A High-performance Solid Oxide Fuel Cell Anode Based on Lanthanum Strontium

Journal of Power Sources 196 (2011) 7488 7494

Contents lists available at ScienceDirect

Journal of Power Sources

jou rna l h omepa g e: www.elsev ier .com

A high- bvanada

Jong-Sun hea Department o , 1910b Department o

a r t i c l

Article history:Received 23 FeAccepted 12 MAvailable onlin

Keywords:Fuel cellsCeramic compCatalystStrontium vanComposite anoLSV

inlr useand Ym ov

poretivityhe abn election o

for p

1. Introduction

Solid oxide fuel cells (SOFC) have great potential as energy-conversion conductorswith high estate-of-thetrolyte matexcellent elin their abilof Ni to cataple, Ni-baseand H2) or but operatimal conditielectrodes hand shut-doforms NiO [

Electrodsome of thedo not catvolumetric expected to

CorresponE-mail add

Unfortunately, the performance of ceramic fuel electrodes is usu-ally very poor because of the following factors. First, it is difcultto achieve good conductivity with oxides under reducing condi-

0378-7753/$ doi:10.1016/j.devices. Because the electrolytes are usually oxygen-ion, SOFC can in principle operate on any combustible fuelfciency and a low output of pollutants. At present,-art SOFC anodes are composites of Ni and the elec-erial. These ceramicmetal (cermet) composites exhibitectrochemical and catalytic performance but are limitedity to operate on hydrocarbon fuels due to the tendencylyze the formation of carbon laments [16]. For exam-d anodes are able to operate on syngas (a mixture of COmethane if sufcient steam is co-fed with the Ni [7];on on larger hydrocarbons is not practical under nor-ons [8]. Furthermore, even the best engineered Ni-basedave only limited tolerance to oxidation during start-upwn cycles due to the expansion that occurs when Ni9].es based on electronically conductive oxides could solve

problems associated with Ni-based electrodes. Oxidesalyze the formation of lamentous carbon and thechanges associated with oxidation and reduction are

be less than that for conversion between Ni and NiO.

ding author. Tel.: +1 215 898 4439; fax: +1 215 573 2093.ress: [email protected] (R.J. Gorte).

tions [10]. Second, the reactivity of most oxides with YSZ preventsone from using the high-temperature processing methods that giveNi cermets their nearly ideal interfacial structure with the elec-trolyte [11]. Third, compared to metals like Ni, oxides are very pooroxidation catalysts [12,13].

To minimize these problems, our groups have been focusing onthe preparation of composite electrodes using inltration meth-ods [1417]. This fabrication procedure involves rst synthesizinga porous layer of the electrolyte that has been pre-sintered to thedense electrolyte, then inltrating the catalytic and electronicallyconductive components into the porous scaffold [14]. With com-posites prepared by inltration, the sintering temperature for theelectrolyte component of the composite electrode can be muchhigher so as to establish better ion-conducting channels from theelectrolyte into the electrode. Furthermore, the composites formedby inltration do not have a random structure, so that sufcientconductivity can be achieved using lower loadings of the electronicconductor [18]. The non-random structure also causes the coef-cients of thermal expansion (CTE) of the composite to be closerto that of the electrolyte scaffold than to the weighted average ofthe components [19,20]. The effects of expansion due to redox ofthe electronic conductor are minimized [17,21]. Finally, catalyticcomponents can be added separately if required [2224].

While we have been able to achieve very good performance withfuel electrodes based on inltrated La0.8Sr0.2Cr0.5Mn0.5O3 (LSCM)

see front matter 2011 Elsevier B.V. All rights reserved.jpowsour.2011.05.028performance solid oxide fuel cell anodete

g Parka, Ian D. Hassonb, Michael D. Grossb, Chen Cf Chemical and Biomolecular Engineering, University of Pennsylvania, Philadelphia, PAf Chemical Engineering, Bucknell University, Lewisburg, PA, 17837, USA

e i n f o

bruary 2011ay 2011e 19 May 2011

osite

adatede

a b s t r a c t

Ceramic composites were prepared bystabilized zirconia (YSZ) and tested fofor solid-state reaction between LSV at 973 K, LSV formed a continuous lto 973 K, with the reduction formingtrodes showed high electronic conduca conductivity of 2 S cm1 at 973 K. In tatively poor performance; however, a973 K in humidied H2 following addistable in CH4 but there was evidenceoxidation./ locate / jpowsour

