development of flexible but                 Air                    O2
                                                                 Electrode                            H2O(g)+ VO•• =O2-+2 Hi•
still efficient SOFC systems in a moderate size
(several kW class) by adopting advanced methods of
                                                                  Electrolyte                h•    H+
feeding fuels to small SOFCs.     The FLEXSYS team                                    O2-                     O2 -
                                                                             VO••                   VO••
thus decided to make the primary focus on the
anode[3] side of the cells and the reforming[4]                  Fuel                       H2O
                                                                 Electrode      e- H2
    The main focuses in the FLEXSYS project are                Fig. 1 Schematic electrochemical reaction mechanism
placed to investigate the direct feeding/oxidation of          with strong interaction with water vapors.

fuels (methane, propane and butane)             in   the
FLEXSYS cells consisting of alternative as well as standard anodes.          The direct oxidation [5] is the interdisciplinary
subjects covering catalyst, electrochemistry, solid state ionics etc. In the FLEXSYS project, this will be investigated
with emphases on the metal/oxide/gas interactions; that is, possible effects of water vapors [6,7] on oxygen exchange
reaction rate and electrode activity as illustrated in Fig. 1; water vapor is one of the reaction product of direct oxidation
of hydrocarbons in SOFCs so that reactions between water vapor with hydrocarbons on metal/oxide interfaces will
become crucial.    So far, the acid-base theory provides basis of understanding the chemical behavior in catalytic
reactions and proton affinities. The FLEXSYS project aims at bridging among different treatments on basicity in
catalyst and solid-state electrochemistry. The FLEXSYS project does not attempt only tests of new ideas in a small
size of experimental cells but also will test the FLEXSYS concept in rather large cells that can be applicable to one kW
class system.   Another focus will be placed on modeling for a small but efficient SOFC system by analyzing the
FLEXSYS cells and the energy flow of those systems containing the FLEXSYS cells.

2 Materials and Methods
    The followings are the important decisions of FLEXSYS teams on the interesting but wide-spread research areas:
1) Since “Fuel Flexibility” is the most important concept in the FLEXSYS project, anodes will be focused but effects on
cathode will not be focused; 2) Model electrodes and model fuels are selected by considering the followings:
Karlsruhe will test the FLEXSYS cells in a large size (10 x 10 cm) with a focus on well characterized materials,
whereas Kyoto and Oslo will use smaller cells to test a wide range of materials.             Tsukuba will focus mainly on the
SIMS investigation on the anode related materials and on the thermodynamic considerations on basicity. 3) As common
materials, scandia-stabilized zirconia (ScSZ), iron-doped calcium titanate and (ZrO2)0.325(CeO2)0.325(YO1.5)0.35 were
    Investigation subjects of the respective groups are summarized in Fig. 2.                     Tsukuba (AIST) will make
measurements on proton solubility, surface reaction rate and electrical properties of the oxides in relation with
alternative anodes and reforming catalysts to be investigated in Kyoto, Oslo and Karlsruhe. Oslo also focuses on the
fundamental properties related with electrocatalytic and catalytic behaviors as functions of water vapors, temperature,
                Karlsruhe Univ.                                                                Kyoto Univ.
                                                                           Exergy analysis
                   Single cell Test                             Fuel Processing
                                          Model Fuels                             Catalytic reforming //
                                       with/without water                         Carbon deposition.
                                                                                  Cermet for internal reforming
                                                                                  Single cell Operation
                                        Model Electrodes
                Microstructure        standard : alternative?
                multilayer anodes                                                               Oslo Univ.
                                                         Surface Exchange Kinetics
                                      Correlation        Transport Parameters
                   AIST                                  Thermodynamics of Water Solubility

             Fig 2. Schematic illustration to show the research topics and their mutual relations in the
             FLEXSYS project.

oxygen partial pressure etc. Kyoto analyzes effects of oxides in cermet (ceramics and metal) anodes and reforming
catalysts.   Since carbon deposition is the key phenomena when direct fuel feeding/oxidation is attempted, Kyoto
clarifies the determining factor of carbon deposition in thick cermet anodes. Some additives are selected among the
basic oxides. In the FLEXSYS project, the basicity can be correlated with the proton solubility, the thermodynamic
activity of basic oxides, surface reaction rate etc. The surface property, surface reaction rate, and the mass transport in
solids will be considered on the basis of many accumulated and newly created knowledge in the FLEXSYS project.
     In order to clarify the system image of the FLEXSYS project, the modeling of cell performance will be made in
Karlsruhe, exergy analysis being in Kyoto.      Test on the FLEXSYS cells will be tested in Karlsruhe using gas and
liquid fuels on the long term behavior, degradation on fuel direct oxidation etc. This test will become the key
experiments in the FLEXSYS project.

