A comparative study of Sm0.5Sr0.5MO3−δ (M = Co and Mn) as oxygen reduction electrodes for solid oxide fuel cells

ww.sciencedirect.com

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 7

Available online at w

journal homepage: www.elsevier .com/locate/he

A comparative study of Sm0.5Sr0.5MO3Ld (M [ Co and Mn)as oxygen reduction electrodes for solid oxide fuel cells

Feifei Dong a, Dengjie Chen a, Ran Ran a, Heejung Park b, Chan Kwak b,**, Zongping Shao a,*a State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemistry & Chemical Engineering,

Nanjing University of Technology, No. 5 Xin Mofan Road, Nanjing 210009, PR Chinab Samsung Advanced Institute of Technology (SAIT), 14-1 Nongseo-dong, Yongin-si, Gyunggi-do 446-712, South Korea

a r t i c l e i n f o

Article history:

Received 22 October 2011

Received in revised form

28 November 2011

Accepted 30 November 2011

Available online 22 December 2011

Keywords:

Solid oxide fuel cells

Perovskite

Sm0.5Sr0.5CoO3�d

Sm0.5Sr0.5MnO3�d

Cathode

* Corresponding author. Tel.: þ86 25 8317 22** Corresponding author. Tel.: þ82 31 280 672

E-mail addresses: [email protected]/$ e see front matter Copyright ªdoi:10.1016/j.ijhydene.2011.11.150

a b s t r a c t

Sm0.5Sr0.5MO3�d (M ¼ Co and Mn) materials are synthesized, and their properties and

performance as cathodes for solid oxide fuel cells (SOFCs) on Sm0.2Ce0.8O1.9 (SDC) and

Y0.16Zr0.92O2.08 (YSZ) electrolytes are comparatively studied. The phase structure, thermal

expansion behavior, oxygen mobility, oxygen vacancy concentration and electrical

conductivity of the oxides are systematically investigated. Sm0.5Sr0.5CoO3�d (SSC) has

a much larger oxygen vacancy concentration, electrical conductivity and TEC than

Sm0.5Sr0.5MnO3�d (SSM). A powder reaction demonstrates that SSM is more chemically

compatible with the YSZ electrolyte than SSC, while both are compatible with the SDC

electrolyte. EIS results indicate that the performances of SSC and SSM electrodes depend

on the electrolyte that they are deposited on. SSC is suitable for the SDC electrolyte, while

SSM is preferred for the YSZ electrolyte. A peak power density as high as 690 mW cm�2 at

600 �C is observed for a thin-film SDC electrolyte with SSC cathode, while a similar cell with

YSZ electrolyte performs poorly. However, SSM performs well on YSZ electrolyte at an

operation temperature of higher than 700 �C, and a fuel cell with SSM cathode and a thin-

film YSZ electrolyte delivers a peak power density of w590 mW cm�2 at 800 �C. The poor

performances of SSM cathode on both YSZ and SDC electrolytes are obtained at

a temperature of lower than 650 �C.

Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

1. Introduction flexibility and low noise and low emission [1,2]. Several

The conversion of chemical energy to electric power via

electrochemical ways is highly attractive because of its

superior efficiency. Solid oxide fuel cells (SOFCs) with oxide

electrolyte and cermet or oxide electrodes operating at

elevated temperature are one type of electrochemical devices

for such a conversion, and they have received particular

attention recently for their additional advantages of fuel

56; fax: þ86 25 8317 2242.1; fax: þ82 31 280 6739.(C. Kwak), [email protected], Hydrogen Energy P

decades of extensive research activities have been conducted

on SOFCs technology, and many progresses have been made

since then, especially toward improving cell power output

[3e7]. Poor reliability and high cost are still the main obstacles

to realize widespread application of SOFCs used in power

generation. People try to reduce the operation temperature of

SOFCs from approximately 1000 �C to an intermediate range of

500e800 �C because of the increased cell lifetime and reduced

du.cn (Z. Shao).ublications, LLC. Published by Elsevier Ltd. All rights reserved.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 74378

materials as well as operation cost [8e10]. Current research on

SOFCs focuses on the cathode since the large cathodic over-

potential results in low power output from the SOFCs at

reduced temperatures [11e15].

An ideal cathode for IT-SOFCs should meet several basic

requirements, such as high electrochemical activity for

oxygen reduction, sufficient electrical conductivity (typically

>100 S cm�1), thermal and chemical compatibility with the

electrolyte (no interfacial reactions and similar thermal

expansion behavior to other cell components) and high

chemical stability under the operation conditions (no reaction

with gases in the environment). La0.8Sr0.2MnO3�d (LSM) is still

themost applied cathodematerial in high-temperature SOFCs

[16,17], which is a perovskite-type crystalline oxide with good

electrical conductivity (>100 S cm�1), favorable thermo-

chemical compatibility with yttria-stabilized zirconia (YSZ)

electrolytes (up to 1200 �C) and sufficiently high activity for

oxygen electrocatalytic reduction at temperatures typically

higher than 850 �C. However, because it has a negligible

oxygen-ionic conductivity, the cathodic overpotential

becomes too large in the intermediate temperature range.

Therefore, a doping or substitution strategy has been widely

applied in developing alternative materials as candidates for

IT-SOFCs cathodes [18e25]. For example, replacement of

manganese in La1�xSrxMnO3�d with cobalt for La1�xSrxCoO3�d

(LSC) perovskite oxides substantially increased the oxygen

surface exchange kinetics, electrical conductivity (up to 2000

S cm�1) and oxygen ionic conductivity [26]. As a result, LSC

have much lower electrode polarization resistance of oxygen

reduction than La1�xSrxMnO3�d, especially at reduced

temperatures [27]. Electrode performance was further

improved by substituting La3þ at the A-site in LSC with Sm3þ

[28], a lanthanide with a smaller cation size than La3þ, with

the development of Sm1�xSrxCoO3�d oxides, which have

substantially improved surface exchange kinetics [29].