ased on lanthanum strontium

na, J.M. Vohsa, R.J. Gortea,

4, USA

tration of La0.7Sr0.3VO3.85 (LSV) into porous scaffolds of yttria- as solid oxide fuel cell (SOFC) anodes. There was no evidenceSZ at calcination temperatures up to 1273 K. For calcinationer the YSZ. The LSV phase reduced easily upon heating in H2s in the LSV and greatly increasing its surface area. The elec-

after reduction, with a 10-vol% LSVYSZ composite exhibitingsence of an added catalyst, the LSVYSZ electrodes showed rel-trode impedance of approximately 0.1 cm2 was achieved atf 0.5 vol% Pd and 2.8 vol% ceria The LSVYSZ composites wereoisoning of the Pd catalyst by V following high-temperature

2011 Elsevier B.V. All rights reserved.

J.-S. Park et al. / Journal of Power Sources 196 (2011) 7488 7494 7489

[22,23], La0.3Sr0.7TiO3 (LST) [25] and SrMoO3 [17] when a sepa-rate catalyst is added, each of these ceramic conductors has seriouslimitations. In the case of LSCM, the conductivity is only marginal;under reducing conditions at 973 K, the conductivity of bulk LSCMis less thanwt% LSCM ilower. Ideal1 S cm1. Tunder norma conductivreduce and will be the [27].

Sr-dopedbecause itsnamic inveindicated t1018 atm aditions for Nconventionhave resulteat temperatvious workgood tolera

In the prof SOFC anoWe will deelectronic cseparate ox

2. Experim

The electri-layer, YSsandwichedof the densporous regicated by la1773 K for 4formulationslurry and telsewhere including sused to predensity of tity of the ladetermined

To prepacursor soluSr(NO3)2 (9acid (one mwater. The imately 3 hsolutions wand drying at either 97chiometry ocompositestion of La(N99%), Mn(Ning of 20-vdescribed ein this studsite the ano1123 K for 4

[37,38]. The anode composites were prepared before the LSFYSZcathodes in all cases except for when the LSVYSZ composites werecalcined to only 973 K.

As a nal step in cell preparation, CeO2 (added as 1 M Ce(NO3)3,sar, 9n, Alce cae voal polculare 2.e fral%, rets, thnles

on ccatal

fuel-ic a

12, S bothanceiostaidi

at 0 Hz, YSZSZ pastinn the

1 mnduc

in hed using Krmo

inv temnd ad red00 mvestiicros

ults

rder, we

the r0.3)Vent wre ofsizesatter(JCPDcturplitte duebtain73 Kple

d thedditir2O7tate 2 S cm1 and composites formed by inltration of 45-n yttria-stabilized zirconia (YSZ) are a factor of 20 timesly, the composite should have a conductivity of at leasthe conductivity of bulk LST is only marginally betteral reduction conditions [26]. While bulk SrMoO3 hasity of nearly 1000 S cm1 [17], SrMoO4 is difcult toSrMoO3SrMoO4 equilibrium data suggest that SrMoO4stable phase under typical SOFC operating conditions

LaVO3 (LSV) is a particularly interesting material conductivity can be very high [2830]. A thermody-stigation of the equilibrium between LaVO4 and LaVO3hat the reduced phase will be stable at P(O2) belowt 973 K [31], which is similar to the equilibrium con-i. Although LSV has been used in SOFC fuel electrodes,

al methods for preparing electrodes from this oxided in electrodes with only modest performance and onlyures well above 973 K [3235]. Interestingly, this pre-

has shown LSV to have a low reactivity with YSZ andnce to sulfur [33,34].esent study, we set out to investigate the performancedes prepared by inltration of LSV into YSZ scaffolds.monstrate that the LSVYSZ composites exhibit highonductivities and good electrode performance when aidation catalyst is added to the composite.

