ww.sciencedirect.com

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 4

Available online at w

ScienceDirect

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

High-performance Y0.9In0.1BaCo3(Zn,Fe)O7 þ d

swedenborgite-type oxide cathodes for reducedtemperature solid oxide fuel cells

Matthew West, Christina Ortiz, Arumugam Manthiram*

McKetta Department of Chemical Engineering, University of Texas at Austin, Austin, TX 78712, USA

a r t i c l e i n f o

Article history:

Received 17 September 2014

Received in revised form

29 October 2014

Accepted 3 November 2014

Available online 3 December 2014

Keywords:

Solid oxide fuel cells

Cathodes

Phase stability

Polarization resistance

* Corresponding author. Tel.: þ1 512 471 179E-mail address: rmanth@mail.utexas.edu

http://dx.doi.org/10.1016/j.ijhydene.2014.11.00360-3199/Copyright © 2014, Hydrogen Energ

a b s t r a c t

With the goal of improving the electrochemical performance while maintaining good

phase stability, the Fe-substituted swedenborgite-type Y0.9In0.1BaCo3Zn1 � xFexO7 þ d series of

oxides have been investigated as cathode materials for solid oxide fuel cells (SOFCs). All the

samples with 0 � x � 0.8 were obtained as single-phase oxides with the P31c space group.

Both the x ¼ 0 and 0.2 samples were stable after 120 h exposure to 600, 700, and 800 �C,

while the x ¼ 0.6 and 0.8 samples showed instability at 700 and 800 �C. The x ¼ 0.4 sample

was stable at 600 and 800 �C, but generated a BaCoO3 impurity phase at 700 �C. The

reversible oxygen absorption tendency of these materials was found to increase with

increasing Fe content. Increasing the Fe content in the range of 0 � x � 0.4 was found to

reduce the polarization resistance in the Y0.9In0.1BaCo3Zn1 � xFexO7 þ d þ Gd0.2Ce0.8O1.9

composite cathodes, and the x ¼ 0.4 sample was found to have performance superior to

that of the well-studied Ba0.5Sr0.5Co0.8Fe0.2O3 � d (BSCF) cathode in the range of 400e700 �C,

making it an attractive cathode candidate for intermediate temperature SOFCs.

Copyright © 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

Introduction

Solid oxide fuel cells (SOFCs) are power generation devices

that offer high energy conversion efficiency and fuel flexibility

without the need for precious metal catalysts, unlike the low-

temperature fuel cells [1e3]. However, SOFCs traditionally

operate at temperatures in the range of 800e1000 �C, which

presents unique materials challenges and hamper the

commercialization efforts. The high operating temperatures

require the use of expensive, specialized materials that can

withstand the high-temperature environment, which

dramatically increases the cost of stack assembly. These

1; fax: þ1 512 471 7681.(A. Manthiram).27y Publications, LLC. Publ

difficulties have generated immense interest in reducing the

SOFC operating temperatures, but the traditional SOFC cath-

ode materials offer poor performance at temperatures below

800 �C. Therefore, one of the main areas of SOFC investigation

is to develop cathode materials that can operate in the

intermediate-temperature (IT) region (600e800 �C) or in the

low-temperature region (T < 600 �C) [1e14].Recently, swedenborgite-type RBa(Co,M)4O7 þ d (R ¼ Y, In,

and Ca; M ¼ Co, Zn, Fe, Ga, and Al) oxides have been shown to

have favorable thermal expansion and electrochemical per-

formance properties in these reduced temperature ranges

[8e16]. However, these materials have traditionally been

observed to decompose after long-term (e.g., 120 h) exposure

ished by Elsevier Ltd. All rights reserved.

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 4 1187

to SOFC operating temperatures in the range of 600e800 �C,but recent work by our group has shown that these materials

can be stabilized through selective substitutions in both the R

and M sites [8e16]. One of the most frequently used sub-

stitutions is the replacement of Co by Zn as the high con-

centration of Co2þ/3þ in the tetrahedral sites is thought to be

the main source of instability in these materials [8]. However,

Zn2þ with a stable 3d10 configuration diminishes the elec-

tronic conductivity and the catalytic activity for the oxygen

reduction reaction (ORR), degrading the overall electro-

chemical performance [8].

Previous work with the swedenborgite-type system for SOFC

cathodes has shown that a composition of YBaCo3ZnO7 þ d has

good electrochemical performance but tends to decompose at

lower temperatures, while InBaCo3ZnO7 þ d has lower perfor-

mance but superior low-temperature stability [8,10,12,13].

Recently, our group has shown that a composition of

Y0.9In0.1BaCo3ZnO7 þ d is stable at all temperatures, but with an

electrochemical performance similar to that of the unsub-

stituted YBaCo3ZnO7 þ d cathode [13]. Another frequently used

R-site cation is Ca2þ, which enhances the electrochemical

performance, but severely destabilizes the phase and tends to

require an increased Zn content to compensate and stabilize

the phase [8,11]. Further investigation of the stabilization ef-

fect with In substitution has shown that it can successfully

stabilize a composition of Y0.5In0.1Ca0.4BaCo3ZnO7 þ d without

requiring an increased Zn substitution [14]. These results

suggest that it may be possible to utilize the stability pro-

moting effect of In to reduce the required concentration of Zn,

thereby improving the electrochemical performance.