Throughout the past decade, Sm0.5Sr0.5CoO3�d (SSC) has been

extensively investigated [30e35], and it is now one of themost

promising cathode materials for IT-SOFCs with doped ceria

electrolytes.

In addition to the intrinsic properties of electrode mate-

rials, the electrode performance is closely related to the elec-

trolyte upon which the electrode is deposited. For example,

Ba0.5Sr0.5Co0.8Fe0.2O3�d (BSCF) is a material with extraordi-

narily high oxygen-ion bulk diffusion rate and oxygen surface

exchange kinetics [36], and it has a very low electrode polar-

ization resistance for oxygen reduction on a doped ceria

electrolyte [37]. However, it performs much worse on a stabi-

lized zirconia electrolyte [38]. The interfacial phase reaction

andmismatch in thermal expansion coefficient (TEC) between

the electrode and electrolyte as well as the electrolyte

composition could all have large effect on electrode perfor-

mance. This suggests that an electrode that performs well on

one electrolytemay not definitely yield a good performance on

another electrolyte. Scandia- or yttria-stabilized zirconia and

gadolinium- or samarium-doped ceria are the electrolytes

most often applied in SOFCs. SSC has been extensively studied

as an electrode on doped ceria electrolytes, while its perfor-

mance on stabilized zirconia electrolytes has not been re-

ported as much. On the other hand, Sm1�xSrxMnO3�d are the

cathodes most often applied to SOFCs with a stabilized

zirconia electrolyte [39,40]; however, there are very few

reports on the application of Sm1�xSrxMnO3�d perovskite

oxides as SOFCs cathodes with doped ceria electrolytes.

In this study, to systematically investigate of the effect of

dopants and electrolytes on electrode performance,

Sm0.5Sr0.5MO3�d (M ¼ Co and Mn) oxides were synthesized,

and their properties and performances as SOFCs cathodes on

samarium-doped ceria (SDC) and YSZ electrolytes were

compared, some interesting results were obtained.

2. Experimental

2.1. Power synthesis and cell fabrication

Both SSC and SSM composite oxides were synthesized using

a combined ethylenediamine tetraacetic acid (EDTA)-citrate

complexing sol-gel process using analytical reagents

Sm(NO3)3$6H2O, Sr(NO3)2, Mn(CH3COO)2$4H2O and Co(N-

O3)2$6H2O as the raw materials. The details for the synthetic

process are available in Ref [41].

A symmetric cell with the configuration of electro-

dejelectrolytejelectrode was applied in the electrochemical

impedance studies. A dense SDC or YSZ pellet with 12 mm in

diameter was prepared by dry-press and subsequently sin-

tering in air at 1400 �C for 5 h. SSC or SSM electrodes was

prepared onto both surfaces of the electrolyte pellet by

spraying deposition with a following calcination at

1000e1050 �C for 2 h under air atmosphere.

Cathodic polarization of the electrodes was evaluated

using three-electrode electrochemical cells. The working

electrode (WE) of SSC or SSM was applied to one side of SDC

electrolyte substrate by spraying deposition with a subse-

quent sintering at 1000e1050 �C in air for 2 h. Porous silver

applied as the counter electrode (CE), which was painted onto

the other side of the SDC pellets as symmetrically as possible

with the WE. Silver paste (DAD-87, Shanghai, China) was

painted into a ring surrounding the CE and was used as the

reference electrode (RE). The gap between the CE and RE was

w4 mm, and the areas of WE, CE and RE were 0.26, 0.26 and

0.30 cm2, respectively.

The performances of the SSC and SSM cathodes were also

evaluated in complete fuel cells. Two types of fuel cells were

adopted: (1) NiO þ SDC (60:40 wt %) anode-supported fuel cell

with SDC electrolyte, and (2) NiO þ YSZ (60:40 wt %) anode-

supported fuel cell with YSZ electrolyte. The NiO þ SDCjSDC

dual-layer pellets were fabricated using a dry-press and co-

sintering technique; for a detailed description of fabrication

processes, please refer to our previous publication [42].

However, the NiO þ YSZjYSZ dual-layer cells were fabricated

using a different method. The NiO þ YSZ anode substrates

were first prepared by tape casting followed by sintering at

1150 �C, and then the YSZ electrolyte layer was prepared via

spraying deposition followed by sintering at 1400 �C in air for

5 h. The SSC or SSM colloidal suspension was spray-deposited

onto the central portion of the electrolyte surface of the as-

fabricated dual-layer cells and then fired at 1000e1050 �C for

2 h to form the complete cells, which had a porous cathode

layer with a w0.48 cm2 geometric surface area. The silver

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 7 4379

paste was used as the current collector. The three-layered

pellets were then used for IeV characterization.

2.2. Electrochemical characterization

The electrode polarization resistance was investigated using

electrochemical impedance spectroscopy (EIS) with a Solar-

tron 1260 frequency response analyzer in combination with

a Solartron 1287 potentiostat. The frequency used for the EIS

measurements ranged from 105e106 to 10�1e10�2 Hz with

a signal amplitude of 10 mV. The samples were tested under

either an open-circuit voltage (OCV) condition or a constant

cathodic current polarization of 600mA cm�2. For the cathodic

polarization measurements, the samples were first polarized

at a constant cathodic current of 600 mA cm�2 for 60 min

before thedatawere acquired at each testing temperature. The

data were collected using the Z-Plot 2.9c software program.

Fuel cell performance was assessed via IeV polarization at

temperatures either between 450 and 650 �C for the cells with

SDC electrolyte or between 600 and 800 �C for the cells with

YSZ electrolyte. During the test, H2 fuel was fed into the anode

side at a flow rate of 80mlmin�1 [STP], and ambient air served

as the oxidant gas. The measurements were performed using

a Keithely 2420 digital source meter, interfaced with

a computer for data acquisition. A four-terminal configuration

was employed for the measurement.