3. Results
     3.1 Water/Oxide Interaction
     In order to understand the electrochemical performance, it is essential to know the interactions of the oxide
component in electrodes and gaseous species such as oxygen, hydrogen, water vapor etc.          Particularly, the interaction
of water vapor with zirconia or ceria has been examined in Tsukuba with an emphasis on the proton solubility, surface
reaction rate of oxygen isotope exchange, effects of water vapor on the surface reaction rate etc. The thermodynamic
analysis has been also made on the doped ceria or zirconia-ceria solid solutions and has clarified that the geometrical
arrangement of oxide ion vacancies relative to the dopant lattice position governs the thermodynamic properties and
also behavior of protons and holes. In Oslo, literature survey has been made on the proton conduction in a large
number of oxides, and the empirical correlations were derived.         Experimental investigation showed no effects of
protons in La2NiO4 and LaFeO3.        These investigations give the physicochemical basis of understanding the roles of
oxides in anodes.
    3.2 Design for Reforming Catalysts and Anodes

                                                                      log (DO* / m2 s-1), log (α / m s-1),
    Scandia-stabilized zirconia (ScSZ), which exhibits                                                        -7
the higher oxide ionic conductivity than YSZ, is                                                              -8
interesting from the oxide/metal interactions in cermet                                                                                               -1
anodes, because Ukai et al.[8] have already shown that                                                                                                -2
Ni/ScSZ cermet is stable in slightly humidified methane                                                      -10
whereas Ni/YSZ cermet is not. Tsukuba revealed that
the surface oxygen isotope exchange reaction rate, α, is                                                                                              -4
                                                                                                                                                    log θ
essentially the same as YSZ as shown in Fig. 3; the
surface exchange reaction rate is enhanced in the                                                            -13

presence of the water vapor and the magnitude of the
enhancement can be correlated with the surface coverage                                                            2             3              4

of water molecule on the ScSZ.          In addition, no                                                                 log(p(H2O) / Pa)

significant effects of Fe doping to ScSZ were observed.        Fig. 3 Oxygen isotope diffusivity (DO*) and surface
                                                               exchange rate constant (α) of ScSZ (open symbols)
A similar insignificant effect was observed on Mn-doped        and Fe-ScSZ (closed symbols) as functions of water
YSZ, although the electronic conductivity is enhanced.         partial pressure. Solid line was derived from the
                                                               concentration of water chemisorbed on YSZ quoted
The present results lead to the following important            from Ras et al.)
considerations that not the surface property, but the high
ionic conductivity in ScSZ is important in avoiding
carbon deposition inside cermet anodes.         In other
wards, the high oxygen flow rate is critical.          In                                                              600       500            400
degradation behavior of the ionic conductivity of ScSZ
                                                                α / m s-1

was measured in Karlsruhe as in an order of 4% / 1000 h.                                            10-9
   The second common material, Ca(Ti0.9Fe0.1)O3,
exhibits the proton conductivity as confirmed in Oslo.                                  10-10
In addition, Oslo also observed that hydrogen or hydride
ions can exist in this material in the same manner as in
                                                                 D*O / m2 s-1

SrTiO3.   This material in the same batch has been
distributed to respective groups.      In Tsukuba, the
surface exchange reaction rate and the oxide ion                                        10-13                          CaTi0.9Fe0.1O3 Praxair
diffusivity, DO*, were measured; as given in Fig. 4, the                                                               p(H218O) = 2 kPa, p(18O2) = 7 kPa
surface reaction rate is high, whereas the oxide ion                                    10-14
                                                                                                                   1.1 1.2 1.3 1.4 1.5 1.6
diffusivity is low when compared with other perovskite                                                                               kK / T
oxides such as lanthanum cobaltites.       In Oslo, the      Fig. 4 Oxygen isotope diffusivity, DO*, and surface
                                                             exchange rate constant, α, of CaTi0.9Fe0.1 O3-δ in
dissociation rate on oxide materials has been measured
                                                             humid atmosphere: ■; DO*, ○; α
for the oxygen as well as hydrogen dissociation reactions
                                                                                                                                                          Fe dopedTiO2
                                                                La0.9Sr0.1FeO3-δ calc
                                       -4                                                                                                                 Nd-doped CeO2
                                  10                             Ce0.9Gd0.1O2-δ
                                                                                                                               -6                         TiO2
                                                                 CeO2-δ                                                   10