As the octahedral-site stabilization energy (OSSE) of Co2þ/

3þ shows a preference for octahedral coordination, it is likely

that replacing Zn with an increasing concentration of Co will

rapidly destabilize the phase, leading to decomposition at the

operating temperatures [5,8]. However, with an OSSE of zero,

Fe3þ is far more stable in the tetrahedral sites, while not

drastically diminishing the electronic conductivity and ORR

activity unlike Zn substitution. Accordingly, we present here

an investigation of the Y0.9In0.1BaCo3(Zn,Fe)O7 þ d series of

cathodeswith lower Zn contentswith the goal of discovering a

stable cathode with enhanced electrochemical performance

for reduced-temperature SOFCs.

Experimental methods

Materials synthesis

The Y0.9In0.1BaCo3(Zn,Fe)O7 þ d samples were synthesized by

conventional solid state reaction (SSR) methods. Stoichio-

metric amounts of Y2O3, In2O3, BaCO3, Co3O4, ZnO, and Fe3O4

required to produce 10 g of product were mixed with ethanol

in an agate mortar and pestle for 1 h, dried, pressed into pel-

lets, and calcined in air at 1000 �C for 12 h [8e14]. The resultant

pellets were then ground into powder, pressed again, and

sintered in air at 1200 �C for 24 h [8e14]. After sintering, the

pellets were annealed in air at 900 �C for 6 h and slowly cooled

to room temperature at a rate of 1 �C min�1 [13,14].

The Y0.2Ce0.8O1.9 (YDC) electrolyte powder was synthesized

by the glycine nitrate process (GNP) [17,18]. Stoichiometric

amounts of Y(NO3)3$6H2O and Ce(NH4)2(NO3)6 required to

produce 3 g of product were dissolved in deionized (DI) water

with glycine (NH2CH2COOH) in a cations: glycinemolar ratio of

1:1 and heated on a stir plate until a gel was formed. The gel

was then heated until combustion and the resultant powder

was calcined in air at 600 �C to ensure removal of any

remaining organic species [13,14].

The Gd0.2Ce0.8O1.9 (GDC) powders were synthesized by SSR

and GNP methods. For SSR, stoichiometric amounts of CeO2

and Gd2O3 required to produce 20 g of product were ball-

milled in ethanol for 36e48 h with 1 mol% (metals basis)

Ni(NO3)2 as a sintering aid [19,20]. Themixture was then dried,

pressed into pellets, and sintered in air at 1550 �C for 24 h. The

GNP process was identical to the synthesis of YDC, but with

Gd(NO3)3$6H2O instead of Y(NO3)3$6H2O.

Materials characterization

After synthesis and annealing, the Y0.9In0.1BaCo3(Zn,Fe)O7 þ d

powders were characterized by x-ray diffraction (XRD) with

Cu Ka radiation. The XRDpatternswere analyzedwith theMDI

Jade program (www.materialsdata.com) and refined to calcu-

late lattice parameters.

The long-term phase stabilities of the Y0.9In0.1BaCo3(Zn,Fe)

O7 þ d samples were determined by long-term exposure to

intermediate-temperature solid oxide fuel cell (IT-SOFC)

operating temperatures. The samples were heated at a rate of

2 �C min�1 in alumina crucibles in a box furnace and allowed

to dwell at temperatures of interest for 120 h before cooling to

room temperature at the same rate. Powders were placed on

top of small pelletized platformsmade of the samematerial to

avoid reaction with the alumina plates. Three separate tests

were conducted at 600, 700, and 800 �C, and the resultant

powders were characterized by XRD to determine phase sta-

bility [8e14].

The oxygen content and oxidation state of (Co,Fe) at room

temperature were determined by iodometric titration [21].

Thermogravimetric analysis (TGA) (Netzsch STA 449 F3) was

utilized to determine the change in oxygen content during

thermal cycling. TGA data were collected during two consec-

utive heating/cooling cycles at a rate of 3 �C min�1 from 80 to

900 following a 15 h dwell at 900 �C during the first heating

cycle to allow for oxygen contents to normalize [13,14].

Electrochemical characterization

Composite cathodes were manufactured by ball-milling equal

parts by mass of Y0.9In0.1BaCo3(Zn,Fe)O7 þ d and GDC powders

in ethanol for 36e48 h [9,13,14]. The composite cathodes were

then dried andmixedwith an organic binder (Heraeus V006) in

a 60: 40 cathode: binder weight ratio to create an ink [13,14].

YDC inks were prepared by mixing the YDC powder with the

binder in a 50: 50 ratio without prior ball milling [13,14].