2.3. Additional characterizations

The crystal structures of the SSC and SSM powders were

determined using X-ray diffraction (XRD; Model D8 Advance,

Bruker, Germany) with Cu Ka radiation at 40 kV and 40 mA.

The experimental diffraction patterns were collected at room

temperature through step scans in the 2q range of 10e90� with

a 0.05� angle step. Rietveld refinements on the XRD patterns

were performed using DIFFRACplus Topas 4 software. The TEC

was measured in air using a Netsch DIL 402C/3/G dilatometer

from room temperature to 1000 �C at a heating rate of

5 �C min�1.

For oxygen temperature programmed desorption (O2-TPD)

experiment, an approximately 0.15 g sample at a 40e60 mesh

size was loaded into a U-type quartz tube. The tube was then

Fig. 1 e The Rietveld refinement

placed in a tubular furnace equipped with a temperature

controller. Pure argon was introduced to the reactor as the

carrier gas at a flow rate of 15mlmin�1 [STP]. The temperature

was increased from room temperature to 950 �C at a ramp rate

of 10 �C min�1. The effluent gas from the reactor was con-

nected to a Hiden QIC-20 mass spectroscope (MS) to monitor

the oxygen concentration modulation in-situ. Occasionally,

the temperature was held at 950 �C for sufficient time to

enable the completion of the oxygen-desorption process.

At room temperature, iodometric titration was used to

measure the oxygen non-stoichiometries of SSC and SSM

samples, which were either freshly prepared or used after the

O2-TPD experiment. A powder sample at approximately 0.1 g

was dissolved in a 6.0 mol L�1 HCl solution under an inert gas

atmosphere (to prevent air oxidation of I� ions) before it was

titrated with a standard thiosulfate (S2O32-) solution. Several

drops of a starch solution were added to discern the reaction

endpoint, which was indicated by an abrupt color change in

an initially transparent solution. The details for this

measurement process are available in Ref. [41].

The electrical conductivity was measured within the

temperature range of 300e900 �C at 10 �C intervals using the

four-terminal DC technique with bar-shaped samples and

2 mm � 5 mm � 12 mm dimensions. The silver paste was

painted on the rectangular cross-section edges and two

circumferential surfaces of the bar to form the current and

voltage electrodes. The current and voltage were applied/

detected using a Keithely 2420 source meter. At each

temperature step, the samples were allowed sufficient time to

stabilize to ensure that the electrical conductivity reached

a steady state.

3. Results and discussion

3.1. Basic properties

Fig. 1shows the room temperature XRD patterns for the SSC

and SSM powders after the sintering at 1000 and 1100 �C,respectively, in air for 5 h. Both samples were well crystallized

in a perovskite-type structure. Based on the Rietveld refine-

ment, SSC had an orthorhombic symmetry with a Pnma space

plots for (a) SSC and (b) SSM.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 74380

group and the calculated lattice parameters were a ¼ 5.366 A,

b ¼ 5.398 A and c ¼ 7.585 A, which agree well with the litera-

ture [43]. SSM had a similar tetragonal lattice structure

compared with La0.5Sr0.5MnO3 with the I4/mcm space group.

The SSM lattice parameters derived from the XRD patterns

were a ¼ b ¼ 5.448 A, c ¼ 7.658 A, which are slightly smaller

than the La0.5Sr0.5MnO3 parameters, which are a¼ b¼ 5.481 A,

c ¼ 7.721 A [44]. This difference can be explained by the

smaller cation size of Sm3þ than La3þ.Fig. 2 shows the thermal expansion curves for SSC and SSM

between 200 and 1000 �C. The thermal expansion data were

obtained by heating the sintered, bar-shaped SSC and SSM

pellets in air at a rate of 5 �C min�1. The TEC is defined as the

slope of the dilatometric curve. The average TECs for SSC and

SSM were 24.0 � 10�6 and 11.8 � 10�6 K�1, respectively. These

values indicate that the B-site cation has a significant effect on

the thermal expansion behavior of the Sm0.5Sr0.5MO3�d elec-

trodes. The apparent thermal expansions of SSC and SSM

were generated in part from the “real” thermal expansion,

which is associated with the bonding energy between ions in

the lattice and the B-site cation spin state transition. Chemical

expansion also contributes to the apparent thermal expan-

sion, which is related to the valence state change of the B-site

cation from the thermally induced reduction of the B-site

cation to a lower valence state. The thermal reducibilities of

SSC and SSM were then examined by O2-TPD analysis. Both

samples were programmatically heated in a flowing argon

atmosphere from room temperature to 950 �C at a rate of

10 �C min�1. The thermal reduction of the B-site cation in the

perovskite oxides accompanies the release of molecular

oxygen into the surrounding atmosphere, which was carried

out by the flow of argon and detected by the on-line MS. As

shown in Fig. 3, there was almost no oxygen released

throughout the range of temperatures investigated

(100e950 �C) for the SSM sample. For SSC, the oxygen was

released from the oxide lattice beginning at 250 �C, and peaked

at 380, 820 and 940 �C. The oxygen desorption was not

completed at 950 �C, which is indicated by the failure of the

oxygen profile to approach the baseline. Clearly, it was much

easier to thermally reduce the cobalt than the manganese.