ental

trochemical cells used in this study were prepared fromZ wafers, in which an 80-m dense electrolyte disc was

between two 40-m porous YSZ layers. The diametere part of each wafer was 1 cm but the diameter of theons was only 0.67 cm. The tri-layer wafers were fabri-minating three green tapes followed by calcination at

h. The tapes used in this study were cast using a slurry based on organic solvents. The composition of thehe procedures used to prepare the tapes are described[36]. Porosity in the sintered layers was achieved byynthetic graphite (300 mesh, Alfa Aesar) in the tapespare the porous YSZ layers. By comparing the apparenthe porous YSZ to the bulk density of YSZ, the poros-yers made from tapes with graphite pore formers was

to be 73 1%.re composites of LSV and YSZ, we rst prepared pre-tions consisting of La(NO3)36H2O (99.9%, Alfa Aesar),9.9%, Alfa Aesar), NH4VO3 (99%, Alfa Aesar), and citricole per mole of metal cations) dissolved in deionizedsolutions were heated to 330 K and stirred for approx-

until the solutions stopped bubbling. The precursorere added to the porous YSZ by multiple inltrationcycles to a nal loading of 10-vol% LSV, then calcined3 K or 1373 K for 1 h. Unless otherwise noted, the stoi-f the oxidized LSV was (La0.7Sr0.3)VO3.85. The LSCMYSZ

were prepared by inltration with an aqueous solu-O3)36H2O, Sr(NO3)2, Cr(NO3)39H2O (Acros Organics,O3)26H2O (Alfa Aesar, 99.98%), and citric acid to a load-ol% LSCM, followed by calcination to 1473 K for 4 h, aslsewhere [23]. La0.8Sr0.2FeO3 (LSF)YSZ electrodes usedy were fabricated by inltrating the porous YSZ oppo-de layer to a loading of 20-vol% LSF, then calcined at

h. The details of this procedure are described elsewhere

Alfa Aesolutioenhanand thresiduThe caites wevolum0.42 voponenions, uductedof the

Fora ceram11209tion atimpedpotentto humsured to 0.01

LSVfrom Ytape cused i1 mm The comentsobtainics) us

Theused toin air. A973 K ation anwas 2also intron m

3. Res

In ophasesshows(La0.7Sconsiststructutallite XRD pLaVO3tal strushow ssonablwere oand 13the samYSZ antains ato La2Zsolid-s9.5%) and Pd (0.45 M tetraammine palladium(II) nitratefa Aesar, 99.9%) were added to the anode composites totalytic activity [22]. Each solution was inltrated twicelume % of ceria and Pd were calculated based on there volumes remaining after adding either LSV or LSCM.ted electrode volume fractions in the LSVYSZ compos-8-vol% CeO2 and 0.50-vol% Pd, while the correspondingctions in the LSCMYSZ composites were 2.4 vol% andspectively. Following the addition of the catalytic com-e cells were calcined to 773 K to decompose the nitrates otherwise indicated. Some experiments were con-ells that were calcined to 973 K following the additionyst.cell testing, cells were attached to alumina tube usingdhesive (Ceramabond 552, Aremco). Ag paste (lot no.PI supplies) and Ag wire were used for current collec-

the anode and cathode. Vi polarization curves and spectra were measured using a Gamry Instrumentst with the cathode exposed to air and the anode exposeded (3% H2O) H2 or CH4. Impedance spectra were mea-.1 V bias voltage in the frequency range from 100 kHz

with a 20-mV AC perturbation. and LSCMYSZ composites were also fabricatedorous slabs prepared from the same slurry used ing the porous layers. The dimensions of the slabs

conductivity and BET-isotherm measurements werem 10 mm and 0.5 cm 0.5 cm 1.8 cm, respectively.tivities were determined using 4-probe, DC measure-umidied (3% H2O) H2. The BET surface areas wereing a Tristar II 3020 surface area analyzer (Micromerit-r adsorption at 78 K.gravimetric analysis (SDT Q600, TA Instruments) wasestigate the reduction of LSV in 4% H2/Ar and oxidationperature ramp of 5 K min1 from room temperature to

gas ow rate of 50 mL min1 was used for both oxida-uction experiments. The weight of the powder sampleg. The physical properties of LSVYSZ composites were

gated using X-ray diffraction (XRD) and scanning elec-copy (SEM).

and discussion

to demonstrate that inltration produced the proper examined the XRD patterns of various samples. Fig. 1apattern for the bulk powder having the stoichiometryO3.85 after calcination in air at 973 K. The pattern isith that which would be expected for the monoclinic

LaVO4, with relatively broad features due to small crys-. After reduction in humidied H2 at 973 K for 5 h, then changed dramatically and became similar to that ofS card No. 81-2436), which has an orthorhombic crys-

e. The major peaks in the pattern for the reduced LSVing, indicative of a tetragonal structure, which is rea-

to the presence of Sr. The XRD patterns in Fig. 1c and ded for LSVYSZ composites after calcination at 1273 K, respectively, and reduction at 973 K. The pattern for

calcined at 1273 K shows only peaks associated with reduced LSV phase. The sample calcined at 1373 K con-onal small peaks at 28.2 and 31.8 2, which are dueand possibly Sr3V2O8, respectively, showing that some

reaction had occurred.