                                                                                        -2 -1



                                                                                          Dissosciation rate / mol X cm

    Dissociation rate / mol O s






                                       -8                                                                                      -9
                                  10                                                                                      10

                                            1.0   1.2     1.4       1.6      1.8                                                    0.8   1.0       1.2        1.4
                                                        1000K/T                                                                                 1000/T (K )
Fig. 5. (a) The oxygen dissociation rate and (b) the hydrogen dissociation rate over different oxides at a total pressure of
20 mbar (2 kPa) as a function of the inverse temperature. CFT=Ca(Ti.9Fe.1)O3.

(Fig. 5). As shown in Fig. 5(b), Ca(Ti0.9Fe0.1)O3 exhibits very high values for the hydrogen dissociation rate; this
should be compared with the rather high surface reaction rate given in Fig. 4. Kyoto examined the catalytic activity
for steam reforming and shift reactions; the catalytic activity of Ca(Ti0.9Fe0.1)O3 is found to be quite low. As a
Ni/Ca(Ti,Fe)O3 cermet anode, several characteristic properties have been investigated in Oslo, Kyoto and Karlsruhe.
The first attempts on Ni/Ca(Ti,Fe)O3 cermet in Kyoto revealed the followings: (1) their anode activity is generally
lowered than Ni/YSZ for H2 and CH4 fuels; (2) heat treatment at a higher temperature than 1300 ºC caused some
degradation. Similar phenomena associated with heat treatment were also observed in Oslo and Karlsruhe. During
the experiments with CH4, low OCV has been observed in Kyoto (Fig. 6). This gave rise to some discussions within
the team, mainly because with hydrogen fuels, it is rare to observe low OCV and as a result, low OCV indicates some
inappropriateness in experimentation. However, with CH4 fuels, methane is not the directly electrochemical active
species but hydrogen or CO will be active. This implies that the OCV value depends on water vapor/CO2 vapor
pressure as well as hydrogen vapor pressure. This makes the situation complicated. The carbon deposition rate on
the Ni/Ca(Ti0.9Fe0.1)O3 cermet was observed to be smaller than that on the Ni/YSZ. This better performance against
the carbon deposition is due partly to the proton solubility in the cermet anode, which is consistent with the mechanism
proposed from AIST Tsukuba. Although Ni/Ca(Ti0.9Fe0.1)O3 cermet is not better than Ni/YSZ, the durability is better
particularly when the Ni to Ca(Ti0.9Fe0.1)O3 ratio of 1/1 is adopted as shown in Fig. 7.
        The metal component of cermet anodes has been investigated in Karlsruhe. Copper was the first candidate for
current collector because copper is inactive to methane decomposition reactions. However, cooper was found to be

                                              Ni/CaTi0.9Fe0.1Oxide calcined at 1200
                         1.2                  Ni/CaTi0.9Fe0.1Oxide calcined at 1300
                                              Ni/CaTi0.9Fe0.1Oxide calcined at 1400

                                                                                              Terminal voltage / V
 Terminal voltage / V

                         1.0                  Ni/YSZ calcined at 1400



                                                                                                                     0.45                                            Ni-FCT (4:1)
                                                                                                                                                                     Ni-FCT (1:1)
                               0    0.1      0.2       0.3       0.4        0.5         0.6                          0.40
                                                                                                                            0             5            10             15            20
                                          Current density / Acm-2                                                                               Operating time / h

                        Fig.6 Current-voltage characteristics of                                                            Fig.7 Terminal voltage vs. operating time for CH4-fueled
                        CH4-fueled       SOFCs     with      Ni-                                                            SOFCs at 1000ºC with Ni-Ca(Ti-0.9Fe0.1)O3 anodes at under
                        Ca(Ti0.9Fe0.1)O3 oxide.                                                                             a constant current load of 200 mA/cm2.