Based on the O2-TPD results, it is likely that the large lattice

Fig. 2 e The thermal expansion curves for the SSC and SSM

samples.

volume expansion, which is associated with thermal reduc-

tion, primarily contributed to the large TEC for SSC. The

oxygen non-stoichiometries of SSC/SSM before and after the

O2-TPD were 0.052/-0.029 and 0.267/-0.020, respectively,

determined using the iodometric titration method. It was

interesting that SSM had an excess of oxygen even after the

O2-TPD process. Similar oxygen over stoichiometry has also

been reported for La0.8Sr0.2MnO3 [45]. The differences in

oxygen stoichiometry before and after the O2-TPD process

were 0.009 and 0.215 for SSM and SSC, respectively. It is well

known that oxygen vacancies are the charge carriers for

oxygen ions; the absence of oxygen vacancies in SSM implies

that it has a very low oxygen-ionic conductivity, while the

large oxygen non-stoichiometry of SSC after the O2-TPD

suggests that it has a high oxygen-ionic conductivity. More-

over, it is well known that SSC is a typical mixed-oxygen ionic

and electronic conductor.

The electrical conductivities of SSC and SSM in air were

measured using a four-probe DCmethod and the temperature

dependent curves were shown in Fig. 4. The electrical

conductivity for SSC reached 1800 S cm�1 at 300 �C, and it

decreased progressively with an increase in temperature,

which suggests a metallic-conduction behavior. SSM behaved

as a semi-conductor because the electrical conductivity

reached approximately 140 S cm�1 at 300 �C, and it increased

slightly with temperature to reach 180 S cm�1 at 900 �C. Theelectrical conductivity of SSM is comparable to a typical LSM

cathode [46]. Although SSM has much lower electrical

conductivity than SSC, it still meets the >100 S cm�1 [47]

requirement to be a potential electrode of SOFCs.

3.2. Chemical compatibility with electrolytes

The chemical compatibilities of the SSC/SSM oxides as elec-

trodes with SDC/YSZ electrolytes were investigated using

a solid-phase reaction in a powder form. SSC (SSM) wasmixed

with SDC (YSZ) at a 50:50 weight ratio and then calcined at

various temperatures in air for 5 h. After the calcination, the

samples were characterized by room-temperature X-ray

diffraction. Fig. 5 shows the XRD patterns of the calcined

SSCþ SDC, SSCþ YSZ, SSMþ SDC and SSMþ YSZmixtures. It

Fig. 3 e The oxygen temperature-programmed desorption

(O2-TPD) profiles for SSC and SSM.

Fig. 4 e The temperature dependence of the electrical

conductivities for SSC and SSM between 300 and 900 �C.

Fig. 5 e The X-ray diffraction patterns for (a) SSC D SDC, (b) SSC

between 900 and 1100 �C for 5 h.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 7 4381

is well known that cobalt-based perovskite oxides are highly

reactive. For the SSC þ SDC mixture, the diffraction peaks

were easily indexed from a physical mixture of SSC and SDC

phases even after the calcination at 1100 �C. This suggests thatSSC and SDC did not undergo a serious phase reaction, which

is consistent with the literature [48]. However, the relative

intensity of the (110) diffraction plane for SSC phase and (111)

diffraction plane for SDC phase decreasedwith the calcination

temperature, which suggests that certain reaction might

appear between SSC and SDC at elevated temperatures.

Previously, we demonstrated that BSCF and SDC undergo

a cation exchange at high temperatures [49]. Similarly, SSC

and SDC may have also undergone a cation exchange reac-

tion. A potential cation exchange reaction between SSC and

SDC was supported by the successful synthesis of a single-

phase of Sr or Co slightly doped SDC with the nominal

composition of (Sm0.2Ce0.8)0.95M0.05O1.9 (M ¼ Sr and Co), as

shown in Fig. 6. It was interesting that Co doping into SDC

electrolyte increased the electrical conductivity of the SDC

electrolyte (Fig. 7). Moreover, an increase in electrical

conductivity upon cobalt doping has also been reported for

D YSZ, (c) SSM D SDC and (d) SSM D YSZ mixtures calcined

Fig. 6 e The X-ray diffraction patterns for (A) SDC, (B)

(Sm0.2Ce0.8)0.95Sr0.05O1.9, (C) (Sm0.2Ce0.8)0.95Co0.05O1.9 and (D)

(Sm0.2Ce0.8)0.95Mn0.05O1.9.

Fig. 7 e The temperature dependence of the electrical

conductivities for (A) SDC, (B) (Sm0.2Ce0.8)0.95Sr0.05O1.9, (C)

(Sm0.2Ce0.8)0.95Co0.05O1.9 and (D) (Sm0.2Ce0.8)0.95Mn0.05O1.9

samples.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 74382

SDC [50]. For the SSC þ YSZ mixture, after calcination at

900 �C, weak diffraction peaks also appeared, which cannot be

assigned to either the SSC or YSZ phase, which suggests that

SSC and YSZ underwent a solid phase reaction. This is

consistent with the result in the literature [43], in which SSC

reacts with YSZ at high temperatures (>900 �C) and there is

a poor chemical compatibility between SSC and YSZ under

operating conditions. With an increase in the calcination

temperature to 1100 �C, peak intensity of the SSC primary

phase became very weak. This suggests that SSC and YSZ

underwent a strong solid-phase reaction at 1100 �C. As shown

in Fig. 5c and d, a negligible phase reaction was observed

between SSM and SDC as well as between SSM and YSZ, as

indicated by the absence of additional diffraction peaks in the

XRD patterns for both the SSMþ SDC and SSMþYSZmixtures

calcined as high as 1100 �C. This suggests that SSM is much

more compatible with electrolytes than SSC, and it is partic-

ularly compatible with the YSZ electrolyte.