The polarization resistances (Rp) of the composite cathodes

were measured on 8 mol % yttria stabilized zirconia (8YSZ)

electrolyte supported symmetric cells in the range of

400e800 �C by AC impedance spectroscopy using a Solartron

1260 FRA impedance analyzer. The thickness of the electrolyte

supports was between 150 and 200 mm (FuelCellMaterials.

com). YDC interlayers between the cathode and electrolyte

Fig. 1 e Room-temperature XRD patterns of selected as-

synthesized Y0.9In0.1BaCo3Zn1 ¡ xFexO7 þ d oxides before

long-term stability testing. All samples with x < 0.8 are

single-phase oxides with the P31c space group. The * in the

x ¼ 0.9 and 1.0 samples denote impurity phases.

Table 2 e Oxygen contents of theY0.9In0.1BaCo3Zn1 ¡ xFexO7 þ d oxides.

x Initial value(25 �C)

After 15 h dwellin TGA at 900 �C

After final coolingin TGA to 80 �C

0.0 7.24 7.09 7.07

0.2 7.25 7.21 7.22

0.4 7.25 7.21 7.22

0.6 7.23 7.21 7.24

0.8 7.27 7.38 7.45

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 41188

were utilized to prevent interfacial reactions [8,22,23]. The

YDC layers were screen printed two times and sintered at

1250 �C for 1 h [13,14]. The composite cathode layers were

screen printed three times on each side and sintered for 3 h at

900 �C [8e14]. The effective area of the electrodes was

0.25 cm2. After sintering the cathode, silver current collectors

were applied by screen printing silver paste diluted with a

dispersant on each cathode, followed by attaching silver wires

and mesh to the screen-printed layer with a small amount of

thick silver paste (FuelCellMaterials.com Ag-I). The current

collectors were sintered for 1 h at 800 �C with a heating rate of

1 �C min�1 followed by a cooling rate of 2 �C [13,14]. The

symmetric cells were allowed to dwell at 400 �C for 1e2 h

before testing. Rp values of the selected Y0.9In0.1BaCo3(Zn,Fe)

O7 þ d composite cathodes were measured on 250 mm thick

GDC electrolytes (FuelCellMaterials.com) for a comparison at

lower temperatures [13,14]. The symmetric cells were manu-

factured similarly to those on 8YSZ electrolytes, although

without the YDC buffer layer due to the lack of side reactions.

After testing, the microstructures of the symmetric cells were

observed by a JEOL JSM-5610 scanning electron microscope

(SEM).

Table 1 e Unit cell parameters, high temperature phase stabiliY0.9In0.1BaCo3Zn1 ¡ xFexO7 þ d oxides and composite cathodes.

x a (A) c (A) Volume (A3)

0.0 6.308 10.274 354.03

0.2 6.312 10.275 354.51

0.4 6.313 10.274 354.59

0.6 6.315 10.278 354.95

0.8 6.312 10.277 354.58

a Values calculated in the range of 433e766 �C.

Results and discussion

Crystal chemistry and phase stability ofY0.9In0.1BaCo3(Zn,Fe)O7 þ d

The room-temperature XRD patterns of selected

Y0.9In0.1BaCo3Zn1 � xFexO7 þ d samples are presented in Fig. 1.

All samples in the range of 0 � x � 0.8 are single-phase ma-

terials with the P31c phase group, analogous to other RBa-

Co3ZnO7 (R¼ Y, In, Ca) materials [10,13,14]. On the other hand,

both the x ¼ 0.9 and x ¼ 1.0 samples were not single-phase

after sintering at 1200 �C for 24 h, with many unidentified

peaks appearing in their XRD patterns. Some of the observed

impurity peaks align well with various barium iron oxide

phases, which correlates well with the previous work where

BaCoO3 was the major impurity phase [10,13], but the number

of peaks and the possible phases with the presence of six

cations make definitive phase identification challenging. Unit

cell parameters of the single-phase samples obtained by the

refinement of the XRD patterns are given in Table 1. The data

indicate that the lattice parameters do not significantly

change with the substitution of Fe for Zn, although a slight

increase in unit cell volume with increasing Fe content is

observed.

The room-temperature oxygen content values of all single-

phase samples obtained by the iodometric titration after

annealing at 900 �C are given in Table 2. All the samples show

oxygen contents of 7.24e7.27. These variations are quite small

and are within experimental error; for a reference, in our

previous work on Y1 � xInxBaCo3ZnO7 þ d (0 � x � 0.5), the

oxygen contents varied between 7.13 and 7.24 [13]. The simi-

larity of oxygen content across the samples can also help

explain the relative similarity in lattice volumes. For the

samples with 0 � x � 0.6, the slight increase in lattice volume

can be attributed to the slightly larger ionic radius of Fe2þ

ty, and activation energies of the

Stability (120 h) Ea (eV)a

600 �C 700 �C 800 �C

✓ ✓ ✓ 1.24

✓ ✓ ✓ 1.28

✓ � ✓ 1.25

✓ � � N/A

✓ � � N/A

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 4 1189

(0.63 A) compared to that of Zn2þ (0.60 A) [24]. As for the x¼ 0.8

sample, its higher oxygen content leads to an oxidation of

Co2þ and Fe2þ (0.58 and 0.63 A) to smaller Co3þ and Fe3þ ion

(0.49 Ǻ for both), resulting in a reduction in lattice volume.