instable in a cell due to the evaporation as CuH vapors. This is consistent with the thermodynamic prediction made in
Tsukuba. In Karlsruhe, tests were made on Ni-Cu, Ni-Fe, Ni-Al. Among them, Ni-Al was found to be promising as
current collector.
                        Redox stability of cermet anodes has been tested in Karlsruhe for conventional anodes with different firing
temperature. It has been found that the microstructure seriously affects the redox stability and the multilayer-structure
anode improves the adhesion of anodes to electrolyte substrate and as a result, long term stability. Furthermore, Sc
stabilized zirconia has been fabricated to produce the electrolyte-support cells after optimizing sintering process by
adopting new control techniques; that is, the rate controlled mass loss and the rate controlled sintering for binder
burnout processes.
                        Fuel flexibility has been examined by Kyoto and Karlsruhe. Kyoto first investigated the effects of precious
metals added to Ni/YSZ cermets for steam reforming reaction and anode reactions. They found that the presence of
Ru or Pt is effective in prompting the reforming reaction and suppressing the carbon deposition. As a result, this is
also effective in improving anode activity for methane and water under the internal reforming condition.
                        Discussion has been made on the roles of oxides and metals in cermet anodes. As shown in Fig. 5, Oslo observed
that the hydrogen dissociation rate is quite high for metals such as Pt and Ni, but quite low for oxides such as YSZ or
ceria. This strongly suggests that as one of anode components, metals are needed. Furthermore, the selection of
oxides should be crucial to obtain those anodes which can be applied to various fuel situations. Tsukuba proposed one
material (ZrO2)0.325(CeO2)0.325(YO1.5)0.35.                                             To examine the appropriateness as the anode component, the electron
conductivity, the chemical volume expansion and other properties related to the anode behavior have been examined in
Tsukuba.                           The phase stability is found to be improved by increasing the content of YO1.5 compared with the
corresponding solid solution with YO1.5 content of 0.2. However, the cell test in Tsukuba showed the worse anode
activity compared with ScSZ cermet anodes. This material was sent to Karlsruhe to make cell tests.
                 3.3 Characterization of FLEXSYS Cells in Experiments and Simulation
                 In Karlsruhe, the FLEXSYS single cells have been tested with conventional fuels, propane and butane. The main
   point of the test with conventional fuel (hydrogen) is to examine the degradation rate at the high current density and at
   high fuel utilization. Some improvement has been obtained by adopting the two layer anode, leading to the decrease in
   degradation from 5%/1000 h to 3%/1000 h.
                 The first attempt on methane without adding water vapor (S/C=0) using the conventional Ni/YSZ cermet revealed
   that the cell performance showed no degradation for more than 1000 h. This is, however, due to the oxygen or water
   vapor that was transported as a result of electrochemical reactions. Under the OCV condition where no such a gas will
   be available, the anode had serious damage on the nickel part due to carbon depositions. When S/C ratio was changed,
   the best performance was obtained at S/C=0.5.
                 For propane and butane fuels, carbon deposition becomes severer and operation with S/C=0 could not be made.
   Thus, steam reforming, partial oxidation and their combination were adopted as fuel processing system.                                                                                     Under the
   condition where carbon deposition can be avoided, namely, S/C (steam to carbon ratio) =2 to 3 or lambda value (air to
   fuel ratio) = 0.2 to 0.3, good performance was obtained without serious problem associated with carbon deposition.
   Furthermore, no difference was observed between fuels, indicating that the fuel processing system worked very well.
   Figures 8 and 9 show the experimental results on the common materials in Karlsruhe. Figure 8 is for the operation for
   hydrocarbon fuels with Ni-ScSZ anode, whereas Fig. 9 is for the methane operation with different S/C ratios on
                 Table 1 summarizes the results of cell tests made in Karlsruhe. From the results for 5% humidified hydrogen, the
   standard anode activity can be judged. From the results at different S/C ratios, it can be derived information on
   durability against carbon depositions.
                 When comparison is made between Ni/YSZ and Ni/ScSZ, the performance of Ni/ScSZ is better for
   hydrocarbon fuels.               However, the operation at S/C=0 was failed due to carbon deposition at 950 °C.                                                                                    This is
   different from Ukai et al.[8].                  The Ni/CaTi0.9Fe.1O3-d cermet anode exhibited poor performance when compared
   with other fluorite-type oxides.                   Even so, this anode can be used for S/C=0 at 950 and 800 °C. The Ni/CZYO