7490 J.-S. Park et al. / Journal of Power Sources 196 (2011) 7488 7494

Fig. 1. XRD pareduction at 91373 K, follow

TGA memine the exa qualitativresults are soxidation astoichiometoxidized staextent of reand was neapartially limOxidation o573 K.

Fig. 2. TGA cuLa0.7Sr0.3VO2.8

Table 1Specic surface areas of LSVYSZ composites fabricated at 973 K or 1373 K, beforeand after reduction.

Fabrication

973 1373

SEM micand 1373 KFig. 3a shohas a sponbetween 1 calcinationa uniform, f973 K, Fig. shown in Fiup into irreYSZ surfacethe LSV stru

The chanlar to whatthat case, it

YSZto beuponre mthe is incllowductdata

elecV calare s2. Intterns of (a) (La0.7Sr0.3)VO3.85 synthesized at 973 K in air; (b) LSV after73 K for 5 h in H2. LSVYSZ composite fabricated at (c) 1273 K and (d)ed by reduction at 973 K for 5 h in H2.

asurements were performed on LSV powder to deter-tent to which the sample is reduced at 973 K and providee estimate of ease with which the LSV is reduced. Thehown in Fig. 2. The reversible, 5.7% weight change uponnd reduction corresponds to a change in the oxygenry of 0.9 (i.e., a change from (La0.7Sr0.3)VO3.85 in thete to (La0.7Sr0.3)VO2.94), which is close to the expected

ing theorder layers ites we973 K, and th5 h. Fowith reThese

Thevol% LSLSCM, ied Hduction. Reduction in 4% H296% Ar began below 873 Krly complete by 973 K. It is likely that the reduction wasited by the rate at which H2 was supplied to the sample.f the reduced LSV began at a much lower temperature,

rves for the reduction of La0.7Sr0.3VO3.85 in 4% H2/Ar and oxidation of5 in air.

with similaslightly higthis study, ature. The results is lipresent woimportant iconductivitthe high-poand Petric [imately 100the high conreasonable.to decrease

Similar LSCMYSZ order to achis shown infor three cethe anodescined at 973The third celyst. All threelectrolytesfrom LSVYopen-circuithe maximuEven with ttemp. (K) Surface area (m2 g1)

Before reduction After reduction

0.13 1.140.14 0.68

rographs of the LSVYSZ composites calcined at 973 K, before and after reduction at 973 K, are shown in Fig. 3.ws the YSZ scaffold before adding LSV. The scaffoldge-like character, with characteristic pore dimensionsand 3 m. Following the addition of 10-vol% LSV and

at 973 K, the entire scaffold appears to be coated witheatureless lm, as shown in Fig. 3b. Upon reduction at3c shows that cracks have formed in the LSV lm. Asg. 3d, calcination to 1373 K caused the LSV lm to breakgular-shaped features that no longer cover the entire. After reduction, Fig. 3e, small cracks again appear inctures.ges observed in the LSV following reduction are simi-

has been observed previously for LSCM in YSZ [39]. In was argued that these cracks were critical for expos-

and creating a long three-phase boundary (TPB). Intter quantify the changes that occurred in the LSV

reduction, the surface areas of the LSVYSZ compos-easured using BET isotherms. Following calcination at

LSVYSZ composite had a surface area of 0.13 m2 g1

reased to 1.14 m2 g1 after reduction in H2 at 973 K foring calcination at 1373 K, the increase in surface areaion, from 0.14 m2 g1 to 0.68 m2 g1, was less dramatic.are summarized in Table 1.trical conductivities of LSVYSZ composites, with 10-cined at 973 K, and LSCMYSZ composites, with 20-vol%hown in Fig. 4 as a function of temperature in humid-

a previous study of inltrated, LSCMYSZ compositesr LSCM loadings [23], the electrical conductivities wereher, near 0.1 S cm1 at 973 K compared to 0.04 S cm1 inbut showed a similar increase with increasing temper-difference between these and the previous LSCMYSZkely due to the fact that the YSZ scaffold used in therk has a higher porosity, 73% compared to 65%. Mores the fact that LSVYSZ composite has a much highery, >2 S cm1 at 973 K, even with only 10-vol% loading inrosity YSZ scaffold. Based on the previous work of Hui29], the conductivity of bulk La0.7Sr0.3VO3 is approx-