The long-term phase stabilities of all the single-phase

Y0.9In0.1BaCo3Zn1 � xFexO7 þ d samples in the range of

0 � x � 0.8 are presented in Fig. 2. As can be seen, both the

x ¼ 0.0 and x ¼ 0.2 samples are stable after exposure to all

temperatures of interest. The x¼ 0.6 and x¼ 0.8 samples show

stability at 600 �C, but decompose at both 700 and 800 �C;significant decomposition is observed, making the identifica-

tion of the phases difficult. The x ¼ 0.4 sample, interestingly,

shows stability at both 600 and 800 �C, but shows a BaCoO3

impurity phase after 120 h exposure to 700 �C. This result was

verified multiple times to ensure repeatability of the results,

and the BaCoO3 phase is a well-known decomposition product

of similarmaterials [8,10,13]. This result affirms that increased

Fe content compromises lattice stability, and a significant Zn

content is required to maintain the phase-stability at all

temperatures of interest. Other work on the phase stability of

similar materials, YBaCo3 þ x(Ge,Al)1 � xO7 þ d series, have

found that samples with Co content > 3.2 tend to show

decomposition or phase impurities [25]. Interestingly in this

work, we find that a (Co,Fe) content of 3.4 in the lattice is stable

Fig. 2 e Room-temperature XRD patterns of the Y0.9In0.1BaCo3Zn

700, and 800 �C. The * in the x¼ 0.4 sample at 700 �C denote the g

x ¼ 0.6 and 0.8 samples at 700 and 800 �C are unidentified due

at both high and low temperatures, in contrast with previous

work where instability occurs either above or below certain

temperature thresholds [8e14]. The reason for the x ¼ 0.4

sample's instability at 700 �C is uncertain, and should be

investigated further in future work.

Thermal properties and electrochemical performance ofY0.9In0.1BaCo3(Zn,Fe)O7 þ d oxides

The variations in oxygen content with temperature of the

Y0.9In0.1BaCo3Zn1 � xFexO7 þ d oxides were determined by TGA,

with all mass loss/gain attributed to variations in oxygen

content. The data gathered with TGA are shown in Fig. 3, with

Fig. 3a giving a full trace of Y0.9In0.1BaCo3Zn0.6Fe0.4O7 as a

reference. The samples were allowed to anneal for 15 h at

900 �C to stabilize the oxygen contents [13,14]; for ease of

reference, these data are listed in Table 2 alongwith the room-

temperature oxygen content values and the oxygen content

values after the final cooling of the samples in the TGA to

80 �C. As can be seen in Fig. 3a, the oxygen contents during the

final heating and cooling cycles are similar in contrast to the

larger variations observed during previous steps of the TGA

measurement. This trend was observed in all samples, and it

correlates well with previouswork [10,13,14]. Fig. 3b shows the

1 ¡ xFexO7 þ d (0 ≤ x ≤ 0.8) oxides after 120 h exposure to 600,

enerated BaCoO3 impurity. The impurities generated in the

to the severity of the phase decomposition.

Fig. 3 e Variations in oxygen content as observed via TGA

of all the single-phase Y0.9In0.1BaCo3Zn1 ¡ xFexO7 þ d

(0 ≤ x ≤ 0.8) oxides: (a) full trace of the x ¼ 0.4 sample for a

reference, with the arrows denoting directionality; (b)

oxygen content with time after 15 h annealing at 900 �C;and (c) final cooling curve of all samples to 80 �C. Thesample curves across all three figures are the same

experimental data, presented in various manners for

clarity.

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 41190

oxygen contents of all single-phase samples (0 � x � 0.8) with

time after the 15 h annealing process. This data show that the

reversible absorption/desorption of oxygen during thermal

cycling tends to increase with increasing Fe content. This

trend can be understood to be due to the replacement of Zn2þ

with a fixed valence by the transition-metal ion Fe2þ/3þ with a

variable valence; the increase in ions with variable oxidation

states allows for an increased oxygen variation during ther-

mal cycling. The increased oxygen absorption capacity of the

samples with a larger Fe content can also be used to explain

their instability at higher temperatures, as previous work has

determined that compounds with larger oxygen absorption

are more likely to decompose at high temperatures [25]. The

reversible adsorption of oxygen in this class of materials has

been investigated in detail before [25e29], and the larger

amount of oxygen adsorption leads to drastic changes in the

coordination geometry of the transition-metal ions, resulting

in greater phase instability.

Fig. 3c shows the final cooling curves of all single-phase

oxides for clarity. In this figure it can be observed that the

oxygen content after the final cooling cycle to 80 �C increases

with increasing Fe content, although the 0.2 � x � 0.6 samples

all show relatively similar oxygen contents. Previouswork has

shown that the interstitial excess oxygen in swedenborgite-

type materials tend to orient itself around the (Co,M)nþ ions in

tetrahedral sites [26]. As the concentration of the fixed-valent

Zn2þ ion decreases, an increase in tetrahedra with variable

oxidation states lead to an increase in oxygen storage capac-

ity. Previous work has also shown that an increasing con-

centration of ionswith a fixed oxidation states (e.g., Zn2þ, Al3þ,Ga3þ) decreases the oxygen storage capacity, which further

supports our observations [8,10,11,25].