                                                                                                                                                 B2704AK.047, B2704AK.048, B2704AK.049, B2704AK.051
                                                           b0901bk.002, b0901bk.004, b0901bk.005


           V                             H2 = 200 ml/min (28 A)
                                         + 60 % H2O                                                                1.0

                   0.7 V                                                                                            V
          0.6                                                                                                      0.7
                                                                                                                                S/C = 0
                                                C4H10 = 15 ml/min (28 A)
          0.4                                   λ = 0,1      S/C = 3                                                            S/C = 1

                                                                                                   0.5                          S/C = 2
                                                C3H8 = 20 ml/min (28 A)
                                                λ = 0,1      S/C = 3                                                            S/C = 3

          0.0                                                                                      0.0             0.0
                0.00       0.05   0.10      0.15        0.20        0.25 A/cm² 0.30                                   0.00                0.05                 0.10        A/cm²         0.15
                                                       Current density                                                                                               current density

                 Fig. 8 Hydrocarbon operation of a Ni/SSZ-anode at                                                           Fig. 9 Ni/Ca(Ti0.9Fe.1)O3 anode operated with
                 T=800°C                                                                                                     methane at different S/C ratios at T=800°C
{(ZrO2)0.325(CeO2)0.325 (YO1.5)0.35} cermet showed about the same performance with other anodes Ni/YSZ,
Ni/ScSZ.    This is slightly different from the result in Tsukuba probably because of better powder processing.
    Gas analysis has been made on the internal steam reaction along the anode gas channel at 800 and 950 °C. At
950°C, essentially the same reforming activity was obtained for Ni/YSZ and Ni/CGO, whereas some dependence on
anode materials appears at 800 °C; a higher methane conversion is obtained for Ni/CGO. This information is needed
to model FLEXSYS cell performance.          Modeling the temperature distribution and the local gas composition at the
anode side has been carried out by computational fluid dynamics (CFD) with the software FLUENT. This model
already took into account the gas flow, chemical reactions (methane reforming, shift), heat transfer and diffusion in
porous media.    Furthermore, the electrochemical oxidation for different species such as hydrogen and CO was
incorporated in the model.
     3.4 Demonstration of FLEXSYS Cells and Stacks and System Analysis
    Conceptual design of the FLEXSYS system and evaluation of efficiency attained by improved fuel flexibility has
been made by Kyoto on the basis of planar cells with taking into account the reforming and shift reactions. After
incorporating this model to the commercial process simulator ASPEN Plus, process simulation was made to clarify the
difference between the internal and the external reforming of methane at the S/C ratio = 2 at 1000ºC and Ufuel= 0.8.
Results showed that the internal reforming leads to the conversion efficiency of 54.7%(HHV) compared with the 49.8%
for external reforming. This is due to better utilization of heat evolved from the cells.
    In Karlsruhe, a five-cell-stack test bench, which was installed initially as a fuel cell diagnosis system, was modified
in the fuel processing system and the fuel supplying system (evaporator for liquid fuel and water). Single cells for the
FLEXSYS stack have been prepared; cells consist of ScSZ electrolyte with a strontium lanthanum manganite cathode
and a Ni/ScSZ anode. Stack tests were carried out in June/July 2004.

4. Discussions
     4.1 Role of Proton Solubility/Conductivity in Anodes with Hydrocarbon Fuels:
   The present FLEXSYS cell is based on the idea that protons in oxides may play critical roles in determining the
anode reaction mechanism as illustrated in Fig. 1. The present investigation revealed that no difference was observed