3.3. Electrode activity for oxygen reduction

The electrode performances of SSC and SSM on the SDC and

YSZ electrolytes were first tested using EIS in a symmetric cell

configuration. TheEISof all the symmetric cellswas conducted

under OCV conditions. As shown in Fig. 8a, SSC had relatively

low polarization resistance on SDC electrolyte with the area

specific resistances (ASRs) of 0.04, 0.08, 0.16, 0.40and1.24U cm2

at 800, 750, 700, 650 and 600 �C, respectively. The results agree

with the literature [51]. As for the SSM electrode, it had a larger

electrode polarization resistance than SSC, as shown in Fig. 8b;

theASRs for the SSMelectrode onSDCelectrolyte reached18.2,

47.1, 124.2, 281.9and602.2U cm2at 800, 750, 700, 650and600 �C,respectively. However, the two electrodes had similar activa-

tion energies (Ea) of 132.7 (SSC) and 136.2 kJ mol�1 (SSM) from

the temperature dependence of ASRs. Fig. 8c and d shows the

ASRs for the SSC electrode on YSZ electrolyte at various

temperatures, where the SSC was fired onto YSZ at 1000 and

1100 �C, respectively. The ASRs reached 29.5, 90.2, 279.4, 805.4

and 2323.2 U cm2 at 800, 750, 700, 650 and 600 �C, respectively,for the SSC electrode fired at 1000 �C. The ASRs were larger

when the firing temperature was increased to 1100 �C, and the

corresponding values were 153.2, 231.5, 422.2, 998.5 and

2756.1 U cm2, respectively, as shown in Fig. 8d. In connection

with the results of powder reaction in Fig. 5, the lower perfor-

mance of SSC on the YSZ electrolyte compared with the SDC

electrolyte is primarily attributed to the interfacial reaction

between SSC and YSZ. With an increase in firing temperature,

the interfacial reaction became stronger, and as a result, the

SSC electrode had a lower performancewhen it was fired onto

theYSZelectrolyteat 1100 �Ccomparedwith theelectrodefired

at 1000 �C. For the SSM electrode on the YSZ electrolyte, as

shown in Fig. 8e, the ASRs were 23.1, 65.4, 192.9, 481.4 and

1449.6 U cm2 at 800, 750, 700, 650 and 600 �C, respectively.Although the corresponding values were all larger than those

for the SDC electrolyte, the difference in performance of the

electrodeon theYSZandSDCelectrolyteswassmaller than the

SSC electrode. Because there was no obvious phase reaction

between SSM and SDC as well as SSM and YSZ, the different

ASRs for SSM on the SDC and YSZ electrolytes were directly

related to the different electrical conductivities of the two

electrolytes. It is well known that the active sites for an elec-

trode are primarily located at the electrode-electrolyte-gas

phase triple phase boundary (TPB) region. A change in the

electrical conductivity of an electrolytewould affect the charge

transfer from the electrode to the electrolyte, and as a result,

the ASRs of the electrode were altered.

For an electrode with a negligible oxygen ionic conduc-

tivity, the electrode reaction is limited strictly to the TPB.

However, for an electrode with a mixed-oxygen ionic and

electronic conductivity, the active sites may extend to the

entire exposed electrode surface, and as a result, the electrode

activity is significantly improved. It is well known that oxygen

vacancies are the charge carriers for oxygen ions; SSC was a

mixed conductor for high levels of oxygen non-stoichiometry,

Fig. 8 e The Nyquist impedance curves for (a) SSC fired onto SDC at 1000 �C, (b) SSM fired onto SDC at 1050 �C, (c) SSC fired

onto YSZ at 1000 �C, (d) SSC fired onto YSZ at 1100 �C and (e) SSM fired onto YSZ at 1050 �C for 2 h at 600e800 �C in air. The

insets are the magnified Nyquist plots from EIS for the above samples at 700e800 �C in air.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 7 4383

Table 1 e The polarization resistance for the SSC and SSM electrodes on an SDC electrolyte before and after the cathodicpolarization at various temperatures.

Temperature (�C) 700 650 600 550 500

Electrode materials SSC SSM SSC SSM SSC SSM SSC SSM SSC SSM

Polarization resistance (U cm2) Before CP* 0.74 126.2 1.91 343.7 5.02 911.7 15.05 2785 54.60 9989

After CP* 0.36 2.65 0.83 3.34 1.69 4.91 3.67 8.46 6.74 17.5

Decreasing amplitude (%) 51.02 97.90 56.71 99.03 66.45 99.46 75.60 99.70 87.66 99.83

CP*: Cathodic polarization.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 74384

while SSM was a purely electronic conductor because of its

negligible oxygen vacancy. The mixed conductivity accounts

for the better performance of SSC compared with SSM on the

SDC electrolyte.

It was reported that the typical LSM electrode performed

better under cathodic polarization because the polarization

promotes an in-situ electrochemical reduction of the Mn4þ in

LSM to Mn3þ with the introduction of oxygen vacancies into

LSM [52]. As a result, LSM changes from a purely electronic

conductor to a mixed-oxygen ionic and electronic conductor.

Oxygen ionic conduction within LSM bulk improves the elec-

trode activity for oxygen reduction because of the increased

number of active sites. Table 1 shows the polarization resis-

tances for the SSC and SSM electrodes on SDC electrolyte

before and after cathodic polarization under a current density

600 mA cm�2 at various temperatures for 60 min. For the SSM

electrode, a sharp decrease in the ASRs was observed after the

Fig. 9 e The IeV and IeP curves for the complete cell based on

Ni D YSZjYSZjSSM and (d) Ni D SDCjSDCjSSM.

cathodic polarization, which suggests successful creation of

oxygen-ionic conduction within the SSM bulk. Before the

cathodic polarization, the ASRs reached 126.2 and 9989 U cm2

at 700 and 500 �C, respectively. However, the ASRs were

reduced significantly to only 2.65 and 17.5 U cm2, respectively,

at 700 and 500 �C after the polarization, which are onlyw2.1%

and 0.17% of the values before the polarization. The

substantially different ASRs values can be interpreted by the

generating oxygen vacancies after the cathodic polarization,

which can account for the enhanced performance of SSM in

the following single-cell test. For the SSC electrode, the

improvement in electrode performance was also observed

after the polarization. However, the degree of improvement

was much smaller than for SSM; for example, the ASRs before

and after the polarization were 0.74 and 0.36 U cm2, respec-

tively, at 700 �C. This can be explained by the mixed conduc-

tivity of SSC even before the polarization. Themore significant

(a) Ni D SDCjSDCjSSC, (b) Ni D YSZjYSZjSSC, (c)

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 7 4385

improvement in electrode performance at the lower temper-

ature is consistent with a decrease in oxygen-ionic conduc-

tivity of SSC as the temperature decreases.