Electrochemical performances of the Y0.9In0.1BaCo3Zn1 � x

FexO7 þ d oxides were characterized by AC impedance spec-

troscopy on cathode jYDCjelectrolytejYDCjcathode symmetric

cells. Composite cathodes were utilized as previous research

has found RBa(Co,M)4O7 þ d type oxides to have relatively low

oxygen transport properties [8,9]. The composite cathodes

were manufactured with 50:50 wt% cathode:GDC, as previous

research found this to be the ideal composition for

YBaCo3ZnO7 þ d [9]. Previous work has found that the perfor-

mance of this class of materials is dependent on the three-

phase boundary (TPB) region, which is the surface area

shared by the cathode, electrolyte, and oxidant [9e14,30].

Accordingly, the powders were ball-milled to maximize ho-

mogeneity and decrease particle size to increase the size of

the TPB. As 8YSZ is known to react with both cobalt and

gadolinium to form non-conductive interface layers, YDC was

utilized as a buffer layer to prevent reaction [13,14,22,23]. Sil-

ver was used for the cathode current collector due to its high

electrical conductivity and oxygen reduction properties

[31e33]. The silver current collector was assembled as a self-

assembling micron-scale mesh as described in our previous

work, due to its high performance [13,14].

Fig. 4 shows the Arrhenius plots of the polarization re-

sistances (Rps) of the more stabilized Y0.9In0.1BaCo3Zn1 � x

FexO7 þ d þ GDC (0 � x � 0.4) composite cathodes on 8YSZ

electrolyte-supported symmetric cells in the range of

400e800 �C. Previous work has shown that the diameter of the

impedance loop can be used to approximately calculate the

Fig. 4 e Arrhenius plots in the range of 400e800 �C of (a)

Y0.9In0.1BaCo3Zn1 ¡ xFexO7 þ d (0 ≤ x ≤ 0.4)þ GDC composite

cathodes and (b) Y0.9In0.1BaCo3Zn0.6Fe0.4O7 þ d þ GDC

composite cathode compared to the single-phase BSCF. All

samples were measured on 8YSZ electrolyte-supported

symmetric cells. Activation energies were calculated from

the best-fit lines in the range of 433e766 �C and are listed

in Table 1.

Fig. 5 e Area-normalized Nyquist plots of the

Y0.9In0.1BaCo3Zn1 ¡ xFexO7 þ d þ GDC composite cathodes

on 8YSZ electrolyte-supported symmetric cells: (a) x¼ 0, (b)

x ¼ 0.2, and (c) x ¼ 0.4. The resistance attributed to the

electrolyte has been removed from all plots for clarity.

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 4 1191

polarization resistances of the cathodes in cells where the

electrolyte has negligible electronic conductivity [34e36]. As

8YSZ has negligible electronic conductivity in this range, the

difference between the low and high frequency interceptswas

used to calculate the polarization resistances of these cath-

odes. Fig. 4 shows that the electrochemical performances of

the Y0.9In0.1BaCo3Zn1 � xFexO7 þ d þ GDC composite cathodes

increase with increasing Fe content. The performance inhib-

iting nature of Zn2þ in this series of oxides is well docu-

mented; its [Ar]3d10 electronic configuration is exceptionally

stable, and as such it cannot be further oxidized or reduced,

prohibiting it from participating in electronic transport and

oxygen reduction [8,10e12,37e41]. By replacing Zn2þ with

Fe2þ/3þ, the material gains additional electrochemically active

sites, significantly boosting the performance. It has also been

shown that both the electrical conductivity and oxygen

reduction reaction kinetics of these materials is improved

with increased oxygen content due to an increase in the

oxidation state of the transition-metal ions [14,27]. As the

substitution of Fe for Zn is shown in Fig. 3 and Table 2 to in-

crease the affinity of these materials for oxygen, this can also

explain why the substitution of Fe for Zn improves the per-

formance. Fig. 4b compares the performance of the x ¼ 0.4

sample with the well-studied Ba0.5Sr0.5Co0.8Fe0.2O3 � d (BSCF)

Fig. 6 e Cross-sectional SEM micrographs of the Y0.9In0.1BaCo3Zn1 ¡ xFexO7 þ d þ GDC composite cathodes on 8YSZ

electrolyte-supported symmetric cells after testing. Performances of these cells can be observed in Figs. 4 and 5.

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 41192

as a reference. While the x ¼ 0.4 sample is in a composite

cathode, the BSCF sample was manufactured as single-phase

due to its mixed ionic and electronic conduction properties. In

the previous work, we investigated a composite BSCF cathode

with the same weight ratio as our swedenborgite-type com-

posite cathodes; the composite BSCF was shown to have

reduced performance in the lower temperature regions

(T < 600 �C) [13]. With its reduced activation energy, the

Y0.9In0.1BaCo3Zn0.6Fe0.4O7 þ d þ GDC composite cathode

significantly outperforms the single-phase BSCF in the

reduced temperature region; however, BSCF begins to

outperform the x ¼ 0.4 sample at around 700 �C. However, it

should be noted that neither the BSCF nor the swedenborgite-

type cathodes investigated in this work are fully optimized,

and that optimization of the synthesis and cell preparation

procedures for both materials may lead to different perfor-

mance values.