S cm1 at 1073 K under reducing conditions, so thatductivities we observe for the LSVYSZ composites are

The conductivity of the LSVYSZ composite was found slightly with increasing temperature.to what was found with SOFC anodes made withcomposites [22,23], it is necessary to add a catalyst inieve high performance with LSVYSZ composites. This

Fig. 5. In Fig. 5a, Vi polarization curves are shownlls operating at 973 K with humidied H2 owing over. Two of these cells had anodes with 10-vol% LSV, cal-

K; but catalysts, Pd and ceria, were added to only one.ll had an anode with 20-vol% LSCM and a Pd/ceria cata-

e cells had identical LSFYSZ cathodes and 80-m thick. The effect of adding the catalyst to the anodes madeSZ was dramatic. While both cells exhibited the propert voltage (OCV), 1.1 V, addition of the catalyst increasedm power density from 90 mW cm2 to 470 mW cm2.he Pd/ceria catalyst, the cell with the LSCMYSZ anode


Fig. 3. SEM imbefore and (e)

performed mum powe

The impdifferencesmined fromages of (a) the initial YSZ scaffold. The LSVYSZ composite calcined at 973 K (b) before a after reduction.

slightly worse than the LSVYSZ cell, showing a maxi-r density of only 370 mW cm2.edance spectra shown in Fig. 5b help to explain the

between the three cells. First, the ohmic losses deter- the high-frequency intercepts with the real axis in the

Nyquist ploSince the co[40], this o80-m eleLSVYSZ cond (c) after reduction. The LSVYSZ composite calcined at 1373 K (d)

t were identical for the two LSVYSZ cells, 0.43 cm2.nductivity of YSZ at 973 K is reported to be 0.019 S cm1

hmic loss is consistent with that expected for thectrolyte. Based on the measured conductivity of themposites in Fig. 4, 2 S cm1, the ohmic contribution


Fig. 4. Electrical conductivities of LSVYSZ () and LSCMYSZ composites () mea-sured in humidied (3% H2O) H2, using the DC 4-probes method.

Fig. 5. (a) Vi polarization curves and (b) impedance spectra of cells at 973 K usingLSVYSZ anode with () and without () CeO2 and Pd. Data for the LSCMYSZ anodewith CeO2 and Pd ().

Fig. 6. (a) Vi anode with Ce

from the 4The additionon-ohmicto 0.2 clar to whatLSCM [22,2also found that the LSthe LSVYS0.1 cm2

The impperformancthis cell wehigher, as e0.04 S cm1

electronic cTo deter

carbon fuelPd and ceriColeCole p(3% H2O) CHilar to thospolarization curves and (b) impedance spectra of cell having LSVYSZO2 and Pd in humidied (3% H2O) H2 () or CH4 () at 1073 K.

0-m anode is expected to be less than 0.002 cm2.n of the Pd/ceria catalyst caused a large decrease in the

losses, from 2.5 cm2 for the cell without the catalystm2 for the cell with the catalyst. This result is simi-

has been reported for anodes made from inltrated3] and LST [25], where the addition of a catalyst wasto be critical for obtaining high performance. GivenFYSZ cathode likely contributes 0.1 cm2 [37,38],Z anode with the Pd/ceria catalyst must contribute.edance data in Fig. 5b also help to explain the lowere of the LSCMYSZ cell. While the non-ohmic losses inre nearly identical, the ohmic losses were 0.1 cm2xpected for a 40-m electrode with a conductivity of. This result points out the benet of achieving higheronductivity with the LSVYSZ composite.mine the stability of LSVYSZ composites in hydro-s, we examined the performance of the cell with thea catalyst in methane. The Vi polarization curves andlots for the cell operating at 1073 K in both humidied4 and H2 are shown in Fig. 6. The results are very sim-

e obtained previously for a cell based on an LSCMYSZ


Fig. 7. (a) Vi H2 for cells ha

composite wboth fuels, 1higher in H510 mW cmwith H2, theThe impedalosses wereohmic lossethere mustthis molecuin CH4 bec[41]; howeof approximthe methan