3.4. Single cell performance

The performances of the SSC and SSM electrodes were further

evaluated in single cells with an anode-supported thin-film

electrolyte configuration. Fig. 9 shows the IeV and IeP polar-

ization curves for the cells with SSC or SSM cathode and SDC

or YSZ electrolyte. All of the cells had a similar electrolyte

thickness of approximately 20 mm. The cells were tested by

applying hydrogen as the fuel and ambient air as the cathode

atmosphere. As shown in Fig. 9a, peak power densities of

approximately 930 and 690 mW cm�2 were reached at 650 and

600 �C, respectively, for the Ni þ SDCjSDCjSSC cell. These

results are comparable to those reported in the literature [53].

However, a poor cell performance was observed for the

Ni þ YSZjYSZjSSC cell, for which a peak power density of

95 mW cm�2 was achieved at 800 �C, and it was only reduced

to 10 mW cm�2 at 600 �C. Consistent with the symmetric cell

results, the interfacial reaction between SSC and YSZ sharply

increased the cathodic polarization resistance. However, the

interfacial reaction also formed the insulating phase, which is

similar to SrZrO3 and might substantially increase the ohmic

resistance of the cell. As a result, the cell had a very low power

output. For the Ni þ YSZjYSZjSSM cell, a much higher cell

performance compared with the similar SSC cathode-

containing cell was observed. As shown in Fig. 9c, the cell

delivered peak power densities of 590, 290, and 120 mW cm�2

at 800, 750 and 700 �C, respectively. This suggests that SSM

may be applied as a SOFCs cathode with the YSZ electrolyte at

an operation temperature of higher than 700 �C. The prom-

ising cell performance can be attributed to the negligible

interfacial reaction between SSM electrode and YSZ electro-

lyte as well as the polarization-induced creation of oxygen-

ionic conduction within the SSM electrode bulk. As for the

Ni þ SDCjSDCjSSM cell, the cell performance was poor with

the peak power densities of 160, 105 and 60 mW cm�2 at 650,

600 and 550 �C, respectively, which are much lower than the

similar SSC cathode-containing cell. Because the interfacial

reaction of both electrodes with the SDC electrolyte is weak,

the lower performance of the Ni þ SDCjSDCjSSM cell

compared with the Ni þ SDCjSDCjSSC cell can be attributed to

the lower oxygen reduction activity of SSM compared with

SSC at reduced temperatures.

4. Conclusion

The B-site cation dopant in Sm0.5Sr0.5MO3�d (M ¼ Co and Mn)

oxides had a strong influence on the phase structure, thermal

expansion behavior, oxygen mobility, oxygen non-

stoichiometry, electrical conductivity and oxygen-reducing

electrochemical activity. It was much easier to thermally

reduce cobalt than manganese, and SSC has a much higher

oxygen vacancy concentration and TEC than SSM at elevated

temperatures. Although both SSC and SSM oxides had high

electrical conductivity values that are suitable for fuel cell

electrode applications (>100 S cm�1), the electrical

conductivity of SSC was more than 7 times that of SSM at

600 �C. SSC and SDC did not undergo a strong solid-state

reaction at the elevated temperature (900 �C); however, the

solid-state reactionwas strong betweenSSC andYSZ. The SSM

electrode underwent a negligible reaction with both the SDC

andYSZ electrolytes, even at 1100 �C. This suggests that SSM is

more thermochemically compatible than SSC with the YSZ

electrolyte, while SSC is compatible with the SDC electrolyte.

Cathodic polarization led to a significant improvement in SSM

electrode performance, while the improvement was only

modest for the SSC electrode because of the favorable oxygen-

ionic conductivity of SSC prior to the polarization. Because of

the strong interfacial reaction between SSC and YSZ, the SSC

electrode is only suitable for the SDC electrolyte, while SSM

may be applied to both the SDC and YSZ electrolytes. The SSM

electrode performed well on the YSZ electrolyte at an opera-

tion temperature of higher than 700 �C, which suggests SSM

may be applied as a SOFCs cathode with the YSZ electrolyte at

high temperatures. The poor performances of SSM cathode on

both the YSZ and SDC electrolytes were observed at

a temperature of lower than 650 �C due to the lower oxygen

reduction activity of SSM at reduced temperatures.

Acknowledgments

This work was supported by the National Science Foundation

for Distinguished Young Scholars of China under contract No.

51025209, the program for New Century Excellent Talents

(2008) and the Fok Ying Tung Education Foundation under

contract No. 111073.

r e f e r e n c e s

[1] Achenbach E, Riensche E. Methane/steam reforming kineticsfor solid oxide fuel cells. J Power Sources 1994;52:283e8.

[2] Shao ZP, Haile SM, Ahn J, Ronney PD, Zhan ZL, Barnett SA. Athermally self-sustained micro solid-oxide fuel-cell stackwith high power density. Nature 2005;435:795e8.

[3] Souza SD, Visco SJ, Jonghe LCD. Thin-film solid oxide fuel cellwith high performance at low-temperature. Solid StateIonics 1997;98:57e61.

[4] Tsai T, Barnett SA. Increased solid-oxide fuel cell powerdensity using interfacial ceria layers. Solid State Ionics 1997;98:191e6.

[5] Huang KQ, Wan JH, Goodenough JB. Increasing powerdensity of LSGM-based solid oxide fuel cells using new anodematerials. J Electrochem Soc 2001;148:A788e94.

[6] Shao ZP, Haile SM. A high-performance cathode for the nextgeneration of solid-oxide fuel cells. Nature 2004;431:170e3.