                                      Table 1: Performance of different anodes

     Temperature     Fuel            Ni/YSZ         Ni/ScSZ         Ni/CGO           Ni/CZYO         Ni/CFT
                                     j at 0.7 V     j at 0.7 V      j at 0.7 V       j at 0.7 V      j at 0.7 V
                                     Standard       Common I                         Common III      Common II
        950°C        H2, 5% H2O       0.64 A/cm2     0.72 A/cm2      0.63 A/cm2       0.68 A/cm2      0.49 A/cm2
                     CH4 S/C=2        0.35 A/cm2           -         0.41 A/cm2             -          0.3 A/cm2
                     CH4 S/C=1        0.49 A/cm2           -         0.54 A/cm2       0.53 A/cm2      0.36 A/cm2
                     CH4 S/C=0        0.51 A/cm2           *               -                *         0.43 A/cm2
        800°C        H2, 5% H2O       0.17 A/cm2     0.35 A/cm2       0.2 A/cm2       0.25 A/cm2      0.16 A/cm2
                     CH4 S/C=2        0.11 A/cm2     0.22 A/cm2       0.1 A/cm2             -         0.06 A/cm2
                     CH4 S/C=1        0.14 A/cm2     0.22 A/cm2      0.13 A/cm2             -         0.09 A/cm2
                     CH4 S/C=0        0.13 A/cm2     0.12 A/cm2      0.17 A/cm2             -          0.11 A/cm2
    * failure: carbon deposition at S/C=0
      CGO=(Ce0.8Gd0.2)O1.9; CZYO=((ZrO2)0.325(CeO2)0.325(YO1.5)0.35; CFT=Ca(Ti0.9Fe0.1)O3
between ScSZ and YSZ in the effect of water vapors on the surface reaction rate whereas apparent difference is
observed in the electrochemical performance as listed in Table 1. This indicates the importance of supplying oxide
ions in the electrolyte on the basis of the proposed mechanism of transferring oxygen from electrolytes to nickel metal
surface in the form of water vapor; this may enable direct electrochemical oxidations of hydrocarbons on the nickel
surfaces. The present results on the CaTi0.9Fe0.1O3 anodes also revealed that stability against carbon deposition is
improved in this anode partly because of its high proton conductivity. The poor performance of Ni/ CaTi0.9Fe0.1O3 can
be ascribed to its peculiar microstructure caused during the high temperature treatment and also to the poor electron and
oxide ion conductivity. This makes it necessary to compare with those oxides having higher proton solubility and
higher electron conductivity.    In view of this, the present results on the Ni/CGO and Ni/(ZrO2)0.325(CeO2)0.325
(YO1.5)0.35 seem to be interesting because of their mixed conductivity and proton solubility. Against this expectation,
these anodes showed worse performance than Ni/ScSZ. One possible explanation is that both anodes exhibit still large
chemical volume expansion associated with the reduction of tetravalent cerium ions, another being that both may have
some chemical or mechanical instability with ScSZ electrolyte.
     4.2 Microstructure of FLEXSYS Cell Anodes
   The present results revealed that although no degradation was observed in the operation with dry methane for more
than 1000 h, carbon deposition became severe for the OCV condition. This strongly suggests that anodes should
consist of several components having their respectively specified functions in the anode-current collection assembly.
In many cases, nickel current collector is used. However, this nickel current collector is weakest against carbon
deposition because this is far from the anode reaction sites from which water vapors will be emitted and this is exposed
to fuels before reforming. The present results showed that Ni-Al alloys will be one of possible current collectors.
From a similar reason, the oxide current collector will be also interesting. One possible candidate is YO1.5 doped
CaTiO3; this is analogous oxide to LaO1.5 doped SrTiO3 that has been intensively investigated in USA. Even so, these
materials have poor chemical stability.   Further thermodynamic considerations lead to ZrO2-CeO2-MO1.5(M=Y or Sc)
solid solutions.   As described above, these materials showed worse or comparable performance as the oxide
component in cermet anodes. This is because the chemical volume expansion associated with cerium reduction is still
large even when the cerium content in solid solutions is reduced by keeping some level of electron conductivity.
5. Conclusions