The effeshown in FH2 at 973 Kor 1373 K. Fmation of LSand calcinerelationshipOCV of 1.1 Vpower dens1373-K celllosses were

i polith P

ite to 9

ancn-ohatesYS

affen-ohB dueed inFig. 8. Vanodes wcompos

performthe noit indicthe LSVicantlythe nothe TPobservpolarization curves and (b) impedance spectra in humidied (3% H2O)ving LSV fabricated at 973 K () or 1373 K ().

ith Pd and ceria added [23]. The OCV was the same for.1 V; but the maximum power density was signicantly2 than in CH4, 960 mW cm2 in H2 compared to only2 in CH4. While the Vi relationship was a straight linere was noticeable curvature near open circuit with CH4.nce data, measured at open circuit, show that ohmic

independent of fuel and 0.2 cm2, but that the non-s were signicantly higher in CH4, demonstrating that

be catalytic limitations to breaking the CH bonds inle. We did not test long-term stability of these anodesause it is known that Pd will form carbon over timever, the electrodes were stable over the testing periodately a day, suggesting that the LSV is not affected bye.ct of LSV calcination temperature on the electrodes isig. 7 through a comparison of fuel-cell performance in

in which the 10-vol% LSV was calcined to either 973 Kor the 1373-K cell, the LSF cathode was added after for-V; in both cells, the Pd and ceria were added separately

d to only 773 K prior to cell testing. For both cells, the Vi in Fig. 7a was a straight line, starting from the expected, although the 973-K cell exhibited a higher maximumity of 470 mW cm2 compared to 410 mW cm2 for the. The impedance data in Fig. 7b show that the ohmic

the same for both cells, 0.42 cm2, and that the lower

We did oPd/ceria caheated in amance, as scalcination973 K decreall of this chcell losses. ScompositesThis is a pothe anode pbased on co

From exit is knowncial catalytiSince vanadit seems likanodes is dthe LSV. Toa series of sof the LSV pLSV was calthen calcinthat vanadithat phase vanadium i

The resuthat the ohtially unchacatalyst caldensities wcompared ations in thnon-ohmic

While thlishing thatin electrodearization curves in humidied (3% H2O) H2 for cells having LSVYSZd/ceria catalysts. Data are shown before () and after () heating the73 K in air.

e of the 1373-K cell is associated with an increase inmic losses to 0.3 cm2. This result is important in that

the impurity phases observed following calcination ofZ composite to 1373 K in Fig. 1 apparently do not signif-ct the conductivity within the electrode. The increase inmic losses is likely caused by a decrease in the length of

to the change in the morphology of the inltrated LSV, Fig. 3.bserve a potential problem of LSV interacting with the

talyst. When an LSVYSZ cell with Pd and ceria wasir to 973 K, there was a signicant decrease in perfor-hown by the Vi polarization curves in Fig. 8. Following

in air, the maximum power density in humidied H2 atased from 470 mW cm2 to 280 mW cm2. As expected,ange was associated with an increase in the non-ohmicimilar treatments for electrodes made from LSCMYSZ

with Pd and ceria had no effect on cell performance.tentially serious problem because the ability to oxidizeeriodically is one of the advantages offered by anodesnducting ceramics.periments conducted with LSCMYSZ electrodes [22],

that the catalytic metal, Pd in this case, is the cru-c component in anodes based on conducting ceramics.ium is a known poison in heterogeneous catalysis [42],ely that the deterioration in performance of the LSVue to poisoning of the Pd catalyst by vanadium from

test this hypothesis, we measured the performance ofimilar fuel cells in which the vanadium stoichiometry

hase, La0.7Sr0.3V1 + O2.85, was varied. In each case, thecined to 973 K prior to the addition of Pd and ceria anded again to 973 K. The idea behind the experiments wasum should be less able to migrate from the LSV phase ifis decient in vanadium and more able to migrate if thes in excess.lts from this experiment are shown in Table 2. Notemic losses were between 0.43 and 0.45 cm2, essen-nged by the vanadium stoichiometry or the increasedcination temperature. However, the maximum powerere higher in the vanadium-decient cell, 330 mW cm2

to the vanadium-excess cell, 250 mW cm2. The vari-e power densities were matched by variations in the

losses in each of the cells.e results in Table 2 are not entirely denitive in estab-

vanadium poisoning is responsible for the decrease performance following the exposure of the anode to


Table 2Maximum power densities of the cells at 973 K in humidied H2, with anode having10 vol% LSV, together with Pd and ceria. In these experiments, the vanadium contentof the LSV phase (La0.7Sr0.3V1 + O2.85) was varied and the cells were heated to 973 Kin air to study the effect of interactions between the LSV phase and the Pd/ceriacatalyst.