[7] Ramakrishna PA, Yang S, Sohn CH. Innovative design toimprove the power density of a solid oxide fuel cell. J PowerSources 2006;158:378e84.

[8] Kim YN, Kim JH, Manthiram A. Characterization of (Y1�xCax)BaCo4�yZnyO7 as cathodes for intermediate temperature solidoxide fuel cells. Int J Hydrogen Energy 2011;36:15295e303.

[9] Yamaguchi T, Shimizu S, Suzuki T, Fujishiro Y, Awano M.Fabrication and characterization of high performancecathode supported small-scale SOFC for intermediatetemperature operation. Electrochem Commun 2008;10:1381e3.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 74386

[10] Brett DJL, Atkinson A, Brandon NP, Skinner SJ. Intermediatetemperature solid oxide fuel cells. Chem Soc Rev 2008;37:1568e78.

[11] Adler SB. Factors governing oxygen reduction in solid oxidefuel cell cathodes. Chem Rev 2004;104:4791e844.

[12] Jiang SP, Leng YJ, Chan SH, Khor KA. Development of (La, Sr)MnO3-based cathodes for intermediate temperature solidoxide fuel cells. Electrochem Solid-State Lett 2003;6:A67e70.

[13] Sasaki K, Wurth JP, Gschwend R, Godickemeier M,Gauckler LJ. Microstructure-property relations of solid oxidefuel cell cathodes and current collectors. J Electrochem Soc1996;143:530e43.

[14] Leng YJ, Chan SH, Khor KA, Jiang SP. Development of LSM/YSZ composite cathode for anode-supported solid oxide fuelcells. J Appl Electrochem 2004;34:409e15.

[15] Liu QL, Khor KA, Chan SH. High-performance low-temperature solid oxide fuel cell with novel BSCF cathode. JPower Sources 2006;161:123e8.

[16] Koep E, Mebane DS, Das R, Compson C, Liu ML. Characteristicthickness for a dense La0.8Sr0.2MnO3 electrode. ElectrochemSolid-State Lett 2005;8:A592e5.

[17] Jiang SP. Development of lanthanum strontium manganiteperovskite cathode materials of solid oxide fuel cells:a review. J Mater Sci 2008;43:6799e833.

[18] Ishihara T, Honda M, Shibayama T, Minami H, Nishiguchi H,Takita V. Intermediate temperature solid oxide fuel cellsusing a new LaGaO3 based oxide ion conductor: I. DopedSmCoO3 as a new cathode material. J Electrochem Soc 1998;145:3177e83.

[19] Zheng Y, Ran R, Shao ZP. Activation and deactivationkinetics of oxygen reduction over a La0.8Sr0.2Sc0.1Mn0.9O3

cathode. J Phys Chem C 2008;112:18690e700.[20] Wei B, Lv Z, Jia DC, Huang XQ, Zhang YH, Su WH. Thermal

expansion and electrochemical properties of Ni-dopedGdBaCo2O5þd double-perovskite type oxides. Int J HydrogenEnergy 2010;35:3775e82.

[21] Zhou W, Shao ZP, Ran R, Jin WQ, Xu NP. A novel efficientoxide electrode for electrocatalytic oxygen reduction at400e600 �C. Chem Commun 2008;44:5791e3.

[22] Niu YJ, Sunarso J, Zhou W, Liang FL, Ge L, Zhu ZH, et al.Evaluation and optimization of Bi1�xSrxFeO3�d perovskites ascathodes of solid oxide fuel cells. Int J Hydrogen Energy 2011;36:3179e86.

[23] Hrovat M, Katsarakis N, Reichmann K, Bemik S, Kuscer D,Hole J. Characterisation of LaNi1�xCoxO3 as a possible SOFCcathode material. Solid State Ionics 1996;83:99e105.

[24] Wei B, Lv Z, Huang XQ, Liu ML, Li N, Su WH. Synthesis,electrical and electrochemical properties ofBa0.5Sr0.5Zn0.2Fe0.8O3�d perovskite oxide for IT-SOFC cathode.J Power Sources 2008;176:1e8.

[25] Bidrawn F, Lee S, Vohs JM, Gorte RJ. The effect of Ca, Sr, andBa doping on the ionic conductivity and cathodeperformance of LaFeO3. J Electrochem Soc 2008;155:B660e5.

[26] Yang YL, Chen CL, Chen SY, Chu CW, Jacobson AJ.Impedance studies of oxygen exchange on dense thin filmelectrodes of La0.5Sr0.5CoO3�d. J Electrochem Soc 2000;147:4001e7.

[27] Takeda Y, Kanno R, Noda M, Tomida Y, Yamamoto O.Cathodic polarization phenomena of perovskite oxideelectrodes with stabilized zirconia. J Electrochem Soc 1987;134:2656e61.

[28] Tsipis EV, Kharton VV. Electrode materials and reactionmechanisms in solid oxide fuel cells: a brief review II.Electrochemical behavior vs. materials science aspects. JSolid State Electrochem 2008;12:1367e91.

[29] Fukunaga H, Koyama M, Takahashi N, Wen C, Yamada K.Reaction model of dense Sm0.5Sr0.5CoO3 as SOFC cathode.Solid State Ionics 2000;132:279e85.

[30] Xia CR, Liu ML. Low-temperature SOFCs based onGd0.1Ce0.9O1.95 fabricated by dry pressing. Solid State Ionics2001;144:249e55.

[31] Yang S, He TM, He Q. Sm0.5Sr0.5CoO3 cathode material fromglycine-nitrate process: formation, characterization, andapplication in LaGaO3-based solid oxide fuel cells. J AlloysCompd 2008;450:400e4.

[32] Wang SZ, Zou YM. High performance Sm0.5Sr0.5CoO3 -La0.8Sr0.2Ga0.8Mg0.15Co0.05O3 composite cathodes. ElectrochemCommun 2006;8:927e31.