   The systematic approach has been adopted to clarify the soundness of the FLEXSYS concept for achieving the
efficient and fuel flexible SOFC system in a moderate size. Experimental analyses have been made on the catalytic
activity, electrical properties, dimensional stability and other characteristics for the oxide components and the metallic
components. For the oxide components, several common oxides are selected in addition to the standard oxides such as
YSZ or CGO from the view point of oxide ion conductivity, proton conductivity, and electron conductivity.          For the
metallic components, Cu and other alloys are examined in addition to Ni. In the present investigation, the best cell
performance in the operation for hydrogen as well as hydrocarbons at various S/C ratios was achieved for Ni/ScSZ that
was selected as the first common material. Although the second common material, Ni/CaTi0.9Fe0.1O3, exhibited worse
performance, its stability against carbon deposition is remarkable, suggesting that some role of proton conduction in
anode mechanism.     The present results on the third common material, Ni/(ZrO2)0.325(CeO2)0.325(YO1.5)0.35 implies that
as the oxide component in cermet anodes, the chemical volume expansion associated with cerium reduction should be
reduced to provide better performance. To demonstrate the feasibility of the FLEXSYS cells, the 5-cell stack based on
the ScSZ electrolyte with Ni/ScSZ anode was operated in June/July 2004.
6. References
[1] S. C. Singhal and K. Kendall ed. “High Temperature Solid Oxide Fuel Cells: Fundamental, Design and
    Application,” Elsevier, (2003).
[2] H. Yokokawa and N. Sakai, “Chapter 13 History of high temperature fuel cell development,” Volume 1, pp.219
    -266, W. Vielstich et al. ed. Handbook of Fuel Cells Fundamentals Technology and Applications, Wiley (2003).
[3] E. Ivers- Tiffée, A. V. Virkir, “Chapter 9 Electrode Polarizations,” in ref (1), pp229-260.
[4] K. Eguchi, “Chapter 75 Internal Reforming,” pp. 1057-1069, in ref(2).
[5] S. A. Barnett, “Chapter 78 Direct hydrocarbon SOFCs,” pp. 1098-1108, in ref(2).
[6] N. Sakai et al. PCCP 5, 2253-2256 (2003).
[7] T. Norby, Nature 410, 877-878 (2001).
[8] K. Ukai, Y. Mizutani, Y. Kume        “Current Status of SOFC development using Scandia Doped Zirconia,” pp.
    375-383, in Solid Oxide Fuel Cells VII, H. Yokokawa and S. C. Singhal ed.. ECS PV-2001-16, (2001)
The list of the most important papers from the project
[1] N. Sakai et al. “Effect of water on oxygen transport properties in electrolyte surface in solid oxide fuel cell,” J.
    Electrochem. Soc. 150, A689-A694 (2003).
[2] A. C. Müller, A. Weber, E. Ivers-Tiffée, “Intrinsic Degradation of Solid Electrolytes based on Zirconia,” Solid State
    Ionics in press.
[3] T. Takeguchi, T. Yano, R. Kikuchi, K. Eguchi, “Effect of precious metal addition to Ni-YSZ cermet on reforming of
    CH4 and electrochemical activity as SOFC anode,” Catalyst Today 84, 217-222 (2003).
[4] H. Yokokawa, T. Horita, N. Sakai, K. Yamaji, M. E. Brito, Y.P. Xiong and H. Kishimoto, “Protons in Ceria and their
    Roles in SOFC Electrode Reactions from Thermodynamic and SIMS Analyses,” Solid State Ionics in press.
[5] T. Norby, “High temperature proton conductors –properties and applications,” British Ceramic Proc. 63, 1-8(2001).
[1] H. Yokokawa, et al. “Study on an Efficient and Flexible SOFC System,” Solid Oxide Fuel Cells VIII, Paris, (2003).
[2] N. Sakai et al. “Significant effect of water on surface reaction and related electrochemical properties of mixed
    conducting oxide (invited),” 14th Int. Conf. Solid State Ionics, Monterey, (2003).
[3] T. Norby, “Hydrogen in oxides,” 78th Bunsen discussion meeting, Vaals, (2002).
[4] T. Takeguchi et al. “Effect of Additive to NI-YSZ cermet on reforming CH4 and electrochemical activity for
    SOFC,” SOFC VIII, Paris, (2003).
[5] D. Fourquet et al. “SOFC Single Cell Test Setup For the Use of Various Hydrocarbons,” SOFC VIII, Paris (2003).
[1] Harumi Yokokawa, Outstanding Achievement Award of High Temperature Materials Division, The Electrochemical
    Society, 2002.10.21, in recognition of his contributions to the practical applications of thermochemistry to high
    temperature materials research and technology, especially in the area of solid oxide fuel cells.

Shared By:
Tags: SOFC