Fig. 5 shows the area normalized Nyquist plots of

Y0.9In0.1BaCo3Zn1 � xFexO7 þ d þ GDC composite cathodes on

8YSZ symmetric cells in the temperature range of

433e600 �C. According to Adler et al. [36,37], impedance arcs

at higher frequencies are attributed to charge transfer across

the cathode/electrolyte interface and electronic transfer

across the cathode, and impedance arcs at the lower fre-

quencies are attributed to the oxygen adsorption and

reduction reactions. At all temperatures, the diameter of the

low-frequency arc decreases with increasing Fe content. As

previously discussed, Zn2þ does not contribute to the oxygen

reduction reaction, so this result is expected. Interestingly,

the diameter of the high-frequency arc is also diminished in

the x ¼ 0.4 sample, compared to the others. Fig. 6 shows the

Fig. 7 e Symmetric cell tests comparing Y0.9In0.1BaCo3Zn0.6Fe0.4solid state reaction and by GNP on GDC electrolyte supports. (a

400e800 �C, and (b) Nyquist plots of these cells at 500 �C. The r

removed for clarity.

cross-sectional SEM micrographs of these symmetric cells

after testing, where all samples show similar microstruc-

tures. As there is no observable difference in microstructure

for the x ¼ 0.4 sample, this reduction in impedance is

attributed to the higher oxygen content at elevated temper-

atures and increased electronic conductivity due to the

increased Fe content. At lower temperatures, the existence of

three distinct arcs can be observed, which correlates well

with our previous work that investigated the equivalent-

circuit modeling of similar electrochemical cells [11,13]. As

the temperature increases, the high-frequency arc reduces

drastically in all samples, signifying that at higher tempera-

tures the oxygen reduction reaction is the primary limitation.

This correlates well to previous work on similar materials

[13,14].

Fig. 7a shows the Arrhenius plots of the Rps of two

Y0.9In0.1BaCo3Zn0.6Fe0.4O7 þ d þ GDC composite cathodes on

GDC electrolyte; one with GDC synthesized by SSR, and

another with GDC synthesized by GNP. Powders synthesized

by GNP have an exceptionally small particle size, greatly

increasing the surface area per mass ratio [17,18]. Previous

work by our group has shown that the inclusion of GNP-

synthesized GDC can reduce the polarization resistance of

these materials by increasing the size of the TPB [13]. Simi-

larly, previous comparisons on SOFC composite cathode ma-

terials containing gadolinia doped ceria have shown that

synthesis and preparation methodology can have a large

impact on electrochemical performance [42e44]. In Fig. 7a, it

can be seen that the inclusion of GDC powder synthesized by

GNP reduces the polarization resistances at higher tempera-

tures, signifying improved electrochemical performance. At

O7 þ d composite cathodes with both GDC synthesized by

) Arrhenius plots of polarization resistance in the range of

esistance attributed to the electrolyte in Figure b has been

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 4 1193

lower temperatures both samples have similar performances,

though the sample with GDC by SSR has slightly reduced

impedance. Fig. 7b shows the area-normalized Nyquist plots

of both samples at 500 �C. The data show that the higher-

frequency arc of the GDC by GNP sample is significantly

expanded, while the low-frequency arc is nearly identical for

both samples. The expansion of the high-frequency arc can be

understood to be due to increased grain boundary resistance

between the electrolyte support and the GDC particles in the

composite cathode due to their reduced size from the GNP

synthesis method. The similarity of the low-frequency arcs,

on the other hand, suggests that at lower temperatures the

expansion of the three-phase boundary region has negligible

impact on the oxygen reduction reaction. Further work should

focus on optimizing the composite cathode ratio as a different

ratio may be ideal due to the reduction of dependence on the

three phase boundary.

Conclusions

The effects of Fe substitution in the Y0.9In0.1BaCo3Zn1 � xFexO7 þ d series of oxides have been investigated sys-

tematically. All samples in the range of 0 � x � 0.8 are single-

phase materials and have similar room-temperature oxygen

contents, although the samples with higher Fe content show a

larger capacity for reversible oxygen absorption. Both the

x¼ 0.0 and 0.2 samples are stable in the 600e800 �C range after

120 h exposure, while the x ¼ 0.6 and 0.8 samples show sig-

nificant phase decomposition at 700 and 800 �C. Interestingly,the x ¼ 0.4 sample is stable at both 600 and 800 �C, although it

generates the BaCoO3 impurity at 700 �C. Electrochemical

performances of themore stable compounds (0� x� 0.4) were

investigated with AC impedance spectroscopy, which shows

that higher Fe content leads to reduced cathode polarization

resistance. The x ¼ 0.4 composite cathode exhibits a signifi-

cant increase in performance compared to the well-studied

single-phase BSCF cathode at temperatures below 700 �C,revealing that the Y0.9In0.1BaCo3Zn0.6Fe0.4O7 þ d composite

cathode is an attractive option for reduced temperature solid

oxide fuel cells.