Non-stoichiometryof LSV

Power density(mW cm2)

Ohmic ASR( cm2)

Non-ohmicASR ( cm2)

= 0.05 326 0.43 0.49 = 0 280 0.45 0.61 = 0.05 254 0.45 0.71

air at high temperatures, there are methods that could be used tominimize contact between the Pd catalyst and the LSV phase. Forexample, it has recently been demonstrated that Pd sintering inLSCMYSZ electrodes could be dramatically reduced through theuse of Pd/ceria, coreshell particles [43], in which Pd nanopar-ticles were surrounded by a thin shell of ceria, prepared usingself-assembly methods [44]. It seems likely that the same oxideshell that mbetween ththat needs t

The mosgreat similaanodes formconductive case suggestronic condby a secondously differand in theirtrolyte phaselectrolyte trated compHowever, itelectrode in

4. Conclus

Ceramicprepared bmaterial redary structurinsufcientaddition of impedances

Acknowled

We thanwork.

References

[1] S. McIntosh, R.J. Gorte, Chem. Rev. 104 (2004) 4845.[2] H. Kim, C. Lu, W.L. Worrell, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 149 (2002)

A247.[3] M.L. Toebes, J.H. Bitter, A.J. van Dillen, K.P. de Jong, Catal. Today 76 (2002) 33.[4] C.W. Sun, U. Stimming, J. Power Sources 171 (2007) 247.[5] S. Helveg, C. Lopez-Cartes, J. Sehested, P.L. Hansen, B.S. Clausen, J.R. Rostrup-

Nielsen, F. Abild-Pedersen, J.K. Norskov, Nature 427 (2004) 426.[6] K. Yamaji, H. Kishimoto, Y. Xiong, T. Horita, N. Sakai, M.E. Brito, H. Yokokawa,

J. Power Sources 159 (2006) 885.[7] S. Singhal, Ind. Ceram. 28 (2008) 53.[8] M. Cimenti, J.M. Hill, Energies 2 (2009) 377.[9] M. Cassidy, G. Lindsay, K. Kendall, J. Power Sources 61 (1996) 189192.

[10] A. Atkinson, S. Barnett, R.J. Gorte, J.T.S. Irvine, A.J. Mcevoy, M. Mogensen, S.C.Singhal, J. Vohs, Nat. Mater. 3 (2004) 1727.

[11] S. de Souza, S.J. Visco, L.C. De Jonghe, J. Electrochem. Soc. 144 (1997) L35.[12] R.J. Gorte, J.M. Vohs, J. Catal. 216 (2003) 477.[13] S. Park, R.J. Gorte, J.M. Vohs, Appl. Catal. A 200 (2000) 55.[14] R.J. Gorte, S. Park, J.M. Vohs, C. Wang, Adv. Mater. 12 (2000) 1465.[15] R. Craciun, R.J. Gorte, J.M. Vohs, C. Wang, W.L. Worrell, J. Electrochem. Soc. 146

(1999) 4019.[16] M.D. Gross, K.M. Carver, M.A. Deighan, A. Schenkel, B.M. Smith, A.Z. Yee, J.

Electrochem. Soc. 156 (2009) B540.. Smith, M.D. Gross, Electrochem. Sol-State Lett. 14 (2011) B1.e, Y.

04) 33uang. Vohs. Busa08) B1im, S

id-Staim, G

(2008. Groee, G. . Fu, Faldych11) 28naba, 21Rqiang heng,. Shaheng, Jheng,te Lettguilar

Ge, S.Hngas6.201Wanguangorre, 09) 10asaki,. Kim,.A Ko

reira, . Kim,ctrochCargn

(201itigates Pd sintering may also minimize interactionse metal and the LSV phase. This is obviously somethingo be investigated further.t important conclusion to be taken from this work is therities in the reported data for all of the ceramic SOFCed by inltration [17,23,25]. Whether the inltrated,

ceramic was LSCM, LST, SrMoO3, or LSV, the data in eacht that the ceramic component simply provides elec-uctivity and that the catalytic function must be supplied

component which is added separately. There are obvi-ences in the level of conductivity of the various ceramics

tendency to undergo solid-state reactions with the elec-e. Interactions between the conductive ceramic and thephase also seem to affect the morphology of the inl-onent, which can in turn affect the length of the TPB.