[33] Yang L, Zuo CD, Wang SZ, Cheng Z, Liu ML. A novelcomposite cathode for low-temperature SOFCs based onoxide proton conductors. Adv Mater 2008;20:3280e3.

[34] Baek SW, Bae J, Yoo YS. Cathode reaction mechanism ofporous-structured Sm0.5Sr0.5CoO3�d and Sm0.5Sr0.5CoO3�d/Sm0.2Ce0.8O1.9 for solid oxide fuel cells. J Power Sources 2009;193:431e40.

[35] Guo YM, Chen DJ, Shi HG, Ran R, Shao ZP. Effect of Sm3þ

content on the properties and electrochemical performanceof SmxSr1�xCoO3�d (0.2 � x � 0.8) as an oxygen reductionelectrodes on doped ceria electrolytes. Electrochim Acta2011;56:2870e6.

[36] Chen DJ, Shao ZP. Surface exchange and bulk diffusionproperties of Ba0.5Sr0.5Co0.8Fe0.2O3�d mixed conductor. Int JHydrogen Energy 2011;36:6948e56.

[37] Zhou W, Ran R, Shao ZP. Progress in understanding anddevelopment of Ba0.5Sr0.5Co0.8Fe0.2O3�d -based cathodes forintermediate-temperature solid-oxide fuel cells: a review. JPower Sources 2009;192:231e46.

[38] Duan ZS, Yang M, Yan AY, Hou ZF, Dong YL, Chong Y, et al.Ba0.5Sr0.5Co0.8Fe0.2O3�d as a cathode for IT-SOFCs with a GDCinterlayer. J Power Sources 2006;160:57e64.

[39] Wen TL, Tu HY, Xu ZH, Yamamoto O. A study of (Pr, Nd,Sm)1�xSrxMnO3 cathode materials for solid oxide fuel cell.Solid State Ionics 1999;121:25e30.

[40] Sakaki Y, Takeda Y, Kato A, Imanishi N, Yamamoto O,Hattori M, et al. Ln1�xSrxMnO3 (Ln ¼ Pr, Nd, Sm and Gd) asthe cathode material for solid oxide fuel cells. Solid StateIonics 1999;118:187e94.

[41] Guo YM, Shi HG, Ran R, Shao ZP. Performance ofSrSc0.2Co0.8O3�d þ Sm0.5Sr0.5CoO3�d mixed-conductingcomposite electrodes for oxygen reduction at intermediatetemperatures. Int J Hydrogen Energy 2009;34:9496e504.

[42] An BM, Zhou W, Guo YM, Ran R, Shao ZP. A compositeoxygen-reduction electrode composed of SrSc0.2Co0.8O3�d

perovskite and Sm0.2Ce0.8O1.9 for an intermediate-temperature solid-oxide fuel cell. Int J Hydrogen Energy 2010;35:5601e10.

[43] Tu HY, Takeda Y, Imanishi N, Yamamoto O. Ln1�xSrxCoO3

(Ln ¼ Sm, Dy) for the electrode of solid oxide fuel cells. SolidState Ionics 1997;100:283e8.

[44] Teng F, Han W, Liang SH, Gaugeu B, Zong RL, Zhu YF.Catalytic behavior of hydrothermally synthesizedLa0.5Sr0.5MnO3 single-crystal cubes in the oxidation of COand CH4. J Catal 2007;250:1e11.

[45] Mizusaki J, Mori N, Takai H, Yonemura Y, Minamiue H,Tagawa H, et al. Oxygen nonstoichiometry and defectequilibrium in the perovskite-type oxides La1�xSrxMnO3�d.Solid State Ionics 2000;129:163e77.

[46] Kuo JH, Anderson HU, Sparlin DM. Oxidation-reductionbehavior of undoped and Sr-doped LaMnO3: defect structure,electrical conductivity, and thermoelectric power. JSolidState Chem 1990;87:55e63.

[47] Carter S, Selcuk A, Chater RJ, Kajda J, Kilner JA, Steele BCH.Oxygen transport in selected nonstoichiometric perovskite-structure oxides. Solid State Ionics 1992;53-56:597e605.

[48] Zhang XG, Robertson M, Yick S, Deces-Petit C, Styles E, QuW,et al. Sm0.5Sr0.5CoO3 þ Sm0.2Ce0.8O1.9 composite cathode for

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 7 ( 2 0 1 2 ) 4 3 7 7e4 3 8 7 4387

cermet supported thin Sm0.2Ce0.8O1.9 electrolyte SOFCoperating below 600 �C. J Power Sources 2006;160:1211e6.

[49] Wang K, Ran R, Zhou W, Gu HX, Shao ZP, Ahn J. Propertiesand performance of Ba0.5Sr0.5Co0.8Fe0.2O3�d þ Sm0.2Ce0.8O1.9

composite cathode. J Power Sources 2008;179:60e8.[50] Ayawanna J, Wattanasiriwech D, Wattanasiriwech S,

Aungkavattana P. Effects of cobalt metal addition onsintering and ionic conductivity of Sm(Y)-doped ceria solidelectrolyte for SOFC. Solid State Ionics 2009;180:1388e94.

[51] Xia CR, Rauch W, Chen FL, Liu ML. Sm0.5Sr0.5CoO3 cathodesfor low-temperature SOFCs. Solid State Ionics 2002;149:11e9.

[52] Lee HY, Cho WS, Oh SM, Wiemhofer HD, Gopel W.Active reaction sites for oxygen reduction inLa0.9Sr0.1MnO3/YSZ electrodes. J Electrochem Soc 1995;142:2659e64.

[53] Zhao F, Wang ZY, Liu MF, Zhang L, Xia CR, Chen FL. Novelnano-network cathodes for solid oxide fuel cells. J PowerSources 2008;185:13e8.