Acknowledgments

Financial support by the Welch Foundation grant F-1254 is

gratefully acknowledged. C. Ortiz was supported by the

Research Experience for Undergraduates (REU) program of the

National Science Foundation Materials Interdisciplinary

Research Team (MIRT) grant DMR-1122603. The authors thank

Soa-Jin Sher for her assistance with some experiments, and

Daeil Yoon for valuable discussions.

r e f e r e n c e s

[1] Brett DJL, Atkinson A, Brandon NP, Skinner SJ. Intermediatetemperature solid oxide fuel cells. Chem Soc Rev2008;37:1568e78.

[2] Yamomoto O. Solid oxide fuel cells: fundamental aspectsand prospects. Electrochim Acta 2000;45:2423e35.

[3] Minh NQ, Takahashi T. Science and technology of ceramicfuel cells. Amsterdam: Elsevier; 1995.

[4] Kim JH, Manthiram A. LnBaCo2O5 þ d oxides as cathodes forintermediate-temperature solid oxide fuel cells. JElectrochem Soc 2008;155:B385e90.

[5] Kim JH, Prado F, Manthiram A. Characterization ofGdBa1exSrxCo2O5 þ d (0 � x � 1.0) perovskites as cathodes forsolid oxide fuel cells. J Electrochem Soc 2008;155:B1023e8.

[6] Kim J, Jun A, Shin J, Kim G. Effect of Fe doping on layeredGdBa0.5Sr0.5Co2O5 þ d perovskite cathodes for intermediatetemperature solid oxide fuel cells. J Am Ceram Soc2014;97:651e6.

[7] West M, Manthiram A. Layered LnBa1exSrxCoCuO5 þ d

(Ln ¼ Nd and Gd) perovskite cathodes for intermediatetemperature solid oxide fuel cells. Int J Hydrogen Energy2013;38:3364e72.

[8] Kim JH, Manthiram A. Low Thermal expansion RBa(Co, M)4O7

cathode materials based on tetrahedral-site cobalt ions forsolid oxide fuel cells. Chem Mater 2010;22:822e31.

[9] Kim JH, Manthiram A. Electrochemical characterization ofYBaCo3ZnO7 þ Gd0.2Ce0.8O1.9 composite cathodes forintermediate temperature solid oxide fuel cells. ElectrochimActa 2010;55:5312e7.

[10] Kim YN, Kim JH, Huq A, Paranthaman MP, Manthiram A.(Y0.5In0.5)Ba(Co,Zn)4O7 cathodes with superior high-temperature phase stability for solid oxide fuel cells. J PowerSources 2012;214:7e14.

[11] Kim YN, Kim JH, Manthiram A. Characterization of (Y1 � xCax)BaCo4 � yZnyO7 as cathodes for intermediate temperaturesolid oxide fuel cells. Int J Hydrogen Energy2011;36:15295e303.

[12] Kim JH, Kim YN, Bi Z, Manthiram A, Paranthaman MP,Huq A. High temperature phase stabilities andelectrochemical properties of InBaCo4 � xZnxO7 cathodes forintermediate temperature solid oxide fuel cells. ElectrochimActa 2011;56:5740e5.

[13] West M, Sher SJ, Manthiram A. Effects of in substitution inY1 � xInxBaCo3ZnO7 þ d (0 � x � 0.5) cathodes for intermediatetemperature solid oxide fuel cells. J Power Sources2014;271:252e61.

[14] West M, Manthiram A. Improved phase stability andelectrochemical performance of (Y,In,Ca)BaCo3ZnO7 þ d

cathodes for intermediate temperature solid oxide fuel cells.Int J Hydrogen Energy 2014;39:19722e30.

[15] Zhou Q, Shi Y, Wei T, Li Z, An D, Hu J, et al. NovelYBaCo3.2Ga0.8O7 þ d as a cathode material and performanceoptimization for IT-SOFCs. Int J Hydrogen Energy2014;39:10710e7.

[16] Zhou Q, Shi Y, Hu J, Wei T, Li Z, Li Y, et al. Preparation,characterization, and electrochemical properties ofYBaCo3.4Al0.3Ga0.3O7 þ d and YBaCo3.2Al0.4Ga0.4O7 þ d cathodesfor IT-SOFCs. Ceram Int 2014;40:13481e5.

[17] Taguchi H, Matsuda D, Nagao M, Tanihata K, Miyamoto Y.Synthesis of perovskite-type (La1 � xSrx)MnO3 (O � 0.3) at lowtemperature. J Am Ceram Soc 1992;75:201e2.

[18] Chick LA, Pederson LR, Maupin GD, Bates JL, Thomas LE,Exarhos GJ. Glycineenitrate combustion synthesis of oxideceramic powders. Mat Lett 1990;10:6e12.

[19] Kleinlogel C, Gauckler LJ. Sintering of nanocrystalline CeO2

ceramics. Adv Mater 2001;13:1081e5.[20] Nicholas JD, De Jonghe LC. Prediction and evaluation of

sintering aids for cerium gadolinium oxide. Solid State Ionics2007;178:1187e94.