appears that one can optimize each component of thedividually to maximize performance.

ions

composites with high electronic conductivity can bey inltration of La0.7Sr0.3VO3.85 into porous YSZ. Thisuces easily and forms an effective three-phase bound-e within the YSZ scaffold. The resulting composite has

catalytic activity for optimal electrode performance buta Pd/ceria catalyst produces electrodes with very low

in both H2 and CH4 fuels.

gement

k the US Ofce of Naval Research for support for this

[17] B.H[18] H. H

(20[19] Y. H[20] J.M[21] A.N

(20[22] G. K

Sol[23] G. K

11 [24] M.D[25] S. L[26] Q.X[27] I. B

(20[28] F. I

R22[29] Shi[30] Z. C[31] P.R[32] C. P[33] Z. C

Sta[34] L. A

17.[35] M. [36] R. K

291[37] W. [38] Y. H[39] G. C

(20[40] K. S[41] J.-S[42] M.D

Fer[43] J.-S

Ele[44] M.

132Huang, J. Regal, M. Boaro, J.M. Vohs, R.J. Gorte, J. Am. Ceram. Soc. 871., K. Ahn, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 151 (2004) A1592., R.J. Gorte, Adv. Mater. 21 (2009) 943.won, D. Sarantaridis, A. Atkinson, Electrochem. Solid-State Lett. 1186.

. Lee, J.Y. Shin, G. Corre, J.T.S. Irvine, J.M. Vohs, R.J. Gorte, Electrochem.te Lett. 12 (2009) B48.. Corre, J.T.S. Irvine, J.M. Vohs, R.J. Gorte, Electrochem. Solid State Lett.) B16.ss, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 154 (2007) B694B699.Kim, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 155 (2008) B1179.. Tietz, D. Stover, J. Electrochem. Soc. 153 (2006) D74.ev, A. Javadekar, D.J. Buttrey, J.M. Vohs, R.J. Gorte, Appl. Catal. A 3947.T. Arima, T. Ishikawa, T. Katsufuji, Y. Tokura, Phys. Rev. B 52 (1995)2224.Hui, Anthony Petric, Solid State Ionics 143 (2001) 275.

S.W. Zha, L. Aguilar, M.L. Liu, Solid State Ionics 176 (2005) 1921., M.M. Khader, J.M. Vohs, R.J. Gorte, J. Phys. Chem. C 112 (2008) 2613..L. Luo, A.R. Sanger, K.T. Chuang, Chem. Mater. 22 (2010) 1032.

S.W. Zha, L. Aguilar, D. Wang, J. Winnick, M.L. Liu, Electrochem. Solid. 9 (2006) A31., S.W. Zha, Z. Cheng, J. Winnick, M.L. Liu, J. Power Sources 135 (2004)

. Chan, J. Electrochem. Soc. 156 (2009) B386., J.-S. Kim, J.M. Vohs, R.J. Gorte, J. Am. Ceram. Soc., doi:10.1111/j.1551-0.04359.x., M.D. Gross, J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 154 (2007) B439., J.M. Vohs, R.J. Gorte, J. Electrochem. Soc. 151 (2004) A646.G. Kim, M. Cassidy, J.M. Vohs, R.J. Gorte, J.T.S. Irvine, Chem. Mater. 2177.

J. Maier, Solid State Ionics 134 (2000) 303. V.V. Nair, J.M. Vohs, R.J. Gorte, Scripta Mater. 65 (2011) 90.rn, D.S.S. do Santos, B. Welz, M.G.R. Vale, A.P. Teixeira, D.D. Lima, S.L.C.Talanta 73 (2007) 1.

Noah M. Cargnello, L. Wieder, R.J. Gorte, P. Fornasiero, J.M. Vohs, J.em. Soc. 158 (2011) B596.ello, N.L. Wieder, T. Montini, R.J. Gorte, P. Fornasiero, J. Am. Chem. Soc.0) 1402.

A high-performance solid oxide fuel cell anode based on lanthanum strontium vanadate1 Introduction2 Experimental3 Results and discussion4 ConclusionsAcknowledgementReferences

Documents

A High-performance Solid Oxide Fuel Cell Anode Based on Lanthanum Strontium