[21] Manthiram A, Swinnea JS, Sui ZT, Steinfink H,Goodenough JB. The influence of oxygen variation on thecrystal structure and phase composition of the

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 4 0 ( 2 0 1 5 ) 1 1 8 6e1 1 9 41194

superconductor yttrium barium copper oxide(YBa2Cu3O7 � x). J Am Chem Soc 1987;109:6667e9.

[22] Duan Z, Yang M, Yan A, Hou Z, Dong Y, 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.

[23] Tsoga A, Gupta A, Naoumidis A, Nikilopoulos P. Gadolinia-doped ceria and yttria stabilized zirconia interfaces:regarding their application for SOFC technology. Acta Mater2000;48:4709e14.

[24] Shannon RD. Revised effective ionic radii and systematicstudies of interatomic distances in halides andchalcogenides. Acta Crystallogr A 1976;32:751e67.

[25] Parkkima O, Yamauchi H, Karppinen M. Oxygen storagecapacity and phase stability of variously substitutedYBaCo4O7 þ d. Chem Mater 2013;25:599e604.

[26] Hao H, Cui J, Chen C, Pan L, Hu J, Hu X. Oxygen adsorptionproperties of YBaCo4O7-type compounds. Solid State Ionics2006;177:631e7.

[27] Hao H, He Q, Cheng Y, Zhao L. Oxygen adsorption andelectronic transport properties of Fe-substituted YBaCo4O7

compounds. Mater Res Bull 2014;53:84e8.[28] Kadota S, Karppinen M, Motohashi T, Yamauchi H. R-site

substitution effect on the oxygen-storage capability ofRBaCo4O7 þ d. Chem Mater 2008;20:6378e81.

[29] Parkkima O, Karppinen M. The YBaCo4O7 þ d eBasedfunctional oxide material family: a review. Eur J Inorg Chem2014;25:4056e67.

[30] O'Hayre R, Barnett DM, Prinz FB. The triple phase boundary. JElectrochem. Soc 2005;152:A439e44.

[31] Shim JH, Kim YB, Park JS, Prinz FB. Silver nanomesh as acathode for solid oxide fuel cells. ECS Trans 2011;35:2209e12.

[32] Huang H, Holme TP, Prinz FB. Oxygen reductioncharacteristics on Ag, Pt, and AgePt allows in low-temperature SOFCs. ECS Trans 2007;32:31e40.

[33] Guo Y, Zhou Y, Chen D, Shi H, Ran R, Shao Z. Significantimpact of the current collection material and method on theperformance of Ba0.5Sr0.5Co0.8Fe0.2O3 � d electrodes in solidoxide fuel cells. J Power Sources 2011;196:5511e9.

[34] Choi S, Yoo S, Kim J, Park S, Jun A, Sengodan S, et al. Highlyefficient and robust cathode materials for low-temperaturesolid oxide fuel cells: PrBa0.5Sr0.5Co2 � xFexO5 þ d. Sci Reports2013;3:2426.

[35] Liu M, Hu H. Effect of interfacial resistance on determinationof transport properties of mixed-conducting electrolytes. JElectrochem Soc 1996;143:L109e12.

[36] Adler SB, Chen XY, Wilson JR. Mechanisms and rate laws foroxygen exchange on mixed-conducting oxide surfaces. JCatal 2007;254:91e109.

[37] Adler SB. Factors governing oxygen reduction in solid oxidefuel cell cathodes. Chem Rev 2004;104:4791e843.

[38] Wang Z, Peng R, Zhang W, Wu X, Xia C, Lu Y. Oxygenreduction and transport on the La1 � xSrxCo1 � yFeyO3 � d

cathode in solid oxide fuel cells: a first-principles study. JMater Chem A 2013;1:12932e40.

[39] Lee YL, Klies J, Rossmeisl J, Shao-Horn Y, Morgan D.Prediction of solid oxide fuel cell cathode activity withfirst-principles descriptors. Energy Environ Sci2011;4:3966e70.

[40] Pavone M, Ritzmann AM, Carter EA. Quantum-mechanics-based design principles for solid oxide fuel cell cathodematerials. Energy Environ Sci 2011;4:4933e7.

[41] Suntvich J, Gasteiger HA, Yabuuchi N, Nakanishi H,Goodenough JB, Shao-Horn Y. Design principles for oxygen-reduction activity on perovskite oxide catalysts for fuel cellsand metal-air batteries. Nat Chem 2011;3:546e50.

[42] Lee SJ, Kim DS, Jo SH, Muralidharan P, Kim DK.Electrochemical properties of GdBaCo2/3Fe2/3Cu2/3O5 þ d e

CGO composite cathodes for solid oxide fuel cell. Ceram Int2012;38:S493e6.

[43] Nielsen J, Jacobsen T, Wandel M. Impedance of porous IT-SOFC LSCF: CGO composite cathodes. Electrochem Acta2011;56:7963e74.

[44] Xia C, Zhang Y, Liu M. Lsm-gdc composite cathodes derivedfrom a solegel process: effect of microstructure oninterfacial polarization resistance. Electrochem Solid-StateLett. 2003;6:A290e2.